text
stringlengths 14
5.77M
| meta
dict | __index_level_0__
int64 0
9.97k
⌀ |
|---|---|---|
#
#
Elliot
du Néant
David Calvo
d'autres images, d'autres textes vous attendent sur
www.lavolte.net
Conception graphique et illustration de couverture : Stéphanie Aparicio
Illustrations intérieures : David Calvo
© Éditions La Volte – 2013
I.S.B.N : 9782917157374
Le Code de la propriété intellectuelle interdit les copies ou reproductions destinées à une utilisation collective. Toute représentation ou reproduction intégrale ou partielle faite par quelque procédé que ce soit, sans le consentement de l'auteur ou de ses ayants cause, est illicite et constitue une contrefaçon sanctionnée par les articles L.335-2 et suivants du Code de la propriété intellectuelle.
« Tout se simplifie et s'explique : c'est le plein qui était vide et le vide qui était le plein. L'éther impondérable est un marbre aux transmissions instantanées. La matière n'est qu'un trou. »
Jean Painlevé
# LE NÉANT (1)
Au début, il y a tout. Le plein, indifférencié, mêlé de lui, soupe trop riche, trop épaisse, bouillon turbulent. Enceint d'infini, ce ventre non délimité : tout y est possible sans contraintes, donc sans forme. Un œil extérieur pourrait y discerner un monde en mouvement, qui se mange puis se régurgite ; mais, hors limites, point d'intérieur, ni d'extérieur : simplement, le bâillement long des possibles, les non-espaces molaires qui détendent et poussent la membrane de ce qui n'est pas. Rien n'est en devenir, tout est déjà là, dans cette boule sans circonférence, dans ce carré sans sinus, qui n'éternue pas. Il n'y a rien dehors, rien qui puisse être défini car dehors n'est pas une notion. Seuls les mots peuvent dégager un semblant d'organisation dans cet ensemble sans début ni fin. Sans mot, tout peut s'y lire, les paradoxes résolus dans un même mouvement, pas même respiration, puisque croissance et décroissance n'existent pas. Tout unidimensionnel puisque sans pont supérieur, tout réduit à un point où chaque regard trace un souvenir, une forme, un motif, des figures qui deviennent silhouettes, puis gestalts détruits par l'absence de sens individuel, sans interaction. Toute possibilité de présence s'anéantit dans la machine à laver de ce déplacement sans motricité, condamné au surplace dans un monde qui ignore la place. Quel son pour ce tout qui, de fait, ne transmet rien ? La musique d'un vrombissement, le rotor éternel d'élytres invisibles. Il faut fermer les yeux et concevoir la chambre d'échos d'un noir profond pour entendre, doucement, l'hymne perpétuel de ce qui n'est.
Essayez. Entendez ces ténèbres qui hurlent.
Mortes eaux / vives eaux
Islande, 6 janvier 1986
Je ne rêve pas. Pour être plus précis : j'ai conscience du rien. Si le sommeil est un canevas de ténèbres sur lequel nous brodons des merveilles, pour moi ce noir, cette absence de tout, est déjà texture. Une terreur ; toutes les nuits, il faut compter les secondes qui me séparent du réveil. Je ferme les yeux puis je m'éteins, sans batteries, sans plus de souffle. Je saisis chaque mesure du temps. Je préfère parfois rester debout, attendre que le temps s'écoule. Tout le monde rêve, à ce qu'on dit. Il suffit juste d'apprendre à se souvenir, il existe des solutions. J'ai essayé. J'ai laissé un verre d'eau près de mon lit. J'ai roulé pour retrouver cette place sur le matelas, dans la vallée de l'oreiller où, passé l'écrin du souvenir, les rêves s'incarnent. J'ai placé des mobiles de planètes, des fleurs séchées dont les pétales souillaient les draps. J'ai bu tous les breuvages de sorcières, avalé tous les séminaires, mangé des runes. J'ai modélisé le tunnel que les rêveurs ont l'habitude d'emprunter pour sortir de l'autre côté, dans le rêve. Mais moi, Bracken, moi, je ne sors jamais. Pour être plus précis : il n'y a pas de tunnel.
KOR
Entendez ! La langue invisible dit le geste d'Elliot
Au Néant descendu pour dérober le Ptyx.
Lo ! Une écume en cette frange !
La spirale du téléphone m'arrache au sommeil. Je dors les pieds contre le mur, mains et pieds en appui sur la paroi. J'ai mal partout, avec l'impression d'avoir passé la nuit dans un tambour de lave-linge. J'ai tendance à baver.
Entre les rideaux filtre une lumière de nuit, à peine lampadaire.
— Grmml ?
Des ténèbres ordinaires du signal émerge une voix fantastique.
— Bracken !
—... Fink ?
Fink, bon sang, Fink. Je croyais en avoir fini avec toi, je croyais en avoir fini avec l'école, ta parano de vieux flic, tes idées folles sur les cailloux. Laisse-moi en paix, je t'en supplie. Laisse-moi sous ma couette, à essayer de rêver.
— J'ai besoin de vous ici, Bracken ! C'est la catastrophe !
Dans la distance, cancane un vol nocturne de canards.
— Je ne reviendrai pas, Fink.
Je vais raccrocher, le vieux insiste.
— Bracken, je vous en supplie. Il s'agit d'Elliot...
— Quoi, Elliot ?
— Je vous envoie un taxi !
— Fink, qu'est-ce qui se passe ?
Je l'entends se mordre la lèvre.
— Mettez des moufles !
Un besoin s'installe, comme on entre dans des chaussons moelleux, depuis longtemps pelés. Me mettre en mouvement, avant que mon corps ne sèche. Prendre ma vie en main, cesser de mourir tous les matins. Avancer, sortir du lit, mettre un pied devant l'autre et agir. Depuis ma démission de l'école, j'ai l'impression de patienter dans la salle d'attente de mon destin. Étranger chez moi, de passage, entre parenthèses, toute une vie comme un paquet-cadeau. Mon futon moisi, ces chaussettes, ces livres, ces bibelots, nouveaux objets technologiques aux formes blanches, mon intimité dans le marbre mou des draps. Qui viendra fouiller ces ruines fragiles pour savoir de quoi ma vie était faite ?
Je tourne les yeux vers la fenêtre. Le ciel noir noir noir au-dessus du port semble vouloir s'enfoncer dans le noir de la baie. Pourquoi suis-je encore ici ? Pourquoi ne suis-je pas parti ? Est-ce que je suis à ce point incapable de prendre une décision concernant ma vie ? Je me déteste. Oh, je sais, se dire les choses à soi-même pour se rassurer, ce n'est pas ce qui nous fait avancer. Mais moi, j'ai le choix, je suis encore jeune, la pourriture est réversible, il me suffit d'un appel d'air, prendre pied et me jeter. Mais je ne peux rien faire. Je veux ma couette, le silence de ces murs. Depuis que j'ai quitté l'école, je ne suis quasiment jamais sorti de chez moi, à part pour me nourrir. Je connais mal leur langue. Même s'ils parlent tous anglais, je suis pour toujours un étranger. J'ai toujours pensé que j'étais trop paresseux pour faire un Islandais. Je ne peux pas retourner chez moi. Ce serait terrible, un aveu d'échec. Je n'ai pas vu mon père depuis dix ans, nous ne sommes même pas fâchés, c'est plus simple de ne pas se parler. Il ne comprendrait pas. De toute façon, il n'y a rien pour moi, là-bas. J'ai encore cette lettre de l'éducation nationale, qui me radie pour mes méthodes iconoclastes, dans un tiroir où je garde les pauvres traces de mon passé : des photos de moi, enfant, prises sur une piste de ski. Enfoncé dans une parka, avec des lunettes en plastique. Des moufles. Quelques fleurs séchées. Un dessin que j'avais fait de maman, avant sa disparition.
Et voilà. J'en suis où ? Je fais quoi ? Il n'y a rien, rien. Rien que le vide, cette sensation de moisi qui m'envahit.
Mais si Elliot, en danger...
Près d'un arbre au bord d'une rivière
Il y a un trou dans le sol
Où un vieil homme d'Aran
Tourne et tourne encore.
Filet de notes échappé d'une fréquence nocturne. Bercé par le roulis du taxi, j'ai trop chaud. Sur ma tête, un bonnet ; à mes mains, des moufles. Une mamie de Stykkishólmur y a tricoté des flocons de neige, une sorcière m'a-t-on dit, une de ces fanatiques des anciens dieux, ceux-là mêmes dont les évêques avaient jeté les effigies dans une chute d'eau. Quand j'avais quitté l'école, pris d'une dernière pulsion de liberté, persuadé que je pouvais triompher du destin en conjurant l'artiste en moi, j'avais entrepris de parcourir le pays à pied. Je n'avais parcouru que le nord, le Snaefflesness, les Westfjords, et, résigné, j'étais rentré en avion d'Isafjordur. Je n'avais pas trouvé l'énergie pour dessiner. J'étais revenu défait. Sans avis. Sans volonté. Il ne me restait rien à accomplir, rien à faire. J'étais vidé de tout, une pelote de laine dévidée sur le parquet, qui me renvoyait l'image de ma misère. Moi, presque au pôle Nord, sans famille, sans travail. Sans espoir. Mais au chaud.
Sous ma couette.
Son esprit est un phare dans le voile de la nuit
D'une étrange et certaine façon.
Le dos du conducteur dépasse du siège avant, comme une tarte d'un moule trop petit. Il me lance un regard dans le rétroviseur.
— Vous n'êtes pas d'ici.
Derrière la fenêtre embuée, des silhouettes zigzaguent dans les rues de cette ville cernée de montagnes, une cité pépiant sous les étoiles, si petite que l'on connaît le nom de tous ses chiens. Entre les maisons de tôles, le froid s'est fait une place ; glacial, il attire les promeneurs, les enroule d'une pointe de vitesse. Toute la ville semble arasée. Au loin, le massif de l'Esja domine les ténèbres d'un sombre plus clair. J'y distingue une trace de lune accrochée à ses falaises, ses éboulis, comme des cheveux tressés d'ombres.
— Je suis Français.
— Ah, Paris.
Il y a le vrai, il y a le faux,
Mais jamais il ne se battra pour toi.
Les faubourgs de Reykjavík disparaissent le long d'une côte déchiquetée, en route vers la proche banlieue, Hafnafjordur, un petit port reconverti en cité-dortoir où vont se loger les familles nombreuses, les vieilles dames et les vikings.
— Vous allez où exactement ?
— L'école.
— Vous travaillez là-bas ?
J'ai démissionné, mais je ne lui dis pas. Je n'étais pas titulaire, car je ne parle pas couramment l'islandais, une langue que je peux comprendre, mais qu'il m'est très difficile de prononcer. À l'école d'Hamarinn, tout le monde doit pouvoir enseigner n'importe quelle matière en islandais ; étant un étranger, j'avais dû opter pour des cours spécialisés, en l'occurrence les arts plastiques.
— Vous avez déjà vu des fées ? demande le conducteur.
Je ne saurai pas dire si les Islandais se trompent quand ils regardent ces rochers pour y voir sculpté, en creux, le mythe d'un peuple parallèle au leur, vivant sa vie dans les failles, discret, affairé à ses cuisines. Je sais que les enfants de l'école fréquentent des fées tous les jours. Avant ma démission de l'école, je fréquentais des enfants tous les jours.
Ce n'est pas moi que tu vois.
— Je ne crois pas.
— Cette école, c'est un endroit bizarre... J'ai un ami qui habite juste sous la falaise. Il me raconte que la nuit, il voit des lumières, que parfois il entend des chansons. Je crois que ce sont des feux follets, mais ma femme, avec ses amies, elles font des pique-niques là-bas, elles laissent des gamelles de grumeau sur les rochers en partant, pour s'attirer les faveurs de la cour royale.
Il ricane. Il n'est pas d'ici non plus.
— Ça les fait rêver, nos femmes, vous savez, c'est comme les Anglais, ils ont un roi, une reine, mais nous, on a quoi ?
Dans le ciel, les autoroutes phosphorescentes d'une aurore boréale s'effacent doucement, ruines d'une architecture céleste. Je cligne des yeux quand la lumière d'un phare, au loin sur la mer profonde, trace un prisme sur la fenêtre. Je me frotte les yeux, la laine me tire des larmes de rien.
Le pli des hommes sages.
Hamarinn est une butte au centre-ville d'Hafnafjordur, en bordure d'un champ de lave. Côté ville, la butte se termine en courte falaise, parsemée de bois miniatures et de petites résidences privées. Côté terres, c'est un large parvis d'anciennes pierres couchées, une mosaïque minérale constellée de pistils turquoises où le pollen se dépose. Un endroit sacré ; une intense sauvagerie au cœur même d'un village où ont poussé les supermarchés, les coffee shops. Entre les interstices éclosent de minuscules fleurs jaunes et mauves, en collerettes le long des buissons. Pendant la récréation, les enfants gravent leurs noms sur les pierres, pour attirer les fées. Ils cachent des feux d'artifice sous les grandes stèles, en désirant les jours de fête.
L'école a été bâtie sur ce plateau, juste avant de devenir _hraun_ , terre durcie. Deux grands bâtiments d'un gris futuriste, reliés par des couloirs extérieurs vitrés, qui semblent posés comme des cubes d'enfant. Le modernisme de l'ensemble tranche radicalement avec le panthéisme bien tempéré dont font preuve les promeneurs sur les falaises, en quête d'une fleur rare, d'un sous-bois de poche où se nichent les énormes chats du voisinage, gras et ronronnants. Dans un de mes premiers dessins réalisés en Islande, où j'étais venu pour retrouver l'inspiration, j'avais tracé l'école à grands traits, sur un chaos végétal sans grand rapport avec la réalité, et un champ de lave bien plus étendu. J'avais voulu souligner la différence, ces deux aspects d'un même pays cohabitant dans l'harmonie de ses paradoxes, sans espace entre eux pour les opposer. Parfois un simple couloir, un pont, vient unir deux éléments à la dérive.
Tout ici est frontière, passage de l'un à l'autre, tension de l'entre-deux.
« Pour la treizième semaine, The Riddle _de Nik Kershaw, en tête de notre top 30 ! Nik Kershaw qui donnera un concert exclusif sur la falaise demain midi, pour célébrer la fin de Noël, en espérant que Gryla ne se pointe pas ! En attendant ce grand rendez-vous, rejoignons l'équipe de sismologues du cratère de Burfell, où l'on nous dit que... »_
Sous le préau de l'entrée, la silhouette d'un oisillon chétif : Fink tape des pieds pour se réchauffer, rabat une pelure sur ses os, une veste en laine trouée, rapiécée, dont toutes les coutures ont déjà été remplacées. Un strato-cumulus s'échappe de ses lèvres quand je descends du taxi.
— J'ai bien cru que vous ne viendriez pas.
Un instant, il esquisse un mouvement vers moi, comme pour me prendre dans ses bras, avec une intimité à peine entamée, peut-être en souvenir de choses anciennes, disparues.
— Salut Fink, dis-je, morbide.
— Si vous saviez comme je suis content de vous voir !
— Donnez-moi une seule raison de rester.
— J'ai fait du café !
Des canards passent sous la voûte étoilée d'une nuit éternelle.
Little generator won't get the spark
motor's in a bad condition, you gotta have
these batteries charged...
Tout est moisi dans le bureau de Fink, imbibé de caféine comme son vieux costume et sa chemise au col élimé. D'un phonographe suinte le sirop d'un siècle enfui : une guitare décharnée, une voix sortie de nulle part. Dans le fouillis, papiers, pages de journal et fougères grimpent jusqu'au plafond pour tenter de s'échapper. La circulation se fait par des chemins entre les piles, la cafetière, le bureau ; par la fenêtre, au-delà du parking abandonné, la lave, la lave à perte de vue ; à la lueur des lampadaires, la roche vire à l'orange.
— Alors, Fink ?
Fink prend son air de coyote affamé.
— Pauvre Elliot.
J'adore Elliot. Tout le monde adore Elliot. Il s'agit du concierge qui va et vient, un vieillard affairé à des tâches imprécises. On dit qu'il est autiste. Je ne connais pas suffisamment d'autistes pour savoir si tous sont comme lui, mais il ne m'a jamais paru plus fou qu'un autre – peut-être plus discret, oui, obsédé par ses rituels, étrangement silencieux. Il porte toujours le même maillot de matelot, avec un gros nœud blanc. Ses grands yeux mi-clos sont soulignés d'un liseré noir, comme d'impossibles cernes. Un sourire fatigué, entre épuisement et ironie.
Elliot est le protégé de Plouffe, le proviseur. Plouffe parle parfaitement français, Elliot est comme son fils. J'avais trouvé auprès d'eux une forme de réconfort. Elliot comprend quand je lui parle en français. Il n'a d'ailleurs pas le type islandais. Elliot ne parle pas. Je suis peut-être le seul à l'avoir presque entendu dire un mot, un matin où, les yeux fixés sur une feuille raturée qu'il n'arrivait pas à transformer en dessin figuratif, il avait tenté d'articuler ce qu'il avait voulu représenter :
Elliot vit ici depuis toujours, placé par l'État pour entretenir les canalisations de l'école. Il est si vieux, il a pu mener mille existences. Un environnement stable lui avait probablement permis de prendre conscience de sa condition dès son jeune âge, mais il était resté confiné dans un monde d'ordre obscur, voué à des tâches d'entretien pacifiques. L'école a des tubes parfaits. Quand j'avais constaté sa précision, je l'avais autorisé à participer à mes cours d'éveil avec les plus petits, pensant que cela aiderait tout le monde de le voir dessiner des petits monstres. Elliot se passionnait pour les catalogues, les escaliers et les macareux. Il aimait par-dessus tout dessiner la mer, les animaux marins, les tortues, les baleines, les hippocampes, les anémones entre les rochers, les algues dans le ressac. Son monde était un vivier où chaque espèce était un sujet d'émerveillement.
— Il est mort ?
Fink finit de servir le breuvage en silence. Son café est une mélasse saturée de cassonade ; il le cuit avec un fer à cheval, porté à incandescence dans une gamelle en fer blanc. Pendant le peu de temps que j'avais passé ici, ce bouillon était devenu ma drogue. Mon métabolisme le réclame depuis mon départ. Je prends la tasse dans mes mains mouflées. Lorsque la gnôle m'emplit la gorge, mon squelette s'illumine, comme sous l'effet d'une injection de fluorescéine, cartographiant mon espace intérieur.
— Peut-être, répond Fink en crachant dans sa tasse.
Je laisse échapper un long soupir, comme si mon âme voulait quitter cette prison, s'envoler dans le froid et laisser mon corps sur le carreau, tout vieilli, tout éteint.
D'une inspiration, je la retiens.
mmm mmm mmm
You ooo ooo ooo
Pendant le trajet jusqu'au sous-sol, Fink me jette des regards inquiets de ses petits yeux mouillés. Il porte une boîte à outils, penche du côté droit.
— Hier soir, comme tous les soirs, je suis descendu voir Elliot pour lui donner ses pilules, c'est le docteur qui a insisté pour qu'il les prenne chaque jour à la même heure. J'ai tapé trois fois, je ne rentre jamais vous savez, mais il faisait un de ces ramdams... Il n'a pas voulu m'ouvrir, alors j'ai attendu toute la nuit devant sa porte puis, il y a environ une heure, il a cessé de faire du bruit, alors je vous ai appelé. Je ne savais pas quoi faire.
— Vous l'aviez vu avant qu'il ne s'enferme ?
— Plouffe l'a emmené faire sa promenade de santé au parc. À leur retour, Elliot s'est enfermé, il portait un sachet en plastique. Plouffe m'a dit qu'il avait acheté deux bébés tortues à une bohémienne. Je ne l'avais jamais vu aussi excité.
— Excité ?
Fink opine.
— Moi aussi ça m'a étonné. La dernière fois que je l'ai vu dans cet état, c'était quand il avait appris que l'Islande allait participer à l'Eurovision.
Passé les murs pastel, les bureaux rétrécis, les tableaux noirs, sous les salles de sieste, sous les petits lavabos et les tout petits porte-manteaux. Demain, c'est kermesse pour le dernier jour de Noël. Les enfants ont travaillé toute la semaine pour décorer les couloirs de banderoles, d'accordéons et de fanions, des lampions, d'étincelles en papier. Des dessins sont accrochés partout sur les murs. On n'y devine qu'un hymne vague à la nature, à la mer, au ciel, au soleil. Il y a là des créatures terribles, une sorcière énorme, un volcan qui illumine de peinture jaune un ciel en flammes. Une série d'icebergs, sur une mer déchaînée.
— Oh lala ! quelle affaire... pleurniche Fink.
J'ai tellement sommeil, je ne cesse de bâiller.
— Et la retraite, Fink ?
— Elliot était déjà là quand j'ai pris mes fonctions. S'il meurt durant mon mandat, je ne me le pardonnerai jamais. Personne ne me le pardonnera, c'est mon boulot.
J'ai dû faire un bruit, car Fink s'interrompt.
— Oh, je sais Bracken, je sais, vous pensez comme Bram, vous êtes de ces grands cyniques qui s'imaginent que nous mourrons seuls. Mais moi, Bracken, je crois à des choses qui vous dépassent, vous et Bram. Et jamais je ne pourrais échapper à mes responsabilités ! Je suis toujours le surveillant général !
Je lève mes moufles en signe d'apaisement.
— Je ne voulais pas vous vexer.
Je sais que Fink est un gentil.
— S'il ne prend pas ses pilules, son petit cœur ne tiendra pas !
— On va trouver une solution.
La chambre d'Elliot est située à l'extrémité d'un couloir sinueux, dans le tréfonds le plus intime de l'école. Elliot a vécu ici toute sa vie, derrière la chaufferie, dans un labyrinthe de tuyaux enchevêtrés. Personne n'a jamais su comment il était installé, son intimité reste une légende. Les enfants s'aventurent parfois ici, croyant dur comme fer qu'il s'agit du toit du monde et que dans ces canalisations nagent de robustes orques épaulards. Mais tout le monde respecte le secret. Nous avons besoin d'entretenir nos secrets, nous sommes en panne de mystère.
La porte bleue de sa dépendance tient lieu de rempart, infranchissable vallée où dansent les brumes d'un âge révolu. Elliot y a punaisé un de ses dessins, tracé de cette main maladroite que je connais si bien : au centre de la feuille blanche, gribouillée, une fente noire.
— Qu'est-ce que c'est ?
Fink hausse les épaules.
— Un vagin ?
Il n'y a pas la promesse d'un enfant dans cette ouverture, ni la possibilité d'un ailleurs. Cette fente ne mène à rien. Elle n'est faite de rien. En regardant plus attentivement, mais il faut se pencher pour le voir, elle repose sur un petit socle tout fin, comme une vasque qui la soutiendrait. L'image fait naître en moi un souffle, quelque chose qui s'ouvre brusquement, comme une pupille : un mouvement d'apparition, magique et sacré. Une fente originelle, une matrice divisant la feuille, créant l'espace.
— Quand a-t-il dessiné ça ?
— Juste après votre départ. Cyldrid avait récupéré Elliot au fond de sa classe, sa présence calmait les enfants. Elle le laissait dessiner pendant les cours mais, depuis quelques semaines, Elliot était malade. Plouffe a appelé le docteur, qui a dit qu'il fallait lui donner des pilules pour que son cœur ne cesse pas de battre. On a aussi essayé de lui demander d'en faire moins mais c'est une vraie tête de mule, il continuait à taper sur ses tuyaux. C'est terrible de le savoir si proche de la mort. Moi-même, je sens que mes forces me quittent...
Il triture ses doigts aux ongles rongés, incapable de choisir entre l'initiative et la résignation : entrer dans cette pièce interdite, ou se laisser mourir sur le palier. Pour la première fois, je le vois tel qu'il doit apparaître aux enfants, peu habitués au quotidien de la vieillesse : un grand costume vide.
— Je suis désolé de vous imposer tout ça, Bracken.
— Ça va, abdiqué-je, las de tout.
— Je n'y arrive plus tout seul.
— N'en faites pas trop.
— Pourquoi êtes-vous parti ?
Une boule de haine s'est formée dans mon plexus, un soleil noir, trop noir.
— J'avais sommeil.
Le verrou est fermé de l'intérieur.
Mettre un coup d'épaule dans la porte.
Elle vacille à peine.
Fink prend son élan.
KOR
Lumières, écailles du Cap,
Ces glissements qui ne cessent d'oublier,
Ce que les mots ne disent.
Un rideau s'est déchiré. En silence, le sanctuaire d'Elliot est brisé. Violé le mystère. Nous n'aurions pas dû faire ça, c'est un pressentiment ; non, c'est une certitude.
Il y a des choses qu'on ne rompt pas.
— _Hey !_
— _La porte !_
— _Ça caille !_
Dans la pénombre : loupiotes sur les murs, guirlandes qui clignotent et crépitent, petite chambre aux murs immaculés, vaguement turquoise. Parfum d'iode. Un éventail japonais. Sur le secrétaire ouvert, un sachet en plastique rempli d'eau.
Deux tortues me dévisagent.
— _Pourquoi il nous regarde comme ça ?_
— _Je vais lui mordre le doigt !_
— _Arrête, tu vas te faire tuer !_
Fink me piétine pour entrer, ses grandes pattes de mouette montées sur ressort. La pièce est vide.
— Où est-il ? demande Fink.
La porte baille, le verrou arraché.
— Étrange, une chambre sans serrure.
— Et fermée de l'intérieur.
— Il doit se cacher.
— Grml.
Fink ouvre l'armoire, défait le lit, regarde sous le matelas, sous le sommier, sous le fauteuil, entre les coussins du fauteuil, sous le repose-pieds, dans le secrétaire. Pas sous le lit, pas dans l'armoire. Pas dans un coin sombre. Pas sous les coussins, ni sous le tapis. Ni replié dans les petits tiroirs du secrétaire.
— Elliot ? You-oo !
Sous les planchers ? Dans les murs ?
— _Mais qu'est-ce qu'ils font ?_
— _On dirait qu'ils cherchent Elliot._
— _Oh lala..._
Fouillée, la pièce. Retournés, les bouts de tapis, le matelas, l'armoire – vide, sans cintres ni sous-vêtements –, les murs, tâtés les interstices, mes doigts de laine quêtant un passage, un mécanisme. Je jette un coup d'œil aux murs délavés, rien de décollé dans cette antique tapisserie, rien qui permettrait à un vieillard, aussi fin soit-il, de s'y cacher sans être vu, sans former une énorme boule, telle une marmotte avalée par un boa. Le plafond s'affaisse au nord-est, formant un angle taillé en biseau. Impossible de se résigner à imaginer que quelqu'un puisse se dissimuler dans cette petite chambre sans fenêtre.
Le secrétaire est une boiserie pleine de termites, qu'on entend ronger patiemment. Les tiroirs sont encombrés de matériel de dessin, crayons de couleurs, règles, agrafes, gommes, trombones, gratouilles. Dans le tiroir du milieu je trouve une pomme, un Walkman sans piles, une cassette, une paire de lunettes cassée et une BD. Les lunettes sont celles d'Elliot. Les verres sont brisés, la monture rompue au niveau du nez, patchée par un bout de Scotch rose électrique. L'étiquette de la cassette est gribouillée de ratures incohérentes.
Je connais cette BD. Don di Rosa, _le Vol du Sampo_ , 1976. Picsou part à la recherche du mystérieux Sampo, avec Donald et ses neveux. Une adaptation libre du Kalevala, recueil de récits mythiques finlandais, pour ce que j'en sais. J'ai souvent vu Elliot la lire pendant mes cours. Je me souviens de ses grandes mains qui tournaient les pages, assis à un bureau trop petit, au milieu des petites têtes mal peignées. Quelque chose a été ajouté à l'album, un détail tellement grossier qu'il passe d'abord inaperçu : tous les espaces entre les cases ont été noircis au feutre épais. Sans dépasser.
— Je n'y comprends rien, avoue Fink.
— Il a pu nous jouer un tour, dis-je en pliant la BD dans ma poche.
— Impossible !
— Fink, vous n'avez quand même pas inventé tout ça pour me faire revenir...
Il y a de la buée, sur ses pupilles blanchies par la cataracte. Ses mains tremblent en triturant la petite boîte de dragées médicinales.
— Elliot est comme un fils pour Plouffe, Bracken, pour nous tous. Il va mourir si on ne le retrouve pas ! Je vous en supplie Bracken, croyez-moi ! Et s'il avait fait une grosse bêtise ?
— Grosse comment ?
— Grosse grosse.
Il va se mettre à pleurer, pauvre Fink.
— Allons, Fink, ne vous inquiétez pas.
Je suis un vrai cœur d'artichaut.
C'est pour ça que je porte des moufles.
— Vous avez dit qu'il faisait du bruit quand vous êtes descendu le voir hier soir. La seule chose qui pourrait faire du ramdam ici, ce sont les meubles. Il faut tout dégager.
Fink s'illumine. Envolées, les larmes.
— Riche idée !
— _Rooo, mais ils vont tout casser !_
— _J'espère qu'Elliot n'a pas volé le Ptyx..._
— _Le quoi ?_
Je prends les mesures de l'emplacement de chaque meuble, en tenant compte des illusions d'optique créées par le peu de hauteur du plafond, sa terminaison en pointe, penchée, qui engendre cet angle spécial. Nous bougeons toute la pièce, rotation de cubes dans des cubes. Sur les lattes du plancher, aucune trace de poussière, sur les meubles non plus. Une chambre sans fenêtre, ni poussière ?
Quand le lit et l'étagère sont placés dans le prolongement du repose-pied, la ligne divisant la chambre semble agrandir le tout, esquisse de labyrinthe menant d'un mur à un autre. Quand l'étagère est rabattue vers le fauteuil, lui-même placé à la diagonale du lit, parallèlement au secrétaire, l'espace est considérablement réduit.
Il y a d'imperceptibles traces au sol, dans le bois, raclures, striures, frottements. En les suivant, un nouvel ordre s'impose : la chaise sur la table basse, l'étagère perpendiculaire au secrétaire, repose-pied et lit parallèles, éventail déplié, l'armoire en pointe et le petit tapis rond couronnant le tout.
— Pourquoi bougerait-il ses meubles ? demande Fink.
— J'ai l'impression que ça libère des coins.
Je marche jusqu'à l'angle en pointe. Quelque chose m'y attire, un repère d'instinctif, presque magnétique, qui aimante ma curiosité.
— Celui-là, peut-être ?
— C'est juste un coin, fait Fink en haussant les épaules.
Je m'agenouille. Aucune aspérité dans cette paroi, aucune prise où forcer un doigt. Je pose ma main dans la fausse fente ; du doigt, je suis la ligne invisible qui en divise la surface.
— _Mais si, tu sais, le Ptyx !_
— _Quel Ptyx ? Elliot est juste descendu au Néant._
— _Elliot ?_
Sur le lit, bredouille, Fink se ronge les ongles. Il me semble capter l'écho d'un vol de canards à travers les murs. Peut-être des oies.
— Comment est-ce que je vais expliquer ça à Plouffe ? se lamente Fink.
— Il comprendra. Ce n'est pas de votre faute.
Fink me regarde, ahuri. Lui et Plouffe sont des ennemis jurés. Quelque chose les avait divisés, dans le passé, deux bords qui s'éloignaient doucement, sans espoir de réconciliation. Pendant mon séjour ici, j'avais souvent servi de pont.
— Les choses ont bien changé depuis votre départ, Bracken.
Fink farfouille dans sa boîte à outils.
— Je ne voudrais pas vous culpabiliser, mais votre démission a foutu une sacrée pagaille. Vous étiez notre soupape. Depuis que Bram s'est installé dans le gymnase avec sa greluche et ses machines, Plouffe n'a plus aucune autorité sur les enfants, ils se font pratiquement cours tout seuls. Je ne sais pas ce qui s'est passé, à un moment, Plouffe a cessé de s'intéresser aux enfants. Il n'en avait que pour Elliot, et Bram en a profité, avec son statut de professeur, il peut prétendre à la succession... Demain, c'est la kermesse, les parents seront là, le maire, la télévision et Nik Kershaw, et moi je sais pas quoi faire, il reste tout à mettre en place, construire la scène, finir les stands et je croyais qu'Elliot et Plouffe m'aideraient mais je suis tout seul ! Ah, je donnerais tout pour revenir en arrière. Quand vous, vous étiez là, et que tout allait bien.
Étrange comme mes souvenirs de l'école semblent s'être éteints après ma démission. Comme si tout ce que j'avais vécu ici, ce que j'avais appris, était enchaîné aux lieux, sans existence en dehors, sans aucune texture. Tout me revient progressivement. Comme si ce qui m'est refusé pendant le sommeil devenait une façon d'être au monde, dans le monde. Ici, tout le monde enseignait tout, mais chacun avait sa spécialité. Plouffe dirigeait d'une main subtile, il avait décidé de laisser les enfants plus libres qu'ailleurs, expérimentant des méthodes discutables qui m'avaient séduit : encouragement à l'échec, hyperparticipation, cours taillés sur mesure, nouvelles disciplines. J'avais trouvé un terrain de jeu à la hauteur de mes ambitions d'enseignant : non plus des cours où l'enfant était passif, mais un enseignement envisagé comme une activité dynamique, certes contrainte par des notions à communiquer, mais dont la forme s'adaptait aux besoins et à l'attention d'une jeunesse en pleine transformation. Peut-être le secret de toute discipline résidait-il dans l'acceptation d'une relation à double sens, peut-être que le rapport maître-disciple était obsolète. Tous les instituteurs n'y étaient pas favorables, certains avaient démissionné, mais d'autres, comme Bram, dont les mathématiques étaient la spécialité, avaient pris le train en marche pour en accélérer l'allure. Une nausée me remonte au souvenir des cheveux sales de Bram qui me regarde en riant de toutes ses dents jaunies, brandissant une souris d'ordinateur au bout de son fil, presque morte.
— Le seul qui comprenait Elliot, c'était vous, soupire Fink.
— Vous parlez de lui au passé.
Ses yeux se perdent dans le vide.
— Je savais qu'il lui arriverait malheur.
— Pourquoi ? C'est quoi, une grosse grosse bêtise ?
Il lève les yeux vers moi.
— Les fées, Bracken.
— Vous n'allez pas recommencer, Fink.
— Bracken, écoutez-moi...
Il se penche et d'un air de conspirateur :
— Je crois qu'Elliot a une affaire avec la reine.
— La reine ?
— La reine des fées.
— Fink...
— Vous ne me croyez pas, vous ne m'avez jamais cru, mais c'est la seule explication. Je ne sais pas si c'est elle qui l'a fait enlever ou si Elliot est parti la rejoindre de son plein gré, mais quelque chose ne tourne pas rond entre ces deux-là et mon petit doigt me dit que si le roi revenait demain, ce serait une sacrée pagaille.
Je sais qu'ici, en Islande, beaucoup de gens croient encore aux fées et aux elfes, parce que pendant longtemps le pays fut très peu peuplé ; la solitude devait être pénible à supporter, et les pierres étaient partout, comme des résidences, attendant d'être occupées. Ce n'est pas un cliché : c'est plus qu'une croyance, plus qu'une tradition. C'est une façon de voir le monde. Les fées n'existent pas vraiment, mais le monde lui, il est là, il est vivant, il respire. Le doter d'une conscience humaine nous renverrait à une obscurité trop difficile à avaler. Alors pourquoi ne pas le combler de petits « nous » qui s'agitent dans l'invisible, rafistolant les fonds de culotte, toujours là pour qui veut les voir.
— Fink, on va le retrouver.
— Si seulement elles pouvaient parler, soupire Fink en lorgnant le sachet en plastique posé sur le lit.
Les tortues, nagent, pépiant quelques filets de bulles. Elles exécutent des pirouettes, si douces, je peux presque les voir sourire, puissante tromperie de la nature.
— On dit que leur mémoire n'excède pas trente secondes.
— _Mais si tu sais, Elliot !_
— _Non je ne sais pas ! Qui êtes-vous ?_
— _Et vous, qui êtes-vous ?_
Fink se raidit soudain, son presque sourire devient grimace : Cyldrid Oup se tient devant la porte défoncée, ses petits poings bien serrés.
KOR
Vite, ces rivières, pressées de flux
Dans leur vallée, dirigent le temps
Des couches singulières de l'expérience.
Cyldrid Oup plisse les lèvres.
— Fink, qu'est-ce que vous avez fait au petit ?
Elle enjambe les débris, se signe en pénétrant dans ce royaume dévasté. Elle porte son tailleur haut sur son ventre rond. Tout son corps semble rigide, comme fait de bois ou de plastique, son chignon est paraît-il plus dur qu'une bobine de fil.
— Qu'est-ce que vous faites ici ? dit Fink en dégainant son tournevis. Je suis le seul à pouvoir entrer par effraction !
— Oh je vois, môssieur se prend pour un héros.
J'essaye de calmer le jeu.
— Cyldrid, ne l'énervez pas, il a eu une nuit difficile...
Elle me dévisage, comme si elle me voyait pour la première fois.
— Qu'est-ce que vous faites là, vous ?
— Fink m'a demandé de revenir.
— Après ce qui s'est passé ?
Fink en lâche son outil. Un instant, peut-être par pitié, je veux lui expliquer les raisons de mon départ, parce que tout le monde sait. Tout le monde sauf lui. Cyldrid sait parce que je le lui ai dit il y a un an, sur le seuil. À l'école, comme en prison, on a nos coutumes. Les autres avaient gardé le silence, c'est la loi : tout le monde est contre Fink, et ç'aurait été accorder trop de respect au surveillant général que de lui expliquer pourquoi j'avais décidé de tout plaquer du jour au lendemain.
— Je suis revenu pour Elliot, dis-je finalement.
— Où est-il ? demande Cyldrid.
— Pas ici.
Elle laisse pendre le verrou au bout de son majeur, perplexe. Cette école n'a pas connu le progrès, elle a vieilli plus vite que tout, déjà vieille au moment de naître, quasiment déjà crevée.
— Vous avez fouillé partout ?
— Oui.
Cyldrid arpente la pièce en marmonnant. Elle refuse l'idée qu'un homme puisse se volatiliser dans une chambre close, sans fenêtres.
— Sous le plancher ?
Elle frappe le sol de ses talons usés.
— Il y a probablement un passage secret, dis-je.
— Plouffe prétend la même chose, dit-elle.
Fink soupire.
— Plouffe est sénile !
Plouffe est persuadé que l'école est truffée de passages secrets. Personne ne l'a jamais cru. Pendant mon séjour ici, je n'ai jamais vu le moindre tunnel. Mais je suppose que ces vieux sols de lave sont parcourus de galeries où vont mourir les moutons.
— Elliot est peut-être en train de suffoquer derrière un mur, suggéré-je.
— Alors il finira par avoir envie de faire pipi, remarque Cyldrid.
— Qu'allez-vous faire quand il sortira, lui administrer une bonne fessée ? demande Fink.
— Et pourquoi pas ?
— J'appelle la police, dit-elle en marchant vers la porte.
— Cyldrid, vous n'y pensez pas ! l'arrête Fink. Vous laisseriez des étrangers se mêler de nos affaires ?
— Au moins, eux, ils sauront vous faire cracher le morceau.
— Le morceau, quel morceau ?
— Vous mentez, dit Cyldrid.
Fink ferme son poing, l'air mauvais.
— Qu'est-ce que vous insinuez, vieille bique ?
— Elliot n'a jamais été dans cette pièce.
Il a l'air sincèrement touché. Dans sa bouche entrouverte, je peux voir trembler ses chicots. Elle en profite.
— Vous l'avez caché et vous vous foutez de nous.
— Pourquoi ferai-je une chose pareille ?
— Parce que vous êtes vieux, et seul ?
Cyldrid a brisé quelque chose en lui, quelque chose de mou. Peut-être suis-je le témoin d'un théâtre désespéré : ces deux épaves se disputent les reliques d'un rêve. Elliot au centre, comme un enjeu, un enfant sur qui veiller quand plus rien n'a d'importance, quand les corps fuient, troués de partout par le temps, par l'usure, par les couloirs qu'on monte et qu'on descend, sans but, par automatisme, par destin.
Fink fait mine de la frapper.
Je lève une main, pleine de moufle.
— On se calme !
— C'est elle qui a commencé !
— Vieux bouc !
Il va l'étrangler. Je les sépare à bout de bras, ils partent chacun dans un coin de la pièce. Combien de fois ai-je accompli ce geste en salle des profs ; moi, l'entre-deux, moi, la frontière qui divise. Le seul moyen que j'ai de faire pression sur les fous de cette école : mon corps comme rempart. Comme cette fente, punaisée sur la porte d'un disparu. Comme cette fine paroi, entre la vie et la mort.
— Je ne crois pas que Fink jouerait avec la vie d'Elliot, dis-je.
— Alors, c'est un miracle, répond-elle.
Fink ricane.
— Cyldrid, voyons.
Elle le regarde, longuement.
— Vous lui préférez vos miniatures ?
— _On a tout oublié..._
— _Oui, il faut trouver un moyen de se souvenir._
— _Mais de quoi ?_
Un silence s'installe dans la pièce, comme un géant qui reprend son souffle après une course au-dessus des pins.
Rompre le malaise en toussant.
— Un miracle, donc ?
— En quoi croyez-vous, Bracken ? dit Cyldrid.
— Heu...
— Vous laissez pas embobiner, Bracken, ce sont des histoires de pêcheurs.
Cyldrid se lève, lui enfonce un doigt dans le plexus, comme pour l'empêcher de penser, pour stopper net sa réflexion.
— Je vous trouve très agressif, vieil homme. Les fées sont des créatures du bon Dieu, elles font partie du troupeau.
— Cyldrid, je ne prétends pas connaître la vérité, ni parler aussi bien que vous, mais si vous croyez que le bon Dieu a pris ce pauvre Elliot, prouvez-le !
Étranges, leurs rixes spirituelles. Ce pays avait été bâti sur une double foi. Dans leur sagesse, les évêques avaient admis les superstitions dans la société, en laissant les hommes vénérer les anciens dieux dans l'intimité de leur foyer. Je me demande en quoi je crois, moi. Croire en rien me paraît insurmontable. Je crois que je suis ici, dans cette pièce vide, sans Elliot, de retour malgré moi. Ce simple fait suffit à me convaincre qu'il n'y a rien en ce monde qui ne soit explicable.
— Dieu, les fées, ce serait donc la même chose ? avancé-je.
Cyldrid croise les mains devant elle, comme une petite fille.
— Si vous me demandez s'ils existent, ma réponse est oui. Si vous me demandez si j'y crois, je vous répondrai que ce sont des brebis du Seigneur dans un royaume éthéré. Le petit peuple caché dans les rochers sont des paroissiens, comme nous le sommes tous.
— Je suis ici depuis longtemps, Bracken, intervient Fink en la coupant. Des choses étranges, j'en ai vues... Des messes, entre les fentes des rochers. Des touristes disparus, tombés dans un trou. Des objets sur les pentes, disséminés, je peux le prouver, tenez, regardez...
Il fouille sa poche, en sort un petit dé à coudre, très petit, tout petit, même pour une fée. Un porte-clés, attaché au trousseau d'un passe-partout.
— J'ai trouvé ça un matin, après le solstice d'hiver. Entre les herbes de la cour de récréation, où Elliot était allé s'asseoir. Regardez la finition. Les fleurs gravées.
Un travail d'artisan, certainement. J'ai déjà vu des bibelots dans le genre, en vente dans les boutiques du centre-ville pour que les touristes se souviennent de cet endroit sans avoir à tricoter un pull.
— Ça ne prouve rien, dit Cyldrid. Moi aussi j'ai vu des lumières, j'ai même entendu des chansons. Les jeunes viennent ici la nuit, quand il ne fait pas trop froid. Et il y a des lucioles.
En retournant le dé entre ses doigts, Fink égrène sa pensée.
— Quelque part dans cette école se cachent la dernière reine et ses valets. On peut accéder à ses appartements par des tunnels secrets dissimulés dans la lave. Elliot y croyait, tout comme Plouffe croit à ses souterrains. Enfant, il devait avoir vu ce palais d'ivoire et de velours ; tous les enfants, s'ils en ont envie, peuvent le voir. Sa maladie a conservé la fraîcheur de son regard. Le palais d'Hamarinn se tient au bord d'une faille, une fine ligne qui sépare le royaume mortel de Féerie. Le couple royal garde cette division, depuis des siècles les deux mondes ne communiquent pas. Mais le roi a disparu, alors...
J'émets un « mrrrl » de frustration.
— Vous me prenez tous les deux pour un idiot.
— Pourtant Elliot a disparu, dit Fink, tout fier.
— Et s'il avait fait une fugue ? demande Cyldrid.
Fink grogne.
— Elliot n'y voyait pas à un mètre, dis-je, et il est parti sans lunettes. Ça ne lui ressemble pas. Ses verres sont cassés, se pourrait-il qu'il se soit battu ?
Leur silence ne me dit rien qui vaille.
Très bien.
— Je vais chercher Plouffe.
— Plouffe ? Vous n'y songez pas ! hurle Fink.
— Ce serait pire ! proteste Cyldrid.
Pour la première fois, ils sont d'accord.
— Vous me cachez quelque chose.
Ils me regardent, ahuris, comme ces tortues. Ou bien est-ce moi la tortue, dans le sac plastique de cette chambre marine ?
— Mais enfin Bracken, pas du tout.
— Où allez-vous chercher tout ça ?
— Vraiment, vous me peinez.
— Je ne sais pas ce qui se passe dans cette chambre, dis-je, mais si le proviseur n'est pas au courant, ça va poser beaucoup de problèmes demain matin. Vous vous rendez compte des conséquences ? Vous n'êtes pas tout seuls ici, c'est un service public !
— Oh lala, la kermesse.
— Oui, et si vous voulez rester ici à parler métaphysique, c'est votre problème, mais moi, je veux savoir !
Respirer.
— En plus, il faut changer l'eau des tortues.
— _Oh oui !_
— _Il est sympa lui !_
— _C'est qui ?_
KOR
Des carrés magiques émane nouvelle nature,
Pure interprétation, herbes régurgitées,
Car ces constructions ne sont de miel.
Dans le vaste évier en aluminium de la cantine déserte, remplir un vase rond, vider le sachet dans le bocal. En le manipulant, je me rends compte que je n'ai pas retiré mes moufles depuis mon retour. Elles font partie de moi. Leur maladresse me donne l'assurance de rester ici sans partir en hurlant. Si je les enlevais, j'aurais peur de ne plus y trouver mes mains.
— _Bracken, il s'appelle Bracken._
— _Merci Bracken !_
— _Il vous entend pas !_
Fermer le robinet.
— Voilà, les tortues, maintenant ça devrait aller.
— _J'en avais marre du caca !_
— _On fait quoi maintenant ?_
— _Caca !_
Cette école a des propriétés qui m'échappent. La nuit, les lieux changent de configuration, prennent la teinte des rêves ; les recoins deviennent volumes, les surfaces se creusent. Je me cogne contre un mur que je ne pensais pas là. Est-ce donc cela, la propriété onirique d'un lieu ? Quand tout change de forme, quand les visages se meuvent, que les murs se déplacent. Je n'ai jamais rêvé, je n'ai aucun souvenir de ce monde. Tous mes souvenirs viennent d'ici, du monde réel, mais mes souvenirs de l'école sont ce qui s'approche le plus d'un rêve.
Je traverse le couloir extérieur qui relie les deux unités de l'école ; la baie vitrée vient d'être nettoyée. Dans le ciel, les étoiles luisent encore, pas tout à fait aube, plus tout à fait nuit. Un monde entre deux, où tout s'étire éternellement, où la mer disparaît dans la couleur de la nuit. De chaque côté luit le _hraun_ , le champ de lave, devenant doucement, vers la ville, la falaise. Un vaste plan gris-bleu, strié de fentes, parsemé de touffes rugueuses. Au loin, des lampes de camps scintillent sur le cratère du Burfell. Il y a des gens là-bas ce soir.
Il me semble voir une lueur, sur les rochers. Un feu follet. Un instant, mon souffle devient si régulier, si profond, que je prends conscience de ma présence ici et maintenant, à travers la douceur de ma respiration.
— _Ma belle, j'ignore encore votre prénom..._
— _Hiiii... Pourquoi pas Diane ?_
— _Oh, Diane,_ cute _._
Plouffe n'est pas dans son bureau. La lampe de chevet est allumée, un mégot de cigarette refroidit dans un cendrier. Il était ici il n'y a pas si longtemps.
— _Et vous ?_
— _Mmm... Roméo !_
— _Hi_ _hi !_
Les étagères sont pleines de livres aux noms compliqués, recueils de poésie, guides, romans, anthologies... Plouffe est francophile. Avant de devenir proviseur, il était professeur de français à l'université de Reykjavík. Il avait passé sa vie à s'intéresser à la poésie française. Plouffe est spécialement un admirateur de Stéphane Mallarmé, un poète que je n'ai personnellement jamais lu. Il possède des dizaines d'ouvrages le concernant, des recueils, des biographies, des analyses. Il y a des carnets ouverts sur le bureau, des cartes en islandais, des diagrammes, des figures géométriques.
Soudain, j'entends des voix. J'entrebâille la porte pour jeter un coup d'œil dans le couloir. Émergent des ténèbres deux adolescentes en baskets et grosses mailles. Elles portent des plateaux-repas.
— Et il me dit : "Non, je ne veux pas utiliser le téléporteur, c'est capricieux."
— Oh !
— Carrément !
— Et il compte s'y prendre comment ?
— Il dit qu'il gardera la balise sur lui, pour la sauvegarde.
— Oh !
— J'ai dit aux macareux de ne pas intervenir, mais toutes les boules de cristal ont dit la même chose : demain matin, ce sera la fin, alors personne ne veut plus rien entendre... Attends, tiens-moi la porte.
Le gymnase s'ouvre ; vives discussions ; j'ai le temps d'apercevoir des tables, des câbles partout, des terminaux d'ordinateur, des chaises et des sachets en papier, puis le battant se ferme, plus rien. Qu'est-ce qui se passe là-dedans ? Pourquoi les ordinateurs ne sont-ils pas en salle informatique, pourquoi les enfants sont-ils là si tard ? Où est Plouffe ? Pourquoi je panique ?
Plus loin dans le couloir, une autre baie vitrée donne sur le _hraun_.
Encore plus loin, si je suis discret, le gymnase et ses vestiaires donnant sur le terrain de foot... Mmmm...
— _Il sort ?_
— _On dirait bien..._
— _Mais il fait froid !_
Dehors, la mousse éteint tout bruit de pas. Un petit avion zonzonne dans la fraîcheur vespérale, simple clignotant vert au firmament. J'ai l'impression de voir danser la nuit. Les masses sombres couronnées par le massif de l'Helgafell bâillent sous les étoiles. Il y a beaucoup de montagnes sacrées ici, toutes des femmes. Ça m'a toujours semblé judicieux.
On peut entrer dans les vestiaires par l'extérieur. Parfois, quand il faisait bon, j'officiais comme arbitre des matchs de foot. J'ai toujours les clés, je ne les ai jamais rendues. Pendant que je restais chez moi, sous ma couette, à regarder mourir le plafond, je continuais à toucher ma paye. Je n'avais jamais officiellement démissionné. Pour l'administration islandaise, j'existais toujours.
J'entre dans les vestiaires enténébrés en poussant des ballons de foot ; je me place derrière la porte, l'oreille tout contre.
— Il me faut la position du Cap, maintenant.
— Bram, ça va prendre des heures pour trianguler.
— Passons par en dessous, nous avons la position de l'iceberg.
— Je ne veux plus attendre.
Bram. Un Viking baryton aux lunettes de vue trop grandes. De mon point d'observation, je ne devine que sa crinière rousse, mal peignée. À mon départ de l'école, Bram utilisait déjà les ordinateurs pour enseigner aux enfants comment faire des calculs, mais pendant mon absence quelque chose était arrivé, quelque chose que je ne comprenais pas. On dirait qu'ils jouent, mais tous ensemble, au même jeu. Se pourrait-il que Bram ait décidé de changer la façon d'enseigner ? Il avait souvent prédit, au cours de nos réunions hebdomadaires, la façon dont les ordinateurs allaient s'emparer de la planète, comment nous allions être tous connectés les uns aux autres. Ces réunions auraient facilement pu tourner au pugilat. Il me raillait, moi et mes crayons, mes feuilles de papier et mes petits mots écrits à la va-vite. Moi, l'analogique.
Je capte un reflet digital, une fenêtre colorée, du texte et des cubes aux coins arrondis, quelque chose d'inconnu, altéré par la bande coulissante du rafraîchissement de l'écran, scintillant.
— J'ai pensé à une pêche aux macareux... glousse une voix.
— Je vais en _rezzer_ deux cents, j'en ai même avec du son.
— Coin coin coin ! cancane un synthétiseur.
— Ah zut, j'ai fait des canards.
Bram est assis devant trois consoles, ses mains courent sur un clavier, il tape des indications dans un phylactère, déplace un personnage, puis le paysage se change en plage, trois étranges soleils se couchent sur la ligne d'horizon. Quelques clics de souris plus tard, un personnage sort une épée flamboyante et commence à tailler dans la masse de canards. Mes compétences en islandais sont trop sommaires, je ne suis pas certain de tout comprendre.
— Et le Mnyx ?
— Toujours stable.
— La balise de l'île est active.
— Mettez le signal en attente.
En reculant, je percute un nouveau ballon, qui va rouler dans les ténèbres jusqu'au mur. Mais il ne rebondit pas. Il s'arrête. Comme si quelqu'un l'avait stoppé net. Je ne fais plus un geste. Ma respiration me semble plus bruyante qu'un camion. Il y a quelqu'un avec moi ici, quelqu'un que je n'ai pas vu en entrant, qui m'a guetté tout ce temps, sans bouger. La terreur me saisit. Je scrute cet angle noir, où tout pourrait se jouer. Si je me mettais à courir maintenant, des mains jailliraient-elles pour me tirer ?
— _J'ai un mauvais pressentiment..._
— _Qu'est-ce qui se passe ?_
— _Nous ne sommes pas seuls, Roméo._
J'avance de quelques mètres. Le ballon ne revient toujours pas. Je tends les bras. L'obscurité est plus compacte. Rien ne me permet de discerner qui est assis sur ce banc. Je sais qu'il y a quelqu'un derrière le rideau d'ombre. Si je cesse de respirer, je peux l'entendre lui, il respire fort, c'est un vieillard. Il est malade. Mais pourquoi est-ce que je ne distingue rien ? Pas un bras, ni une jambe, ni même un côté de son visage.
— Elliot ?
Dans l'ombre, quelque chose remue.
Soudain, d'un coup de nageoire, une des tortues m'éclabousse.
Dans l'ombre, un mouvement s'esquisse.
Le ballon revient vers moi.
— _Bracken !_
— _Bracken est en danger ?_
— _Bracken ! Il faut partir !_
Il fait tellement froid dehors que mon souffle se cristallise instantanément. Je traverse le champ de lave sans un regard en arrière, mais je sais que quelque chose me suit. Quelqu'un est là, derrière moi. Il est sur mes talons, je peux l'entendre, le sentir. De la lave jaillissent des ombres, comme de la fumée. Est-ce de la brume, ou bien une armée invisible se réveille-t-elle pour m'emmener ?
Longer l'école, en manquant de trébucher dans les trous, en glissant sur une mousse qui pulse d'une nouvelle lueur, comme respiration. Monte une odeur de pluie, de crotte de chèvre fermentée, de lait caillé. J'entends des clochettes, des carillons, presque un galop. Combien sont-ils ? Je n'ose me retourner ; tout droit, tout droit, il faut filer tout droit. Vite, la terrasse, le béton, mes pieds qui claquent et la baie vitrée. Poser les tortues, vite, refermer la baie.
Vite, la verrouiller.
— _C'était moins une !_
— _Vous avez vu quelque chose ?_
— _Non, j'ai fermé les yeux..._
Je reprends ma respiration, je n'ai jamais couru aussi vite. J'ai bien cru mourir, sans savoir pourquoi, ni aux mains de qui. Je manque d'exercice, je n'ai pas senti mon corps depuis longtemps. Je colle mon nez sur la vitre. Rien. Rien que la lave, à perte de vue, disparaissant dans les ténèbres. J'essaye de deviner la silhouette de mon poursuivant, mais tout se confond avec les formes torturées des cheminées de lave. Une sensation m'étreint, comme si le paysage me regardait. Comme si, dans son silence, il m'observait.
On s'agite derrière moi, à l'entrée du gymnase, quelqu'un semble-t-il a fait une découverte, des jeunes crient « Trix ! Trix ! ». Plusieurs enfants accourent des salles d'études pour entrer dans le gymnase, tout excités. Combien sont-ils cette nuit, à ne pas être rentrés chez eux ? Les parents sont-ils au courant ? Quelle pagaille, Fink a raison, quelque chose ne tourne pas rond. Je me fais discret, tout petit, tout replié, je longe les couloirs, en marchant en crabe.
— _Trix, Trix, ça me dit quelque chose..._
— _C'est pas une sorte de moule à tarte ?_
— _Ou alors c'est un marsouin !_
Plouffe n'étant pas dans son bureau, j'en déduis qu'il est aux toilettes handicapés, son fief aménagé.
— Allô, allô ? fais-je, en toquant à la porte.
— C'est occupé ! répond une voix éraillée.
— Papy Plouffe ?
— Bracken ?
Bruit d'un cadenas qu'on déverrouille, puis la porte s'entrebâille sur une broussaille de sourcils.
— Bracken, bougre d'idiot, qu'est-ce que vous foutez là ?
— Je suis revenu, papy Plouffe !
Plouffe est une couverture en patchwork dans un vieux fauteuil roulant en bois. Il se tient voûté, tel un prophète annonçant la venue d'un autre monde, encastré dans le nôtre. Sa moustache lui mange la moitié du visage. Quand il parle, ses lèvres en mâchouillent la frange.
— Papy Plouffe, j'ai besoin de vous.
— Après nous avoir abandonnés ?
— Je suis désolé papy. Elliot... Elliot a disparu.
Plouffe ferme un œil.
— Disparu ? Il n'est plus dans sa chambre ?
Son accent français est un subtil mélange d'alsacien et de portugais. Il prononce avec emphase chaque mot de plus de deux syllabes, tout en avalant les adverbes et les articles. Il a plus de vocabulaire qu'un bréviaire.
— Non.
— Il a pris ses pilules ?
— Non. Et Fink ne l'a pas vu sortir. Il dit qu'Elliot a été enlevé par les fées.
— Vous avez cherché ?
— Oui.
— Partout ?
— Oui.
— Partout partout ?
— Comment ça, partout partout ?
— Grand Dieu, roulez-moi en bas !
KOR
Si nous devions compter les pas
Qui séparent l'ombre de son ombre,
Pourrais-tu me dire où s'en va le vent ?
Je pousse Plouffe jusqu'au lobby, puis je le porte dans l'escalier. Je le laisse allongé sur le plancher pendant que je remonte chercher son fauteuil. Pendant tout le voyage, il ne cesse de radoter, sombre, prophétique dans sa moustache en balai-brosse.
— Fink n'aurait jamais dû le laisser seul ! À quoi est-ce qu'on le paye ? Il est incapable de prendre ses responsabilités, c'est pourtant pas compliqué d'abolir son petit moi pour le bien de la communauté !
— À quoi est-ce qu'ils jouent, dans le gymnase ?
— Comment voulez-vous que je le sache ? Je suis directeur, pas informaticien ! C'est Bram qui a convaincu le pays qu'il fallait qu'il continue ses expériences sur une plus large population. Croyez-moi, si j'avais eu le choix, ça ne se serait pas passé comme ça, ah ça, non ! J'ai essayé de m'interposer, mais personne n'a rien voulu entendre : c'est la marche du progrès, tous des robots ! Ah, le requin... Il a profité que mon attention soit accaparée par Elliot pour avancer ses petits pions, ah, mais je vais pas me laisser faire, on ne peut pas éduquer les enfants avec des jeux, il faut un peu de rigueur tout de même, des cadres, des règles, ces jeunes n'ont aucune idée de ce que c'est que d'apprendre dans un livre !
Je porte Plouffe à bout de bras pour le remettre sur sa chaise, je le fais glisser jusqu'au fond du couloir.
— Je vous le dis Bracken, tout ça finira mal, je ne veux pas sonner comme un vieux rabat-joie, mais avant, au moins, on avait des valeurs !
Dans la chambre, Fink et Cyldrid ont essayé de nouvelles configurations de meubles, sans y trouver quelque sens. Quand Fink aperçoit Plouffe, il devient si rouge que je crains qu'il ne meure sur le tapis, là, tout de suite.
— Plouffe !
Il se tourne vers moi, furieux.
— Tartuffe !
— Restez en dehors de ça, Plouffe ! dit Cyldrid.
— C'est le problème du collectif, réplique le proviseur en roulant au centre de la pièce.
Il fait demi-tour dans son fauteuil, puis hurle :
— Tu peux sortir maintenant, Elliot !
Il tousse.
— Elliot ! Ce n'est pas drôle !
Personne ne bouge. Plouffe nous regarde. Il tousse.
— Elliot ! Sors de ta cachette !
— Gardez votre souffle, Plouffe.
Il grogne.
— Vous allez me faire croire qu'il aurait disparu sans laisser de trace ?
— Les fées, murmure Fink.
— C'est un enfant, dit Plouffe.
Je ne peux m'empêcher de tiquer.
— Elliot a plus de quatre-vingts ans !
Plouffe me regarde doucement.
— Mon petit, il y a des choses dont vous ne savez rien, et eux non plus. Nous sommes tous responsables d'Elliot.
— Vous plus qu'un autre, fait Fink.
— Vous avez beau râler, ça ne changera rien. Vous étiez censé le surveiller. Depuis combien de temps a-t-il disparu ?
— Quelques heures.
— Alors il faut faire vite. Il faut trouver où il a réussi à se cacher.
— Mais enfin, vous voyez bien qu'il n'est nulle part...
— Il y a des endroits que nous ne voyons pas.
— Que voulez-vous dire ?
— Oui, soyez plus précis, renchérit Cyldrid.
— Il faut fouiller mieux.
— _Qu'est-ce qu'il a dit, le jeune, dans le couloir ?_
— _Fryx !_
— _Non, Trix !_
Plouffe a demandé à Fink de retirer les lattes du plancher avec ses outils, pour voir si Elliot ne s'y cacherait pas. Le surveillant général s'exécute en râlant, pourtant je sais que le projet lui plaît. Cyldrid est assise dans le fauteuil, elle a tiré des aiguilles de son sac et, en paix avec elle-même, elle nous regarde faire en tricotant un pull bleu. Moi non plus, je n'ai rien de spécial à faire. J'ai déjà fouillé, et je suis fatigué.
J'ai choisi très jeune la facilité : j'aimais dessiner, puis j'ai aimé voir les autres le faire. Je n'ai jamais voulu exercer d'autorité mais il me fallait un domaine, et ce domaine, c'était le trait. On m'a soupçonné d'être un manipulateur, mais la vérité, c'est que j'aime voir des gens se concentrer sur un point. J'ai un rapport au dessin qui n'est pas simplement esthétique, ou même hygiénique. C'est un sens métaphysique de la forme, en tant que chariot de l'âme ; sans cela, elle retombe. J'ai appris que pour survivre dans ce monde, il était plus facile d'enseigner que de faire la guerre. C'était ce que je faisais le mieux, communiquer ma passion, transmettre les gestes intimes et originaux sous-tendant ma pratique du dessin, je pouvais les faire émerger des corps de ces enfants, qui savaient à peine marcher. Prendre une main, la lever pour eux. Leur apprendre à respirer, avec leurs petits poumons. Décrisper leurs petits doigts. Tout devait être revu : mes exercices d'adulte étaient devenus de simples mouvements, doux, imperceptibles. La vie avec de jeunes enfants est un ralenti, tout y devient très lent, très long. Le temps, les distances changent de nature. J'avais découvert un nouveau monde, un monde de couleurs, de pastels, de fluides. Un monde où les traits prenaient vie. Où les lignes nées des figures mal maîtrisées pouvaient changer la pensée. J'aimais les voir vivre, ces dessins d'autres mains que les miennes. Comme des mondes à apprivoiser, à visiter. Le mystère comme un jeu d'exploration.
— _Gnix !_
— _Pouix !_
— _Bouh, que c'est ennuyeux..._
Plouffe roule jusqu'à moi.
— Bracken...
— Oui, papy ?
— Bracken, mon petit Bracken, asseyez-vous un instant.
Je m'accroupis pour me mettre à sa hauteur. Une chaleur émane de lui, elle irradie comme une aura. Son costume rapiécé me rappelle une fête auquel nul invité ne se serait rendu.
— J'ai besoin de quelqu'un sur qui je puisse compter. Je ne peux pas faire confiance aux autres. Vous me paraissez plus pragmatique.
— Moi ?
— Je sais que vous allez croire que je suis fou...
— Quelle idée.
— Elliot avait un secret. Un secret terrible. Il ne l'a jamais dit à personne, il ne le partageait qu'en code, et je crois que cette chambre est le code, au bout du code. Il était votre élève, vous le connaissiez à votre manière : qu'est-ce que vous savez ?
— Pas grand-chose, papy. Je n'ai que ses dessins, et c'est juste naïf. Il n'y a pas de secret.
Elliot avait besoin d'exprimer des formes et des couleurs. Il n'avait pas de références. Et il n'a jamais dessiné cette chambre, ni aucun de ses objets. Il était parfois très abscons, je ne suis jamais parvenu à en tirer du sens. Il était fermé, comme peuvent l'être les autistes les plus évolués, les enfants sans parents, les vieillards qui se savent finis.
— Il ne vous a jamais parlé ?
— Non.
Je me retiens de lui rapporter la phrase qu'Elliot m'avait confié dans un moment d'impuissance devant son dessin. Pour la première fois, je comprends le plaisir d'Elliot à retenir ses mots.
— Nous n'avons que ces murs pour le trouver, dit Plouffe.
— À quoi pensez-vous ?
— À un inventaire complet. Il faut l'envisager comme... Voyez ce bleu partout, n'avez-vous pas l'impression d'être...
— Dans le ventre d'une baleine ?
Plouffe applaudit faiblement.
— Une chambre marine... Elle ne nous a pas tout dit. Prenez ces crayons, ce papier, notez tout, dessinez. Rien ne doit nous échapper. Le moindre objet peut constituer un indice.
Je regarde ces instruments, ces crayons, ces feuilles, comme des artefacts extraterrestres. Je n'ai pas touché un crayon depuis que je suis parti. J'ai perdu toute envie de tracer une ligne.
— Très bien, papy.
Il me tape sur l'épaule.
— Atta boy !
— _Et si on jouait à cache-cache ?_
— _D'accord ! Je me cache, vous comptez jusqu'à trois !_
— _OK ! 1... 2... 3... J'arrive ! Oh ! Vu !_
J'avais oublié. Comme un ami imaginaire, enfermé dans un placard pendant trop longtemps, contre lequel on peut s'aplatir pour ne plus sentir le poids de la gravité, du monde. Un ailleurs, juste enlevé d'une pointe habile, légère dans le poignet. Soudain, restitué, le monde là, sous le trait, qui devine et devient, par instinct, un modèle de ce que je vois, par l'œil établi, le froid clinique d'une architecture, d'un squelette que personne d'autre que moi ne peut voir, vomi tout en style. J'ai du mal à dessiner avec des moufles.
Inventaire de la chambre d'Elliot, établi selon mes relevés précis, entre les coups de pioche, les « Atchoums » et les disputes : un lit avec couverture, un matelas ; une armoire, vide ; un dessin et une punaise ; un secrétaire où sont désormais stockés les objets, un repose-pied, un éventail japonais, un tapis rond, du matériel de dessin, un Walkman et son unique cassette, trois guirlandes de loupiotes bleues, des lunettes cassées, une BD. Rien de neuf.
La pièce fait à peu près vingt-cinq mètres carrés. Elle paraît plus grande ou plus petite au gré de nos mouvements et l'angle du plafond n'arrange rien. Pour la première fois, en le déplaçant, je prête attention à l'éventail. Fait d'un papier laiteux tendu entre deux baguettes laquées de noir, on y voit un dragon enlaçant une haute montagne sombre sur laquelle brille une lumière. Un chemin serpente tout autour, ou plutôt une sorte d'escalier de bois aux marches branlantes qui émerge d'une brume lourde et menaçante. Le dragon lui-même ne ressemble pas à l'idée que je me faisais d'un dragon japonais. Il semble plus abstrait, moins brutal et animal.
Je me perds dans une forêt de traits quand Fink pousse un cri :
— Bougrefoutre...
Il semble qu'il ait trouvé une cachette. Sous le plancher, le sol, non pas cimenté mais bien rocheux laisse apparaître un trou suffisamment vaste pour s'y tenir accroupi.
— C'est volcanique, dit Plouffe, penché sur les accoudoirs de sa chaise. Du basalte, on en trouve partout par ici.
— On aurait mis des murs autour d'un rocher ? dit Cyldrid.
— Cela signifierait que nous sommes dans une grande coulée de lave, dis-je. Quand est-ce que cette école a été bâtie ?
— Eh bien, répond Plouffe, je dirais à la fin du dix-neuvième siècle. C'était un orphelinat. Il a été transformé dans les années cinquante par la mairie, juste avant que je ne prenne mes fonctions. Elliot, lui, était déjà là...
— Aménager un rocher pour y vivre me semble une très bonne idée, dit Fink. Des elfes l'ont fait avant nous.
Je saute dans le rectangle. De l'intérieur, la roche semble bleutée elle aussi. S'agit-il d'un reflet minéral, ou est-ce de la peinture ?
— Ça me rappelle... dis-je.
— Quoi ?
— Amityville...
— De quoi parlez-vous, Bracken ?
— Une histoire de maison hantée. Une imposture, mais à laquelle j'ai longtemps cru. Je ne me souviens plus des détails, mais il me semble que sous l'escalier de la cave, se trouvait une petite pièce, non répertoriée sur le plan de la maison. Un endroit où deux personnes assises pouvaient tenir ensemble, une pièce aux parois entièrement peintes en rouge. On avait soutenu que le criminel qui avait sévi dans la maison y pratiquait des rituels sauvages, sataniques. La pièce en question a été complètement réinventée dans le film, le réalisateur en a fait une sorte de caverne semblable à celle-ci, aux parois peintes en rouge. Tout le monde prétendait que cette pièce était une porte de l'enfer.
Plouffe hoche la tête.
— Ulysse, pour se rendre en enfer, fit creuser un trou rectangulaire. Il couvrit le trou de farine, puis de sang qu'il fit couler de ses mains. Les fantômes vinrent lui rendre visite. Les passages vers l'enfer ne sont pas des tunnels, ni des routes. Ce sont des trous où l'on se place pour attendre la mort, dans un simulacre de descente vers le centre de la Terre. Pour les anciens sortis des hautes cavernes, l'enfer était simplement ce qui se trouvait au-dessous – jungles, forêts – d'où des sons inconnus, terrifiants, remontaient. L'enfer est un point de vue.
— L'enfer est peut-être bleu, dis-je.
— _J'ai l'impression de ne servir à rien..._
— _On a sûrement connu des choses avant, non ?_
— _Avant quand ?_
Une fois la pièce retournée de part en part, le moindre de ses recoins débarrassé de toute poussière, une fois le trou bleuté méticuleusement arpenté, dessiné, documenté, le moindre de ses recoins tâté avec soin, il est temps pour nous de procéder à un premier bilan des fouilles. Nous nous réunissons autour du lit éventré pour écouter Plouffe.
— Elliot est plus malin que nous tous. Si ce qu'affirme Fink est exact, si Elliot n'a pas quitté la pièce et qu'il a bien disparu ici, quasiment sous nos yeux, alors la solution la plus improbable est la vérité, comme dirait Nestor Burma.
Je veux lui faire remarquer que la citation est de Sherlock Holmes, mais je n'ai pas envie de nuancer sa foi en l'intelligence française.
— Aurait-il une raison quelconque de faire une fugue ?
— Sa vie était ici, Bracken. Soit il a été enlevé, soit il est parti de son plein gré.
J'ouvre la bouche comme un poisson mais sans produire la moindre bulle.
Fink noue un mouchoir autour de sa tête.
— Les fées.
— Je pensais à quelque chose de plus concret, Fink.
— Un miracle, dit Cyldrid, qui a fini une manche de son pull.
— Un truc de magicien ? dis-je, espérant en finir avec les hypothèses. Il doit bien y avoir une fente, un trompe-l'œil que nous ne voyons pas.
— Je ne comprends pas, dit Fink.
— Oui, soyez plus précis, insiste Cyldrid.
— Bracken a raison, conclut Plouffe. Il faut travailler à la surface des choses, pas autour d'elles. Et la surface qu'il nous reste à explorer, c'est le mur, n'est pas ?
KOR
Fente, ouverture,
Fissure d'iris en clarté,
Perds-tu de vue la Voie lactée ?
Plouffe demande à Cyldrid et Fink d'entasser tous les meubles dans un coin afin que nous puissions examiner la tapisserie dans son ensemble. Il lui paraît important de trouver des traces d'occupation antérieure. Cyldrid a achevé une seconde manche. Ça m'amuse de voir tous ces vieux se démener dans un placard.
— Nous savons qu'Elliot habite cette chambre depuis toujours, et qu'elle a dû subir des rénovations successives. Elliot nous a précédé, il a peut-être connu la première administration de l'école. S'il était ici depuis tout ce temps, alors la pièce doit en porter les stigmates, car toute une vie ne se conserve pas sans poussière. C'est dans ces marques que nous trouverons peut-être son secret, cette cachette abolie que nos yeux ne peuvent déceler.
Je note que Plouffe semble goûter la notion d'abolissement. C'est un terme typique de ces poètes qu'il lit sans cesse tout en les citant à moitié.
— Chaque meuble est comme un mot. Il règne une profondeur ici. Nous devons peigner, peler chaque couche. Si nous voulons connaître Elliot, il nous faut un canevas. Il s'agit d'une chasse au trésor en deux dimensions, les amis. Derrière ces murs se trouvent d'autres tapisseries, d'autres strates de son occupation. Tout comme sous le plancher il y avait un autre sol.
— Je ne suis pas votre ami, crache Fink.
— Nous n'avons pas d'instrument pour décoller la tapisserie, fais-je observer.
— On fera des carottages.
— _Vous vous souvenez d'un avant ?_
— _Toujours pas, pourtant j'essaye._
— _Poussez plus fort !_
Le coin de tapisserie que j'ai choisi a l'air assez souple pour me permettre de frotter avec les instruments rudimentaires dont je dispose, c'est-à-dire la plume d'un stylo et quelques couteaux à peinture.
Le papier est très épais. Une fois grattée la première couche, une nouvelle apparaît, d'une couleur plus foncée. Toujours unie. Je continue à gratter. Une autre couche, cette fois semée de fleurs, à peine écloses. Je gratte encore : des vagues apparaissent. Suit une autre épaisseur et encore une autre, puis une dernière. Ah non... Encore une... Il y a combien de vies là-dessous ? On dirait que le temps ne s'est pas écoulé correctement, ou alors trop vite. J'ai l'impression d'être aux prises avec un fantôme fétichiste du motif de tapisserie. Le papier bleu semble désormais turquoise. Plus je remonte le temps, plus les motifs semblent compacts, une névrose faite papier peint. Il serait parti d'une effroyable complexité du moment présent pour en arriver à la pureté du vide, de l'encre invisible, des motifs implicites. Jusqu'où vais-je remonter ? La tapisserie devient mauve, orange, puis bleu-vert, bleu nuit, bleu électrique, peuplée de petites tortues stylisées, un vivarium de papier. Une boursouflure de vie, une éternité d'enluminures. Peut-on lire la personnalité de quelqu'un sur ses murs ? Voici la légende d'Elliot, comme ces tertres qu'on arase strates après strates pour y lire les motifs d'une existence, d'une circulation. Une légende de peu, enfermée dans un corps souriant dont il est impossible de forcer la serrure. Une chambre, une citadelle, où luit juste un œil, ou pas même le vide d'un repli sur soi, barricadé de l'intérieur. Un parchemin craquelé, vierge de toute écriture, comme une peau pelée sous le soleil d'une plage écrasée par le...
— _Des tortues sur la tapisserie, on aura tout vu._
— _On ne sert à rien._
— _Moi, ça me frustre._
Quelque chose, soudain, m'aveugle. Un flash me traverse. Une fulgurance visuelle, pas une réalisation, une vrille qui fait un trou dans mon âme. Aucune évidence. La tache s'étend sur mon iris, comme un virus, puis s'éteint lentement, résidu sur rétine, où se peint doucement la chambre, ses occupants, et Plouffe, qui me scrute d'un sourcil inquiet.
— Bracken ?
— Rien... Je... dis-je en portant la main à mes yeux.
— Parlez, mon petit...
— J'ai cru... Une lumière...
Je tends le doigt vers l'angle. Ce même angle qui m'avait attiré plusieurs fois, cette plissure. Les larmes ne s'arrêtent pas de couler sur la laine de mes moufles, comme si la lumière avait déclenché en moi une fontaine, une source perpétuelle. Je ne me sens pas triste, mais mon corps m'échappe, et je ne peux retenir ces rivières. Cyldrid va chercher un mouchoir dans sa sacoche puis, de ses doigts d'ange, elle écarte délicatement mes paupières pour essuyer mes larmes. Puis elle effleure délicatement ma joue de la main. Je ne sais pas ce que ça me rappelle, peut-être une grand-mère que je n'ai pas connue. Certains gestes conditionnent-ils notre rapport au monde ? Une séquence d'intimité enregistrée par des récepteurs autour de nous, dans l'air, invisibles, un schéma de bien-être, à jamais ressuscité quand surgit le drame du quotidien.
Un doudou sensuel.
— Atchoum, éternue Cyldrid.
Elle a couiné doucement, presque désolée de nous déranger. Sur le moment, je ne me pose pas la question, mais quand elle recommence je me dis qu'il y a un courant d'air.
— C'est très humide ici...
Contre toute attente, Fink lui tend le chiffon qu'il avait noué sur son crâne. Un geste doux. Elle le prend en souriant. Quand elle rayonne ainsi, malgré l'âge, les os brisés, la morve, les pets d'un corps enfui, je peux voir l'enfant, assise sur les marches d'un escalier. Et Fink, assis près d'elle, il y a longtemps.
— Merci, dit-elle en baissant les yeux.
Fink rougit.
— _Elle est bien trop vieille pour lui !_
— _Roméo, croyez-vous que l'âge signifie quelque chose ?_
— _Il va grandir, elle va vieillir._
— Selon Bracken, il y a une source de lumière quelque part derrière cette tapisserie, dit Plouffe. C'est peut-être Elliot.
— Après la tapisserie, il y a le mur, et ensuite on est dehors, dit Fink.
— Mais y a-t-il une couche avant l'extérieur, mais après le mur ?
— Euh...
— Je ne comprends pas.
— Oui, soyez plus précis.
— Qu'avez-vous vu exactement, Bracken ?
Je ferme les yeux, incapable de trancher. Est-ce que tout ceci est réel, ou suis-je prisonnier d'un monde que rejette ma raison ?
Une tache persistante. Qui ne veut pas s'éteindre. Je pense au dessin d'Elliot, punaisé sur sa porte. Cette même fente, dont ne sort pourtant aucune lumière. Qu'ai-je vu ? Suis-je à ce point désespéré, pour me construire des illusions qui me contraignent à rester ici, parmi eux, à chercher un disparu ? Ne suis-je pas en train de me mentir ? Cette lumière était un appel, comme une chanson venue de très loin, de plus loin que le monde, de plus loin que la pensée elle-même. Serait-ce Elliot qui m'appelle, de son endroit vide, de cette absence de lieu ?
En allant le déterrer sous la porte, je tire sur le dessin, qui se déchire à moitié. La fente coupée en deux dans le sens de la longueur, presque parfaitement. Pourtant la ligne est si fine... c'est un miracle que la découpe puisse être aussi précise. Je le tends à Plouffe.
— C'est ça que j'ai vu.
Plouffe fourrage dans sa moustache en serrant le bout de papier ; sans attendre, Cyldrid lui arrache le dessin des mains.
— Je me souviens quand il a dessiné ça. C'était en cours, un matin. C'était toujours impressionnant de l'avoir parmi nous dans ces moments-là : il était si grand, habillé tout en noir, parmi ces petits enfants qui le regardaient comme un extra-terrestre. Je leur avais demandé de me dessiner la première image qui leur venait quand ils fermaient les yeux. Il m'a rendu ça. Il semblait content.
Je me glace soudain. Elle le remarque.
— Bracken, je ne voulais pas vous vexer. Je ne suis pas professeur de dessin, mais je ne voulais pas qu'il perde l'habitude d'être avec d'autres enfants. J'espère ne pas avoir mal agi.
Non, ce n'est pas ça. Sur le seuil, Bram vient d'apparaître. Il se tient courbé dans l'embrasure de la porte enfoncée, ricanant.
— Tiens tiens tiens... murmure Bram.
On dirait qu'il n'a pas dormi depuis des jours. Une barbe hirsute, des doigts souillés d'encre, des taches blanches sur son pantalon noir froissé, le tout pieds nus. Un naufragé sur la plage de cette chambre.
Plouffe roule jusqu'à lui et le défie en islandais.
— C'est plutôt à vous de nous l'expliquer ! Encore une de vos conspirations à la noix ?
Bram me fixe en fronçant les sourcils.
— Qu'est-ce qu'il fait là ?
— Il nous aide, lui !
Bram s'avance vers moi. Il fait à peu près ma taille. Il porte un casque de Walkman autour du cou. Dans ses yeux, derrière ses lunettes jaunies, je lis une lassitude amusée, comme un ballon crevé joué au pied par des enfants.
— Je pensais que tu ne reviendrais pas.
Il parle un anglais parfait, méprisant. Pour lui, j'ai toujours été un étranger.
— Qu'est-ce que tu veux, Bram ?
— Je dois parler à Elliot.
— Pourquoi ?
Il m'ignore. Il traverse la pièce en se grattant le menton, inspecte chaque meuble. Il s'arrête sur les tortues, dans leur bocal. Ses yeux se remplissent d'inquiétude.
— _Oh, je n'aime pas ce type..._
— _Il me dit quelque chose..._
— _Ce ne serait pas... ?_
— Alors il est parti, conclut Bram.
— Où ? dis-je.
Il sourit d'un air entendu.
— Vous vous prenez bien trop au sérieux ; que vous le vouliez ou non, maintenant, vous êtes tous dans le jeu.
— Un jeu, Bram ?
— Le secret d'Elliot, me répond-il en riant.
Ses dents jaunies luisent, enfoncées dans ses gencives.
— Qu'est-ce que tu sais ?
Un instant, une flamme traverse son regard.
— Vous êtes en retard, mais je serai beau joueur. Il vous reste une chance de l'emporter, chacun à votre manière.
— Ce n'est pas un jeu, Bram.
— Ah non ?
— C'est la vie d'Elliot que nous risquons !
— Tu es plus idiot que je ne le pensais.
— Tu ne me connais pas.
— Tu serais encore prêt à me défier, après la dernière fois ?
Je serre les poings.
— Quelle dernière fois ? demande Fink, piqué par la curiosité.
— Le fils prodigue ne vous a pas raconté pourquoi il avait quitté l'école ?
— Tais-toi, dis-je, les poings serrés.
Bram étire un rictus.
— Il a perdu.
En moi monte une musique, un larsen. Un frisson familier me parcourt l'échine, une vibration, mais Plouffe s'interpose, gesticule, fulmine.
— Vous êtes un malade, Bram ! Regardez ce que vous avez fait de cette école, tout le monde ne pense qu'à jouer, et Elliot était le plus puéril de tous ! Tout cela, c'est de votre faute. Vous l'avez poussé à se dépasser pour ne pas être rattrapé. Vous vouliez en faire votre cobaye et maintenant il est en danger !
Bram explose d'un rire strident, un oracle, une caverne pleine de bulles qui éclatent comme des sentences.
— Elliot est parti parce que son père adoptif est sénile !
— Assez !
Plouffe bondit de son fauteuil pour tenter d'agripper la chemise de Bram qui esquive, laissant choir le pauvre vieux sur le plancher dans un bruit sourd.
— Dieu du ciel ! hurle Cyldrid.
Je me précipite. Sa tête a heurté le sol, laissant une araignée de veinules éclatées s'épanouir sur son front. Il respire difficilement. Nous le couchons. J'exécute quelques mouvements simples pour le maintenir en vie. Je ne m'y prends pas très bien.
— Laissez-le mourir, dit Bram. Il va disparaître avec son siècle.
Plouffe sue à grosses gouttes. Fink l'essuie avec le chiffon déjà imbibé de nos larmes. Dans cette pièce paranoïaque, enfermé, cerné par toute cette vieillesse, je me sens courbé, diminué. Cette faiblesse m'est insupportable. Comment pourrais-je continuer à vivre, en sachant que moi aussi, un jour, je partirai, petit bout par petit bout ? Comme un bobo qui s'infecte et ne guérit plus. Obsédé par le trait, je vois en chaque ride un sillon que la mort trace dans nos rêves pour y semer les graines de l'oubli. En ces yeux vidés, usés par les chirurgies successives, devenus semblables à de petites billes, je lis la terreur d'un ailleurs sans nom, dont nous sommes tous voués à fouler le sol.
KOR
Si nous devions nommer cette poussière
Que le temps n'a commis, non de poids
Mais de pure granularité.
Je vais frapper Bram, juste pour me détendre.
— Oh, le preux Bracken.
Je l'empoigne.
— Vas-y, fais-toi plaisir, use tes petites mains sur mon visage. Dis-moi, Bracken, est-ce que tu penses toujours qu'elles sont ton bien le plus précieux ?
Fink abaisse mon bras, doucement.
— Laissez Bracken, il n'en vaut pas ma peine.
Bram me scrute, insondable derrière ses lunettes. Je m'approche très près de son visage, bien en face, penché, à deux doigts du coup de boule.
— Crache le morceau.
Il tire une cigarette de sa chemise à carreaux, en mordille le bout, le recrache.
— Pourquoi est-ce que je te dirais quoi que ce soit ?
— Parce que tu ne perceras pas le secret d'Elliot tout seul, tu as besoin de moi.
Il éclate de rire et allume sa cigarette.
— Tu prendrais le risque ?
Ses yeux reviennent dans les miens, ne les quittent plus.
— À quoi jouez-vous dans le gymnase ? dis-je.
Il ricane, secoue la tête, l'air navré.
— _Habitat_.
— C'est quoi ?
— Ce que j'ai trouvé de plus utile pour parler à Elliot.
— Je ne comprends pas.
— Je ne m'attends pas à ce que tu comprennes.
— Il va mieux, intervient Cyldrid.
Entre ses bras, contre sa lourde poitrine, Plouffe reprend des couleurs. Il semble toujours sur le point de s'effondrer, château d'allumettes tenu par je ne sais quelle glue.
— Feilu... marmonne Plouffe.
— Feiluleikur ? suggère Fink.
— Cache-cache, dit Cydrid.
Sa couverture glisse, je la ramasse. En la remettant, je vois ses jambes pour la première fois : un brouillon de membres, maigres et tordus à l'extrême, presque enroulés sur eux-mêmes. Mon cœur se serre, le pauvre homme n'a pas dû changer de pantalon depuis la dernière décennie.
Cyldrid émet un gloussement.
— Il se réveille.
Le vieux nous regarde tous les trois, comme des animaux nouveaux, derrière les barreaux de sa pupille embuée. Il semble si fragile, plus fragile encore que Fink. Il tend la main vers mon visage.
— Bracken, mon petit...
Cyldrid soupire, rassurée.
— Papy Plouffe, j'ai bien cru qu'on ne vous avait perdu...
Plouffe sourit faiblement, puis tousse trois fois.
— Creuser... Creuser la tapisserie... Plus loin...
Si ce sont les dernières paroles d'un vieil homme, alors je serai celui qui prend ses responsabilités. Je retourne à la tapisserie, devant le trou de la dernière couche qui laisse entrevoir les tortues sur fond bleu.
— Il y a des tortues, Plouffe. Après, c'est juste de la roche. Il n'y a rien derrière.
Il porte une main brunie à son front, contemple les tortues dans leur bocal, sourit faiblement.
— Oui... les tortues... au parc...
— Le parc ?
— Ils sont allés à Hellisgerdi avec Elliot, me rappelle Fink.
— Vous les avez achetées là-bas ?
Plouffe se redresse sur un coude.
— Non, non.
— C'est quoi, cette histoire de tortues ? questionne Bram.
Plouffe tousse. Ses yeux se perdent dans le vide.
— « Il y a des lieux interdits, des lieux qui ne doivent pas être. Ces lieux sont des rêves impossibles, des rêves de sueur, de terreur muette dans l'espace de ces mondes réduits. Il y a des lieux qui sont les témoins d'une tragédie, d'autres d'une consécration. Certains, ces lieux dont nous parlons, ces lieux qui ne veulent pas dire leur nom, sont des lieux qui ne signifient rien, qui n'ont pas de bâtisseurs, ni d'architecte. Ces lieux sont des cases vides, l'antique preuve d'un danger, d'une attente, enceinte d'une révélation. Leur présence est insoutenable dans notre réalité. Pourtant, en ces lieux, la terre glisse. »
Il tousse.
— C'est une citation. Stein Steinarr.
Bram lève les yeux au ciel.
— Vous auriez pu le citer en islandais, râle Fink.
— Où voulez-vous en venir, Plouffe ? demande Cyldrid.
— Oui, soyez plus précis.
Plouffe pose son regard sur nous, un par un, puis il se met à parler doucement, ses fines lèvres humides comme une vulve ouverte sur un temps inconcevable, sur des horizons inconnus.
— Hier, en fin d'après-midi, nous sommes descendus au jardin dès les premiers rayons du soleil. Nous avons remonté la rue comme nous le faisons toujours, une fois par semaine. Elliot poussait ma chaise et je lui parlais de la vie, de Mallarmé. Ah, ce qu'il aimait Mallarmé ! Nos promenades sont toujours l'occasion de parler, même s'il ne dit rien, mais je sais qu'il écoute, qu'il comprend. Hier, il s'est assis sur le banc près du bassin et il a levé les doigts vers le soleil, pour le pincer. Je n'ai pas compris son geste mais depuis quelques jours il était obsédé par les objets lointains. La lune, les étoiles, le soleil, le sommet des montagnes. Il me les montrait, comme s'il pouvait voir leurs silhouettes et en modifier les contours du bout des doigts. La veille, il était excité et je savais qu'il voulait aller au parc, c'est son endroit préféré, un jardin occupé par les elfes qui n'ont pas le rang requis pour loger au palais d'Hamarinn. Je l'ai toujours accompagné dans ses jeux, je n'ai jamais vraiment compris, je n'y participais pas, mais ça lui permettait de s'ouvrir un peu aux autres, et ça me suffisait. Assis là, alors que le jour se levait, il a commencé à pleurer. Ça lui arrive de temps en temps, je pense qu'il n'a pas grandi depuis qu'il a été abandonné et placé à l'orphelinat. Il doit encore penser à sa maman. Chaque fois qu'il pleure, le seul moyen de le faire cesser, c'est de jouer à cache-cache, _Feiluleikur_. J'ai dit que j'allais compter jusqu'à cent, et il est parti se cacher. Quand j'ai rouvert les yeux, je suis tout de suite parti à sa recherche. Il se cache toujours au même endroit, dans un défilé de lave, derrière une pierre. Il n'y était pas. Mais pour la première fois j'ai remarqué quelque chose qui m'avait échappé jusqu'ici, faute d'avoir jamais vraiment regardé. La lave présentait une longue excroissance rectangulaire, comme une sorte de caisson accolé à la roche, et qui semblait s'y fondre. Les couleurs se complétaient, l'illusion d'optique était parfaite. Il n'y avait rien d'autre, rien que cette aberration n'évoquant aucune période archéologique, aucun repère historique, aucune habitude maçonnique. Même dans un endroit aussi dévasté, cette œuvre anonyme trahissait un plan trop pur, terrifiant de simplicité. Ce n'était pas le produit d'une éruption volcanique. Le vent n'érode pas la roche aussi précisément. Pas un signe, pas une marque, pas une inscription. Aucune trace de doigt sur les parois qui aurait pu attester de la technique employée pour obtenir une telle construction. Les lignes étaient parfaites, si parfaites que le bâtiment tout entier, long d'une cinquantaine de mètres, semblait directement taillé au laser dans un bout de basalte. Pas un bruit, pas un son, pas un animal, pas un souffle de vent. Le ciel lui-même semblait suspendu dans cet instant d'éternité, attentif à ce qui allait se passer. Peut-être est-ce une tombe, ai-je pensé, mais qui aurait-on enterré ici ? Dans le mur un trou béant avait été ménagé, sur les ténèbres du rectangle d'argile.
J'ai pu deviner l'amorce d'un couloir. J'ai roulé à l'intérieur : le plafond semblait haut, plus haut que le volume extérieur semblait le permettre. Le sol était fait d'une boue solidifiée. Le poids des siècles, des millénaires m'oppressait, une force venue de plus loin que tout, une terreur sourde, aveugle, sur laquelle je n'avais aucune prise. Un coude puis, occupant toute la longueur du bâtiment, un corridor. Je n'étais éclairé que par la lumière du jour, dans le dos. Les parois défilaient lentement, leur nudité constituant la plus sournoise des mises en garde : n'allez pas plus loin, il n'y a rien pour vous ici, vous n'êtes pas de taille à comprendre. Tout cela vous dépasse. J'ai roulé les vingt mètres qui me séparaient de l'extrémité du couloir. Tout semblait si net que de cette perfection naissait une terreur primitive que je n'avais jusqu'ici jamais éprouvée. Au bout du tunnel, il n'y avait rien. Juste un autre mur. Il était froid. Le plafond avait perdu sa régularité et s'inclinait vers le fond en biseau. Au pied du mur il y avait semble-t-il un tas de feuilles mortes. J'avais le sentiment que quelque chose allait apparaître au fond du tunnel, quelque chose d'énorme, de silencieux, dépourvu de forme ou de silhouette mais dont la présence me semblait certaine. C'était plus vieux que tout, c'était minéral et végétal à la fois, c'était l'origine de toutes choses ici, qui réveille l'amibe la plus basique, qui effraie la biologie elle-même, qui engendre une peur contre laquelle les cellules sont impuissantes, devant laquelle les atomes eux-mêmes s'enfuient. Je me suis penché pour remuer les feuilles, découvrant un orifice dans le sol. À même la roche, creusé avec les mains. Dans le trou, quelque chose, tassé. Dans les ténèbres, on distinguait un bruit de reptation. Un vent se levait, je sentais le tunnel se refermer sur moi, les ténèbres m'envahir. Le sol se dérobait sous mon fauteuil, l'air fuyait mes poumons. J'ai ressenti la fin de tout. La fin de moi.
— _Il me fait peur, Roméo..._
— _Venez plus près de moi, Diane._
— _Vous êtes tout froid._
— Qu'avez-vous vu dans le trou, Plouffe ? questionne Fink.
Les yeux du directeur s'emplissent soudain d'une brume de détresse.
— Elliot.
Il expire. J'ai l'impression qu'il va mourir.
— Doux Jésus, fait Cyldrid.
— Elliot s'est retourné vers moi, il a tendu ses mains boueuses. Chacune contenait une petite tortue, apparemment tout juste éclose. Un jeune touriste nous a donné un sachet en plastique pour les ramener.
— Les tortues étaient dans le sol ?
— Quelle affaire !
— Elliot voyait des choses invisibles aux autres humain. En partie à cause de son handicap, de son obsession du détail, mais aussi du fait de ses croyances.
— Mais qu'est-ce qu'il cherchait ?
— Comment savoir ? Un autre monde, dont il faudrait trouver les clés ?
La chambre m'apparaît soudain plus menaçante. Je sais qu'il ne s'agit pas d'un rêve, car je ne rêve pas, et pourtant : si je n'ai jamais rêvé, comment savoir si ce que j'expérimente à l'instant n'est pas le produit d'un sommeil profond ? Comment trouver quelque chose dont on ignore l'existence ? Si l'apprentissage suppose l'aide d'un maître, où vais-je trouver la sagesse nécessaire pour sortir de ce dédale onirique où se retire notre âme le sommeil venu ?
Bram s'éclaircit la voix.
— Bracken, vous l'avez dit, c'est du côté de l'illusion qu'il nous faut chercher. Et si nous sommes confrontés à une illusion, alors nos yeux nous jouent des tours. Peut-être que les siens pouvaient percevoir des choses qui échappent aux nôtres.
— Mais comment voir des choses qui ne sont pas là ? dit Fink.
— Dites plutôt : comment ne pas voir ce qui est pourtant là ? Un espace par défaut. Ni positif, ni négatif. Zéro. Une tache aveugle.
— Mais Elliot était très propre ! riposte Cyldrid.
— J'ai été suivi par un orthoptiste pendant toute mon enfance, continue Bram. La tache aveugle correspond à la partie de la rétine où s'insère le nerf optique. L'œil n'est pas une caméra, ce que nous percevons est une interprétation constante du réel. Les informations transmises sont traitées par le cerveau qui crée une perception relative du monde conforme à ses attentes. Le cerveau interprète les signaux et construit une « image » de ce réel dont nous prenons conscience et que par habitude nous interprétons comme réelle, alors qu'il ne s'agit que d'une interprétation de la réalité.
— L'œil est incapable de se voir lui-même, murmure Cyldrid.
Plouffe roule jusqu'au mur, pelé, sondé et carroté jusqu'à sa plus tendre intimité, juste avant le rocher. Il pose sa main sur la tapisserie, cette vieille tapisserie quadrillée de rien, avec ses tortues dessinées qui nagent dans un monde de silence.
— C'est possible. Peut-être Elliot est-il sous nos yeux depuis le début. Il suffit d'apprendre à voir comme lui. Ce flash que vous avez subi, Bracken... Peut-être Elliot est-il captif et essaie-t-il de communiquer par des signaux lumineux, s'il nous voit et que nous ne le voyons pas.
Bram l'interrompt.
— De son point de vue, nous n'appartenons plus à son champ de vision. Ses signaux éventuels s'adressent à d'autres que nous. Ils appartiennent au monde où il se trouve désormais, retenus par leur gravité.
Je ne comprends pas ce qu'il tente de nous expliquer.
— Si Elliot ne voit pas les choses comme nous, dit Bram exaspéré, il se peut qu'il ait trouvé une cachette qui nous soit à tous invisible.
J'ai l'impression que tout se brouille. Comme un appel d'air, qui m'empêche de voir clair. J'ai cette étrange sensation de sortir de mon corps, de flotter au-dessus, de nous voir là, dans cette chambre aux angles tordus, moi, et tous ces Islandais bizarres aux raisonnements obliques, affairés autour de cette disparition, chacun posant ses pièces sur un plateau de jeu qui me dépasse. Je ne suis pas dans le jeu, je ne suis pas comme eux. De mon nouveau point de vue, plus haut, en plongée, je comprends qu'ils mélangent une pensée déductive à la pensée du rêve, une pensée qui m'est étrangère, car je ne rêve pas, mais dont je peux saisir les contours. Ils ne sont pourtant pas fous : c'est juste qu'ils fonctionnent ainsi. Leur pensée fait une place à l'invisible autant qu'au visible. Est-ce que cette prise de perspective est ma façon de sombrer avec eux, de basculer dans leur entendement double ? Si je me laisse aller à épouser leur pensée, à tolérer un tel paradoxe, à permettre à mon esprit de fonctionner en maintenant deux visions du monde concurrentes, soudain, je ne vois plus rien. Un terrible brouillard monte, noyant les formes, ne conservant que de vagues silhouettes, fantômes flottant dans le mystère opaque de ces lieux maudits.
— La double vue, dit Fink. Seule la double vue permet de voir les fées.
— Est-ce inné ? questionne Bram. Ou bien y a-t-il un moyen de l'acquérir ?
— Kirk a affirmé que tout le monde pouvait y parvenir.
— Qu'est-ce que Star Trek vient faire là-dedans ? dis-je, atterrissant brusquement, tandis que la pièce reprend sa texture solide, ses racines.
Ravi d'étaler sa science, Fink nous explique que Robert Kirk était un révérend qui avait été enlevé par les fées. Il avait écrit un ouvrage, le _Commonwealth Secret_ , où il détaillait précisément leur mode de vie, somme toute très similaire au nôtre. Il y abordait surtout la notion de double vue, ce pouvoir permettant de distinguer ce monde parallèle. Selon lui, la double vue était un don de naissance, mais pouvait s'acquérir par un entraînement de la rétine, joint à une foi indéfectible en l'existence d'une société parallèle. Le peuple invisible était présent, à tout moment, en tous lieux de cette île, dans chaque caillou, dans chaque interstice.
— Et vous pensez qu'Elliot a ce don ?
— Je l'ai souvent vu porter ses lunettes, je croyais qu'il était myope, mais non.
— Astigmate, je crois, c'est ce qu'a dit le docteur.
— Hum. Quelqu'un d'autre ici est astigmate ? demande Bram.
— Moi, réponds-je.
Tout le monde me regarde. Ils sont vieux, mais aucun ne porte de lunettes. Oui, je suis astigmate. Et non, ce n'est sûrement pas une coïncidence. Elliot et moi partageons ce défaut. Il a laissé ses lunettes derrière lui à mon intention. Il veut que je voie ce qu'il a vu. Je vais chercher ses lunettes. Tendre ces verres brisés vers la lumière des guirlandes. Ils ne sont pas cassés, comme je l'avais cru tout d'abord.
— C'est un kaléidoscope.
— Elliot devait renforcer ses altérations oculaires.
— Et nous n'avons rien vu...
Je retourne la monture entre mes doigts. Elle semble fabriquée dans un ivoire brun vert que je ne parviens pas à identifier. J'imagine Elliot, ici, tout seul, désespéré de ne pas trouver une fée pour le guider, à se meurtrir les yeux en tentant de voir quelque chose que lui dicte son imagination. Un ultime effort, dans sa chair, pour franchir le voile invisible qui le sépare d'un monde impossible.
Très bien. Chaussons ces lunettes.
— _Donc, si j'ai bien compris, ce mystérieux Elliot est dans le mur ?_
— _J'ai l'impression que c'est plus compliqué._
— _Ah. Entre le mur ?_
Ils me suivent à la queue leu leu, tels des randonneurs amateurs en cordée avec ma seule parole pour guide dans les ténèbres du mystère. La chambre d'Elliot compte quatre coins. Les deux premiers encadrent la porte ; les autres sont respectivement derrière le lit et derrière le secrétaire. Des coins libres de tout objet, de simples coins. Une arête, une jonction sur le plancher, deux nouvelles lignes détourant le reste de la pièce, doublées de parallèles au plafond.
Ce n'est qu'au quatrième coin, près du lit, que je m'immobilise. Celui où mon premier instinct m'avait guidé par deux fois, comme si j'avais moi aussi senti quelque chose, dissimulé dans cette absence. Le souvenir du flash.
— Pourquoi cette pièce paraît-elle plus grande qu'elle n'est ? demande Plouffe.
J'ai, moi aussi, eu cette impression – d'avoir passé des jours à arpenter cet espace – un royaume, un monde en soi, entre les quatre murs de ce caveau. J'ai l'envie soudaine de sortir respirer.
— Qu'est-ce qui se passe, Bracken ?
J'approche de l'angle, lève la main, n'ose pas tâter. Je la laisse en suspens, comme pour absorber une présence ou deviner une veine d'eau sans l'aide d'une baguette de sourcier. Trois lignes convergeant vers un point unique. Mettre le bout d'un doigt dans le coin, comme dans une prise de courant.
— J'ai cru voir...
— Quoi ?
—... un mouvement.
Soudain apparaissent des traits sur le mur, autour du coin, jusqu'au plafond. Je ne sais s'il s'agit de glyphes ou de simples indications topographiques. Peut-être les deux. Comme des vagues. Des motifs très serrés, pourtant séparés par du blanc. Je peux les faire bouger par inattention. La tapisserie se confond avec le reste de la pièce depuis le début notre enquête, sans autre valeur qu'un canevas sur lequel dérouler nos questions. Tout le long du coin, sur une largeur d'environ quinze centimètres, courent désormais des entrelacs de traits formant des formes subtiles qui nous ont échappé jusqu'ici, tant il faut en être proche pour les remarquer.
— Des signes, dis-je.
— Un effet d'optique ?
Ces arabesques forment des spirales, telles des tresses de fleurettes sauvages, toutes différentes, sans cesse renouvelées, aux formes si délicates, si finement qu'il est impossible de les délier sans manquer de tout faire basculer. Plusieurs strates de dessins sont superposées les unes aux autres. Il a fallu des années pour faire tout cela... Je me demande pourquoi seul ce coin est couvert. Comme s'il y avait ici quelque chose de spécial. Je pense à ces détenus qui pour s'évader d'une cellule passent leur vie à creuser la roche avec une simple cuillère. Quelle détermination faut-il pour ainsi passer son existence à creuser dans l'espoir ténu de revoir un jour le ciel, d'entendre à nouveau chanter les oiseaux. Renaître au monde, puis le quitter, le corps épuisé d'avoir tant voulu vivre.
— On dirait les traces qu'on trouve sur les pierres, dans la lave refroidie...
J'approche mes doigts de moufle.
—... ou les fientes d'oiseaux sur les ruines de Londrangar...
— Bracken, attention ! s'écrie Plouffe.
De l'électricité statique fait frémir l'extrémité de mes phalanges sous les moufles. Un gratouillis, des chatouilles. Une odeur monte à mes narines. Non pas d'ozone, mais plutôt vaguement cyanurée, de pâte à modeler, piquante et fade à la fois. Je frôle le papier. Mon doigt s'arrête sur un filament et glisse : l'encre vient avec elle, en une longue trace. Entre les mailles de la laine, se fait sentir le froid.
— Vous suez ? demande Plouffe.
— La tapisserie est humide.
Je porte le doigt à mes lèvres : une folie, mais je reste un primitif, et pour comprendre, j'ai besoin de goûter.
De l'eau de mer, teintée d'une légère amertume en plus.
— Bracken, regardez !
Mon doigt a activé quelque courant sur cette surface, une douce luminescence irrigue les veines du dessin. Dans un flash, celui-ci se superpose à la pièce, presque en trois dimensions, visible de tous, telle une signature, un souvenir, la trace d'un passage. Une migraine ophtalmique enfle soudain, explose sous mon crâne, une vrille si violente que je ne peux garder les yeux ouverts. J'arrache les lunettes.
— Ouch !
Les dessins luisent faiblement sur le mur puis s'éteignent. Comme l'image résiduelle d'un écran de télévision, ou un effet de persistance rétinienne. Une électricité impalpable reste en suspension au milieu de la pièce bleu marine, comme un souvenir d'éternité. Tout le monde l'a vu.
— Une vision ! s'exclame Cyldrid.
— Le Palais... balbutie Fink.
— Quel palais ?
Fink s'énerve.
— Le palais de la reine d'Hamarinn. Le hall d'entrée, les globes... Ça correspond à toutes nos gravures...
— Non, dit Plouffe, c'est une langue.
— Comment ça ?
— Toute création d'espace est engendrée par un mot. Le mot est la chose. Si Elliot s'est réfugié dans un monde, ce monde a un nom.
— Ce ne sont pas des mots, Plouffe, ce sont des lignes.
— Vous pensez qu'Elliot est dans la tapisserie ? dis-je. Physiquement ?
— Il est peut-être camouflé dans les vagues, dans les dessins, mais personnellement je ne le crois pas.
Il regarde disparaître les dernières lignes, qui achèvent de s'effacer.
— Tous les motifs étaient différents mais tracés de façon identique. Il y avait des traces de rupture, quand le flux s'est interrompu pour laisser place à une autre phrase. S'il a enluminé ces murs, c'est pour signifier quelque chose qui soit compréhensible y compris pour des yeux profanes, comme les nôtres.
— Il s'agirait donc d'un panneau de signalisation ?
— En quelque sorte.
— Mais qu'est-ce qui vous fait dire que ce n'est pas le Palais ? boude Fink.
— Il y a forcément des correspondances entre ces mots et notre langue. Les mots DISENT le monde, créent la pensée, qui en retour sculpte l'univers. C'est pour cela que nos illusions ont une existence propre, toutes ces choses que nous créons auxquelles nous accordons une valeur culturelle, sociale ou politique. Ici, le mot est un dessin, car c'est ainsi que vivait Elliot. Il suffit d'interpréter le texte.
— Concrètement, intervient Bram, vous affirmez qu'Elliot aurait créé une langue à partir de lui-même, qu'il aurait tracé ces signes sur les murs pour donner une existence à son univers mental, puis qu'il s'y serait réfugié, tout en nous laissant des indices ?
— Ça me semble plausible.
Bram fait un drôle de bruit avec sa langue. Une lueur est apparue dans ses yeux, une flamme blanche qui vacille, celle de la folie, de la peur, ou des deux mêlées dans une danse de mort, comme un refuge où se recroqueviller quand le monde se décompose, quand plus rien ne fait sens. Dieu, peut-être. Et encore : Dieu lui-même doit se poser des questions sur sa solitude.
— Et maintenant ? demande Cyldrid.
Plouffe esquisse un geste de déni.
— Il nous sera très difficile de décoder cette fresque. Et je ne vois pas ce qui pourrait nous tenir lieu de pierre de Rosette. Il s'agit d'un langage primitif. Comme celui des nouveau-nés. Pour nos anciens, le monde autour d'eux était le langage. Toute chose avait un nom, et le nom devenait verbe. Rien n'était inventé, tout était lié aux référents, le signifiant et le signifié ne faisaient qu'un. Ce n'est qu'avec le temps que nous avons appris à donner des sens différents aux mots, puis les cultures ont achevé de désintégrer les relations entre le monde et le mot. Puis le mot devient un verbe, comme le font les enfants et les primitifs. Il rend le monde actif.
— Le passé est un animal grotesque, dit Bram, dédaigneux.
— Vous auriez tort de croire qu'ils étaient stupides. Ils avaient un accès direct à l'essence de ce monde. Tout comme Elliot, qui en vieillissant s'était rapproché de l'origine. C'est sa langue de Babel à lui. L'origine, ne serait-ce pas la langue en tant qu'organe capable de parler toutes les langues sonores ? L'enfant qui apprend les premiers phonèmes est capable de tous les prononcer, ce n'est qu'ensuite qu'il perd cette capacité, qu'il se spécialise et devient le locuteur d'une ou de quelques langues en particulier. Le moment où tout en lui parle la même langue... Mais le monde était plus pur, on ne le confondait pas avec le ciel. Chaque chose était à sa place, et chaque place à sa chose. Leur référent était un Eden. Nous ne connaissons que confusion. Elliot voulait recréer cette enfance du monde.
Je repense à ce qu'Elliot avait essayé de me dire. À sa façon de parler, incompréhensible. Il s'était déjà, _de facto_ , réfugié dans son langage, dans son monde. Arrivé au terme de son existence, avait-il trouvé un moyen pour y accéder physiquement ?
KOR
Plat, le plan de la surface
Reflet monotone d'un monde incrusté
– Dont rien ne cerne plus les bords –
Debout, face au mur, à chercher l'impression du dessin dans le réel, sans lunettes, sans artifice. Mais non, rien.
Dans un éclair de lucidité, je dis :
— De quelle sorte d'encre pourrait-il s'agir ?
Plouffe mordille sa lèvre inférieure.
— Une encre qu'on ne peut discerner que d'un certain angle ?
— Ça existe ?
— Une encre que notre œil connaît intuitivement mais que nous n'avons jamais eu l'occasion de percevoir. Une encre qui fait partie intégrante du monde, comme autant de couleurs qu'on ne devine pas, comme des sons audibles des seuls animaux, des sens réservés aux seuls enfants. Ce qui semble logique, si tout ici est l'enfance d'Elliot.
Je rechausse les lunettes, dans l'espoir de trouver quelque chose de nouveau. Toutes les arabesques réapparaissent. Perdu dans les entrelacs du dessin, la migraine bat au rythme de mon cœur, je sais que je pourrais y passer ma vie. Comme il est difficile de restituer l'essence d'une vitesse avec des mots, comme la grâce – beauté en mouvement – est inaccessible au vocabulaire limité de notre pensée formatée. Même ces images qui interviennent dans le silence de l'esprit, ces illusions criardes qui chassent le rationnel, ne sont que des articulations de concepts, des mots transformés en borborygmes, puis en signes, puis en images. Il n'y a rien d'autre que le verbe.
Ou alors. Peut-être Plouffe se trompe-t-il ?
Ces dessins n'ont rien de vocables, ils semblent exprimer quelque chose de plus ancien que les mots, une émotion en tant que mythe. Comme ces portes sans couloir que j'avais vues, à la section égyptienne d'un musée, plus jeune, quand j'étais encore en France : aménagée dans les recoins les plus secrets d'une tombe, une porte vers l'au-delà comme but ultime, un rectangle taillé dans la pierre, débouchant sur rien. J'imagine que pour le mort cette sépulture pouvait l'avaler tout entier dans un rayon solaire puis se refermer sur les mortels laissés derrière lui. Ce qui m'avait terrifié dans cette surface, c'était sa simplicité. La pure terreur du vide. Il me manque le bagage scientifique pour assimiler clairement les motifs qui relieraient l'ensemble ; je me contente de ressentir une peur, un affolement général devant l'absence de sens, dans le dénuement de la mort, quand le plat se mue en relief.
Sous mes yeux filtrés par les lunettes, le dessin semble luire à nouveau, puis s'éteindre. Je prends conscience de mon souffle : là, je respire. Ça me calme un peu de savoir que là, maintenant précisément, je respire. Je sais que je suis là, que je tiens quelque chose que je n'avais jamais vu auparavant, une terreur et une excitation simultanées, le frisson d'un ailleurs véritable étranger à toute image, à toute pensée, à tout langage. Étranger car primordial, avant tout le reste, juste avant que le choix n'apparaisse, avant que la différentiation du jugement s'accomplisse. L'espace entre pensée et non-pensée, juste là.
Désormais, rendu à la perception de mon environnement immédiat, à ses détails, j'essaye de suivre une des arabesques d'Elliot jusqu'à son terme, pour finalement revenir à mon point de départ, égaré dans une forêt de fleurs faite de fibres. Il n'y a pas de chemin possible. Ce cadre détoure du rien, il s'agit peut-être d'une porte, mais où se trouve le levier ? Est-ce une Moria, ouverte d'une incantation ? Comment prononcer l'imprononçable ? Sur mes lèvres se forme un hymne, quelque chose qui évoque, par la majesté de ses intonations, les nuées d'une boucle de présent, le passé enfui, pour replonger dans le virage d'un ensemble esquissé : le dessin, serpent dans le serpent, accrochés aux branches du mur, les pleins et les vides comme des grottes sans fin où, lovés dans l'absence, les dragons rêvent. Quelle puissance dans ces danses, quelle résorption du moi...
— Bracken, attention, pas trop longtemps.
Cyldrid m'a réveillé, d'un frôlement de coude, d'une caresse, écho d'une maman hâtivement tirée du lit un matin de cauchemar. Je cligne des yeux.
— Vous étiez perdu, mon petit.
Je tourne ma langue sept fois dans ma bouche.
— Cyldrid...
— Oui ?
— J'ai besoin de ses dessins.
— Les dessins d'Elliot ?
— Tous ceux qu'il a réalisés pendant mon absence. Pendant vos cours.
— Mais...
— Vite, Cyldrid.
Elle jette un coup d'œil à Plouffe, à Fink, puis elle sort d'un pas rapide. Elle n'a pas l'air contente. Mais j'ai besoin du contexte. J'ai besoin de plus d'informations. Je lève la main vers le mur, un geste qu'il me semble avoir exécuté si souvent qu'il en devient essentiel au fonctionnement de l'âme : d'une main agile, d'un ongle, je fais tomber un fragment d'encre solide, non diluée, dans ma paume. La trace qu'il laisse est un décalcomanie de rien, un vide plein de son contour. Je retire mes lunettes, mes tempes pulsent. Il n'y a rien sur le papier que je puisse voir sans elles. De l'ongle, je retourne l'invisible. J'ouvre la bouche, tire la langue, y dépose le précieux morceau de rien.
— Oh lala ! gémit Plouffe.
— Bracken, ne faites pas ça !
— Mmh... On dirait, du... une algue.
— Une algue ? Ici ?
— À moins qu'il ne s'agisse d'encre de seiche.
— Où Elliot se serait-il procuré une encre pareille ?
Je pense brusquement aux tortues dans leur bocal. Toute cette histoire sent la tortue, l'iode mariné dans une soupe d'écailles dans une chambre turquoise. Une image, légère, passe à travers moi : une lagune peu profonde, d'énormes nuages d'eau salée qui enflent sous un ciel marin, mes pieds sur le sable humide, entre les flaques, un filet, des anémones de mer qui s'ouvrent en palpitant. Je me penche et mon visage dans l'eau trouble me renvoie l'image d'un ange.
— _Vous sentez, Diane ?_
— _Oui, comme un frémissement..._
— _Le vent se lève._
Soudain, j'ai mal partout. Mon cou, mon dos. J'ai l'impression d'être une conscience enfermée dans une armure d'os et de chair. Mes yeux voient à travers ses jointures, où se glissent parfois des fluides, des larmes. Ma tête pulse. Je m'assois par terre, pris dans une convulsion de doutes et de douleur. Est-ce cela, vieillir ? Est-ce que rester au contact de ces vieillards, dans un espace clos, ne me transforme pas malgré moi en croûte ? Une terreur me saisit : quand tout est derrière, et devant, juste la promesse d'un choc sur le sol qui brisera cette coquille. Glisser le long d'une pente, sans espoir de remonter. Je ne sais pas si mon âme survivra à la perte du corps, je ne me suis pas préparé. Jusqu'à présent, je ne m'étais jamais posé la question de la mort, pas aussi clairement en tout cas. Pour la première fois, je peux la sentir, elle a l'odeur d'une chambre refermée sur son bleu, sans espoir, peuplée de fantômes affairés sur leur douleur, sur leurs questions. Aucune réponse à espérer en ces lieux, aucune respiration. Plus que le doute, qui devient quête sans retour. Vieillir est une malédiction dont nous sommes tous les victimes. On n'échappe pas à la dégradation. On n'échappe pas au corps.
— _Je crois qu'ils sont en danger..._
— _Mais que pouvons-nous faire ?_
— _On est des tortues ! On peut tout faire !_
Il me faut quelques secondes pour retrouver mon équilibre, un horizon. J'ai l'impression d'avoir tourné sur moi-même pendant des heures. Je suis pris de l'envie de retourner voir le dessin, d'essayer d'en savoir plus, mais la peur du mal est la plus forte. Quelque chose coule en moi, comme un poids mort, comme les pieds d'un criminel dans le béton au fond d'un lac, un renoncement, que dans ma jeunesse j'avais perdu de vue. C'est ici que je suis, maintenant, et je n'irai nulle part ailleurs. Je regarde mes mains, mes moufles et les tortues. Tout est ici, avec moi. C'est avec ça que je dois faire quelque chose. Ma vie dehors était infinie. Trop de possibilités. Je n'ai jamais pu choisir. Mais ici, rien ne relève d'un choix. Tout est. Il n'y a pas d'échappatoire à cette chambre. J'aurais pu courir dans le champ de lave, m'enfuir très loin de cette histoire qui est train de me rendre fou, que ne semble avoir ni début ni fin. Je ne veux pas qu'Elliot meure, je veux savoir ce qu'il voulait me dire, je veux prouver à Bram que je ne suis pas un bon à rien, inutile et paresseux. Moi aussi je peux gagner, moi aussi je peux créer et jouer sans compétition. Je ne me suis jamais laissé le temps d'être moi-même. Après quoi est-ce que je cours ? Pourquoi est-ce que je me pose autant de questions ?
Je n'aurai peut-être pas dû avaler cette algue...
— _Il faut vous y résoudre, Roméo, on ne sert à rien._
— _Si seulement nous avions la mémoire._
— _De quoi ?_
Je veux ma couette. Là, maintenant, tout de suite. Je n'en peux plus, je me sens si vieux. J'étais tellement bien sous ma couette. Je mangeais du chocolat, parfois j'écoutais la radio toute la journée, un flux continu de chansons et d'infos en islandais. Parfois j'imitais leurs formules, en roulant les « r ». Je ne comprenais pas le dixième de la culture dans laquelle je baignais. Oublieux de mon passé hexagonal, incapable de m'intégrer à mon nouvel environnement. Sans mon artisanat, sans ma petite discipline de chaque jour. Sans foi. Voilà. Finalement, c'était mieux ainsi. Finalement, il y avait quelque chose de vrai là-dedans : quand je suis sous la couette, je suis un humain misérable, qui ne sert à rien, qui ne vaut guère plus. Ces deux tortues valent plus que moi. Je sais qu'il est stupide de faire des généralités sur le genre humain, mais sous une couette, le monde apparaît soudain plus simple. Comme une clarté, le matin, quand quelque chose de moelleux vous attrape et ne vous lâche plus. Lorsque la lumière change doucement, que chaque heure est une couleur, et que tout glisse comme de la pâte, dans les ronflements du lit. Je ne veux plus de trous, de chambres closes, de tapisseries et d'avant l'avant. Je ne veux pas savoir ce qu'il y a entre l'avant-dernier et le dernier. Je veux être avant, avant l'avant-dernier, dans ce moment gros de choix où rien ne se passe. Une éternité, et moi, sous la couette, vivant pour toujours, respirant rapidement. Je donnerai tout pour sortir de ce rêve et retrouver ma couette, mon insipide réalité. Peut-être que si je fermais les yeux, je m'éveillerais. Je ne risque pas grand-chose à essayer.
— _Bon, j'abandonne._
— _Moi aussi !_
— _Foutons le camp._
Quand je rouvre les yeux, Fink est en train de se battre avec Bram.
— Voleur ! Voleur ! hurle le vieux.
— Qu'est-ce qu'il a volé ? dis-je en m'interposant.
— La cassette d'Elliot !
Bram repousse le vieillard. Dans sa main, la cassette audio que j'avais trouvée dans un des tiroirs du secrétaire. Pour la première fois, je fais le lien avec son casque de Walkman.
— Alors c'est pour ça que tu es venu...
Bram soupire.
— Il ne faut pas m'en vouloir d'avoir essayé. J'ai vu la cassette pendant que vous étiez tous là à tenter de sauver le vieux. J'ignorais son existence.
Il me rend la cassette. Plouffe semble soudain très intéressé.
— Quand a-t-il enregistré ça ? Il n'y a pas d'étiquette.
— Je ne sais pas, dit Fink, j'ignorais qu'il possédait un magnétophone.
Bram a placé la cassette dans le Walkman, une grosse chose carrée rose fluo qui détonne sur le bleu de la chambre d'Elliot. Je prends le casque, personne ne s'oppose à ce que j'écoute en premier, pas même Bram, qui scrute mon visage en quête d'une information dont j'ignore la nature. Je ne sais pas à quoi je dois m'attendre. Je presse _Play_. Le set commence par une chanson à la guitare. Desséchée. Celle que j'avais entendue dans le bureau de Fink, en arrivant quelques heures plus tôt.
— Robert Johnson ? demande Fink, ahuri, collé à mon oreille.
— Hum.
La chanson meurt doucement, puis le silence se fait.
Des bruits d'ambiance. Des oiseaux, des voitures, un avion qui traverse dans le ciel. Des pas, l'enregistrement est haché. Puis s'arrête. Il reprend quelques secondes plus tard.
Une respiration.
Puis :
« gagaghhhaammm hhaaauummm kjlloodiii hhhprimmm, ttuuut ptyx ptyx da la purrr, tttre ptyx da la purrr et de na po ja kolo popa popo kilikili pilitopli, malako malaki nepadelria jeerrodhh jjee ptyx dukjillpodddh jkrrroo sst dhun gagahimm humhumi toalal ahddppz jdkko loppi lpoppi ptyx nabala nabala nabalu. »
La voix d'Elliot. Personne ici à part moi ne l'avait entendue.
L'enregistrement s'arrête très vite, Bram fait avancer la bande, rien. Bram tourne la cassette. Sur la face B, plusieurs enregistrements de _The Riddle_ , le tube de Nick Kershaw, tirés de ses nombreux passages à la radio.
— Heu...
— Eh bien, comment dire...
— Ce mot, qui revient... dit Plouffe en comptant sur ses doigts.
— Ptyx ?
— _Ah ah ! Ptyx !_
— _Enfin, un indice vital !_
— _Qu'est-ce que c'est ?_
— Qu'est-ce qu'un ptyx ? demande Fink.
Plouffe fixe un point devant lui, pris dans ses pensées.
— En grec, ça signifie pli, dit-il. Mallarmé en a parlé.
Je n'ai jamais lu Mallarmé. Ou alors je ne m'en souviens plus. Mais je me souviens de ce qu'avaient dit les jeunes de Bram, devant leurs écrans, puis dans le couloir. Mnyx. Trix. Ptyx.
— Ptyx... répété-je.
— Un pli. Un repli. Un endroit où se cacher.
Rompant le silence, nos respirations. Je peux sentir Plouffe se tendre, impatient, comme incapable d'admettre que non, Elliot n'est plus ici, qu'il a disparu dans cette chambre close, et que seul un sortilège peut expliquer sa disparition. Mais imaginer une intervention miraculeuse, c'est plus difficile encore que de la refuser.
Plouffe déclame soudain :
Sur les crédences, au salon vide : nul ptyx,
Aboli bibelot d'inanité sonore,
(Car le Maître est allé puiser des pleurs au Styx,
Avec ce seul objet dont le Néant s'honore.)
— Qu'est-ce que vous baragouinez ? bougonne Fink.
— C'est un sonnet de Mallarmé qui explique comment un poète descend en enfer pour y chercher un objet magique.
— Quel objet magique ?
— Dans son enfance peut-être, une sorte de talisman.
— Bram, je vous ai entendu, dans le gymnase. C'est ce que vous cherchez ? Vous cherchez le Ptyx ?
Il hausse les épaules.
— Tout jeu requiert un objectif autour duquel se structurer. Toute quête est magnifiée par son propre objet. Nous avons conservé un principe d'abstraction car le digital, ce ne sont que des pixels. Il nous fallait un objectif, nous avons pris celui d'Elliot, même si nous en ignorons tout.
Il éclate de rire. Personne n'a compris.
— Le Ptyx est un bibelot aboli, dit Plouffe. Caché dans un pli du même nom. Peut-être un Graal, un Sampo, une corne d'abondance, la vie éternelle, la jeunesse.
Il s'interrompt, sourit faiblement. Il me parle soudain en français, et je me rends compte que depuis tout ce temps, en public, il me parlait tantôt en anglais, tantôt islandais sans que je fasse la différence.
— Elliot cherche la grâce.
— De quelle façon ?
— Il est parti de lui-même vers la mort. De lui-même. Comme Orphée descendant aux enfers, bien vivant, pour aller y sauver Thésée et Eurydice. Pour lui, c'est simplement un moyen d'en finir avec l'existence.
Peut-être que le coup qu'il a reçu lui a définitivement fait perdre la raison, que nous sommes ici dans l'école à sa merci. Bram croise les bras, très calme.
— Le plus grandiose des jeux : comment passer de l'autre côté tout en demeurant vivant.
Plouffe claque des doigts.
— Si Elliot est parti dans un pli, alors c'est dans le Néant qu'il faut le chercher ! Et le Néant était là avant lui, le Néant est entre toute chose, donc entre la vie et la mort. Elliot, en tant qu'idée, était déjà présent dans cette pièce avant même d'être né. C'est dans cette absence qu'il faut le chercher, dans cette préconception de lui-même où il s'est réfugié. Elliot était un être à part, extrêmement secret, il se passionnait pour les poèmes que je lui lisais, il a baigné dans ce Néant. Oui, je pense qu'il avait ce pouvoir, il est allé plus loin que je n'ai jamais été, il a trouvé le Néant réel, il a compris comment passer, peut-être grâce à ces tortues.
— _Roméo, Vous avez entendu ?_
— _J'essaye de faire un trou dans le bocal !_
— _L'ancêtre a parlé du Néant !_
— Le Néant est un lieu ?
— Pourquoi pas ? dit Plouffe.
Fink échange son anxiété avec la mienne. Je le sens se refermer. Il accuse son grand âge, l'heure matinale. Pour un vieux, l'absurde est quotidien. Certains ont tendance à vouloir oublier de quoi le réel est fait pour se construire un nid douillet dans le déni.
— Ne prenez pas ça au pied de la lettre, Bracken, soupire Fink. Ce n'est qu'une expression. C'est comme de dire, je ne sais pas, « mourir ». Entrer dans la beauté de Dieu. Dans le Lieu, vous dirait cette pimbêche de Cyldrid. Moi je vais vous dire ce que ça veut dire, aller au Néant : aller en Féerie, que nous avons tous vu sur les murs de cette chambre. Nous savons tous pertinemment qu'Elliot a été enlevé ! Il est là-bas, là où se trouve le palais, la garde royale et les chats mauves qui ne voient que d'un œil !
Plouffe soupire.
— Bon sang Fink, on est en 1986, l'Islande est à l'Eurovision ! Grandissez un peu ! Vous croyez que la nature est vivante, qu'elle a une âme, ou, comme Cyldrid, qu'un créateur ordonne toute chose. Ce sont des conceptions délirantes, motivées par la paresse. La dualité caractérise notre civilisation, mais l'invisible n'a pas à s'y soumettre. Je vous parle de quelque chose de beaucoup plus subtil, qui relève presque de la science religieuse. Nous le savons tous, dans notre pays, la vérité se situe quelque part entre le ciel et le rocher. C'est physique.
Moi, je n'entends rien à la philosophie. Mais je sais que le Néant, c'est le Néant. Alors je me trompe probablement, mais je crois que le Néant c'est juste rien.
— On ne peut pas aller dans le Néant, dis-je. C'est impossible. Si vous me dites que c'est ni l'enfer, ni Féerie, qu'est-ce que c'est ?
— Rien.
— Donc ?
— S'il n'y a rien, il y a quelque chose. C'est un mot, c'est déjà beaucoup.
Il vient de me casser la tête. Il joue sur les mots. Je me souviens d'un épisode de _la Quatrième Dimension_ , dans lequel une petite fille disparaissait dans une dimension parallèle sous son lit. Ses parents et un ami physicien la cherchaient partout.
Moi, je sais que le Néant, c'est du vent.
Je l'ai appris. Je ne sais plus où.
— Vous me rendez fou avec vos histoires, Plouffe. Elliot est dans un trou, caché dans la roche. Il n'a pas pu se rendre dans... rien !
— Il y a des _koan_ qui prétendent le contraire.
— Des koalas ?
— Les _koan_ sont des énigmes zen, Bracken. Connaissez-vous celle du mât de cocagne ? Trois moines zen grimpent un mât de cocagne bien graissé, le gagnant sera celui qui réussira à atteindre le sommet avant les autres pour en décrocher la clochette. Tout le monde glisse, mais un des moines réussit à toucher le sommet, sans pour autant obtenir le _satori_. Pour cela, il lui faudra continuer à grimper.
— Mais s'il n'y a plus de mât ?
— Il n'y a jamais eu de mât. Il n'y a ni début, ni fin, car ce sont des repères temporels. Nous n'atteindrons jamais l'éternité d'Elliot en avançant dans le temps. Cela signifie qu'on n'atteint pas l'éternité par une prolongation indéfinie du temps, pas plus qu'on n'obtient une image en deux dimensions par le prolongement d'une ligne unidimensionnelle. Les deux, éternité et temps, sont par définition discontinus dans le sens mathématique du terme, composés d'éléments séparés.
— Je ne comprends rien, de quoi est-ce que vous parlez ?
— Bracken, le Néant est une obsession chez tous les poètes, continue Plouffe. Et tout spécialement en Islande.
— Ah, c'est donc ça ! Des notions de poésie appliquées à la physique ? Et vous croyez que vous valez mieux que l'Église ou que les elfes qui font de la soupe ?
— Vous n'êtes pas d'ici Bracken, alors vous ne pouvez pas comprendre notre façon de parler. Je vous parle en français, en anglais, les autres vous parlent en anglais, parfois nous parlons tous islandais. Votre cerveau est une bouillie d'informations, vous ne pouvez pas assimiler. Notre langue est si vieille, elle est encore empreinte de tant d'images que vous ne connaissez pas, des mythes subtils qui ont forgé notre pensée, sans jamais la changer réellement. Nous avons ceci en nous depuis toujours, et le monde autour de nous est devenu ce que nous pensions. La question de savoir s'il l'a toujours été ou si nous l'avons rendu, perçu ainsi, est la question métaphysique essentielle de notre foi.
Il roule en cercles dans sa chaise, comme pour tracer sa pensée dans l'espace, sans jamais s'interrompre.
— En Islande, le Néant ne se dit pas. Très peu d'artistes ont réussi à articuler cette notion, un des rares à avoir accédé à la postérité fut le poète Stein Steinarr, dont je vous ai parlé, un homme qui avait voué sa vie à l'abîme. Le Néant est là pour nous, mais il ne doit jamais être dit, au risque d'y perdre son sens.
En me rendant étranger à leur pensée, à leur culture, ces gens avaient fait de moi un être indépendant, articulé dans sa droiture. J'ai l'habitude. Dans mon pays natal, j'ai toujours été un étranger.
— Vous parliez de foi, dit Bram.
— La foi n'interdit pas le Néant, dit Plouffe, je dirai même plus : la foi existe à cause du Néant. Car en vérité, si je ne le vois pas, il permet tout le reste.
— Mais si le Néant n'existe pas...
— Le Néant n'est pas le vide, Bracken. Si je bouche un verre avec ma main, je crée un verre de vide, mais le Néant où est-il ?
Je ne veux pas dire une bêtise.
— Dans l'absence de verre ?
— Pourtant le verre est bien là, non ? Non, le Néant est une matière volatile, qui échappe à tout contrôle car elle n'est pas localisée. Il ne faut pas la chercher en trois, mais en deux dimensions. Voir en une seule, pour pouvoir approcher le plus possible du zéro. Je vous dirai que le Néant, c'est la surface de contact entre le verre et le reste du monde, l'endroit où se forme l'interdépendance de toute chose.
— Il ne s'agit donc pas de la mort, ou de l'enfer ?
— Voyez-ça plutôt comme un au-delà pragmatique. Un endroit où l'on se rend pour vivre une transition. Un entre-deux.
— Cette chambre ?
— Et tout ce qui la précède, et tout ce qui la relie au reste de l'univers, nos pensées, ces murs, ces couloirs. Le Néant, Bracken, c'est l'autre côté, le côté plein du vide de ce qui nous entoure. Vous n'avez jamais vu que la matière, vous pensez que nous sommes pleins, mais si ce qui nous constituait était creux ? Que tout autour de nous, une matière nous définissait.
Je pense : il est fou, il est vieux, il se cherche une foi pour accepter la mort. Mais mon autre hémisphère cervical me dit le contraire. J'essaye de me remémorer Elliot, son visage. Tout de lui m'est étranger. Comme si je n'avais jamais été ici. Comme si je ne l'avais jamais connu. Je ne reconnais pas Elliot dans cette chambre, ni sa silhouette, ni sa trace. Si tous ces objets, ces meubles, forment les traits de son visage, où sont ses yeux ?
— Plouffe, c'est bien gentil tout ça, dit Fink, mais on parle d'Elliot, pas d'un magicien ou d'un philosophe !
— Peut-être simplement d'un poète.
— Un poète !
— Orphée était poète, et il avait de grands pouvoirs magiques. Il chantait, et les cailloux venaient l'écouter. Il pouvait soulever des vents terribles, et guérir des plaies. Elliot dessinait, vivait une vie dédiée à son ouvrage, à cette école, à son monde. Il avait compris toutes les nuances de cet univers riche, il s'en satisfaisait et il l'exprimait, à sa façon. Cette chambre est son œuvre.
— C'est votre faute, Plouffe ! C'est vous qui lui avez fourré toutes ces sornettes dans la tête, vous et vos Ptyx en plastique !
— Le Néant est bien réel pour un poète, Fink. Le Néant existe tout autant que la matière. Pour lui, les deux sont indissociables, interdépendants et, par le rythme auquel ils s'enlacent, ils créent la continuité de la vie. Elliot pensait peut-être qu'il était possible de pénétrer dans le Néant, tout comme nous vivons dans la matière. Il voulait faire sa catabase.
Je lève les yeux vers Bram, espérant de l'aide, mais ses yeux sont vides de toute expression. Le Néant nous a pris, tous, il est là, je peux le sentir, en son absence.
— À quoi pourrait bien ressembler le Néant ? demandé-je.
— Comment le savoir ? Personne ne l'a jamais vu. L'Islande est un poste d'observation privilégié, à la jointure de deux plaques tectoniques ouvertes sur l'intimité de la planète. Entre les deux, se trouve peut-être ce Néant.
— Comment quelque chose qui n'existe pas pourrait-il avoir une profondeur ?
— Le Rien est enceint de tout, mon petit Bracken. Le Néant est omniprésent, dans tous nos gestes, dans toute notre vie. Le Néant est partout autour de nous, entre chaque chose, définissant le contact entre les surfaces. Il abolit toute forme.
Une inspiration monte en moi : ENTRE les surfaces ! L'image me fait brutalement penser à une ligne noire cernant un personnage de bande dessinée ; à l'espace entre deux cases.
— L'ellipse incarnée, fais-je.
Il voit que je frémis.
— Qu'est-ce qui se passe, Bracken ?
Je sors la BD de ma poche.
— Où l'avez-vous trouvée ? dit Plouffe en m'arrachant l'ouvrage des mains.
— Dans son secrétaire...
Plouffe regarde la grille raturée, grommelle.
— Si Elliot peut voir les interstices, il peut lire entre les lignes pour atteindre le Néant. Peut-être son handicap l'avait-il préparé à tout ça. Les autistes peuvent être hypersélectifs, ils n'ont pas d'intelligence systémique, ils se braquent sur un détail et en font un linéament de la réalité.
— Il était plombier ou poète ? demande Fink.
— C'est la même chose ! Elliot était un poète, ses gestes étaient motivés par une soif de grâce. Nous n'y comprenons rien car il était fermé, il ne partageait rien. Elliot était un spécialiste de la plomberie... Il excellait dans la pose des joints ! Qui scellent les deux touts ! Ah, ah ! Voilà la clé, il faut trouver un moyen de construire un pont, peut-être en récitant des incantations, ou en tournant des vers épiques et spirituels, il faudrait commencer par mettre des bougies...
Bram s'étire doucement.
— Plouffe, je vous avais mal jugé.
Il s'avance vers le vieil homme.
— Vous me croyez ? Vous Bram, le scientifique ?
— Oh, tout cela est très scientifique, Plouffe. Je ne crois pas à vos gris-gris. Elliot est allé au Néant, OK, pour prendre un objet magique qui s'appelle le Ptyx.
— Comment savez-vous cela ?
— Ma définition du Néant et la vôtre se rejoignent. Tout comme vous parliez français à Elliot pour en faire votre _über_ -fils ou je ne sais quoi, moi, je lui parlais.
— Tu as entendu Elliot parler ? dis-je, proprement stupéfait.
— Techniquement je l'ai lu.
— Où ?
— Dans notre jeu, _Habitat_.
— Mais c'est quoi ce jeu ?
— Disons que c'est une simulation de vie, nous avons un système de messages en temps réel, et Elliot nous parlait par cet intermédiaire.
Plouffe devient tout rouge.
— Il jouait sur un de vos ordinateurs ?
— Je ne sais pas comment il s'y prenait, mais il me parlait, par messagerie interposée, sans terminal d'accès.
— Et qu'est-ce qu'il vous disait ?
— Il était obsédé par une seule chose : voler le Ptyx.
— Elliot, un voleur ? Impossible !
— Et pourtant. Vous nous parlez de littérature ou de dessin, mais la vérité c'est qu'Elliot était avant tout un joueur. C'est ce que vous refusez d'admettre, parce que ça le réduirait à rien et que vous avez d'autres ambitions pour lui, n'est-ce pas, Plouffe ?
Plouffe reste coi.
— Vous le pouponnez pour qu'il croie à vos théories, à votre Mallarmé, que sais-je, mais il les pratique à sa manière, et ça lui donne, je pense, une sacrée avance sur vous. Elliot se fout bien de la façon dont vous le voyez. Il agit, c'est un joueur, et si nous parvenons à comprendre ses règles, alors nous pourrons le suivre. Souvenez-vous de vos parties de cache-cache au parc : vous l'avez dit vous-même, il joue tout le temps. La littérature est ici impuissante, Plouffe. À travers le jeu, Elliot a trouvé ce que nous cherchons tous en nous creusant la tête. C'est ça, l'art. Car le jeu n'est pas un art : c'est l'art qui est un jeu. Le jeu précède l'art. Il en est la condition. Le rituel, la liturgie, l'écriture, tout cela n'est possible que si l'on sait se placer dans un ailleurs. Et l'ailleurs improvisé est permis par le jeu. La foi elle-même est un jeu.
Je n'ai plus de mots pour décrire ce que nous sommes en train de vivre, ici, dans cette chambre. J'ai l'impression de glisser dans un monde nouveau, un monde qui me terrifie, dont j'ignore tout, dont je ne veux rien savoir.
— Nous sommes au seuil d'une découverte fondamentale, dit Bram. Si nous voulons retrouver Elliot, si nous voulons comprendre les règles, il nous faut remonter plus loin encore, à l'intérieur même de la matière, entre les atomes qui la constituent, là où plus rien n'a d'existence. Là où se trouve la règle.
— Vous êtes allé trop loin avec ces gamins, regrette Plouffe.
— Il a raison, dis-je soudain, émergeant de ma terreur. Qu'est-ce que tu sais du jeu d'Elliot ? Il y a un rapport avec vos ordinateurs ?
— Non. Elliot joue un jeu solitaire. J'ai simplement tenté de le copier, de l'émuler. À ma façon.
— Pourquoi ?
Bram hausse les épaules.
— Être le premier à percer le secret.
— Toujours être le premier, n'est-ce pas ?
— C'est le sens même du jeu.
— De toute façon, il va étouffer, dit Fink. Ça ne sert plus à rien d'en discuter.
— Étouffer dans le Néant ? questionné-je.
Plouffe roule jusqu'à l'angle.
— Si je comprends bien, Elliot essayait de trouver l'interstice et vous Bram, vous faites la même chose avec vos programmes ? Tout ce que je vous ai laissé entreprendre, parce que vous aviez ma confiance.
— Vous avez bien fait. Mais je ne pensais pas qu'Elliot réussirait à quitter notre dimension de lui-même, sans mon aide. Je veux savoir comment il s'y est pris.
J'ai besoin de retrouver un fil rationnel pour poursuivre cette discussion.
— Attendez un peu, cet interstice, il est entre quoi et quoi ?
— Cette chambre lui sert de rituel, quoi qu'en pense Plouffe. Le Néant est un fluide, un placenta. Presque une membrane. Elliot veut revenir avant la naissance, retrouver le tout primordial. Il va retourner à la vie d'avant la vie, à la non-vie. La mort, qui précède toute existence. Il va aller là-bas, dans ce monde interdit où rien ne se passe, il va y trouver l'inspiration de la poésie et se transcender.
— Et s'il se réincarne ? dis-je.
— Je n'y compterai pas trop.
Voilà. Je sais qu'il y a longtemps que je suis ici, avec eux, dans cette pièce, et que nous nous parlons. Nous tremblons ensemble, nous nous posons ces questions, et quand nous pensons détenir un fragment de réponse, le monde nous échappe et nous redevenons des enfants, émerveillés des possibilités que nous offre le monde. Je nous regarde nous ébattre entre ces quatre murs, sans point de fuite. Je me vois, moi Bracken, tout de noir vêtu, bretelles et moufles incluses, avec mes cheveux trop longs, je sais que je ressemble à l'image qu'un fan de science-fiction peut se faire d'un gardien de but. Pourquoi est-ce que j'accepte d'être à la merci des autres ? Est-ce encore par paresse que j'ai choisi de m'en remettre à eux ? Mais alors, pourquoi suis-je si en colère ? Y a-t-il quelque chose au fond de moi qui gronde de vouloir s'imposer, alors que tout mon être aspire au calme, à la passivité, au secret ?
Bram et Plouffe se figent, me dévisagent.
— Bracken...
— Bracken, qu'est-ce qui t'arrive ?
— Quoi ?
Plouffe pivote lentement et s'immobilise à un mètre de moi.
— Bracken, vous paraissez plus petit.
Je n'ai pas l'impression d'avoir rétréci.
— Allons bon, dis-je en faisant un pas dans leur direction.
Ils reculent.
— Bracken, à présent vous avez grandi !
Je regarde mes mains, dans le cocon de leurs moufles. Elles ont conservé les mêmes proportions.
— Attendez un peu... ajoute Plouffe en fronçant les sourcils. Bon sang, vous avez raison, Bracken, dirigez-vous vers le mur...
Je fais machine arrière. Leurs bouches s'arrondissent de stupéfaction.
— Vous êtes tout petit, Bracken, s'écrie Cyldrid.
Je recule encore. Plouffe agite les bras.
— Stop ! Stop !
Je regarde Bram. Il n'a pas cessé de sourire.
— Bracken le nain...
Je me retourne vers l'angle du mur. Et soudain, je réalise : tout me paraît plus grand, comme déformé par le grand-angle d'un appareil photo. Il faut se tenir face au coin de la pièce pour s'en rendre compte, mais c'est indéniable : les perspectives ont changé, étirées vers un point de fuite invisible, les parois du mur se dressent devant moi comme des monolithes, formant le portail d'un monde sacré.
— Oh.
— Voilà comment Elliot s'y est pris pour quitter la pièce ! jappe Plouffe. Il est bel et bien passé par l'angle !
— Mais enfin, comment est-ce possible ?
— Il a dû, tout comme vous, ingérer ces algues, et il s'est approché de l'infini. Voyez-vous, je suis certain que cet angle n'a pas de fin. Vous allez diminuer au fur et à mesure, mais pour nous, vous serez toujours là.
J'entends Bram soupirer.
— C'est un trou noir. Tu vas t'approcher de la singularité du point entre le mur et le sol, toujours plus près, à jamais sur la route extensible à l'infini du point zéro.
— Juste un trou noir ?
— Pourquoi pas ? Si Elliot est parvenu à renouer avec la préhistoire de sa pensée, ce qui précédait son essence, pour s'y cacher et gagner notre jeu, alors il a forcément déchiré quelque chose.
— Je croyais qu'un trou noir était un affaissement, dit Plouffe.
— Ça dépend du côté où vous vous situez.
Difficile de le contredire, vu que je n'y connais rien. Un trou noir, au fond d'une école, dans une chambre. Qui exerce une attraction. Depuis le début, je me dirige vers cet angle, je ne peux en détacher les yeux, j'y reviens sans cesse, mon attention est captivée. D'une certaine manière, j'ai été pris dans son champ de gravité. Malgré moi.
J'y suis tout entier.
KOR
De rien se fait l'essentiel du cygne
De ses ailes, frappant l'eau merveille
Des concentriques, il tire un Mi.
Quand Cyldrid revient, tout le monde est silencieux.
— C'est à n'y rien comprendre, les placards sont vides, il n'y a plus aucun dessin, quelqu'un les aura pris, volés, je ne sais pas... Ce n'est jamais arrivé, et je me demande si Elliot n'a pas... Oh !
Elle me voit, adossé au mur, les bras croisés.
Elle porte la main à la bouche de surprise.
— _Bon, alors, Néant, Ptyx..._
— _Il manque un indice..._
— _Oui... Je me demande... Et si..._
— Mon Dieu, Bracken...
Fink la retient.
— C'est peut-être contagieux.
Après avoir lâché un soupir, Plouffe prend la parole, comme s'il devait faire ses adieux au monde terrestre.
— Nous pensons que Bracken est plus loin dans l'espace que nous le sommes.
— Comment ça ?
— Qu'il est plus proche de la fente.
— Le Ptyx ?
— Ce qu'il y a dans l'angle du mur, tout au fond. Le lieu et l'objet, confondus. Ensemble, car le Néant est tout.
Plouffe désigne le dessin déchiré.
— Quand tout se retourne. Et que l'on passe de l'autre côté.
— Que voyez-vous ? me demande Cyldrid.
— Les murs se divisent, ils forment un seuil. Je crois qu'il y a encore plus d'espace en bas.
— Bracken, revenez, cessez de faire l'imbécile !
Je reviens vers elle, mais je demeure à un mètre, pour ne pas la toucher.
— Il faut que quelqu'un y aille.
— Mais si vous disparaissez vous aussi ?
— Avons-nous le choix ?
Pourquoi jouer la vie d'un homme contre celle d'un autre ? N'est-ce pas là ce qui fait la vanité de notre existence, toujours à vouloir échanger une valeur contre une autre, sans se soucier ni des besoins, ni des désirs ? Nous bradons le monde pour avoir le choix. La vie est un jeu à somme nulle.
— Emmenez les tortues, dit Plouffe en me tendant le bocal.
— Pourquoi ?
— Elliot les a ramenées, je pense qu'elles viennent de l'autre côté. Elles vous aideront en cas de problème, j'en suis certain.
Je regarde les tortues. Elles nagent en silence, on dirait qu'elles sont malades.
— Très bien.
Plouffe éternue.
— Prenez soin de vous, Plouffe.
— Vous me survivrez !
Cyldrid enserre mes mains crispées autour du bocal avec les siennes et me regarde tendrement en une muette bénédiction.
— Gardez vos moufles, Bracken.
Je remue mes mains, dans leur cocon de laine. Un bouclier.
— Toutes les églises elfiques d'Islande sont couronnées d'un portail double, m'indique Fink en me tendant les cachets d'Elliot dans un petit sac en cuir. Tenez Bracken, au cas où.
Il hésite un instant, l'air triste, puis comme affaissé sous le poids des ans :
— Dites-moi Bracken...
— Oui ?
— Pourquoi êtes-vous parti ?
J'ai toujours su, quand j'ai accepté de revenir, que je devrai le lui expliquer. Certaines choses doivent être gardées pour plus tard, quand les âmes seront des fleurs ouvertes capables de retenir la rosée sans la boire. Mais peut-être est-il aussi trop tard pour retenir ce qui doit sortir. Pourquoi suis-je ici, dans ce cachot de misère, à tenter de recoller les fragments de ma vie ? Je n'arrive pas à trouver une seule raison satisfaisante, tout effort en ce sens semble vain. Passés les quelques instants de doute, je me rends compte que toute ma vie semble pétrie de cette incertitude, de ce non-choix. Je n'ai pas de stratégie, aucune carte en main. Je me contente de prendre ce que me donne la pioche, à chaque nouveau tour. Parfois, c'est drôle. Parfois triste. C'est souvent déconcertant. Et tout le temps frustrant. Si on ne s'attend pas l'inattendu, il n'arrive jamais rien. Moi, je peux tout concevoir, mais le destin ne m'offre aucune nouvelle clé. Je n'escompte pas spécialement comprendre comment gérer une vie correctement à mon âge. Pas d'attaches, aucun principe. Je suis prêt à tout. Donnez-moi une raison de vivre, de rester ici parmi vous, de continuer à respirer. Voilà la réponse : je suis ici.
— J'ai perdu un duel contre Bram.
— Un duel de quoi ?
— De dessin, dit Bram sans nous jeter un regard.
— Vous dessinez, Bram ?
— Non. Mais avec un programme, un ordinateur et une souris, on fait des merveilles. Bracken pensait qu'il serait impossible d'aller aussi vite que lui. De faire aussi bien. Nous étions convenus que le perdant quitte l'établissement. C'est que moi, je joue pour gagner.
C'était la vérité. J'avais été déchu du seul talent qu'on me prêtait, un talent que j'avais laissé à l'abandon, entre des doigts étrangers. J'avais cru, dans ma prétention, que j'étais le garant et l'héritier d'une tradition que rien ne remettrait jamais en cause. Je pensais que, même si le monde évoluait, il resterait toujours des hommes comme moi, attachés aux valeurs fondamentales du trait, de ce lien direct entre la pensée et la main. Avec un bout de plastique et un écran, Bram m'avait humilié, devant ses élèves, devant Cyldrid et Plouffe. Je n'avais qu'une seule parole et je savais que nos mains allaient être remplacées. J'étais la première victime. J'étais mort.
Je me retourne, sans un mot pour Bram qui, les bras croisés, attend contre le mur. Je prends une inspiration et je me dirige résolument vers l'angle.
— _Et si on imaginait que tout ça n'était jamais arrivé ?_
— _Oui, si seulement Elliot n'était pas allé voler le Ptyx..._
— _Je me demande si... hein ?_
Tout autour de moi, quelque chose s'est mis en mouvement. Je peux sentir les objets se rapprocher, comme un étrange origami dont ma seule présence orienterait les plis. Se peut-il qu'un peu d'algue ait suffi à changer à ce point ma perception de l'espace ? Pour la première fois depuis mon retour, je me demande si je ne rêve pas. Mais comment l'affirmer, puisque je n'ai jamais rêvé ? Je ne peux en déduire que tout ce qui distingue d'une vie stérile constitue nécessairement un rêve. Tout a l'air si réel... mais peut-être un rêve est-il aussi réel que la vie ? Peut-être les rêves sont-ils une vie pleine et intelligible, qui n'acquiert sa consistance onirique qu'au réveil, quand notre cerveau essaye de rabouter des fragments en un tout narratif cohérent.
— _Diane, vous vous souvenez !_
— _Oui ! Le pauvre Elliot ! Au Cap du Néant..._
—... _pour devenir Maître ! Oh lala !_
La tapisserie brille, quelque chose la recouvre, comme un prisme qui dévie la lumière, la fait vaciller. Il y a une couche sur les meubles, un vert léger, transparent. Les algues qui ont servi à tracer la tapisserie, sédimentées après tout ce temps. Mais cette poussière est nouvelle, elle ne se voit qu'une fois très près du pli. De la condensation se forme sur la laine de mes moufles. Quand je darde la langue pour y goûter, c'est salé. Une odeur nauséabonde envahit mes narines, un serpent de malaise, une infâme bouillie d'algues et de poissons pourris, échoués sur le sable... Le Néant monte, le Néant est une marée qui obéit à ses propres lois, poussée par le jeu complexe d'astres inconnus, de principes qui violent les lois de notre science, et dont je suis le preux découvreur. Ou le concepteur : suis-je en train de défricher un terrain déjà existant, ou crée-je ce que je vois à mesure que je l'invente ?
KOR
Arrache le Ptyx des mains du Poète,
Fais vibrer leurs silhouettes par l'écaille.
Le retour sera fin pour les habitants du plein.
Imperceptiblement, la lumière change. Le bleu devient plus profond, forme un véritable nuage. Je me retourne et derrière moi, très loin, après le lit, par-delà les monts du plancher et le ciel du plafond, où les loupiotes deviennent constellations, je vois mes amis, ils me regardent partir. Je n'ai pas l'impression d'avoir rétréci. Je lève les bras pour leur faire signe. Je regarde tout autour de moi, dans cette montée de Néant que je sens pourtant, une bouffée de joie et de tristesse entremêlées qui me prend tout entier, qui monte toujours plus haut puis se retire, jusqu'à ce que son mouvement m'enveloppe complètement.
Le monde se transforme en gruau de tapisserie, de moulures et d'objets, morphés les uns dans les autres pour former une nouvelle texture, un tourbillon, incohérente mélasse qui s'agglutine autour de la moindre aspérité. Comme si tout se dissolvait pour redevenir un, réintégrer le chaos originel, indifférencié. Je suis témoin, sans rêver, en explorateur, aventurier inconscient du danger. Je n'ai rien, ni outil, ni arme. J'entre nu dans la marée du Néant.
— _Comment ça, vous vous souvenez ?_
— _Je me souviens des mains d'Elliot, de son plan..._
— _Attendez... son plan pour voler le Ptyx ! Mais oui !_
Je me souviens de mes vacances, tout jeune, sur une plage infinie parcourue de sillons et recouverte de bois flotté, comme des amarres pour empêcher le sable de glisser dans la mer, ou destinées à casser les vagues. Je me souviens des forteresses au loin, endormies dans la brume iodée. À marée basse, nous allions pieds nus, pantalons relevés jusqu'aux genoux, pour explorer ce royaume mystérieux entre la mer et la roche ; un monde subtil aux couleurs discrètes, où tout éclate dans un détail, un minuscule écosystème, si fragile qu'un seul orteil aurait suffi à déranger toute une colonie. Je me souviens des anémones dans l'eau brouillée, des petits coquillages vaillamment cramponnés, des fientes de mouette traçant d'indéchiffrables arabesques entre les failles ; et de la frontière mouvante entre la mer et la plage, comme le contour translucide d'une décalcomanie comme un bronzage finissant net sur le maillot. Tout le bleu du monde, et maman qui disparaît entre les rochers, à la recherche d'un hippocampe. Soudain la mer qui rote, la mer qui pète, écœurante, comme une digestion, et maman tout entière avalée puis recrachée, décomposée.
— _Oui ! On lui avait dit pourtant, toucher c'est voler..._
— _Il a volé le Ptyx, c'est le nouveau Maître !_
— _Et c'est notre faute !_
L'iridescente tapisserie se change en une jungle de lianes phosphorescentes, qui poussent avec luxuriance en irriguant leurs filaments de flashes, de pulsations. Un rythme tribal monte de cette jungle virtuelle, qui se peint sur mon cristallin puis s'évanouit lentement, entretenue par la persistance rétinienne. Je ne discerne plus les limites, et d'étranges formes flottent et dansent alentour, silencieuses anguilles dans un aquarium sans parois. Noir. Si noir. Pourquoi le noir est-il si dense ? Qu'y a-t-il dans ces ténèbres de si pâteux, qui forme une telle mélasse ? Je n'ai jamais su pourquoi le noir nous permettait de convoquer les pires images de nos mondes intérieurs. Souvent la nuit, quand j'oublie de rêver, quand je dors les yeux grands ouverts, je dessine les images d'un monde perdu dans les recoins à peine éclairés de cette pièce bleue qui me sert d'habitation. Les plantes et les rochers que je trace sur ces surfaces invisibles sont d'une telle plasticité qu'ils se transforment en de terribles prédateurs aux mâchoires béantes qui reviennent me hanter sitôt qu'autour de moi se trame le rideau du sombre. Comment concevoir l'écologie du Néant ? Si sa texture est un brouillon sans limites, aux vagues reflets bleutés, quelles terribles et impossibles créatures peuvent donc en jaillir ? Comment les affronter ? Je ne suis pas prêt. Je ne suis qu'un prof de dessin qui ne dessine plus, je n'ai aucun talent de rechange. Je ne sais même pas rêver, quelque chose que tout le monde sait faire, même les méchantes gens, dépourvus de scrupules qui nous marchent dessus sans un regard, je ne vois pas comment je pourrais les aider. Je voudrai juste retrouver Elliot. Je ne veux pas qu'il meure, je veux savoir ce qu'il voulait me dire, moi qui ai peut-être entendu les seuls mots qu'il ait jamais prononcés. Il aurait pu me sauver et m'apprendre à rêver, lui qui vivait enfermé dans un autre monde. Pourquoi ne l'ai-je pas écouté ? Pourquoi ai-je préféré défier Bram en duel ? Si seulement je me connaissais un peu mieux, je pourrais m'écouter.
— Rêve pas, mon loulou.
— _Mais qu'est-ce qu'on a fait ?_
— _On lui a dit où était le Cap !_
— _Moi j'ai rien dit ! C'est votre faute !_
Bram est soudain à côté de moi.
— Tu n'imaginais quand même pas que j'allais te laisser y aller seul.
Je me retourne, les autres ne sont qu'à quelques mètres mais j'ai l'impression de ne plus pouvoir les rejoindre. Ils sont très gros, très loin.
— Comment as-tu fait pour venir jusqu'ici sans lunettes ?
— Tu laisses une trace dans le Néant, comme un sillon, qui ne se referme pas. Et je pense que tu as pris le sentier d'Elliot, que toi seul peux voir, avec tes yeux. J'ai confiance en ton instinct. Les autres ont dit que j'étais fou, mais je ne trouve pas que ç'ait l'air si dangereux. Tu n'as pas entendu notre conversation ?
— Tu as dit qu'il s'agissait un trou noir.
— Ça l'est, d'une certaine façon. Ce que je ne m'explique pas, c'est pourquoi la chambre n'a pas été attirée tout entière. Peut-être que quelque chose sort du pli, mais n'y entre pas. Elliot a forcé le passage, mais c'est trop petit.
— Elliot... et nous !
— Espérons que nous ne soyons pas trop gros.
Le sol glisse, comme s'il penchait. Je cherche un appui. Les murs suintent, gouttent. Je peux le sentir, comme un reflet sur la tapisserie. Plus je marche vers le mur, plus il semble s'éloigner, plus l'angle du pli semble s'élargir. Se pourrait-il que par le Ptyx, le Néant monte et descende ? Peut-être alors que ce que nous avons trouvé dans la chambre marine, ces objets, ces énigmes enfermées dans des mystères, sont les sédiments déposés par l'océan, les traces d'un monde invisible, qui n'apparaît qu'entre les choses, à travers leurs relations, le rapport que nous entretenons avec elles. Toutes ces dépendances forment un visage, une personne. Elliot.
— _Ils vont tout déchirer !_
— _Mais qu'est-ce qu'il va se passer ?_
— _J'espère qu'ils ne vont pas essayer de voler le Ptyx._
Entre mes jambes file un courant invisible, rapide, si rapide, que je ne peux résister. Bram s'accroche aux meubles. La fente semble plus large désormais, une valve qui frétille, veut tout avaler. Le point de jonction entre les deux murs, le plafond et le plancher se réduit doucement, puis s'ouvre, comme une porte, libérant un filet de rien qui semble tournoyer, ou peut-être suis-je simplement en train de l'imaginer. Passer de l'autre côté équivaut à inverser les perspectives. Ce qui était sommet devient crevasse, les triangles se recoupent. Bram me retient alors que j'avance vers la fente. Il me retient, car à part moi, qui lui servira d'adversaire ? Je ne céderai pas à ses sirènes. Le jeu n'est pas pour moi. Tout ceci est très sérieux. Elliot est là-bas, et je dois le ramener. Peut-être est-il déjà mort. Fink a dit que sans ses pilules il ne pourrait pas survivre plus de quelques heures. À quoi bon tout ceci, si c'est pour ramener un cadavre ? Pourtant, impossible de retourner en arrière, ils comptent tous sur moi et même si je ramène un corps ce sera la preuve de ce qu'il a accompli, de cette traversée réussie. Je lui dois bien ça, à Elliot. Après tout, je ne l'ai jamais remercié pour ce qu'il m'a appris, dans le silence capitonné de sa gentillesse.
— _Bracken va passer !_
— _Il ne sait pas ce qu'il est en train de faire !_
— _On ne doit pas le laisser tomber !_
Par un effort surhumain, j'entame un crawl dans le vide, des mouvements lents, ralentis par la pâte invisible du Néant, la purée du rien qui me retient de glisser entre les murs. Entre mes mains, le bocal se brise et l'eau se met à flotter sous mon nez, comme une pâte à tarte ondulant. Les deux tortues se mettent à battre des nageoires maladroitement dans l'air trop épais. J'ai l'impression de voir leurs petits corps se disloquer, et je me dis que je les suivrai de peu. Tout mon corps se replie suivant des angles impossibles, mécaniques, sans déchirer les muscles. Je suis témoin de ma propre désarticulation. Ça n'a rien de métaphysique. Sans un bruit, tout lâche, le corps, la réalité.
— _La marée !_
— _Adieu, Belle Diane..._
— _Adieu Roméo..._
Les contours de la pièce vacillent, comme un reflet remué dans la soupe. Les formes coulissent les unes dans les autres, et je sens mes jambes m'abandonner. Brusquement, le plan s'incline, et ce que je croyais être le haut devient le bas, la gauche, la droite, tout s'inverse, se renverse ou devient miroir. Le reflux de l'écume que j'entends sans le voir, dont je perçois simplement l'effervescence, et me voilà dégringolant, aspiré par le trou, par le vide, par le sombre conduit. J'essaye de m'accrocher à quelque chose de concret, qui ne soit pas l'une de ces créations fantasmagoriques, mais tout est faux, simple dessin, motif ou angle et tout n'est qu'illusions et trompe-l'œil. Il n'y a pas de prise, mes doigts se referment sur le vide, et je sais que c'en est fini de moi, que la corde a lâché, et que rien ne m'empêchera plus d'être aspiré par le vortex infernal, invisible, tapi tout au fond du Ptyx. Je hurle, mes mots semblent se délier, se confondre avec la tapisserie, voyelles, consonnes et syntaxe désintégrées, bouillie de pensée qui n'a plus de sens, fin du possible, commencement du doute. Avant de disparaître pour toujours, Bram esquisse un dernier geste désespéré. Sa main se tend telle une étoile scintillante dans le tourbillon que devient ma vie, tout s'éteint doucement, doigt après doigt.
KOR
Fermer les yeux, pour ne dire que l'eau
Qui entre ces lèvres, coule de source
Et dans l'informe évidence, se noie.
Bulles. Bulles.
Tourbillons.
Je me retiens de mourir. J'ai l'impression d'être dans une machine à laver. Écartelé, retourné, tourneboulé. Je dérive dans un courant furieux. C'est ainsi que le souffle nous quitte : trop longtemps retenu dans l'armure du corps, il s'enfuit et laisse ouvert le gouffre béant d'un vide que la nature remplit d'une eau sournoise. Passée la brutalité de la première gorgée, interminable, se noyer, paraît-il, est indolore, voire agréable. Nous redevenons fœtus, imbibé de tout le liquide de la mère. Elliot est donc mort noyé, redevenu plus petit que l'enfant qu'il n'a cessé d'être, et j'ai suivi son chemin, sans comprendre ce qui m'attendait. Mourir, ça s'apprend tout petit, comme une naissance à l'envers. Puis on grandit et on ne cesse de mourir. Avec le recul, je prends conscience de tous ces instants de mort, de toutes ces occurrences qui ont fait de ma vie ce qu'elle est devenue. Il ne s'agit pas de prétendre qu'on devient plus fort après chaque défaite, au contraire, chaque échec nous ôte un morceau de nous-mêmes. Petit bout par petit bout, on devient plus petit, on s'approche du rideau qui nous sépare des ténèbres.
Lâchant prise, je laisse le Néant m'envahir.
# LE NÉANT (2)
Entraîné par le poids de son inertie, ce qui n'existe pas s'emballe. Épais, lourd, il pèse sur sa non-existence qui le force à occuper une place dans un espace non fini : il crée par son défaut le nécessaire emplacement de son absence. Cette friction avec lui-même passe encore et encore par un même emplacement, la seule présence définie dans un vaste océan de rien. Toute réalité se fragilise en ce point unique, passé et repassé jusqu'à l'usure. Ce qui n'est pas s'engendre en répétant sa disparition. Peu à peu, la trace devient flèche, courant continu à ampérage constant. Elle devient sens, dans les deux sens. Cette division primordiale, cette poutre travaillant le zéro divise l'espace, créant une droite, une gauche, un haut, un bas, si tant est que quelqu'un soit là pour en juger. Pourtant, sans perception, la traînée s'affirme, persiste. Impossible de l'ignorer plus longtemps. C'est la preuve qu'une force invisible peut devenir visible par le frottement qu'elle engendre. Mais la texture de cette présence adverse, sur laquelle elle imprime sa trace, ce résidu, n'existe pas plus. Ou bien : si tout est compris dans le ventre du rien, la friction s'exerce en dedans. S'automutile. Une scarification de l'intérieur, créant de fait une première limite à ce qui n'en avait pas. Le non-être ne peut intervenir que s'il se connaît, s'il a conscience de lui-même. Ce procédé de génération spontanée s'opère donc par la douleur. En se marquant, par la rotation éternelle de son inanité, il se justifie à ses propres yeux. Le son lui tranche une forme, un point de départ, infinitésimal, d'où les limites peuvent naître. Ainsi, irritée, cette blessure originelle est la matrice de tous les mondes, de tous les mythes, de tout ce qui engendre la création. Car, entaillée, la chair du vide forme une fissure, puis des lèvres qui, pour un mot, laissent filer la fertile semence. Dans ce déplacement, où deux chaînes nulles transportent de l'énergie, la matière peut naître, en dehors du rien, de l'autre côté de la tranchée, où sur ces franges déchirées s'accrochent encore, tels des parasites, les déchets de ce premier jaillissement de vie, cet éveil primal qui ne devient que désert abrutissant.
La ligne, entre le vide et le plein, où dansent les ruines du cosmos.
Ginnungagap
Ár var alda
það _er ekki var,_
vara sandur né sær
né svalar unnir ;
jörð _fannst æva_
né upphiminn,
gap var ginnunga
en gras hvergi.
Je ne sais pas combien de temps je reste inconscient. Un million d'années, des éons, juste une pression sur le front, mes lèvres en ventouse. Quand j'ouvre les yeux, le monde reprend forme, une trace qui doucement se détache d'un infini qui se compose, matière, peinture, l'espace entre mon corps et les choses, espace qu'on ne saurait qualifier de discret, pourtant : et si tout n'était que fragments, si le tissu même de nos pensées, de toutes les galaxies qui nous composent, de nos amis, de nos papas et de nos mamans, décalques d'une vie, séparés les uns des autres, et moi, vous, nous tous qui occupons la conscience dans l'entre-deux, la bande entre deux cases, le rien terrible qui vous regarde tout le temps, toute la vie, jusqu'à la fin ; et si tout, tout était pulvérisé ? Au profit d'une nouvelle idée, née du noir le plus profond. Comprendre que la continuité apparente des choses résulte d'un affinement en escalier, de barres qui composent tout, devenues, par l'exercice des sens, une ligne droite prétendument jamais croisée.
Mmmmh...
J'ouvre un œil, mes lèvres recrachent quelques bulles salées. Je suis tout cassé. Ma tête est une essoreuse. J'essaye de fixer un détail devant moi pour stabiliser mon mal de mer. Un grain de sable, deux grains de sable, trois, puis cent, puis mille, puis l'univers. J'accommode peu à peu, j'entends ma respiration, je prends une longue bouffée pour me sentir vivant, le cœur reflue, je me retiens de vomir, me contente de roter. Je laisse échapper un rire nerveux qui dit juste mon plaisir d'être en vie.
Je me redresse sur le rivage, jeté comme un naufragé. Un moment d'hésitation quand, dans la clarté laiteuse, je prends la mesure du temps. Vertige. M'asseoir dans la fraîcheur du ressac, reprendre un souffle que je croyais éteint. Retrouver toutes les parcelles de mon corps, une par une, fonctionnelles. La laine de mes moufles est gorgée d'eau, je frissonne.
Dans le ciel, d'énormes nefs croisent en silence, sans effort, glissant vers un ailleurs fait de noires profondeurs, perspectives de collines rêvées, des lueurs, des étoiles, pâles signaux qui clignotent pour moi, ou pour rien. Un instant de répit pour me sentir au complet, les jambes croisées, le corps trop lourd mais en vie, empli de moi-même. Les vagues laissent leur écume autour de moi, une silhouette blanche détourant ma personne dans ce lieu désert, où quelques traits gris attestent dans le ciel d'un vol de mouettes, planantes, portées par le vent.
La plage étire son long ruban de sable noir à perte de vue. Comme une ligne sombre, tracée entre une mer trouble et un ciel d'ébène. Aucun horizon, du moins rien de lisible ou de très net, entre ces renflements, un azur si vaste qu'on le dirait courbe. Il ne semble pas y avoir de fin, ni de falaise ou de village. Pas de touristes, pas de baigneurs. Aucune voile sur les flots. Pourtant au large apparaît une constellation de taches sombres, des rochers déchiquetés, tels les crêtes de créatures pétrifiées. Il me semble reconnaître cette plage. Quand j'ai débarqué en Islande, Fink m'avait emmené au sud, une fois, pour me montrer les plages de sable noir. Ce n'était pas si loin de la capitale, quelques heures de route pour arriver nulle part. J'avais déjà l'impression d'être sur une autre planète, retiré de toute vie sociale, laissant derrière moi la culture, les soucis, les envies. Un monde nouveau, extra-terrestre, pour moi seul, où ma petite personne trouverait sa place. Peut-être s'agit-il de la même plage ? Ce serait si simple à expliquer : entre la chambre et le moment présent, un long évanouissement, qui ne tarderait pas à recevoir une explication rationnelle.
Je pourrai rester dans cet entre-deux, sans aucun but, si ce n'est celui de retrouver le chemin de l'école. Évanouissement ou pas, il s'est passé quelque chose, dans cet angle. Comme une aspiration. J'aimerai savoir. Ou pas. Après tout, je pourrai en profiter pour prendre la fuite, loin d'eux, de leurs théories, de leur folie trop islandaise à mon goût. Peut-être devrais-je juste rentrer chez moi. Me recoucher sous la couette, respirer les odeurs douces des draps, placer ma tête contre les coussins, dans le creux charnu de leur masse.
En me levant, je remarque la trace que j'ai laissée, dans le sable noir.
Un cercle rompu.
Dériver le long du rivage. Le ressac lèche mes pieds. L'eau, toujours, insondable, mystère qui ne dit aucun mot, ne lâchant pas le sens. Le sable est si noir qu'il ne retient aucun reflet : en l'envahissant, l'eau qui lèche la berge est bue instantanément, comme épongée par les ténèbres. Quand je me penche pour en saisir une poignée, ce sable semble plus fin que la poussière, terrible symbole du temps qui passe, qui me glisse entre les doigts et retourne combler ces trous. C'est ainsi que l'énergie retrouve sa forme, jamais arrachée, extensible, s'aplatissant.
Quelques rochers orphelins affleurent comme une forteresse sortie des flots. L'eau s'en retire, révélant des galaxies. Timides, ils dressent leurs pointes tels des ailerons, tout au bord, vaillants éclaireurs des premières frontières où la terre rencontre la mer. De cette étreinte naît une zone inter marées, un monde en soi, minuscule poche de vie au sein de laquelle la dynamique puise son sens. Entre ces interstices je devine une activité, des algues, des poches et des ventricules qui s'ouvrent et se ferment dans la clarté de l'eau, minuscules labyrinthes de lames acérées, alternant avec des trous sans fond où se terrent de petits yeux, des mouvements fugitifs de crustacés invisibles. Tout y prend son temps, indolent dans les remous. Quelques bruits percent parfois, fuites de gaz lorsqu'une patte invisible agite l'onde, qu'une goutte glisse le long d'une lèvre de roche, jusqu'à rejoindre la matrice et se confondre en auréoles concentriques. Entre deux fissures se faufilent des fleurs étranges, ondule une masse de pédoncules qui se rétractent quand je pointe mon doigt dans leur direction. Mon majeur est soudain aspiré. Je le libère dans un petit bruit de succion. Motifs ornementaux, déliés et spirales aux efflorescences parfaites, scintillant de vitesse et pourtant immobiles, s'entrelacent, zigzaguant d'une veine à l'autre. Accumulées en couches sédimentaires, ils racontent un passé. Tout éteinte que soit cette vitalité, elle parle autant qu'une autre, s'arrache à l'inertie par la simple beauté de son relief. Son évidence est ancrée dans sa texture, comme une croûte séchée d'une blessure obtenue par courage.
Ces multiples traces m'entraînent dans un monde où tout répond à ce qui le suit et le précède, où la moindre délicatesse est un événement. Cette miniature, cette maquette du tout, c'est l'incarnation de ce qui manque au visible. Les coulisses du réel, les extensions, les petites touches qui donnent la profondeur, étranges dans leurs perspectives, insondables dans leurs rituels. Je contemple le spectacle silencieux de cet univers en expansion et pourtant stable, paisible comme peut l'être une patience, image même de la solitude.
Reste cette étendue terrible, ce ruban perpétuel qui, au loin, semble s'effacer dans un nuage de brume. À quelle géographie vais-je être soumis ? Dois-je fantasmer ces atomes d'événements pour en déduire un monde complet ? Je ne suis plus là où j'étais. J'ai basculé ailleurs, dans quelque limbe, ou dans le flou d'un rêve. Si je dors, alors pour la première fois je comprends le songe, et ce sentiment de ne pouvoir dominer le flux. Je me laisse emporter par mes pas, par la brise, par la certitude que tout finira par faire sens, quand le réveil sonnera et que je serai devenu comme les autres un rêveur, l'esprit encombré de ses souvenirs fragmentés.
Je devine les ruines d'un château de sable, gigantesque cité visiblement détruite par le vent, non la marée. Ses remparts restent encore debout, témoignant du génie architectural de leur constructeur. Difficile d'interpréter de ces murailles, ces fossés, ces tours effondrées. Il y a un trou dans le sol, à présent rempli d'eau, qui devait ouvrir sur un tunnel. Plusieurs ponts détruits, dont ne subsistent que les piliers. L'édifice central devait être massif, presque une tour défiant la gravité, à en juger par la masse de sable qui demeure. La cour principale est assez grande pour qu'un homme adulte puisse s'y tenir. Je ne sais pas le temps qu'il a fallu pour bâtir un tel miracle. La mer elle-même n'a pu en venir à bout. Le sable d'ébène qui le compose miroite au soleil. Il est par endroits si tassé qu'on ne peut l'effriter du bout de l'ongle. Un instant me prend l'envie de m'établir en archéologue en châteaux de sable. Après tout, j'ai déjà été archéologue d'une chambre d'enfant, remontant avant la naissance de son occupant. Il y a peut-être une carrière à mener : mettre au jour les secrets des enfants en fouillant les couches successives correspondants à leurs activités. Mais est-ce réellement un enfant qui a construit ce château ?
Il n'y a nulle trace de pas alentour, ce qui témoigne de la délicatesse de son éventuel bâtisseur. Cet inconnu est déjà un dieu pour moi, un être savant dont je tente de saisir quelques éléments afin de donner un sens à ma présence en ce lieu. D'où venait-il ? De la mer ? Comme moi. C'est ainsi que l'on arrive ici : en naufragé. Forcément. Il n'y a rien. C'est un mythe bien répandu dans l'inconscient collectif de nos civilisations : l'arrivée dans le rien par la mer, expulsé du ventre maternel, nous naissons au vide dans un remugle d'algue et de sel, encore baignés du liquide qui nous a nourri pendant des mois. Engorgés. Avec pour unique objectif ces traces que des explorateurs ont peut-être gravées avant nous, et que la marée n'a pas effacées. Ces prédécesseurs deviennent nos dieux et nous interprétons à l'infini leurs actes indéchiffrables une fois sortis de leur contexte.
Suivre les pas. Quelque chose attire mon œil, à l'extrémité de la plage. Peut-être un vague relief qui me permettrait de prendre un peu hauteur afin de m'orienter. Dans ma précipitation, je manque de trébucher. J'aperçois mon ombre, noire sur le sable noir, qui se détache encore par je ne sais quel miracle. Je pense à Peter Pan et à son ombre qui se fait la malle. Ma jeunesse, mon enfance, tout, se condense dans cette silhouette impossible à attraper. Une masse émerge, une sombre paroi, bien solide, bien tangible sous mes deux mains. Un mur. Une limite, la première à cette plage, si j'oublie les traces que laisse l'écume sur le sable en refluant, comme autant de cristaux de lumière qui s'éteignent doucement.
Le mur se prolonge dans la mer jusqu'au large, et plus loin encore, semble en fixer l'extrémité. Il n'y a pas d'autre côté. On dirait une conque. Impossible d'en distinguer le sommet. La roche est compacte et constellée de petits cristaux d'où pulse une lumière blafarde. Je ne me laisserai pas arrêter par un mur. Je me remets en route, toujours en suivant les traces, cette fois vers l'intérieur des terres, en longeant la roche. Étouffés, le flux et le reflux des vagues ne forment plus qu'un simple murmure. Au-dessus, les nuages se sont écartés, laissant l'éclat surnaturel des étoiles me guider. Je ne reconnais rien à leur configuration, mais je n'ai jamais été très fort pour reconnaître la moindre constellation. Je n'ai pas eu de papa pour me les montrer du doigt.
Les étoiles me suivent, innombrables, mais la nuit autour de moi semble soudain plus épaisse. Comme un plafond, partout, qui soutiendrait la voûte céleste. Sûrement la fatigue. Après tout, je suis presque mort. Mais quelque chose sonne faux, et je comprends très vite quoi. Ce ne sont pas des étoiles. Il y a bien un plafond. C'est une caverne. Une gigantesque caverne, qui enferme la plage. À la surface de la roche scintillent des pépites. Le plafond descend doucement jusqu'au fond de la grotte, sur lequel je bute. Pas d'issue, pas de route. Peut-être y a-t-il une passe plus loin ?
Courir le long du mur, dans l'espoir d'y trouver quelque chose qui me permette de sortir de cet enfer, mais je comprends vite l'horrible vérité. Il n'y a pas de sortie. C'est une caverne, et cette mer est une mer intérieure. Si j'avais parcouru la plage dans l'autre sens, j'aurai rencontré le même mur, qui m'aurait ramené au point que j'occupe actuellement. Je remarque devant moi que du sable noir a été remué, comme un nid dans le sol, creusé par un oiseau trop gros. Le naufragé s'est agenouillé ici. Il y a un petit tunnel, à la base de la roche. C'est tout petit. Trop petit. Je m'accroupis pour regarder de l'autre côté : je distingue une lueur tout au fond d'un goulot, et je sens l'odeur d'une grotte aquatique, sortie d'une enfance improbable. Je prends conscience de la faim, qui me tiraille, de la soif, de la fatigue et de cette panique qui monte en moi.
Les pas ne continuent pas. Le naufragé est passé par ce tunnel. Est-il si mince, si fluet ? Elliot, lui, était grand, très grand et mince. Allons, d'où venait-il ? De la mer, d'un endroit que j'aurais vu si j'étais parti dans l'autre sens, sur la plage ? Une seule façon de le savoir. Je fais demi-tour, en pivotant lentement, un moment dont je prends la mesure car la panique enfle en terreur : un énorme bruit ravage la structure même de la caverne. Un grondement qui fait trembler le quartz et provoque de petits éboulements. Je me tourne vers le rivage. Une brume dense s'élève dans la chaleur intense. La mer qui monte.
La marée me rattrape au galop. On dirait un chariot tiré par un troupeau de chevaux islandais, mèches d'écumes sur des yeux féroces, plissés dans un relent de vague. Où que je regarde, il n'y a aucune d'issue : déjà les traces de pas ont disparu, je suis dos au mur. Une pointe menaçante s'enfonce au creux de mes reins. En hâte, à genoux, je tente d'élargir le tunnel, dans l'espoir de rejoindre ces souterrains qui dansent à la lisière de mon champ de vision. Plus je creuse, plus le sable devient dur, difficile à collecter, et bientôt mes moufles sont des pâtés noirs agitant des masses sombres. Je les enlève sans réfléchir, pour creuser plus vite, plus profond. Un coup d'œil derrière moi : la mer est presque là, sans humour, sans recul, prête à m'engloutir. Poussé par l'énergie du désespoir, je redouble d'effort mais je heurte quelque chose d'encore plus solide : de la roche. Le tunnel restera trop étroit, cet autre monde ne veut pas de moi, je suis trop gros.
Les vagues m'encerclent sur un îlot, envahissent le tunnel puis, dans un soupir et un glouglou, se referment brusquement sur mes pieds, puis mes mollets, mes cuisses et très vite mon torse est submergé. Il ne faudra pas longtemps pour qu'elle m'emporte à nouveau, cette mère aux enfants morts. On la connaît bien, elle donne et prend la vie des pêcheurs, avec indolence, elle peut être la plus cruelle des maîtresses et soudain vous couver avec tendresse. Elle est la source de nos malheurs, de nos bonheurs. Elle insuffle l'espoir tout en prospérant sur nos détresses humaines. Elle exige un abandon.
Calotté, je dérive, une lame me retourne. Bulles dans le courant, je ne vais pas résister longtemps, je n'ai rien à quoi m'accrocher, je vais me laisser emporter, je n'ai pas le courage, pas la force. Je n'en peux plus. Alors que le reflux m'emporte, je parviens à garder la tête hors de l'eau assez longtemps pour apercevoir, au loin, une tache fugitive.
Je lève les bras.
— Au secours !
Je suis à nouveau aspiré, je sombre dans le mou, dans le malléable, dans le secret des va-et-vient. S'il y a une marée, il y a une lune, mais je ne vois que ces cristaux qui semblent se rire de moi quand, dans un remugle, perce à la surface mes yeux rougis par le sel, irrités. Mes bras sont lourds, je perds pied et entraîné par le fond, je me prépare à mourir.
Si près. Mais si près de quoi ? Peut-être suis-je châtié par le destin pour avoir osé défier le temps. Arrogance de celui qui croit qu'il vivra éternellement, dans un palais illuminé où résonnent les échos des fêtes passées, quand les espaces étaient purs, les choix infinis, les responsabilités inexistantes. Je n'ai pas été à la hauteur de mes aspirations. La paresse a fini par l'emporter, comme cette mer qui a raison de tous mes rêves, comme de mes efforts. Je préfère le sommeil d'une noyade à un combat perdu d'avance.
Quelque chose glisse sous moi. Quelque chose d'énorme, de lisse, qui s'entortille. Un instinct primaire vomit du tréfonds de ma personne, la terreur de ce qui nage en dessous, là où mes jambes sont libres. Cette fosse, qui grandit alors que je m'éloigne. Un requin, des piranhas, quelque poisson qui me veut du mal. J'essaye de chasser cette présence invisible en nageant plus vite, plus fort, mais je la sens me rattraper, je devine un sillage sur l'eau devant moi, peut-être celui d'un petit aileron cartilagineux, qui disparaît à quelques centimètres de moi, tandis qu'une pointe s'enfonce dans ma main. Je hurle, je donne un coup de poing, je sens un corps visqueux s'éloigner. Une douleur m'engourdit le doigt, pourtant je ne saigne pas. J'essaye de me réchauffer, je pèse une tonne. J'ai du mal à remuer, je me laisse aller, dérivant avec le courant. Mes paupières sont lourdes, si lourdes, j'ai envie de me recroqueviller, de dormir sous la couette liquide de ces vagues qui montent et descendent. L'eau gronde et me renverse comme une gaufre trop molle. Elle me prend tout entier, m'emplit les narines, j'ai tellement peur que j'ouvre la bouche pour hurler et je bois la tasse. C'est ainsi que tout commence, la fin, une rasade. Dans un dernier sursaut, mes doigts tentent d'accrocher le vide.
Et se prennent dans des cordages. On me hisse.
— Un sacré gros poisson !
Je crache. Ça brûle.
Je vais mourir.
Des mains tendues me redressent la nuque, leurs doigts au goût de sel s'infiltrent en moi, nouvelle invasion, la demi-conscience comme des chaînes entre le noir et la lumière. Remonte la bile aqueuse, torrent qui déferle de mes entrailles pour se répandre en flaque au fond de la barque.
— Laisse sécher tes rêves, et laisse-nous ramer.
Les transitions sont des moments qu'il convient de décortiquer, ce sont les plus délicats, car dans le dédale qui mène au réveil, ils sont ceux sur lesquels la conscience se focalise le plus. Une nouvelle fois, des formes se peignent sur ma rétine, structures fluides qui prennent parfois l'aspect d'un réel enfui, de silhouettes fugaces. J'essaye d'accommoder ma vision, c'est peut-être ainsi que l'on apprend à rêver : tout seul, en testant ce qui marche. Forcer la forme, la modeler pour lui donner un sens. Le souffle de la création, sculptant le néant en cercles concentriques, infinis qui ne cessent de fondre et de réapparaître. Hypnotisé par ces possibles, je fais des bulles en attendant de les voir se stabiliser. L'une dans l'autre, petite brique par petite brique, construisent une maison, puis un ciel, puis un nid, et des oiseaux, des canards. Imperceptiblement, sur cet océan sombre où rien ne bouge, deux cents macareux tournent leur bec coloré dans ma direction.
J'ai chaud, quelque chose de mou posé sur mon corps, confortable cocon dans lequel je transpire en rivières de sommeil. Saisir le sens, lentement : allongé dans un lit, sous une énorme couverture rapiécée, véritable cabane.
Je me redresse brusquement, aux aguets. Où suis-je ? Une longue chambre sous les combles, plusieurs lits en bois épais, des habits suspendus, une lampe à huile qui veille sur la table de chevet. Je suis en chemise de nuit. Aucune trace de mes moufles. Elles ont dû couler là où je me suis noyé.
Il faut sauter pour descendre du lit. Après quelques étirements, je retrouve ma souplesse coutumière. Mes habits sèchent sur le dossier d'une chaise. Je m'habille en prenant soin de ne pas me froisser un muscle. La pénombre jette des ombres tordues sur les cloisons de bois. La seule fenêtre donne sur la nuit étoilée. Je m'efforce d'y deviner un paysage, mais les ténèbres ont avalé toute trace de relief. Une douleur me lance. La morsure au doigt. On m'a pansé la main, soigneusement. J'écarte les couches du bandage, prêt à affronter mon infirmité.
Un cercle, rompu, à peine irrité, apparaît sur la dernière phalange de mon majeur. Comme une morsure toute ronde, mais sans trace de dents. Comme si un bébé avait tenté de me manger, de ses terribles gencives.
Un bruit sous mes pieds, une horrible tête apparaît sur le plancher. Un instant, je songe à quelque fantôme décapité, mais quand l'énorme morse s'adresse à moi depuis l'ouverture d'une trappe, d'une voix de stentor, je comprends que tout est vrai.
— Kafi !
Le café ressemble étrangement à celui de Fink avec tellement de cassonade qu'il n'y a plus de place pour le café. Le morse y ajoute du lait en marmonnant dans sa moustache un islandais incompréhensible. Je bois silencieusement, en essayant de faire taire les hurlements de mon estomac.
Tout est suspendu dans cette cabane de pêcheur, filets, fils de pêche, vareuses, poissons séchés. Une longue table en bois noir, un plafond trop bas, des fenêtres à croisillons. Le cliché de l'Islande, que je n'avais pourtant jamais encore expérimenté. N'ont-ils jamais évolué depuis le début du siècle ?
Je souffle sur le café chaud.
— Merci, dis-je.
Le morse me fixe sans ciller. Il est si vieux, sa peau est comme un parchemin desséché venu d'un pays antique. Une bise pourrait l'effriter. Sa moustache est un balai poussiéreux, et ses deux immenses défenses ont jauni au point d'épouser la teinte d'un morfil ancien, presque antique. J'étale une confiture noire sur un pain bis, dur comme la pierre. Quand je l'avale, la mélasse a une arrière-goût de sel, de poivre, d'algue et de bonbon. J'ai tellement faim que je pourrai avaler du ciment.
— Mmm... C'est... Bon.
Le morse me regarde, on dirait qu'il n'a pas compris. Je parle mal islandais, mais je reste intelligible à la plupart des autochtones. Serait-il simple d'esprit ? Il lève une grosse nageoire vers mon visage. Caresse lentement mes joues, comme s'il prenait plaisir à éprouver leur douceur, leur jeunesse. Il soupire, puis va à la fenêtre ; en me tournant le dos, il attrape des hameçons emmêlés ; il se parle à lui-même, une série de borborygmes gutturaux où je distingue des mots de norrois, mais aussi peut-être d'espéranto. Il défait la tresse, comme pour en isoler les trous. À chaque hameçon retiré, il pousse un grognement amusé. Au fur et à mesure les mailles disparaissent, pour ne laisser qu'un vaste trou.
— Qui êtes-vous ? dis-je.
La porte s'ouvre brusquement, puis se referme. Je n'ai vu personne entrer, à moins qu'il ne s'agisse d'un nain. Je me penche sous la table. Rien. Quand je retourne à mon bol, un macareux s'est perché sur le tabouret d'en face. Il me scrute attentivement.
— Mon frère, dit le morse. Il est muet.
Le macareux se penche, picore une biscotte, secoue ses plumes.
— Votre frère ?
— Nous n'avons pas la même mère.
Le morse me ressert du café, énorme dans cette pièce si petite. Comment a-t-il fait ne serait-ce que pour passer la porte ?
— Egill a insisté pour qu'on sorte, une chose à ne jamais faire par ici pendant les vives eaux.
Je ne suis pas très sûr de savoir comment m'y prendre.
— Merci.
Le macareux claque-claquette.
— Moi, c'est Hinrik, dit le morse en tendant sa nageoire.
Elle est énorme, rugueuse, pleine de poils. Elle sent le varech moisi.
— Bracken.
— Briki.
— Non, Bracken.
— Briki. Vous n'êtes pas Islandais.
— Frönsku.
— Ah, Paris !
Il exécute une pirouette et sur une minuscule mandoline qu'il décroche du mur, il grattouille une chanson dans un français impeccable. Je n'ai pas revu Paris depuis très longtemps. Je déteste cette ville. Je l'ai quittée pour venir en Islande, refaire ma vie dans un monde qui ne connaissait rien de moi. Paris m'a laissé partir, sans un adieu. Aujourd'hui, sur cette plage du bout du monde, elle me manque terriblement.
Une fois le spectacle terminé, Hinrik ébouriffe affectueusement le macareux. Il continue à me parler français :
— Egill et moi, on n'a plus l'habitude de voir des étrangers par ici. C'est bien de pouvoir discuter. Nous sommes très pauvres, nous n'avons pas de quoi voyager. Mais même si nous le pouvions, je ne pense pas que nous pourrions quitter la plage.
— On est où, ici ?
Hinrik désigne un point, sur la mer, par la fenêtre.
— Tout près de Ginnungagap, au bout de la mer.
Je suppose qu'il parle d'une île. Il me semble pourtant qu'ici toutes les îles ont un nom qui se termine par ey.
— Je dois téléphoner.
— Nous n'avons pas de téléphone.
— Où est le village le plus proche ?
Il hausse les épaules.
— Vous ne voulez pas rester avec nous ?
— Écoutez, Hinrik, je vous remercie vraiment pour votre aide, sans vous je serai peut-être mort, mais je dois vraiment trouver un moyen de contacter mon école. Il... Un ami a disparu.
— Vous venez de l'autre côté.
— De l'autre côté de quoi ?
— De Ginnungagap.
— Qu'est-ce que je fais là ?
— Comment savoir ? Vous êtes un prétendant ?
— À quoi ?
— Vous voulez devenir Maître du Néant ?
— Qui est le Maître du Néant ?
— Celui qui tend le Ptyx.
— C'est quoi, le Ptyx ?
Le vieux morse ferme un œil.
— Dites donc, vous êtes bien sûr que ça va ?
— Non.
Il sourit ; deux défenses énormes fendent sa moustache.
L'espace entre deux moments présents est un long souffle qui tend vers le zéro, approximativement replié sur lui-même par une force qui appuie fort, très fort. Ce sont des petits paquets de perception mis les uns derrière les autres et dont les frontières s'effacent au fur et à mesure que nous avançons vers le grand Rien. Prendre conscience de ces choses a beau être un exercice pénible, peu gratifiant, il n'en est pas moins nécessaire. Comme l'agonie qui précède la mort, il s'agit une condition non négociable. C'est ce qui en fait la cruauté, et ce qui nous permet d'en tirer quelque maigre profit : comme découvrir la respiration profonde, qui monte, fait une pause, et descend. Cet intervalle, qui n'existe que parce que nous prenons le temps de l'envisager, c'est le fondement de l'interdépendance. Ce qui permet à l'amour d'être lié à la haine, la lumière aux ténèbres. Si toute chose est duale, si chaque point de passage existe par les surfaces qui le précèdent, alors l'intervalle mérite de prendre forme pour nous permettre, nous, simples âmes, de connaître l'importance de ce soupir imparfait qui s'étire entre deux souffles.
Hinrik et Egill se sont assis l'un en face de l'autre, les pieds bien tanqués sur le sol. Nageoires dans les ailes, ils vont et viennent l'un vers l'autre, se tirant à tour de rôle, sans bouger. Egill émet un murmure, presque un gémissement, tandis qu'Hinrik raconte d'une voix tonitruante. À chacune de ses paroles, la luminosité de la cabane varie. Parfois, je peux entendre la voix d'Egill qui se mêle à celle de son frère à plumes, n'existant que pour la soutenir. Une voix d'enfant, qui peine à s'exprimer au-delà des mots.
En travers de Ginnungagap,
Le Maître vient au Néant
Abolir le Ptyx.
Mortes ou vives eaux,
Le Maître défait le Kor ;
Il joue la mélodie du Vide.
Noir, séparant les mondes,
Pleine est l'enceinte,
D'où il tire l'essence.
Au sommet du Cap,
Il ordonne les lignes,
Qui disent le mot.
Qui saura l'abolir ?
Remplaçant la Pénultième
Pour excuser la tortue.
Ils s'arrêtent. S'étirent en détendant leurs bras. Hinrik se ressert du café, doucement, sans me regarder. Le macareux hoche la tête au moindre de ses mouvements. Leur connivence pourrait déplacer les montagnes, c'est bien ce qui m'apparaît, dans son élégante simplicité : l'empire des mots et des mythes sur la matière, ordonnant le chaos, créant du sens là où il n'y avait rien.
Je cligne des yeux.
— Je n'y comprends rien.
— Vous n'êtes pas d'ici.
— Vous avez déjà vu le Maître ?
— Oh, oui. Tous les Maîtres passent par ici. Personne ne réussit jamais, à part le Pénultième et le dernier. Le Pénultième a eu le temps de nous apprendre sa langue. Le dernier, lui, il a très peu vécu avec nous, quelques jours seulement, c'était au début des mortes eaux. Puis il est parti pour le Cap, afin de voler le Ptyx.
Je me ressers un café.
— Et qu'est-ce qu'il est devenu ?
— Un jour, les nuages ont cessé de bouger au-dessus du Cap, et même l'oracle est resté muet. Le Néant est demeuré tel qu'il a toujours été, quand il n'y avait pas de Maître.
— D'où venait-il ?
— De la mer.
— Qu'est-ce qu'il a fait ici, ce Maître ?
Du manche de sa pipe, Hinrik désigne un dessin sur le mur. Je ne l'avais pas remarqué jusqu'ici, tant son évidence crevait les yeux. Une fente noire, sur fond blanc, reposant sur un mince socle.
— Le Maître a dessiné ça ? dis-je, abasourdi.
— Oui.
— Elliot ?
Hinrik écarquille grand les yeux. Le macareux couine.
— Vous connaissez le dernier Maître ?
— Elliot est le Maître du Néant ?
Ses grands yeux tristes versent une larme de crocodile.
— Un piètre Maître, hélas...
Elliot ! Elliot est vivant ! Vivant ! En moi explose une joie, un rythme doux, du lait onctueux, qui me remplit, m'envahit. Elliot est vivant ! Je vais peut-être pouvoir lui donner ses pilules. Il ne va pas mourir. Je vais le ramener. Mais... mais comment est-ce possible ? Elliot était encore à l'école hier soir... Et moi, comment suis-je arrivé ici, sur cette côte ? Mmmm...
Par la fenêtre, les ténèbres s'éternisent. Alors je lui raconte : l'école, les cours de dessin, la chambre turquoise, le parc, la disparition d'Elliot, le Ptyx, les algues, la marée. Quand j'ai fini, Hinrik fait trois ronds de fumée, en fronçant ses sourcils poivre et sel.
— Si vous êtes un ami du Maître, ça se fête !
Il dégaine une bouteille de tafia. Je ne comprends pas comment Elliot peut être Maître depuis si longtemps. Nous avons constaté sa disparition de sa chambre hier soir. À moins que le temps ne s'écoule pas ici comme de l'autre côté. Mais où suis-je ? Dans ce Néant dont parlait Plouffe, ou dans ce jeu qu'évoquait Bram ? J'ai du mal à me souvenir des détails. Tout s'efface, les visages se mélangent. Il n'y a plus qu'ici et maintenant, ce morse qui me parle, et ce macareux qui me fixe.
Hinrik verse le rhum dans le café, peut-être pour le rendre plus imbuvable encore. Il peine à soulever la bouteille. Il soupire. Le poids des ans cloue sa lourde carcasse, trop lourde pour ce vieux corps tout flasque, étriqué dans un costume serré au cou par un petit nœud papillon. Je devine un tatouage, juste sous la manche relevée de sa chemise : une tortue.
— Parlez-moi de Ginnungagap, dis-je.
— C'est au bout de la mer.
— Vous dites que c'est de là que viennent les Maîtres, mais c'est aussi là-bas qu'ils retournent ?
Hinrik hoche la tête, ses deux défenses lustrées de frais. Il a gravé des signes dans l'ivoire, un tracé délié qui ne rappelle aucune langue de ma connaissance.
— Mais il y a bien un autre endroit où aller, non ?
— Non, Briki.
— C'est quoi, cet oracle dont vous avez parlé ?
— Les cousines d'Egill, de vraies pipelettes. Ce sont les seules à savoir ce qui se passe au Cap. Parfois, nous leur demandons des nouvelles.
Je repense à ce trou, au bout de la plage, aux odeurs enivrantes de merveilles marines, grotte sublime où passer une éternité en contemplation.
— Ce n'est pas compliqué, Briki, le rituel est le même depuis toujours. Les prétendants échouent sur la plage pour trouver l'inspiration. Ils espèrent tous voler le Ptyx. Mais il n'y a jamais eu que deux Maîtres. Le Pénultième et le dernier, à savoir Elliot.
— Le Pénultième ?
— L'avant-dernier.
— Et moi ?
— Vous serez peut-être le prochain, faisant d'Elliot le Pénultième, et du Pénultième l'Antépénultième. Si vous réussissez à lui voler le Ptyx.
— Je ne veux pas être Maître. Je veux juste...
— Vous voulez quoi ?
Je ne me suis jamais posé la question de savoir ce que je désire vraiment. Prof, ça me suffit à traverser la vie sans me poser de questions. Mais peut-être n'ai-je pas été à la hauteur de mon destin. Trop feignant.
— Ce Cap, c'est quoi ?
— Une forteresse, bâtie sur les restes d'un volcan. Du sommet, on peut voir l'ensemble du Néant, la plage, le gouffre. Enfin, c'est ce qu'on dit. Je n'y suis jamais allé. Personne n'en est jamais revenu non plus.
Le macareux saute de son tabouret pour quérir un poisson, qu'il avale en trois coups de bec.
— Il faut que j'aille au Cap.
— Pourquoi donc faire une chose pareille si vous ne voulez pas devenir Maître ?
— Je veux rentrer chez moi.
— Ah ! Jeunesse...
Il a sorti des assiettes et commence à mettre le couvert. Je l'aide, pauvre morse !
— Vous ne pouvez être sûr de rien, Briki. C'est ainsi que va le Néant, depuis toujours, immuable. Je ne me souviens même plus de mon enfance, lorsque j'étais un petit morse et qu'Egill n'était pas encore né. Les visages de nos parents se sont effacés, comme s'ils n'avaient jamais été là. Nous sommes ici, nous ne pouvons pas partir. Il faut se résigner. Le seul moyen de prendre notre destin en main ce serait de devenir Maître, mais nous n'avons rien de poètes.
— Hinrik, je ne peux pas rester ici ! Il faut que je retrouve Elliot et que je le ramène à la maison, dans sa maison !
— Contentez-vous de ce que vous avez, on vit mieux ainsi, croyez-moi... Si vous allez là-bas, il sera impossible de revenir. Et comment comptez-vous y aller seul ? Moi, j'ai un travail. Si vous saviez tout ce qu'il y a à faire, ici ! Tenez, passez-moi les serviettes.
Je lui passe les serviettes. Je me rends compte qu'il met la table pour de nombreuses personnes. De petites personnes. Des enfants ?
— Qu'est-ce que vous faites ?
— Je sers le repas.
— À qui ?
Il lève vers moi de petits yeux fatigués, rongés par le sel.
— À l'âme de la mer.
Un coup sourd ébranle les murs de la cabane. Puis un deuxième. Hinrik ne semble pas y prêter attention. J'ouvre la porte. L'eau est partout. Plus aucune plage, nulle part. Tout est noir sous une pluie d'étoiles éternelles, dans cette caverne intérieure aux frontières invisibles. Sur la véranda, un macareux vient de s'écraser. Je le remets sur pied, la pauvre bête à des spirales dans les yeux, il est tout sonné. Il s'ébroue, claque du bec, puis entre dans la cabane.
— Vous devriez rentrer, dit Hinrik.
— Pourquoi ?
Dans le ciel viennent d'apparaître deux cents macareux, en formation d'atterrissage. Je cours me cacher alors que leur escadrille s'abat sur la cabane. Les pauvres oiseaux ont des ailes trop petites, ils traversent le ciel comme des bolides, incapables de maintenir une assiette leur permettant d'atterrir sans dommage. Ils finissent par entrer en file indienne, titubant sans me voir. Vêtu d'un tablier, Hinrik les fait asseoir, parfois sur la table. Il en déplace quelques-uns, et m'attrape par l'épaule.
— Allez, allez, dit Hinrik en me tirant de sous un meuble.
— Laissez-moi tranquille !
— Vous n'allez quand pas rester là planté à ne rien faire ?
— Et pourquoi pas ?
Je repense à la mansarde, au confort de la couette. À l'odeur enivrante d'une sieste en pleine mer.
Hinrik me fourre un seau dans les mains.
— Au boulot, matelot !
Servir à manger à deux cents macareux est une expérience qui manquait à mon CV, je ne sais pas si j'en sors grandi, mais je suis épuisé. Ils ont d'abord écouté Hinrik jouer du violon, une vieille mélodie islandaise, puis ils ont papoté, comme au restaurant. Nous avons fini par faire le service, circulant difficilement entre les boules de plumes, en cognant parfois, je ne savais pas que ces volatiles étaient si légers. Ils ont tout mangé, la soupe d'algue, les poissons gluants, le porridge, les petits gâteaux amoureusement cuits par Hinrik dans son antique four. Je suis couvert de détritus, d'une mousse étrange, de plusieurs poissons – des cadeaux apparemment, ou un pourboire. Je les raccompagne tous, ils se prélassent sur la véranda, continuant à papoter. Je ne parle pas le macareux. Je ne sais pas de quoi ils peuvent bien discuter, parfois ils rient. Peut-être de moi.
Nous nettoyons la cabane en silence. Je retire du balai un bon kilo de plumes. Le balai me rappelle à la moustache d'Hinrik.
— Vorace, l'âme de la mer.
— Le ventre est vaste.
— Hinrik, dis-je en jetant le torchon. Je ne peux pas rester. Je n'ai rien à faire ici. Je dois rejoindre le Cap, trouver un moyen de rentrer chez moi. Aidez-moi.
— Et Elliot ?
— Je ne sais plus, Hinrik. Je suis là, au bout du monde, à parler avec un morse, je ne sais plus ce que je dois faire. J'aimerais juste que ça s'arrête.
— Vous ne pourrez pas vous rendre là-bas seul.
— Prêtez-moi votre barque.
— Jamais ! Comment je vais faire pour pêcher ?
— Vous êtes un morse ! Et Egill est un oiseau marin !
— Ah, mon ami, je ne nage plus, je suis bien trop gros, je coule. C'est plus simple de ramer. Quant à Egill, le pauvre petit, voyez, ses ailes sont trop petites, et il mange trop. Il n'a pas volé depuis au moins deux bonnes marées...
Le macareux ronfle béatement, son ventre plein de poissons.
— Comment Elliot a-t-il fait pour aller au Cap ? À la nage ?
— Non Briki, ce serait impossible ! Il avait une barque.
— Une barque ? Il l'a construite lui-même ?
— Il est arrivé à marée basse, ça aide, et beaucoup de prétendants viennent accompagnés, parfois par des flottes entières, avec tout un équipement pour faire des relevés afin de trouver le Ptyx, mais aucun ne devient jamais Maître. Elliot est arrivé avec une barque, un simple rectangle de bois. Il a dit, je m'en souviens très bien : « Je l'ai faite avec un trou. » Là, nous sommes à marée haute et, à moins de construire une barque avec le bois de cette cabane, je ne vois pas comment faire.
Je me renfrogne. Déprimé.
— Comment puis-je vous convaincre ?
Hinrik tire de sa pipe d'énormes blobs de gris.
— Briki, nous sommes des gens simples. Nous sommes très pauvres mais nous avons une fonction, nous nourrissons l'âme la mer.
Son regard vagabonde sur les flots.
— Croyez-moi, vous ne voulez pas vraiment aller là-bas. C'est un lieu maudit. Défier le Maître, c'est la pire chose que vous puissiez faire. Je sais que vous voulez rentrer chez vous, mais le seul moyen, c'est de devenir Maître.
Soupirer, résigné. S'il n'y a qu'une seule façon de quitter ces lieux, faut en prendre son parti. J'ai trop fui, trop attendu. C'est ici et maintenant qu'il faut se décider.
— Très bien. Alors je serai le prochain Maître.
Il se décompose, terrifié.
— Briki, vous n'y pensez pas ! Personne n'a jamais réussi !
— Et Elliot ?
— C'est parce que le Pénultième est mort ! Le Ptyx était libre. Bien sûr, il faut le mériter, mais c'est bien plus simple. Tenter de voler le Ptyx à un Maître, c'est la mort assurée ! Ne faites pas l'enfant, restez ici, avec nous. J'ai besoin d'aide.
Soudain réveillé, Egill remonte sur la table. On ne l'avait pas vu depuis le début du repas. Il est visiblement saoul d'avoir trop mangé. Il se met à caqueter hargneusement, contre Hinrik.
— Mais enfin, Egill...
Le macareux ne veut rien entendre. Il commence à taper du bec dans la panse du morse. Hinrik lui donne un coup de patte, l'envoyant bouler dans un coin de la pièce.
— Il n'a pas l'air de votre avis.
— Que sait-il de tout ça ? Il est si jeune...
Le macareux reste à bouder dans son coin.
— Oh, Egill, voyons... ne le prends pas comme ça.
Le morse s'approche du petit volatile. Il lui murmure des choses à l'oreille, ils discutent. J'ouvre la fenêtre. L'air du large entre à grandes bouffées. Tout scintille, tout est illuminé. Je suis ici, maintenant, dans ce monde qui n'existe pas, ou si peu, à peine un trait, une anecdote. Qui sont ces êtres, en qui je remets mon destin ? Comment obtenir une réponse qui ne soit pas une énigme ? Elliot est ici, quelque part sur cette mer, et je dois le trouver, je l'ai promis aux autres, je me le suis promis à moi-même. Je n'ai jamais rien fait de ma vie. J'ai passé mon temps à glander. Je suis un feignant, un tueur de temps. J'ai appris à le voir passer, sans bouger, sans chercher à modifier son cours. Je connais tous ses recoins, toutes ses petites manies. Mais il me connaît aussi et, dans notre danse, nous nous sommes enfermés. J'ai toujours repoussé ce moment où je devrai assumer, enfin, d'être ce que je suis. Et ce que je suis, aujourd'hui, c'est cet homme à la recherche d'un autre, pour aucune autre raison que le sens de la quête. Qu'importe le Ptyx. Qu'importe le Néant. Je respire. Lentement. Mon doigt me lance, je le frotte sans réfléchir, j'en tire une douleur et un plaisir mêlés. Une nouvelle force bat dans mon cœur. Peut-être le temps dans le Néant est-il différent du temps réel. La mort ici n'exerce pas d'empire. Je comprends mieux ce qu'a fait Elliot : pour échapper à la mort, il est venu jusqu'ici, au bord de son imagination, pour y régner éternellement. Il a brandi la source de son pouvoir pour exister à jamais dans cet espace entre la vie et la mort. Je ne me l'explique pas, pas plus que je ne sais comment il s'y est pris pour franchir toutes les couches de la tapisserie de sa chambre et se réfugier dans un repli de son enfance. Tout ceci est physique, ça n'a rien de psychanalytique. Il ne s'agit pas d'une métaphore, d'une abstraction.
Dans le salon, le morse s'est assis, épuisé.
— Ah... jeunesse.
Il soupire. Un énorme soupir d'énorme morse.
— Très bien, nous partirons pour le Cap.
Le macareux bondit sur la table, se met à caqueter.
— Il vous a convaincu ?
— Je crois qu'il a trop bu, ou que votre présence le rend nerveux. Il veut partir. Il croit en vos chances. Il pense que vous serez le nouveau Maître, celui qui vient pour prendre le Ptyx. Même si vous ne le souhaitez pas, vous allez devoir prouver votre valeur et, qui sait, quelque chose de bon en sortira peut-être. Si Egill croit en vous, pour une raison dont j'ignore tout, je ne peux rien pour le retenir. Et moi, je n'ai rien à lui offrir. Je ne suis qu'un pauvre morse. Et je suis égoïste, en voulant vous garder pour moi. Vous m'êtes si sympathique... Je ne peux pas vous retenir. Moi, je ne veux pas rester tout seul ici. Peut-être que si je pars avec lui, je ne serai pas si malheureux. Je ne suis qu'un vieux morse.
— Hinrik, je suis désolé.
— Ce n'est pas votre faute.
Il se lève.
— Venez.
Nous sortons sur la terrasse. Hinrik me désigne un point, au loin. Un phare apparaît, disparaît dans la houle. À son sommet, jaillit parfois un pinceau lumineux qui balaie le paysage avec délicatesse.
— Le Cap.
— Il a l'air si proche.
— Il n'est pas si loin, là où la mer devient courbe.
— Quand partons-nous ?
Hinrik m'explique qu'à marée haute les vagues vont grandir et grandir, qu'elles vont former un mur infranchissable, que sa modeste barque pourrait bien y rester, et nous avec.
— Quand, alors ?
— Il faudra attendre la marée basse.
— Combien de temps ?
Il compte sur ses nageoires boudinées, rongées par le sel.
— Si tout va bien, peut-être dix.
— Dix heures ?
— Dix ans.
Quelque chose d'immense bascule en moi, de lourd, d'infiniment douloureux. Je sais que je suis coincé ici, sur ce bout de sable au bout de nulle part. Je sais que ma seule chance de m'en sortir, c'est de rejoindre cette tour, là-bas, quelque part.
— Je ne peux pas attendre dix ans, Hinrik.
— Si vous voulez devenir Maître, il le faudra.
Dressé sur ses pattes, le macareux me regarde.
Il déploie ses ailes pour me faire un câlin.
# LE NÉANT (3)
Des lèvres se sont formées sur ce visage aveugle et sourd. Une fente, qui ne s'infecte pas, qui prend du dehors et fait sortir du dedans. Ce double mouvement, réflexe plus qu'habitude, devient respiration, et dans sa délicate altération, modifie l'espace alentour. Propagée, l'onde vibratoire, à chaque goulée, devient cercle concentrique, créant progressivement des cordes, des filaments qui s'enroulent les uns dans les autres pour former des nœuds, puis des pelotes. Denses, ces boules forment une ronde, entraînées par le courant d'un remous qui les aspire. Baignées par l'humidité, elle-même engendrée par l'incessant va-et-vient du rien sur le vide, chaque pelote se déroule suivant un ballet orchestré par les bonds subtils de chaque anneau. Dédoublées, puis triplées, ces formations d'arabesques se nouent en nouvelles formes, longues et sensuelles embrassades, autour desquelles circulent des gouttes, des constellations transparentes, réfléchissant la lumière venue de derrière la fente. Prismes, ils irisent les nouvelles surfaces aménagées entre les espaces de chaque tresse. Tous ces miroirs se renvoient la même image, à l'infini, le regard de l'un perdu dans la pupille de l'autre. De nouveaux plans se multiplient, s'encastrent et s'imbriquent en dédales, en couloirs. Échafaudés, architecture cristalline nourrie d'un liquide coagulé, ces châteaux se dressent vers un ailleurs incomplet, où se nichent déjà des astres, d'autres séries de spirales, de tournoiements infinis, échos des uns, vertige des autres. Tout, dans cet ensemble sacré, vibrant d'un même son venu de l'origine, incarne le principe éternel d'un même chiffre, d'un même mot, d'une même tonalité. C'est ainsi que naît la matrice, c'est ainsi que naît, dans le bain, dans la soupe, les possibles constructions où s'érigeront les sens, les directions, puis les chemins, les sentiers, les terriers qui mènent au plus profond du réel, où tout se délite en fragile dentelle.
Tohu wa-bohu
KOR
Mmmmh...
Retour, éternel retour,
Revenir aux couleurs qui se fendent.
— Bracken ? Vous m'entendez ?
Ce filet de voix me tire du vide où je dérive.
— Gniii...
— Vivant ! Il est vivant !
Fink m'aide à me redresser.
— On a bien cru vous avoir perdu, mon petit !
En clignant des yeux, je retrouve la chambre, composée. Tout a changé. Les meubles sont renversés, placés différemment. Il y a des algues partout, de nouveaux objets entassés, des fragments de bateaux, des frigos, des flaques. La pièce semble avoir rétréci. Ce n'est plus une chambre, c'est un couloir.
— Qu'est-ce qui s'est passé ?
— Vous avez glissé...
— Sur quoi ?
Cyldrid marche difficilement dans le chaos, ses talons l'empêchant de rester debout trop longtemps. Elle s'accroche à une étagère brisée. Son chignon s'est défait, ses cheveux sont couverts d'algues. Plouffe est allongé sur le sol, les jambes en accent circonflexe. Sa chaise est encastrée contre le mur. Toute la chambre pulse d'une lueur turquoise.
— Je ne sais pas, dit Plouffe, comme si quelque chose était venu du Néant, avec tous ces objets. Puis était reparti.
— J'ai vu...
— Bracken, qu'avez-vous vu ?
J'essaye de sonder au plus profond de moi. Là où je ne vais que lorsque j'ai mal, quand je suis incapable de trouver des réponses à ma douleur. Un endroit où je m'enferme pour être jeune à jamais.
— Une fente.
Je m'étire, je fais tout craquer dans ma nuque, dans mon dos. J'ai l'impression d'avoir été tassé dans une toute petite boîte. J'ai besoin d'étendre les bras, de sentir que j'occupe un espace. Là, devant cet angle infranchissable, j'ai besoin de sentir mon corps.
— Attendez...
Tout le monde se regarde.
— Où est passé Bram ?
KOR
Il est en nous maintenant,
Il parle au nom de nous,
Nous sommes lui qui parle.
Nous fouillons toute la pièce. Nous avons déjà perdu Elliot, et maintenant c'est le tour de Bram. Je remue les amas de bric et de broc qui entravent ma progression. Le trou d'Elliot est rempli de coquillages, de fossiles, d'écailles. Le couloir se termine sur le mur par lequel nous avons tenté de passer le Ptyx. Je prends soin de ne pas aller tout au fond, là où l'angle diminue.
Plouffe gémit.
— Il était là, avec vous, Bracken, quand vous avez glissé... Nous ne vous avons jamais perdu de vue.
— Je n'ai pas disparu ?
— Disparu ?
Les autres font non de la tête.
— Mais Bram, oui ?
— Et bien, il n'est plus là.
— Comment aurait-il pu faire pour passer ?
— Passer où ?
— Là où je suis allé.
— Où êtes-vous allé, Bracken ?
C'est vrai, où suis-je allé ? Nulle part. J'ai glissé dans le grand angle et puis j'ai vu... la fente. Le gouffre. Juste un tourbillon, qui s'ouvre pour m'avaler.
— Dans le chas de l'aiguille.
Ils me regardent tous, attentifs. Soudain, je les vois clairement délimités dans le réel, leurs silhouettes noires si vives détachées sur les couleurs du chaos.
— Le point d'origine, dis-je, presque par automatisme, essayant de récolter les fragments de sens. Le siphon d'où filtre le Néant pour détourer notre monde.
Plouffe hoche la tête. Je sais qu'il comprend.
— Les mythes lui donnent un nom. Ginnungagap. Khaos. Mais le Khaos n'a pas de nom. Le Khaos n'est pas. C'est la séparation, le Ptyx. C'est une somme et rien simultanément. Tout compressé en une ligne qui devient point.
— Et pourquoi ai-je vu le Khaos ?
— Parce que vous avez traversé la ligne du mur, mais vous n'êtes pas passé de l'autre côté. Vous êtes juste passé en travers d'elle, dans son épaisseur. C'est-à-dire que vous n'êtes allé nulle part. Vous êtes resté ici, avec nous et vous avez retrouvé votre taille normale.
— Mais Bram est passé.
— Il a dû faire quelque chose que vous n'aviez pas prévu. Il a eu le temps de se préparer.
— Grumpf.
— Et les tortues ? demande Fink.
Je ne retrouve pas les tortues. Je crois qu'elles ont rejoint leur monde, d'où Elliot les avait tirées pendant son rituel à Hellisgerdi. Elles étaient sympas, ces tortues. Elles vont me manquer.
Cyldrid tremble, elle croise les bras, frigorifiée.
— Un courant d'air ? s'enquiert Fink en la réchauffant.
— Non... quelque chose est avec nous.
— Ici ?
Elle fronce le nez, en quête d'un discret parfum.
KOR
Oui, nous sommes là,
Par le passage, par le petit trou
Venus reprendre le Ptyx.
Elle regarde autour d'elle, visiblement inquiète. La chambre. Les murs. Pris par nos conversations, nous ne l'avions pas remarqué : toute la dernière couche de la tapisserie a été décollée. Elle ballonne par endroits, flotte à d'autres. Certains pans sont tombés, d'autres s'accrochent encore par lambeaux. Pas de colle, aucun raté, une perfection.
— Ça alors...
— Comme si quelque chose était passé...
— Entre le mur et le papier.
— Rooo...
— Et la colle aurait fondu ?
— Quelque chose d'assez chaud pour faire fondre une colle qui a près d'un siècle, le tout sans endommager le papier ?
Je passe ma main. Rien ne pègue. Tout est sec.
— Nous devrions tout arracher, dit Fink. Et voir ce mur. Après tout, c'est la fin du voyage. Elliot n'a pas pu aller plus loin.
Nous nous mettons au travail. Ces derniers fragments de vie semblaient participer à la structure même des murs, mais nous nous trompions. Il y a de la roche derrière. Personne ne s'y cache, aucun passage, aucun nouveau trou, aucune figure d'Elliot enfouie sous les décombres. Le mystère est entier.
— On pourrait casser le mur.
Fink redresse Plouffe pour le reposer dans son fauteuil.
— Cette pierre n'a pas l'air tendre, mais on va essayer.
— Nous ne retrouverons jamais Elliot...
— Ne soyez pas défaitiste.
Cyldrid couine. Quelque chose en elle a changé, subtilement, dans ses yeux, comme du cristal fondu. Je m'approche d'elle et je l'entoure de mes bras, délicatement.
— Dites-moi.
— Il y a une présence, c'est le diable qui est entré ici. C'est arrivé quand vous êtes passés dans le chas de l'aiguille.
— Mais je ne suis pas passé. Je suis là, avec vous. Plouffe a dit que j'avais juste glissé, n'est-ce pas Plouffe ?
— Certainement.
— Gryla, dit Cyldrid.
— Gryla ? reprends-je. C'est quoi, Gryla ?
Fink lève les yeux au ciel.
— Bracken, nous sommes le six.
— Et ?
— C'est le jour de l'année où Gryla vient enlever ses proies. Gryla, l'ogresse des montagnes. Gryla, le monstre qui hante les rêves de tous les enfants.
— Et Gryla est revenue ?
Cyldrid me regarde étrangement, comme si elle pouvait voir par-delà ma personne. Elle me prend les mains, elle regarde mes moufles, elle les retourne.
— Vos moufles...
— Quoi, mes moufles ?
— Vous ne les enlevez jamais ?
— Non.
Un élément sonne faux dans ce que je viens de dire. À mes propres oreilles. Comme si je me cachais quelque chose à moi-même.
Cyldrid cligne des yeux, rapidement, plusieurs fois. Je crois entendre tinter des verres, des flûtes, des triangles symphoniques.
— Gryla... Elle est venue par le chas, elle est passée entre les murs, et maintenant elle est avec nous.
— Dans l'école ?
Elle ferme les yeux. On entend toujours le même son cristallin.
— Attendez...
Nous nous pressons autour d'elle, comme des hommes des cavernes qui verraient le feu pour la première fois. Sa petite bouche s'ouvre et se ferme comme un minuscule mollusque.
— La terrible Gryla qui vient prendre les enfants, qui fauche des vies pour satisfaire son appétit. C'est notre peur qui la nourrit. Elle descend des montagnes poussée par la faim. Personne ne peut l'arrêter : quand le roi des Elfes était là, il pouvait défendre la communauté, mais depuis sa disparition ses féaux sont livrés à eux-mêmes. Personne n'empêchera Gryla de cuire en ragoût tous les enfants qu'elle aura jugé méchants... Elle porte des cornes et des sabots et son ouïe est si fine qu'elle détecte les mauvaises herbes à travers l'île entière. Personne ne peut l'arrêter.
Elle s'assoit sur la pile de coquillages, épuisée.
— Mais enfin, Cyldrid, dis-je en lui prenant les mains, Gryla est un personnage de conte...
Plouffe pontifie, balzacien.
— Conte dont notre gouvernement a interdit l'usage. Si Gryla existait, elle ne pourrait même pas approcher cette école. Nous sommes sur un terrain gouvernemental...
Elle lève vers lui des yeux pleins d'espoir, du moins à ce qu'il me semble. J'y lis surtout une folie venue de très loin, de très profond, là où la raison n'a plus de prise, comme le nouveau-né réduit à la passive contemplation d'une indéchiffrable agitation aux confins de son champ de vision faite de couleurs trop criardes, de sons trop perçants – manèges, girafes en caoutchouc au couinement terrifiant.
— Cyldrid...
J'essaye de me rappeler ce que j'ai vu, dans l'entre-deux, fugacement pour la convaincre qu'il n'y a pas de quoi avoir peur. Mais je ne me souviens de rien. Comme si un masque oblitérait ma mémoire, comme si une brume s'était levée pour me protéger d'une vérité ultime, de quelque terrible secret enfoui à tout jamais et devant à tout prix le rester.
— Il n'y a pas de danger, Cyldrid, assuré-je.
Elle se tourne vers moi, mais son regard est opaque.
— Aux quatre coins du monde, dit-elle, comme possédée, se répandra le mal venu des interstices. Celui qui détient le Ptyx pourra jouer sur les cordes qui tiennent le monde. L'architecture s'écroule quand sont retirées les lignes qui tracent les arches et les colonnes.
Elle se dresse sur le lit.
— Cyldrid, qu'est-ce que vous faites ?
Ses deux bras se replient autour de son torse.
— La vieille dame est arrivée.
Elle danse, désarticulée.
— J'ai vu l'ombre, l'ombre sortir et danser entre les zones, celle qui ôte la différence et résout les paradoxes en les repliant.
Elle impose les mains sur ses paupières. Nous sommes trop terrifiés pour faire quoi que ce soit. Tout se transforme en cauchemar, je n'ai plus de prise sur ce rêve, il va partir en vrille.
Cyldrid tend à présent les bras.
— Voyez, mes précieux.
Dans ses paumes, quelque chose brille.
Fink me retient.
— Mon garçon, attendez.
— Ses yeux... regardez ses yeux... dit Plouffe.
Ses yeux ne sont plus que deux globes blancs. Sur le moment, je me dis qu'elle les a révulsés, comme ces sorcières dans les mauvais films qui passent tard le soir. Mais je comprends très vite que rien de connu ne pourra expliquer ce qui se passe en ce moment même dans cette pièce, dans le sous-sol d'une école au bord de l'effondrement collectif.
— Cyldrid, donnez-moi ça.
Elle fixe le plafond, hagarde, mais retourne ses paumes sur les miennes. Je sens un contact froid, un peu comme du verre. Puis elle s'effondre, comme un château de cartes sous l'action du vent. Sur le lit, Fink s'est rué pour s'occuper d'elle, il essaye de contenir ses spasmes. Je n'arrive pas à y croire. Là. Dans le creux de ma main, me fixent par-delà le voile d'un impossible gouffre, non pas, comme je le croyais au début, ses verres de contact, non, mais bien ses rétines.
KOR
Nous savons comment voir les lignes,
Nous nageons sur ces vibrations,
Nous pouvons séparer ce qui était attaché.
Deux rétines, parfaitement décollées. Transparentes et conservant toute la splendeur des yeux qui jadis les portaient. Comment est-ce possible, comment décoller une rétine de son globe oculaire ? Un frisson terrible me ravage la colonne vertébrale. J'étouffe un cri d'angoisse, que j'essaye de partager muettement avec Plouffe, lequel m'interroge du regard. Je lui montre ma paume. Il secoue la tête, incapable de comprendre ou d'admettre de quoi il retourne. Plouffe tente d'approcher un doigt, le retire aussitôt.
— Bracken, quelle horreur...
— Quelque chose est sorti du Néant, dit Plouffe. Quelque chose de terrible.
— Qu'allons-nous faire ?
Soudain, au-dessus de nos têtes, on entend un raffut, une cavalcade.
— Oh lala, gémit Fink.
Plouffe devient blême.
— Les enfants !
On ne résout pas des paradoxes en évitant de les formuler. Ne sachant pas si à l'instant présent je me souviens de tout, je peux essayer de combler les trous pour modérer la part la plus improbable. Il s'est passé tant de choses, entre les moments où je devais saisir la portée de chacun de mes gestes. Oui, c'est un paradoxe, mais un paradoxe est aussi une forme de poésie, je suppose, et c'est bien de cela qu'il s'agit : comment je me suis levé pour devenir poète, moi, celui qui n'a rien entre les mains.
Hinrik, où es-tu né ?
Dans la mer salée.
As-tu essayé de partir ?
Non, jamais.
Pourquoi ?
Quelqu'un doit nourrir la mer.
Hinrik n'avait pas menti. Il fallut dix ans à la marée pour redescendre. Dix ans... Toute une vie. Il ne se passe pas grand-chose dans le Néant, il ne s'y passe pour ainsi dire rien. Si sa mixture d'algues et de crabes ne cessait de se modifier, l'inéluctable marche du temps semblait retenue par un jet d'inertie. Une grille, peut-être, enfermant toute pensée, toute géographie. Ma première année fut un supplice. Comme lorsque je dors, que je compte ces secondes de vide qui s'étiolent, je suis resté dans la cabane un an durant sans en sortir, parfois debout, parfois assis, à attendre que quelque chose se passe. Juste quelque chose. Immobile à la fenêtre, à regarder l'eau danser doucement, clapotant juste sous notre porte. Hinrik entrait et sortait, revenait de la pêche, contant des mythes et des histoires de morse. Hinrik me racontait comment le Néant était né, d'une boule de rien où tout le chaos fusionnait sans retenue. Comment l'érosion avait créé des ouvertures, comment de ces ouvertures avait filtré le possible. Dans cet esquif immobile, nous jouions aux questions, sans pour autant que je croie quoi que ce soit de ce qu'il me disait. Après tout, c'était juste un morse. J'ai compté les jours sur un morceau de bois, et chaque marque formait comme une ride supplémentaire sur mon visage, comme une cicatrice à moitié effacée par le temps.
Hinrik, qui a créé le Néant ?
Le Néant existe depuis toujours.
Qui a créé cette plage ?
Le Pénultième.
Hinrik, qui est le Pénultième ?
L'avant-dernier, qui invoqua la belle tortue.
Je compris que la subtile mythologie du Néant se trouvait précisément dans ce qui n'était jamais dit : si tout était répété, les sons créaient malgré tout un appel d'air entre eux, comme des cases vides qu'il fallait remplir avec soi-même, comblant les trous de nouveaux principes, connus de moi seul, relevant de la sphère intime. Il n'en restait pas grand-chose, à part ce qu'ils avaient appris à chanter, avec d'infinies modulations. C'était là la clé : quand tout se répète inlassablement, le plus petit changement, le moindre son nouveau devient un trésor précieux. En somme : un mythe intime. J'ai lu tous les livres de la bibliothèque, les sagas essentiellement, j'ai beaucoup progressé dans leur langue. Egill m'aidait en me signalant mes erreurs d'un coup de bec sur la main. Je développai une forme d'amitié avec le volatile, rapidement plus qu'avec le morse, qui restait un mystère opaque refermé sur un secret hermétique. Egill dormait en boule sur la couette, tout en gardant sur moi un œil curieux. Il était ma seule source de réconfort, véritable peluche vivante dans toute cette régression infantile. Rien ne pouvait me tirer de cette éternité de contemplation, à mes yeux, le moindre grain de poussière devenait monde. Le sommeil et l'éveil finirent par former une bouillie de sens, sans commencement, ni fin. Comme, j'imagine, on l'éprouve en prison, la résignation à compter les années devint un défi.
Hinrik, qu'est-ce que le Ptyx ?
L'instrument pour changer le Néant.
D'où vient le Ptyx ?
De la belle tortue.
D'où vient la belle tortue ?
Le Pénultième l'a invoquée.
Le temps passant, j'éprouvai le besoin de quitter la mansarde, d'explorer ce monde stérile autour de moi. Au début hésitant, Hinrik m'apprit à pêcher les crabes du Néant. Nous sortions tous les jours, au milieu des hautes vagues noires, véritables cathédrales. Nous devions lâcher nos caisses, nos filets, et attendre. Parfois Hinrik plongeait, énorme morse à la recherche de quelque proie. Nous restions des heures en mer, c'était difficile, il fallait ramer, se stabiliser, braver le froid, la pluie, parfois la grêle. Une nuit perpétuelle nous enveloppait. Revenus à la cabane, nous mangions de la soupe d'algues, quelques coquillages, et tous les jours du crabe. Les conques étaient étonnantes, car je n'arrivais pas à les voir totalement, comme si leur coquille spiralait à l'infini. Les énormes mollusques, eux, s'ouvraient comme des feuilles, couche après couche de carapace molle qu'il fallait peler pour atteindre le cœur, une sphère immense remplie de vide, parfois impossible à manger dans son absence. Ne restaient sur la table que d'improbables squelettes aux incompréhensibles labyrinthes, que nous transformions pour égayer nos soirées en jeux de société aux règles ésotériques compréhensibles de nous seuls. Comment s'extraire d'un dédale en un minimum de coups de dés, dés au reste taillés dans le bois de la cabane. Le jeu nous aidait à tuer le temps, à attendre que dehors la mer daigne se retirer pour nous laisser approcher l'horizon.
Comment le Pénultième a-t-il invoqué la tortue ?
Il a changé le Néant.
Mais ne faut-il pas le Ptyx pour changer le Néant ?
C'est la condition.
Si la tortue donne le Ptyx, qui donne la tortue ?
Le Mystère.
Je me consolais en me disant que vieillir n'était pas la pire des choses. Je mûrissais, je dessinais tous les jours, et ma barbe devenait forêt. Je me regardais dans le miroir fendu de la mansarde, je pensais être encore jeune, beau. Progressivement, pour effacer les traces de l'âge, j'appris à ne voir que les lignes qui composaient mon visage, les fentes qui apparaissaient au coin de mes yeux, de ma bouche. La peau qui, légèrement, perd sa densité pour devenir plus élastique, presque flasque. Je n'avais pas peur. Je savais que si je me concentrais, je pourrais rester dans cet état de maturité toute ma vie. La mort n'était pas une option. Elle ne l'avait jamais été. Dans le salon, je regardais cette peinture au mur, ce socle et cette fente, et je me demandais : le Ptyx est-il la fente ou le socle ? Le Ptyx est-il l'objet ou le passage ? Cette abstraction exprimait une nouveauté dont je devais trouver l'écho quelque part en moi. Je devais avant tout organiser une pensée du Néant. Peut-être Hinrik disait-il vrai, quand il affirmait qu'on ne pouvait le traverser. Que le Néant était une fin en soi, qu'il n'y avait pas d'autre côté, autre que celui duquel on était venu. Je devais donc accepter que je fusse arrivé au fond de la bouteille, et que le seul moyen d'en sortir était de m'armer de patience. J'avais décidé d'envisager que le Néant fût comme ces traces sur le sable, quand le ressac se relâche, lignes sur lignes, couche après couche. Parfois, quand l'eau était moins froide, je me baignais, juste devant la cabane, et je regardais dans le sel trouble comment le sable formait des petites dunes au fond, sculptées par le remous. Toutes ces lignes me confortaient dans mon interprétation : le Néant était un feuilleté tendant vers l'infini, un simple escalier éternel, où chaque marche était un pas de plus vers l'infime, sans jamais changer de palier. Ou bien étais-je coincé à un palier ? Comment savoir si une réalité plus grande ne m'attendait pas derrière ce trou, dans la paroi. L'espoir est humain, et c'est peut-être là-dessus que nous avons bâti nos conceptions spirituelles : sur la capacité de l'humain à rendre cette réalité bien plus intéressante qu'elle ne l'est réellement. Devais-je donc amplifier ce que je vivais, coincé dans cette cabane du bout du monde, pour en faire une symphonie des sens ? Moi, qui avais passé ma vie à ne rien vouloir subir, à ne rien vouloir vénérer sous peine de devoir m'engager dans une voie ou dans l'autre, je devais contempler ce que j'avais sous les yeux, l'apprécier à sa juste valeur et en faire un royaume.
Hinrik, comment voler le Ptyx ?
Il faut vaincre le Kor.
Hinrik, c'est quoi, le Kor ?
Un monstre ancien.
Comment vaincre le Kor ?
En volant le Ptyx.
Le dessin devint une hygiène quotidienne de survie. Comme si tous les traits que je pouvais trouver sur une page blanche étaient l'assurance que je finirai par me rendre Maître de ce territoire. Au Néant, ma vision radioscopique s'était décuplée : tous les objets, toutes les silhouettes, se superposaient au réel comme des décalcomanies. Tout acquérait une qualité visuelle proche du dessin animé, de la bande dessinée : en mouvement, en séquence continue, dénuée de discrétion. Moi qui avais toujours envisagé le dessin comme des plages de couleurs créant des surfaces, envisager le trait comme primordial me demandait un effort de coordination, facilité par l'omniprésence du découpage qui se dépliait sous mes yeux. Il me suffisait de reprendre les mesures. La marée règle le Néant comme du papier-musique, ses lois sont cruelles pour qui ne sait pas lire entre ces lignes de vertige. J'ai appris à tracer tous les traits qui pouvaient m'unir à la simple figure géométrique du Néant, cette pure ligne infinie, la mer, le ciel, les déplacements de nos gestes, signifiés dans le réel par des traits censés rendre la vitesse. Ce ne fut pas facile, car j'avais perdu l'habitude de dessiner pour moi-même. Découvrir une perspective dans un espace froid était plus enthousiasmant que tout. J'appris à regarder la page comme du marbre à tailler, où les pleins et les vides s'enlaçaient pour créer de nouveaux motifs, des espaces impossibles. Parfois, je commençais un dessin en me disant que c'était un objet bien précis, pour finalement me rendre compte que c'était tout autre chose. Je me découvris une passion pour l'improvisation. Quand Hinrik me demandait de le dessiner, je lui disais que cela m'était impossible, mais peut-être qu'en dessinant la fenêtre, je finirai par le trouver, lui, dans cet enchevêtrement de courbes, de lignes et de points. Et ce fut bientôt le monde que je regardais à travers ce prisme nouveau.
Comment voler le Ptyx si on ne peux le voler qu'en l'ayant volé ?
Seul le Kor connaît la réponse.
Hinrik, le Néant sera-t-il toujours ainsi ?
Il ne peut changer que si le Maître le veut.
Hinrik, pourquoi le Maître ne change-t-il rien ?
On ne remplit pas le Vide avec du Rien.
J'appris à me nourrir de détails, à regarder chaque vaguelette comme un monde en soi, similaire et pourtant différent. À célébrer le plus infime changement de luminosité comme un événement sans précédent. Le rien devenait tout et à chaque crabe ouvert, à chaque algue entamée, à chaque sortie en mer, à chaque partie de donjons, à chaque câlin d'Egill, je me retrouvais à réagir d'un frisson nouveau. Tout devint aventure. L'écho virtuel de ce monde inter marée, que j'avais exploré à mon arrivée sur la plage me confortait dans mon idée que ce mini-univers était porteur d'une infinité de possibles, qu'il me faudrait toute une existence pour en saisir les subtilités. C'était lui que je voulais faire vivre sous mon crayon. C'était lui que je souhaitais célébrer ici, perdu sur ce radeau minéral. Bientôt, je me levais le matin, frais, heureux, presque rassuré d'être ici, sans avoir à choisir où me rendre, sachant que nous sortirions peut-être affronter les vagues, rire sous le crachin, respirer l'iode ou regarder les tourbillons entraîner l'écume en farandoles. Hinrik se moquait, il riait de me voir apprécier sous un jour nouveau cette vie qui avait été la sienne pendant tant d'années, et je crois qu'il était fier que je puisse un jour prendre sa succession, dans cette modeste cabane, au milieu des filets et des pattes de crabes. Je me nourrissais de ce mets somme toute raffiné. Je n'avais plus envie de me rendre au Cap, que je voyais disparaître, petite lumière sous les vagues. Les souvenirs de l'école s'enfuyaient, Elliot, Fink, Plouffe devenaient des figures anonymes dans les fractales de vagues. Je trouvais un souffle, une source, dans cette répétition. Un contact direct avec mon enfance, mes amis imaginaires. J'y trouvais la patience, une certaine forme de sagesse. J'interrompis le décompte des jours et la vie redevint ce long ruban déroulé en douceur, dans la joie des matins qui se répètent. Pour la première fois, je trouvai la paix.
Puis un matin, je me suis levé, la lumière dans la chambre avait changé imperceptiblement. Egill était dressé sur le lit, la plume froissée. Il éternua. Je sus immédiatement que la mer était basse.
KOR
Ce n'est pas ce jeu qui te peine,
De n'être que passif, tu souffres,
Et dans le solaire, tu renonces.
Anoraks fluorescents, moufles et bonnets rebondissent dans les couloirs au son d'un incessant bourdon.
— Oh lalala, fait Fink.
Nous avons laissé Cyldrid aux bons soins de Plouffe. J'ai posé ses rétines sur le bureau, nous avons décidé de gérer la situation dans la mesure de nos moyens. Fink avait l'air de vouloir s'activer à tout prix, terrifié par ce que nous venions de vivre. Il se tord les mains. Il veut se changer les idées, participer à quelque chose sur lequel il peut avoir un semblant de prise. Sans aide, nous pourrions être très vite dépassés, car personne ne semblait s'être déplacé pour encadrer la kermesse.
— Fink, on ne va jamais y arriver... Où sont tous les autres instituteurs ?
— Ils ne sont jamais rentrés de vacances.
— Il n'y a que nous ici pour gérer cet enfer ?
— Il faudrait retrouver Bram.
Je regarde partout, comme si Bram pouvait apparaître d'un coup, là, providentiellement, mais je ne vois que ces dessins sur les murs, ces couleurs qui explosent sous mes yeux, comme libérées de la page. Étrange, on les dirait changeants, mouvants, ces dessins, comme si le fusain fondait, ou bien... Je m'approche, le nez collé à la feuille, presque comme si j'étais myope ; je pourrais jurer avoir vu cette maison s'enfoncer dans une forêt, et ces nuages coulisser vers l'est, alors que la fumée de la cheminée serpente jusqu'à...
— Bracken, gare ! hurle Fink.
Les enfants courent partout, suivis par une foule de parents, d'étudiants, des gens venus de la capitale. Tout le monde est sur la falaise – par extension, les terrasses de l'école. Nous sommes noyés dans la masse, emportés par la vague dans le couloir. Je me souviens de ces matins, lorsque j'enseignais encore, quand la vue de ces visages rosissants m'emplissait d'une joie infinie que rien ne pouvait ternir. La voici, la vie perpétuelle, la fontaine de jouvence. Ce sont eux avec leurs jeux, leurs mots si doux, à peine compris, prononcés comme des vœux. Cette liberté, cette improvisation. Tout fait sens en si peu de mots.
Une grappe d'enfants se pressent autour de moi, certains m'attrapent la jambe, d'autres me tirent les mains.
— Bracken ! Bracken est revenu !
— Salut les kids !
Prissy me fait un énorme câlin. Elle m'a manqué, cette petite. Elle était la plus brillante de toutes : passionnée par le hula-hoop et le chewing-gum, pouvant remuer les hanches et mâcher en même temps. J'ai à peine le temps de former son prénom sur mes lèvres que la marée nous emporte à nouveau, plus dense au point que j'en ai le souffle coupé. Tout autour, les enfants se déguisent les uns les autres, tirent des objets de leurs casiers, occupant le moindre espace disponible, empêchant toute traversée. Nous sommes de grands adultes, nous pourrions nous tailler un chemin, mais à quel prix ? Ne pas les blesser, ne pas les brusquer.
Nous devons nous laisser porter.
Tous ensemble, ils chantent :
— GRYLA ! GRYLA !
Gryla... La vision de Cyldrid. À moitié enseveli sous les mains des enfants qui tirent sur mes moufles, je m'accroche à Fink, qui avance en diagonale. Nous sommes vite acculés contre un mur, tout aplatis.
— Je croyais qu'on ne fêtait plus Gryla !
— Tous ces films d'horreurs américains ont semé la zizanie !
— Et qu'a fait Gryla pendant tout ce temps ?
— Elle était dans sa grotte, dans les montagnes, à râler.
— Elle doit avoir très faim.
La masse reflue, des dizaines de petits bras hystériques qui nous happent. Il y a de fausses antennes qui dodelinent sur leurs ressorts, des chapeaux pointus, des ailes coloriées au feutre, des sarbacanes.
— Bracken, accrochez-vous !
La foule nous entraîne vers l'extérieur, le long des couloirs décorés, des salles ouvertes sur des bureaux abandonnés, des tableaux immaculés. Une faible lueur troue la nuit, cette nuit qui ne quitte jamais ce ciel, ou alors juste pour une heure, à midi... Si les nuages sont cléments, nous pourrons voir le disque solaire iriser l'horizon d'un trait d'espoir. Nos souffles froids y tracent des formes singulières, qui disparaissent en s'élevant.
La marée humaine nous laisse échoués sur le parking. Les plus jeunes ont couru gambader entre les rochers de la falaise, jouant à cache-cache, à la recherche de quelque fée. Les plus âgés se bécotent dans les rochers. Les parents, qui ont garé leurs voitures plus bas, se regroupent près des baies vitrées. Fatigués en cette heure matinale. Frigorifiés. Visiblement très inquiets. Ils boivent du café dans des Thermos verts. Une forte poitrine se détache du groupe et nous rejoint d'un pas décidé, martial.
— Môssieur Bracken.
— Erla, dis-je en reconnaissant la mère de Prissy, une valkyrie.
Elle regarde ma moufle tendue comme une sorte de monstruosité.
— Où est la pêche aux canards... Et les stands de chichis ?
— Heu...
— Et la barbe à papa ?
— Eh bien...
— Concernant la chorégraphie de midi, avez-vous prévu une animation ? Et au niveau du son ? On dirait que rien n'a été fait. Où est la kermesse ?
Je regarde autour de moi. Je ne vois que le champ de lave, les enfants qui s'amusent dans la mousse, les parents, sur la terrasse. Je n'oublie pas que Fink m'avait dit que Plouffe avait perdu toute emprise sur l'école. Que personne ne s'occupait plus de rien. Plouffe a laissé les choses aller à vau-l'eau. Si je veux reprendre le contrôle de l'école, il faudra que je fasse preuve d'un peu d'autorité. Les bonnes paroles, fussent-elles sensées, n'y suffiront pas...
— Fink, demande-t-elle en comprenant que je ne l'aiderai pas, où est Plouffe ?
— Il est... aux cabinets.
— L'école s'est engagée à s'occuper des festivités ! C'est Plouffe et monsieur Bram eux-mêmes qui ont...
— Monsieur Bram ? l'interromps-je.
— Il avait pris ses dispositions pour la venue d'un chanteur populaire.
Quelque chose cliquette lentement dans ma tête.
— Nik va arriver d'ici une heure, dit Erla. Vous ne pensez tout de même pas qu'il va chanter sans micro ?
Fink paraît soudain très, très vieux. Il se tourne vers moi, comme si je pouvais trouver instantanément une solution qui arrangerait tout.
— Eh bien, dis-je, nous avons tout ce qu'il faut dans la réserve du gymnase, y compris une petite estrade, nous pouvons tout mettre en place d'ici là, ne vous inquiétez pas. Nous avons quelques soucis cette nuit.
— Des soucis ?
— C'est que nous avons perdu...
Je la regarde et, brusquement, la réalité s'effrite : comme si son visage venait de changer de forme. Très légèrement. Sur la gauche, décalé, mollement.
— Eh bien...
— Bracken ? s'enquiert Fink en se penchant vers moi.
Erla fronce ses sourcils, dans des angles impossibles. Ce n'est pas seulement son visage. Tout son corps vient de se froisser, subtilement, comme bousculé. Mais ce n'est pas son corps, non, c'est la surface, juste les plans, les à-plats. Je ne comprends pas. Je plisse les yeux, et soudain, je discerne quelque chose autour d'elle, une aura.
Sa silhouette. Qui vacille.
— Bracken ?
Je cligne des yeux : sa silhouette retrouve une stabilité.
— Excusez-moi Erla, je fatigue.
Je me pince le nez.
— Fink, voilà la situation : d'ici midi, toute la ville va se presser sur cette terrasse et il n'y aura rien pour les divertir. Nous contribuons tous à cette école, et les parents d'élèves sont une force avec laquelle il faut compter. Vous ne pouvez pas nous snober.
— Nous pouvons mettre tout le monde au travail, propose Fink, rassurant. En attendant, les enfants peuvent s'amuser avec un rien, vous les connaissez, je vais demander à madame Oup de les gérer.
Erla s'éloigne en grognant rejoindre le groupe de parents d'élèves. Je vois son épaisse carcasse osciller de droite à gauche, puis trembloter. Comme les parois d'une amibe dans la soupe primordiale. Je conserve l'image d'une algue sous l'eau, trouble.
Mes yeux. Mal.
— Bracken, tout va bien ?
— Non.
— Qu'est-ce qui se passe ?
— Je ne sais pas. Je ne comprends pas. Je vois...
Fink scrute mon visage, inquiet.
— Depuis que vous avez glissé, vous semblez différent...
Il ferme un œil.
— Êtes-vous bien sûr de ne rien avoir vu, là-bas ?
La plage. Les châteaux, la grotte, trop petite. Et la tour noire au loin, trop loin, qu'on ne peut atteindre qu'en tendant le pouce...
— Non. Rien. Il n'y a rien.
Je ne lui dis rien car je n'ai rien à lui dire. Je ne suis pas sûr d'avoir vu, vécu tout ceci. En moi, quelque part, peut-être y suis-je toujours, à chercher des réponses aux questions que nous, ici, avons renoncé à nous poser. Ainsi m'ont été données, à moi, en moi, les clés pour voir ce monde depuis la perspective du Néant. Cet espace entre les surfaces, tout autour de nous, comme la ligne claire autour de Tintin, comme...
KOR
Nous ne sommes pas tous les carnivores,
En toi, Chevalier, naît la lecture,
De ces mots invisibles qui disent le tout.
La silhouette des gens.
— Je vois le Néant, Fink, partout.
— Le Néant ?
— Comme dans les bandes dessinées. Le trait noir qui cerne les personnages.
— Vous le voyez autour de moi ?
En effet, je peux voir son corps occuper une place définie dans son cocon longiligne. Je croyais que seuls les objets inanimés étaient affectés. Mais non. Toute créature aussi.
— Oui.
— Et autour de vous ?
Je n'y avais pas pensé. Je regarde mon corps. J'ai un trait noir tout autour de moi, qui me situe dans l'espace. Sauf...
— Il n'y a pas de contours sur mes moufles.
Il ouvre grand les yeux, tape dans ses mains noueuses.
— Vous avez acquis la double-vue ! Vous voyez l'autre monde !
— Non Fink, je ne vois rien de plus que ce qui est.
— Vous ne comprenez pas, Bracken : c'est la ligne qui est le monde !
Je retourne mes mains, c'est incompréhensible. Comme si mes moufles formaient une sorte de repoussoir. Mais pourquoi ? Ces moufles existent pourtant. Elles sont un plan, un espace, en contact avec le reste. Elles devraient générer ce Néant dont Plouffe a fait la description.
— Bracken, continue Fink, Plouffe dois vous examiner, il doit savoir ce qui se passe ! Laissez-le faire ses tests, Bracken, vous avez sûrement dû voir quelque chose là-bas, pendant ce moment d'égarement qui vous a fait chanceler devant l'angle. Bracken, vous ne vous souvenez peut-être pas, mais vous avez vécu là-bas, vous savez ce que nous ne savons pas et nous devons savoir !
Bram.
Bram, l'absent.
Bram, qui cherchait à décrypter le jeu d'Elliot. Bram, qui était passé avec moi par le Ptyx. Bram, qui n'est pas là, qui a complètement disparu, alors que tout le monde s'occupait de moi quand j'ai glissé.
— Il faut retourner au gymnase.
KOR
Cours, petit homme, cours
Nous, dans ton sillage,
Comme les cheveux d'une suite.
Il n'y a plus personne dans le gymnase. Tout est abandonné. Les ordinateurs ronronnent, en veille, leurs écrans noirs.
— Où sont-ils tous ? demande Fink.
Je vais regarder dans les vestiaires. Sur les gradins. Personne, nulle part. Il y a des confettis par terre, comme s'ils avaient fêté une victoire.
— Qu'est-ce qui s'est passé ici ?
Les échos de la fête au-dehors nous parviennent étouffés, comme des rythmes venus d'un autre continent. Je me rends compte de la pression qui s'exerce sur nous, sans en saisir la teneur exacte. J'essaye de conjurer mon instinct, de mettre fin à ce statu quo dans mon rapport avec moi-même, mais je ne capte rien, que de la neige, comme les écrans de télévision après la fin des émissions. Il n'y a plus rien en moi, ni souvenir, ni désir.
Je suis simplement là, tout entier.
Je m'assois sur un gradin.
— Qu'est-ce qui m'arrive, Fink ?
Je les regarde, lui et sa silhouette, je regarde le gymnase, tout entier dessiné, enfermant les couleurs dans des surfaces précises.
Fink pose sa main sur mon épaule.
— Vous avez vécu quelque chose, Bracken. Vous êtes passé en travers de la réalité, en suivant Elliot. Vous avez vu ce qu'il y avait là-bas, et ça vous a changé à jamais. Vous avez vu Féerie.
— Mais je ne veux pas changer !
— C'est trop tard, il me semble.
Je ne vais pas me mettre à pleurer, mais une force enfle dans ma gorge, un sanglot étouffé, enroulé sur lui-même. L'envie de me mettre sous ma couette, immobile, de rester bien au chaud.
— Mais je ne suis allé nulle part. Vous l'avez dit, j'ai juste glissé.
— Essayez de vous souvenir, Bracken. Il y a forcément une partie de vous-même qui a recueilli un enseignement. Il doit rester, dans vos os, dans votre cœur, dans vos réflexes, une trace de ce que vous avez vu. Cherchez.
Je ferme les yeux, pour éviter toutes ces lignes qui définissent le monde. Elles s'impriment en négatif sur ma rétine puis disparaissent, petite grille mal définie. Quand j'ouvre à nouveau mes paupières, une trace blanche s'efface doucement, un cercle à moitié fermé, presque tracé par un pinceau.
KOR
Doux petit oiseau revenu du silence,
Si tu savais les choses qui échappent
Au sens de la marche.
Fink me passe la main sur le front, souriant faiblement.
— Essayons de comprendre comment ces enfants ont pu quitter cet endroit...
— Deux pièces fermées, deux mystères.
Fink regarde la rangée d'ordinateurs éteints.
— Comment est-ce que ça marche, tout ça ?
Il s'assoit devant un moniteur, ses semelles trouées répandent les confettis qui paraissent plus lourds que du papier.
— Ah, je ne sais pas, dis-je. Bram a parlé d'un jeu vidéo.
Fink farfouille sur un bureau.
— Il y a des notes, là.
Je feuillette un classeur de feuilles volantes.
— Il y a un mois, dit Fink, Bram avait convaincu Plouffe d'utiliser les ordinateurs dans un cadre éducatif plus large, il soutenait que pour le bien de l'école, il fallait enseigner aux enfants comment penser en termes de systèmes. Comment coder, quoi que cela signifie. Il les a laissés faire parce que c'est un vieil homme dépassé par le monde moderne et que notre île a besoin d'un coup de pouce. Il m'avait expliqué pendant un conseil de classe, que ces enfants auraient une chance de s'en sortir si on leur apprenait à se servir des outils nouveaux et ainsi qu'ils n'auraient pas à élever des moutons ou à pêcher la morue.
Ce que je lis dans ces notes confirme ce que je pensais depuis le début : Bram avait concocté une sorte de superjeu, très régulé, superposé à celui, plus improvisé, d'Elliot et au design d'un jeu américain, _Habitat_. Un jeu fondé sur une série de lieux, créés spécialement pour l'occasion par l'équipe de Bram à partir d'un set de pièces déjà définies par cette compagnie basée à San Francisco. Si j'en croyais les diagrammes, Bram avait recréé l'école dans le jeu, plus quelques endroits spéciaux : une plage, une cabane, un rocher, une mer... Une tour noire qu'on ne pouvait déverrouiller qu'à la condition d'avoir obtenu un objet spécifique. Le Ptyx. Tout cela était lié. Est-ce qu'Elliot avait rejoint le monde de Bram, ou bien Bram avait-il trouvé un moyen d'incarner le monde d'Elliot dans le virtuel, afin de l'y rejoindre ?
Fink pianote sur un clavier ; il fronce les sourcils.
— Bracken, venez voir.
Sur l'écran, il a comme des traces de doigts. Les graphismes que j'avais entraperçus quand je les espionnais depuis les vestiaires ont tous disparus. Pourtant les ordinateurs sont allumés, ronronnant. J'ai soudain une idée, venue de très loin, suscitée par l'image d'un monde de plans, de surfaces, comme un papier tenu à la verticale, forcément plat, ou si mince que ses deux faces se confondent en ligne.
KOR
Doucement, nous passons, entre,
Sans déchirer, juste séparons,
Nous avons le pouvoir de diviser.
Sur la semelle de mes chaussures, les confettis se sont agglutinés, comme adhésifs. J'en détache un soigneusement. Ce n'est pas un confetti. C'est un fragment de calque. Je baisse les yeux.
— Attendez un peu.
Le gymnase est constellé de carrés et de rectangles transparents, de teintes variées. Une mosaïque, un vitrail fracassé.
— Qu'est-ce que c'est ?
Je brandis le filtre dans la lumière ce que j'avais pris pour un confetti.
— Une petite membrane !
— Une peau.
— Un pixel.
— Un quoi ?
— Une unité visuelle du jeu. Ce qui le compose, petit bout par petit bout. Il s'est décollé de l'écran.
— Mais... comment est-ce possible ?
Je le regarde, plongeant dans ses yeux tristes l'évidence d'une folie en devenir.
— Comment une tapisserie se décolle-t-elle ?
—...
— Comment des rétines se décollent-elles ?
Fink soudain se braque, se prend la tête entre les mains. Il n'accepte pas. Il ne veut pas voir. Comment le pourrait-il... depuis le début, il n'a cessé de contrer toutes nos théories, les fées, le Néant pour lui rien d'évident. Elliot a disparu, aussi simplement que le font les maris cruels qui quittent leur femme. Il n'y a là rien de surnaturel. Pour les gens comme Fink, c'est la nature qui est surnaturelle.
— Fink, vous avez vu les rétines de Cyldrid...
— C'étaient ses verres de contacts... Je sais que des gens portent ces choses, ce n'est pas de la fantaisie... C'était du verre, voilà tout...
— Et ses yeux. Vous avez vu ses yeux...
— Révulsés ! Il n'y a pas de malédiction ! Tout a une explication...
— Quelque chose passe entre les murs, passe entre le pixel et l'écran. Quelque chose qui est entre toutes choses et qui les libère.
— Quelque chose ?
— Ou quelqu'un.
Il me dévisage.
— Écoutez Bracken, je crois que Plouffe vous a rongé la cervelle avec ses théories. Je vous dis que tout ça, c'est la faute des fées depuis le début et qu'il suffit de faire en sorte que les traditions soient respectées pour arranger la situation. Il faut trouver un roi, sinon les problèmes vont continuer, il fait que la reine trouve un roi ! Je vais aller m'occuper de ces pauvres gamins, il faut bien que quelqu'un garde la tête sur les épaules dans cet asile !
Il quitte le gymnase en courant.
— Et ses élèves, où sont-ils ?
Mais il est déjà parti. Je regarde tout autour. Personne. Aucune trace. Le gymnase semble immaculé. Il brille de nouvelles teintes, on dirait du chrome, peut-être ce qui m'avait aveuglé quand j'étais entré. Le reflet du soleil sur les pixels morts. Du coin de l'œil, je saisis un ensemble de lignes, cette architecture entre chaque élément, qui désormais structure ma vision périphérique. Une vue diagonale, qui me permet de tout appréhender dans sa complexité.
Je plisse les yeux, j'essaye de m'en défaire, mais je suis changé pour toujours. Que s'est-il passé ? Où suis-je allé pour acquérir un tel talent ? Comme si mes yeux dessinaient au crayon la ligne claire du monde. Comme si le réel devenait bande dessinée. Il me faudrait un...
Quelque chose vient de faire du bruit.
Sous les gradins.
KOR
Enfin le barde trouve la clé,
Nous n'avons que le reste qui tend,
Ces cratères fument, dragons.
Sous les gradins, il y a des tiroirs destinés à entreposer les ballons et le matériel de gymnastique. Je ne suis pas sûr d'y avoir jeté un coup d'œil quand nous avons fouillé les lieux. Le bruit que j'ai entendu était sourd, comme un coup porté sur du bois. En me souvenant de la présence que j'avais sentie derrière moi, sur le champ de lave, je fais attention à ne pas trop approcher mon visage quand je tire le tiroir vers moi. Repliée entre les ballons thérapeutiques, une jeune fille en larmes.
KOR
Pauvre enfant,
Le témoin sans relais
Qui ne passe que le flou.
Elle m'ignore. Je tourne doucement son menton vers moi pour la forcer à affronter mon regard. Elle tremble. Je la frictionne avec mes moufles. C'est la petite amie de Bram, je l'ai aperçue amener des hot-dogs, il y a déjà si longtemps... Elle a probablement vu quelque chose.
— Mademoiselle... Qu'est-ce qui s'est passé ? Où sont-ils tous ?
Sa pupille s'élargit imperceptiblement. J'essaye d'y lire à l'envers quelque chose de gravé, une présence qui se serait inscrite, un terrible souvenir, mais je ne vois que le bleu, et mon propre reflet immuable. Ses lèvres s'entrouvrent.
— Tous nos macareux... partis...
— Des macareux ? Où ça ?
— Les lumières, dans les écrans, et puis la scie, et la division...
Elle plonge son visage entre ses bras, hoquette. Je ne sais que faire pour la calmer. Je choisis d'attendre, patiemment.
— Qu'est-ce qu'il s'est passé ici ? Où sont vos amis ?
— Il a dit qu'il voulait un système de sauvegarde mais que ça pourrait demander beaucoup d'énergie, il a dit qu'il y avait un danger pour nos corps, car nous sommes aussi faits d'électricité, et que quand le système serait activé, il pourrait y avoir des sacrifices, mais il a dit qu'il ne fallait pas s'inquiéter, que le système n'était là qu'en cas d'urgence. Il avait confiance, il était persuadé qu'il allait réussir. Nous avions tout prévu, tout, ça fait des mois que nous avons modélisé nos propres salles sur le serveur, je ne pensais pas que nous pourrions échouer.
— Échouer en quoi ?
Ses grands yeux s'ouvrent sur une dimension folle, inconnue.
— Nous avons placé toutes les balises sur les îlots autour du Cap, pour qu'il puisse arriver directement sans passer par la plage, mais nous ignorions les dangers. Nous l'avons vu mangé tout cru, avalé tout rond, je voulais qu'il rentre, la porte était encore ouverte, il pouvait revenir, il pouvait passer. Nous l'avons tous vu, il est allé sur l'iceberg, puis il a tenté s'emparer du Ptyx, mais quelque chose s'est déréglé, le système de sauvegarde s'est activé, nous avons pu récupérer le signal, mais quand il est revenu, il avait changé, il n'était plus lui-même, il était dedans, il a pris tout le monde, il a mangé tout le monde, parce qu'il avait besoin d'électricité pour exister dans notre monde, vous comprenez ? S'il veut reprendre le Ptyx au dernier Maître...
— Quel est le dernier Maître ?
Ses yeux rencontrent les miens, puis elle se met à hurler.
Elle regarde mes moufles.
— Lâchez-moi !
Elle se débat et d'un coup se libère de mon étreinte, court vers la porte. Je la manque de peu, mes doigts se referment sur le vide. Je veux la rattraper, lui dire que tout va bien, que je ne lui veux aucun mal, mais soudain, un écran d'ordinateur tombe à terre, un grondement sourd ébranle l'école. Tout se met à trembler.
Je garde mon appui sur le sol. J'attends que ça passe. Il y a des tremblements de terre tout le temps ici, ils font partie de la vie quotidienne, comme un métro qui circule sous l'appartement et auquel on s'habitue. Tous les tremblements de terre sont différents. Je n'en avais jamais connu un aussi puissant. Il dure quelques secondes, assez pour que je comprenne que chacune d'entre elle est une éternité. Je prends conscience de chaque fragment du gymnase, de la mosaïque de pixels, sur le sol, qui compose une nouvelle image : une sorte de vertige, un gouffre peut-être, comme un passage. Tout est passage : Elliot qui part, moi qui glisse, Bram qui disparaît. Cette fille, qui me parle de venue, Cyldrid qui sent quelque chose de nouveau. Quelque chose serait entré.
Quand les secousses cessent, un ballon roule jusqu'à moi. Je cherche d'où il est sorti. Personne ne l'a lancé, il semble avoir jailli tout seul de son tiroir. Je me souviens de cette présence dans les vestiaires, avec moi. Qui m'épiait.
Une brise m'accueille en riant, la porte s'ouvre sur un grand-angle iodé, la mer m'avale tout entier, moi et ma joie. Sur les marches, Egill pointe de l'aile un coin de ciel. J'y devine trois mouettes : des traits flous, sans silhouette. En dix ans ici, je n'en ai jamais vu une de près.
Les étoiles factices scintillent comme des boules disco, poilade astrale penchée sur ma vie. Il est si étrange de marcher sur la terre ferme, j'ai perdu le sens de l'orientation, de l'équilibre.
La barque repose sur le sable, la mer s'est retirée d'un kilomètre, laissant derrière elle une architecture d'algues en deux dimensions. Le morse s'affaire déjà.
— Vite..., crie Hinrik.
Je ne bouge pas. Le morse fronce les sourcils.
— Briki, qu'est-ce qui ne va pas ?
— Je ne sais pas.
Le morse me fixe de ses petites billes sombres.
— Vous avez changé d'avis ?
— Non.
À mes pieds, le macareux lève son bec vers moi. Je ne sais pas ce qui se passe dans mon ventre. À la vue de ces algues, l'envie d'aller revoir le monde intermarée me saisit. Je voudrais aller explorer ses moindres recoins, me nourrir de sa pico-activité, maintenant que j'ai appris à voir en chaque détail un événement, maintenant que la patience est devenue ma façon naturelle d'interagir avec le monde. Peut-être est-ce là mon destin, basculer dans ce petit univers, vivre à son échelle, dans le néant de l'intérêt personnel, pour respirer tout entier, de concert avec un ensemble de poche.
Je n'ai pas envie de renoncer à ce que je suis devenu ici, dans le rien. À dessiner le rien, qui est, qui n'est pas. Le trait comme seule raison de vivre. En dix ans, j'ai eu le temps d'oublier pourquoi je cherchais Elliot. L'ai-je jamais su ? Après tout, si Elliot était Maître ici, en quoi pouvait-il représenter un danger pour nous, de l'autre côté ? Il y avait eu un autre Maître avant lui.
Egill se cale sur ma jambe. Il fait toujours ça. Juste être là, tout contre, posé comme un gardien. Simplement posé. Soudain, à côté lui, à côté de cet ami qui a vieilli avec moi et qui est à présent si proche de la mort, je comprends à quel point je suis égoïste. Si Hinrik avait accepté, il y a dix ans, de m'emmener là-bas, c'est parce qu'Egill voulait savoir d'où il venait. Lui qui est incapable de voler. Lui qui a besoin de nous pour le mener à sa famille.
Hinrik attend.
— Allons-y, dis-je en renfilant mes moufles, après dix ans mains nues.
Hinrik pousse la barque en souriant. Il a besoin d'aide : c'est un vieux morse obèse qui a du mal à se déplacer sur la terre ferme, même si le sable est humide. Au cours de la dernière année, il a pris un gros coup de vieux, m'obligeant à assumer le quotidien de la maisonnée. Entre ses rides apparaissent des corps nouveaux, des colonies de poussière. Son regard est devenu vitreux, sa moustache, si noire quand je l'ai rencontré, était un champ de neige parsemé de petits rochers, affleurant sous la glace. Il respire lourdement. Je sais qu'ils pense que nous ne reviendrons pas de ce voyage : nous n'avons pas pris de réserves, juste les filets, pour assurer notre survie au cas où nous dériverions trop loin. C'est trop dur de le regarder mourir. Il faut qu'il périsse en mer. Jamais il n'avouera sa douleur de quitter la vie harmonieuse que nous avons menée ici ensemble pendant dix ans. Peut-être est-il déçu. J'ai lu dans son regard un dépit teinté de résignation. Malgré la paix que j'avais trouvée dans mon existence, je devais aller jusqu'au bout. L'entre-deux est toujours un moment à passer, dont on prend peut-être conscience, mais aucun être humain qui éprouve du désir ou de l'attachement ne peut y demeurer éternellement. Moi, je suis un être humain.
Je dis mentalement adieu aux petites anémones, aux royaumes entre les algues, en leur promettant de revenir, comme ces chevaliers courtois qui juraient à leur pucelle de revenir un jour demander leur main. Je dis adieu à la cabane. Ces dix ans vont bientôt devenir un souvenir, comme l'est devenue ma vie avant la plage, quand j'étais un professeur de dessin, que ces mains enseignaient aux enfants comment représenter un oiseau. Après toutes ces années à m'autoéduquer dans la nature, je suis redevenu un de ces enfants à qui il faut expliquer le fonctionnement de l'anatomie. Je suis un souvenir de moi-même, et je ne me ressemble plus.
Une fois la barque mise à flot, Hinrik se hisse en ronchonnant. Egill s'est posté à la proue, les yeux fixés sur l'horizon : une ligne que je n'arrive jamais à lire, confondue dans le noir du ciel, ce profond miroir d'étoiles sinistre par son immobilité. J'enfile le ciré, les bottes puis, coiffé d'une casquette, je me penche vers l'eau pour vérifier que je suis bien là, comme je le fais chaque fois que nous partons en mer. Les noirs remous ne me rendent aucun reflet.
Une vaguelette s'écrase mollement sur la proue. Hinrik me passe les rames ; leur bois semble irréel, presque trop froid, trop brun, trop taillé. Il pousse le bateau jusqu'aux premières vagues, se place au gouvernail ; relâchant le cordon qui nous reliait aux pilotis, il contemple l'infini. Je jure avoir vu un nouveau vent se déplacer dans sa moustache, aiguiser son regard, quand sur cette mer d'huile, sans surface, il devine des profondeurs que je ne connais pas.
Emmené par le courant, je regarde s'éloigner la plage, sachant que je ne la reverrai jamais. J'aurai pu aller voir la caverne, vérifier si la marée ne l'aurait pas élargie d'une façon ou d'une autre, pour y respirer les odeurs d'une enfance impossible, parmi les fleurs sous-marines. En dix ans, j'ai appris que certaines choses, quand elles nous donnent la paix, doivent rester confinées dans cette zone étrange entre la certitude et l'impossible. Je ne saurai jamais qui est entré dans ce tunnel, menant vers l'ailleurs de l'ailleurs, l'autre côté de l'autre côté, l'au-delà du Néant. Je n'aurai jamais accès à ces mondes merveilleux où l'on glisse sur l'onde en chantant avec les anémones. Je n'ai plus que cette mer, cette nuit de roches.
Pendant les premiers kilomètres, Hinrik n'a pas dit un mot. Parfois, il se penche pour tailler une fente sur le bois, sa façon à lui de compter les distances. La côte a entièrement disparu, nous sommes désormais en eaux inconnues, bien trop profondes. Souvent, je me suis demandé quelles créatures hantaient ces fonds, où semble somnoler toute la menace d'un mythe. Je n'ai jamais rien vu qui eût pu me faire croire que de tels monstres existent, mais j'ai vu un épaulard, une fois. Il avait émergé des flots comme une sublime sphère noire, dans le blanc de ses yeux immenses, je voyais se refléter toute l'histoire du Néant, son apparition et sa disparition dans la lumière, ses scintillations.
Nous doublons des îlots de roches sombres aux brisants acérés, parfois nous y jetons l'ancre. La tour elle-même apparaît parfois, comme un phare, loin devant, sans jamais sembler se rapprocher. Je me surprends à penser que nous allons passer notre vie sur cette étendue liquide, sans espoir de jamais rentrer chez nous. Mon aveuglement avait conduit notre maisonnée à sa chute et aucun de nous n'en sortirait vivant.
Pour me calmer, je sors des feuilles. Les vagues, toutes les vagues, une par une. Cataloguées. Dessiner avec des moufles est une expérience amusante : ma main était devenue plus légère, après toutes ces années d'entraînement. Mes figures sont différentes. Je comprends qu'il est important de changer ses habitudes. Mon trait glisse, se prend dans un détail du ciré d'Hinrik, le morse incapable de nager. De là, je dévie vers une forme qui pourrait être son énorme nageoire, toute molle, comme une main de papa géant, puis je sors de son corps pour dessiner la barque. Par la magie du dessin, toutes les formes fusionnent. Il est cependant possible de les séparer par des lignes de démarcation. J'éprouve la différence entre la zone et la forme, entre la forme et la silhouette. Je me dis qu'après toutes ces années à avoir oublié qui j'étais, je me retrouve en paix, ici au milieu des eaux, assis sur ce banc de bois, entouré d'amis patients. Je ne redoute plus d'être celui que je suis, ni ce que j'ai été puis oublié d'être, de peur d'avoir à justifier un point de vue.
Je cligne des yeux. Dans le clapotis, vertige quand mon champ de vision se met à tanguer. Affaissé sur mes rames, en plein milieu d'un sombre océan, sous un ciel gris-bleu au regard noir qui me renvoie le tourbillon d'un orage sans pluie, sans tonnerre, sans foudre. Juste une masse informe qui tourne, vortex de noire barbe à papa. Il n'y a plus d'étoiles, ce n'est pourtant pas le jour. Je n'arrive pas à saisir la ligne d'horizon, à peine un trait fuligineux comme du pastel entre les nuages et les vagues. Douces vagues, qui jamais ne nous renversent. Un long mouvement balance, me berce, puis les nuages changent de forme, s'effilochent et s'étiolent, irisés d'une nouvelle lumière venue d'outre-espace. Soudain, la mer se creuse, découpée dans un monde composé de petits cubes. Cet océan discontinu n'offre aucun repère. À perte de vue, rien que la mer, et l'écume en dents de scie.
Après des semaines de navigation, nous approchons d'un attroupement de macareux, dodelinant sur les vagues.
— L'Oracle, m'annonce Hinrik.
Leurs palabres nous accompagnent depuis longtemps, elles explosent en fanfare de salon quand nous traversons leurs rangs. Ils nous regardent passer, sans bouger, s'écartant à peine de la barque, curieux, presque ironiques. Parfois ils s'envolent, battant maladroitement de leurs petites ailes et s'écrasent dans les flots un peu plus loin, ce qui comme je l'ai dit constitue leur façon d'atterrir. Hinrik m'explique que ces macareux sont la conscience de Ginnungagap, les mots qui ne disent rien, qui ne veulent rien dire. Ils sont la matérialisation de ces pépites d'attention, une grammaire aquatique régulant sa propre orthographe.
Egill s'est juché sur le plat-bord, ses deux petites pattes bien crochetées ; de son bec, il claque un mot, pour demander des nouvelles du temps. D'un seul mouvement, les macareux se déplacent pour former des signes sur l'onde, leur code oiseau indiquant que peut-être va se lever une terrible tempête, qu'il vaudrait mieux que nous empruntions certaines ravines de vagues plutôt que d'autres. Ils indiquent à grand renfort de battements d'ailes que traduit Hinrik, que le Cap est abandonné, que le Maître est parti. Hinrik semble nuancer leur propos : le Maître est paresseux, c'est pour cela qu'il ne se passe rien. Nous quittons les volatiles, je ne sais pas ce que tout cela peut signifier mais j'ai peur qu'un volcan ne s'éveille. Personne n'est très agréable au réveil.
Si long, ce voyage. Sans rien d'autre qu'un trait, au loin, et ces vagues, si proches. Hinrik n'a pas cessé de chanter la geste d'Elliot, il compose des accords sur sa mandoline, mélodies qu'il improvise pour agrémenter la répétition. Comment Elliot s'est levé pour gagner les derniers étages du Cap, comment il a triomphé du Kor et volé le Ptyx laissé vacant par le Pénultième, pour régner sur le Néant. Rien de nouveau, toujours les mêmes motifs, encore et encore, tricotés en farandole de superlatifs. J'ai fini par admettre que le mythe du Néant n'avait que la consistance que lui accordait un homme venu d'un monde inconnu pour échapper à la mort. Je ne sais toujours pas ce qu'est exactement le Kor, le Ptyx, ni pourquoi le premier habitant du Néant était une belle tortue, paradoxalement invoquée par l'habitant suivant.
Après un mois de voyage, nous accostons un rocher en pleine mer afin de bivouaquer. Avec la marée basse, certains affleurements apparaissent miraculeusement. C'est la première fois que nous posons le pied sur quelque chose. J'ai l'impression d'être rouillé par le sel. J'en fais le tour au petit trot pour me dégourdir les jambes, aussi dans l'espoir de retrouver quelque colonie marine qui pourrait me divertir.
Au sommet du rocher, qu'il me faut gravir à plat ventre, se trouve une torche. Plantée entre deux rochers, assez haute pour préserver sa flamme dans la marée. Elle semble presque neuve. En plastique, comme ces gadgets pour illuminer les jardins la nuit, et attirer les feux follets. Elle me semble si peu adaptée à ces lieux, tellement anachronique. Comme ce château de sable, sur la plage, mais en plus concret. Comme une vision plus mûre, plus aboutie, de quelque chose qui n'avait été jusque-là qu'entrevu. Je reste assis là, tout en haut, replié sur moi-même, écoutant le grondement lointain, un ronronnement qui s'amplifie depuis quelques milles. J'ai l'impression d'être au bord du monde, au pied d'une flamme marquant le début d'une ligne de démarcation. Soudain, je sens la secousse. La terre tremble. La mer semble pencher. Au loin, un flash de lumière, puis une trace d'orange se peint comme du pastel sur l'horizon, teintant chaque nuage, tous les recoins du ciel. Je redescends à la barque en courant. Le visage d'Hinrik est à présent si pâle, que je peux voir au travers.
— Hinrik ?
— Le ciel se déchire.
— Un volcan sous-marin ?
— Ginnungagap.
Au moment d'enjamber le flanc de la barque pour m'asseoir sur le banc, pour la première fois j'ai mal au dos. Je me rends compte, là, sur ce caillou, que j'ai désormais dix ans de plus et que ma vie est derrière moi. Comme si ce passage obligé avait été balisé par la torche.
Un nuage de cendre bouche le ciel, enfermant la mer dans un cercle scintillant. Nous dérivons. Egill s'est assis pipe au bec et marmotte d'obscures sentences dans sa moustache, effilochées en miettes de tabac. Le bateau dérive lentement, attiré par l'horizon. Tout autour de nous, le monde devient violet.
Hinrik attire mon attention sur un détail, à bâbord.
— Nous sommes sur la faille. Regardez.
La mer a changé d'aspect. Je n'ai jamais réellement fait attention à sa texture, mais elle m'apparaît soudain visqueuse, tel du miel ou de la confiture. Une vaste étendue de gelée, monstrueuse de densité. Je me penche sur le bord pour regarder l'eau se troubler, me laissant entrevoir les méandres du terrible Ginnungagap, là, tout en bas, cette fente qui m'a entraîné ici et que je ne peux pas approcher, à moins d'en mourir. Je crois que mes yeux me trompent, qu'une chose pareille ne peut exister, mais alors que la lumière se fait progressivement douce, je prends conscience de l'immensité autour de moi, sous moi : un véritable gouffre, si profond que l'océan lui-même semble s'y écouler. Nous sommes si petits, tout petits, au sommet d'un monde en train de naître, ou de finir. Je ne trouverai pas de mots pour décrire ce vertige qui me prend quand la zébrure de la faille s'étale sous mes yeux, serpentant sous une eau devenue cristalline. Une bande de noir, qui divise le monde en deux, tout en bas, là où le sable semble gris. Une balafre qui danse dans le remous. Je suis pris d'une terreur primitive. Là se trouve l'impossible vérité de ma condition : je me sens abandonné au bord du vide, livré à moi-même.
Le morse pose sa nageoire sur mon épaule. Je sens qu'il essaye de me consoler, de me montrer à sa façon qu'il comprend ma détresse. Mais je ne peux plus rien exprimer, ni reconnaissance, ni compassion, que ce soit pour lui ou pour moi-même. Je suis inscrit dans un entre-deux, avec la sensation d'y pourrir pour toujours. J'ai envie de me mettre en mouvement, là, maintenant, de faire des exercices, de me sentir vivant, de ne plus m'arrêter de bouger. Mais en moi la fureur s'est éteinte pendant toutes ces années, je suis resté muet, passif, à attendre que l'envie reflue, que la mort reprenne son empire sur mon corps. Je laisse mon doigt dans l'eau, juste pour établir un contact entre la mer et le ciel, avec moi entre les deux. Je creuse un sillon dans l'eau épaisse, sombre, insondable, où mes rêves prennent substance, ici dans l'imagination matérielle.
— Diane !
— Oui ?
— Vous voyez ce que je vois ?
Un grondement éventre le ciel. À la barre, Hinrik lève des yeux inquiets. Au loin, un immense front orageux s'avance, marée du ciel, et nous, juste en dessous, pauvres coquillages sans rochers attendons d'être emportés. Nous sentons le vent qui précède la pluie, puis l'eau nous fauche, de terribles rafales qui nous repoussent, toujours plus loin de notre but, charriant dans leurs mèches invisibles un déluge d'eau, puis de grêle. Nous nous accrochons à ce que nous pouvons, craignant d'être balayés comme on essuie une larme d'un visage, d'une main songeuse. Pourtant, j'ai l'impression que c'était le Néant lui-même qui regimbait contre notre présence en ces eaux, dans sa chevelure aquatique. Je me sens parasite.
— Ce serait pas le petit Bracken ?
— Hourra ! C'est bien lui !
— Il va tous nous sauver !
Monstrueuses, les vagues enflent, se brisent sur la barque. Nous allons y rester, couler là, si près du but. Au loin, quand une vague nous soulève sur sa crête, j'aperçois le phare, qui nous guide.
La vérité s'y trouve, peut-être même Elliot. Je ne sais plus ce qui me pousse à le retrouver, je ne sais même pas si j'ai jamais eu une raison de partir à sa recherche. Il me semble, dans l'engourdissement de ma volonté, qu'une motivation toute simple suffirait à me sortir la tête de l'eau. Là, en pleine tempête, à moitié mort, je réalise soudain que ma vie n'est qu'indolence, que la quête absurde que je me suis fixée n'est qu'un prétexte pour ne pas mourir.
— _Nous sauver de quoi, Roméo ?_
— _Mais vous savez très bien, Diane..._
— _Heu... Non._
Nous luttons contre les vagues qui, tels des monstres fluides, grondent autour de nous, dragons ouvrant leurs mâchoires pour nous avaler. Nous les transperçons de la proue, dans les éclaboussures. J'ai enfoui Egill sous ma vareuse, son corps tout chaud contre moi, dont je sens la respiration et parfois le bec qui sort pour me picorer la barbe en signe d'encouragement. Hinrik empale les vagues de sa gaffe, l'eau explose tout autour alors que dans le ciel tourbillonne l'œil borgne d'un cyclone. Je me surprends à rire au milieu de ce chaos, me sentant pour la première fois vivant dans toute ma démesure et ma vaine persistance.
— _Mais si, vous savez, le Ptyx !_
— _De quoi parlez-vous ? Qui êtes-vous ?_
— _Qui je suis ?_
Le Cap apparaît à travers la tempête, faible lueur orange nous faisant signe d'approcher. Depuis que nous avons pris la mer, nous ne l'avons pas vu grandir : son apparence est demeurée celle qu'il avait déjà sur la terrasse de la cabane, lorsque je le regardais apparaître et disparaître dans la brume. Il était déjà le symbole de notre aveuglement, de notre folie, de notre audace. Il se dresse à présent pour nous rappeler notre destin ultime, à savoir périr noyés, nous tous, homme, animal qui ne nage pas, oiseau qui ne vole pas. Tous autant que nous sommes, nous voilà condamnés.
— _C'est bien, la vie de poisson, on n'a pas de soucis._
— _Non, c'est vrai, tout est toujours nouveau !_
— _Regardez, plein de nouveaux objets !_
Les ténèbres s'emparent du monde, en lames épaisses qui remplissent les silhouettes. Les vagues forment des parois opaques, sans issues. Les bras me tirent à force de ramer, nous ne pourrons plus tenir très longtemps sans secours. J'hurle à Hinrik que je n'en peux plus. Il me relaie aux avirons, je passe à l'arrière, et je me cramponne au gouvernail, essayant de déchiffrer le monde rugissant, déchaîné, bien trop rapide pour moi. Parfois la barque se hisse à la verticale, comme à flanc de montagne, pour retomber lourdement dans un ravin liquide. Tout s'accélère, nous glissons à nouveau pour recommencer, en un manège infernal qui me donne la nausée.
— _Sympa, cette boum en pleine mer !_
— _On va se faire des copains !_
— _Et des copines !_
Quelque chose choit dans la barque. L'un de ces drôles d'oiseaux que j'avais vu tourner en nuée quelques heures auparavant, brusquement assommé par le vent. En fait d'oiseau, ça n'a rien d'organique. On dirait plutôt un morceau de bois brûlé, du pur carbone, mais très vite, je comprends de quoi il s'agit. Ces mouettes que j'avais vu, à plusieurs reprises, sans jamais les voir approcher. Ces traits dans le ciel. C'est bien ce qu'ils sont réellement : des traits dans le ciel. Littéralement. Sans pliure pour voler. Sans aucun signe de vie. Juste un trait au fusain, épais, réel. Tombé là car décroché de sa voûte. Tout ici est illusion, mais illusion réelle. Les trompe-l'œil ne sont pas là pour tromper. Ils sont devenus la nature.
— _Sacrée tempête !_
— _Ça n'augure rien de bon !_
— _J'espère que ça va pas gâcher la boum..._
La barque se disloque. Le bois craque, l'eau commence à pénétrer, je sens que la fin est proche. Nous écopons, même Egill tente de nous aider, mais le pauvre oiseau est si maladroit qu'il fait rentrer plus d'eau qu'il n'en sort. Je suis désespéré. Hinrik cesse de s'agiter, me regarde, haletant.
— Je suis désolé Bracken.
— C'est pas grave, c'est de ma faute.
Brusquement, je me fige. À travers le rideau de pluie, j'entrevois soudain la masse sombre d'une montagne en mouvement, devant nous. Elle nous dépasse très vite, c'était un gigantesque iceberg, déchiqueté. Qui sait quelle formidable masse ce bout de glace peut cacher sous les flots. Quelque chose bouge tout en haut, une forme, une silhouette, peut-être un homme, je n'arrive pas à distinguer. Il nous fait des signes, ou bien il danse, tel un fou au bord d'une falaise. Nous le regardons passer, ahuris, comme des vaches regardant passer des trains, des forteresses sur la pelouse. De quelle banquise vient-il ? L'écosystème du Néant est donc plus complexe que je ne l'avais envisagé. Même s'il fait du mensonge une réalité, le Néant reste cohérent avec lui même, presque prévisible.
Une vague de titan se hisse devant nous, massive, suspendue, séparant les flots avec une suprême majesté. Elle nous regarde de ses yeux froids et profonds, derrière lesquels je peux deviner la farandoles des poissons, des tortues et des algues, habituel ornement de ces cathédrales aquatiques. Un moment de silence, d'éternité, pendant lequel m'échappe la maîtrise de mon destin. Mais l'homme survit, l'homme résiste à la tentation de la béatitude. L'homme est faillible, l'animal l'est aussi.
Hinrik se jette à l'eau, tandis qu'Egill tente de s'envoler. Je saute au moment précis où la vague fracasse la barque.
— _Ah ben tiens, en voilà un !_
— _Bonjour copain !_
— _Yo !_
Ouvrir les yeux dans une eau si froide que je suis immédiatement saisi de convulsions. Je sais qu'il me reste peu de temps à vivre. Il faut que je bouge. Sous l'eau, tout est bleu. Je m'attendais aux ténèbres, mais je prends conscience de l'immensité et de la perfection de ce tapis, infini. Je suis seul. Les vagues roulent au-dessus en étranges motifs déliés, réguliers. J'occupe le moins d'espace possible, tout mon corps est en apesanteur, épousant les mouvements ondulants de mes vêtements. Si j'enlevais mes moufles, je pourrais voir mes doigts si pâles, dans ce bleu sans limites, dans ce cube sans arête. Je regarde droit devant moi, dans toutes les directions, mais je ne vois ni Egill, ni Hinrik. Se sont-ils déjà noyés ? Cette brusque angoisse me fait me retourner dans la masse fluorescente de cette mer étrange, et je me retrouve face aux abysses qui semblent venir à ma rencontre à la vitesse d'une locomotive. Je m'atomise, je me renie, j'essaie d'être inexistant, me faire si petit que tout me traverse sans dommage. Brisé, rompu, l'espace bleu devient une plaie béante, une tranchée dans le sol, galopant vers l'inconnu. Je vois mes pieds, gesticulant dans le rien. Je vois la fente. Ginnungagap béant sous mes pieds, à des kilomètres de profondeur, et pourtant, j'ai l'impression de marcher sur ses flancs. L'océan entier glisse doucement par cette fente, tout le Néant est aspiré, avalé pour être recraché de l'autre côté, peut-être dans la chambre d'Elliot, peut-être ailleurs. Je ne pense pas pouvoir survivre à un nouveau transfert. Je n'ai plus assez d'air dans mes poumons, ni le moindre espoir de rester sain d'esprit. Tout est trop vaste pour être assimilable. Tout va bien au-delà de ma pensée. J'ai beau me souvenir de l'illusion concrète, des voiles à déchirer, une terreur me mure en moi-même. Cette bouche est immense, elle va m'avaler, je sais que je vais vers elle, elle m'attire, elle me tire, elle a peut-être déjà pris mes amis. Si je pouvais dire ce qui se passe en moi, désormais, tandis que ce ventre s'ouvre sur la misère de ma présence, je le ferai uniquement en dessin. Mais mon cahier se noie sous mes yeux, spiralant vers le gouffre. Je bats des pieds pour grimper, mais je sais qu'une tempête m'attend, et que je finirai noyé. Des yeux, dans le flou salé, je cherche la masse de la barque renversée, mais je ne la trouve pas. Je suis aspiré, je vais encore me noyer pour la troisième fois... ce sera peut-être la bonne.
— _Dis donc, il pourrait dire bonjour !_
— _C'est parce qu'il coule !_
— _Mordez-y le doigt !_
Avant de fermer les yeux pour toujours, je vois danser devant moi deux petites formes. Je manque d'en rire de démence. Pour la première fois depuis mon plongeon, je trouvais un repère plus petit que moi auquel m'accrocher. Il me faut une demi-seconde pour comprendre. Je forme leur nom sur mes lèvres, une série de bulles s'échappent, moqueuses. Les tortues ! Les tortues d'Elliot ! Elles flottent devant moi, leurs petites nageoires déployées, souriant presque. Je les avais perdues, elles étaient là depuis tout ce temps, peut-être veillaient-elles sur moi, qui sait. Peut-être est-ce un mirage. Je tends la main pour les toucher, pour m'assurer de leur réalité à travers mes moufles. Et ça fait mal.
— _Bravo !_
— _Gnrrururrr !_
— _À moi, maintenant !_
Elles me mordillent les mains. Leurs petites bouches entaillent à peine la laine gorgée d'eau de mes moufles. Elles me tirent, exerçant une force suffisante pour m'extraire du courant menant à la faille. Ou bien me suis-je persuadé que le gouffre m'attirait, alors que c'était moi qui désirais l'atteindre, sans me l'avouer. Mon corps passe à l'horizontale. Elles ne me font pas remonter. Elles m'entraînent vers le perpétuel ruban bleu qui se délite doucement en bandes turquoise irisées de marine où dansent les pâles lambeaux d'un azur de miel. J'ai mal partout. Tout mon souffle brûle. Mes poumons deviennent écarlates, jurant dans ce camaïeu de bleus, je suis phosphorescent... mais qu'est-ce que c'est ?
— _Mmrrhrrhrh !_
— _Kjjfflmmmddd !!!_
— _Hjk !_
Dans l'immensité, brusquement, vient d'apparaître un mur. Plus terrible encore que cette fente diabolique qui fait respirer le grand fond. Est-ce le bord de la caverne, qui englobe tout ? Ou bien... Non, c'est une patatoïde, flottante. L'iceberg. Si vaste, une montagne qui plonge dans l'abysse, presque à toucher ses lèvres. Un vaisseau croisant telle une magnifique baleine grise, lisse et sans aspérité. Des formes dansent devant mes yeux, la fatigue m'envahit. La blessure à mon doigt me fait mal, elle commence à me lancer d'une douleur nouvelle. Quelque chose la réveille. Est-ce la mort ? Ma conscience qui s'enfuit, des nénuphars, des chevaux qui galopent vers moi puis disparaissent. Mes pensées montent et descendent, semblent s'éteindre pour toujours. La paroi s'approche, monstrueuse, constellée de petits orifices. J'aimerai m'y infiltrer, passer le cap de ces créneaux, si petits qu'une mouche ne pourrait y... suis-je en train de me transformer en plancton ? Non, je ne veux pas, je vais m'écraser, j'ai besoin de respirer, d'ouvrir la bouche, je dois... Non, c'est trop petit, je ne passerai jamais, je ne passerai pas, c'est trop petit, l'espace est trop réduit, et moi, trop gros, bien trop gros, j'essaye de reculer mais les tortues sont les plus fortes, elles me tirent, j'entre tout entier, je deviens minuscule, dans le goulot, dans le couloir, puis vers le ciel, vers une tache blanche, qui se déploie en corolle de lumière.
Alors que ma cage thoracique menace d'exploser, que mes tympans sont déchirés par un atroce son de cloches, je garde à l'esprit qu'ici il se peut que la lumière soit solide.
KOR
La fin, sans fin,
Pourtant s'approche
D'une fin
J'ai perdu la fille. Les couloirs sont déserts. Les enfants sont tous dehors, je les vois danser derrière les baies vitrées. Personne n'a paniqué après le tremblement de terre. C'est comme si rien ne s'était passé. Le ciel s'éclaircit, les énormes paquebots de nuages se rangent, s'écartent pour libérer une mer d'azur aux reflets de nacre. Un soleil froid apparaît timidement, rieur dans la joie d'un jaune fauve. Les enfants lèvent les bras vers le ciel et dans le souffle blanc de leur parole, je lis l'invocation sacrée du dernier jour de Noël.
Une douleur me lance dans la main droite. Fugitive, elle s'évanouit aussi vite qu'elle est venue. J'ai pu la sentir naître, grandir, puis diminuer, jusqu'au néant. Intéressant de voir comme je peux désormais, sans même y penser, me faire le témoin de chaque microévénement intervenant dans mon corps ou dans ma tête. Comme si chaque idée, chaque nerf, était un émetteur dont je connaissais d'avance la fréquence. Et que je pourrai peut-être apprendre à tripatouiller, afin d'améliorer la réception. C'est effrayant d'être à ce point réveillé.
Je retourne dans la chambre d'Elliot, pour prendre des nouvelles de Cyldrid. Il n'y a plus personne. Sur la table basse, plus trace des pupilles. Où sont-ils tous passés ? Le chaos est toujours là, comme une empreinte séchée, comme des algues qui prendraient la forme de meubles. La tapisserie est toute entière décollée, laissant la roche du mur apparaître. Les interstices sont noirs, noirs d'avoir vieilli. Je passe la main doucement, espérant trouver un trou, quelque orifice où m'infiltrer, pour me convaincre que tous ces gens qui disparaissent ne sont que des illusions proprement humaines. Fink a peut-être raison : l'école est bâtie sur un ancien royaume féerique, et peut-être que tout ce que je vois n'est qu'un écho, une illusion. Du glamour, comme disent les vieux. Peut-être suis-je aussi une fée. Peut-être ai-je tout oublié de ma véritable nature. Je regarde la chambre en panique, cette même chambre que je connais par cœur, recoin par recoin, et cet angle tout au bout, vers lequel, inéluctablement, je suis attiré. Je m'y assois. Il ne se passe rien. Ici, étrangement, je n'arrive pas à lire les lignes. Je ne vois que des surfaces chaotiques, des enchevêtrements d'objets, un dégueulis tout en fouillis.
Je retourne aux toilettes pour handicapés.
— Toc, toc !
J'entends rouler derrière la porte.
— Qui est-ce ?
— C'est moi, dis-je, fatigué.
Les loquets sont ouverts, un à un, et voici Plouffe, la mine réjouie.
— Bracken, quelle chance !
— Qu'est-ce que vous faites là-dedans ?
— Des recherches. Et... Je me cache.
— Pourquoi ?
Il regarde à droite, à gauche dans le couloir, puis me tire par la manche.
— Venez.
Il a mis son appartement sens dessus dessous. Je sais qu'il vit autour de la cuvette, parce qu'il est malade, qu'il a besoin d'être là souvent, que la lumière le rassure. Avant, il vivait dans une petite mansarde, mais, la mort approchant, il cherche un confort.
— J'ai consulté tous les textes, et je peux vous garantir que le Ptyx existe bel et bien. Bram et Elliot ont raison ! Tenez, regardez, c'est un ex-libris de Mallarmé.
Il me montre un texte en français, « La Pénultième est morte », dont je comprends parfaitement les mots, mais dont je ne saisis pas le sens.
— C'est un de ses rares textes en prose. Lisez donc : "[...] je reconnus en le son nul la corde tendue de l'instrument de musique, qui était oublié et que le glorieux Souvenir certainement venait de visiter de son aile ou d'une palme ; et, le doigt sur l'artifice du mystère, je souris et implorai de vœux intellectuels une spéculation différente."
— C'est de la poésie, Plouffe, ce sont juste des mots. En quoi est-ce que ça va nous aider à retrouver Elliot ?
Il lâche le livre.
— Ça peut nous aider à trouver le Ptyx. Cet objet qui nous permettrait de changer le Néant. Comprenez, tout ceci est une mythologie, c'est comme ça qu'il faut la lire, et comme toutes les mythologies elle peut s'interpréter littéralement. Vous allez encore dire que je suis fou, mais après tout ce que nous avons vu, et vécu, je crois que j'ai le droit d'y croire... Il y a un Maître au Néant, qui ordonne l'invisible du monde autour de nous, qui nous influence, et ce Maître ne peut être Maître que s'il détient le Ptyx.
— Comment l'invisible pourrait-il changer le visible ?
— Si le Néant change, alors la réalité change, puisque c'est un écho de ce qui se meut dans le Néant. Tout est là : comment le Pénultième qui garde le Néant dans un état stable est remis en cause par le dernier, Elliot.
— Attendez, quoi, comment Elliot pourrait-il être Maître ?
— Il a volé le Ptyx, non ?
— Nous n'en savons rien, c'est ce que nous dit Bram, et Bram a disparu, alors c'est peut-être lui, le nouveau Maître ! Pour autant que je puisse en juger, Bram avait tout prévu, avec ses machines et ses balises et ses systèmes de sauvegarde. Il a eu accès au même endroit qu'Elliot. Il aurait très bien pu gagner.
Plouffe soupire.
— Ce serait terrible.
Je contemple son débarras, ses vieilles affaires, tous ces livres entassés, jaunis, comme les journaux dans le bureau de Fink. Qui est cet homme qui abandonne son école pour se mettre en chasse d'une image poétique ?
— Où est Cyldrid ? Vous étiez censé la surveiller, non ?
— Oh, elle s'est levée d'un bond, puis elle est partie !
— Elle a repris ses yeux ?
Plouffe hoche la tête.
— Elle les a gardés dans la paume de sa main, et elle est partie. Elle chantait.
— Quoi ?
— Elle chantait le Néant, Bracken.
Plouffe prend une inspiration puis, d'une voix de fausset, tente de reproduire la mélodie.
— "Venu de l'entre-deux, le Kor, derrière, pour réparer le pli, n'aura de cesse de filtrer, quand le Maître du Néant dira le mot qui frémit, comme le dernier avant lui a pris le Ptyx au Pénultième..." C'est ça aussi qui m'a mis sur la piste de Mallarmé. Dans ce texte que je vous ai montré, « La Pénultième est mort » est une phrase mystérieuse, qui s'impose à lui sans la comprendre, sans la résoudre. Un mystère. Je crois que Mallarmé avait des visions, que le mot "Pénultième" était une obsession, comme un mantra, dont il devait trouver la solution. Comment un tel mot se retrouve-t-il dans la bouche de Cyldrid ? Si mes théories sont exactes, alors Elliot s'est rendu dans un lieu connu de Mallarmé seul, un lieu auquel nous pouvons accéder ici, en Islande, en cet entre-deux. Un lieu où le mot Pénultième devient l'objet d'un pouvoir sans limites, peut-être même un miroir de son âme. Le Maître est le Pénultième, et Elliot a pris sa place.
— C'est quoi, un Kor ?
— Ça veut dire "Chœur", en islandais. Comme un chœur de théâtre.
— OK, donc, laissez-moi comprendre : Elliot s'est rendu au Néant pour obtenir un trésor qui lui permettrait de changer la réalité...
— Techniquement, il change le Néant, mais si l'on part du principe que...
— Il a pris ce trésor des mains d'un poète qui semble-t-il se faisait appeler Pénultième...
— Stéphane Mallarmé lui-même. L'homme qui a trouvé le Néant en creusant le vers.
— Mais vous avez dit qu'il est mort il y a presque un siècle...
— Essayons de reprendre son parcours. Il a écrit ce poème sur le Ptyx, il descend au Néant prendre le sceptre et il change la réalité, en tant que Maître Pénultième.
— Qu'est-ce qu'il a changé ?
Plouffe ouvre la bouche, puis la referme. Il ne sait pas.
— L'une des élèves de Bram m'a dit que Bram était parti à la suite d'Elliot pour voler le Ptyx. C'est un jeu, une compétition. Je sais qu'Elliot joue seul, mais Bram est derrière lui. Admettons qu'Elliot soit donc le dernier. Qu'est-ce qui vient après le dernier ?
— Bram.
Soudain, j'ai très chaud. Trop chaud. Je m'assois sur une pile d'annuaires cornés, épuisé par tout ça, renonçant à comprendre. Je regarde Plouffe et toutes les traces qui le composent, sa silhouette, détourée sur la mosaïque de livres à demi-ouverts, sur les lignes régulières de la faïence, des murs. J'arrive à voir la lumière se décomposer en traits, comme des flèches tirées d'une petite sphère.
— Qu'est-ce qui ne va pas ?
Inquiet, Plouffe me passe la main sur le front.
— Vous avez de la fièvre.
— Je ne sais pas ce que j'ai, je... Fink pense que je suis allé quelque part.
— Quand vous avez glissé ?
— Oui. Et je vois...
— Que voyez-vous, Bracken ?
Je prends mon souffle, comme un plongeur en eau profonde. Dans cette aspiration, qui a le goût d'un caramel enrobé de chocolat au lait, d'une enfance reculée, disparue dans l'ombre d'un arbre, il y a tout le malheur de ma vie d'assisté, d'assistant, incapable de prendre en main ma destinée et de me conduire en individu.
— Je vois entre les formes, Plouffe.
— Soyez plus précis.
— Un plan. Je vous vois cerclé d'un trait noir. Comme quand je dessine.
— Et tout est ainsi ?
— Oui, sauf quand c'est trop complexe, quand les formes passent les unes dans les autres, comme votre cafoutche, là.
— Depuis que vous avez glissé dans l'angle ?
— Oui.
— Mmm...
Il penche la tête sur son torse, semble s'endormir. Après un instant de réflexion, il reprend.
— Et si...
— Oui ?
— Et si vous étiez vraiment allé là-bas ?
— Où ?
— Au Néant. Et que vous y aviez disputé une partie contre Bram et que le destin pourrait être changé si nous décidons que ce moment vous appartient, qu'il continue à se dérouler alors même que je vous parle, qu'il influe sur ce qui se passe ici. C'est là-bas que se noue l'intrigue dont nous ne recueillons ici que les échos, que des oracles à interpréter, comme de la bouillie de rêve.
Tout se remet à trembler. Je m'accroche à son fauteuil, de peur de retomber, mais la secousse persiste, semble différente.
— Une réplique ?
Plouffe lève le nez vers le plafond.
— Nik Kershaw.
KOR
La dernière ligne tracée
Enfin, dominée,
Par le semblant de sens.
Nik Kershaw sort d'une modeste coccinelle jaune fluo, drapé dans un grand manteau noir, comme des ailes de bébé dragon ; la crête peroxydée qu'il porte sur la tête oscille en vagues quand il sourit. Suivent trois grosses Islandaises avec à leur tête Erla qui trace son chemin parmi les fées pour se pâmer devant lui, ivre de glamour.
— Oh, Nik...
Nik regarde les falaises où courent les enfants, la nuit qui coule. Il claque des doigts : un secrétaire se dépliant hors l'habitacle du minuscule véhicule, lui tend un chewing-gum à la cannelle.
— Le soleil devrait se montrer dans une heure ou deux, dit Erla, le temps de vous mettre à l'aise et de profiter de nos installations...
Fink a fait ce qu'il a pu pour organiser la kermesse, mais le champ de lave ressemble surtout à un terrain vague : des jeux de ballons approximatifs, quelques fanions, les gâteaux artisanaux encore emballés de feuilles d'aluminium, un trampoline, quelques parents qui échafaudent une estrade, des lampes-torches montées sur ressort. Pas de pêche aux canards, semble-t-il, alors que je traîne Plouffe vers le parking en hélant le chanteur :
— Nik ! Nik !
Il se retourne vers nous, intrigué.
— Nik, dis-je, haletant. Je suis Bracken.
Il bâille. Erla me foudroie du regard.
— Et voici le proviseur, continué-je, indomptable.
— Monsieur Kershaw, dit-il en lui serrant la main. C'est un honneur de vous avoir parmi nous, sachez que nous ferons tout pour vous épargner les...
Je lui tape dans le dos en riant, coupant net sa diatribe.
— Sacré farceur ! dis-je, en lui intimant des yeux de se taire. On vous épargnera trop d'enfants, bien sûr.
Je me penche à l'oreille de Plouffe.
— Bon sang, ne dites rien, sinon il va prendre la fuite !
— Compris !
Nik ferme un œil, croise les bras.
— Et voilà Fink, dis-je en voyant le surveillant escalader les rochers dans notre direction gesticulant, le corps couvert de punaises, de lambeaux de tissu, un seau sur la tête.
— Mes hommages, monsieur, dit le vieux, qui menace de succomber à une crise d'apoplexie, rouge comme un coquelicot.
Je ne saurai dire ce qui se passe dans la tête de Nik Kershaw en ce moment précis, mais il ne paraît pas contaminé par notre folie. Il pose les mains sur ses hanches et s'enquiert :
— Où est Elliot ?
Je le regarde, lui, presque mon double en négatif : il porte des gants de cuir et des bretelles sous son manteau.
— Vous connaissez Elliot ?
— C'est pour lui que je suis là.
— Lui ? Mais pourquoi ?
— Oh lalala... dit Fink.
— Nous devons vous parler, monsieur Nik, c'est de la plus haute importance !
— Dites donc, gémit Erla, et les enfants ?
— Fink va s'en occuper !
Je traîne Nik par la main à l'intérieur, alors qu'une meute de fées, carnets ouverts, sollicitent un autographe. Fink est emporté par le tsunami. Sans réfléchir, j'entre dans le bureau de Plouffe, pendant que ce dernier, roulant comme un damné, ferme la porte.
— Qu'est-ce que ça veut dire ? demande Nik.
— Elliot a disparu, dit Plouffe.
— Disparu ?
— Peut-être dans un autre monde.
Nik s'assoit, visiblement très concerné.
— Nik, comment le connaissez-vous ?
— Eh bien, c'est... comment dire... Nous avons organisé un concours, il y a deux ans. À la suite de notre vidéo-clip.
— Quel vidéo-clip ?
— _The Riddle_. Nous avions... Écoutez, je ne comprends pas, vraiment, si je dois tout vous expliquer...
— Nik, dis-je, je ne peux pas vous forcer à parler, mais vous devez me croire quand je vous dis qu'il en va de la sécurité de cette école. Elliot a disparu, personne ne sait où il est et quelque chose est à l'œuvre dans ce bâtiment, quelque chose de très dangereux, qui est lié, nous le pensons, à la disparition d'Elliot. Alors dites-nous ce que vous savez.
Il s'enfonce dans son siège, nous scrute, peut-être pour être sûr d'avoir affaire à des gens sains d'esprit, puis soupire.
— Oh, et puis je suppose que ça n'a aucune importance. Quand j'ai écrit cette chanson, j'avais commencé par chanter des mots sans queue ni tête, juste pour coller au rythme, vous savez, on appelle ça une piste temporaire, mais quand je suis arrivé en studio, je me suis rendu compte que tous les gens qui l'écoutaient étaient d'une certaine façon envoûtés par le sens occulte des phrases, et comme la chanson est une énigme, les paroles en devinrent une pour les auditeurs, lorsque j'ai décidé de les conserver et d'enregistrer la chanson telle quelle. Quand nous avons fini le clip, qui apportait de nouveaux éléments à la réflexion, mon manager a eu l'idée d'organiser un concours, pour proposer la clé de l'énigme.
— Mais, vous avez expliqué vous-même que ça ne voulait rien dire.
— Ce n'est pas parce que ça ne veut rien dire que ça n'a pas de sens. Souvent quand nous improvisons des choses avec beaucoup de concentration et d'intensité, certains appellent ça de la maîtrise, elles finissent par opérer une sorte d'alchimie secrète, quelque chose qui dépasse le créateur. Une sorte d'autonomie si vous le voulez, et cette chanson, ma foi, je crois que c'était la preuve que ces choses-là pouvaient arriver. Appelez ça de la magie si vous voulez, en un sens c'en est. La chanson est devenue numéro un partout, je suis devenu riche, et je ne pense pas devoir son succès et le mien à ses seules qualités artistiques.
— Mais quand Elliot intervient-il dans cette histoire ?
— Elliot avait la réponse la plus folle que nous ayons jamais entendue. Pour lui, tout avait un sens, le moindre élément du clip, la moindre parole, il réussissait à leur trouver une logique que personne n'avait envisagée. Jusque-là, nous pensions juste que cette chanson parlait du devenir de l'homme, en permanence déchiré entre le bien et le mal, et que certains gardiens nous empêchaient de basculer. Mais pour Elliot, la texture de la chanson était infiniment plus complexe. Il nous a livré une véritable mythologie de l'univers dans laquelle la chanson prenait place, argumentée, rationnelle. Et il essayait de comprendre le rôle exact du mystérieux personnage qui m'avait attiré dans ce piège.
Dans le vidéo-clip de _The Riddle_ , succès culte en Islande depuis plus d'un an et qu'on voit régulièrement passer avant le journal de 20h, le chanteur bascule dans une sorte de dimension parallèle où tout devient illusion d'optique. Il y a un passage avec un angle, où les personnes semblent plus petites à mesure qu'ils s'approchent du fond. Maintenant que j'y pense, tout dans ce clip trouve un écho ici : la collection d'objets, le vieil homme disparu, la pièce éternelle qui n'en finit pas d'exister, même mes bretelles.
Et Nik Kershaw porte des gants.
— Dans le clip, il y a cette figure étrange, un mystère, un homme masqué, en collant, qui rappelle le Riddler de Batman. Quand le clip commence, je cambriole un bureau à la recherche de documents secrets, mais il se passe quelque chose et je suis avalé dans un monde parallèle, entièrement régi par cet homme. À la fin du clip, nous comprenons que le monde est un point d'interrogation, littéralement, dont l'homme mystère s'empare.
Je ne sais quoi dire.
— Elliot avait envisagé que l'objet du clip, et de la chanson, était de faire comprendre que tout est surface et que mon œuvre était une clé pour accéder à une sorte de monde dans le monde, SUR le monde, comme une couche invisible. Un peu comme si nous devions marcher sur les lettres des pays figurant sur une carte. Ces lettres existent, elles sont là. La carte devient le monde réel. C'était tellement fou que nous l'avons déclaré vainqueur, avec comme récompense une visite spéciale de notre équipe. Il a insisté pour que ce soit aujourd'hui, à cette date exacte, en ce lieu précis.
— Pourquoi l'Islande ?
— Dans la chanson, je précise que le trou qui permet d'accéder à l'autre monde est gardé par un vieux pêcheur d'Aran, l'île irlandaise. Elliot pensait qu'il s'agissait en vérité de l'Islande, et que cela constituait la seule approximation de la chanson. Nous nous sommes dit que nous pouvions bien fermer les yeux sur ce détail, surtout quand un gamin de huit ans venait de gagner.
Un silence s'installe dans le bureau, brutal. Glacé.
— Huit ans ?
— C'est ce que j'ai dit.
— Mais...
On tape à la porte, et Erla entre, suivie de sa cohorte de grosses, bouffant de leurs énormes doigts des choux au chocolat.
— Bracken ! Maintenant ça suffit.
Entrent trois journalistes et une énorme caméra grise.
— Voilà les gens du journal, et la Sjónvarpið.
— Bonjour bonjour, dit Nik en se levant. Désolé pour cette petite pause, je devais vérifier quelque chose avec ces messieurs.
Erla me tire en arrière, furieuse.
— Dites donc Bracken, à quoi est-ce que je vous jouez ?
— Eh bien...
— Nik Kershaw a été invité par la ville, pas par l'école, il n'est pas à vous. Les gens du _Frettabladid_ vont arriver pour l'interviewer, et je ne veux pas passer pour celle qui a privé le peuple islandais de la parole d'un tel artiste.
Elle le prend par le bras, les journalistes l'assaillent de questions en anglais, une lumière est braquée sur son jeune visage. Avant de disparaître dans le couloir, Nik se retourne vers moi, sourcils froncés.
— Trouvez Elliot.
Puis il est emporté. Nous restons silencieux tous les deux. Quand tout le monde est parti, Plouffe se met à rouler en cercles dans le bureau.
— Bon sang, mais c'est bien sûr !
— Eh ?
Il débusque un vieux tome roussi, l'ouvre pour chercher un article, s'arrête brusquement sur une photo.
— Bracken, regardez !
Sur la photo, des enfants sur une pelouse font des mouvements incongrus. Chacun différent. Au centre, une balle de croquet, des dinosaures en plastique et une peluche de phoque.
— Je ne comprends pas.
Il cligne de l'œil droit, comme s'il voulait me dire quelque chose de nouveau, quelque chose de subtil que les mots ne sauraient traduire.
— Paidia !
— Paia ?
— Le jeu improvisé, sauvage. Le jeu des enfants dans la cour de récréation. Le jeu des fées et des loutres. Un état de grâce, où tout est possible, où toutes les contradictions se résolvent dans la joie de l'imaginaire.
— Que voulez-vous dire, qu'Elliot voulait...
— Si Elliot voulait aller au Néant, là où le sens n'existe pas, il ne pouvait pas avoir les réponses ! Il les créait au fur et à mesure ! C'est comme ça qu'il est entré dans la tapisserie, qu'il a remonté plus loin que sa naissance, rejoignant l'idée qui avait précédé... Il a organisé un jeu, à l'échelle de sa vie, puis de l'école, qui devait servir à créer une sorte de continuum de _nonsense_ , de jeu en évolution... Les objets de la chambre sont une série d'indices sans but, qui n'existent que pour former un sens émergent, lui permettant de trouver le sens progressivement, en l'adaptant à ses besoins.
— Ce même jeu qu'essayait de pratiquer Bram ?
— Avec ses ordinateurs et ses tralalas ? Je ne sais pas, peut-être qu'il essayait de suivre Elliot, de l'imiter, mais qu'il lui manquait quelque chose pour avancer.
— Mes yeux. Moi, dans l'angle. Il comptait sur Elliot pour l'emmener, mais Elliot voulait être Maître, donc Bram s'est servi de moi. Sans moi, il n'aurait jamais pu vivre ce jeu sans écran. Un ordinateur, c'est limité, ça ne fait que ce qu'il est programmé pour accomplir.
— Et Bram a réussi.
— Et ce serait à cause de cela que tout se décolle ?
— Je ne sais pas. Je ne pense pas. Peut-être est-ce un effet secondaire.
— Et moi ? Pourquoi je vois des silhouettes ?
— Les silhouettes sont le Néant, Bracken. C'est la ligne invisible qui soutient toute chose. C'est ce qui sépare le ciel de la mer, c'est l'horizon que nous n'atteignons jamais. Peut-être qu'en ayant voyagé dans le Néant vous avez appris à le voir partout.
— Mais je n'ai été nulle part !
— Vous ne vous en souvenez pas, c'est tout.
— Je n'ai pas bougé !
Il soupire, visiblement très agacé.
— Je vous l'ai dit Bracken, le Néant n'est rien, il ne contient pas de temps, pas tel que nous le connaissons. Il peut prendre la forme d'un moment étiré à l'infini. Ou n'être qu'un micro-instant. Vous avez peut-être passé une vie entière là-bas, une vie dont vous avez tout oublié. Mais qui vous reviendra. Vous avez hérité d'elle. Elle vous a changé. Elle continuera à le faire.
Je plonge en moi pour trouver un écho à ses paroles. Je conjure des images simples de plages et de rochers, ce que j'ai vu ici, dans ce pays, ce cliché que l'on sert comme un ailleurs, un bout de monde. Une plage de sable noir, une bande, une ligne entre le ciel et la mer. Ai-je vécu sur la ligne d'horizon ? Ai-je été le touriste d'un monde impossible, si simple à décrire qu'il suffirait d'une seule ligne dans le roman de ma vie, cette vie que j'observe comme étranger à moi-même, parlant comme à quelqu'un qui n'existe pas, à moi qui m'écoute penser ?
Quelque chose me lance brusquement dans le doigt.
— Aïe, aïe, aïe...
— Quoi, quoi ?
— Ça fait mal !
Sous mes moufles. Quelque chose. Qui me brûle.
— Enlevez votre petite moufle, Bracken !
— Non !
— Voyons, ne faites pas l'enfant !
Il me saisit la main, tire sur la moufle. Je ne veux pas la retirer. Elle m'a protégé jusque-là. C'est une sécurité, je me sens bien avec des moufles, comme ça, je ne peux pas... Mes moufles, c'est ma vie, mes moufles, c'est moi, elles sont tricotées avec amour, elles me ressemblent tellement, elles...
Plouffe ôte la moufle. Ma main est toute blanche.
Sur la dernière phalange, il y a à mon majeur, une callosité, comme une petite bosse. Rougie, irritée, formant un cercle non refermé. Une blessure en train de cicatriser.
— Où est-ce que vous vous êtes fait ça ?
— Je ne sais pas. On dirait...
...m'empêchent de dessiner. Comment dessiner avec des moufles ? C'est impossible, on ne peut pas, ou alors, on devient naïf, on devient débile. On n'a pas le contrôle, le trait ne va pas là où l'on veut. Je porte des moufles parce que j'ai peur de dessiner. Parce que j'ai peur de me confronter à moi-même.
Je passe ma main dessus. Doucement, la blessure s'illumine. Elle n'a pas de silhouette, pas de forme. Juste une couleur, qui pulse, qui fait de moi...
Quelqu'un de nouveau.
Les yeux de Plouffe brillent.
— Alors, c'est fait, dit-il. Elliot a lâché son trône.
Il me regarde, presque en pleurant.
— Vous pouvez devenir Maître.
Il laisse éclater sa joie en roulant dans la pièce.
— De quoi parlez-vous Plouffe ?
— Oh, si vous saviez, mon ami... Je ne sais comment c'est arrivé, mais c'est la preuve que j'attendais. Bram n'a pas gagné. Il reste une possibilité.
— Laquelle ?
— Que vous ayez déjà remplacé Elliot au Néant.
— Je n'ai remplacé personne.
— Vous ne le savez pas, vous ne vous en souvenez pas, mais vous l'avez peut-être déjà fait.
— Quand ?
— Dans ce moment imperceptible où vous avez glissé, dans le pli du mur.
Il a peut-être raison. Je sais qu'il s'est passé quelque chose. Je sais que je suis revenu d'un lieu où je ne suis pas allé. Une boule de terreur me remonte dans la gorge. Une boule de haine, qui gonfle, menace de m'étouffer.
— Mais alors, Plouffe... Vous saviez ?
— Ah, j'aimerais vous dire que je sais tout, que c'est moi qui aie tout manigancé, la disparition d'Elliot, votre passage, Bram et ses jeux stupides... Mais non. Simplement, comprenez : Mallarmé est une obsession, j'ai toujours voulu mettre au jour ce qu'il a vu, ce jour où il affirmait, dans ses lettres, que sa "pensée s'était pensée". Je ne suis pas un dévot, Bracken, mais je crois que la poésie est la forme la plus pure de spiritualité, à la hauteur de l'homme et à hauteur d'homme. J'ai essayé d'enseigner tout ceci à Elliot, et il est allé beaucoup plus loin que je ne l'ai fait moi-même. Il a fait d'un poème un mythe, il a accédé physiquement à ce mythe, car il était suffisamment doué pour voir ce que nous autres ne pouvons déceler. Bram a essayé de le copier.
Et moi, pendant tout ce temps, c'est dans les livres que je cherchais la réponse, sans me douter que tout était beaucoup plus simple.
Sa moustache rebique, il tousse, puis il poursuit dans un français parfait, théâtral :
— Depuis que j'ai commencé à travailler sur l'œuvre de Stéphane, que j'ai compris que la Pénultième était la clé du sonnet en yx. Sur un simple plan littéraire, Mallarmé y raconte comment naît l'inspiration, sous la forme d'un mot qui devient un monde et qui ordonne les figures, les allégories, les métaphores. D'où vient ce mot ? L'enfant en moi a toujours su qu'il existait, quelque part, un lieu où les poètes allaient puiser leur inspiration. Non pas visités par les Muses, mais visitant l'île des morts, où les cyprès chantent la vérité de la matière. Dans leur esprit, peut-être, pendant leurs rêves, après leur mort. J'ai écrit des pages et des pages sur le sujet, et j'ai cherché ce lieu partout. Voyez tous ces livres, ces carnets. Mes échecs. Je suis un piètre poète. Je suis un raté, Bracken. J'ai essayé d'être un enseignant mais encore une fois ce n'était que pour poursuivre mes chimères. Peut-être ai-je voulu trop essayer. Peut-être, finalement, n'y ai-je jamais vraiment cru. Après tout, un monde magique ne se trouve pas en claquant des doigts. Au fond de moi, j'ai toujours désiré le Ptyx, j'aurai voulu le brandir moi aussi. Je voulais être le nouveau prince des poètes. Je voulais me rendre au Néant et trouver le sceptre qui me permettrait de montrer les choses sans les dévoiler, sans les dire, une par une, comme des révélations offertes au monde. Je voulais me servir de tout ça pour expérimenter de nouvelles méthodes éducatives. Je voulais rendre aux enfants une pensée archaïque, loin de l'arrogance raisonneuse, de la pensée scientifique. Je voulais être capable de leur montrer le vrai soleil, là où s'incarne la vie divine, la nature enceinte d'elle-même. Mais moi, je ne suis pas poète, non. Elliot, lui, l'était. J'ai reporté tous mes espoirs en lui. Je ne savais pas qu'il réussirait à trouver le chemin du Néant. Peut-être le saviez-vous, Mallarmé avait un fils, Anatole. Le pauvre petit est mort très jeune, après que Mallarmé eut découvert le Néant. Je crois que si Anatole avait vécu, il aurait pu être Elliot. Ce même Elliot dont j'ai farci la tête de Néant et de Ptyx. De contrôle. Pour un autiste comme lui, le contrôle est la clé de la stabilité. Ça le rassure. Il en a fait une quête personnelle, et moi, je n'ai rien vu, ou je n'ai rien voulu voir, tellement j'étais sans doute jaloux.
Je pense : « Il est fou. »
Puis : « Je l'ai toujours su. »
— C'est une métaphore, Plouffe.
Il sourit.
— J'oubliais, vous ne pensez pas en mots, mais en images. Oui, je dois vous avouer que je me sens dépassé. Regardez-moi : je suis si vieux dans ma chaise avec ma vieille poésie et mes vieilles habitudes. Je n'ai même pas compris ce que voulait faire Bram. Je l'ai laissé œuvrer dans les ténèbres, passer un pacte tacite avec Elliot. Et moi, je suis resté dans mes livres, à essayer de comprendre comment suivre le même chemin que Stéphane, mais c'est impossible, n'est-ce pas, un véritable Maître doit posséder sa propre vision, c'est à ça qu'on le reconnaît... Elliot, lui, a compris, et Bram aussi, d'une certaine manière. Ils ont chacun procédé à leur façon. Elliot a communiqué ce qu'il savait à Bram, qui vous a utilisé pour suivre le chemin qu'Elliot avait creusé dans sa chambre pour quitter de cette réalité. Vous êtes allé là-bas, il s'est passé quelque chose. Vous êtes revenu. Pas Bram. Tout ce que je sais, c'est que moi, je n'ai rien fait. Je suis resté sur le rivage, je mérite mon sort.
Je le regarde et je comprends sa douleur. Comme Elliot, il se cherche un but, un dernier effort à accomplir, quelque chose à léguer avant de disparaître, de glisser dans l'angle de la vie et de se replier.
— Ne m'en voulez pas, Bracken. Si je vous avais tout dit, vous m'auriez pris pour un fou. Et même moi, je n'étais pas sûr de ce qui avait pu se passer.
— Vous êtes fou, mais vous n'êtes pas méchant.
— Vous me croyez ?
— Je ne sais pas. Je ne sais plus.
— Bracken, prenez mes mains.
Il me les tend. J'hésite, je n'ai plus mes moufles.
— Bracken, je vous en conjure...
Je les prends.
— Regardez, Bracken.
Ces vieilles lignes, dans ses paumes, comme des rides. Non, comme des traits, des visages, affaiblis, diminués, mais toujours là. Comme une dédicace. Elles se croisent et se décroisent telles des routes vers nulle part, des destins avortés. Je n'y lis aucun nom, aucun mythe, aucun passé, pas plus que je ne lis de personnalités dans les constellations. Ce ne sont pas des noms humains qui permettent d'envisager leur puissance. Les aplatir sur une surface n'en fait pas une image. Et ces lignes, sans dimensions, à peine un sillon, sont bien plus que des souvenirs. Ce sont des parfaites sections d'arabesques enlacées, rieuses, des tresses à la douce texture. Je pourrai les fixer des heures, y bâtir des cabanes. Si j'osais, je les déplacerai.
J'avance un doigt.
Faiblement, Plouffe sourit.
— Je me suis trompé, Bracken.
— Trompé ?
— Vous êtes déjà le Maître.
Je le vois réellement peiné. Comme si son ultime plan lui échappait. Il ne s'attendait pas à ce que je dise non. Le pouvoir ne m'intéresse pas. Je ne suis pas maître de quoi que ce soit. Je ne suis même pas maître de moi-même.
— Il y a des choses que vous savez faire et dont vous ignorez tout. Ces mains, Bracken, regardez.
Sous mes yeux, les lignes de vie pulsent, s'illuminent. Certaines me gênent. Certaines demandent à être effacées pour être renforcées. Je vois la surface entre elles, de nouvelles formes apparaissent, le plein, le vide, tout s'inverse, et la nature s'entrouvre, me laissant une tentative d'interprétation.
— Allez-y Bracken, m'encourage Plouffe.
Du pouce, j'en repousse une, elle s'entortille dans une autre, créant un nouveau motif, plus séduisant. Puis, de l'index, j'efface le trop-plein. Reconfigurées, certaines nouées, les lignes composent désormais un paysage que je ne connais pas, une fausse perspective qui se construit à l'intérieur de la main, une fenêtre ouverte sur un nouvel horizon, que j'aimerai parcourir, libre, enfin libre.
— Mmmmh..., fait Plouffe.
Ça n'a pas l'air de lui faire mal. Je déroule, je trie, je repousse, j'efface. Je veux en savoir plus sur ces ruisseaux, sur ces bosquets, sur ces montagnes au loin, couronnées d'ombres et de brumes, ces forêts denses de sapins, ces sentiers qui rejoignent le soleil, je veux tirer vers moi la perspective, la possible...
Soudain, à vouloir aller trop vite, une ligne casse. Simplement, avec un petit « crac ». Tout le reste se dévide, comme une bobine. Pour bien faire, il faudrait l'enlever complètement, cet excédent. Je tire dessus, comme on enlève un poil trop noir, un cheveu trop blanc. Mais la ligne n'a pas de fin, elle croise d'autres traits puis fusionne en un trait noir qui détoure le majeur. Puis l'index, puis le pouce.
Sans le savoir, en silence, je défais Plouffe.
— _Bon ça va, je crois qu'il est sauvé._
— _Il était moins une !_
— _Hip hip !_
Allongé sur le dos, empesé d'eau glacé, je tremble convulsivement. Je me roule en boule pour avoir moins froid, mais c'est Egill, tout chaud, qui vient me couver en roucoulant. Je n'arrive pas à faire le point, un flou devient doucement contexte. J'interromps le claquement de mes dents en me mordant la langue. Je trouve la force de me ramasser, pour m'asseoir par terre. Je prends Egill dans mes bras. Je pose ma tête contre sa fourrure, pour retrouver le chaud d'une intimité, quelque chose de doux, qui me sortirait de tout ça.
Une vaste caverne est creusée dans l'iceberg. Un trou m'a permis d'y accéder. Un trou plus grand qu'il n'y paraissait, à moins que je ne sois devenu encore plus petit ? Je crois comprendre. J'effectue un voyage vers l'infiniment petit, pour approcher ce qui est caché dans l'épaisseur du trait, au plus profond de celui-ci. Un voyage au bout de moi-même, année après année, jusqu'à la plus petite instance de moi, la mort, ce mystère insondable, qui prend toutes les formes qu'on lui prête dans le secret du sommeil, dans la paix des cimetières.
Un terrible grondement roule derrière les parois transparentes de l'iceberg. Je vois défiler des formes sombres, peut-être des orques. Près d'un trou dans le sol, étendu tel un pacha sur le miroir de glace de la pièce, rond, vaste, si vaste, qu'on n'en voit pas le bout, Hinrik avale goulûment un poisson en retirant l'arête entière de sa bouche. Je suis si heureux de le voir en vie.
Lui aussi a droit à son câlin.
— Oh lala ! Briki, quelle aventure ! chantonne Hinrik.
— Comment avez-vous fait ?
— J'ai nagé, j'ai embarqué le petit avec moi.
— Hinrik, petit cachottier !
Le morse ronchonne. La caverne forme une cathédrale, miniature de cette grotte où la mer et la plage se rencontrent, où j'ai passé les dix dernières années de ma vie. Combien de temps vais-je rester ici ? Le double ? Une éternité ? Vais-je trouver le Cap ? Je vais finir dans un trou, un tout petit trou. Une reproduction de tout le reste. De ma vie, de ce trou que j'ai creusé pour m'y enfouir.
Hinrik se pose sur ses fesses, se grattouille la moustache.
— Attention où vous marchez, Briki !
Je regarde sous mes pieds et j'avise les deux petites tortues, toutes joyeuses. Deux tortues sorties de mon passé, il y a dix ans, quand j'avais quitté cette chambre marine pour atterrir ici. Deux adorables créatures aux secrets impénétrables.
— Les tortues !
— Il faut croire que vous avez beaucoup d'amis, mon gars !
— _Oh, de rien Monsieur !_
— _Il s'appelle Bracken !_
— _C'était sympa, cette fête !_
Je les prends dans mes mains.
— Merci les tortues.
Je crois deviner leurs cils, qui battent quand elles me regardent, presque avec amour, mais je sais combien la nature est traîtresse. J'aimerai leur donner de l'eau, mais je n'ai aucun récipient sous la main. En attendant de trouver une solution, je les glisse dans ma poche.
— Ah, Bracken, ces petites bêtes sont stupides, ne vous embarrassez pas d'elles.
— Elles m'ont sauvé la vie, je ne les abandonnerai pas.
— Où voulez-vous aller ?
— _Il me dit quelque chose..._
— _Mais qu'est-ce qu'il fait ?_
— _Il nous met au chaud... Qu'il est chou !_
Dans ma poche, je sens mon crayon tout mordillé, tellement taillé qu'il n'en reste qu'un tout petit bout. Je ne sais pas ce que je ferai quand il sera totalement usé. Juste de la poussière pour dessiner. L'envie me prend de tracer de grandes lignes sur ces surfaces pures, cette glace réfléchissante, plongée et contre-plongée sur nous, seuls au centre de ce grand tout. Mais je ne veux pas user la mine.
— Où sommes-nous ?
— Allez savoir, rote Hinrik. Je n'ai jamais vu d'iceberg de cette taille. À chaque nouvelle marée, des fragments de banquise se disloquent de Ginnugagap. Il y a parfois des ours blancs qui dérivent dessus, ils viennent s'échouer sur la plage. Ils finissent mangés par les crabes.
Je repense au château de sable, me disant que, peut-être, ces architectures ont pu être bâties entre les côtes d'un ursidé. Je n'en suis plus à une déduction près. Mais d'où viennent les ours blancs ? D'où serait venue la première tortue, la belle tortue ? Si le Néant à besoin d'un Maître pour exister, qui était le Maître avant le premier Maître ? Qui a pu créer ces choses qui échappent à notre entendement ? L'Antépénultième... J'ai souvent posé la question à Hinrik, pendant toutes ces années à m'imprégner de leurs coutumes. La réponse était toujours la même : le mystère.
Je remarque une tache plus sombre dans la paroi. En m'approchant, je comprends que mes yeux m'ont joué un tour : ce n'est pas une ombre, c'est une torche. Comme celles en plastique que j'avais vues sur l'îlot avec un feu découpé dans le papier. Quelqu'un a placé ces torches et elles ne correspondent à rien de ce que j'ai pu voir dans ce monde-ci. Ni par sa texture, ni par sa fonction. Pourquoi placer des torches là où il n'y a rien à éclairer, _a fortiori_ si elles ne produisent aucune lumière ?
« — Passons par en dessous, nous avons la position de l'iceberg.
— _Je ne veux plus attendre. »_
Je crispe les poings. Bram. Bram et ses images, sur l'ordinateur. Il était avec moi, devant le Pli. Il est passé mais il ne s'est pas échoué. Se pourrait-il qu'il ait pu s'accrocher à cet îlot, comme une bouée, parce qu'il avait pris soin d'y planter un repère, une borne... un totem... un relais... une balise. Quelque chose qui lui aurait permis de... Non, c'est impossible. Qu'aurait-il fait quand la marée est montée, à peine quelques minutes après notre arrivée ? Il n'aurait pas pu survivre... À moins que... que l'iceberg lui-même ait pu le localiser...
— Briki ? demande Hinrik. Tout va bien ?
— BRAM ! ! ! hurlé-je.
— _Mais qu'est-ce qu'il a ?_
— _Je crois qu'il a mal._
— _Il brame._
Je hurle à nouveau.
— BRAM ! ! ! !
L'écho de la caverne ne me renvoie qu'une rauque voix d'adulte, et non celle de cet adolescent attardé que je croyais être encore. Tout me revient à présent. Ils ont joué à mettre en place ces choses dans leur univers virtuel, comme des jalons, des bornes auxquelles se raccrocher en cas de naufrage. Mais je ne comprends pas comment ces objets ont pu atterrir ici, au Néant. Il s'agit d'un jeu américain très éloigné de cette aride abstraction poétique, sans plage, ni morses. Non, un monde presque réel, pour le peu que j'avais pu en voir. Des villas dans une résidence, une place ornée d'une fontaine... une plage avec ses palmiers. Comment des objets disposés dans un jeu pourraient-ils soudain apparaître ici, dans ce monde où je vis, dans ce monde où rien n'est virtuel, où tout est si réaliste, où tout à l'air plus ridé, plus défini. Ce n'est pas un monde virtuel. Ce n'est pas un monde de pixels où, passé un certain point, il est impossible de zoomer. Ces choses ne sont pas un assemblage de carrés. Ce sont des réalités continues, qui nous permettraient d'aller jusqu'à l'atome, et au-delà.
Le morse traîne sa carcasse jusqu'à moi.
— Briki, calmez-vous, il n'y a personne.
Non. Au loin, dans la caverne, il y a une tache verte qui bouge, minuscule, et qui est apparue quand j'ai hurlé. Bram. J'en suis sûr. C'est lui. Il a répondu. Il savait que quelqu'un viendrait.
Je me précipite, mais, zip zip, je glisse sur la glace, elle m'emporte au loin, je me prends un des murs. Ça fait très mal.
— Aïe !
Le morse a galopé, l'oiseau sur son dos. Dans ma poche, les tortues s'accrochent pour ne pas tomber.
— Briki, enfin, tenez-vous !
Je pointe la tache verte, qui s'agite.
— Hinrik, là-bas, regardez !
Le morse met sa nageoire en visière, bougonne. Egill s'agite, perd quelques plumes.
— Allons voir ça. Montez Briki !
Je grimpe sur le dos du morse, qui s'élance sur son ventre. Je prend son chapeau avec les tortues. Rapides, nous filons vers l'extrémité de la grotte, la glace défile réfléchissant notre image en accéléré, des lignes qui strient tout, deviennent pure expression de vitesse, forment un torrent. Le morse prend appui sur ses nageoires pour se propulser, notre ménagerie tient bon mais nous nous écrasons quand même, et nous valsons comme les quilles d'un jeu de bowling. Je compte les oiseaux qui voltigent autour de ma tête, il n'y en a qu'un, un macareux qui mâchouille, et deux tortues, qui rient aux éclats.
— _Ah ah ah ah !_
— _Qu'est-ce qu'on s'amuse !_
— _Fun ! Fun ! Fun !_
En me relevant, je constate que la tache n'était pas une illusion. Il y a une porte, au fond de la grotte. Une porte ronde, en bois, peinte en vert. J'aurai juré la voir se refermer sur l'étrange silhouette verte.
— On dirait que cet iceberg est habité.
Soudain, le morse panique.
— Oh lala ! fait-il.
— Quoi ?
— Et si nous étions déjà au Cap ?
— Je croyais que le Cap était une forteresse sur un volcan.
— Et pourquoi pas un phare au sommet d'un iceberg... Qui sait ? Nous n'avons jamais vu le Cap, il peut prendre toute sorte de formes...
— Ce serait pour ça qu'il n'est jamais au même endroit ! dis-je, comprenant soudain que mon intuition concernant Bram est fondée. Il dérive !
Je me penche pour toquer à la porte.
— Toc, toc !
Personne ne répond.
— Je vais ouvrir, reculez.
Hinrik et Egill se retirent pour me laisser pousser la poignée.
— Attention, Briki...
— Attention à quoi ?
— Si nous sommes au Cap, alors...
J'ouvre la porte. Elle n'est pas verrouillée. Elle ne possède ni serrure, ni loquet.
Un courant d'air. Sombre. Lourd.
— Alors le Kor est ici, avec nous.
Je ne sais toujours pas ce qu'est le Kor. Après dix ans passés en compagnie de leurs légendes, de leurs mythes et de leurs rêves, je me suis fait dans l'idée qu'il pouvait s'agir d'une sorte de cerbère, de monstre terrible tout en griffes et en dents, une créature à la fourrure noire, sortie d'un cauchemar grec, défendant un trésor appelé Ptyx. L'équivalent de cette horrible chimère, ou de cet agrégat de terreur que formaient les Euménides quand elles devaient accomplir une vengeance. Je me vois soudain dépassé par la taille formidable d'un adversaire auquel je ne suis pas préparé. Je n'avais pas pris tout cela au sérieux. Si nous sommes bien au Cap, là, tout au bord de Ginnungagap, là où Elliot a fait son nid, où il exerce son pouvoir sur le Néant, alors il va me falloir devenir plus fort. Plus solide.
— _Kor, Kor, ça me dit quelque chose..._
— _C'est pas un genre de pingouin ?_
— _Non... je crois que c'est beaucoup plus gros._
Ténèbres.
— Quelqu'un a un briquet ?
— J'avais des allumettes dans le bateau, répond Hinrik. Mais il a coulé.
— Bon.
Je passe en premier, j'entends les petites pattes d'Egill derrière moi. J'avance à tâtons, pour chercher un interrupteur, même si je ne pense pas qu'il puisse y avoir de circuit électrique dans un iceberg. Ni de porte non plus d'ailleurs. Dans le grand hall, nous pouvions apercevoir la lumière du jour à travers la glace. Mais ici, il y a comme des tentures sur les parois.
Ma main frôle un tissu, je l'écarte. Un rai de lumière envahit l'espace révélant de la roche sur les murs et des cristaux dissimulés derrière l'étoffe, qui nimbent désormais les lieux d'une lueur lactée. Une chambre, qui ressemble étrangement à celle que j'ai quittée il y a des années. La chambre d'Elliot. Mais différente. Ce n'est pas une chambre, c'est un long couloir, encombré de meubles, de fioritures, d'objets incongrus.
— Comment vais-je faire pour entrer ? demande Hinrik, la tête dans l'embrasure de la porte.
— Hum.
J'essaye de le tirer. Il pue. Il est énorme. À mes côtés, Egill tente de m'aider, mais un macareux n'a pas de mains, ses ailes trop petites parviennent tout juste à agripper la moustache du morse.
— Allons bon, bougre d'oiseau de malheur !
Il soupire, constatant que nous n'y arriverons jamais.
— Je n'irai pas plus loin, mon petit.
— Je ne peux pas vous laisser, Hinrik.
Je le serre dans mes bras. Il est si gros que ça fait « schpouik ».
— Prenez soin de vous mon petit Bracken. Laissez-moi mourir...
— Vous êtes mon ami, Hinrik.
Il rougit.
— Poussez-vous.
Il décoince sa tête. Je sors, je me place derrière lui, je le pousse du pied. Il est si gros que sa masse doit se contracter pour passer entre les montants, mais bientôt, avec le bruit d'un bouchon de champagne qui saute, son postérieur passe l'obstacle. J'entre à sa suite, il prend toute la place, il a du mal à tenir dans le couloir. J'ai l'impression d'en faire trop, à me trimbaler ma ménagerie, tous mes amis, dans le même espace vital que moi, comme si j'avais besoin d'eux pour exister, ici, chacun correspond à une partie de moi, de ma personnalité. Là, ici, au bout du Rien, la pire des choses serait d'être séparé d'eux. Jamais je ne pourrais me résoudre à les perdre. Le destin ou la quête n'ont rien à voir là-dedans. Si je dois retourner d'où je viens, et pêcher jusqu'à la fin de mes jours pour demeurer avec eux, je le ferai. J'ai été stupide d'imaginer que je devais à tout prix retrouver Elliot. Je n'ai jamais su pourquoi, j'ai toujours couru ainsi après quelque chose dont je n'avais pas vraiment besoin. Si je suis ici, aujourd'hui c'est parce que je me suis menti si fort que je l'ai cru. Je me suis persuadé que j'avais besoin d'Elliot. De ma vie d'autrefois. J'ai cru qu'il y avait quelque chose en moi qui dépendait de lui. Mais peut-être est-ce là ma leçon : si Elliot est le vieil homme que je me refuse à devenir, lui courir après, c'est rester toute sa vie un enfant à la poursuite d'une lubie.
— _Attendez, je me souviens, le Kor..._
— _Quoi ? Quoi ?_
— _Le dévoreur de tortues !_
Nous traversons le couloir en silence, hallucinés par les bibelots entassés. Des loupes, d'étranges statues orientales, des sortes de bouddhas, des jouets, des bilboquets, des yeux en cristaux sur de petits piédestaux. Le plafond incliné donne aux lieux un air de catacombes, toutes de roche cristalline, parfois comblées de stuc blanc dépoli, très propre. Des gravures représentent des fées, d'innombrables petites fées qui dansent autour de fougères, de rochers. Des lumières pastel, sur des ciels inquiétants. Plus loin, un sombre volcan, qui fume et crache vers le haut ses miasmes de ténèbres, la suie, les cendres qui collent, la pluie qui achève de tout transformer en encre. Nous devons traverser une toile d'araignée en fil de soie, il y a des coussins géants, cousus de fils de toutes les couleurs, il y a des mannequins de couture et des tubes de laboratoire, des cages vides, des maquettes de montagnes, des bougies fondues, quelques bustes d'hommes illustres...
— _Oh ! Vous avez raison !_
— _Le Kor, qui a mangé toutes les tortues du Néant !_
— _Le monstre horrible aux dents de lait !_
Egill s'est juché sur une armoire, face au portrait d'une tortue dessinée avec de longs cils. L'arrière-plan est un crépuscule merveilleux, dans les tons rose et bleu. J'ai l'impression que des trous sont pratiqués dans les yeux. Comme pour m'espionner. Mais s'il y a eu de vrais yeux derrière ces orbes, ils n'y sont plus.
— La Belle Tortue, commente Hinrik, plein de sagesse.
— _Mais, mais..._
— _C'est..._
— _Maman ! ! !_
Je cligne des yeux.
— Vous voulez dire...
— Elle a été invoquée par le Pénultième, le premier Maître du Néant. La Belle Tortue au sourire fendu. On raconte que le Pénultième a écrit sur sa carapace, avec un bout de bois brûlé.
— Ce serait sa carapace, le Ptyx ?
Hinrik hausse les épaules. Ça me rappelle le mythe d'Hermès, dont la première lyre fut taillée dans une carapace de tortue. Mais écrire sur une carapace...
— Écrire quoi ?
— Allez savoir...
— _Comment ça, maman ?_
— _C'est ma maman !_
— _Non, c'est la mienne !_
Dans un renfoncement du couloir, je découvre une large alcôve, où le plafond se rétrécit. Il y a là un confortable fauteuil anglais sur lequel est posé un châle. À côté d'une table basse, entièrement dégagée. Le fauteuil loge à peine dans l'angle. Il y a un autre portrait contre le mur. Un homme portant barbichette, enroulé dans un châle, le même que celui posé sur le fauteuil. Brodé d'un prénom : Stéphane.
— Le Pénultième ! s'écrie Hinrik.
Je me penche sur le tableau. Une main sort du châle pour écrire. Une main avec une moufle.
— Stéphane, fais-je soudain. Il s'appelait Stéphane.
— Vous le connaissiez lui aussi, Briki ?
Je me souviens des mots de Plouffe au sujet du sonnet en yx. De la description d'un poète abandonné à ses rêves de pureté. Je me souviens de mon pays, que j'ai quitté.
— C'est un homme célèbre dans mon pays.
J'ignorais que Mallarmé portait des moufles.
— _Vous voulez dire que..._
— _Si c'est votre maman, et la mienne aussi..._
— _Oh lala !_
Plouffe avait donc raison depuis le début. Stéphane Mallarmé, le grand poète du Néant, qui avait parcouru les abysses à la recherche du sens de l'essence. C'était donc lui, le Pénultième. Il avait trouvé cet endroit, il avait brandi le Ptyx. Et c'est lui qu'Elliot avait remplacé. Le sonnet en yx était un carnet de bord. Stéphane était descendu, avait conçu le Ptyx, aboli bibelot. Mais pour en faire quoi ? À quoi bon régner sur ce qui n'est pas ? Pourquoi tant d'acharnement à vouloir être quelque part où il n'y a rien, à part des morses, des oiseaux maladroits et des lignes, toujours des lignes ? J'aimerais savoir, j'aimerais comprendre. Pourquoi la tortue a- t-elle été invoquée par la poésie de Stéphane, alors que la seule chose qui lui permettrait d'introduire du nouveau au Néant était le Ptyx, ce même Ptyx donné par la tortue ? Le paradoxe est trop épais. Il doit y avoir un chemin. Il me faut trouver la clé de ce dernier mystère, qui me donnera enfin le sens de la démarche.
Dans le salon, sur les crédences, sont disposés encore d'autres objets. Un éventail aux motifs chinois, reprenant la figure montagneuse de l'éventail d'Elliot. Mais le dragon est ici remplacé par un simple cygne, majestueux calice blanc aux courbes à la fois sûres et alanguies. Il y a un miroir sur lequel apparaît une licorne ruant contre un feu animé. Une coupe toute simple, sans aucun sertissage, semble attendre qu'on plonge dans son eau encore claire des lèvres assoiffées.
Moi, j'y plonge les tortues.
— _Aaah, fraîcheur..._
— _N'en profitez pas pour changer de sujet !_
— _Bon, vous êtes ma sœur, et alors ?_
Et ce châle.
Je prends le châle entre mes mains, le déplie. J'ai envie de m'y blottir. Je le tourne et le retourne. Les motifs se croisent en mosaïque, en petite grille, tel un vitrail de laine. Sur cette grille, entre ces traits, je lis la carte d'un monde magique, aux frontières incertaines. Du doigt, je suis une couture, qui mène au croisement d'un fil et d'une impureté. Brusquement, mû par je ne sais quel sens du destin, j'ouvre le châle, je l'étends sur mes épaules. Sa présence me réchauffe, je souris d'être lové en lui, comme un œuf dans sa coquille. Il m'a manqué ce châle, même si j'en ignorai l'existence.
Egill vient se frotter contre ma jambe. Il lève vers moi un regard plein de questions, ou d'on ne sait quoi, comme s'il redoutait ce qui allait bientôt advenir. Je le prends dans mes bras, j'ébouriffe affectueusement son plumage, il claque du bec.
— _Ça veut dire qu'il ne pourra jamais rien se passer entre nous._
— _Qu'est-ce qui vous fait croire qu'il aurait pu se passer quelque chose ?_
—...
Je sens le morse s'ébrouer dans l'espace réduit du carrefour.
— Briki !
Au fond d'un des couloirs qui mènent au carrefour vient d'apparaître, hirsute, un homme vêtu d'une combinaison de Lycra verte. Il porte des charentaises, un masque à la Zorro sur ses yeux. Il s'est rasé la barbe, il est plus vieux de dix ans, ses yeux ont un éclair de démence, mais je n'ai aucun mal à l'identifier.
— Bram !
— Bracken ! Ah, ah, ah, ah !
Je me rue vers lui, le châle voletant comme une cape. J'essaye de l'attraper, mais il s'enfuit dans le dédale. Il gesticule devant moi, il rebondit sur les murs en riant, comme un chewing-gum. Il a terriblement maigri. Ses fesses flasques ballottent sous son collant. Je ne sais pas pourquoi il me fuit. Pourquoi il ne veut pas m'affronter. S'il est resté ici dix ans lui aussi, enfermé dans un iceberg, alors qui sait ce qui sera advenu de sa raison...
— Bram ! Arrête !
— Qui me suit le cochon d'Inde sans le principe des cordes qui tendent à n'être qu'harmonies dans la bêtise ! Ah, ah, ah !
— Qu'est-ce que tu racontes ?
J'essaye de le retenir par la cheville, il s'esquive, laissant derrière lui une pantoufle. J'ai le temps d'apercevoir la plante de son pied, noire de croûtes.
— Va va va, moi aussi je creuse le vers, je creuse si bas, tellement bas, tout au fond, nous tombons, là où les vers mangent le vers, qui rime avec lui-même, et qui se mange la queue, tout au fond, dans la fosse, avec les vers qui grouillent et qui se mangent la queue, et qui riment entre eux !
Il fait de la poésie, me dis-je soudain, comprenant qu'il a peut-être tenté de devenir Maître en faisant des rimes. Je ne connais rien à la métrique. Je ne sais même pas si c'est faire de la poésie. En tout cas, il fait son possible pour s'en approcher, privé du secours de son ordinateur.
— Bram ! Je ne te veux aucun mal !
Il se retourne. Une lueur passe soudain dans ses yeux.
— Nous sommes deux, Bracken. Toi, moi, ici, dans le couloir, à la recherche du Ptyx. Mais sais-tu seulement pourquoi tu le cherches ?
— Je veux juste rentrer chez moi !
— Ou retrouver Elliot ?
— Tu sais où il est ?
— Tu as regardé dans tous les petits trous ?
Je me souviens, de ce trou dans la paroi, sur la plage. Tout au fond de la caverne, là ou une odeur de pâtisserie m'avait attiré. Un trou trop petit pour moi. Mais peut-être pas pour Elliot. Aurait-il déserté sa fonction de Maître ?
— Si Elliot est parti avant la marée, pourquoi ne pas avoir pris sa place avant, dis-moi Bram, pourquoi c'est pas toi le Maître ? Qu'est-ce que t'as fait pendant dix ans ?
Il rit, encore. Puis son sourire s'efface.
— Le lag !
— Le lag ?
Encore un mot islandais que je ne comprends pas. Je sais que Bram prend un malin plaisir à me faire sentir ma différence. Mais nous ne sommes plus dans son île, nous sommes au Néant, un terrain neutre où la langue n'a pas de frontière.
— La différence de temps entre ici et là-bas, en tenant compte du décalage avec les serveurs aux États-Unis.
— Mais de quoi est-ce que tu parles ?
— Qu'est-ce que ça peut bien te faire, nous sommes là, c'est ce qui compte et, que tu le veuilles ou non, tu es aussi un prétendant. Te cacher derrière Elliot ne sert plus à rien. Tu dois assumer.
— Pourquoi ?
— Elliot est parti, il a abandonné le Ptyx.
Ainsi, l'oracle avait dit vrai, malgré ce qu'avait cru Hinrik.
— Elliot est arrivé au Cap, il a vaincu le Kor, il a régné une journée, puis il est reparti. Il n'avait pas vraiment envie d'être Maître. C'était prévisible. Il n'était pas fait pour ça. Il désirait autre chose : la vie éternelle, dans un monde affectueux. Il ne voulait pas créer, il voulait être en paix. Il a rejoint l'autre côté du Néant.
— Sa chambre ?
— L'autre autre côté, de l'autre côté du mur de la caverne.
— Je ne comprends pas, Bram.
— Tu n'as jamais rien compris.
Je m'assois, éreinté. L'image d'une couette molle danse devant mes yeux. Glisser dans l'éternité sans se poser de questions, sans mystère, sans quotidien, sans charge ni devoir. J'ai envie de me mettre à pleurer, mais je ne suis pas assez triste et puis il y a trop d'eau autour de moi !
— Bram, quelle folie.
Il penche la tête sur le côté.
— Folie ?
Il s'accroupit près de moi, regarde à droite, à gauche, en murmurant.
— Je joue au fou. C'est la seule solution sinon le Kor va me prendre, et il te prendra aussi, s'il pense que tu es capable de voler le Ptyx.
— Je ne veux rien voler du tout !
— Je croyais que tu voulais trouver Elliot...
Terrassé par sa logique, je ne sais plus quoi dire. Il fait claquer sa langue, puis émet un « tt-tt »...
— Bracken, Bracken...
Il pose une main sur mon épaule.
— Je t'ai battu une fois. Et je te battrai encore. C'est bien que tu sois là, parce que tout seul, je m'ennuyais un peu. Sans toi ça n'a aucun sens. Dans la vie, il y a ceux qui jouent seuls, comme Elliot qui écrivent leurs propres règles et il y a ceux, comme moi, qui ont besoin d'adversaires. Ou de compagnons, même si c'est souvent la même chose... que tu le veuilles ou non, à présent tu es avec moi, ce sera notre dernier défi. Je t'ai attendu dix ans, même si pour moi ça ne fait pas de différence.
Il arrache mon châle, d'un coup que je n'ai pas vu venir.
— Bram ! Non !
Il se replie sur lui-même, recule dans le mur, et referme une porte que je n'avais pas vue. Je l'entends rire, derrière le bois.
— Bram, bon sang...
— Tu es venu jusqu'ici, tu es bien courageux... tu veux devenir le Maître ? Ce sera moi, le dernier Maître ! Je n'ai pas attendu tout ce temps pour me laisser griller ! C'est un jeu où il n'y a pas de second !
Il rit plus fort encore.
— Rendez-vous devant le Ptyx ! Gare aux coins !
Son rire cesse soudain. Comme s'il s'éloignait, à l'intérieur de l'armoire. Il n'est bientôt plus là. Hinrik et Egill se ruent vers moi.
— Briki, c'était qui ?
— Bram.
— Un ami ?
Coincé ici, le Bram. Comme moi. En ce lieu interdit, me revient un écho que j'ai connu jadis et dont je ne me souviens plus.
Dans leur coupe, les tortues n'en finissent pas de tourner, dans le sens inverse des aiguilles d'une montre.
— _Vraiment, vous n'avez aucun sentiment pour moi ?_
— _Ce n'est pas ça, c'est juste que..._
— _C'est parce que je suis votre frère ?_
Le morse me relève.
— Il est parti, dis-je, épuisé. Il a pris le châle.
— Ah, le félon ! Si je l'attrape...
Il fait mine d'étrangler quelque chose.
— C'est ma faute.
Je m'assois par terre. J'ai l'impression que le monde vient de se replier sur moi.
— Je ne sais pas pourquoi je vous ai emmenés jusqu'ici, mes amis. Je suis désolé. Je ne pensais qu'à moi. Je voulais tellement mettre un mot sur ce monde, dire à quoi il pouvait rimer, ce que tout cela signifiait, ce rien, cette plage, ces traits. Je voulais trouver une raison de vivre, peut-être, en entraînant tout le monde à ma suite. Je suis obsédé par moi-même. Il n'y a que cela qui m'intéresse. Je pensais que dessiner me renverrait au monde, m'obligerait à voir ce qui est, de mes propres yeux. Mais je ne fais qu'interpréter.
Hinrik se pose sur ses énormes fesses.
— Allons Briki, ne vous torturez pas...
Son frère se hisse sur sa tête, lui martèle le front avec ses petites pattes.
— On vous aime Briki. C'était notre choix de venir avec vous. Même si on comprend pas.
Il baisse la tête, d'un air presque coupable.
— On voudrait que vous soyez le nouveau Maître, Briki.
— _Laisse-toi faire !_
— _Parce que... Si, je vous aime, je..._
— _Pourquoi tu me vouvoies toujours ?_
Soudain, la vérité me transperce. Pourquoi j'ai si peur de perdre mon châle. Pourquoi j'ai perdu mes moufles. Je regarde mon doigt, cette callosité sur la phalange où se niche le bout de crayon. Serais-je en train de suivre un chemin qu'on a tracé pour moi, une ligne, un rail, que je n'ai jamais voulu voir, perdu dans mon nombril, dans l'obsession de mon petit moi-même ? Je suis sur un podium, que je n'avais jamais remarqué avant. Je me suis levé pour recevoir une récompense dont je ne connaissais pas l'existence. Autour de moi, mes amis m'ont toujours soutenu, m'ont sauvé la vie à de multiples reprises. Et qu'ai-je fait pour eux ? Rien. Car je suis le Néant en mouvement.
Je me relève.
— Il faut trouver la sortie.
Hinrik applaudit, Egill fait des petits bonds en tous sens.
Puis, le sol se met à trembler. Les murs, tout. Une vibration secoue la nature même du réel, quand, dans le couloir biscornu, résonne un bruit de trompette.
Non, celui d'un cor de chasse.
KOR
Ce que nous n'avons pas,
Nous le prenons dans tes bras,
Quand nous nous penchons pour lire la sève.
J'essaye de concentrer en un point unique les derniers événements. J'essaye de me souvenir de ce que je suis. De ce que j'ai fait. L'ai-je fait sciemment ? Est-ce que quelque chose m'a échappé ? Je suis un autre, à présent. Je suis un criminel sans haine. Je ne veux rien. Je veux juste retrouver ces traits, les manipuler, me tenir contre eux, bien au chaud. Ces traits sont les signes de mon enfance, quand tout était encore possible, que je n'avais pas compris la finitude du corps. Il est tard désormais, trop tard pour apprendre à maîtriser les rêves. Tout ce que j'ai, c'est ce que je n'ai pas vu. Cette vie que je n'ai pas vécue. Ce désir quotidien de régénération. Se réinventer chaque jour, quand le matin se lève et qu'on ne peut rien y faire. Juste accepter d'être là, avec ce que nous sommes, faire du mieux qu'on peut. Je n'ai pas fait de mon mieux. Je suis la paresse incarnée. J'en paye aujourd'hui le prix, quand le pouvoir m'échappe. J'aimerai passer un contrat avec moi : désormais être toujours ce que je suis.
KOR
N'oublie pas,
Tu vois le plan,
Mais il reste des coins.
Entre mes mains, filandreuse, la silhouette de Plouffe est une bouillie de spaghettis qui se dissolvent. Toutes les couleurs de son corps ont glissé sur le parquet. Des zones de couleur plus pures, sans définition, sans silhouette. Des flaques, de l'huile qui irise la lumière d'une façon nouvelle. J'approche le doigt de l'une d'entre elles où se fondent à la fois du turquoise, du mandarine et du mauve, afin de toucher, cette texture qui évoque une peinture très épaisse.
La porte du bureau s'ouvre, Fink entre en trombe.
— Bracken ! Enfin, vous voilà !
Je vois sa silhouette, découpée dans le réel, rejoindre toutes les autres lignes de la pièce, cette radiographie instantanée du moment présent. En hâte, j'enfile mes moufles. Mes moufles. Si je devais toucher la silhouette de Fink à mains nues, sans être protégé par cette laine qui depuis le début forme un écran entre le Néant et ma peau, alors je deviendrai un destructeur de monde. Ou un Dieu.
Ou les deux.
— Où est Plouffe ? s'enquiert-il en parcourant la pièce du regard.
Comment lui dire ? J'aimerai articuler : « J'ai détruit Plouffe. J'ai défait son corps, j'ai tiré une ligne de sa main, sans savoir que toutes les lignes qui composent un corps sont liées entre elles. Tout est venu à la suite, sous mes yeux, Plouffe a fondu. Il s'est répandu en couleurs, cette flaque, là, à vos pieds. »
— Je l'ignore.
Fink cherche dans le bureau, les livres, les taches de couleurs. Il n'arrive pas à faire le lien. Comment le pourrait-il ?
— Bon, Bracken, écoutez. Nik commence à chanter dans cinq minutes, j'ai toutes les mamans à mes basques. Et pour couronner le tout, je crois que Cyldrid se prend pour Gryla !
Il me prend par la main. Dans le couloir attendent trois femmes en uniforme orange fluorescent.
— Vous êtes le proviseur ?
— Heu... Je...
Fink me donne un coup de coude.
— Oui, oui. Je suis... Plouffe.
— Il va falloir évacuer l'école, monsieur Plouffe.
— Pourquoi ? Il y a une bombe ?
— Le volcan. Il se réveille.
— C'est impossible, il est éteint depuis des siècles.
— N'empêche que tous nos relevés indiquent une éruption imminente.
Je me presse les sinus. Fink me repousse.
— On va gérer ça ! conclut-il en me jetant dehors, laissant les sismologues ahuris dans le couloir, avec leurs instruments, leurs craintes et leurs responsabilités.
— Fink, dis-je sur le parvis, si le volcan se réveille, ce sera...
— Ah, ne les écoutez pas, ils disent n'importe quoi. Regardez...
Il me désigne le Burfell, au loin, masse sombre dans la nuit factice d'un matin à peine lumineux. Sous un ciel laiteux, piqué d'étoiles, un soleil peine à se lever derrière l'horizon. Sur le champ de lave, les montagnes noires se découpent comme des ongles mal taillés, griffant les nuages, déchirant des trouées.
Je cligne des yeux. Je respire dans le silence. J'ai envie d'ôter mes moufles et d'explorer le nouveau pouvoir qui est le mien. Je peux lire dans chaque repli du réel une ligne, un trait, qui ne demandent qu'à bouger, qu'à être déplacé. Je bloque sur un aspect du ciel qui m'avait échappé, une nuance de jour qui point sous ce tapis sombre, une ligne striée d'autres lignes où se compose un mille-feuille d'atmosphère, de gaz et de poussière, redéfinissant ce qui doit être d'un côté ou de l'autre. Soudain, le plein m'apparaît vide, de l'œil, j'y trace un monstre sublime qui s'élève en hurlant, sa gueule ouverte sur des rangées de dents métalliques. Ici, ses naseaux, ici, ses écailles qui reflètent les...
— Bracken ?
Fink me tire de ma rêverie.
— Bracken, vous êtes où, là ?
— Pardon, je...
— Concentrez-vous un peu sur ce que je dis, Bracken.
— Très bien.
— Erla a organisé un concours de déguisements, juste avant le rappel, quand Nik va chanter _The Riddle_. Elle veut trouver un fiancé pour la reine, puisque le roi ne revient pas. C'est un peu une tradition, vous savez, même si je ne suis pas certain que cela l'aide vraiment la reine...
— Où est-elle ?
— Qui ?
— Eh bien, la reine...
— Elle est...
Il a l'air perdu, le pauvre vieux. Moi aussi, j'avoue. Je me sens si loin. Les scientifiques sortent de l'école, leurs talkies-walkies vissés à l'oreille. Peut-être aurions-nous dû les écouter. Je regarde le Burfell, noyé dans un océan de brumes : il ne semble pas fumer. Est-ce une blague de circonstance, en cette période fêtes ?
— Je ne sais pas où est la reine, mais je suis sûr qu'elle nous regarde et que ça la touche, tout ce que nous faisons.
— Comment ça ?
— Eh bien, ces enfants sont des princes, non ? En eux, la féerie s'incarne tous les jours, encore et encore. Une énergie qu'aucune religion n'est parvenu à tarir, qu'aucune résignation n'a entamé. Je ne suis pas fou, Bracken, je mesure la portée symbolique de ce que nous faisons. Nous lui proposons un roi pour lui donner la force d'attendre le retour du véritable souverain. La reine se nourrit de cette joie, de ce potentiel, de ce glamour. Nous lui donnons vie, vous comprenez ?
— Je comprends.
— Vraiment ?
— Oui.
Il a l'air visiblement impressionné.
— Bracken, vous avez changé.
— Ah bon ?
— Depuis que vous êtes allé dans le coin. Quand vous avez glissé. Je vous l'ai déjà dit. Je vous trouve changé. Au début, je pensais que vous alliez devenir encore plus pénible, mais j'ai l'impression que vous avez enfin ouvert votre cœur.
— À quoi ?
Il lève les bras.
— À tout ça ! À l'invisible !
— Dieu ?
— Ah, ne parlez pas de choses qui fâchent.
— Pourquoi ? Dieu, les fées, c'est la même chose...
— Bon sang, vous parlez comme Cyldrid... Où est-elle cette ogresse ?
Il scanne la foule.
— Ah, la voilà, venez Bracken, on ne sera pas trop de deux.
Ils ont construit un grand podium, juste devant la scène. Des enfants déguisés se pressent sur les marches d'un escalier en carton. Des petites filles pleurent sur le côté, des petites fées déprimées de ne pas pouvoir participer. Des mamans les consolent, en les adjurant de retenir leurs larmes, les elfes n'aiment pas les larmes et puis, si on fait tout ça, c'est dans le but de rendre la reine heureuse. Avec un peu de chance, l'année prochaine, on organiserait un concours de courtisanes. À cette perspective, certaines reprennent confiance, tapent dans leurs mains pleines de paillettes, éparpillées puis mouillées de pleurs. D'autres continuent leurs lamentations assises sur un rocher, leurs petites ailes déchirées, leurs bas tire-bouchonnés ; ces petits visages maculés deviennent de petites perles isolées dans les ténèbres, des feux follets.
Erla est au côté de Cyldrid, qu'elle tient par le bras. On a noué un bandeau sur ses yeux. Son costume est souillé de boue. Fink lui prend la main.
— Allons Cyldrid, comment ça va ?
Erla grogne.
— Elle va bien, ne vous inquiétez pas. Nous avons appelé une ambulance pour ses yeux. Elle va bien finir par arriver. En attendant, elle effraye les enfants, c'est assez drôle.
Cyldrid, tonne, d'une voix terrible.
— Mmm ! Ça sent la chair fraîche ! Qui veut venir faire un câlin à Gryla ?
Les enfants hurlent de terreur et se réfugient dans les jupes de leur maman.
— Gryla ! Gryla ! scandent les neuf jeunes princes debout sur l'estrade.
— La reine va recevoir son simulacre de roi. Nous devons attendre, et Gryla est descendue pour exiger son dû.
Le concert noie l'essentiel de ses paroles.
— Où l'avez-vous trouvée ?
Erla me répond avec morgue.
— Elle errait dans le champ de lave. C'est le manager de Nik qui l'a retrouvée. Elle délirait, elle s'est visiblement fait mal aux yeux.
Fink me lance un regard qui signifie : « Restez coi. »
Cyldrid tourne son visage vers moi.
— Revenu du Néant, Bracken ?
Elle pose ses mains sur mon visage.
— Bracken, l'enfant roi...
Elle me prend les mains.
— Votre cœur caché sous cette laine. Ce vrai pouvoir, qui vous prend tout entier, que vous ne pouvez renier, qui vous a sorti de l'entre-deux.
— Un pouvoir ? questionne Fink.
— Elle divague, dis-je en lui retirant mes mains.
Elle lève le nez, humant l'air ambiant.
— Soudain, le vent se lève, libérant les effluves de l'autre voie, sans cesse fragmentée, derrière le miroir. En lui se drape l'essence de ce qui devrait être mort et qui reste en vie malgré tout, malgré les voiles, car le Maître est parti rejoindre ses rêves, et ne reste qu'un monde maudit, à la recherche d'un nouveau médiateur.
Elle se tourne vers moi.
— Et le Kor, venu de là-bas pour chercher son dû.
— Le Kor ? demande Fink.
Ce qui est venu du pli. Ce qui a décollé la tapisserie, les rétines, les pixels. Erla cligne des yeux. Je prends Cyldrid par la main, l'emmenant à ma suite. Nous traversons la fête, nous passons le concert, les minettes hystériques, les parents en rangs d'oignons.
Sur scène, le groupe de Nik compose une constellation de néons, collants de spandex, guitares en plastique, claviers de marque italienne ; leurs chevelures ondulent, cristallisées par le gel dans les projecteurs, en piques, en arêtes, en vagues mouvantes. On dirait des super-héros. Leurs silhouettes sont faites de traits hachés, rapides, comme tracés au marqueur. Pour la première fois, je me demande : qui a créé ces surfaces ? D'où viennent ces traits ? Je pensais que le monde existait par lui-même, qu'il se créait ses propres frontières, alors qu'en lui se multiplient les organismes. Mais si une limite est posée, alors, qui est le dessinateur ? Qui décide de placer un trait ici ou là, de la plume à employer, de l'épaisseur ? Il y a bien quelqu'un, quelque part, qui a créé ces choses que je défais.
J'entraîne Cyldrid vers un rocher, sur lequel elle s'assoit. Fink s'agenouille à ses pieds, plus inquiet que je ne le pensais. Se pourrait-il que derrière leurs éternelles chamailleries se cache une relation plus profonde ? Peut-être que si Plouffe m'a menti sur ses motivations, ces deux-là sont aussi dans le coup, comme je l'ai supposé en tout premier lieu.
— Parlez-moi du Kor, Cyldrid. J'ai besoin d'en savoir plus. Je sais que c'est lui qui vous a fait ça, mais je ne sais pas ce que c'est. Et si je dois le combattre, autant savoir de quoi il est fait.
— Le Maître, Bracken, le Maître qui joue les accords du Néant sur la carapace de la tortue, le Maître est menacé à tout instant par la somme. C'est un monstre terrible, au sein duquel est gardé l'objet de tous les désirs. Dans le pli du pli, là où les mondes se touchent, au-dessus de la faille, il garde le passage.
Elle lève le doigt.
— Mais il est entré ici. Nous l'avons tous senti. En prenant mes yeux, il m'a rendu la clairvoyance. Il m'a fait ce don précieux ; il m'a permis d'entendre, moi qui, saisie par la force de ce moment, n'avais plus en moi que l'amour.
Un silence, puis :
— Bracken, il est ici, avec nous.
Soudain, je le sens.
KOR
Oui, nous sommes là.
Ici, venus du Rien,
Tout comme toi.
Autour de moi, il pourrait être partout. Comme une fumée, comme une ombre de mort, qui recouvre la vie. Je le cherche des yeux, dans les contours, dans les replis, dans ce monde qui, à chaque minute qui passe, me devient plus étranger.
Cyldrid s'agite, ses petits bras flasques s'activent. J'essaye de la calmer, je la tiens par les épaules. J'aimerai toucher son squelette chétif d'oiseau malade. Je peux sentir sa frêle structure, je peux voir son châssis vaciller, il me suffirait d'un rien pour le rétablir. Si je peux détacher, alors puis-je également peut-être renouer ?
Une clameur explose autour de moi. Des enfants dévalent la pente en quête de Cyldrid. Le soleil ne va pas tarder à se montrer. Il va rester une heure au-dessus de l'horizon, peut-être plus. Il ne fait pas si froid, mais il va tout réchauffer. L'astre du jour. Le prince des cieux.
— Gryla ! Gryla !
Elle rit.
— On se calme, intime Erla, qui suit la farandole. On dit merci à Gryla.
Erla la prend par le coude.
— Par ici, Cyldrid. Ça va aller ?
— Inquiétez-vous plutôt pour les autres.
En partant, Cyldrid se retourne vers moi, toujours sans me voir.
— Petit Bracken, tu dois être fort, mais je te sens faible, si faible. Pourquoi gardes-tu toujours tes moufles ?
— Je... J'ai peur, Cyldrid.
— La peur n'est rien face à ce qui nous attend si tu ne fais pas ce que tu es né pour faire. Admets ton échec. Plonge dans le vide.
Elle s'éloigne, une nuée de petites fées l'acclame, une _ola_ merveilleuse, presque une haie d'honneur au fin fond d'une forêt. Tout scintille. Des enfants vont lancer des feux d'artifices dans les buissons. Paroxystique, le concert s'emballe, nappes de synthés sur cordes distordues, un solo de guitare éternel qui résonne dans la nuit comme les trompettes du Jugement dernier. Au micro, Nik se contorsionne, véritable dandy en latex, ses cheveux blonds en forêt sur son crâne, impossibles à modifier. J'aimerai tellement leur donner une autre forme. Peut-être qu'en y allant délicatement...
J'ôte ma moufle droite. Je libère mes doigts. Du pouce, levé devant moi comme pour évaluer la distance, je force une mèche à rentrer dans le rang, j'en étire une autre. Nik ne se rend compte de rien. Je peux changer sa coiffure à ma guise, peut-être puis-je aussi déformer sa plastique. Essayons... En comblant la courbe de ses hanches, je lui donne une nouvelle posture. Derrière lui, le batteur semble avoir noté la différence, je lui étire la main, il en lâche une baguette, rate une mesure, mais la chanson y gagne en originalité ; la ferveur de la foule le confirme.
Bon, très bien. Et sur des rochers ? Toutes ces pointes, alignées, irrégulières, j'aimerai leur donner un rythme, leur communiquer un peu de joie, les sortir de leur torpeur, me faire leur berger. Allez, par ici, hop hop. En ramenant toute ma main autour d'une aspérité, je change la forme d'un rocher qui, déséquilibré, roule sur le côté, changeant l'enchaînement. Les autres rochers bougent à sa suite, comme des dominos. En un clin d'œil, j'ai changé toute la configuration de la ligne. Je pourrais en faire plus, je pourrais essayer de modifier des visages peut-être, en jouant sur la frontière entre les ombres et la peau, là où le néant s'installe, trait subtil qui sépare sans apparaître, et que moi, pourtant, je perçois.
Allons. Peut-être sur Erla, son nez est courbe. Je suis certain qu'elle apprécierait, cette peste, de pouvoir respirer correctement.
— Bracken, qu'est-ce que vous faites ?
Vite, renfiler ma moufle.
— Rien, rien, je...
Fink ferme un œil.
— Qu'est-ce que vous me cachez ?
À l'horizon, un rayon de soleil vient d'apparaître, les enfants chantent leur allégresse. Le concert s'achève enfin, les cheveux fous de Nik s'embrasent dans un déluge de couleurs, puis Erla monte sur scène, prend un micro.
— Allez les enfants, montez !
En rang serré, les petits princes montent un par un les marches de l'escalier branlant et se rangent sur la scène entre les instruments, sous l'œil blasé des musiciens. Dans un coin, Nik essaye de remettre sa chevelure en place : il ne comprend pas ce que vient faire cette nouvelle pointe au milieu, cet épi définitivement rebelle.
Plusieurs voitures se rangent sur le parking. De la plus grosse sort un homme en écharpe, que je reconnais vaguement comme le maire. Une enfilade d'apparatchiks le suivent en trottinant, blottis les uns contre les autres. Le pauvre soleil tisse leurs ombres en nuances sur le sol. Je les contemple et je bloque à nouveau, obsédé par leur danse, par la façon dont elles glissent et coulent sur la pierre, distendues puis rétractées, formes élastiques que j'ai soudain envie de manipuler. Car dans cette surface, je lis désormais une profondeur. Dans les ténèbres, je vois les strates, successives, une par une, ligne par ligne, qui composent le Rien. Y aménager un sens et le peupler d'images, d'amis nouveaux, de plusieurs singularités serait si simple.
Les élus montent à leur tour, saluant la foule du plat de la main, un sourire faux en guide de masque. Comprennent-ils ce qui se joue ? Des caméras de télévision sortent soudain de leurs camions, d'autres reporters s'affairent autour de leurs enregistreurs. Je n'ai vu personne pendant le concert. Il a suffi qu'un officiel pointe son nez pour qu'ils sortent tous.
— Mes amis, mes amis, quel plaisir d'être là ce soir...
Je n'écoute pas. Absorbé par le reste, par tout le reste, tout ce qui est autour, tout ce qui est entre, tout ce qui me sort de moi-même, enfin. Ce moment présent que je cherchais à toucher du doigt est devenu concret. Le voilà, le mystère : l'acte de création en temps réel, qui s'actualise à chaque seconde, à chaque instant. Tenir à jour l'évolution subtile de la réalité au fur et à mesure de sa construction instantanée. Tout me menait à ça. Plus rien n'a d'importance à présent. Il n'y a plus rien à trouver, à chercher. Elliot a disparu, j'ai tué Plouffe, Cyldrid est folle et Fink... Eh bien Fink, je le regarde, il applaudit, je pense qu'au fond de lui il souhaite que l'un de ces princes soit le véritable roi, qu'il puisse remplir ses fonctions et honorer la reine, dans son malheur, elle qui, sans souverain, ne peut régner. C'est pour cela que les fées n'apparaissent plus, m'avait confié Fink. Si la reine était comblée, alors, à chacun de ses orgasmes, son glamour pourrait se répandre sur le monde, comme une confiture depuis trop longtemps conservée. J'aime cette image de la féerie qui coule. Elle me rappelle ces couleurs qui ont coulé de Plouffe, quand j'ai défait sa silhouette. Je n'ai pas fait exprès, mais je pense que c'est ce qu'il voulait. Se changer en pur glamour.
— Et c'est ainsi que...
J'aimerai le faire taire, faire disparaître ces lignes qui composent ses lèvres, mais j'ai peur que Fink ne me voit opérer. Je dois attendre. Après son discours, le maire fait défiler les jeunes princes dans leurs habits chamarrés, bandeaux de papiers, tissus plissés, ceintures de plastique, bottes, tout cela prenant enfin l'allure d'une vraie célébration atour de ces figures patriarcales, politiques, dans leurs manteaux trop longs, trop gris. À mon côté, Fink croise les mains, peut-être en une prière adressée à l'un de ces dieux païens, ou bien fait-il le vœu de revoir enfin le merveilleux envahir le monde, le transformer, d'assister à une restauration que nous devrions tous souhaiter, nous tous pour qui le sens prime sur le confort. Je trouve désormais un intérêt à ces structures archaïques car, prenant conscience que chaque chose a sa place, et que chaque place a sa chose, je deviens l'instrument d'une volonté qui me dépasse, investi d'une nouvelle responsabilité.
KOR
Si tu savais ce que tu as fait,
Si seulement tu avais le souvenir,
D'avoir enfin volé le Ptyx.
Le maire a quitté la scène, serrant chaleureusement la main de Nik au passage. Les sismologues l'approchent, accompagnés de plusieurs agents de police. Erla les a rejoint, toujours prête à incarner une figure d'autorité. La petite amie de Bram est avec elle, en larmes. Ils discutent tous ensemble, sans que nous puissions entendre quoi que ce soit. Cyldrid se tient au pied de la scène, drapée dans une couverture de survie marmonnant je ne sais quelle prophétie. Fink se crispe à vue d'œil.
— Bracken, ils vont nous évacuer !
— Ma foi, pourquoi pas ?
Il me saisit par le col de ma parka.
— Bracken, vous ne comprenez pas, il faut un roi !
Bon. Il est fou lui aussi. Je ne sais pas ce qui a déclenché tout cela. Comme si quelque chose avait chamboulé leur cerveau, comme si... Attends, attends Bracken. Il y a quelque chose qui ne va pas, là. Quelque chose qui ne tourne pas rond. Comme tous ces étudiants qui ont disparu, comme cette fille qui m'avait appelé Maître, qui savait quelque chose que Bram lui avait expliqué. Il s'était passé quelque chose dans ce gymnase. Quelque chose avec lequel j'avais peut-être un lien. Et si j'avais commencé à user de mon pouvoir avant, sans m'en rendre compte ? Et si j'avais joué avec les silhouettes, bouleversant la façon dont les gens pensent, modifiant les neurones ? Simplement en bougeant leur structure. Mais comment ? Pourquoi ? Qui m'a donné ce pouvoir, qu'est-ce qui a fait de moi ce que je suis ? Quel est donc ce pouvoir, caché dans la bosse de mon majeur ?
La copine de Bram me montre du doigt.
Tout le groupe me regarde.
— Laissez-moi faire, dis-je à Fink.
— Oh merci Bracken, merci ! Si je le pouvais, je vous embrasserais !
Je m'approche du groupe, les mains dans les poches. Fink est derrière moi, comme un petit chien.
— Messieurs ?
Cyldrid éclate de rire, puis s'exclame :
— Il est là, le Maître qui a vaincu le Kor et volé le Ptyx. Il ne vous épargnera pas, vous ne pourrez le vaincre, mais le Kor est venu par le pli pour empêcher l'instauration.
KOR
Oh, vieil oracle,
N'as-tu pas faim,
d'enfin taire la fin ?
— Où est le proviseur ? me demande un flic.
Leur groupe est formé d'une ligne unique qui détoure chaque individu, reliée à des points stratégiques. Il me suffirait de les rompre pour m'en débarrasser. Je voudrais que Fink ait raison, finalement. Je voudrais que les fées reviennent, je voudrais que le monde redevienne comme il l'était, quand nos cerveaux étaient câbles pour admettre le merveilleux. C'est peut-être ainsi, que je pourrai me rendre utile : restituer à tous la capacité de rêver, perdue faute de la bonne connexion. Il nous manque la chirurgie.
— Par là, dis-je en les laissant passer.
Perché un peu au-dessus de nous, Nik nous regarde d'un air interrogateur. Je lui fais un signe qui signifie : « Continue. » Je dis à Fink de ne pas bouger.
— Occupez-vous de Cyldrid.
Gryla sourit faiblement.
— Le Maître a compris.
Elle me touche le bras.
— Attention de ne pas perdre le Ptyx, le Kor est en colère.
Je lui baise la main, puis j'entraîne le groupe à ma suite. Je ne les compte pas. Pour moi, c'est un paquet complet. Indivisible. Une entité. Comme une tumeur, qui va empêcher le monde de s'accomplir, de s'enfanter. Ils n'arrêtent pas de jacasser : à quel point Plouffe est irresponsable, comment ils vont fermer l'école, et pourquoi les enfants sont-ils laissés libres de faire ce qu'ils veulent ? Qui organise les activités ? L'histoire et la géographie sont-elles correctement enseignées ? Et l'orthographe si précieuse, notre langue ? Et pourquoi engager des étrangers ? J'ai envie de leur répondre, j'ai besoin de me défendre, mais tout au fond de moi je sais que je suis désormais l'instrument de quelque chose qui me dépasse.
Je les emmène tout au fond de l'école, au dernier niveau, dans les ruines marines de la chambre d'Elliot. Où je sais que personne ne nous entendra.
— Voilà.
— Là-dedans ?
Je les fais entrer un par un. Quand je suis seul dans le corridor, j'enlève mes moufles. Ils sont surpris par la pagaille dans la chambre. Erla roule de grands yeux, mains sur les hanches. Il est temps.
Je lève ma main, je localise un des nœuds qui réunit leur concept dans l'espace : eux en tant que groupe venu empêcher le monde d'avancer, l'histoire de se construire. Il vient tout seul.
Ils hurlent, ils souffrent. Plouffe avait choisi de mourir, il attendait la mort, il s'est délecté de cette façon de partir. Eux, ils ne comprennent pas. Ils sont terrifiés. Ils se voient fondre, les uns dans les autres, en une galaxie de couleurs, de surfaces coulissantes. Pire, je pense que leur incompréhension augmente encore leur souffrance. Ils ne savent pas, ils ne connaissent pas. Ils se débattent, espérant repousser leur tourment, mais ces affres ne sont pas de celles qu'on peut arrêter avec les mains.
Je détisse leur chaîne et leur trame, fil à fil.
KOR
Voici le Maître à l'œuvre
Vas-tu défaire le monde,
Toi qui cherches ton âme ?
Je baigne dans une mare de couleurs, irisée par les reflets de la lumière artificielle, les bras ballants. Ils sont partis. La voie est libre. Il ne reste rien qui puisse s'opposer à la célébration d'un monde meilleur. Cette contemplation du présent doit être répandue. Tous les moyens seront bons. Dieu, les fées, les anciens dieux, la source vive au fond de nous, tous les espoirs, toutes les prières, tout a un sens si nous voulons voir le monde d'une autre façon. Et je serai le petit soldat de cette vision, le mécanisme.
KOR
Oui, oui, regarde
Nous sommes avec toi,
Nous te protégerons.
Quelque chose apparaît brusquement autour de moi, comme une nuée d'insectes. Un chorus de voix, une ronde de mains tendues. Des visages qui se fondent les uns dans les autres en une danse frénétique, hallucinante. Je ne sais pas de quoi il s'agit, mais je sais qu'entre toutes ces formes il y a des angles, et dans les angles des plis, et dans ces plis, cachés tout au fond, d'autres plis, à l'infini, et dans cet infini, tous ces mots, ces vers, ces rêves, il y a un dernier point où s'articule l'origine, et au fond de ce point, niché comme un chat dans un châle, il y a mon âme qui attend, éveillée, et le Ptyx, qui fait vibrer l'espace entre les lignes.
— _On en a parlé..._
— _Je sais._
— _C'est pas possible._
— Ces lieux sont maudits ! hurle Hinrik.
Nous galopons dans ces couloirs depuis des heures, pourchassés par des cors de chasse dont nous n'identifions pas l'origine. Nous tournons en rond. Le couloir forme un nœud. Il n'y a rien ici, et à nouveau, nous nous retrouvons face au fauteuil vide de Stéphane Mallarmé. Cette fois, nous arrêtons de courir.
— C'est inutile, dis-je en me retournant.
— Le Kor va nous attraper !
J'essaye de me souvenir du clip de Nik Kershaw. À la fin, me semble-t-il, on comprend que le labyrinthe a la forme d'un point d'interrogation. Nous ne pouvons pas sortir d'une question. Il nous faut une réponse.
Je me tiens debout face au couloir, tandis que les notes sombres se rapprochent. Une douleur me lance dans le majeur, là où le mystérieux poisson m'a mordu. Hinrik se range à mon côté, l'oiseau sur sa tête. Je tiens les tortues dans la coupe, dressée devant moi comme un talisman.
— Qu'est-ce que vous faites, Briki ?
— Aucune idée.
— _Mais personne n'en saura rien !_
— _Tu es ma sœur !_
— _Et alors ? ? ?_
Les cors de chasse tonnent, tout le couloir tremble. Voici le Kor qui s'avance devant nous, enfin révélé, lui le gardien de ces lieux. Devant nous, le monde se décolle de lui-même, comme un calque. Ce fruit de rien se déplie en origami, une série d'angles imparfaits dont chaque recoin est une mâchoire. Formidable machine à nuances, miroirs se reflétant entre eux à l'infini. Pour la première fois, je réalise que depuis la minute où nous sommes peut-être à l'intérieur du Kor, nous avons pris pieds dans cet iceberg. Le Cap est-il réel ? Ou bien n'y a-t-il qu'une série de trompe-l'œil qui ne mènent nulle part ?
— _Dis, tu crois pas qu'on devrait plutôt aider Bracken ?_
— _Ne change pas de sujet, veux-tu !_
— _Mais enfin, regarde, c'est le Kor !_
Impuissants, nous regardons cette ombre colossale nous submerger, sur le point de nous engloutir. Elle s'ébroue, massive, s'apprêtant à nous absorber tous autant que nous sommes. Hinrik se couvre les yeux, tout craintif.
— Bracken, faites quelque chose ! Il va nous manger !
Il murmure une litanie de formules toutes faites. Des échos, des séries de monstruosités sémantiques, une rhétorique de l'indicible, sans forme, sans mots distincts. Elle tente de m'entraîner là où tous les autres avant moi sont passés. Je peux les sentir, les âmes perdues, qui ont tenté de dérober le Ptyx pour devenir Maître à leur tour. Mais moi, je ne veux pas être Maître. Je suis ici parce que je cherche Elliot. J'ai emprunté ce sentier car il a laissé de petits cailloux derrière lui. Il a fait de moi son héritier, sans que je ne comprenne pourquoi. Je ne suis pas poète, je suis...
Je suis...
— _Le Kor, je m'en fous !_
— _Mais alors, qu'est-ce que tu veux ?_
— _Maman !_
Alors que le volet du Kor se referme sur nous, Bram surgit de nulle part, saute sur les plaques articulées qui feuillettent l'armure du démon, le chevauchant comme une bête de rodéo. Un fou chantant sur une masse abstraite qui prend forme puis se déforme pour redevenir l'illusion qu'elle était. Le Kor tonne, hurle. Je comprends à présent que Bram nous attendait pour le surprendre pendant qu'il fondrait sur une autre proie que lui.
Nous avons servi d'appâts.
— _Maman est morte, Roméo..._
— _Morte ?_
— _Oui. Nous étions sur la plage..._
Hinrik se pousse sur le côté au moment le monstre se rue sur nous, faisant voler les meubles, les bibelots et tout ce qui entrave son passage. De solide, le Kor devient liquide, il se mire en lui-même, son volume décuple. Trompettes, encore, des vents glacés qui soufflent par des orifices qui s'ouvrent et se ferment de concert, créant de nouvelles bouches, de nouveaux trous où Bram glisse les doigts, tire sur la matière, la retournant comme de la pâte à modeler.
— Aidez-moi ! hurle-t-il, à moitié désarçonné.
Je me jette dans la mêlée, enfonçant mes doigts là où je peux, car de ma main droite, je tiens encore la coupe, dont l'eau s'échappe chaque fois qu'elle tangue. Sur la crête, Bram plonge ses deux bras dans la turbulence, empalant le vortex qui se meut sous lui, enflant dans le corridor. Autour de nous, Egill volette, ses ailes battant le rythme de notre danse de mort des plumes arrachées jonchent le sol de nouveaux reflets.
— Montez ! ordonne le morse.
Je l'enfourche, il se rue sur le fleuve de strates, donne des coups de nageoire en psalmodiant de vieilles incantations, ou aussi bien des chansons à boire. Je ne sais pas si nos coups portent, mais à chaque trouée refermée, j'ai l'impression que la chimère ralentit, sans s'amenuiser. Ce ralentissement me permet de saisir les jointures qui délimitent chacune de ses parties. J'ai appris à lire les lignes, dans le sable, sur les vagues. Je sais deviner ce qui est à la rencontre de deux surfaces, mon œil est entraîné, il sait déceler l'aspérité là où il tout semble plat. Seul un dessinateur peut trouver la faille pour s'y glisser. Voilà pourquoi Bram ne pouvait rien faire seul et la raison pour laquelle il nous a attendus. Il n'a jamais su dessiner, il avait besoin d'être assisté par un ordinateur. Il n'a pas compris ce que le dessin pouvait offrir comme jouissance analogique.
— _Je ne peux pas le croire._
— _Tu ne t'en souviens pas ?_
— _Et toi, tu t'en souviens ?_
Je lève ma main droite, la douleur enfle dans mon majeur. J'attrape une plaque, la fais pivoter sur elle-même. Je peux entendre quelque chose hurler autour de moi, comme un chœur de voix qui s'entrelacent et chantent : "Race immémoriale, dont le temps qui pesait est tombé, excessif, dans le passé, et qui pleine de hasard n'a vécu, alors, que de son futur. – Ce hasard nié à l'aide d'un anachronisme, un personnage, suprême incarnation de cette race, – qui sent en lui, grâce à l'absurde, l'existence de l'Absolu, a, solitaire, oublié la parole humaine en le grimoire, et la pensée en un luminaire, l'un annonçant cette négation du hasard, l'autre éclairant le rêve où il en est. Le personnage qui, croyant à l'existence du seul Absolu, s'imagine être partout dans un rêve (il agit au point de vue Absolu) trouve l'acte inutile, car il y a et n'y a pas de hasard – il réduit le hasard à l'Infini – qui, dit-il, doit exister quelque part. »
— _Oui, je me souviens, elle est morte d'un mystère._
— _Un mystère l'a tué ?_
— _Tole._
Poussé par la rage, je saisis une vitre, je la brise d'un coup sec, le Kor rugit, Bram en tombe à la renverse. J'accélère mes gestes, je fais coulisser l'entité sur elle-même, je localise tous les carrefours de circulation, les points nodaux où infiltrer mes images. Dans ces interstices, je trouve de nouveaux pleins, de nouveaux vides, et je la transforme, cette horreur, cette terreur antique, toute cette frustration qui ne demande qu'à être dissipée. Une par une, je brise toutes ses fenêtres, tous ses points de vue ; les bris de glace pleuvent autour de moi, lacérant nos chairs. Hinrik émet des grognements de douleur, mais il a tant de graisse qu'il ne souffre pas plus que ça. Egill s'est caché sous lui et Bram se relève, ahuri par ce qu'il voit.
— Eh bien, la conscience du jeune prétendant s'éveille !
Il jubile, s'anime, gesticule. Je n'ai pas le temps.
Je dois finir.
— _Tole, c'est qui ?_
— _Le fils du Pénultième._
— _Où est-il ? Je vais te me le..._
Infernale, cette série d'angles qui menace à tout instant de s'affaisser sur elle-même. Je dois trouver un dernier espace pour tout faire basculer, mais à chercher trop loin, trop profond, je ne fais que créer un effet centrifuge. Tout vient à la suite, château de cartes transparentes en train de s'effondrer. Une spirale se dévide, infinie, dévoilant le cœur du Kor, le point de vue originel, le premier frottement du Néant. Il a le goût d'un murmure, une longue tirade étirée qui ne cesse de se retourner, boucle rebouclée aux accents de boléro.
Soudain en la matrice s'ouvre une perspective.
De l'intérieur, un nouveau couloir, en travers du précédent.
— Par là !
Sans un mot, le morse se précipite par l'embrasure. Il a compris. C'est le seul chemin : en travers du chemin, en sa diagonale. L'oiseau le suit. Je me retourne pour attraper les tortues, et je vois Bram, le regard perdu, à la fois luisant d'espoir et désespéré de n'avoir pu accomplir tout cela seul. Je lui tends la main. Ce geste, une fois de plus. Comme si moi aussi j'avais besoin de lui pour clore mon combat. Il hésite une fraction de seconde, puis se jette dans le trou, sans un regard en arrière. Moi, je ne sais plus, je n'ai pas le choix, si mes amis sont là-bas, alors j'irai. Je réclamerai ce qui doit être réclamé, même si je n'ai ni raison ni de but.
Ou bien est-ce cela même qui fait de moi un élu ?
— _Tu ne le retrouveras pas !_
— _Pourquoi ?_
— _Tole est mort._
Ténèbres.
Grondements.
— Briki ? Briki, vous êtes là ?
— Oui oui, calmez-vous.
— Que s'est-il passé ?
— Nous avons traversé le Kor.
Peu à peu, les ténèbres sont parcourues de veines lumineuses, comme de la sève, inondant les formes torturées d'une crête rocheuse. L'air est chargé de gouttelettes, comme pulvérisées par un brumisateur. Je sais que nous sommes tout près de Ginnungagap, plus près que nous ne l'avons jamais été. Quand cessera ce jeu de poupées russes ? Est-il seulement possible d'atteindre le Néant sans jamais le toucher, en se contentant de le frôler, toujours prisonnier d'une pensée, d'un mot, d'une croyance ou d'une interprétation ?
— Je sais à quoi tu penses, me susurre Bram à l'oreille.
Le morse se tourne vers lui, s'allonge d'un coup de nageoire et l'écrase de toute sa masse.
— Espèce de ! ! ! !
— Hinrik !
— Il nous a piégés !
Bram étouffe sous la masse du morse, Egill lui picore les pieds.
Je m'agenouille, presque satisfait.
— On dirait bien que je t'ai encore aidé à franchir une étape. Quand cesseras-tu d'être un assisté ? Je croyais que l'analogique ne valait plus rien.
— Crève, Bracken !
Je lui ôte son masque ridicule.
— Tu sais ce qu'est le Kor, n'est-ce pas ?
— Pourquoi est-ce que je te le dirai ?
— Parce qu'autrement, le morse te mange.
Bram jette un regard aux deux énormes canines d'Hinrik et se décide à parler.
— C'est un système de sécurité. Mis en place par le premier Maître, le Pénultième. Vois-tu, techniquement le Kor obéit au Maître, à celui qui contrôle le Ptyx, mais vu qu'il n'y a plus de Maître, le Kor dévore tous les prétendants sans distinction. Ceux qui échouent sont ingérés et intégrés. Une bouillie de poètes, de musiciens, de peintres, de sculpteurs qui réclament vengeance et sont prêts à avaler tout ce qui pourrait prétendre à la beauté. Elliot a réussi parce qu'il se fichait bien de la poésie, c'était un joueur de la plus belle espèce, dont le moindre geste était un acte ludique. Et moi, je peux réussir. Je ne cherche pas le beau, je cherche le vrai.
— Dans le virtuel ?
— Le virtuel n'est qu'un moyen, Bracken. Tout comme Elliot a utilisé un jeu d'enfant. Toute entrée dans le Néant se fait par le jeu. Y demeurer, par contre, ça demande du SENS.
— Comment t'y es-tu pris pour créer des objets dans le Néant sans le Ptyx ? Toutes ces machines, ces torches...
— Oui, c'était bien vu, mais ce n'est efficace que pour de toutes petites choses. Nous avons trouvé comment utiliser le lag pour créer une instance dans un monde virtuel, nous avons pu nous glisser dans la fente.
— Quel monde ?
— Je t'ai dit, _Habitat_. C'est un jeu en ligne californien.
— En ligne ?
— Oui, qui se joue par l'intermédiaire du réseau téléphonique. Ils ont un partenariat avec un opérateur téléphonique aux USA qui permet au jeu d'être joué par plusieurs personnes simultanément, en différents points du pays, voire de la planète. C'est de la belle ouvrage, c'est la firme de George Lucas qui produit ça.
— Et ça marche en Islande ?
— Ça coûte très cher, il faut une ligne dédiée. L'université de Reykjavík a un accès réduit, mais les militaires de la base de Keflavik ont tout ce qu'il faut... Je jouis de quelques privilèges. Il suffisait d'y penser, c'était si simple : tellement de joueurs, tellement de potentiels, il fallait simplement trouver le moment où tout était stoppé, où le temps n'existait plus. Tu vois Bracken, pendant un lag, tu continues à jouer de ton côté, sans que ce soit enregistré par le serveur. Quand la distance est réduite, tu retournes d'où tu viens. Toutes tes actions sont effacées. Ont-elles réellement existé ? Eh bien, je peux te répondre : oui. Nous avons trouvé un moyen d'enregistrer nos actions laggées. Et d'y travailler. Sans le savoir, nous étions en train d'ajouter des choses au Néant, ces petites choses qui m'ont permis de trouver le chemin du Cap. Tu vois, je me débrouille parfois très bien sans toi...
— Sans moi, tu n'aurais jamais trouvé l'angle, chez Elliot. Sans moi, tu ne serais jamais passé. Tes ordinateurs, c'est bien joli, mais quand il s'agit du corps, il n'y a plus rien qui marche. Il n'y a que les autres corps.
— _Pourquoi ce soupir ?_
— _Parce que je ne me souviens pas. Ça fait mal._
— _Mon pauvre frère. Laisse-moi te raconter._
Egill s'agite soudain.
— Chut, fait le morse en levant la tête.
Il n'y a que le remous de la faille, son glouglou obsédant, qui ronfle. Egill continue de voler en rond, dans ce qui semble être une nouvelle caverne. Il y a un trou dans le fond, avec des marches d'escaliers taillées dans la roche luminescente. Des particules minérales dérivent en flottant autour de nous. Il y a quelque chose d'intense, de presque retenu, qui menace à tout moment d'exploser. Une attente insupportable, un éclat muet qui menace.
— Hinrik, laisse-le.
Le morse me regarde, sans comprendre.
— Mais enfin, Briki...
— Laisse, je te dis.
Le morse se relève, bougonnant et, sans surprise, Bram détale en riant. Il galope vers l'escalier, s'y enfonce, sa folie disparaissant.
— Je ne comprends pas, Briki.
— Il va trouver le Ptyx avant nous. Il va échouer. Laissons-le s'imaginer qu'il peut s'en sortir seul.
Egill vient s'assommer contre la paroi et retombe mollement en tourbillonnant comme une feuille morte. Hinrik le prend entre ses nageoires et le cajole tendrement. Il y a des étoiles autour de la tête du frêle volatile.
— Pauvre enfant...
— Quittons cet endroit si nous le pouvons, autrement nous allons tous devenir fous.
Je m'approche des escaliers. Ils sont particulièrement raides et glissants d'algues et de varech. Dans cette pénombre, dans la précipitation, il y a de quoi se rompre le cou, si nous ne faisons pas attention.
Allons, courage ! C'est la dernière ligne droite.
— Le Néant a toujours existé. Mais personne ne l'ayant vu, nul ne pouvait lui accorder une valeur. Il était là sans être là. Puis le Pénultième l'a découvert en creusant le vers, et en comprenant qu'il n'y avait rien. Il est arrivé dans le Néant, il a failli se noyer dans l'eau trouble. Dans son désespoir, il s'est mis à composer des vers spirituels, une sorte d'invocation où il est allé puiser l'espoir d'une sauvegarde. De ce geste sublime qui devait servir de modèle au concept du Ptyx naquit une belle et grande tortue qui nagea majestueusement jusqu'au Pénultième pour le porter. Il embrassa sa carapace et voyagea de longues nuits dans ce monde déserté. Le Pénultième n'ayant toujours pas de berge où prendre pied, la Belle tortue, qui seule pouvait créer du mouvement dans le Néant, remua le limon des fonds marins et fit émerger une plage, une mince bande grise et noire, sur laquelle le Pénultième s'allongea. Il s'étonna de trouver quelqu'un au Néant, et la Belle tortue lui dit que le Néant ne pouvait pas créer quoi que ce soit lui-même, il lui fallait un guide, une onde pour lui donner forme.
Une marche, puis une autre. Je descends dans l'obscurité. J'entends le bruit de l'abîme qui gronde, qui ébranle les fondations. Nous sommes au Cap désormais. Il n'y a plus de trappe, plus de piège. Je sais que le Ptyx est au bout du chemin, là, tout près. Parfois, j'entends le rire de Bram, loin devant moi, qui s'éteint, puis renaît. Comme si nous nous éloignions, puis nous rapprochions. Peut-être s'agit-il d'une illusion acoustique ou alors ces lieux forment-ils des labyrinthes qu'il convient de décoder. Je n'ai pas le sentiment de résoudre quoi que ce soit. Tout ce que je fais, c'est d'agir en maîtrisant les formes sur lesquelles ma pratique du dessin me donne une prise. Je n'ai pas de prise sur ces murs de roches suintantes tout glisse, je ne vois pas de lignes.
— _C'en était trop pour le Pénultième, qui ne supporta pas que le Néant fût aussi vide. Il voulait lui donner une forme acceptable, en faire un écrin où poser ses valises. Il avait compris que lui seul avait su invoquer la Belle tortue, mais il l'avait fait instinctivement, sans réfléchir. Il devait désormais retrouver son pouvoir créateur, en le formalisant. Il eut l'idée de se fabriquer un instrument. Sur la plage, le Pénultième s'est penché sur la carapace de maman, il a décroché une écaille, et maman lui a dit : ceci est la dernière écaille de ma carapace. Quiconque la brandira sera le Maître du Rien. Autour de vous, ces lignes, ces nœuds, naissent des remous que créent mes nageoires dans la matière. Qui pourra les ressentir sera le Maître des lieux, à ma place. Il sera celui qui jouera la mélodie du vide. Mon écaille sera son sceptre, la première idée de ce monde infini. Elle sera le chant de tous les possibles. Avec elle, la volonté sera manifeste. Alors le Pénultième leva l'écaille, le Ptyx, il créa le Cap pour l'y déposer. Il devint le premier Maître. Il fit venir son fils malade, pour lui offrir un monde._
Après deux cents marches, – je les compte toutes, – je réalise que nous sommes toujours au sommet de l'escalier. Ça suffit maintenant. Je n'en peux plus, je suis épuisé, j'ai envie de tout abandonner. Je n'y arriverai jamais. J'ai l'impression que cela ne prendra jamais fin. Qui doit décider ? Il y a tant de choses à gérer, tant de paramètres, je n'ai plus la patience, plus l'énergie, quelque chose doit céder.
J'ai l'impression d'avoir raté une marche.
Ou alors est-ce que je me suis trompé ?
— Nous sommes perdus, dis-je.
— On ne peut pas se perdre dans un escalier..., me répond Hinrik, l'oiseau sur le dos, mort ou endormi, je ne sais.
Il faut que je lâche quelque chose. Je suis trop lourd. Je suis plus lourd que l'air, plus lourd que tout. J'ai l'impression de peser cent tonnes. Il me reste peu de temps, car je ne peux pas tabler sur un échec de Bram pour nous sauver. Il y a forcément quelque chose à faire. Mes mains ne me servent à rien. Je sais qu'éternellement, nous reviendrons à ce commencement, encore et encore, qu'il convient de polir sans cesse. Comme ces pas, sur les marches, qui ont fini par les éroder. Combien de prétendants sont passés ici ?
Peut-être est-ce la clé. Est-ce que je suis un prétendant ? Est-ce que je n'en ai pas assez d'être ce que je ne suis pas ? Je dois prendre une décision. Ce Ptyx, il me le faut. Je suis devenu quelqu'un, je sais dessiner, je ne suis peut-être pas le grand poète que fut Mallarmé qui pourrait l'égaler, mais l'essentiel de ce que j'ai vécu ici, je l'ai vécu dans la célébration du moment présent, sans poser de questions. En acceptant mon sort. En comprenant les lignes. En aimant des amis, qui m'aiment en retour. En acceptant d'être simplement ce que je suis, depuis toujours, et que j'ai refusé jusqu'ici... Pourquoi Mallarmé porte-t-il des moufles sur son portrait ? Je portais moi-même des moufles avant de venir ici. Lui aussi, peut-être, était-il en train de se mentir. On n'écrit pas des vers en moufle, à moins de s'empêcher volontairement d'écrire les vers qu'on souhaite, ou alors... on écrit ainsi les vers qu'aucun autre ne pourrait composer.
Dans le bruit de mes pensées, sans me soucier du sens de la marche, je me rends compte que je suis descendu plus bas. J'aurai bientôt dépassé la deux centième marche, peut-être, ou pas, qu'importe ! Ce qui compte, c'est d'être là, sur cet escalier, et d'en voir le fond. Je sais que nous sommes sur la bonne voie. Nous descendons tout en bas, au tréfonds. Je peux sentir, s'approchant, un vent frais, venu du large.
Enfin, nous arrivons.
— L'enfant s'appelait Tole. Le Maître lui offrit la tortue. Ils étaient inséparables. Ils vécurent une belle histoire, et comme toutes les belles histoires, elle alla trop loin. C'est ainsi que nous sommes nés, toi et moi, dans ce ventre, fruits d'une union sacrée. C'est ainsi que les choses se font, entre les hommes qui disent le mot, et la Nature, qui le reçoit. Tout change, tout s'entrelace, tout est possible. Nous sommes nées d'un amour impossible, d'un pur paradoxe, et maman nous a déposées dans un trou, sur la plage, près du Cap, là où la mer, la terre et le ciel se rencontrent et tracent une onde qui sépare toute chose d'une autre. Elle nous a couvées. L'amour de maman était infini, elle nous regardait de ses grands yeux bleus, et moi, j'y lisais la mer, le ciel, toute l'âme des nuages qui s'y reflétaient. Mais le Maître était triste. Il nous regardait vivre, et Tole s'amusant à bâtir des châteaux de sable ; parfois, il s'éloignait sur la plage pour aller voir le trou que le Maître avait interdit, tout au bout, le bout qu'on s'était promis, toi et moi, d'aller voir un jour, cette grotte où nous n'avons jamais osé pénétrer.
Des vagues se dressent, se cassent, lames de fond qui lèchent les flancs du pic où nous nous trouvons. Nous avons grimpé, malgré la descente. Nous sommes tout en haut. Je dois m'asseoir pour prendre la mesure de toute la majesté de l'ensemble. Ginnungagap. Trente monstrueuses cascades, qui plongent dans un vide éternel, sous un ciel sombre où s'enroulent d'autres cascades de fumées, de cumulus enrobés d'éclats de miel. Des pulsations chantent dans la nuée, irisant les crêtes des vagues sur la mer déchaînée, puisant dans le maelström de nouvelles textures réfléchissantes. Ocre et jaune s'enlacent dans un gouffre griffé de noir, peinture à l'huile dont on peut voir l'épaisseur enfler puis couler comme du fromage fondu. Des orgues basaltiques crénellent ces failles.
— Mais les belles histoires finissent mal, et la beauté est toujours perdante. Tole aimait maman d'un amour sincère, ensemble ils nous regardaient grandir dans ce trou, en espérant que jamais nous ne serions des adultes. Tole pria le Maître de nous épargner le voyage des ans, et il fit de nous ce que nous sommes à présent, des enfants pour toujours. Maman n'apprécia pas que Tole demande aussi au Maître d'exaucer ses volontés : l'idée de changer le Néant ne peut venir que de celui qui brandit le Ptyx. Mais Tole protesta, il demanda à maman comment il était possible qu'elle existât, puisque c'est de sa carapace que le Pénultième tira le Ptyx. En remettant en cause sa nature, maman fut trahie, et son chagrin l'emporta. Échouée sur la plage, elle perdit sa consistance et le ressac l'emporta, dispersant le labyrinthe de son corps dans les méandres du remous. Seul, Tole décida de rentrer chez lui pour y continuer sa vie d'enfant. Le Pénultième ne put s'y opposer. Mais dans le monde de l'autre côté, la maladie de Tole n'était pas limitée par le temps du Néant. Tole continua de décliner. Rapidement. Le Maître s'absenta pour aller le soigner. Et puis un jour le Maître revint et annonça que Tole était mort.
De l'abîme, une lourde masse émerge soudain, fracture la surface de l'eau. Elle explose autour de nous, comme une supernova, quand jaillit la terrible masse d'une baleine, dressée vers le ciel. La baleine gronde, tournoie sur elle-même ; lentement, sa mâchoire s'ouvre, libérant ses filaments de dents, ses fanons en fanions, drapeaux de mousse et d'écume, les flots tumultueux de krills et de planctons, vomis, déversés. Sur sa langue, déliée, déroulée depuis l'immense caverne de sa bouche, une pièce argentée apparaît dans toute son absurdité : une plate-forme, parfaitement circulaire, quasi chirurgicale. Au bord d'un précipice, entre les chutes d'eau venues des tréfonds d'une déesse marine, cet ongle dédie ses reflets aux astres invisibles, au mouvement lent qui ralentit, jusqu'à cesser. Tout au bord de l'abîme, portée par le cétacé, la plate-forme s'affine en rotonde, de métal froid. La citadelle intérieure. Forteresse vide.
Au centre de la rotonde, il est là, sur son socle fin.
Je le vois.
Le Ptyx.
— Le Maître a dit qu'il n'admettrait jamais qu'un autre le devienne à sa place, Tole excepté. Mais Tole était parti, et le chagrin du Maître était inconsolable. Alors le Maître devint enragé, car il n'ignorait pas, maman le lui avait confié, que le Néant ne pouvait rester sans Maître. Lui seul avait été capable de trouver et de brandir le Ptyx. Alors il fit du Kor le gardien de la tour du Cap puis il quitta le Néant, n'y laissant que quelques créations primitives, des oiseaux, des créatures aquatiques. Le Néant est resté silencieux tout ce temps. Mais quelque part, un nouveau Maître s'est levé. Il nous a invoqués, tout comme le Pénultième avait invoqué maman dans la nuée. Nous l'avons retrouvé dans son bateau vide, dans la boue retournée. Nous l'avons guidé jusqu'ici, au Néant, mais il a abandonné son poste. Et ça n'a pas d'importance, car il n'a jamais voulu du Ptyx. Il n'a pas voulu être Maître. Il est passé par la grotte des merveilles pour rejoindre la joie éternelle sans nous et il n'est jamais revenu.
On dirait un médiator de guitare. Un triangle d'onyx, comme une perle, une goutte, ou l'iris d'un œil, ouvert sur des profondeurs abyssales. Il est plus beau que tout. Il est tout à la fois : le pinceau parfait, la plume idéale, le crayon rêvé. Je le veux. Il dit tout, il montre tout. Il m'attire, c'est plus fort que moi, c'est du fétichisme. Plus rien d'autre n'existe. Il me le faut. C'est pour lui que je suis là. Je suis prêt. J'ai traversé toutes ces épreuves pour être à la hauteur de cet enjeu. Mais moi, je ne parle pas. Ces mots sont mes pensées. Je ne veux pas les transcrire. Je veux les dessiner. Elles doivent couler de moi comme jaillies de ces chaos de traits dont j'ai été le témoin. Même si ça veut dire tourner le dos à mon passé, à ce que j'ai construit. Car je sais que pour retrouver cette ligne, il me faudra revenir au tout début, avant tout ça.
— Voilà, Roméo.
— Diane... Mais alors...
— Quoi ?
Je regarde mes amis, Hinrik et Egill, me dire adieu, ailes et nageoires en berne. J'ai passé ma vie avec eux, je les regarde, je peux sentir leur chaleur, tout ce qu'ils m'ont donné, à commencer par la foi en mes mains, en mon dessin, en ma vision. La foi en mon cœur d'aventurier, à l'assaut des limites, jamais en repos, et cependant terrifié par le Néant, qui ne cesse de gronder derrière avec une énergie capable de tout effacer. Amis, vous m'avez donné la force de me prendre en main, de ne plus épiloguer sur les ténèbres qui chaque jour prennent mon cœur, noient mon âme. Vous m'avez donné la certitude de tenir debout, seul, poussé en avant par ce thorax de lumière. Vous m'avez accompagné jusqu'ici pour trouver ce Graal dont personne n'avait deviné la teneur. Il existe, il est là, entre mes mains, alors que je le décroche de son piédestal pour le tenir devant mon visage. La puissance de sa gloire épuise toute volonté de lui résister. Il est ce but dont je n'avais pas idée. Il est cette présence dans l'azur de cette tour décapitée, qui me prend tout entier, m'envahit pour faire de moi le nouveau Maître.
— Si l'amour de Tole et de la tortue était possible...
— Oui ?
— Alors le nôtre...
Le moment ne dure pas huit instants. Hinrik pleure à chaudes larmes de morse. Bram se tient déjà devant le Ptyx. Il a pris une avance décisive. Il se retourne un sourire dément lui déchire la face. Il se dresse face au Ptyx, ses yeux glacés d'envie. Il sue, je peux voir les gouttes glisser doucement sur son front, illuminées pour une étrange foudre. Le Ptyx est là. Devant lui, offert. Ses deux mains nues vont le saisir. Il attendait notre confrontation, le dernier duel de son jeu maudit, de cette régression. Je n'ai pas le choix, je dois jouer. Je ne peux pas le laisser gagner. Avec le Ptyx, je pourrai faire ce que je veux. Je pourrai rentrer chez moi. Garder mes amis. Tout sera possible. À la seule condition que je joue. Ça me paraît peu. Je vais me lancer, mais soudain le Kor se matérialise. Aucun de nous ne l'avait vu. Tapis derrière le mince socle, il attendait. Quand les doigts de Bram touchent le Ptyx, un éclat de lumière m'éblouit et je l'entends hurler.
Aveuglé, je me précipite.
— Roméo !
— Diane !
— Je t'aime !
KOR
Là où la gravité se meurt,
Il n'y a que nous quand, dans le pli,
Nous prenons un fil pour le tendre.
Je sais qu'il est temps d'en finir. Ce monde bascule, grésille, il a perdu sa stabilité. Désormais plus rien ne doit m'échapper. C'est ainsi, nous décomposons le présent en soixante-cinq instants, couche sur couche feuilletant le Néant, puis vient le moment où nous réalisons que la vie est passée, que nous sommes vieux. Qu'il ne reste qu'une ultime aventure, la mort, le dernier mystère dont on ne revient pas, que notre jeunesse est perdue pour toujours, que d'autres viendront bientôt l'usurper, nous toiser ; la jalousie nous consumera, nous hurlerons de malheur et de frustration sans pouvoir espérer de nouveaux horizons. Pourtant, dans ces méandres, le trait fatal du destin se déroule ; notre capacité à voir les synchronicités est un coup de pinceau dans le liquide du réel. Ces lignes sont des pistes à suivre, à saisir d'un frôlement d'aiguille, ces encres nous fixent dans le mouvement. Si tout est transformation, la continuité analogique que la pensée affine à partir des fragments discrets des sens, cette route rectiligne doit être improvisée, contrôlée, puis libérée, puis découverte, baptisée d'un mot qui dise le sens.
KOR
Nous voilà dans l'accordéon,
Cette musique des fentes
Le vent sans souffle.
La foule rugit. Les basses font vibrer les murs, ou bien est-ce encore une secousse ? Je cligne des yeux. Tout est ligne, tout est pli. Tout étincelle. Sans m'en rendre compte, ou bien simultanément, j'ai fait le choix d'être la main qui tisse ce monde. En moi s'est renversée la coupe, mon âme a coulé. Baignant dans la mer de couleur de ces intrus, comme j'ai défait Plouffe, je sais ce que je suis devenu, ce que, depuis le début, j'avais fixé comme enjeu.
Je suis né à moi-même dans l'improvisation de cette quête. Peut-être Elliot le savait-il en faisant de moi le dépositaire de sa clé. Il voulait que je quitte ces moufles. Il voulait que je devienne ce que je suis. Lui, il a quitté ce monde pour cette grotte des merveilles au-delà du Néant, cet ailleurs qui m'est barré pour toujours, je le crains. À moins que la vieillesse ne m'apporte une paix qui me permettra de le rejoindre. Je tombe à travers l'horloge des siècles, les rêves s'écartent sur la scène de mon existence pour me dévoiler des coulisses impossibles aux perspectives faussées, où les lunes sont déformées comme des œufs.
Tout me revient. Tout ce que j'ai vécu là-bas, au Néant. Le contact avec le dessin. Mes amis, Hinrik, Egill. La Tortue, Stéphane. La poursuite, dans le dédale du Cap, Bram et ses mocassins, notre duel au centre du Kor, tout en bas de l'escalier, en haut, sur le jet d'eau jaillissant du dos de la baleine, dans le vide où se niche le Ptyx. Et Bram, qui hurle.
Bram hurle. Tout son être est dissocié, semblable à ces assemblages de pixels qu'il alignait sur son écran. Tous les pores de sa peau se creusent, le réseau fibreux qui le reliait, qui maintenait la cohésion intime de son corps, se tend, se distord, se plie. Il se désintègre sous mes yeux sans que j'intervienne. Je me tiens seul devant le socle, à présent. Le Kor finit de digérer Bram et je sais que je n'ai qu'une fraction de seconde pour agir. Tout m'a mené à ça : un moment suspendu où, avant de devoir affronter la terreur de l'infini, je peux me saisir de l'objet de mon désir. Je peux devenir Maître du Néant, ici et maintenant, puisque le dernier prétendant expire en hurlant. Je peux avoir ce courage car cette ligne, ce pli fatal, je le vois, exercé par toutes ces années de dessin au trait, ce dont Bram était incapable, enfermé dans la discrétion de sa perception. J'entends battre mon cœur. Par l'interstice, j'élève les mains vers le Ptyx. Je sens irradier son énergie, à la fois brûlante et glacée. Il est nimbé d'un halo de glace fondue, où la neige et la lave, enlacées, boivent aux lèvres l'une de l'autre. Je n'ai pas peur. Je ne suis pas venu ici pour reculer.
Comment m'y suis-je pris ? Pourquoi suis-je incapable de me souvenir de ce que j'ai fait au sommet de cette tour, quand mes doigts se sont tendus vers le Ptyx ? Tout le reste, toutes ces années ne sont qu'un intervalle entre ces instants de poursuite, à tenter d'empêcher la destruction de la réalité. Pourquoi un seul moment continue-t-il à m'échapper ?
Je regarde autour de moi. La chambre d'Elliot, où j'ai commis un massacre. Je me lève, je déplace d'un doigt toutes les silhouettes officielles, fondues les unes dans les autres. Finies énigme et les chausse-trappes. Il n'y a plus de mystère. Je suis à présent tout entier ici, sans mes moufles, sans mes doutes. La clé de la dernière cassette, c'est moi. Depuis le début, j'étais appelé à être celui qui transformerait la matière.
J'empile tous les meubles, les uns sur les autres, convergents, lit, armoire ainsi que les derniers objets. Dans l'angle, le pli où Elliot s'est enfui forme un clin d'œil, que j'ouvre en grand pour y placer la pièce elle-même. Il n'y a bientôt plus que la roche, et ce son toujours plus lourd, ces infrabasses qui pilonnent ; le son d'une cérémonie, celle de ma naissance, de mon ascension, car je suis revenu du Néant et je suis Maître de tous les mondes qu'il autorise, ces cercles concentriques que mes dessins créent dans le sable et qui s'encombrent de matière pour devenir stérile réalité où s'embourbent nos ego, nos âmes fatiguées. Quand j'ai fini de détruire la pièce, il n'en reste qu'une ligne, écrite sur le fragile espace où la terre rencontre la mer.
Au sol, l'éventail déplié donne à voir cette même image de montagne, enroulée de nuages, dont les formes fantastiques laissent deviner dragons, chimères, nixes, sirènes ailées, hideux polypes. Chacun de ses soufflets, légèrement ombré, est une perspective, elle s'étire dans ma vision comme les voiles d'une ondine entre deux eaux, à peine noyée, les yeux grands ouverts sur le ciel. Elle plonge dans mes yeux. J'ai l'impression qu'Elliot me regarde.
KOR
Nous sommes ton miroir brisé
Dans chaque fêlure, tu vois une lame
Qui fend le voile.
Sur le parvis d'Hamarinn, Nik scande son refrain à une foule féerique. Des buvettes et des stands de jeux parsèment l'esplanade, pointillée de couleurs dans le midi. Le soleil se lève doucement à l'horizon, sourire discret déjà presque figé. Tout ce qui naît tend à disparaître. Ici, dans ce pays à cheval sur deux plaques tectoniques, ça n'a jamais été aussi vrai.
Toutes les silhouettes tremblent. Toutes les lignes claires qui détourent le monde forment des accordéons, vibrations qui s'harmonisent les unes avec les autres. Je n'entends pas le son du Néant, ce son du silence, énorme, qui a rompu la membrane originelle pour couler et former la matière différenciée, mais je peux sentir sa présence, vaste chambre d'écho. Je pense à Bram, à ses yeux, à Stéphane, puis à Elliot, à tous ceux qui ont tenté d'approcher le secret du Ptyx pour se retrouver assimilés dans la masse informe de l'essence éternelle. Moi, je me suis échappé, moi, j'ai trouvé la force de m'en aller.
KOR
Voilà, enfin : tu as compris
Tous ces traits dont tu es témoin
Tu es leur mère.
Fink me rejoint, haletant.
— Quoi ? dis-je.
— Cyldrid. Elle se meurt.
Je le vois, si vieux, son réseau de lignes, de rides et d'interstices plus épuisés encore, courbes, crevasses, abysses de lui-même.
Il m'emmène près d'une roche couchée, un de ces lits de pierre où les fées alanguies des contes étaient censées attendre les adolescents montés pour les prendre. Cyldrid est allongée, entourée de petites fées en anorak. Ses pauvres yeux béants fixent un ciel aux couleurs changeantes, si rapide, illisible. Elle ânonne.
— Un nouveau Maître. Un nouveau temps.
Un petit sylphe lui tient la main.
— S'il vous plaît Gryla, ne mourez pas.
— Où est le maire ? me demande Fink. Où sont les sismologues ?
Je ne peux pas lui dire. Je ne saurai pas comment lui dire.
— Partis.
— Vous voulez dire que... l'école est sauvée ?
— Oui. Si le volcan n'explose pas.
Il sourit.
— J'aurai aimé qu'elle vive pour voir ça.
Il tient Cyldrid tout contre lui. Je comprends. Depuis combien de temps se connaissent-ils ? Peut-être étaient-ils alliés, dans leur combat pour faire respecter la parole des fées. Puis elle s'est détournée du paganisme et s'est ralliée à la parole divine, tandis que Fink continuait seul son chemin.
Fink se penche vers moi.
— Mais vous savez... Les fées. Elles vont venir la sauver.
Je ne sais pas comment lui expliquer qu'il n'y a pas de fées. Je n'en ai pas le courage. Je ne sais pas si je veux.
— Vous avez raison, Fink. Elles vont venir.
Ses vieux yeux s'embuent, il est rassuré. Entre ses bras, Cyldrid est secouée de spasmes.
— Ma mère disait que Dieu dort dans les pierres, rêve dans les fleurs, s'étire dans les animaux, et se réveille dans l'homme.
Ses yeux se mouillent.
Je le regarde, je regarde Cyldrid. Je regarde les enfants. Nik Kershaw monte sur scène, gabardine et cheveux dans le vent. Il entonne _The Riddle_.
Un moment. Que je puisse découper. Je ne sais pas combien d'instants vont composer ce moment. Mais je sais que c'est comme ça que je fonctionne. Je peux tout expliquer. Car il n'y a pas de fées, ni de Dieu. Il y a moi.
Et ma main, qui peut guider.
Soudain je comprends pourquoi je porte des moufles. Bram pensait que le châle de Stéphane le protégerait, mais non. Tous ceux qui sont venus ici prendre le Ptyx voulaient le toucher de leurs doigts d'humains, la peau, les sens, les nerfs, tout exacerbés par le désir de vouloir sentir le vase de la gloire secrète où le jus du Rien coule à jamais. Ils voulaient, dans leur folie analogique, éprouver l'humain en eux face aux périls qui hantent les vastes étendues de l'univers. Ils étaient grands, mais fiers, trop fiers. Ils étaient arrivés jusque-là pour ne jamais reculer, et devenir Maître. Mais le prix à payer était trop lourd. Stéphane avait scellé le destin de tous les prétendants au Ptyx. Personne ne pouvait s'en saisir à mains nues. Comme un athlète s'entraîne avec des poids aux pieds pour éprouver plus tard une certaine légèreté. Il ne voulait pas que quelqu'un puisse s'emparer du Ptyx sans en avoir compris l'essence. Car le Ptyx est l'essence. Et Bram, lui, a tenté d'attraper l'impossible entre ses mains d'humain. Peut-être qu'en restant derrière son écran, il serait parvenu à saisir le pli. Avec du temps. Beaucoup d'argent, du travail. Il aurait pu matérialiser l'espace entre les pixels. Ainsi, il aurait approché le Ptyx avec ses moufles digitales. Mais il est entré ici en âme analogique. Il n'a pas su rester fidèle à sa démarche. C'est ça, au final, que nous dit Stéphane : chacun sa méthode, chacun ses moufles. Lui les vers, moi le dessin, Bram le pixel. Et Elliot, le jeu de cache-cache.
Je fais glisser les traits qui composaient la boule à mon majeur, cette callosité bizarre. Entre mes doigts, elle devient une petite écaille en forme de médiator.
Fink arrondit les yeux.
— Mais comment ?
— Le Ptyx, dis-je en brandissant le médiator.
— Le Ptyx, mais... mais... où l'avez-vous trouvé ?
— Au Néant.
Je lève le Ptyx et je déplace tous les nuages. Un ciel d'un bleu intense fait surface entre les vagues moutonneuses. La foule soudain s'enflamme. Jamais le soleil n'avait ainsi dardé la ville de ses rayons à pareille saison. On a oublié sa couleur, son odeur, sa chaleur onctueuse comme de la confiture étalée sur le canevas d'un azur infini. Les enfants ont levé leurs mains vers l'astre, les parents en lâchent leurs assiettes pleines de tête d'agneau. Une liesse s'empare du parvis, de toute l'esplanade, et bientôt on entend monter de la ville une clameur. Un frisson physique qui me fait vibrer, qui me rappelle...
— Racontez-moi, me supplie Fink. Je veux tout savoir. Vous avez trouvé Elliot ?
— Non. Elliot est parti pour toujours.
— Où ?
— Sous la mer émerveillée.
— Où est-ce ?
— De l'autre côté du Néant.
Du Ptyx, je fais le ménage dans les montagnes, afin de libérer un point de vue dégagé vers l'horizon. Seul le cratère du Burfell me résiste. La perspective ne me permet pas de saisir sa circonférence. Mais la fumée qui s'en échappe n'augure rien de bon. Je crois percevoir dans le sol plusieurs sortes de vibrations nouvelles. Toute vibration peut par définition se visualiser. J'essaye de faire la part de la musique, des vivas, de la clameur, du grondement sourd du volcan qui se réveille. Prendre une onde et la tordre. Peut-être que je pourrais faire changer le ton du volcan, le réduire au gazouillis d'un d'oiseau. J'aimerai faire de ce monde un pays nouveau. Cette terre minérale, battue par les vents. Cet exil où mon âme s'est fait un nid. Peut-être que je n'aurai pas la chance d'être Elliot, et de partir loin. Je ne serai pas non plus comme Stéphane, seul dans ma tour. Je suis revenu ici pour accomplir une tâche.
La mienne.
Alors que je m'empare du Ptyx, tout se replie depuis cette fente qui est le tout, sur ce piédestal si fin qu'il existe à peine et parfois s'efface. Au centre de cette impossible fissure, je deviens l'origine et le point de vue. Je sais qu'ainsi un monde pourra naître, dont je serai le Maître. Moi, le porteur de moufles, je serai celui que le Kor n'aura pas. Je ne sais pas si Elliot a abandonné le Cap car il se savait condamné. Il redoutait peut-être une trop lourde responsabilité et ce qu'il allait advenir de lui, une fois absorbé dans les méandres de ce pouvoir cosmique. Ou encore craignait-il les peurs les plus minuscules, celles qui se nichent entre le sol et le lit, dans cet espace sombre où les monstres font leur tanière. La peur de ce qui n'est pas, de ce qui ne saurait être. Non pas la peur du vide ou de l'inconnu, mais la peur de ce qui s'échappe, de ce que l'on croit connaître et qui finit par violer nos idées les plus pures, les plus décemment formées. Elliot est parti, il a déserté cette tour pour s'enfoncer dans le petit trou sur la plage, celui qui mène à l'au-delà, où la pensée s'abolit. Elliot a cessé d'être le Démiurge, car il savait que tout démiurge finit toujours par être remplacé par un autre. Les Maîtres ne sont pas Dieu, ils sont de simples intendants. Et comme je suis le nouveau, lorsque je prends le Ptyx le Kor s'efface, s'incline révérencieusement, me jurant fidélité. À moi. Qui peut décider de tout. Surtout. De.
... faire et défaire. Je suis le maître d'œuvre du monde réel, à présent, revenu du Néant pour trouver un sens à ce qui n'en a plus. Tout ce que j'ai pu expérimenter sans m'en rendre compte, masqué à moi-même, prend désormais tout son sens : je peux changer le monde, sa définition. Je peux voir sa silhouette et armé du Ptyx, je peux changer ces silhouettes. Tout ce que je découpe est emporté dans l'intervalle, replié dans le Rien, pour reprendre substance, ailleurs, plus loin, plus profond là où se niche l'origine. Tels des ronds dans l'eau, créés par le remous d'un caillou, j'agis là-bas pour instruire ici. En ce moment suspendu où je crée le monde à mon image, je décide d'y incorporer des fées. Je veux que tout le monde soit content. Je veux être celui qui offrira le bonheur d'un monde meilleur. J'ai le pouvoir. Je suis le porteur de la flamme, de cette magie qui fait danser les lignes du Vide pour en tirer de nouveaux sons. Je suis revenu de l'envie et du doute pour incarner un idéal de force, entre le chant et le trait, entre l'amour et la compassion, tirant de moi-même tout un continent, tout un univers où nicher la pensée, où incarner la joie de retrouver la grâce du trait.
— Bracken ? demande Fink.
Je l'entends à travers un rideau d'ego, toutes mes pensées, confuses, qui dictent ce que je deviens, obsessionnel, totalement soumis à mon propre élan. Il semble vouloir me dire quelque chose, je sens ses mains se poser sur moi, mais je ne saisis pas le sens de sa démarche. Je suis trop absorbé. Toutes les lignes s'assemblent pour créer de nouveaux principes que j'avais ignorés jusqu'ici. Rien ne paraît simple. Il y a des aspérités partout. Des angles qui m'ennuient. Des saillies, des profondeurs, des taches. Pourquoi toutes ces taches ?
— Bracken, je ne comprends pas, pourquoi êtes-vous revenu ? Pourquoi n'êtes-vous pas au Néant ?
— C'est ici, le Néant.
KOR
Voilà.
Le danger.
Toujours.
D'abord, leur fabriquer des fées, histoire de les calmer. De les éloigner. Quelques reflets dans une mare de pluie me suffisent à créer de nouvelles particules, dont j'étire les ailes. Je les fais battre et je les suspends, en galaxie de petites choses qui attirent l'œil. Les enfants se ruent sur elles, je vois leurs mots se détacher, les impacts de leurs exclamations deviennent des bulles, puis des onomatopées. Ça ne me gêne pas. Au contraire, je peux jouer avec. M'en servir pour y percher de petits nids, fabriqués avec le creux des rochers, comme une toile d'araignée, enroulée sur elle-même encore et toujours. Toutes les petites étoiles que je trouve, les scintillements de quartz, les flammèches, les touchantes intentions qui enfièvrent les yeux des enfants, je les transforme en fées. Je fais déferler une féerie sur ces lieux mornes. Tout s'illumine. Brusquement, l'Islande s'anime : elle qui depuis toujours avait nourri l'imagination de ses habitants du silence de ses pierres, elle se met à sourire. Partout dans les rues, on sort pour assister à l'avènement. À la restauration. La féerie revient. Elle a toujours été là. Sur une butte, non loin du port d'Hafnafjordur, je fais apparaître une cavalerie, composée à partir d'une petite herbe locale dont j'ai étiré tous les pétales. Sur le plus grand des chevaux, un roi revient sur ses terres, en fait un simple oiseau dont j'ai modifié la colonne vertébrale. De ses amis, j'ai fait des elfes chevaliers. Leur cavalcade aimante soudain tous les regards. Dans la ville, tout le monde se lève. J'ai pris la poussière des rues, j'en ai drapé des voilages pour composer des étendards. C'est jour de fête, jour de révolution. Gryla se meurt sur un rocher. Il nous faut célébrer l'instauration. Tous les derniers accords sont des petites notes, dont je me sers pour donner une voix au monde, aux plantes, aux herbes. J'ai besoin d'un héraut, pour réveiller la reine. Ce petit garçon fera l'affaire : je le réduis de trois quarts et efface tous ses traits, pour ne conserver que sa bouche.
KOR
Trop loin
T'emmène l'ego.
Reste.
Fink tente de m'arrêter.
— Bracken !
Il me retient, ses yeux emplis de terreur. Je ne comprends pas.
— Vous vouliez voir les fées, je vous les offre !
— Mais enfin Bracken, vous ne pouvez pas ! Je vous vois faire le chef d'orchestre, Bracken, vous changez, vous devenez plus grand, qu'est-ce qui se passe ? Qu'est-ce que vous êtes en train de nous faire ?
— Je suis le Maître, Fink. Cette réalité m'appartient !
— Vous êtes Maître au Néant ! Pas ici ! Comment êtes-vous revenu, pourquoi nous faites-vous ça ?
— Vous ne voulez pas du merveilleux ?
— Je... Je ne sais pas.
— Le Néant commande cette réalité. C'est depuis le Néant que je change tout. Je suis dans l'interstice, Fink. Vous me voyez mais je ne vieillis pas, je suis immuable, je suis dans l'entre-deux, et chacun de mes gestes affecte votre monde.
— C'est mal, Bracken !
— Mal ? Qui êtes-vous pour me contredire ?
Une nuée de fées vient l'encercler, lui tirer les cheveux. Il se débat dans l'essaim. Je ne sais pas ce que je vais faire de lui. Soudain, il m'ennuie. Il ne comprend pas que j'ai fait tout ça pour lui.
KOR
Non.
Pas pour lui.
Pour toi, toi, toi.
Très bien, puisque c'est ce que tu veux, je vais te rendre heureux. Je vais te donner la merveille des merveilles, la double-vue, la pensée magique.
— Bracken, je vous en supplie.
Je le tords, je le noue, je le retaille. Je tire tous ses traits vers un point de fuite pour lui donner le visage d'un homme mûr, revenu d'une longue chasse. Je le cloue sur un de mes chevaux, sur la butte, signalé par un feu follet. Je prends ensuite Cyldrid, je lui tire les oreilles, les yeux, puis de la courbure de ses omoplates, je déplie deux longues ailes. Elle sera ma reine, et d'une autre mère, je fais sa courtisane. Voilà. J'entends des hurlements. Je ne me suis pas rendu compte de ce que je lui ai fait, mais quand j'ai commencé à travailler ses seins pour en faire des couronnes, j'ai déchiré un bout de sa personne. Pas le temps de s'en occuper maintenant. Je me redresse, immense, au-dessus de l'école. Les acclamations explosent en feux d'artifices. Le concert s'achève en riff, sinueux, dont je pitche chaque note. Les cheveux de Nik deviennent une forêt où vont s'enfoncer les derniers larsens d'une guitare que je transforme en lyre. Soudain, alors que mes doigts s'emballent, que la création devient frénétique, que je ne sais plus où m'arrêter, quelque chose en moi s'immisce, une autre voix, venue de très loin. Comme le terrible murmure d'une maman dans une nuit sans étoiles.
KOR
Non
Stop
Arrête.
Qui es-tu pour m'arrêter ? Le Kor ? L'ancien garde-fou du Pénultième ? Comment as-tu fait pour me suivre jusqu'ici ? Oh, je sais. Il y a un peu de Bram en toi, n'est-ce pas ? Bram et ses tours de passe-passe en pixels, ses torches en plastique. Qu'est-ce que c'était, une sorte de lien avec tes anciens ordinateurs, c'est pour cela que tous tes étudiants ont fondu ? Tu les as utilisés jusqu'au bout, tu as sucé leur énergie vitale pour alimenter ta survie ici, dans ce monde. Je suis impressionné, mais ne crois pas que ça te sauvera. Tu n'es qu'une bouillie, le mauvais souvenir de toutes mes erreurs, de toutes ces heures passées à ne pas rêver, à ne pas me laisser guider par l'instinct qui me poussait à partir au loin, à être un autre. Combien de temps ai-je perdu, enfermé dans ma tête, sous ma couette, à te craindre, Kor, toi et tout ce que tu représentes ! Regarde ce que moi je suis devenu ! Je suis ton Maître maintenant ! Tu feras ce que je te dis !
KOR
Non.
Je vais
T'arrêter.
Non, tu ne feras rien. Car regarde, je te pense en mots, tu me parles en mots. Mais ma pensée elle-même se pense à présent. Je suis en dehors de toi. Tu crois que tu m'entoures, mais je suis au-delà. Je peux te replier, comme un mouchoir, dans ma poche, car c'est tout ce que tu es : l'image d'un mouchoir. Tu n'as pas d'épaisseur, plus rien n'en a. C'est moi qui donne l'horizon, la perspective. Je te dénie ce droit à exister sur un plan. Je fais de toi un simple autocollant, collé de travers dans un album d'enfant. Il n'y a rien que tu puisses faire pour m'arrêter. Il est trop tard. Tous ces poètes que tu as engloutis ne sont que des voix du passé, qui tentent de retenir quiconque voudrait donner un nouveau ton. Un nouveau sens. Je suis plus fort désormais. Le Ptyx et moi ne faisons plus qu'un, je suis venu à travers le voile pour rendre ce monde meilleur. Véritablement magique. Ça t'étonne ? Pourquoi ? C'est vrai, je ne crois en rien, mais je ne me suis pas laissé entraîner sur cette pente sans penser que, tout au bout, il y aurait ce but, ce rêve. Moi, Bracken, l'homme qui ne rêve pas, qui vient de vivre un rêve éveillé dont il est sorti transformé. Voilà, ce que tu es : un souvenir, au réveil, dans les brumes.
KOR
Non.
Je veux
le bien.
Le bien ? Le bien de qui ? Le tien ? Celui de Stéphane ? Il est mort depuis cent ans, et son héritier a tout abandonné. C'est moi maintenant, il n'y a rien d'autre. Tu feras ce que je te dis. Je t'ai vaincu, toi et tes stratagèmes. Tu es venu ici avec moi, depuis le Néant. Je ne t'ai pas vu. C'est toi qui as décollé la tapisserie. C'est toi qui a fait ça à Cyldrid. Je ne t'ai pas vu car tu étais dans l'éventail. Un éventail a deux faces. Il se déplie. Il se replie. La cachette parfaite. Et maintenant, tu es là, à me regarder, sans pouvoir me toucher, sans pouvoir m'empêcher d'être Maître ici, dans le Plein et non dans le Vide. Et tu crois que ta morale m'importe ? Si tu es la mémoire de toute la poésie qui a échoué à changer le monde, si tu es cette humble soupape dont Stéphane se gargarisait, alors moi, je te fais sauter. Un poète n'a plus à chanter dans le désert, il peut à présent être ici et tout changer. Tu aimerais savoir pourquoi je ne suis pas resté là-bas, quand je me suis emparé du Ptyx ? Il était en mon pouvoir de rester, comme Stéphane. Mais je suis là, et tu m'as suivi comme un toutou parce que tu savais ce qui risquait de se passer si je comprenais l'étendue de mon pouvoir. Tu n'aurais pas dû. Regarde, je te prends, je te plie, je te mets dans ma poche, petit mouchoir. Reste là. Tu me serviras plus tard, pour empêcher les autres de me ravir mon trône.
KOR
Non, ne laisse pas
(...)
nous envahir...
Je souffle. La nuée s'efface, comme des chiures de gomme sur une page blanche. Bientôt, le clair-obscur redevient calme et froid. Mon soupir se cristallise dans le froid, je le disperse du pouce. Seul, enfin. Personne pour m'entraver. C'est effrayant, ce qu'on peut faire quand on peut tout faire. Donner des pichenettes aux silhouettes d'enfant pour les faire vibrer, ça les chatouille, ça les fait rire, et ces rires, ces éclats, j'en fais des fruits que je suspends aux arbres que je fais pousser à partir de rien. Il n'y a pas d'arbres ici, seulement de minuscules petits sapins, mais il y en aura, parce que j'en veux, parce que j'en ai décidé ainsi.
Voilà la véritable raison de mon retour, pourquoi je ne suis pas resté au Néant. Parce que le mot ne m'intéresse pas. Parce que le Néant est un endroit pour les Maîtres qui se vautrent dans le stupre du verbe. Moi, les mots, je n'en veux plus. Je veux ce monde, ces images. Je veux m'abandonner au flux des formes, des compositions. Ma pensée même est une insulte car, alors même que je pense ces mots, ils restent des mots. Si je pouvais penser en images alors, peut-être... Mais la clé ne se trouve-t-elle pas dans mon expérience ? Peut-être que méthodiquement, si je poussais un peu, je trouverais l'extase dans ces nouvelles images que je conjure à partir de cette matière que je croyais morte. Toutes couleurs fondues, enfants, adultes, paysage, transformés en bouillie, tout et rien, grosse d'une perfection à venir. Il me suffit de bien regarder.
La fête s'est transformée en panique. Tout est confus, des grappes d'écoliers en fuite, des voitures coincées sur les routes... qu'est-ce qui se passe ? En scrutant le territoire, je comprends que le cratère crache d'énormes volutes de fumée. Rapides, ces formes changeantes me troublent : où est leur point d'accroche ? Où sont leurs frontières floues, qui ne cessent de s'effacer ?
Le volcan me menace peut-être, physiquement. Il serait trop bête de périr ainsi, aux portes de l'immortalité, de la toute puissance. Il a peut-être senti ce qui se passait. Hamarinn est construit sur une faille, entre deux mondes, comme l'Islande est déchirée entre deux continents. Peut-être, quand Elliot est passé, le Pli s'est-il agrandi, laissant la terre s'ouvrir, se fondre, comme quand j'ai vu les couleurs glisser hors des corps déliés de Plouffe, puis des officiels. Et voilà la réponse de la Nature à cette aberration. C'est contre la Nature que le combat décisif se jouera.
Le volcan bouillonnant ne cesse de vomir. Je n'arrive pas à attraper son flot, tout va trop vite, comme cette mer, déchaînée, que je vois refluer dans la baie. Il me faut quelque chose pour boucher. Je pourrais créer quelque chose qui serait plus lourd, plus dense, avec toute cette roche, mais elle risquerait de fondre elle aussi. Une brusque chaleur sur mon visage me fait comprendre qu'il y a un détail que je n'ai pas pris en compte.
Je lève les mains vers le ciel ; agiles, mes doigts tricotent une figure autour du soleil. Le soleil tourne sur lui-même, souriant, entonnant presque un chant, dans sa splendeur circulaire. Le donneur de vie. Je trace sur sa bouche un sourire pour le faire chanter, ses notes sont des arpèges joués sur la température déclinante avec la buée en guise de portée où s'accrochent des fragments de rayons, des illuminations. La symphonie qui prend son ampleur est une coda emportant tout sur son passage, les restes de vie dissociés, que j'entasse en pyramides de détails, mes derniers doutes, mes espoirs pour ce monde. Je suis là pour tout changer. Je suis passé, je ne suis pas simplement Maître au Néant, je suis Maître ici.
Je mesure l'instant, l'aplatissement du moment présent qui me demande, là, maintenant, de faire un choix. Suspendu dans l'intervalle, je me donne le temps d'hésiter. J'y trouve le confort, alors que les clameurs de la fin d'un monde secouent le voile de la réalité. Voici l'éternité : la conscience simple d'un monde sans choix, où toutes les décisions sont prises ensemble, le mot et le dessin, jaillis d'un même élan. Mes yeux cillent, alourdis, bruit d'une grille qui se referme sur les possibles, sur tout le reste qui n'est pas moi.
Dans un soupir, mes doigts se saisissent de l'ordre solaire ; le cercle, pincé par le Ptyx. La musique enfle en paradoxe, en désaccords, en stridences et en hurlements quand fondent les dernières formes de roches et de corps en boules de matière incandescente. L'astre se décroche, je brise son cercle, son miel se répand sur le monde. C'est une confiture, c'est une noyade. Tout redevient indifférencié, uni dans un jaune sublime qui s'irise de l'intérieur. Dans ce grand tout, seule habite ma conscience, ma respiration. Je suis la frontière, je suis la ceinture. Tout me remplit, tous mes vides, mes doutes, mes choix. En brûlant son chemin, le feu me ravage, il tourne, centrifuge, cerné par mes limites. Je le tiens en moi, il s'étire comme un enfant sur le point de naître, dans cette caverne qui est le mot, que sont toutes les lettres qui composent cette dernière conscience, cette prise de pouvoir. Dans mon ventre, dans mon intimité, je contiens le monde, je suis ce qui le sépare du reste, des autres, des yeux qui scrutent. Je suis la dernière marche avant la chute, l'escalier qui devient bâtiment. Je suis entre deux pensées, où se niche la nuance des paradoxes, réunies pour toujours, liées pour dire ce que je suis, composées par le défaut, pour dire qui a le mot qui dit et le trait qui fait. Pour dire le nom du Maître d'après, quand le dernier s'efface, qu'il laisse derrière lui, en creux, encore un trou dont une seule âme est l'écrin.
Moi,
# Notes
Le sonnet en -yx de Stéphane Mallarmé n'est pas reproduit dans son intégralité, par choix éditorial. Stéphane lui-même m'a autorisé en rêve à utiliser le terme de Ptyx, d'une chaleureuse poignée de main. Qu'il en soit remercié. Sa découverte du Néant peut être suivie à travers son poème Igitur. Les travaux d'Yves Bonnefoy sur l'imaginaire mallarméen m'ont été précieux J'ai eu la chance de tenir le châle de Mallarmé entre mes mains, dans le musée départemental aménagé dans sa maison de Valvins (Vulaines-sur-Seine). Stéphane était pudique. Son fils Anatole, « Tole », mourut en 1879 à l'âge de huit ans. Tole avait un jour demandé à son père de lui ramener une baleine en cadeau d'un de ses voyages. Tole porte un pull marin sur les photos que l'on connaît de lui, avec lequel Stéphane l'a enterré.
Le terme de Paidia a été inventé par Roger Caillois dans son livre Man and play.
Habitat fut un projet de courte durée de Lucasfilms games développé en 1986 en partenariat avec l'opérateur américain Quantum Link. Merci à Noah Falstein pour les détails techniques.
The Riddle, titre de Nik Kershaw sorti en 1984, était, selon les mots de Kershaw lui-même, une invraisemblable logorrhée sans cohérence, sur laquelle les fans ont greffé leurs interprétations. Dont acte, même si, pour les besoins de l'intrigue, j'ai assimilé l'île d'Aran à l'Islande. Pardonnez mon audace.
Ce projet a ses anges gardiens : merci Mathias Echenay, Norbert Merjagnan, Célia Chazel, Fleur Marty, Manon, Stéphanie Aparicio, Laure Afchain, merci aux Voltés et à Jean-Claude Dunyach.
Merci à vous !
# Bibliographie de David Calvo chez d'autres éditeurs
Romans
**Nid de Coucou** (nouvelles)
Les moutons électriques, 2007
**Minuscules flocons de neige depuis dix minutes**
Les moutons électriques, 2006
**Acide Organique** (nouvelles)
Les moutons électriques
**Sunk**
Les moutons électriques, 2005
**Wonderful**
Bragelonne, 2001
**Atomic Bomb**
Le Bélial, 2002
**La nuit des Labyrinthes**
J'ai Lu, 2003
**Délius, une chanson d'été**
Mnémos, 1997
Bandes dessinées
**Constellations**
avec Popcube
Ankama, 2010
**La guerre civile mondiale**
pour le collectif anAmnesie
**Le grand doudou**
avec Krassinsky
Dargaud, 2010
**Vorax**
avec Gabriel Delmas
Quadrants, 2010
**Akhenaton**
avec Thomas Azuelos
Carabas Révolution, 2006
**Télémaque**
avec Thomas Azuelos
Carabas, 2004
**AK**
avec Krassinsky et Schwendimann
Carabas, 2006
**Kaarib**
avec Krassinsky
Dargaud, 2002
<http://www.morethancoral.org>
# Autres ouvrages parus aux éditions La Volte
Collectif
**Aux limites du son** (2006)
Nouvelles autours des Vertus de l'inuadible
avec la bande originale du livre (collectif).
Collectif
**Ceux qui nous veulent du bien** (2010)
17 mauvaises nouvelles d'un futur bien géré
Collectif
**Le Jardin schizologique** (2010)
Vous sur une rive, nous sur l'autre, nous resterons étrangers
Yvan Améry
**Âme sœur** (2006)
Découverte de la rentrée littéraire 2005
avec la bande originale du livre de Toog.
Stéphane Beauverger
Le triptyque **Chromozone** , un univers furieux et pessimiste
**Chromozone** (2005)
**La Mémoire du crime** (2009)
**La Cité nymphale** (2006)
avec la bande originale du livre de Hint.
**Le Déchronologue** (2009)
Grand Prix de l'Imaginaire 2010
Prix Européen Utopiales 2009
Jacques Barbéri
**L'Homme qui parlait aux araignées** (2008)
Intégrale raisonnée des nouvelles (1)
**Le Landau du rat** (2011)
Intégrale raisonnée des nouvelles (2)
Le cycle **Narcose** , trois romans déjantés de la Cité-sphère
**Narcose** (2008)
avec la bande originale du livre, direction J. Barbéri et L. Pernice
**Les Noctivores** (2005)
**Le Tueur venu du Centaure** (2010)
Mallarmé en moufles, l'Islande et la kermesse de l'école.
Philippe Curval
**L'Homme qui s'arrêta** (2009)
Journaux ultimes (nouvelles)
Alain Damasio
**La Horde du Contrevent** (2004)
Grand Prix de l'Imaginaire 2006
avec la bande originale du livre d'Arno Alyvan
**La Zone du Dehors** (2007)
Prix européen Utopiales 2007
**Ceux qui nous veulent du bien** (2010) - Collectif
17 mauvaises nouvelles d'un futur bien géré
**Le Jardin schizologique** (2010) - Collectif
Vous sur une rive, nous sur l'autre, nous resterons étranger
**Aucun Souvenir Assez Solide** (2012)
Dix volumes d'air en attendant les furtifs
Valerio Evangelisti
Le cycle romanesque de Nicolas Eymerich.
L'inquisiteur pourfendant sans relâche hérésies et phénomènes étranges.
**Nicolas Eymerich, inquisiteur** (2011)
**Les Chaînes d'Eymerich** (2011)
**Le Château d'Eymerich** (2012)
**Le Corps et le Sang d'Eymerich** (2012)
**Le Mystère de l'inquisiteur Eymerich** (2012)
Léo Henry
**Rouge gueule de bois** (2011)
Derniers jours de Fredric Brown
**Le diable est au piano** (2013)
Fantastiques nouvelles fantastiques
Momus
**Le Livre des blagues** (2009)
Roman freaks
Laurent Rivelaygue
**Poisson-chien** (2007)
_Couché dans ton bocal, Albert Fish !_
Version ePub par LEC Digital Books.
|
{
"redpajama_set_name": "RedPajamaBook"
}
| 8,556
|
Nethergill Club & Day Visitors
Eco Footprint
Nethergill Nature Barn
Nethergill Farm was the result of just three years searching to achieve a lifelong ambition to own land and create a wildlife haven and long term conservation project. Situated in Langstrothdale at the headwaters of the Wharfe on the Dales Way, it is no ordinary farm…
Set in 400 acres between 1,200 and 1,600 feet above sea level, Nethergill nestles quietly into the hillside below the road to Hawes. With 70 acres of 'in bye' or grazing land round the house and the rest peaty moorland, the land is perfect for our rare breed Dalesbred ewes and White Shorthorn cattle. With a grass fed diet, Nethergill meat is very nutritious and is used in our own range of 'ready-to-go' meals, available to all self-catering guests.
The large Victorian Farmhouse and traditional farm buildings were built in 1871 and the original occupant a Mr Woodd was cousin by marriage to the famous naturalist Charles Darwin.
The very beginning
From a college romance (Seale Hayne Agricultural college, Newton Abbot), Chris and I started our farming careers renting land and rearing outdoor pigs on the Loseley Park Estate near Guildford, Surrey. Our pigs enjoyed the organic pastures and chocolate ice cream 'left overs' from the Loseley Dairy. After 10 years of 'hands on' and long days we successfully sold the business and retrained in Graphic Design and Marketing, always with the goal of owning our own farm. We bought Nethergill in November 2005 and in 2014 we finished the development of 2 self-catering haylofts, the field centre and 2 observation hides, as well as installing 'Bertha' our biomass boiler to provide all our heat and hot water and solar panels to provide 75% of our electricity needs.
Wildlife Enhancement
To further enhance the wildlife across the farm and to particularly encourage Red Squirrels, we have planted an area of 30 acres of moorland with mixed species woodland totalling 20,000 trees in conjunction with the National Parks and Forestry Commission. The planting area runs close to Greenfield Forest which has been declared the 17th Red Squirrel reserve in Britain. This area of woodland planting has been named Ellbeck Wood after our daughters Ella and Becky. Further woodland areas have been planted with a further 8,000 trees and shrubs on the moorland and gill sides. The tree planting on the farm has had a dramatic effect on the biodiversity of the land and has enticed Black Grouse back to the area after an absence of 40 or so years!
The White Shorthorns
We have built up the herd to a total of 12 cows and their followers.
Considered to be one of the most docile and gentle of breeds they also are some of the hardiest. They are able to live out on grass alone all winter at great heights and on all types of grazing. Originally from the North, they thrive here at Nethergill.
Very free range chickens
Nethergill chickens have access to the whole farm and produce delicious eggs of all colours and sizes. Self-catering guests are welcome to purchase eggs (subject to availability in winter).
Other team members
Pip our wonderful working collie; strong, loving and good with cows and sheep.
On-going conservation projects
Upper Wharfe River Restoration Project - We are working with the Yorkshire Dales River Trust and Wharfedale Naturalist Society to reduce river erosion along the banks of Oughtershaw Beck running across Nethergill land. Willow bundles have been laid along the banks, interspersed with trees. By reducing the erosion it will reduce the level of silt in the water which will benefit trout breeding grounds, help to reduce flooding lower down the Wharfe and enrich the habitat all along the river, encouraging Water Voles and Otters amongst others.
Grip blocking carried out in 2016 by Yorkshire Wildlife Trust - Peat Partnership.
Self-catering and Field Centre
We are now getting closer to our long term vision and have built two self-catering apartments (or haylofts) and a Nature Barn within our existing Shippon and barn, now ready for guests to enjoy. We aim to demonstrate how food, farming and the environment can all work together and provide a wonderful 'get away from it all' haven for visitors to enjoy. The haylofts will be as in keeping with this aim and are as eco-friendly as possible; providing luxury and comfort within a stunning environment. The Nature Barn is available for groups, courses and individuals to relax, see, observe, learn and enjoy the amazing wildlife sights on the farm.
Bertha the Biomass boiler
To reduce our carbon footprint we installed a biomass boiler in February 2012 to provide heating and hot water to all the self-catering accommodation as well as the farmhouse. The boiler sources timber from the woodland on our boundary and other forestry within the Yorkshire Dales National Park. It operates using a gasification process which means that the energy conversion is very efficient. We use a firewood processor on the farm to process the timber into short lengths for firewood.
To further reduce our carbon footprint we installed solar panels on our timber storage barn making us 75% self sufficient in energy.
Bird Hides
Two bird hides - River Hide and Fell Hide were built in April 2013. Guests are able to use the hides which overlook the scrape, the moss, both up and down the Oughtershaw Beck and into the woodland areas. The hides are made out of recycled plastic bags.
As a keen painter before I moved to Nethergill, I am now surrounded and inspired by the unique atmosphere of the Dales. Mixed media including paper and wool and more recently oils are favourites of mine. As winner of the Pro-Arte competition run by Craven District I enjoy producing commissions for both locals and visitors to the Dales. I am also a member of the local art gathering - Buckden Art Group - exhibiting August Bank Holiday each year.
Prints both limited and unlimited copies, framed and unframed are available of all artwork, as well as A5 cards. Click here to see paintings.
Awarded GOLD by Green Tourism 2015, 2016, 2017, 2018 and now 2019
Nethergill has given us a unique opportunity to live in one of the most beautiful parts of the country, within which we can strive to enhance, develop and protect the surrounding countryside. Our long-term aim is to let nature take the lead and reduce our stocking rates, introducing a rewilding concept to the farm and thus creating a wildlife haven combined with sustainable farming practice, which can both financially support us and be enjoyed by other like-minded countryside and wildlife enthusiasts.
Nethergill Associates
Nethergill Associates was set up in 2005 - a business consultancy aimed initially at helping farmers develop their businesses. Farmer advice has developed into many SMEs using our services and now in 2019 many conservation bodies and individual farmers are using us to prepare for the post Brexit environment.
Blog prepared by Chris Clark for The Green Alliance
Blog prepared by Fiona Clark for the Yorkshire Peat Partnership
Nethergill Farm, Oughtershaw, Buckden nr Skipton, BD23 5JS. T: 01756 761126 E:fiona.clark@nethergill.co.uk
Privacy Policy | Terms & Conditions | Accessibility Statement
Designed and Built by Purple Creative Studio - Login
Copyright Nethergill Farm 2019
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 9,006
|
\section{Introduction}
The Snyder noncommutative spacetime \cite{Snyder:1946qz} is the earliest and one of the most studied proposals for describing quantum properties of spacetime at the Planck scale.
It is generally understood as an interesting case of spacetime whose coordinates have discrete spectrum\cite{Lu:2011it,Amelino-Camelia:2014mea},
as one can most rigorously establish within a
manifestly-covariant analysis \cite{Rovelli:1989jn, Reisenberger:2001pk, Gambini:2000ht, Halliwell:2002th} on the kinematical Hilbert space.
However, this discreteness leaves no trace\footnote{Similar conclusions were also reached in studies \cite{Mignemi:2013aua} of the semiclassical limit of the Snyder model.} on the physical Hilbert space \cite{Amelino-Camelia:2014mea}, obtained by enforcing
the on-shellness requirement for physical particles. In an appropriate sense it is a spacetime whose unphysical/abstract points are quantized but whose physical events (crossing of worldlines of on-shell particles) are indistinguishable from those of ordinary (commutative) Minkowski spacetime.
Here we provide further results in support of this picture
by analyzing the thermodynamical properties\footnote{Some thermodynamical implications of Snyder noncommutativity
were previously studied in \cite{Nozari:2015iba}, but only considering
a ``nonrelativistic" (Galileian) regime and restricting the noncommutativity to three-dimensional Euclidean space.}
of a gas of photons living on Snyder spacetime. We also show that the same conclusions are reached in
its generalization introduced in \cite{Quesne:2006fs,Quesne:2006is} as a covariant extension of the deformed model of Heisenberg phase space proposed in \cite{Kempf:1996fz}.
Our main objective, however, is to
propose the Snyder model and its generalizations as a rather natural arena for
testing
different proposals for the notion of dimensionality of a quantum spacetime:
in light of the observations we just summarized,
physically meaningful notions of dimensionality
should describe such spacetimes as trivially four-dimensional, without any running with scales.
In order to make our case for this, we use the (generalized) Snyder model to compare the behavior of two different notions of dimensionality that can be applied to quantum spacetime. The thermal dimension, first proposed in \cite{Amelino-Camelia:2016sru}, is computed by exploiting the fact that some thermodynamical properties of a gas of massless particles living on standard Minkowski spacetime scale with temperature in a way that depends on the dimensionality of the spacetime itself. One can then associate an effective dimension to the quantum spacetime by looking at the scaling with temperature of the same quantities, evaluated for a gas of massless particles living on it. The spectral dimension is a geometrical notion of dimensionality that has been used in several approaches to quantum gravity research \cite{Ambjorn:2005db, Benedetti:2009ge, Benedetti:2008gu, Calcagni:2013vsa, Litim:2003vp, Horava:2009if, Modesto:2008jz, Amelino-Camelia:2013cfa, Arzano:2014jfa}. It is deeply linked to the properties of the Euclideanized d'Alembertian of the theory, and can be thought of as the effective dimension probed by a fictitious diffusion process on the Euclideanized spacetime.
In the following Section \ref{sec:SnyderDef} we present the Snyder model and its generalizations, defined by the deformed Heisenberg commutators of phase space coordinates. We construct the kinematical and physical Hilbert spaces, as first developed in \cite{Amelino-Camelia:2014mea} for the Snyder case. We show that the nontrivial spacetime properties encoded in the kinematical Hilbert space (\emph{e.g.} coordinate discreteness) do not survive in the restriction to the physical Hilbert space, that behaves as that of undeformed Minkowski spacetime. To make this point stronger, and emphasize that it is a significant result, in Subsection \ref{sub:kMHS} we contrast it to the nontriviality of both the kinematical and physical Hilbert spaces of the $\kappa$-Minkowski model for spacetime noncommutativity \cite{majid1994bicrossproduct, lukierski1995classical}. The fact that the (generalized) Snyder geometry is discrete, but this discreteness is not observable, leads us to argue that indeed the model is a good test ground for evaluating the sensitivity to unphysical spacetime properties of different notions of dimensionality. In particular, we expect that any physically relevant notion of dimensionality should not be affected by the unobservable discreteness characterizing the (generalized) Snyder spacetime. In Section \ref{sec:thermaldimension} we show that indeed the thermal dimension is not sensitive to unphysical properties, since it is constantly equal to the standard value of four for both the Snyder model and its generalizations. As a cross-check, we show that however the thermal dimension does show a running if we modify the physical properties of the model by taking higher-order functions of the energy-momentum invariant to characterize the dispersion relation. So the thermal dimension does flag physically relevant departures from standard Minkowski spacetime. The spectral dimension is dealt with in Section \ref{sec:spectraldimension} and it demonstrates to be dominated by the unphysical geometrical properties of the (generalized) Snyder spacetime, with its value running from the standard value of four in the infrared to smaller values in the ultraviolet. Matters become even worse if we modify the on-shell condition as done for the thermal dimension. In fact, in this case the spectral dimension it not always well defined, because of divergences introduced by the Euclideanized d'Alembertian.
\section{The Snyder model and its generalizations}
\label{sec:SnyderDef}
The Snyder model is characterized by the following deformed Heisenberg relations between phase space coordinates:
\begin{eqnarray}
\left[x^\mu,\, p^\nu\right]&=& -i \left[ \eta^{\mu \nu}-\lambda^{2} p^\mu p^\nu \right],\nonumber\\
\left[x^\mu,\, x^\nu\right]&=&-i \lambda^{2}\left(p^\mu x^\nu - p^\nu x^\mu \right), \label{eq:Snyder}\\
\left[p^\mu,\,p^\nu\right]&=&0,\nonumber
\end{eqnarray}
where $\lambda$ is the length scale of coordinates noncommutativity, $\eta^{\mu\nu}=\text{diag}[1,-1,-1,-1]$ and we adopted units such that $c=1$ and $\hbar=1$.
A generalization of this was proposed in \cite{Quesne:2006fs,Quesne:2006is}, that introduced further corrections to the commutators, depending on a new length scale $\bar\lambda$, in principle independent from $\lambda$:
\begin{eqnarray}
\left[x^\mu,\, p^\nu\right]&=&-i \left[ (1-\bar\lambda^2 p_\rho p^\rho)\eta^{\mu \nu}-\lambda^{2} p^\mu p^\nu \right],\nonumber\\
\left[x^\mu,\, x^\nu\right]&=&i \frac{2\bar\lambda^2-\lambda^2-\bar\lambda^2(2\bar\lambda^2+\lambda^2)p_\rho p^\rho}{1-\bar\lambda^2 p_\rho p^\rho}\left(p^\mu x^\nu - p^\nu x^\mu \right),\nonumber\\
\left[p^\mu,\,p^\nu\right]&=&0,\label{eq:genSnyder}
\end{eqnarray}
such that the Snyder model is recovered for $\bar\lambda=0$.
While the Snyder commutation rules in Eq.~\eqref{eq:Snyder} are compatible with standard generators of relativistic symmetries, for the more general model in Eq.~\eqref{eq:genSnyder} a deformed representation on phase space of rotation generators $\mathcal{L}_{ij}$ and of boost generators $\mathcal{L}_{i0}$ is required for compatibility with the symplectic structure of Eq.~\eqref{eq:genSnyder} and with undeformed commutators of the Lorentz sector of the Poincar\'e algebra:
\begin{equation}
\mathcal{L}_{\mu\nu}= \left( 1-\bar\lambda^2 p_\rho p^\rho \right)^{-1} \left( x_\mu p_\nu - x_\nu p_\mu \right).\label{eq:Lorentz}
\end{equation}
Given that for both the Snyder model and its generalization the algebra of symmetry generators is undeformed, in both cases the Casimir takes the usual special-relativistic form,
\begin{equation}
\mathcal C= p_{\mu}p^{\mu}\,.\label{eq:Casimir}
\end{equation}
The construction of the kinematical Hilbert space of the (generalized) Snyder model relies on the representation of the noncommuting coordinates $x^{\mu}$ on the operators $\{q^{\mu},p^{\mu}\}$ that satisfy standard Heisenberg relations,
\begin{align}
\left[ {q}^{\mu}, {q}^{\nu} \right]&=\left[ {p}^{\mu}, {p}^{\nu} \right]= 0, \nonumber \\
\left[ {q}^{\mu}, {p}^{\nu} \right]&=-i\eta^{\mu\nu}\,, \label{eq:pregeometry}
\end{align}
and are the phase-space operators of covariant quantum mechanics \cite{Rovelli:1989jn, Reisenberger:2001pk, Gambini:2000ht, Halliwell:2002th}. One can easily find a family of viable representations,
\begin{equation}
x^{\mu}=(1- \bar\lambda^{2} p_{\nu}p^{\nu}) q^{\mu}-\lambda^{2}p^{\mu}p^{\nu}q_{\nu}+ i \gamma p^{\mu},
\end{equation}
spanned by the new parameter $\gamma$, with dimensions of a length squared.\footnote{Note that this parameter has no role in the definition of the model itself, since it does not enter in the commutators of the phase space coordinates, Eq.~\eqref{eq:genSnyder}, and is only relevant in the mapping between the noncommuting coordinates and the auxiliary operators $\{q^{\mu},p^{\nu}\}$.}
This representation reduces to the one found in \cite{Amelino-Camelia:2014mea} for the Snyder model in the limit $\bar\lambda=\gamma=0$. The states of the kinematical Hilbert space are normalizable wave functions, whose scalar product can be written in the momentum space representation as
\begin{equation}
\langle \psi|\phi\rangle = \int d\mu(p) \psi^* (p) \phi(p)\,.
\end{equation}
The measure $d\mu(p)$ is defined by the requirements that the coordinates $x^{\mu}$ be Hermitian operators on the kinematical Hilbert space (so that the kinematical Hilbert space actually encodes the geometrical properties of the noncommutative spacetime) and that the scalar product is invariant under the action of the Lorentz generators in Eq.~\eqref{eq:Lorentz}.
These conditions select the following measure \cite{Quesne:2006is}:
\begin{equation}
d \mu (p) = d^{4}p \left[1-(\bar\lambda^2+\lambda^2)p^\mu p_\mu\right]^{-\alpha}\,, \label{eq:dmu}
\end{equation}
with $\alpha= \frac{2\bar\lambda^2+5\lambda^2-2\gamma}{2(\bar\lambda^2+\lambda^2)}$.
Because of this deformed measure, the kinematical Hilbert space of the (generalized) Snyder model is nontrivial and correspondingly encodes nontrivial properties of the spacetime coordinates $x^{\mu}$, such as a discrete spectrum \cite{Amelino-Camelia:2014mea}. However, physical observables are not defined on the kinematical Hilbert space, but on the physical one, which is the restriction of the Hilbert space to states that satisfy the Hamiltonian constraint $\mathcal C-m^{2}=0$. And, as we show in the following, in a non-interacting theory (generalized) Snyder noncommutativity leaves no trace on the physical Hilbert space.
The physical constraint (along with the restriction to positive-energy states) can be enforced at the level of the scalar product \cite{Amelino-Camelia:2014mea}:
\begin{equation}\label{eq:physHilbSp}
\langle \psi|\phi\rangle_{\text{phys}}=\int d\mu (p) \delta(p^{\mu}p_{\mu}-m^{2})\Theta(p_{0})\psi^{*}(p)\phi(p)\,.
\end{equation}
Once on-shell, the deformed integration measure \eqref{eq:dmu} reduces to the trivial measure multiplied by a function of the mass, which can be reabsorbed in the state normalization. Then, after integration over the energy, the physical scalar product takes the standard form,
\begin{equation}
\langle \psi|\phi\rangle_{\text{phys}}\sim \int \frac{d^{3}\vec p}{2 \omega_{\vec p}} \psi^{*}(\omega_{\vec p},\vec p)\phi(\omega_{\vec p},\vec p)\,,
\end{equation}
where $\omega_{\vec p}=\sqrt{|\vec p|^{2}+m^{2}}$. And not only is the scalar product undeformed, but also any observable on the physical Hilbert space is so. \footnote{This was demonstrated in \cite{Amelino-Camelia:2014mea} for the Snyder model. The proof can be easily extended to the generalized model of Eq.~\eqref{eq:genSnyder} by noting that the modification of the Lorentz generators described in Eq.~\eqref{eq:Lorentz} reduces to a constant multiplicative factor once the on-shellness constraint is enforced.} This brings us to the conclusion we anticipated in the Introduction, that while the nontrivial properties of (generalized) Snyder spacetime are relevant in the unphysical kinematical Hilbert space, they leave no trace on the physical Hilbert space (of free particles). For this reason the (generalized) Snyder model is an ideal test ground for proposals to characterize the dimensionality of quantum spacetime. In fact, the results just exposed motivate the requirement that any physically relevant notion of dimensionality should also be insensitive to the coordinate discreteness, and in particular it should evaluate the dimension to be constantly equal to the standard value of four.
Before proceeding to analyze the behavior of different notions of dimension applying them to the (generalized) Snyder model, we find it useful to pause and compare the results on the kinematical and physical Hilbert spaces of the (generalized) Snyder model to the corresponding ones valid for another much studied model of spacetime noncommutativity, known as $\kappa$-Minkowski \cite{majid1994bicrossproduct, lukierski1995classical}. This comparison will turn out to be very insightful, since we will show that the triviality of the physical Hilbert space is peculiar to the (generalized) Snyder model, while in $\kappa$-Minkowski the traces of coordinates noncommutativity are not lost in going from the purely geometrical description of spacetime provided by the kinematical Hilbert space to the physical description encoded in the Hilbert space. For this reason, the $\kappa$-Minkowski model would not be a good case study for testing the physical relevance of different notions of dimensionality, since one would not be able to discern whether any nontrivial property that is eventually found is to be ascribed to the physically relevant on-shell theory or is just an unphysical artifact. So it is exactly the triviality of the physical Hilbert space of the (generalized) Snyder model
that makes it a good candidate to evaluate the meaningfulness of different proposal for a notion of quantum spacetime dimension.
\subsection{Aside: kinematical vs. physical Hilbert space of $\kappa$-Minkowski noncommutative spacetime}
\label{sub:kMHS}
The $\kappa$-Minkowski spacetime is characterized by the following commutators between spacetime coordinates \cite{majid1994bicrossproduct, lukierski1995classical}:
\begin{equation}
\left[ {x}_i, {x}_0 \right] = \frac{i}{\kappa} {x}_i; \qquad \left[ {x}_i, {x}_j \right] = 0,
\end{equation}
where $i=\{1,2,3\}$ and $ \kappa^{-1}$ is a length scale. The construction of the kinematical and physical Hilbert spaces can be performed in a completely similar way as discussed above. The representation of the noncommuting coordinates on the phase space operators of Eq.~\eqref{eq:pregeometry} reads \cite{amelino2013relative, Amelino-Camelia:2013nza}:
\begin{equation}
{x}_0 = {q}_0, \qquad {x}_i = {q}_i e^{ {p}_0/\kappa}.
\end{equation}
The relativistic symmetries of $\kappa$-Minkowski are deformed, and are described by the $\kappa$-Poincar\'e algebra \cite{majid1994bicrossproduct, lukierski1995classical}. The action of the symmetry generators on the noncommutative spacetime in terms of the phase space coordinates \eqref{eq:pregeometry} was given in \cite{amelino2013relative}. For our purposes it is sufficient to note that
the Casimir of the $\kappa$-Poincar\'e algebra can be represented as
\begin{equation}
\mathcal{C}_{\kappa} = \left(2 \kappa\right)^2 \sinh^2\left( \frac{ {p}_0}{2\kappa} \right) - e^{- {p}_{0}/\kappa} |\vec{p}|^2,
\end{equation}
and is used to define the on-shell constraint for point particles of mass $m$:
\begin{equation}
\mathcal{C}_{\kappa}-m^2=0\,.\label{eq:HamiltonContraint}
\end{equation}
The kinematical Hilbert space is again equipped with the scalar product
\begin{equation}
\langle \psi|\phi\rangle = \int d\mu(p) \psi^* (p) \phi(p)\,,
\end{equation}
where now the measure on momentum space is of course required to be invariant under the $\kappa$-Poincar\'e relativistic symmetries:
\begin{equation}
d\mu(p)= d p_{0} d^{3} \vec p e^{-3 p_0/\kappa}.\label{eq:kMmeasure}
\end{equation}
The physical Hilbert space, obtained after enforcing the Hamiltonian constraint \eqref{eq:HamiltonContraint}, is characterized by the scalar product
\begin{equation}
\langle \psi|\phi\rangle_{\text{phys}} = \int d\mu(p) \delta(\mathcal{C}_\kappa - m^2)\Theta(p_0)\, \psi^* (p) \phi(p)\,.\label{eq:physHSkM}
\end{equation}
In contrast to what happens in the (generalized) Snyder case, where the fact that the integration measure \eqref{eq:dmu} only depends on momenta via the Casimir turned out to be crucial in establishing the triviality of the physical Hilbert space, it is easy to convince oneself that the scalar product of Eq.~\eqref{eq:physHSkM} is still nontrivial. In fact, not only does the momentum space measure \eqref{eq:kMmeasure} not reduce to the standard one when evaluated on-shell, but the Hamiltonian constraint is itself deformed.
For this reason, the fuzziness characterizing $\kappa$-Minkowski coordinates \cite{amelino2013relative} leaves observable traces in the fuzziness of trajectories of freely-propagating particles \cite{Amelino-Camelia:2013nza}.
One would then expect that any notion of distance that is sensitive to physically relevant properties of spacetime would give nontrivial result when applied to the $\kappa$-Minkowski models. However, it would do so just as a notion of distance that is sensitive to off-shell properties would. For this reason, models such as the $\kappa$-Minkowski one are not ideally suited to select physical notions of dimensionality.
\section{Thermal Dimension of the (generalized) Snyder model}
\label{sec:thermaldimension}
Having motivated the appropriateness of the (generalized) Snyder model in testing the physical relevance of different notions of dimensionality, in this Section we analyze the behavior of the thermal dimension.
In a gas of standard special-relativistic massless particles, both the Stefan-Boltzmann law, describing the scaling of the total energy of the gas $U$ with its temperature $T$, and the equation of state parameter, relating the energy density $\rho$ and pressure $P$, depend on the dimension of spacetime in a well-defined way. Specifically, if the gas of photons lives in a $D+1$ dimensional spacetime, then the Stefan-Boltzmann law reads
\begin{equation}
U\propto T^{D+1}\,,\label{eq:SB}
\end{equation}
while the equation of state parameter $w\equiv P/\rho$ equals
\begin{equation}
w=\frac{1}{D}\,.\label{eq:EOS}
\end{equation}
The thermal dimension of a quantum spacetime can be computed by considering a gas of photons living on such a space and comparing the behavior of the Stefan-Boltzmann law and of the equation of state to those described by Eqs.~\eqref{eq:SB}-\eqref{eq:EOS}. In this way, one can associate an effective dimension to the quantum spacetime model \cite{Amelino-Camelia:2016sru}. In the following, we will focus on a gas of photons whose phase space and symmetries are described by the (generalized) Snyder model defined in the previous Section.
The thermodynamics of a gas of massless particles is encoded in the partition function. The relevant possibly nontrivial ingredients that contribute to its evaluation are the on-shell relation and the measure of integration on momentum space \cite{Amelino-Camelia:2016sru, Gorji:2016gfr}. As was shown in the previous Section, for the (generalized) Snyder model only the integration measure is deformed (see Eq. \eqref{eq:dmu}), while the on-shellness is codified by the standard Casimir, Eq. \eqref{eq:Casimir}. So the associated partition function can be written in a covariant form as:
\begin{eqnarray}
\log Q &=& -\frac{V}{(2\pi)^3} \int d\mu (p)\Big[ \delta(\Omega)\Theta (p^{0})\cdot\nonumber\\ && \cdot2p^{0} \,\log \left( 1 - \text{e}^{- p^{0}/k_{B}T} \right)\Big]\,,
\end{eqnarray}
where $k_B$ is the Boltzmann constant and $\delta (\Omega)$ enforces the on-shell relation $\mathcal C=0$. We can see already at this level that there is no reason to expect the thermal dimension to behave nontrivially. In fact, the only deformation with respect to the standard case resides in the momentum space measure, that however enters in the picture in the same way as it did in the definition of the physical Hilbert space, Eq.~\eqref{eq:physHilbSp}. And, as discussed in that context, once the on-shell condition is enforced on the measure, its nontrivial contribution becomes an irrelevant constant factor. This is in fact what happens also in the partition function.
After integrating out the energy and enforcing the on-shell relation, the partition function reads
\begin{equation}
\log Q\propto -\frac{V}{(2\pi)^3} \int d^3 \vec p \log \left( 1 - \text{e}^{-|\vec p|/k_B T} \right),
\end{equation}
which is the same expression as for a standard gas of photons. So the thermodynamical properties of (generalized) Snyder photons are undeformed and the thermal dimension is constant and equal to four. We would like to emphasize that the fact that the deformed momentum space measure is trivial on-shell played a crucial role in reaching this result.
In order for the reader to compare this case where the thermal dimension is undeformed to one where the model has physically relevant nontrivial properties, and verify that the thermal dimension does flag these, we modify the physical constraint of the Snyder model, using a function of the Casimir \eqref{eq:Casimir} to define the on-shellness condition:
\begin{equation}
\Omega_\xi(p)\equiv\mathcal{C}-\ell^{2\xi}\mathcal{C}^{1+\xi} = 0, \label{eq:CasimirXi}
\end{equation}
where $\xi$ is a positive integer and $\ell$ is a positive constant with the dimension of a length, in principle independent from the other parameters of the model. Note that, being it given by a function of the Casimir, this constraint is still compatible with the (standard) relativistic symmetries of Snyder spacetime.
The partition function then reads:
\begin{eqnarray}
\log Q_{\xi} &\propto& -\frac{V}{(2\pi)^3} \int d\mu (p)\Big[ \delta(\Omega_{\xi})\Theta (p^{0})\cdot\nonumber\\
&&\cdot2p^{0} \,\log \left( 1 - \text{e}^{- p^{0}/k_{B}T} \right)\Big]\,.\label{eq:OmegaXi}
\end{eqnarray}
This is completely analogous to the partition function of the Asymptotic-Safety-inspired model studied in \cite{Amelino-Camelia:2016sru}. As discussed in more detail there, the partition function receives contributions from the two positive-energy solutions of the on-shell condition:
\begin{equation} \label{eq:delta2}
\delta (\Omega_\xi) = \frac{\delta (p^{0}- |\vec p|)}{2|\vec p|} + \frac{\delta (p^{0}- \sqrt{|\vec p|^2 + \frac{1}{\ell^2}})}{2\xi \sqrt{|\vec p|^2 + \frac{1}{\ell^2}}}\,.
\end{equation}
At low temperatures only the first term is relevant, while at super-Planckian temperatures the two contributions are effectively equivalent. So in these two regimes the thermal dimension takes the standard value of four. \footnote{Because of the doubling of equivalent contributions, at high temperatures the numerical value of the energy density is twice as much the value at low temperatures, but is scales in the usual way with the temperature, so it does not affect the thermal dimension.} It is only at intermediate scales that the two on-shell relations are relevantly different and both contribute to the partition function. And in fact at intermediate scales, around the Planck scale, the Stefan-Boltzmann law and the equation of state have a nontrivial behaviour, as shown in Figures \ref{fig:SB1} and \ref{fig:SB2}. Accordingly, at intermediate scales the thermal dimension takes values higher than four, as shown in Figure \ref{fig:SB3}.
\begin{figure}[h]
\centering
\scalebox{0.65}{\includegraphics{Total_energy_thermal.pdf}}
\caption{\label{fig:SB1} Temperature dependence of the energy density of a gas of photons whose partition function is given by Eq. \eqref{eq:OmegaXi}, with $\xi=1$. The temperature is in Planckian units, while the energy density is in arbitrary units. }
\end{figure}
\begin{figure}[h]
\centering
\scalebox{0.65}{\includegraphics{W_thermal.pdf}}
\caption{\label{fig:SB2} Temperature dependence of the equation of state parameter of a gas of photons whose partition function is given by Eq. \eqref{eq:OmegaXi}, with $\xi= 1$. The temperature is in Planckian units.}
\end{figure}
\begin{figure}[h]
\centering
\scalebox{0.65}{\includegraphics{D_ther_modCas.pdf}}
\caption{\label{fig:SB3} Temperature dependence of the thermal dimension of a gas of photons whose partition function is given by Eq. \eqref{eq:OmegaXi}, with $\xi= 1$. The temperature is in Planckian units.}
\end{figure}
\section{Spectral dimension of the (generalized) Snyder model}
\label{sec:spectraldimension}
As mentioned in the Introduction, the spectral dimension is the effective dimension probed by a fictitious diffusion process on the Euclideanized spacetime. It is determined by the dependence of the average return probability on the (fictitious) diffusion time. The average return probability related to the diffusion process is computed as
\begin{equation} \label{eq:AvRetProb}
P(s) \propto \int d\mu^{(E)}(p) e^{ s \,\Omega^{(E)} (p)}\,,
\end{equation}
where $s$ is the diffusion time, $d\mu^{(E)}(p)$ is the Euclideanized measure on momentum space and $\Omega^{(E)} (p)$ is the Euclidean d'Alembertian. The spectral dimension is then defined as:
\begin{equation}
d_s(s)= -2 \frac{\partial \log P(s)}{\partial \log s}\,.
\end{equation}
The ultraviolet value of the spectral dimension, which is sensitive to the (ultra-)Planck-scale properties of the theory, is obtained in the $s\rightarrow 0$ limit. It was already suggested \cite{Amelino-Camelia:2016sru, Carlip:2017eud} that, while the spectral dimension could possibly be useful in characterizing some geometrical properties of quantum spacetime, is it however unphysical, being it sensitive to the off-shell modes of the theory. In the following we expose this by computing the spectral dimension of the (generalized) Snyder model, for which we have shown above that all nontrivial features are unphysical, at least at the level of the free theory.
Using the standard choice of Casimir, Eq. \eqref{eq:Casimir}, the average return probability reads:
\begin{eqnarray}
P(s)&\propto& \int d^4 p \left[1+(\bar\lambda^2+\lambda^2)(p_0^2 +|\vec p|^{2})\right]^{-\alpha} e^{-s (p_{0}^{2}+|\vec p|^{2})}\nonumber\\
& \propto& \int d r \left[1+(\bar\lambda^2+\lambda^2)r^2\right]^{-\alpha}r^3 e^{-s r^2} \,.\label{eq:Psspherical}
\end{eqnarray}
Note that we have performed a Wick rotation $p^0 \rightarrow i p^0$ on both the integration measure and the D'Alembertian and in the second step we have introduced spherical coordinates $r^2\equiv p_0^2+ |\vec p|^{2}$ (the angular variables can be integrated out and do not contribute to the spectral dimension).
The behavior of the spectral dimension is shown in Figure \ref{fig:Sp_dim_alpha} for different values of $\alpha$, keeping $\bar\lambda^2=\lambda^2=1$. As expected, in the low-energy ($s\rightarrow \infty$) limit the spectral dimension always takes the standard value of four. In the UV regime ($s\rightarrow 0$), for any $\alpha\neq 0$ (in which case the integration measure becomes trivial) the spectral dimension runs to lower values, that tend to zero as the value of $\alpha$ increases. The value of $\alpha$ characterizing the canonical Snyder model, $\alpha=\frac{5}{2}$, is the lowest value of $\alpha$ for which the spectral dimension goes to zero in the UV. We also plot the behavior of the spectral dimension in the standard Snyder model ($\bar\lambda^2=\gamma=0$), see Figure \ref{fig:Snyder_sp_dim}.
\begin{figure}[h]
\centering
\scalebox{0.55}{\includegraphics{SpecDim_alphamodzero.pdf}}
\caption{Spectral dimension deduced from the average return probability in Eq. \eqref{eq:Psspherical} as a function of $s$ and for different values of $\alpha$, with $\bar\lambda^2=\lambda^2=1$.}\label{fig:Sp_dim_alpha}
\end{figure}
\begin{figure}
\centering
\includegraphics[scale=0.55]{Snyder.pdf}
\caption{Running of the spectral dimension in Snyder model, as deduced from the average return probability in Eq. \eqref{eq:Psspherical} with $\bar\lambda^2=0$, $\lambda^2=1$ and $\alpha=\frac{5}{2}$.}\label{fig:Snyder_sp_dim}
\end{figure}
While the results just presented already signal that the spectral dimension is not to be trusted in characterizing the physical properties of quantum spacetime models, it is quite suggestive that we have to face even bigger difficulties when accounting for the possibility that the on-shell relation is given by some function of the Casimir, as in Eq. \eqref{eq:CasimirXi}. In fact, not all of the possible choices of Casimir that we have considered for the thermal dimension, Eq. \eqref{eq:CasimirXi}, admit a well-defined spectral dimension. This is not surprising, since the spectral dimension is defined on a Euclideanized version of the model, and in general the Euclidean and Lorentzian versions of a quantum gravity model can be profoundly different \cite{Carlip:2015mra, Carlip:2017eud}. Adopting this new on-shell condition, the Wick-rotated form of the D'Alembertian is
\begin{equation}
- (p_{0}^{2}+|\vec p|^{2}) -\ell^{2\xi} (-1)^{1+\xi}(p_{0}^{2}+|\vec p|^{2})^{1+\xi}\,,
\end{equation}
and it is immediate to see that for even values of $\xi$ the return probability diverges so that it is not possible to compute the spectral dimension.
If we restrict to odd values of $\xi$, the average return probability is
\begin{eqnarray}
P(s)\propto \int d r \left[1+(\bar\lambda^2+\lambda^2)r^2\right]^{-\alpha}r^3 e^{-s r^2 (1+\ell r)^{2\xi}} \,,\label{eq:PssphericalXi}
\end{eqnarray}
and one finds a running of the spectral dimension that has similar features as the ``standard Casimir case'' studied above. The role of $\xi$ is merely to further shift the UV value of the spectral dimension towards zero.
\section{Closing remarks}
Recent research in quantum gravity has devoted quite some effort to describing the dimension of quantum spacetime. We were here concerned with the identification of notions of dimensionality that are informative with respect to the physical content of the theory.
We argued that the Snyder model for noncommutative spacetime (and its generalizations) is an ideal test ground to evaluate the physical relevance of different proposals for describing dimensionality in the quantum gravity regime. In fact, the (generalized) Snyder model is quite peculiar in having a highly non-trivial geometry, such that spacetime coordinates have a discrete spectrum, but it behaves as standard Minkowski spacetime once the physical constraint is imposed.
As a first investigation along this line, we compared the behavior of the thermal and spectral dimension associated to (generalized) Snyder spacetime. Quite interestingly, we found confirmation of earlier claims that the former is a good indicator of the presence of genuinely nontrivial physical properties of quantum spacetime models, or lack thereof. In fact, we found that the thermal dimension does not present any running and is constantly equal to the standard value of four. On the other hand, analysis of the spectral dimension showed that its behavior is dominated by off-shell properties of the model, so that for the Snyder spacetime it runs from the value of four in the infrared to zero in the ultraviolet. This confirms previous concerns regarding the physical relevance of such notion of dimensionality.
\section*{Acknowledgements}
F.G. acknowledges support from the John Templeton Foundation, Project 60609, ``Quantum Causal Structures'', from the research platform ``Testing Quantum and Gravity Interface with Single Photons'' (TURIS), and the Austrian Science Fund (FWF) through the project I-2526-N27 and the doctoral program ``Complex Quantum Systems'' (CoQuS) under Project W1210-N25. This publication was made possible through the support of a grant from the John Templeton Foundation. The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the John Templeton Foundation.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 4,716
|
.class public Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;
.super Ljava/lang/Object;
.source "ProgramGenerator.java"
# instance fields
.field private TAG:Ljava/lang/String;
.field private TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
.field private mFragmentShader:Ljava/lang/String;
.field private mProgram:I
.field private mVertexShader:Ljava/lang/String;
.field private maPositionHandle:I
.field private maTextureHandle:I
.field private muAlphaHandle:I
.field private muBoundHandle1:I
.field private muBoundHandle2:I
.field private muDirectionHandle:I
.field private muMVPMatrixHandle:I
.field private muTurnAlphaHandle:I
.field private muTurnNumberHandle:I
# direct methods
.method public constructor <init>(Ljava/lang/String;Ljava/lang/String;Lcom/samsung/android/srib/virtualtourAPI/Transition;)V
.locals 3
const/4 v2, -0x1
invoke-direct {p0}, Ljava/lang/Object;-><init>()V
const-string/jumbo v0, "ProgramGenerator"
iput-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TAG:Ljava/lang/String;
iput-object p1, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mVertexShader:Ljava/lang/String;
iput-object p2, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mFragmentShader:Ljava/lang/String;
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mVertexShader:Ljava/lang/String;
iget-object v1, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mFragmentShader:Ljava/lang/String;
invoke-direct {p0, v0, v1}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->createProgram(Ljava/lang/String;Ljava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
iput-object p3, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
if-eqz v0, :cond_3
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "aPosition"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetAttribLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->maPositionHandle:I
const-string/jumbo v0, "glGetAttribLocation aPosition"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->maPositionHandle:I
if-eq v0, v2, :cond_4
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "aTextureCoord"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetAttribLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->maTextureHandle:I
const-string/jumbo v0, "glGetAttribLocation aTextureCoord"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->maTextureHandle:I
if-eq v0, v2, :cond_5
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uMVPMatrix"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muMVPMatrixHandle:I
const-string/jumbo v0, "glGetUniformLocation uMVPMatrix"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muMVPMatrixHandle:I
if-eq v0, v2, :cond_6
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->WALK:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_7
:cond_0
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->TURN:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_8
:cond_1
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->HELPER:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_b
:cond_2
return-void
:cond_3
return-void
:cond_4
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting aPosition"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_5
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting aTextureCoord"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_6
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uMVPMatrix"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_7
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "utext_alpha_out"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muAlphaHandle:I
const-string/jumbo v0, "glGetUniformLocation utext_alpha_out"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muAlphaHandle:I
if-ne v0, v2, :cond_0
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting utext_alpha_out"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_8
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uAlpha_value"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnAlphaHandle:I
const-string/jumbo v0, "glGetUniformLocation uAlpha_value"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnAlphaHandle:I
if-eq v0, v2, :cond_9
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uBound1"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle1:I
const-string/jumbo v0, "glGetUniformLocation uBound1"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle1:I
if-eq v0, v2, :cond_a
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uDirection"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muDirectionHandle:I
const-string/jumbo v0, "glGetUniformLocation udirection"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muDirectionHandle:I
if-ne v0, v2, :cond_1
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting udirection"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_9
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uAlpha_value"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_a
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uBound1"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_b
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uAlpha_value"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnAlphaHandle:I
const-string/jumbo v0, "glGetUniformLocation uAlpha_value"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnAlphaHandle:I
if-eq v0, v2, :cond_c
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uBound1"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle1:I
const-string/jumbo v0, "glGetUniformLocation uBound1"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle1:I
if-eq v0, v2, :cond_d
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uDirection"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muDirectionHandle:I
const-string/jumbo v0, "glGetUniformLocation udirection"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muDirectionHandle:I
if-eq v0, v2, :cond_e
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uBound2"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle2:I
const-string/jumbo v0, "glGetUniformLocation uBound2"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle2:I
if-eq v0, v2, :cond_f
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
const-string/jumbo v1, "uTurnNumber"
invoke-static {v0, v1}, Landroid/opengl/GLES20;->glGetUniformLocation(ILjava/lang/String;)I
move-result v0
iput v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnNumberHandle:I
const-string/jumbo v0, "glGetUniformLocation uTurnNumber"
invoke-direct {p0, v0}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnNumberHandle:I
if-ne v0, v2, :cond_2
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uTurnNumber"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_c
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uAlpha_value"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_d
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uBound1"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_e
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting udirection"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
:cond_f
new-instance v0, Ljava/lang/RuntimeException;
const-string/jumbo v1, "Error while getting uBound2"
invoke-direct {v0, v1}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v0
.end method
.method private checkGlError(Ljava/lang/String;)V
.locals 4
invoke-static {}, Landroid/opengl/GLES20;->glGetError()I
move-result v0
if-nez v0, :cond_0
return-void
:cond_0
iget-object v1, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TAG:Ljava/lang/String;
new-instance v2, Ljava/lang/StringBuilder;
invoke-direct {v2}, Ljava/lang/StringBuilder;-><init>()V
invoke-virtual {v2, p1}, Ljava/lang/StringBuilder;->append(Ljava/lang/String;)Ljava/lang/StringBuilder;
move-result-object v2
const-string/jumbo v3, ": glError "
invoke-virtual {v2, v3}, Ljava/lang/StringBuilder;->append(Ljava/lang/String;)Ljava/lang/StringBuilder;
move-result-object v2
invoke-virtual {v2, v0}, Ljava/lang/StringBuilder;->append(I)Ljava/lang/StringBuilder;
move-result-object v2
invoke-virtual {v2}, Ljava/lang/StringBuilder;->toString()Ljava/lang/String;
move-result-object v2
invoke-static {v1, v2}, Landroid/util/Log;->e(Ljava/lang/String;Ljava/lang/String;)I
new-instance v1, Ljava/lang/RuntimeException;
new-instance v2, Ljava/lang/StringBuilder;
invoke-direct {v2}, Ljava/lang/StringBuilder;-><init>()V
invoke-virtual {v2, p1}, Ljava/lang/StringBuilder;->append(Ljava/lang/String;)Ljava/lang/StringBuilder;
move-result-object v2
const-string/jumbo v3, ": glError "
invoke-virtual {v2, v3}, Ljava/lang/StringBuilder;->append(Ljava/lang/String;)Ljava/lang/StringBuilder;
move-result-object v2
invoke-virtual {v2, v0}, Ljava/lang/StringBuilder;->append(I)Ljava/lang/StringBuilder;
move-result-object v2
invoke-virtual {v2}, Ljava/lang/StringBuilder;->toString()Ljava/lang/String;
move-result-object v2
invoke-direct {v1, v2}, Ljava/lang/RuntimeException;-><init>(Ljava/lang/String;)V
throw v1
.end method
.method private createProgram(Ljava/lang/String;Ljava/lang/String;)I
.locals 7
const/4 v6, 0x1
const/4 v5, 0x0
const v4, 0x8b31
invoke-direct {p0, v4, p1}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->loadShader(ILjava/lang/String;)I
move-result v3
if-eqz v3, :cond_1
const v4, 0x8b30
invoke-direct {p0, v4, p2}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->loadShader(ILjava/lang/String;)I
move-result v1
if-eqz v1, :cond_2
invoke-static {}, Landroid/opengl/GLES20;->glCreateProgram()I
move-result v2
if-nez v2, :cond_3
:cond_0
:goto_0
return v2
:cond_1
return v5
:cond_2
return v5
:cond_3
invoke-static {v2, v3}, Landroid/opengl/GLES20;->glAttachShader(II)V
const-string/jumbo v4, "glAttachShader"
invoke-direct {p0, v4}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
invoke-static {v2, v1}, Landroid/opengl/GLES20;->glAttachShader(II)V
const-string/jumbo v4, "glAttachShader"
invoke-direct {p0, v4}, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->checkGlError(Ljava/lang/String;)V
invoke-static {v2}, Landroid/opengl/GLES20;->glLinkProgram(I)V
new-array v0, v6, [I
const v4, 0x8b82
invoke-static {v2, v4, v0, v5}, Landroid/opengl/GLES20;->glGetProgramiv(II[II)V
aget v4, v0, v5
if-eq v4, v6, :cond_0
iget-object v4, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TAG:Ljava/lang/String;
const-string/jumbo v5, "Could not link program: "
invoke-static {v4, v5}, Landroid/util/Log;->e(Ljava/lang/String;Ljava/lang/String;)I
iget-object v4, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TAG:Ljava/lang/String;
invoke-static {v2}, Landroid/opengl/GLES20;->glGetProgramInfoLog(I)Ljava/lang/String;
move-result-object v5
invoke-static {v4, v5}, Landroid/util/Log;->e(Ljava/lang/String;Ljava/lang/String;)I
invoke-static {v2}, Landroid/opengl/GLES20;->glDeleteProgram(I)V
const/4 v2, 0x0
goto :goto_0
.end method
.method private loadShader(ILjava/lang/String;)I
.locals 5
const/4 v3, 0x0
invoke-static {p1}, Landroid/opengl/GLES20;->glCreateShader(I)I
move-result v1
if-nez v1, :cond_1
:cond_0
:goto_0
return v1
:cond_1
invoke-static {v1, p2}, Landroid/opengl/GLES20;->glShaderSource(ILjava/lang/String;)V
invoke-static {v1}, Landroid/opengl/GLES20;->glCompileShader(I)V
const/4 v2, 0x1
new-array v0, v2, [I
const v2, 0x8b81
invoke-static {v1, v2, v0, v3}, Landroid/opengl/GLES20;->glGetShaderiv(II[II)V
aget v2, v0, v3
if-nez v2, :cond_0
iget-object v2, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TAG:Ljava/lang/String;
new-instance v3, Ljava/lang/StringBuilder;
invoke-direct {v3}, Ljava/lang/StringBuilder;-><init>()V
const-string/jumbo v4, "Could not compile shader "
invoke-virtual {v3, v4}, Ljava/lang/StringBuilder;->append(Ljava/lang/String;)Ljava/lang/StringBuilder;
move-result-object v3
invoke-virtual {v3, p1}, Ljava/lang/StringBuilder;->append(I)Ljava/lang/StringBuilder;
move-result-object v3
const-string/jumbo v4, ":"
invoke-virtual {v3, v4}, Ljava/lang/StringBuilder;->append(Ljava/lang/String;)Ljava/lang/StringBuilder;
move-result-object v3
invoke-virtual {v3}, Ljava/lang/StringBuilder;->toString()Ljava/lang/String;
move-result-object v3
invoke-static {v2, v3}, Landroid/util/Log;->e(Ljava/lang/String;Ljava/lang/String;)I
iget-object v2, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TAG:Ljava/lang/String;
invoke-static {v1}, Landroid/opengl/GLES20;->glGetShaderInfoLog(I)Ljava/lang/String;
move-result-object v3
invoke-static {v2, v3}, Landroid/util/Log;->e(Ljava/lang/String;Ljava/lang/String;)I
invoke-static {v1}, Landroid/opengl/GLES20;->glDeleteShader(I)V
const/4 v1, 0x0
goto :goto_0
.end method
# virtual methods
.method getAlphaHandle()I
.locals 2
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->WALK:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_0
const/4 v0, -0x1
return v0
:cond_0
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muAlphaHandle:I
return v0
.end method
.method getBoundHandle1()I
.locals 2
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->TURN:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-ne v0, v1, :cond_1
:cond_0
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle1:I
return v0
:cond_1
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->HELPER:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_0
const/4 v0, -0x1
return v0
.end method
.method getBoundHandle2()I
.locals 2
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->HELPER:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_0
const/4 v0, -0x1
return v0
:cond_0
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muBoundHandle2:I
return v0
.end method
.method getMVPHandle()I
.locals 1
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muMVPMatrixHandle:I
return v0
.end method
.method getPositionHandle()I
.locals 1
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->maPositionHandle:I
return v0
.end method
.method getProgram()I
.locals 1
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->mProgram:I
return v0
.end method
.method getTextureHandle()I
.locals 1
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->maTextureHandle:I
return v0
.end method
.method getTurnAlphaHandle()I
.locals 2
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->TURN:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-ne v0, v1, :cond_1
:cond_0
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnAlphaHandle:I
return v0
:cond_1
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->HELPER:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_0
const/4 v0, -0x1
return v0
.end method
.method getTurnDirectionHandle()I
.locals 2
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->TURN:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-ne v0, v1, :cond_1
:cond_0
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muDirectionHandle:I
return v0
:cond_1
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->HELPER:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_0
const/4 v0, -0x1
return v0
.end method
.method getTurnNumberHandle()I
.locals 2
iget-object v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->TRANSITION:Lcom/samsung/android/srib/virtualtourAPI/Transition;
sget-object v1, Lcom/samsung/android/srib/virtualtourAPI/Transition;->HELPER:Lcom/samsung/android/srib/virtualtourAPI/Transition;
if-eq v0, v1, :cond_0
const/4 v0, -0x1
return v0
:cond_0
iget v0, p0, Lcom/samsung/android/srib/virtualtourRenderer/ProgramGenerator;->muTurnNumberHandle:I
return v0
.end method
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 7,847
|
MacArthur Bible Studies
Ezra & Nehemiah: Israel Returns from Exile
© 2016 by John MacArthur
All rights reserved. No portion of this book may be reproduced, stored in a retrieval system, or transmitted in any form or by any means—electronic, mechanical, photocopy, recording, scanning, or other—except for brief quotations in critical reviews or articles, without the prior written permission of the publisher.
Published in Nashville, Tennessee, by Nelson Books, an imprint of Thomas Nelson. Nelson Books and Thomas Nelson are registered trademarks of HarperCollins Christian Publishing, Inc.
Originally published in association with the literary agency of Wolgemuth & Associates, Inc. Original layout, design, and writing assistance by Gregory C. Benoit Publishing, Old Mystic, CT.
"Unleashing God's Truth, One Verse at a Time®" is a trademark of Grace to You. All rights reserved.
Thomas Nelson titles may be purchased in bulk for educational, business, fund-raising, or sales promotional use. For information, please e-mail SpecialMarkets@ThomasNelson.com.
Scripture quotations are taken from the _New King James Version_. Copyright © 1982 by Thomas Nelson, Inc. Used by permission. All rights reserved.
Some material in the Introduction, "Keys to the Text," and "Exploring the Meaning" sections are taken from _The MacArthur Bible Commentary,_ John MacArthur, copyright © 2005 Thomas Nelson Publishers.
ISBN 978-0-7180-3479-5
ISBN 978-0-7180-3492-4 (eBook)
First Printing April 2016
# EBOOK INSTRUCTIONS
In this ebook edition, please use your device's note-taking function to record your thoughts wherever you see the bracketed instructions [Your Notes] or [Your Response Here]. Use your device's highlighting function to record your response whenever you are asked to checkmark, circle, underline, or otherwise indicate your answer(s).
# CONTENTS
_Ebook Instructions_
_Introduction_
1. Return from Exile
_Ezra 1:1–11_
2. The Work Begins
_Ezra 2:1–3:13_
3. Trouble with Samaritans
_Ezra 4:1–5:17_
4. Choosing God's Priorities
_Haggai 1:1–2:9_
5. Obedience over Observance
_Zechariah 7:1–8:23_
6. The Second Wave of Exiles
_Ezra 7:1–8:36_
7. Unequally Yoked
_Ezra 9:1–10:44_
8. Rebuilding the Wall
_Nehemiah 1:1–2:20_
9. The Opposition Increases
_Nehemiah 3:1–4:23_
10. Leading by Example
_Nehemiah 5:1–6:19_
11. Worshiping God
_Nehemiah 9:1–38_
12. Reviewing Key Principles
# INTRODUCTION
The Lord God had warned the nation of Israel that He would send them into foreign captivity if they persisted in idolatry. Yet the people had ignored Him—and the prophets He sent to warn them—and had run headlong into disobedience. True to His word, God used the Assyrians and the Babylonians to evict Israel from its land, destroy the city of Jerusalem, and plunder the temple. With the people in captivity, it appeared Israel had no hope and no future.
However, God had also promised the captivity of Israel would last for only seventy years. At the end of that time, He raised up men such as Zerubbabel, Ezra, and Nehemiah to lead His people back to Jerusalem and begin rebuilding the city. He then sent several prophets, including Haggai and Zechariah, to encourage His people to continue the rebuilding efforts. Under this godly leadership, the people of Israel got a new start. Unfortunately—and predictably—they failed again and went back to their former idolatrous ways.
In these twelve studies, we will examine the events that led to the people being allowed to return to their homeland and what took place during the first two waves of their return. We will see how God raised up leaders among the exiles to not only champion the rebuilding efforts but also to protect the people from outside attack and turn them back to the worship of the Lord. We will discover what it means to put God first and see that apart from the new covenant and the Messiah, none of us would be able to be righteous before the Lord.
Through it all, we will learn some precious truths about the character of God, and we will see His great faithfulness in keeping His promises. We will learn, in short, what it means to follow Him wholeheartedly, walk by faith, and remain committed to His Word.
## THE BOOK OF EZRA
Even though Ezra's name does not enter the account of Judah's post-exilic return to Jerusalem until 7:1, the book bears his name (which means "Jehovah helps") as a title. This is because both Jewish and Christian traditions attribute authorship to this famous scribe-priest. New Testament writers do not quote from the book of Ezra.
## AUTHOR AND DATE
Ezra was most likely the author of both Ezra and Nehemiah, which might have originally been one book. After his arrival in Jerusalem in 458 BC, he changed from writing in the third person (Ezra 1–6) to writing in the first person (Ezra 7–10). In the earlier section, it is likely that he used the third person because he was quoting from his memoirs.
As a scribe, Ezra had access to the myriad of administrative documents that are found in the books of Ezra and Nehemiah. Few people were allowed access to the royal archives of the Persian Empire, but Ezra proved to be the exception. He was a strong and godly man who lived at the time of Nehemiah, and tradition says he was founder of the Great Synagogue, where the complete Old Testament canon was first formally recognized.
## BACKGROUND AND SETTING
In 605–586 BC, God chastened His people because of their continued unfaithfulness to His covenant and allowed the Babylonians to sack and nearly depopulate Jerusalem. However, God promised that after seventy years of captivity the people would be allowed to return. The fulfillment of this prophecy began in 539 BC when Cyrus the Great, king of Persia, overthrew the Babylonians. Ezra begins a decree from Cyrus one year later that allowed the Jews to return to Jerusalem. The book chronicles the reestablishment of Judah's national calendar of feasts and sacrifices, including the rebuilding of the second temple.
Just as there had been three waves of deportation from Israel into Babylon (605 BC, 597 BC, and 586 BC), so there would be three returns to Jerusalem during a nine-decade span. Zerubbabel led the first return in 538 BC. Ezra led the second return in 458 BC. Nehemiah led the third return thirteen years later in 445 BC. However, the Jewish people's uncontested political autonomy never returned. The prophets Haggai and Zechariah preached during Zerubbabel's time, beginning about 520 BC.
## HISTORICAL AND THEOLOGICAL THEMES
The Jews' return from the Babylonian captivity seemed like a second Exodus, sovereignly patterned in some ways after Israel's first redemption from Egyptian bondage. The return trip from Babylon involved activities similar to those of the original Exodus: (1) the rebuilding of the temple and the city walls; (2) the reinstitution of the law, which made Zerubbabel, Ezra, and Nehemiah collectively seem like a second Moses; (3) the challenge of the local enemies; and (4) the people's temptation to intermarry with non-Jews, resulting in idolatry.
The primary message of Ezra is that God orchestrated the past grim situation (captivity) and would continue to work through a pagan king and his successors to give Judah hope for the future (return). Ezra's message was that God's administration overrode that of any kings of this world, and He would continue to extend His covenant grace toward Israel. Another prominent theme is opposition from the local Samaritan residents, whose ancestors had been imported from Assyria. After being shunned from participating in the rebuilding of the temple, these enemies hired counselors to conspire against the Jews, but the Lord (through the preaching of Haggai and Zechariah) rekindled the spirit of the people and their leaders.
## INTERPRETIVE CHALLENGES
There are a number of interpretive challenges in the book of Ezra. First, how do the post-exilic historical books of 1 and 2 Chronicles, Ezra, Nehemiah, and Esther relate to the post-exilic prophetic books of Haggai, Zechariah, and Malachi? The two books of Chronicles serve as a reminder of the promised Davidic kingship, the Aaronic priesthood, and appropriate temple worship. Haggai and Zechariah prophesied during the period of Ezra 4–6, when temple construction was resumed. Malachi wrote during Nehemiah's revisit to Persia.
Second, what purpose does the book serve? Historically, Ezra reported the first two of three post-exilic returns to Jerusalem from the Babylonian captivity (Zerubbabel's in 538 BC and Ezra's in 458 BC). Spiritually, Ezra reestablished the importance of the Aaronic priesthood, encouraged the rebuilding of the second temple, and dealt with the people's gross sin of intermarriage with foreigners. Most importantly, he described how the sovereign hand of God moved kings and overcame opposition to reestablish Israel in the land.
Third, how does the command for the Jews to divorce their foreign spouses correlate with the fact that God hates divorce (see Malachi 2:16)? Ezra does not establish the norm but deals with a special case in history. It seems to have been decided that the lesser wrong of divorce would be preferable to the greater wrong of the Jewish race being polluted by intermarriage so that the messianic line of David would not be ended by being mingled with Gentiles. To solve the problem this way magnifies the mercy of God, in that the only other solution would have been to kill all those involved.
## THE BOOK OF NEHEMIAH
Nehemiah was a famous cupbearer who never appears in Scripture outside of this book. As with the book of Ezra, the book recounts selected events of his leadership and was thus titled after him. Both the Greek Septuagint (LXX) and the Latin Vulgate named this book Second Ezra. Even though the two books are separate in most English Bibles, they may have once been joined together in a single unit (as currently in the Hebrew texts).
## AUTHOR AND DATE
Though much of this book was clearly drawn from Nehemiah's personal diaries, both Jewish and Christian traditions recognize Ezra as the author. This is based on the evidence that Ezra and Nehemiah were originally one book and on internal evidence, such as the recurrent "hand of the Lord" theme, that dominates both books.
The events depicted in the opening chapter of Nehemiah began late in the year 446 BC, the twentieth year of Persian King Artaxerxes' reign. The book follows chronologically from Nehemiah's first term as governor of Jerusalem (c. 445–433 BC) to his second term, which possibly began in 424 BC. Ezra would have written Nehemiah sometime during or after Nehemiah's second term, but no later than 400 BC.
## BACKGROUND AND SETTING
The book of Nehemiah chronicles the third return of the exiles to rebuild the wall around Jerusalem. At that time, the Persian Empire's administration of Judah, though done with a loose hand, was mindful of disruptions or any signs of rebellion from its vassals. Rebuilding the walls of conquered cities posed the most glaring threat to the Persian central administration, and only a confidant of the king could be trusted for such an operation. At the most critical juncture in Judah's revitalization, God raised up Nehemiah to fulfill this role for His people.
Several historical notes are of interest. First, Esther was Artaxerxes' stepmother, and thus could have influenced him to look favorably on the Jews. Second, Daniel's prophetic "seventy weeks" began with Artaxerxes' decree in 445 BC. Third, Egyptian documents dated to the late fifth century BC support the account of Nehemiah by mentioning Sanballat being governor of Samaria and Nehemiah being replaced as governor of Jerusalem by Bigvai. Finally, Nehemiah and Malachi represent the last of the Old Testament canonical writings, both in terms of the time the events occurred and when they were recorded.
## HISTORICAL AND THEOLOGICAL THEMES
A constant theme of Nehemiah is the people's need to be diligent in reading God's Word in order to perform His will. The spiritual revival came in response to Ezra's reading of "the Book of the Law of Moses" (8:1), after which Ezra carefully explained its meaning to the people. So deep was the concern to abide by God's revealed will that the leaders took "a curse and an oath to walk in God's Law" (10:29). When the marriage reforms were carried out, they acted in accordance with that which "they read from the Book of Moses" (13:1).
A second theme involves the methods God used to bring His people back into obedience. He worked through Nehemiah's obedience, but also worked through the wicked hearts of His people's enemies. Those enemies failed not so much as a result of the success of Nehemiah's strategies but because God brought their plots to nothing. God used the opposition of Judah's enemies to drive His people to their knees in the same way that He used the favor of Cyrus to return His people to the land.
A third theme is the people's opposition. Judah's enemies started rumors that God's people had revolted against Persia, with the goal of intimidating the people into stopping reconstruction of the Jerusalem walls. However, in spite of opposition from without and heartbreaking corruption and dissension from within, the people completed the walls in only fifty-two days, experienced revival after the reading of the law by Ezra, and celebrated the Feast of Tabernacles (c. 445 BC).
## INTERPRETIVE CHALLENGES
There are a number of interpretive challenges in the book of Nehemiah. First, because much of Nehemiah is explained in relationship to Jerusalem's gates (see Nehemiah 2; 3; 8; 12), it is important to have an understanding of how those gates were laid out in the city walls (see the map "Jerusalem in Nehemiah's Time" for an orientation).
Second, it is important to note that the timeline of Nehemiah 1–12 encompassed about one year (445 BC), followed by a long gap of time (more than twenty years) before the events continue again in Nehemiah 13. (See the chart "Timetable of Events.")
Finally, it must be recognized that Nehemiah actually served two governorships in Jerusalem, the first from 445–433 BC, and the second beginning possibly in 424 BC and extending to no longer than 410 BC. During his absence the people returned to their former ways, led by the high priest Eliashib, causing Nehemiah to return and institute some needed reforms. It was during Nehemiah's absence that Malachi wrote his prophetic book indicting both priests and people for their sinful defection. Nehemiah 13 was the last portion of the Old Testament to be written.
## TIMETABLE OF EVENTS
**Event** | **Approximate Date ( BC)**
---|---
Judah carried to Babylon | 586
Zerubbabel returns with first wave | 538
Haggai and Zechariah begin to prophesy | 520
Temple rebuilt | 515
Reign of Ahasuerus | 486–465
Ezra returns with second wave | 458
Nehemiah returns with third wave | March–April 445
Nehemiah starts the wall | July–August 445
Nehemiah completes the wall | August–September 445
The wall is dedicated | September–October 445
Nehemiah serves his first term as governor | 445–433
Nehemiah returns to Persia | 433–424
Nehemiah serves his second term as governor | 424–410
# **1**
# RETURN FROM EXILE
# _Ezra 1:1–11_
## DRAWING NEAR
What are some common distractions today that can keep people from following God's will for their lives? How does a person overcome these distractions?
[Your Response Here]
## THE CONTEXT
In 722 BC, God allowed the powerful Assyrian Empire to invade the kingdom of Israel, conquer all its territory, and carry its people away into slavery. The author of 2 Kings sums up the sad reason for these events: "The children of Israel had sinned against the LORD their God. . . . Therefore the LORD was very angry with Israel, and removed them from His sight" (17:7–18). A little more than a century later, in 605 BC, God allowed the Babylonians to invade Judah for the same reason and begin to take its people into captivity. About twenty years after this time, Nebuchadnezzar carried the remaining Jews into Babylonian captivity.
From an outside perspective, it appeared that God's people would be forever lost to history. Yet these events did not take God by surprise—nor did they occur without warning. Furthermore, in spite of the people's sins, Jeremiah prophesied the Jews would remain in captivity for only _seventy_ years. God said, "This whole land shall be a desolation and an astonishment, and these nations shall serve the king of Babylon seventy years. Then it will come to pass, when seventy years are completed, that I will punish the king of Babylon and that nation, the land of the Chaldeans, for their iniquity" (Jeremiah 25:11–12).
This study opens at the end of this seventy-year period, when God promised that His chosen people would begin to return to Jerusalem to rebuild His chosen city. However, by this time the mighty empire of Babylon had collapsed and been absorbed by another powerful empire: Persia, under the reign of Cyrus "the Great." The year was 538 BC.
## KEYS TO THE TEXT
The author of Chronicles (who most likely was Ezra) provides some background information on why the Jewish people were exiled. Read his words in 2 Chronicles 36:11–21, noting the key words and phrases indicated below.
> _T HE LAST KING OF JUDAH: Our studies open with a look back at the Babylonian invasion of Judah—and the events leading up to it._
36:11. ZEDEKIAH WAS TWENTY-ONE: Zedekiah became king in Judah around 597 BC. He was the last king of Judah and ended a succession of wicked leaders. The nation of Israel had been divided into two kingdoms (Israel and Judah) after the reign of Solomon hundreds of years earlier. As mentioned in the introduction to this study, Israel had long since gone into captivity.
12. HE DID EVIL IN THE SIGHT OF THE LORD: Many of the kings of Judah—and most of the kings of Israel—had led God's people away from His Word into idolatry. The Lord had warned the Jews repeatedly that He would send them into captivity if they worshiped false gods.
JEREMIAH THE PROPHET: Jeremiah wrote the book of Lamentations to mourn the destruction of the temple, which occurred when Zedekiah was king (586 BC).
13. KING NEBUCHADNEZZAR: Nebuchadnezzar was the king of Babylon at this time. Under his leadership, the empire grew to its greatest extent.
14. ALL THE ABOMINATIONS OF THE NATIONS: These abominations included idolatry, child sacrifice, sexual perversion, denial of God as the Creator, and many more sins. All these pagan practices are abundant in modern Western society.
DEFILED THE HOUSE OF THE LORD: God's people had defiled His temple by incorporating pagan practices into their worship. This resulted in the Lord's swift judgment.
> _G OD'S FAITHFULNESS: The people of Judah have been constantly unfaithful to the Lord, but God has proven completely faithful to them—even at cost to Himself._
15. RISING UP EARLY AND SENDING THEM: The Lord was constantly faithful to His people in spite of their unfaithfulness to Him. The figure of speech here suggests that He went out of His way to lead them back to Him, going to great lengths—and at immense cost—by sending prophets again and again to urge Israel and Judah to obey His Word.
BECAUSE HE HAD COMPASSION ON HIS PEOPLE: Here we learn why God went to such lengths to turn His people back to Him: because He loved them! We will see more of God's compassion in the events leading up to the people's return to Jerusalem and find that His hand of discipline was always tempered with grace. Yet His compassion is most clearly expressed in the life and person of Jesus Christ. God sent His only Son expressly to die for our sins—and there can be no greater expression of love than that.
16. MOCKED . . . DESPISED . . . SCOFFED AT: The world has always treated the things of God with contempt, and still does today, but God's own people committed this sin. It was this contempt that ultimately brought His discipline on the nation of Judah.
17. THE KING OF THE CHALDEANS: This king is Nebuchadnezzar. Chaldea had been absorbed into the Babylonian Empire, yet the empire was interchangeably referred to by both names.
19. BURNED THE HOUSE OF GOD, BROKE DOWN THE WALL OF JERUSALEM: In 597 BC, King Nebuchadnezzar carried 10,000 people into captivity from Judah, including the prophet Ezekiel. (He had actually carried captives away from Judah in several waves, beginning in 605 BC.) Zerubbabel and Ezra would later lead people back to Jerusalem to begin rebuilding the temple, and Nehemiah would lead the work of rebuilding the city walls.
20. UNTIL THE RULE OF THE KINGDOM OF PERSIA: Cyrus the Great conquered Babylon in 539 BC. He allowed Jews to begin returning to Jerusalem the following year.
21. UNTIL THE LAND HAD ENJOYED HER SABBATHS: The Lord had commanded His people to allow their land to lie fallow every seven years. On this seventh year, they were to neither plant crops nor reap (see Leviticus 25:4–5). Evidently, they had failed to obey this command beginning around the time when Eli was high priest (c. 1107–1067 BC). The Lord had warned the Jews that He would enforce the Sabbath rest on His Promised Land if they failed to keep it (see Leviticus 26:27–46). Jeremiah later prophesied that the people would remain in captivity for seventy years—one year for every Sabbath they neglected (see Jeremiah 25:1–11). The exact number of Sabbath years was 490 years, the period from Saul to the Babylonian captivity.
Read Ezra 1:1–11, noting the key words and phrases indicated below.
> _T HE PROCLAMATION: Seventy years later, the Lord raises up a Gentile leader—Cyrus, the king of the Persian Empire—who will begin to send the Jews back home._
1:1. IN THE FIRST YEAR OF CYRUS: The Lord had prophesied through Isaiah, saying of Cyrus, "He is My shepherd . . . saying to Jerusalem, 'You shall be built,' and to the temple, 'Your foundation shall be laid' " (44:28). These events occurred c. 538 BC.
BY THE MOUTH OF JEREMIAH: As previously mentioned, Jeremiah had prophesied the return of the exiles after a seventy-year captivity in Babylon. This period of exile likely began during the fourth year of the reign of King Jehoiakim of Judah, c. 605 BC, when Jerusalem was captured and the temple treasures were taken. It ended with the decree of Cyrus to allow the Jews to return, thus spanning a period from 605/604 BC to 536/535 BC.
THE LORD STIRRED UP THE SPIRIT OF CYRUS: Throughout these studies, we will see how God used the deeds of men—both good and evil—to accomplish His sovereign purposes. Even when the circumstances seemed dark and hopeless, He was still in control and was still working out His promises for His people.
MADE A PROCLAMATION: This was the most common form of public communication and usually came from the central administration. The king would dispatch a herald, perhaps carrying a written document, into a particular city. The messenger would address the people by either going to the city gate (where people often congregated for social discourse) or gathering the crowds together in a square, occasionally by the blowing of a horn. The herald would then read the proclamation. In 1879, archaeologists recovered one such document called the Cyrus Cylinder, which was evidently some sort of general policy from Cyrus that commissioned people from many lands to return to their cities to rebuild the temples to their gods. Whether or not this document was an extension of the proclamation made to the Jewish exiles in this passage must remain a matter of speculation.
2. THE LORD GOD OF HEAVEN HAS GIVEN ME: Cyrus evidently recognized the sovereign hand of God in his life. He acknowledged that he himself held power in Persia only because the Lord had given it to him—even though he probably did not worship Yahweh as the only true God. Josephus, a Jewish historian, would later write that Daniel was Cyrus's prime minister and that he read prophecies to the king that mentioned Cyrus by name—more than a century before Cyrus was born (see Isaiah 44:28). According to Josephus, this led Cyrus to make his decree allowing the Jews to return to Jerusalem.
TO BUILD HIM A HOUSE: This refers to the second temple, which the Jews would begin to build after they returned to Jerusalem in the first wave led by Zerubbabel. During the first century AD, King Herod would commission a massive reconstruction effort in which he greatly expanded the temple mount. This second temple would stand until 70 AD, when, after the people rebelled, the Roman general Titus seized the city and destroyed the structure.
4. LET THE MEN OF HIS PLACE HELP HIM: Cyrus commanded the neighbors of the returning Jews to assist them with finances and goods for their trip and for the work of rebuilding the temple. This was reminiscent of the Israelites' preparations for the exodus from Egypt, when the Lord had them ask their neighbors for gold, silver, and clothing (see Exodus 12:35–36).
> _T HE EXILES RETURN: In response to Cyrus's proclamation, many of the Jewish exiles decide to return to Jerusalem. It will not turn out to be a simple endeavor._
5. ALL WHOSE SPIRITS GOD HAD MOVED: The Hebrew word here literally means "to rouse up or awaken," and it is the same expression used of Cyrus in verse 1. The Lord stirred the hearts of His people and made many of them restless and unsettled with the knowledge that His temple was lying in ruins back in Judah. The Jews were not enslaved in Persia but were permitted to live as all other Persians lived, and many of them had risen to prosperity and influence. It is possible, therefore, that God's people had become comfortable and complacent with their lot, so the Lord stirred their hearts to be grieved over the desecration of His holy temple and His chosen city. The important thing to note, however, is that God motivated and directed this work of rebuilding, not men or any one charismatic leader. "Unless the LORD builds the house, they labor in vain who build it" (Psalm 127:1).
6. ALL THAT WAS WILLINGLY OFFERED: Throughout the books of Ezra and Nehemiah we find many parallels to the story of the Exodus. One can hear faint echoes of the Egyptians supplying treasures in order to provide splendor for the tabernacle (see Exodus 11:2). However, the Jews there had been slaves, and the Egyptians had despised them, so the Israelites "plundered the Egyptians" (Exodus 12:36). But in this case there is a hint of goodwill involved, which suggests the Jews' fellow countrymen were glad to assist them in rebuilding the Lord's house. Other nations around Israel were called to contribute as well. They were also assisted by some of their captive countrymen who had been born in Babylon and chose to remain, and perhaps by some Babylonians and Assyrians who were favorably disposed to Cyrus or the Jews.
7. THE ARTICLES OF THE HOUSE OF THE LORD: These were the vessels that Nebuchadnezzar had removed when he sacked the temple (see 2 Kings 24:13) God had preserved them with the Babylonians for the Jewish people's return as prophesied by Jeremiah (see 27:22).
8. SHESHBAZZAR THE PRINCE OF JUDAH: Nothing else is said about this man in the Bible, but most likely he was a political appointee of Cyrus to oversee Judah. He is not to be confused with Zerubbabel, who was the leader recognized by the Jews and by the Lord. While Zerubbabel did not serve as king, he was in the Davidic line of the Messiah (see Matthew 1:12).
9. THIS IS THE NUMBER OF THEM: The 2,499 articles counted in verses 9 and 10 are representative of the total of 5,400 mentioned in verse 11.
11. THE CAPTIVES WHO WERE BROUGHT: These were the individuals King Nebuchadnezzar had taken into Babylonian captivity from Jerusalem, and their return probably occurred early during the reign of Cyrus (c. 538/537 BC). The journey from Babylon to Jerusalem would have taken three to five months.
## UNLEASHING THE TEXT
1) Why did God send the Jews into captivity? Why did He limit the time of their captivity to seventy years?
[Your Response Here]
2) How had God's people drifted away from Him into idolatry? What contributed to their sin? What did God do to turn them back to Himself?
[Your Response Here]
3) In what ways did God show compassion to the Jews over the centuries? In what ways was He showing compassion by sending them into captivity?
[Your Response Here]
4) What does Ezra mean when he says the Lord "stirred up" the hearts of the Jews? Why was this needed? What was God trying to accomplish?
[Your Response Here]
## EXPLORING THE MEANING
**_God keeps all His promises._** God had promised His people, "If you diligently obey the voice of the LORD your God, to observe carefully all His commandments which I command you today, that the LORD your God will set you high above all nations of the earth" (Deuteronomy 28:1). Under the reign of David, the nation of Israel obeyed the Lord's commands, and God kept that promise, subduing all the nation's enemies and giving Israel peace and prosperity. But God had made another promise in that same passage: "But it shall come to pass, if you do not obey the voice of the LORD your God . . . you shall beget sons and daughters, but they shall not be yours; for they shall go into captivity" (verses 15, 41).
Beginning with the reign of King Solomon and extending until the time of the exile, the nation's kings gradually led God's people away from obedience and into idolatry. As a result, God proved faithful to His second promise and sent them into captivity. However, when this time came, the Lord then promised His people that they would return to Jerusalem in seventy years—and He kept that promise as well. He first raised up Nebuchadnezzar to carry the nation into Babylon, and then He raised up Cyrus to absorb Babylon into Persia and set His people free. He holds sovereign control over all human affairs, and He can raise up an empire or throw it down as He sees fit. But whatever happens, He always keeps His promises.
God has not changed since the time of Ezra. He has given many promises to Christians, both of blessing and of discipline, and He still keeps those promises. One of the most important promises of all is this: "For God so loved the world that He gave His only begotten Son, that whoever believes in Him should not perish but have everlasting life. For God did not send His Son into the world to condemn the world, but that the world through Him might be saved. He who believes in Him is not condemned; but he who does not believe is condemned already, because he has not believed in the name of the only begotten Son of God" (John 3:16–18). God will keep this promise of eternal life to anyone who accepts Jesus, but He will also keep His promise of eternal judgment for anyone who rejects Christ. If you do not know Jesus as your Savior, claim God's promise of salvation right now—because God keeps _all_ His promises.
**_The Lord sends discipline to help us, not to harm us._** The Lord sent the Babylonian army into Judah to carry His people into captivity when they turned away from Him and pursued nonexistent pagan gods. But two factors in this are important for us to remember. First, the people had been unfaithful to God for centuries (with brief periods of revival during that time), which demonstrates God's great patience and grace. Second, the Lord sent the Jews into captivity to discipline them and turn them back to Himself, _not_ to destroy them. This was a loving but stern Father's hand of correction, not a harsh judge's verdict of condemnation.
The author of Chronicles made it clear that the Lord had patiently endured hundreds of years of disobedience in His people, "rising up early and sending" His prophets to remind them of His Word again and again, "because He had compassion on His people and on His dwelling place" (2 Chronicles 36:15). Even in the midst of discipline, the Lord demonstrated His compassion and love by raising up a new king who would send His people back to Jerusalem and by restricting the captivity to a relatively short period of time. Jeremiah prophesied concerning this time: "For thus says the LORD: After seventy years are completed at Babylon, I will visit you and perform My good word toward you, and cause you to return to this place. For I know the thoughts that I think toward you, says the LORD, thoughts of peace and not of evil, to give you a future and a hope" (29:10–11).
The whole reason the Lord sent His people into captivity was to turn their hearts back to Him. "Then you will call upon Me and go and pray to Me," He said, "and I will listen to you. And you will seek Me and find Me, when you search for Me with all your heart" (verses 12–13). His ways have not changed today. The Lord may send discipline into our lives, but He does so in order to draw us toward Himself—not to push us away. When hardship enters our lives, we seek His face, and He has promised that we will find Him.
**_This world is not our home._** The people of Judah had been carried away to captivity, but they had not been made slaves as they were many centuries before in Egypt. Instead, they were allowed to establish relatively normal lives within the new land, and many of the Jews had risen to levels of power and prosperity. Daniel, for example, served at least three different kings as a close personal counselor. Yet this relative freedom brought a danger the Israelites had not faced when they were slaves: _complacency_. Many of God's people had become quite comfortable in captivity and were fitting in to the society and doing well.
The problem was that God did not intend for His people to make their permanent home outside of Judah. Their real home was there, and the Lord did not want them to put down roots anywhere else. God's temple in Jerusalem was in ruins, the city's walls lay in rubble, and the Lord grieved over that situation. He wanted His people to share His priorities and to long to return to their proper land where they would worship and serve Him as He had ordained. The world in which they had grown so content was _not_ their home.
This is equally true for Christians today. This world is not our home! It is not wrong to pursue a career or to establish a home, but the Lord does not want us to lose our eternal focus. He wants us to remember that the things of eternity are what matter most, not the things of this world. Paul wrote, "Set your mind on things above, not on things on the earth. For you died, and your life is hidden with Christ in God" (Colossians 3:2–3). Paul was reminding us that by being born again into the salvation of Christ, we have died to the things of this world. And if we are dead to this world, there is no purpose in trying to make our home here. Our existence is with Christ in eternity, and that is where our focus needs to remain.
## REFLECTING ON THE TEXT
5) What promises did God keep in these passages? What part did the behavior and attitudes of the Jews play in God's promises?
[Your Response Here]
6) Why did God choose to use Nebuchadnezzar and Cyrus to fulfill His plans? What does this suggest about His sovereignty?
[Your Response Here]
7) Why did God need to "rouse" the hearts of His people to return to Jerusalem? What worldly elements tend to lull Christians into complacency?
[Your Response Here]
8) When has the Lord sent discipline into your life? What was He trying to accomplish? How did you respond to that discipline?
[Your Response Here]
## PERSONAL RESPONSE
9) Have you accepted Jesus as your Lord and Savior and know that you will see the eternal home He has prepared for you? If not, what is preventing you from doing so right now?
[Your Response Here]
10) What might the Lord be "stirring up" your heart to do at present? What things of this world might be distracting you from sharing His priorities?
[Your Response Here]
# **2**
# THE WORK BEGINS
# _Ezra 2:1–3:13_
## DRAWING NEAR
What are some of the things that get in the way of people setting the right priorities? In what ways are Christians often guilty of not setting the priorities that God wants them to set?
[Your Response Here]
## THE CONTEXT
During the exodus from Egypt, the Lord demonstrated His presence with His people in a variety of tangible ways. He showed that He was with them by providing heavy cloud cover during the day to shield the people from the desert sun. By night, He provided a dramatic pillar of fire to illuminate the darkness. Beyond these manifestations, the Lord also commanded the people to construct a portable temple, which they were to set up whenever they made camp. At the center of this tabernacle was the ark of the covenant.
The ark represented God's chosen place of meeting with His people—a physical manifestation of His presence wherever they went. Many generations later, King Solomon built a magnificent permanent temple in Jerusalem to house the ark, and this became the Lord's chosen place for His people to assemble for corporate worship. After the nation split, Jeroboam set up alternate sites of worship in the north—where he set up calves made of gold—but the true site of the Lord's worship remained in Jerusalem in the south. The first temple existed until King Nebuchadnezzar conquered the land and carried the Jews into captivity. At that time he destroyed that temple and absconded with the ark.
In this study, the Jews have returned to Jerusalem from exile, and it is time for them to begin work on rebuilding the temple. They gather together as one—but then, instead of starting the work on the actual temple, they rebuild the altar that will go _inside_ the temple. At first glance this might seem like building one's living room furniture before building the house, but in reality it demonstrates a correct order of priorities in God's prescribed worship. We will discover that even without the ark of the covenant, God was still with His people—because it is people who comprise God's ultimate temple, not a building made of bricks and mortar.
## KEYS TO THE TEXT
Read Ezra 2:1–3:13, noting the key words and phrases indicated below.
> _C OUNTING THE CAPTIVES: Ezra's chronicle of the Jewish exiles' return under Zerubbabel begins with a comprehensive listing of the people who made the journey._
2:1. WHO CAME BACK FROM THE CAPTIVITY: This list is given almost identically in Nehemiah 7:6–73. The "province" refers to Judah, which by this time had been reduced from an illustrious, independent, and powerful kingdom to an obscure servile province of the Persian Empire. The returning Jews were still considered subjects of Cyrus living in a Persian province.
2. THOSE WHO CAME WITH ZERUBBABEL: As previously mentioned, this man was the rightful leader of Judah in that he was of the lineage of David through Jehoiachin (see 1 Chronicles 3:17–19). His name means "offspring of Babylon," which indicated his place of birth. Zerubbabel, rather than Cyrus's political appointee Sheshbazzar, led Judah according to God's will.
JESHUA: The high priest of the first return, whose name means "Jehovah saves," was called _Joshua_ in Haggai 1:1 and Zechariah 3:1. His father, Jozadak, had been among those taken into exile. He came from the lineage of Levi, Aaron, Eleazar, and Phinehas, and thus was legitimately in the line of the high priest (see Numbers 25:10–13).
NEHEMIAH . . . MORDECAI: These were not the same men listed in Nehemiah or Esther.
3. THE PEOPLE OF PAROSH: Ezra lists various Jewish families in verses 3 to 20 who were part of what would be known as the "first wave" of Jewish exiles returning to Jerusalem.
21. THE PEOPLE OF BETHLEHEM: Ezra lists returning exiles from various Judean cities in verses 21 to 35.
36. THE PRIESTS: The record of the priests and Levites in Nehemiah 12:1–9 lists three generations of high priests beginning with Jeshua. Ezra lists priests and Levites in verses 36 to 42.
43. THE NETHINIM: These were temple servants—the descendants of the Gibeonites who performed servile duties at the temple. Ezra lists them in verses 43 to 54.
55. THE SONS OF SOLOMON'S SERVANTS: These are the descendants of Solomon's servants, whom Ezra lists in verses 55 to 57.
59. AND THESE WERE THE ONES: In verses 59 to 62, Ezra lists several families whose genealogical information could not be verified.
63. CONSULT WITH THE URIM AND THUMMIM: Because these people's lineage could not be determined, they were banned from priesthood and could not "eat of the most holy things" until a priest could consult with the Urim and Thummim. These objects, kept in the breastplate of the high priest, were used to determine God's will. "You shall put in the breastplate of judgment the Urim and the Thummim, and they shall be over Aaron's heart when he goes in before the LORD" (Exodus 28:30).
64. THE WHOLE ASSEMBLY TOGETHER: The number of returning exiles listed here is 12,000 more than the particular numbers given in the catalog, when added together. Reckoning up the smaller numbers, they amount to 29,818 in Ezra 2 and to 31,089 in the parallel chapter in Nehemiah 7. Ezra also mentions 494 persons omitted by Nehemiah, and Nehemiah mentions 1,765 not noticed by Ezra. If Ezra's surplus is added to the sum in Nehemiah, and Nehemiah's surplus to the number in Ezra, they both become 31,583. Subtracting this from 42,360, there is a deficiency of 10,777. These were omitted because they did not belong to Judah or Benjamin or to the priests, but to the other tribes. The servants and singers, male and female, were reckoned separately so that, putting all of them together, the number of all who went with Zerubbabel amounted to 50,000, with 8,000 beasts of burden.
69. GOLD DRACHMAS . . . MINAS OF SILVER: Drachma probably refers to a Persian coin, the _daric_ , named after Darius I. This quantity would have amounted to approximately 1,100 pounds of gold. A mina weighed about one and one-quarter pounds, so this would represent three tons of silver.
> _B UILDING THE ALTAR: The returning Jews soon gather in Jerusalem, where they will rebuild the temple. However, they begin the project by first building the altar._
3:1. THE SEVENTH MONTH: After the Jews' arrival in their homeland, they were at first occupied with their own dwellings in and around Jerusalem, but after that work was done, they turned to building the altar of burnt offering. In the "seventh month" (September–October 537 BC), the Jews celebrated three important annual observances: the Feast of Trumpets, the Day of Atonement, and the Feast of Tabernacles (see Numbers 29:1–38). Such an assembly had not convened for seventy years. More than ninety years later, Nehemiah and Ezra would lead a similar celebration (see Nehemiah 8:13–18).
THE PEOPLE GATHERED TOGETHER AS ONE MAN: This refers again to the people who had chosen to return to Jerusalem. Their unity of purpose here demonstrated that they were deeply concerned with restoring the forms of worship that the Lord had commanded under Moses. It also suggests that the Spirit of God was at work in their hearts, stirring them up to obedience just as He had done with Cyrus and others in Ezra 1:5.
2. THEN JESHUA . . . AND ZERUBBABEL: It is significant that these two important men—Jeshua, the high priest, and Zerubbabel, the head of the tribe of Judah—led the Jews in returning to Jerusalem. The kings and priests of Judah and Israel had previously led the people _away_ from God, but now the nation's leaders, both at the civic and spiritual levels, were leading the people _back_ to obedience to His Word.
BUILT THE ALTAR OF THE GOD OF ISRAEL: It is also significant that the people's first act of rebuilding was the altar rather than the temple or the city walls. This demonstrated that the returning exiles' first priority was in offering sacrifices and repenting of their sins, and that this took precedence over any other act of worship (and even over their own physical safety). The people were trusting God to protect them while they obeyed His commands.
3. FEAR HAD COME UPON THEM: The settlers who had come to occupy the land during the seventy years of Israel's absence were deportees brought in from other countries by the Assyrians and the Babylonians. They saw the returning Jews as a tremendous threat to their possessions, land, and way of life, for they viewed the land as their own by right of seventy years of occupation, while the Jews viewed it as theirs by virtue of God's command. We can still see this conflict occurring in the Middle East today.
SET THE ALTAR: This was all that was needed to reestablish temple worship. The people reset it on its old foundation (or _bases_ ) so it occupied its sacred site. The burnt offerings they sacrificed to God were the most common offerings for sin.
> _L AYING THE FOUNDATION: Having established obedience to God's Word, the people now turn their attention to the mechanics of building the temple._
4. THE FEAST OF TABERNACLES: This festival, also called the Feast of Booths, commemorated the Israelites' wandering in the wilderness during their exodus from Egypt (see Leviticus 23:33–43).
5. A FREEWILL OFFERING TO THE LORD: In addition to reinstituting the sacrifices and feasts that God had commanded, the people began to offer voluntary tithes—all prior to rebuilding the temple. This demonstrated that worship of the Lord consists of confession and repentance of sins, worship, and giving—but not meeting inside a building. The building came later, but only after proper worship and obedience had been reestablished.
6. THE FOUNDATION OF THE TEMPLE OF THE LORD HAD NOT BEEN LAID: In a spiritual sense, the people were laying a firm foundation for the temple by obeying God's Word. Obedience is more important to God's church than bricks and mortar.
7. THEY ALSO GAVE MONEY TO THE MASONS AND THE CARPENTERS: The process of rebuilding the temple is similar to the original construction under Solomon (see 1 Kings 5–6). The Jews gave freely of their possessions, time, and skills to the construction and paid laborers and manufacturers the proper rates. Note that they did not ask foreigners to donate. The Jews would have accepted those gifts if offered, but they did not expect that of foreigners, nor did they try to get things for free. As David said, "I will not take what is yours for the LORD, nor offer burnt offerings with that which costs me nothing" (1 Chronicles 21:24).
PEOPLE OF SIDON AND TYRE . . . CEDAR LOGS FROM LEBANON . . . TO JOPPA: The workmen loaded cedar logs onto ships in the cities of Tyre and Sidon to the north, on the coast of the Mediterranean (or Great Sea), and then sailed them south to the city of Joppa, which was approximately forty miles from Jerusalem.
8. IN THE SECOND MONTH OF THE SECOND YEAR: That is April–May 536 BC. This officially ended the seventy-year captivity that began in 605 BC. Once again, we see that construction of the temple did not even begin until all the elements of obedience and worship were in place.
> _W EEPING AND REJOICING: The people gather in unity and respond to the preliminary construction—some with weeping, and others with shouts of joy._
ALL THOSE WHO HAD COME OUT OF THE CAPTIVITY: Once again, we see the unity and wholeheartedness of God's people. They joined together as one to participate in the building of the temple, just as they had shown unity in obedience.
11. THEY SANG RESPONSIVELY: This song of praise is evidently from Psalm 136:1. The priests might have sung, "Oh, give thanks to the LORD," and then one group of the people would have responded with the words, "For He is good!" The second group would then respond, "For His mercy endures forever toward Israel."
12. OLD MEN . . . WEPT WITH A LOUD VOICE: These old men had been young when carried into captivity. They would have remembered Solomon's glorious temple that stood in Jerusalem at the time. They grieved over the lost splendor of Solomon's structure, for this second temple did not begin to match the grandeur of the first temple, nor did the presence of God reside within it. It was apparent that the new nation was small and weak—and that the new temple was smaller and less beautiful by far. Gone were all the riches and glory as in the days of David and Solomon. However, most of all, the old men wept over the fact that the ark of the covenant was gone, and with it the manifestation of the Lord's presence.
YET MANY SHOUTED ALOUD FOR JOY: However, in spite of the loss of the ark, many of the people recognized that God was still with His people—which alone was cause for joyful shouting and praise (see Zechariah 4:7–10). It is possible that Psalm 126 was written for this occasion and sung at the celebration.
13. THE SOUND WAS HEARD AFAR OFF: The obedient praise and worship of God's people carries a profound testimony to the world around us. It draws people toward the Lord just as curiosity might have drawn neighboring people toward Jerusalem on this day.
## UNLEASHING THE TEXT
1) Why does this passage reiterate that the people gathered "as one man" (Ezra 3:1)? What does this indicate? What drew the people together?
[Your Response Here]
2) Why did the returning exiles build the altar before the temple that would house it? Why did they build it before they constructed the city walls that would protect them? What does this indicate they knew about God's priorities?
[Your Response Here]
3) Why had fear come on the people (see Ezra 3:3)? What did this have to do with rebuilding the altar?
[Your Response Here]
4) What was involved in the actual building of the temple? What roles did the people play in the process? What did the construction cost them?
[Your Response Here]
## EXPLORING THE MEANING
**_God dwells in believers, not in a temple._** Cyrus the Great of Persia allowed the Jews to return to Jerusalem for the express purpose of rebuilding the Lord's temple there. However, as we have seen in this study, the people did not immediately turn their attention to that particular construction project. Rather, the first thing they did was rebuild the altar—even though that altar would eventually be housed within the temple. Their top priority was being personally and corporately obedient to God's Word, not constructing a building in which to worship.
The people's priorities demonstrated that they understood God's priorities. They knew that "to obey is better than sacrifice, and to heed [God's Word] than the fat of rams" (1 Samuel 15:22). Yet in those days the temple represented the Lord's presence, and its completion was important to the proper worship by His people. This emphasis on personal and corporate sacrifice underscored the fact that the Lord is more concerned with our obedience than with our outward shows of worship.
When Jesus was crucified, God abolished the need for the temple in Jerusalem. He tore the curtain that separated the Most Holy from His people to demonstrate that His presence would no longer be found inside a building. God's presence now resides in the very people who belong to Him, because God has reconciled them to Himself through their faith in Jesus Christ. God said this to John in the book of Revelation: "Behold, the tabernacle of God is with men, and He will dwell with them, and they shall be His people. God Himself will be with them and be their God" (21:3). There is no longer a need for a temple simply because _we_ are His temple!
**_God's people are unified through their obedience._** The devil seeks to divide believers. He works ceaselessly to separate what God has brought together, whether that is marriages, churches, or any other relationship in which unity and commitment are essential. Ironically, the devil also works tirelessly to _remove_ separations and divisions that God has established: right from wrong, darkness from light, good from evil . . . and on and on. The evil one's goal is to create unity in wickedness and disunity in righteousness.
It is interesting to note that the nations of Israel and Judah were fairly unified in pursuing wickedness for many generations. This was the sort of unity that the devil breeds, not the unity that God requires. The Lord calls His people to be unified in obedience. When we pursue disobedience, or when we allow disunity to separate Christians, we are following the paths of wickedness rather than the ways of righteousness.
The exiles who returned to Jerusalem drew together with one accord. They all worked together as one to reestablish godly worship—and the Lord blessed their efforts tremendously. We will see clearly as we go through these studies how the Lord stymied the efforts of others who sought to disrupt His work and how He preserved and blessed His people as they did that work. When Christians unite in obedience to God's Word, nothing can stand in their way.
**_The Lord will put an end to mourning._** When the returning exiles laid the foundation for the temple in Jerusalem, there was a mixed response from the people. Many shouted with joy at seeing the Lord rebuilding what had been utterly destroyed. At the same time, many others wept with deep grief over their memories of what had once been. Ironically, both responses were valid—but in the long run, the joyful shouting drowned out the tears.
The people of Israel had good reason to mourn, because it was their own sin and stubbornness that had led to the destruction of the temple and the loss of all it contained—including the ark of the covenant. Yet God had not abandoned His people, and He had not forgotten His promises of faithfulness and blessing. The Lord was still at work, and what mattered most was that Israel still existed, due to the promise of God. Tears had a place, but the shouts of joy would last far longer.
When we sin, we often do grave damage to ourselves and to others. We do well when we grieve and mourn over our own sinful behavior. We also do well when we repent and recognize the depth of the damage we have caused. But we must also remember that God has not abandoned us, even in the midst of deliberate sin. He may choose to bring discipline into our lives to urge us toward repentance, but He will never forsake us and will never disinherit those who are His children through the redemption of Jesus Christ. In the eternal kingdom, "God will wipe away every tear from their eyes; there shall be no more death, nor sorrow, nor crying. There shall be no more pain, for the former things have passed away" (Revelation 21:4).
## REFLECTING ON THE TEXT
5) Why did some people shout with joy when the foundation was laid? Why did others weep? If you had been there, how do you think you would have responded?
[Your Response Here]
6) How would you have reacted if you had been outside Jerusalem at the time of this loud shouting? How does your worship of God influence the people around you?
[Your Response Here]
7) Why was unity among the people so important to the effort to rebuild the temple? In what ways today is unity among Christians important?
[Your Response Here]
8) What did the apostle Paul mean when he said a Christian is "the temple of God" (1 Corinthians 3:16)? What implication does this have in your life?
[Your Response Here]
## PERSONAL RESPONSE
9) When have you grieved over your past sins? When have you rejoiced over God's faithfulness and blessings? When has God turned your tears into joy?
[Your Response Here]
10) Which takes higher priority in your life: personal obedience to God's Word or regular attendance at church? Which is God's higher priority?
[Your Response Here]
# **3**
# TROUBLE WITH SAMARITANS
# _Ezra 4:1–5:17_
## DRAWING NEAR
What are some famous deals—whether in business, politics, or other—that seemed good on the surface but in the end proved to be disastrous to one of the parties?
[Your Response Here]
## THE CONTEXT
As previously mentioned, in 722 BC the Assyrian king Shalmaneser V invaded the northern kingdom of Israel and quickly conquered the land. However, the people in the capital city of Samaria—due to the city's strong walls, internal water supply, and storehouses of food—were able to resist the Assyrian attack for three years. When the city finally fell and the people were carried away, the Assyrians repopulated the area with people from other lands they had conquered. Those foreigners settled in Samaria and intermarried with the few Israelites who were still living there, and their descendants became known as Samaritans.
These Samaritans did not fear God and, as a result, the Bible tells us that as a form of discipline, "the LORD sent lions among them, which killed some of them" (2 Kings 17:25). The people recognized this as the hand of God, but they thought He was only the God of Samaria—they did not understand that His control extended to the entire world. So the Lord sent a priest to live with them, and he "taught them how they should fear the LORD" (verse 28). The Samaritans embraced the worship rituals the priest taught, but they did not forsake their other gods. They believed they could simply add the God of all creation to their long list of idols, and so invented their own syncretistic religious practices.
When Zerubbabel and the people of Judah returned to Jerusalem, they discovered that the Samaritans were still living nearby. Soon the two groups would begin to interact—and those encounters would not lead to pleasant outcomes for the Jews. In this study, we will see what happens when obedient believers like the exiles are confronted by pious frauds.
## KEYS TO THE TEXT
Read Ezra 4:1–5:17, noting the key words and phrases indicated below.
> _A F RIENDLY OFFER: The Samaritans, the neighbors of the Jews, come to Jerusalem and offer to help build God's temple. But the people of Judah reject the offer._
1. THE ADVERSARIES OF JUDAH AND BENJAMIN: These _adversaries_ were the Samaritans. As previously noted, these foreigners had intermarried with the remaining Israelites and brought their own brands of idolatry with them.
2. WE SEEK YOUR GOD AS YOU DO: This statement was true from the Samaritans' point of view, but it was absolutely false from God's perspective. The Lord had sent a priest to Samaria to teach His truth to those who had been transported there by the Assyrians during the Israelite captivity. The people had embraced the Lord in part, thinking He was "the God of the land" (2 Kings 17:26) where they had been transported, and they wanted to appease Him. However, they had not completely forsaken their false gods in the process; rather, they had merely attempted to add the one true God to their pantheon of idols.
3. WE ALONE WILL BUILD TO THE LORD GOD OF ISRAEL: This was an important and costly decision on the part of Judah's leaders. They were not arbitrarily excluding outsiders from building the temple out of some spirit of snobbery, nor were they excluding others merely out of obedience to Cyrus's commands. Rather, they were taking a firm stand that only God's obedient worshipers were permitted to participate in building His temple, and they were avoiding contact with the very idolatry that had caused the Lord to send them into captivity in the first place. The note from Cyrus gave authority to their refusal.
> _S HOWING THEIR TRUE FACE: The Samaritans respond by showing their true motives and begin an ongoing attempt to stop God's work in Jerusalem._
4. TRIED TO DISCOURAGE THE PEOPLE OF JUDAH: Discouragement is one of the devil's favorite tactics in trying to thwart the work of God. His goal is to persuade God's people to give up on obedience and turn to the easier temptations of the flesh. God had previously warned His people not to give in to discouragement: "Look, the LORD your God has set the land before you; go up and possess it, as the LORD God of your fathers has spoken to you; do not fear or be discouraged" (Deuteronomy 1:21). The same warning is relevant for us today—God wants us to choose instead to trust Him for the outcome.
5. HIRED COUNSELORS AGAINST THEM: The Samaritans took some form of legal action against the Jews. They hoped to bog down the Jews in litigation if they couldn't prevent the rebuilding outright. God's enemies still use this tactic today by staging attempts to thwart the spread of the gospel through laws and lawsuits.
FRUSTRATE THEIR PURPOSE: The legal experts the Samaritans hired were successful in causing a sixteen-year delay of the rebuilding effort (c. 536–520 BC). In the meantime, the people took more interest in their personal affairs than spiritual matters (see Haggai 1:2–6). This delay lasted "all the days of Cyrus" and into the days of Darius, who ruled Persia from 521–486 BC.
6. THEY WROTE AN ACCUSATION: The Hebrew word for _accusation_ here is related to the word for _Satan_. Satan is the accuser of the brethren (see Revelation 12:10) and tirelessly brings accusations against God's people. The enemies of God's people are like their father, the devil (see John 8:44), and thus persistently do the same.
7. ARTAXERXES KING OF PERSIA: Ezra chronicles some of the continued opposition the Samaritans waged against the Jewish people. The letters sent to King Artaxerxes occurred later, during the time when Nehemiah was ministering in Jerusalem. Note that two different words for _letter_ are used in this passage. The first is an official document as opposed to a simple form of correspondence. The second is the generic term for letter. The context verifies the choices of two different terms, since two different letters are indicated.
10. THE GREAT AND NOBLE OSNAPPER: This probably is another name for the Assyrian king Ashurbanipal (c. 669–633 BC), who resettled Samaria with foreigners. The obsequious tone of the enemies' letter is typical of those who use flattery and attempt to ingratiate themselves to those in power in the hopes of accomplishing their own personal agendas. The Assyrians were noted for their cruelty and barbarity, and their kings were anything but "great and noble."
12. THE REBELLIOUS AND EVIL CITY: God's view of Jerusalem was quite different: "Beautiful in elevation, the joy of the whole earth, is Mount Zion on the sides of the north, the city of the great King" (Psalm 48:2). Yet this illustrates another favorite tactic of God's enemies: to accuse His people of the very wickedness that they commit themselves. As Isaiah wrote, "Woe to those who call evil good, and good evil; who put darkness for light, and light for darkness; who put bitter for sweet, and sweet for bitter!" (Isaiah 5:20).
12. THE JEWS THAT CAME UP FROM YOU: The name _Jews_ for the people of God was generally used after the captivity, because the exiles who returned to the land were mainly from Judah. Most of the people of the ten northern tribes had been dispersed, and the largest number of returnees came from the two southern tribes.
> _S PEAKING THE KING'S LANGUAGE: The enemies of God finally find a sympathetic ear with King Artaxerxes—when they tell him that he will lose money—and he stops the work._
13. THEY WILL NOT PAY TAx, TRIBUTE, OR CUSTOM: The Samaritans told the Persian king that the Jews would not pay any taxes or tributes to him after the work on the city and its walls had been completed. Of course this was not true, yet the accusation hit a chord in the mind of King Artaxerxes. Judah had historically at times refused to pay tribute to foreign kings prior to the captivity.
THE KING'S TREASURY WILL BE DIMINISHED: Here is the crux of the Samaritans' real argument. If King Artaxerxes permitted the Jews to continue their rebuilding projects, he would end up losing money. Rulers in Ezra's day were no different from those in our own time, and this argument proved all too effective.
14. IT WAS NOT PROPER FOR US TO SEE THE KING'S DISHONOR: Here again we see the hypocrisy and dissembling of God's enemies. They pretended to have high and lofty motives, yet they were merely pursuing their own selfish gain.
15. THE BOOK OF THE RECORDS: This was an administrative document called a "memorandum" kept on file in the royal archives. The enemies of the Jews noted that in this document the king would find a record of Jerusalem's destruction by the Babylonian king Nebuchadnezzar (c. 586 BC).
19. REBELLION AND SEDITION HAVE BEEN FOSTERED IN IT: This refers to the rebellions of Kings Jehoiakim (see 2 Kings 24:1), Jehoiachin (see verses 9, 12), and Zedekiah (see verse 20)—all of whom ruled in Judah prior to the captivity. These rebellions were against Babylon, not against Persia.
19. AND I GAVE THE COMMAND: This line might better be translated, "I established a decree." In other words, this was no simple routine order given by the king to one person, but a major edict to a large group of people.
21. NOW GIVE THE COMMAND: With this decree the king called to a halt not the efforts of one or two workers but the efforts of 50,000 workers. The king was commissioning a decree of great significance, and it would not lose its authority until the king established a new decree.
23. KING ARTAXERXES' LETTER: Another official document, as opposed to a generic letter, came from Artaxerxes transferring authority to the regional leaders to establish the decree. Without the king's official administrative correspondence, the decree could not have been established.
BY FORCE OF ARMS MADE THEM CEASE: The enemies of God's people may succeed temporarily in outlawing obedience to His Word, but God is always in control and no power on earth or in hell can prevent the fulfillment of His plans. In a later study, we will see that the people of Jerusalem eventually completed the temple in spite of the efforts of their enemies.
24. THE WORK . . . CEASED, AND IT WAS DISCONTINUED: It was this decree that halted the Jews' rebuilding efforts for sixteen years.
> _T HE WORK RESUMES: God speaks through His prophets and instructs the people to continue the work on the temple—in spite of the Persian king's decree._
5:1. HAGGAI AND ZECHARIAH: The book of Haggai is styled as a "royal administrative correspondence" sent from the Sovereign King of the universe through the "the LORD's messenger," who was Haggai (1:13). The prophet addressed part of the message specifically to Zerubbabel, the political leader, and to Joshua, the religious leader, telling them to "be strong . . . and work" on the temple because God was with them (see 2:4). This prophet, along with Zechariah, gave severe reproaches and threats to the people if they did not return to the building, but promised national prosperity if they continued the work. Not long after the exiles heard this message, the temple work began once again.
2. PROPHETS OF GOD: These men who helped in building the temple would be in addition to Haggai and Zechariah.
3. TATTENAI THE GOVERNOR: He was most likely a Persian official. He basically asked the people, "Who issued you a royal decree to build?"
5. BUT THE EYE OF THEIR GOD WAS UPON THE ELDERS: God's hand of protection, which had led this endeavor, allowed the work to continue while official communication was going on between Tattenai and Darius, the Persian king.
> _T HE GOVERNOR'S LETTER: Tattenai reports to King Darius that the Jews are violating the official decree that is in place—but he also includes the Jewish people's response._
8. HEAVY STONES, AND TIMBER: This technique of using beams and stone blocks was a well-known form of wall construction. It appears the governor mentioned this in his letter to tell Darius that the Jews were preparing for conflict or battle. Including this piece of information would have served as a threat to the Persian king, who wanted no such conflict.
11. THEY RETURNED US AN ANSWER: In other words, the Jews sent back a report (an official document for the archives). The "great king of Israel" the people mention was Solomon, who built the first temple (c. 966–960 BC).
12. GAVE THEM INTO THE HAND OF NEBUCHADNEZZAR. This expression was used commonly in royal administrative correspondence when a more powerful administrator, such as a king, relinquished some of his authority to an underling and yet kept the lower administrative official under his command. The point here was that God—as the sovereign King of the universe—had satisfied His wrath by relinquishing the authority for this administrative action to Nebuchadnezzar. The greatest king the ancient Near East had ever known was just a petty official in the administration of the sovereign Lord.
13. IN THE FIRST YEAR OF CYRUS: This refers to the initial proclamation that Cyrus had made in Ezra 1:2–4, which initiated the Jewish people's return to Jerusalem.
16. SHESHBAZZAR CAME AND LAID THE FOUNDATION: This seems to contradict the statement in Ezra 3:8–10 that Zerubbabel, Jeshua, and the Jewish workmen laid the foundation, but in truth it does not. Sheshbazzar was the political appointee of the Persian king over the Jews, and he was thus given official credit for work actually done by them.
## UNLEASHING THE TEXT
1) If you had been present when the Samaritans offered to help, how would you have responded? Why did the Jews reject their offer?
[Your Response Here]
2) The Samaritans told the Jews, "We seek your God as you do" (Ezra 4:2). In what ways was this partially true? In what ways was it false? Why did it matter?
[Your Response Here]
3) What tactics did the Samaritans use to discourage the people of Judah? What turned out to be the most effective tactic they used against God's people?
[Your Response Here]
4) How did the Jewish people respond when the Persian governor asked them why they had defied an official decree? How would you have responded if you were in their place?
[Your Response Here]
## EXPLORING THE MEANING
**_One cannot serve God and other "gods" at the same time._** The people of Samaria had been taught the truths of God's Word, and they understood (at least outwardly) the proper worship of Yahweh. They even embraced that worship and made sacrifices to God on a regular basis. The problem was that they did not renounce their false gods. They hoped to placate the true God while not repenting of their idolatry. To the casual observer they might have appeared as worshipers of Yahweh, but they did not obey His command to turn away from false gods—and God does not share His glory with any mythical pantheon.
People still make this mistake today. Many have been instructed in the truths of God's Word, and they may even regularly attend a local church—but church attendance will not bring salvation. We must _serve God_ and God alone _—_ not both God and the flesh. Jesus was clear on this when He said, "No one can serve two masters; for either he will hate the one and love the other, or else he will be loyal to the one and despise the other. You cannot serve God and mammon" (Matthew 6:24). ( _Mammon_ refers to the love of material possessions and comforts, but this is only one of the many forms of idolatry which people engage in today.)
Another trap the devil uses is _syncretism_ , where people attempt to comingle elements of many different religions into their own personal brand of worship. This is also very common today, as people attempt to amalgamate many contradictory religious systems into their own notions of "transcendence." Even some who purport to teach God's Word are adulterating it with worldly notions of evolution, psychology, and many other false teachings. But Jesus stated clearly, "I am the way, the truth, and the life. No one comes to the Father except through Me" (John 14:6). There is no other way to salvation than through Jesus Christ, and the Lord will not tolerate any adulteration of His Word.
**_The devil wears many disguises._** The Samaritans approached the leaders of Judah with a friendly offer to help them build the temple. They even pointed out that because they worshiped the same God, they should all be working together toward their common goal. To refuse such kindness would be rude at best—a refusal to get along with others, even a self-righteous hypocrisy—yet that is precisely what God's people did. The Jewish leaders refused to permit the Samaritans to participate in the work that God had given to them.
The Jews' refusal, however, was not due to self-righteousness or rudeness but because they recognized the evil one was behind the offer. The Jews did not judge the Samaritans' hearts but simply compared their behavior with God's revelation and recognized the Samaritans did not worship the Lord the way He had commanded. That was enough to cause Judah to reject their offer, whatever spirit lay behind it. Subsequent events proved the offer was made in a spirit of duplicity. The Samaritans were pretending to be friendly while actually being sinister. Solomon warned of such people when he wrote, "Faithful are the wounds of a friend, but the kisses of an enemy are deceitful" (Proverbs 27:6).
Paul further warned us, "Satan himself transforms himself into an angel of light. Therefore it is no great thing if his ministers also transform themselves into ministers of righteousness, whose end will be according to their works" (2 Corinthians 11:14–15). In other words, Christians must be on guard against those who pretend to be godly but are not, because there are many who try to "transform themselves into ministers of righteousness" by wearing a false disguise. Jesus warned us, "Behold, I send you out as sheep in the midst of wolves. Therefore be wise as serpents and harmless as doves" (Matthew 10:16). Christians should never practice hypocrisy or false pretense, but the devil always does.
**_God commands us to resist discouragement._** The enemies of God's people attempted to interrupt their works of obedience by causing them to become discouraged. The _King James Version_ renders Ezra 4:4: "The people of the land weakened the hands of the people of Judah," which captures the essence of discouragement—to become weak, to sink down, to lose the ability to carry on, and to let God's projects drop from despair.
Fear is at the root of discouragement. We are suddenly faced with a circumstance that is beyond our control, and we quickly begin to fear that it is beyond _God's_ control as well. And if it is beyond God's control, we might as well just give up now—which is, of course, precisely what Satan is hoping that we will do! However, God commands us to not give in to fear but to strengthen our hands when they become weak (see Isaiah 41:10; Hebrews 12:12).
The best way for us to do this, wrote the author of Hebrews, is to "consider Him who endured such hostility from sinners against Himself, lest you become weary and discouraged in your souls" (12:3). We must remember that Jesus Himself faced immense opposition—more severe than any we will ever face. Yet He overcame it all through the faithfulness of God and through utter confidence in and reliance on God's sovereignty.
## REFLECTING ON THE TEXT
5) Why were the Samaritans not qualified to participate in the work of building the temple? What was wrong with their worship of God? Why are such matters important to God?
[Your Response Here]
6) What is the difference between discerning false motives and judging other people? What does it mean to be "wise as serpents and harmless as doves" (Matthew 10:16)? How is this done?
[Your Response Here]
7) How is discouragement related to fear? How is fear related to faith? What is the solution to discouragement? What role do a person's deliberate choices play?
[Your Response Here]
8) What elements of syncretism (blending false religions with God's Word) are being taught today? How can a Christian discern such false teachings?
[Your Response Here]
## PERSONAL RESPONSE
9) Is someone or something discouraging you from doing the will of God? How can the example of Jesus give you courage and perseverance? What can other Christians do to help you?
[Your Response Here]
10) Are you trying to serve two masters? What is competing with your loyalty to Christ? What will you do to remove that from your life?
[Your Response Here]
# **4**
# CHOOSING GOD'S PRIORITIES
# _Haggai 1:1–2:9_
## DRAWING NEAR
What are some comparisons people make when it comes to a home, a car, or success? What traps do people fall into when they compare their situation to the situation of others?
[Your Response Here]
## THE CONTEXT
Haggai was a prophet of the Lord in Jerusalem during the time of Zerubbabel. Essentially nothing is known about him apart from his own writings and the mentions made in Ezra, yet he is specific as to the dates of his prophecies: a four-month period during 520 BC. He most likely had returned from Babylon to Jerusalem with Zerubbabel eighteen years earlier, in 538 BC, and he might even have been old enough to have been carried into captivity with Daniel in 586 BC.
The name Haggai means "festival." At first glance this name may seem quite ironic, given the disciplinary content of his message from God. But the Lord had sent His people to Jerusalem specifically to reinstate His prescribed worship practices and to rebuild His chosen city, which _should_ have been a festive occasion. If there was any loss of festivity, it was because the Lord's people had lost His perspective.
As we have seen, the people had become discouraged by the opposition of their neighbors and had wrongly concluded it was not yet time for them to rebuild the temple. So instead, they turned to expanding their own wealth. Like their ancestors who had been exiled before them, they ceased caring about God's kingdom and started building their own. With a biting query, the Lord reminded them it was not right for them to live in paneled houses while the temple lay in ruins. He urged them to consider the consequences of their indifference.
In this study, we will learn that the Lord wants His people to share His priorities, and we will discover what He sometimes does to get our attention. The central message of Haggai is vital to us today: consider your ways!
## KEYS TO THE TEXT
Read Haggai 1:1–2:9, noting the key words and phrases indicated below.
> _M AKING EXCUSES: The Jews in Jerusalem have stopped building the temple, at first by compulsion but now by choice. So the Lord sends His prophet with an important message._
1. THE SECOND YEAR OF KING DARIUS: Not to be confused with Darius the Mede (see Daniel 5:31), Darius I (Hystaspes) became king of Persia in 521 BC, having ascended to the throne after the death of Cambyses. As an officer of Cambyses and the great-grandson of Cyrus the Great's brother, Darius retained the loyalty of the Persian army and thereby defeated other contenders for the throne. He reigned until his death in 486 BC.
IN THE SIXTH MONTH, ON THE FIRST DAY: This first day of the month of Elul in the Jewish calendar would correspond to August 29, 520 BC.
HAGGAI THE PROPHET: Little is known about Haggai apart from what he tells us in this book. He apparently traveled with Zerubbabel in the first wave of Jews returning to Jerusalem.
ZERUBBABEL . . . GOVERNOR OF JUDAH: As previously mentioned, Zerubbabel was the grandson of King Jehoiachin and in the Davidic line. Although he is not to be identified with Sheshbazzar (see Ezra 1:8), his role as civil leader and overseer of the temple rebuilding project is certain. He reestablished the Davidic throne, though it would not again be occupied until the time of the Messiah.
JOSHUA . . . THE HIGH PRIEST: Spelled _Jeshua_ in Ezra 3:2, this man was a descendant of Jehozadak (see 1 Chronicles 6:15) and the religious leader of the exilic community that returned to Jerusalem. He reestablished the high priestly line of Aaron though Eleazar. His father, Jehozadak, had been one of Nebuchadnezzar's captives.
2. THE TIME HAS NOT COME: Haggai began his message by quoting a popular expression of the people who said it was not time to build the temple. Although they had endured the hostile opposition of their neighbors and the lack of economic prosperity, the roots of their reluctance lay ultimately in their selfish indifference to the Lord. Notice also that God referred to the Jews as "this people" rather than "My people." It would be comparable to an angry man referring to his spouse as "this woman" rather than "my wife."
4. TIME FOR YOU YOURSELVES TO DWELL IN YOUR PANELED HOUSES: Walls and ceilings overlaid with cedar were common in wealthy residences. There is a biting irony in the Lord's words. If Jewish exiles' own houses were falling apart, they would not casually say, "The time just hasn't come yet for us to repair the roof." The Lord's people had made their own comfort a high priority while caring nothing about the Lord's house. They were building luxury onto their houses, while the Lord's house was in ruins.
> _G ET BUSY: The Lord commands His people to stop making excuses, make His priorities their priorities, and resume the work on rebuilding the temple._
5. CONSIDER YOUR WAYS: This is the central message of Haggai's prophecies. He called on God's people to take stock of their lives, ensure they were living in obedience to God's Word, and make sure they were sharing His priorities as their own top priorities.
6. YOU HAVE SOWN MUCH, AND BRING IN LITTLE: Ironically, the people had stopped working on the temple to save riches for themselves, yet the Lord was disciplining them by sending drought and famine. The more they saved, the less they had. However, Haggai says, if they would make God's priorities their own, He would once again send them fullness and prosperity.
8. GO UP . . . AND BRING WOOD AND BUILD: Three relatively simple steps outline what the Lord required of His people in this project. They were to go to the forests (which had regrown during the captivity), make lumber, and start building. By putting God first, they would honor Him in their worship, and they would be blessed in the secondary matters of life. There was no mystery surrounding the Lord's will, and the people did not lack knowledge of what His plan was for them. What they lacked was simple obedience.
THAT I MAY TAKE PLEASURE IN IT: The Lord's pleasure would not come from the building itself. In fact, this temple was to be a pitiful structure compared to the magnificent temple that Solomon had built (see 2 Chronicles 2–5). Rather, the Lord would take pleasure in seeing His chosen people worshiping Him in the way He had prescribed, and His name would be glorified when the world saw them doing so.
9. I BLEW IT AWAY: Literally, "I sniffed at it!" The Lord snorted with contempt when He saw His people's priorities as they strove after the things of this world that would not last.
EVERY ONE OF YOU RUNS TO HIS OWN HOUSE: That is, the Jews were busy running to and fro, looking after their own interests, while the house of the Lord lay in ruins. Jesus addressed this same conflict of priorities when He said, "Do not worry, saying, 'What shall we eat?' or ' What shall we drink?' or 'What shall we wear?' For after all these things the Gentiles seek. For your heavenly Father knows that you need all these things. But seek first the kingdom of God and His righteousness, and all these things shall be added to you" (Matthew 6:31–33).
11. ON ALL THE LABOR OF YOUR HANDS: The price of the people's disobedience had come in the form of an economic catastrophe resulting from God withholding the summer dew. Grain, wine, and oil—the primary crops of the land—could not grow, and the cattle also languished because of the absence of the people's spiritual health. Yet the Lord sent not only drought and famine on the people but also futility and frustration on all their labors. He longed to bless their work with success, but first their work needed to be _His_ work.
12. THE PEOPLE FEARED THE PRESENCE OF THE LORD: The exiles who had returned from Persia took Haggai's message to heart. They realized the words of the prophet were from the Lord and "obeyed" and "feared," knowing that God was present. They renewed their commitment to Him and to seeking His presence. The people's repentance had come less than two weeks after Haggai began prophesying on September 21, 520 BC.
13. I AM WITH YOU: As soon as the people repented of their false priorities and embraced God's priorities once again, the Lord relented His hand of discipline in their lives and began to pour out His blessings. These words are among the most joyful and encouraging words of Scripture. They serve to continually remind us that the almighty God is always with His people in whatever situation they face.
14. STIRRED UP THE SPIRIT: The Lord energized the leaders and the people through His Word to carry on the work of rebuilding the temple. God had sovereignly moved in the heart of Cyrus the Great sixteen years earlier, and now the people's response of repentance and obedience allowed God's Spirit to also energize them for the task.
> _T HE LATTER TEMPLE: The Lord sends another prophecy through Haggai, this time concerning events in the distant future, when God would build His most glorious temple._
2:1. IN THE SEVENTH MONTH: This twenty-first day in the Jewish month of Tishri corresponds to October 17, 520 BC. According to Leviticus 23:39–44, this would be the final day of the Feast of Tabernacles, in which the people celebrated God's provision for the Israelites during their forty years of wilderness wanderings and gave thanks for a bountiful harvest. On this occasion, the Lord gave Haggai another message to give to the people.
2. TO THE REMNANT OF THE PEOPLE: Haggai's first message had been directed toward the leaders, Zerubbabel and Joshua. But in this message the prophet includes the remainder of the exiles who returned from Babylon.
3. IS THIS NOT IN YOUR EYES AS NOTHING: The older Jews would have remembered the magnificence and splendor of Solomon's temple, and this rebuilt one must have seemed like a hovel when compared to that structure. Yet the Lord was not discouraged, nor did He want His people to succumb to discouragement. He had a far more glorious temple in the works that no man had ever dreamed of—one not made by human hands.
4. BE STRONG: On the surface, this may sound like a platitude, similar to "cheer up." But the Lord was actually commanding His people to be strong. Strength is generally something that is not a matter of choice—a person is either physically powerful or not, according to the bodily structure with which he or she was born. Yet the strength the Lord commands _is_ a matter of choice: the choice to set one's heart on steadfast obedience regardless of the cost.
5. DO NOT FEAR: Closely related to the previous command is the command not to fear. The Israelites would have been discouraged by the small size of this temple, compared to their previous one, but God did not want them to fear the future because He had a plan that was greater than anything they could imagine. His covenant commitment and the promise that His Spirit would be with them as "when you came out of Egypt" would have been most reassuring.
6. I WILL SHAKE: The shaking of the cosmic bodies and the nations would go beyond the historical removal of kingdoms and the establishment of others, such as the defeat of Persia by Greece (see Daniel 7). Rather, the text looks to the cataclysm in the universe described in Revelation 6–19, the subjugation of the nations by the Messiah, and the setting up of His kingdom that will never be destroyed (see Hebrews 12:25–28; Revelation 6–9).
7. DESIRE OF ALL NATIONS: While some view this phrase as referring to Jerusalem (see Ezra 6:3–9), it seems preferable to view it here as a reference here to the Messiah, the deliverer for whom all the nations ultimately long.
I WILL FILL THIS TEMPLE WITH GLORY: God had filled the first temple built during the days of Solomon with His Shekinah glory, but there no evidence from Scripture that indicates His glory ever did come to Zerubbabel's temple. Nor can this glorification refer to Christ's physical presence in Herod's temple, because the events of verses 6–9 cannot be accounted for historically. The context speaks of the establishment of God's earthly, Davidic, millennial kingdom and His presence in the temple during that kingdom (see Ezekiel 43:5).
8. THE SILVER IS MINE, AND THE GOLD IS MINE: God reassured the people, who were economically destitute, that He was the possessor of all things.
9. THIS LATTER TEMPLE: Haggai's audience might have thought that the temple they were building was "the latter temple," but as we've seen, the Lord's prophecy went far into the future. This temple would vastly surpass any man-made structure in glory, and the nations of the earth would come to it for worship.
I WILL GIVE PEACE: This peace is not limited to the peace that God gives to believers (see Romans 5:1), but looks ahead to that ultimate peace when Christ returns to rule as the Prince of Peace on the throne of David in Jerusalem (see Acts 2:30).
## UNLEASHING THE TEXT
1) Why did the Jews stop building the temple in the first place? How did that gradually grow into complacency over time?
[Your Response Here]
2) God does not dwell in any physical structure, so why was He angry the people had not rebuilt His house? What were the larger issues involved?
[Your Response Here]
3) How did the Lord get His message across to His people? What circumstances did He use? What people did He use? How does He do similar things today?
[Your Response Here]
4) What does it mean to be strong? What does it mean to not fear? How are these things done voluntarily?
[Your Response Here]
## EXPLORING THE MEANING
**_It is wise to consider your ways._** The Lord sent His prophet Haggai to bring one clear message to His people: obey God even when the circumstances are difficult. The prophet proclaimed, "Consider your ways!" (Haggai 1:5). This command might be translated, "Set your mind on your way of life." In other words, it was a call for the people to consider their ultimate priorities. In effect, God proclaimed, "Make _your_ priorities _My_ priorities!"
This is an important discipline that God's people need to exercise on a regular basis. The priorities and perspectives of the world are _not_ God's views, yet they have a way of seeping into our thinking without our even being aware of it. As Christians, we need to reassess our priorities and views daily, even moment by moment, to ensure that we are thinking the way God thinks. We do this by spending time daily studying and meditating on His Word, by seeking His wisdom through prayer, and by fellowshipping with other believers.
Paul warns his readers that this process is vital if we are to understand God's will for our lives. "I beseech you therefore, brethren, by the mercies of God, that you present your bodies a living sacrifice, holy, acceptable to God, which is your reasonable service. And do not be conformed to this world, but be transformed by the renewing of your mind, that you may prove what is that good and acceptable and perfect will of God" (Romans 12:1–2). Christians cannot hope to understand God's perfect will unless they are constantly renewing their minds by the Word of God. If we forget to consider our ways, we will wind up conformed to this world.
**_God may use obstacles and setbacks to get your attention._** The Jews living in Jerusalem had returned to God's chosen city in order to obey His commands. They had reestablished His prescribed worship practices and started rebuilding the temple and walls. Yet as time went along, they found themselves suffering hardship. They were faced with drought and famine, nothing seemed to prosper, and everything they undertook was met with frustration and failure. Surely this couldn't be God's will for His people who were doing His work!
But that, of course, was exactly the case. The problem was that the people were not doing _God's_ work but their _own_ work. They were pursuing their own goals, and the Lord needed to send hardship into their lives to get their attention. This is the flip side to a principle we considered previously that Christians should expect opposition when doing the Lord's work. The caveat to that principle is that we should not ignore opposition in case the Lord is using it to get our attention. And, in this study, we see how He sometimes does that.
This principle holds as true for us today as it did for the people in Haggai's day. When God's people refuse to share His priorities and to view the world from His perspective, He will send hardship, frustration, and obstacles intended to make us stop and look up. The great irony is that when the people stopped pursuing God and focused on their own needs, their own needs were not met. The more they put in their own pockets, the less they had for themselves. Likewise, when we turn away from pursuing Christ, we are consumed by the love of the world, which can never satisfy. True satisfaction can only be found in living for Jesus. So we must always consider our ways and renew our godly perspective, for the Lord might send chastening to us in order to keep us from conforming to the world.
**_There awaits a future glorious kingdom._** The Jews of Haggai's day were discouraged. They had returned from exile and labored diligently on the temple. Even when they were distracted from their task and the prophet had to rebuke them, they responded faithfully and resumed the Lord's work. However, with the finished temple now before them, they wept. The temple was small, and the few in the crowd who were old enough to remember Solomon's temple—destroyed seventy years earlier—were dismayed by the comparison. For many of the Israelites, all that was lost in the exile was finally starting to sink in.
The Lord used Haggai to encourage this beleaguered group. He pointed them to the future—to a time when a new temple would stand in Jerusalem. The very spot where they were now weeping would one day be the site of a temple the likes of which the world has never seen. It would be so glorious that the nations of the world would bring their gold and silver to Jerusalem, and the people would worship there (see Haggai 2:6–8). But the glory of this temple, which is also described in Ezekiel 40–48, is not found in the gold and the silver, or even in the people. Rather, Haggai said, from this temple the Lord will give peace to the world.
We find ourselves in a similar place as the Jews in Haggai's day. The gospel has gone forth, and there are Christians around the world. Nevertheless, the world is dominated by wars and rumors of wars. Injustice is the norm, and hunger and poverty are rampant. Yet we, too, can be encouraged by the promise of God that Jesus is going to return to earth and fulfill every one of His promises. We know that He will establish His glorious kingdom on earth, and from that temple God will return peace and order to the world.
## REFLECTING ON THE TEXT
5) What exactly is complacency? What causes it and what are its results? How can a Christian guard against it?
[Your Response Here]
6) What does it mean to consider one's ways? How is this done, in practical terms? Why is it so vital for Christians to do?
[Your Response Here]
7) Why do people often lose sight of the Lord's priorities in the midst of their busy lives? How do the demands of daily life crowd out concern about the things of God? How should Christians guard against this?
[Your Response Here]
8) Why were the Jews distressed when they saw their new temple? How did Haggai encourage them? How are Haggai's words encouraging to us today?
[Your Response Here]
## PERSONAL RESPONSE
9) Are you facing hardship or obstacles in your life at present? Are they in opposition to the Lord's work or the Lord's attempt to get your attention? Explain.
[Your Response Here]
10) What priorities and perspectives do you share with God? Which priorities or perspectives are more like those of the world? What will you do this week to renew your mind?
[Your Response Here]
# **5**
# OBEDIENCE OVER OBSERVANCE
# _Zechariah 7:1–8:23_
## DRAWING NEAR
What are some holidays or other observances that today have lost connection to what they were originally intended to celebrate? Why does this happen over time?
[Your Response Here]
## THE CONTEXT
Zechariah was a priest who—much like the prophet Haggai—traveled back to Judah with the first wave of exiles under the leadership of Zerubbabel. Tradition holds that he was also a member of the Great Synagogue, a council of 120 men originated by Nehemiah and presided over by Ezra. (This council later developed into the ruling elders of the nation, called the Sanhedrin.) He is occasionally referred to as the son of his grandfather, with whom he traveled to Judah, so it is thought that his father may have died when Zechariah was young.
Zechariah was a contemporary of Haggai and began his ministry two months after Haggai gave his first prophecies. The Lord used Haggai to begin a revival, and He subsequently used Zechariah to keep it going strong. Following the time of Zechariah, the Lord would raise up one final prophet named Malachi to rebuke and condemn the exiles for the abuses they committed during the period of Nehemiah's absence (c. 433–424 BC). After Malachi, prophecy from the Lord would fall silent for 400 years, until a man named John the Baptist arrived on the Judean scene and "came baptizing in the wilderness and preaching a baptism of repentance for the remission of sins" (Mark 1:4).
The Lord used Zechariah during a crucial time in His people's history to bring a rich outburst of promise for the future and sustain the faithful remnant. But first He asked His people to examine their hearts and motives. Fasting and sacrifice have a place, but as should be clear by this point in the study, God is far more interested in something else: _obedience_.
## KEYS TO THE TEXT
Read Zechariah 7:1–8:23, noting the key words and phrases indicated below.
> _A Q UESTION OF RELIGIOUS OBSERVANCE: Now that the temple is being rebuilt, some Jews travel to Jerusalem to seek the Lord's will about whether they should continue their annual fasts._
1. IN THE FOURTH YEAR OF KING DARIUS: Zechariah was living in Jerusalem at the same time as Haggai and had joined Zerubbabel in the first wave of exiles returning to Judah. He gave the prophecies of this chapter in November–December 518 BC, about two years before the temple was completed.
2. TO THE HOUSE OF GOD: While the phrase "house of the LORD" is used of the temple some 250 times in the Old Testament, nowhere else does Bethel (which means "house of God") refer to the temple. The word is thus best viewed as a reference to a city and not the temple. These men came _from_ rather than _to_ Bethel, a town twelve miles north of Jerusalem. Since the return from Babylon, the Jews had rebuilt and reinhabited that city.
3. SHOULD I WEEP: These men had journeyed from Bethel to Jerusalem to ask the priests whether they needed to continue observing a regular schedule of fasts. The only fast required by God's Law was the Day of Atonement (see Leviticus 23:27), though God called for other occasional fasts (see Joel 1:12–14). The Jews remembered the fall of Jerusalem by four fasts during the fourth, fifth, seventh, and tenth months. Because the temple had been burned in the fifth month (July–August), that fast was considered the most serious and, thus, the delegation uses it as the test case. They had kept this wailing and fasting for "many years," but now it seemed a wearisome ritual in light of their present prosperity.
> _G OD'S FIRST RESPONSE: In the first of four responses to the question, the Lord asks the people to examine their hearts. What is their real reason for fasting?_
5. SEVENTH MONTHS: This fast mourned the death of Gedaliah, the governor appointed by Nebuchadnezzar (see 2 Kings 25:22–26) after the fall of Jerusalem in 586 BC.
DID YOU REALLY FAST FOR ME: The Lord sent a total of four responses to the question of fasting, with this first one being a rebuke. The Jews, He pointed out, were not actually fasting from repentant sorrow but only out of self-pity (see Isaiah 58:3–9).
6. DO YOU NOT EAT AND DRINK FOR YOURSELVES: The Lord was asking His people to truthfully examine their hearts before Him, laying bare the true motives behind both their fasting and their eating. If their eating was for their own pleasure, the fasts were likely also driven by selfish motivations. If the people had genuinely repented of the idolatry that led them into captivity, they could stop their regimen of fasting, as the temple was now being rebuilt. But if their fasting was not from true repentance, it had been a waste of time.
7. SHOULD YOU NOT HAVE OBEYED: God was stating that if the people of Judah had obeyed His Word in the first place, there would never have been any call for their self-imposed fasts. The important matter was not ritual but obedience. In the past it had been obedience to God's Word that brought great joy, peace, and prosperity to Israel during the time of David and Solomon. If the present generation in Zechariah's time were to substitute ritual for obedience, they would lose the joy, peace, and prosperity they were enjoying. God wanted them to show their repentance through obedience, not through pious religious observances.
THE SOUTH AND THE LOWLAND: A reference to the area south of Beersheba and the Mediterranean coastal plain, encompassing the land from south to west.
> _G OD'S SECOND RESPONSE: The Lord now moves to more practical matters in His answer, calling His people to concern themselves with obedience more than with fasting._
9. EXECUTE TRUE JUSTICE: The Lord here offered some practical examples of the type of obedience He desired. Widows, orphans, and foreigners who lived among the Jews had no advocate to plead their cause and were easy targets for political and financial oppression. God instructed his people to show mercy and compassion to these individuals and to treat them as brothers. For the Lord, true justice was not influenced by a person's wealth or status, nor was it guided by any political agenda. Those who looked for every opportunity to advance themselves would only end up planning evil against others, hoping to maneuver themselves ahead. The Lord said to repent of such things, and they would have no call to be fasting.
11. BUT THEY REFUSED TO HEED: The Lord next pointed out that the Jews' forebears had been told these exact same things—yet they had deliberately and steadfastly refused to listen. The implication of this was a warning to the present generation not to repeat the stubbornness of the previous generation but to yield their hearts and minds in obedience to God's Word. They were to demonstrate their godliness through loving obedience rather than outward displays of piety.
12. THUS GREAT WRATH CAME: The Holy Spirit had served a vital function in the revelation and inspiration of God's Word through human authors, but the people of Judah and Israel had not listened to them. The Lord was gently reminding His people that since He had not spared previous generations from stern discipline for their stubborn sinfulness, He would not spare them either if they refused to listen.
13. I WOULD NOT LISTEN: The previous generation had plugged their ears and hardened their hearts deliberately in order to not hear the Lord's voice, so He responded by doing the same to them. When they cried out for deliverance from captivity, He did not answer.
14. I SCATTERED THEM: This of course refers to the captivity and dispersion of the people, in addition to the desolation of the land in their absence.
> _G OD'S THIRD RESPONSE: The Lord now turns His eyes forward to describe a future time of blessing that will extend far beyond the Jews' immediate concerns._
8:2. THUS SAYS THE LORD OF HOSTS: This is the Lord's third response to the Jews' question concerning their fasting. Zechariah would now contrast Israel's past judgment with the promised future restoration. In light of past captivity, the nation was to repent and live righteously; in light of promised future blessings, Israel was to repent and live righteously.
I AM ZEALOUS FOR ZION WITH GREAT ZEAL: This strong language, emphasizing the Lord's zeal with a triple repetition, expressed the idea that God could not bear the estrangement from His chosen people that had been brought about by their sin. He ached and yearned to be reconciled with them and to reveal His glory among them once again. His love for Israel was so great that He would come in full presence to them again and dwell with His people.
3. I WILL RETURN TO ZION: This promise had many levels. Most immediately, the Lord promised that He would establish His temple once again in Jerusalem. (Zion was the mountain on which ancient Jerusalem was built, which became a name for the city.) But the day would also come when God would enter Zion in the physical form of Jesus Christ. Ultimately, Christ will return to Jerusalem on a future day to establish His thousand-year kingdom on earth.
4–5. OLD MEN AND OLD WOMEN . . . BOYS AND GIRLS: The Lord had commanded His people to look after those who were weak and easily oppressed. However, in the coming day there would be no enemy to threaten any of His people, as the Lord Himself would rule over the whole earth, being physically present in Jerusalem.
6. IF IT IS MARVELOUS IN THE EYES OF THE REMNANT: That is, the Jews of Zechariah's day might have thought these promises sounded marvelous or hard to believe, but their lack of belief and understanding did not limit God in any way. There is nothing too hard for the Lord, and He always keeps His promises—even those that seem far-fetched (see Matthew 1:23).
8. THEY SHALL DWELL IN THE MIDST OF JERUSALEM: The context assures that this return speaks of a worldwide regathering at the Second Advent of Christ. The return from Babylon cannot be in view, as Israel had not been scattered to the west until the diaspora engineered by the Romans during the first century AD.
> _G OD'S FOURTH RESPONSE: The Lord continues to expound on the countless blessings He intends for His people. In the end, His answer to their question is simple: obey Me._
9. LET YOUR HANDS BE STRONG: Here again the Lord commands His people to be resolute in their obedience. Strength of character and faith are signs of devotion to the Lord, and here that devotion was seen specifically in the rebuilding of His temple.
9. BY THE MOUTH OF THE PROPHETS: This refers to Haggai and Zechariah, and there were possibly some other non-writing prophets included as well.
10. BEFORE THESE DAYS: Zechariah recalled the immediate years prior to 520 BC (described in Haggai 1:6–11) when the people's hassles and intrigues with the Samaritans and their love of ease and comfort created an indifference in them toward building the temple. As we have seen, this resulted in divine punishment in the form of great economic distress. However, since the people had again started to build the temple, God would not treat the people as He had those described in this verse.
12. THE SEED SHALL BE PROSPEROUS: These promises of blessing applied in some measure to the Jews of Zechariah's day, but they also pointed to the future day of Christ's millennial kingdom, when God will restore the nation of Israel to full fellowship and reconciliation.
15. DO NOT FEAR: Notice how frequently God gave this command in Zechariah's prophecies. Fear is the enemy of faith and causes a person to doubt that God is in control in the face of overwhelming situations. The Lord was reiterating the important truth of His absolute sovereignty over all events, as well as His unshakable commitment to show forth His blessings and faithfulness to His people. This same truth is vital for Christians of all ages, and we are commanded by God to resist fear through faith in His sovereignty.
16. THESE ARE THE THINGS YOU SHALL DO: As always, the promised blessings were connected with obedience to God's righteous standards. Such obedience can only be brought about by the power of the Spirit in the life of a person who has been transformed by God's grace through faith. These standards are reminiscent of Psalms 15:1–5; 24:4; and Proverbs 6:20–22.
17. ALL THESE ARE THINGS THAT I HATE: As God's people learn to share His priorities, they will grow to hate lies and evil as much as the Lord does.
19. THE FAST OF THE FOURTH MONTH: This was the Lord's fourth and ultimate answer to the Jews' question concerning their self-appointed fasts: turn the fasts into feasts of joy to rejoice over God's blessings and faithfulness. They were to remember, whether in feasting or fasting, to "love truth and peace," for obedience was better than any form of religious observance.
FAST OF THE TENTH: Two other fasts were held in addition to the fasts of the fifth and seventh months (see Zechariah 7:3–5). In the fourth month, the Jews commemorated the breaching of the wall of Jerusalem. In the tenth month, they remembered the beginning of the final siege of Jerusalem, which began in 588 BC.
22. MANY PEOPLES AND STRONG NATIONS: Israel's restoration in millennial glory would be the means of blessing the entire world. Gentiles from around the world would make a pilgrimage to Jerusalem to entreat the Lord. This signifies salvation of people from all over the world during the kingdom, fulfilling the words of Psalm 122.
23. IN THOSE DAYS: In the days in which the messianic kingdom on earth is inaugurated, the Jews will truly be God's messengers as originally intended and will bring multitudes to Christ. The ten-to-one ratio represents a vast number of Gentiles who will come. The Messiah, in the midst of millennial Israel, will be the attraction of the world. When these Gentiles see the Jews being so blessed in their kingdom, they will demand to go to meet the Savior King.
## UNLEASHING THE TEXT
1) Why had the Jews observed this cycle of fasting for so many years? What did it represent in their view? What did it represent in God's view?
[Your Response Here]
2) What is the difference between repentance of sin and sorrow over the consequences? How does a Christian distinguish between them? What is the outward result of each?
[Your Response Here]
3) Define the following in your own words, giving practical examples of each.
Mercy:
[Your Response Here]
Compassion:
[Your Response Here]
Oppression of widows or the fatherless:
[Your Response Here]
Oppression of foreigners or the poor:
[Your Response Here]
Planning evil in your heart:
[Your Response Here]
4) In what sense are religious observances a waste of time? When are they valuable? How can a Christian distinguish between these two situations?
[Your Response Here]
## EXPLORING THE MEANING
**_Obedience is better than sacrifice and fasting._** The Jews traveled to Jerusalem to seek the Lord's guidance concerning an annual series of fasts they had observed since the beginning of their captivity seventy years earlier. The fasts brought a degree of hardship on the people, as they involved times of mourning and self-denial as well as abstinence from food. The fasts were intended to represent the people's mourning over the Lord's judgment when He permitted the Babylonians to destroy Jerusalem, but now that the city was being rebuilt and temple worship was being reestablished, they wondered if God would object if they ended these observances.
As we have seen, the Lord was more concerned about the people's obedience than with their outward displays of religion. It is true that the Jews demonstrated a desire to obey by asking the Lord's guidance on the matter, but God probed deeper into their hearts, asking them where their deepest priorities lay. Were they truly committed to obeying His Word, or were they mostly concerned with their own comfort? Would they serve Him with whole hearts, as their predecessor David did? Or would they be only halfhearted, like Solomon, and eventually wander back to foreign gods? Were they more concerned with obedience to God or with outward demonstrations of religious regulations?
The Lord had addressed these questions in more detail through His prophet Isaiah. The people, He said, had performed fasts and acts of penitence on the outside, while their hearts were still wicked on the inside. The Lord wanted full obedience to His Word, not mere showmanship during occasional religious observances. "Is this not the fast that I have chosen," the Lord asked, "to loose the bonds of wickedness, to undo the heavy burdens, to let the oppressed go free, and that you break every yoke?" (Isaiah 58:6). Samuel gave Saul similar counsel: "Has the LORD as great delight in burnt offerings and sacrifices, as in obeying the voice of the LORD? Behold, to obey is better than sacrifice, and to heed than the fat of rams" (1 Samuel 15:22). Obedience to God's Word is the only sacrifice the Lord wants from His people.
**_The Lord yearns for our fellowship._** The Lord told His people, "I am zealous for Zion with great zeal; with great fervor I am zealous for her" (Zechariah 8:2). Many modern Bibles translate "fervor" as "wrath," which captures the burning passion the Lord was expressing. "I burn for you with a jealous fire," the Lord was effectively saying, "a zealous and jealous love that will tolerate no competition." The Lord yearned so deeply for the fellowship of His people that it was like a consuming fire, destroying anything that prevented their full reconciliation.
It is interesting that the Lord referred to the zeal of His love in this context. The Jews had been zealous themselves in maintaining a cycle of fasts over a period of seventy years during their captivity. "Should I weep in the fifth month and fast as I have done for so many years?" they asked (Zechariah 7:3). Their question implied that they had not permitted anything to prevent them from keeping this religious observance, but the Lord called them to search their hearts and question why they were truly fasting. Were they mourning the loss of God's close fellowship, which they had once enjoyed as His chosen nation? Or were they mourning over their own suffering and misfortune? The two were not the same!
The Lord wanted His people to be zealous in seeking His face and entering His holy presence. Now, this might well be associated with certain spiritual activities, such as commemorating the Lord's Supper as Jesus commanded (see Luke 22), but those activities themselves are hollow and meaningless if one's heart is not fully committed to living in fellowship with God. The Lord has a burning zeal for our fellowship, for our company, and He wants us to have that same zeal for Him.
**_Jesus is coming again, and the day may be very soon._** The Lord's prophecies of blessing in Zechariah 7–8 had some immediate application to the Jews of Zechariah's day, but they also looked forward to future events. For example, His promise to return to Zion (8:3) indicated that He would rebuild His temple in Jerusalem during their day, yet it also looked ahead to the wonderful day when He would enter Jerusalem in physical form through His Son, Jesus Christ. And the promise has yet another level of fulfillment still to come, when Jesus returns to earth to establish His thousand-year earthly kingdom centered in Zion.
The Bible is clear that Jesus is coming again—but this time not as a helpless babe lying in a manger. He will return in His strength and glory, and He will establish a kingdom that rules all the nations of the earth with a "rod of iron" (Psalm 2:9; Revelation 2:27; 19:15). In that day, the nation of Israel shall once again be the seat of blessing to all peoples (see Zechariah 8:13), while the enemies of God and of His people shall be broken without remedy.
We live in the age of grace, when God's gift of salvation is readily available to anyone through faith in the gospel. But the day is coming when the Lord "will gather out of His kingdom all things that offend, and those who practice lawlessness, and will cast them into the furnace of fire. There will be wailing and gnashing of teeth" (Matthew 13:41–42). The day of Christ's return is drawing close. For the Christian, this means to "look up and lift up your heads, because your redemption draws near" (Luke 21:28). But for those who have not received salvation and eternal life through Jesus Christ, the Lord calls out, "Behold, now is the accepted time; behold, now is the day of salvation" (2 Corinthians 6:2).
## REFLECTING ON THE TEXT
5) What things does God hate? Why does He hate them? How does your own list of "hates" compare with God's list?
[Your Response Here]
6) The Jews asked a fairly straightforward question, but the Lord's answer was lengthy and complex. What deeper truth was He trying to dig out?
[Your Response Here]
7) What did God mean when He said, "I was zealous for Zion with great zeal; with great fervor I am zealous for her" (Zechariah 8:2)? How does God's zeal for your fellowship compare with yours for His?
[Your Response Here]
8) What are some signs that Jesus' return is imminent? How does this affect your daily life? How might your priorities change in the coming week if you took this to heart?
[Your Response Here]
## PERSONAL RESPONSE
9) Does your daily life reflect true obedience to God's Word, or do you tend to "go through the motions" of religious observances? What area of obedience might the Lord be calling you to this week?
[Your Response Here]
10 Have you received God's gift of salvation through Jesus Christ? If not, what is preventing you from doing so right now?
[Your Response Here]
# **6**
# THE SECOND WAVE OF EXILES
# _Ezra 7:1–8:36_
## DRAWING NEAR
What are some "shortcuts" that people take when it comes to following the rules? Why is it often so hard for people to obey the letter of the law?
[Your Response Here]
## THE CONTEXT
Ezra was descended from a long line of priests that dated all the way back to Aaron himself, who was the first high priest of Israel (see Exodus 28:1). The Lord had appointed Aaron, Moses' brother, to take on that sacred role, and He had stipulated that only men descended from Aaron should ever follow in his steps. In this alone, Ezra would have been a good choice to lead the Jews back to Jerusalem, as it would have indicated that the people were reestablishing the Lord's prescribed worship in the temple.
However, that was not Ezra's sole qualification as a leader. He also had spent his life studying God's Word and memorizing it—and, more important, he had spent his life _obeying_ it. He had become an expert in the law of God, a teacher of teachers, and a teacher by example as well as by word. The Lord had appointed him to take the lead on this important expedition because he was well qualified in obedience as well as knowledge.
In previous studies, we saw how many of the Jews had returned to Jerusalem during the "first wave" of exiles under the leadership of Zerubbabel (c. 538 BC). Now, Ezra himself would gather together the next wave to return to the people's homeland (c. 458 BC). As before, the trip would be long and fraught with danger, but the king had issued a decree giving his blessing, and he had loaded the people with gold and silver to fund the work. Everything was lined up and ready to go—but then Ezra discovered that some people were missing!
In this study, we will look more closely at the life of Ezra and see what made him a godly leader. We will also discover what it means to be a follower of God who does not cut corners.
## KEYS TO THE TEXT
Read Ezra 7:1–8:36, noting the key words and phrases indicated below.
> _T HE LORD CALLS EZRA: Ezra is a priest, a scribe, and an expert in the Word of God. The Lord calls him to move to Jerusalem, and he obeys God's command._
1. AFTER THESE THINGS: Zerubbabel led the Jews in rebuilding the temple in Jerusalem, which was finally completed in 515 BC during the reign of King Darius of Persia. The events in Ezra 7 take place many years later, under the reign of King Artaxerxes, when Ezra led a second wave of Jews back to Jerusalem. During the sixty-year interval between Zerubbabel and Ezra, the events of the book of Esther took place in Persia.
6. A SKILLED SCRIBE IN THE LAW OF MOSES: Ezra was descended from a long line of priests, which he traced back all the way to Aaron, the brother of Moses. He was also a scribe, one who studied and transcribed the Law of Moses. This dual role made him a valuable asset to the returning Jews, who were in desperate need of instruction concerning God's Word. Tradition claims that Ezra could write out the entire law from memory.
ACCORDING TO THE HAND OF THE LORD HIS GOD UPON HIM: Once again, we are reminded that all the events in the book of Ezra took place according to the sovereign will of God. Without God's constant intervention and protection, Jerusalem could never have been rebuilt—and for that matter, the Jews themselves would have remained in exile.
7. THE PRIESTS, THE LEVITES . . . AND THE NETHINIM: The priests were descendants of Aaron who served as the nation's spiritual leaders and oversaw the worship and sacrifices. The Levites were members of the priestly tribe of Levi who assisted the priests. The Nethinim were not members of the priestly class but were descendants of the Gibeonites. They were responsible for looking after the temple building. "Joshua made them woodcutters and water carriers for the congregation and for the altar of the LORD" (Joshua 9:27).
8. EZRA CAME TO JERUSALEM: The second wave of exiles' four-month journey from Babylon to Jerusalem, covering almost 1,000 miles, started in March–April and ended in July–August.
10. SEEK . . . DO . . . TEACH: Ezra set an example of the process by which a person grows to spiritual maturity and godliness. First, he "prepared his heart" in a fixed determination to understand the Word of God, probably by spending his youth in the study and meditation of the Scriptures. This process naturally fit together with immediate application, as Ezra strove to do what God's Word commanded, putting it into practice in his own life first. This, finally, led to a natural role as a teacher of others—for one cannot teach what he has not first practiced. Ezra's life of obedience enabled him to teach God's Word; but more important, "the good hand of his God upon him" (verse 9) strengthened and equipped him for that role.
11. COPY OF THE LETTER: The original was usually kept as an official record. This particular letter was addressed to Ezra because the decree recorded therein was the critical administrative document. In essence, it authorized the document into Ezra's hands so he could carry it and read it to its intended audience.
EZRA THE PRIEST, THE SCRIBE, EXPERT: This is a powerful testimony to both Ezra's character and his grasp of God's Word. He had studied and memorized the Lord's commands to the point that he had become a teacher of teachers. It is also significant that he used these roles as his description, rather than "envoy of the king" or some other title.
> _T HE KING'S DECREE: King Artaxerxes writes a decree giving the Jews permission to return to Jerusalem and rebuild God's temple._
12. ARTAXERXES, KING OF KINGS: It was true that Artaxerxes ruled over other kings, but Jesus Christ is the true King of kings and Lord of lords (see Revelation 19:16). Jesus alone can make that claim, for He will rule over all earthly powers in His coming kingdom (see 11:15).
13. I ISSUE A DECREE: Decrees were commonly written as official documents and then reiterated in personal letters that gave an emissary authority in his travels. Ezra would have read this letter to the Jews in Jerusalem to demonstrate that he had the king's backing. This is a remarkable decree that evidences God's sovereign rule over earthly kings and His intent to keep the Abrahamic, Davidic, and New Covenants with Israel.
14. SEVEN COUNSELORS: This number was according to the Persian tradition.
15. FREELY OFFERED TO THE GOD OF ISRAEL: It is interesting that King Artaxerxes chose to make a freewill offering to the Lord, even though he did not worship Him. The king may have been hoping to appease "the God of Judah" in hopes of avoiding future trouble from that region, as the people during that time thought that various gods ruled in specific geographical regions. The fact that Ezra was carrying a bounty of the king's gold and silver would make his journey to Judah that much more dangerous.
16. THE FREEWILL OFFERING OF THE PEOPLE AND THE PRIESTS: The freewill gifts of God's people were an important part of the rebuilding effort in Jerusalem, both for the temple and the walls.
17. NOW THEREFORE: The royal decree protocol recorded in the opening words of verses 13 to 16 lead up to the section introduced by these words.
21. THE GOD OF HEAVEN: The king's use of this phrase suggests that he had some insight into the true nature and character of God, but his repeated references to "your God" and "the God of Jerusalem" indicate that he had probably not bowed himself before the true God of all creation. Knowing who God is does not make a person a Christian—one must also submit to Him as the only God and accept His free gift of salvation. The king's question, "Why should there be wrath against the realm of the king and his sons?" (verse 23), suggests that he viewed the Lord as yet another "god" whose power was limited geographically. Again, his gifts probably were motivated by the hope of appeasing this unknown god of the Jews.
22. ONE HUNDRED TALENTS: This represents nearly four tons in weight; "one hundred kors" is approximately 750 bushels; and "one hundred baths" is 600 gallons.
25. SET MAGISTRATES AND JUDGES . . . AND TEACH: The king had turned to Ezra in a demonstration of administrative trust and granted him permission to appoint magistrates and judges for the region. The effect of this decree gave a measure of autonomy to the Jews in Judah—exactly what their enemies were striving to avoid. Yet this element of spiritual shepherding was exactly what Ezra had a heart to do. He knew the Jews returning from captivity needed instruction in the Word of God, and this became the focus of his ministry.
27. WHO HAS PUT SUCH A THING AS THIS IN THE KING'S HEART: Once again, Ezra recognized the king's generosity was due to the Lord's intervention. He was grateful to King Artaxerxes for his benevolence, but he gave the final glory to God.
> _P REPARING FOR THE JOURNEY: Ezra gathers together the necessary people and supplies, and they prepare to make the dangerous and lengthy trip to Judah._
8:1. WENT UP WITH ME FROM BABYLON: The list that follows no doubt included those who lived in the surrounding areas. The total number of males in this section was 1,496 plus the men named, so with the addition of the women and children, the number easily approached 7,000 to 8,000. Just as with the first group of returnees, many Jews remained behind in Babylon after this group had departed, content with their established and comfortable lifestyle.
15. THAT FLOWS TO AHAVA: An unknown location where a canal or river flowed into the Euphrates River. This was in Babylon, and Ezra chose it as the place where the returning Jews would render vows for several days in preparation to leave.
THE SONS OF LEVI: That is, men from the tribe of Levi. This tribe had been set apart by God to be the priestly class, and all who acted in a priestly role were descendants of Levi. The high priest was always a descendant of Aaron, a subset of the tribe of Levi, but other Levites also served in other priestly functions. Apparently, no descendant of Levi had joined Ezra for this return to Jerusalem, though others had returned previously with Zerubbabel. Ezra was deeply concerned about this situation, because he had no one qualified to serve in the temple.
17. THEY SHOULD BRING US SERVANTS FOR THE HOUSE OF OUR GOD: To remedy this situation, Ezra took the time and effort to find qualified men from the tribe of Levi. Others might have been tempted to look for some more expedient method—perhaps not being too concerned about meeting the Lord's qualifications for the priestly class—but Ezra was determined to obey the Lord's Word in all things. He especially wanted to obey God's Law in something as important as selecting the nation's spiritual leaders.
18. THEY BROUGHT US A MAN OF UNDERSTANDING: Only thirty-eight Levites were willing to join Ezra in this second return to Jerusalem. However, this remnant of men whom the Lord had raised up had understanding and wisdom and recognized that even the mundane tasks in God's service were of infinitely more value than the most glorified projects of the world. This also helps us recognize the sacrifice that Ezra and Nehemiah paid in leaving behind their great Persian positions to undertake the arduous task of rebuilding Jerusalem.
> _F INAL PREPARATIONS: Before Ezra leads the exiles back to Jerusalem, he holds a fast to pray for God's protection and appoints priests to carry the gifts for the temple._
21. I PROCLAIMED A FAST: The Jews' four-month trek back to Judah would be fraught with dangers, because thieves who robbed people for survival frequented the roads. Even messengers traveled with caravans to ensure their personal safety. However, Ezra and the people did not want to confuse the king about their trust in God's protection, so they entreated the Lord for safety with a prayerful fast. They also wanted their hearts and lives to be pure before Him. God honored their prayer of faith with His protection.
22. AN ESCORT OF SOLDIERS AND HORSEMEN: The presence of the great gift of King Artaxerxes' gold and silver only made the caravan a more attractive target for bandits and enemies. Yet Ezra was "ashamed to request" an escort from the king because he had previously told the Persian monarch that the Lord would protect the Jewish people. He was concerned about the Lord's glory, not about their safety, and he did not want the Persians to think that his God could not live up to His promises. Later, Nehemiah _would_ accept his king's offer of a military escort due to the same concerns for his safety.
26. I WEIGHED INTO THEIR HAND: Once again, Ezra carefully followed the Word of God as he gave the sacred treasure into the hands of the priests (see Leviticus 3).
28. YOU ARE HOLY TO THE LORD: The Lord had set apart the tribe of Levi to be His holy priesthood, and as such they were consecrated to His service. He had said to Moses, "Take the Levites instead of all the firstborn among the children of Israel, and the livestock of the Levites instead of their livestock. The Levites shall be Mine" (Numbers 3:45). In the same way, the gold, silver, and other precious objects had been set aside for the Lord's use, and in that sense they were holy in His sight.
31. THE HAND OF OUR GOD WAS UPON US: Once again, Ezra reminded his readers that the Lord is completely faithful to His promises and His people. The Jews had no need of a military escort in spite of the fact that they were carrying a king's ransom in the caravan. The important fact for Christians to understand is that the Lord's hand is always on His children, and He must be the sole source of our security.
31. TWELFTH DAY OF THE FIRST MONTH: The twelve-day delay occurred because of the three-day delay searching for more Levites and the fast.
36. THEY DELIVERED THE KING'S ORDERS: The plural of the word _orders_ may account for the change in terminology. This would include the decrees plus other orders in the official correspondence that Artaxerxes gave to Ezra to deliver to support the Jews and their building of the temple.
## UNLEASHING THE TEXT
1) According to Ezra 7:10, what three things characterized Ezra's life? How are these things done? Why are they important?
[Your Response Here]
2) How had Ezra become an expert in the Word of God? What effect did that expertise have on his ministry? What impact did it have on the people around him?
[Your Response Here]
3) Why did Ezra feel ashamed to ask the king for a military escort? What motivated his decision? What does this reveal about his character?
[Your Response Here]
4) Why did Ezra take time to find some members of the tribe of Levi before leaving for Jerusalem? What did this demonstrate about God's Word? What did it demonstrate about his level of obedience to God's Law?
[Your Response Here]
## EXPLORING THE MEANING
**_There are no shortcuts in obedience to God's Word._** Ezra made many preparations for the move to Jerusalem. He gathered a large body of Jews to join him, and each of those families made all the necessary preparations involved in making a life-changing move. The king had given Ezra his full blessing on the trip and provided him with a letter of authority to reestablish a Jewish community and rebuild the Lord's temple. The king had also handed Ezra a huge sum of money and treasure, and Ezra probably felt a sense of urgency to get that money where it belonged. A host of people and plans were ready to go. But then Ezra made the startling discovery that there were no members of the tribe of Levi with him on the journey.
Human wisdom would suggest a "work-around" measure at this point. So many people were standing around, waiting to get started, and the king's money was just sitting there waiting for theft. Surely, under such circumstances, prudence would dictate an "ad hoc" alternate plan. But Ezra refused to begin rebuilding the temple without the leadership and assistance of God's selected priests because God's Word commanded it. Ezra may have been caught by surprise, but he knew that the Lord was not. God wanted him to follow His prescribed methods, and He would take care of the timetable.
There are no shortcuts to obeying God's Word. The Lord does not call His people to find "work-arounds" or "emergency interim methods." His Word gives clear guidance in our daily lives, the correct approach to worship and church structure, roles of authority and submission, and much more that often goes contrary to what the world believes today. When it comes to the clear teachings in Scripture, there is no substitute for obedience.
**_The Lord calls us to seek His guidance in every undertaking._** Ezra and the Jews were ready to move forward on their expedition back to Jerusalem. They had spent months preparing for the long journey and made arrangements for a permanent change of address. This was no small undertaking. It represented a complete change of lifestyle for everyone concerned, and it is unlikely that anyone had made the decision on the spur of the moment. Yet when the big day had nearly arrived, Ezra stopped everything and called the people to join him in fasting and prayer to seek the Lord's guidance and blessing.
This might seem like an odd approach. After all, this was no last-minute realization that they had forgotten to seek the Lord's guidance. The people in general, and Ezra in particular, had undoubtedly been in _much_ prayer during the months of preparation. Nevertheless, when everything was prepared and the people were ready to begin, Ezra once again spent time before the Lord. Ezra had made it a habit of his life to bathe every undertaking in prayer. He constantly sought the Lord's guidance to ensure that he was going in the right direction and not overlooking anything important.
This is the pattern for all God's people, for we are to keep entering His presence deliberately and without distraction. Ezra added the element of fasting in order to dedicate the prayer time fully to the Lord, blocking out all the distractions of daily life to concentrate on His voice. The person who seeks the Lord's direction for every step will never go far from His chosen path.
**_Study, obey, and teach._** In Ezra 7:10 we read, "Ezra had prepared his heart to seek the Law of the LORD, and to do it, and to teach statutes and ordinances in Israel." This statement reveals a great deal about the character of the man: he had "prepared his heart," meaning that he had consciously made it a high priority in his life, and he had stuck to that priority for many years. He had sought the Lord by studying His Word, meditating on it daily, and looking for ways to apply it practically in his life. He probably had also taken great pains to memorize it, as history claims that he could write the Law of Moses from memory. However, beyond even these intellectual pursuits, Ezra had consciously applied God's Word to his life, deliberately obeying His commands on a daily basis for a long period of time.
These habits are what equipped Ezra to become a great teacher. A person cannot hope to teach others how to play a musical instrument until he or she has gained some degree of mastery of that same instrument. In the same way, a person cannot teach others the Word of God unless he or she is already living by it. Not every Christian is a gifted teacher, and not all Christians are called to preach the Word from the pulpit, but every Christian _is_ called to teach by example. In fact, some of the most powerful teaching is done simply by living according to God's Word, whether or not any words of explanation are ever given to those watching.
The psalmist summarized this process nicely: "How can a young man cleanse his way? By taking heed according to Your word. With my whole heart I have sought You; oh, let me not wander from Your commandments! Your word I have hidden in my heart, that I might not sin against You" (Psalm 119:9–11). The process illustrated by Ezra is threefold: study God's Word, obey God's Word, and teach God's Word. This is the process all Christians are called to emulate.
## REFLECTING ON THE TEXT
5) If you had been in Ezra's place, what would you have done on discovering that you had no priests with you? How would your response have been similar to or different from Ezra's?
[Your Response Here]
6) What shortcuts do some Christians take when it comes to God's Word? What is the danger of such an approach? What shortcuts have you taken in the past?
[Your Response Here]
7) Why did the people take time to fast and pray before leaving on their long journey? Why are the "little ones" mentioned in this context (see 8:21)? What effect would this action of praying and fasting have on their children?
[Your Response Here]
8) Why is it so important for a Christian to spend time alone studying God's Word? What role does fasting play in this process?
[Your Response Here]
## PERSONAL RESPONSE
9) What are you teaching to people around you each day concerning God's Word and His character? What might a casual observer learn from watching your life?
[Your Response Here]
10) What decisions are you facing at present? In what ways are you submitting to God's will and casting your worries on Him? How can you do that more this week?
[Your Response Here]
# **7**
# UNEQUALLY YOKED
# _Ezra 9:1–10:44_
## DRAWING NEAR
How does our culture tend to view marriage between people? How do you think these views are similar to or different from what God thinks about marriage?
[Your Response Here]
## THE CONTEXT
We now return to Ezra and his fellow Jews as they embark on their journey to Jerusalem. Ezra reports, "The hand of our God was upon us, and He delivered us from the hand of the enemy and from ambush along the road. So we came to Jerusalem" (8:31–32). Ezra immediately set to work to put everything in order: disseminating the money and treasure the king had given him, appointing spiritual leaders, guiding the people in worship to the Lord, and all the other myriad details involved in establishing a new home for himself and others.
Within four months or so, Ezra was able to complete all of these tasks. It seemed the time was right for him to now "buckle down" and get busy with the project of rebuilding the temple. But no sooner had Ezra gotten things in order than he was confronted with an unexpected problem: many of the Jews had married foreign spouses! This must have seemed incomprehensible to Ezra, since that was the very sin that had led the nation of Israel into captivity in the first place.
However, as we have seen, human nature is always prone to wander from the will of God. Solomon wrote, "As a dog returns to his own vomit, so a fool repeats his folly" (Proverbs 26:11). In this lesson, we will see that while the Jews may have gotten a new start, they were still in need of the new covenant and the Messiah to change their hearts. We will also look at marriage from God's perspective and discover that He takes it very seriously.
## KEYS TO THE TEXT
Read Ezra 9:1–10:44, noting the key words and phrases indicated below.
> _R ETURNING TO FORMER SINS: Ezra has just arrived in Jerusalem with a group of returning exiles, and almost immediately the people fall into old patterns of sin._
1. WHEN THESE THINGS WERE DONE: "These things" refers to the implementation of the different trusts and duties that the king had commissioned Ezra to complete. The Jewish people had not been in the land for long when the following events transpired.
THE PRIESTS AND THE LEVITES: As was the case during the Assyrian deportation of Israel and the Babylonian deportation of Judah, the spiritual leadership defaulted along with the people.
HAVE NOT SEPARATED THEMSELVES: The Lord had commanded His people to keep themselves apart from the people of Canaan when they arrived in the Promised Land. In fact, He had even ordered them to make war against many of the pagan nations, destroying their altars and religious sites dedicated to false gods. If the people failed to do this, the Lord warned, they would wind up becoming just like their pagan neighbors—intermarrying and adopting their idolatrous ways (see Exodus 34:10–16; Deuteronomy 7:1–6). Unfortunately, the Israelites did not obey the Lord's commands in this regard, and the Lord sent them into captivity. It was a sad statement on the human condition to find the returning Jews immediately beginning to repeat their former sins. It is compounded by the fact that the priests and Levites—the nation's spiritual leaders—had joined in.
2. TAKEN SOME OF THEIR DAUGHTERS AS WIVES FOR THEMSELVES: Malachi added that some of these people had even divorced their Jewish wives in order to marry foreigners (see Malachi 2:11, 14–16).
THE HOLY SEED IS MIXED WITH THE PEOPLES OF THOSE LANDS: This is a powerful expression that can help us understand why God does not want His people to intermarry with unbelievers. The Lord had chosen Abraham and his descendants to be the race into which He would bring Jesus, the Messiah, His holy and sinless Son. As such, the Lord wanted Abraham's descendants to be set apart as a holy people, distinct from the world around them, rather than blending in. This same principle applies to God's people today, for the Christian church is the bride of Christ. The Lord still demands that His bride keep herself apart from the world.
THE HAND OF THE LEADERS AND RULERS HAS BEEN FOREMOST: Once again, it was bad enough that the people of Judah had fallen immediately back into their former sins, but it was unconscionable and inexcusable that the spiritual leaders were actually leading the way. When the leaders go in the wrong direction, the rest of the people can only be expected to follow.
> _E ZRA GRIEVES: Ezra's immediate response is to fall before the Lord in deep mourning. Yet, as God's chosen leader, he identifies himself with the people of God._
3. I TORE MY GARMENT AND MY ROBE: Ezra's response was a threefold display of the tremendous grief he felt when he learned of the people's sin. He knew that their wanton disregard of God's Word would lead to sorrow at both the personal and national levels—for the sin that brought God's discipline in the past could only bring it again in the future.
4. EVERYONE WHO TREMBLED: There were still some in Judah, including Ezra, who took the Lord's Word seriously. They sat with Ezra until the gathering of the people for the evening sacrifice, during which there was surely public prayer and confession as Ezra fasted, lamented, and prayed in an effort to lead the leaders and people to repent. Their action of gathering with Ezra also made a public statement that they were not joining together with those who were in sin. If God's people refuse to separate themselves from the world, then those who still fear God's Word will separate themselves from them.
6. OUR INIQUITIES . . . OUR HEADS . . . OUR GUILT: Notice how Ezra identified himself with the people who were in sin, even though he had not participated in that sin. He recognized that sin is like leaven: a little bit of it can infect the entire nation. But he was also demonstrating an important element of spiritual leadership by identifying himself with the entire body of God's people and remembering that he, too, was a sinner.
7. WE HAVE BEEN VERY GUILTY: Ezra also understood that God takes sin seriously, and that included the sin of intermarriage and divorce.
8. A PEG IN HIS HOLY PLACE: That is, the remnant of God's people in Judah had been firmly fixed in God's holy place as a peg is firmly fixed in a wall.
9. TO GIVE US A WALL: The Jews as a people group had been scattered all over the Fertile Crescent and were vulnerable to the nations. However, in Judah, with God as protector, they would be safe. The _wall_ did not exclude the walls of Jerusalem yet to be built, but it speaks more broadly of God's provision for protection.
10. WE HAVE FORSAKEN YOUR COMMANDMENTS: This was not a quotation from Ezra of any single text of Scripture but rather a summation of God's commands on the subject.
14. NO REMNANT OR SURVIVOR: Ezra feared that this sin would provoke the ultimate judgment of God and the abrogation of the Lord's unconditional covenants. However, while God judges sin, the coming of the Messiah and Paul's insights on God's continued faithfulness in His promise to the Jews assures us that God's calling of Israel as a beloved people and nation is irrevocable (see Romans 11:25–29).
15. NO ONE CAN STAND BEFORE YOU: All the people were reckoned guilty and had no right to stand in God's presence, yet they would penitently seek the grace of forgiveness.
> _P UTTING AWAY THE WIVES: Ezra's solution to the problem is to call on the people to put away any unbelieving spouses. It is imperative that God's people be separate from the world._
10:1. THE PEOPLE WEPT VERY BITTERLY: When confronted with their sin, the people of Judah did the right thing: they repented and mourned over their wickedness. This is what God desires to see when His people commit sin.
2. SHECHANIAH THE SON OF JEHIEL: This leader, who was not involved in the mixed marriages, was bold and chose to obey God rather than please his relatives.
3. TO PUT AWAY ALL THESE WIVES: Shechaniah called for the people and leaders to accomplish the specific action of divorcing their wives and children and acknowledged that Ezra has counseled a course of action consistent with Scripture (see 2 Chronicles 29:10). We must remember, however, that Ezra is a book of history and records actual historical events. The fact that a person did something in Israel's history does not mean it is a model of God's ideal plan for us to follow, for God hates divorce (see Malachi 2:16).
4. YOUR RESPONSIBILITY: Here again, Shechaniah acknowledged that Ezra was the chief spiritual leader and had appropriate divine authority and the responsibility to take on the execution of this formidable task of dealing with divorces for so many.
6. HE ATE NO BREAD AND DRANK NO WATER: This was an unusually stringent fast, for people generally only fasted from food but not from water. It indicated the depth of Ezra's grief and consternation over this situation, which underscores the seriousness of marrying an unbeliever.
7. THEY ISSUED A PROCLAMATION: A proclamation was delivered orally by a herald and often had the force of law, as did this particular proclamation. Not participating in the assembly, as some might have been tempted to do, meant not just losing your property but also being ostracized from Israel.
8. HE HIMSELF WOULD BE SEPARATED FROM THE ASSEMBLY: The people of Judah needed to separate themselves from anyone who had married an unbeliever, lest the unbeliever in their midst would lead others into idolatry. The people were required to respond within seventy-two hours. Given that only the territories of Judah and Benjamin were involved, the greatest distance for the proclamation to travel would have been no more than forty to fifty miles.
> _T HE PLAN PUT INTO EFFECT: The leaders of Judah agree to Ezra's plan, but they have an issue with how it will be carried out. So they suggest a way to make it more feasible._
13. THERE ARE MANY PEOPLE: This statement demonstrates how widespread the sin was among the people. Due to the large number of people to be processed (in addition to the heavy rains), the whole operation could go on for a long time. So the people made an administrative suggestion for dealing with the magnitude of the problem: for each unlawful marriage, a questioning or court session could be conducted locally.
15. JONATHAN . . . AND JAHAZIAH . . . OPPOSED THIS: It is unclear whether these two, along with Meshullam and Shabbethai, opposed the delay in dealing with the situation or opposed dealing with the sin at all. It was, however, a good plan and brought about a reasonably fast resolution. It took three months for the people to rectify the situation in all cases, after which they were prepared to celebrate Passover.
17. THEY FINISHED QUESTIONING ALL THE MEN: This verse indicates that Ezra and the other spiritual leaders considered each case individually, and it is likely that each foreign wife was given an opportunity to openly forsake her false gods and embrace the only true God. Such repentance would change the marriage from an unequal yoking to a proper marriage between believers. The list of those who divorced their pagan wives seems rather short, considering the frequent references to "many of us" (see verse 13), so it is possible that at least some of the unbelieving spouses repented of their pagan ways and embraced God's truth.
18. THE SONS OF JESHUA THE SON OF JOZADAK: At the head of the list of those who had intermarried were the descendants and other relatives of the high priest who first returned with Zerubbabel and led in the temple reconstruction. They set the example for all the people in giving the appropriate trespass offering (see verse 19).
20. ALSO OF THE SONS OF: Given the fact it took three months to resolve the situation, this list of 113 men could represent only those in leadership. There were apparently more violators among the people, and even though Ezra and the Jewish leaders dealt with the problem directly, it would eventually reappear (see Nehemiah 9–10).
44. BY WHOM THEY HAD CHILDREN: An appropriate provision was doubtlessly made for the divorced wives and the children.
## UNLEASHING THE TEXT
1) Why did Ezra respond so vehemently to the Jews' marrying foreigners? Why was that such a big problem for God's people?
[Your Response Here]
2) Why had the Jews married foreigners in the first place? Considering that they had just recently moved to a new home far from Persia, what factors might have made such marriages seem acceptable to them at the time?
[Your Response Here]
3) Why does God condemn marriage between a believing and an unbelieving spouse? What dangers does it pose to the believer? To the children? To the church as a whole?
[Your Response Here]
4) How do some Christians justify marriage to unbelievers as "no big deal"? What arguments do they use? What does God's Word say on the matter?
[Your Response Here]
## EXPLORING THE MEANING
**_God hates divorce._** Ezra's solution to the dire problem of intermarriage was to call on the people to divorce their unbelieving spouses. His reason for such an extreme approach was that the sin was extreme. It was intermarriage with Israel's pagan neighbors that had led God's people into the deadly sin of idolatry in the first place. Desperate circumstances sometimes call for desperate measures, and this solution may well have been the best.
Yet we must also understand that divorce is _not_ God's desire for His people. Indeed, God uses strong language on the subject. " 'Take heed to your spirit, and let none deal treacherously with the wife of his youth. For the LORD God of Israel says that He hates divorce, for it covers one's garment with violence,' says the LORD of hosts. 'Therefore take heed to your spirit, that you do not deal treacherously' " (Malachi 2:15–16). Notice the words that God used to describe divorce: _treachery_ and _violence_. This is because when Christians divorce their spouse, they are treacherously betraying the vows they made before the Lord, and they are violently tearing asunder a couple who are one flesh.
Jesus expounded further on this subject. "Have you not read that He who made them at the beginning 'made them male and female,' and said, 'For this reason a man shall leave his father and mother and be joined to his wife, and the two shall become one flesh'? So then, they are no longer two but one flesh. Therefore what God has joined together, let not man separate" (Matthew 19:4–6). Jesus did make an exception in the case of an adulterous spouse, who has already been treacherous and violent toward the sacred union. As a general rule for believers, however, divorce is not God's plan.
**_Christians should not marry unbelievers._** To the modern reader, Ezra's reaction to the Jews marrying foreign women might seem extreme. However, we must take care not to thrust modern sensibilities onto the text of God's Word. Ezra's response to the situation demonstrated how seriously the Lord took the sin of intermarriage. Indeed, the fact the Jews were returning from captivity and rebuilding the temple and walls indicated that God did indeed take intermarriage seriously—for it was that very sin that had led the nation of Israel into idolatry in the first place.
The Bible refers to Christians who are married to non-Christians as being "unequally yoked." The word picture is drawn from a pair of oxen that a farmer would use to pull his plow. The two animals would be connected with a firm collar, or yoke. If the two oxen had different ideas concerning their roles as plow pullers, they would pull the plow in different directions, with one trying to go left and the other pulling toward the right. The result to the farmer and his plow is self-evident: the work would come to a standstill and the plow itself might be destroyed in the contest of wills. Paul warned clearly, "Do not be unequally yoked together with unbelievers. For what fellowship has righteousness with lawlessness? And what communion has light with darkness?" (2 Corinthians 6:14).
Two further principles are given for a Christian already married to an unbeliever. First, he or she is not permitted to seek a divorce even though the marriage is unequal because God hates divorce. "If any brother has a wife who does not believe, and she is willing to live with him, let him not divorce her. And a woman who has a husband who does not believe, if he is willing to live with her, let her not divorce him" (1 Corinthians 7:12–13). Second, if an unbeliever is married to a Christian, and the unbeliever wants a divorce, the Christian should let the unbeliever go. Paul wrote, "If the unbeliever departs, let him depart; a brother or a sister is not under bondage in such cases. But God has called us to peace" (verse 15).
**_A new start is not a new heart._** The Jews returned to the land with such promise. They had put away their idols, and only those who wanted to work on the temple returned to the land. This was a new beginning for Israel, and the sins of the previous generations of Jews must have seemed a distant memory. Yet as soon as they were in the land, they began imitating their parents' sins. It was as if the exile had never happened!
The main lesson of Ezra (and Nehemiah, Haggai, and Zechariah) is that apart from the new covenant, true obedience to God's Word is impossible. The Israelite history bears testimony to this fact. God repeatedly started over, and the Israelites repeatedly failed. He started over with Noah and his family, and sin gripped them immediately when they stepped off the ark. He started over with Moses and a new nation fleeing Egypt, but ended up leaving every one of them (except Joshua and Caleb) to die in the wilderness. Now God had again purified Israel, removing their entire nation for seventy years, but they were back into their sinful ways again.
This illustrates the fact that a new start does not necessarily indicate a new heart. This is the lesson that Jesus gave Nicodemus, who was a leader of the Jews. Jesus told him that unless a person is born again, he cannot see the kingdom of God (see John 3:3). Nicodemus thought Jesus was telling him to start his life over, and he wondered how that was even possible. But Jesus told him that without a new heart, fresh starts would simply produce fresh failures. The Jews were learning the lesson that they did not need their land, their temple, or their wall—what they needed was their Messiah.
## REFLECTING ON THE TEXT
5) Why does God hate divorce? Why did Jesus make an exception in the case of adultery (see Matthew 5:32)? How are adultery and divorce similar?
[Your Response Here]
6) What does it mean to be unequally yoked? What can you learn about intermarriage from the "ox and plow" metaphor? How does this principle apply in other areas of life?
[Your Response Here]
7) Why did Paul counsel believers not to divorce an unbelieving spouse? What did he mean when he said, "God has called us to peace" (1 Corinthians 7:15)? How is peace an important issue in an unequal marriage?
[Your Response Here]
8) Why did the Jews fall back into sin again when they returned to the land? What does this teach us about the human heart? What is the solution for their sin?
[Your Response Here]
## PERSONAL RESPONSE
9) How do these passages affect your views of marriage? How do they affect your views on divorce or singleness? How do these teachings apply in your own life?
[Your Response Here]
10) How does living with a new nature affect your struggle with sin? Do you find yourself like the Israelites, trapped in sin, or do you experience the freedom that comes with a relationship with Jesus? Explain.
[Your Response Here]
# JERUSALEM IN NEHEMIAH'S TIME
# **8**
# REBUILDING THE WALL
# _Nehemiah 1:1–2:20_
## DRAWING NEAR
What are some steps governments take today when a disaster has struck an area? What are some things that get in the way of a quick rebuilding effort?
[Your Response Here]
## THE CONTEXT
We now move forward in time approximately 100 years to 446 BC, when a man named Nehemiah was serving King Artaxerxes of Persia. Nehemiah held an important post in the king's court as his personal cupbearer, or "food taster." He was responsible for ensuring the king did not ingest any poison, which was a fairly common way of assassinating an unpopular monarch at the time. For Nehemiah, the position came with a great degree of trust and intimacy with the king, and probably brought him great wealth, comfort, and influence at court as well.
Meanwhile, the Jews who went to Jerusalem with Zerubbabel (and their descendants) were struggling against opposition and hatred from their enemies. As we saw previously, the projected rebuilding had been halted when the enemies of the Jews successfully persuaded King Artaxerxes to forbid completion of the city wall, which was the most important part of defense for not only the temple but also the people themselves.
Judah was far away from Susa, the capital of Persia where Nehemiah served the king, and it took a full three months to travel between the two cities. Yet when Nehemiah learned of the Jews' plight, he immediately made plans to forsake his comfortable situation and join his people in their work to rebuild the wall—even at great cost to himself.
In this study, we will look at the importance of seeking God first in prayer, what it means to have a heart for the things that matter to Him, and what is required to follow His will.
## KEYS TO THE TEXT
Read Nehemiah 1:1–2:20, noting the key words and phrases indicated below.
> _M OURNING WITH THE MOURNERS: Nehemiah, the king's cupbearer, learns of the situation in Jerusalem and responds with deep grief._
1. NEHEMIAH THE SON OF HACHALIAH: Nehemiah was the cupbearer to King Artaxerxes of Persia. His name means "Yahweh comforts." Nehemiah's father, Hachaliah, is mentioned again in Nehemiah 10:1 but nowhere else in the Old Testament.
THE TWENTIETH YEAR: This would be November–December 446 BC.
SHUSHAN THE CITADEL: This city was also known as Susa. It was located east of Babylon, approximately 150 miles north of the Persian Gulf. The book of Esther took place in this city, not long before the events of Nehemiah. Esther may well have been still alive at the time.
3. IN GREAT DISTRESS AND REPROACH: In a previous study we examined some of the distress and reproach suffered by the Jews in Jerusalem as they experienced the betrayal and opposition of the Samaritans (see Ezra 4). Yet their suffering was greater than mere human opposition. In Nehemiah's day, pagans thought the walls and fortifications that surrounded a god's temple reflected his strength. The lack of city walls in Jerusalem reflected badly on the character of God in the eyes of Judah's neighbors, and it also left the temple unprotected against further attacks and desecration.
4. I SAT DOWN AND WEPT: Nehemiah was deeply moved by the plight of his fellow Jews in Jerusalem. He did not just express shallow sympathy but reacted as though he were right there, in their midst, suffering with them.
FASTING AND PRAYING: Nehemiah's mourning and grief were not impotent. He did not just sit in Persia's palace, wringing his hands and wondering what to do, but took action to help those who were suffering. He began with the most important and effective step: going before the Lord in prayer and fasting to seek His will and intervention.
> _R ESPONDING WITH PRAYER: Nehemiah does more than sit and weep—he goes before God and asks for His help and guidance._
5. YOU WHO KEEP YOUR COVENANT AND MERCY: In this beautiful prayer, Nehemiah revealed an important fact concerning God's character: He is completely faithful to His promises. The people of Jerusalem didn't need any walls to protect the temple or themselves, for they were shielded by God's complete faithfulness (though it was still God's will for the walls to be rebuilt). Nehemiah also recognized God's absolute sovereignty over all the affairs of mankind, calling him "LORD God of heaven" who rules in the lives of kings and paupers alike.
6. I PRAY . . . DAY AND NIGHT: Nehemiah prayed persistently over a period of days, not simply on this one occasion.
THE CHILDREN OF ISRAEL . . . MY FATHER'S HOUSE AND I: Notice the progression of Nehemiah's confession. He acknowledged first that his nation had sinned against the Lord by forsaking Him and pursuing false gods. He then confessed that his "father's house" had not been faithful to the Lord, which encompassed his family and those for whom he was responsible. (This might be comparable in modern times to praying for one's immediate family and local church.) Finally, he acknowledged that he was personally guilty of sin and failure before the Lord. He did not omit his own areas of culpability but recognized that his own sinful nature had contributed to the unfaithfulness of his tribe and nation. When one begins by focusing on God's faithful and holy nature, such recognition of sin—both personal and corporate—naturally follows.
8. REMEMBER, I PRAY: Nehemiah was asking the Lord to act on His promises, not suggesting that He had forgotten something. The Lord loves to have us quote His Word in an expectant way, which indicates we have read His promises and have faith that He will keep them. And He always does.
9. I WILL GATHER THEM FROM THERE: As we have already seen, the Lord had promised His people that He would scatter them to the four winds if they refused to heed His Word (see Deuteronomy 4:25–28). However, He had also promised that, at the end of seventy years, He would gather a remnant and return them to His chosen city of Jerusalem (see verses 29–31). The imagery used here is of sowing seed (which was scattered randomly by hand) and reaping a harvest. The Lord had scattered His people when they were unfaithful, but He had now reaped a harvest of obedient and loving children.
11. THE KING'S CUPBEARER: Nehemiah held one of the most trusted and intimate positions in Persia. His job was to ensure the king's safety by personally tasting his food first to make certain that no one poisoned him. This situation provided Nehemiah with a unique opportunity to speak confidentially with the king, and that was what he was preparing to do as he prayed and fasted. ("This man" refers to the king, whom Nehemiah was about to petition.)
> _P REPARING TO APPROACH THE KING: Nehemiah next determines to ask the king for a leave of absence—but this was a dangerous request to make._
2:1. THE MONTH OF NISAN: This would be March–April 445 BC, approximately four months since Nehemiah first learned of the situation in Jerusalem.
I TOOK THE WINE AND GAVE IT TO THE KING: Nehemiah's chief responsibility was to ensure the king's enemies did not poison him, so he was basically putting his life on the line every time he sampled the king's food and drink. This role enabled him to establish a strong bond of trust with Artaxerxes, and he would now draw on that relationship as he made his petition.
I HAD NEVER BEEN SAD IN HIS PRESENCE: Court etiquette in ancient times required that subjects appear cheerful before their king, and a sad or grim countenance could actually bring down royal wrath (see Esther 4:2). Happy subjects suggested the king was a good and wise ruler, while tears suggested otherwise. As cupbearer, Nehemiah would have needed to observe such customs of courtly behavior very carefully.
2. SINCE YOU ARE NOT SICK: The king might have been referring to Nehemiah's general state of health, which is the usual understanding here. However, he might also have been suggesting that because Nehemiah had clearly not ingested poison from the king's cup, he should be pleased. Middle Eastern monarchs in general did not take time to consider the personal lives of their subjects.
I BECAME DREADFULLY AFRAID: From a human standpoint, Nehemiah had good reason to be afraid. As already stated, it was risky business for a subject to appear unhappy in the presence of the king. But even more, Nehemiah was planning to request permission to leave the king's service, which would have been seen as a terrible insult from someone who held such a high and trusted position. He was also going to ask to rebuild the walls of Jerusalem from the very king who had ordered that building to stop. Nevertheless, Nehemiah was not acting from a human perspective, and he knew that the Lord had control over the king's decisions (see Proverbs 21:1).
4. SO I PRAYED TO THE GOD OF HEAVEN: Nehemiah was a man of prayer. He had fasted and prayed for several days prior to making his request to the king. Even now, in the midst of taking action, he was speaking to the Lord.
6. THE QUEEN ALSO SITTING BESIDE HIM: Esther had been queen of King Ahasuerus (Xerxes), who was Artaxerxes' father. It is possible that her role as the present king's stepmother had made him predisposed to show favor to the Jews—particularly if she was still alive.
> _W HATEVER YOU NEED: The Lord moves the king's heart to grant Nehemiah's request for a trip to Judah, and he gives him much more besides._
7. LET LETTERS BE GIVEN TO ME: These letters, which bore the royal seal of Artaxerxes, commanded all government officials to assist Nehemiah on his trip to Judah and with his work in Jerusalem. The king was effectively giving Nehemiah full authority as one of his officials in Judah.
8. HE MUST GIVE ME TIMBER: In addition to granting Nehemiah governing authority and safe passage, the king also provided much of the raw materials the Jews would need to rebuild the city gates and wall—and Nehemiah's own house as well. Lumber was a precious commodity at the time, and the ruling monarchs guarded forests carefully.
ACCORDING TO THE GOOD HAND OF MY GOD UPON ME: Nehemiah was undoubtedly grateful to King Artaxerxes for his generosity, but he recognized that his success was due solely to God's sovereign hand. It was the Lord who moved the king's heart, even after Artaxerxes had previously forbidden any such reconstruction projects in Jerusalem.
9. THE KING HAD SENT CAPTAINS OF THE ARMY AND HORSEMEN: This was a wise human precaution. It was a three-month journey from Susa to Jerusalem, and the roads along the way were fraught with banditry. The letters Nehemiah carried bore the king's seal, and as such were immensely valuable because they conferred on the bearer the full authority of the king of Persia. That alone would have put Nehemiah's life in extreme peril on such a long journey, to say nothing of the many peoples along the way who were enemies of the Jews. Yet from God's perspective, the military guard was merely superfluous. He had determined that Nehemiah should go to Jerusalem, and no power on earth could have prevented his safe arrival.
> _M ORE OPPOSITION: Initiating the project of rebuilding is merely a beginning. Satan's agents soon reappear to continue to oppose God's work at every step._
10. SANBALLAT THE HORONITE AND TOBIAH THE AMMONITE: Sanballat was the governor of Samaria and probably also a Moabite—an ancient enemy of Israel. Tobiah was governor of the region east of the Jordan River, and he was an Ammonite—another ancient enemy of God's people. These two men were probably also behind the opposition against Zerubbabel that had previously stopped the work in Jerusalem.
DISTURBED THAT A MAN HAD COME: This attitude revealed the true nature of these men and their cohorts: they were committed enemies of God and His people. It is particularly revealing that they were not so much opposed to the authority that Nehemiah wielded but were opposed to anyone who sought "the well-being of the children of Israel."
12. I TOLD NO ONE: Nehemiah needed to know the current condition of the wall, what areas needed the most immediate attention, what would be required to rebuild it, and so forth. He gained this information in secret because he knew that his enemies were watching him, and he didn't know yet who could be trusted. He probably suspected that even some of the Jews in the city were in league with Sanballat and Tobiah, so to be safe he did not even tell "the Jews, the priests, the nobles, the officials, or the others who did the work" (verse 16).
17. LET US BUILD: Nehemiah next invited all of God's people to join him in the work. It was not Nehemiah's project but the Lord's, and He had called all His children to be involved in His work. This also would enable Nehemiah to discover who was fully committed to the plans of God, as those would be the people who threw themselves into the rebuilding.
THAT WE MAY NO LONGER BE A REPROACH: This was Nehemiah's motivation in undertaking this huge project: he was concerned about the glory of God. The destruction of the temple and city of Jerusalem had made a mockery of God's name in the eyes of Judah's enemies, so Nehemiah called on the people of God to restore His glory to the world around them.
18. I TOLD THEM OF THE HAND OF MY GOD: It is good to tell others of the ways God has blessed you and how you have seen Him working faithfully in your own life. This encourages others who might be facing discouragement, and it brings glory to God's name.
## UNLEASHING THE TEXT
1) Why was Nehemiah so upset when he learned of the plight of the Jews in Jerusalem? What was his response?
[Your Response Here]
2) What was the "great distress and reproach" that the Jews were facing in Judah? In what ways did this reflect on God's glory?
[Your Response Here]
3) Why did Nehemiah survey the wall around Jerusalem in secret? Why did he then call the people to join him in the rebuilding? What does his example teach about leadership?
[Your Response Here]
4) What can you learn about prayer from Nehemiah's example (see Nehemiah 1:5–11)?
[Your Response Here]
## EXPLORING THE MEANING
**_Service is crucial, especially when it's costly._** Nehemiah lived in Persia, the greatest and wealthiest nation of its day. Furthermore, he lived in Susa, the nation's capital and one of the richest and most comfortable of the Persian cities. To top this off, he was the cupbearer to the king himself, a position of high trust and influence. He was undoubtedly rich and high in the ranks of the most powerful nation on earth. Jerusalem was far away—a journey of three full months. It would have been easy for Nehemiah to forget about his people.
Yet when he heard about the plight of his fellow Jews in far-off Judah, he mourned, wept, fasted, and prayed. What's more, he determined in his heart to forsake all the blessings and comforts of Persia and exchange them for hard work, rough living conditions, and constant hatred and opposition from God's enemies. In fact, Nehemiah was so determined to help with the work in Jerusalem that he risked his own life to get there. He put himself in peril of the king's wrath by making the request to leave the king's service, and then he embarked on the dangerous and uncomfortable trip to Judah.
It was not wrong or sinful for Nehemiah to enjoy the comforts of Persia and the king's court. The Lord had placed him in this position, and Nehemiah was being faithful to the tasks God had given him. But the Lord had placed him there specifically so that he might be positioned to help the Jews at this moment of crisis, just as He had placed Esther where she could save the Jews from annihilation a generation earlier (see Esther 4:14). The Lord was calling Nehemiah to voluntarily forsake all these blessings to participate in an important project, and the blessings that came from his obedience far surpassed all the comforts of the king's palace. When the Lord calls us in a similar way, we must always heed the call, for our obedience will bring glory to His name and great blessings to us.
**_Our protection is found in God alone._** The Jews in Jerusalem were anxious to rebuild the city walls. They wanted a strong defense against the many enemies who surrounded them and would have been glad to see them carried off into captivity once again. King Artaxerxes himself recognized the danger in the region, which is why he provided Nehemiah with a powerful military escort to protect him on the dangerous journey. These were treacherous times for the Jews, for there were many who hated them and longed to participate in their destruction.
Yet the Lord did not need stone walls or well-armed soldiers to protect His people. His omnipotent hand was sufficient, and His faithfulness to His promises ensured that nothing could ever touch His servants without His permission. Ezra had previously chosen to trust in the Lord's protection and make the same long journey without any military escort (see Ezra 8:22)—and his faith had proved sound. This does not mean, of course, that Nehemiah had less faith than Ezra. The king had offered the escort, and Nehemiah saw that as God's provision at the time. Nevertheless, had the king _not_ offered any soldiers, Nehemiah would still have arrived safely through the Lord's faithfulness.
However, we must not overlook the important element of obedience in the Lord's protection. Notice what Nehemiah prayed in 1:5–6: "O great and awesome God, You who keep Your covenant and mercy with those who love You and observe Your commandments, please let Your ear be attentive and Your eyes open." The Lord is always faithful to His promises, and faithful to protect His children from the enemy, but He also expects us to be faithful to His Word. When we sin, we make ourselves vulnerable to the attacks of the evil one and hinder God's hand of blessing in our lives. Our job is to obey His Word and trust Him for our needs.
**_God loves to see the faith of His people._** When Nehemiah spent time in fasting and prayer before approaching the king with his risky request, he reminded God of the promises he had made concerning Judah. He pointed out that the Lord had kept His promise to send the people into captivity if they persisted in idolatry, and he then reminded God of His further promise to return them to Judah after seventy years. He was reminding the Lord of these things not out of a spirit of accusation but in a spirit of expectation.
There is another layer to this prayer, however, which we might easily overlook. In order for Nehemiah to remind God of His promises, he had to _know_ those promises in the first place. Nehemiah's knowledge of the limitation on the captivity indicated that he had spent time reading and meditating on the Word of God. This is one reason why the Lord is pleased when we quote His Word to Him in prayer, because He wants us to read it, memorize it, and meditate on it. But He also wants us to make His Word part of our own lives and ask ourselves how it applies to us personally in our daily routines.
It is good for us to remind the Lord of the promises He has made to us. Of course, the Lord will keep His promises even if we do not remind him of them, simply because He never fails. Yet He delights when His children remind Him of His Word, because it shows that we have faith in His character and His faithfulness. It also strengthens us and deepens our faith when we remind _ourselves_ of His promises.
## REFLECTING ON THE TEXT
5) What risks did Nehemiah take to help the Jews in Judah? What were his motivations in taking these risks? What would you have done in his situation?
[Your Response Here]
6) Why did Ezra not take a military escort on his trip to Judah (see Ezra 8:22)? Why did Nehemiah accept the king's guard on this trip? What do these men's attitudes reveal about God's protection?
[Your Response Here]
7) Why did Nehemiah remind God of His promises in his prayer? What effect might this have had on Nehemiah? How do these principles apply to your prayer life?
[Your Response Here]
8) What did Nehemiah tell the people in Jerusalem after he surveyed the wall (see Nehemiah 2:18)? How did the people respond? How might these insights apply to your witness for Christ?
[Your Response Here]
## PERSONAL RESPONSE
9) What are you presently depending on for your security? In what areas of your life might you need to place more faith in God's protection?
[Your Response Here]
10) When has another Christian served you at great cost to himself or herself? Why did he or she do so? Who might be the one the Lord is calling you to serve this week?
[Your Response Here]
# **9**
# THE OPPOSITION INCREASES
# _Nehemiah 3:1–4:23_
## DRAWING NEAR
What tends to happen to people's enthusiasm for a project when it drags on longer than they expected or they encounter problems they didn't expect?
[Your Response Here]
## THE CONTEXT
Nehemiah had made the arduous journey from Susa to Jerusalem, had personally surveyed the damage to the city wall, and had called on the officials of the city to join with him in the rebuilding effort. The people had responded in faith, committed themselves to the project, and set to work on the walls and gates with great vigor and zeal. Through the Lord's blessing and help, they made good progress, and soon the wall was nearly half built.
When Sanballat, Tobiah, and Geshem—the leaders of those who opposed the Jews—observed the people's progress, they laughed and said, "What is this thing that you are doing? Will you rebel against the king?" Nehemiah responded by telling them the God of heaven Himself would guide and prosper their efforts. He added, "You have no heritage or right or memorial in Jerusalem" (Nehemiah 2:19–20).
In this study, we will find these enemies responding with as much zeal and vigor to _halt_ the rebuilding as the Jews have shown in pursuing the work. They will stop at nothing to prevent the completion of the walls and will resort to all forms of verbal abuse and scandal—all while planning a mass murder of the Jews. Nehemiah will demonstrate that God's people must expect opposition and always be prepared to face it.
The most important lesson we will gain from this study is that we must always trust in God's promises as we accomplish the tasks that He has given to us to complete. As we lean on His strength and depend on His provision, we will see Him come through in miraculous ways.
## KEYS TO THE TEXT
Read Nehemiah 3:1–4:23, noting the key words and phrases indicated below.
> _D IVIDE AND CONQUER_ **:** _The reconstruction effort begins in Jerusalem, with different groups assuming responsibility for the repairs on different parts of the wall and gates._
3:1. ELIASHIB THE HIGH PRIEST: He was the grandson of Jeshua, the high priest from Zerubbabel's era (see Ezra 2:2). The events in this chapter took place on the fourth day of the Hebrew month of Ab (July–August) in 445 BC.
THE SHEEP GATE: This gate was located in the northeast section of Jerusalem. The narrative moves around the perimeter of Jerusalem in a counterclockwise direction.
TOWER OF THE HUNDRED . . . TOWER OF HANANEL. This northern section of Jerusalem opened up to the central Benjamin plateau, where enemy forces could attack most easily from the north. The rest of the perimeter of the city was protected by the natural valley topography.
3. FISH GATE: So named because merchants sold fish on the northern side of Jerusalem. The people of Tyre and other coastal towns routinely brought fish to sell.
5. THEIR NOBLES DID NOT PUT THEIR SHOULDERS TO THE WORK OF THEIR LORD: One explanation for this, beyond just the laziness of the rich people in the city, is that these nobles had been pledged to Tobiah for personal gain.
6. THE OLD GATE: This is believed to have been in the northwest corner of Jerusalem.
8. THE BROAD WALL: Located on the western side of the northern sector.
11. TOWER OF THE OVENS: Located on the western side of Jerusalem.
13. THE VALLEY GATE: Located on the western side of Jerusalem.
14. THE REFUSE GATE: Also known as the Dung Gate. A common sewer ran to the Kidron Brook into the Valley of Hinnom at the southern tip of the city.
15. THE FOUNTAIN GATE: Located on the southeast corner of Jerusalem.
THE KING'S GARDEN: Located in the southeast sector.
16. TOMBS OF DAVID: Presumably located in the southeast sector.
HOUSE OF THE MIGHTY: This location probably was associated with David's mighty men (see 2 Samuel 23:8–39).
19. THE ARMORY: Located on the eastern side of Jerusalem.
26. WHO DWELT IN OPHEL: An area located south of the temple mount, near the Water Gate, where the Nethinim lived.
THE WATER GATE: Located south of Gihon Spring on the east side of Jerusalem.
28. THE HORSE GATE: Located in the eastern sector.
29. THE EAST GATE: Possibly located to the east of the temple mount.
31. THE MIPHKAD GATE: Located in the northeast sector.
32. THE SHEEP GATE: Having traveled around Jerusalem in a counterclockwise direction, the narrative ends where it began.
> _M OCKERY AND DISDAIN: The enemies of God's people see the progress the Jews are making on the walls and gates of Jerusalem and are filled with fury._
4:1. FURIOUS AND VERY INDIGNANT: On the surface, Sanballat, the governor of Samaria, was outraged that the Jews had overridden his authority in the region by gaining the support of the king. But the deeper truth is that he was an enemy of God—and therefore an enemy of God's people. He could not abide the thought that the Jews were going to prosper and rebuild their beloved city of Jerusalem.
2. HE SPOKE BEFORE HIS BRETHREN: For mockery to have any effect requires a sympathetic audience, so Sanballat attempted to draw others into his plans to belittle the Jews. His "brethren" would have been others of like mind who hated God's people as well. It is possible that he also intended to provoke "the army of Samaria" into action, as that would have swiftly brought the Persian overlord down on Samaria. Harassment and mockery would prove to be the primary strategy the enemy would use to prevent the reconstruction of the walls.
4. HEAR, O OUR GOD: Once again, Nehemiah's first response to the danger and opposition was to bow down before the Lord in prayer. Notice that Nehemiah did not respond to the taunts and mockery of his enemies. Instead, he poured out his heart to the Lord and allowed God to deal with the opposition.
5. DO NOT COVER THEIR INIQUITY: Nehemiah was acknowledging that his enemies were attacking God rather than him. It was the Lord whom they despised, and Nehemiah was calling on Him to defend His glory before the world. Nehemiah recognized that forgiveness is granted not to those who oppose the will of God but to those who eagerly submit to His will.
6. SO WE BUILT THE WALL: Nehemiah effectively turned the situation over to the Lord in his short prayer, and then he returned his focus to the work the Lord had given him. He did not pretend the opposition did not exist, but he did trust that God would be their defender.
THE PEOPLE HAD A MIND TO WORK: The people had been enthusiastic and eager to participate during the early phases of the building project. They had divided the work among themselves and been excited to see the progress as the wall came together. However, human nature tends to cause people to lose enthusiasm as time passes—and the taunts of the enemy were not helping to keep the people motivated. It is during such times in our own lives that determination becomes important in obeying and completing the Lord's work, for distractions and discouragement will multiply as our enthusiasm dissipates.
> _E NEMIES ON ALL SIDES: Groups of people who would not normally be allies with each other now band together on all four sides of Jerusalem to hinder God's work._
7. THE ARABS, THE AMMONITES, AND THE ASHDODITES: The Samaritans (under Sanballat) were to the north, the Arabs (under Geshem) were located to the south, and the Ammonites (under Tobiah) were to the east. Added to the list of enemies were the dwellers of Ashdod to the west—thus effectively surrounding the people of God on every side. Apparently, the Ashdodites had come to the point where they were at least contemplating a full-scale attack on Jerusalem because of the rapid progress of the wall. Ironically, these groups would not have cooperated under normal circumstances—what drew them into this alliance was a mutual hatred of God and His people. This same trend can be seen today in parts of the Middle East.
THEY BECAME VERY ANGRY: Here the enemies tipped their hand, showing the true nature of their opposition. They had pretended to care about the king's taxes and about national security, and had acted as if the Jews could not construct anything lasting, but in the end they were only angry at the thought of the Jews being able to defend themselves. God's enemies criticize, manipulate, and persecute God's people today just as they did then.
8. CREATE CONFUSION: Another of the devil's favorite tactics is to create confusion and chaos wherever unity exists among God's people. For this reason it is not uncommon for the people of God to be tempted toward disunity, especially during times of critical work on behalf of the Lord. The enemies of the Jews did not care what the people believed. They simply wanted to stop the progress being made on rebuilding the wall.
9. WE MADE OUR PRAYER . . . WE SET A WATCH: Note the people's twofold approach to the danger they faced: they turned to God in prayer, and they took steps to guard themselves. The Jews recognized their protection was from God alone, yet that did not exempt them from carrying out their basic human responsibilities. It would have been negligent not to establish watchmen to guard the wall (and their homes) at night. The Lord wants His children to trust Him for their security, but He also expects them to be responsible.
> _D ISCOURAGEMENT AND VIOLENCE: God's enemies now resort to threats of murder, and the Jews begin to grow weary and discouraged._
10. THE STRENGTH OF THE LABORERS IS FAILING: The Jews' initial enthusiasm had worn off and fatigue was setting in. This was a critical point in Nehemiah's administration, for the enemy's attacks were beginning to have their desired effect of halting the work. Nehemiah needed to take steps to renew the people's enthusiasm for the project to be a success.
SO MUCH RUBBISH: This term literally means _dust._ The people were referring to the rubble or ruins of the prior destruction (in 586 BC), which they had to clear away before they could make significant progress on rebuilding the walls.
11. KILL THEM AND CAUSE THE WORK TO CEASE: Part of the strategy of the enemy coalition was to frighten and intimidate the Jews by making them think an army would surprise them with a massive force that would engulf them. God's enemies will stop at nothing, even mass murder, in hopes of preventing the fulfillment of His plans. This is a futile hope, however, as nothing whatsoever can prevent God's plans from coming to pass. The destruction His enemies plan will come on their own heads (see Proverbs 26:27).
12. THE JEWS WHO DWELT NEAR THEM: These were Jews who had not gotten involved in the rebuilding project. They probably meant well by warning the others of an impending attack. However, they had inadvertently become tools of the enemies by yielding to fear and potentially spreading that fear to the people who were actually working. As we have seen throughout this study, fear is the enemy of God's people.
> _F OILED AGAIN: The Lord prevents His enemies from attacking His people during the work of rebuilding, for the battle belongs to Him alone._
13. THEREFORE I POSITIONED MEN: God had exposed Sanballat's strategy of mustering the army of Samaria by revealing it to the nearby Jews so they would report it to Judah's leaders. Although Nehemiah and those he led were vigilant, armed, and ready, they consistently gave God the glory for their victories and construction successes.
14. DO NOT BE AFRAID OF THEM: Once again we are given this important injunction to not be afraid. God continually tells His people not to yield to fear because fear leads a person away from God rather than toward Him.
REMEMBER THE LORD . . . AND FIGHT: Here again we see Nehemiah taking both a spiritual and practical approach to the problem. Remembering the Lord involves meditating on His Word and His character as well as turning to Him in prayer. The fighting that Nehemiah had in mind was quite literal, as demonstrated by the Jews' willingness to take up arms to defend themselves.
15. GOD HAD BROUGHT THEIR PLOT TO NOTHING: Nehemiah never lost sight of the fact that God was in control of all events. The plot was not foiled because the Jews had learned about it, nor because of their readiness to fight back, but because God prevented it.
16. HELD THE SPEARS, THE SHIELDS, THE BOWS, AND WORE ARMOR: Notice that half of these weapons were offensive (spears and bows) and half were defensive (shields and armor). Paul would later use such military weaponry to describe the Christian life, including the shield of faith, the breastplate of righteousness, and "the sword of the Spirit, which is the word of God" (Ephesians 6:14–17). Yet on a practical level, Nehemiah took strong precautions to guard against the threatened physical attack of the enemy by dividing the workers into two teams. One team would guard while the other team worked. The effect of the threats, however, was to cut the work force in half, and even those who worked carried weapons in case of attack.
18. THE ONE WHO SOUNDED THE TRUMPET: Among other functions, the trumpet would have sounded a loud call to arms that alerted all the Jews of an attack. Nehemiah kept a trumpeter at his side always so the alarm could be sounded immediately. He took it on himself, as the Lord's appointed leader, to stand guard vigilantly on behalf of those under his leadership.
20. OUR GOD WILL FIGHT FOR US: Once again, Nehemiah reminded his coworkers that the battle belonged to the Lord, not to the strength of their arms.
## UNLEASHING THE TEXT
1) How did Nehemiah divide the work on the city walls and gates? What were some of the benefits in adopting this approach?
[Your Response Here]
2) What tactics did the enemies use to thwart the rebuilding project? How do God's enemies today use similar tactics?
[Your Response Here]
3) Why did Nehemiah command his workers to always be armed? What effect did this have on the work? What effect did it have on the enemies?
[Your Response Here]
4) What role did prayer play in the events of this chapter? What practical steps did the people take? How does a Christian blend prayer and practical common sense?
[Your Response Here]
## EXPLORING THE MEANING
**_Expect opposition, but don't ignore it._** Zerubbabel, Ezra, and Nehemiah all faced determined opposition from powerful foes. They understood that life is filled with spiritual battles, for the enemy of our souls continuously strives to hinder the work of God. These men expected opposition to the Lord's great project of rebuilding Jerusalem, and they were not caught off guard when it came. We must not assume every hardship, opposition, or obstacle is an indication that we are acting according to the Lord's will. God _does_ use practical hindrances at times to force us to recognize when we are not going in the right direction.
Consider Balaam, for example, a pagan shaman who practiced sorcery and other wicked acts for profit. He had been offered a great sum of money to cast a curse on the Lord's people after their exodus from Egypt, but the Lord had warned him, "You shall not go . . . you shall not curse the people, for they are blessed" (Numbers 22:12). Nevertheless, Balaam saddled his donkey and headed out—ostensibly to bless God's people, yet in disobedience to the Lord's command. As he traveled along, the Lord sent an angel to hinder his donkey's progress. He even caused the donkey to speak (who showed more wisdom than his foolish master) in order to warn Balaam that he was going the wrong way.
When we do the Lord's work, we will face spiritual opposition from the enemies of God's people, and we must expect it. But when there's any doubt, we shouldn't assume that opposition is a "green light" on our actions. Rather, we must turn immediately to the Lord in prayer and meditate on His Word. The Lord never commands His people to do anything contrary to the principles of Scripture, and He might be using obstacles to open our eyes to a wrong decision.
**_When enemies assail you, talk to the Lord._** The Jews faced a variety of threats and opposition from their enemies as they were rebuilding the wall. They were openly mocked and derided, falsely accused of evil motives and rebellious plans, subtly infiltrated by men who sought to discourage them and hinder their work, and even threatened with murder. However, in every instance under Nehemiah's leadership, we see God's people turning to the Lord in prayer and expecting to find their safety and security in Him.
This is an excellent example that Christians would do well to follow. Most of us instinctively try to resolve problems on our own, either by trying to devise our own schemes and solutions, or by simply running away. However, notice how Nehemiah responded to those who were mocking him and slandering his reputation. He did not defend himself, nor did he retaliate with sarcasm and verbal abuse. When those same enemies planned to attack the Jews violently, Nehemiah did not strike out with his own sword. In both cases, the Lord's people turned to their God in prayer.
Turning to God in prayer is not the same as abdicating responsibility. We still need to fulfill our responsibilities—as the next principle will illustrate—and one of those responsibilities is to use common sense while not giving in to fear and discouragement. But in the big picture, the battle belongs to the Lord, and we are wise to turn the fighting over to Him.
**_We must be alert and prepared for battle._** Nehemiah didn't take the enemy's threat lightly when he learned they were planning to slaughter the Jews. He knew, of course, that the true protection of God's people lay solely in God's hands, but he also understood that he had a responsibility for the safety and welfare of those under his authority. Consequently, he took strong steps to prepare for the threat of battle. What's more, he expected all those working on the Lord's project to also remain alert and well armed. It must have been a real hindrance to the work of building, which was strenuous enough without having only one hand available, but Nehemiah felt the work of self-defense was as important as the work of rebuilding the wall.
Most of us do not face the threat of physical violence for our faith (though there are many Christians in parts of the world who do), but this principle applies at least as much on the spiritual level. Even if our neighbors are not threatening to attack us, we all face an enemy who is deadlier than those who opposed the Jews. For this reason, God commands us through His Word to be constantly on guard against the forces of wickedness, and He also commands us to go everywhere well armed. As Paul instructs, we are to carry with us "the sword of the Spirit, which is the word of God" (Ephesians 6:17). The writer of Hebrews gives this further detail: "For the word of God is living and powerful, and sharper than any two-edged sword, piercing even to the division of soul and spirit, and of joints and marrow, and is a discerner of the thoughts and intents of the heart" (Hebrews 4:12).
Well-armed and vigilant Christians spend time reading and meditating daily on the Word of God. And like all well-trained soldiers, believers in Christ also stay in close contact with the Commanding Officer through prayer and obedience. The military analogy, in fact, is apt in life, because we live in a battle zone where the enemy is constantly trying to destroy us. Therefore, we must always "be sober, be vigilant; because your adversary the devil walks about like a roaring lion, seeking whom he may devour. Resist him, steadfast in the faith, knowing that the same sufferings are experienced by your brotherhood in the world" (1 Peter 5:8–9).
## REFLECTING ON THE TEXT
5) Why did Nehemiah ask the Lord to not cover the iniquity of his enemies (see Nehemiah 4:5)? How is this different from harboring a vindictive spirit?
[Your Response Here]
6) What does it mean to resist the devil (see 1 Peter 5:9)? How is this done, in practical terms?
[Your Response Here]
7) Why did Nehemiah make no response to the verbal abuse of the enemies? Why did he make a practical response to their threats of violence? In what circumstances is each response most wise?
[Your Response Here]
8) When have you faced opposition or problems caused by an enemy? When has the Lord used difficulties to turn you away from doing wrong? How can you tell the difference?
[Your Response Here]
## PERSONAL RESPONSE
9) How do you normally respond to threats or verbal abuse? How did Jesus respond to such things? In what ways do you need to become more like Him?
[Your Response Here]
10) What trials or hardships are you trying to resolve in your own strength? List them below, and then spend time each day this week asking the Lord to take over the battles.
[Your Response Here]
# **10**
# LEADING BY EXAMPLE
# _Nehemiah 5:1–6:19_
## DRAWING NEAR
What does it mean to "practice what you preach"? Why is it important for a person in leadership to exhibit this trait?
[Your Response Here]
## THE CONTEXT
Even though Nehemiah's workforce had essentially been cut in half—with half the men working while the other half stayed on security duty—progress on the wall continued. Yet at the end of a typical day, Nehemiah's workers had neither the time nor energy left to attend to many of their own personal responsibilities. Many of those responsibilities were important, such as paying their taxes and gathering enough food to feed their families.
This led many of the people into financial difficulties. They had debts that could not be paid right away, or they had to borrow to pay their taxes. Some of them had to go outside the city to work as hired hands harvesting other people's crops. Many people in Jerusalem and the surrounding areas were going hungry, and they ended up being exploited by those who were more fortunate.
The Lord had given clear instructions to His people on how to be generous with the poor. However, Nehemiah discovered that despite the work on the wall and the temple, the people were not obeying the laws that God had given Israel. In spite of the new start in the land, many Jews were still reluctant to live in the way God had instructed, and they were taking unfair advantage of the poor for personal gain.
## KEYS TO THE TEXT
Read Nehemiah 5:1–6:19, noting the key words and phrases indicated below.
> _T HE PEOPLE CRY OUT: Work on the wall is suddenly interrupted by a great outcry from the people. They cannot pay their bills, and they are being sold into slavery._
5:1. THERE WAS A GREAT OUTCRY: Notice the outcry was from "the people and their wives," which suggests entire families of workers had been affected. The sad irony is that the outcry was "against their Jewish brethren." The workers had no complaint against the Persian government (indeed, they had reason to be grateful to the king), nor against the foreigners in their midst. The problem came from the Jews themselves, for the wealthy were taking advantage of those doing the rebuilding. These may have been the same Jews who earlier refused to join the work of rebuilding, probably having formed an alliance with Sanballat and Tobias.
2. LET US GET GRAIN: There were three groups facing financial crisis. The first group had spent all their time and energy rebuilding the wall and had no time left to work for their own food. These people may have been among the poorest—those who owned no property and normally would have earned a living by harvesting the fields of others.
3. WE HAVE MORTGAGED OUR LANDS: The second group consisted of property owners who had been forced to take mortgages in order to pay their bills. To make matters worse, there had been a famine, and these people were having trouble feeding their families.
4. WE HAVE BORROWED MONEY FOR THE KING'S TAX: The third group had been hard hit with Persian taxes both on property and on produce. This group was the hardest hit, for they had not only mortgaged their lands but also sold their children into slavery.
5. FORCING OUR SONS AND OUR DAUGHTERS TO BE SLAVES: People who could not pay their debts might sell themselves (or their children) into slavery. God's Word contains specific stipulations concerning this practice—stipulations these wealthy Jews were not obeying. First, the Lord forbade His people to charge interest to one another on loans (see Exodus 22:25). Second, the debtor was to be treated with dignity and respect as a fellow member of God's chosen people (see Leviticus 25:39–40). Third, such debts might include indentured service, but it did _not_ include a person's property. The Lord had given the Promised Land to all His people, and they were not permitted to buy or sell that land, even to one another (see Numbers 36:7–9).
NOT IN OUR POWER TO REDEEM THEM: Even the indentured servant-hood was limited to seven years, after which the entire debt was to be canceled and the servants returned to their homes and families. (There was also a special Year of Jubilee every fifty years in which all debts were automatically canceled.) Neglecting this was one of the sins that had brought on the exile, and the Jews had already returned to it.
> _R IGHTEOUS INDIGNATION: Nehemiah responds to this situation with great anger, but his actions in response are somewhat surprising._
6. I BECAME VERY ANGRY: Nehemiah demonstrated the difference between sinful anger and righteous indignation. His anger was kindled by the fact that the Jews were not obeying God's Word. However, he did not express his anger in retaliation but instead addressed the underlying sinful behavior that was causing the problem.
7. I REBUKED THE NOBLES AND RULERS: Nehemiah took time to reflect and seek the Lord's wisdom in how to handle the situation before rebuking the nobles and rulers. The commitment of this group to the reconstruction project had been negligible, and their loyalty to the Jews' enemies had added to their opportunistic attitudes, which placed them close to the status of opposition. They had become the enemy from within. The Hebrew phrase here for _rebuke_ implies a strong contention, even taking legal action against someone. Nehemiah did not overreact, but he did not gloss over the nobles' and rulers' sin either.
8. WE HAVE REDEEMED OUR JEWISH BRETHREN: Once again, Nehemiah set an example by his own behavior. He first denounced with just severity the evil conduct of selling a brother by means of usury. Nehemiah then contrasted it with his action of redeeming (with his own money) some of the Jewish exiles who through debt had lost their freedom in Babylon.
THEY WERE SILENCED: Nehemiah's rebuke was unanswerable because he had founded it on two things: God's Word and his own personal example. The fact that he had obeyed the Lord at cost to himself in a similar matter left no room for his opponents to argue back.
9. BECAUSE OF THE REPROACH OF THE NATIONS: Once again, Nehemiah's paramount concern was the glory of God before the rest of the world. The actions of these nobles and rulers had placed their fellow Jews in danger, but it was far worse that they had slandered God's name before their pagan neighbors.
> _N EHEMIAH'S STAND: Nehemiah continues to lead by example by refusing to accept provisions that were lawfully due to him in order to avoid further burdens on the people._
10. I . . . AM LENDING THEM MONEY AND GRAIN: Nehemiah expected God's people to lend to those in need without expectation of profit. If a person was destitute, those who had the means were to give to them as a free gift. If the debtor was able to repay the money, no interest was to be charged. Such generosity was a mark of true godliness (see Psalm 15:1–5).
11. RESTORE NOW TO THEM: Nehemiah commanded those who had taken advantage of their brethren to restore all the property they had confiscated and to restore all the interest they had charged. He called the people to obey the Lord's commands, but he did not go beyond that to mete out punishment against those who had disobeyed. Once they had restored what they had taken, he considered the matter to be closed.
14. NEITHER I NOR MY BROTHERS ATE THE GOVERNOR'S PROVISIONS: Nehemiah was entitled to collect taxes from the people in Judah because he had been officially appointed as the king's representative. Yet he didn't. He refrained from taxing the Jews in order to serve as an example to them of self-sacrificial love.
16. I ALSO CONTINUED THE WORK ON THIS WALL: Nehemiah dealt with issues that came up, but his central focus was always on completing the work the Lord had called him to do. He was not there for personal profit, and he did not waste his time buying land or striving to further his career. He was there to build the wall . . . and build he did.
17. AT MY TABLE: Nehemiah's role as governor carried certain social obligations, just as any important political official is expected to entertain influential guests today. The costs were quite high for such obligations, and previous governors had not hesitated to make the Jews pay the bill through taxes. But Nehemiah met his obligations out of his own pocket, refusing to add to the financial burden of his fellow Jews.
> _A N OPEN LETTER: The enemies of the Jews have been unable to stop the work through the threat of military force, so now they turn to more subtle means._
6:2. SENT TO ME: This suggests that a messenger had delivered either a letter or an oral message to Nehemiah. Sanballat, Tobiah, Geshem, and the others had evidently realized they could not prevent Nehemiah's project from succeeding by open military engagement, so they now attempted to overcome him by deception.
IN THE PLAIN OF ONO: This plain was located south of Joppa on the western extremity of Judah along the seacoast.
3. SO I SENT MESSENGERS: Nehemiah knew his enemies were luring him into a trap, so he sent representatives to meet with them. This action itself was risky, as these representatives themselves might have been killed or imprisoned for ransom.
5. OPEN LETTER: Official letters were typically rolled up and sealed with an official signet by the letter's sender or one of his assisting officials. An open or unsealed letter was not only a sign of disrespect and open criticism but also suggested the information therein was public knowledge. The goal of this document was to intimidate Nehemiah into stopping the work.
6. IT IS REPORTED AMONG THE NATIONS: The letter suggested that Nehemiah's intent to revolt was common knowledge that would get back to the king of Persia if he did not come to the requested conference. This information, had it been true, would have brought Persian troops against the Jews. Even though Judah had a reputation for breaking its allegiances with its overlord kings, on this occasion that was not the case.
THAT YOU MAY BE THEIR KING: As we have seen, Artaxerxes had commissioned the rebuilding of the wall based on his relationship of trust with Nehemiah. Once the project was accomplished, the king expected Nehemiah to return to Susa. These allegations that Nehemiah was fortifying the city so he might be made king would seriously violate the Persian king's trust—if not create a war. The plot was an attempt to intimidate Nehemiah with the idea that a wedge was being driven between him and Artaxerxes. In this way they hoped to provoke Nehemiah to come to the meeting—which would have resulted in his death.
7. APPOINTED PROPHETS TO PROCLAIM: If there were such prophets, Sanballat had actually hired them to feed incorrect information, thus generating the false rumor. By dispatching such prophets to make public proclamations that Nehemiah had made himself king, he hoped to make it appear that the Jews were supplanting the Persian imperial rule.
10. SECRET INFORMER: When the letter failed to intimidate Nehemiah, his enemies decided to try intimidation from within. They hired a false prophet named Shemaiah to tell Nehemiah there was a plot against his life in an effort to lure him into the Holy Place in the temple. For Nehemiah to enter and shut himself in the Holy Place would have been a desecration of the house of God and would have caused people to question his reverence for God.
> _T HE WALL IS COMPLETED: In spite of all the enemies' tactics and strategies, the work on the wall continues and is actually completed in just fifty-two days._
15. TWENTY-FIFTH DAY OF ELUL: The wall was finished in August–September of 445 BC.
16. THIS WORK WAS DONE BY OUR GOD: While modern readers might be tempted to exalt the leadership qualities that brought the work to completion, Nehemiah viewed it as a result of the work God had done among them. God works through faithful people, but it is God who works.
17. THE NOBLES OF JUDAH SENT MANY LETTERS TO TOBIAH: Nehemiah added a footnote that in the days of building the wall, the nobles of Judah who refused to work were actually in alliance and correspondence with Tobiah, because—though his ancestors were Ammonites—he had married into a respectable Jewish family.
18. MANY IN JUDAH WERE PLEDGED TO HIM: According to Nehemiah 13:4, the high priest, Eliashib, was allied with Tobiah. The meddling of these nobles, by trying to play both sides through reports to Tobiah and to Nehemiah, had only widened the breach as Tobiah had escalated the efforts to frighten Nehemiah.
## UNLEASHING THE TEXT
1) What financial crisis befell the laborers in Jerusalem? What caused it? How had they become vulnerable to such a crisis?
[Your Response Here]
2) How did God expect the nobles to respond to the needs of their brethren? Why do you think the Jewish leaders did not follow God's Law?
[Your Response Here]
3) What tactics did Sanballat, Tobiah, Geshem, and the other enemies employ to stop the work? How did Nehemiah respond to each attack against him?
[Your Response Here]
4) In what ways did Nehemiah demonstrate the qualities of a strong, godly leader in this passage? What leaders have you known who were like him?
[Your Response Here]
## EXPLORING THE MEANING
**_God is compassionate, particularly toward the poor._** Many people often assume the Old Testament describes God as being wrathful and vengeful. However, the most common word used in the Old Testament to describe God is _compassionate_. God cares about people, and He particularly cares about the poor and needy who belong to Him. In this case, the poor who were being exploited were the ones who had thrown themselves wholeheartedly into the work of rebuilding the walls and had stood firm in spite of opposition.
These individuals had demonstrated a strong commitment to the Lord's work, and one would have expected their fellow Jews to be grateful. Unfortunately, they were not rewarded with gratitude but were exploited. The nobles and rulers who were taking advantage of them assumed that God simply would not notice or not care about this exploitation of the weak. These were powerful and influential men—with money and property—who had the capability of taking advantage of an opportunity and did not hesitate to do so. These men had forgotten that because God is compassionate, he hears the cry of the oppressed (see Exodus 22:27).
The Lord calls His people to be generous with those who are less fortunate and give freely with no expectation of repayment. Solomon wrote, "He who has pity on the poor lends to the LORD, and He will pay back what he has given" (Proverbs 19:17). Jesus gave us practical instruction on how to do this: "Give to everyone who asks of you. And from him who takes away your goods do not ask them back. And just as you want men to do to you, you also do to them likewise. . . . Love your enemies, do good, and lend, hoping for nothing in return; and your reward will be great, and you will be sons of the Most High. For He is kind to the unthankful and evil. Therefore be merciful, just as your Father also is merciful" (Luke 6:30–31, 35–36).
**_A godly leader leads by example._** Nehemiah was a great leader. He organized more than one trip from Persia to Judah, gathering all the materials and resources needed to undertake the mighty work of rebuilding the city walls. He also governed with wisdom, leading God's people through some difficult circumstances. But one of the strongest elements of his leadership was his own life, as he demonstrated through his own example what it meant to be a godly man.
When some of the people sinned and took advantage of their brothers, Nehemiah took a firm stand by confronting the wrongdoers and leading them back into obedience. The key word, however, is _leading._ Nehemiah did not _force_ them into submission but _led_ them into godliness through his own example. Nehemiah not only told the nobles and rulers to stop charging usury but also demonstrated the right way to give generously and lend freely. He set that example _before_ it became a problem among the people.
A good teacher does not simply give lectures on how to master a skill but also demonstrates that mastery by performing the skill in front of the students. In this way, the students can learn by _imitation_. Jesus did this to perfection with His disciples, both teaching them and setting an example for them to imitate. This is also the prime quality of a godly leader. The godly leader does not say, "Go forth and do it," but, " Watch me, and do the same."
**_Prayer should precede action._** Nehemiah was filled with anger when he learned his co-laborers had given their all for the Lord's work—sacrificing home, security, daily responsibilities, and even food—and been rewarded by being made paupers and sold into slavery. What's far worse, their own brethren had committed this crime! The nobles and rulers of the land were the very people who should have been setting an example of living by God's Word, but they were living more like the pagans around them for the sake of personal gain. If ever there was a time for righteous anger, this was it.
However, Nehemiah's first response was to give the matter some serious thought. He didn't need to think about whether their behavior was wrong, but he did need to think about what the Lord would have him do to correct that wrong. We can be confident, based on Nehemiah's responses to other crises, that he spent ample time on his knees before the Lord, perhaps even fasting before acting. He knew that if he responded in anger, he would be following the lead of his flesh—and he wanted to follow the lead of the Holy Spirit instead.
The psalmist wrote, "The LORD is gracious and full of compassion, slow to anger and great in mercy. The LORD is good to all, and His tender mercies are over all His works" (Psalm 145:8–9). Solomon gave us this practical advice: "A soft answer turns away wrath, but a harsh word stirs up anger. The tongue of the wise uses knowledge rightly, but the mouth of fools pours forth foolishness. . . . Better is a dinner of herbs where love is, than a fatted calf with hatred. A wrathful man stirs up strife, but he who is slow to anger allays contention" (Proverbs 15:1–2, 17–18). When anger arises—even righteous indignation—we must take time to seek the Lord in prayer and reflect on how He would have us respond.
## REFLECTING ON THE TEXT
5) Why did Nehemiah give serious thought to the situation with the nobles and rulers before addressing them? In practical terms, how is this done? Why is it important?
[Your Response Here]
6) What were "the governor's provisions" (Nehemiah 5:14)? Why did Nehemiah not take advantage of what was his by right? How did this influence his rebuke of the nobles?
[Your Response Here]
7) How did Nehemiah demonstrate both courage and wisdom in repelling all of the enemies' strategies against him? What do you learn from his example?
[Your Response Here]
8) When has someone taken advantage of your misfortunes? When have you been guilty of doing that to others? What did Jesus teach on this subject?
[Your Response Here]
## PERSONAL RESPONSE
9) When you find yourself in an emotionally charged situation, how do you usually respond? How does your normal response compare with Nehemiah's? How does it compare with Christ's response?
[Your Response Here]
10) What leadership roles do you hold (parent, teacher, boss, group leader)? Do you tend to lead more by example or by command? Explain.
[Your Response Here]
# **11**
# WORSHIPING GOD
# _Nehemiah 9:1–38_
## DRAWING NEAR
What are some of the ways today that people picture God? What qualities and attributes do they tend to assign to Him?
[Your Response Here]
## THE CONTEXT
Now that the walls of Jerusalem were completed, the people of Judah could finally settle back into their homes. In celebration, the people gathered together less than a week after finishing the walls to observe the Feast of Tabernacles (see Leviticus 23:33–44). At the end of that feast, the people gathered once more to worship God as a reunited nation, humbling themselves and acknowledging their dependence on Him.
The Israelites had certainly endured many trials. They had been attacked by enemies from without and put into slavery by greedy nobles from within. Yet they had endured because of God's protection and provision for them. In the end, we find Ezra's public reading of the Word of God pricking the hearts of these weary wall-builders. They are reminded of God's greatness and their own failure to adequately live for His glory. Together, as a people, they confess their sins and worship the Lord their God.
This final study will require some self-examination on our part, which can certainly be a humbling and painful experience. Yet if we are to fully understand God's grace and mercy, we must also understand our own condition as sinners. And if we are to praise Him for His love and mercy, we must also remind ourselves of all He has done on our behalf. Just as the people of the newly reconstituted nation understood this truth, so must we as followers of Christ.
## KEYS TO THE TEXT
Read Nehemiah 9:1–38, noting the key words and phrases indicated below.
> _T HE PEOPLE PREPARE FOR WORSHIP: The Jews assemble in Jerusalem for a time of national worship and repentance. Before they begin, however, they prepare themselves._
1. ON THE TWENTY-FOURTH DAY OF THIS MONTH: This event took place in the Jewish month of Tishri, which would be September–October of 445 BC.
WITH FASTING, IN SACKCLOTH, AND WITH DUST ON THEIR HEADS: These were outward demonstrations on the people's part of deep mourning and sadness for their sins. Their actions seem to have been performed in the spirit of the Day of Atonement, which was observed on the tenth day of this same month.
2. SEPARATED THEMSELVES FROM ALL FOREIGNERS: It was necessary to again call the Jews to divorce all lawful wives taken from among the heathen because the previous call—prompted thirteen years before by Ezra—had been only partially successful. Many among the people had escaped the required action of divorce and kept their pagan wives. Perhaps new defaulters had now also appeared and were confronted for the first time with this necessary action of divorce. Nehemiah's efforts were successful in removing this evil mixture.
3. READ FROM THE BOOK OF THE LAW . . . CONFESSED AND WORSHIPED: God's Word is the foundation of the worship process. Its truths confront the sin in our lives, and at the same time they reveal the character of God. Apart from God's Word there is no way to know Him, and so the rest of this prayer is built on the foundation of the study of the Word.
5. STAND UP AND BLESS THE LORD: The Levites led the people in a corporate outpouring of confession and worship. It was a time of national humility and public confession, set against the backdrop of God's great mercy and forgiveness. The end result of the three-hour worship service was a national promise of obedience to God in the future (see verse 38).
> _B EFORE THE BEGINNING: The Levites lead the people in worship chronologically through the history of Israel. They begin at the best place—before the beginning._
6. YOU ALONE ARE THE LORD: The worship service began with a fundamental truth: there is only one God, and He alone rules all things in heaven and on earth. It was important for the people to begin with this foundational understanding, for they had already been yielding to the temptation to embrace the beliefs of the world, which included a vast pantheon of make-believe gods. The Levites also were leading the people through a chronological remembrance of who God is and what He had done for them, so it was appropriate for them to begin before the world was even created, when God alone existed.
YOU HAVE MADE HEAVEN: This was a summary of God's creation, which encompasses not merely heaven and earth but also absolutely all that exists. God created everything, and He did so in six twenty-four-hour days. The Jews were reminding themselves that the world's teachings were utterly false concerning the origins of man.
7. WHO CHOSE ABRAM: It was the Lord God who had called Abraham, not Abraham who had sought out God. The Lord chose Abraham simply because He wanted to set His love on him, and He had been faithful ever since to the promises He had made (see Genesis 12; 17).
8. YOU FOUND HIS HEART FAITHFUL BEFORE YOU: Abraham had demonstrated a determination to obey the Lord's commands, as most clearly illustrated in his willingness to sacrifice his only son at God's request (see Genesis 22). Unfortunately, Abraham's descendants did not exhibit the same level of faithfulness to God's Word.
YOU HAVE PERFORMED YOUR WORDS: One of the themes of this song of praise is that God always keeps His promises, and He alone is righteous.
> _T HE EXODUS: The prayer now moves forward in time to the period when the Israelites were slaves in Egypt and follows them on their exodus to the Promised Land._
10. SIGNS AND WONDERS AGAINST PHARAOH: The Lord had sent a series of ten plagues against the people of Egypt for their hardness of heart in refusing to allow His people to leave. The final plague was the death of every firstborn son in every household within Egypt, except those that were covered by the blood of a lamb on the doorposts. This event led to the annual celebration of Passover, and it also pictured the eventual sacrifice for sin offered by Christ on the cross.
11. YOU DIVIDED THE SEA BEFORE THEM: The Lord allowed His people to encounter an immovable obstacle—the Red Sea—while witnessing their enemies hard on their heels behind them. The Lord miraculously parted the sea before them, allowed them to walk across on dry ground, and slammed the sea shut on the heads of those same powerful enemies (see Exodus 14). God sometimes leads us through difficult circumstances in order to demonstrate His omnipotent sovereignty in our lives.
12. CLOUDY PILLAR . . . PILLAR OF FIRE: The Lord made His presence known to the people of Israel night and day during their exodus from Egypt. During the day, He provided an immense cloud covering that shielded them from the hot desert sun. By night, He provided some sort of flaming element in the sky above, which allowed them to see in the utter darkness. These miraculous displays served the Israelites, but they also made known God's presence to the world around.
13. JUST ORDINANCES AND TRUE LAWS: The Lord also came among His people during their journey and taught them how to live according to His plan. The Law that God gave to Moses was designed to make the Jews noticeably different from the rest of the world, so that other nations would see the goodness and uniqueness of Israel's God.
15. BREAD FROM HEAVEN . . . WATER OUT OF THE ROCK: Both the manna from heaven (see Exodus 16) and the water from the rock (see Exodus 17) miraculously provided for the Israelites' physical needs in a dry and barren land. But they also provided small pictures of God's ultimate plan of salvation through His Son Jesus, who is the "bread of life" (John 6:35).
> _A M ISERABLE CONTRAST: The Levites now turn their attention from God's character to their own as they consider the ungodly behavior of their forefathers and themselves._
16. THEY AND OUR FATHERS ACTED PROUDLY: The comparison between God's character and the character of the people did not paint a pretty picture. Interestingly, the Levites used the same phrase "acted proudly" to describe Pharaoh and the Egyptians (see verse 10), which implies that God's chosen people had proved to be no different from the world. They had hardened their necks just as Pharaoh had hardened his heart and stubbornly refused to obey God's commands.
17. THEY WERE NOT MINDFUL OF YOUR WONDERS: The Lord had performed countless dramatic miracles to demonstrate His power and presence, yet the people had ignored them. It was not even that they were not paying attention and didn't notice, but that they had deliberately refused to acknowledge God's faithfulness. They had made themselves blind by choice.
THEY APPOINTED A LEADER: The Hebrew of this statement is almost a repeat of Numbers 14:4, which records the dissatisfaction of the people with God's plan and Moses' leadership.
BUT YOU ARE GOD: Two of the most joyful words in Scripture are _but God_. The Israelites had deliberately hardened their hearts, but God had remained faithful. They had sinned in every way, but God had remained ready to pardon them. They had wandered away from Him repeatedly, but God had not forsaken them. The entire human race deserves God's wrath, but God provides the free gift of salvation.
18. THEY MADE A MOLDED CALF FOR THEMSELVES: The human heart is constantly seeking to make gods from the material world (see Exodus 32)—the world that was created by the only true God. This trend has not changed in modern times.
20. YOUR GOOD SPIRIT TO INSTRUCT THEM: Christians can say this with an even greater significance, as God sends His Holy Spirit to reside in the life of every believer (see John 14:26).
21. THEY LACKED NOTHING: Even during an extended time of discipline for stubborn sin, the Lord still met the people's every need. They wandered in the wilderness for forty years, but their clothes and shoes never wore out!
> _P OSSESSION TO PRESENT: The Levites now move through the time when Israel took possession of the Promised Land and continue up to the present moment in Judah._
23. MULTIPLIED THEIR CHILDREN: Another aspect of the promise God made to Abraham was that he would be the father of a great nation (see Genesis 12:1–3). God told Abraham that his seed would be like the stars of heaven (see Genesis 15:5), and multiplication of the children of Israel in the land of Egypt had been nothing short of miraculous (see Exodus 1:1–3).
24. SUBDUED BEFORE THEM: Moses had once said, "The LORD is a man of war" (Exodus 15:3). As Israel's military leader and king, God had led them into battle to defeat their enemies and take the land.
26. NEVERTHELESS THEY WERE DISOBEDIENT: _But God_ are two joyful words in Scripture, but the word _nevertheless_ here is a sorrowful one. The Lord had poured out blessing after blessing on His people and demonstrated His power, sovereignty, faithfulness, and love. _Nevertheless_ , the people remained rebellious and disobedient. Humankind is no different today.
WHO TESTIFIED AGAINST THEM: God's prophets had brought the people into God's court to be judged by His law. This theme is repeated throughout the prayer (see verses 29, 30, 34).
**32.** NOW THEREFORE: The Levites, having reviewed the faithfulness of God to the Abrahamic covenant throughout Israel's history, now turn to the present. They begin by confessing the people's unfaithfulness to God.
KINGS OF ASSYRIA . . . THIS DAY: This statement sweeps across a summary of Assyrian, Babylonian, and Persian domination of the nation for almost four centuries up to that time.
33. YOU ARE JUST IN ALL THAT HAS BEFALLEN US: The people asked the Lord to help them in their present distresses, yet they still recognized those distresses had befallen them because of their sin. They did not blame God, but they also did not refrain from asking for His mercy.
**36.** HERE WE ARE, SERVANTS IN IT **:** The Levites rejoiced that the Jews had been returned to the land, but grieved that Gentiles still ruled over them.
37. MUCH INCREASE TO THE KINGS: Because God's people continued in widespread sin, enemy kings enjoyed the bounty that would have belonged to Israel.
38. BECAUSE OF ALL THIS, WE MAKE A SURE COVENANT: The history of God's faithfulness, in spite of Israel's unfaithfulness, formed the ground of a pledge and promise the people made to obey God and not repeat the sins of their fathers. A covenant was a binding agreement between two parties—a formalized relationship with commitments to loyalty. In this case, the nation of Israel initiated this covenant with God.
## UNLEASHING THE TEXT
1) What preparations did the people make before spending time in worship? What did these things mean? Why were they important? What parallels might they have in your life?
[Your Response Here]
2) Why did the Levites work chronologically through the history of Israel in their worship? How might this help a person recognize both God's goodness and his or her sin?
[Your Response Here]
3) Why did the worship begin with God's preexistence and role as creator of the universe? In what sense are these truths fundamental to an accurate understanding of His character?
[Your Response Here]
4) What miraculous provisions did God make for His people throughout Israel's history? What miraculous provisions has He made in your life?
[Your Response Here]
## EXPLORING THE MEANING
**_God is the creator and sustainer of the universe._** Before anything existed, God was present. He is beyond and apart from everything, self-sustaining and self-sufficient. Even His name "I AM" suggests this (Exodus 3:14). God created all things that exist in just six days, and created everything with the mere power of His Word: "God said . . . and there was" (Genesis 1:3). He called the stars into existence and fixed them in place, created the sun and moon to give light to the earth, and created the earth and the heavens to manifest the glory of His character.
But God did not create the universe and then walk away to do something else. He continues to maintain His creation and demonstrate His love for everything He made—especially people. This is what the Levites had in mind when they said to God, "You preserve them all" (Nehemiah 9:6). God preserves and sustains His entire creation. He controls all events in a constant unfolding of His perfect plan through His absolute sovereignty—a plan that He had in mind before the beginning, before the creation of the universe.
God is both eternal and immutable. He has always existed, with no beginning and no ending, yet He has not changed in any way since before creation. He is the same yesterday, today, and forever (see Hebrews 13:8). His plan to bring salvation to all people has been unfolding since before the beginning of time, and His plan to complete His work of salvation will continue through the end of the ages. This is one of the many reasons we join the Levites in finding God worthy of worship and praise!
**_God is gracious, even with the presumptuous._** The constant theme of the Levites' praise was the graciousness of God. In fact, the entire prayer is a litany of God's acts of compassion toward His people. Ever since the first sin plunged all of humanity into sin, God has been graciously working to bring peace and reconciliation to the earth. After Adam sinned, God told the serpent that from Eve's "seed" would come a Man who would crush Satan (see Genesis 3:15). From that point forward, history—much of which was recounted by the Levites in this chapter—has been leading up to that Person.
In the meantime, the world continued to rebel, and God continued to respond to each rebellion with a mixture of punishment, discipline, and graciousness. He destroyed the world, yet started over with one family. He scattered the nations when He shattered the Tower of Babel, but He then chose one new nation from whom He would bless all the families of the earth. When that new nation, Israel, rebelled against Him in the wilderness, God disciplined them but also guided them through their wanderings by way of a cloud and supernatural light.
God was faithful and gracious to His people during Moses' day, and He is still faithful and gracious to His people today. The difference is that now God has sent His Son in the ultimate act of His graciousness. Jesus came as a sacrifice for sin, and in the consummate act of kindness, He gave His life for people who hated Him and wanted Him dead. As a result, the news of the gospel is really the news that God's graciousness has conquered the hearts of presumptuous sinners. Jesus' grace is stronger than the hardest of sinners' hearts.
**_Praise God for His attributes._** Meditate on some of God's attributes below, and then spend time in prayer thanking God for who He is.
> **_Omniscience:_** God knows all things, and there is nothing hidden from Him (see Revelation 2:23).
> **_Alpha and Omega, Beginning and End_ :** God described Himself as "I AM WHO I AM" (Exodus 3:14; Revelation 1:8). He sustains all things in the universe that He created, and without Him nothing could exist.
> **_Ready to pardon:_** Jesus demonstrated this quality at the very point of His death, calling on the Father to forgive the ones who crucified Him (see Luke 23:34).
> **_Gracious and merciful:_** God pours out mercy and tenderness on His creation, even when we deserve His judgment (see Hebrews 2:17).
> **_Slow to anger:_** Throughout the ages, God's people have frequently put His grace to the test and persisted in stubborn sin despite His goodness—yet He has held back His hand of judgment (see Psalm 103:8).
> **_Keeps covenant and mercy:_** The Levites recognized this attribute of God's character as they worked their way through the history of Israel. The Lord's people frequently violated their covenant with God, but God never failed to keep it and to show mercy on them (see Nehemiah 9:32).
## REFLECTING ON THE TEXT
5) In what ways did Israel betray God throughout its history? Why did God continually forgive and restore the people? What does this reveal about God?
[Your Response Here]
6) Consider God's triune nature: Father, Son, and Holy Spirit. How is each person unique? How are all three perfectly united? How might these truths influence your worship?
[Your Response Here]
7) Why is worship important to individual Christians? Why is it important to the church as a whole? What is involved in worshiping the Lord?
[Your Response Here]
8) What are some other attributes of God's nature that you could add to the above list?
[Your Response Here]
## PERSONAL RESPONSE
9) What are some shortcomings and failures that you are willing to confess in your life? Spend some time in self-reflection to acknowledge your sins and repent of them.
[Your Response Here]
10) How do you see God's graciousness triumphing over sin in your own life?
[Your Response Here]
# **12**
# REVIEWING KEY PRINCIPLES
## DRAWING NEAR
As you look back at each of the studies, what is the one thing that stood out to you the most? What is one new perspective you have learned?
[Your Response Here]
## THE CONTEXT
During the course of these studies, we have watched as God's people made their way on the arduous move from Persia to Jerusalem, and we have considered the sacrifices they made to rebuild temple and the city walls. Indeed, many of the Jewish exiles had grown comfortable and prosperous in Persia and chose not to make the move back to Judah simply because the cost was too high. But there were others—like Ezra, Nehemiah, Haggai, and Zechariah—who did not hesitate to pay whatever cost was required to obey God's commands.
And hardship there was. God's people faced strong opposition from powerful enemies and endured mockery, sabotage, false friendship, and even death threats. Beyond that, the godly leaders came to realize that apart from the new covenant and the Messiah, the people would never be able to obey the Lord's commands. Yet through it all, God's character shines out. We are reminded at every turn that He is completely sovereign over all affairs of mankind, and we see again and again that He is faithful and true to His Word.
Here are a few of the major principles we have found during our study. There are many more we don't have room to reiterate, so take some time to review the earlier studies—or, better still, to meditate on the Scripture that we have covered. As you do, ask the Holy Spirit to give you wisdom and insight into His Word. He will not refuse.
## EXPLORING THE MEANING
**_This world is not our home._** The people of Judah had been carried away to captivity, but they had not been made slaves as they were many centuries before in Egypt. Instead, they were allowed to establish relatively normal lives within the new land, and many of the Jews had risen to levels of power and prosperity. Daniel, for example, served at least three different kings as a close personal counselor. Yet this relative freedom brought a danger the Israelites had not faced when they were slaves: _complacency_. Many of God's people had become quite comfortable in captivity and were fitting in to the society and doing well.
The problem was that God did not intend for His people to make their permanent home outside of Judah. Their real home was there, and the Lord did not want them to put down roots anywhere else. God's temple in Jerusalem was in ruins, the city's walls lay in rubble, and the Lord grieved over that situation. He wanted His people to share His priorities and to long to return to their proper land where they would worship and serve Him as He had ordained. The world in which they had grown so content was _not_ their home.
This is equally true for Christians today. This world is not our home! It is not wrong to pursue a career or to establish a home, but the Lord does not want us to lose our eternal focus. He wants us to remember that the things of eternity are what matter most, not the things of this world. Paul wrote, "Set your mind on things above, not on things on the earth. For you died, and your life is hidden with Christ in God" (Colossians 3:2–3). Paul was reminding us that by being born again into the salvation of Christ, we have died to the things of this world. And if we are dead to this world, there is no purpose in trying to make our home here. Our existence is with Christ in eternity, and that is where our focus needs to remain.
**_God commands us to resist discouragement._** The enemies of God's people attempted to interrupt their works of obedience by causing them to become discouraged. The _King James Version_ renders Ezra 4:4: "The people of the land weakened the hands of the people of Judah," which captures the essence of discouragement—to become weak, to sink down, to lose the ability to carry on, and to let God's projects drop from despair.
Fear is at the root of discouragement. We are suddenly faced with a circumstance that is beyond our control, and we quickly begin to fear that it is beyond _God's_ control as well. And if it is beyond God's control, we might as well just give up now—which is, of course, precisely what Satan is hoping that we will do! However, God commands us to not give in to fear but to strengthen our hands when they become weak (see Isaiah 41:10; Hebrews 12:12).
The best way for us to do this, wrote the author of Hebrews, is to "consider Him who endured such hostility from sinners against Himself, lest you become weary and discouraged in your souls" (12:3). We must remember that Jesus Himself faced immense opposition—more severe than any we will ever face. Yet He overcame it all through the faithfulness of God and through utter confidence in and reliance on God's sovereignty.
**_It is wise to consider your ways._** The Lord sent His prophet Haggai to bring one clear message to His people: obey God even when the circumstances are difficult. The prophet proclaimed, "Consider your ways!" (Haggai 1:5). This command might be translated, "Set your mind on your way of life." In other words, it was a call for the people to consider their ultimate priorities. In effect, God proclaimed, "Make _your_ priorities _My_ priorities!"
This is an important discipline that God's people need to exercise on a regular basis. The priorities and perspectives of the world are _not_ God's views, yet they have a way of seeping into our thinking without our even being aware of it. As Christians, we need to reassess our priorities and views daily, even moment by moment, to ensure that we are thinking the way God thinks. We do this by spending time daily studying and meditating on His Word, by seeking His wisdom through prayer, and by fellowshipping with other believers.
Paul warns his readers that this process is vital if we are to understand God's will for our lives. "I beseech you therefore, brethren, by the mercies of God, that you present your bodies a living sacrifice, holy, acceptable to God, which is your reasonable service. And do not be conformed to this world, but be transformed by the renewing of your mind, that you may prove what is that good and acceptable and perfect will of God" (Romans 12:1–2). Christians cannot hope to understand God's perfect will unless they are constantly renewing their minds by the Word of God. If we forget to consider our ways, we will wind up conformed to this world.
**_The Lord yearns for our fellowship._** The Lord told His people, "I am zealous for Zion with great zeal; with great fervor I am zealous for her" (Zechariah 8:2). Many modern Bibles translate "fervor" as "wrath," which captures the burning passion the Lord was expressing. "I burn for you with a jealous fire," the Lord was effectively saying, "a zealous and jealous love that will tolerate no competition." The Lord yearned so deeply for the fellowship of His people that it was like a consuming fire, destroying anything that prevented their full reconciliation.
It is interesting that the Lord referred to the zeal of His love in this context. The Jews had been zealous themselves in maintaining a cycle of fasts over a period of seventy years during their captivity. "Should I weep in the fifth month and fast as I have done for so many years?" they asked (Zechariah 7:3). Their question implied that they had not permitted anything to prevent them from keeping this religious observance, but the Lord called them to search their hearts and question why they were truly fasting. Were they mourning the loss of God's close fellowship, which they had once enjoyed as His chosen nation? Or were they mourning over their own suffering and misfortune? The two were not the same!
The Lord wanted His people to be zealous in seeking His face and entering His holy presence. Now, this might well be associated with certain spiritual activities, such as commemorating the Lord's Supper as Jesus commanded (see Luke 22), but those activities themselves are hollow and meaningless if one's heart is not fully committed to living in fellowship with God. The Lord has a burning zeal for our fellowship, for our company, and He wants us to have that same zeal for Him.
**_There are no shortcuts in obedience to God's Word._** Ezra made many preparations for the move to Jerusalem. He gathered a large body of Jews to join him, and each of those families made all the necessary preparations involved in making a life-changing move. The king had given Ezra his full blessing on the trip and provided him with a letter of authority to reestablish a Jewish community and rebuild the Lord's temple. The king had also handed Ezra a huge sum of money and treasure, and Ezra probably felt a sense of urgency to get that money where it belonged. A host of people and plans were ready to go. But then Ezra made the startling discovery that there were no members of the tribe of Levi with him on the journey.
Human wisdom would suggest a "work-around" measure at this point. So many people were standing around, waiting to get started, and the king's money was just sitting there waiting for theft. Surely, under such circumstances, prudence would dictate an "ad hoc" alternate plan. But Ezra refused to begin rebuilding the temple without the leadership and assistance of God's selected priests because God's Word commanded it. Ezra may have been caught by surprise, but he knew that the Lord was not. God wanted him to follow His prescribed methods, and He would take care of the timetable.
There are no shortcuts to obeying God's Word. The Lord does not call His people to find "work-arounds" or "emergency interim methods." His Word gives clear guidance in our daily lives, the correct approach to worship and church structure, roles of authority and submission, and much more that often goes contrary to what the world believes today. When it comes to the clear teachings in Scripture, there is no substitute for obedience.
**_Service is crucial, especially when it's costly._** Nehemiah lived in Persia, the greatest and wealthiest nation of its day. Furthermore, he lived in Susa, the nation's capital and one of the richest and most comfortable of the Persian cities. To top this off, he was the cupbearer to the king himself, a position of high trust and influence. He was undoubtedly rich and high in the ranks of the most powerful nation on earth. Jerusalem was far away—a journey of three full months. It would have been easy for Nehemiah to forget about his people.
Yet when he heard about the plight of his fellow Jews in far-off Judah, he mourned, wept, fasted, and prayed. What's more, he determined in his heart to forsake all the blessings and comforts of Persia and exchange them for hard work, rough living conditions, and constant hatred and opposition from God's enemies. In fact, Nehemiah was so determined to help with the work in Jerusalem that he risked his own life to get there. He put himself in peril of the king's wrath by making the request to leave the king's service, and then he embarked on the dangerous and uncomfortable trip to Judah.
It was not wrong or sinful for Nehemiah to enjoy the comforts of Persia and the king's court. The Lord had placed him in this position, and Nehemiah was being faithful to the tasks God had given him. But the Lord had placed him there specifically so that he might be positioned to help the Jews at this moment of crisis, just as He had placed Esther where she could save the Jews from annihilation a generation earlier (see Esther 4:14). The Lord was calling Nehemiah to voluntarily forsake all these blessings to participate in an important project, and the blessings that came from his obedience far surpassed all the comforts of the king's palace. When the Lord calls us in a similar way, we must always heed the call, for our obedience will bring glory to His name and great blessings to us.
**_We must be alert and prepared for battle._** Nehemiah didn't take the enemy's threat lightly when he learned they were planning to slaughter the Jews. He knew, of course, that the true protection of God's people lay solely in God's hands, but he also understood that he had a responsibility for the safety and welfare of those under his authority. Consequently, he took strong steps to prepare for the threat of battle. What's more, he expected all those working on the Lord's project to also remain alert and well armed. It must have been a real hindrance to the work of building, which was strenuous enough without having only one hand available, but Nehemiah felt the work of self-defense was as important as the work of rebuilding the wall.
Most of us do not face the threat of physical violence for our faith (though there are many Christians in parts of the world who do), but this principle applies at least as much on the spiritual level. Even if our neighbors are not threatening to attack us, we all face an enemy who is deadlier than those who opposed the Jews. For this reason, God commands us through His Word to be constantly on guard against the forces of wickedness, and He also commands us to go everywhere well armed. As Paul instructs, we are to carry with us "the sword of the Spirit, which is the word of God" (Ephesians 6:17). The writer of Hebrews gives this further detail: "For the word of God is living and powerful, and sharper than any two-edged sword, piercing even to the division of soul and spirit, and of joints and marrow, and is a discerner of the thoughts and intents of the heart" (Hebrews 4:12).
Well-armed and vigilant Christians spend time reading and meditating daily on the Word of God. And like all well-trained soldiers, believers in Christ also stay in close contact with the Commanding Officer through prayer and obedience. The military analogy, in fact, is apt in life, because we live in a battle zone where the enemy is constantly trying to destroy us. Therefore, we must always "be sober, be vigilant; because your adversary the devil walks about like a roaring lion, seeking whom he may devour. Resist him, steadfast in the faith, knowing that the same sufferings are experienced by your brotherhood in the world" (1 Peter 5:8–9).
**_God is the creator and sustainer of the universe._** Before anything existed, God was present. He is beyond and apart from everything, self-sustaining and self-sufficient. Even His name "I AM" suggests this (Exodus 3:14). God created all things that exist in just six days, and created everything with the mere power of His Word: "God said . . . and there was" (Genesis 1:3). He called the stars into existence and fixed them in place, created the sun and moon to give light to the earth, and created the earth and the heavens to manifest the glory of His character.
But God did not create the universe and then walk away to do something else. He continues to maintain His creation and demonstrate His love for everything He made—especially people. This is what the Levites had in mind when they said to God, "You preserve them all" (Nehemiah 9:6). God preserves and sustains His entire creation. He controls all events in a constant unfolding of His perfect plan through His absolute sovereignty—a plan that He had in mind before the beginning, before the creation of the universe.
God is both eternal and immutable. He has always existed, with no beginning and no ending, yet He has not changed in any way since before creation. He is the same yesterday, today, and forever (see Hebrews 13:8). His plan to bring salvation to all people has been unfolding since before the beginning of time, and His plan to complete His work of salvation will continue through the end of the ages. This is one of the many reasons we join the Levites in finding God worthy of worship and praise!
## UNLEASHING THE TEXT
1) Which of the concepts or principles in this study have you found to be the most encouraging? Why?
[Your Response Here]
2) Which of the concepts or principles have you found most challenging? Why?
[Your Response Here]
3) What aspects of "walking with God" are you already doing in your life? Which areas need strengthening?
[Your Response Here]
4) To which of the characters that we've studied have you most been able to relate? How might you emulate that person in your own life?
[Your Response Here]
## PERSONAL RESPONSE
5) Have you taken a definite stand for Jesus Christ? Have you accepted His free gift of salvation? If not, what is preventing you from doing so?
[Your Response Here]
6) What areas of your life have been most convicted during this study? What exact things will you do to address these convictions? Be specific.
[Your Response Here]
7) What have you learned about the character of God during this study? How has this insight affected your worship or prayer life?
[Your Response Here]
8) What are some specific things you want to see God do in your life in the coming month? What are some things you intend to change in your own life during that time? (Return to this list in one month and hold yourself accountable to fulfill these things.)
[Your Response Here]
If this is the first of these studies that you have completed, read the previous titles in this series. They will greatly enhance your knowledge of the Old Testament—not to mention your walk with God.
# Table of Contents
1. Cover Page
2. Title Page
3. Copyright Page
4. Ebook Instructions
5. Contents
6. Introduction
7. 1. Return from Exile: Ezra 1:1-11
8. 2. The Work Begins: Ezra 2:1-3:13
9. 3. Trouble With Samaritans: Ezra 4:1-5:17
10. 4. Choosing God's Priorities: Haggai 1:1-2:9
11. 5. Obedience Over Observance: Zechariah 7:1-8:23
12. 6. The Second Wave of Exiles: Ezra 7:1-8:36
13. 7. Unequally Yoked: Ezra 9:1-10:44
14. 8. Rebuilding the Wall: Nehemiah 1:1-2:20
15. 9. The Opposition Increases: Nehemiah 3:1-4:23
16. 10. Leading By Example: Nehemiah 5:1-6:19
17. 11. Worshiping God: Nehemiah 9:1-38
18. 12. Reviewing Key Principles
## List of Pages
1. i
2. ii
3. iii
4. iv
5. v
6. vi
7. vii
8. viii
9. ix
10. x
11. xi
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
|
{
"redpajama_set_name": "RedPajamaBook"
}
| 4,047
|
\section{Acknowledgments}
This material is based on research sponsored in part by NSF under grant No.\@ 2047899, DARPA under contract number HR001121C0168, and an Adobe Data Science Research Award.
\bibliographystyle{ACM-Reference-Format}
\section{Introduction} \label{sec:intro}
Estimating heterogeneous treatment effects (HTEs) is essential in many fields.
HTE estimation aims to find subpopulations whose causal effects differ from the effects of the population as a whole.
For example, if the treatment is a drug, some individuals may have adverse reactions, and some individuals may benefit from treatment~\cite{shalit-icml17}.
Similarly, public policy may affect different sociodemographic groups differently~\cite{grimmer-pa17}.
HTE analysis allows the discovery of these different subgroups of the population.
\begin{figure}
\centering
\newcommand{0.275}{0.3}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\node(c) at (0.0, -0.75) [label=right:$L$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (y);
\path (t) edge (c);
\path (y) edge (c);
\end{tikzpicture}
\caption{Collider}\label{fig:collider_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (m) at (0.0, 0) [label=below:$M$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (m);
\path (m) edge (y);
\end{tikzpicture}
\caption{Mediator}\label{fig:mediator_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (a) at (-1.5, 0) [label=left:$A$,point];
\node (t) at (-0.75, 0) [label={below left}:$T$,point];
\node (b) at (-0.75, -0.75) [label={below}:$B$,point];
\node (x) at (-0.75, 0.75) [label={above}:$X$,point];
\node (c) at (0.0, 0.75) [label={above}:$C$,point];
\node (d) at (0.0, 0) [label={below left}:$D$,point];
\node (e) at (0.0, -0.75) [label={below}:$E$,point];
\node (f) at (0.75, 0.75) [label={above}:$F$,point];
\node (y) at (0.75, 0.0) [label={right}:$Y$,point];
\node (g) at (0.75, -0.75) [label={below}:$G$,point];
\path (a) edge (t);
\path (x) edge (t);
\path (t) edge (b);
\path (x) edge (y);
\path(t)edge(d);
\path(c)edge(d);
\path(d)edge(e);
\path (d) edge (y);
\path (f) edge (y);
\path (y) edge (g);
\end{tikzpicture}
\caption{Multivariable}\label{fig:marloes_scm}
\end{subfigure}
}
\caption{Causal models where data-driven HTE estimation methods may perform poorly.
Some variables (e.g., $L$, $M$, and $E$) are not valid adjustment variables for estimating the effect of $T$ on $Y$.
}\label{fig:fail_examples}
\end{figure}
The accumulation of large amounts of data in many domains has stimulated the increased development of algorithms tailored for automated \emph{data-driven} HTE estimation, relying on supervised machine learning~\cite{athey-pnas16,johansson-icml16,grimmer-pa17,shalit-icml17,kennedy-arxiv20,kunzel-pnas19,tran-aaai19,athey-annals19}.
The goal of data-driven HTE estimators is to estimate the conditional average treatment effect using the features available in the data.
While it is appropriate to use all possible features for general supervised learning problems, unbiased estimation of \emph{causal effects} requires knowledge of the causal mechanisms underlying the data and selecting an appropriate subset of variables from the data~\cite{pearl-book09}.
One way to select this subset is by modeling the underlying causal mechanisms through structural causal models (SCMs)~\cite{pearl-book09} and selecting an \emph{adjustment set} --- a set of variables that are necessary for unbiased estimation of a causal effect of interest~\cite{pearl-book09,perkovic-uai15,henckel-arxiv19}.
However, the causal model structure is rarely known for real-world data, and data-driven methods do not take it into consideration.
Our paper addresses this problem and improves data-driven HTE estimation methods through causal feature selection under such structure uncertainty.
A common assumption used by data-driven HTE estimators is the assumption of \emph{strong ignorability} which states that the treatment assignment is independent of the potential outcomes given the variables in the data~\cite{rosenbaum-bio83,shalit-icml17}.
This untestable assumption implies that all potential confounders, variables that can influence both the treatment and the outcome, are observed in the data.
In turn, accounting for them would remove spurious associations and allow for causal effect identification.
However, hidden confounding is not the only problem leading to biased HTE estimation.
Even if all potential confounders are observed in the data, not all variables should be used for adjustment.
Another important and sometimes implicit assumption is that all variables in the data are \emph{pre-treatment} variables, i.e., the variables that were measured before and thus were unaffected by the treatment.
When the temporal order of observations is unknown, many practitioners would simply use all variables in the data~\cite{montgomery-ajps18}.
To illustrate the problem of using all variables in data-driven methods, consider the SCMs in Figure~\ref{fig:fail_examples} where $X$ is a confounding variable.
According to SCM theory, we need to adjust for \( X \) when estimating causal effects, but should not include variables that will bias effect estimation, such as descendants of the treatment or instrumental variables (e.g., \( L, M, A, B, D, E, G \)).
To address these deficiencies of data-driven methods, we define the problem of causal feature selection for accurate data-driven HTE estimation under structure uncertainty.
One vanilla solution to this problem is to first learn the causal structure from the data~\cite{henckel-arxiv19,spirtes-book00,ogarrio-pgm16} and then use an existing identification algorithm to find the adjustment set~\cite{shpitser-uai06,van-uai14,perkovic-uai15,henckel-arxiv19}.
However, structure learning is computationally expensive when the space of feature is large, and it may not work well when there is uncertainty in the learned structure~\cite{heinze-stat18,chickering-jmlr04,vowels-arxiv21}.
Instead, we develop a feature selection method that reduces this complexity by considering the value of each feature for improving HTE estimation and learning only the relevant parts of the causal structure from data.
We provide strong empirical evidence that our method improves existing data-driven HTE estimation methods under arbitrary underlying causal structures.
We evaluate the benefits of our feature selection algorithm in reducing HTE estimation error through synthetic, semi-synthetic, and real-world dataset experiments.
\section{Introduction}
Estimating heterogeneous treatment effects (HTEs) is an important problem in many fields.
The goal of HTE estimation is to find subpopulations for which causal effects differ from the population as a whole.
For example, if the treatment is a drug, some individuals may have adverse reactions, and some individuals may benefit from treatment~\cite{shalit-icml17}. Similarly, public policy may affect different sociodemographic groups differently~\cite{grimmer-pa17}.
HTE analysis allows the discovery of these different subgroups of the population.
Several supervised machine learning techniques have been developed for the problem of HTE estimation in recent years~\cite{athey-pnas16,johansson-icml16,grimmer-pa17,shalit-icml17,kennedy-arxiv20,kunzel-pnas19,tran-aaai19,athey-annals19}. We refer to these methods as \emph{data-driven} HTE estimators. Given a dataset, the goal of HTE estimators is to estimate the conditional average treatment effect using the features available in the data.
In supervised learning, algorithms use all possible features for accurate predictions.
In contrast, estimating causal effects requires knowledge of the causal mechanisms underlying the data and accounting for a correct subset of variables for accurate and unbiased causal effect estimation~\cite{pearl-book09}.
\begin{figure}
\centering
\newcommand{0.275}{0.275}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\node(c) at (0.0, -0.75) [label=right:$L$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (y);
\path (t) edge (c);
\path (y) edge (c);
\end{tikzpicture}
\caption{Collider}\label{fig:collider_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (m) at (0.0, 0) [label=below:$M$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (m);
\path (m) edge (y);
\end{tikzpicture}
\caption{Mediator}\label{fig:mediator_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (a) at (-1.5, 0) [label=left:$A$,point];
\node (t) at (-0.75, 0) [label={below left}:$T$,point];
\node (b) at (-0.75, -0.75) [label={below}:$B$,point];
\node (x) at (-0.75, 0.75) [label={above}:$X$,point];
\node (c) at (0.0, 0.75) [label={above}:$C$,point];
\node (d) at (0.0, 0) [label={below left}:$D$,point];
\node (e) at (0.0, -0.75) [label={below}:$E$,point];
\node (f) at (0.75, 0.75) [label={above}:$F$,point];
\node (y) at (0.75, 0.0) [label={right}:$Y$,point];
\node (g) at (0.75, -0.75) [label={below}:$G$,point];
\path (a) edge (t);
\path (x) edge (t);
\path (t) edge (b);
\path (x) edge (y);
\path(t)edge(d);
\path(c)edge(d);
\path(d)edge(e);
\path (d) edge (y);
\path (f) edge (y);
\path (y) edge (g);
\end{tikzpicture}
\caption{Multivariable}\label{fig:marloes_scm}
\end{subfigure}
}
\caption{Causal models where data-driven HTE estimation methods may perform poorly.
Some of the variables (e.g., $L$, $M$, and $E$) are not valid adjustment variables for estimating the effect of $T$ on $Y$.
}\label{fig:fail_examples}
\end{figure}
Structural causal models (SCMs) are a class of graphical models that model such causal mechanisms~\cite{pearl-book09}.
SCMs enable the selection of an \emph{adjustment set} -- a set of variables that are necessary for unbiased estimation of a causal effect of interest.
Consider the SCMs in Figure~\ref{fig:fail_examples} where $X$ is a confounding variable.
\new{SCM theory tells us that we need to adjust for \( X \) when estimating causal effects, but should not include variables that will introduce bias in effect estimation, such as descendants of the treatment or instrumental variables~\cite{myers-aje11} (e.g., \( L, M, A, D \)).
}
\new{However, SCMs are rarely known in real-world settings, so the correct adjustment variables cannot be found.
This makes data-driven HTE estimators an attractive alternative for practical applications.}
A common assumption used by data-driven HTE estimators is the assumption of \emph{strong ignorability} which states that the treatment assignment is independent of the potential outcomes given the variables in the data~\cite{rosenbaum-bio83,shalit-icml17}.
This assumption implies that there is no hidden confounding and HTE can be identified based on the given data.
However, hidden confounding is not the only problem that can lead to biased HTE estimation.
\new{
In practice, data-driven methods typically consider variables to be \emph{pretreatment} features, features which are measured before treatment.
However, this requires knowledge of the temporality of measurements, which may not be available.
Features in the data may actually be mediating variables, which should not be included in estimation.
}
To address the deficiencies of data-driven methods, we study the problem of causal feature selection for accurate data-driven HTE estimation when the SCM is unknown.
We develop two classes of algorithms for \emph{causal feature selection}, one based on local structure learning and another based on heuristic HTE evaluation metrics, neither of which has been studied previously in the context of HTE feature selection.
Through synthetic dataset experiments, we show that our feature selection algorithms improve existing data-driven methods and reduce error in causal effect estimations.
\section{Related Work}
We provide a brief overview of related work on estimating heterogeneous treatment effects and structure learning.
\textbf{Data-driven methods for HTE estimation.}
Many methods have been developed for estimating heterogeneous treatment effects (HTEs) from data.
Several methods rely on recursive partitioning using tree-based methods, which naturally partition the population into subgroups that may contain heterogeneous effects~\cite{athey-pnas16,athey-annals19,tran-aaai19,su-jmlr09,zeileis-jcgs08}.
Besides tree-based methods, methods have been developed that use machine learning to predict causal effects, such as meta-learners~\cite{kunzel-pnas19,kennedy-arxiv20} and neural networks~\cite{johansson-icml16,shalit-icml17,schwab-perfect18}.
However, these methods do not consider the structure of the underlying causal mechanisms that generate the data, introducing bias in estimation.
\textbf{Causal structure learning.}
Many causal structure learning algorithms have been developed for recovering SCMs from data~\cite{heinze-stat18,vowels-arxiv21}.
Heinze-Deml et al.\ provide a recent survey on causal structure learning algorithms and evaluate their performance in various settings~\cite{heinze-stat18}.
However, learning the full causal structure from data can be computationally expensive for high-dimensional datasets~\cite{chickering-jmlr04,vowels-arxiv21}.
Related to causal structure learning is local structure learning~\cite{aliferis-jmlr10}.
Aliferis et al.\ presented an algorithmic framework for learning local causal structure around target variables, the parent-children set (\( PC \)) and the Markov Blanket (\( MB \)).
Yu et al.\ survey and evaluate several local causal structure learning methods that learn both \( PC \) and \( MB \) of target variables~\cite{yu-arxiv19}.
They study local structure learning under the context of causality-based feature selection with the goal of improving supervised learning.
In contrast, we study the problem of causal feature selection.
Local structure learning has been studied in the context of average treatment effect (ATE) estimation, but not HTE estimation~\cite{cheng-arxiv20}.
This work also does not consider excluding mediators or descendants, which we do in our work.
\section{Background}\label{sec:background}
We provide an introduction to heterogeneous treatment effects under the potential outcomes framework~\cite{rubin-74} and structural causal models~\cite{pearl-book09}.
\subsection{Heterogeneous treatment effects}
Let $(\mathbf{X}_i, Y_i(0), Y_i(1), T_i) \sim \mathcal{P}$ be a distribution with $N$ units which are independently and identically distributed (i.i.d.).
For any unit $i$, define $\mathbf{X}_i \in \mathbb{R}^d$ be a $d$-dimensional feature vector,
$T_i \in \{0, 1\}$ to be the treatment assignment indicator, $Y_i(0)$ to be the potential outcome when $i$ is not treated, and $Y_i(1)$ to be the potential outcome when $i$ is treated.
The individual treatment effect (ITE) is the difference in potential outcomes: \( \tau_i = Y_i(1) - Y_i(0) \).
However, both potential outcomes $Y_i(1)$ and $Y_i(0)$ cannot be observed at the same time and observed data $\mathcal{D} = (\mathbf{X}_i, Y_i, T_i)$ contains only $Y_i = Y_i(T_i)$. This is known as the fundamental problem of causal inference.
Since $\tau_i$ can never be measured directly, its estimate is expressed through the conditional average treatment effect (CATE):
\begin{equation}
\label{eq:cate}
\hat{\tau}(\mathbf{x}) = E[Y(1) - Y(0) \mid \mathbf{X}_i].%
\end{equation}
The average treatment effect (ATE) of the population is: $\text{ATE} = E[Y(1) - Y(0)]$. If CATE and ATE are not the same, then $\mathbf{X}_i$ induce heterogeneous treatment effects (HTE).
Given an observed dataset $\mathcal{D}$, the goal of an \emph{HTE estimator} is to estimate CATE, $\hat{\tau}$, using the available features $\mathbf{X}$ in the data.
HTE estimators estimate CATE in a variety of ways.
Some HTE estimators rely on recursive partitioning using tree-based methods~\cite{tran-aaai19,athey-pnas16,athey-annals19}.
Other HTE estimators use predictions from machine learning methods to train a causal effect estimator~\cite{kunzel-pnas19,kennedy-arxiv20} or to predict counterfactual outcomes for estimation of effect~\cite{shalit-icml17,johansson-icml16}.
A common assumption used by data-driven HTE estimators is the assumption of \emph{strong ignorability} which states that the treatment assignment is independent of the potential outcomes given a set of variables $\mathbf{X}$ ~\cite{rosenbaum-bio83,shalit-icml17}: $Y_1, Y_0 \perp T \mid \mathbf{X}$, and $0 < p(T=1 \mid \mathbf{X}) < 1$ for all $\mathbf{X} = \mathbf{x}$.
One of the implications of this assumption is that there is no \emph{hidden confounding}.
However, hidden confounding is not the only problem that can lead to biased HTE estimation. In practice, it is often assumed that all observed variables in the data are the set of variables $\mathbf{X}$ that meets strong ignorability.
This is problematic because some of the variables may make $T$ and $Y$ dependent or introduce new spurious dependencies~\cite{pearl-book09}.
Some variables, such as mediators, should not be included in the estimation. Moreover, when the temporal order of variables is unknown, practitioners may consider all features in the data as pre-treatment variables~\cite{montgomery-ajps18}.
Selecting the right variables is the goal of our work.
\subsection{Structural Causal Models}
A structural causal model (SCM), $M$, consists of two sets of variables, $\mathbf{U}$ and $\mathbf{V}$, and a set of structural equations, $\mathbf{F}$, describing how values are assigned to each \emph{endogenous} variable $V_i \in \mathbf{V}$ based on the values of $\mathbf{v}$ and $\mathbf{u}$~\cite{pearl-book09}: \( v_i = f_i(\mathbf{v}, \mathbf{u}) \).
A causal graph \( G \) is a directed acyclic graph that captures the causal relationships among the variables.
The variables in $\mathbf{U}$ are considered \emph{exogenous}.
It is assumed that every endogenous variable has a parent in $\mathbf{U}$.
The causal graph may omit representing $\mathbf{U}$ explicitly.
We denote the parents, children, ancestors, and descendants of $X$ in the graph as: \(pa(X), ch(X), an(X), de(X)\).
Every node is a descendant and ancestor of itself.
When the SCM underlying the dataset is known, its graph G supports the use of graphical criteria for variable selection for unbiased causal effect estimation, known as adjustment set selection~\cite{shpitser-uai06,shpitser-uai10,van-uai14,perkovic-uai15,henckel-arxiv19}.
One such criterion is the backdoor criterion~\cite{pearl-book09}, which aims to find a set that blocks the paths from treatment to the outcome, starting with an arrow into treatment.
For example, in Figure~\ref{fig:collider_scm}, there is one backdoor path, namely, $ T \leftarrow X \rightarrow Y $, and so the set $ \{ X \} $ sufficiently satisfies the backdoor criterion.
Several adjustment criteria have been developed, such as the adjustment criterion~\cite{shpitser-uai06,van-uai14}, generalized back-door criterion~\cite{maathuis-annals15}, generalized adjustment criterion~\cite{perkovic-uai15}, and optimal adjustment set (O-set)~\cite{henckel-arxiv19}.
These adjustment sets defined by the various adjustment criteria can be found through the ID algorithm~\cite{shpitser-jmlr08}, the framework by Van der Zander et al.~\cite{van-uai14} or the pruning procedure by Henckel et al.~\cite{henckel-arxiv19}.
Once an adjustment set is found, they can be used with an HTE estimator to estimate heterogeneous effects.
When the SCM for a dataset is unknown, and therefore there is \emph{structure uncertainty} of the causal mechanism, the adjustment criteria cannot be applied.
An alternative approach is to use causal structure learning~\cite{heinze-stat18} to learn a causal graph from the data and then apply an adjustment criterion for feature selection.
However, learning the full structure is computationally expensive for high-dimensional datasets~\cite{chickering-jmlr04,heinze-stat18,vowels-arxiv21}.
Instead, we propose reducing the feature space by assessing the fit of each variable for HTE estimation and learning the local structure, which we describe in Section~\ref{sec:cfs}.
\subsubsection{HTEs in SCMs}
In SCMs, heterogeneous effects are represented by an \emph{interaction} or \emph{effect modification} through ancestors of $Y$~\cite{pearl-book09}.
For example, suppose $Y$ is generated by the following structural equation corresponding to the SCM in Figure~\ref{fig:marloes_scm}:
\begin{equation}
Y = \alpha \cdot T + \beta \cdot X + \gamma \cdot T \cdot f(X) + U_Y.
\end{equation}
Here, there is a fixed effect \( \alpha \) of the treatment \( T \) and a heterogeneous effect \( \gamma \) due to the interaction between \( T \) and \( X \).
Interaction can occur with any parent of \( Y \) (e.g., \( X, F \) in Figure~\ref{fig:marloes_scm}) or any parent of a mediator (e.g., \( C \) in Figure~\ref{fig:marloes_scm})~\cite{vanderweele-epid07}.
\subsubsection{C-specific effects}
\new{
Shpitser and Pearl provide a nonparametric estimand for HTEs with respect to some feature assignment $C=c$, called $c$-specific effects~\cite{shpitser-uai06,pearl-smr17}.
}
One issue with estimating $c$-specific effects is that it requires knowledge of the SCM and which variables, C, exhibit heterogeneous effects.
When the SCM is known, the O-set~\cite{henckel-arxiv19} can be used on the learned structure since it implicitly includes all potential heterogeneity-inducing variables, while some identification algorithms may miss them. When the SCM is unknown, the identifiability of \( c \)-specific effects cannot be established.
Moreover, when the set of heterogeneity-inducing variables $\mathbf{C}$ is unknown, then all combinations of features and feature values would need to be checked for heterogeneity in effect, which is exponential in the number of valid features.
Circumventing this exponential search is one of the main reasons why data-driven methods are popular in practice.
\section{Problem Setup}
Our goal in this work is to augment data-driven methods to accurately predict heterogeneous effects when both the SCM and the variables that explain heterogeneity are unknown.
More specifically, the goal is to find a subset $\mathbf{X^{(\ell)}}$, to which we refer as the \emph{causal feature set{}}, that improves data-driven estimation.
In general, the goal in HTE estimation is to obtain accurate estimation of CATE:
\begin{problem}\label{prob:hte_problem}
(Heterogeneous treatment effect (HTE) estimation)
Given a dataset $\mathcal{D}=(\mathbf{X}_i, Y_i, T_i)$ of N instances with some true ITE $\mathbf{\tau}$, estimate CATE $\hat{\tau}(\mathbf{X})$ such that the mean squared error is minimized:
\begin{equation}\label{eq:cate_objective}
\frac{1}{N} \sum_{i=1}^N (\tau_i - \hat{\tau}(\mathbf{X}_i)) ^ 2.
\end{equation}
\end{problem}
Since the true ITE is unknown in real-world scenarios, several heuristic HTE evaluation metrics have been developed for evaluating the performance of HTE estimators~\cite{schuler-model18,schwab-perfect18,saito-cv19}.
Instead of minimizing the true estimator error, we can instead minimize one of these metrics, such as the \( \tau \)-risk~\cite{schuler-model18,nie-rlearner17}.
A \emph{causal feature set{}} is the feature set that minimizes the error in HTE estimation.
\begin{problem}\label{prob:causal_feature_selection}
(Causal feature selection (CFS) for HTE estimation) Given a dataset $\mathcal{D}=(\mathbf{X}_i, Y_i, T_i)$ and a data-driven estimator $\hat{\tau}(\mathbf{X})$, find a \emph{causal feature set{}} $\mathbf{X}_i^{(\ell)} \subseteq \mathbf{X}_i$ such that $\hat{\tau}(\mathbf{X}^{(\ell)})$ results in the lowest MSE:
\begin{gather}
\argmin_{\mathbf{X^{(\ell)}}} \frac{1}{N} \sum_{i=1}^{N} {\Big( \tau_i - \hat{\tau}\big( \mathbf{X}^{(\ell)}_i \big) \Big)}^2.
\end{gather}
\end{problem}
\noindent A data-driven estimator typically estimates $\hat{\tau}(\mathbf{X})$ using all available data.
Our goal in this work is to select a causal feature set{} $\mathbf{X}^{(\ell)}$ such that the error of the estimator is minimized.
In this work, we assume that it is unknown a priori whether all observed features in the data are pre-treatment variables and whether all of them meet the strong ignorability assumption.
Instead, we use the following more general assumptions:
\begin{enumerate}
\item Causal sufficiency: there are no latent variables.
\item Causal faithfulness: if $X_A, X_B$ are conditionally independent given $\mathbf{X_Z}$, then $A,B$ are d-separated in the causal graph by $\mathbf{Z}$.
\item Acyclicity: the underlying causal graphs are directed acyclic graphs (DAGs).
\item There is no selection bias.
\end{enumerate}
Understanding how our methods need to be adjusted when these assumptions do not hold is left for future work.
\section{Causal Feature Selection for HTE Estimation}\label{sec:cfs}
When the SCM that generated the data is unknown, a vanilla solution for finding an adjustment set for HTE estimation is to learn the causal structure and then find the optimal adjustment set (O-set)~\cite{henckel-arxiv19}.
However, structure learning is computationally expensive in high-dimensional data, both in theory~\cite{chickering-jmlr04} and in practice~\cite{heinze-stat18,vowels-arxiv21}.
Instead of doing full structure learning, we propose Heterogeneous Treatment Effect Feature Selection (HTE-FS){}, which relies on two ideas to make feature selection more practical.
First, instead of considering all features, it assesses the value of each feature in reducing the HTE estimation error and selects only the features that contribute to error reduction.
Second, instead of learning the full structure, it focuses on learning the parts of the structure that give sufficient information about selecting a valid adjustment set.
Figure~\ref{fig:glcfs_fig} shows a high-level overview of HTE-FS{}, and we present its main components, HTE-Fit{} and Structure-Fit{}, next.
\input{04_01_heuristic.tex}
\input{04_02_local.tex}
\input{04_03_global_local.tex}
\subsection{HTE-Fit{}}
The first component of our framework, HTE-Fit{}, proposes the use of \emph{heuristic HTE evaluation} metrics to sequentially select features that are likely to reduce the HTE estimation error.
This is shown as the first step in Figure~\ref{fig:glcfs_fig}.
It has two variants: forward (\globalmethodname{}-F{}) and backward (\globalmethodname{}-B{}).
Since the ground truth of HTE effects in unknown, several heuristic HTE evaluation metrics have been developed to approximate the errors of HTE estimators, including $\tau$-risk~\cite{schuler-model18}, Nearest Neighbor Precision in Estimating Heterogeneous Effects (NN-PEHE)~\cite{schwab-perfect18}, Plug-in $\tau$~\cite{saito-cv19}, and Counterfactual Cross-validation (CFCV)~\cite{saito-cv19}.
The $\tau$-risk takes as input the estimator $\hat{\tau}$ and approximates the objective defined in~\eqref{eq:cate_objective}~\cite{schuler-model18}.
The second metric, NN-PEHE, uses nearest neighbor matching for imputing treatment effects for training~\cite{schwab-perfect18}.
The plug-in $\tau$ uses an arbitrary estimator for imputing effects for computing error.
The fourth metric, CFCV, uses a doubly robust formulation with the plug-in $\tau$ for imputing effects for evaluation~\cite{saito-cv19}.
These metrics have been shown to correlate with HTE estimator performance~\cite{schuler-model18,saito-cv19}.
Here, we propose to use these metrics as causal feature selection criteria, which, to the best of our knowledge, has not been studied in previous research.
\subsubsection{\globalmethodname{} --- Forward{}}
\globalmethodname{}-F{} starts with a feature with the best (lowest) error metric, then iteratively adds features that improve the metric.
If there are no more features or improvements, the selection stops.
We will use $\tau$-risk as an example metric to make this procedure more concrete. $\tau$-risk has been shown to lead to a consistent high-performing model~\cite{schuler-model18} and to have quasi-oracle error bounds on the CATE~\cite{nie-rlearner17}.
The $\tau$-risk, denoted as $R$, takes as input the estimator $\hat{\tau}$ and approximates the objective defined in~\eqref{eq:cate_objective}~\cite{schuler-model18}:
\begin{equation}\label{eq:tau_risk}
\tau\text{-risk}(\hat{\tau}) = R(\hat{\tau}) = \frac{1}{N} \sum_{i=1}^N \Big(
(Y_i - \hat{m}(\mathbf{X}_i)) - (T_i - \hat{p}(\mathbf{X}_i)) \hat{{\tau}}(\mathbf{X}_i)
\Big)^2.
\end{equation}
Here, \( \hat{m}(\mathbf{X}_i) \) is an outcome estimate for unit $i$ (i.e., \( E[Y_i | \mathbf{X}_i] \)), and \( \hat{p}(\mathbf{X}_i) \) is an estimate of the propensity score, (i.e., \( E[T_i | \mathbf{X}_i] \)).
Both $\hat{m}$ and $\hat{p}$ are learned through supervised techniques and optimized for predictive accuracy.
Suppose the feature vector of individuals has $d$ dimensions, $\mathbf{X} = (X_1, \dots, X_d)$.
Given the $\tau$-risk, we can compute the feature fit for one feature, say $X_j \in \mathbf{X}$, by estimating $\hat{\tau}(\{X_j\})$ and inputting the estimator into the $\tau$-risk.
For \globalmethodname{}-F{}, the first feature found is the feature that minimizes $R(\hat{\tau}(\{X_j\})$, and we define the initial set of found features as $\mathbf{X}^{(g)} = \{X_j\}$.
Now that the first feature is found, we need to add another to $\mathbf{X}_{(g)}$.
For each feature $X_k \in \mathbf{X}, X_k \notin \mathbf{X}^{(g)}$, we estimate a new $\hat{\tau}(\mathbf{X}^{(g)} \cup \{X_k\})$ which is the current set of found features $\mathbf{X}^{(g)}$ combined with a new feature.
The next feature added to $\mathbf{X}^{(g)}$ is the feature that minimizes $R(\hat{\tau}(\mathbf{X}^{(g)} \cup \{X_k\})$.
This process continues until there is no feature that, when combined with $\mathbf{X}^{(g)}$, minimizes the $\tau$-risk or when there are no more features to add.
\subsubsection{\globalmethodname{} --- Backward{}}
Backward feature selection (\globalmethodname{}-B{}) works similar to forward feature selection.
The difference is that it starts with all features, $\mathbf{X}^{(g)} = \mathbf{X}$, and it
iteratively removes features until either the error metric does not improve or there is only one feature left for estimation.
Using $\tau$-risk, it would first compute $R(\hat{\tau}(\mathbf{X}))$ to get an initial score on the entire feature set.
For each $X_j \in \mathbf{X}$, it would compute $R(\hat{\tau}(\mathbf{X} \setminus \{ X_j \})$, and remove the feature that results in the lowest $\tau$-risk when removed.
\subsection{Structure-Fit{}}
Now that we have selected the features that can contribute to HTE estimation error reduction, the next step is to select only the features which are a valid adjustment set. The second part of our framework,
Structure-Fit{}, iteratively applies local structure learning~\cite{aliferis-jmlr10,yu-arxiv19} to find a path from the treatment, $T$, to the outcome $Y$, in order to remove mediators and descendants of treatment.
Step 2 of Figure~\ref{fig:glcfs_fig} represents this part, Structure-Fit{}, where local structures from $T$ to $Y$ are learned iteratively, and connected. A local structure is a subgraph of the causal graph centered on a target variable.
For example, in Figure~\ref{fig:glcfs_fig}, the local structure of $T$ are the edges and variables connected to $T$ (e.g., $X_r^{(g)}, X_p^{(g)}, X_j^{(g)}$).
Since $Y$ is the outcome variable and we are only interested in paths from $T$ to $Y$, Structure-Fit{} terminates and the final feature set is found by removing mediators and descendants (crossed out in red).
This approach has two main advantages compared to full causal structure learning.
First, it only considers local structure around causal nodes and does not need to consider all variables if they are not connected to the causal path from \( T \) to \( Y \).
Second, it aims to only select variables which do not bias causal effect estimation and allow the HTE estimator to use features as appropriate for estimation.
\subsubsection{Discovering local structure around target variables}
Multiple structures can encode the same conditional independences in the data, making the task of finding the parent sets of $T$ and $Y$, $pa(T), pa(Y)$, nontrivial.
For example, $X_1 \rightarrow T \rightarrow X_2, X_1 \leftarrow T \leftarrow X_2$, and $X_1 \leftarrow T \rightarrow X_2$, all encode that $X_1 \perp X_2 | T$ and in one of them $X_1$ is a parent of $T$ whereas in the other two $X_1$ is a child.
We use two different ways to find the local structure.
The first way relies on finding the parent-children ($PC$) set around our target variables, $T$ and $Y$~\cite{aliferis-jmlr10}, collider discovery, and edge orientation methods to find causal directions~\cite{daniusis-uai12,fonollosa-book19,blobaum-aistats18}.
Several $PC$ discovery algorithms have been developed~\cite{buhlmann-bio10,pena2005scalable,aliferis-jmlr10,yu-arxiv19} and have mainly been used for improving classification methods~\cite{yu-arxiv19}.
To distinguish between parents and children, we first discover the parents of each target variable through collider discovery.
If two variables are parents of the target variable (e.g., \( X \) and \( A \) are parents of \( T \)), then they are dependent if conditioning on the child and can be discovered through conditional independence tests.
However, if there is zero or one parent, we will not leverage collider discovery.
Instead, we orient the remaining edges using an edge orientation algorithm~\cite{daniusis-uai12,fonollosa-book19,blobaum-aistats18} and then check if there is a parent in the set.
Using these algorithms, we can check for any variable in $PC$ whether an arrow is pointing in or away from the target variable (i.e., $T$ and $Y$) and thus identify $pa(T), pa(Y), ch(T)$, and $ch(Y)$.
The second way to find local structures is to utilize more recent local structure learning algorithms, such as PCD-by-PCD~\cite{yin-causation08}, MB-by-MB~\cite{wang-csda14}, and Causal Markov Blanket~\cite{gao-neurips15}
These algorithms discover the $PC$ or Markov Blanket ($MB$) while partially orienting edges.
We can find a partially oriented local structure and orient edges using edge orientation methods.
\subsubsection{Learning partial structures}
Using the process to discover local structure around target variables, we propose an iterative algorithm to remove post-treatment variables from HTE-Fit{}, which we call Structure-Fit{}.
We apply local structure learning on the treatment variable, $T$, and then iteratively apply local structure learning on children of $T$ in a breadth-first-search manner until we reach the outcome variable $Y$.
Then, local structures are put together to form a subgraph of the causal graph over the variables selected by HTE-Fit{}.
After learning the local structure, we can apply an identification algorithm to find an adjustment set, such as the O-set~\cite{henckel-arxiv19}.
Another option is to only remove all descendants of treatment.
We demonstrate Structure-Fit{} using Figure~\ref{fig:marloes_scm}.
We first find $PC(T)=\{A, X, B, D\}$ and $ch(T)=\{B, D\}$.
We proceed to iteratively find the $PC$ on each child, $B, D$.
Since $B$ has no children, we move to $D$ and find $PC(D)=\{T, C, E, Y\}$ and $ch(D)=\{E, Y\}$.
Finally, $E$ has no children, and we have reached the outcome variable, $Y$ and find $PC(Y)=\{X, C, F, G\}$.
Using the local structures, we find the causal graph shown in Figure~\ref{fig:marloes_scm}.
In practice, we only use the output of HTE-Fit{} in Structure-Fit{}, so not all variable will be used.
\section{Experiments}
To evaluate whether and how much feature selection improves existing data-driven HTE algorithms, we generate synthetic datasets which contain ground-truth causal effects.
We also use a semi-synthetic dataset and two real-world datasets.
\subsection{Experimental setup}
In our experiments, we make a distinction between four classes of estimation methods: \emph{base learners}, \emph{baseline feature selection}, \emph{our feature selection}, and \emph{oracle}.
A base learner is an HTE estimator without feature selection, and it uses all features for estimation.
We consider several prominent HTE estimators as base learners:
Causal Tree Learn~\cite{tran-aaai19}, Causal Forest~\cite{athey-annals19},
Meta-learners (S, T, XLearners)~\cite{kunzel-pnas19}, Doubly Robust learner (DRLearner)~\cite{kennedy-arxiv20}, Balancing Neural Network (BNN)~\cite{johansson-icml16}, and TARNet~\cite{shalit-icml17}.
The baseline feature selection methods use local structure learning~\cite{yu-arxiv19} and causal structure learning~\cite{heinze-stat18,henckel-arxiv19} to select an adjustment set.
The first baseline, \emph{PC Simple (Parents)}, uses a local structure learning method on the treatment to find parents.
PC Simple has been used for selecting variables for classification~\cite{yu-arxiv19} but not for HTE estimation.
We employ PC Simple to learn the local structure on the treatment variable and select the treatment parents as an adjustment set.
Since our setting only considers one treatment variable, the parents of treatment are a valid adjustment set~\cite{williamson-resp14}.
Two other baselines we consider are based on causal structure learning~\cite{heinze-stat18,henckel-arxiv19}, which we refer to as Full Structure Learning (FSL){}.
We use GFCI~\cite{ogarrio-pgm16} to learn a causal structure and apply two adjustment criteria: optimal adjustment set (O-set) and removing post-treatment variables only (Valid).
Then a base learner uses the adjustment set for estimation (e.g., TLearner).
For our feature selection methods, we consider our framework HTE-FS{} and each component, HTE-Fit{} and Structure-Fit{}, separately.
For HTE-Fit{}, we use both variants: \globalmethodname{}-F{} and \globalmethodname{}-B{}.
We use four HTE fit metrics: $\tau$-risk~\cite{schuler-model18}, Nearest Neighbor Precision in Estimating Heterogeneous Effects (NN-PEHE)~\cite{schwab-perfect18}, plug-in $\tau$~\cite{saito-cv19}, and Counterfactual Cross-validation (CFCV)~\cite{saito-cv19}.
We report results using \globalmethodname{}-F{} and $\tau$-risk since this combination performed the best in our experiments.
For Structure-Fit{} we used PC-Simple~\cite{buhlmann-bio10}, GetPC~\cite{pena2005scalable}, and HITON-PC~\cite{aliferis-jmlr10}, CMB~\cite{gao-neurips15}, and PCD-by-PCD~\cite{yin-causation08}.
For pairwise edge orientation, we use Conditional Distribution Similarity Statistic (CDS)~\cite{fonollosa-book19}, Information Geometric Causal Inference (IGCI)~\cite{daniusis-uai12}, and Regression Error based Causal Inference (RECI)~\cite{blobaum-aistats18}.
In order to find an adjustment set, we only remove mediators and descendants, and allow data-driven HTE estimators to decide which variables to use in estimation.
We report results using PC Simple with RECI (PC, RECI) and CMB since these two combinations represent different variants of Structure-Fit{} and perform the best in our experiments.
For HTE-FS{}, we report results using \globalmethodname{}-F{} with $\tau$-risk and Structure-Fit{} with CMB, since this combination performed the best in experiments.
We compare against using Oracle (Known Structure), which uses the true optimal adjustment set.
\subsection{Datasets}
We use three types of datasets.
The first dataset is a set of synthetic datasets, in which random SCMs are generated, and data is generated following each SCM.
The second dataset is based on the Infant Health and Development Program (IHDP) dataset~\cite{hill-jcgs11}.
The third type are real-world datasets, League of Legends and Cannabis.
\subsubsection{Synthetic dataset}
We generate SCMs by following a procedure similar to one for evaluating causal structure learning algorithms~\cite{heinze-stat18}.
SCMs and their corresponding datasets differ in the following characteristics: the number of variables \( d \); the probability of edge creation \( p_e \); the variance in the noise \( \sigma \); the strength of the noise term in non-source nodes \( \rho \).
We add parameters to control confounding \( \gamma \), mediation \( m \), and heterogeneity in treatment effects \( p_h, m_p \).
We initialize a set of parameters as follows:
\begin{enumerate}
\item Number of variables: \( d \in [10, 20, 30] \)
\item Edge probability parameter: \( p_e \in [0.1, 0.3, 0.5] \)
\item Noise variance: \( \sigma \in [0.2, 0.4, 0.6] \)
\item Magnitude in noise: \( \rho \in [0.1, 0.5, 0.9] \)
\item Confounder: \( \gamma \in [\text{True, False}] \)
\item Mediator chain length: \( m \in [0, 1, 2] \)
\item Number of HTE inducing parents: \( p_h \in [0, 1, 2] \)
\item HTE from mediating parents: \( m_p \in [\text{True, False}] \)
\end{enumerate}
For each parameter value, we randomly sample from all other parameter values \( 500 \) times to generate SCMs.
For each of the $11,500$ SCMs, we create one dataset of size \( 10,000 \).
Root node variables are generated from a normal distribution, and non-root nodes are generated linearly from parents.
The detailed dataset generation description is available in the Appendix in Section~\ref{sec:data_gen_desc}, and the link for the code is available in Section~\ref{sec:code}.
\subsubsection{Semi-synthetic dataset: IHDP}
The covariates in the Infant Health and Development Program (IHDP) dataset come from a randomized experiment studying the effects of specialist home visits on future cognitive test scores~\cite{hill-jcgs11}.
Outcomes can be simulated using the covariates.
We simulate 1,000 instances of simulated outcomes and average all results\footnote{Outcomes are generated using the JustCause package:\\https://github.com/inovex/justcause}.
\subsubsection{Real-world datasets}
\kdd{
We use two real-world datasets: \emph{League of Legends} and \emph{Cannabis}.
League of Legends (LoL) is a popular multiplayer online battle arena (MOBA) game.
Riot Games, the developer of LoL, regularly creates software patches to the game that update various parts of the game, affecting balance and how players decide to play the game.
We use a dataset consisting of player, match, and patch information over time~\cite{he-fdg21}.
The treatment variable of interest is a software patch, and the outcome of interest is the number of \emph{kills} of a team in a single match, which is a post-match statistic.
We average results over patches in the dataset.
}
\kdd{
\emph{Cannabis} is a dataset consisting of tweets about the e-cigarette Juul and cannabis-related topics~\cite{adhikari-icwsm21}.
English tweets related to the e-cigarette Juul were collected from 2016 to 2018.
For the users who mention Juul, cannabis-related tweets are collected from 2014 to 2018.
In this dataset, tweets' stances towards Juul and cannabis are crowdsourced, and a model is trained for stance detection.
In this work, we are interested in how users' stance on Juul affects their stance on cannabis.
To do this, we find users who have Juul-related tweets occurring before a cannabis-related tweet and detect their stance on Juul and cannabis using the classifier from~\cite{adhikari-icwsm21}, where features are the word embeddings for tweets using BERT~\cite{devlin-nacl19}.
The treatment variable of interest is a user's stance on Juul, in favor (1) or against (0).
The outcome variable is their stance on cannabis, in favor (1) or against (0).
Since this dataset contains only one group, we bootstrap the estimators to compute the standard deviation.
}
\input{figures/learner_compare.tex}
\input{figures/winner_table.tex}
\subsection{Evaluation}
We use the mean squared error (MSE) for evaluating HTE estimations since the true causal effect is known with synthetic and semi-synthetic data: $\text{MSE} = \frac{1}{N} \sum_{i=1}^N (\tau(\mathbf{X}) - \hat{\tau}(\mathbf{X}_i))^2$.
For each dataset, we split the data into 80/20 training and test sets.
Since different SCMs will result in different MSEs, we first rank all estimation methods based on their MSE for a particular SCM.\@
Then we average the rank of each method across SCMs and report on the average rank.
To study their performance under specific settings, we also report on average ranks for subsets of parameter settings.
For real-world datasets, we report results using $\tau$-risk~\cite{nie-rlearner17,schuler-model18}, which we defined in eq~\eqref{eq:tau_risk}.
Similar to MSE, we report rankings based on $\tau$-risk.
\kdd{
In addition, for datasets in which we can decide which variables \emph{post-treatment} variables, we compute the percentage of included by each feature selection method, which we denote as \emph{inclusion error (IE)}.
Given the \emph{post-treatment} set, $\mathcal{F}$, and the predicted causal feature set $\mathbf{X}^{(\ell)}$, we compute the inclusion error as:
\begin{equation}
\text{IE} = \frac{| \mathbf{X}^{(\ell)} \cap \mathcal{F} |}{|\mathcal{F}|}.
\end{equation}
Some variables occur after treatment for real-world datasets, such as other post-match statistics in LoL and tweets related to cannabis after the initial cannabis-related tweet in the Cannabis dataset.
}
\subsection{Evaluation on synthetic data}
\kdd{
We first present the results on synthetic datasets.
We first show how our feature selection method improves multiple base learners.
Then we show how feature selection improves base learners across multiple structural configurations.
}
\subsubsection{Feature selection helps all base learners}
Figure~\ref{fig:learner_compare} shows the MSE comparisons between several HTE estimators and feature selection methods on SCMs which are generated using: 30 variables, confounding, mediator chain length of 2, 2 HTE inducing parents, and HTE induced by mediating parents (i.e., \( (d,\gamma,m,p_h,m_p) = (30, True, 2, 2, True) \)).
We average MSE over all other parameters (i.e., \( p_e, \sigma, \rho \)).
\kdd{
We show the best performing Structure-Fit{}, \globalmethodname{}-F{}, HTE-FS{}.
This figure shows that adding feature selection improves MSE for all HTE estimators.
In addition, feature selection reduces the error bars for all base learners.
In general, HTE-FS{} (Tau risk, CMB) performs the best, compared to using \globalmethodname{}-F{} or Structure-Fit{} alone.
Structure-Fit{} with CMB performs better than Structure-Fit{} with PC Simple and RECI, although the latter still reduces error overall compared to using the base learner on all features.
}
\subsubsection{Feature selection improves base learners for different types of structures}
\kdd{
Table~\ref{tab:mse_rankings} ranks the feature selection methods using TLearner as the base learner vs.\ several other HTE estimators (without using feature selection) under a range of structural scenarios.
}
\kdd{
We choose TLearner as the base learner since it is a relatively simple model and performs consistently in all settings.
}
One experimental parameter that affects the causal structure is fixed in each column while averaging over the other parameters.
The top section shows the average rank for base learners without feature selection, the second section shows baseline feature selection methods, the third section shows our feature selection methods, and the bottom section shows the Oracle method.
Bolded results indicate the best rank in the column (ignoring the Oracle estimator).
The large variety of underlying structures explains the high variance in rank.
\kdd{
The table shows that our feature selection methods improve base learners for different structures, perform better than baseline feature selection methods, and perform better than other base learners without feature selection.
HTE-FS{} performs the best overall.
In two settings, with no confounding and a low number of variables, Structure-Fit{} with CMB performs the best.
While our methods perform better than baseline methods, baseline methods still improve base learners, compared to no feature selection.
}
\input{figures/wrong_variables.tex}
\subsubsection{Feature selection reduces inclusion of wrong variables}
\kdd{
In addition to performance in terms of MSE, we investigate the \emph{inclusion error (IE)} of feature selection methods.
Figure~\ref{fig:wrong_variables} shows the inclusion error for each feature selection method, averaged over all experiments, where lower is better.
The grey line represents the case of randomly selected features.
The first set of bars shows the average percentage of overall synthetic datasets.
All feature selection methods do better than random feature selection.
HTE-FS{} has the lowest chance to include wrong variables, at about 8\% inclusion rate.
This result supports the results shown in Table~\ref{tab:mse_rankings}, that HTE-FS{} has the lowest error, followed by Structure-Fit{} with CMB and \globalmethodname{}-F{}.
}
\subsection{Evaluation on IHDP dataset}
\input{figures/ihdp_table.tex}
\kdd{
Table~\ref{tab:ihdp_rankings} shows the average rankings on the IHDP dataset using MSE and $\tau$-risk.
}
We used TLearner as the base learner and showed the same variations in feature selection as in Table~\ref{tab:mse_rankings}.
We see that all feature selection methods perform better on average than the base estimators, with HTE-FS{} performing the best.
Since TLearner is a linear model close to the data generation process for this dataset, it already performs very well on it.
Since this dataset contains pre-treatment features only, we do not evaluate inclusion error.
\subsection{Evaluation on real-world datasets}
\kdd{
Finally, we show results on real-world datasets.
Since these datasets do not contain ground truth effects, we evaluate methods using $\tau$-risk rankings.
}
\subsubsection{LoL dataset}
\kdd{
The average rank for each method in terms of $\tau$-risk for the LoL dataset is shown in Table~\ref{tab:ihdp_rankings}.
Overall, HTE-FS{} achieves the lowest $\tau$-risk rank among all estimators.
We see that TLearner performs the best among base learners and has a lower rank than some feature selection methods.
One reason could be that some methods are not suitable for high dimensional datasets, such as Structure-Fit{} with PC Simple.
However, Structure-Fit{} with CMB performs much better than PC Simple, \globalmethodname{}-F{}, and FSL{}.
}
\kdd{
In the LoL dataset, there are pre-match and post-match features.
Since our outcome \emph{kills} is a post-match feature, other post-match features should not be included in the estimation, which is the \emph{post-treatment} set in this setting.
The second set of bars in Figure~\ref{fig:wrong_variables} show the inclusion error for the LoL datasets.
Overall, each feature selection method has a higher IE than in the synthetic dataset.
HTE-FS{}, Structure-Fit{} with CMB, and \globalmethodname{}-F{} reduce the inclusion error rate on average, while Structure-Fit{} with PC-Simple and FSL{} do not reduce the inclusion error more than random selection.
}
\subsubsection{Cannabis}
\kdd{
The rank in terms of $\tau$-risk for the Cannabis dataset is shown in Table~\ref{tab:ihdp_rankings}.
Since this dataset contains only one set of data, we bootstrap each estimator.
In this dataset, HTE-FS{} achieves the lowest $\tau$-risk, followed by FSL{} and \globalmethodname{}-F{}.
Here, XLearner achieves the lowest $\tau$-risk among base learners and performs close to our proposed methods.
}
\kdd{
In the Cannabis dataset, the post-treatment features are tweets after the first cannabis-related tweet.
The features that should be included are only tweets related to the stance towards Juul and the tweet related to the first cannabis tweet.
Figure~\ref{fig:wrong_variables} shows the inclusion error for the Cannabis dataset in the rightmost set of bars.
In most cases, our feature selection methods reduce the number of wrong variables included, except Structure-Fit{} with PC-Simple.
}
\section{Conclusion}
In this work, we addressed the problem of improving data-driven heterogeneous treatment effect (HTE) estimation when the underlying structural causal model (SCM) is unknown.
We developed HTE-FS{} for causal feature selection for HTE estimation which consists of two components: HTE-Fit{} and Structure-Fit{}.
We show that HTE-FS{} reduces HTE estimation error of existing data-driven methods under a variety of causal model assumptions when compared to using their vanilla versions with all available features.
Our research suggests several interesting avenues of future research, including %
the development of causal feature selection algorithms in the presence of latent variables and the consideration of the full do-calculus, not just backdoor adjustment.
\section{Appendix}
Here we provide additional information for Heterogeneous Treatment Effect Feature Selection (HTE-FS){} and the synthetic data generation.
\subsection{HTE-Fit{}}\label{sec:hte_fit_desc}
Here we present pseudocode for the two HTE-Fit{} algorithms.
The \globalmethodname{} --- Forward{} algorithm is shown in Algorithm~\ref{alg:forward_selection}. Lines 1-8 first find the one feature that returns the largest score (computed by $R$) and adds it to \( \mathbf{X}^{(g)} \).
Then, for each remaining feature \( X^{(j)} \), we compute the score when combining \( X^{(j)} \) to \( \mathbf{X}^{(g)} \), and concatenate the feature that returns the best score.
This repeats until the score does not change, or there are no more features to add.
\globalmethodname{} --- Backward{} works similarly to \globalmethodname{}-F{} and the pseudocode is shown in Algorithm~\ref{alg:backward_selection}.
We start by computing the score using all features in Line 1.
In Lines 9-15, we iteratively add features to \( \mathbf{K} \), which are features that when removed improve the score.
The final set returned is the whole feature set without \( \mathbf{K} \).
\input{99_algorithms.tex}
\subsection{Structure-Fit{}}\label{sec:structure_fit_desc}
Algorithm~\ref{alg:lcfs} details Structure-Fit{}.
In lines 1-2, we initialize two variables: $dn$, which are discovered nodes, $\mathbf{Z}$, a queue of nodes to traverse.
The discovered nodes are initialized to $ch(Y)$, so we do not traverse past $Y$.
In lines 3-5, we initialize a graph $\hat{G}$ which will be the local structure from $T$ to $Y$, and $ \mathbf{Z} $ is initialized to be the children of $ T $.
Lines 6-10 show the iterative process in which we consider the nodes in \( \mathbf{Z} \).
For every node \( Z \in \mathbf{Z} \), find \( pa(Z) \) and \( ch(Z) \).
$Z$ is added to the discovered nodes, is removed from $\mathbf{Z}$, and the children of $ Z $ are added to $ \mathbf{Z} $.
Then we connect $Z$ to its children.
In the end, a graph, $\hat{G}$ with causal paths from $T$ to $Y$ is constructed, and we can construct the valid set.
\input{04_02_algorithm}
\subsection{Complexity of HTE-FS{}}\label{sec:complexity}
The complexity of HTE-FS{} depends on the complexity of training and prediction of the HTE estimator and the local structure learning algorithm.
Given an HTE estimator with complexity $ O(H) $ training and $ O(P) $ prediction, HTE-Fit{} takes $ O(d^2 \cdot (H + P) ) $.
If HTE-Fit{} outputs $ s $ variables and we use CMB for Structure-Fit{}, then the worse case scenario happens when we find local structures for all $s$ variables: $ O(s \cdot 2^{|PC|}|U|^2|PC|) $~\cite{gao-neurips15}.
\subsection{Synthetic data generation}\label{sec:data_gen_desc}
Here, we describe how we generate synthetic datasets used in the experiments.
We first generate an adjacency matrix, \( \mathbf{A} \) over the \( d \) variables.
Assuming that the variables \( \{1, \dots, d \} \) are causally ordered.
For each pair of nodes \( i, j \), where \( i < j \) in the causal order, we sample from Bernoulli(\( p_e \)) to determine if there is an edge from \( i \) to \( j \).
With the adjacency matrix, we sample the corresponding coefficients, \( \mathbf{C} \), from Unif(\( -1, 1 \)).
Noise terms \( \epsilon_i \) are generated by sampling from a \( d \)-dimensional zero-mean Gaussian distribution with covariance matrix \( \Sigma \), where \( \Sigma_{i,i} = 1 \) and \( \Sigma_{i,j} \) = \( \sigma \).
We use \( \rho \) to tune the magnitude of the noise.
Next, we choose the treatment and outcome nodes based on two parameters, \( \gamma \) and \( m \). \( \gamma \) controls whether there is confounding in the generated dataset and \( m \) controls the length of the mediating chain (\( 0 \) means no mediator).
First, we select all pairs of nodes which have a \emph{directed} \( m \) hop path.
Then, if \( \gamma \) is True, we filter all pairs that do not contain a \emph{backdoor path}.
Otherwise we filter out pairs that contain a backdoor path.
Finally, we sample a pair of nodes from the filtered set to obtain the treatment and outcome nodes.
If no pairs of nodes satisfy the criteria for the given parameter setting, we resample the graph up to 100 times until a graph is found, or we discard the parameter setting.
A source node \( i \) (node with no parents) takes the value of the sampled noise terms, \( \epsilon_i \).
For a non-source node, \( i \), we consider several additive factors.
First is the base term from the coefficient matrix: \( C_{pa(i), i} X_{pa(i)} \), where \( C_{pa(i), i} \) are the coefficients from \( pa(i) \) to \( i \) and \( X_{pa(i)} \) are the values generated for \( pa(i) \).
Next, are the interaction terms, which are determined by the mediating chain length \(m\), number of HTE inducing parents \( p_h \), and whether there is HTE from parents of mediators \( m_p \).
For mediators, we have: \( \mathbbm{1}_{m}(i) \mathbbm{1}_{m_p} X_{pa_m(i)} X_{pa_n(i)} \), where \( \mathbbm{1}_{m} \) is an indicator for if \( i \) is a mediator, \( \mathbbm{1}_{m_p} \) is an indicator for if there is HTE induced from mediator parents, and \( pa_m \) and \( pa_n \) represent mediating and non-mediating parents.
For outcomes, the interaction is defined similarly: \( \mathbbm{1}_{y}(i) X_{pa_m(i)} X_{pa_n(i)} \), where \( \mathbbm{1}_{y}(i) \) indicates whether the node is the outcome node.
The number of HTE parents determines how many non-mediating parents (\( pa_n \)) the interaction occurs with.
A value of zero means there is no heterogeneity.
So the final data generating equation is:
\begin{align}
X_i & = C_{pa(i), i} X_{pa(i)}
\nonumber \\
& + \mathbbm{1}_{m}(i) \mathbbm{1}_{m_p} X_{pa_m(i)} X_{pa_n(i)}
\nonumber \\
& + \mathbbm{1}_{y}(i) X_{pa_m(i)} X_{pa_n(i)} + \epsilon_i
\end{align}
\subsection{Code}\label{sec:code}
Code for HTE-FS{} and the synthetic data generation is available at \href{https://github.com/edgeslab/causal_feature_selection}{https://github.com/edgeslab/causal\_feature\_selection}.
\section{Conclusion}
In this work, we addressed the problem of improving data-driven heterogeneous treatment effect (HTE) estimation when the underlying structural causal model (SCM) is unknown.
We developed HTE-FS{} for causal feature selection for HTE estimation which consists of two components: HTE-Fit{} and Structure-Fit{}.
We show that HTE-FS{} reduces HTE estimation error of existing data-driven methods under a variety of causal model assumptions when compared to using their vanilla versions with all available features.
Our research suggests several interesting avenues of future research, including %
the development of causal feature selection algorithms in the presence of latent variables and the consideration of the full do-calculus, not just backdoor adjustment.
\section{Background}\label{sec:background}
We provide an introduction to heterogeneous treatment effects under the potential outcomes framework~\cite{rubin-74} and structural causal models~\cite{pearl-book09}.
\subsection{Heterogeneous treatment effects}
Let $(\mathbf{X}_i, Y_i(0), Y_i(1), T_i) \sim \mathcal{P}$ be a distribution with $N$ units which are independently and identically distributed (i.i.d.).
For any unit $i$, define $\mathbf{X}_i \in \mathbb{R}^d$ be a $d$-dimensional feature vector,
$T_i \in \{0, 1\}$ to be the treatment assignment indicator, $Y_i(0)$ to be the potential outcome when $i$ is not treated, and $Y_i(1)$ to be the potential outcome when $i$ is treated.
The individual treatment effect (ITE) is the difference in potential outcomes: \( \tau_i = Y_i(1) - Y_i(0) \).
However, both potential outcomes $Y_i(1)$ and $Y_i(0)$ cannot be observed at the same time and observed data $\mathcal{D} = (\mathbf{X}_i, Y_i, T_i)$ contains only $Y_i = Y_i(T_i)$. This is known as the fundamental problem of causal inference.
Since $\tau_i$ can never be measured directly, its estimate is expressed through the conditional average treatment effect (CATE):
\begin{equation}
\label{eq:cate}
\hat{\tau}(\mathbf{x}) = E[Y(1) - Y(0) \mid \mathbf{X}_i].%
\end{equation}
The average treatment effect (ATE) of the population is: $\text{ATE} = E[Y(1) - Y(0)]$. If CATE and ATE are not the same, then $\mathbf{X}_i$ induce heterogeneous treatment effects (HTE).
Given an observed dataset $\mathcal{D}$, the goal of an \emph{HTE estimator} is to estimate CATE, $\hat{\tau}$, using the available features $\mathbf{X}$ in the data.
HTE estimators estimate CATE in a variety of ways.
Some HTE estimators rely on recursive partitioning using tree-based methods~\cite{tran-aaai19,athey-pnas16,athey-annals19}.
Other HTE estimators use predictions from machine learning methods to train a causal effect estimator~\cite{kunzel-pnas19,kennedy-arxiv20} or to predict counterfactual outcomes for estimation of effect~\cite{shalit-icml17,johansson-icml16}.
A common assumption used by data-driven HTE estimators is the assumption of \emph{strong ignorability} which states that the treatment assignment is independent of the potential outcomes given a set of variables $\mathbf{X}$ ~\cite{rosenbaum-bio83,shalit-icml17}: $Y_1, Y_0 \perp T \mid \mathbf{X}$, and $0 < p(T=1 \mid \mathbf{X}) < 1$ for all $\mathbf{X} = \mathbf{x}$.
One of the implications of this assumption is that there is no \emph{hidden confounding}.
However, hidden confounding is not the only problem that can lead to biased HTE estimation. In practice, it is often assumed that all observed variables in the data are the set of variables $\mathbf{X}$ that meets strong ignorability.
This is problematic because some of the variables may make $T$ and $Y$ dependent or introduce new spurious dependencies~\cite{pearl-book09}.
Some variables, such as mediators, should not be included in the estimation. Moreover, when the temporal order of variables is unknown, practitioners may consider all features in the data as pre-treatment variables~\cite{montgomery-ajps18}.
Selecting the right variables is the goal of our work.
\subsection{Structural Causal Models}
A structural causal model (SCM), $M$, consists of two sets of variables, $\mathbf{U}$ and $\mathbf{V}$, and a set of structural equations, $\mathbf{F}$, describing how values are assigned to each \emph{endogenous} variable $V_i \in \mathbf{V}$ based on the values of $\mathbf{v}$ and $\mathbf{u}$~\cite{pearl-book09}: \( v_i = f_i(\mathbf{v}, \mathbf{u}) \).
A causal graph \( G \) is a directed acyclic graph that captures the causal relationships among the variables.
The variables in $\mathbf{U}$ are considered \emph{exogenous}.
It is assumed that every endogenous variable has a parent in $\mathbf{U}$.
The causal graph may omit representing $\mathbf{U}$ explicitly.
We denote the parents, children, ancestors, and descendants of $X$ in the graph as: \(pa(X), ch(X), an(X), de(X)\).
Every node is a descendant and ancestor of itself.
When the SCM underlying the dataset is known, its graph G supports the use of graphical criteria for variable selection for unbiased causal effect estimation, known as adjustment set selection~\cite{shpitser-uai06,shpitser-uai10,van-uai14,perkovic-uai15,henckel-arxiv19}.
One such criterion is the backdoor criterion~\cite{pearl-book09}, which aims to find a set that blocks the paths from treatment to the outcome, starting with an arrow into treatment.
For example, in Figure~\ref{fig:collider_scm}, there is one backdoor path, namely, $ T \leftarrow X \rightarrow Y $, and so the set $ \{ X \} $ sufficiently satisfies the backdoor criterion.
Several adjustment criteria have been developed, such as the adjustment criterion~\cite{shpitser-uai06,van-uai14}, generalized back-door criterion~\cite{maathuis-annals15}, generalized adjustment criterion~\cite{perkovic-uai15}, and optimal adjustment set (O-set)~\cite{henckel-arxiv19}.
These adjustment sets defined by the various adjustment criteria can be found through the ID algorithm~\cite{shpitser-jmlr08}, the framework by Van der Zander et al.~\cite{van-uai14} or the pruning procedure by Henckel et al.~\cite{henckel-arxiv19}.
Once an adjustment set is found, they can be used with an HTE estimator to estimate heterogeneous effects.
When the SCM for a dataset is unknown, and therefore there is \emph{structure uncertainty} of the causal mechanism, the adjustment criteria cannot be applied.
An alternative approach is to use causal structure learning~\cite{heinze-stat18} to learn a causal graph from the data and then apply an adjustment criterion for feature selection.
However, learning the full structure is computationally expensive for high-dimensional datasets~\cite{chickering-jmlr04,heinze-stat18,vowels-arxiv21}.
Instead, we propose reducing the feature space by assessing the fit of each variable for HTE estimation and learning the local structure, which we describe in Section~\ref{sec:cfs}.
\subsubsection{HTEs in SCMs}
In SCMs, heterogeneous effects are represented by an \emph{interaction} or \emph{effect modification} through ancestors of $Y$~\cite{pearl-book09}.
For example, suppose $Y$ is generated by the following structural equation corresponding to the SCM in Figure~\ref{fig:marloes_scm}:
\begin{equation}
Y = \alpha \cdot T + \beta \cdot X + \gamma \cdot T \cdot f(X) + U_Y.
\end{equation}
Here, there is a fixed effect \( \alpha \) of the treatment \( T \) and a heterogeneous effect \( \gamma \) due to the interaction between \( T \) and \( X \).
Interaction can occur with any parent of \( Y \) (e.g., \( X, F \) in Figure~\ref{fig:marloes_scm}) or any parent of a mediator (e.g., \( C \) in Figure~\ref{fig:marloes_scm})~\cite{vanderweele-epid07}.
\subsubsection{C-specific effects}
\new{
Shpitser and Pearl provide a nonparametric estimand for HTEs with respect to some feature assignment $C=c$, called $c$-specific effects~\cite{shpitser-uai06,pearl-smr17}.
}
One issue with estimating $c$-specific effects is that it requires knowledge of the SCM and which variables, C, exhibit heterogeneous effects.
When the SCM is known, the O-set~\cite{henckel-arxiv19} can be used on the learned structure since it implicitly includes all potential heterogeneity-inducing variables, while some identification algorithms may miss them. When the SCM is unknown, the identifiability of \( c \)-specific effects cannot be established.
Moreover, when the set of heterogeneity-inducing variables $\mathbf{C}$ is unknown, then all combinations of features and feature values would need to be checked for heterogeneity in effect, which is exponential in the number of valid features.
Circumventing this exponential search is one of the main reasons why data-driven methods are popular in practice.
\section{Problem Setup}
Our goal in this work is to augment data-driven methods to accurately predict heterogeneous effects when both the SCM and the variables that explain heterogeneity are unknown.
More specifically, the goal is to find a subset $\mathbf{X^{(\ell)}}$, to which we refer as the \emph{causal feature set{}}, that improves data-driven estimation.
In general, the goal in HTE estimation is to obtain accurate estimation of CATE:
\begin{problem}\label{prob:hte_problem}
(Heterogeneous treatment effect (HTE) estimation)
Given a dataset $\mathcal{D}=(\mathbf{X}_i, Y_i, T_i)$ of N instances with some true ITE $\mathbf{\tau}$, estimate CATE $\hat{\tau}(\mathbf{X})$ such that the mean squared error is minimized:
\begin{equation}\label{eq:cate_objective}
\frac{1}{N} \sum_{i=1}^N (\tau_i - \hat{\tau}(\mathbf{X}_i)) ^ 2.
\end{equation}
\end{problem}
Since the true ITE is unknown in real-world scenarios, several heuristic HTE evaluation metrics have been developed for evaluating the performance of HTE estimators~\cite{schuler-model18,schwab-perfect18,saito-cv19}.
Instead of minimizing the true estimator error, we can instead minimize one of these metrics, such as the \( \tau \)-risk~\cite{schuler-model18,nie-rlearner17}.
A \emph{causal feature set{}} is the feature set that minimizes the error in HTE estimation.
\begin{problem}\label{prob:causal_feature_selection}
(Causal feature selection (CFS) for HTE estimation) Given a dataset $\mathcal{D}=(\mathbf{X}_i, Y_i, T_i)$ and a data-driven estimator $\hat{\tau}(\mathbf{X})$, find a \emph{causal feature set{}} $\mathbf{X}_i^{(\ell)} \subseteq \mathbf{X}_i$ such that $\hat{\tau}(\mathbf{X}^{(\ell)})$ results in the lowest MSE:
\begin{gather}
\argmin_{\mathbf{X^{(\ell)}}} \frac{1}{N} \sum_{i=1}^{N} {\Big( \tau_i - \hat{\tau}\big( \mathbf{X}^{(\ell)}_i \big) \Big)}^2.
\end{gather}
\end{problem}
\noindent A data-driven estimator typically estimates $\hat{\tau}(\mathbf{X})$ using all available data.
Our goal in this work is to select a causal feature set{} $\mathbf{X}^{(\ell)}$ such that the error of the estimator is minimized.
In this work, we assume that it is unknown a priori whether all observed features in the data are pre-treatment variables and whether all of them meet the strong ignorability assumption.
Instead, we use the following more general assumptions:
\begin{enumerate}
\item Causal sufficiency: there are no latent variables.
\item Causal faithfulness: if $X_A, X_B$ are conditionally independent given $\mathbf{X_Z}$, then $A,B$ are d-separated in the causal graph by $\mathbf{Z}$.
\item Acyclicity: the underlying causal graphs are directed acyclic graphs (DAGs).
\item There is no selection bias.
\end{enumerate}
Understanding how our methods need to be adjusted when these assumptions do not hold is left for future work.
\section{Introduction} \label{sec:intro}
Estimating heterogeneous treatment effects (HTEs) is essential in many fields.
HTE estimation aims to find subpopulations whose causal effects differ from the effects of the population as a whole.
For example, if the treatment is a drug, some individuals may have adverse reactions, and some individuals may benefit from treatment~\cite{shalit-icml17}.
Similarly, public policy may affect different sociodemographic groups differently~\cite{grimmer-pa17}.
HTE analysis allows the discovery of these different subgroups of the population.
\begin{figure}
\centering
\newcommand{0.275}{0.3}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\node(c) at (0.0, -0.75) [label=right:$L$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (y);
\path (t) edge (c);
\path (y) edge (c);
\end{tikzpicture}
\caption{Collider}\label{fig:collider_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (m) at (0.0, 0) [label=below:$M$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (m);
\path (m) edge (y);
\end{tikzpicture}
\caption{Mediator}\label{fig:mediator_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (a) at (-1.5, 0) [label=left:$A$,point];
\node (t) at (-0.75, 0) [label={below left}:$T$,point];
\node (b) at (-0.75, -0.75) [label={below}:$B$,point];
\node (x) at (-0.75, 0.75) [label={above}:$X$,point];
\node (c) at (0.0, 0.75) [label={above}:$C$,point];
\node (d) at (0.0, 0) [label={below left}:$D$,point];
\node (e) at (0.0, -0.75) [label={below}:$E$,point];
\node (f) at (0.75, 0.75) [label={above}:$F$,point];
\node (y) at (0.75, 0.0) [label={right}:$Y$,point];
\node (g) at (0.75, -0.75) [label={below}:$G$,point];
\path (a) edge (t);
\path (x) edge (t);
\path (t) edge (b);
\path (x) edge (y);
\path(t)edge(d);
\path(c)edge(d);
\path(d)edge(e);
\path (d) edge (y);
\path (f) edge (y);
\path (y) edge (g);
\end{tikzpicture}
\caption{Multivariable}\label{fig:marloes_scm}
\end{subfigure}
}
\caption{Causal models where data-driven HTE estimation methods may perform poorly.
Some variables (e.g., $L$, $M$, and $E$) are not valid adjustment variables for estimating the effect of $T$ on $Y$.
}\label{fig:fail_examples}
\end{figure}
The accumulation of large amounts of data in many domains has stimulated the increased development of algorithms tailored for automated \emph{data-driven} HTE estimation, relying on supervised machine learning~\cite{athey-pnas16,johansson-icml16,grimmer-pa17,shalit-icml17,kennedy-arxiv20,kunzel-pnas19,tran-aaai19,athey-annals19}.
The goal of data-driven HTE estimators is to estimate the conditional average treatment effect using the features available in the data.
While it is appropriate to use all possible features for general supervised learning problems, unbiased estimation of \emph{causal effects} requires knowledge of the causal mechanisms underlying the data and selecting an appropriate subset of variables from the data~\cite{pearl-book09}.
One way to select this subset is by modeling the underlying causal mechanisms through structural causal models (SCMs)~\cite{pearl-book09} and selecting an \emph{adjustment set} --- a set of variables that are necessary for unbiased estimation of a causal effect of interest~\cite{pearl-book09,perkovic-uai15,henckel-arxiv19}.
However, the causal model structure is rarely known for real-world data, and data-driven methods do not take it into consideration.
Our paper addresses this problem and improves data-driven HTE estimation methods through causal feature selection under such structure uncertainty.
A common assumption used by data-driven HTE estimators is the assumption of \emph{strong ignorability} which states that the treatment assignment is independent of the potential outcomes given the variables in the data~\cite{rosenbaum-bio83,shalit-icml17}.
This untestable assumption implies that all potential confounders, variables that can influence both the treatment and the outcome, are observed in the data.
In turn, accounting for them would remove spurious associations and allow for causal effect identification.
However, hidden confounding is not the only problem leading to biased HTE estimation.
Even if all potential confounders are observed in the data, not all variables should be used for adjustment.
Another important and sometimes implicit assumption is that all variables in the data are \emph{pre-treatment} variables, i.e., the variables that were measured before and thus were unaffected by the treatment.
When the temporal order of observations is unknown, many practitioners would simply use all variables in the data~\cite{montgomery-ajps18}.
To illustrate the problem of using all variables in data-driven methods, consider the SCMs in Figure~\ref{fig:fail_examples} where $X$ is a confounding variable.
According to SCM theory, we need to adjust for \( X \) when estimating causal effects, but should not include variables that will bias effect estimation, such as descendants of the treatment or instrumental variables (e.g., \( L, M, A, B, D, E, G \)).
To address these deficiencies of data-driven methods, we define the problem of causal feature selection for accurate data-driven HTE estimation under structure uncertainty.
One vanilla solution to this problem is to first learn the causal structure from the data~\cite{henckel-arxiv19,spirtes-book00,ogarrio-pgm16} and then use an existing identification algorithm to find the adjustment set~\cite{shpitser-uai06,van-uai14,perkovic-uai15,henckel-arxiv19}.
However, structure learning is computationally expensive when the space of feature is large, and it may not work well when there is uncertainty in the learned structure~\cite{heinze-stat18,chickering-jmlr04,vowels-arxiv21}.
Instead, we develop a feature selection method that reduces this complexity by considering the value of each feature for improving HTE estimation and learning only the relevant parts of the causal structure from data.
We provide strong empirical evidence that our method improves existing data-driven HTE estimation methods under arbitrary underlying causal structures.
We evaluate the benefits of our feature selection algorithm in reducing HTE estimation error through synthetic, semi-synthetic, and real-world dataset experiments.
\section{Introduction}
Estimating heterogeneous treatment effects (HTEs) is an important problem in many fields.
The goal of HTE estimation is to find subpopulations for which causal effects differ from the population as a whole.
For example, if the treatment is a drug, some individuals may have adverse reactions, and some individuals may benefit from treatment~\cite{shalit-icml17}. Similarly, public policy may affect different sociodemographic groups differently~\cite{grimmer-pa17}.
HTE analysis allows the discovery of these different subgroups of the population.
Several supervised machine learning techniques have been developed for the problem of HTE estimation in recent years~\cite{athey-pnas16,johansson-icml16,grimmer-pa17,shalit-icml17,kennedy-arxiv20,kunzel-pnas19,tran-aaai19,athey-annals19}. We refer to these methods as \emph{data-driven} HTE estimators. Given a dataset, the goal of HTE estimators is to estimate the conditional average treatment effect using the features available in the data.
In supervised learning, algorithms use all possible features for accurate predictions.
In contrast, estimating causal effects requires knowledge of the causal mechanisms underlying the data and accounting for a correct subset of variables for accurate and unbiased causal effect estimation~\cite{pearl-book09}.
\begin{figure}
\centering
\newcommand{0.275}{0.275}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\node(c) at (0.0, -0.75) [label=right:$L$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (y);
\path (t) edge (c);
\path (y) edge (c);
\end{tikzpicture}
\caption{Collider}\label{fig:collider_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (t) at (-0.75, 0) [label=left:$T$,point];
\node (m) at (0.0, 0) [label=below:$M$,point];
\node (x) at (0.0, 0.75) [label=above:\(X\),point];
\node(y) at (0.75, 0.0) [label=right:$Y$,point];
\path (x) edge (t);
\path (x) edge (y);
\path (t) edge (m);
\path (m) edge (y);
\end{tikzpicture}
\caption{Mediator}\label{fig:mediator_scm}
\end{subfigure}
}
\resizebox*{0.275\columnwidth}{!}{
\begin{subfigure}[b]{0.325\columnwidth}
\centering
\begin{tikzpicture}
\node (a) at (-1.5, 0) [label=left:$A$,point];
\node (t) at (-0.75, 0) [label={below left}:$T$,point];
\node (b) at (-0.75, -0.75) [label={below}:$B$,point];
\node (x) at (-0.75, 0.75) [label={above}:$X$,point];
\node (c) at (0.0, 0.75) [label={above}:$C$,point];
\node (d) at (0.0, 0) [label={below left}:$D$,point];
\node (e) at (0.0, -0.75) [label={below}:$E$,point];
\node (f) at (0.75, 0.75) [label={above}:$F$,point];
\node (y) at (0.75, 0.0) [label={right}:$Y$,point];
\node (g) at (0.75, -0.75) [label={below}:$G$,point];
\path (a) edge (t);
\path (x) edge (t);
\path (t) edge (b);
\path (x) edge (y);
\path(t)edge(d);
\path(c)edge(d);
\path(d)edge(e);
\path (d) edge (y);
\path (f) edge (y);
\path (y) edge (g);
\end{tikzpicture}
\caption{Multivariable}\label{fig:marloes_scm}
\end{subfigure}
}
\caption{Causal models where data-driven HTE estimation methods may perform poorly.
Some of the variables (e.g., $L$, $M$, and $E$) are not valid adjustment variables for estimating the effect of $T$ on $Y$.
}\label{fig:fail_examples}
\end{figure}
Structural causal models (SCMs) are a class of graphical models that model such causal mechanisms~\cite{pearl-book09}.
SCMs enable the selection of an \emph{adjustment set} -- a set of variables that are necessary for unbiased estimation of a causal effect of interest.
Consider the SCMs in Figure~\ref{fig:fail_examples} where $X$ is a confounding variable.
\new{SCM theory tells us that we need to adjust for \( X \) when estimating causal effects, but should not include variables that will introduce bias in effect estimation, such as descendants of the treatment or instrumental variables~\cite{myers-aje11} (e.g., \( L, M, A, D \)).
}
\new{However, SCMs are rarely known in real-world settings, so the correct adjustment variables cannot be found.
This makes data-driven HTE estimators an attractive alternative for practical applications.}
A common assumption used by data-driven HTE estimators is the assumption of \emph{strong ignorability} which states that the treatment assignment is independent of the potential outcomes given the variables in the data~\cite{rosenbaum-bio83,shalit-icml17}.
This assumption implies that there is no hidden confounding and HTE can be identified based on the given data.
However, hidden confounding is not the only problem that can lead to biased HTE estimation.
\new{
In practice, data-driven methods typically consider variables to be \emph{pretreatment} features, features which are measured before treatment.
However, this requires knowledge of the temporality of measurements, which may not be available.
Features in the data may actually be mediating variables, which should not be included in estimation.
}
To address the deficiencies of data-driven methods, we study the problem of causal feature selection for accurate data-driven HTE estimation when the SCM is unknown.
We develop two classes of algorithms for \emph{causal feature selection}, one based on local structure learning and another based on heuristic HTE evaluation metrics, neither of which has been studied previously in the context of HTE feature selection.
Through synthetic dataset experiments, we show that our feature selection algorithms improve existing data-driven methods and reduce error in causal effect estimations.
\subsection{Structure-Fit{}}
Now that we have selected the features that can contribute to HTE estimation error reduction, the next step is to select only the features which are a valid adjustment set. The second part of our framework,
Structure-Fit{}, iteratively applies local structure learning~\cite{aliferis-jmlr10,yu-arxiv19} to find a path from the treatment, $T$, to the outcome $Y$, in order to remove mediators and descendants of treatment.
Step 2 of Figure~\ref{fig:glcfs_fig} represents this part, Structure-Fit{}, where local structures from $T$ to $Y$ are learned iteratively, and connected. A local structure is a subgraph of the causal graph centered on a target variable.
For example, in Figure~\ref{fig:glcfs_fig}, the local structure of $T$ are the edges and variables connected to $T$ (e.g., $X_r^{(g)}, X_p^{(g)}, X_j^{(g)}$).
Since $Y$ is the outcome variable and we are only interested in paths from $T$ to $Y$, Structure-Fit{} terminates and the final feature set is found by removing mediators and descendants (crossed out in red).
This approach has two main advantages compared to full causal structure learning.
First, it only considers local structure around causal nodes and does not need to consider all variables if they are not connected to the causal path from \( T \) to \( Y \).
Second, it aims to only select variables which do not bias causal effect estimation and allow the HTE estimator to use features as appropriate for estimation.
\subsubsection{Discovering local structure around target variables}
Multiple structures can encode the same conditional independences in the data, making the task of finding the parent sets of $T$ and $Y$, $pa(T), pa(Y)$, nontrivial.
For example, $X_1 \rightarrow T \rightarrow X_2, X_1 \leftarrow T \leftarrow X_2$, and $X_1 \leftarrow T \rightarrow X_2$, all encode that $X_1 \perp X_2 | T$ and in one of them $X_1$ is a parent of $T$ whereas in the other two $X_1$ is a child.
We use two different ways to find the local structure.
The first way relies on finding the parent-children ($PC$) set around our target variables, $T$ and $Y$~\cite{aliferis-jmlr10}, collider discovery, and edge orientation methods to find causal directions~\cite{daniusis-uai12,fonollosa-book19,blobaum-aistats18}.
Several $PC$ discovery algorithms have been developed~\cite{buhlmann-bio10,pena2005scalable,aliferis-jmlr10,yu-arxiv19} and have mainly been used for improving classification methods~\cite{yu-arxiv19}.
To distinguish between parents and children, we first discover the parents of each target variable through collider discovery.
If two variables are parents of the target variable (e.g., \( X \) and \( A \) are parents of \( T \)), then they are dependent if conditioning on the child and can be discovered through conditional independence tests.
However, if there is zero or one parent, we will not leverage collider discovery.
Instead, we orient the remaining edges using an edge orientation algorithm~\cite{daniusis-uai12,fonollosa-book19,blobaum-aistats18} and then check if there is a parent in the set.
Using these algorithms, we can check for any variable in $PC$ whether an arrow is pointing in or away from the target variable (i.e., $T$ and $Y$) and thus identify $pa(T), pa(Y), ch(T)$, and $ch(Y)$.
The second way to find local structures is to utilize more recent local structure learning algorithms, such as PCD-by-PCD~\cite{yin-causation08}, MB-by-MB~\cite{wang-csda14}, and Causal Markov Blanket~\cite{gao-neurips15}
These algorithms discover the $PC$ or Markov Blanket ($MB$) while partially orienting edges.
We can find a partially oriented local structure and orient edges using edge orientation methods.
\subsubsection{Learning partial structures}
Using the process to discover local structure around target variables, we propose an iterative algorithm to remove post-treatment variables from HTE-Fit{}, which we call Structure-Fit{}.
We apply local structure learning on the treatment variable, $T$, and then iteratively apply local structure learning on children of $T$ in a breadth-first-search manner until we reach the outcome variable $Y$.
Then, local structures are put together to form a subgraph of the causal graph over the variables selected by HTE-Fit{}.
After learning the local structure, we can apply an identification algorithm to find an adjustment set, such as the O-set~\cite{henckel-arxiv19}.
Another option is to only remove all descendants of treatment.
We demonstrate Structure-Fit{} using Figure~\ref{fig:marloes_scm}.
We first find $PC(T)=\{A, X, B, D\}$ and $ch(T)=\{B, D\}$.
We proceed to iteratively find the $PC$ on each child, $B, D$.
Since $B$ has no children, we move to $D$ and find $PC(D)=\{T, C, E, Y\}$ and $ch(D)=\{E, Y\}$.
Finally, $E$ has no children, and we have reached the outcome variable, $Y$ and find $PC(Y)=\{X, C, F, G\}$.
Using the local structures, we find the causal graph shown in Figure~\ref{fig:marloes_scm}.
In practice, we only use the output of HTE-Fit{} in Structure-Fit{}, so not all variable will be used.
\section{Experiments}
To evaluate whether and how much feature selection improves existing data-driven HTE algorithms, we generate synthetic datasets which contain ground-truth causal effects.
We also use a semi-synthetic dataset and two real-world datasets.
\subsection{Experimental setup}
In our experiments, we make a distinction between four classes of estimation methods: \emph{base learners}, \emph{baseline feature selection}, \emph{our feature selection}, and \emph{oracle}.
A base learner is an HTE estimator without feature selection, and it uses all features for estimation.
We consider several prominent HTE estimators as base learners:
Causal Tree Learn~\cite{tran-aaai19}, Causal Forest~\cite{athey-annals19},
Meta-learners (S, T, XLearners)~\cite{kunzel-pnas19}, Doubly Robust learner (DRLearner)~\cite{kennedy-arxiv20}, Balancing Neural Network (BNN)~\cite{johansson-icml16}, and TARNet~\cite{shalit-icml17}.
The baseline feature selection methods use local structure learning~\cite{yu-arxiv19} and causal structure learning~\cite{heinze-stat18,henckel-arxiv19} to select an adjustment set.
The first baseline, \emph{PC Simple (Parents)}, uses a local structure learning method on the treatment to find parents.
PC Simple has been used for selecting variables for classification~\cite{yu-arxiv19} but not for HTE estimation.
We employ PC Simple to learn the local structure on the treatment variable and select the treatment parents as an adjustment set.
Since our setting only considers one treatment variable, the parents of treatment are a valid adjustment set~\cite{williamson-resp14}.
Two other baselines we consider are based on causal structure learning~\cite{heinze-stat18,henckel-arxiv19}, which we refer to as Full Structure Learning (FSL){}.
We use GFCI~\cite{ogarrio-pgm16} to learn a causal structure and apply two adjustment criteria: optimal adjustment set (O-set) and removing post-treatment variables only (Valid).
Then a base learner uses the adjustment set for estimation (e.g., TLearner).
For our feature selection methods, we consider our framework HTE-FS{} and each component, HTE-Fit{} and Structure-Fit{}, separately.
For HTE-Fit{}, we use both variants: \globalmethodname{}-F{} and \globalmethodname{}-B{}.
We use four HTE fit metrics: $\tau$-risk~\cite{schuler-model18}, Nearest Neighbor Precision in Estimating Heterogeneous Effects (NN-PEHE)~\cite{schwab-perfect18}, plug-in $\tau$~\cite{saito-cv19}, and Counterfactual Cross-validation (CFCV)~\cite{saito-cv19}.
We report results using \globalmethodname{}-F{} and $\tau$-risk since this combination performed the best in our experiments.
For Structure-Fit{} we used PC-Simple~\cite{buhlmann-bio10}, GetPC~\cite{pena2005scalable}, and HITON-PC~\cite{aliferis-jmlr10}, CMB~\cite{gao-neurips15}, and PCD-by-PCD~\cite{yin-causation08}.
For pairwise edge orientation, we use Conditional Distribution Similarity Statistic (CDS)~\cite{fonollosa-book19}, Information Geometric Causal Inference (IGCI)~\cite{daniusis-uai12}, and Regression Error based Causal Inference (RECI)~\cite{blobaum-aistats18}.
In order to find an adjustment set, we only remove mediators and descendants, and allow data-driven HTE estimators to decide which variables to use in estimation.
We report results using PC Simple with RECI (PC, RECI) and CMB since these two combinations represent different variants of Structure-Fit{} and perform the best in our experiments.
For HTE-FS{}, we report results using \globalmethodname{}-F{} with $\tau$-risk and Structure-Fit{} with CMB, since this combination performed the best in experiments.
We compare against using Oracle (Known Structure), which uses the true optimal adjustment set.
\subsection{Datasets}
We use three types of datasets.
The first dataset is a set of synthetic datasets, in which random SCMs are generated, and data is generated following each SCM.
The second dataset is based on the Infant Health and Development Program (IHDP) dataset~\cite{hill-jcgs11}.
The third type are real-world datasets, League of Legends and Cannabis.
\subsubsection{Synthetic dataset}
We generate SCMs by following a procedure similar to one for evaluating causal structure learning algorithms~\cite{heinze-stat18}.
SCMs and their corresponding datasets differ in the following characteristics: the number of variables \( d \); the probability of edge creation \( p_e \); the variance in the noise \( \sigma \); the strength of the noise term in non-source nodes \( \rho \).
We add parameters to control confounding \( \gamma \), mediation \( m \), and heterogeneity in treatment effects \( p_h, m_p \).
We initialize a set of parameters as follows:
\begin{enumerate}
\item Number of variables: \( d \in [10, 20, 30] \)
\item Edge probability parameter: \( p_e \in [0.1, 0.3, 0.5] \)
\item Noise variance: \( \sigma \in [0.2, 0.4, 0.6] \)
\item Magnitude in noise: \( \rho \in [0.1, 0.5, 0.9] \)
\item Confounder: \( \gamma \in [\text{True, False}] \)
\item Mediator chain length: \( m \in [0, 1, 2] \)
\item Number of HTE inducing parents: \( p_h \in [0, 1, 2] \)
\item HTE from mediating parents: \( m_p \in [\text{True, False}] \)
\end{enumerate}
For each parameter value, we randomly sample from all other parameter values \( 500 \) times to generate SCMs.
For each of the $11,500$ SCMs, we create one dataset of size \( 10,000 \).
Root node variables are generated from a normal distribution, and non-root nodes are generated linearly from parents.
The detailed dataset generation description is available in the Appendix in Section~\ref{sec:data_gen_desc}, and the link for the code is available in Section~\ref{sec:code}.
\subsubsection{Semi-synthetic dataset: IHDP}
The covariates in the Infant Health and Development Program (IHDP) dataset come from a randomized experiment studying the effects of specialist home visits on future cognitive test scores~\cite{hill-jcgs11}.
Outcomes can be simulated using the covariates.
We simulate 1,000 instances of simulated outcomes and average all results\footnote{Outcomes are generated using the JustCause package:\\https://github.com/inovex/justcause}.
\subsubsection{Real-world datasets}
\kdd{
We use two real-world datasets: \emph{League of Legends} and \emph{Cannabis}.
League of Legends (LoL) is a popular multiplayer online battle arena (MOBA) game.
Riot Games, the developer of LoL, regularly creates software patches to the game that update various parts of the game, affecting balance and how players decide to play the game.
We use a dataset consisting of player, match, and patch information over time~\cite{he-fdg21}.
The treatment variable of interest is a software patch, and the outcome of interest is the number of \emph{kills} of a team in a single match, which is a post-match statistic.
We average results over patches in the dataset.
}
\kdd{
\emph{Cannabis} is a dataset consisting of tweets about the e-cigarette Juul and cannabis-related topics~\cite{adhikari-icwsm21}.
English tweets related to the e-cigarette Juul were collected from 2016 to 2018.
For the users who mention Juul, cannabis-related tweets are collected from 2014 to 2018.
In this dataset, tweets' stances towards Juul and cannabis are crowdsourced, and a model is trained for stance detection.
In this work, we are interested in how users' stance on Juul affects their stance on cannabis.
To do this, we find users who have Juul-related tweets occurring before a cannabis-related tweet and detect their stance on Juul and cannabis using the classifier from~\cite{adhikari-icwsm21}, where features are the word embeddings for tweets using BERT~\cite{devlin-nacl19}.
The treatment variable of interest is a user's stance on Juul, in favor (1) or against (0).
The outcome variable is their stance on cannabis, in favor (1) or against (0).
Since this dataset contains only one group, we bootstrap the estimators to compute the standard deviation.
}
\input{figures/learner_compare.tex}
\input{figures/winner_table.tex}
\subsection{Evaluation}
We use the mean squared error (MSE) for evaluating HTE estimations since the true causal effect is known with synthetic and semi-synthetic data: $\text{MSE} = \frac{1}{N} \sum_{i=1}^N (\tau(\mathbf{X}) - \hat{\tau}(\mathbf{X}_i))^2$.
For each dataset, we split the data into 80/20 training and test sets.
Since different SCMs will result in different MSEs, we first rank all estimation methods based on their MSE for a particular SCM.\@
Then we average the rank of each method across SCMs and report on the average rank.
To study their performance under specific settings, we also report on average ranks for subsets of parameter settings.
For real-world datasets, we report results using $\tau$-risk~\cite{nie-rlearner17,schuler-model18}, which we defined in eq~\eqref{eq:tau_risk}.
Similar to MSE, we report rankings based on $\tau$-risk.
\kdd{
In addition, for datasets in which we can decide which variables \emph{post-treatment} variables, we compute the percentage of included by each feature selection method, which we denote as \emph{inclusion error (IE)}.
Given the \emph{post-treatment} set, $\mathcal{F}$, and the predicted causal feature set $\mathbf{X}^{(\ell)}$, we compute the inclusion error as:
\begin{equation}
\text{IE} = \frac{| \mathbf{X}^{(\ell)} \cap \mathcal{F} |}{|\mathcal{F}|}.
\end{equation}
Some variables occur after treatment for real-world datasets, such as other post-match statistics in LoL and tweets related to cannabis after the initial cannabis-related tweet in the Cannabis dataset.
}
\subsection{Evaluation on synthetic data}
\kdd{
We first present the results on synthetic datasets.
We first show how our feature selection method improves multiple base learners.
Then we show how feature selection improves base learners across multiple structural configurations.
}
\subsubsection{Feature selection helps all base learners}
Figure~\ref{fig:learner_compare} shows the MSE comparisons between several HTE estimators and feature selection methods on SCMs which are generated using: 30 variables, confounding, mediator chain length of 2, 2 HTE inducing parents, and HTE induced by mediating parents (i.e., \( (d,\gamma,m,p_h,m_p) = (30, True, 2, 2, True) \)).
We average MSE over all other parameters (i.e., \( p_e, \sigma, \rho \)).
\kdd{
We show the best performing Structure-Fit{}, \globalmethodname{}-F{}, HTE-FS{}.
This figure shows that adding feature selection improves MSE for all HTE estimators.
In addition, feature selection reduces the error bars for all base learners.
In general, HTE-FS{} (Tau risk, CMB) performs the best, compared to using \globalmethodname{}-F{} or Structure-Fit{} alone.
Structure-Fit{} with CMB performs better than Structure-Fit{} with PC Simple and RECI, although the latter still reduces error overall compared to using the base learner on all features.
}
\subsubsection{Feature selection improves base learners for different types of structures}
\kdd{
Table~\ref{tab:mse_rankings} ranks the feature selection methods using TLearner as the base learner vs.\ several other HTE estimators (without using feature selection) under a range of structural scenarios.
}
\kdd{
We choose TLearner as the base learner since it is a relatively simple model and performs consistently in all settings.
}
One experimental parameter that affects the causal structure is fixed in each column while averaging over the other parameters.
The top section shows the average rank for base learners without feature selection, the second section shows baseline feature selection methods, the third section shows our feature selection methods, and the bottom section shows the Oracle method.
Bolded results indicate the best rank in the column (ignoring the Oracle estimator).
The large variety of underlying structures explains the high variance in rank.
\kdd{
The table shows that our feature selection methods improve base learners for different structures, perform better than baseline feature selection methods, and perform better than other base learners without feature selection.
HTE-FS{} performs the best overall.
In two settings, with no confounding and a low number of variables, Structure-Fit{} with CMB performs the best.
While our methods perform better than baseline methods, baseline methods still improve base learners, compared to no feature selection.
}
\input{figures/wrong_variables.tex}
\subsubsection{Feature selection reduces inclusion of wrong variables}
\kdd{
In addition to performance in terms of MSE, we investigate the \emph{inclusion error (IE)} of feature selection methods.
Figure~\ref{fig:wrong_variables} shows the inclusion error for each feature selection method, averaged over all experiments, where lower is better.
The grey line represents the case of randomly selected features.
The first set of bars shows the average percentage of overall synthetic datasets.
All feature selection methods do better than random feature selection.
HTE-FS{} has the lowest chance to include wrong variables, at about 8\% inclusion rate.
This result supports the results shown in Table~\ref{tab:mse_rankings}, that HTE-FS{} has the lowest error, followed by Structure-Fit{} with CMB and \globalmethodname{}-F{}.
}
\subsection{Evaluation on IHDP dataset}
\input{figures/ihdp_table.tex}
\kdd{
Table~\ref{tab:ihdp_rankings} shows the average rankings on the IHDP dataset using MSE and $\tau$-risk.
}
We used TLearner as the base learner and showed the same variations in feature selection as in Table~\ref{tab:mse_rankings}.
We see that all feature selection methods perform better on average than the base estimators, with HTE-FS{} performing the best.
Since TLearner is a linear model close to the data generation process for this dataset, it already performs very well on it.
Since this dataset contains pre-treatment features only, we do not evaluate inclusion error.
\subsection{Evaluation on real-world datasets}
\kdd{
Finally, we show results on real-world datasets.
Since these datasets do not contain ground truth effects, we evaluate methods using $\tau$-risk rankings.
}
\subsubsection{LoL dataset}
\kdd{
The average rank for each method in terms of $\tau$-risk for the LoL dataset is shown in Table~\ref{tab:ihdp_rankings}.
Overall, HTE-FS{} achieves the lowest $\tau$-risk rank among all estimators.
We see that TLearner performs the best among base learners and has a lower rank than some feature selection methods.
One reason could be that some methods are not suitable for high dimensional datasets, such as Structure-Fit{} with PC Simple.
However, Structure-Fit{} with CMB performs much better than PC Simple, \globalmethodname{}-F{}, and FSL{}.
}
\kdd{
In the LoL dataset, there are pre-match and post-match features.
Since our outcome \emph{kills} is a post-match feature, other post-match features should not be included in the estimation, which is the \emph{post-treatment} set in this setting.
The second set of bars in Figure~\ref{fig:wrong_variables} show the inclusion error for the LoL datasets.
Overall, each feature selection method has a higher IE than in the synthetic dataset.
HTE-FS{}, Structure-Fit{} with CMB, and \globalmethodname{}-F{} reduce the inclusion error rate on average, while Structure-Fit{} with PC-Simple and FSL{} do not reduce the inclusion error more than random selection.
}
\subsubsection{Cannabis}
\kdd{
The rank in terms of $\tau$-risk for the Cannabis dataset is shown in Table~\ref{tab:ihdp_rankings}.
Since this dataset contains only one set of data, we bootstrap each estimator.
In this dataset, HTE-FS{} achieves the lowest $\tau$-risk, followed by FSL{} and \globalmethodname{}-F{}.
Here, XLearner achieves the lowest $\tau$-risk among base learners and performs close to our proposed methods.
}
\kdd{
In the Cannabis dataset, the post-treatment features are tweets after the first cannabis-related tweet.
The features that should be included are only tweets related to the stance towards Juul and the tweet related to the first cannabis tweet.
Figure~\ref{fig:wrong_variables} shows the inclusion error for the Cannabis dataset in the rightmost set of bars.
In most cases, our feature selection methods reduce the number of wrong variables included, except Structure-Fit{} with PC-Simple.
}
\section{Related Work}
We provide a brief overview of related work on estimating heterogeneous treatment effects and structure learning.
\textbf{Data-driven methods for HTE estimation.}
Many methods have been developed for estimating heterogeneous treatment effects (HTEs) from data.
Several methods rely on recursive partitioning using tree-based methods, which naturally partition the population into subgroups that may contain heterogeneous effects~\cite{athey-pnas16,athey-annals19,tran-aaai19,su-jmlr09,zeileis-jcgs08}.
Besides tree-based methods, methods have been developed that use machine learning to predict causal effects, such as meta-learners~\cite{kunzel-pnas19,kennedy-arxiv20} and neural networks~\cite{johansson-icml16,shalit-icml17,schwab-perfect18}.
However, these methods do not consider the structure of the underlying causal mechanisms that generate the data, introducing bias in estimation.
\textbf{Causal structure learning.}
Many causal structure learning algorithms have been developed for recovering SCMs from data~\cite{heinze-stat18,vowels-arxiv21}.
Heinze-Deml et al.\ provide a recent survey on causal structure learning algorithms and evaluate their performance in various settings~\cite{heinze-stat18}.
However, learning the full causal structure from data can be computationally expensive for high-dimensional datasets~\cite{chickering-jmlr04,vowels-arxiv21}.
Related to causal structure learning is local structure learning~\cite{aliferis-jmlr10}.
Aliferis et al.\ presented an algorithmic framework for learning local causal structure around target variables, the parent-children set (\( PC \)) and the Markov Blanket (\( MB \)).
Yu et al.\ survey and evaluate several local causal structure learning methods that learn both \( PC \) and \( MB \) of target variables~\cite{yu-arxiv19}.
They study local structure learning under the context of causality-based feature selection with the goal of improving supervised learning.
In contrast, we study the problem of causal feature selection.
Local structure learning has been studied in the context of average treatment effect (ATE) estimation, but not HTE estimation~\cite{cheng-arxiv20}.
This work also does not consider excluding mediators or descendants, which we do in our work.
\section{Acknowledgments}
This material is based on research sponsored in part by NSF under grant No.\@ 2047899, DARPA under contract number HR001121C0168, and an Adobe Data Science Research Award.
\bibliographystyle{ACM-Reference-Format}
\section{Appendix}
Here we provide additional information for Heterogeneous Treatment Effect Feature Selection (HTE-FS){} and the synthetic data generation.
\subsection{HTE-Fit{}}\label{sec:hte_fit_desc}
Here we present pseudocode for the two HTE-Fit{} algorithms.
The \globalmethodname{} --- Forward{} algorithm is shown in Algorithm~\ref{alg:forward_selection}. Lines 1-8 first find the one feature that returns the largest score (computed by $R$) and adds it to \( \mathbf{X}^{(g)} \).
Then, for each remaining feature \( X^{(j)} \), we compute the score when combining \( X^{(j)} \) to \( \mathbf{X}^{(g)} \), and concatenate the feature that returns the best score.
This repeats until the score does not change, or there are no more features to add.
\globalmethodname{} --- Backward{} works similarly to \globalmethodname{}-F{} and the pseudocode is shown in Algorithm~\ref{alg:backward_selection}.
We start by computing the score using all features in Line 1.
In Lines 9-15, we iteratively add features to \( \mathbf{K} \), which are features that when removed improve the score.
The final set returned is the whole feature set without \( \mathbf{K} \).
\input{99_algorithms.tex}
\subsection{Structure-Fit{}}\label{sec:structure_fit_desc}
Algorithm~\ref{alg:lcfs} details Structure-Fit{}.
In lines 1-2, we initialize two variables: $dn$, which are discovered nodes, $\mathbf{Z}$, a queue of nodes to traverse.
The discovered nodes are initialized to $ch(Y)$, so we do not traverse past $Y$.
In lines 3-5, we initialize a graph $\hat{G}$ which will be the local structure from $T$ to $Y$, and $ \mathbf{Z} $ is initialized to be the children of $ T $.
Lines 6-10 show the iterative process in which we consider the nodes in \( \mathbf{Z} \).
For every node \( Z \in \mathbf{Z} \), find \( pa(Z) \) and \( ch(Z) \).
$Z$ is added to the discovered nodes, is removed from $\mathbf{Z}$, and the children of $ Z $ are added to $ \mathbf{Z} $.
Then we connect $Z$ to its children.
In the end, a graph, $\hat{G}$ with causal paths from $T$ to $Y$ is constructed, and we can construct the valid set.
\input{04_02_algorithm}
\subsection{Complexity of HTE-FS{}}\label{sec:complexity}
The complexity of HTE-FS{} depends on the complexity of training and prediction of the HTE estimator and the local structure learning algorithm.
Given an HTE estimator with complexity $ O(H) $ training and $ O(P) $ prediction, HTE-Fit{} takes $ O(d^2 \cdot (H + P) ) $.
If HTE-Fit{} outputs $ s $ variables and we use CMB for Structure-Fit{}, then the worse case scenario happens when we find local structures for all $s$ variables: $ O(s \cdot 2^{|PC|}|U|^2|PC|) $~\cite{gao-neurips15}.
\subsection{Synthetic data generation}\label{sec:data_gen_desc}
Here, we describe how we generate synthetic datasets used in the experiments.
We first generate an adjacency matrix, \( \mathbf{A} \) over the \( d \) variables.
Assuming that the variables \( \{1, \dots, d \} \) are causally ordered.
For each pair of nodes \( i, j \), where \( i < j \) in the causal order, we sample from Bernoulli(\( p_e \)) to determine if there is an edge from \( i \) to \( j \).
With the adjacency matrix, we sample the corresponding coefficients, \( \mathbf{C} \), from Unif(\( -1, 1 \)).
Noise terms \( \epsilon_i \) are generated by sampling from a \( d \)-dimensional zero-mean Gaussian distribution with covariance matrix \( \Sigma \), where \( \Sigma_{i,i} = 1 \) and \( \Sigma_{i,j} \) = \( \sigma \).
We use \( \rho \) to tune the magnitude of the noise.
Next, we choose the treatment and outcome nodes based on two parameters, \( \gamma \) and \( m \). \( \gamma \) controls whether there is confounding in the generated dataset and \( m \) controls the length of the mediating chain (\( 0 \) means no mediator).
First, we select all pairs of nodes which have a \emph{directed} \( m \) hop path.
Then, if \( \gamma \) is True, we filter all pairs that do not contain a \emph{backdoor path}.
Otherwise we filter out pairs that contain a backdoor path.
Finally, we sample a pair of nodes from the filtered set to obtain the treatment and outcome nodes.
If no pairs of nodes satisfy the criteria for the given parameter setting, we resample the graph up to 100 times until a graph is found, or we discard the parameter setting.
A source node \( i \) (node with no parents) takes the value of the sampled noise terms, \( \epsilon_i \).
For a non-source node, \( i \), we consider several additive factors.
First is the base term from the coefficient matrix: \( C_{pa(i), i} X_{pa(i)} \), where \( C_{pa(i), i} \) are the coefficients from \( pa(i) \) to \( i \) and \( X_{pa(i)} \) are the values generated for \( pa(i) \).
Next, are the interaction terms, which are determined by the mediating chain length \(m\), number of HTE inducing parents \( p_h \), and whether there is HTE from parents of mediators \( m_p \).
For mediators, we have: \( \mathbbm{1}_{m}(i) \mathbbm{1}_{m_p} X_{pa_m(i)} X_{pa_n(i)} \), where \( \mathbbm{1}_{m} \) is an indicator for if \( i \) is a mediator, \( \mathbbm{1}_{m_p} \) is an indicator for if there is HTE induced from mediator parents, and \( pa_m \) and \( pa_n \) represent mediating and non-mediating parents.
For outcomes, the interaction is defined similarly: \( \mathbbm{1}_{y}(i) X_{pa_m(i)} X_{pa_n(i)} \), where \( \mathbbm{1}_{y}(i) \) indicates whether the node is the outcome node.
The number of HTE parents determines how many non-mediating parents (\( pa_n \)) the interaction occurs with.
A value of zero means there is no heterogeneity.
So the final data generating equation is:
\begin{align}
X_i & = C_{pa(i), i} X_{pa(i)}
\nonumber \\
& + \mathbbm{1}_{m}(i) \mathbbm{1}_{m_p} X_{pa_m(i)} X_{pa_n(i)}
\nonumber \\
& + \mathbbm{1}_{y}(i) X_{pa_m(i)} X_{pa_n(i)} + \epsilon_i
\end{align}
\subsection{Code}\label{sec:code}
Code for HTE-FS{} and the synthetic data generation is available at \href{https://github.com/edgeslab/causal_feature_selection}{https://github.com/edgeslab/causal\_feature\_selection}.
\section{Causal Feature Selection for HTE Estimation}\label{sec:cfs}
When the SCM that generated the data is unknown, a vanilla solution for finding an adjustment set for HTE estimation is to learn the causal structure and then find the optimal adjustment set (O-set)~\cite{henckel-arxiv19}.
However, structure learning is computationally expensive in high-dimensional data, both in theory~\cite{chickering-jmlr04} and in practice~\cite{heinze-stat18,vowels-arxiv21}.
Instead of doing full structure learning, we propose Heterogeneous Treatment Effect Feature Selection (HTE-FS){}, which relies on two ideas to make feature selection more practical.
First, instead of considering all features, it assesses the value of each feature in reducing the HTE estimation error and selects only the features that contribute to error reduction.
Second, instead of learning the full structure, it focuses on learning the parts of the structure that give sufficient information about selecting a valid adjustment set.
Figure~\ref{fig:glcfs_fig} shows a high-level overview of HTE-FS{}, and we present its main components, HTE-Fit{} and Structure-Fit{}, next.
\input{04_01_heuristic.tex}
\input{04_02_local.tex}
\input{04_03_global_local.tex}
\subsection{HTE-Fit{}}
The first component of our framework, HTE-Fit{}, proposes the use of \emph{heuristic HTE evaluation} metrics to sequentially select features that are likely to reduce the HTE estimation error.
This is shown as the first step in Figure~\ref{fig:glcfs_fig}.
It has two variants: forward (\globalmethodname{}-F{}) and backward (\globalmethodname{}-B{}).
Since the ground truth of HTE effects in unknown, several heuristic HTE evaluation metrics have been developed to approximate the errors of HTE estimators, including $\tau$-risk~\cite{schuler-model18}, Nearest Neighbor Precision in Estimating Heterogeneous Effects (NN-PEHE)~\cite{schwab-perfect18}, Plug-in $\tau$~\cite{saito-cv19}, and Counterfactual Cross-validation (CFCV)~\cite{saito-cv19}.
The $\tau$-risk takes as input the estimator $\hat{\tau}$ and approximates the objective defined in~\eqref{eq:cate_objective}~\cite{schuler-model18}.
The second metric, NN-PEHE, uses nearest neighbor matching for imputing treatment effects for training~\cite{schwab-perfect18}.
The plug-in $\tau$ uses an arbitrary estimator for imputing effects for computing error.
The fourth metric, CFCV, uses a doubly robust formulation with the plug-in $\tau$ for imputing effects for evaluation~\cite{saito-cv19}.
These metrics have been shown to correlate with HTE estimator performance~\cite{schuler-model18,saito-cv19}.
Here, we propose to use these metrics as causal feature selection criteria, which, to the best of our knowledge, has not been studied in previous research.
\subsubsection{\globalmethodname{} --- Forward{}}
\globalmethodname{}-F{} starts with a feature with the best (lowest) error metric, then iteratively adds features that improve the metric.
If there are no more features or improvements, the selection stops.
We will use $\tau$-risk as an example metric to make this procedure more concrete. $\tau$-risk has been shown to lead to a consistent high-performing model~\cite{schuler-model18} and to have quasi-oracle error bounds on the CATE~\cite{nie-rlearner17}.
The $\tau$-risk, denoted as $R$, takes as input the estimator $\hat{\tau}$ and approximates the objective defined in~\eqref{eq:cate_objective}~\cite{schuler-model18}:
\begin{equation}\label{eq:tau_risk}
\tau\text{-risk}(\hat{\tau}) = R(\hat{\tau}) = \frac{1}{N} \sum_{i=1}^N \Big(
(Y_i - \hat{m}(\mathbf{X}_i)) - (T_i - \hat{p}(\mathbf{X}_i)) \hat{{\tau}}(\mathbf{X}_i)
\Big)^2.
\end{equation}
Here, \( \hat{m}(\mathbf{X}_i) \) is an outcome estimate for unit $i$ (i.e., \( E[Y_i | \mathbf{X}_i] \)), and \( \hat{p}(\mathbf{X}_i) \) is an estimate of the propensity score, (i.e., \( E[T_i | \mathbf{X}_i] \)).
Both $\hat{m}$ and $\hat{p}$ are learned through supervised techniques and optimized for predictive accuracy.
Suppose the feature vector of individuals has $d$ dimensions, $\mathbf{X} = (X_1, \dots, X_d)$.
Given the $\tau$-risk, we can compute the feature fit for one feature, say $X_j \in \mathbf{X}$, by estimating $\hat{\tau}(\{X_j\})$ and inputting the estimator into the $\tau$-risk.
For \globalmethodname{}-F{}, the first feature found is the feature that minimizes $R(\hat{\tau}(\{X_j\})$, and we define the initial set of found features as $\mathbf{X}^{(g)} = \{X_j\}$.
Now that the first feature is found, we need to add another to $\mathbf{X}_{(g)}$.
For each feature $X_k \in \mathbf{X}, X_k \notin \mathbf{X}^{(g)}$, we estimate a new $\hat{\tau}(\mathbf{X}^{(g)} \cup \{X_k\})$ which is the current set of found features $\mathbf{X}^{(g)}$ combined with a new feature.
The next feature added to $\mathbf{X}^{(g)}$ is the feature that minimizes $R(\hat{\tau}(\mathbf{X}^{(g)} \cup \{X_k\})$.
This process continues until there is no feature that, when combined with $\mathbf{X}^{(g)}$, minimizes the $\tau$-risk or when there are no more features to add.
\subsubsection{\globalmethodname{} --- Backward{}}
Backward feature selection (\globalmethodname{}-B{}) works similar to forward feature selection.
The difference is that it starts with all features, $\mathbf{X}^{(g)} = \mathbf{X}$, and it
iteratively removes features until either the error metric does not improve or there is only one feature left for estimation.
Using $\tau$-risk, it would first compute $R(\hat{\tau}(\mathbf{X}))$ to get an initial score on the entire feature set.
For each $X_j \in \mathbf{X}$, it would compute $R(\hat{\tau}(\mathbf{X} \setminus \{ X_j \})$, and remove the feature that results in the lowest $\tau$-risk when removed.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 14
|
Management Secrets Of Barack Obama
Why is there not
In the land of Obama
only more than
A hint of drama?
Barack Obama's presidential campaign is a sight to behold for many reasons, but among the least appreciated is its allergy to the disunity and spatting that generally convulses even the most successful of endeavors.
Obama, who never managed so much as a newsstand, has turned out to be a fairly remarkable leader of a what's becoming a billion dollar enterprise.
Not that there haven't been disputes; not that there haven't been rollicking disagreements about strategy and message and tactics; not that people haven't been fired; not that personalities haven't clashed; some high-ranking Obama advisers really embraced the concept of multiple town hall meetings with John McCain while others were wary; after some half-hearted stabs at negotiation, the idea kicked the bucket. (By the way: kicking the bucket is not a term of endearment for vegans; a dollar to the person who can explain why.)
So what principles are responsible for the discipline, the coherence and relational tightness of the campaign? The question is relevant because they tell us something about how Obama might set up his White House, and the principle of function following form applies.
1. A challenging, clearly-defined mission.
Win the Democratic nomination. Ok, that's obvious. But in November of 2006, Obama had a handful of wealthy Jewish donors from Chicago ready to raise money, a few Democratic strategy types who stood ready to put a campaign together, and not much of anything else. The challenge was immense, and the folks who joined the Obama campaign early on -- this was when Hillary Clinton was the frontrunner -- came aboard because they believed in Obama the challenge, not because they expected glory or material rewards. Salaries weren't competitive with the Clinton campaign's either. Obama attracted a large number of ideological Democrats who either had reason to dislike the Clintonian influence over the party or who believed that Obama stood at the crossroads between history and hope.
2. Clear lines of authority, with budget power appropriately vested in the campaign manager.
David Plouffe was named campaign manager and given the portfolio of essentially building an airplane as it was speeding down the runway. Plouffe's authority did not derive from his personal relationship with Barack Obama; indeed, Plouffe really didn't interact much with Obama before the campaign. Instead, it derived from the decisions he made that Obama later ratified. A corollary: Obama's best friend, Marty Nesbitt, was named campaign treasurer. That meant that Nesbitt, too, had oversight functions about spending and could make sure that Obama's personal values were reflected in the decisions made by the campaign.
David Axelrod would handle strategy, always a kind of messy and intrusive portfolio, but the strategy was pretty clear from the start: this is a change election, and Obama's the change agent.
Robert Gibbs would handle the press; and handle is a good word, because Obama's relationship with the press has been fascinating as it has evolved.
Their subordinates were generally given tasks and clear lines of authority.
3. The tone from the top.
Much has been written about Plouffe's preternatural unflappability, but his attitude helps confirm a basic principle of psychology: when the boss is freaked, employees get freaked. When the boss is calm, employees tend to remain calm.
But there's Obama's tone, too. Generally happy. "No drama." That's made clear to new employees, who feel the presence of the dictum as they go about their work. Drama disappoints Obama. No one wants to disappoint Obama. So -- if there are conflicts -- staffers are advised to work them out by themselves. A social cuing effect helps to cement the bond here; employees see that everyone else is getting along and feel pressure to get along, too.
4. Protecting Obama's public image at all costs
This one's also obvious, but it has been a priority from day one. The type of people drawn to the campaign are the type of people who would be angry at themselves if they somehow distorted the carefully crafted image that Obama and guru Gibbs et. al. had worked to create.
5. The personality cult
Not meant as an insult or a reference to messianic fervor; I mean more a reflection of Obama's personality. Obama is an occasionally fastidious perfectionist. He expects a high level of professional integrity from his staff; he expects a higher level of execution. When things go wrong, Obama doesn't just leave the aftermath to his subordinates, he actively helps to clean it up.
6. Decisions made quickly and decisively.
Occasionally, Obama temporizes. But just as often, he endorses quick, clear decision-making.
7. The freedom to do things differently.
Let the Clintons run their campaign the traditional way. Obama had nothing to lose and everything to gain, and so his campaign could experiment; they were trying to accomplish something that no challenger had every before accomplished, so they could afford to try new things, to throw out the old rulebooks, to ignore established habits of mind, and take some risks.
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 1,516
|
\section{Introduction}
The growth in computing power together with data science has allowed the emergence of data-driven computational mechanics in which one aims at e.g. material parameter identification, improving or complementing standard computational mechanics approaches such as FEM, or governing equations discovery, etc. \citep{HDAZ2020, OISHI2017327, KIRCHDOERFER201681, Raissi2018}.
Over the past few years, Artificial Neural Networks (ANNs) \citep{Rosenblatt58theperceptron}, which provide input-output transfer means for complex problems, have been applied successfully in this field. One major issue of using this approach is the training procedure which can be considerably time-consuming and sensitive to several ANN-related parameters. In the latter process, the learnable parameters (the ones such as its \emph{weights} and \emph{biases} that determine the ANN output) are tuned by minimising a distance function called the cost function which is calculated based on, in the simplest case, a dataset including the exact solutions to be interpolated.
Furthermore, ANN training sometimes is only a one-time calculation \citep{HDAZ2020} which allows the choice of non-efficient approaches.
However, for \emph{strong solution} of mechanical problems (i.e. the solution of the strong-form of the problem) described by a system of Partial Differential Equations (PDEs) a training dataset of exact solutions is not available in which case the cost function is constructed based on the governing PDEs (so-called Physics Informed ANNs) \citep{Raissi2018}. In fact, in this concept, solving a problem such as a linear elastic deformation problem by ANN means training a network which provides a space-displacement relationship satisfying the strong form of governing physical laws in the form of PDEs which is enforced via the cost function. For a better understanding of the general methodology one could assume that ANN is a complex ansatz, relating the coordinates of interest to the corresponding displacements, whose parameters are tuned by minimising the mentioned distance function.
The provided solution (displacement field), as well as its spatial gradient (strain/stress field), are continuous, directly differentiable, and free of space discretisation. The latter, results in increased flexibility and less complexity of the problem implementation. For example, for obtaining the mechanical response at a specific coordinate one does not need to discretise and solve the entire domain, which will be shown in the results. However, the training procedure can be challenging and inefficient as we need to compute the third order gradients to solve every problem entirely by ANN without having a training dataset that provides the exact results.
There are several training methods derived from Back Propagation (BP) such as classical Stochastic Gradient Descent (SGD), Adam, and Adagrad \citep{Robbins2007ASA,kingma2014adam, Duchi2011} which train the network based on a graph of the derivatives of the cost function with respect to each learnable parameter. Besides, in the present problem, the cost function is obtained by calculating the second order partial derivatives of every output with respect to every input via chain rule which implies calculation of the third order derivatives for the Back Propagation (BP) process. Having a large number of training points and intermediate/ANN parameters, SGD does not provide acceptable efficiency and, most of the time, it does not provide an acceptable outcome for 3D problems.
Moreover, it is shown that using the SGD-based approaches, the high number of iterations necessary to reach an acceptable accuracy is a major issue in 1D and 2D problems \citep{Dockhorn2019}. This issue imposes a limit on the choice of the size of training inputs and the network's parameters which determines the accuracy and robustness of the solution if we make use of the mentioned training approaches. Another important issue is that the SGD based algorithms often get trapped in local minima as a consequence of the lack of global search. Moreover, they are sensitive to the initial choice of the network's parameters which are, normally, chosen randomly. On the other hand, in this application, these methods have a better potential of the local search compared to Evolutionary Algorithms (EAs) such as Genetic Algorithm (GA).
EAs are population-based optimisation and search algorithms inspired by natural/biological evolution mechanisms such as survival of the fittest \citep{Yao1999, Vikhar2016}. These methods have been widely employed for global optimisation problems where the connectionist approaches such as SGD are not effective. EAs include several evolution strategies such as Evolutionary Programming (EP) and Genetic Algorithm (GA) \citep{Fogel1998, Yao1999, Holland1992, Goldberg1989}. These methods are more likely to avoid local minima because of the strategies such as \emph{mutation} that are responsible for carrying a global search and maintaining the diversity of the population. Traditionally, they are not gradient-based methods although, due to the type of the presented problem, the gradient plays a role here (to compute the loss). Moreover, EAs are less sensitive to the initial choice of the parameters. However, in the case of ANN's weights and biases training these approaches always search for a globally optimal solution and are usually inefficient in finding fine-tuned settings. In practice, in the case of ANN training, EAs are efficient in finding good initial parameters but a local search algorithm is needed to find the optimal setting.
Hybrid training algorithms have been preferred and successfully applied in several scenarios of interest from civil engineering and material science to pattern, strategy, image, and text recognition \citep{Adeli1994, Skinner1995, taha1995evolutionary, Lee1996,Aguilar}. In these methods, usually, GA is used with the goal to find a set of initial learnable parameters that is in the basin of attraction of the global minimum and a connectionist search algorithm such as SGD is employed to take fine steps to \emph{go down the valley} and minimise the distance function. In other words, GA identifies the deepest \emph{valley} and SGD determines a setting at or close to the minimum. In practice, in several studies, the hybrid algorithms are reported to outperform either GA or BP alone \citep{Belew90evolvingnetworks, TOPCHY1997240, yan1997}. The first part of the hybrid methods (GA) is based on choosing several (usually hundreds) random initial settings of learnable parameters (each one is called an individual which together form a population), compute the error/cost function for each one, affix a score called \emph{fitness} based on their distance from the target, select a certain number of individuals (surviving individuals), and construct a new population that inherits features from the latter. This procedure is repeated until a population with acceptable fitnesses is achieved, which will be explained in details. As an output, an initial setting of learnable parameters is obtained, which has a smaller distance from the target which explains the better efficiency of the algorithm. However, we show that starting from the latter initial point in the space of learnable parameters does not necessarily lead us to a smaller final distance function. Moreover, in the present problem, as we need to compute the second gradients to evaluate an individual, it is a time-consuming procedure to calculate the fitnesses for a population in each iteration. Consequently, there is a need to introduce a new training method for an accurate and efficient problem-solving procedure based on a small network with a small number of training points and iterations which is able to solve problems in 3D.
A novel hybrid method called Modified Genetic Algorithm-Multilevel Stochastic Gradient Descent (MGA-MSGD) is introduced in the present study which is, for the first time, able to provide the continuous solution to the strong form of 3D PDEs in solid mechanics (here, a linear elastic problem) efficiently and accurately. In this framework, the parameters that cause the "error explosion" (sharp unbound increase in cost function) in the training procedure are constantly modified so that a large learning rate can be used which helps to avoid several local minima. In the MGA part, the standard GA is modified in which, for example, the individuals are the binary representation of each learnable parameters and the population consists of all the weights and biases so that a direct manipulation of the parameters is done. We introduce a new way of selection based on the computation of the \emph{importance} of the individuals. In this method, for better efficiency, we only consider the qualified offsprings to produce a new generation so that every generation is necessarily \emph{better} than the previous one. The population qualification procedure is based on a coarse-level SGD training which indicates if we are at a basin of attraction of a global (or near global) minimum. We show that this approach is considerably more reliable than the conventional evaluation of the individuals and population. At the end of MGA-CSGD (Coarse-level SGD), we introduce a fine-scale SGD to step closer to the minimum with considerably smaller steps which completes the procedure of MGA-MSGD method. This approach is less sensitive to the initial parameters, the number and distribution of the training points, and learning rates. We compare the results of MGA-MSGD with the ones of a simple SGD training to demonstrate its effectiveness and we measure its accuracy using the approximate response provided by the Finite Element Method (FEM).
In the following section, the ANN setup including its architecture, activation and cost functions, and the implementation of the physical laws (in the form of PDEs) and Dirichlet and Neumann Boundary Conditions (BCs) etc. is explained. The details of the novel training approach MGA-MSGD are presented in Section \ref{sec_MGAMSGD}. A thorough sensitivity analysis follows this in Section \ref{sec_sensitivity}, which provides a deep understanding of each parameter, its effects on the results, and the advantages of the proposed training method. Finally, in Section \ref{sec_conclusion} the concluding remarks are provided.
\section{ANN and continuum mechanics} \label{sec_ANNsetup}
In this section, we provide details of the ANN for continuous solution of a continuum mechanics problem which is assumed to be a 3D linear elastic solid deformation problem in a displacement-based formulation.
The inputs of the ANN are the spatial coordinates $(x,y,z)$ of every point of interest inside the domain $\Omega \subset \mathbb{R}^3$ and the outputs are their corresponding components of the displacement vector $\vett u =(u_x,u_y,u_z)$. Such a setting is shown in Figure \ref{fig_ANN} and can be written in the form of the following transfer function
\begin{equation}
f{:}( x, y, z) \rightarrow u_x, u_y, u_z
\end{equation}
\begin{figure}
\includegraphics[width=12cm]{ANNsketch1}
\centering
\caption{Schematic representation of ANN setup for solving PDEs in solid mechanics with three spatial dimensions. Note that $\vett u_d$ represents the displacement vectors of the points located on $\Gamma_d$ and $\vett{t}_n$ indicates the Cauchy stress vector of the points on $\Gamma_n$.}
\label{fig_ANN}
\end{figure}
Here, we make use of a network with $N_h$ hidden layers and $N_{nh}$ neurones in each layer. The following activation function
\begin{equation}
ELU(\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,) = \max (0,\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,)+ \min (0, \alpha * (\exp (\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,) - 1))
\end{equation}
with the coefficient $\alpha = 1$ as we need the 2nd order derivative of the outputs (displacements) with respect to the inputs (coordinates) as well as the 3rd order derivative with respect to the learnable parameters for the ANN training procedure.
The latter starts from, in this case, a linear Feed Forward expressed as
\begin{align}
\psi^{(i)}_k&=w^{(i)}_{kj}a^{(i-1)}_j+b^{(i)}_k \label{ANN1} \\
a^{(i)}_k&=ELU(\psi^{(i)}_k) \label{activeF}
\end{align}
where $a^{(i)}_k$ is the value of the $k$ -th neurone in the $(i)$ -th hidden layer with $a^{(0)}_k$ being the coordinates of interest ($\vett a^{(0)} = (x,y,z)$). $ i \in(1,..,N_h)$, $k$ and $j$ indicate the number of the neurones in the $(i)$ -th and $(i-1)$ -th layers, respectively. The ANN's outputs are then calculated by
\begin{equation}
u_k = w^{(N_h+1)}_{kj} a^{(N_h)}_j. \label{ANN3}
\end{equation}
Equations \eqref{ANN1}-\eqref{ANN3} provide a complex ansatz relating spatial coordinates to their corresponding displacements
Having a prediction of the displacements ($u_k$), we need to introduce an appropriate distance function to quantify the deviation of the solution provided by ANN from the exact/unknown displacements satisfying the following continuum mechanics governing equations.
\paragraph{Residuals of the strong-form governing equations in the domain}
The mechanical deformation problem considered here is based on the classical continuum mechanics of linear elasticity. The differential form of the governing equations are presented here as residual expression assuming the presence of approximate solutions to the unknown physical fields: displacement vector $\vett u(\vett x)$, strain tensor $\vett{\varepsilon}(\vett x)$, and stress tensor $\vett \sigma(\vett x)$ in the domain, and displacement vector $\vett{u}(x)$ and stress vector $\vett t(x)$ on the boundary.
The residual of the linearised strain tensor (assuming infinitesimal strain)
\begin{equation}
\vett r_1(\vett x) = \vett{\varepsilon}(\vett x) - \frac{1}{2} \left(\nabla \vett u(\vett x) + \nabla \vett u^T(\vett x)\right) \quad \quad \textrm{in} \quad \quad \Omega
\end{equation}
is a function of the (approximate) displacement state $\vett u(\vett x)$.
The residual of the linear momentum equation, that describes the equilibrium, is
\begin{equation}
\vett r_2(\vett x) = \nabla \cdot \vett \sigma(\vett x) - \vett f(\vett x)\quad \quad \textrm{in} \quad \quad \Omega
\end{equation}
in which $\vett \sigma(\vett x)$ denotes the Cauchy stress and $\vett f(\vett x)$ is the vector of external forces.
Since a linear elastic and isotropic material is considered here, one can write the residual of the Saint Venant-Kirchhoff law as
\begin{equation}
\vett r_3(\vett x) = \vett \sigma(\vett x) - \mathbb C {:} \vett{\varepsilon}(\vett x) \quad \quad \textrm{in} \quad \quad \Omega
\end{equation}
with $ \mathbb C = \mathbb C(E,\nu)$ being the spatially constant 4-th order elasticity tensor depending on the material constants $E$ (elasticity modulus) and $\nu$ (Poisson's ratio).
At the boundary the residual of the Cauchy principle is
\begin{equation}
\vett r_4(\vett x) = \vett t(\vett x) - \vett \sigma(\vett x) \cdot \vett n(\vett x) \quad \quad \textrm{on} \quad \quad \Gamma
\end{equation}
with $\vett u(\vett x)$ as the outward unit normal vector at the boundary of the domain.
The physical boundary conditions of the underlying Boundary Value problem are stated in their residual form. This leads to
\begin{equation}
\vett r_5(\vett x) = \vett u(\vett x) - \vett u_0(x) \quad \quad \textrm{on} \quad \quad \Gamma_d
\end{equation}
on the boundary $\Gamma_d$ where the displacement state is given (Dirichlet Boundary Conditions).
The residual of the Neumann boundary condition
\begin{equation}
\vett r_6(\vett x) = \vett t(\vett x) - \vett t_0(x) \quad \quad \textrm{on} \quad \quad \Gamma_n
\end{equation}
is established on the boundary $\Gamma_n$ with prescribed boundary stress vector state $\vett t_0(\vett x)$.
In the following we decide to fulfil the residuals $\vett r_1$, $\vett r_3$,and $\vett r_4$ exactly, such that
\begin{align}
\vett r_1(\vett x) &= \vett 0 \rightarrow \vett{\varepsilon}(\vett x) = \frac{1}{2} \left(\nabla \vett u(\vett x) + \nabla \vett u^T(\vett x)\right) \quad \quad \textrm{in} \quad \quad \Omega \\
\vett r_3(\vett x) &= \vett 0 \rightarrow \vett \sigma(\vett x) = \mathbb C {:} \vett{\varepsilon}(\vett x)\quad\quad \quad\quad \quad\quad \quad \quad \,\,\,\, \textrm{in} \quad \quad \Omega \\
\vett r_4(\vett x) &= \vett 0 \rightarrow \vett t(\vett x) = \vett \sigma(\vett x) \cdot \vett n(\vett x)\quad \quad \quad\quad \quad \quad \,\,\,\,\, \textrm{on} \quad \quad \Gamma
\end{align}
and resulting relations can be used directly. On the other hand we assume that the residuals $\vett r_2$ (equilibrium), $\vett r_5$ (Dirichlet boundary conditions) and $\vett r_6$ (Neumann boundary conditions) are not a priori zero and only fulfilled in an approximate sense.
It is possible to train the ANN by defining appropriate cost function which enforces above mentioned governing equations considering the BCs on $\Gamma_d$ and $\Gamma_n$ \citep{SIRIGNANO20181339,Raissi2018}. Let us assume that we have $n_d$ points on $\Gamma_d$, $n_n$ points on $\Gamma_n$, and $n$ points in $\Omega$. The cost function representing the loss of the ANN output is considered as the point-wise evaluated weighted residuals
\begin{align}
\mathit{MSE}_e = & \frac{1}{n}\sum_{i}^n \vett w_2(\vett x_i) \cdot \vett r_2(\vett x_i) \,\,\, \quad \quad in \quad \quad \Omega,\label{balance}\\% {\lVert} \nabla\cdot \vett \sigma^i{\rVert}^2 \quad \quad \quad \quad in \quad \quad \Omega,\label{balance}\\
\mathit{MSE}_d = & \frac{1}{n_d}\sum_{i = 1}^{n_d} \vett w_5(\vett x_i) \cdot \vett r_5(\vett x_i) \quad \quad \,on \quad \quad \Gamma_d, \label{EqDBC}\\% {\lVert} \vett u^i - \vett{u}_0^i{\rVert}^2 \quad \quad \,on \quad \quad \Gamma_d, \label{EqDBC}\\
\mathit{MSE}_n = & \frac{1}{n_n}\sum_{i = 1}^{n_n} \vett w_6(\vett x_i) \cdot \vett r_6(\vett x_i) \quad \quad on \quad \quad \Gamma_n,\label{neumann}\\% {\lVert}\vett t^i - \vett t_0^i{\rVert}^2 \quad \quad on \quad \quad \Gamma_n,\label{neumann}\\
\mathit{MSE} = & \mathit{MSE}_d+ \mathit{MSE}_n + \mathit{MSE}_e,\label{eqMSE}
\end{align}
in which the weighting functions are chosen as the respective residual functions
\begin{align}
\vett w_2(\vett x) &= \vett r_2 (\vett x)\\
\vett w_5(\vett x) &= \vett r_5 (\vett x)\\
\vett w_6(\vett x) &= \vett r_6 (\vett x)
\end{align}
and thus leading to a cost function similar to the least-square method, however, they are not based on a space discretisation using a mesh. In other words, one does not need to discretise the whole spatial domain in order to find the solution at a specific point. Basically, the only requirement is to place sufficient data points on $\Gamma_d$ and $\Gamma_n$ which will be studied in details in Section \ref{sec_sensitivity}.
\newpage
Having calculated the cost function, we can choose an optimisation algorithm to minimise it and train the ANN. Here, we develop a method called "hybrid MGA-MSGD" to efficiently obtain a reliable trained ANN which provides us with the continuous solution (displacements and their gradients) of the problem. The abbreviation MGA indicates Modified Genetic Algorithm and MSGD stands for Multi-level Stochastic Gradient Descent.
\section{Hybrid MGA-MSGD training algorithm} \label{sec_MGAMSGD}
A Genetic Algorithm (GA) is a stochastic population-based optimisation approach which is considered as a type of Evolutionary Algorithm (EA) \citep{Goldberg1989, Holland1992}. The advantage of this method is that it performs a global search with the expectation to find a basin of attraction of the global minimum of the cost function. However, it is observed that it is not effective as a local search engine which is the reason for combining it with SGD-based algorithms such as BP \citep{Belew90evolvingnetworks, Yao1999}.
In this section, we introduce a Modified GA approach enhanced by the Multi-level SGD suitable for both global and local search for better identification of learnable parameters of ANNs.
First, we introduce each element of GA and compare it with the ones of MGA in the following.
\paragraph{Population}
Both algorithms start with constructing a \emph{population}. In GA, a population consists of individuals that each represents a potential solution to a problem while in MGA, the population is a set of individuals that together provide one potential solution. In the presented framework, the population is a list of the binary representation of the network's learnable parameters (weights and biases) so that, for fixed network architecture, they characterise the output of the network. The reason for such arrangement is the high dimension of the learnable parameters and time-consuming procedure of distance function calculation. It also allows us to modify the most important parameters that cause error explosion so that we save computational effort. Furthermore, the binary representation is used so that the strength of mutation, which will be defined, can be controlled and principles of GA can be exploited.
\paragraph{Chromosomes and genes}
The basic definitions of these elements are the same in GA and MGA. Each individual of a population (learnable parameters of ANN) is called a \emph{chromosome} which consists of several \emph{genes}. The genes can adopt letters, numbers, vectors etc. Here, a chromosome is the binary representation of one learnable parameter, and a gene can adopt one integer of this representation (which is 0 or 1). In other words, a population is a high dimensional array of the binary representation of learnable parameters such as weights and biases. We also consider the sign of the parameter as a gene so that it can be flipped with a given chance maintaining the diversity in the learnable parameters' space.
\paragraph{Fitness and Importance}
In GA, \emph{fitness} is a value assigned to each chromosome which is the relative success of the corresponding individual on the problem. For example, if we choose this value between 0 and 100, the individual that provides the correct value has the fitness 100, and the one that has no similarity with the correct one adopts 0. Due to the fact that in MGA the individuals do not provide a solution to the problem (the entire population can only provide it) they can not be directly examined and assigned a fitness value. However, using gradient-based decomposition approaches such as Back Propagation we can quantify the relative sensitivity of the cost function with respect to the learnable parameters (i.e. $\frac{\partial C}{\partial \,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,}$, where $\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,$ is one learnable parameter and $C$ is the cost function) showing how "important" is each individual. We highlight that the most important individuals are the ones that cause error explosion enforcing the choice of small learning rates. The latter can be seen as one reason of getting trapped into local minima which are avoidable using larger learning rates, e.g. the ones in Zone 1 of Figure \ref{fig_LocalM}. This non-trivial dependency of the results on the learning rate is proven via sensitivity analysis and shown in Figure \ref{fig_lrF}.
Consequently, we replace the term "fitness" in GA with the term "importance" in MGA.
\begin{figure}
\centering
\includegraphics[width = 12cm]{LocalM}
\caption{A schematic sketch of the error ($e$) associated with every location in learnable parameter's space $[w,b]$(e.g. weights and biases). Note that the dimension of $[w,b]$ is way higher than what is shown here. Getting trapped in local minima shown in Zone 1 is avoidable choosing a relatively large learning rate that allows larger steps in the horizontal direction. Using SGD, the local minimum in Zone 2 will be the final response even with a large learning rate if we start from its basin of attraction (e.g. Point A). The latter can be tackled by means of MGA which is able to carry out a global path-independent search and step from Point A to Point B. We highlight that although Point A as an initial location has less error, the Point B is favourable as it is in the neighbourhood of "global" minimum.}
\label{fig_LocalM}
\end{figure}
\begin{figure}
\centering
\includegraphics[width = 8cm]{lrF}
\caption{The profile of learning rates vs average $\mathit{MSE}$ shows that having smaller learning rates does not necessarily result in higher accuracy. The average $\mathit{MSE}$ (or other variables in the plots) is its averaged value of several analyses under equivalent conditions so that decreasing the effect of randomness and indicating the role of the parameter under study.
This plot is produced experimentally by performing training tests with different learning rates. Every point is the average of the value of interest ($\mathit{MSE}$ and time) of 100 tests.}
\label{fig_lrF}
\end{figure}
\paragraph{Selection and offsprings}
In GA, some chromosomes (according to their fitness scores) are chosen in the process of \emph{selection}. Usually, the chromosomes with higher fitness score have a higher chance of being selected. These chromosomes are called "parents" and are chosen to produce offsprings from which the next generation is constructed which is to be more similar to the selected chromosomes while maintaining the diversity.
In MGA, the chromosomes to be modified are selected based on their importance via Tournament Selection procedure (\cite{Miller1995GeneticAT}). The selected parameters for minimisation of the cost function (the most important chromosomes), as explained before, are the ones that cause problems such as error explosion enforcing the choice of very small learning rate which, in turn, increases the chance of getting trapped into a local minimum. The fact that GA tends to modify the parameters with high sensitivity (importance) helps us to avoid this problem. These individuals, subsequently, go through some operations out of which we hope to acquire a better arrangement of the population. We highlight that, here, the unselected individuals are the surviving ones.
Using the activation function $ELU$ in Equation \eqref{activeF} and expanding the BP gradient formula $\frac{\partial C}{\partial \,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,}$ we understand that the sensitivity of a parameter does not only depend on itself. In fact, it depends on all the parameters that are directly or indirectly in contact with that parameter. Consequently, in order to maintain the diversity of selected individuals (i.e. avoiding the similar choices in several MGA iterations), we introduce the condition that MGA can not select the same parameters more than a specific times (here, twice). This restriction is released once all of the parameters in the population are previously selected (in order to avoid empty selection) by clearing the history of the selected individuals.
\paragraph{Operations}
There are two well-known operations of GA called \emph{crossover} and \emph{mutation}. The idea of the former is to produce some offsprings by combining the genes of two parents in order to generate new individuals that provide new solutions. The produced offsprings are partially similar to their parents. However, when we use MGA for ANN training, there is no need for the similarity between the selected individuals (learnable parameters) as the inheritance from the previous population is obtained via the unselected individuals tuned only by the connectionist optimiser gradient-based CSGD. So, as expected and from our observations as well as the literature \citep{Yao1999, Heimes1997, Sarkar1997, BORNHOLDT1992327}, the crossover does not perform well in ANN training. On the other hand, the second operation (mutation) is very useful in the global/local search for the optimal point as it maintains the genetic diversity by flipping the value of one or more genes \citep{Goldberg1989}. Here, we also introduce the sign of the chromosome as a gene that can be flipped by a given chance. The sign and the first following genes are responsible for the global search while the last genes are responsible for small changes performing a local search. In other words, if we flip the first/last genes of the chromosome, a large/small change is imposed thus a global/local modification on the selected parameters takes place. We introduce three factors controlling global/large scale, medium scale, or local/small scale mutation probability. The first one (global mutation) allows us to, for example, step from Point A to Point B in Figure \ref{fig_LocalM} while the others are responsible for smaller steps.
\paragraph{Offspring, population qualification, and new generation}
The obtained chromosomes from mutation are, then, converted back to the float representation constructing the \emph{offspring}. The latter is the link to produce a new population which we hope to fall into a better point in the parameters' space. In fact, we replace the selected parameters by their corresponding offspring with the expectation that this new arrangement is "better" than the previous one. In this case, a "better" population is the one that leads us nearer to the global minimum (accuracy) at fewer increments and time (efficiency). The new population serves as the initial parameters for a SGD optimisation process. At this point, the important question is that if a population that creates a smaller initial cost function $\mathit{MSE}_i$ is a better one or it is possible that we achieve, after performing SGD, a more accurate final result (shown by $\mathit{MSE}_{min}$) from a larger $\mathit{MSE}_i$? Figure \ref{msecomp} shows that having smaller $\mathit{MSE}_i$ does not necessarily result in a smaller $\mathit{MSE}_{min}$, which is obtained via performing a SGD optimisation with a fine learning rate to find the nearby minimum.
\begin{figure}
\centering
\includegraphics[width = 8cm]{msei-msemin}
\caption{The plot of $\mathit{MSE}_{min}$ vs $\mathit{MSE}_i$ shows that although $\mathit{MSE}_i$ can be an important parameter, it is not sufficient as the only measure to identify a "better" population.}
\label{msecomp}
\end{figure}
At this stage, we can introduce the \emph{Coarse-level} SGD (CSGD or a SGD with a large learning rate) so that we can both examine the new population and get near the objective accuracy. In general, if we have a good arrangement of the learnable parameters such that the output/cost function does not largely depend on some specific small group of parameters with very high gradients ($\frac{\partial C}{\partial \,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,}$) it is possible to reach an accurate result with large learning rate and small number of iterations avoiding local minima in a short time. In fact, achieving such population arrangement, we use the full potential of our ANN in which all the neurones play a considerable role in the final outcome. Consequently, we perform a CSGD optimisation based on the new population. Then, we can base the population qualification on $\mathit{MSE}_c$ as follows
\begin{numcases}
{if} \mathit{MSE}_c^{(j)}<\mathit{MSE}_c^{(j-1)}: \quad\textrm{qualified population} \\
\mathit{MSE}_c^{(j)}>\mathit{MSE}_c^{(j-1)}: \quad \textrm{unqualified population},
\end{numcases}
where $j$ is the current MGA iteration number. If the population gets qualified we call the obtained arrangement (the one after CSGD) a new \emph{generation} and base the following MGA iterations on it otherwise we neglect it and we use the previous population/generation. Due to the complexity of the BP (using the third order partial derivatives) we might need to impose a break condition when the cost/distance divergence starts in order to avoid stepping away from the minimum \citep{Duckett2008}. The latter condition, as well as the convergence one, is shown in Figure \ref{fig_div_conv}.
\begin{figure}
\centering
\begin{subfigure}{6.6cm}
\includegraphics[width = 6.7cm]{convergence}
\caption{Two conditions under which the cost function converges to a minimum with sufficiently small learning rates.}
\label{fig_conv}
\end{subfigure}\quad
\begin{subfigure}{4.8cm}
\includegraphics[width = 5.1cm]{Divergence}
\caption{The divergence of the cost function from the minimum due to a large learning rate (large horizontal step).}
\label{fig_div}
\end{subfigure}
\caption{A schematic representation of convergence and divergence of the cost function. Although a large learning rate helps to skip several local minima, it causes divergence at a distance from the target minimum. To tackle this condition, we introduce a fine level SGD after achieving the final MGA-CSGD result.}
\label{fig_div_conv}
\end{figure}
Finally, at the end of the MGA analysis with CSGD based qualification, we perform a \emph{Fine-level} SGD (FSGD or a SGD with small learning rate) in order to achieve a more accurate result with smaller $\mathit{MSE}$. The reason is shown in Figure \ref{fig_div_conv}. This completes the hybrid MGA-MSGD training procedure.
\section{Sensitivity analysis and numerical results} \label{sec_sensitivity}
The above-described framework has several non-mechanical/framework parameters that may considerably affect the efficiency of the procedure and accuracy of the results. Given the parametric uncertainty, it seems necessary to employ a Sensitivity Analysis method (SA) to identify important model parameters and to calibrate them. Furthermore, due to the stochastic nature of the ANN training procedure, every piece of data provided for the sensitivity analysis (e.g. mean $\mathit{MSE}$, mean time, standard deviation, variance, etc.) should be obtained by averaging the results of several analyses which brings high computational expense and should be considered as a constraint when adopting a SA method. Therefore, the Morris method \citep{Morris1991}, which is a simple and efficient approach based on changing one-factor-at-a-time (OAT) is employed.
\paragraph{\textbf{Statement of the test problem}}
The deformation behaviour a unit cube of homogeneous isotropic linear elastic material with zero displacements in $\vett x$ direction ($u_{x0} = 0$) on $\Gamma_d$ located at $x=0$ and a nonzero uniformly distributed $\vett t_0 = (-0.1, 0, 0)$ (see Figure \ref{fig_cube}). The uniqueness conditions
\begin{align}
\frac{1}{n} \sum_{i=1}^n u_y = 0 \label{uniquex} \\
\frac{1}{n} \sum_{i=1}^n u_z = 0 \label{uniquey}
\end{align}
are also imposed.
\begin{figure}
\includegraphics[width=7cm]{cube}
\centering
\caption{A representative sketch of the model. Zero displacements in $\vett x$ direction ($u_{x0}=0$) as the Dirichlet BC is applied on $\Gamma_d$ and $\vett t = (-0.1, 0, 0)$ Neumann BC is applied on $\Gamma_n$.}
\label{fig_cube}
\end{figure}
The density and distribution of the training dataset can be controlled by the number of the sampling points in $\vett x$, $\vett y$, and $\vett z$ directions, which are shown by $N_{\vett x}$, $N_{\vett y}$, and $N_{\vett z}$, and the number of points located on the boundaries with Dirichlet/Neumann constraint $N_b$ to the one of the whole volume (the total number) $N_V$. We also introduce the variable $\beta_i$ that modifies the value of $\frac{N_b}{N_V}$
\begin{equation}
\frac{N_b}{N_V} = \frac{N_{bu} + \beta_i N_{bu}}{N_{Vu} + \beta_i N_{bu} }
\end{equation}
Where $N_{Vu}$ and $N_{bu}$ are the total number of the points and the number of the ones exactly on the boundaries at the uniform distribution of sampling points, repectively. Note that, as this method is free of space discretisation (mesh-free), one can randomly choose the sampling points up to the condition that there must be some points on the surfaces $\Gamma_d$ and $\Gamma_n$ so that we can enforce Dirichlet/Neumann BCs in the training procedure. However, we choose to have an entirely controlled point distribution in order to understand how it affects the outcome of the analysis.
In Morris method there are two parameters showing the sensitivity of the output to a variable which are the standard deviation $\sigma$ and mean elementary effect $\mu$ computed by
\begin{align}
\sigma = \sqrt{\frac{\sum_{i=1}^{n}(x_i - \bar x)^2}{n-1}}\\
\mu = \frac{ \sum_{i=1}^{n}\lvert(x_i - \bar x) \rvert}{n-1}
\end{align}
Generally, the more the value of $\sigma$ and $\mu$ the more important the corresponding parameter.
We study the effects of the non-mechanical/framework parameters shown and explained in Table \ref{MorrisSensitivity}, where surviving population fraction is the fraction of unselected population which are the ones that remain unchanged in the next population. The results are provided in Table \ref{MorrisSensitivity1}.
\begin{table}
\caption{}
\centering
\scalebox{0.85}{
\begin{tabular}{|c| c| c| c|}
\hline
Parameter & Notation & Distribution & Base value\\
\hline
Coarse scale learning rate & $lr_c$ & $U(0.5, 1)$ & $0.7$\\
\hline
Number of MGA iterations & $N_{GAi}$ & $U(10,60)$ & $30$\\
\hline
Number of ANN hidden layers & $N_h$ & $U(2,6)$ & $3$ \\
\hline
Number of neurones in each hidden layer & $N_{nh}$ & $U(3, 20)$ & $10$\\
\hline
Surviving population fraction & $P_{sf}$ & $U(0.9, \approx 1)$ & $0.98$\\
\hline
Number of sampling points in $\vett x$ direction & $N_{x}$ & $U(5, 14)$ & $10$ \\
\hline
$ N_{\vett x} - N_{\vett y} = N_{\vett x} - N_{\vett z}$ & $N_{x} - N_y$ & $U(0, 8)$ & $0$ \\
\hline
Sampling point distribution parameter & $\beta_i$ & $U(0, 2)$ & $0$\\
\hline
Global mutation probability & $M_g$ & $U(0.1, 0.5)$ & $0.1$ \\
\hline
Median mutation probability & $M_m$ & $U(0.1, 0.5)$ & $0.1$ \\
\hline
Local mutation probability & $M_l$ & $U(0.1, 0.5)$ & $0.1$ \\
\hline
\end{tabular}}
\label{MorrisSensitivity}
\end{table}
\begin{table}
\caption{}
\centering
\scalebox{0.85}{
\begin{tabular}{|c| c| c| c| c| c| c| c| c|}
\hline
\multirow{2}{*}{Notation} & \multicolumn{2}{|c|}{Average $\mathit{MSE}$} & \multicolumn{2}{|c|}{Minimum $\mathit{MSE}$} & \multicolumn{2}{|c|}{Average time} & \multicolumn{2}{|c|}{Minimum time} \\ \cline{2-9}
& $\mu$ & $\sigma$ & $\mu$ & $\sigma$ & $\mu$ & $\sigma$ & $\mu$ & $\sigma$ \\
\hline
$lr_c$ & $2.75e{-06}$ & $9.14e{-06} $ & $8.35e{-07}$ & $2.64e{-06}$ & $9.84$ & $31.1$ & $1.73$ & $5.47$\\
\hline
$N_{GAi}$ & $2.21e{-06}$ & $7e{-06}$ & $1.07e{-06}$ & $3.38e{-06}$ & $13.6$ & $43.16$ & $5.9$ & $18.66$ \\
\hline
$N_h$ & $2.36e{-05}$ & $4.71e{-05}$ & $9.14e{-06}$ & $1.82e{-05}$ & $10.6$ & $21.34$ & $4.1$ & $8.17$ \\
\hline
$N_{nh}$ & $7.17e{-06}$ & $1.9e{-05} $ & $1.3e{-06}$ & $3.4e{-06} $ & $8.67$ & $22.9$ & $12.58$ & $33.3$ \\
\hline
$P_{sf}$ & $1.92e{-06}$ & $6.4e{-06}$ & $8.8e{-07}$ & $2.86e{-06}$ & $3.36$ & $11.13 $ & $2.22$ & $7.37$\\
\hline
$N_{x}$ & $1.34e{-06}$ & $4.1e{-06}$ & $1.34e{-06}$ & $4e{-06}$ & $30.9$ & $92.8$ & $6.08$ & $18.2$\\
\hline
$N_{x} - N_y$ & $8.18e{-07}$ & $2.31e{-06} $ & $1.24e{-06}$ & $3.5e{-06} $ & $20.29$ & $57.4$ & $4.59$ & $13$\\
\hline
$\beta_i$ & $1.1e{-06}$ & $3.32e{-06}$ & $1.18e{-06}$ & $3.54e{-06}$ & $21.7$ & $65.1$ & $4.19$ & $12.56$\\
\hline
$M_g$ & $8.8e{-07}$ & $1.76e{-06}$ & $7.48e{-07}$ & $1.5e{-06}$ & $11.5$ & $23$ & $19.6$ & $39.3$\\
\hline
$M_m$ & $1.51e{-06}$ & $3.03e{-06}$ & $7.07e{-07}$ & $1.41e{-06}$ & $59.4$ & $118.8$ & $52.3$ & $104.7$\\
\hline
$M_l$ & $5.53e{-07}$ & $1.1e{-06}$ & $5.64e{-07}$ & $1.13e{-06}$ & $39.5$ & $79$ & $20.1$ & $40.3$\\
\hline
\end{tabular}}
\label{MorrisSensitivity1}
\end{table}
We plot the normalised values of average $\mathit{MSE}$ and training time with respect to their mean value (which is provided in the line labels) in Figures \ref{propss}-\ref{props} in order to tune the algorithm's parameters. In general, the training time depends on several factors including the dimension of population, number of generations, number of sampling points etc. Although the plots in the provided figures are somehow noisy, one can understand the overall trends. The goal is to decrease $\mathit{MSE}$ at the lowest possible cost/increase in time. For example, according to Figure \ref{NGAiAVG}, although by decreasing the coarse level learning rate ($lr_c$) the average $\mathit{MSE}$ decreases, the training time increases sharply, so we choose a moderate value such as $lr_c =0.6$. The inverse of the latter profile, somehow, holds for the number of MGA-CSGD iterations shown in Figure \ref{NGAiAVG}. We again choose a median value for this parameter $N_{GAi} = 30$. The most important parameters that also have the greatest values of Morris factors for average $\mathit{MSE}$ are the number of hidden layers $N_h$ and neurones in each layer $N_{nh}$. By increasing the former, as shown in Figure \ref{NhAVG}, the accuracy of the results decreases showing that the provided training method performs the best with a small number of hidden layers. It is noteworthy that the decrease in average time at increasing $N_h$ is due to the fact that we have fewer generations (qualified populations) when we have more hidden layers.
Usually, the loss divergence in unqualified populations happens at a low number of CSGD iterations (sometimes at the first try) indicating the advantage that we mostly spend time on suitable arrangements of the learnable parameters which are most probable to result in high accuracy.
Figure \ref{NnhAVG} shows that, at least, seven neurones are required in each hidden layer for an acceptable accuracy ($N_{nh}=10$ is chosen). In fact, this shows that using hybrid MGA-MSGD training method we exploit a high potential of the assigned hidden layers and neurones. Figure \ref{PsfAVG} shows a noisy profile and indicates that the surviving population fraction $P_{sf}=0.97$ is an appropriate choice. The profile of $N_x$, $N_x - N_y$, and $\beta_i$ that are, respectively, shown in Figures \ref{NxAVG}, \ref{Nx-NyAVG}, and \ref{betaAVG} demonstrate the low sensitivity of average $\mathit{MSE}$ to the number and distribution of sampling points. This is a crucial advantage of the provided method when ANN is applied to solving mechanical problems because, as expected, the average training time grows exponentially at growing number of sampling points. Note that, as in Figure \ref{Nx-NyAVG} the number of $N_x$ is fixed at its base value, by changing the value of $N_x-N_y$ both distribution and number of sampling points are changed.
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{LRcAVGn}
\centering
\caption{Coarse level learning rate ($lr_c$) has a significantly larger impact on the efficiency compared with the accuracy.}
\label{LRcAVG}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{NGAiAVGn}
\centering
\caption{The number of iterations of MGA-CSGD ($N_{GAi}$) also has a significantly larger impact on the efficiency compared with the accuracy.}
\label{NGAiAVG}
\end{subfigure}
\caption{The dependency of the output accuracy and training time on $lr_c$ and $N_{GAi}$ for parameter tuning purpose. The chosen values in this study are $lr_c=0.6$ and $N_{GAi}=30$. Each one of the "average $\mathit{MSE}$" and "average time" values is divided by a specific value shown in the labels which are their mean values (the mean "average time" and "average $\mathit{MSE}$").}
\label{propss}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{NhAVGn}
\centering
\caption{The highest dependency of the result's accuracy is on the number of hidden layers indicating that a network with two hidden layers produces the lowest $\mathit{MSE}$ at an affordable time which agrees with the observations in \citep{Dockhorn2019}. Note that deeper networks can also produce better results, but they need more MGA-CSGD iterations etc.}
\vspace{0.6cm}
\label{NhAVG}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{NnhAVGn}
\centering
\caption{The profile of the number of neurones in each hidden layer ($N_{nh}$) vs normalised $\mathit{MSE}$ and time indicates that before a specific point the accuracy of the results depends on the number of the neurones to a great extent which decreases sharply after it. In fact, this point can be interpreted as the minimum required number of neurones that are able to interpolate the specific problem. For this problem, we choose $N_{nh}=10$.}
\label{NnhAVG}
\end{subfigure}
\caption{The number of hidden layers and neurones in each are of the most critical parameters to be tuned. Obtaining an acceptable accuracy at such a small number of neurones and layers demonstrates the efficiency of the presented method.}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{PsfAVGn}
\centering
\caption{Surviving population fraction $P_{sf}$ is considerably larger than the ones in typical applications of GA as the target here are only those demanding small learning rates. As it is shown here, average $\mathit{MSE}$ decreases and average time increases at increasing $P_{sf}$ up to 0.99.}
\vspace{1.13cm}
\label{PsfAVG}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{NxAVGn}
\centering
\caption{$N_x$ represents the density of the training/sampling points. It has a profile similar to the one of the number of neurones in each hidden layer in terms of average $\mathit{MSE}$ meaning that it should be more than a specific number from where on the accuracy is not highly sensitive on it. This can be seen as one significant achievement of the proposed approach as increasing it further the required time increases exponentially due to the gradient calculation.}
\label{NxAVG}
\end{subfigure}
\caption{The profile of the surviving population fraction and the density of training points vs training time and obtained $\mathit{MSE}$ as a measure of accuracy. Each one of the "average $\mathit{MSE}$" and "average time" values is divided by a specific value shown in the labels which are their mean values (the mean "average time" and "average $\mathit{MSE}$").}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Nx-NyAVGn}
\centering
\caption{The profile of $N_x-N_y$ represents the training/sampling points' distribution in different directions showing a relatively low dependency of the obtained result on it. Note that $N_x=10$ is fixed at its base value so that the decreasing time at increasing $N_x-N_y$ is due to the lower number of training points.}
\label{Nx-NyAVG}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{betaAVGn}
\centering
\caption{$\beta_i$ determines the distribution of training points in a way that increasing it the number of the points exactly on the boundaries increase.}
\vspace{1.15cm}
\label{betaAVG}
\end{subfigure
\caption{The small sensitivity of the results on the two factors ($N_x-N_y$ and $\beta_i$) controlling the distribution of the training points results in the choice of uniform distribution.}
\label{Nx-Ny_beta}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Mg}
\centering
\caption{Global mutation probability around $M_g=0.3$ provides acceptable results. This factor is responsible for large scale steps in the learnable parameters' space.}
\label{MgAVG}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Mm}
\centering
\caption{$M_m$ controls the median steps in ANN training. In this case, a probability of 0.3 is chosen.}
\vspace{0.6cm}
\label{MmAVG}
\end{subfigure
\label{MgMm}
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Ml}
\centering
\caption{A small path-independent step together with SGD (path-dependent) guarantees that we take the full advantage of the local search. $M_l=0.3$ is shown to be an acceptable value.}
\label{MlAVG}
\end{subfigure}
\caption{The probability of mutation for large, medium, and small steps has a considerable effect on the training result.}
\label{props}
\end{figure}
For more detailed analysis of sampling point distribution, we consider two different distribution scenarios Case 1: $(N_x, N_y, N_z) = (30,2,2)$ and Case 2: $(N_x, N_y, N_z) = (5,5,5)$ with their results ( average $\mathit{MSE}$, minimum $\mathit{MSE}$, average time, and minimum time) provided in Table \ref{tab:T2}. The results show that for a smooth mechanical problem, the choice of uniform distribution of data points provides the best $\mathit{MSE}$ and training time. On the other hand, the small difference between the results of Case 1 and Case 2 together with Figure \ref{Nx-Ny_beta}, reveals that being mesh-free and directly differentiable, this approach (solving mechanics problem via ANN) is able to provide an approximate response of specific coordinates independent of additional points inside the body provided that the BCs are appropriately enforced. For example, Case 1 has points only on four sides of the cube, not inside the volume. This feature together with the flexibility and straight forward implementation and providing continuous spatial gradients can be seen as advantages of approximating mechanical problems via ANNs.
\begin{table}
\caption{}
\centering
\scalebox{0.8}{
\begin{tabular}{|c| c| c| c| c|}
\hline
sampling point distribution & Average $\mathit{MSE}$ & Minimum $\mathit{MSE}$ & Average time & Minimum time \\
\hline
$(N_x, N_y, N_z) = (30,2,2)$ & $3.35e{-05}$ & $2.2e{-05}$ & $47$ & $24.1$\\
\hline
$(N_x, N_y, N_z) = (5,5,5)$ & $3.18e{-05}$ & $2.15e-{05}$ & $38.2$ & $18.9$\\
\hline
\end{tabular}}
\label{tab:T2}
\end{table}
In summary, the chosen setting for this problem is $(N_x, N_y, N_z) = (5,5,5)$, $lr_c=0.6$, $N_{GAi}=30$, $N_h=2$, $N_{nh} = 10$, $P_{sf} = 0.97$, $\beta_i=0$, and $lr_f = 1e{-5}$. To authors' experience, the parameters $lr_c$, $(N_x, N_y, N_z)$, and $N_{nh}$ are the ones that should be available at a higher programming level to the users for tuning purposes.
In the following subsection the results of the presented methodology are compared with the standard training approaches.
\subsection{Training performance}
The most important points to be considered for computational methods are efficiency and accuracy. Here, we choose two standard well celebrated training algorithms, namely classical SGD (\cite{Robbins2007ASA}) and Adam (\cite{kingma2014adam}), to assess the relative success of the presented framework. Here, for the sake of clarification and reader's convenience, the update strategy for SGD and Adam is provided. The former has simple update terms as below
\begin{equation}
\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}:=\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}- n \frac{\partial \mathit{MSE}}{\partial \,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}}
\end{equation}
where $n$ is a coefficient that determine the learning rate (size of steps) and $\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,$ is the updated parameter.
In the case of Adam optimiser the formulation is more complex where the parameters are updated via
\begin{equation}
\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}:=\,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}- n \frac{\bar{m}}{\sqrt{\bar{v}}+ \bar{\epsilon}} \label{laststep}
\end{equation}
where $\bar{\epsilon}$ is a term to improve the numerical stability. $\bar{m}$, $\bar{v}$ are defined and updated in each iteration via
\begin{equation}
\bar{m} := \frac{m}{1-\beta_1^t} \quad \quad \quad \quad \bar{v} := \frac{v}{1-\beta_2^t}
\end{equation}
where, $t$ is the number of iteration and $\beta_1$ and $\beta_2$ are exponential decay rates for the moment estimates. The terms $m$ and $v$ are defined and updated in each iteration by
\begin{align}
m&:=\beta_1 m +(1-\beta_1) \frac{\partial \mathit{MSE}}{\partial \,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}} \\
v &:=\beta_2 v +(1-\beta_2) \left(\frac{\partial \mathit{MSE}}{\partial \,\begin{picture}(-1,1)(-1,-3)\circle*{3}\end{picture}\;\,^{(i)}_{kj}}\right)^2
\end{align}
We choose the initial values of $m$ and $v$ equal to zero.
Figure \ref{fig_Performance1} shows that MGA-MSGD converges to a minimum closer to the global minimum in remarkably smaller times. To our experience, the key to the higher efficiency is the use of a large learning rate which has become feasible by the help of the provided Modified Genetic Algorithm (MGA).
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{PerformanceN}
\centering
\caption{The large scale picture of overall performance shows the efficiency of MGA-MSGD training algorithm and indicate that the major competitor is Adam optimiser}
\label{fig_Performance}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Performance_zoom}
\centering
\caption{The limited plot frame shows that even at very long times such as 1000[s] and 2000[s] MGA-MSGD outperforms the standard training approaches.}
\label{fig_Performance_zoom}
\end{subfigure
\caption{A performance comparison between the presented framework and two standard optimisers showing considerable improvement in both accuracy and efficiency.}
\label{fig_Performance1}
\end{figure}
\subsection{Analysis of local displacement/output error}
Apart from the statistical measures, it is important to see if the results make sense from a mechanical point of view as well. Figure \ref{fig_wrong} is the contour plot of the displacements in $\vett y$ direction of an analysis with acceptable $\mathit{MSE}$ loss, which, from a statistical viewpoint, is acceptable. However, from a mechanical viewpoint, to obtain the accurate response a rigid body motion is required which is also observed in similar works in the literature \cite{Raissi2018}.
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UyWrong}
\centering
\caption{}
\label{fig_UyWrong}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UzWrong}
\centering
\caption{}
\label{fig_UzWrong}
\end{subfigure
\caption{The contour plot of the resultant displacement components ($u_y$ and $u_x$) without enforcing Equations \eqref{normalisation} and \eqref{gamma}. We observe that the Balance of linear momentum in Equation \eqref{balance} is satisfied but the Dirichlet BC and the uniqueness condition in Equations \eqref{EqDBC}, \eqref{uniquex}, and \eqref{uniquey} are not imposed correctly. In other words, the exact results (provided in Figures \ref{fig_FEMUy} and \ref{fig_FEMUz}) can be obtained by a rigid body motion.}
\label{fig_wrong}
\end{figure}
In fact, although the balance of linear momentum seems to be enforced appropriately, the Dirichlet BC and uniqueness conditions are not represented correctly. The reason roots in the definition of $\mathit{MSE}$ (which is not a normalised/dimensionless value) as the values of stress components and their divergence can be considerably higher than the displacements, so the ANN training procedure only tends to satisfy the balance of linear momentum to a considerably higher extent than Dirichlet BC and uniqueness conditions. This problem can be solved by normalising the stress $\vett{\sigma}$ as below
\begin{equation}
\hat{\vett{\sigma}} = \frac{\vett{\sigma}}{\lambda+2G}\label{normalisation}
\end{equation}
where, $\lambda$ and $G$ are the material parameters (Lam\'e constants).
Moreover, in order to ensure a strong/exact enforcement of Dirichlet BC we introduce the weight $\gamma$ as follows
\begin{equation}
\vett w_5(\vett x) = \vett r_5 (\vett x)*\gamma \quad \quad \,on \quad \quad \Gamma_d, \label{gamma}
\end{equation}
where $\gamma=0.05 n=6.25$ is chosen experimentally based on Figure \ref{fig_GMDB}. The use of the total number of points in the volume ($n$) in obtaining the parameter $\gamma$ is to ensure that the Dirichlet BCs are considered while it does not undermine other terms in the cost function $\mathit{MSE}$.
The mentioned conditions result in a more accurate mechanical response which is shown (of the same displacement components) in Figures \ref{fig_UYcomp} and \ref{fig_UZcomp}.
At this stage, we solve the problem with the Finite Element Method (FEM) in order to obtain a reference solution that allows to study the local displacement error. This provides us with a criterion to study the resultant mechanical response. In fact, having the reference mechanical response of the problem ($\vett u$ in $\Omega$) we are able to quantify the distance function directly by
\begin{equation}
\mathit{MSE}_u = \frac{1}{n}\sum_{i = 1}^{n} \vett w_7(\vett x_i) \cdot \vett r_7 (\vett x_i) \quad \quad \,in \quad \quad \Omega \label{eqMSEu
\end{equation}
where
\begin{align}
\vett r_7(\vett x) &= \vett u(\vett x) - \vett u_r (\vett x) \quad \quad \,in \quad \quad \Omega\\
\vett w_7(\vett x) &= \vett r_7(\vett x) \quad \quad\quad \,\,\quad \quad \,in \quad \quad \Omega
\end{align}
and where ${\vett{u}}_{r}(\vett x)$ is the reference results, to ensure that our results are accurate. Note that the latter is only used to test the results with no role in the network training procedure.
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{gamma_MSEDisp_DBC}
\centering
\caption{An increase in $\gamma$ results in considerable decrease in $\mathit{MSE}_d$.}
\vspace{0.3cm}
\label{fig_gamma_MSEDisp_DBC}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{gamma_MSE}
\centering
\caption{The increase in $\mathit{MSE}$ is due to the multiplication of the distance function with the coefficient $\gamma$.}
\label{fig_gamma_MSE}
\end{subfigure
\caption{The parameter $\gamma$, which is a weight for strong enforcement of Boundary Conditions has a significant effect on the accuracy from a mechanical viewpoint, although it might cause an increase in the general $\mathit{MSE}$.}
\label{fig_GMDB}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Ux}
\centering
\caption{$u_x$ from ANN.}
\label{fig_Ux}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{FEMUx}
\centering
\caption{Reference $u_x$.}
\label{fig_FEMUx}
\end{subfigure
\caption{The mechanical response provided by ANN trained via MGA-MSGD compared to the reference one provided by FEM.}
\label{fig_UXcomp}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Uy}
\centering
\caption{$u_y$ from ANN}
\label{fig_Uy}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{FEMUy}
\centering
\caption{Reference $u_y$.}
\label{fig_FEMUy}
\end{subfigure
\caption{The ANN results are in a good agreement with the reference ones showing the obtained accuracy.}
\label{fig_UYcomp}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{Uz}
\centering
\caption{$u_z$ from ANN}
\label{fig_Uz}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{FEMUz}
\centering
\caption{Reference $u_z$.}
\label{fig_FEMUz}
\end{subfigure
\caption{Displacements in $\vett z$ direction from ANN and FEM as a mean of comparison.}
\label{fig_UZcomp}
\end{figure}
A comparison between the solution provided by ANN and the reference displacements is shown in Figures \ref{fig_UXcomp}, \ref{fig_UYcomp}, and \ref{fig_UZcomp} shows that a good approximation of a 3D mechanical problem can be calculated by an ANN using the provided training strategy with an acceptable efficiency which was not possible via the conventional training algorithms.
In order to further evaluate the potential of the presented framework we solve the same boundary value problem with $\vett u_0 = (0,0,0)$ on $\Gamma_d$ which results in slightly more complex deformation. In this case, the displacements in $\vett y$ and $\vett z$ directions decrease approaching $\Gamma_d$.
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UxFixed}
\centering
\caption{$u_x$ from ANN.}
\label{fig_Ux}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UxFEMFixed}
\centering
\caption{Reference $u_x$.}
\label{fig_FEMUx}
\end{subfigure
\caption{
The mechanical response imposing Dirichlet BC $\vett u_0 = (0,0,0)$ on $\Gamma_d$ (a) provided by ANN trained via MGA-MSGD (b) the reference one provided by FEM.}
\label{fig_UXcomp}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UyFixed}
\centering
\caption{$u_y$ from ANN}
\label{fig_Uy}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UyFEMFixed}
\centering
\caption{Reference $u_y$.}
\label{fig_FEMUy}
\end{subfigure
\caption{The ANN results are in a good agreement with the reference ones highlighting its feasibility.}
\label{fig_UYcomp}
\end{figure}
\begin{figure}
\centering
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UzFixed}
\centering
\caption{$u_z$ from ANN}
\label{fig_Uz}
\end{subfigure} \quad
\begin{subfigure}{5.8cm}
\includegraphics[width=5.8cm]{UzFEMFixed}
\centering
\caption{Reference $u_z$.}
\label{fig_FEMUz}
\end{subfigure
\caption{Displacements in $\vett z$ direction are also in agreement with the reference one.}
\label{fig_UZcomp}
\end{figure}
\begin{figure}
\centering
\includegraphics[width = 8cm]{MSEu}
\caption{Local displacement error $\mathit{MSE_u}$ calculated via Equation \eqref{eqMSEu}}
\label{fig_MSEu}
\end{figure}
Finally, we visualise the local displacement error $\mathit{MSE_u}$ over the domain in Figure \ref{fig_MSEu} which shows that although the error is acceptable it exists mostly on the corners and where the displacement gradients are higher leaving the possibility to further refinements of the method in future works.
\section{Conclusion} \label{sec_conclusion}
We have introduced the hybrid MGA-MSGD training approach for ANNs with application to solving mechanical problems in three spatial dimensions described by elliptic PDEs. The latter application requires third order derivatives of the outputs with respect to the inputs and each network parameter, which is considerably time-consuming, especially, if we need a dense training data set/points and a large number of network's parameters.
This novel training approach includes modified GA adjusting the learnable parameters' components which cause the error explosion so that we can employ large learning rates (CSGD) avoiding several local minima. This is followed by a Fine-scale SGD training procedure to obtain the most accurate results. The method introduces some new non-mechanical parameters to be tuned via experiments and sensitivity analysis for a timely procedure. We show that the obtained results are less sensitive to the number and distribution of data points, and learning rates which are crucial achievements. Furthermore, this method allows us to exploit a larger potential of the network hence obtaining accurate results from small networks. The effectiveness of the training procedure is compared with two competitors, namely the classic SGD and another SGD-based optimiser (Adam optimiser), showing a significant improvement in both accuracy and efficiency.
The obtained results can be accurate up to a rigid-body motion. We have introduced the stress normalisation and displacement boundary condition enforcement weight in order to eliminate the need for any post-processing such as the mentioned rigid-body motion.
The same problem is then solved via FEM which, having sufficiently fine space discretisation, is considered as the reference solution. The accurate final mechanical response obtained via the presented training method highlights its reliability. This approach considers the strong form of the governing equations providing results that are directly differentiable and free of space discretisation (mesh-free). In practice, in order to obtain the response of a specific point, one does not need to discretise the whole domain and obtain the response of all the nodes (path-independent response). This advantage reduces the complexity of the problem implementation while it provides us with more flexibility. The authors highlight that this method should not be taken as a replacement for approaches such as FEM and finite difference method but complementing them.
The potential of the presented framework is further studied by solving a slightly more complex problem highlighting that it can be applied to a broad range of the scenarios of interest such as poroelasticity \citep{DEHGHANI2020103996, HdehghaniThesis, Hdehghani}, by including time as the 4th dimension, and non-linear elasticity with residual stress \citep{DEHGHANI201951, Njfiber, JHA201979} by using the corresponding governing equations and measures.
\section*{Acknowledgements}
We acknowledge the support of this research work via the framework of DTU DRIVEN, funded by the Luxembourg National Research Fund (PRIDE17/12252781), and the project CDE-HUB, funded by the Luxembourg Ministry of Economy (FEDER 2018-04-024).
\bibliographystyle{plain
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 2,664
|
import express from 'express';
import Authenticate from '../middleware/Authenticate';
import Authorize from '../middleware/Authorize';
import RolesCtrl from '../controllers/Roles';
/**
* Roles route class, for handling request to the routes
* @class Roles
*/
class Roles {
/**
* Creates an instance of the Route controller
* Setup Middleware methods
* @method constructor
*/
constructor() {
this.authenticate = Authenticate.route;
this.authorize = Authorize.route;
this.router = express.Router();
this.rolesCtrl = new RolesCtrl();
this.baseRoute();
this.baseRouteParam();
}
/**
* Returns a router to be expose other routes created
* @method route
* @return {Object} Router object
*/
route() {
return this.router;
}
/**
* Handle request made to the role root route
* @method baseRoute
*/
baseRoute() {
this.router.route('/')
.get(this.authenticate, this.authorize, (req, res) => {
this.rolesCtrl.all(req, res);
})
.post(this.authenticate, this.authorize, (req, res) => {
this.rolesCtrl.createRole(req, res);
});
}
/**
* Handles request with query string parameters
* @method baseRouteParam
*/
baseRouteParam() {
this.router.route('/:id')
.get(this.authenticate, (req, res) => {
this.rolesCtrl.getRole(req, res);
})
.delete(this.authenticate, this.authorize, (req, res) => {
this.rolesCtrl.deleteRole(req, res);
})
.put(this.authenticate, this.authorize, (req, res) => {
this.rolesCtrl.updateRole(req, res);
});
}
}
export default new Roles().route();
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 9,617
|
{"url":"https:\/\/artofproblemsolving.com\/wiki\/index.php?title=1957_AHSME_Problems\/Problem_19&oldid=150539","text":"# 1957 AHSME Problems\/Problem 19\n\n(diff) \u2190 Older revision | Latest revision (diff) | Newer revision \u2192 (diff)\n\nThe base of the decimal number system is ten, meaning, for example, that $123 = 1\\cdot 10^2 + 2\\cdot 10 + 3$. In the binary system, which has base two, the first five positive integers are $1,\\,10,\\,11,\\,100,\\,101$. The numeral $10011$ in the binary system would then be written in the decimal system as:\n\n$\\textbf{(A)}\\ 19 \\qquad \\textbf{(B)}\\ 40\\qquad \\textbf{(C)}\\ 10011\\qquad \\textbf{(D)}\\ 11\\qquad \\textbf{(E)}\\ 7$\n\n## Solution\n\nNumbers in binary work similar to their decimal counterparts, where the multiplier associated with each place is multiplied by two every single place to the left. For example, $1111_2$ ($1111$ in base $2$) would equate to $1 * 2^3 + 1 * 2^2 + 1 * 2^1 + 1 * 2^0 = 8+4+2+1 = 15$.\n\nUsing this same logic, $10011_2$ would be $1*2^4 + 1*2^1 + 1 * 2^0 = \\boxed{\\textbf{(A) }19}$.","date":"2021-06-25 10:33:30","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 10, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.5568840503692627, \"perplexity\": 772.0234622042499}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2021-25\/segments\/1623487630081.36\/warc\/CC-MAIN-20210625085140-20210625115140-00363.warc.gz\"}"}
| null | null |
# THE FIRST-TIME
HOMEOWNER'S
SURVIVAL GUIDE
# THE FIRST-TIME
HOMEOWNER'S
SURVIVAL GUIDE
A Crash Course in Dealing with Repairs,
Renovations, Property Tax Issues,
and Other Potential Disasters
**SID DAVIS**
CONTENTS
Preface
Chapter 1Your Home's Electrical System 101
Chapter 2Maintaining and Fixing Your Home's Plumbing System
Chapter 3How to Create Great-Looking Walls and Ceilings
Chapter 4Interior Improvements That Add Value
Chapter 5Exterior Maintenance and Improvements That Add Value
Chapter 6Your Home's Landscaping and Curb Appeal
Chapter 7Hiring and Working with Contractors
Chapter 8Tax Aspects of Owning a Home
Chapter 9Insurance Matters for Homeowners
Chapter 10Avoiding the Ten Most Costly House Problems
Appendix ADealing with Disaster
Appendix BAverage Life Expectancy of Your Home's Components
Appendix CEnergy Savings and a Greener Home
Index
A comprehensive glossary along with printable checklists and worksheets are available at www.sid-davis.com by clicking on the book banner.
## PREFACE
Your home is likely to be the biggest investment you'll ever make. You need to care for it and nurture it, not only because it's where you live, but because your financial future and emotional well-being may well depend on it. It's therefore critical to keep it in good condition. That way it will be a joy to live in and also give you maximum return when you sell. How to do this is the goal of this book.
If you're a first-time homeowner who has just moved in and feels overwhelmed, or you want to brush up on the nuts and bolts of making the most of your home, this book is for you. Some of the important items contained in the following chapters are:
•Easy-to-follow explanations and charts that illustrate how different systems and components work to make your home a marvel of comfort and convenience.
•Tips and how-to instructions for handling easy repairs on your own, even if you're not sure how a screwdriver works.
•Suggestions for hiring and working with handymen and contractors on those projects for which you need professional help, along with tips on how not to get ripped off.
•Tips on which upgrades add to the value of your home and which ones don't.
•Checklists at the end of each chapter to remind you of things to do to keep your home's components working smoothly. These checklists will also keep you on top of what needs to be done to make sure your home weathers the seasonal changes.
•How to avoid losing thousands of dollars in unneeded repairs and scams that target homeowners.
•Money-saving tips on buying insurance, replacing appliances, choosing floor coverings and other materials, and making upgrades.
In other words, this is a book you'll want to keep in a handy place so you can refer to it often when pesky homeowner problems come at you fast, demanding prompt action and/or a credit card number.
Also, for a comprehensive glossary along with printable checklists and worksheets, go to www.sid-davis.com, and click on the book banner.
# THE FIRST-TIME
HOMEOWNER'S
SURVIVAL GUIDE
## CHAPTER 1
## YOUR HOME'S ELECTRICAL SYSTEM 101
In this chapter you'll learn:
**How your home's electrical system works**
**Potential electrical problems to look out for**
**Quick and easy electrical fixes you can handle on your own**
**When you should call an electrician and how to hire one**
**A troubleshooting guide that will help you track down and solve problems**
Few things are more feared, less understood, and more discouraging to a new homeowner than the electrical system. Fortunately, it doesn't have to be that way. If you can chew gum and tie your running-shoe laces at the same time, you can understand and perform common electrical maintenance. It's that straightforward! Even when you're hiring professionals you need to have a basic understanding of your electrical system, if for no other reason than to understand what you're being charged.
### HOW TO READ THE POWER METER AND MAKE SURE YOU'RE NOT OVERCHARGED
One question many homeowners ask about their electrical service is how they can be sure the power company is not overcharging them. Meter readers are in a hurry and can misread the meter. Sometimes that can be in your favor, other times the bill goes up dramatically and you wonder why. You can solve the uncertainty be reading the electric meter dials yourself and keeping a log of your usage.
The power company charges you for power used in blocks of 1,000 watts used for one hour, or one kilowatt-hour (kwh). If you look closely at your electric meter, you'll notice a small spinning wheel in the middle. When there's current flowing, the small wheel is spinning, which means it's costing you.
There are two types of meters for recording electrical usage. One is a meter similar to the odometer on your car, where you can read the kilowatts used just as you would read the miles driven. The other meter has five dials numbered like a clock face, except that the numbers run from 0 to 9 counter-clock wise (see Figure 1-1).
To read the dials, simply start with the left-hand dial and jot down the numbers. If the pointer lies between two numbers, always use the lower number even if the hand is almost on a higher number. For example, in Figure 1-1 the meter reading would be 06573.
If you want to verify the power company's reading, read the meter the same day the meter reader stops by, and again the following month. Then compare the usage to what's shown on your power bill.
_Figure 1-1_.
For example, the typical power bill in Figure 1-2 shows the previous and current meter readings, 5142 and 6075. Subtracting the two readings shows that 933 kilowatts were used this billing period. If your meter reading varies more than a few kwh from what the billing shows, call your power company's customer service number on your statement.
Some utilities charge more per kilowatt-hour over a certain amount, and/or for usage during peak hours (1 P.M. to 8 P.M., for instance).
In Figure 1-2 for example, the first 400 kilowatts cost just over 6.9 cents per kwh. The next 533 kilowatts cost over 7.8 cents per kwh, for a total of $69.70 for the billing period.
Of course, power costs vary depending on the area you live in, but check your bill for the power company's website. They usually have special rates and programs that can help you lower your power costs.
_Figure 1-2_.
### THE DANGER OF WORKING WITH ELECTRICAL POWER
Broadly speaking, the electricity that powers your MP3 player, computer, or refrigerator is the same as lightning; the only difference is that one is controlled and the other is raw, untapped electrical energy.
When the power company generates electrical energy and distributes it over a long distances, it needs to maintain a high voltage in the lines. It's helpful to compare voltage to water pressure. A lot of pressure (lbs. per square inch) is needed to push water through the pipes to the those homes on the system that are farthest removed. Likewise, lots of electrical pressure (volts)is needed to get power out to the power company's farthest removed customers.
Another important electrical measurement is _amps_ , which is the amount of electrical power you have to work with. Using the water analogy, the amount of water in a system is measured in gallons, whereas the amount of power coming into your home is measured in amps. For example, to show the difference between volts and amps, shuffling your feet across the carpet may generate a charge of about 400 volts. That this charge will jump from you to another object shows there's a lot of pressure. But it's not dangerous because there are so few amps involved, just a lot of voltage.
But if you were to up the amps to the 15 or so that you find in a 120 volt household circuit, it's a different story—now it's enough to kill you if you get careless.
The good news is that electricity follows natural laws and is highly predictable. The downside is if you don't follow the rules there can be unpleasant consequences.
So it's important to keep in mind the following safety rules as you work on your home's electrical system.
### IT ALL STARTS WITH THE SERVICE PANEL
Safety starts with knowing where the service panel (sometimes called "breaker box") or fuse box (if you have an older system) is located. This is where you can cut off power to different parts of the house. If you see a smoking or sparking receptacle box, the service panel is where you need to run to and cut off the power. Details on how this works is covered in the next section.
The panel is usually located in the basement, a closet, near the back door, or even outside near the electric meter. So your first step is to locate your service panel or fuse box and open it up. You'll notice a row or double row of on/off switches or round fuses. Each on/off switch (fuse) in the box has a wire or circuit going from it to somewhere in your house, whether it's could your living room wall receptacles or your refrigerator. If you were to flip that switch to off, that particular circuit would be dead, and it would be safe to work on.
### Working Safely
With that introduction, the following tips will help you work with your electrical system safely:
•Always shut off the power to the circuit you plan on working with. If you're not sure which breaker (or fuse) goes to that circuit, plug in a light and start flipping breaker switches until the light goes out. It doesn't hurt to tape the "off" breaker (or remove the fuse) to prevent it from accidentally being turned on while you're working on the circuit. One home renovator—who shall remain nameless—was working on some wiring when a helper flipped all the breakers on. Luckily the owner wasn't hurt, but the short destroyed the electric meter and the owner got a don't-you-know-what-you're-doing lecture from the power company rep.
•Make sure all of the breakers are correctly labeled. (How to label the circuits in the breaker box is explained later in the Mapping Your Electrical System section.)
•Use a circuit tester to make sure wires are not hot (meaning there's no electricity flowing through them). You can find inexpensive circuit testers in any hardware store. A tester is simply two probes (wires) connected to a light or buzzer (see Figure 1-3). When the wire is hot, the light will glow or the buzzer will sound.
•Never stand on a wet or damp floor when working with electricity. Use dry boards or a rubber mat.
•Avoid touching plumbing, fixtures, radiators, or metal ducts while working with electricity. If you accidentally touch a live wire and a metal pipe, you could get seriously hurt.
•Never touch the wires going into your service panel. They are still live even if you have pulled the main disconnect or flipped all the breakers off.
•If you have overhead wires, be careful moving aluminum ladders around them. Aluminum is a great electrical conductor, and touching an overhead wire with the ladder can give you a bad shock.
•Use tools with rubber or plastic covered handles.
_Figure 1-3_.
### THE SERVICE PANEL: HOW IT WORKS
Power from the utility company goes from their pole or underground line to the electric meter and from there to the service panel or fuse box (see Figure 1-4).
_Figure 1-4_.
If your power comes from an overhead line, you'll see a grey cylinder or transformer mounted on the pole. If the power comes from underground wires, there'll be a large metal box somewhere on your property that contains the transformer.
The transformer's job is to reduce the main line's voltage down to what your house can handle. Three wires go from the transformer to your service entrance and meter then to your service panel or breaker box. From there, wires or circuits run throughout your house to the switches and outlets.
The first thing you want to become familiar with is the location of the main disconnect, which cuts all power to your house. Some breaker boxes have a main switch at the top of the panel; others don't and you have to look outside to where the meter is, as in Figure 1-5.
If you have a fuse box, the main disconnect is likely a double fuse block with a handle you pull out to cut the power, as illustrated on the right in Figure 1-4.
Both the breakers and the fuses have the same job, which is to protect the electrical system. For example, if lightning were to strike close by and overload the system it would flip the breaker off or cause the metal connector in the fuse to melt. To restore power you just push the breaker back to the "on" position. But if you have a fuse-protected system, you would need to replace the fuse. It's always a good idea to keep several spare fuses where they're easy to find, since having no fuses on hand means no video games and melted ice cream.
Some other conditions that can trip a breaker or blow a fuse are:
•Putting too many items on a circuit.
•Overloading a circuit with an appliance that pulls more power than the circuit is designed for. Common examples are hair dryers, heaters, and frying pans. Shorts from frayed electric cords can also trip a breaker.
_Figure 1-5_.
•A spike in the power company's power grid. (This is why you need a surge protector for your electronics.)
•An electric motor or appliance malfunctions, water getting into in a receptacle or switch, or a loose wire touching a metal object.
HELPFUL ELECTRICAL HOW-TO WEBSITES
www.selfhelpandmore.com/switchoutlet/index.htm
www.electrical-online.com/Replacingabreakerinyourpanel.htm
www.doityourself.com/stry/typesfusescircuits
### Mapping Your Circuits
Once you know how to cut off the main power, the next step to becoming buddies with your electrical system is knowing what circuits control which receptacles and switches. If the electrician who wired the home or the previous owner labeled the circuits similar to those in Figure 1-4, you're ahead of the game. But if it's unlabeled, you'll need to take the time to map the system and label each circuit.
_A word of caution:_ If you're a new homeowner or haven't paid too much attention to your electrical system, you should verify that the labels on the breakers or fuses are accurate. With older homes especially, previous owners may have changed wiring around and not noted it on the panel.
Five reasons why it's important to map your system are:
1.When you replace a receptacle or switch, cutting power to only that circuit rather than the whole house saves you the frustration of resetting all the clocks, and avoids having to turn off the electronics and interrupt others using power in the house.
2.If a circuit becomes overloaded and flips a breaker, you can narrow down the problem fast.
3.You can keep track of what items you put on a circuit so you don't exceed it's capacity. For example, if you turn on a hair dryer and a breaker flips, you'll know at once which circuit has too much of a load.
4.If you need to add outlets, you can tell whether the nearest circuit can handle an increased load safely.
5.By knowing the total capacity (amps) of your system and what you're using, you can calculate whether you can add additional circuits. For example, say you want to put in a new patio and add power for lights and a sound system, will your present system handle it?
Before you start mapping your system, you should know how many amps you have to work with, or your system's electrical capacity.
### DETERMINING YOUR ELECTRICAL SYSTEM'S CAPACITY
As mentioned above, the starting point for all electrical maintenance and improvement projects is the fuse box or the circuit breaker panel. When there's trouble on a circuit, such as a short or overload, the breakers or fuses save the day by shutting off the current.
Some older homes, however, can have a smaller, separate fuse box or breaker box added to the main electrical system to protect it when they're adding a high load appliance, such as central air-conditioning. This satellite box doesn't add to the total amps available, but adds a layer of protection.
Opening the panel, you'll notice that there are three wires coming into your fuse or breaker box from the meter: a black, a red, and a white wire. The red and black wires carry 120 volts, and the white is neutral.
Smaller circuits that power lights, receptacles, and small appliances will use the black and white wires. The white or neutral wires for all the circuits are connected to a metal strip, called the neutral busbar. The black or hot wires are connected to the fuse or breaker to power that circuit. For example, if you were to add a new 15-amp circuit, you would run a plastic sheathed 3-wire cable from where you want the power back to the panel. The white wire and bare copper ground wire would connect to the neutral busbar while the black wire would connect to the fuse or breaker.
For most circuits the black and white wires furnish about 15 to 20 amps at 120 volts. However, high load appliances, such a dryers, range/ovens, air conditioners, or shop welders require more voltage. These appliances require both the 120 volt black and red wires (totaling 240 volts) and may draw up to 50 amps.
Here's the really important part:
**Your electrical panel can handle only so many amps.**
Older homes that haven't been upgraded in sixty years probably have only a small 30-amp to 60-amp fuse box. That's a real underachiever system by today's standards. Some later systems have 90-amp services, which you may still find in older homes that haven't been upgraded. Although 100-amp services are still common in many homes, with today's power-hungry appliances and electronics, 150 to 200 amps are better.
However, if your home has a gas range/oven, a gas water heater, and a gas clothes dryer, you could get along nicely with 90 amps or less.
Somewhere in your panel or main disconnect will be the number of amps your service is rated for. For example, in Figure 1-5, the amp rating—150—is molded onto the end of the breaker switch.
Once you know how many amps are available, the next step is mapping your system to find out how many amps you're now using and how they're allocated among the different circuits.
### HOW TO MAP YOUR ELECTRICAL SYSTEM
Two people are needed to do a thorough job. One person works at the main service panel flipping breakers or removing fuses, while the other tests outlets, appliances, and switches with a circuit tester, similar to the one shown in Figure 1-3. The following five easy steps show how:
1. _Sketch the floor plan_. Draw a floor plan of each room and mark all the receptacles, switches, and light fixtures. Sketch in appliances that are connected directly to the panel, such as electric ranges, water heaters, dryers, and air conditioners.
2. _Number the circuits_. At the main service panel, number each circuit breaker or fuse with stick-on labels. Some panels may already be numbered as in Figure 1-4.
3.Set up the test. Work on one room at a time. Turn on all the lights, lamps, and appliances in the room. If the room has double receptacles, plug a light into each. Don't turn on appliances that have their own circuit.
4. _Record what's on the circuit_. The person at the service panel turns off the first breaker or fuse. On the floor plan the helper writes down the number of the circuit or fuse and which fixtures, switches, and receptacles are affected. For example, on circuit # you list everything that went out, for example, the ceiling light, hall light, and three of the five receptacles in the living room.
One homeowner plugged her hair dryer into the receptacles rather than using a circuit tester to see which ones were hot; another one used a small lamp. Both worked great in identifying what was working.
5. _Repeat the process for the other circuits_. Flip circuit #1 back on and repeat the process for all the breakers or fuses in the panel. When you're through, each circuit should have a label stuck next to it detailing what it controls. Do the same for any 240-volt double breakers or fuses that control the range, dryer, water heater, or air conditioner.
Now that you've mapped your electrical system and understand it, if a fuse blows or breaker trips from a short or overload, it won't be panic time. You'll be able to go right to the affected circuit and troubleshoot. However, there's a big red warning flag waving here if this happens often. Blown fuses mean your electrical system may have problems and is not meeting your needs. It could be time to think about upgrading.
### DOES YOUR SYSTEM HAVE THE CAPACITY IT SHOULD?
As mentioned above: _Amps_ are how much current a conductor can carry, and _volts_ are electrical pressure. The third important measurement, _watts_ , tells you:
•How much load or what appliances you can put on a circuit.
•How many watts, or 1,000-watt bundles called kilowatts, the power company uses to bill you for electricity used.
•How efficient your appliances are by comparing how many watts they use. This is disclosed on tags or labels on all electrical appliances.
To calculate the wattage a circuit can handle, multiply amps times the volts. (In the United States, voltage is standardized at 120 and 240.) For example, a standard 15-amp circuit times 120 volts equals 1,800 watts.
Most kitchen circuits are 20 amps, so multiplying by 120 volts equals 2,400 watts—enough to allow you to run the coffee maker and toaster at the same time.
If your kitchen circuit is only a 15-amp circuit, however, you know that the total appliances you can use at the same time can't consume more than 1,800 watts. For example, if you're using a hair dryer on turbo setting that uses 1000 watts, while at the same time warming up an iron that consumes another 1000 watts, and both on the same circuit, you're going to burn out a fuse or flip a breaker. The simple math is: 1000 + 1000 = 2000 watts on a 1,800 watt circuit.
Table 1-1 gives you an idea of some common appliance wattage ratings.
Ideally, the wattage of any one appliance should not be more than 80 percent of a circuit's total capacity. As you can see from the figures in Table 1-1, when you have too few circuits, overloading them with extensions and appliances will cause problems.
_Table 1-1. Typical wattage ratings_.
### UPGRADING YOUR ELECTRICAL SYSTEM
You should seriously consider upgrading your service if:
•The electrical system has two-wire service. Outlets take two prong plugs only and the fuse box has no bare copper ground wire. It's likely this system dates earlier than the 1940s and is a prime candidate for trouble unless you upgrade.
•You have too few outlets. Wall receptacles should be no more than 12 feet apart.
•Receptacles are not located where you need them and you have to run extension cords.
•Not all bathroom and outdoor receptacles are protected by ground-fault circuit interrupters (GFI).
•You have to unplug some appliances to run others.
•Using a high-amp appliance, such as a hairdryer, toaster, or waffle iron, flips a breaker.
•You want to add more receptacles or another circuit but there's no room in the panel.
•Your system is protected by fuses rather than circuit breakers.
•The system's amp rating is less than 100 amps. (When you're tripping breakers, it may be a good investment to go 150 or 200 amps.)
•Some of the wiring shows its sheathing is old and deteriorating from age, rodents, or environmental damage.
•You have aluminum wiring mixed with copper-wire additions.
The cost of upgrading your electrical system typically runs from $2,500 to $12,000, depending on how extensive the remodeling job is. A beginning point to determining what it will cost for your home is talk to several electrical contractors. (Tips on finding one are covered later on in the chapter.)
In most areas, upgrading the electrical system will increase your home's value equal to or more than your cost. This is an important upgrade for buyers. A new _electrical service entrance_ (where the power company's lines enter your home) and a 150- or 200-amp breaker box with room for expansion are great selling points.
To determine exactly how much you will recoup, talk to a couple of appraisers who work in your area. Or have a Realtor pull up on their MLS (Multiple Listing Service) database sales of homes with upgraded electrical systems, compared with sales of homes with older systems. This should give you a good idea of what your rate-of-return will be.
It's also important to note that if you have a house fire, the insurance company usually pays to restore your home to what it was, not according to the current code. That means if your home's electrical system is not updated, you'll get a paltry check for what it would take to restore it. But the real kicker lies with the city building codes. You won't be able to rebuild your home unless you bring the electrical (and usually everything else) up to code. It can be a case of pay me now or pay me later.
While you're thinking about it, it's a good idea to contact your homeowner's insurance agent and see how your policy handles this.
### WHAT ELECTRICAL REPAIRS YOU CAN DO AND WHEN TO CALL A PROFESSIONAL
Even though your electrical system may be straightforward and you don't need a physics degree to handle repairs, lack of experience or lack of the right tools means you'll occasionally need professional help. Still there are a number of problems you can handle yourself and save a few bucks, such as replacing switches, outlets, blown fuses, and worn-out breakers.
### Replacing Switches
Because switches take a lot of abuse, it's amazing they last as long as they do. But when they go out, replacing them isn't difficult.
Basically, a single pole switch has two terminals. The hot wire (black) is hooked to one terminal screw and the outgoing wire (again black) is hooked to the other one. The white wire remains uninterrupted or is connected by a wire nut.
You can tell when a switch goes bad if you flip the toggle and there's a pop, flash, or burned plastic smell, and the switch no longer works.
Replacing a malfunctioning switch is easier than learning how to doubleclick the left mouse button. Here's how:
1.Turn off either the main power disconnect or the circuit you're working on.
2.Remove the face plate. You may have to run a razor blade along the edge to free the plate from a painted wall.
3.Remove the two screws that hold the switch to the box and pull the switch out from the box far enough to work on the connectors.
4.Notice how the black and white wires are connected to the switch. Draw a diagram so you can repeat the connections when you hook up the new switch.
5.Disconnect the wires from the switch. The wires will be connected one of two ways: either by pushing the wire into a spring loaded clip in the back, or the wire will be bent around a screw terminal and tightened (see Figure 1-6).
6.If the wire is attached with a spring clip, simply insert a screw driver tip as illustrated in Figure 1-7 and release the clip. Take the old switch to an electrical supplier or hardware store and get a similar switch.
7.Connect the wires to the switch and replace the retaining screws. Note that many switches have both screw terminals and spring clips to attach wires. If you go the screw terminal route make sure the wire loop tightens as you turn the screwdriver clockwise as in Figure 1-6.
8.If you use the spring clip to attach the wires, simply remove about a half-inch of insulation and insert the wire end into the hole in the back of the switch as in Figure 1-7.
9.Replace the mounting screws and the face place, turn on the breaker and give it a test. It should work!
Throughout the house are more complicated switches, such as two- and three-way switches, and you can replace those too. Simply follow the previously outlined steps and draw a detailed diagram of where the black and white wires go. Then take the bad switch to a supply house, match it, and hook up the wires the same way you disconnected them.
### Replacing Outlets
When outlets malfunction it can be more dramatic than when a switch goes. For example, you can plug or unplug an appliance and get a bright flash, the smell of burning insulation, or a shower of sparks and a tripped breaker.
Replacing an outlet is similar to replacing a switch. Disconnect the main breaker or fuse and remove the face plate. You'll notice there can be a lot more wires. That's because outlets are sometimes wired in a series, or wires run from one outlet to several others.
_Figure 1-6_.
Again, as you did with the switch wiring, draw a diagram of where the black and white wires go. Take the broken outlet to a electrical supply store and match it, making sure you get the right amp rating, 15 or 20. While you're at it, buy some extra outlets, because they don't always malfunction during business hours.
Replace the screws and test it. It should work.
### Blown Fuses
Blown fuses can be a symptom of a worn out switch or outlet, or a short. The first two are easy to replace, but a short may be more difficult to track down. But under no circumstances replace a fuse with one that has a higher amp rating than the circuit is rated for, not even long enough to finish your favorite TV program.
_Figure 1-7_.
For example, replacing 15-amp fuses with 20-amp fuses to keep them from blowing is asking for trouble. _Over-fusing_ , as it's often called, is a major contributor to over 40,000 electrically caused fires and 350 deaths a year, according to the U.S. Consumer Product Safety Commission.
If you replace a fuse and it burns out again, follow these steps:
•Determine what's on the dead circuit. Here's where your circuit map comes in handy. You may have overloaded the circuit in an unthinking moment by plugging in a high wattage item or an extension cord into the nearest outlet.
•Check the switches and outlets for scorch marks or melted connectors.
•If you have appliances on the circuit, plug them into another circuit and see if one of them has shorted out and is causing your fuse to blow. If this is the culprit, replace the appliance.
•If possible, check the wiring from the box to the outlets. A rodent may have chewed the insulation and caused a short.
If after trying these suggestions the circuit still blows fuses, it's time to call an electrician and hunt for your checkbook.
### Malfunctioning Breaker Switches
Although breaker boxes come in many different styles, one thing they have in common is that when a circuit overloads it trips a toggle switch. The tripped toggle will typically flip all the way off or to a center position. Some designs have a red indicator on the side of the toggle that is visible when the toggle is tripped; with others you have to look for the toggle that's out of line with the rest of the toggles.
Some designs will let you reset the breaker by simply pushing it to the "on" position, while others require you to push the toggle to the "off" position first, and then to "on." If the breaker keeps tripping when you try to reset it, then you'll need to troubleshoot the problem as outlined for fuses above.
If after troubleshooting, the breaker won't reset when you push the toggle to "on," it's possible that you have a faulty breaker. To replace the breaker switch simply:
1.Cut off the power at the main disconnect.
2.Remove the screws that connect the panel cover to the box and remove the cover.
3.Loosen the screw and disconnect the black wire connected to the faulty breaker.
4.The breaker may have a screw that holds it in place. If so, remove the screw and the breaker should come out easily. Sometimes breakers that have been in for a long time tend to stick and you may have to tap on it with a screwdriver to loosen it.
Some breakers have a clip instead of a retaining screw. In that case, pull the breaker toward you and it should come free. If it's a close fit, you may have to lightly pry the breaker free with a screw driver.
5.Take the faulty breaker to a electrical supply or hardware store and get an exact match.
6.Install the new breaker, flip the toggle to the "off" position, and replace the panel.
7.Turn on the main disconnect and flip the new breaker toggle to "on." If it snaps off, you've still got an serious electrical short on that circuit and you'll need to call a professional electrician to track down the problem.
As you replace a few switches, outlets, and do some troubleshooting, your confidence and knowledge of your home's electrical system will grow and you'll also save quite a few bucks. Check out the info on the websites listed in this chapter and you can soon add fixtures and additional circuits on our own. For more complicated electrical problems hire an electrician.
### HOW TO FIND AN ELECTRICAL CONTRACTOR
If your home's electrical system needs upgrading, it's important that it's done according to the national electrical code (NEC) or your local electrical codes. Yes, you can cut costs by having your brother-in-law, who got a C in high school shop, to do the work. But in the long run that can come back to bite you. How?
1.If the wiring wasn't up to code and you have a fire, it could affect your insurance settlement.
2.When you decide to sell your home, the home inspector, appraiser, or lender may require copies of permits, inspections, or work orders.
3.Substandard wiring can prolong a sale, kill a sale, or force you to reduce your asking price—often by more than what it would have cost you to do it professionally in the first place.
4.Electrical wiring isn't something you cut corners on. You and your home's safety are too important.
The first step to finding a electrician or contractor is ask around for referrals, and check with contractors, building inspectors, and Realtors. The goal is to end up with a short list of three contractors (licensed electricians) who will look at your home and give you a detailed written bid.
Incidentally, one way to cut costs is to find a licensed electrician (verify they have a current and active license) who works for a contractor, builder, or even an inspector, and who does jobs on the side for a little extra income. But hiring one of these people doesn't mean you cut down on the detailed paperwork. As with any business deal, the paperwork is your only protection.
No matter whom you hire, here is a list of items you need to have in writing:
•A detailed written description of the work agreed upon.
•An itemized list of materials to be used: number and cost of outlets, receptacles, wiring, breaker boxes, and so on. You can also have the electrician give you a list of materials, which you can then buy. That way you pay for installation only, control the quality and cost of materials, and always know you aren't being overcharged.
•A beginning date and completion date on the paperwork.
Never pay for the job up front. If it's an extensive remodel, you may want to set up a payment schedule as certain parts of the job are completed.
Finally, check with your homeowner's insurance agent and make sure you're covered in case the electrician or any helpers are injured on your property.
To wrap up this chapter, check the guide in Table 1-2 when you have an electrical problem. Electrical components wear out, malfunction, and are misused. Unfortunately the potential for serious injuries or worse is high, so it's important to stay of top of problems and correct them as soon as possible.
_Table 1-2. Electrical troubleshooting guide_.
Homes older than 15 years | Get an inspection by an licensed electrician or electrical inspector.
---|---
Any arcing, sparking, or smell of burning insulation | If it's outside the house, call the power company fast. Inside, cut the circuit disconnect. If you don't know which circuit, flip the main disconnect. Call an electrician to check the wiring.
Lights flicker and the TV screen distorts, especially during bad weather. | Could be a bad connection where power comes into the electrical service entrance. Call the power company immediately.
Fuse blows or breaker trips on a circuit. | Check and see what appliances or extension cords are on that circuit and make sure you haven't exceeded the wattage capacity.
Inspect the wiring if possible for damaged wiring.
Circuit capacity | 15-amp circuits can carry about 1500 watts. 20-amp kitchen circuits can carry about 2000 watts. Add up the wattage of appliances on circuit and take care not to run appliances with too high wattage simultaneously.
One appliance repeatedly blows fuse or trips breaker | The appliance is probably defective, don't use. Repair or replace.
Switch, light, or receptacle blows fuse or trips breaker when use. | Isolate what's on the circuit and check each item for damage. Replace after cutting off main or affected circuit disconnect.
Overamping a circuit. | Never use a higher amp fuse than the circuit is rated for. Example: Using a 20-amp fuse in a 15-amp circuit is a deadly problem waiting to happen.
Lamps or lighting fixtures are overheating or show signs of heat discoloring. | Check the light bulb for too high wattage for the fixture. Example: 100 watt bulb in 60 watt lamp. Match the fixture to the correct bulb. Recessed lighting may have too much insulation and heat can't dissipate; call an electrician to check recessed lighting problems.
You get a shock touching a metal faceplate of an outlet or switch, a lamp socket, or a metal part. | Cut off the circuit and look for a short or miswire in the receptacle or lamp. If you can't locate the problem, check with a qualified electrician.
Receptacles or switches feel hot when touched. | You may have more power than the circuit is designed for. Check the fuse for too high amp rating for circuit. Also, check for loose connections. Call a electrician if it isn't a fuse or connection problem.
Aluminum wiring. | Have an electrician make sure compatible connections are used in outlets and switches. Incorrect connections have caused fires.
You get a shock when touching the outside of metal appliances. | Your wiring system or appliance may not be grounded. Call an electrician ASAP. Shorts and grounding problems are dangerous.
## CHAPTER 2
## MAINTAINING AND FIXING YOUR HOME'S PLUMBING SYSTEM
In this chapter you'll learn:
**How your plumbing system works.**
**What problems you can solve on your own.**
**What problems are best handled by a professional plumber.**
**How to upgrade your plumbing system. What pays, what doesn't.**
**How to solve problems using the Troubleshooting Guide at the end of the chapter.**
Along with the home's electrical system, plumbing ranks near the top of homeowner's and homebuyer's concerns. But the good news is there's much you can do on your own to fix and maintain your home's plumbing system. You can save a lot of money on service calls and take care of many problems fast without calling a plumber.
Your home's plumbing system is not that complicated; it's just a matter of taking the time to understand how it works and what components tend to wear out. In fact, if you can stand on one foot and pat your head at the same time, you can handle most of the plumbing problems that occur.
Also important is that upgrading and keeping your plumbing system in good condition will make a big difference when you sell your home. And this chapter shows which improvements give you the best return.
Basically, your home's plumbing system is a fairly simple dual system. One part handles water under pressure coming into the home from a municipal system or private well. Components of this part are:
•Your home's supply line, main shutoff valve, pressure regulator, and water meter that ties into the city water main. If you have a private well, the main supply line connects to the well's pressure tank instead of the city water main.
•Cold water pipes that supply water to faucets, toilets, washer, and other appliances on the system.
•The faucets and valves that control water flow until you need it.
•Typically a 3/4-inch pipe supplies cold water to the water heater. From the water heater, 1/2-inch hot water lines supply faucets, dishwasher, washer, showers, tubs, and wherever else you have hot water. In Figure 2-1 the dashed lines are cold water pipes and the solid lines are hot.
•If the city water pressure is over 100 pounds per square inch (psi), a high pressure regulator—usually near the turnoff valve—reduces it to less than 80 psi.
•Of course, for city water the water meter records the number of gallons you use during the billing period. If you forget to pay the bill, the city can shut off the water at the main turnoff, usually located in the street or parking strip.
MONEY-SAVING TIP
Read your meter monthly and keep track of the gallons used. This ensures the water company doesn't make a mistake at your expense. Also, if a water line leaks or breaks, you'll know it before you get hit with a huge bill.
The other part, called the DWV (drain, waste, vent) system, is gravity powered and removes water and waste from the home to a municipal sewer system or private septic system. Its main components are:
_Figure 2-1_.
•1½-inch to 1¾-inch drain pipes that carry wastewater from fixtures to the 3-inch or 3½-inch diameter, plastic or cast-iron main soil stack that connects to the city sewer or private septic system.
•Drain or "U" traps that connect to the drains of all fixtures. Figure 2-2 shows how these connect to the main soil stack. Drain traps are necessary on all fixtures that connect to a drain because they trap water in the bottom of the U. This creates a water seal that prevents sewer gas (very bad smelling) from leaking into the home and spoiling your dinner party.
•When water moves through a pipe it creates a vacuum and wouldn't flow efficiently if there weren't vents to let outside air into the system. The main soil stack creates one vent, but other secondary vents are needed for each branch that connects to the main stack. Figure 2-3 shows the roof of a typical home where several branch drains vent through the roof.
Once you have an idea of how your home's plumbing system works, regular maintenance will keep it working smoothly and save big bucks on service calls. The next section will detail how the supply system works and how to maintain it.
_Figure 2-2_.
### YOUR HOME'S WATER SUPPLY SYSTEM
As mentioned previously, water under pressure piped into your home usually comes from a municipal source or a private well. If you buy water from the city, a branch pipe delivers it from the main water line to your service entrance, which includes a shutoff valve and a water meter.
Just as the main switch cuts off electrical power to your home, as we discussed in Chapter 1, the service entrance gate valve cuts off water flow to your home. The main water valve is usually located close to where the branch pipe connects to the main. In cold climates, the water main and service lines are buried several feet underground below the frost line. The turnoff valve is typically located at the bottom of a round pipe fitted with a metal cap.
_Figure 2-3_.
USEFUL PLUMBING WEBSITES
www.hometime.com/Howto/projects/plumbing/plum_7.htm
www.doityourself.com/scat/plumbing
www.allabouthome.com/directories/dir_plumbing.html
www.friendlyplumber.com/plumbing101.html#septic%20system
In most homes, there are also shutoff valves to all the sinks, toilets, and faucets. However, on some older water systems the main valve is the only way to shut off the water.
If your home has only one shutoff valve, updating the plumbing system and bringing it up to code should be at the top of your improvement list. In most areas, these upgrades will increase your home's value by more than the cost of your investment.
One important plumbing item is faucets. Upgrade them or keep them in good leak-proof condition—not only to save water and keep it from staining sinks, but to enhance your home's value.
### FAUCET SAVVY 101
Many new homeowners feel that fixing or replacing leaking faucets is something you have to call a plumber for, with credit card in hand. Actually, you can do it yourself. It's not that difficult, all you need is a few tools and a PMA (positive mental attitude).
For starters, plan a shopping trip to your favorite hardware store or home center to pick up a few tools that you may not have. If you're unsure, ask someone in the plumbing section to help you select the right tools. The list should include:
•Adjustable or slip joint pliers, regular and needle-nose pliers.
•An 8-inch or 10-inch adjustable wrench.
•Set of screw drivers that has both flat and Phillips head blades.
•Basin wrench. This is a must-have tool for turning nuts that hold faucets to sinks. It looks like a metal bar with a swivel head and jaws on one end and a slip bar on the other to give you leverage turning the nut. In Figure 2-4 the length of the tool in the foreground adjusts to allow it to fit into tight places.
•Set of Allen, or hex-head, wrenches. These are small, L-shaped, six-sided wrenches that fit Allen or set screws found on faucet handles.
•If you have an older home with separate handles for hot and cold water, you'll need a valve-seat wrench. These look like oversized Allen wrenches with slightly tapered ends to remove valve seats.
•In areas where water has a high dissolved-mineral content (hard water), handles can almost bond to valve stems. So rather than force the handle off the stem and damage it, use an inexpensive screw-type handle puller to break it free.
•Utility knife and extra blades.
_Figure 2-4_.
This is a basic set. However, you may need a few other tools depending on the faucet types you have in your home.
DECORATING TIP
If you are renovating an older home or are remodeling, pick out a quality manufacturer, such as Moen, Delta, Peerless, or American Standard. Go with a specific line or style to ensure that all faucets are similar. This will help tie the home's décor together and make it more attractive.
### Fixing the Five Most Common Faucet Types
Before fixing a faucet, don't forget to first turn off both hot and cold water lines to the sink.
1. _Two-Handled Faucets_. Two-handled faucets are commonly equipped with washers (compression style) or ceramic stem cartridges; you'll know which type when you disassemble them. If you live in a hard-water area, the seats and washers will need replacing more often than a soft-water area. To fix a compression-style faucet:
a.Pop the cap off the top of the faucet handle (on some models the screw may not be covered with a cap) with the tip of a utility knife to access the handle screw.
b.Remove the screw and pull off the handle. If the handle is corroded and won't come off easily, use a handle puller.
c.Use an adjustable wrench to loosen the stem retaining nut and remove the stem (see Figure 2-5).
d.On the bottom of the stem is a rubber washer retained by a screw. Replacement kits usually contain two washers, brass screws, and rubber O-rings. First time around, you'll probably need to take the stem to a plumbing supply for a correct match. If the stem is badly worn, replace it too. When replacing the stem and seat, take care not to over tighten.
e.Another part you'll probably need to replace is the valve seat, especially if you have hard water. Hard water under pressure can cut a small U in the seat causing a steady drip. The valve seat has threads like a screw, so you'll need to use a special tool—a valve seat wrench—to remove it (see Figure 2-6).
_Figure 2-5_.
_Figure 2-6_.
f.Reassemble the faucet with the new parts and test. If the faucet is worn or corroded, you're probably better off replacing it.
2. _Ceramic-Disc Faucets_. Ceramic-disc faucets consist of a cylinder with a moveable disc that rotates to control the hot and cold water mix. In this type, the seals wear and allow the faucet to leak. To disassemble and replace the seals:
a.Remove the small set screw in the handle—it will likely take an Allen wrench or a small flat-head screwdriver.
b.Next, unscrew the cap that houses the cylinder and disks.
c.Remove the screws holding the cylinder in the faucet body and pull out the cylinder. Use pliers if it sticks.
d.Turn the cylinder upside down, remove the seals, and clean the base and seats.
e.Install new seals and reassemble. You may have to take the old seals to a plumbing supply store or home center to ensure the right match.
3. _Ball Type with One Handle_. These faucets are usually inexpensive, and sometimes it's better to replace such faucets rather than try to fix them. However, if you think the faucet is worth fixing you can buy repair kits. To repair, remove the set screw in the handle to access the faucet parts. When you dissemble it, note how the parts fit. Replace the old parts with those in the kit (including the O-rings) and reassemble.
4. _Tipping Valve, Single Handle_. The heart of this faucet is a cam that rotates when you turn the handle, allowing the hot or cold side to open and close. These are reliable designs, but sometimes particles stick to one of the seats, causing a drip. So before grabbing the pliers, try flushing the problem away by opening and closing the faucet quickly a few times. If this doesn't work, the problem is likely damaged valve seat. To access the seat:
a.Start by unscrewing and removing the spout. Take care that you don't scar the spout with the pliers when unscrewing the base. Putting masking tape on the pliers jaws can help prevent this.
b.With the spout off, remove the cover and you'll see a slotted nut on each side of the faucet body. Remove the nut and you can pull out the valve. If the valve is corroded, replace it; otherwise replace the seat only. You may want to take the valve assembly to a plumbing supply store to get the right parts. Also, replace the O-ring on the base.
c.Screw the base back into the faucet body and the leak should be no more.
5. _Cartridge Style with One Handle_. This type is sometimes more difficult because the retaining clip holding the cartridge in the faucet is harder to find. It may look like a small ridge projecting out of the exterior under the handle. Needle-nose pliers are the tool of choice to pull it out. You may also have to pop a cap off the top of the handle to remove a screw that holds the handle on. Once the handle is off, you can find the retaining clip that holds the cartridge in. Take the old cartridge to a plumbing supply store for the right match.
These faucets are some of the more common types, but if you have one that is different, don't panic. The basics are similar: 1) Find the screw that retains the handle and remove to get to the inside. 2) Remove the stem or cartridge and replace along with any O-rings or seats.
If you decide to replace an old or ugly faucet rather than fix it, the next section tells how.
### INSTALLING A NEW FAUCET
Of course, before installing a new faucet you need to remove the old one. Sometimes this is the hardest part of the project, because mounting and coupling nuts that connect faucet to sink are often corroded from being in a humid environment. To find out, look up under the sink with a flashlight and see if there's corrosion, stains, or dirt where the retaining nut threads onto the faucet. If it looks bad, spray the connection with WD40 or another penetrating lubricant and let it set for an hour or so while you catch the game's second half.
After turning off the hot and cold water supply, the following steps show how to remove the old faucet:
1.An adjustable wrench does a good job loosening the water line and faucet mounting nuts if you can get to them. Otherwise, you'll need to use a basin wrench, as illustrated in Figures 2-7 and 2-4. Remember, turn the wrench counter-clock wise to loosen the nuts.
2.If the faucet has a lever that works the pop-up sink stopper, you'll need to loosen the screw that connects the lift rod and pull it out. On kitchen faucets, unscrew the spray hose if it has one.
3.Remove the old faucet and gasket if it has one, and clean up any corrosion or old putty left on the sink.
4.There are two ways to mount the new faucet: 1) Use the gasket that comes with the faucet, or 2) apply a bead of plumber's putty around the base plate's perimeter. In some environments, the gasket may dry out faster than putty, but using the gasket is less messy than spreading a bead of plumber's putty with your finger.
_Figure 2-7_.
While you've got the water turned off and the lines disconnected, check them to make sure they're in good condition. If not, replace with new braided flex lines and connectors. Also, check the shutoff valves. If they're corroded or the handles are hard to turn, replace them as well. Don't forget to use a turn or two of white Teflon tape on the thread end of metal-to-metal connections. Teflon tape, which comes in small spools, not only helps prevent connection leaks, but it makes connection nuts easier to remove even after years of service. As an alternative to Teflon tape, you can apply pipe joint compound to the threads.
When the lines are connected, turn on the water and look for leaks. If you find a connection that's leaking, tighten the nut a little at a time until the leak stops.
### TOILET SAVVY 101
No, this section isn't about potty-training kids, which is light-years beyond this home renovator's expertise! But more in the comfort zone is how to keep the potty functioning smoothly, and how to replace parts as they wear out.
HOME VALUE-INCREASING TIP
If you have an older home you're fixing up, start with the kitchen and then next tackle the bathrooms. Make all the bathroom fixtures match so you create a flow. Try not to have mismatching toilets or faucets.
Fixing or keeping toilets in good condition is not difficult. But it's important to stay on top of problems and do as much preventive maintenance as possible. Remember, the humid and chemical environment of a bathroom is unfriendly to metal and plastic parts.
Basically, toilets consist of two systems: water under pressure in; water under gravity out. Simple, but it all has to work together to get the job done. Figure 2-8 illustrates basic toilet design.
Note that a plastic or metal closet flange screws into the wood subfloor and connects to the waste pipe. This provides a solid mounting for the toilet. A very important wax gasket connects the waste pipe to the toilet base. This gasket prevents leaks when the bowl is flushed and water, with the help of gravity, goes down the waste pipe.
Eventually, the wax gasket loses shape or hardens and allows water to escape from under the base. If not corrected, the leak will rot the subfloor. When that happens a section of floor around the toilet will have to be cut out and replaced. This is often the cause of spongy floors in bathrooms and is high on the list of things property inspectors look for. It's also a good reason not to make an offer on a house if you're shopping for one.
Other important maintenance items are the two bolts mounted in the closet flange, shown in Figure 2-8. The toilet base fits over the bolts and is secured by a nut and washer. For cosmetic reasons, a plastic cover hides the bolt and nut.
_Figure 2-8_.
Because these securing bolts and nuts corrode in the humid bathroom environment, you'll want to pop the cap off and check the bolts every few months to make sure they're tight. Loose bolts can also distort the wax gasket and cause leaks.
The water supply is another area of concern. If you need to cut the water supply to the toilet fast, there needs to be a shutoff valve under the tank where the water supply line comes out of the wall.
If there is no valve, hire a plumber to install one. An overflowing toilet can do a lot of damage in the time it takes you to find and cut off the main water supply.
For existing turnoff valves, make sure they turn easily and there are no leaks. If a handle and stem are corroded, replace immediately. These valves are inexpensive and easy to replace. Simply turn off the main water supply, drain the tank, unscrew the water supply line nuts, and remove the old valve. Wrap the threads of the new valve with a couple of turns of Teflon tape and replace.
FOUR WAYS TO LOOSEN A STUCK NUT
1.Tighten the nut slightly. This sometimes breaks the bond between corrosion and nut.
2.Tap the nut with a hammer or punch. Be careful not to damage the threads.
3.Heat the nut with a propane torch or heat gun. Make sure there's nothing flammable close by. Try turning the nut before it cools down.
4.Clean off the corrosion and soak the nut with WD40 if it's rust, or with vinegar if the problem is mostly caused by hard water. Sometimes it takes a day or two of soaking to loosen the hard water buildup.
Two other areas where a leak can happen are the tank mounting bolts and the spud washer that connects the bowl to the tank. If the tank mounting bolts corrode, they can become loose and leak. Loose mounting bolts can also damage the spud washer and leak water behind the bowl. So check the tank mounting bolts every few months to make sure they're in good condition (see Figure 2-8).
If water or a brown waxy liquid is oozing from under the bowl, you probably have a wax gasket that needs replacing. To replace the gasket or closet bolts:
•Turn off the water supply to the toilet and empty the tank by flushing it and sponging up any water remaining in the bottom.
•Disconnect the line either at the turnoff or the bottom of the tank.
•Remove the closet nuts on both sides of the bowl. If they're corroded, use WD40 or a similar lubricant. Tools of choice are a 7/16-inch (or whatever bolt-size you have) boxend wrench or ratchet. An adjustable wrench also works, but you have to be more careful not to round the nut corners.
•Lift the toilet (two people or one NFL linebacker are needed) off the closet bolts and tip it over on its side. Sometimes it's easier to remove the tank first by unscrewing the tank bolts and lifting it off the bowl (see Figure 2-8).
•Remove the old wax ring and clean up the area with a putty knife and rag. Gases from the open sewer line won't be fun to work around, so stuff a rag in the opening—just don't forget to remove it.
•While you've got the bowl on its side, inspect the openings under the rim. This is where water from the tank enters the bowl when you trip the handle. If the openings are restricted by grime or hard water buildup, the flushing action won't be optimum. Use a narrow putty knife, Scotch-Brite pad, and vinegar or other mineral dissolver to thoroughly clean the openings. In areas of hard water, this can be a reoccurring maintenance problem.
_Figure 2-9_.
•Install the new wax or rubber gasket and reseat the bowl (see Figure 2-9). Tighten the closet bolts until you feel light resistance; go another turn and stop. The nut should be snug, but not too tight.
Reassemble the bowl and tank, connect the water line, and it should work like new.
DIY PLUMBING WEBSITES
www.toiletology.com
www.stemdoctor.com/toilet_repairs.htm
www.fluidmaster.com/usa.html
www.hometips.com/content/toilets_intro.html
### MAINTAINING THE WATER IN-TAKE
The next area of irritating toilet malfunctions is the water in-take system that fills the tank and releases it to flush the bowl (see Figure 2-10). Common problems are:
1.Water continues to trickle into the bowl after the flush cycle. This is often the result of a deteriorated or misaligned flush valve (rubber ball or flapper that seals the valve seat). Check and make sure it's aligned for a good fit with the valve seat. If the valve is old or deteriorated, replace it.
2.The valve seat may be dirty and not allow the flush valve to form a water tight seal when it drops down after the tank starts to refill after a flush.
Try emptying the tank and cleaning the valve seat with vinegar. If the problem persists you'll need to replace the valve seat. To do this, turn off the water and disconnect the water line. Remove the two bolts that hold the tank to the bowl (Figure 2-8). Use a spud wrench (available at home center plumbing sections) to unscrew the valve seat (see Figure 2-11). Replace both the seat and washer and reassemble.
_Figure 2-10_.
3.Water into the tank doesn't shut off. When the tank empties after a flush, the ballcock or float travels upward as the water level increases. When the water level nears the top of the tank, the lever attached to the float shuts off the water with a plunger valve. If this valve malfunctions, water will continue to run into the tank by way of the overflow tube. Worn out seals inside the valve can be replaced, but if it's an old unit, it's better to replace it than try to track down replacement parts.
_Figure 2-11_.
Replacement is easy: Simply turn off the water and empty the tank. Then remove the ballcock assembly by undoing the retaining nut on the bottom of the tank, disconnecting the chain to the plunger valve and lifting the unit out. At the home center, you'll find several different replacement types at different prices. They'll all work, but get an employee to explain the different models and pick the one you like best. It's just like shopping for jeans.
4.The tank doesn't empty completely. This is often caused by the chain that raises the flush valve having too much slack. Shorten the chain and flush the tank to see if this solves the problem. You may have to try a couple of chain length adjustments to get it right.
In some areas with high humidity, a toilet tank filled with cold water can sweat enough water to do serious damage to floor and walls. There are two ways to solve this problem:
• _First Option:_ Buy a toilet tank insulation kit and install it on the inside of the tank, or try an external cover that insulates the cold tank from the moisture laden air.
• _Second Option:_ Have a professional plumber run a water line from the hot water tank to the back of the toilet where it and the cold water line are connected to a mixing valve that brings water up to room temperature before it enters the toilet tank.
True, the second option is a more expensive way to go, but if sweating is a big problem it's cheaper than replacing the subfloor every few years.
It also helps to have a ventilation fan that vents humid bathroom air to the outside. Incidentally, make sure fans don't vent into the attic. You don't want humid air transferring problems from the bathroom to the attic or anywhere else.
After toilets, the next important area that you need to be proactive on are the pressurized water supply pipes and the waste-removal pipes. If you have a home that's older than twenty years, it's important to inspect the plumbing system at least once a year for potential problems.
### PIPES: CARE, CLEANING, AND CLOGS
Where water is piped under pressure, Murphy's Law applies: Anything that can go wrong will. Take a proactive approach to keeping your plumbing system in good condition.
If you live in an area with cold winters that freeze pipes, extra care is needed to prevent potentially expensive problems. Many new homeowners move in during warm months, and no one tells them what below freezing temperatures can do to their water system in the months to come. So, here are a few tips on how to keep your pipes from freezing.
•Unheated basements and crawl spaces are prime candidates for cold-snap problems. Possible solutions are:
1.Wrap the pipes with insulation strips or use foam tubes that slip over the pipes.
2.Install a thermostatically controlled space heater.
3.Wrap pipes with electrical heat tape. Install a dedicated outlet rather than running an extension cord.
4.Install a small vent from your home's heating ductwork.
•Pipes that pass through outside walls are especially vulnerable to freezing. They need to be insulated with foam or insulation blown into the wall around the entry area. Using heat tape in addition to insulation is a good idea, especially if the outside wall faces north or is exposed to direct cold winds.
•For outside spigots, install a turnoff far enough into the basement or crawlspace to protect it from freezing. You can then turn off the water and drain the pipe from the turnoff to the outside spigot.
•Pipes that run up or along exterior walls need to be protected by having a professional insulation company insulate the entire wall. A split or frozen pipe inside a wall during a cold snap is guaranteed to cost you pain and big bucks.
In the event you do get a frozen pipe, warm up the frozen section with a hair dryer or heat gun. Have someone stand near the water turn off in case the pipe has a split and you need to cut off the water fast. If you're not sure exactly where the frozen spot is, feel along the pipe until you come to a cold spot and heat that section. If the water stars flowing, you've guessed right. Of course, it's best not to let pipes freeze, because weakened joints or sections can and do develop leaks at the most inopportune time.
### Winterizing the Pipes
Winterizing or draining the pipes is critical if you're closing up a vacation home for the winter or planning to be gone from your home for more than a couple of weeks during cold weather. Leaving a home unattended for very long can have serious results if there's a power failure and the furnace shuts down. Some winterizing tips are:
•Drain the water at the lowest point in the system or as close to the main shutoff valve or pump as possible after you've turned off the water where it enters the house or at the street turnoff.
•Turn off the water heater. For gas units, turn off the main gas shutoff on the meter. Electric units can be disconnected at the fuse or breaker box. Drain the tank or the holding tank on a well.
•Open all faucets indoors and out, and empty the toilet tanks of all water. Since there's water left in the toilet S traps, pour a couple of cups of automotive antifreeze into the bowl. Make sure some of the antifreeze spills or flushes into the S trap.
•Outdoor spigots may have knobs that when loosened allow trapped water to drain.
•Pour a cup of antifreeze into all the drains in the home. Don't forget the less obvious ones, such as floor drains or washing machine drain.
Some homeowners like to hire a professional to blow out all the drains with air pressure so there's no water left in the system to freeze. This is the best solution if the home is going to be left unattended for several months of cold weather.
### Water Hammer
Water hammer is a common annoyance homeowners experience when water is turned off by a dishwasher, washer, or even a spigot. This loud noise is caused by a lack of air chambers to cushion water under pressure when it's turned off.
Air chambers are short extensions (12 to 18 inches) of capped pipes that are installed in the water line. These extensions contain an air bubble that cushions water in the pipe when it's abruptly turned off. If there's no cushion, flowing water can recoil with a loud bang when you turn it off abruptly. Look at your water pipes and see if you have these extensions.
Some older homes don't have air cushions, but they can be added by cutting the line and fitting a T joint with the capped extension. Unless you're handy with a propane torch, this is a project for which you may want to hire a plumber. Also, your water system may have the extensions but one or more pipes may have lost their air bubble.
If you have this problem, try turning off the water at the main cutoff and allowing the system to drain. Hopefully, the water trapped in the air chambers will drain too. Turn the water back on, and as the pipes fill, air in the extensions should compress to restore the cushion.
### Clogged Pipes and Drains
Clogged pipes and drains are the bane for many homeowners, and it can get expensive calling a plumber every time you have a clog. Luckily, there are several things you can do to solve these problems.
The first and most important thing is prevention. Sink drains are designed to handle liquids. But when hair, soap particles, food scraps, and greasy products are carelessly washed into the drain, they often accumulate until they form an obstruction. When this happens you can try a chemical additive that may dissolve the clog, but if that fails, more aggressive action is needed. Some suggestions are:
•Start with a plunger. Block all openings, such as overflow drains or the second drain hole of a double sink, with a rag. (If a dishwasher drains into the sink, disconnect and block that drain as well.)
•Spread a thin layer of petroleum jelly on the rim of the plunger for a better seal.
•Fill the basin with about three inches of water and place the plunger's rubber cup over the drain.
•Push down on the plunger with forceful downward strokes to create a pulse of water that hopefully will break up the clog.
•If the drain clears, run hot water down it for about a minute to clear out all the residue.
Suppose you plunge away for twenty or so strokes and nothing happens, what then? The next step is to remove the U trap under the sink, because these can be notorious clog sites that defy a plunger.
To remove the trap, slip pliers are the tool of choice. Put a bucket under the trap and unscrew the nuts a few turns until the trap slips off the drain tube.
Clean the trap out (a wire coat hanger works great) and remove all the grime with a stiff brush. Heavy deposits sometimes build up on the inside of U traps (especially metal ones) and create rough spots that trap hair, soap scum, and other goop. When this happens, it's better to replace it with a new PVC trap. Take the old one to a plumbing supply and match it up.
_Figure 2-12_.
When you take the U trap off and find that it's not the culprit, the clog is probably further into the waste pipe. You'll then need to escalate your attack with an auger (see Figure 2-12). Augers are inexpensive and available at home centers or plumbing supply houses. If you have frequent clogs, an auger that attaches to an electric drill is a good investment.
If the auger fails to restore water flow, the clog is more serious and you'll need to find your credit card and call a professional. With older homes especially, buildup over the years can create choke points where clogs often occur. If your home has these, it's a good idea to replace that pipe section or run an auger through it regularly to prevent problems.
### Toilet Clogs
Toilet clogs can be a real pain, especially when small children have tossed things into the bowl. Some suggestions are:
•First try using a plunger with an extended rubber flange that creates a good seal with the trap.
•Empty the bowl of water and use a small mirror and flashlight to see if a toy, hair spray cap, or other item is lodged in the trap.
•If there's nothing caught in the trap, use a plumber's snake or closet auger (flexible auger) to try and remove the clog.
•Still clogged? Then more drastic action is called for. Remove the toilet bowl and turn it upside down. (How to remove the bowl is explained earlier in the chapter) Yes, you'll feel foolish when a toy or plastic cap falls out. A proactive approach to prevent this is keep the lid down so items don't get knocked into the bowl.
Sometimes the problem is in the pipe between your waste system and the main sewer line in the street or the septic tank in the back yard, and replacing it is expensive. In this case, hiring a plumber to clean that section every so often is the most economical way to go.
Because tree roots that invade the waste line as it runs from your home to the main line in the street can create a reoccurring problem, it's a good idea to identify the route your waste line takes across your yard to the main sewer connection and not plant trees near that route.
On the day she moved into her new home, one homeowner had a backup in the basement caused by roots having invaded the line and creating a bottleneck. The offending tree in the front yard had been cut down a few years prior, but the roots that had infiltrated the line had remained, partially blocking it. No one suspected the problem existed until the new homeowner turned on all the spigots in a bathroom at once to flush out the drain. Normally, this shouldn't cause a problem, but in this case the partially blocked sewer pipe backed up, filling the finished basement with a foot of smelly water.
Luckily, the new owner had a good homeowner's policy that covered the cost for a clean-up company to repair the damage.
### UPGRADING YOUR PLUMBING SYSTEM
If you have an older home with galvanized or copper pipes, and you want to upgrade or add on, there's some good news. Plastic flexible and rigid pipe, along with simpler ways to attach fittings, make upgrading easier and cheaper than ever.
For example, one homeowner who upgraded the plumbing on his 1971 ranch opted to go with a manifold system. With this type of system, water from the city or pump supply is connected to a manifold, or "switch box." From the manifold, flexible water supply lines branch out to the various faucets, toilets and appliances (see Figure 2-13). Each line has a shutoff; red shut-offs are hot water lines and blue shutoffs are cold. The system has the look of an electrical breaker box.
This brings up an obvious question: If you spend the bucks to upgrade your plumbing, what's the payback when you sell? Of course the real estate market is not the same in all areas, but this is one upgrade that should give you an above average return.
If the upgrade is part of a well-thought-out remodel, it should return above 90 percent in most areas. There's also the indirect return that if the plumbing is not upgraded, the home will sell for less and stay on the market longer. The bottom line is that in most cases, upgraded electrical and plumbing systems will pay back the investment and often more.
_Figure 2-13_.
Also, keeping the plumbing system in good working order is critical to maintaining your home's value. In one case, a couple made a full-price offer on a home because it was in an area and of a style they wanted badly. When the home inspection report came back a few days later, the inspector noted that a downstairs bath didn't work. The previous winter a water pipe that supplied the downstairs bath had frozen and created a minor flood in the finished basement. The homeowner's insurance paid to fix the water damage and replace the carpeting, but the sellers capped the pipe rather than replace it.
There were also a few other deferred maintenance items that the buyers would have ignored, but the unresolved water supply pipe problem created in the buyer's minds a what-other-problems-have-the-owners-ignored thinking mode. Since the offer was subject to a satisfactory inspection, the buyers decided to back out of the deal.
Plumbing is an emotional thing with many buyers, more so than carpeting, paint, décor, and landscaping. When they find something leaking or not working, overall credibility suffers.
In the above case, it would have cost the homeowners less than $150 to replace the pipe. Instead it cost them thousands of dollars because it was weeks before another offer came in, and it was much less than the first. By then, time had run out for the sellers; they had to move and were forced to take the lower offer.
To keep your plumbing in good working order, use the Troubleshooting Guide in Table 2-1 to help you solve most messy problems.
_Table 2-1. Plumbing troubleshooting guide_.
**Problem** | **Possible Solutions**
---|---
Faucet drips into the sink. | The seals or cartridge inserts need replacing. Remove and take to a plumbing supply for a match.
Water leaks around the faucet base. | Faucet water supply nuts can be loose, the seal between faucet and sink may need replacing. Inside seals leak.
Low water flow from faucet. | If the faucet has a aerator, make sure the screen is not full of sediment. Check water turn-on-valve and lines from water supply to house for leak.
Water leaks around screw on fittings. | Turn off water supply and make sure threads are coated with pipe joint compound wrap a few turns of Teflon tape on threads and reinstall.
Water leaks around copper fitting. | Turn off water and drain pipe. Use propane torch to heat the joint and separate pipes. Clean ends with abrasive cloth, coat with flux, slide pipe into fitting and heat while letting solder melt and run into the joint.
Water leaks from the bottom of water heater. | Check drain spigot to make sure it's closed. Look at the bottom of heater with mirror and flashlight for corrosion. If the heater is 8 years old or older you probably need to replace it.
Toilet sweats from cold water in tank. | 1. Get inside tank insulation kit for your toilet model.
2. Best long-term solution is having a plumber install a mixing valve that heats tank water to room temperature.
Water leaks into toilet bowl. | Flush valve seat is dirty or worn out. Rubber flush valve is not sitting directly on valve seat and needs replacing. Float may need adjusting to lower water level so water doesn't flow down overflow tube.
Water leaks from under the toilet bowl. | The seal between toilet and waste pipe needs to be replaced. This is either a wax or rubber seal. Remove toilet from bolts and replace.
Toilet bowl is loose. | Remove the plastic caps that cover the mounting bolts on both side of the bowl. Tighten nuts with wrench (usually 7/16) until snug, but don't overtighten.
U trap under sink leaks. | Try tightening locknuts first (hand tight only). You may have to remove trap and replace the washer that the locknut fits over to create a seal. Do not tighten these locknuts with a wrench.
Water leaks from the tank. | Bolts that hold tank to bowl may be corroded and need replacing, or rubber seal (spud washer) may have failed. Drain tank and check. Also check connection where water supply line connects to the water intake valve.
Water from tank continues to run into bowl after flush. | The rubber flapper or flush valve is not seating. Make sure it's positioned above the outlet. Replace if needed.
Toilet is clogged. | First try a plunger. If that doesn't work, try a plumber's snake or auger. Remove water from tank and check with mirror and flashlight for object that may be blocking opening. Finally, if all else fails remove the toilet, turn it upside down, and clean outlet. From now on keep lid closed to prevent objects from falling in bowl.
Water is slow to drain. | Check vent pipes on roof for obstructions that cut off air supply if you can't locate any clogs in the pipe.
Waterlogged soil in the front yard. | You've probably got a leak in the waterline from the street to your home. Call your water company to locate the line. You may have to hire a contractor to replace the water line.
## CHAPTER 3
## HOW TO CREATE GREAT-LOOKING WALLS AND CEILINGS
Your home is one of your biggest investments, and it's important to keep it in top condition and protect it so that its value grows. This chapter shows you how easy and fun it is doing those projects that not only enhance the value of your home, but which make it uniquely yours.
Some of the exciting and important things covered in this chapter are:
**How to maintain and repair walls and ceilings**
**Painting tips and techniques the pros use for outstanding results**
**How to pick out paint colors**
**Decorating and money-saving tips with relevant websites for more ideas**
**Worksheets at the end of the chapter that help you calculate the amount of paint you'll need to complete your project**
### ALL ABOUT WALL CARE
Walls are important components of your home's interior. They not only divide space, but they say a lot about you and your tastes from the way you decorate.
As visitors enter your home the first thing they zero in on are the walls. If the walls are in good condition, tastefully decorated, and have harmonizing colors, the visitors form a positive first impression, and your decorating reputation skyrockets.
On the other hand, dings, nail holes, door knob holes, bad paint, and chipped corners all lower the value of your home and your morale. If the home is for sale, buyers will form a what-else-is-wrong-with-this-dump mentality, and that's not good for full-price offers.
If you're a new homeowner, you're probably excited about the unlimited decorating possibilities. Even more exciting, there's no landlord to tell you your security deposit is at risk if you act on your artistic impulses. But unless it's a new home, there are probably a few holes and other flaws you'll need to take care of before popping off that first paint can lid.
The next few sections show you how to take care of wall problems, along with painting and decorating techniques the pros use that'll make your home look its best with the least cash outlay.
### Anatomy of Walls
Interior walls are typically framed with 2 × 4 inch lumber. The vertical members, called studs, are spaced 16 inches apart, center to center (see Figure 3-1). In fact, this is such a universal standard that just about all tape measures have the 16-inch position clearly marked (see Figure 3-2). Sometimes homeowners or unscrupulous contractors try to cut corners and space studs 24 inches apart, but this creates an unsafe and less rigid wall.
_Figure 3-1_.
Basically, walls come in two types: drywall (sometimes called sheetrock or wallboard) and plaster. In some higher-end homes, plaster can still be the wall of choice, but it takes a skilled crew to install it. You rarely find plaster walls in homes built after sheetrock came along in the 1950s. Sheetrock is much easier, faster, and cheaper to install. It commonly comes in 1/2-or 5/8-inch thick, 4 foot by 8 foot or larger sheets, and installers use battery-powered driver/drills and coarse-threaded drywall screws.
Unfortunately, years ago (during the dark ages of home building) nails were commonly used to hang sheetrock, which over time tended to pop out, creating small, unsightly dimples in walls and ceilings. How to fix these dimples is covered later in this chapter.
_Figure 3-2_.
### Wall Care
One wall problem is caused by shelves and pictures. It's amazing how often nails are left sticking out of walls when people move out of a home and the new buyer is left with a lot of patch work.
In one case, the sellers had a lot of pictures on the walls, and the home looked great. However, when the buyer entered the home a few hours after closing, she found the walls a mess. Nails were sticking out everywhere, and there were several quarter-sized holes left by wallboard anchors that had been carelessly removed. Also, the previous owners had painted around some of the larger pictures, which when removed left large white rectangles on pale blue walls. Of course, the buyer was upset and angrily called her agent, but the sellers were long gone, and not much could be done about it.
WALL-MOUNTING INFORMATION WEBSITES
www.govart.com/faq.html
www.ab.seweb.uci.edu/FIMA-Picture_safety.pdf
www.ehow.com/how_9936_hang-shelf.html
When mounting pictures, shelves, and other items, it's better to tap wall anchors into a stud rather than into drywall. True, there are wall anchors that work in drywall, but the load limits are small, and if one pulls loose, you've got a sizeable hole to patch.
As for hanging pictures, there are many different choices at your local home center. The key to picking one picture hanging system over another is to think ahead to when you'll need to remove whatever you're hanging. Bigger holes mean bigger patch jobs.
One of the first things many new homeowners want to do before or soon after they move in is paint the interior. Although they have high hopes for great looking walls, too often the results are disappointing. This is usually the result of poor prep work. In many cases, painting can make flaws even more noticeable. The first step in creating a great looking interior is prep the walls by getting rid of dings, holes, popped sheetrock fasteners, settling cracks and uneven spots.
Actually, prepping walls and repairing holes is an important homeowner skill to develop. It's not difficult, it's just a matter of attention to detail and a little practice and your home will always be in top condition.
### Fixing Holes, Dings, and Other Repairs
The first step to becoming a drywall ding expert is assembling a tool kit, as illustrated in Figure 3-3. Tools you'll need are:
•6-inch taping or spackle knives
•Utility knife and spare blades
•Assorted grit drywall sand paper and sanding block
•Drywall screws
•A roll of fiberglass mesh drywall joint tape
•An electric drill with Phillips bit for setting drywall screws
•Premixed all-purpose joint compound (referred to as mud by installers)
_Figure 3-3_.
•Straight edge
•Claw hammer
A common problem with wallboard in walls and ceilings is that as the house settles, seams open and fasteners pop out just enough to create a noticeable bump. You'll want to correct these bumps before painting or wallpapering.
To repair popped fasteners, first hammer the nail lightly until the head is just below the surface but not breaking the paper wallboard cover. If the fastener is a screw, give it a few turns with a Phillips screwdriver until it's also just below the surface, as shown in Figure 3-4.
Next, drive a drywall screw about 2 inches above or below the popped fastener to keep it from popping out again. If it's loose and the area around it is damaged, remove it. Then cut a length of fiberglass mesh tape large enough to cover both dimples.
HINT
If you're using a cordless drill to set drywall screws, practice on a piece of scrap until you know which setting will set screws just barely below the surface yet doesn't break the paper cover.
Fill the dimples with joint compound and run the flat edge of the trowel over the area to even out the fill. Let the first coat dry and then add another coat. When that coat dries, sand lightly and feather or brush the edges so they blend into the surrounding surface.
_Figure 3-4_.
To make sure a patched area is completely feathered, hold a bright light next to the repair and look at it from the side. If you see a bump, sand lightly until the surface blends perfectly with the rest of the area. Feathering takes patience and a little practice, but this is critical to ensure that the repaired area doesn't stand out when you paint it.
Small cracks and nail holes can be filled with one or more coats of joint compound as needed and then sanded and feathered. Holes from large nails, drywall anchors and dings can be covered with a piece of mesh tape and two or three coats of compound—and then sanded and feathered, of course.
PATCHING WALLBOARD WEBSITES
www.drywallschool.com/bigpatch.htm
www.thriftyfun.com/tf551724.tip.html
www.wisegeek.com/what-is-drywall.htm
www.alsnetbiz.com/homeimprovement
Bigger holes from doorknobs hitting the wall, moving furniture, and rough-housing kids needs a more aggressive approach. Crisscross the hole with layers of mesh tape and then force joint compound into the mesh with a six-inch trowel and scrape off the excess. Add two more coats and when dry, sand and feather to match the surrounding area.
Another approach on especially large holes, is to cut out the wallboard until you expose half of the stud on both sides of the hole. Cut a piece of wallboard to fit the hole. Secure the patch by driving wallboard screws at each end into the studs; space the screws about two inches apart. Then tape, mud, and sand the seams to blend.
_Figure 3-5_.
Since you're going to have a bare section of wallboard, you'll need to brush off or vacuum the sanded area to remove dust. Then prime and paint to match the rest of the wall.
Still another hole-patching solution that doesn't require you to cut all the way to the studs on either side is to insert 1-inch by 3-inch furring strips into the hole. The wood strips will extend about two inches beyond the edge of the sheetrock. Drive screws through the sheetrock into the wood strips, giving you a base to pack the hole with joint compound. After the packing dries, cover with mesh tape and add two or three more finishing coats. When dry, sand and feather to match surrounding wall (see Figure 3-5).
DECORATING TIP
Flat paints are more forgiving of wall imperfections but are harder to maintain. Eggshell, satin, and gloss paints are easier to wash but tend to show seams from careless taping and sanding.
Cracks should be filled with mud and sanded. Sometimes it's easier to cut the crack into a V with a utility knife, and then to fill with compound and sand smooth when it dries.
If you've got a really bad wall—with patches, gouges, bad mud, or poor taping—you may want to consider hiring a professional to skim the wall. Basically, the installer applies a thin, even coat of mud over the entire wall. After drying, the wall is sanded smooth and a coat of primer-sealer is applied.
For older homes with plaster walls, fixing a hole is not that difficult. It's a matter of applying several coats of fill, and then waiting for them to dry before applying a final coat of joint compound. The following steps show how:
1.Scrape and chip all loose material from around the edge of the hole and vacuum.
2.Undercut the edge of the plaster so it bevels inward towards the metal or wood strips (called lath) that form the backing. This is to create a seat for the patch.
3.Fill the hole with patching plaster available at home centers. Scrape off any excess and score the patch in a crisscross pattern so the final coat will adhere better and then let dry. The final step is to spread joint compound over the patch and smooth with the flat edge of trowel. Let dry and apply additional coats until the dried patch is roughly even with the rest of the wall. Sand and feather the patch until it's smooth and blends in.
ENVIRONMENTAL CAUTION
Before removing or working with old plaster walls, have a sample tested for asbestos. If the test is positive, contact a licensed abatement contractor for removal and disposal. Because asbestos is a hazardous material, you don't want to breathe in contaminated dust.
Sometimes if the plaster walls are in bad shape, it's easier to hire a professional to install drywall over the plaster.
### Water Damage Problems
In one case the homeowners were suddenly awakened by the sound of a huge crash coming from the back of house. Running towards the sound and flipping on lights as they went, the owners came upon a huge mess in the kitchen's dining area. Most of the ceiling had given way and soggy chunks of wallboard covered the table and floor.
For months a yellow stain had spread slowly across the ceiling. Water, which leaked in from a carelessly installed evaporative cooler mounted on the roof, had saturated the ceiling wallboard until it gave way. Luckily, it was in the middle of the night and not during mealtime. What would have been a $10 repair job had now escalated into several thousands of dollars in restoration costs.
Water damage to drywall happens most often to ceilings. A leaking roof or leaking plumbing in upstairs bathrooms are common culprits. Water runs along rafters, or the top surface of the drywall ceiling, until it comes to a seam or corner. Eventually, the water will run down the wall or corner onto the floor.
The first step, of course, is to find the leak and fix it. After that's done, peel away the insulation from the drywall so it can dry out. If the water damage isn't too bad, the drywall may not have to be replaced, although it may have to be refastened and/or retaped and mudded in a few places.
Next, seal the stained areas with an oil-based, stain-covering paint. After the paint is dry, patch with tape and two to three coats of mud. Sand and feather the seams. If no water stains have bled through the covering paint, apply the ceiling top coats.
Another solution is to texture the ceiling with a spray-on or hand-applied finish. These finishes are not only good at hiding minor problems, but create attractive ceilings. There's a wide variety of finishes and patterns available. However, it's probably best to hire a professional to do the texturing; it can be messy.
ENVIRONMENTAL CAUTION
Older "cottage cheese" ceilings sprayed on before 1980 may contain asbestos. The only way to find out for sure is have a lab test. Usually these ceilings are not hazardous to those living in the home. However, there may be some hazard if you remove the finish by scraping it off dry. If the lab test is positive for asbestos, you may want to have a professional remove it especially if all the rooms in your home have that type of ceiling.
After ceilings and walls are in good condition the next step to making your home stand out is a super paint job. Painting is not difficult, and if you can spell your name and write cursive, you can do a top-notch paint job. The next section shows how.
### HOW TO CREATE A GREAT PAINT JOB
The joy of living in a well-decorated home, that feeling of pride when company comments on how nice your home looks, and, of course, the boost a great interior gives your market value—these are great reasons to learn how to do a super job with brush and roller.
Painting can be fun and rewarding, especially when it's all done and you step back and see outstanding results. Many people who say they hate painting and don't want anything to do with it may ultimately admit that they once tried to paint a room and it turned out terrible.
Fortunately, terrible paint jobs don't have to happen, and they won't if you learn and apply the easy steps to follow. Admittedly, there's some grunt work involved, but the end result is more than worth the effort.
Actually, a great paint job is one of the easiest and best things you can do to increase the value of your home. If you're thinking of selling in the near future, you'll want to get started on the painting about ninety days prior to planting a For Sale sign, so it doesn't turn into a rush job.
However, if you've just bought a home and are looking at dismal or uninspiring contractor-white walls, this section will show you how to change your environment for the better.
First, plan on about two days per room. No, you can't knock out painting the house over the weekend; not even the Labor Day weekend. Even though slumlords and painters with power sprayers can coat a new house in a day, that's not for you. So allocate the time to do it right, which means a room at a time.
Basically, painting a room consists of three parts:
1. _Prepping the Room_. This encompasses getting the drywall in perfect condition, because a paint job can be no better than the prep work. Remove or mask all the fixtures, doorknobs, and whatever you can't remove from the room.
2. _Priming the Walls_. A great-looking top coat starts with a great primer coat.
3. _Applying the Top Coat_. This is the main event. Using good technique, tools, and paint will guarantee you a lot of compliments at your next party.
### The Prep Work
To begin, move everything out of the room, even large furniture. Prep dust, spackle, and paint have little homing devices that target anything left in the room, even when covered in plastic. Also remove doors, light fixtures, and hardware, and then use blue painter's masking tape to mask everything that's left. Table 3-1 at the end of this chapter is a checklist you can use to help you make sure you have the proper equipment on hand as you begin your project.
HOW-TO PAINTING WEBSITES
www.alsnetbiz.com/homeimprovement/painting.html
www.paintingyourhouse.info
www.house-painting-info.com
www.behr.com/behrx/expert
www.thisoldhousepaints.com
www.ppg.com/ppgaf/pittsburgh/ptechniques.htm
Cover the floor with good canvas or paper-backed plastic drop cloths. Avoid old bed sheets or other thin fabrics because paint will easily soak through them. Plastic sheeting provides a good cover, but it is slippery, and spilled paint makes it an even more unsafe working surface.
The next step is to inspect the wall for any holes, cracks, dings, or other problems. Even if the wall looks good, shine a strong light—an old lamp with a bare bulb works great—next to the wall and you'll quickly notice any imperfections.
You can repair damaged woodwork or molding with painter's putty or a wood filler, such as Minwax's High Performance Filler. Both products sand to a smooth finish.
Next, follow these tips to put your walls in painting shape:
•Lightly sand the entire wall to make sure there are no glossy spots.
•Using a soft bristle brush, go over the wall to remove any sanding dust.
•Wash the walls with Trisodium Phosphate (TSP) available at home centers. If smokers live or have lived in the home, use 50/50 bleach and water solution. Use this solution also if you live in a mildew problem area to kill any spores hanging around.
•Unless you're great with a cutting brush, use blue painter's tape to tape the molding and other surfaces you don't want painted. To prevent paint from bleeding behind the tape, run a putty knife along the edge to seal it.
•At any home center or paint store, buy an inexpensive caulking gun and several tubes of latex paintable caulk. Make a diagonal cut to remove about a quarter of an inch from the clear plastic tip of the tube latex—just enough to get a controllable bead. Run the wire attached to the gun down through the tip and puncture the cartridge's seal. Place the cartridge in the gun and squeeze the trigger a few times until caulk starts flowing (see Figure 3-6).
•Run a bead around window and door casings, floor molding and other gaps. Moisten a finger and smooth out the bead so it blends smoothly with the wall surface. If you have old caulking that's in bad shape, remove it with a putty knife and recaulk.
_Figure 3-6_.
### How to Pick the Primer
Primers are formulated to create a solid base, seal stains, and help bond the top coat to the wall. Both alkyd and latex primers give good coverage and do a great job. Which one you choose will depend on the top coat you have chosen.
Also, go with good quality, brand-name primers and paints, such as Sherwin-Williams, Pratt and Lambert, Benjamin Moore, or Pittsburgh Paints. Better yet, go to a professional paint supplier and talk to the knowledgeable employees. They can give you valuable tips and help you get the paint system (primer and top coat) that will work best in your climate and for the job you're doing.
Some important priming tips are:
•If the walls and ceilings have been heavily prepped and the first coat didn't quite do the job and there's a light bleed through, put on a second coat.
•Some climates, especially those close to salt water, need a second primer coat.
•Work closely with your paint supplier. Some primers are better for wetter rooms such as bathrooms and laundries. In other words, one primer doesn't necessarily fit all conditions.
•Older homes that were built without vapor barriers will need a primer that seals and prevents moisture from getting under the paint and causing peeling.
•Drywall and plaster walls are different and usually require different primers. Talk to your paint dealer for the best type to match your walls and conditions.
•Interior wood trim can also need different primers. Your paint dealer can also point you in the right direction in this area.
•If you've primed over unpainted drywall or wood trim, you'll probably need to lightly sand after the primer has dried. Primer tends to raise the fibers or grain so a light sanding is needed before painting the top coat. Be sure to run a tack cloth over the sanded area so no dust is left behind.
•Never spot prime problem areas on walls that you've had to go back and work on. Those areas will often show through. Correct the problems and then recoat the entire wall. Fortunately, ceilings are more forgiving and don't show spot priming as much. There are also special paints formulated for ceilings that cling to textured surfaces better than wall paints. Check with your paint supplier for what works best on your type of ceiling.
•If you're going with a darker top coat, tint the primer coat to match. It may save you having to apply a second top coat.
### How to Pick the Top Coat
A trip to the home center or paint supplier can be confusing. There are dozens of paint types, methods of application, and special effects to choose from. The following tips will help you sort out the confusion:
•Go with a high quality latex primer and top coat. That means you'll spend $20 to $35 per gallon. Top quality paint goes on better, lasts longer, and looks better than cheap discount paint. Consider it an investment in your home's value.
•Gloss and semi-gloss paints are more stain-resistant and washable, but highlight any wall problems. Flat paints are great for living and dining rooms. Eggshell gloss is a good all-around finish for hallways, kids' rooms, kitchens, and bathrooms.
•Some pros feel flat paint is the best way to go for ceilings, others like a bright white gloss that reflects more light into the room. If your home tends to be on the dark side, gloss ceilings reflect more light and makes the room feel larger.
•Plan on two top coats for the best looking job and figure on about 400 square feet per gallon. Also, make sure you have about a half gallon left over for touch-ups.
•If you live in a dry climate, you may want to include an additive to slow down drying time and make the paint more workable. Adding a few ounces per gallon of Floetrol (www.flood.com/Flood/Products/Interior/PaintAdditives) or another similar additive can make the job go easier.
•Check out mildew-inhibiting paints if you live in a humid area, and also for baths, kitchens, and laundry rooms. These paints will not kill mildew that's already present, but they'll prevent mildew from forming later on.
Worksheet 3-1 is a handy checklist to fill out and take to the paint supplier so you don't forget anything important, and Worksheet 3-2 helps you calculate how much paint you'll need for your project. Both worksheets are located at the end of this chapter for ease in making copies of the forms.
WEBSITES TO HELP YOU CHOOSE PAINT COLORS
www.interiordec.about.com
www.sherwin-williams.com/doityourself
www.paint.org/con_info/choosing.cfm
www.purecontemporary.com/FeatureArticle/article/103
www.architecture.about.com/cs/repairremodel/a/pickcolors.htm
www.home.ivillage.com/decorating/color/topics/0,,4tjs,00.html
Once you've decided on the paint system, the next step is to decide on the colors and tints. For some homeowners this is the project's fun part, for others it's divorce material. Luckily, there are six shortcuts that will help you decide:
1.Take your time and check out builder's open houses to get ideas of what's currently trendy in your area.
2.Look at color schemes in home magazines and check out the websites listed in the box.
3.Collect paint chips of colors you like and narrow it down to those that you think will go with your home's other colors.
4.Remember, color choice should take into consideration your home's architecture, ceiling height, and exterior and interior colors.
5.Once you've narrowed it down to three or four colors, get pint samples from the paint store and roll four foot swaths on one of the walls you're planning on painting. Live with it for a couple of days and see which color grabs you.
6.Don't be afraid of making a big mistake. What's the worst that can happen? You can always repaint!
Remember that although this is your home, it's also your biggest investment. At some point you will likely want to sell, either because you're moving up or relocating. The average American only stays in their home seven years. So keep this in mind when you paint. Don't follow the example of two homeowners who painted every room of their home a different bright color: blue, green, salmon, yellow, and pink! The homeowners said they wanted to express their artistic sides and felt confident others would feel the same way and their home would sell fast.
Reality, however was much different. The home sat on the market for months and not even a generous painting allowance stirred up interest. As one homebuyer said after looking at the home, "The owners must think I'm dumb to buy this home and put a huge amount of work into painting it when for the same price I can buy another home that hasn't been ruined."
In the end the home finally sold, but the owners lost many thousands of dollars that investing a few hundred dollars could have prevented. In another similar situation, the owners painted all the rooms in their home a bright fire-engine red. When they needed to sell, the result was almost predictable: The home stayed on the market a long time and ended up selling for a steep discount. Why did these homeowners paint all the rooms red? According to them, "We wanted to be cool and impress our friends. . . ."
The lesson learned here is that if you want to go over the top into bizarre color choices, it's going to eventually cost you. Either you'll have to repaint before selling, or you won't get top dollar. Good taste increases value, nonmainstream decorating will cost you unless you want to do a lot of redecorating when you decide to sell.
So, if you're not confident in your color choices, talk to the people at paint stores the pros use: They know the current trends. Other sources are magazines, builder open houses, and home fix-up programs on TV.
Now that you've got all those buckets of paint lined up on the paint store's counter, the next step is to mosey over to the painting tools section and get the hardware you'll need to do a pro job.
### How to Choose the Right Painting Tools
After spending a lot of time and energy prepping the walls and ceilings and buying a top quality paint, you don't want to skimp on the tools that make a super job happen. Many people who hate to paint and decorate developed that attitude because they once tried to paint with poor tools and cheap paint and got disappointing results.
Investing in good quality tools is less expensive in the long run than going with bargain brands that are thrown away after the job is done. Because good tools last for many paint jobs, the cost becomes much less when spread over several redecorating jobs. They also usually help speed up a job. A word of caution: Don't lend your good painting tools to your friends or relatives. You won't get them back!
The tools you'll need are:
•Drop cloths with canvas backing—they absorb paint drops and don't allow paint spills to soak through. Plastic sheeting works, but spilled paint soon makes it slippery and that's a safety concern.
•A couple of 5-gallon buckets. Since colors can vary slightly from gallon to gallon, it's a good idea to pour all the gallon cans into one of the buckets, mix thoroughly, and work from the 5-gallon container (see Figure 3-7).
•Several 2-quart buckets to pour small paint quantities into for brush work.
•A coarse screen that sits inside a 5-gallon bucket to roll excess paint off. Notice that Figure 3-6 shows no roller pan. That's because a shal low pan is easier to spill, easier to put your foot into, and a pain to keep refilling.
_Figure 3-7_.
•One telescoping pole for each painter. This allows you to do both walls and ceilings without ladders or boxes to stand on.
•Nine-inch rollers are the best all-around size. Get good quality 1/2-inch lamb's wool rollers for walls and ceilings. If you want to cover up some imperfections, a longer nap gives you a slight texture.
•Quality paint brushes are a must. If you must give up lunches for a month to get the best brushes, it's worth it. Get a couple of 2-inch or 2 1/2-inch brushes with angled soft nylon bristles (cutting brushes), and a couple of 4-inch brushes with straight bristles. Also look for tapered bristles with split ends that help work an edge and spread the paint smoothly. Plan on spending about $25 plus for a good brush that can last for a long time.
•Have lots of rags on hand for spills and dabbing up mistakes. Whatever you have around the house works.
Before you start slinging paint, you need to make a decision. Do you want to use a brush to paint the line between wall and ceiling or use masking tape? If you're a newbie to painting, you probably should use masking tape.
MAKE-LIFE-EASIER TIP
You can buy paint at K-Mart, Wal-Mart, and other discount stores, but you don't want your walls to look like blue-light specials, do you? So find out where the pros buy their paint. Look for vans with ladders on top taking up most of the parking places. The people selling paint at these stores know their product and can advise you. You can find them in the phone book under Paint Dealers.
Start with a roll of blue 2-inch-wide painter's tape and go around the room's perimeter with the top edge abutting the ceiling. After the perimeter is taped, run the edge of a spackle or taping knife along the top edge. This seals the tape to the wall so paint won't bleed under the edge.
With all the tools and paint sitting in the middle of an empty room and the wall masked, it's now showtime. The ceiling comes first, and those painting poles are going to earn their keep. However, when you do the brush work around the edges, you'll need a short stepladder or two milk crates and a stout plank.
### Painting Techniques: How the Pros Do It
In doing the ceiling, the brush work comes first. Fill a 1-quart or 2-quart container about half-full of ceiling paint to work from. Using a 3-inch or 4-inch brush, dip it about an inch or two into the paint, and tap both sides of the brush against the container side to remove the excess. Start at one corner and lay an even, smooth coat—just thick enough to cover—about 1½-inches wide to the other corner and then on around the room.
Feather the paint towards the middle of the ceiling so you don't get a smooth hatband effect around the ceiling edges. Some pros, however, prefer to roller the ceiling first and then use a brush to paint the half inch or where the roller can't get around the edges. If there's crown molding, do that next. The key is work from the top down.
That's all there is to it; after a while you'll develop a rhythm that makes the work go faster and smoother. As your brush talents develop you may find you can dispense with masking tape and cut a fine line just like the pros.
Once the ceiling's perimeter is painted, slip a roller cover onto the handle and screw it onto a painter's pole. With the 5-gallon bucket about half full of paint and the grid in place, dip about a quarter of the roller into the paint and run it over the grid to work paint into the nap. Cover the ceiling with broad, overlapping strokes, taking care not to leave any ridges of paint. You want to create a thin even coat. Don't try to load up paint to get a one coat coverage—two coats give better looking results.
Once the ceiling coats are dry, lightly run the edge of a drywall knife or spackle knife along the masking tape's paint edge. If you try to rip the tape off without doing this first, you'll pull some of the dried paint with it, creating a ragged edge.
DECORATING TIP
When you choose your paint color, match it to your floor colors, appliances, and furniture. The key to an outstanding home décor is picking a color that runs throughout the house and keying everything else to it. It's like matching up suit, shirt, and tie—or dress, shoes, and purse.
Next in line is crown molding. Tape the wall below the molding using blue painter's tape. A 2-inch or smaller brush works well for painting molding, and use a container that holds about a quart of paint. When the molding is dry remove the tape as before, scoring it along the paint edge before pulling it off.
Walls are next, but before you mask the ceiling make sure it's completely dry. You don't want ceiling paint sticking to the tape when you remove it. Also mask the base molding and anything else that you don't want paint on.
The pole extensions make painting walls a snap. No step ladders or boxes to get in the way, and you can stay far enough way from the wall to spot runs or other problems.
After the first primer coat is dry, give the wall a light test for any irregularities. Then lightly sand the wall with 320 grit sandpaper and remove any dust with a tack cloth. If the wall looks great, roll on the top coat and again give it the light test. Apply the final top coat and let dry before scoring and removing tape.
The last item to be painted in a room is the base molding. Again, you'll need to tape the wall to avoid ruining a great-looking wall.
Window sashes can be a challenge. You need to paint each one with a high degree of control, which using 1-inch brushes allows you to do. As for paint on the glass, you can either tape the panes or scrape off the paint when it dries with a single-edge blade. Doors are best removed to where you can paint or stain them while they rest on sawhorses.
Since painting projects can last for several days, you can put brushes and rollers into a plastic bag, which you can seal and store in the refrigerator between coats. At the end of the project, brushes should be cleaned thoroughly in warm soapy water and hung bristles down to dry. Wrap brushes in old newspaper and store until the next project calls. Roller covers are usually not worth cleaning and can be thrown away.
When painting or repairing your walls and ceilings, use the guide in Table 3-1 to help you solve most problems.
With wall and ceilings in top shape, many homeowners next cast an eye on upgrading their floors, counters, and kitchens. How to do that is coming up in Chapter 4.
_Table 3-1. Common wall and ceiling problems and how to solve them_.
**Problem** | **Possible Solution**
---|---
Plaster wall is in bad condition. | Have a professional install drywall over the plaster. After mudding, taping, and sanding the wall, you can paint it. This is often easier than trying patch plaster holes.
Drywall is in bad condition. | If it's just a few dings and holes, they can be patched. If that isn't practical, you can drywall over the existing wall and finish.
Door knob punched hole in wall. | 1. You can buy shock absorbers at home centers that mount over the hole and protect it.
2. Use drywall tape to crisscross the hole and then mud, tape, sand, and refinish.
3. Put a wood backing in the home and fill it will compound. Then sand and refinish.
Wall has a lot of shallow dings or an uneven surface. | Hire professional drywall finisher to skim the entire wall with joint compound and then sand it smooth. You may need an extra coat of primer to make the wall look super.
First coat of primer made the wall rough. | On new or poorly painted drywall, the first primer coat often raises fibers in the paper covering. Sand the wall with 220 grit sandpaper, vacuum well or use a tack cloth, and roll on a second coat of primer. That coat may need sanding also.
Prepping walls painted with gloss paint. | Home centers sell liquid deglossers, or you can sand the wall with 220 grit sandpaper and clean off dust with a tack cloth.
Pulling masking tape off the wall tears the paint and leaves a ragged edge. | Whenever you're applying masking tape, run a taping blade over the paint side edge to seal it tightly against the surface so paint won't bleed under it. When you remove tape, run the blade's edge lightly over the same edge to cut the paint-to-tape bridge or bond so the tape will lift of cleanly.
Getting rid of cottage cheese or popcorn ceilings. | If the texture is non-asbestos, you can it spray lightly with water and scrape it off with a wide putty knife. Cover the floor with plastic because there's going to be a big mess. After scraping the texture off, sand smooth and apply a primer coat or two before painting the top coat.
_Worksheet 3-1. Painting project equipment checklist_.
_Worksheet 3-2. Job quantity and cost estimator_.
**Directions for using the cost estimator table:**
Suppose you're painting the living room and dining room. The living room walls measure 18 feet × 8 feet for two walls (1 and 2), and 16 feet × 8 feet for the other two walls (3 and 4). There are two doors at 3 feet × 7 feet and two windows at 8 feet × 4 feet each.
Wall 1 = 18 feet × 8 feet = 144 square feet
Wall 2 = 18 feet × 8 feet = 144 square feet
Wall 3 = 16 feet × 8 feet = 128 square feet
Wall 4 = 16 feet × 8 feet = 128 square feet
Total = 544 square feet
Less square footage for doors and windows:
Doors: 3 feet × 7 feet × 2 doors = 42 square feet
Windows: 8 feet × 4 feet × 2 windows = 64 square feet
Total to be subtracted = 106 square feet
Square footage to be entered in table is 544 − 106 = 438 square feet.
Estimating 400 square feet coverage per gallon or 438 divided by 400 = about 2 gallons for two coats. By using this formula for each room, and filling in the table, you can estimate what it will cost you to paint the inside of your home.
## CHAPTER 4
## INTERIOR IMPROVEMENTS THAT ADD VALUE
One of the biggest and most perplexing challenges for new and even seasoned homeowners is deciding which improvements add value and which ones just end up costing them big bucks.
Basically, home upgrades and improvements can be loosely divided into two areas: those that are needed to maintain the home, and those that are elective and may not return your investment. Sometimes there's a blend, such as an upgrade that returns only a fraction of the cost but one you enjoy while living in the home, and that's payback enough.
Some examples of the first category would be repairing or replacing a roof, furnace, water heater, or carpeting. If you don't make these types of improvements, the home's value and your enjoyment of it suffers. This is why replacing a roof usually doesn't increase the value of your home. Buyers expect a good roof as part of the purchase price.
Elective upgrades are often area specific, in that what is routine in one area may not be so in another. A prime example is Phoenix or Los Angeles, where a swimming pool is almost standard; without one value suffers. In Grand Rapids, Minnesota, a pool would create a hard-to-sell property. There you would need to stay in the home for enough years to make a pool worth the loss when selling.
The key to upgrading and improving your home is knowing what to improve and what may not return your investment. That's the focus of this chapter, as well as the following:
**Tips on upgrading floors, kitchens, baths, and other rooms**
**How to shop for energy efficient appliances and upgrades**
**How to save money on your upgrades**
**How to hire and work with contractors and remodelers so you don't get taken.**
**Useful websites that point you to more specific information**
Many new homeowners start out expressing their decorating and artistic talent by painting their walls and ceilings, as was covered in Chapter 3. Once that's done, they often start thinking of upgrading floors. And like choosing paint, flooring options have exploded with exciting choices and new materials. The next section shows how to shop for and pick out what flooring will work best for your home.
### MAINTAINING AND UPGRADING FLOOR COVERINGS
Floor coverings are an important part of enjoying your home and maintaining its value. Some types are trendy and exciting, but when after a few years their owners tire of them, they end up in the landfill. Examples of these are avocado, gold, and red long-shag carpets and their matching appliances.
The key to picking flooring is go with a type that has stood the test of revolving-door fads and trends. Oak, for instance, is a classic. Other woods and finishes bubble to the surface for awhile and then fade away, but oak seems to keep coming back because few finishes can match its beauty.
MONEY-SAVING TIP
You can save a lot of money on tile floors if you take a class on tiling from adult education sources, home centers, or tile suppliers. Laying tile is easy and fun, and if you shop around you can find incredible bargains on discontinued patterns, styles, or overstocks.
Basically, floors are covered in either carpet, tile, vinyl, wood, or combinations. You get the best results when matching the flooring material to the room and environment.
### Flooring Options for Different Rooms
Starting with kitchens, tile is the number one choice. True, wood looks great in a kitchen and the wood laminates are water resistant. But nothing outwears tile, and you can save a lot of money installing it yourself. There's zillions of patterns, styles, colors, and textures, and they are continually changing. Even more exciting is that you can go with different grout colors to create a truly unique floor.
DECORATING TIP
If you just bought a used home and are planning on fixing it up, paint the house first, then match floor coverings to wall color, and finally choose the furniture, appliances, and accessories to compliment the wall colors.
You can also rent tile cutters and other tools cheaply from most home centers and tile suppliers. If you end up doing a lot of tiling, buying the tools is not that expensive, and it's a good investment because they last a long time. Just don't lend them to your brother-in-law.
Tile is also great for bathrooms, for the same reason it's great for kitchens: it's long wearing, waterproof, big-dog proof, kid proof, and so on. Other great places for tile are laundry rooms, entryways, and borders around family rooms. Developing a pattern or color and carrying that theme throughout all the rooms you tile helps tie your décor together.
Wood floors are so understandably popular that your options can get confusing. Table 4-1 gives some interesting comparisons:
One homeowner had a couple of large dogs that destroyed a vinyl floor and scratched and gouged a natural wood floor. He finally installed a good quality laminate. The dogs can't get traction and they spin out trying to corner on the slick finish, but after two years the floor still looks great. Other homeowners with active pets have also had good results with tile floors.
_Table 4-1. Popular wood-floor options_.
Along with tile and wood floors, carpet is a main-stream floor covering. There's a bewildering array of fiber, colors, and pattern choices. Carpeting, however, isn't for high traffic, wet, or hard use areas, such as entryways, kitchens, baths, and laundry rooms. It works best in bedrooms, living rooms, and family rooms. However, some homeowners like carpeting in hallways and on stairs for its noise-dampening ability.
When you go shopping for carpeting, look for these three characteristics that define quality:
1. _Pile_ —the height of the fibers
2. _Density_ —how many fibers there are per square inch
3. _Texture_ —how the fibers are looped, twisted, or cut
The more dense a carpet is, the longer it's likely to last. As for texture and pile, that tends to be a more individual choice. You run your hand through the samples and narrow it down to the ones you like the feel of.
Nylon is probably one of the best carpet choices, with costs from $12 to $24 per square yard. If you have allergies, wool may be a better choice, although it doesn't do as well in humid climates and its cost is around $25 to $80 a square yard. Still, wool is the fiber all others are compared to. If you have the climate and budget, wool is hard to beat.
HELPFUL FLOOR-COVERING WEBSITES
www.carpetguru.com/const200.htm
www.thisoldhouse.com/toh/knowhow/interiors/article/0,16417,229547,00.html
www.splise.com/carpet/category_2387.html
www.familyresource.com/lifestyles/home-and-garden/what-to-look-for-when-picking-out-a-carpet
www.duro-design.com
Polyester is the at the bottom for durability and cost. It's costs about $8 to $12 per yard. If you go with polyester carpet, get a better quality pad, which means thicker and denser than the usual pad paired with this type of carpet.
There are so many patterns and colors to choose from that it's easy to suffer from sensory overload, but here are some tips to help narrow it down.
•Paint a bold swab of your wall paint on a piece of white paper and take it with you carpet shopping (after it dries, of course).
•If you're leaning toward a patterned carpet, the dominant color should match the wall color.
•For solid colors, you can contrast, compliment, or match. Which way you go depends on your home's architecture, how light or dark it is, and your taste. Darker carpet makes a room look smaller and shows lint and pet hair more than lighter colors. Rooms with lots of windows and high ceilings can go with darker carpet better than a home that doesn't get much light.
•Light carpet colors tend to make rooms look larger, especially if the home is a ranch or bungalow. And if you have a blond Lab, it won't show the pet hair.
•Narrow your choices down and take samples home to see how they match in your house. Showrooms are likely to have different lighting from what your home has, so you need to see what the colors look like in your setting.
•Also keep in mind that you're going to be living with the carpet for a few years, and that classy plaid or bright purple may not be the rage for long. It's better to stay with classic colors and patterns that won't date your purchase.
Decorating your home can be a work-in-progress journey. In reality, that's what makes homeownership fun. You decorate it; if you don't like it you can redo it. You don't have to live with the landlord's decorating!
Once you have the walls and flooring to your liking, decorating the bath(s) to your expectations can do a lot to turn your house into a home. As one new homeowner said when the Realtor handed her the condo keys: "Once I get the bath fixed up the way I want, the rest will fall into place."
### MAINTAINING AND UPGRADING BATHS
For many homeowners who have older homes, the bath(s) is one of their biggest maintenance challenges. Leaks, clogs, water damage, and cleaning demand time and treasure to keep them "house guest" clean.
MONEY-SAVING TIP
It's often much cheaper to refinish older tubs than go with a tear-out and replace. Check the yellow pages for companies that refurbish tubs onsite. Two interesting websites on refinishing are:
www.bobvila.com/wwwboard/messages/48204.html
www.permaglaze.com
Chapter 2 covered how to handle common plumbing problems. This section will cover how to maintain the rest of the bathroom: cleaning, upgrades, floors, fixtures, and caulking.
One highly important bathroom maintenance item is caulking. Why is that so important? Here's six reasons why:
1.Bad caulking around tubs allows water to seep down the wall to the floor and damage both.
2.A shower pan-to-wall caulking failure can allow water to seep under the pan and cause floor rot and mold growth.
3.Caulking around sinks keeps water from damaging the counter, wall, and flooring.
4.Depending on the floor, caulking can prevent water from seeping down along the edges to the subfloor.
5.Water can leak through damaged caulking and cause damage to cabinets or flooring. You'll notice that all these reasons for caulking have one thing in common: keeping water from getting to the subfloor and costing you a lot of money.
6.Bad caulking makes a home look like a dump. If you're selling, it can make the difference between sale or no sale, full price or low offer.
### All About Bathroom Caulking
Check out the caulking section of any home improvement center and you'll be amazed at the many different caulks. The different formulations are designed for different applications. Latex caulk, for example, is designed for interior sealing and can be painted to match.
HOW-TO-CAULK WEBSITES
www.naturalhandyman.com/iip/infxtra/infcau.shtm
www.repair-home.com/how_to/caulk.htm
www.bathtubdoctor.com/services/caulk.php
www.ehow.com/how_114916_seal-wall-joints.html
Table 4-2 gives you a broad overview of the different types and their applications:
Most manufacturers put a lot of info on their tubes, so read the labels before you buy. For bathroom tubs and showers, pick a product that is designed for the type of surface you have. Many professionals use silicone caulk, and it does a good job, but the downside is you can't clean up spills or get it off your hands. Unless you're a professional with caulking skills, it's better to go with Polyseamseal Tub and Tile Caulk or similar latex, mildew-proof products. They work great and clean up with soap and water.
_Table 4-2. Caulk types and their uses_.
However, if your tub is fiberglass with ceramic tile walls or fiberglass sections with caulked joints, silicone is recommended.
To recaulk the seams in your bathroom, follow these six easy steps:
1.Patching caulked areas doesn't work too well, so it's better to remove all the caulk and redo the entire tub or shower. Remove all the old caulk, clean the area thoroughly, and then vacuum. Any loose material or dust can compromise your joint.
2.Degrease and clean the joint with alcohol, but if there's mildew present treat the joints with a fungicide and then thoroughly dry.
3.Use a hair dryer or small heater to dry the joint completely before caulking or it may not stick. You can also use a hair dryer to soften old caulking and make it easier to remove.
4.If you have only one or two bathrooms you may want to buy 5.5 oz. plastic tubes of caulking instead of a caulking gun and cartridges. There's less waste, and beginners often find tubes easier to work with.
5.Cut the tip of the cartridge or tube at a 45-degree angle and start at the far end of the joint and work toward you. Squeeze out a uniform bead of caulk into the joint as you move the tip along the joint. When you get to end, put your finger over the tip or insert a nail to stop the caulk from flowing.
6.The next step is to tool the joint, andthe best way to do that is with your finger. That can get messy, so you'll want to get a package of latex gloves—the kind your dentist, doctor, and car mechanic use—especially if you use silicone caulk. You can also use a putty knife with a rounded end, plastic spoon, wood tongue depressor, or butter knife. Go over the joint only once, don't dab at it but run your finger (dipping your gloved finger in water helps) along the joint, forcing the caulk into the gap. You want to create a smooth wedge of caulk over the joint. Yes, it's art work: The more you do it the better you get—just like painting. If you mess up a joint, clean it out and do it over; no big deal. Let the caulk cure for about 24 hours and clean up with a razor blade.
Areas in your bathroom that are especially important and you should inspect every couple of months are:
•Around bathtubs and floor-to-shower seams.
•Around sinks and where the vanity top meets the wall.
•Around the base of the tub or shower enclosure where it meets the floor.
•Along the baseboard to floor seam.
•Anywhere water could get through to the subfloor.
•Do not caulk around where the toilet bowl meets the floor. If the wax seal fails (see Chapter 2), you want to know about it fast so it doesn't damage the subfloor.
Although caulking can prevent water from seeping where it not supposed to, additional strategies are needed to keep bathroom mold under control—such as frequently checking under sinks, installing good venting systems that exhaust moist air to the outside, and installing watertight shower enclosures. Homebuyers tend to get uptight when they or their inspectors find bathroom mold.
### Bath Remodeling: What Pays, What Doesn't
There are two ways to measure the return on bathroom improvements or any home improvement for that matter:
1.If you don't do the upgrades, what will it cost you in lower offers and more days on the market. Home sellers usually lose more money this way than from too much improvement.
2.The second measurement is more direct. Suppose you spend $7,000 to remodel your bath. Will it increase the value of your home by the same amount? Looking at national statistics can be misleading because real estate values vary from neighborhood to neighborhood. The best way is to have a friendly Realtor run a list of what similar homes to yours have sold for with and without the improvements. Local appraisers also have a good feel for what improvements have good payback for your area.
Typically, the following bathroom improvements will more than pay for themselves:
•Replacing a beat up vanity with a pedestal sink or new vanity, counter-top, and sink.
•Replacing vinyl flooring with tile especially if you do the install.
•New tub or shower surround if it's consistent with your price range. In-other-words, tearing out your tub-shower combo and replacing it with a double-size imported marble shower-steam room wouldn't have a good payback. You want to keep upgrades consistent with what other homes in your price range have.
•An attractive light bar above sink. Good bathroom lighting is a must.
•If it's an older home, updating wiring and outlets with a GFI outlet.
•An attractive paint scheme that ties the bathroom into the rest of the home.
In addition to upgrading baths, adding a new bathroom is often a good investment that adds more to the value of your home than the cost. This is especially true if you have an older home with one bath and everyone else on your block has two baths.
Adding a bath when you finish off a basement, attic or part of the master bedroom is usually a good investment. Again, before spending big bucks, make sure the neighborhood will support the investment.
BATHROOM REMODELING WEBSITES
www.remodeling.hw.net/content/CvsV/CostvsValue.asp?articleID=211765§ionID=173
www.hgtvpro.com/hpro/di_bathrooms/article/0,2618,HPRO_20173_3727707,00.html
www.design.hgtv.com/bath/sub_topic_03.aspx?t=2&st=8
In addition to bathrooms, kitchens are at the top of the list of what buyers zero in on first. If these two meet their expectations, the home goes on their short list. Even if you're not planning on selling in the near future, these rooms are what family and friends see most. Kitchens are where everyone wants to be: that's where the action, the refrigerator, and the food are. It's no wonder then that kitchens are the most remodeled, upgraded, fretted over, and decorated areas in the home, and the most covered in TV shows and magazines.
The next section will cover kitchens and how to make them look their best along with how to save some serious bucks.
### HOW TO MAKE YOUR KITCHEN LOOK ITS BEST
Like bathrooms, how much of your remodeling dollar increases the value of your home depends on the area you live. That means you'll want to do some homework before writing checks at the nearest home center.
First step is go through homes for sale and open houses in your area to see what upgrades your neighborhood will support. For instance, are homeowners upgrading counters to Corian™, granite, concrete, laminate, or whatever is fashionable? How about cabinets, what's the current rage: oak, stressed aspen, or dark cherry?
A good way to find out what's current is go through builder's open houses. They spend a lot of money and time making their model homes appealing to the largest number of people.
Builder's decorators strive to create shock and awe with color and furnishings, so that when you walk back into your own home you're dissatisfied with your humble dwelling and start crunching numbers.
But what's exciting is that you can take ideas not only from open houses but from magazines, TV shows, and homes for sale—to personalize your home while adding to its value.
A good place to start making your kitchen look its best is the floors. As previously discussed, tile and wood laminates wear and look good. If you do the work yourself, you get a lot more bang for the buck. Most home improvements you can make yourself, and developing new skills will give you a lot of satisfaction.
MONEY-SAVING TIP
If you can hold a pencil and draw a reasonably straight line you can do a great tile job. Many home centers have free classes and rent the tile cutters and tools. Tiling your kitchen, bath, and entryway floors make your home look great, and it's the best way to add value without a lot of cost.
After floors, cabinets are key features of a kitchen, and there are a lot of options for making drab cabinets attractive. In fact, one home renovator with a twenty-year track record says she has never yet replaced kitchen cabinets on any of their rehabs. They've been able to refinish and rebuild cabinets so they're attractive. It must work because the homes they renovate sell quickly and for top price.
WEBSITES FOR MORE REFINISHING HOW-TO
www.finefurniturefinishing.com/newsletter10.html
www.inin.essortment.com/cabinetrefinish_rgqc.htm
www.lowes.com/lowes/lkn?action=howTo&p=HomeDecor/CabFceLft.html#4
www.thisoldhouse.com/toh/knowhow/kitchen/article/0,16417,202424,00.html
www.doityourself.com/stry/kitchencabinets
Here are some alternatives to laying out big money for all new kitchen cabinets: (this also applies to sinks):
•If there are not enough cabinets in your kitchen and you can't find additional matching ones, you can still keep the old cabinet frames. Replace the doors and hardware on the old cabinets to match the new ones; refinish both old and new to match. However, because the old and new wood frames may not match, painting both a gloss white or other light color is often the best way to go.
•Strip the cabinet's finish, sand, and refinish with new hardware. For example, if you have 1980s dark wood cabinets, you should be able to strip the finish, sand, and stain a lighter color. With new hardware, you wouldn't know it was the same cabinets. If the finish won't come off, you can always paint the wood a light gloss color.
•It's amazing how many complete sets of cabinets are discarded by remodelers. Many of these are in excellent condition or can be refinished. Check the yellow pages for salvage yards and second-hand outlets.
Check with cabinet shops for seconds, unfinished units, and cancelled orders.
•Look for discontinued lines and sales at home centers. It's amazing the deals you can get if you shop around.
•Sometimes a good cleaning, finish coat and new hardware can make your cabinets look new. Also consider adding molding on the doors to create a custom look.
•There are also companies who can glue a thin wood veneer or plastic laminate to cabinets that creates a new look for less than replacing them.
KITCHEN DECORATING TIP
Give your kitchen personality by varying the height and depth of the cabinets. For example, create a staggered line or put a taller cabinet over the stove. Glass doors on one or two units break up a monotonous line of identical doors. Adding crown molding and cabinet-to-wall trim also works well.
If you would rather start over and buy new cabinets, there are a lot of exciting options to choose from. Basically, there are three ways you can go:
1. _Stock cabinets_ are those you find at home centers ready for delivery. It's usually what-you-see-is-what-you-get. Typically you'll pay $50 to $200 per linear foot.
2.Buying _semi-custom cabinets_ is like buying a car: Take the base model and add what accessories you want from the options list. Cost typically runs from $200 to $600 per cabinet foot.
3. _Custom cabinets_ are made specially for your kitchen layout with the latest cutting edge design and options. Custom cabinet shops usually charge from about $500 per foot on up to what a Ferrari costs.
The following shopping tips can help you keep costs in suborbit:
•Consider going with oak or white laminate cabinets. These are less expensive than cherry, pine, or exotic woods.
•Go with better cabinets and cut costs elsewhere, such as countertops and appliances; replacing those later on could help stretch the budget.
•Consider going granite on an island and laminate on all the other counters.
•Choose veneered doors instead of solid wood.
•Keep existing appliances and refinish to match your new color décor.
Along with cabinets, countertops are a kitchen's focal point and help define the personality of a kitchen. There are lots of countertop options, each with its pros and cons. It boils down to how much abuse the counter is going to get and what surface you really like.
To get the most functionality and economy plus create a dramatic effect, don't overlook mixing styles and materials. For example, the island top could be granite, marble, wood or whatever works best for your lifestyle. If you're a dynamite pastry chef, the island could be marble and the counters another material, such as stainless steel or laminate. You can be as creative as you like: no more straight line of cabinets and a boring length of laminate counter from the kitchen dark ages.
Most popular countertop options are shown in Table 4-3.
An important touch to an impressive kitchen is lighting. As in countertops and flooring, there are a lot of options. If you have recessed lighting, new broad spectrum bulbs that approximate natural daylight can make a big difference.
For replacing light fixtures, consider a low-voltage cable-and-rail system. The rails are attached to the ceiling and can be bent into shapes. Two parallel cables are typically stretched from wall to wall, and light fixtures are attached to the wires. They can be moved anywhere along the cable.
In practice, you can have a couple of lights drop down over a island, dining table of other areas that need lighting. You may want to check out lighting supply stores that have working displays and help you tailor a system to fit your home.
MORE INFORMATION ON TRACK AND CABLE-AND-RAIL SYSTEMS
www.signonsandiego.com/uniontrib/20061022/news_1hs22ceiling.html
www.lightwavesconcept.com/home.php?cat=251&cm_re=Menu-_-LightRail-_-Systems
www.con-techlighting.com/catalogs/OdysseyLineVoltageRail.pd
Also pay special attention to the cutting-edge lighting in new-home open houses and home decorating magazines. Lighting is an inexpensive way (typically under $500) to add high tech to your kitchen.
_Table 4-3. Countertop materials: pros and cons_.
Of course, no kitchen can look better than its appliances, and sometimes it's a challenge trying to decide whether to keep or replace them. Typical life expectancy for kitchen appliances is:
Electric Range | 17 years
---|---
Gas Range | 19 years
Refrigerator | 14 to 17 years
Microwave | 11 years
Garbage Disposal | 10 years
Dishwasher | 10 years
Some appliances may last a lot longer, while others die the day after the warranty expires. In reality, there are several ways you can go when deciding on what to upgrade and what not to upgrade:
1.If the color or finish of the appliance isn't a problem, keep it and spend the funds on other upgrades; possibly replace it later on.
2.When the new kitchen remodel changed colors so that the appliances no longer fit the décor, you can save quite a few bucks by having a professional paint the appliances.
3.Bite the bullet and replace the appliances now. If your appliances are at or over the age range listed above, there's an additional consideration: energy efficiency. New appliances consume less power than older models.
ENERGY-SAVING TIP
When shopping for appliances, look for the EPA's ENERGY STAR label. Appliances with that label use up to 30 percent less power than those sold as recently as 2001. That means you can help out the environment and save money too, often enough to justify replacing older appliances.
With kitchen and baths looking good, the next area to consider is living and family rooms.
### LIVING AND FAMILY ROOMS THAT ADD VALUE
Living rooms have taken a back burner position the last few years, with emphasis more on family and great rooms. In new homes, living room size has often shrunk to where it's almost an afterthought. Still, they're important because visitors will pass through or by the living room on the way to the kitchen, and you want to create a favorable impression. This is especially important if you have an older home where the entryway opens into the living room.
Some living-room decorating suggestions are:
•Of course, a dynamite paint job goes a long way to making the room impressive. This is also a room you can be more expressive with. For example, one homeowner painted three walls a light beige and the wall facing the entry door a dark red. She then included that color in the furniture, accessories, and trim for an impressive effect. The owners could also have carried the dark red color throughout the house to tie the décor together.
•If the entryway door opens into the living room, you can tile or install wood flooring in a half circle or square in front of the door. This creates an attractive entryway. A light decorative screen can block the view into the kitchen or rest of the living room.
•Window treatments are an important part of a room's décor. Small rooms should avoid heavy drapes; go with light colored blinds or shutters instead. The more light a room gets the better it looks.
•Carpets, rugs, and furniture should be on the light side.
•Scale is also critical. In a smaller room, less is often more. Avoid oversize furniture, large pictures, over-grown plants, and anything that overwhelms the room. The super-sized sectional sofas, armchairs, and coffee tables look better in a larger family room.
Family rooms or great rooms have taken center stage from living rooms the past few years because they are close to or part of the kitchen, where the action (food) is. Many family rooms also have vaulted or cathedral ceilings that open many exciting decorating possibilities.
In fact, so important are family rooms that, according to _Remodeling_ magazine's annual "Cost vs. Value Report" for 2006, family room additions returned 113.3 percent, or $63,556, on an average cost of $56,111 to build.
Does this mean you can bump out a wall, tack a family room onto your kitchen, and get more than the cost back when you sell? Not always. Remember, real estate is local, so before you draw up plans, check out what other homes similar to yours have done in the area. Look at what homes have sold for that have added family rooms versus homes that haven't. Some areas can support all sorts of home improvements while other neighborhoods just a few blocks away can't.
Other less costly family-room options are finishing off a basement or attic. Here, too, you'll need to look at recent sales in the area to determine the payback. If you do add a family room, consider adding a bath at the same time: It's cheaper to add one when construction is underway than later. Also, another bath may add enough value to your home to justify the cost.
As for decorating family rooms, it's more relaxed. However, if the family room connects to the kitchen, coordinating the colors and trim adds to the flow.
### DON'T FORGET BEDROOMS AND LAUNDRY ROOMS
Unlike family rooms, visitors don't often see the bedrooms, and decorating them often takes a back seat. But there are some things you can and should do enhance your home's value:
•If possible use the same floor covering in all bedrooms. A common color or decorating theme ties all the rooms together.
•Closets are a big item with some people. If you have walk-in closets, organize them and make sure they're well lighted. Older homes with small closets will need more creativity. Decluttering and adding organizers are a good start. Also, painting them a bright or gloss white makes them appear larger. Adding lighting can also help.
•A bedroom light/fan combo can add a decorator touch. Match the fan size to the bedroom. For example, a large bedroom with vaulted ceilings can accommodate a large fan where a small bedroom would need a scaled-down model.
•Window treatments should also fit the room. Blinds and shutters make a room look larger than heavy drapes will.
•Some homeowners like to decorate the master bedroom with creative colors and patterned wall coverings. However, when you decide to sell, it's better to restore the décor back to vanilla neutral.
INTERESTING BEDROOM-DECORATING WEBSITES
www.interiordec.about.com/od/bedroomdecor
www.creativehomemaking.com/home_decorating/bedrooms.shtml
www.home.ivillage.com/decorating/bedrooms/0,,8sgq,00.html
www.homeandfamilynetwork.com/Decorating.html
www.decorate-redecorate.com/decorate-bedroom.html
www.realsimple.com/realsimple/package/0,21861,1013123,00.html
For many homeowners the laundry room is further down the upgrade list than the bedrooms. But in reality, the laundry room has come up the food chain of importance the last few years. To see how far, check out the laundry rooms in new home construction. Some rival small bedrooms in size.
LAUNDRY ROOM WEBSITES
www.interiordec.about.com/od/laundryroom1/
www.servicemagic.com/article.show.7-Essential-Laundry-Room-Solutions.8288.html
www.ehow.com/how_10820_jazz-laundry-room.html
If your laundry is in the basement, you probably have the space to enclose and make it an impressive room without paying a lot. Also, some basements have plumbing stubbed in for a bath, which can be used to create a half or three-quarter (with shower) bath/laundry room combo. This is an economical way to add a bath to your home's stats.
As covered in previous sections, developing tiling skills can pay big dividends. If you're looking for a place to start, this is it. A downstairs bath/laundry combo is just the right size for a first project.
***
The five biggest mistakes homeowners make when trying to improve their homes' interiors are listed in Table 4-4.
You could spend a lifetime upgrading and improving your home's interior. But to maintain value and keep your home looking its best you need to spend time on the exterior as well: it's what keeps the rain off your head. How to do that is detailed in Chapter 5.
_Table 4-4. The biggest interior-improvement mistakes homeowners make_.
**Mistake** | **Suggested Action**
---|---
Keeping old inefficient appliances and fixtures | Keeping old appliances is often not cost- or energy-wise. Replace with EPA's ENERGY STAR-rated appliances for maximum savings.
Not keeping the interior in good condition | It's cheaper to fix or upgrade items along the way. If you decide to sell, you won't have a huge get-ready-to-sell bill.
Not coordinating the home's décor | To make a home stand out, the colors and décor should flow from room to room. Pick a theme or accent color and carry it throughout the home.
Adding a room or addition that overvalues the home | Adding on to a home is a good idea if the neighborhood will support it. Have other homes in your area added on? If you do, will you get your investment back?
Not decorating to the home's style | Make sure improvements like door styles or cabinet styles go with your home's style. For example, you wouldn't want contemporary appliances in a Craftsman-style home.
## CHAPTER 5
## EXTERIOR MAINTENANCE AND IMPROVEMENTS THAT ADD VALUE
Maintaining and improving your home's exterior is especially important, because much of a home's value and pride of ownership comes from curb appeal. It's usually not expensive or difficult, but it does entail staying on top of potential problems to keep the home weather-tight.
Once a home's exterior components are compromised by damage or neglect, or are just worn out, water starts doing its thing. When that happens, the dollar-calibrated repair meter spins faster than a cabbie's mileage meter. It's much cheaper to head off problems before you feel raindrops falling on your head during dinner.
Unfortunately, too many homeowners wait until they want to sell before thinking about exterior maintenance, and in the rush it often doesn't get done. This is reflected by telling comments in their ads and MLS remarks, such as:
"Must see the inside . . ."
"Bigger than it looks . . ."
"Just needs a little TLC . . ."
"Roofing allowance . . ."
And so on, but you get the idea. Owners and agents hope to minimize the problems when in reality they are throwing away thousands of dollars needlessly. Buyers aren't dummies, and their pencils are sharp.
But there is good news in all this: If you can turn a screwdriver and tell your right hand from your left, you can do most of the repairs needed to keep your home's exterior in good condition. And as a bonus, you can improve and increase its value significantly.
Some of the critical items covered in this chapter are:
**How to make sure your home's outside components are watertight**
**How to upgrade your home's curb appeal**
**How to keep gutters draining**
**Wells and septic systems**
**How to keep concrete and asphalt looking good**
### PREVENTING WATER DAMAGE
Whenever you talk about a home's exterior you're really discussing the weather and how to keep it from getting into the home. A critical first line of defense against water infiltrating the home or causing exterior damage is the gutter or roof drainage system.
Some homes don't have gutters and rely on the roof extending beyond the wall to divert water a couple of feet out from the foundation, but this creates serious problems. If the ground doesn't have enough slope to drain water away from the foundation, water can damage it and can seep into the crawl space or basement.
Also, water spilling over the eaves can kick up mud onto the siding, which leaves an ugly stain that's not easy to remove. These stains and possible siding damage can lower your curb appeal and your home's value considerably. It's a red flag that tells all who see this home that it's not well taken care of.
### Rain Gutter Savvy
The first step in weatherproofing your home is install a gutter system or make sure the existing one is in good shape and is handling the runoff. If your home needs a new system, Table 5-1 lists some common options.
_Table 5-1. Common rain-gutter options_
A new vinyl or aluminum gutter system for a typical 2,500 sq. ft. home would cost approximately $700 to $1,200 installed. In addition, you would want to add inexpensive extensions to the down spouts to route water at least six feet from the foundation. The costs involved here are much less than replacing siding or repairing interior damage from rain water entering through a window well. Figure 5-1 shows how a slip-on extension can route water safely onto the lawn and away from foundations and window wells. For more a permanent low profile, the extension in Figure 5-2 runs underground.
HELPFUL RAIN-GUTTER WEBSITES
www.doityourself.com/stry/vinylraingutters
www.hometips.com/hyhw/shell/128gut.html
www.thisoldhouse.com/toh/knowhow/repair/article/0,16417,487931,00.html
If you have a gutter system already installed, inspect it twice a year (fall and spring) to make sure it's working. Use the maintenance checklist in Table 5-2.
Yes, cleaning out gutters twice a year is a pain, but beware of gutter-cleaning silver bullets that promise to eliminate it. "There ain't no shortcuts, unfortunately," according to this author / homeowner / gutter cleaner who lives in the middle of a grove of deciduous trees and is still looking for a quick and easy solution. However, there are homeowners who have had good results with the "Waterfall"-type gutter guards, and they may be worth checking out for your area.
_Figure 5-1_.
Once your gutter system is fine-tuned and able to divert roof run-off away from the house, it's important to take a quick look at the landscaping. Does the landscaping divert water away from the home? If it does, you're in good shape. But if it doesn't, you'll need to move some dirt around until it does.
MONEY-SAVING TIP
Many homeowners find out too late that run-off water from outside the home is not covered by homeowners insurance. For this kind of coverage you'll need additional Federal flood insurance available through most insurance companies.
_Table 5-2. Gutter maintenance checklist_.
**Job** | **Action needed**
---|---
Clean out the muck. | Remove any screens or debris guards and clean out the sludge with a plastic scoop.
Remove the elbows. | Use a drill to remove the elbow fasteners. If they're attached with rivets, use a 1/8-inch bit to drill them out. Clean out the goop.
Scrub the gutters. | Using a stiff brush clean out the gutters and elbows and flush with water.
Replace elbows. | Reattach the elbows with 3/8-inch hex-head metal screws (zip screws). Replacing the rivets with hex-heads makes the twice-yearly cleaning much easier.
Check straps. | Downspouts have metal straps at each end. If they're loose, reattach or replace if needed.
Channel the water. | Put 4-foot to 5-foot extensions on the downspouts to channel water away from the foundation and window wells (see Figure 5-1).
Fix sagging gutters. | Probably a nail has come loose letting that section sag. Use a hammer and pry bar to remove the nails and clips. Replace the nails and clips or hangers. Spacing should be about 24 inches. Roof-mounted gutters are fixed by replacing the straps anchored to the roof with roofing nails.
Prevent water from damaging fascia. | Slide a length of aluminum drip edge (available at home centers) under the roofing or shingles and over the gutter's back edge. Lift the shingles and nail the drip edge to roof deck (not the shingles) with roofing nails every 24 inches.
Recaulk leaking seams. | Use a heat gun to make sure the surface is perfectly dry. Apply a generous amount of aluminum joint sealer or butyl caulking evenly over the seam. Use latex gloves; these sealants are hard to get off hands.
If you've just bought the home, it's a good idea during the next hard rain to walk around the yard with an umbrella and observe where the water is going and how well your gutter system is working.
The next line of defense against the elements is the home's siding. Its job is to keep water in liquid and vapor form outside with the bugs and spiders.
### MAINTAINING AND REPLACING SIDING
_Wood siding_ is commonly used in older frame homes and some newer ones. Typically it can last hundreds of years if it's maintained properly. But it's the word "maintained" that's critical here: Wood siding has to be painted and caulked every few years, or water, mold, and insect infestation will quickly destroy it.
_Figure 5-2_.
It's far cheaper to maintain the siding than getting an equity loan to replace it, or get stuck having to reduce your price when a home inspector finds serious problems. Use the following checklist every spring and fall to keep your siding in top shape:
•Look for signs of paint failure, such as blistering, peeling, and worn areas. The question you should continually ask yourself is: "Can water get through here?" A caulking gun, sandpaper, and paint brush are your best friends to keep the weather and critters at bay.
•Also look for mold, mildew, and blackening on the paint. If you find any of these, remove them immediately with a fungicide or bleach solution. Make a note to use a paint with a fungicide next time you recoat.
•To correct paint failure in only one or two areas, scrape and sand the bad area and apply a matching coat of quality, fungus-resistant paint.
•Carefully inspect where pipes and wiring enter the house and reseal with a good exterior caulk or canned foam insulation. Also look at where the siding butts up to chimneys, inside corners, and any other dissimilar surfaces. Scrape out old caulking and reapply to gaps or cracks that are not watertight.
•Checked and peeling paint are signs of water vapor getting behind the painted surface. Look for sources of infiltration: poor caulking, cracks, seams, and joints.
### Cedar Siding
Cedar siding comes in several different applications, with tongue-and-groove, beveled, vertical shiplap, shingle, or shake being the most common. Some homeowners prefer to use a semi-transparent stain to protect and keep the wood looking new; some use a pigment stain while others let the wood weather to a natural silver gray.
INTERESTING INFORMATION ABOUT SIDING
www.wrcla.org/specifications_and_publications/default.htm
www.askthebuilder.com/066_Bleeding_Redwood_And_Cedar_Siding_Stains.shtml
www.siding4u.com/vinyl-siding-cost-calculator.php
Although the natural grey is attractive, letting cedar weather without any treatment doesn't always work out. Mold can gain a foothold and stain the surface with dark splotches. If not treated with a fungicide or chlorine solution, mold can ruin the siding.
Upkeep on cedar siding entails replacing split boards, keeping mold at bay, sealing empty knotholes, and replacing rusting nails or screws with stainless steel or other nonrusting fasteners.
### Vinyl/Aluminum Siding
Vinyl/aluminum siding is popular because it's economical, has a long lifespan, maintains its color, and doesn't require painting every few years. It's a good choice for homeowners who are tired of painting or staining wood siding because it can often be installed over the existing siding.
Maintenance is a strong point because when grime starts to build up, power washing is all you have to do to keep it looking good.
### Fiber-Cement Siding
Fiber-cement siding is another good choice for new or replacement siding. Made out of cement, cellulose fibers, and other additives, it's a stable, durable, and economical option. You can paint it any color or it can be ordered pre-painted. Even better, it's available in styles that resemble stucco, wood clapboards, or even cedar shingles, and it's highly fire-resistant.
With fiber-cement siding, about the only maintenance required is to repaint it every ten years or so.
### Stucco Siding
Stucco siding is a mix of Portland cement, sand, and other fibers. It's durable, economical, comes in a large array of colors and patterns, and can also be painted.
Unfortunately, one type of synthetic stucco, known as EIFS (Exterior Insulation and Finish Systems), has been associated with moisture problems. The underlying wood on EIFS-sided homes may suffer rot damage. However, other types of synthetic stucco are quite durable. Still, it's wise to hire a professional to inspect the home if you're not sure what type you have. Early detection of water seepage can save big bucks later on.
Table 5-3 compares the most common options for replacing siding:
Because there are so many siding options, it's important to ask lots of questions about the product and the installing contractor, so you can compare bids accurately. Also, it's critical to use a reputable contractor experienced in the specific siding you're installing. (Chapter 6 details how to find and work with siding contractors.) The manufacturer's warranty can be voided if the product isn't installed properly. In addition, don't forget to factor in other costs, such as tearing off old siding, adding insulation wrap, and replacing any damaged wood.
Also important is removing any shrubs, trees, and vines that are growing against the home. Create at least a two-foot space between siding and plant growth; this will help control mildew and rot, and avoid damage from branches rubbing against the siding.
If you have wood siding or trim, you need to protect it by painting or staining every few years; it's part of homeownership. The next section shows how to protect your investment.
### EXTERIOR PAINTING TIPS
Basically, there are two ways to coat exterior wood: stain or paint. Both are formulated with pigments for color, and a water or oil base with binders and hardening agents. Stains have less pigment and resin than paints, which makes them a better choice if you want to retain more of the wood color and texture.
The biggest difference is in how stain and paint react with the wood surface. Stains penetrate raw wood like warm mustard spilled on your favorite shirt or blouse. Paint, on the other hand, bonds to wood surfaces and creates a thin film with little penetration into the wood fibers.
_Table 5-3. Siding comparisons_.
Advantages of stains on previously stained or unpainted wood are:
•Moisture trapped behind stain coatings usually bleeds through and evaporates on the surface. With paint, water cannot bleed through, so blistering and peeling result.
•Surface prep is easier. All you have to do is clean off dirt and mildew and re-stain.
•Stains are available with UV blockers, fungicides, and wood preservatives.
•You can get stains in semitransparent and solid colors to match the effect you want and the surface to be treated.
The bottom line is: If you have cedar, redwood, or other wood siding that is new or previously stained, you're probably better off staying with stain. But if you have siding that has been previously painted, you need to stick to paint, preferably acrylic latex or similar high quality products.
Although exterior house paint has a flat, nonreflective finish that tends to hide minor imperfections, you still have to prep the surface for good results.
HOUSE PAINTING HOW-TO WEBSITES
www.easy2diy.com/cm/easy/diy_ht_index.asp?page_id=35720475
www.architecture.about.com/od/paint/paint_your_house_tips_for_choosing_exterior_colors.htm
www.servicemagic.com/article.show.exterior-painting-techniques.8598.html
www.doityourself.com/scat/decoratingexterior
Also, color choices are endless, and for some homeowners that creates a problem. Paint choices made from color chips at the paint store won't look the same on your house. It's best to paint a test section first. Brush a wide swath on the back of the house first and see what it looks like in the real world, especially if you're changing colors. The pale blue color on a paint chip can look nearly bright blue when applied. You've undoubtedly driven by homes and wondered what the owners were thinking when they picked out their house colors.
One couple, for example, who wanted their home painted went to a local hardware store and picked out what appeared to be pale yellow from the paint chip display. They bought the paint, stacked the gallon cans on their porch, hired a handyman-painter and left for a ten-day vacation. When they returned, they were shocked to find their home's new color was closer to a mid-yellow. The neighbors called it the mustard house and before the owners could arrange for a repaint, the weather turned wet and cold for a few weeks. The home did get repainted, but the owners had to live with a color they didn't like for about six weeks, not to mention paying the costs of the new paint and a professional painter.
Lessons learned are to test the color before you order it in gallons, and to stay involved if you hire a painter, so you know the job is done the way you want.
However, before you paint, the critical first step is surface prep. Skipping this step is like forgetting to add eggs to the cake mix; you won't like the result. Use the following steps to guide you through surface prep:
1.Examine every square foot of your home's exterior and look for mildew, peeling, or checked paint. Remove mildew with a solution of one part bleach to two parts water.
2.If the paint is peeling, it's likely that water is getting under it and causing the problem; correct this before painting. Likely sources are failed caulking, bad insulation letting moist interior air seep through, or leaking flashing or gutters. For an older home, you may have to add insulation and a vapor barrier. Otherwise look for areas that water could get into and have your caulking gun at the ready.
3.For seriously deteriorated paint, you can sometimes clean most of it off the siding with a pressure washer. However, you'll need to keep the pressure under 2,000 psi (pounds per square inch) and the spray head about a foot away from the surface to avoid damaging the siding.
4.Depending on how bad the siding surface is, you may have to sand it. Palm and orbital sanders work well with medium grit (#100) sandpaper. The goal is to remove any loose paint and feather the patches of old paint into a smooth surface.
5.Make sure the siding surface is thoroughly dry before painting.
After you've completed the prep work, the semi-final step is to apply a coat of primer. Many homeowners skip this part, but if you want all your grunt work to last as long as possible, go with a primer coat. It's best to get the primer and top coat from the same manufacturer so you have complete compatibility.
The primer and final coat can be applied with a 4-inch synthetic bristle brush or a 6-inch or 9-inch roller. Some homeowners use an airless spray gun, which is a fast and efficient way to cover large areas. Before you go the sprayer route, however, be sure to mask the trim, windows, and everything else that you don't want painted.
Sprayers can be rented at home centers, along with ladders and scaffolding.
The cost of painting your home yourself is likely to be under $500; the cost of hiring a professional to paint it could be $6,000 to $8,000, depending on the area and season.
Just as sound and great-looking siding is critical to attractive curb appeal, a water-tight and attractive roof is equally important. Unfortunately, many first-time homeowners buy a home whose roof has only a few years left. Some appraisers require roofs to have a remaining lifespan of only three years or more. That means you'll need to start thinking of roof replacement right after you move in!
The following section lays out some typical roofing options.
### ROOFS: KEEPING THE WATER OFF YOUR HEAD
A newer roof adds curb appeal and value to your home. Although putting a new roof on just prior to selling will likely not get you a dollar-for-dollar return, it will help sell the home. According to _Remodeling_ magazine's 2006 "Cost vs. Value" report, nationally the average cost recouped of putting on a new roof is 73.9 percent.
As you might expect, not replacing an aging roof can easily lower the value of your home by more than the replacement cost. Once buyers see a problem area they go into a what-else-is-wrong-with-this-dump attitude, and chances of a good offer nosedive.
A far better strategy is to replace the roof or keep it in good condition before you decide to sell. Then you won't have to spend a lot of money to get the home in selling condition. And who knows, your roof may weather that once-every-500-years storm that leaks through your neighbor's roof.
INFORMATIVE ROOFING WEBSITES
www.usinspect.com/Roof/RoofMaterials.asp#Asphalt
www.nachi.org/roofs.htm
www.doityourself.com/scat/roofing
www.owenscorning.com/around/roofing/newroof.asp
www.asphaltroofing.org/resources_steep.html
www.nrca.net/consumer/fyi.aspx
There are lots of roofing options, some more expensive that others. But the first question is: How long are you likely to be in the home? Shingles are sold by the length of their warranty (e.g., 20-year, 25-year, 30-year, and more). The differences are the weight and design of the shingles.
If you're planning to move in a few years, 20-year shingles can be a good choice. But if you're planning on sticking around for a long time, the focus changes. For example, in Table 5-4, you can see that a higher initial cost often translates into a longer roof lifespan. But this is true only up to a point. Climate is a great modifier. For example, a 20-year asphalt shingle roof in a moderate climate that lasts 25 years would only last 14 years in the hot southwest or the high snowfall area of Vermont.
If you need a new roof, check out what type works best in your area. For instance, in fire-prone areas you would want to go with Class A shingles (the most fire resistant). In warm or humid climates, zinc- or copper-coated granules added to shingles can help protect against algae. Whatever type roof you need, get at least three written bids.
For all those roofs between new and nearing replacement, the following tips will help you extend their life and prevent costly water damage:
•One of the most important things you can do is inspect your roof every few months. (Do not climb or walk on the roof; use binoculars instead.)
_Table 5-4. Typical roofing options_.
Three things you especially want to look for with your binoculars besides birds are:
1.Brittle and curling shingles, especially along the edges
2.Protective granules wearing off, leaving bare spots on the shingles
3.Rusting or corroding flashing and cracks in sealant around pipes
•Look for gravel blowing or washing off onto the ground around tar and gravel roofs. Wear is hard to spot from the ground on these types of roofs, so it's important to get a professional roofing inspection every few years.
•After high winds or heavy rain storms, check for missing or damaged shingles.
•Make sure you have adequate attic ventilation: louvers, ridge, and soffit vents should total one square foot of ventilation for every 150 square feet of attic floor space. Moisture trapped under the roof decking can lead to mold and water damage. Check the underside of the roof decking for leaks and mold from inside the attic periodically.
•Look for rodent or squirrel damage, especially around fascia and soffit boards. If you live in squirrel or woodpecker country, adding a metal soffit and fascia covering can prevent a lot of damage.
•Make sure you have metal drip caps to protect the fascia boards under the edge of the roof.
•And a biggie that can get really ugly: a sagging ridge beam or a wavy roof. This likely means the roof and decking have to be torn off, and the damaged rafters and ridge beam need to be replaced along with any other damaged wood.
As you've probably deduced already, the battle cry for homeowners should be "keep out the water at all costs!" Vigilance will save you a lot of money in repairs and replacements.
### WINDOWS DO MORE THAN LET IN LIGHT
Along with the roof, windows are your home's most costly and high-profile component. They also define the home's personality and usually match (but not always) its style. We've all seen homes where owners have upgraded and changed the window's size and type, resulting in a dramatic improvement.
ENERGY- AND MONEY-SAVING TIP
Replacing old aluminum or wood single-pane windows with energy-efficient, double-pane windows will save you serious bucks on your heating bill. You may also get a tax credit for upgrading your windows. Check out www.energystar.gov for more information.
Replacing windows may or may not give you dollar-for-dollar payback when you sell. If you have older single-pane aluminum windows and replace them with energy-efficient vinyl windows, your payback will be high. Replacing windows just to create the latest look will probably not return near your investment. Buyers often have a hard time justifying paying extra for something they consider nonessential or cosmetic.
Basically, there are two ways you can go when upgrading your home's windows:
1.You can replace the existing windows with upgrades without altering the opening. This, of course, is the easiest and most economical way to go.
2.You can alter the window openings for bigger or smaller windows and give your home a new look. This is an approach that you want to think through carefully to make sure the new windows are in harmony with the whole house; you may even want to consult an architect or designer. Sometimes this approach can really bump up the value of the home and justify the extra cost.
CURB-APPEAL TIP
On many homes the siding surrounds the windows so they don't have picture-type framing. When you upgrade your windows consider adding molding and trim, which can be painted or stained a different color from the siding.
If you do decide to replace windows, there are some exciting styles, coatings, and fillers that can make a big difference in your energy bills. Some typical window styles for upgrading are:
• _Double-hung windows_ have two sashes that slide up and down in the jamb. In the past a system of weights and pulleys counterbalanced the sashes. These have been replaced by pressure-fit vinyl jambs that are more weather proof and work smoother. Some models tilt inward for easier exterior-side cleaning. These are probably the oldest style and most common windows around, and with double panes they are energy efficient.
• _Casement windows_ typically have hand cranks that control opening and closing. These windows are common on Craftsman, Tudor, and California Ranch style homes. Modern materials and designs have made these a good choice for replacement windows.
• _Slider windows_ are a common style in homes built in the 1970s. If your home still has these single-pane aluminum sliders, replacing them with double-pane windows should be at the top of your list. The new windows are more energy efficient, and the money you save over five to ten years should easily cover the replacement cost.
• _Fixed windows_ or _picture windows_ are great for views and bringing light into the home. However, older single-pane windows leak a lot of energy, and replacing them with double- or even triple-pane glass should be a top priority.
One homeowner who replaced her old aluminum single-pane windows was amazed at the difference in comfort and noise levels. The large picture window in her front room had made it almost impossible to heat adequately and the traffic noise was difficult to get used to.
The owner spent $6,500 to replace all the windows in her 975-square-foot 1970s ranch-style home, including the picture window in her living room. The double-pane vinyl windows saved her about $55 a month or $660 a year in heating and cooling costs. If you divide $660 in savings into the $6,500 window replacement costs, it will take the owner about ten years to break even. However, colder climates like Minnesota or Vermont could even shorten that break-even point.
• _Bay or bow windows_ are multiple windows that bump out from a wall. Typically, two double-hung windows extend out at a 45-degree angle from the wall, with a larger, fixed window in the middle. Bay windows make a great upgrade on single-level or ranch-type homes. They open up the living room or kitchen dining area to more light and room as well as add value to the home.
WINDOW-REPLACEMENT WEBSITES
www.efficientwindows.org/nfrc.cfm
www.vinyl-replacement-windows.com/vinyl-replacement-window-basics.html
www.online.consumerreports.org/Windows
www.b4ubuild.com/links/windows.shtml
www.thisoldhouse.com/toh/knowhow/tools/article/0,16417,419997,00.html
Shopping for windows isn't always easy. There are lots of options, a wide range of prices, and an incredible amount of hype. Typically, quality windows cost about $200 to $250 or more per window. A 12-window house with patio sliding door, for example, would cost around $4,500 to $6,000 installed.
The following tips will help your thread your way through the buying maze:
•Look for windows that are double-paned and argon-filled with a low E-coating, that is, ENERGY STAR–certified windows. They cost about 10 to 15 percent more but reduce energy loss by 30 to 50 percent.
•Window coatings are climate specific. You won't have the same coatings in Phoenix that you would in northern Michigan. So shop at local suppliers who know what your area needs are.
•Look for windows that have a NFRC (National Fenestration Rating Council) sticker on the window. This discloses facts about the windows that allow you to compare different models. Go to www.nfrc.org/fenestrationfacts.aspx for a complete guide to deciphering the numbers on the sticker.
•Don't buy windows from salesmen knocking on your door offering you a good deal because they happen to be in the area. Deal only with reputable window dealers or contractors.
•Double-pane insulating windows cut energy costs in hot climates as much or more than cold climates. The specific glass coating for your area will make the home much more comfortable than single-pane windows.
•Once you've decided on what windows you want, get three install bids in writing and compare them.
### OTHER EXTERIOR FEATURES TO KEEP AN EYE ON
Once you've got your home watertight and storm-proof, the next area to focus on are items that are important but not critical. Important in that they need to be done to preserve curb appeal or keep from getting worse, but not critical as when you'll hear a drip, drip in the middle of the night.
### Fireplaces
Fireplaces are almost a necessity for some buyers, while others could care less. How much value it gives your home depends on the area and home buyer preferences.
Still, if your home has a fireplace, it's important to keep it maintained, not only for its safe operation but to preserve your home's resale value.
INFORMATIVE FIREPLACE WEBSITES
www.homebuying.about.com/cs/fireplacesafety/a/wood_fireplace.htm
www.csia.org/homeowners.htm
www.oldhouseweb.com/stories/detailed/10151.shtml
On homes ten years old or older, start out by getting a fireplace inspection (wood burning stoves included). It's important to know what condition your fireplace is in. Typically this will cost around $50 for an inspection and $250 for cleaning. If the firebrick liner or metal flue is damaged or worn out, you'll need to replace it. Depending on the type wood you burn, how often you build a fire and the climate, liners usually last anywhere from 12 to 50 years. Cost to replace liners and structural repairs can run from $1,500 to $3,000.
To keep your fireplace in top condition and enjoy it safely, check out the following tips:
•Use the fireplace for short-duration fires, five hours or less.
•Open glass doors or vents to allow air to be drawn up to cool the chimney.
•Keep screens closed to prevent sparks from jumping out onto the carpeting.
•Never leave a fire unattended when children are around.
•Open a window when using the fireplace so air can circulate up the chimney.
•Make sure your chimney has a cap to keep out rodents, birds, and nests, and a spark arrestor to keep sparks from torching your neighbor's roof.
•Fireplace coals can remain hot long after the fire has burned down. Remove ashes with a metal shovel or scoop (never a vacuum) and store in a metal container away from anything flammable.
•If you're a heavy fireplace user, have a chimney sweep check it each fall before you use it. This is important because obstructions or problems can cause carbon monoxide to back up into the home, and that's something you don't wake up from. Adding carbon monoxide alarms should be a must near the furnace room and close to fireplaces.
•Smoke detectors in every room are a building code requirement in most states. If you have an older home that doesn't have them, they're cheap and easy to install, and they run off 9-volt batteries that have to be replaced about every six months.
•If you convert your fireplace to use another type of fuel (natural gas, wood, propane, pellet, etc.) make sure your conversion meets local codes and have it installed or inspected by a pro. There's no room for error here because carbon monoxide (CO) fumes from an improper install are lethal.
### Swimming Pools
In some parts of the country, swimming pools are as important as the family dog, and a home may be a hard sell without one. In other areas, a home may be a hard sell with one.
Regardless of what area you live in, if you have a pool, it's a big investment. To keep that investment in good shape, good maintenance is a must. For most new pool owners the first step is to find a pool maintenance contractor. Likely, the previous owners used a maintenance company, and if they were happy with the service, you may want to continue with the same outfit. Because pool maintenance is a highly competitive business, you can be sure they'll be contacting you soon after you move in. Otherwise, use the standard procedure for finding a contractor: minimum of three written bids, testimonials, and personal interviews to find one you're comfortable with.
INFORMATIVE POOL-OWNER WEBSITES
www.home.howstuffworks.com/swimming-pool.htm
www.cdc.gov/healthyswimming
www.swimmingpools101.com
www.home.comcast.net/~hot_tub
www.learnaboutpools.com
You'll also want to contact your homeowner's insurance provider and make sure you're covered for additional liability on the pool.
### Garage Doors
Garage doors are an important part of your home's curb appeal and should be maintained in as good a condition as the rest of the exterior. That means keeping the opening system, door, and bottom weather strips in good order. This is important not only to prevent energy loss but to maintain home security.
GARAGE-DOOR INFORMATION WEBSITES
www.hometips.com/cs-protected/guides/garage.html
www.cpsc.gov/cpscpub/pubs/523.html
www.nhc.noaa.gov/HAW2/english/retrofit/garage_doors.shtml
www.home.howstuffworks.com/how-to-repair-a-garage-door.htm
If your garage door is beyond maintenance and it's time to replace, consider these shopping tips:
•The first decision is the style; obviously it has to fit your home's style. In other words, a carriage-house design probably won't look good on a California Ranch-style home. You want the style and color to fit the home.
•The next decision is the color and texture. That too must blend or contrast with the home. For ideas look at new homes and see what other homeowners have done.
•It may cost slightly more, but shop for doors that have sections made of wood or steel that include a sandwich of foam insulation. Insulating the garage door results in significant energy savings.
When you're shopping for a door, you'll find many options; it can get confusing. Table 5-5 can help.
To new homeowners it may seem that maintenance and upgrading are a never ending hassle. And to a certain extent that's true. Although it's nice to be able to call the super or landlord to take care of problems, there's much more satisfaction in acquiring the skills and know-how to maintain your own home.
The appeal of fixing up a house, decorating it, and making it uniquely theirs is what motivates most new homeowners to fire the landlord. And the good news is, it gets better and more fun the more you learn and experiment.
The checklist in Table 5-6 is designed to help you keep the outside of your home in good condition. After a while, looking for problem areas becomes automatic as you gain experience and join the ranks of seasoned homeowners.
_Table 5-5. Typical garage door options_.
_Table 5-6. Exterior seasonal and maintenance checklist_.
**Item** | **Action Needed**
---|---
Electrical Wires | Trim tree limbs that are threatening power lines to the electrical service entrance so that ice storms, winds, and heavy snow don't down a line.
Outside Electrical Panel | Check outside electrical panel for wasp nests or loose wires.
Cracks in Concrete or Asphalt | Fill all cracks before freezing weather sets in so water doesn't infiltrate and freeze, causing more damage. Home centers sell concrete and asphalt crack sealers that prevent this.
Retaining Walls | Make sure there are no bulges or failing areas that can break loose in a heavy snow or rain.
Wood Porches | Check all the wood for rot, failed paint, and loose boards. Repair immediately to prevent further damage.
Siding | Look for areas that need caulking. Check for signs of paint failure or where water can get behind the siding.
Soffits and Fascia | Look for peeling paint, loose boards, rodent or bird damage. Look for anywhere water is going where it shouldn't be.
Around Chimneys | Make sure flashing and siding-to-chimney contact is in watertight condition. Caulk or replace flashing if deteriorated.
Gutters | Make sure laterals and downspouts are free of debris and are not leaking. Attach extensions to channel water away from the house.
Foundations | Look for cracks or places where water can get in, such as utility pipes and wiring.
Termites | If you live in termite country, get a professional inspection to make sure your home isn't being devoured by the ravenous bugs.
Grade | Look at the grade around the foundation and make sure it routes water away rather than into the crawl space or basement.
Roof | Look for missing shingles, cracks in flashing around pipes, and valleys.
Moss | If you find moss, eliminate it with a fungicide or 1 part bleach to 2 parts water.
Well | Have the water tested yearly for contamination.
Septic System | Have it pumped as needed or set up a maintenance contract with a reputable dealer. Check for soggy ground or lush plant growth around the septic field. That could mean a full tank or a failing system.
Sump Pump | Test by pouring water into the well; pump should turn on automatically.
Weather Stripping
Roof Vent Pipes | Check around windows and doors for wear.
Make sure they are free of leaves, bird nests, and other debris.
## CHAPTER 6
## YOUR HOME'S LANDSCAPING AND CURB APPEAL
Curb appeal to a home is what packaging is to the products you buy. The better it's packaged, the more successful it is. Even if you don't plan on selling for awhile—the national average for living in a home is slightly over seven years—it's critical to keep your landscaping and yard in top condition. You'll not only enjoy your home more, but when the time does come to sell you won't have to pay a lot to catch up.
You don't have to have a green thumb to have a great yard and enjoy it. Like decorating the interior, you can create your own little world and have fun doing it. True, there are a few new skills to develop, but they're not difficult.
This chapter gives you some basic tips on how to make your yard look good and keep it that way. Still, you'll probably want to take landscaping your yard to a higher level by reading magazines, landscaping books, and checking out new subdivisions for cutting-edge ideas.
To get you started, some important tips covered in this chapter are:
**Dressing up your yard for maximum curb appeal**
**Creating and keeping an attractive lawn**
**Using trees, bushes, and hedges in your landscaping**
**Keeping driveways, walks, and fences in good condition**
We all want to make a good impression on friends and family when they come to visit us, and a good way to do this is with killer curb appeal.
### HOW TO CREATE KILLER CURB APPEAL
Step one is to walk across the street from your home with a digital camera and a note pad. Take pictures of your yard from several angles and note the problems you see.
Examples of problems to look for are:
•Balance and symmetry. Does one side have something that makes the front view look unbalanced? For example, are there trees, large bushes, hedges, or a fence running along one side with nothing on the other side to balance it?
•Is there a utility box on the property line that creates a distraction? Some streets with underground power lines have large transformer boxes a few feet from the curb.
•Are there problem areas that need attention, such as an old fence, bare or dead spots in the lawn, or out-of-control bushes and trees?
•Are there foundation bushes that are not trimmed below the windows and out just a couple of feet from the home?
•Do your driveway and walks have cracks or spalling (deteriorating) surfaces?
•Are there sunken walkway sections, especially where they tie into the porch? These can also be serious safety hazards.
•How do your curb appeal photos look? Eventually if you put your home on the market, similar photos on a website are likely to be a buyer's first introduction to your home. If there are glaring problems, or one side of the house is particularly ugly, you should address these first.
Once you've put together a problems list, order them in a way that you can tackle and complete them. For example, if you were planning on selling the home soon, time would be an important factor and the order wouldn't matter too much, just so you get it all done. But if your project is to make your home look its best, then order the list by cost, by what you can do personally, or by what's important to you.
Step two is to develop a plan from your list, pull on the gloves, and have fun.
INTERESTING CURB-APPEAL WEBSITES
www.hgtv.com/hgtv/pac_ctnt_988/text/0"HGTV_22056_32482,00.html
www.homebuying.about.com/cs/sellerarticles/a/curb_appeal.htm
www.popularmechanics.com/how_to_central/home_clinic/1548397.html
www.dannylipford.com/home_improvement.php?cat=lawn
www.the-landscape-design-site.com/directory
### LANDSCAPING BASICS
To plan an attractive organized yard, you don't have to spend big bucks to hire a landscape architect. The key is to develop a simple plan and gradually add to it over time. Don't make it too complicated or overwhelming.
Some homeowners like to start out drawing a to-scale landscaping plan of their yard on a large sheet of graph paper. Others prefer to buy a software program and plan with a screen and mouse. If you go the software route, there are many programs on the market—some good, others not so good. Two programs worth checking out are _Your Complete Landscape and Garden Designer_ and _Better Homes and Gardens Landscaping and Deck Designer_. One advantage of a software program is that you can import photos of your yard and see how changes impact your landscaping ideas.
Some basics for developing your landscaping plan are:
•Create a base plan. Draw an scale outline of the property on a large sheet of graph paper showing the dimensions, where the house is located, easements, setback distances from other homes, sidewalks, and streets.
•Add in walls, walks, fences, decks, and other appurtenances.
•Draw in drainage patterns, sources of noise or privacy problems, and views you want to preserve.
•Attach an overlay of tracing paper to the base plan and pencil in the changes or additions that you want to make.
•Try to blend the finished landscaping with the neighborhood. Curb appeal is lost when your home and landscaping are out of character with the rest of the neighborhood.
When you're working on your landscaping plan, keep in mind the following seven design elements that every landscaping should have:
1.Good landscape design starts with composition. Trees, shrubs, flower beds, and walks should fix together like pieces in a puzzle. Landscaping features should be grouped into related components. For example, an elongated area enclosed by concrete curbing could have a grouping of three trees on one end. Colorful flowers could be planted in the middle, and another grouping of smaller trees (flowering fruit trees, for instance) could anchor the other end. The key is to create a sense of balance.
2.When you stand on the curb looking at a home's landscaping and feel there's something missing or isn't quite right, it's usually from lack of balance. Balance is created by placing trees, flower gardens, fences, bushes, and hedges in harmony with each other. When you develop a landscaping plan first, you can move these elements around on paper (or screen) until you achieve a balance. If you start planting without a plan, as many homeowners do, achieving balance is difficult.
3.Scale is the next element to creating killer curb appeal. Driving around any subdivision in the country, it would appear that most homeowners ignore scale when planting trees and shrubs. They don't realize that trees and plants grow up and that the cute fir tree planted by the house isn't going to stay that way. The home in Figure 6-1 shows what happens when a homeowner forgets that a blue spruce can grow up and create problems. If this home were to go on the market, the owner would need to remove the tree to get top dollar.
4.In nature there's a natural ratio of 3 to 5 (often called the Golden Mean or ratio) that resonates with most people. Because we identify with this ratio and find it comfortable, it's commonly used in architecture, advertising layouts and countless other applications. In landscaping your yard, you can use this ratio to design flower beds, plant groupings and so on. For example, if you were to design an garden area enclosed by curbing using this 3:5 ratio, the size could be 9 feet by 15 feet, or 12 feet by 20 feet, and so on. (In other words, 3 feet × 5 feet doubled is 6 feet × 10 feet, tripled it's 9 feet × 15 feet, and so on.) This is also why you often see plants and trees grouped in threes and fives.
_Figure 6-1_.
5.Because a yard with only grass can be boring, it's important to add color. This is a good opportunity to add your favorite colors to the landscaping. Flowers, ground cover, window boxes, and planters are a great way to add colorful personality to your yard.
6.Another important attribute of landscape design is texture. A good way to use this is to group like-textured plants together. For instance, you wouldn't plant a red lacey-leafed maple next to scrub pines.
7.If all the landscaping elements discussed thus far are used, the result should be harmony. Your yard will not only be attractive, but you and your guests will enjoy being there. As Realtors are fond of saying, it has emotional appeal.
Once you've got your landscaping plan put together, the first step in implementing it is to make sure the lawn is thick and green. It's the backdrop for the rest of the landscaping.
However, if you live in the arid southwest, the rules are quite a bit different. Instead of lawns, it's sand, rocks, and drought-resistant plants. Colorful flowers in planters and pots give you a great opportunity to be creative. The same design principles apply; only the medium changes.
For the rest of the lawnmower-owning homeowners, it all starts with the turf.
### LAWN CARE 101
The two house hunters had looked at over a dozen homes that day and found none of them appealing. Some houses had yards that looked like they would support a buffalo or two, while others were just dirt. The shoppers were nearing the end of the list when they drove up to a home with a great-looking lawn.
Not having time before getting transferred to do much more than put in the lawn, the owners had decided to make it a selling point at the urging of their agent. It was a carpet of manicured lush green. The buyers took a quick look through the home, walked across the lawn, and told their agent to write up a full-price offer; they didn't want to chance losing it. As the buyers said later, "The yard doesn't have any trees or other landscaping, but that's good. We can do what we want because the hard part is done."
LAWN-CARE WEBSITES
www.yardcare.com
www.allaboutlawns.com
www.lawn-care-tips.com/lawncaretips.html
www.backyardgardener.com/lawn
www.eartheasy.com/grow_lawn_care.htm
www.gardening.cornell.edu/lawn/lawncare/index.html
Many people feel intimidated by putting in a lawn or making an existing one the envy of the neighborhood. The following tips will help you turn brown into lush green:
•For a killer lawn you need several inches of well-drained, noncompacted soil. Otherwise you'll have to bring in about four inches of top soil and mix in organic material. Cow, chicken, or turkey manure work well, as does leaf compost and peat moss.
•Testing your soil's pH level is an important step. If the soil is too acid or too alkaline, you won't get good results. If it tests acid (less than 7) or alkaline (over 7) you may need to compensate. Typically, lawns thrive best on a pH level between 6 and 7 or slightly acidic. Soil testing kits are available at garden centers, or in some areas from the U.S. Department of Agriculture, whose website is: www.nrcs.usda.gov/feature/backyard/NutMgt.html
•The soil should be raked into a smooth seedbed, with no rocks or compacted areas. Any drainage problems need to be corrected so a rainstorm won't wash out your new lawn.
Once the seedbed is prepared, there are three ways you can go:
1.The most economical way is to buy the seeds and spread them yourself using a wheeled spreader or hand-cranked broadcaster, which can also be used to apply fertilizer when the lawn is established. Spring and fall are the best times to plant lawn seed. Follow the directions on the seed package for best coverage.
2. _Hydro-seeding_ , though more expensive than applying seeds yourself, is fast and efficient. Typically the cost is between $250 and $400 for a 1,000 to 3,000 square-foot lawn. With this process a slurry of grass seed, wood fibers, and fertilizer is sprayed on the seedbed. You keep the area moist, and in a few weeks you have a lawn. Check under lawn installation in the yellow pages for hydro-seeders in your area.
3.Slightly more expensive than hydro-seeding is buying live sod from a sod farm. A big advantage is you have an instant lawn that's ready for the kids to play on in a few days. It's also a quick way to upgrade your lawn when you're planning to sell. Depending on the grass type and where you live, sod delivered on a pallet typically costs $.25 to $.40 per square foot.
Regardless of the way you go, it's important to keep the lawn watered frequently until it's completely established.
FERTILIZER FACTS
Most areas fertilize their lawns three to four times a year. A local nursery can give you the best mix for your lawn type. Bagged fertilizers have labels with three numbers, such as 20-10-5.
The first number gives you the nitrogen content. A higher number means more nitrogen.
The second number is phosphorus, which is needed for a healthy root system.
The last number is potash, which provides amino acids and proteins for lush plant growth.
### Edging
The next step in taking your lawn and yard to the next level is installing edging. This is the icing on the cake: It not only gives your lawn a manicured look, but makes it easier to maintain. There are several ways you can create great looking edging:
•Use a gas- or electric-powered edger to cut clean lines between lawn, walks, driveways, and flowerbeds.
•Hire a professional landscaper to curb around lawns and enclose flowerbeds. You can even get several different patterns and colors to create a customized look. Curbing costs typically between $2 to $3 a running foot.
•Install vinyl, wood, or metal edging you can buy at home and garden centers. Bricks, pavers, and narrow strips of crushed stone or gravel also work well.
By this time you've probably spent a lot of time and money getting your lawn up to speed, or if you're a new homeowner you undoubtedly look at the mortgage balance and feel every blade of grass is worth a lot of money.
The following ten _Do's_ and _Don'ts_ summarize and show you how to protect your investment and keep it green.
### 1. Testing Your Soil
You'll want to know the levels of potassium, magnesium, calcium, sulfur, nitrogen, and organic matter. The U.S. Department of Agriculture has county extension offices that can direct you. For the office nearest you, check out www.csrees.usda.gov/extension.
SOIL-TESTING WEBSITES
www.outsidepride.com/store/catalog/Soil-Test-Kits-p-1-c-268.html
www.agr.state.nc.us/cyber/kidswrld/plant/soiltest.htm
www.ohioline.osu.edu/hyg-fact/1000/1132.html
www.drgoodearth.com/?src=overture
### 2. Mowing at the Proper Height
Each time you mow some of the plant leaf is removed. If you remove too much, the plant can't produce the nutrients it needs to survive. Mowing too short also weakens, thins out the grass blades, and allows weeds to gain a foothold.
Every ⅛-inch increase in mowing height increases the leaf surface 30 percent. In the Northern states where grasses are predominately bluegrass, fescue, and rye grasses, mow 3 inches or higher during mid to late spring and early fall. During the hot mid-summer weeks setting the mower blade 4 inches high helps your lawn get through the heat.
Southern grasses, namely St. Augustine and bahia, should be mowed during summer and early fall at 3 to 4 inches. Bermuda, centipede, and zoysia grasses can tolerate 2- to 3-inch mower settings.
### 3. Cutting Too Much at Once
Lawns should be mowed frequently. Cutting too much off in one mowing, however, forces the grass to put all its energy into replacing lost leaf area. This stops vital root growth and food storage.
_Follow the ⅓ rule:_ Mow as often as needed, but don't remove more than a third of the grass blade. In other words, if you mow at 3 inches, the grass length shouldn't be more than 4½ inches.
### 4. Watering Wisely
Many homeowners with automatic sprinkler systems set them to come on every day for 10 or 15 minutes. Though frequent, the water output isn't enough to penetrate into the root zone where it's needed. As a result, the plants tend to develop shallow root systems that dry out quickly during hot weather.
To give your lawn the right amount of water where it does the most good, set your sprinkler's timer to water for an hour or so. Test the water penetration by pushing a spade about six inches into the soil. If it's moist down to that depth, you're on the right track.
Water heavily once a week, or twice if the weather is hot and windy. Early morning is the best time to water to minimize evaporation loss. Don't overwater, that can be as damaging as too little water. If the grass becomes wilted with a bluish hue, it needs water ASAP.
### 5. Keeping Mower Blades Sharp
A dull mower blade shreds instead of cutting. This can leave ragged grass ends with an unattractive gray color, and it also invites insect problems.
Start the mowing season with a sharp blade and resharpen whenever it starts to dull. Check the lawn for sticks, rocks, or anything that will nick the blade or cause a safety problem before you mow. Some homeowners have two blades so when one becomes dull they can replace it and drop off the dull one to be professionally sharpened. (Hey, Saturdays are too important for lawn-mowing downtime!)
Mulching blades are especially sensitive to maintaining a good edge because they must slice grass several times. Unless you're good at sharpening, it's best to have these blades re-edged by a professional.
### 6. Clippings
If you mow frequently and use a mulching blade, leave the clippings on the lawn where they will break down and reduce the need for fertilizer. It's also more environmentally responsible to keep as much material out of landfills as possible.
If you wait too long between mowings and end up cutting grass longer than an inch, the cuttings can bunch up on top of the lawn. This results in unsightly rows of brown grass. Unfortunately, whenever you let the lawn grow too long the only way to prevent these rows of dead grass is to use the bag attachment. You can feel guilty over this and vow not to let it happen again, but then, when the fish are biting. . . .
### 7. Improving Your Soil Quality
Good quality soil is highly active with earthworms, microbes, and other organisms that break down clippings and improve soil texture. It's a big, complex world at the root level, which you need to maintain at a healthy level.
Unfortunately, using too much insecticide, fungicide, fertilizer, and snow-melting salt can cause problems with these beneficial organisms. To keep your soil healthy:
•Use soil activators (available at all garden centers) that increase the level of beneficial organisms in your lawn.
•When you need to use insecticides and fungicides, treat only the area affected to prevent damaging unaffected parts of your lawn. Depending on the problem, you may be able to use nonchemical remedies.
•If the underlying soil is starting to compact, hire an aerating service or rent an aerating machine. Aerating machines punch 2-inch-deep holes into the turf every 6 inches and remove a small plug of soil. This allows water and nutrients to penetrate into the root zone. Spring is the best time to aerate a lawn.
### 8. Using Fertilizer
Many fertilizers are synthetic and high in nitrogen. They're fast-acting and can overstimulate plant growth. This weakens the plants and opens to door to fungus and insect problems.
For most lawns, organic and synthetic slow-release fertilizers work best. They release nutrients slowly and reduce the chance of burning (killing the grass). If you want to use organic products, manure, fish meal, and kelp work well.
Fertilize no more than three times a year and don't apply more than 1 pound of nitrogen per 1,000 square feet of lawn at a time.
### Fertilizing at the Right Time
Timing is important; if the grass is brown, weeds will flourish and take over the lawn.
Fertilize Northern lawns in early fall and again later on when growth has stopped but before the lawn has turned brown. This helps nurture the grass through the winter months and gives it a boost for the spring growing season. The next feeding should be in the spring, just before it gets hot.
Southern lawns thrive in summer heat and can be fertilized then. Best times for this climate are early spring, summer, and early fall.
### 10. Keeping Thatch Under Control
Thatch is the matting of dead grass, roots, and plant matter above the soil surface. It forms a barrier that prevents water and fertilizer from penetrating into the root zone. This makes your lawn a target for all those critters that prey on grass.
Three things you can do to minimize the buildup of thatch are:
1.As discussed earlier, aeration can break through thatch. However this is a temporary solution, because thatch needs to break down and decompose on a continuing basis to keep your lawn healthy.
2.A long-term solution is to encourage earthworms and microbes to take up residence in the thatch zone and be happy. You do this by maintaining the pH level of the soil between 6.5 and 7 (7 is the neutral zone; neither acid or alkaline) and watering on a consistent schedule.
3.If the thatch is over ½ inch you can rent a power rake or dethatcher. Be aware that mechanical raking can be destructive because lawn roots that are growing in the thatch zone will tear out easily. Because over-seeding is usually required after power raking, it's best to go this route in late August for optimum seed germination.
USEFUL WEBSITES ABOUT THATCH
www.urbanext.uiuc.edu/lawnchallenge/lesson5.html
www.extension.umn.edu/distribution/horticulture/DG1123.html
www.uky.edu/Ag/Entomology/entfacts/trees/ef402.htm
www.ohric.ucdavis.edu/Newsltr/CTC/ctcv34_1.pdf
When you get your lawn green and when the worms, friendly bugs, and microbes are happy, it's time to turn your attention to bigger game.
### TREES AND SHRUBS
One huge problem many homeowners have is allowing their landscaping to get out of control. Foundation bushes grow up and cover windows, and trees planted too close to the house cause siding and roofing damage. Other curb-appeal destroyers are too many trees and bushes in front of the home giving it a back-to-nature appearance.
Ideally, trees should frame a home, not block its view from the street or prevent light from entering the home. Rooms that are dark and dreary from lack of light substantially lower a home's value.
INTERESTING GUIDES TO PRUNING AND CHIPPING
www.aggie-horticulture.tamu.edu/extension/pruning/pruning.html
www.na.fs.fed.us/spfo/pubs/howtos/ht_prune/prun001.htm
www.extension.umn.edu/distribution/horticulture/DG0628.html
www.arborday.org/Trees/pruning
www.extension.umn.edu/distribution/naturalresources/DD3949.html
If your trees and shrubs have made your yard look like a jungle, taming them will do wonders for your curb appeal and pride of homeownership. The following tips will get you started:
•Note what hedges, bushes, and trees obscure your home and need to be trimmed or removed. Trimming is a big project because piles of cuttings build up quickly. You'll need to arrange for a large rental dumpster or other means of getting the refuse to a landfill. Another more green approach is to rent a chipper and turn the cuttings into mulch.
•Don't feel you're damaging plants by pruning excess growth, it usually benefits most trees and shrubs. Older, overgrown plants may be reinvigorated if they are thinned to allow more light and air.
•For thinning shrubs, find a bud facing the direction you want a branch to grow and cut above it. If you pinch or cut off growing tips in the spring, it will make the shrub bushier and more compact, something you may want to do with foundation plants.
•Foundation plants should be trimmed down so they don't grow above the bottom window sill. Nor should any tree limbs obscure the view from any windows.
•Leave about a 2-foot space between the house and any plants. You want to allow air and light to circulate and to discourage mold and fungus growth on the siding.
•To get ideas of how plants can be used to increase your home's attraction, check out new homes and subdivisions. Builders often hire professionals to landscape their model homes. Note the type of plants these landscapers use and how they're grouped. Also, as you drive around different subdivisions, notice what you like and don't like about what other homeowners have done with their yards.
Staying on top of landscaping isn't difficult, it just takes a consistent effort to stay ahead of the plant's growing cycles and being aware that at certain times you'll need to get out the pruning shears. The following are the five most common pruning mistakes homeowners make:
1.Never top or remove large parts of the tree. It can either kill the tree or create an ugly tree with limbs trying to grow out the trunk to compensate. One appraiser estimated that a well-placed mature tree is worth anywhere from $1,000 to $10,000, depending on the area.
2.Avoid pruning more than a quarter of the live branches at one time, or more than a third in a year. Stimulating too much growth creates additional work for you later on. This is why letting bushes get out of control and then having a pruning party is counter-productive.
3.Improper pruning leaves stubs too far from a dormant bud or another branch. These stubs invite decay and disease.
4.Never coat pruning wounds with paint, shellac, or tar. This can damage trees by holding in moisture and causing decay.
5.Perhaps the biggest mistake is not pruning at all, and letting everything grow until it reaches a panic point. These homeowners devalue their homes and the rest of the neighborhood as well.
In one extreme case, a homeowner refused to mow the lawn or take care of his yard. The neighbors got together and offered to do it for him on weekends when they did theirs.
Yes, the neighbors could have gone the legal route, but that probably wouldn't have solved the problem. Because it was a small yard, the homeowners along the street agreed that as long as the slob neighbor agreed to it, spending a few minutes every week or so mowing his lawn was their best short-term solution. (It's been going on now for six years!)
To their credit, the homeowners in the neighborhood took a pragmatic approach: They recognized that an out-of-control yard can adversely affect the value of the entire neighborhood.
### FROM GREEN TO ASPHALT AND CONCRETE
Along with landscaping, driveways and walks are a big part of what makes up curb appeal. You can have a great a great lawn, but if the driveway surface is weathered and flakey (spalling), it calls attention to an unsolved problem. Lush lawn and flower beds won't even be noticed by visitors or buyers if they see a driveway with cracks, holes, or other signs of neglect.
If you want to know how important driveways are to your home's value, check out what builders are installing at new homes, such as concrete stampings that look like brick, or stone along with added color and weather sealing.
MORE INFORMATION ON ASPHALT DRIVEWAYS
www.hotmix.org/driveways.php
www.drivewayimpressions.com
www.pavementpro.org/homeowners_questions.htm
www.naturalhandyman.com/iip/infdrivewaysealer/infdrivewaysealer.shtm
_Asphalt_ makes attractive driveways and has a few advantages over concrete. For example:
•Asphalt is more flexible than concrete and can be installed over more uneven surfaces.
•In cold climates asphalt's dark color absorbs heat better than concrete.
•Asphalt does not require formwork or long curing times
•If it's damaged, asphalt is easy to repair.
•For long runs over 25 feet, asphalt can cost much less than concrete or brick.
However, you can't put in an asphalt driveway on the cheap and expect it hold up. Some do's are:
•Asphalt drives need to be installed by a paving contractor. It's not a do-it-yourself project because of the large equipment needed.
•You should install 6 inches of road base gravel for the asphalt base. Some homeowners try to shortcut this step, which eventually causes the driveway to sink in sections or develop potholes.
•Just like a road or highway, the driveway needs a 1 percent slope or crown to let water drain off. This is especially important in cold climates so that water won't penetrate the asphalt and freeze.
•Depending on the climate and usage, your driveway needs to be resealed every few years to keep it looking good and preventing water from penetrating. Always use a reputable contractor. Cost should be in the $250 to $400 range.
CONSUMER ALERT
For some reason scams involving asphalt resealing or repair are too common. Don't go with a door-to-door salesman, and don't fall for the "leftover mix" scam. Also beware of a salesman who:
•Pushes you for a quick decision.
•Offers you a better deal if you pay cash.
•Offers no written contract.
•Whose offer appears too good to be true.
_Concrete_ is another great driveway option. It's durable and there are lots of options if you want to jazz it up. It can be tinted, embedded with decorative stone and brick, textured, formed to look like bricks, tile, slate, and so on.
One problem many homeowners have is not keeping their concrete driveways clean and free of oil stains. Few things can detract from curb appeal more than dark stains on the driveway from oil-leaking vehicles. There are number of oil-stain-removing products available from auto parts stores. For an environmentally friendly cleaner, try a solution of baking soda and water. Make a paste, apply it to the stain and cover with plastic for a day or two, and then rinse.
If you have a concrete driveway and it's not sealed, this is a project that should go to the top of your to-do list. Sealing a driveway requires just two easy steps:
1.Thoroughly wash the concrete surface with a TSP (tri-sodium phosphate) solution according to package directions.
2.Apply a concrete sealer (available from home centers) with a paint roller, squeegee, or sprayer according to product directions and allow to dry.
Sealing a driveway can extend its life considerably because it prevents water from penetrating the concrete—especially through cracks and seams—and freezing.
Cracks and minor problems should be fixed promptly before the damage gets worse. Spending less than $25 for patching material beats spending several thousand dollars replacing a driveway destroyed by weather.
The following steps outline how to repair minor damage:
•Remove the damaged and loose concrete with a cold chisel. After you've cleaned out all the damaged material, bevel or undercut the edges of the opening you've made with the chisel slanted slightly away from you. This is so the patching material will flow under the edge and create a more stable patch.
•Clean and hose out the area you're patching. Sponge up the water in the damaged area, but leave it a little damp.
•In a small bucket, mix a cement slurry (pancake-batter consistency) of Portland cement and water. Thoroughly coat the area you're patching with this slurry; it acts like a primer for the main patch filling.
•In a large bucket mix one part Portland cement, three parts sand, and enough water to make a stiff paste. Pack this into the area you're patching with a putty knife or small trowel. Mix up the filler to make sure you don't leave any air pockets.
•Level the cement fill with a straight-edged trowel and allow the patch to set for about an hour. Then float or trowel the finish to match the surrounding surface.
•Allow the patch to set for a couple of hours and then cover with plastic. The cover will stay on for about a week while patch cures. Meanwhile weigh down the plastic so it will stay in place.
•The final step is to sprinkle a little water on the patch daily to keep it moist while it cures.
In lieu of buying a bag of Portland cement and a bag of sand, you may be able to get a bag of concrete patch mix; just add water like to a cake mix and you're ready to go.
CHECK OUT THESE PATCHING AND RESURFACING WEBSITES
www.askthebuilder.com/ConcreteDefects.shtml
www.acehardware.com/sm-repairing-and-patching-concrete—bg-1283399.html
www.pavepatch.com/?gclid=CMqI0vW9sYkCFQLYYgod7hneVQ
www.concretenetwork.com/jim_peterson/spray_top.htm
www.interstateproducts.com/concrete.htm
Asphalt driveways need resealing every few years. You can hire a professional for a few hundred dollars, or you can buy sealer at any home center; it's a messy job, though. Also available are tubes of crack sealer you use in a caulking gun. Make sure water doesn't penetrate underneath the asphalt and freeze, causing the familiar potholes you bounce through on roads.
### Patching Steps and Walkways
_Concrete steps_ and walks are high-visibility items. No one can enter your home without noticing problem areas. Chipped corners and damaged steps are ugly—not something you want messing up your curb appeal.
You've really got two options: 1) Hire a contractor to tear out the problem and redo, or 2) if it isn't too bad, save a lot of money and patch the holes and chipped steps yourself. If you go the number two route, you can save a lot of money, and it isn't that difficult a job.
Remember the formula for patching the driveway: one part Portland cement mixed with three parts sand and enough water to make a stiff paste? This same mix and technique works for damaged walkways.
Repairing steps and porches uses this same concrete mix, but the technique is different. For fixing damage on the step edge:
•Using a cold chisel and small sledge hammer (don't use a claw hammer for this, remove the damaged concrete from the stair lip back far enough to get all the damaged area. The chipped out area will look like a rough L-shaped channel along the front of the step. Spray the area you chipped out with a hose to clean out all the debris.
•Make a plywood form the height of the step riser and brace it against step face with 2 × 4s to hold the repair mix in place. (Duct tape also works well for holding the form in place.)
•Mix a batch of concrete mix and a small amount of the slurry mix. Coat the area to be repaired with slurry mix.
•Fill in the channel you chipped out behind the plywood form with concrete mix and smooth it with a trowel so it's even with the rest of the step.
•Let the mix set up for about an hour. Then, using a concrete edging trowel (a trowel with a small lip it for edging concrete) create an edge on the step that matches the others.
•Keep the patch moist by lightly spraying it with a hose and keeping it covered for about a week until it cures.
_Chipped corners_ are fixed the same way: Make a plywood form to contain the patch and use the above procedure. In fact, you can repair most chips and cracks in concrete with this technique.
One problem that's more difficult to solve is getting the new patch to match the weathered concrete. If it's the porch you've done a lot of restoration on, the best solution is stain or paint the entire porch to match. Paint stores have a wide variety of concrete coatings. Light grey and tan, for example, are attractive.
Other solutions that have proven successful—depending on your climate—are covering the surface with tile, slate, wood decking, or other durable materials. One enterprising homeowner rented a saw with a concrete cutting blade and thin-sliced brick about ½-inch thick. He then laid the red brick slices like tile over the walkway and porch using white grout. An ugly stained and cracked porch became an attractive focal point for the entryway. The only limit is making sure it looks good and is durable.
### Dry-Laid Walkways
A popular alternative to concrete walkways is laying pavers over a sand bed. Home centers sell these concrete pavers in a wide variety of interlocking and lattice patterns. The lattice patterns are sometimes installed so grass will grow up through the openings creating an interesting walkway or RV parking pad.
The decision you need to make when putting in a paving system is the type of edging you want to use. Dry-laid pavers need a edging system to contain the pavers and keep them from slipping apart. Typical edging materials are brick, treated wood, concrete edging, metal, or plastic. Local home centers usually carry various edging options that work with their pavers.
Dry-laid pavers are a great way to go for homeowners who want a DIY project that doesn't require a lot of equipment or expense. It's easy to put in a good looking walkway. The steps are:
1.Outline the walkway you want with chalk, paint, or a chalk-line and stakes. It can be straight or curvy, and you can add bump-outs for benches, yard lights, a flagpole, potted plants, or whatever.
2.Remove any grass and excavate the walkway about 10 inches deep so you end up with a smooth bed free of rocks, roots, or debris.
3.Install the edging.
4.Install a bed of gravel or crushed stone about four inches deep for a base.
5.Lay landscaping fabric over the base. Many homeowners skip this step and end up with weeds and grass growing up through the pavers.
6.Over the bed and fabric lay about two inches of sand.
7.Lay the pavers on the sand bed and tamp down lightly with a rented tamping tool. Once you've got the walkway the way you want it, tamp sand into the joints or spaces between the pavers.
You can also use this system to create patios, garden walkways, or edging along fences, driveways, and the foundation. Your imagination is the only limit to the projects you can create.
### FENCES AND HEDGES
Another important component of landscaping and curb appeal is fencing or other dividers. It's often said that good fences make good neighbors; sometimes that's true but other times they become flash points for neighborly misunderstandings.
Some important things you should know about fences and boundaries are:
•Fence heights and setbacks are often regulated by zoning laws. Before you build or renovate a fence, check with zoning. Many times existing fences don't conform to current laws.
•Materials and fence type may be regulated by historical commissions, homeowner's associations, deed restrictions, and subdivision covenants (called CC&Rs).
•Hedges and trees that grow on lot boundaries are also regulated.
•In some areas, it's customary for neighbors to share the cost of maintaining the fence; in other areas it isn't. It's always a good idea to work with a neighbor and get their input when considering putting up a fence or renovating an old one.
If you're thinking of putting up a fence or a boundary hedge, look first at what's popular in your area. For example, you may love cedar fences but if everyone else has installed vinyl, you may want to follow the crowd. The value of your home also comes from neighborhood appeal. A neighborhood with a mix of fence styles, colors, and upkeep isn't as attractive on one that ties together attractively.
Also, older fences that are eyesores can be stained or painted, and rotting posts and broken slats can be replaced or even shortened to make them attractive again. It's amazing what stain or paint and a little imagination can do in restoring a wood fence.
Hedges are often attractive, but they must be trimmed often and neatly or they quickly become an eyesore that can significantly detract from a home's curb appeal.
In one situation, an investor bought a home for a rental that had a rusty chain link fence along one side, which was an eyesore. Not wanting to dig out the posts, which were set in concrete, the owner planted several Virginia creeper vines along the fence and let nature take its course. Within a year the vines had covered the fence and transformed it into an attractive hedge. Vines, ivy, and other creeping plants can make great cover-ups for eyesores.
Table 6-1 is a roundup of curb appeal killers and pet peeves gleaned from Realtors and appraisers, as well as from comments by home buyers.
_Table 6-1. The ten biggest curb appeal mistakes homeowners make_.
**Curb Appeal Problem** | **Suggested Action**
---|---
Yard Clutter | Don't leave hoses, toys, tools, or yard equipment lay about the yard. Invest in a shed to store these items. Most Realtors and buyers agree that the garage is not a good storage area for these items.
Small Items That Really Count | Mail boxes, house numbers, porch lights, and door knockers or bells are important to curb appeal. Visitors and home buyer will notice a banged-up mailbox or a doorbell handing by its wires. Make a to-do list and tackle it.
Bad Driveways | Appraisers knock down value, Realtors hesitate to show, and friends notice and wonder when you're going to fix it. The longer it goes the more it's going to cost you. This should be a top priority before the big screen TV.
Lawn in Poor Condition | The neighbors notice and hope you'll solve it because it affects the neighborhood. A local nursery can advise you on how to solve most lawn problems.
Runaway Landscaping | Appraisers and Realtors find this to be a real problem. It not only lowers value, but makes a home harder to sell. A couple of weekends and sharp cutting tools will do wonders. Neighbors will appreciate it too.
Peeling Paint | This too is a problem that gets more expensive the longer it goes. Failed paint lets water infiltrate, causing more damage to the house, fences, and other wood surfaces. Paint as soon as possible.
Trees Obscuring Home | A very common mistake. When you plant a tree, plan for the size it will be, not what it is now. Trees should frame a home, not tower in front hiding the home.
Fences in Bad Condition | Fences are an important part of curb appeal. Peeling paint, broken or missing slats, and rotting posts need to be fixed and replaced.
Landscaping Not Graded | This an important area with appraisers and home inspectors. Improper grading will channel water to the foundation, and that's not good. Consult a landscaping professional if needed. This problem can lower a home's value big time.
Walks, Decks, and Porches in Bad Condition | These can hold up an appraisal, result in a bad inspection report, and cause buyers to walk away from making an offer. It's also a safety issue if a visitor is injured. Patch or redo as per suggestions in this chapter.
## CHAPTER 7
## HIRING AND WORKING WITH CONTRACTORS
Few things cause homeowners more frustration, headaches, and money than dealing with contractors, subcontractors (called subs), and small job handymen. Horror stories of botched jobs and misunderstandings abound. But, interestingly, they're mostly the result of communication failures on both sides and of homeowners having unrealistic expectations.
Even top-notch builders face the same homeowner frustrations as the rest of us. Yes, they have dripping faucets and cracked driveways too, but pulling a plumber off the job that's a week behind to fix their problem isn't going to happen. They know they have to deal with their client's problems first.
The gritty truth is that no homeowner has the skills or time to fix or replace all the home's components that fail or wear out. So we must rely on skilled (we hope) tradesmen to come to the rescue. How to find and work with these contractors, subs, and handymen is the focus of this chapter.
If the thought of dealing with a contractor intimidates or leaves you in a cold sweat, you've come to the right place. The following pages will show you how the entire process can be made much easier and save you a lot of money. Some of the things you'll find exciting are:
**Tips on how to find a contractor or tradesman**
**What questions you should ask**
**How to pay contractors and control costs**
**How to fire a contractor who isn't performing**
**The ten biggest mistakes homeowners make and how to avoid them**
### HOW TO SHOP FOR A CONTRACTOR OR TRADESMAN
Your first step is to find out whether the contractor is experienced and licensed in the type of work you want done. Every state has different licensing requirements. To find out what your state requirements are, go to www.contractorslicense.org and click on your state. If you have a specific contractor in mind, there's a link to verify that the person is qualified and licensed in your state. Three reasons why you don't want to shortcut this step are:
1.Obviously, you don't want someone who isn't experienced doing the work you want done. Some apprentices may tell you they're masters or journeymen, when they're not.
2.Contractors who are licensed will want to protect their license status and are less likely to skip out on a job.
3.Making sure the contractor is bonded and insured guarantees that if they or one of their employees are injured on the job you won't get a hefty bill or lawsuit.
Your second step is to build a list of possible contractors to contact. Some sources of referrals are:
•Friends, neighbors, and coworkers can be a good resource if you ask them for names of contractors they've used and had good experiences with.
•Architects, mortgage lenders, Realtors, title people, and building supply or home centers can also point you in the right direction.
•Local building associations can sometimes offer a list of members in good standing.
WEBSITES FOR MORE TIPS ON FINDING A CONTRACTOR
www.ftc.gov/bcp/conline/pubs/services/homeimpv.htm
www.hiringacontractor.com/en/default.asp
www.naturalhandyman.com/iip/infcontractor/infhiringcontr.shtm
www.bbb.org/alerts/article.asp?ID=223
www.oldhouseweb.com/stories/how-to/hiring_a_contractor
When Ron and Teri decided to add on to their home, a 1980s Ranch, and create a master suite with a nursery by bumping out a wall, they hired an architect to work with them. It would cost more (about 15 percent) but they wanted the addition to blend in with the rest of the home and not look like a tacky add-on.
These homeowners felt that the extra cost of having professional help in the design and building of their addition would more than add to the homes value. Other similar homes in their area had been upgraded or added on to, and recent sales showed that they could expect almost a 100 percent return if down the road they decided to sell.
With plans in hand and a rough idea of what it would cost, Ron and Teri's next step was to find a contractor to make their dream happen. They started out with two referrals from their architect and added another from a neighbor who had recently added on to her home. Over the following week they collected five more names from friends, contractor signs on job sites in the area, and one from a local cable TV ad.
This gave Ron and Teri eight contacts on their short list to call and try and set up appointments. The next week was frustrating. Out of the eight, two had declined to bid saying they were too busy with current jobs. Three more said they were interested but it would be five to seven months before they could take on new jobs. That left three who agreed to drop by and take a look at the job and work up bids.
Of course before these interviews took place, Ron and Teri had to:
1.Check with their state contractor's license bureau or city building department websites, or call to verify that the contractor's license was active and in good standing.
2.Check with their state consumer protection agency or local Better Business Bureau to see whether any complaints had been filed. One complaint may not be a problem, but several would eliminate that contractor from consideration.
3.Verify that the contractors were bonded and insured.
As Ron and Teri learned, you may find that getting to that short list isn't going to be easy and may take persistence. The best contractors are busy and may not even return your call or may decline to talk to you about your project.
CONSUMER FYI
The best contractors and trades people usually have all the work they can handle. They don't have to go door-to-door to find business. Plus, when they do a job they seldom have material left over. Sloppy estimating is a fast track to going broke. So when someone knocks at your door and says they'll give you a good deal because they have material left over from a job nearby, don't fall for it.
You can sometimes gain a few points by telling contractors up front who recommended them. Many business people are more responsive if they were recommended by a past client or by someone whom they know.
It's a truism that the best contactors work mostly on referral; word of good work travels fast and it's not long until you'll have to stand in line for their services. Even so, it's sometimes hard to identify the right tradesman for your job, but you can narrow down the odds in your favor by doing your homework.
Once you've got several contractors lined up who have agreed to take a look at your project and give you bid, set up a time to meet with each one.
Plan to have extra sets of plans, specs, photos, drawings, and anything else you can give the contractors to use in working up a bid. You won't make a decision at this meeting: That won't happen until you've gone over your notes, talked it over with your partner, listened to your gut feelings, and looked over the bids when the contractors get back to you.
### QUESTIONS TO ASK DURING YOUR INTERVIEWS
Don't hesitate to ask about anything you don't understand, especially about terms and builder jargon. Remember, the only stupid question is the one you didn't ask. The following are some important questions you want answered up front, so be prepared to take lots of notes:
1.Ask how many years they have owned the business? Contractors have a high failure rate the first several years, and it's a good sign if they have survived and prospered more than five years.
2.Ask for a copy of their construction contract. A well-written and professional contract is a big plus. However, you may still want to have an attorney look it over (there are attorneys who specialize in construction contracts), especially if it's a big job.
3.Ask for details of their insurance coverage. Ask to see proof of workers' compensation coverage, liability, and umbrella policies.
4.Ask each contractor to take you to see or set up a time for you to see a couple of completed jobs similar to yours. The relationship between the contractor and past clients is good indication of how the job went. If the job is currently ongoing, notice how the living conditions for the owners are maintained. Are they able to live there as comfortably as possible?
5.Get a list of the past five to ten jobs, along with the home owners' names, addresses, phone numbers, or e-mail addresses.
After you've talked to the three contractors on your short list, the next step is call several of their past clients and see what their experiences were. And no, these past clients won't be offended when you call or e-mail them. Most people are flattered when others seek out their opinions and they have a chance to help out a fellow homeowner. The most common result is that you won't be able to get off the phone, and some will even invite you to look at their remodel.
To make it easier for you, here's a list of questions you can choose from (pick out those that you feel comfortable with) to ask when you call a contractor's past clients:
•Did you enjoy working with ____________?
•What did you like best about working with ___________?
•Did the job turn out as you envisioned?
•Were they consistent and on time?
•Did they return telephone calls promptly?
•Did you have any changes, and if so, how well did the contractor handle them?
•Did the project start and finish as promised?
•Are there any unresolved issues?
•Would you hire them again for another project?
•Would it be possible for me to take a quick look at your remodel?
Granted, going this route does take some time, but then ask yourself: _What's the alternative?_
After Ron and Teri went through the interview process and called over a dozen people referred by the three builders and visited several past jobsites, a pattern started to emerge.
The contractor who advertised on local cable TV got the job done, but he wasn't consistent and sometimes pulled workers off the job to meet deadlines elsewhere. The contractor referred by a neighbor also did a good job, finished on time and budget, and his past clients were happy with their experience. The third contractor, referred by Ron and Teri's architect, also had happy past customers, although he tended to recommend changes that upped the total cost.
### DECIDING ON WHICH BID TO ACCEPT
You may think that choosing among competing bids is going to be like computing a moon-shot trajectory, but in reality they are seldom grouped that close. Here are three tips that'll help you through the process:
•Low bids are not likely to be top quality or the best choice. Perhaps the estimator missed something, or the contractor will end up coming back for more money to complete the project. Be suspicious of any bids that are substantially lower than the rest.
•The highest bid isn't a guarantee of good work or the best choice either. Some builders submit high bids when they don't want the job; it's a good way to say no without offending anyone.
•Most often bids in the middle are more realistic and the ones you should focus on.
In Ron and Teri's remodel, the three contractors got back to them within a week with their bids. The first contractor's bid (the one who advertised on local cable) was lowest at $86,747. The second bid, from the neighbor's referral, came in at $97,453. The third bid was $110,500.
It appeared that the architect's referral wasn't serious about getting the job, and the lowest bid left the homeowners uncomfortable, especially when a couple of past clients complained of inconsistent work flow. Ron and Teri decided the $97,453 bid from the contractor who did their neighbors remodel was the one they would seriously consider.
### GOOD PAPERWORK IS CRITICAL
The next step before making a commitment is to look over the contractor's contract. Putting everything is writing is so critical that you should chisel the following on a granite tablet:
**If it isn't in writing, it doesn't exist.**
That means that all extensions, change orders, and new agreements should be written on addendums, signed by both parties, dated, and numbered 1 of X.
A good contract spells out everything clearly and should include the following:
•Timing—how long the job will take. You should have starting and ending dates clearly spelled out. Also, find out whether the contractor is working on multiple projects and will use unfamiliar subs on your job.
•Who gets the building permits and pays the fees? Usually the builder takes out the permits and you reimburse the fees, or they can be part of the bid.
•Who will supervise the subs, handle problems, and be boots-on-the-project to make sure everyone is on the same page?
•If you're using an architect, has the contractor worked with that person before? A good relationship is important to a smooth job.
•You should have a paragraph spelling out how the contractor will protect the home from construction dust and debris, such as using 6-mil plastic sheeting or tarps supported by a framework enclosing the work area.
•Change orders are a biggie. What happens if you see something that looked great on paper but doesn't look great as the project takes shape?
•Does the builder have a workmanship guarantee and for how long is it valid? What does it cover, specifically?
•What happens if there are misunderstandings. Does the paperwork have a provision for binding arbitration.
•Are the work and materials spelled out in detail. For example, never go with a contract that simply says: "Tear out old deck, dispose of debris, and build new 12 foot by 14 foot deck." It's much better to be specific:
**Remove old deck, dispose of debris, install new footings, posts and handrails. Decking to be 2″ × 6″ cedar, custom knotty grade, and finishedwith two coats of Sherwin Williams DeckScapes semi-transparent, medium-oak stain. Work and materials to be in compliance with all local building codes.**
•How the money is to be paid is, of course, important. There's usually a deposit that shouldn't be more than 10 percent to seal the deal and get the ball rolling. Plan on three to six payments payable as project benchmarks are completed, with the last 15 percent due upon your approval of the job. All payments should be by check or credit card; never, never pay with cash.
•A critical part is proof that the contractor has in force personal liability, worker's compensation, and property-damage policies. Copies of the coverage certificates should be attached to the paperwork.
•You should also have a waiver that frees you from responsibility for litigation or judgments arising from the project. This is important to keep you protected from liability arising from on-the-job injuries and subcontractors' liens (in many states if the contractor doesn't pay his subs, they can place a lien on the property).
MORE INFORMATION ON CONTRACTS
www.b4ubuild.com/resources/contract/index.shtml
www.njsbf.org/njsbf/publications/construction.cfm#2
www.soundhome.com/topics/topic_badcontracts.shtml
www.getvendors.com/household/Residential-Contractor.htm
### PAYING THE CONTRACTOR AND MANAGING COSTS
It's not always easy handling the money side of your project. Both you and the contractor are probably wary of getting burned, so it's important to work out a payment plan you're both happy with.
One option is called a _cost-plus_ , or time-and-materials, contract, It specifies that you pay for all materials and the contractors are paid an hourly fee in addition to 15 to 35 percent markup for the contractor's overhead and profit.
The final bill would list hours worked as well as paid material invoices from all the suppliers. This is good for the contractors because they're guaranteed to make a profit, which is built into the paperwork. You tend to get good work this way, but costs can easily escalate over budget.
From your end, the biggest downside is that you don't have control over the quality of materials or the costs. To make this arrangement work best, you would need to track all the bills and stay on top of the project to make sure the contractor is working efficiently. There's no real incentive for the contractor to save money.
CONSUMER FYI
If a contractor asks you to pay for a job up front and in cash, you're being scammed. Some states regulate how much deposit a contractor can request. If your state doesn't, never go more than 10 to 30 percent. You may need to go the higher amount if expensive custom materials need to be ordered.
In reality, cost-plus works best for small or handyman type jobs that run a few thousand dollars at most.
Another alternative is the _fixed-fee_ contract. With this arrangement, there are usually no surprises at the end. There's usually a mutually agreed-upon allowance for unforeseen changes—$1,000, for example—and everything over this amount you pay for at the time of change. In new construction, these are called _change orders_ , and the clients typically pay for these up front or before the change order is completed.
From the contractor's side, if the job is completed for less than bid, the profit goes up. On the flip side, unforeseen problems or sloppy bid-work may result in the contractor losing money.
For the homeowner the pluses are that you know what the costs are and that changes or additions are handled up front for a known cost. The biggest downside is when the contractor underbids or runs into problems, he may cut corners or rush the project to avoid losing money.
As you might expect, this is one reason why you should be wary of bids that come in significantly lower than the others.
Hybrid contracts that combine features from cost-plus and fixed-fee payment plans are sometimes called a _capped cost-plus_ , with or without a split. For example, you can add a cap to a cost-plus contract. This option is a time-and-materials agreement with a guaranteed maximum price. If costs stay under the max, the homeowner saves money; if the costs run over, the contractor pays the difference.
The downside to this type of hybrid is that the homeowner gets the savings, which means the contractor has no incentive to cut costs. However, you can create an incentive by giving the contractor 35 to 50 percent of any savings. In reality, this gives both sides an incentive to keep costs down, and in the final tally the homeowner would not be likely to lose any money but could end up with a better job. However, it does entail a higher level of involvement on the homeowner's part. This is not an arrangement where you give the contractor the keys and go off on vacation for a month.
### ADDING SWEAT EQUITY TO THE PROJECT
If you have the skills and time you can save some money by doing demo, painting, or sheetrocking yourself. The contractor probably won't go along with you doing code work, such as electrical, plumbing, or roofing; his license is on the line and if you screw up it can cause problems. However, the grunt work is a good pace to save a few bucks.
You'll need to negotiate with the contractor up front exactly what work you'll do. This should be listed specifically on an addendum attached to the construction contract so there'll be no dispute as the project progresses about who does what.
The best time to talk to the contractor about sweat equity is during the bid meeting. Have him work up two bids, one with your participation and the other without it. One contractor who receives many requests from clients wanting to add their sweat equity, says most end up backing out when they look at the bids and realize the time and commitment needed. Still, if you're committed and have the time, you may be able to save up to 25 percent of the total cost, depending on the project and your skills.
INTERESTING SWEAT-EQUITY WEBSITES
www.doityourself.com/stry/sweatequity
www.askthebuilder.com/017_Homeowner_Sweat_Equity_Jobs.shtml
www.ezinearticles.com/?FSBO-Sweat-Equity&id=390433
www.make-my-own-house.com/index.htmlaltytimes.com/rtcpages/20060403_sweatequity.htm
Supplying materials may seem like another good idea to shave off a few bucks, but in reality this rarely saves money and can be counterproductive. It's important that the materials and fixtures fit and install as planned. The contractor is responsible, but he can't warrant whatever you supply. Buying fixtures at close-out sales or at a demo yard isn't always the best way to go.
You may also be tempted to talking the contractor into hiring your out-of-work sheetrocker brother-in-law. But the contractor is likely to nix that idea; not because he doesn't like your brother-in-law, but he doesn't know his work. Most contractors use only subs and suppliers they know. That not only makes the work go smoother, but it lowers their liability and possibly yours.
### TIPS THAT MAKE YOUR PROJECT GO SMOOTHER
Because you're paying the contractor a lot of money, you want the job to go as smoothly as possible without a lot of additional costs. But all too often homeowners are their own worst enemies. The following six tips will help you keep the work going smoothly:
1.Never approve details you don't understand or are fuzzy on. Sometimes it's difficult to visualize what a home addition looks like from the blueprints, so keep asking questions until you can picture it clearly. The last thing a contractor wants is clients who panic because what they see isn't what they want. If you have a problem visualizing a project, contact the architectural department of a local university and hire a student to build a model from the plans. This is much cheaper than being unhappy with a remodeling job. Also, a good computer program that gives you 3-D views is available at www.punchsoftware.com/index.htm.
2.Don't continually change your mind. If you present a moving target to your contractor, you may well end up with a prolonged and way-over-budget project. Do all the necessary planning and shopping for fixtures and materials before the job starts, not afterwards. Keep in mind that change orders are a contractor's profit center and your cost-overrun.
3.Respect the chain of command. If you need to change or add something to the project, go to the contractor. Trying to work through a tradesperson or sub can come back to bite you. If a problem arises, the contractor may not warrant it. Remember, all changes need to be in writing.
4.Have a regularly scheduled meeting with the contractor. Depending on the complexity and size of the job, you may want to have brief meeting at the end of each day, or a weekly update. It's important for you to stay on top of what's happening along the way so you won't have surprises at the end.
5.One thing you want to avoid is letting a disagreement become confrontational. You may have to go the extra mile to maintain your cool, but it's worth it. That a roofer accidentally drops tar on your new Lexus is not license to throw a tantrum. True, you need to be firm and make sure the contractor keeps his promises as per the contract, but letting a disagreement dissolve into a confrontation can sabotage your project.
6.Once you've hired a contractor, don't second-guess your decision by asking third parties their opinion of the job. Many times there's more than one way to do a job, and getting conflicting opinions muddies up the water and your project. Stay the path you've committed to unless there's good reason believe a problem exists. Good communication between you and your contractor is key to dispelling unwarranted misgivings.
### WRAPPING IT UP WITH A WALK-THROUGH AND A PUNCH LIST
No matter how smooth the project's first 95 percent goes, it's the last 5 percent that carries the most emotional weight. Likely, the contractor is in the process of moving on to the next job, and you just want it wrapped up so you can live in a construction-free home.
The key to the final phase is to schedule a walk-through with the contractor so you can work up a punch list. A _punch list_ is a list of things that need to be done to finish the job before you sign off on that final 10 or 15 percent of bid amount.
MORE INFORMATION ON PUNCH LISTS
www.servicemagic.com/article.show.Punch-List.13261.html
www.remodeling.hw.net/industry-news.asp?sectionID=302&articlelD=211757
www.ownerbuilder.com/punch-list.shtml
www.soundhome.com/topics/topic_building.shtml
If you've stayed on top of the project and solved problems along the way, the punch list should be short, and the problems should be quickly taken care of.
A long messy punch list is often the result of not reviewing each day's work, not communicating regularly with the contractor, or letting problems slide until the end. Certainly, you don't want to be a pest on the job, but a quick inspection with the contractor at the end of each day or phase yields big dividends at the end.
However, before you do that walk-through you need to do some homework:
•Before the meeting, review your contract and blueprints and make sure you have copies of all change orders. Create a simple checklist to make sure all the change orders have been completed.
•Walk through the jobsite and jot down anything that needs to be completed, any problems, and any questions that you have.
•Look closely at the finished work. If it's done to your satisfaction and looks good, it probably is. Good finish work is usually the trademark of a good contractor. If you don't trust your judgment, you can hire a professional home inspector to walk through with you.
•Other items to check out are: kitchen backsplash to counter fit, painting dings, trim around doors and windows, counter-tops, and appliances.
•Are the premises left clean and construction debris hauled off?
•Make sure you have a least a half of gallon of paint left for future touch up.
•If there's been an electrical upgrade, have the circuits labeled so you know what the breakers control.
When you meet with the contractor and have done your homework, the punch list walk-through should go quickly. You'll probably find a few minor things that need fixing, which can be done in a day or two. But suppose it's more serious and you find quite a few problems that aren't so minor. In that case, keep the 10 to 15 percent of the bid amount in your checking account until the punch list is completed to your satisfaction.
Suppose everything checks out except for a cabinet that was damaged during installation? In this case you might want to hold out twice the cost of the cabinet and installation. This gives the contractor incentive to follow through or you enough funds to have it done by someone else.
Once the job is complete, you need to create an organized file for all the paperwork. You'll need it when you sell the property, for tax filings, insurance claims or policies, appraisals, and so on.
### WHAT IF YOU HAVE TO FIRE YOUR CONTRACTOR?
Antonio and Jennifer hired a local contractor to enlarge the kitchen and create a family room add-on to their 1980s bi-level. They picked this contractor because they liked him and thought he would do a good job, and because the references he supplied said that he did good work. Feeling that they had done adequate homework and eager to get started, they signed a contract. They planned to have everyone over for Jennifer's parents' anniversary in six weeks.
The rough framing appeared to go smoothly and the roof addition and siding blended with the rest of the home. However, once the work moved to the inside, things quickly fell apart. Interior walls weren't straight and the city's building inspector failed the rough plumbing and electrical upgrade.
The homeowners tried to pin down why they were having problems, but the contractor was defensive and blamed it on his subs, who appeared to be cheap labor hired at a local home improvement parking lot.
The pace also slowed down to a crawl: the week without a kitchen, as promised by the contractor, turned into five weeks of take-out and sandwiches. Finally, Antonio and Jennifer reached the breaking point and told the contractor that he was finished and to leave the premises and not return. They next contacted their attorney and asked him to file suit.
Eventually, the homeowners found out the contractor was not licensed in their state and had no assets they could go after. A short time later he disappeared with $6,000 of their deposit, leaving a stack of bills from suppliers for materials not yet paid for.
INTERESTING FIRE-YOUR-CONTRACTOR WEBSITES
www.doityourself.com/stry/contractorfireh2
<http://ths.gardenweb.com/forums/load/remodel/msg0921272327406.html?8>
www.contractorsfromhell.com
Unfortunate as these situations are, they can often be prevented by following the guidelines in this chapter and by doing your homework.
When you reach that stage, your first option is try to work with your contractor. Perhaps arbitration or a neutral third party can help. Changing builders in the middle of the project is expensive and disruptive. But if all else fails and you need to replace your contractor, your first step is talk to an attorney. You are, in fact, breaking a contract, which can expose you to legal liabilities. If you've been paying only for work done, you're ahead in that the contractor can't disappear with your funds.
However, if it does come to replacing the contractor consider the following:
•Even though the work a contractor does has been unsatisfactory, a lawsuit can tie up the contract amount in escrow, and the job could sit in limbo for longer than you want to live with.
•Try arbitration first, and if that doesn't work, seek for a written termination and full-release agreement with the help of your attorney.
•Make a formal complaint to your state licensing board or local building department. This can sometimes give you some leverage if the contractor faces action from them as well. Some states also have funds that cover losses suffered by homeowners due to wayward contractors.
Finding a replacement contractor presents a whole new set of challenges. Finding a top-notch contractor to take on a botched job isn't going to be easy, and the costs are likely to go up.
First, contractors are going to be wary of the job and of you. They're going to wonder whether you're the client from hell, and they're fully aware that it's expensive to fix someone's else's mess. Plus, they can't predict what it will take to correct the problems, so you may have work on a cost-plus basis.
Still, it's not impossible to find a competent contractor to work with you. Here are some ideas:
•Tell your story, but stick to the facts. Don't rant and rave about how bad the previous contractor was. Stay calm, professional, and focused on the work that needs to be done. Also, make available any paperwork that can help the new contractor access the situation.
•Work up a list of what needs to be done before you solicit bids. You may want to hire an architect, home inspector, or engineer to help you put the list together.
•It's likely that contractors will want to work on a time-and-material basis because of the unknowns, but that leaves you with blue sky over what your projects is ultimately going to cost you. A possible alternative would be split up the project, with payment on an hourly basis for the part that's uncertain and a fixed price for the remainder.
•Learn from your mistakes: Don't shortcut the hiring process discussed earlier in the chapter and make sure the contract has a good termination clause.
### HOW TO BE A SMART BUYER OF HOME IMPROVEMENT SERVICES
The first step in becoming adept at hiring home improvement services is to educate yourself on material, products, and how building components are put together. Good sources are magazines, books, home shows, TV programs, and new subdivisions where you can see craftspeople at work. The more you know about what you're hiring people to do, the better job you'll do.
As discussed earlier, it's important to do your homework when picking a contractor. If you find the right contractor, he will supply the subs for the different job components. This makes the project go smoother, but it will be more expensive.
WEBSITES FOR FINDING AND HIRING TRADESPEOPLE
www.mastertrade.co.nz/MAG_WIRED/index.php?edition=6&article=84
www.money.scotsman.com/scotsman/articles/articledisplay.jsp?section=Home&article_id=2444202
www.tradesecrets.org/forms_publications/other_resources/pdf/hire_certified_tradespeople.pdf
www.realsimple.com/realsimple/gallery/0,21863,1569249,00.html
www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2003/11/22/HO15941.DTL
Suppose you want to hire people to do certain jobs without going through a contractor or large remodeling company? Sometimes this is a good way to go; other times it isn't. But there's a hard and fast rule you should never break, and that's to match the craftsperson to the job. For example, building a deck and adding a wing to your home takes two different levels of expertise.
This is one of the mistakes Lee and Sherri made when they hired their contractor. He had done decks and small remodeling jobs, but had never tackled a large home addition. When the homeowners called a few of the contractor's references they didn't ask whether the project was an addition; they simply asked whether he had done a good job and got it done on time.
The bottom line is this: If you want to save bucks and hire tradespeople to do specific jobs, you need to do the following homework:
•What is the total job you want to do?. If it's a remodel or addition that will require several different trades, you may want to consider a general contractor who can pull everything together and coordinate it, like the conductor of a symphony. A single trade job, such as roofing, windows, or plumbing is another matter. That you can usually handle directly.
•How much of your living space will be affected by the project? If it's considerable, you may want go with a big company that has the resources to put a lot people on the job to get it done faster so you won't miss the Super Bowl.
•How complex is the job? Do you need a specialized company that handles your type of project, such as when repairing fire or flood damage? Specialized remodeling often requires several different trades, and you may be better off going with a contractor with a track record rather trying to hire out the work piecemeal.
•What's the time frame in which you need to job done? If you're not in a rush, a small company may work out. But if you need the job done fast, a larger company usually has more clout with subs and suppliers.
You need to look at the trade-offs: A small one- or two-person company, where you're working with the owner, can give you great customer service and good value. The flip side is that if you need deep pockets, greater breadth of experience, and more resources, then a larger company is a better choice, although it will probably cost more.
### PICKING A PRO FOR EMERGENCIES
There comes a time when we all need a pro wearing a tool belt. These are people who practice a single trade, have experience, and have seen about everything in their trade—and that's exactly what we want when water is gushing or an electrical outlet is sparking.
Every homeowner should develop relationships with three trades: a plumber, an electrician, and an HVAC (heating, ventilation, and air-conditioning) contractor before a panic situation arises. The benefits of doing this are many. For example, an HVAC contractor who is familiar with your heating system can respond quickly and know what parts are likely needed. And it's no secret that a call from a familiar customer will take precedence over an unknown call-in.
Finding these trusted tradespeople takes some trial and error. You can minimize the frustration by looking in the right places. Some suggestions are:
•Word-of-mouth is still the best way to find the top performers. Collect a list of names from friends and coworkers to start with.
•Get references from other trades people. For example, plumbers sometimes need electricians when they do a job. Ask the pros for names of other good pros.
•Contractors will sometimes give you names of subs they use. It's in their best interest to keep their stable of subs happy in slow times or during the off-season.
•Even when you get a dynamite referral, don't shortcut the check-'em-out procedure. That means checking for current license and insurance coverage.
•Check out a tradesperson with a small job first, such as a toilet leak or running an outlet to the patio. If they do well and you like their work, add them to your speed dial.
Of course it's important to build and maintain a good relationship once you find a tradesperson you can trust who also does good work—a rare combination. Here are seven ways to help you get into their preferred customer database.
1. Treat contractors and their helpers with the respect you would give clients or other professionals.
2. Be courteous to office help or family members who take your panicky call. You want them to go the extra mile getting you the help you need, instead of putting you on hold or losing your sticky note message because you were rude.
3. If a tradesperson gives you his personal cell phone number, don't abuse the trust and bug him with minor stuff, or give it out to other people.
4. Don't make every call an emergency. If a repair can wait, let them pick the time to do the work.
5. Show respect for the tradesperson's time, by making the job site accessible so she can get right to work, especially if you promised you would do that.
6. Pay the bill promptly and you will be on the contractor's short list of good customers. Paying when the work is completed is best, but sometimes the contractor says he will figure it out and bill you later. Don't betray that trust; pay the invoice promptly when it arrives.
7. Don't keep good tradespeople a secret. Refer them to your friends and those who you know will be good clients.
### THE TEN BIGGEST MISTAKES
The ten biggest mistakes homeowners make when hiring and working with contractors are these:
1. Hiring a contractor without checking to see if he is actively licensed, bonded, and his insurance coverage is up to date.
2. Not making sure the contractor has the job experience for the project. For example, building dynamite decks doesn't translate into the ability to handle additions or kitchen remodeling.
3. Not having a good contract and failing to put everything in writing. For substantial jobs, letting a construction attorney look over the paperwork is critical.
4. Failing to get change orders in writing and signed by all parties, detailing the material, time frame, and cost.
5. Not having a good termination clause in the contract.
6. Failing to do your homework and know exactly what you want so you don't create chaos with the builder's timetable and go over budget because of change orders.
7. Not having the last 10 to 15 percent due upon final inspection and approval of the project.
8. Trying to get the contractor to use your materials or hire people he is not familiar with, such as pressuring the contractor to use your out-of-work cousin who happens to be an electrician. You're likely to get a chilly response because contractors have their trusted subs, who they know can perform.
9. Asking the contractor or subs to do little jobs not specified in the paperwork. Resist the temptation of asking for "while you're at it can you . . ." additions or changes without consulting with the contractor and doing the necessary change order.
10. Not staying on top of the project and having regularly scheduled meetings with the contractor to catch problems early. Waiting until the end to address problems can create a long punch list and delay closing out the project.
## CHAPTER 8
## TAX ASPECTS OF OWNING A HOME
It's no great secret that homeowners get some pretty good tax breaks. Your mortgage lender, Realtor, and numerous magazine and newspaper article tout the advantages of owning versus renting.
That the first $500,000 is exempt from capital gains for a couple is an enticing incentive to buy a home, live in it for a few years, and then sell and move up to bigger home. This in part fuels the high demand for houses and homeownership, which is good for the economy.
But there's also a flip side: Now that you own a home, you become subject to a hefty property tax on your home's value. How much depends on the state you live in and it can vary widely. Because this tax depends on the value of your home according the county assessor, you want to make sure your home is not overvalued. If you think it is, then you can protest and present your case.
How these and other taxes affect you as a homeowner is the focus of this chapter, where you'll learn:
**All about property taxes and how to protest a tax bill that's too high**
**How to turn your property into a tax break**
**All about capital gains taxes that affect owning and selling property**
**Which improvements are tax deductible and which ones aren't**
**What tax credits Uncle Sam wants to give you for being energy efficient**
### PROPERTY TAXES AND HOW TO KEEP FROM PAYING TOO MUCH
When tax payments become part of the house payment, the tendency is mail a check the first of the month and forget about it. As result, it becomes an out-of-sight/out-of-mind situation. But if you pay property taxes separately, you develop a slightly different perspective. You feel the pain of writing a sizable check on or before November 30th every year, and when the assessment goes up, you are strongly motivated to challenge it. And you should, since according to the International Association of Assessing Officers (IAAO) more than half of the homeowners who protest their assessments get a reduction.
It's important for you to check out how your county handles assessments. Not all states and counties have the exact same rules or quirks. For example, some counties keep your assessment the same if a home is rebuilt on the old foundation (which means you can end up paying the older rate for a nearly new home), while others assess homes based on current market worth.
MORE INFORMATION ON UNDERSTANDING PROPERTY TAXES
www.sctax.org/Tax+Information/property/property.htm
www.ntu.org/main
www.csmonitor.com/2004/1203/p01s01-usec.html
www.usatoday.com/news/nation/2006-08-24-states-property-taxes_x.htm
www.law.freeadvice.com/tax_law/property_tax_law
Understanding how property taxes work is fairly straightforward. There are four simplified steps the county goes through to arrive at the amount they bill you for:
_Step 1_. The assessor's office determines the value of all the properties in the county. They do this by reviewing appraisals, sold records, computer modeling, and building permits. The assessor then adds up all the real estate values in the county to get a grand total.
_Step 2_. If your county uses the full-value approach, the assessor simply totals up the appraised values. But if the county uses an assessment ratio—say, one-half—then you multiply the total value of all taxable properties by .50.
_Step 3_. Each year the county comes up with a new tax rate. This is simply the budget for the money the county needs for the coming year divided by total value of all the real estate in the county. If, for example, next year's county budget is $3,806,000 and the total real estate value for the county is $500,800,000, then dividing next year's budget by the total taxable real estate value in the county gives you a tax rate of .0076.
_Step 4_. In the final step, the county clerk mails out tax notices to the owners of all the property in the county. The amount you owe the tax collector is the tax rate times the assessment of your house. If, using the above example, the tax rate is .0076 and your house is assessed at $300,000, multiplying $300,000 by .0076 equals $2,280, which is your property tax.
If you think your taxes are too high, there are two approaches you can work with to lower them. One, you can go to the county or city budget hearings and challenge how the government spends the money. If enough people get upset a referendum can put a tax cap on the ballot like what happened in California, Texas and other states.
The other approach is make sure your home's assessment is as low as possible. It the tax notice shows a value you think is too high, you can appeal it.
QUICK CHECKLIST FOR EVALUATING YOUR TAX ASSESSMENT
•Make sure you've received all the deductions you're entitled to.
•Check the assessor's math and your home's property description for accuracy.
•Compare your assessment with three to five similar properties.
•Make adjustments between your home and the comparable properties.
If you feel you have a case after doing this homework, make an informal appeal to the assessor.
### APPEALING YOUR TAX BILL
Suppose your tax notice values your house at $200,000. You feel that this is high because you had an appraisal when you refinanced two months ago for $185,000. Then you happen to talk with your neighbor whose home is a bit larger and has a bigger yard and his tax is $200 less than yours. You wonder how that could happen?
Easy. According to Consumer Reports, tax records show an error rate of 40 percent in estimating property taxes. Also, the National Taxpayers Union writes that as many as 60 percent of all homeowners are over-assessed on their home's value. If this is the case, then most likely you're paying more property tax than you should. Unfortunately, to correct this and put some dollars back in your pocket, you have to be proactive and prove to the county that they owe you money.
### What Are the Grounds for an Appeal?
First, an assessment appeal is not a complaint about higher taxes. It's an attempt to prove that your property's estimated market value is either inaccurate or unfair. For a successful appeal you'll need to prove at least one of three things:
1.Items that affect value are incorrect on your property record. For instance, you have one bath, not two. You have a carport, not a garage. Your home has 1,600, not 2,000 square feet, and so on. This is an easy one, you just compare the assessor's record to what you've really got and submit proof such as a recent appraisal or photos.
2.You have evidence (comparables) that similar properties have sold for less than the market value of your property.
3.The estimated market value of your property is accurate but inequitable because it is higher than the estimated value of similar properties. If your home is in a subdivision of similar homes and your neighbors' taxes are less, then you have a case for appeal. A title company, county recorder, Realtors, or websites like Zillow.com all can give you data upon which to build a case. When you receive your assessment notice, read it for instructions about deadlines and filing procedures. They vary from state to state. If they're not clear, call the assessor's office for information. A missed deadline or incorrect filing can cause an appeal to be dismissed.
MORE INFORMATION ON APPEALING PROPERTY TAXES
www.money.aol.com/basics/3canvas/_a/taxes-and-the-homeownerprotesting-your/20050225133909990011
www.firstam.com/list.cfm?id=5588
www.google.com/search?q=how+to+protest+property+taxes&hl=en&lr=&start=70&sa=N
www/realtytimes.com/rtcpages/19990601_proptax.htm
The first step in an appeal is usually an informal meeting with someone in the assessor's office. (Sometimes this informal review is handled by telephone or mail.) Information on the mechanics and deadlines for setting up an appointment should be included with your assessment notice, along with additional information for the entire appeals process.
### Preparing for the Appeal
Find your property identification number on your assessment notice and use this number to get a copy of your property record from the assessor's office or title company.
Next, review the facts on the property record. Is the architectural style correctly stated? If not, a recent photo of your home will help correct the information. Check the living area of your home, lot size, number of bathrooms and bedrooms, garage or finished basement, construction materials, condition, and so on.
Gather as much information as you can on similar properties in your neighborhood. Ask the assessor's office, call your friendly Realtor who sold you the house, or ask someone you know who has access to the MLS. They can print out a list of comparable sold homes in minutes that you can use for ammunition in your appeal.
Compare your home with the assessed values of similar properties and put together a simple chart comparing the homes feature for feature. Some MLS databases can organize the data from comparable homes into columns and make it easier for you to make comparisons. The key is to know what you're talking about and have the proof to back it up.
### Next Step: The Meeting
Follow the instructions on your tax notice and make an appointment with the tax assessor. The purpose of this informal review—which is not yet an appeal—is to verify the information on your property record form and make sure you understand how your value was estimated. Also, the meeting is also used to discover whether the value is fair compared with the values of similar properties in your neighborhood, and to find out whether you qualify for any exemptions.
The person conducting the meeting will probably review your property record form with you and along with information you have about comparable properties. At this time you can present the information you've gathered.
You may not get a commitment for a change in value at this meeting, even though you have uncovered an error or the assessment appears to be inequitable. The decision to change a valuation may have to be made by someone else in writing. If so, find out when you can expect a decision.
In the meeting, view the person you're talking to as an ally, not an adversary. If you're calm, polite, and professional they will likely be more helpful and concentrate on giving you the information you need for an appeal, if it comes to that.
It's possible this is as far as you'll have to go if you get a favorable ruling. If not, then you'll have to proceed to the next step, which is a formal appeal.
### The Formal Appeal
Residential appeals are often settled at the local level. If you are not satisfied with the results of your informal review, you have several more opportunities. The first level of formal appeal is usually to a local board. Here you'll need to again present your evidence and point out the similarity between your property and the comparables on your list. Include a recent appraisal of your property if you have one. Close by asking the board to reduce your assessment to what you think is fair based on your data.
Remember to keep it focused and professional. The appeal board is interested only in the fairness and accuracy of your assessment. Don't go off on tangents about how your aging mother needs a lifesaving operation or why you think property taxes are too high.
If you disagree with the local board's decision, additional administrative or legal remedies are available that vary from state to state. Information about these are available from your assessor's office.
### Beware of Tax Scams
Around tax time, you'll probably get letters offering to reduce your taxes for a percentage of the reduction. Many legitimate companies offer this service. Typically, they'll charge 30-50 percent of the savings. If you don't have time to protest your tax bill yourself, this can be a better way than not protesting at all. As usual, talk to three companies and get bids and references.
However, beware of scammers who charge a fee up front and promise to lower your taxes or get you a rebate. No one can predict whether and what you can save without going through the system. Always be suspicious if there's an up-front fee involved.
### FEDERAL TAXES AND DEDUCTIONS
Over the years you own a home, you'll do a lot of improvements, fix-ups, and repairs. Some of these improvements add to your home's tax basis while others may not. See IRS publication 523 for determining your home's tax basis, or go to www.irs.gov.
Usually, items you replace or repair to keep the home in good condition are not considered improvements that add to its value for tax purposes.
Typical improvements that generally increase your home's value or tax basis are listed in Table 8-1.
Examples of repairs that maintain the home in good condition or prolong its life but don't add to its value are:
_Table 8-1. Improvements that generally add value to your home_.
Repainting the home inside and out
Fixing floors
Fixing rain gutters
Repairing leaks
Repairing walls or ceilings
Replacing broken widows
Replacing carpeting
Any normal maintenance items.
Note that replacing windows with ENERGY STAR-certified upgrades may qualify you for a tax credit.
### TAX ADVANTAGES OF WORKING AT HOME
Thanks to broadband technology and laptop computers, more people can work at home and leave the commuting to the rest of the world. To make this a more viable working option, Congress passed the Taxpayer Relief Act of 1997, which contained a modification of the IRS definition of principal place of business.
Beginning in 1999, the new rules allow those who don't have offsite office space to deduct the expenses of a home office. Contractors, sales reps, consultants, and other who perform their services outside their office but need a home office can benefit. However, you must use the office exclusively and regularly for business. In addition, if you use part of your home for business, such as storing records, inventory, or samples, you may be entitled to a deduction.
You can also convert an existing structure or build a separate structure that's not attached to your house. Detached garages, carriage houses, sheds, or small barns make great home office conversions.
### What You Can Deduct
Depending on the percentage of your home you use for business, you should be able to deduct portions of utility bills, mortgage interest, repairs, depreciation, cost of a second phone line, office equipment, and any other related expenses. You may also be able to depreciate computers, equipment, and office furniture. However, it's important to set aside the area you use for business. You can't put a filing cabinet in the family room that's used for watching TV and call it an office. The area must be used for business only.
If you decide to claim a home-office deduction, you should keep meticulous records of all your expenses and be prepared to back them up if you're asked to by the IRS.
MORE HOME BUSINESS INFORMATION
www.businessweek.com/smallbiz/0001/tx3666150.htm
www.bankrate.com/brm/news/biz/adviser/20020228a.asp
www.homebusinesstaxsavings.com
www.smallbusiness.yahoo.com/r-subCat-m-1-sc-9-getting_started_home_business-i
www.homebiztools.com/tax-breaks.htm
There's also a downside to creating a home office. If and when you decide to move, you may have spent $20,000 to create a great home office, but chances are you won't be able to add all of that cost to the price of the home.
Before you draw up plans and get bids for that dream home office you saw in a magazine article, do some "what if" thinking. If you turned a bedroom into an office, can it be restored back to a bedroom without too much trouble or expense.
### TURNING EXTRA SPACE INTO TAX BREAKS AND INCOME
When Sharee divorced she ended up with the house and a $1,900 mortgage payment. There wasn't a lot of equity, so selling the house was not an option, nor was letting the home go into foreclosure and ruining her credit.
The home had a finished basement with an outside entrance, two bedrooms, a family room, a full bath, and a wet bar. Sharee figured she could easily covert the wet bar into a kitchenette and rent the basement for $800 a month.
Renting extra space in the attic or basement has spread to more affluent urban areas as well as suburban neighborhoods. Empty nesters, one-income families, widows, and widowers often take in tenants to help pay their mortgages.
Others are adding basement or attic apartments when they build a new home. If you're anticipating caring for parents, or for kids who may come home a second time around, consider adding a finished apartment. If that's not feasible now, at least add the electrical and plumbing rough-in. This is a lot cheaper than retrofitting later on.
### FIRST, CHECK OUT THE ZONING
Your first move should be a call to the town zoning department. In most subdivisions, zoning laws limit houses to single-family occupancy. That doesn't mean you can't take in a boarder or two, although that may violate local ordinances. If the tenant is a relative you shouldn't have a problem with adding an in-law apartment. Most zoning departments, insurers, and mortgage lenders will go along with that kind of addition.
But if you want to create an apartment and rent to a nonrelative, check the zoning first. If that is prohibited you may be able to get a conditional use permit or variance. This entails going to the zoning department and filling out the paperwork. There may be restrictions on how utilities are set up as well as inspections for building-code compliance. A hearing may also be required to give the neighbors a chance to protest.
True, this is a more expensive and time-consuming way to go, but in the end it'll be worth it. Some homeowners ignore zoning rules because the city or county is lax in enforcement, and they know they can get away with it. They possibly can for awhile, but it usually comes back to haunt them. Administrations change and all properties eventually are put up for sale. If they sell or refinance, a mortgage lender may want to see the permits. Or the insurance company may deny a claim if you've violated the policy terms and your home is a pile of ashes.
Also, if you're buying a home with an apartment, office, or garage that has been added on or remodeled, you'll want to verify that building permits exist and the house is zoning compliant. Otherwise, you can end up with fines and expensive upgrades. You can also get extended title insurance coverage at closing that insures the previous owners have complied with all zoning and building permits.
Brandon and Julie had this problem when they made an offer on a home with a basement apartment in a great neighborhood near a university. The apartment had been rented out for years, and several other homes on the street also had basement apartments. No one questioned whether the area was zoned for a rental. However, when the appraiser checked the zoning she found it was zoned single family and adjusted her appraisal down.
Although the city zoning department said they hadn't and wouldn't enforce the zoning in this case, they still wouldn't approve the house and apartment for duplex zoning.
This created a sticky situation. As a single family house the appraisal came back $12,000 under sales price. The sellers were upset because now the genie was out of the bottle and the home couldn't be sold as a duplex, lowering the value considerably. And no mortgage lender would finance an illegal duplex. The buyers, rather than walking away from the deal, offered to buy at the reduced price. Reluctantly, the sellers accepted.
MORE INFORMATION ON RENTING OUT AN APARTMENT
www.insurance.com/Article.aspx/Protect_Yourself_When_Renting_Your_Home/artid/170
www.homes101.net/home-buying-tips/bt74
www.secondsuites.info/Upspercent20andpercent20downspercent20ofpercent20renting.htm
www.archhousing.org/adu2/renting.htmlwww.ezinearticles.com/?cat=realestate:leasing-renting
The bottom line is: Before you buy a house with an apartment or think of adding one to your home, check the zoning and verify what permits are required. It can save you costly tax, insurance, and zoning headaches later on.
### TAX CREDITS FOR MAKING YOUR HOME ENERGY EFFICIENT
When shopping for windows, doors, appliances, and other energy-related items, the IRS will give you a tax credit of up to $500 if you buy products with an ENERGY STAR label. This program is good through 2007, but may be extended.
Even if you don't get a tax credit, the energy-efficient products that qualify for the ENERGY STAR label will save you some serious bucks and make you a more responsible homeowner.
The following products qualify as energy-saving products:
• _Windows/Skylights_. These should have double glazing, low-E coating with insulating gas between the panes. Look for the ENERGY STAR label. Possible tax credit of up to $200.
• _Storm Windows_. These should have double weatherstripping, and have a National Fenestration Rating Council (NFRC) label. Up to $200 tax credit.
• _Exterior Doors_. Look for a polyurethane core and fiberglass or metal skin, foam-filled rubber weather stripping. Look for the ENERGY STAR label. Up to $500 tax credit.
• _Storm Doors_. This should have aluminum frame or vinyl skin over a solid wood core and felt weather stripping. Look for the NFRC label. Up to $500 tax credit.
• _Metal Roofing_. This should include snow guards to control excess runoff and a two-year warranty minimum. Look for the ENERGY STAR label. Up to $500 tax credit.
• _Water Heaters_. The tag should disclose an insulation rating of R-25 and an energy factor of 0.80. Up to $300 tax credit.
• _Insulation_. Applies to all types of home insulation. You'll need the manufacturer's efficiency label. Tax credit of up to $500.
With the emphasis on energy-saving products and the increasing cost of heating and cooling, it's important to look at your home's components and consider replacing older items. It's amazing how often the energy savings catch up with the initial cost in only a few years.
MORE INFORMATION ON ENERGY STAR SAVINGS
www.energystar.gov/index.cfm?c=about.ab_index
www.energy.gov/energyefficiency/energystar.htm
www.energytaxincentives.org
Many states and cities are also jumping on the energy-saving bandwagon by offering their own tax credits and incentives. So before you spend, check and see what's available in your area.
### KEY FACTS FOR HOMEOWNERS
Here are some factoids that you need to be aware of as homeowner:
•The IRS defines your main home as the one you live in most of the time. It can be a house, houseboat, mobile home, cooperative apartment, or condo.
•You must have lived in the home for at least two of the last five years.
•If you have more than one home, you can exclude gain only on the sale of your main home, the one you live in most of the time.
• _Gain_ is defined as the selling price (line 101 on the HUD settlement statement).
• _Amount realized_ is the selling price minus the selling expenses.
• _Selling expenses_ are commissions, advertising, legal fees, and loan charges paid by you for the seller, such as points or closing costs.
• _Basis_ is determined by how you got the home: What did it cost if you bought it, or what is the fair market value if you got it by inheritance or gift?
Adjusted basis can be increases or decreases to your basis depending on what you did to the home or what happened while you owned it. Improvements that have a useful life of more than one year generally increase your basis. Credits for losses, energy-improvement costs, and so on decrease your basis.
### Recordkeeping
Of course, keeping records and receipts of your home improvement projects are critical.
What you should keep are:
•Proof of the home's purchase price and related costs. (These are all on the HUD Settlement Statements you signed at closing.)
•Keep a specific file for improvements, additions, and other items that affect your adjusted basis—that is, the items that add value. This file may also come in handy as a selling tool to document the cost of improvements to a buyer.
•Keep the worksheets (found in IRS publication 523) you used to figure adjusted basis when you sell your home.
•Also keep any other worksheets or paperwork related to your previous home's sale for at least three years, or longer if you have postponed gain.
•If you don't live in the home continuously, you need to have proof of the time you did live in it.
Keeping records also saves you money at tax time. For example, one home-buyer bought an older home that needed a lot of major work. His plan was to live in the home, fix it up, and sell it in a couple of years for a substantial profit.
Knowing that keeping careful track of improvements and costs would be important, the owner-remodeler used Quicken, an accounting program that allowed him to keep track of expenses on his laptop. He also set up a simple system that filed basis-allowable receipts in one folder and nonallowable receipts in another, and he went over his system with a certified public accountant to make sure he didn't miss any allowable deductions.
When the homeowner eventually sold his home, he was able to claim the maximum deductions allowed because he maintained accurate records. In this case, it was critical because the home sold for well over the $250,000 exclusion. This meant that a substantial amount of money was exposed to capital gains taxes, and any missed expenses would have cost the owner money.
TAX-INFORMATIVE WEBSITES
www.bankrate.com/brm/itax/news/20030207a1.asp
www.kiplinger.com/features/archives/2005/01/taxguide2.html
www.quickenloans.com/refinance/articles/homeowner_tax_tips.html
www.energystar.gov/index.cfm?c=products.pr_tax_credits
www.realtytimes.com/rtcpages/20061030_taxwriteoffs.htm
It's very important to keep good paperwork on home-related improvements and upgrades. It's amazing how much money people lose because they don't keep records of their improvements.
### THE USE TEST FOR PRINCIPAL RESIDENCE
The rules for meeting the use test are extremely liberal. To take the tax exclusion of a principal residence, all you must do is show that you owned and lived in the property as your main home for either 24 full months, or 730 days during the five-year period, which ends on the date of the sale.
Short temporary absences for vacations or other seasonal absences, even if you rent out the property during these times, are counted as periods of use.
If you own a cooperative apartment, you must have owned the stock for at least two years and lived in the apartment for at least two years.
There are exemptions to the above rules if you have to sell because of health, employment, unforeseen circumstances, and so on (see the worksheets in IRS Publication 523).
You may also want to keep your home as an investment, Renting it can be a good strategy when the market is slow and you can't sell the home. For example, Dan and Sheila tried to sell their home of twenty years when he got a better job offer in another state. A few months prior, they had refinanced the home and ended up with a loan that was more than what they were able to sell their home for.
Not wanting to give up his job offer, Dan and Sheila decided to rent their home for a year or two in hope of the market improving to where they could break even. After renting the home for over two years, the local market did go up. They were able not only to sell the home but to walk away with a good profit. Because they had lived in the home two out of the last five years, they were able to sell the home without a taxable gain.
Another way to defer taxable gain is with a 1031 exchange. It's not a way to avoid taxes, but it is a great tool for building equity and transferring it from property to property.
### 1031 TAX-DEFERRED EXCHANGES
This underutilized financial tool can make a big difference when:
1.You want to keep your starter home for a rental when you move up.
2.You've rented part of your home and it's subject to capital gains.
3.You want to trade your single family up to a duplex or fourplex.
4.Someone has an investment property (single-family rental, duplex, or even a newly built home) you would like, but if they sold, capital gains would kick in.
The possibilities are endless for creating win-win deals and deferring capital gains to a time when the tax bite is not so painful.
Unfortunately, paying capital gains taxes keeps many owners from selling single-family homes and condos that have increased in value past their exemptions. Not wanting to go through the pain of fixing up the property and putting it on the market, many owners continue living with the problem and procrastinating instead of doing something proactive. As equity grows, the problem grows.
Luckily, a 1031 exchange may be able to solve these problems by getting owners into something more suited to their interests.
Putting an exchange together is fairly straightforward, but it may require the expertise of an exchange intermediary, accountant, and a title/escrow company, depending on the number of properties and the complexity involved. The exchange intermediary is the neutral party that handles the nuts and bolds of the exchange. To find one, look in the yellow pages or check the Internet under "real estate exchange." Better still, Realtors and title companies who do 1031 exchanges will be able to recommend good intermediaries. You'll also need a title or escrow company to handle title work and funding.
The exciting thing about 1031 exchanges is you don't have to have two property owners who want to exchange straight across, you can bring in other buyers and sellers with their properties to add to the mix.
Here's a simplified example: You find a buyer for the property you want to get rid of (relinquished property) and the sale goes in escrow. You have forty-five days to find a property you want to buy (replacement property) and that goes into the escrow. The buy/sell mix closes and you end up with the property you want. The party with the least equity can use cash or financing to make up the difference.
### Profiting from an Exchange
When Norm and Susan were transferred from Utah to Georgia, they planned on returning in a few years. Therefore, so they rented their home and a smaller property they had bought from an estate. However, five years later Norm and Susan decided to stay in Georgia and sell the homes. Unfortunately, they no longer qualified for owner-occupied status, so they decided to invest in local rental property and defer capital gains taxes.
Their strategy was to put the proceeds from the sales into a 1031 exchange escrow with a title company in Utah and find a rental property in Georgia. When they closed on the property where they lived, the funds would be released from escrow and used for a down payment. The balance of the purchase price would come from a nonowner-occupied mortgage.
Luckily, the tenants in the smaller home wanted to buy it and were able to qualify for the mortgage payments, which were $80 less than their rent payment. The sale closed and the $97,000 proceeds—after paying off a small mortgage and selling costs—went into the title company exchange department's escrow account.
While the paperwork for the rental sale was going forward, Norm and Susan were out looking for rentals in their area. The market was tight and they didn't find anything they liked until two weeks after their home in Utah had closed. (The IRS allows 45 days to identify a property and up to 180 days to close the deal.) They made an offer on a two-bedroom condominium in a good area for $185,000, and it was accepted.
The $97,000 in escrow was used for a 20 percent down payment, and the balance was financed with a nonowner-occupied mortgage. The equity from one rental home in Utah was transferred to Georgia with no capital gains taxes.
Norm and Susan's other rental has six months to go on a lease. If the tenants can't or don't want to buy it, it'll go on the market and the process will be repeated.
As you can see, the 1031 exchange is a great way to transfer equity from one area to another and defer capital gains taxes. If you're in a difficult market, need to move, and can't sell your home, you can rent it until the market improves and still build equity.
### Rules for a 1031 Exchange
Here is a quick summary of the rules for a 1031 exchange:
•The IRS requires the exchange to be like-kind, which it identifies as "real estate for real estate." You can exchange a duplex for bare land, office building, warehouse, or whatever. Just as long as it's real estate that is not used as a primary or secondary residence.
•From the date of closing on the sale of the relinquished property, you have 45 days to find the replacement property(s) and 180 days to close.
•You must insert a clause into all sale contracts that identify the transactions as a 1031 exchange. The IRS needs to see an easy-to-follow paper trail.
MORE INFORMATION ON 1031 EXCHANGES
www.irs.gov—download Publication 544, Form 8824.
www.realtor.org/libweb.nsf/pages/fg408
www.homebuying.about.com/cs/1031exchange/a/1031_exchanges.htm
www.homebuying.about.com/od/1031exchange
www.realtyexchangers.com
www.wave.net/immigration/lawyer/tax_avoid.html
John and Angie went the exchange route when they decided they no longer wanted the demands of being a landlord. They owned a duplex that had about $80,000 in equity, and they didn't want to pay out a big part of their equity in taxes. Although they didn't want to exchange for more rental property, undeveloped land appeared to be a good way to go because there is low maintenance, no rent to collect, and no late-night plumbing problems to fix.
Finding a buyer for their duplex was easy, and the sale closed with the proceeds going into escrow. Their Realtor found a ten-acre parcel for sale that appeared to be in the path of eventual development.
Since the land cost $139,000, John and Angie needed about $59,000 to make a deal. They decided to take out a ten-year, low-interest equity line of credit on their home for the funds needed to complete the deal. The second leg of the 1031 exchange closed and everyone was happy.
As a result of the exchange, a young couple starting out was able to buy a duplex they had been searching for. John and Angie won't have to collect rents or do maintenance on their days off. Everyone wins and the tax man has to wait for another day to collect his due.
In another case, an investor had several rental homes that he and a friend had owned for several years, The values had increased substantially along with the tenant headaches. Wanting to get rid of these headaches and buy some land, they found a ten-acre parcel that had promise of going up in value over the next few years.
The owners made an offer on the land subject to their rental properties selling over the next sixty days. However, the landowner had been thinking the deal over and decided he wanted to be a landlord so he offered to take one of the homes as a part of the deal. That left three homes that needed to be sold and the proceeds escrowed for a 1031 exchange.
Everything went as planned, the threes homes sold and the exchange closed. The former landlords ended up with a parcel of land where the only upkeep was cutting the weeds once a month.
Table 8-2 lists the typical steps involved in making a 1031 tax-deferred exchange.
As you can see, exchanges are a great way of turning a stale real estate investment into something more exciting for all parties with the taxes deferred.
### FEDERAL TAX CONSIDERATIONS WHEN SELLING
Currently the tax code allows an individual $250,000 and a married couple filing jointly $500,000 tax-free when selling a principal residence. A few years ago this didn't seem like a big problem in many areas of the country. However, homes prices have escalated to the point that many more homeowners will be going over the limits and end up reporting capital gains on Schedule D (Form 1040).
When you upgrade and improve your home to sell, keep track of what you do and keep the receipts. Some of your expenses in showcasing your home may increase your basis and may be a tax break.
_Table 8-2. Steps to a 1031 exchange_.
**Steps** | **What's Involved**
---|---
List property for sale and line up a property exchange intermediary. The intermediary can be a title company or an attorney who is experienced in exchanges. | Include a notice in the listing and sales documents that the property is part of a 1031 exchange. As a seller you will assign the role of grantee or transferee of the deed to the intermediary.
A buyer for property is found. | The intermediary prepares an assignment assigning the role of seller to the intermediary along with the other exchange paperwork, which goes to the closing agent
The sale is closed and equity funds are put in escrow. | Exchanger and buyer sign assignment agreement, which assigns intermediary the role of seller in the sale. The 45-day clock starts ticking on identifying an exchange property.
The hunt for a replacement property should be well on its way by this time. | 45 days to find property and identify it in writing by street address or legal description. This is faxed to the intermediary.
Exchanger makes an offer on the property. | Included in the purchase agreement is a notice that the deal is part of a 1031 exchange with the required assignments. This usually is not a problem with the sellers, they just want their money and to be on their way.
In case of multiple properties and multiple exchangers: | All the legs of the exchange are put into escrow and closed with each party getting its designated property at the about the same time. Closing and funding has to be within the 180 days.
File tax forms. | Exchangers file Form 8824 with the IRS and file any other state-required forms.
Even if you others do your tax return, look over IRS publication 523, _Selling Your Home_ , to get an idea of what tax breaks you can take.
Other IRS publications that may affect your tax situation are:
521, Moving Expenses
527, Residential Rental Property
530, Tax Information for First-Time Homeowners
544, Sales and Other Dispositions of Assets
547, Casualties, Disasters, and Thefts
551, Basis of Assets
587, Business Use of Your Home
936, Home Mortgage Interest Deduction
### THE FIVE BIGGEST TAX MISTAKES HOMEOWNERS CAN MAKE
1.Not keeping good records and receipts of upgrades and improvements that affect your home's basis. This is especially critical when your home's value is over $500,000 (for a married couple) and you have capital gains exposure.
2.Not looking at the property tax statement and checking to make sure you are credited with all the exclusions possible and that the math is correct. Many homeowners whose property taxes are paid by the lender fail to do this. It's an out-of-sight/out-of-mind situation.
3.Failing to check the property tax assessment against what similar properties are assessed. It's been estimated that nearly 75 percent of homeowners who protest get a reduction. The burden is on you, not the assessor, to make sure you're not overcharged.
4.Not following through with a protest during the window period (typically three to six weeks). Check your county website for this window period.
5.Failing to compare your home's assessment on your tax statement to the purchase price. If the assessment is higher, protest and take in your closing documents to verify the selling price.
## CHAPTER 9
## INSURANCE MATTERS FOR HOMEOWNERS
For many new homeowners, homeowner's insurance is one of those back-burner items that appeared on the closing statement as just another charge they had to pay. Actually, depending on the area you live in, there were likely several different insurance policies listed in the buyer's column, including homeowner's insurance, flood insurance, private mortgage insurance (PMI), title insurance, homeowner warranties, and sometimes even mortgage life insurance.
Tucked away in the thick folder of closing documents, these policies are largely forgotten until a problem occurs, when suddenly they become very important.
With so many unknowns that can carry a big price tag, homeowner's insurance and some of the other insurance options become an investment, not an expense. Remember that: _It's an investment_.
While insurance can be as boring as watching paint dry, ask homeowners from the Louisiana or Mississippi coasts how they feel about insurance after Katrina washed away or flooded their homes. Boring, yes—but financially critical to you as a homeowner. In this chapter you'll learn:
**All about homeowner's insurance and what it covers**
**How to shop for a policy and get the best deal**
**What records you need to keep so you can prove your claim**
**What other types of insurance you may want to have**
**All about title insurance and why you need it**
### WHAT HOMEOWNER'S INSURANCE DOES AND DOES NOT COVER
Carrie found this out what homeowner's insurance covers just one day after closing on an updated bungalow in an upscale but older area. During move in, she turned on the tub and another faucet to flush out the drains but got distracted for a few minutes and let the water run. Later when she went downstairs, she found to her horror a foot of water in the basement. Water had backed up through a floor drain in the laundry and flooded the finished basement.
A plumber checked out the line to the street and found tree roots had invaded the sewer pipe and created a partial obstruction. Normal flow of a toilet or faucet was not a problem, but the line couldn't handle the larger volume when two faucets were opened full.
Although Carrie had a professional inspection of the property before closing, the inspector didn't find any problems because the roots that caused the obstruction came from a tree that had been removed several years earlier. The tree stump had also been removed and grass planted over the area; there was no way to tell a tree once grew there.
The damage to the basement was considerable, with soaked sheetrock, carpets, and some furniture. Luckily, Carrie had a good homeowner's policy that paid for the damage, less her $500 deductible.
A good homeowner's policy protects you from the double whammy of having your home damaged or destroyed and still having a mortgage to pay off. The mortgage people know this and they want to protect their security. That's why many lenders collect one-twelfth of the yearly premium each month with your payment. Each year when the policy comes due the mortgage company pays the bill from the funds collected and held in an escrow account.
INFORMATIVE HOMEOWNER'S INSURANCE WEBSITES
www.iii.org/individuals/homei
www.pueblo.gsa.gov/cic_text/housing/12ways/12ways.htm
www.consumeraction.gov/caw_insurance_homeowner_renter.shtml
www.akc.org/insurance/homeowners_inscenter.cfm
www.insuremyhouse.com
www.iii.org/media/facts/statsbyissue/homeowners
If you have a conventional loan with a down payment of 20 percent or more, you can sometimes handle the insurance on your own. FHA- and VA-insured programs, however, still require the monthly escrow for taxes and insurance on all their programs.
If you're handling it on your own, don't forget to renew your policy. If you let it lapse, your former insurer will notify the mortgage company, and they will cover your home with a single vendor policy. These policies protect the mortgage company's interest only and are expensive, costing you about double what a homeowner's policy would. So don't let your homeowner's lapse. It can be very expensive.
### Policy Options
Homeowner's insurance evolved in the late 1950s, when the insurance industry needed a single comprehensive policy to cover not only the house, but the contents and liability. The standard policy has two parts: property insurance and personal liability.
The most common policy, HO-3, covers the house and other structures for everything except floods, earthquakes, and other policy exclusions. This is the policy that most mortgage lenders require you to carry as a loan condition.
Other options are HO-1, a bare-bones policy that is not available in most states, and HO-2, which covers only risks that are specifically insured. HO-4 is designed for renters, and HO-6 covers condominium and co-op owners.
HO-8 is designed for older homes and reimburses you for damage on an actual cash value basis. That means replacement cost minus depreciation. Depending on the area and house, full replacement cost policies may not be available for some older homes. In fact, they are getting harder to find in just about all areas. The different policy options and what each type of policy covers are summarized in Tables 9-1 and 9-2.
_Table 9-1. Different types of homeowner policies_.
**Type of Policy** | **What It Covers**
---|---
HO-1 | Bare bones police. No longer available in most states.
HO-2 (broad) | Covers most perils. Also a available for mobile homes.
HO-3 | The most popular policy. Protects from all perils except those excluded in the fine print. Typical exclusions are floods, earthquakes, war, etc.
HO-4 | Renters policy. Protects possessions only against the same perils as HO-3.
HO-6 | Homeowner's policy for condo or co-op owners. Covers what you own against the same perils as a HO-3. A master insurance policy covering the structure is included in your Homeowner Association fees.
HO-8 | A policy for older homes. Reimburses you for damage on an actual cash value, which means replacement cost minus depreciation. This policy will not usually reimburse you for the costs of bringing a home up to code.
**Policy Options to the Above**
Actual Cash Value | Will replace home and possessions minus a deduction for depreciation.
Replacement Cost | Pays the cost of rebuilding/repairing the home or replacing possessions without a deduction for depreciation.
Guaranteed or Extended Replacement Costs | Offers the highest level of protection. Pays whatever it costs to rebuild your home to what it was before the disaster. You may need an Ordinance or Law rider to pay for any costs to bring home components up to code if the home is dated. This policy may not be available for older homes. (Some companies will go only 20–25 percent over the policy limit on this type of coverage.)
### STANDARD COVERAGE
The standard policy (HO-3) typically covers damage to both structures and personal property from fire, lightning, windstorms, hurricanes, tornadoes, hail, explosions, aircraft, vehicles, smoke, theft, vandalism, falling objects, damage from ice, snow or sleet and freezing pipes. However, as many homeowners found out after Hurricane Katrina, how insurance companies define flood damage versus other hurricane-related damage can give you problems when you file a claim. Since damage from floods is not covered in homeowners policies, you need to read the fine print and make sure you understand exactly what isn't covered and add additional coverage if needed.
_Table 9-2. What your policy covers and doesn't cover_.
Note: Additional and specific perils can be added to your policy, such as coverage for earthquakes, mudslides, or extra valuables.
Personal liability is also covered if you or your property injures someone. Just about everything is covered unless specifically excluded. Typical exclusions are floods, earthquakes, neglect, intentional loss, earth movement, power failure and damage caused by war. Also, if you have a loss and the building codes have changed, increasing the repair or replacement costs, you'll pay the difference.
For example, if you have a fire and your home's electrical system is an older 60-amp-fuse system, you'll end up picking up the cost to upgrade it as well as all other upgrades needed to bring the home into code compliance.
Also, there are limits on the losses that can be claimed for items such as cash, furs, jewelry, or hobby collections. You'll need to decide if you want buy supplemental coverage to increase your protection.
One worthwhile supplemental item is coverage for living expenses if your home is destroyed or damaged and you have to move out for awhile. It covers hotel bills, restaurant meals, and other living expenses incurred while your home is being rebuilt. Coverage for additional living expenses differs from company to company. Many policies provide coverage for about 20 percent of the insurance amount on your house. You can increase this coverage, however, for an additional premium. Some companies sell a policy that provides an unlimited amount of loss-of-use coverage but for a limited time. It doesn't take much damage to a home for you to be glad you added this option to your policy.
In addition to standard coverage, there are three levels of coverage: replacement, extended, and guaranteed replacement. They differ in the amount of money a claim pays and what is covered. Of course, the better the coverage, the higher the cost of the policy. Highlights of these three coverage levels are:
### Replacement Coverage
Most insurance companies offer replacement cost coverage. For an additional 10 percent or so, insurers will pay what it costs to replace your home and belongings up to the amount of your coverage. For example, if your TV set that cost you $700 is damaged in a fire, you'll get the full cost covered. But under standard coverage you would get replacement cost _minus_ depreciation. Because the depreciation schedule for electronics is steep, you would be lucky to get 50 percent, or $350, for the TV. Or if your $50,000 tile roof, which is rated for twenty years, is destroyed in a fire when it's only ten years old, you would get only a $25,000 reimbursement under standard or actual cash-value coverage.
### Extended Coverage
The next step up in coverage is extended coverage. You insure the home for the appraised value, and the policy will pay up to 125 percent to cover unforeseen problems. Extended coverage gives is the minimum you should carry if you have a high LTV (loan-to-value) mortgage. With standard coverage, a major fire would most likely leave you with an insurance check many thousands of dollars less than your mortgage balance. This is why many mortgage lenders require you to have this higher level of protection, but it will cost you more.
### Guaranteed Replacement
But the best protection is guaranteed replacement cost which has no present limit on what it'll take to replace your home and contents. However, because of big losses in recent years, many insurance companies have dropped this policy. You may have to shop around. If you can find it, it's definitely worth considering, especially if you have recently bought your home and don't have much equity.
### YOU'LL NEED THIS RIDER IF YOU BUY AN OLDER HOME
Don't assume that if your policy reads "Guaranteed Replacement" that you're fully covered in case of a disaster. The guarantee can mean different things at different companies, but it usually applies if the cost of rebuilding a home is higher than the face value of your policy.
If you buy an older home chances are some of the wiring, plumbing, heating/cooling systems, and structure may no longer meet newer building codes. If you have a fire, the insurer is required to replace what you had, not what you're required to have in order to rebuild.
So here's the catch: When you rebuild you'll need to include the newer building codes and upgrades. That expense isn't included in your policy. And those upgrades can cost you a bundle.
You can solve this potentially expensive problem by getting something called Ordinance and Law coverage. This is a rider to your policy that applies to the costs of upgrading your home to meet existing building codes. However, keep in mind that this rider will pick up the tab only for bringing the damaged part of the house up to code. It will not pay for bringing the undamaged part of structure up to code. In other words, you could be better off if the house were totaled than only partially damaged.
Rick and Andrea had this happen when they bought a 1940s brick bungalow with the dream of restoring it. They loved the wood floors and trim, the brick construction, and the wide front porch.
The home still had the old-style wiring through a fuse box, which should have been the first upgrade project. But Rick and Andrea had decided to start on the floors, wood trim, and interior decorating. Sometime during the second week of restoration, debris got into an electrical receptacle after the faceplate was removed. The old wiring sparked, igniting the partially stripped wallpaper and engulfing the wall in flames. Fortunately, the fire department was nearby and reacted quickly. Only the front part of the house was gutted.
After getting together with the insurance adjuster and their agent, Rick and Andrea were shocked to find out it would cost them over $12,000 out of their pocket to restore the house. Their insurance policy would restore the home to "as was" condition, but not pay to bring it up to code. And they could not get a building permit unless the house incorporated building code upgrades.
So what's the bottom line? If you buy an older home, make sure the home inspector gives you a list of items that don't meet current building codes. Wiring and heating systems are often at the top of the list for potential fires.
As your budget allows, first upgrade those items on the list that don't meet code. By doing this, you'll not only make your home safer, but increase its value. If disaster strikes it will be replaced to "as was" condition and you won't be out the money you spent on improvements.
Also, be sure to keep all the contracts, receipts, and paperwork in a safe place so you can document these improvements. This is where photos or a video can be worth thousands of dollars to you.
### THREE COMPONENTS OF A HOMEOWNER'S POLICY
It's important to look at your basic policy components—structure, personal property, and liability—to see if you have enough protection. Even if you don't live on the San Andreas fault or collect antique firearms, you'll probably need to get additional coverage for peace of mind.
In Texas, policies vary somewhat from those in most other states. Check out www.tdi.state.tx.us for more information.
### Structure
Many common problems, such as earthquakes, floods, failed sump pumps, and backed-up sewers, aren't covered in basic policies. To get this additional coverage you'll need to add endorsements or riders. The higher a certain risk is for your area, the more important it is to add this coverage. Coverage for sewer clogs is usually less than $50, but it's a must, especially for older homes. A sewer backup is especially expensive because it takes special cleanup procedures due to contamination.
This is why Carrie had to call in a disaster cleanup company to handle her sewer backup mentioned earlier in the chapter, and it added a couple of thousand dollars to her clean up costs.
### Personal Property
All polices include coverage for the contents of your house, but often the amount isn't enough. Basic plans commonly pay 50 to 70 percent of the policy amount. For example, a $175,000 policy would likely give you anywhere from $87,000 to $122,000 for the contents. This may sound like a lot, but when you go through your home and total up everything you've got, it'll be a shock. The value adds up fast when you total furniture, electronics, wardrobes, power tools, stamp collections, and so on.
If you have expensive items, such as gun or art collections, antiques, jewelry, or computer equipment, you may want to consider extra coverage based on their actual value.
### Liability
Most policies have $100,000 minimum liability coverage that protects you in case someone is injured on your property. Whether that's enough depends on much you have to lose if someone sues you. You can pay for additional protection, and you can also add an umbrella policy that covers you for incidents away from home.
When you have a home business or office, and you have people coming to your home, your liability can skyrocket. Adding a $1 million umbrella policy for about $200 a year makes good sense.
If you own a condo or co-op, you'll be dealing with two policies. The homeowners association or co-op board will have a master policy that covers the common areas—roof, basement, elevator, boiler, and walkways—for both liability and physical damage. You are responsible for the other policy—the HO-6—that covers your personal possessions, structural improvements to your apartment, and additional living expenses. It also covers you for fire, theft, and the other disasters listed in your policy, as well as provides liability protection.
To adequately insure your apartment, it is important to know what structural parts of your home are covered by the condo/co-op association and what aren't. You can find out by reading your association's bylaws and/or proprietary lease. If you have questions, talk to your association or board as well as to an insurance professional.
Some associations insure the individual condo or co-op units, as they were originally built, including standard fixtures. In that case, the owner is responsible only for alterations to the original structure of the apartment, like remodeling the kitchen or bathtub. Sometimes this includes not only improvements you make, but those made by previous owners.
In other situations, the condo/co-op association is responsible only for insuring the bare walls, floor, and ceiling. The owner must insure kitchen cabinets, built-in appliances, plumbing, wiring, and bathroom fixtures.
Other coverage options you may be able to get, depending on area, association, or board are:
1. _Unit Assessment_. This reimburses you for your share of an assessment charged to all unit owners as a result of a covered loss. For instance, if there were a fire in the lobby, all the unit owners would be charged for the costs of repairing the loss.
2. _Water Back-Up_. This insures your property for damage by the backup of sewers or drains. Make sure that water back-up is included in the policy.
3. _Umbrella Liability_. This is an inexpensive way to get more liability protection and broader coverage than is included in a standard condo/co-op policy.
4. _Flood or Earthquake_. If you live in an area prone to these disasters, you will need to purchase separate flood and earthquake policies. Both flood and earthquake insurance can be purchased through your insurance agent.
5. _Floater or Endorsement_. If you own expensive jewelry, furs, or collectibles, you might consider getting additional coverage since standard policies generally have a $1,000 to $2,000 limit for theft of jewelry.
When you're buying insurance, it's important to find an agent or company that specializes in condominiums or co-ops. You can reduce your rates by raising your deductibles and by installing a smoke- and fire-alarm system that rings at an outside service. Also don't forget to check on discounts if you insure your unit with the same company that underwrites your building's insurance policy. Shop around and get at least three quotes, because costs can vary considerably.
### IMPORTANCE OF A HOME INVENTORY
In case of loss, would you be able to remember all the possessions you've accumulated over the years? Having an up-to-date home inventory will help you get your insurance claim settled faster and verify losses for your income tax return.
HOME RECORD-KEEPING WEBSITES
www.libertystreet.com/asset-home-inventory.htm
www.iii.org/media/publications/brochures/homeinventory
www.onlineorganizing.com/usefullinkssubcategory.asp?subcategory=homeinventory
www.pueblo.gsa.gov/cic_text/money/keeprecords/keeprecords.htm
www.iii.org/individuals/homei
Rocky and Lisa's home was completely destroyed a few days after Christmas. A flue in their wood-burning stove developed a crack that allowed hot gases to ignite a wood beam in the ceiling.
Luckily, they had gotten a video camera for Christmas and decided to tape their house and possessions after reading a magazine article about fire safety. The camera was one of the few items they saved, and the video was still in it along with some now precious family footage.
The videotape saved these homeowners a lot of problems and thousands of dollars. There's no way they could have remembered or proven everything they lost in the fire without it.
The lesson learned here is to keep copies of vital records, pictures, and videos at a safe and/or remote location.
### How to Inventory
Start by making a list of your possessions, describing each item and noting where you bought it and its make and model. Add any sales receipts, purchase contracts and appraisals you have. For clothing, count the items you own by category—pants, coats, shoes, for example—making notes about those that are especially valuable. For major appliance and electronic equipment, record the serial numbers usually found on the back or bottom of the item.
If you've just bought a home, work up a list as you are moving in and unpacking. Use the following guidelines:
•Valuable items like jewelry, art work, and collectibles may have increased in value since you received them. Check with your agent to make sure that you have adequate insurance for these items. They may need to be insured separately.
•Take pictures of rooms and important individual items. On the backs of the photos, note what is shown, the make, and where you bought it. Don't forget things that are in closets or drawers.
•Videotape your home. Walk through your house or apartment videotaping and describing the contents. Or do the same thing using a tape recorder.
•Use your PC to make your inventory list. Personal finance software packages often include a homeowner's room-by-room inventory program.
•Regardless of how you create the inventory (written list, CD, photos, videotape, or audio tape), keep it along with receipts in your safe deposit box or at a friend's or relative's home. That way you'll be sure to have something to give your insurance representative if your home is damaged.
### HOW TO PICK AN INSURANCE COMPANY
Many homebuyers don't take the time to shop for the best insurance deal. Their mortgage lender calls and tells them they need to have an insurer send a HO-3 binder so the deal can close. Letting the lender choose or asking friends or coworkers who they use is not the best way to find the best insurance deal.
Four things you need to look for when shopping for an insurance company:
1. _Price_. Insurance policies and prices vary greatly from one company to another, so it pays to shop around. Get at least three price quotes from companies, agents, or from the Internet. Check with your state insurance department, because they may publish a guide that shows what insurers charge in the various parts of your state.
2. _Insurer Stability_. Make sure that the company you buy from is financially stable and likely to be around to pay any claims. Few things are more devastating to your financial future than to have your home burn down and then find out that the insurance company can't or won't handle your claim.
3. _Service_. The insurance company and its representatives should answer your questions and handle your claims fairly, efficiently, and quickly. You can get a feel for this by talking to other customers who have used a particular company or agent. Also check with your state insurance department to see whether they have a complaint ratio that compares the number of valid complaints with the company's share of policies in your state.
4. _Availability_. Whether you buy from a local agent, directly from the company by phone, or from the Internet, you should be able to contact the company or agent easily. If you can't, consider getting another company. Fast and easy claim service is one of the basic things you're paying for.
WEBSITES FOR PICKING AN INSURANCE COMPANY
www.ezinearticles.com/?How-To-Pick-The-Best-Home-Insurance-Company&id=381585
www.iii.org/individuals/homei/hbs/pickco
www.money.aol.com/insurance
Some of the factors an insurance company uses to determine the price of your policy are:
•Your credit rating score.
•The square footage of the house and any additional structures.
•Building costs in your area.
•Your home's construction, materials, and features.
•Amount of crime in your neighborhood.
•The likelihood of damage from natural disasters, such as hurricanes and hail storms.
•The proximity of your home to a fire hydrant (or other source of water) and to a fire station; whether your community has a professional or volunteer fire service; and any other factors that can affect the time it would take to put out a fire.
•The condition of the plumbing, heating, and electrical system.
The price you pay for your homeowner's insurance can vary by hundreds of dollars, depending on the above data and the company. So shop around and get at least three price quotes. You can call companies directly or access information on the Internet. Your state insurance department may also provide comparisons of prices charged by major insurers.
### SHOPPING TIPS FOR GETTING THE BEST INSURANCE DEAL
• _Get quotes from different types of insurance companies_. Some sell through their own agencies with the same name as the insurance company; others sell through independent agents, who offer policies from several insurance companies; and some don't use agents. You can even find insurers who sell directly to consumers over the phone or through the Internet.
• _Consider going with a higher deductible_. The deductible is the amount of money you have to pay toward a loss before your insurance kicks in. The higher your deductible, the more money you save on your premium. A deductible of $500 or $1,000 may save you as much as 25 percent.
• _Add the necessary riders_. If you live in a disaster-prone area, don't forget to factor in the cost of additional riders you'll need to cover your specific threat. For instance, if you live near the East Coast, you may need a separate windstorm deductible. In other areas you may need a separate deductible for hail storms or tornados. Sometimes when the cost of riders is added in, a policy is not as competitive as it looks at first.
• _Buy your home and auto policies from the same insurer_. Most companies that sell homeowner's insurance also sell auto and umbrella liability insurance. (An umbrella liability policy will give you extra liability coverage.) Some insurance companies will reduce your premium by 5 to 15 percent if you buy two or more insurance policies from them. But make certain this combined price is lower than buying coverage from different companies.
• _Make your home more disaster resistant_. Find out from your insurance agent or company representative what you can do to make your home more resistant to windstorms and other natural disasters. You may be able to save on premiums by adding storm shutters and shatter-proof glass, reinforcing your roof, or buying stronger roofing materials. Older homes can be retrofitted to make them better able to withstand earthquakes. In addition, consider modernizing your heating, plumbing, and electrical systems to reduce the risk of fire and water damage.
• _Insure the house, not the land_. You'll need to subtract the value of the land when you calculate how much homeowner insurance to go with. Too many homeowners insure their home for the appraised value that includes the land as well as the improvements. Insurance agents should point this out, but many times they don't because it increases your premiums and their profits.
• _Check out discounts for home-security devices_. You can usually get discounts of at least 5 percent for smoke detectors, burglar alarms, or dead-bolt locks. Some companies may cut your premiums by as much as 15 percent or 20 percent if you install a sprinkler system and a fire or burglar alarm that rings at a monitoring station. These systems aren't cheap, and not every system qualifies for a discount. Before you buy one do the math. Find out what kind your insurer recommends, how much the device would cost, and how much you'd save on premiums.
• _Ask what other discounts are available_. Companies don't all offer the same type or amount of discounts in all states. Ask your agent or company representative about discounts available to you. For example, if you're at least 55 years old and retired, you may qualify for a discount of up to 10 percent. Or if you've completely modernized your plumbing or electrical system recently, you may get a price break.
• _Check out group coverage_. Check whether a homeowner's policy is available through your employer, and is a better deal. Also, professional, alumni, and business groups may offer insurance packages at a reduced price.
• _Ask about loyalty discounts_. If you've been insured with the same company for several years, you may receive a discount for being a long-term policyholder. Some insurers will reduce premiums by 5 percent if you stay with them for more than three years, and by 10 percent if you're a policyholder for six years or more. Still, rates and policies can and do change, so compare every couple of years to make sure you're getting the best deal possible.
• _Review policy limits and the value of your possessions annually_. You want your policy to cover any major purchases or additions to your home. But you don't want to spend money for coverage you don't need. If your five-year-old fur coat is no longer worth the $5,000 you paid for it, you'll want to reduce or cancel your floater (extra insurance for items whose full value is not covered by standard homeowner's policies).
### Earthquake Insurance
Standard homeowner's, renters, and business insurance policies do not cover damage from earthquakes. However, coverage is available with an endorsement or as a separate policy through most companies. Unlike flood insurance, earthquake coverage is available from private insurance companies—except in California, where homeowners can also get coverage from the California Earthquake Authority (CEA).
The deductible for earthquake insurance is most often 2 percent to 20 percent of the replacement value of the structure rather than a set amount. Fox example, if it takes $100,000 to rebuild a home with a 2 percent deductible, you would be responsible for the first $2,000. Insurers in states like Washington, Nevada, and Utah, with higher than average risk of earthquakes, can have minimum deductibles of around 10 percent.
EARTHQUAKE INSURANCE INFORMATION
www.insurance.wa.gov/factsheets/factsheet_detail.asp?FctShtRcdNum=20
www.answers.com/topic/earthquake-insurance
www.money.cnn.com/2006/10/16/pf/saving/toptips/index.htm?postversion=2006101615
Premiums differ widely by location, insurer, and the type of structure. Generally, older buildings cost more to insure than new ones. Wood-frame structures have lower rates than brick buildings because they tend to withstand quake stresses better. The cost of earthquake insurance is calculated on per-$1,000 basis. For instance, a frame house in the Pacific Northwest might cost between $1 to $3 per $1,000, while on the East Coast it may cost less than fifty cents per $1,000.
### Flood Insurance, Who Needs It?
Homeowner's policies don't cover flooding. You can only get flood insurance from the federal government's National Flood Insurance Program (NFIP). It boils down to this: If you don't have an NFIP policy, you don't have flood coverage.
Even though you may live outside Special Flood Hazard Area (SFHA) boundaries—also called one-in-100-years flood elevation—the low cost of a NFIP policy may still be worthwhile.
Storm drains may overflow and flood adjacent areas, canals may break, and new developments may channel water where it's never gone before: your basement.
In fact, 25 percent of the 595,000 claims the Federal Insurance Administration has paid out since 1978 have been to people outside the flood zones.
FLOOD INSURANCE WEBSITES
www.fema.gov/nfip
www.floodsmart.gov/floodsmart/pages/index.jsp
www.info.insure.com/flood-insurance/who-needs-flood-insurance.htm
www.pueblo.gsa.gov/cic_text/housing/natlflood/insurance.htm
www.ambest.com/guide/flood.html
To find out if your home is in a flood zone, contact your local building or planning department and ask to see the flood insurance rate map published by the Federal Emergency Management Association, or FEMA. If your zone designation begins with an A or V, you're in a flood plain, and to obtain an FHA, VA, or conventional financing you'll need proof of flood insurance prior to closing. Also, coverage is not limited to homeowners; tenants can purchase their own flood insurance policies covering contents.
Flood insurance can be purchased in any community that has agreed to adopt flood plain management programs. Currently, about 18,000 of the nation's 22,000 cities, towns, and counties are members.
Average premiums in high risk areas are about $300 a year. The rate goes up according to the value of the property and its location. But premiums in low and moderate risk areas are as low as $85 a year. Coverage tops out at $250,000, with an additional $100,000 for contents. The policies also covers up to $500 for removing contents to a safe location, and up to $750 for sandbagging, pumping, and other preventive costs.
Flood insurance policies are sold through local insurance agents. The company that handles your homeowner's policy can probably add this coverage for you.
### MORTGAGE AND TITLE INSURANCE
Mortgage life insurance, title insurance, and private mortgage insurance (PMI) are the three types of insurance policies you're likely to need if you bought your home with less than 20 percent down.
### Mortgage Life Insurance
Many people confuse mortgage life insurance with private mortgage insurance (PMI). Private mortgage insurance, discussed below, pays the mortgage lender in case of default. If you take out a mortgage with less than 20 percent down payment, the lender will require PMI to protect them from default. Although it sounds similar, mortgage insurance that insures you in case of disability or death is optional.
Within a few days of closing, you'll probably get offers and brochures from your mortgage company or an affiliate offering different types of mortgage life insurance. Some policies will make the payments if the borrower becomes disabled, others will pay off the mortgage upon death. The question is, are these policies a good way to go.
MORE INFORMATION ON MORTGAGE LIFE INSURANCE
www.money.cnn.com/2003/12/19/pf/expert/ask_expert/index.htm?postversion=2003121909
www.lifeinsure.com/lifeinsurance/mortgage.asp
www.insweb.com/learningcenter/articles/life-mortgage.htm
www.wtnh.com/Global/story.asp?S=3599578
If you would sleep better at night knowing that if you became sick or disabled your mortgage payment would be paid, then this type of policy may be worth pursuing. Compare quotes from several insurers including the company that has your homeowner's policy. Rates will vary widely depending on area, age, and amount.
Another option available is the policy that pays off the mortgage if you die. Basically this is called decreasing term insurance. In theory, the rates should go down as you pay off the mortgage.
Many financial experts say a standard term policy is not only cheaper but has more flexibility. For example, mortgage insurance would pay off the mortgage balance automatically if you die. A regular term policy would pay the survivors, but if they don't want to pay off the mortgage they don't have to. Here again, the best way is to check with several companies and compare rates. The company that has your homeowner's policy may have this program and give you a cost break.
### Title Insurance 101
In many states the purpose of the title company is threefold: 1) To sell title insurance, 2) to handle escrow funds, and 3) to handle the actual closing, where you go in and sign the paperwork. Other states have attorneys or escrow companies that do the actual closing. But regardless of who does the closing, title companies still provide the title insurance coverage on your property.
MORE INFORMATION ON TITLE INSURANCE
www.firstam.com
www.mtgprofessor.com/title_insurance.htm
www.homebuying.about.com/od/homeshopping/qt/TitleInsurance.htm
www.money.cnn.com/2006/01/11/real_estate/title_insurance_exposed/index.htm?cnn=yes
Title insurance is one of the biggest yet least understood costs in buying a home. It's one of those fees that you don't get involved in directly. And, like homeowner's insurance, it can have a big impact when the need suddenly complicates your life.
Charles and Kristen found this out when they thought they had bought a home on a half acre along a river. What the seller didn't tell them was that a few months earlier the county had bought an easement along the river for a parkway. This reduced the back property line by 40 feet and somehow the title search had missed the easement.
When Charles and Kristen were landscaping their backyard, a county parks employee came by to stake out the jogging trail. He told them they were encroaching on the future parkway. Understandably, the new homeowners were upset and called the title company. Luckily, their insurance was with a good company and they ended up with a fair settlement.
Typically, you (or the seller) pay a one-time fee at closing to insure the property against the following problems:
•Forgery and impersonation
•Lack of competency, capacity, or legal authority of a party
•Deed not joined in by a necessary party (co-owner, heir, spouse, corporate officer, or business partner)
•Undisclosed (but recorded) prior mortgage or lien
•Undisclosed (but recorded) easement or use restriction
•Erroneous or inadequate legal descriptions
•Lack of a right of access
•Deed not properly recorded
•Claims for adverse possession or prescriptive easement
•Deed to land with buildings encroaching on land of another
•Incorrect survey
•Silent (off-record) liens (such as mechanics' or estate tax liens)
•Pre-existing violations of subdivision laws, zoning ordinances, or CC&Rs (covenants, conditions, & restrictions)
•Post-policy forgery
•Forced removal of improvements due to lack of building permit
•Post-policy construction of improvements by a neighbor onto insured land
•Location and dimensions of insured land
As you can see, these are some heavy-duty problems that can ruin your day if they pop up unexpectedly.
In another sticky situation, Jack and Carolyn bought an older home in a small town. The tax notice described the property dimensions as 110 feet by 140 feet, which had been the accepted dimensions for seventy-five years and six buyers.
The new buyers decided to put up a fence, but when they measured 110 feet from the east corner they found their lot went about 7 feet into their neighbor's living room. This was a potentially messy situation.
Jack and Carolyn contacted the title company that insured their sale and they had a surveyor check out the property corners. As it turned out, the old town survey had a few problems. All the lot lines along the street had to be readjusted on the plat, and new property descriptions worked up and recorded in the county recorders office. It turned into a several-thousand-dollars project, which was covered by the title insurance.
Title insurance costs vary from state to state. Some states set the rates, others require that insurers file their rates with their state department of insurance, and some don't regulate the fees.
In Iowa, home buyers typically purchase a title-warranty certificate from the Iowa Finance Authority. You get the same coverage as title insurance at a fraction of the cost. Ask your Realtor or mortgage lender about the norm in your state.
### Private Mortgage Insurance
Private mortgage insurance (PMI) is the third type of insurance policy you're likely to have if you bought your home with less than 20 percent down. Private mortgage insurers are usually separate companies that specialize in insuring mortgages. They're not connected to a government agency, mortgage company, or investors.
PMI insures the lender against you failing to make your payments and they having to foreclose. The monthly premium is calculated on a sliding scale. With a 5-percent down payment the premium will be higher than with a 10-percent down payment, and it is calculated as a percentage of the loan amount.
For instance, a $250,000 home with a 5-percent down payment might have a .70 percent premium. or $138.54 per month. A 10-percent down payment could have a .50 percent premium, or $93.75 per month. Your monthly rate can also depend on your credit rating as well as the loan amount.
Normally PMI is supposed to drop off when the loan is paid down 20 percent, but until recently many homeowners had been taking advantage of lower interest rates and rising house values to refinance and get rid of PMI. On FHA-insured loans, however, the PMI doesn't drop off. So a good strategy is get into the house with an FHA loan and then refinance to a conventional 80-20 loan as soon as possible.
### THE SEVEN BIGGEST INSURANCE MISTAKES MANY HOMEOWNERS MAKE
1. Not keeping careful and complete records of improvements, possessions, and other items in case your home is destroyed or damaged.
2.Failing to review and upgrade your homeowner's policy yearly. Home values are going up so fast in some areas that if you had a disaster would your policy cover it?
3.Not keeping a duplicate set of records and photos offsite so your data won't be destroyed if there's a problem.
4.If you have an older home, not having a plan to upgrade those items that are not up to code. In case of a fire, for example, a homeowner's policy typically pays to rebuild to "as was" condition. Before you can get a certificate of occupancy, all systems will have to be up to current codes. It can be a case of "pay me now or pay me later."
5.Failing to consider other threats in your area and the need for additional policies. If there's a problem, will you be stuck with a big mortgage and a damaged home without coverage?
6.Not shopping around for the best rates or taking advantage of discounts offered by some insurers when you have your homeowner's and other policies with the same company.
7.Failing to do a risk audit of your home and yard. This is where you look at your property and make a list of possible problems you need to correct. For example, let's say an old tree in the front yard is diseased and weakened. If an ice storm hits, will it end up crashing through your roof?
Some things you would look for when doing a risk audit are:
•Look at the trees around your home can that could fall and damage your home. You might want to hire an arborist to check them out if they're old or appear to have problems.
•Prune trees and branches that are too close to power lines or your roof.
•Are there areas around your home that could funnel water into your yard in case of a 500-year storm, a breached canal, or an overflowing creek or river?
•Are there a lot trees and bushes close to your home that pose a fire hazard? Would it be better to clear-cut everything away from the home to create a fire zone?
•Look at your home and yard with a "what if" mindset and note how you can make your home safer should the worst happen.
## CHAPTER 10
## AVOIDING THE TEN MOST COSTLY HOUSE PROBLEMS
In this chapter you'll learn:
**The ten most common problems that homeowners encounter while buying or living in a home**
**Suggestions on how to correct these problems**
**Tips on preparing your home for sale**
If you were to poll home inspectors, appraisers, Realtors, and remodelers on what problems they encounter most frequently, chances are their lists would be similar. The same problems appear so consistently that it's not too difficult to pick out the top ten.
Some of these problems are so common that many homeowners have learned to live with them; they no longer consciously register, not even the dangerous ones that can kill you.
You may want to check your home against the ten problems outlined in this chapter and see if you need to do some preventive maintenance. It's much cheaper to fix problems before they're flagged by an inspector or appraiser. Also, getting into a time crunch to get things fixed before closing usually costs you a lot more bucks.
For example, one home seller, who was a little on the lazy side, let maintenance slide for a few years. When he decided to put the home up for sale, it languished on the market for weeks until a handy couple made a low offer.
Unhappy with the offer but running out of time, the seller accepted it. However, when the home inspection report came back, it listed some wiring and roof problems the buyers insisted be fixed before closing. The seller had no choice but do the repairs if he wanted to keep the sale from falling apart. In the end, the seller sold the home for $12,000 under market and had to spend $3,700 in electrical and roof repairs. He walked away from closing with $15,700 less than he should have had.
Fortunately, these ugly equity-reducing situations don't have to happen, and you can avoid them by checking out the following ten most common house problems and their solutions.
### PROBLEM ONE: FAULTY WIRING
Many electrical problems are caused by homeowners who try to cut corners by not using a licensed electrician or by trying to do the work themselves. This is one area where mistakes can have consequences you don't want.
One homeowner, for example, got a nasty jolt when she walked across a wet, freshly mopped floor and touched an ungrounded electric stove. The metal clamp that attached the grounding wire to a metal pipe had rusted through, creating an ungrounded system.
Luckily she wasn't badly injured, but she was one unhappy homeowner when she called an electrician soon after.
Common electrical mistakes to look for are:
•Open junction, switch, or receptacle boxes. Without faceplates, wires are exposed to kid's exploring fingers. Accumulations of dirt and lint can also cause a short. Face plates that were removed when a room is painted or remodeled need to be replaced promptly. Also, look in basements, attics, garages, or out-of-the-way places for electrical boxes that don't have faceplates.
•Dangling hot wires are especially dangerous. Homeowners sometimes remove light fixtures or reroute wires and forget to wire-nut the hot wires. Always test any dangling wires to make sure they're not hot. Remove unused dead wires and label hot wires (safely wire-nutted in a box) so you know what breaker they're attached to.
You can download an excellent electrical safety manual free in PDF format from: www.cdc.gov/niosh/pdfs/02-123.pdf.
•Especially if you have an older home, check and make sure the service cable, breaker panel, and main disconnect have the same amp rating. Owners sometimes try to upgrade wiring themselves and fail to match up all the system components. If you have an older system that's rated for less than 100 amps, you may want to consider upgrading to 150 or 200 amps.
•Make sure the grounding wire/clamp is attached to a metal pipe or stake. This is usually located next to the electrical service entrance. In older homes these clamps sometimes rust off, leaving the home ungrounded. That can shock you if you touch a metal pipe or appliance. Replace the clamp if it's rusty or not making good contact.
•If your home was built between 1964 and 1974 and you haven't upgraded your electrical, you should check for aluminum wiring. You can tell aluminum wiring by its dull gray color where it connects to the neutral bus bar in the breaker box. Unlike copper wiring, aluminum shrinks and swells, loosening splices and connections. A fire hazard can result if a wire becomes loose and shorts. As a result, aluminum wiring requires a special crimp connector (called a COPALUM connector) to connect to switches and outlets. Have a professional electrician inspect your wiring to make sure you have these types of connectors. (Oh, and keep the paperwork to show buyers the wiring is sound when you sell.)
For more information on aluminum wiring go to: www.inspect-ny.com/aluminum.htm.
### PROBLEM TWO: GRADING AND DRAINAGE PROBLEMS
Appraisers and inspectors often zero in on landscaping conditions that can allow water to drain toward the house or into a basement or crawl space. Damage caused by this condition can not only lessen the value of the home but be very costly, since the restoration costs won't be covered by homeowner's insurance. You have to have government flood insurance to cover damage caused by an exterior water source.
Water damage from these conditions is often caused by:
•A home built on a hillside with poor grading funneling water to the foundation or low spots near the home.
•Poor drainage allowing landscaping bark or loose spongy soil to retain water, which will leak into the basement or crawl space.
•A valley location allowing water to concentrate near the foundation and leak into the basement.
•Careless backfill when the home is built, allowing water to easily penetrate the soil around the foundation.
•Soil settling or eroding around the foundation and allowing water to penetrate rather than run off.
•Downspouts and gutters not properly diverting water away from the foundation.
•Window wells not rising above grade.
Cures are to regrade the ground around the foundation for about 10 feet and slope it at least 4 inches. If the surrounding terrain is higher than the home, you can divert water by excavating swales and incorporating them into your landscaping.
INTERESTING WEBSITES ON GRADING
www.landscaping.about.com/od/sitegradingdrainage
www.taunton.com/finehomebuilding/pages/bh0028.asp
Also make sure gutters are functioning and cleaned out. You may need to add extensions to your downspouts to divert water away from the foundation.
Basement windows should be protected by wells that rise at least four inches above grade. On older homes, galvanized metal wells often rust through and need to be dug out and replaced. Water can also leak through basement windows when downspouts are carelessly placed too near a window well, or when sprinkler heads are not adjusted.
During the first good rainstorm after you move in (or before, if possible), grab an umbrella walk and around the house, looking for possible drainage problems. It's important to be proactive in this area, because water leaking into the house from an outside source is not covered by your homeowner's insurance policy.
### PROBLEM THREE: FAULTY GUTTERS AND DOWNSPOUTS
When rain gutters, downspouts, and extensions work as designed, they do a great job keeping water away from the foundation and out of basements and crawl spaces. But if problems arise because components are missing or aren't installed correctly, it can be expensive to repair the water damage.
One homeowner failed to notice that a downspout strap had come loose, which allowed the leader to channel water into a basement window well. A heavy rainstorm put over a foot of water in the homeowner's finished basement.
Incidentally, because the water came from outside the home, the insurance company refused to pay for the damage. The homeowner ended up paying nearly $10,000 to replace carpets, sheetrock, and furniture, and to cover redecorating costs.
Clearly, constant vigilance is needed to make sure your gutters are and stay in good condition.
Some common gutter and downspout problems to check for are:
•Damaged sections that need repair or replacing
•Leaves and debris blocking water from flowing to the downspout
•Gutters too small to handle the runoff volume given the size of your roof
•No extensions on the downspouts to carry water away from the foundation
•Improperly installed gutters that fail to drain
RAIN-GUTTER WEBSITES
www.hometips.com/csprotected/guides/gutters.html
www.rainhandler.com/checkout.htm
www.home-ideas.org/Rain-Gutter-Installation.htm
To keep your gutter system working properly, clean out them out in the fall, and make sure water flows toward the downspouts. Seal small holes with silicone and larger ones with flashing and roofing cement.
If your home needs new or replacement gutters, check out contractors in the yellow pages. As with any other improvement project, get three bids before you commit.
### PROBLEM FOUR: WATER IN THE CELLAR
A high percentage of homes either currently have or will have basement or crawl space water problems. Some have problems consistently, others only at certain times of the year because of seasonal conditions.
Whatever the source, you'll want to solve water problems as soon as possible to minimize damage and eliminate a potential deal killer when you sell your home.
Common water infiltration clues to look for are:
•Dark stains on floor or walls
•Water marks, or a pale powdery residue called _efflorescence_ , on the walls
•Musty or unpleasant odors
•Mildew
•Recurring wet spots
Cures can be simple, or complex and expensive, depending on the problem. But before you panic and call a contractor, check out these possible solutions:
•Make sure the gutters, downspouts, and extensions are routing water away from the foundation.
•If the crawl space has a dirt floor, you may need to cover it with a plastic vapor barrier to keep water from migrating into the house.
•For small amounts of moisture migrating through walls, get waterproofing coatings that can be applied on the inside with a brush or roller. These are available at home centers and mail order.
•For bigger amounts of infiltrating water, you may need to install a _sump pump_. These are small pumps mounted at the bottom of a sump or shallow well. The pump has a float switch that kicks on the motor when water accumulates in the well. Since heavy storms are more likely to cause a power failure—which is when you need the pump working—you may want to consider adding a battery backup.
•Digging a trench around the foundation 18 to 24 inches deep, lining the bottom with crushed stone and burying a perforated pipe (curtain drain) can be effective.
•Cold water pipes, ducts, and water tanks can sweat a considerable amount of water. Insulating these moisture sources can reduce the problem considerably.
•Check for mold and keep it under control with bleach or other fungicide.
•If you have a floor drain, shine a light down it and make sure you can see water. If not, add water until you can. This is a vapor lock that keeps sewer gases out of your basement. Over time, the water can evaporate and allow unpleasant odors to escape.
•Never vent dryers into a basement or install a shower unless it has a fan that vents to the outside.
MORE DATA ON MOLD
www.traskresearch.com/mold.htm
www.epa.gov/mold/moldguide.html
www.cdc.gov/mold
### PROBLEM FIVE: PROBLEM ROOFS
Few homeowners give their roof much attention until they discover a leak. Unfortunately, by then the damage is often extensive and expensive, as one owner found out when the kitchen ceiling crashed down onto the table. They had ignored a spreading stain in the ceiling, and it didn't take long until the wet sheetrock gave way.
Asphalt shingles, the dominant roofing material, have an average life span of 15 to 30 years depending on climate, exposure, and shingle quality. Because roof longevity can vary significantly even between homes that were roofed at the same time, it's critical to inspect your roof every few months.
Binoculars are good way to check the roof; there's no need to climb up and risk damaging shingles or falling off.
If you're thinking of selling your house and the roof is more than five years old, it's a good idea to get a professional roof inspection before putting it on the market. You don't want an inspector or appraiser finding problems after you've found a buyer. That can kill a sale or put you in a defensive position when a buyer threatens to walk unless you reroof.
Some common warning signs that your roof needs attention are:
•Asphalt shingles are curling and/or brittle and the valleys look worn. This means the roof is nearing replacement time.
•Flashings around vents, skylights, chimneys, and pipes are cracked and worn.
•Stepped flashing (flashing that is interwoven with the shingles) around the chimneys is curling and worn.
•Rubber collars that seal pipes at the roof line are cracked or worn. If roofing cement has been used to seal these joints, check for cracks or peeling.
MORE INFORMATION ON ROOFING
www.roofhelp.com/roof_repairs.htm
www.hometips.com/content/builtup_tp.html
www.hgtv.com/hgtv/rm
If you find problems in your roof inspection, some solutions are:
•New shingles can be applied over existing ones to cut costs. However, you should only do this once. More than two layers create excessive weight, especially in snow areas.
•A roof is only as good as its flashings. Make sure seals are replaced or recoated as soon they start to show problems. You may have to do this every year or two, depending on climate. Spending $50 or less to recoat is much cheaper than spending big bucks fixing water damage.
•If the roof looks like it may need replacing, get three bids from reputable contractors and compare.
•Don't go with 30-year shingles when you plan on living in the home only a few years. A buyer is unlikely to pay more for a 30-year roof than for a 20-year one.
•If you want to change roofing types—such as changing from asphalt shingle to metal, tile, or other type—check other homes around the neighborhood first. Spending thousands of dollars upgrading to tile when no one else has means you're unlikely to get your investment back when you sell.
### PROBLEM SIX: FOUNDATION FLAWS
Older homes and even some newer ones can settle unevenly. Clues to uneven settling are:
•Obvious cracks in the foundation walls
•Cracks above door jambs and windows
•Sloping floors and ceilings
•Sticking or hard-to-open doors and windows
If a house has stabilized (you'll want a structural engineer to verify this), the cracks can be filled with hydraulic cement or similar caulking compounds that expand and seal. But if the problem is more serious, it could entail:
•Excavating the foundation and sealing cracks from the exterior side.
•If cracks go all the way through the foundation, sealing the inside as well.
•Replacing part or all of an unsound foundation. This can get expensive because the house would need be jacked up, the new foundation poured, and the home re-leveled. Obviously, this is going to be expensive, but the problem will have to be corrected before you can live in the home safely or sell it.
FOUNDATION INFORMATION WEBSITES
www.askthebuilder.com
www.inspect-ny.com/structure/foundation.htm
www.concretenetwork.com
One homebuyer did not hire a professional inspector because he worked in the building trades and felt he knew as much as any inspector. After the closing he was moving stuff into the basement, and when he moved some cardboard that was leaning against the wall he discovered a large crack. Water had started to leak through and discolored the concrete around the crack and down the wall. An inspector would certainly have caught this problem because this what they're trained to do.
Lesson learned: Even highly competent contractors can benefit from a professional inspection when they buy a home.
### PROBLEM SEVEN: DEFERRED MAINTENANCE
If the home or yard isn't in good condition, it's not only a safety issue but a selling problem as well. Poorly maintained homes attract bargain hunters or investors who make steeply discounted offers. Homes that are in good condition attract top dollar and full price offers.
Keeping your home in top condition is not a cost, but an investment that will return more than you spend. To spread the costs out, some homeowners find fixing up and updating as they go along easier than having to max out a credit card when it's time to sell.
It's also much cheaper to catch problems up front before they get away from you and demand expensive solutions when you least expect it. This happened to one homeowner who had let maintenance items slide for several years. When health problems required him to sell the house, he couldn't do the necessary fix-up himself, and he didn't have the money to hire a professional to do it. As a result, the home sold for many thousands of dollars less than it should have. It was a good deal for an investor, but a bad deal for the homeowner who sorely needed all the money he could get from the sale.
Important maintenance items you'll want stay on top of are:
•Paint rooms, ceilings, and trim as needed to keep your home looking good and send a message that you're serious about maintaining its quality.
•Replace carpet or vinyl floors coverings when worn or out-of-date, This ranks second in maintaining a home's value. But if wood or tile floors are popular in your area, consider going that route for an up-to-date look.
•Keep the caulking in showers and around tubs and fixtures watertight and in good condition. This will save you money by not having to replace a shower or tub enclosure because of water damage.
•If a toilet starts leaking around the bottom, fix immediately. Damage from water seeping under the floor is a costly repair job.
•Replace or fix leaking faucets. They not only waste water, but can stain sinks as well.
•Keep the home's exterior in good condition. Caulk, paint, or fix sheathing problems immediately so insects and water don't invade and cause damage.
•Keep hedges, bushes, and trees trimmed. One homeowner let his trees and bushes grow out of control for a few years. When he decided to sell he had to pay an arborist $1,700 to cut and prune the overgrown tangle.
•Seal and maintain driveways so water doesn't penetrate. When it freezes, it will destroy the surface and/or widen the cracks.
### PROBLEM EIGHT: FAULTY PLUMBING
Water damage tops the list of homeowner frustrations and cost. But it doesn't have to be that way. It's important to remember that the water supply system is under pressure, and that water under pressure does interesting things, such as leave deposits in pipes and fixtures, corrode fixtures, and leak wherever there's a weak spot in the system.
Vigilance is the key to taming the water supply system. Keep a regular eye on the following:
•Water lines and fittings. Check the system every few months for leaks and hard water buildup on faucets and toilet components.
•If you have an older home and the pressure diminishes markedly as a toilet fills, the fixture or bathroom may be tied into a too small supply line. You may have to reroute the line to a better source. Old galvanized pipes could also have mineral buildup that constricts the water flow.
•If your water comes from a private well, you may need to check the pressure tank and pump. Sometimes a bigger tank can compensate for a low flow well.
•Replacing corroding pipes with newer lines is a good investment. New flex tubing makes replacing old water lines easier and cheaper than upgrading with copper.
•Trace any leak to its source immediately. Ongoing leaks ruin floors, wall studs, sheetrock, and other components that gobble up cash flow fast.
•If you're in a cold winter state, make sure pipes exposed to cold air are wrapped in insulation and/or heat tape. Check the heat tape connection daily. One duplex owner who shall remain nameless failed to do this and a pipe froze. By mid-afternoon when the pipe thawed, it split and water flooded the interior. The result was a costly restoration and unhappy tenants.
HELPFUL HOME-PLUMBING WEBSITES
www.plumbing-today.com
www.homeandfamilynetwork.com
www.hometips.com
The other half of a home's water system consists of the drain-waste-vent (DWV) pipes. Problems with this system show up when water drains slowly from a sink, tub, or shower. Causes can be:
1.Hair, food, or other material blocking the drain; it's an easy fix.
2.Mineral build-up in the drain pipes is a little more serious. Replacing the pipes with ABS pipes is the best long-term solution.
3.A vent pipe on the roof may be blocked, creating a vacuum that slows draining water. You'll need to get on the roof and make sure there's no obstructions in the pipe.
One serious indicator of a leak is water stains on a ceiling. A leaky shower pan, toilet seal, pipe or fixture has caused or is causing problems. For homebuyers this should be a red flag that signals potential water problems: Proceed with caution.
On the flip side, if you're a homeseller, and the water damage is evident, do what's necessary to fix the leak and restore floor, ceiling, or sheetrock to prior condition, pronto.
Once buyers see water damage—even though it's been fixed—they become skeptical, and your chances of a good offer go down the drain.
### PROBLEM NINE: POOR VENTILATION
Although ventilation is one of those out-of-sight/out-of-mind home components, its effects on the house and your health are huge. For example, a poorly vented attic can heat up to 150 degrees in the summer, making it more expensive to cool. In the winter, water vapor condenses on insulation, rafters, and roof sheathing, causing rot and mold.
MORE VENTILATION INFORMATION
www.oikos.com/esb/39/VentOpt.html
www.askthebuilder.com/Ventilation.shtml
www.epa.gov/iaq/homes/hip-ventilation.html
Ventilation is an area that inspectors and appraisers will look at closely, and you don't want to have to deal with costly mold, roof rot, or adding vents a week before closing.
To prevent this:
•Keep gable vents open and free of nests, leaves, and debris.
•Inspect your attic every fall and spring for potential problems.
•If you finish your attic or add insulation, make sure you don't block soffit vents.
•Many older homes don't have enough attic ventilation, so you may need to install roof vents, wind turbines, or powered vents to keep outside air circulating. It's important to keep heat build-up in the attic to less than 15 percent of the outside temperature.
•A big no-no is venting bath, kitchen, dryer, or laundry fans into the attic (or basement). Fans should vent to the outside to prevent damaging water build-up.
### PROBLEM TEN: DEFECTIVE HEATING SYSTEMS
As the Realtor walked through the home on a pre-listing inspection, he noticed the furnace was about twenty years old, and installed when the home was built. Knowing this was going to be an issue with buyers and inspectors, the agent suggested that the sellers get a furnace inspection ASAP.
A couple of days later the sellers called their agent all upset. The furnace contractor found the combustion chamber had cracked and that deadly carbon monoxide could escape into the home. This odorless and lethal gas could have easily caused serious problems for the homeowners. In fact, hundreds of people die from carbon monoxide poisoning every year in the United States.
In this case, the sellers replaced the furnace, the home sold a week later and the inspector found only a few minor problems. Lesson learned here is to have your furnace inspected yearly when it's more than five years old. Also important is when the sellers replaced their furnace, the new model was smaller, quieter, more efficient, and used less natural gas. Had they not sold their home, their fuel savings would have paid for the new furnace in less than seven years.
MORE INFORMATION ON FURNACES
www.aceee.org/consumerguide/topfurn.htm
www.energystar.gov/index.cfm?c=furnaces.pr_furnaces
www.furnacecompare.com/buying_a_furnace.html
Typically, inspectors and appraisers check gas or oil forced-air furnaces for cracks in the heat exchanger or combustion chamber. In a steam system, they look for cracks in the water jacket. If they find any, it usually means the furnace or boiler needs replacing.
Where the flue from gas water heaters and furnaces connect to the chimney is also a critical area, which should be checked every few months. A failure in the seal can leak toxic gas into the home.
Inspectors also check water heaters for bottom leaks—an indication of a rusted tank—and whether the pressure release valve is in good shape.
What are some things you can do to prolong your heating system?
•Consider an annual service contract that ensures your furnace or boiler is properly cleaned and adjusted for the winter season.
•Check chimney and flue junctions and make sure they are sealed and leak-free.
•Replace the anode in the water heater. This is a device that attracts corrosion and prolongs the water heater tank life.
If you're thinking of selling your home, or if you have just bought a home and aren't sure about the condition of the furnace or water heater, contact an HVAC (heating, ventilation, air-conditioning) contractor for an inspection. Make sure they leave a sticker or tag on the furnace showing the date it was inspected.
Finally, if you have a forced-air furnace, it's a good idea to have the ducts professionally cleaned. It's amazing what collects around the vents and in the ductwork that you probably don't want to know about.
All ten problems discussed in this chapter are summarized in Table 10-1.
_Table 10-1. Quick checklist for avoiding the ten most costly house problems_.
**Problem** | **Suggested Action**
---|---
Faulty Wiring | Use a licensed electrician for upgrades and additions. Upgrade an older fuse box or 60-amp system ASAP. Look for uncovered boxes, dangling hot wires, and a poorly grounded system.
Grading and Drainage Problems | Make sure water drains away from the house, not towards it. Add extensions to downspouts to route water away from foundation and window wells.
If needed, add clay or dense soil to create a slope away from the foundation. Perforated pipe buried in a ditch filled with crushed stone a few feet out from the foundation can help drain water.
Gutters and Downspouts | Clean and repair damaged gutters and downspouts. Seal holes. Make sure water isn't pooling around the foundation.
Cellar Moisture | Check drains for water in trap to prevent bad odors.
Condensation problems from cold air may require heating the basement.
Basement showers and dryers should be vented to the outside. Consider installing a sump pump for bad water leaks.
Leaking roofs | Reshingle, but not more than two layers.
Check flashing around chimneys and valleys. Replace worn rubber collars that seal pipes and vents where they come through the roof.
Foundation Problems | Look for cracks in foundation walls, above doors and windows. Also look for sloping floors and sticking doors. Cracks and settling should be fixed by professionals ASAP before further damage occurs.
Poor Maintenance | Not keeping a home in good condition can lower its value thousands of dollars more than the cost of upkeep. To maintain its value, keep it painted, caulked, and keep the yard in good condition. It's also cheaper to correct problems as they occur rather than maxing out a credit card to do it all at once.
Faulty Plumbing | Check for low water pressure. Turn on a faucet and flush a toilet; if pressure falls significantly it may mean the supply line is inadequate. Look for leaks, especially around toilet and under sinks. Repair promptly or costly damage will follow. Keep drains clean and free of blockages. Check shower pan and tub caulking for leaks.
Inadequate Ventilation | Make sure attic has good ventilation. A rule-of-thumb is 1 square foot of vents for every 150 square foot of attic floor. If needed, install power vents or wind turbines. Vent bath, kitchen, and laundry fans outside and not into the attic or basement.
Furnace Problems | Shop around for an annual service contract, especially on furnaces older than five years. Check the water heater for corroded anode and replace if needed. Make sure chimney-flue connections are sealed and don't leak.
## APPENDIX A
## DEALING WITH DISASTER
Why do bad things happen to good homeowners? We don't know why, but when they do happen we can't call the super or landlord to bail us out; we must handle it on our own.
Although there are thousands of things that can go wrong, the following five are common problems millions of homeowners encounter every year. Also included are some suggestions to help you get started thinking about how to keep bad things from happening.
### BAD THING #1: DEALING WITH A FIRE
You're on your way home from a late afternoon shopping trip when you round the corner of your street and see the flashing red lights of emergency vehicles in front of your home. Fire hoses snake from a hydrant down the street and two heavy-jacketed and helmeted figures are spraying water on a back corner of your house.
After fighting down a wave of panic, you run toward the house, trying to control the fear that everything inside is destroyed. Fortunately, you were lucky this time. One of the helmeted figures sees you standing there and walks over as he unbuckles his chin strap. He tells you the fire started in the laundry and only destroyed that room, the adjoining bath and some of the family room.
Once the firefighters have left, you'll have to decide quickly how much of the cleanup you're willing to do yourself. If the damage is extensive, leave everything as is and call your insurance agent immediately so the damage can be documented.
However, if the damage is light and limited to one room, you likely can start cleaning up on your own after calling your insurance agent. Some ways to clean up smoke damage are:
•Start out with a wet/dry shop vacuum to clean up soot and chemical foam from extinguishers. Open windows and door to circulate air through the home. Carpets, clothing, and furniture will require professional cleaning to remove the smoke damage.
•Although the insurance company will refer you to companies that specialize in these kinds of cleanup, it's likely they won't be able to get started for a few days. If you can get a head start and make the home livable, you'll minimize the disruption to your life.
•Dry cleaning sponges work well for removing soot particles from latex paint surfaces. Start on the ceilings and work down. When the sponge surface becomes saturated, simply scrape it off. For oil-based paints, vinyl wallpaper, and greasy kitchen fires, use a sponge and a grease-dissolving cleaner or detergent.
•If the home is unlivable, you'll need to board up broken windows and other holes. Remove as many valuables as you can and consider hiring a security service to keep an eye on the home.
Check out The Institute of Inspection, Cleaning, and Restoration Certification (IICRC) at www.certifiedcleaners.org, or 800-835-4624, for a list of professionals who handle interior damage in your area. For info on how to file an insurance claim go to: www.iii.org/individuals/homei/help/howclaim.
### BAD THING #2: HANDLING A WATER LEAK
Have you or someone you know ever come home from a vacation, or even after being away just a few hours, to find water running out the front door or a basement rapidly filling up?
Water problems can come from failed washer supply hoses, pipe joints leaking, splitting water lines, or even from someone forgetting to shut off a faucet.
Obviously, the first step is to try and turn off the water to the leaking line or fixture. If that proves difficult, you'll need to shut off the main water valve. Some suggestions are:
•In cold areas, you'll likely find the house shutoff valve in the basement, on the wall facing the street or where the supply pipe enters. If you have a private well, either shut off the valve on the supply line coming out of the pressure tank, or shut off the power to the pump.
•In warmer areas, the valve could be outside. If you can't close or find the valve, call the city water department immediately.
•Make sure the water hasn't come in contact with any outlets or cords before you go wading in, it could be carrying an electrical current. Shut off the main breaker, but if you can't get to it, call the power company and wait for them to cut the power.
•If the water damage is extensive, call your insurance agent. You may also need to call a cleanup company that handles water damage.
•If you have access to power, use a wet/dry vacuum, but make sure the circuit is connected to one with a ground-fault interrupter.
•For flooded basements, you'll need to rent a sump pump or gas powered unit that mounts outside and uses intake and discharge hoses.
•If the water is sewage contaminated, it's best to have a professional company disinfect the flooded area. Sheetrock and flooring may have to be replaced. Everything that got wet has to be thoroughly dried out to prevent mold.
•Don't throw away anything (soaked carpet, damaged furniture, etc.) until the insurance adjuster has had a chance to document your loss.
### BAD THING #3: HANDLING A ROBBERY
The panic of finding your door ajar when you return home can blind you to the reality that you could be in danger. The thief/thieves could still be in the home. Before you enter, call 911 on your cell phone or go to a neighbor's home and wait for the police. Some tips to handle a robbery are:
•Don't go in or touch anything until the police arrive and process the home for fingerprints or other evidence.
•Be prepared to give the police a list of items missing and any receipts you may have. You'll also need this information for the insurance claim.
•Once the police find the point of entry, you need to repair any damage and beef up security where the burglars entered.
•Go through your home and do a security check; making sure you have deadbolts that meet ANSI Grade 2 standards on all entry doors. Also, check sash locks on windows to make sure they're secure. Clear out trees branches and bushes that can give thieves cover.
•If you haven't already done so, make up a list of valuables along with photos or video clips.
For some downloads on home security, check out this website: www.ncpc.org/publications/brochures/protecting.php.
### BAD THING #4: DEALING WITH A POWER OUTAGE
Fortunately, most outages don't last too long.and may only interrupt your computer. But in case of a blizzard or ice storm, a blackout is likely to last a long time. If that happens, here are some suggestions:
•Turn off any electrical devices you were using and pull plugs on electronic equipment. Disconnecting the main breaker can help keep surges from damaging anything when the power come back on.
•Fill containers with water. If the blackout is widespread, water pressure may soon drop. On a private well, keep the water in the pressure tank for drinking only. You may also want to fill bathtubs and containers for flushing toilets and nondrinking water uses.
•Never use a gas oven, kerosene heater, or outdoor cooker for heat unless you have good ventilation, like a partially opened window. Keep in mind that any flame heat source produces carbon monoxide, which can kill you fast.
•If the weather is below freezing, keeping water pipes from freezing is a major concern. When no other heat source is available, open faucets enough to allow a pencil thin stream of water to run. This should prevent pipes from freezing.
•Food will stay cold for up to three days, even during the summer, as long as you keep the refrigerator and freezer doors closed.
•Don't run generators in the garage or an enclosed area. Also, before kicking on the generator, shut off the main breaker so you don't send power into the main lines.
•Most important, put together a power outage kit of several five gallon containers of water, propane or Coleman camping type stoves, two or three propane lanterns, flashlights, extra food, and a couple of propane heaters.
### BAD THING #5: DEALING WITH A DOWNED TREE THAT HIT YOUR HOUSE
When a heavy windstorm topples a tree onto your house, resist the urge to climb up on your roof, especially if it's steeply pitched. This is a job for pros with the right equipment. Otherwise:
•Inspect the underside of the roof from the attic. If the plywood decking or rafters are damaged, you're going to need a roofer. Your first concern is keeping water out of your home and preventing further damage.
•If the roof isn't damaged and is still water tight, then it's a matter of carefully cutting up and removing the limbs with pruning shears and bow saws.
•Chain saws should not be used on a ladder or on roofs where the footing is unstable.
•You'll likely find some damaged shingles. To patch these areas, use a piece of aluminum flashing to cover and slip under the shingles. You may have to remove some nails so the patch will fit. Apply asphalt roofing cement to the edges of the patch where you've inserted it under the shingles. This should hold it in place and keep out water.
•You can also use heavy plastic sheeting for a temporary patch, but you'll need to secure it by nailing furring strips around the edges. Sometimes a bead of roofing cement around the edges of the sheeting will hold it in place, unless you're expecting high winds and rain.
•It a good idea to assemble a roof-patching kit to have on hand. When disaster strikes you may not be able to get to a hardware store. Your kit should have several tubes of roofing cement, heavy 8–12 mil plastic sheeting, a couple of squares of aluminum sheeting, 20 to 30 feet of furring strips, roofing nails, claw hammer, and hacksaw.
## APPENDIX B
## AVERAGE LIFE EXPECTANCY OF YOUR HOME'S COMPONENTS
In the United States, few components wear out before they end up in the trash or at a demolition yard (see Table B-1). Features, styles, and efficiency of use change constantly, and homeowners upgrade every few years to keep up. Your home's components seldom have chance to wear out.
However, one great reason to upgrade is to replace a less energy-efficient appliance or building component with a one that is considerably more efficient. For example, if you have a refrigerator that is fifteen years old or older, you would probably save energy and money by replacing it with a new unit that uses 30 percent less power. The same can be said of water heaters, windows, doors, insulation, and flooring.
So, if you have a component that is approaching the end of its average life span, you should consider replacing it without feeling too guilty. You can save serious money if you keep your eye on sales and closeouts for the best deals, rather than waiting until an appliance fails.
_Table B-1. How long you can expect your home's components to last_.
**Item** | **Average life**
---|---
Dishwasher | 10–12 years
Refrigerator | 12–16 years
Clothes dryer | 10–14 years
Microwave oven | 8–10 years
Electric range | 15–20 years
Gas range | 18–20 years
Cast iron tub | 50 years
Fiberglas tub/shower | 10–15 years
Toilet | 50 years
Laminate countertop | 10–15 years
Ceramic tile countertop | 50 years+
Wood countertop | 15–20 years, if cared for
Granite countertop | 20 years+
Interior hollow-core door | 20–25 years
Interior solid-core door | 30+ years
Exterior door, protected | 60+ years
Exterior door, unprotected | 20+ years (exposed to weather)
Garage door | 30+ years
Garage door opener | 10 years
Wood floors: oak or pine | 50+ years (depending on care)
Slate flooring | 50+ years
Vinyl flooring | 10–20 years (less if you have a dog)
Central air conditioner | 15 years
Electric water heater | 12–15 years
Gas water heater | 10–14 years (less for hard water)
Gas furnace | 15–20 years
Heat exchanger | 20–25 years
Wood deck | 15 years, if you use a water protectant
Brick or concrete patio | 24+ years, if sealed and no frost damage
Asphalt driveway | 10+ years, if resealed every few years
Wood fence | 12 + years, if recoated every few years
Asphalt shingle roof | 15–30 years, depending on shingle quality
Wood shingles | 15+ years, depending on climate
Tile roofing | 50+ years
Slate roofing | 50–100 years
Sheet metal roofing | 25+ years
Wood siding | 10 to 100 years, if it's recoated every few years and there is no water damage
Metal siding | 50+ years
Aluminum siding | 20–50 years
Vinyl siding | 50 years
## APPENDIX C
## ENERGY SAVINGS AND A GREENER HOME
An important part of improving your home's interior is making it as energy efficient and green as possible. Table C-1 lists what you can do around the house to make that happen without spending big bucks.
_Table C-1. Checklist for energy savings and a greener home_.
**Action** | **What to Do**
---|---
Buy energy efficient appliances. | Check out www.energystar.gov for a list of energy-efficient appliances and home improvement information.
Replace single pane with double pane widows. | Shop for windows that are ENERGY STAR-rated. You may be eligible for a federal tax deduction of 10 percent of the purchase price.
Insulate pipes. | Insulate attic, ductwork, pipes, and water heater
Lower water usage. | Replace shower heads with low flow (1.5 gallons per minute or better) heads. Replace older toilets with newer, low-use models. Replace leaking faucets promptly.
Install bathroom fan (to reduce mold and mildew). | Fan should move 50 to 100 cubic feet of air per minute to the outside. Vent should be at least 4 inches in diameter to be effective. Also consider putting a timer on the fan.
Vent kitchen. | ENERGY STAR fans use up to 65 percent less energy and are quieter as well.
Clean your home. | Vegetable-based, environmentally friendly cleaning products and homemade cleaners can get the job done without causing skin and respiratory problems. Check the labels at grocery stores for these kinds of products.
Reduce power consumption. | Use compact fluorescent bulbs instead of incandescent bulbs, which use 50 percent more power.
Prevent heating and cooling losses. | Make sure doors and windows are sealed. Check outlet boxes, bedroom closets, and attic insulation.
Keep your furnace working efficiently. | Replace the furnace filter once an month.
Adjust your thermostat. | Make your home 1° to 2° cooler and you won't notice the change except when the utility bill arrives.
Decorate with indoor plants. | With today's airtight homes, indoor air quality can be improved with plants. Check with a local nursery for plants that work best in your area.
## INDEX
The index that appeared in the print version of this title was intentionally removed from the eBook. Please use the search function on your eReading device for terms of interest. For your reference, the terms that appear in the print index are listed below
amps
defined
required by appliances
_vs_. volts
apartment additions
websites on renting
and zoning restrictions
appliances
amps required by
calculating wattage rating
with ENERGY STAR label
life expectancy
asbestos
in "cottage cheese" ceilings
in plaster walls
asphalt driveways
_vs_. concrete
resealing scams
specifications
websites
augers
basements
mold in
water problems in
bathroom remodeling
caulking
matching fixtures
return on cost
websites
bay windows
bedrooms
redecorating options
websites
breaker box, _see_ service panel
breakers
function of
replacing switches
tripping of
troubleshooting
cabinets
custom
low-cost options
second-hand
semi-custom
stock
tips for refinishing
carpeting
characteristics
cost
purchasing tips
casement windows
caulking
inspecting for
professional techniques
types of caulk
ceilings
testing for asbestos
tips for painting
water damage
cellars, _see_ basements
circuits
calculating maximum wattage
identification of
importance of mapping
circuit tester
clogs
augers used for
in pipes and drains
in toilets
concrete driveways
minor repairs
proper sealing
websites
concrete steps
fixing chipped corners
repairs
condos and co-ops
insurance for
umbrella liability
unit assessment
contract
capped cost-plus (hybrid)
cost-plus option
fixed-fee option
key elements of
protection against liability
website information
contractors
accepting a bid
checking past clients
contract with, _see_ contract
hiring replacements
interviewing
mistakes made in working with
payment of
referral sources for
replacing
shopping for
websites for finding
countertops
as kitchen's focal point
options
crown molding, painting of
curb appeal, _see also_ landscaping
homeowners' mistakes
key elements of
double-hung windows
drains
unclogging
driveways
asphalt
concrete
drywall
priming
repairing
water damage
DWV (drain, waste, vent)
earthquake insurance
cost calculation
deductible in high-risk states
websites
electrical contractor, finding the right
electrical service entrance
electrical system
cost of upgrading
cutting the power
dealing with power outage
determining capacity of
faulty wiring
mapping of
probing for hot wires
problem solving
reasons for upgrading
safety tips
troubleshooting guide
websites
electrical usage
calculating cost of
verifying usage
energy-saving products
ENERGY STAR label
for appliances
tax credit for
websites
for windows
exterior improvements
adding value through
maintenance checklist
preventing water damage
exterior painting
choosing colors
cost
_vs_. staining
surface prepping
websites
family rooms
redecorating options
return on cost
faucet repair
of cartridge-style faucets
of one-handled faucets
of tipping valve faucets
tools needed for
of two-handled faucets
faucet replacement
mounting the new faucet
removing the old faucet
tools needed for
faucets
brands
types of
federal taxes
deductions for home office
helpful IRS publications
improvements effect on tax basis
IRS rules effect on tax basis
mistakes by homeowners
principal residence use test
record keeping as key
tax credit for energy-related improvements
tax-deferred exchanges, _see_ 1031 tax-deferred exchanges
websites
fences
regulations for
sharing cost with neighbors
fertilizer
fire damage
tips for handling cleanup
websites
fireplaces
maintenance checklist
websites
flood insurance
average premiums
FEMA rate map
NFIP policy as only coverage
websites
flooring
carpeting
matching to wall color
tile
upgrading of
websites
wood
foundations
uneven settling of
websites
fuse box, _see_ service panel
fuses
blown
function of
overfusing
troubleshooting
garage doors
curb appeal
options and cost
websites
gaskets, replacement of
grading
water damage created by poor
websites
ground-fault circuit interrupter (GFI)
gutters, _see_ rain gutters
heating systems
and carbon monoxide leaks
tips for prolonging life of
websites
hedges
need for trimming
regulations for
home improvement services
hiring small _vs_. large company
hiring tradespeople for
home inventory
guidelines for creating
using PCs and videotape
websites
home office
liability insurance for
tax deductions for
websites
homeownership problems
checklist for avoiding
deferred maintenance
with foundation
with grading and drainage
with gutters
with heating system
with plumbing
with roof
with ventilation
with water in cellar
with wiring
homeowner's insurance
as an investment
for condos and co-ops
discounts
extended coverage
factors that determine cost
guaranteed replacement
inventory as key, _see_ home inventory
liability coverage
mistakes by homeowners
for older homes
Ordinance and Law rider
personal property coverage
policy options
replacement coverage
standard coverage
structure coverage
water back-up rider
websites
home security
dealing with a robbery
website
hydro-seeding
insurance, _see_ earthquake insurance; flood insurance; homeowner's insurance; title insurance
insurance company
considerations in choosing
tips for getting best deal
websites to help pick
interior improvements, _see also_ individual rooms
adding value through
mistakes by homeowners
interior painting
applying the top coat
basic steps
of ceilings
of crown molding
equipment checklist
prep work
priming
professional techniques
tools needed for
using masking tape
of walls
websites
IRS Publication 523
kitchen remodeling adding value
appliances
cabinets
countertops
decorating tips
light fixtures
landscaping, _see also_ exterior improvements; lawn care; trees and shrubs
basic plan
and Golden Mean
grading and drainage problems
key design elements
websites
laundry room
upgrade options
websites
lawn care
aerating
controlling thatch
edging
fertilizing
mowing
seeding
soil testing
tips
watering
websites
licensed electrician, finding the right
life expectancy
of appliances
of home's components
of roofs
light fixtures
cable-and-rail systems
websites
living rooms, redecorating options
main disconnect
MLS (multiple listing service)
mold
in basements
websites
mortgage life insurance
coverage
_vs_. private mortgage insurance
_vs_. term insurance
websites
national electrical code (NEC)
National Flood Insurance Program (NFIP)
nylon carpeting
outlets, replacing
paint
buying where the pros shop
cost calculation
estimating quantity needed
picking the right top coat
tips for choosing colors
websites for choosing colors
paint brushes
cleaning of
quality as key
painting, _see_ exterior painting; interior painting; paint
pavers, installing a walkway
pipes
unclogging
unfreezing frozen
winterizing
plaster walls
priming of
repairing
testing for asbestos
plumbing system
cost of upgrading
DWV (drain, waster, vent)
incoming components
reasons for upgrading
shutoff valve
"switch box" for
troubleshooting guide
watching for water damage
websites
polyester carpeting
power meter
reading the
types of
power outage
priming
picking the right primer
tips
private mortgage insurance (PMI)
coverage
premiums
property tax bill appeal
formal appeal with local board
grounds for
meeting with tax assessor
preparing for
websites
property taxes
appealing, _see_ property tax bill appeal
assessment determination
checklist for evaluating
mistakes by homeowners
tax scams
websites
pruning
common mistakes
tips
punch-list walk-through
checklist for
websites
rain gutters
faulty gutters problem
gutter guards
maintenance checklist
options and cost
regular inspection
websites
remodeling project, _see also_ contractors
contributing sweat equity
punch-list walk-through
tips for smooth operation
roofs
extending life of
inspecting for water damage
lifespan
problem-solving tips
return on replacement cost
types and cost
warning signs
websites
service entrance, water valve
service panel
function
location
sheetrock, _see_ drywall
siding
cedar
comparison of types
EIFS (synthetic stucco)
fiber/cement
maintenance checklist
stucco
vinyl/aluminum
websites
wood
slider windows
soil testing
websites
staining
advantages
_vs_. painting
surge protectors, purpose of
sweat equity
negotiating specifics of
websites
swimming pools
liability insurance for
maintenance of
websites
switches, replacing
1031 tax-deferred exchanges
key steps in
rules for
use of intermediaries
websites
when to use
taxes, _see_ federal taxes; property taxes
thatch
control of
websites
title insurance
coverage
websites
toilets
cause of leaks
components
fixing water in-take problems
insulating the tank
loosening stuck nuts
maintenance of
repair of leaks
replacing gaskets
shutoff valve for
unclogging
tradespeople
developing relationships with
hiring the right
websites
trees and shrubs, pruning of
"U" traps
ventilation
need for adequate
websites
vents
volts
_vs_. amps
defined
walkways
concrete
dry-laid pavers
wall-mounting
tips
websites
wall repair
covering popped fasteners
patching cracks and holes
problem-solving checklist
tools for
websites
walls
anatomy of
caring for
painting of
preparing for painting
types of
water damage
in basement
checklist for handling leak
from deferred maintenance
from faulty gutters
from faulty plumbing
from leaks in roof
from poor drainage
from poor ventilation
water hammer
water in-take system, in toilets
water usage, tips for conserving
watts
defined
typical ratings of appliances
wind damage
dealing with downed tree
need for roof-patching kit
window replacement
available styles
energy savings
options
return on cost
shopping tips
tax credit for
websites
window sashes, painting of
wood floors
cost of
types of
wool carpeting
**Look for These Exciting Real Estate Titles at
www.amacombooks.org/gorealestate**
_A Survival Guide for Buying a Home_ by Sid Davis $17.95
_A Survival Guide for Selling a Home_ by Sid Davis $15.00
_Are You Dumb Enough to Be Rich?_ by G. William Barnett II $18.95
_Everything You Need to Know Before Buying a Co-op, Condo, or Townhouse_ by Ken Roth $18.95
_Make Millions Selling Real Estate_ by Jim Remley $18.95
_Mortgages 101_ by David Reed $16.95
_Mortgage Confidential_ by David Reed $16.95
_Real Estate Investing Made Simple_ by M. Anthony Carr $17.95
_Real Estate Presentations That Make Millions_ by Jim Remley $18.95
_The Complete Guide to Investing in Foreclosures_ by Steve Berges $17.95
_The Consultative Real Estate Agent_ by Kelle Sparta $17.95
_The Home Buyer's Question and Answer Book_ by Bridget McCrea $16.95
_The Landlord's Financial Tool Kit_ by Michael C. Thomsett $18.95
_The Property Management Tool Kit_ by Mike Beirne $19.95
_The Real Estate Agent's Business Planner_ by Bridget McCrea $19.95
_The Real Estate Agent's Field Guide_ by Bridget McCrea $19.95
_The Real Estate Investor's Pocket Calculator_ by Michael C. Thomsett $17.95
_The Successful Landlord_ by Ken Roth $19.95
_Who Says You Can't Buy a Home!_ by David Reed $17.95
_Your Successful Career as a Mortgage Broker_ by David Reed $18.95
_Your Successful Real Estate Career, Fifth Edition_ , by Kenneth W. Edwards $18.95
**Available at your local bookstore, online, or call 800-250-5308.**
**Savings start at 35%** on bulk orders of 5 copies of more!
**Save up to 55%!**
For details, contact AMACOM Special Sales
Phone: 212-903-8316. E-mail: SpecialSls@amanet.org
_Special discounts on bulk quantities of AMACOM books are available to corporations, professional associations, and other organizations. For details, contact Special Sales Department, AMACOM, a division of American_
_Management Association_ ,
_1601 Broadway, New York, NY 10019_.
_Tel: 212-903-8316. Fax: 212-903-8083_.
_E-mail:specialsls@amanet.org_
_Website:www.amacombooks.org/go/specialsales_
_To view all AMACOM titles go to:www.amacombooks.org_
_This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold with the understanding that the publisher is not engaged in rendering legal, accounting, or other professional service. If legal advice or other expert assistance is required, the services of a competent professional person should be sought_.
_REALTOR® is a registered collective membership mark that identifies a real estate professional who is a member of the National Association of REALTORS® and subscribes to its strict Code of Ethics. AMACOM uses these names throughout this book in initial capital letters or ALL CAPITAL letters for editorial purposes only, with no intention of trademark violation_.
_Library of Congress Cataloging-in-Publication Data_
_Davis, Sid_.
_The first-time homeowner's survival guide : a crash course in dealing with repairs, renovations, property tax issues, and other potential disasters / Sid Davis_.
_p. cm_.
_Includes bibliographical references and index_.
_ISBN: 978-0-8144-3462-8 (hardc)_
_ISBN: 978-0-8144-7372-6 (pbk)_
_ISBN: 978-0-8144-0058-6 (ebook)_
_1. Home ownership—United States.I.Title_.
_HD7287.82.U6D38 2007_
_643—dc22 2007003725_
_© 2007 Sid Davis_.
_All rights reserved_.
_Printed in the United States of America_.
_This publication may not be reproduced_ ,
_stored in a retrieval system_ ,
_or transmitted in whole or in part_ ,
_in any form or by any means, electronic_ ,
_mechanical, photocopying, recording, or otherwise_ ,
_without the prior written permission of AMACOM_ ,
_a division of American Management Association_ ,
_1601 Broadway, New York, NY 10019_.
_Printing number_
_10 9 8 7 6 5 4 3 2 1_
|
{
"redpajama_set_name": "RedPajamaBook"
}
| 2,758
|
The City of Worthington provides refuse, recycling and yard waste collections to all single-family, two-family and condominium residences within the City. Collection occurs each Friday between 6:00 am and 6:00 pm, unless otherwise noted due to holidays or natural disasters. Residents are not billed directly for this extensive service.
All items are to be placed at the curb or edge of the paved roadway no later than 6:00 am on Friday. By City ordinance, no items may be placed curbside before 6:00 am on Thursday, and empty containers must be removed by 6:00 pm on Saturday. Crews are not required to pick up any solid waste set at the curb late.
The City of Worthington has contracted the following collection services with Local Waste Services. They can be contacted by calling 614-409-9375 or by visiting the Local Waste Services website.
If you experience any problems with your refuse, recycling or yard waste services, please contact the Worthington Service & Engineering Department at (614) 431-2425. Calls received after 4:30PM on Friday will be addressed Monday morning. You may also file your service request electronically on the by visiting the Local Waste Services website.
If a container is damaged by fault of the collection contractor, please report the incident directly to Local Waste Services by calling 614-409-9375. Damaged recycling bins may be exchanged for new bins at no additional fee.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 2,954
|
If you or a loved one suffered from a heart attack or stroke while using AndroGel you may be entitled to damages. We provide free no-obligation Mississippi testosterone lawsuit case review. We do not charge attorney fees unless you receive compensation and you are under no obligation after your initial consultation.
AndroGel and other similar testosterone replacement products have been linked to a two- to three-fold increase in the risk of heart attack for men. Numerous studies published in recent years have contributed to researchers' understanding of the testosterone heart attack risk. Advertising for drugs such as AndroGel is criticized for making false claims about the product's efficacy while glossing over the dangers of heart attack.
Filing a Mississippi testosterone lawsuit may be the best recourse consumers in Mississippi have to push for corporate accountability for the injustices done to men who have suffered heart attacks while taking AndroGel. This page includes information about testosterone replacement therapy in general, the findings on AndroGel heart attacks, and information about testosterone heart attack warning statements for Mississippi residents.
The FDA has only approved the use of testosterone products for medical complications caused by genuine hypogonadism. Today, many clinicians and researchers warn against the broad use of AndroGel because of the risk of heart attack. The FDA has issued a testosterone heart attack warning, suggesting that the risk of heart attack associated with AndroGel and similar products outweighs the drug's proven benefits.
The FDA AndroGel Heart Attack Warning was issued on January 31, 2014. Citing numerous recent studies, the FDA heart attack warning states that federal regulators have undertaken an investigation of the link between testosterone replacement drugs and heart attack, in addition to other health problems. While the FDA testosterone warning recommends that patients continue taking the drug until consulting with their care providers, doctors are cautioned to weigh the pros and cons of the drug before prescribing it. Additionally, the FDA AndroGel warning encourages clinicians and patients in Mississippi and throughout the country to make reports to the FDA's MedWatch program to contribute to the body of knowledge on testosterone heart attack and death risks.
The heart attack risk is traced to the fact that testosterone supplementation increases red blood cell production. Red blood cells can clump together or coagulate, so an abundance of them can thicken the blood. AndroGel heart attacks and other cardiovascular problems are thought to originate from this thickening of the blood.
Hormone replacement therapies have a history of causing medical complications. For a time, estrogen replacement therapy was thought to resolve a range of symptoms of menopause in women, and a rash of prescribing took place. The overprescribing of estrogen replacement therapy was reined in by findings that hormone replacement in women increases the risk of heart attack, breast cancer and stroke. Today, estrogen is used sparingly and only in extreme cases. Experts speculate that as the testosterone heart attack risk becomes public knowledge, the number of prescribers will nosedive.
Research specifically indicating a link between testosterone and heart attacks first emerged in 2010, when a small study conducted at Boston University's School of Medicine (and later published by the New England Journal of Medicine) was forced to end early due to exceptionally high heart attack rates for subjects taking Testim 1%. The findings from that research suggested men taking testosterone products were five times as likely to have heart attacks or another "serious heart event" as men not taking the supplemental hormone.
The two notable AndroGel heart attack studies that directly resulted in the FDA warning are very recent. Research conducted through the Veterans Affairs health service examined 8,709 men through a type of heart imaging called coronary angiography and found that taking supplemental testosterone increases the risk of heart attack, stroke and death by 29%. The average age of men involved in this study was 60 years old. This study was published by JAMA (the Journal of the American Medical Association) in November of 2013.
The second study cited in the FDA AndroGel warning examined 56,000 middle aged and older men for their first 90 days taking testosterone replacement products. This study concluded that men taking AndroGel face a two- to three-fold risk for heart attack, depending on their age, immediately after starting the prescription. Citing a tendency for hypogonadism and obesity to present with identical symptoms, researchers suggested that many men are being prescribed with testosterone drugs unnecessarily, meaning that many men are facing the risk of heart attack unnecessarily. This most recent AndroGel heart attack research was funded by the National Institutes of Health, conducted by a group of endocrinologists and endocrine research experts in California, and published by PLoS ONE in January of 2014.
Testosterone replacement therapy (TRT) is a hormone replacement method for men suffering from hypogonadism, a medical condition in which a man's body produces little or no serum testosterone. Hypogonadism can be congenital or acquired; in fact, a decline in testosterone is a natural part of aging. Men over the age of 30 have, on average, a 1% loss of testosterone on an annual basis. The natural decline of testosterone is one aspect of andropause, a term that refers to the male aging process.
While small declines in testosterone are normal and typically don't cause severe medical conditions, hypogonadism is an endocrine disorder characterized by a lack of muscle and body hair, low libido, and sometimes more severe medical complications. Hypogonadism can be caused by varying medical conditions, including an inability of the testicles to produce testosterone as a result of genetics or chemotherapy, or a problem with the hypothalamus and pituitary portions of the brain that control hormone production. AndroGel and other similar products are approved to treat genuine hypogonadism only, which must be diagnosed through blood testing.
Testosterone products are available in a variety of forms including topical gel, transdermal patch, buccal system (applied on the upper gums or inner cheek), and injection. Numerous companies produce hormone replacement products, the most popular of which is AndroGel by AbbVie.
Drug companies have adopted the name "Low T" to refer to hypogonadism; Low T has become a catch-all phrase for a range of ailments facing aging men. Critics assert that aggressive advertising campaigns launched by AbbVie and other companies, misleadingly labeled as "Disease Awareness" campaigns feature broad and general questionnaires that anyone can fail, resulting in a huge number of men self-diagnosing for Low T. These advertising campaigns suggest that products like AndroGel can be used to increase energy, boost libido, resolve erectile dysfunction (ED), improve cognitive thinking, build muscle mass and bone density, enhance athletic performance, and shed extra weight, among other claims. Due to a lack of research on the topic, no one knows if testosterone products really can achieve each of these claims.
Despite the emergence of AndroGel heart attack evidence in recent years, AbbVie and other producers of testosterone products have $107.3 million in advertising in 2012 (compared to $14.3 million in 2011) to promote hormone replacement therapies to aging men suggesting it will help them feel youthful. This surge in advertising suggests manufacturers of AndroGel and similar products saw a huge profit opportunity in testosterone. The number of prescriptions issued for testosterone has tripled since 2001. In 2012, 3 million prescriptions were written for AndroGel alone. The sales of testosterone replacement drugs amassed $2 billion on profits during 2012, and are projected to climb to $5 billion by 2017.
Some men who have suffered from heart attacks have already filed testosterone lawsuits, alleging that AbbVie failed to adequately warn consumers of the heart attack risk. Testosterone heart attack lawsuits are expected to grow in number as the heart attack risk becomes public knowledge.
Consumer Reports warns men of the dangers of testosterone. Dr. John Santa, quoted in the Consumer Reports testosterone heart attack warning, believes many men are using AndroGel and other comparable products unnecessarily and that most men don't need hormone replacement therapy. Citing side effects that include heart attack and other cardiovascular events, Consumer Reports reminds consumers that testosterone therapy is not necessarily effective. In particular, this AndroGel heart attack warning notes the treatment has not been approved for or shown to resolve erectile dysfunction (ED). The American Urological Society has included testosterone therapy on its list of overused or dangerous treatments, echoing the warning that AndroGel is not useful for treating erectile dysfunction, and noting that male hormone treatment is accompanied by too many severe risks such as heart attack.
Our attorneys handling testosterone heart attack lawsuits in Mississippi on behalf of men and their family members who suffered from a heart attack while taking AndroGel or another comparable product are drug safety litigation experts. They will work tirelessly on your behalf to assure you get the compensation you deserve. If you or your father, spouse, or other loved one suffered from a heart attack and took testosterone, a Mississippi AndroGel heart attack attorney can help by answering your questions and giving you a sense of your legal options.
Mississippi men who suffered one or more heart attacks and have taken AndroGel may have grounds for a Mississippi AndroGel heart attack lawsuit. We provide legal representation for Mississippi testosterone heart attack lawsuits on contingency basis, meaning that we charge no fee unless we win compensation on your behalf.
Read here for AndroGel Testosterone Stroke Information.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 6,487
|
\section{Introduction and main results}
In recent decades, much attention has been paid to extremal problems involving eigenvalues, and in particular to shape optimization problems in which the unknown is the domain where the eigenvalue problem is solved (see e.g. \cite{Henrot2006,henrot2017shape} for a survey).
The study of these last problems is motivated by stability issues of vibrating bodies, wave propagation in composite environments, or also on conductor thermal insulation.
In this article, we are interested in studying a particular extremal eigenvalues problem, involving a drift term, and which comes from the study of mathematical biology problems; here, we can show that the problem then boils down to a "two-phase" type problem, meaning that the differential operator whose eigenvalues we are trying to optimise has $-\nabla \cdot(A\nabla)$ as a principal part, and that $A$ is an optimisation variable, see Section \ref{BiologicalMotivations}. The influence of drift terms on optimal design problems is not so well understood. Such problems naturally arise for instance when looking for optimal shape design for two-phase composite materials \cite{DambrineKateb,Laurain,MuratTartar}. We expand on the bibliography in Section \ref{SSE:BackgroundH} of this paper, but let us briefly recall that, for composite materials, a possible formulation reads: given ${\Omega}$, a bounded connected open subset of $\textrm{I\kern-0.21emR}^n$ and a set of admissible non-negative densities $\mathcal{M}$ in ${\Omega}$, solve the optimal design problem
\begin{equation}\tag{$\hat P_\alpha$}\label{Pg:TwoPhase}
\inf_{m\in \mathcal M} \hat\lambda_\alpha(m)
\end{equation}
where $\hat \lambda_\alpha(m)$ denotes the first eigenvalue of the elliptic operator
$$
\hat{ \mathcal L}_\alpha^m:W^{1,2}_0({\Omega})\ni u\mapsto -{\nabla} \cdot \left((1+\alpha m){\nabla} u\right).
$$
Restricting the set of admissible densities to {\it bang-bang} ones (in other words to functions taking only two different values) is known to be relevant for the research of structures optimizing the compliance. We refer to Section \ref{Pr:NonExistence} for detailed bibliographical comments.
Mathematically, the main issues regarding Problem \eqref{Pg:TwoPhase} concern the existence of optimal densities in $\mathcal M$, possibly the existence of optimal {\it bang-bang} densities (i.e characteristic functions). In this case, it is interesting to try to describe minimizers in a qualitative way.
In what follows, we will consider a refined version of Problem \eqref{Pg:TwoPhase}, where the operator $\hat{ \mathcal L}_\alpha^m$ is replaced with
\begin{equation}\label{def:Lmalpha}
\mathcal L_m^\alpha:W^{1,2}_0({\Omega})\ni u\mapsto -{\nabla} \cdot \left((1+\alpha m){\nabla} u\right)-mu.
\end{equation}
Besides its intrinsic mathematical interest, the issue of minimizing the first eigenvalue of $ \mathcal L_m^\alpha$ with respect to densities $m$ is motivated by a model of population dynamics (see Section~\ref{BiologicalMotivations}).
Before providing a precise mathematical frame of the questions we raise in what follows, let us roughly describe the main results and contributions of this article:
\begin{itemize}
\item by adapting the methods developed by Murat and Tartar, \cite{MuratTartar}, and Cox and Lipton, \cite{CoxLipton}, we show that the first eigenvalue of $ \mathcal L_m^\alpha$ has no regular minimizer in $\mathcal M$ unless $\Omega$ is a ball;
\item if $\Omega$ is a ball, denoting by $m_0^*$ a minimizer of $ \mathcal L_m^0$ over $\mathcal M$ (known to be {\it bang-bang} and radially symmetric), we show the following stationarity result: $m_0^*$ still minimizes $ \mathcal L_m^\alpha$ over radially symmetric distributions of $\mathcal M$ whenever $\alpha$ is small enough and in small dimension ($n=1,2,3$). Such a result appears unexpectedly difficult to prove. Our approach is based on the use of a well chosen path of quasi-minimizers and on a new type of local argument.
\item if $\Omega$ is a ball, we investigate the local optimality of ball centered distributions among all distributions and prove a quantitative estimate on the second order shape derivative by using a new approach relying on a kind of comparison principle for second order shape derivatives.
\end{itemize}
Precise statements of these results are given in Section \ref{sec:mainRes}.
\subsection{Mathematical setup}
Throughout this article, $m_0$, $\kappa$ are fixed positive parameters.
Since in our work we want to extend the results of \cite{LamboleyLaurainNadinPrivat}, let us define the set of admissible functions
$$
\mathcal M_{m_0,\kappa}(\Omega)=\left\{m\in L^\infty({\Omega})\, , 0\leq m\leq \kappa\, ,\fint_{\Omega} m=m_0\right\},
$$
where $\fint_{\Omega} m$ denotes the average value of $m$ (see Section~\ref{sec:notations}) and assume that $m_0<\kappa$ so that $\mathcal M_{m_0,\kappa}({\Omega})$ is non-empty.
Given $\alpha\geq 0$ and $m\in \mathcal M_{m_0,\kappa}({\Omega})$, the operator $\mathcal L_m^\alpha$ is symmetric and compact. According to the spectral theorem, it is diagonalizable in $L^2({\Omega})$. In what follows, let $\lambda_\alpha(m)$ be the first eigenvalue for this problem. According to the Krein-Rutman theorem, $\lambda_\alpha(m)$ is simple and its associated $L^2({\Omega})$-normalized eigenfunction $u_{\alpha,m}$ has a constant sign, say $u_{\alpha,m}\geq 0$. Let $R_{\alpha,m}$ be the associated Rayleigh quotient given by
\begin{equation}
R_{\alpha,m}:W^{1,2}_0({\Omega})\ni u\mapsto \frac{\int_{\Omega} (1+\alpha m)|{\nabla} u|^2-\int_{\Omega} mu^2}{\int_{\Omega} u^2}.\end{equation}
We recall that $\lambda_\alpha(m)$ can also be defined through the variational formulation
\begin{equation}\label{Eq:Rayleigh} \lambda_{\alpha}(m):=\inf_{u\in W^{1,2}_0({\Omega})\, , u\neq 0} R_{\alpha,m}(u)=R_{\alpha,m}(u_{\alpha,m}).
\end{equation}
and that $u_{\alpha,m}$ solves
\begin{equation}\label{Eq:EigenFunction}
\left\{
\begin{array}{ll}
-{\nabla} \cdot \Big((1+\alpha m){\nabla} u_{\alpha,m}\Big)-mu_{\alpha,m}=\lambda_\alpha(m)u_{\alpha,m}&\text{ in }{\Omega},
\\u_{\alpha,m}=0\text{ on }\partial {\Omega}.
\end{array}
\right.
\end{equation}
in a weak $W^{1,2}_0({\Omega})$ sense.
In this article, we address the optimization problem
\begin{equation}\label{Pb:OptimizationEigenvalue}\tag{$P_\alpha$}\fbox{$\displaystyle
\inf_{m\in \mathcal M_{m_0,\kappa}({\Omega})}\lambda_\alpha(m).$}
\end{equation}
This problem is a modified version of the standard two-phase problem studied in \cite{MuratTartar,CoxLipton}; we detail the bibliography associated with this problem in Subsection \ref{SSE:BackgroundH}. It is notable that it is relevant in the framework of population dynamics,
when looking for optimal resources configurations in a heterogeneous environment for species survival, see Section \ref{BiologicalMotivations}.
\subsection{Main results}\label{sec:mainRes}
Before providing the main results of this article, we state a first fundamental property of the investigated model, reducing in some sense the research of general minimizers to the one of {\it bang-bang} densities. It is notable that, although the set of {\it bang-bang} densities is known to be dense in the set of all densities for the weak-star topology, such a result is not obvious since it rests upon continuity properties of $\lambda_\alpha$ for this topology. We overcome this difficulty by exploiting a convexity-like property of $\lambda_\alpha$.
\begin{proposition}[weak {\it bang-bang} property]\label{Th:BangBang}
Let ${\Omega}$ be a bounded connected subset of $\textrm{I\kern-0.21emR}^n$ with a Lipschitz boundary and let $\alpha>0$ be given. For every $m\in \mathcal M_{m_0,\kappa}({\Omega})$, there exists a {\it bang-bang} function $\tilde m\in \mathcal M_{m_0,\kappa}({\Omega})$ such that
$$\lambda_\alpha(m)\geq \lambda_\alpha(\tilde m).$$
Moreover, if $m$ is not {\it bang-bang}, then we can choose $\tilde m$ so that the previous inequality is strict.
\end{proposition}
In other words, given any resources distribution $m$, it is always possible to construct a {\it bang-bang} function $\tilde m$ that improves the criterion.
\paragraph{Non-existence for general domains.}
In a series of paper, \cite{CasadoDiaz1,CasadoDiaz2,CasadoDiaz3}, Casado-Diaz proved that the problem of minimizing the first eigenvalue of the operator $u\mapsto -{\nabla} \cdot(1+\alpha m){\nabla} u$ with respect to $m$ does not have a solution when $\partial {\Omega}$ is connected. His proof relies on a study of the regularity for this minimization problem, on homogenization and on a Serrin type argument. The following result is in the same vein, with two differences: it is weaker than his in the sense that it needs to assume higher regularity of the optimal set, but stronger in the sense that we do not make any strong assumption on $\partial {\Omega}$. For further details regarding this literature, we refer to Section \ref{TwoPhase}.
\begin{theorem}\label{Th:NonExistence}
Let ${\Omega}$ be a bounded connected subset of $\textrm{I\kern-0.21emR}^n$ with a Lipschitz connected boundary, let $\alpha>0$ and $n\geq 2$. If the optimization problem \eqref{Pb:OptimizationEigenvalue} has a solution $\hat m \in \mathcal M_{m_0,\kappa}({\Omega})$, then this solution writes $\hat m=\kappa \mathds{1}_{\hat E}$, where $\hat E$ is a measurable subset of ${\Omega}$. Moreover, if $\partial \hat E$ is a $\mathscr C^2$ hypersurface and if ${\Omega}$ is connected, then ${\Omega}$ is a ball.
\end{theorem}
The proof of this Theorem relies on methods developed by Murat and Tartar, \cite{MuratTartar}, Cox and Lipton, \cite{CoxLipton}, and on a Theorem by Serrin \cite{Serrin1971}.
\paragraph{Analysis of optimal configurations in a ball.}
According to Theorem \ref{Th:NonExistence}, existence of regular solutions fail when ${\Omega}$ is not a ball. This suggest to investigate the case ${\Omega}={\mathbb B(0,R)}(0,R)$, which is the main goal of what follows.
Let us stress that proving the existence of a minimizer in this setting and characterizing it is a hard task. Indeed, to underline the difficulty, notice in particular that none of the usual rearrangement techniques (the Schwarz rearrangement or the Alvino-Trombetti one, see Section~\ref{TwoPhase}), that enable in general to reduce the research of solutions to radially symmetric densities, and thus to get compactness properties, can be applied here.
\begin{center}
\textsf{\large{The case of radially symmetric distributions}}
\end{center}
Here, we assume that ${\Omega}$ denotes the ball $\mathbb B(0,R)$ with $R>0$. We define $r_0^*>0$ as the unique positive real number such that
$$\kappa \frac{\left|\mathbb B(0,r_0^*)\right|}{\left|\mathbb B(0,R)\right|}=V_0.
$$
Let
{\begin{equation} \label{eq:m0*}m_0^*=\kappa \mathds{1}_{\mathbb B(0,r_0^*)}=\kappa \mathds{1}_{E^*_0}\end{equation} }
be the centered distribution known to be the unique minimizer of $\lambda_0$ in $\mathcal M_{m_0,\kappa}({\Omega})$ (see e.g. \cite{LamboleyLaurainNadinPrivat}).
In what follows, we restrict ourselves to the case of radially symmetric resources distributions.
\begin{theorem}\label{Th:RadialStability}
Let ${\Omega}=\mathbb B(0,R)$ and let $\mathcal M_{rad}$ be the subset of radially symmetric distributions of $ \mathcal M_{m_0,\kappa}({\Omega})$. The optimization problem
$$\inf_{m\in \mathcal M_{rad}}\lambda_\alpha(m)$$ has a solution. Furthermore, when $n=1,2,3$, there exists $\alpha^*>0$ such that, for any $\alpha<\alpha^*$, there holds
\begin{equation}
\min_{m\in \mathcal M_{rad}}\lambda_\alpha(m)=\lambda_\alpha(m^*_0).\end{equation}\end{theorem}
The proof of the existence part of the theorem relies on rearrangement techniques that were first introduced by Alvino and Trombetti in \cite{AlvinoTrombetti} and then refined in \cite{ConcaMahadevanSanz}.
The stationarity result, i.e the fact that $m_0^*$ is a minimizer among radially symmetric distributions, was proved in the one-dimensional case in \cite{CaubetDeheuvelsPrivat}. To extend this result to higher dimensions, we developed an approach involving a homogenized version of the problem under consideration. The small dimensions hypothesis is due to a technical reason, which arises when dealing with elliptic regularity for this equation.
Restricting ourselves to radially symmetric distributions might appear surprising since one could expect this result to be true without restriction, in $ \mathcal M_{m_0,\kappa}({\Omega})$. For instance, a similar result has been shown in the framework of two-phase eigenvalues \cite{ConcaMahadevanSanz}, as a consequence of the Alvino-Trombetti rearrangement. Unfortunately, regarding Problem \eqref{Pb:OptimizationEigenvalue}, no standard rearrangement technique leads to the { conclusion}, because of the specific form of the involved Rayleigh quotient. A first attempt in the investigation of the ball case is then to consider the case of radially symmetric distributions. It is notable that, even in this case, the proof appears unexpectedly difficult.
Finally, we note that, as a consequence of the methods developed to prove Theorem~\ref{Th:RadialStability}, when a small amount of resources is available, the centered distribution $m^*_0$ is optimal among all resources distributions, regardless of radial symmetry assumptions.
\begin{corollary}\label{Th:Sketch}
Let ${\Omega}=\mathbb B(0,R)$ and $m_0^*$ be defined {by (\ref{eq:m0*})} There exist $\underline m>0$, $\underline \alpha>0$ such that, if
$m_0\leq \underline m$ and $\alpha<\underline\alpha$, then the unique solution of \eqref{Pb:OptimizationEigenvalue} is $m^*_0=\kappa \mathds 1_{E^*_0}.$
\end{corollary}
\begin{center}
\textsf{\large{Local stability of the ball distribution with respect to Hadamard perturbations of resources sets}}
\end{center}
In what follows, we tackle the issue of the local minimality of $m^*_0$ in $\mathcal M_{m_0,\kappa}({\Omega})$ with the help of a shape derivative approach. We obtain partial results in dimension $n=2$.
Let ${\Omega}$ be a bounded connected domain with a Lipschitz boundary, and consider a {\it bang-bang} function $m\in \mathcal M_{m_0,\kappa}({\Omega})$ writing $m=\kappa \mathds{1}_E$, for a measurable subset $E$ of ${\Omega}$ such that
$\kappa |E|=m_0|{\Omega}|$. Let us write $\lambda_\alpha(E):=\lambda_\alpha\left(\mathds{1}_E\right)$, with a slight abuse of notation.
Let us assume that $E$ has a $\mathscr C^2$ boundary. Let $V:{\Omega}\rightarrow \textrm{I\kern-0.21emR}^n$ be a $W^{3,\infty}$ vector field with compact support, and define for every $t$ small enough, $E_t:=\left(\operatorname{Id}+tV\right)E$.
For $t$ small enough, $\phi_t:=\operatorname{Id}+tV$ is a smooth diffeomorphism from $E$ to $E_t$, and $E_t$ is an open connected set with a $\mathscr C^2$ boundary.
If $\mathcal F:E\mapsto \mathcal F(E)$ denotes a shape functional, the first (resp. second) order shape derivative of $\mathcal F$ at $E$ in the direction $V$ is
$$
\mathcal F'(E)[V]:=\left.\frac{d}{dt}\right|_{t=0}\mathcal F(E_t) \quad \left(\text{ resp. }\left.\frac{d^2}{d t^2}\right|_{t=0} \mathcal F(E_t)\right)
$$
whenever these quantities exist.
For further details regarding the notion of shape derivative, we refer to \cite[Chapter 5]{HenrotPierre}.
It is standard to write optimality conditions in terms of a sort of tangent space for the measure constraint: indeed, since the volume constraint $\operatorname{Vol}(E_t)=m_0\operatorname{Vol}({\Omega})/\kappa$ is imposed, we will deal with vector fields $V$ satisfying the linearized volume condition $\int_E {\nabla} \cdot V=0$. We thus call \textit{admissible at $E$} such vector fields and introduce
\begin{equation}\label{Eq:X}\mathcal X(E):=\left\{V\in W^{3,\infty}(\textrm{I\kern-0.21emR}^n;\textrm{I\kern-0.21emR}^n)\, , \int_E {\nabla} \cdot V=0\, , \Vert V\Vert _{W^{3,\infty}}\leq 1\right\}.
\end{equation}
A shape $E\subset {\Omega}$ with a $\mathscr C^2$ boundary such that $\kappa |E|=m_0|{\Omega}|$ is said to be critical if
\begin{equation}\label{Eq:FOO}
\forall V \in \mathcal X(E),\quad \lambda_\alpha'(E)[V]=0.
\end{equation}
or equivalently, if there exist a Lagrange multiplier $\Lambda_\alpha$ such that $ \left(\lambda_\alpha'-\Lambda_\alpha \operatorname{Vol}'\right)(E)[V]= 0$ for all $V \in W^{1,\infty}({\Omega})$, where $\operatorname{Vol}:{\Omega}\mapsto |{\Omega}|$ denotes the volume functional.
Furthermore, if $E$ is a local minimizer for Problem \eqref{Pb:OptimizationEigenvalue}, then one has
\begin{equation}\label{Eq:SOO}
\forall V \in \mathcal X(E), \quad \left(\lambda_\alpha''-\Lambda_\alpha \operatorname{Vol}''\right)(E)[V,V]\geq 0.
\end{equation}
In what follows, we will still assume that ${\Omega}$ denotes the ball $\mathbb B(0,R)$ with $R>0$.
\begin{theorem}\label{Th:ShapeStability}
Let us assume that $n=2$ and that ${\Omega}=\mathbb B(0,R)$. The ball $E=\mathbb B(0,r_0^*)=\mathbb B^*$ satisfies the shape optimality conditions \eqref{Eq:FOO}-\eqref{Eq:SOO}. Furthermore, if $\Lambda_\alpha$ is the Lagrange multiplier associated with the volume constraint, there exist two constants $\overline\alpha>0$ and $C>0$ such that, for any $\alpha\in [0,\overline \alpha)$ and any vector field $V\in \mathcal X({\mathbb B(0,R)}^*)$, there holds
$$\left(\lambda_\alpha''-\Lambda_\alpha \operatorname{Vol}''\right)({\mathbb B(0,R)}^*)[V,V]\geq C\Vert V\cdot \nu \Vert _{L^2(\mathbb S^*)}^2.$$
\end{theorem}
\begin{remark}The proof requires explicit computation of the shape derivative of the eigenfunction. We note that in \cite{DambrineKateb} such computations are carried out for the two-phase problem and that in \cite{KaoLouYanagida} such an approach is undertaken to investigate the stability of certain configurations for a weighted Neumann eigenvalue problem.
The main contribution of this result is to shed light on a monotonicity principle that enables one
to lead a careful asymptotic analysis of the second order shape derivative of the functional as $\alpha \to 0$. It is important to note that, although this allows us to deeply analyze the second order optimality conditions, it is expected that the optimal coercivity norm in the right-hand side above is expected to be $H^{\frac12}$ whenever $\alpha>0$, which we do not recover with our method. The reason why we believe that in this context the optimal coercivity norm is $H^{\frac12}$ is that in \cite{DambrineKateb}, precise computations for the two-phase problem \eqref{Pg:TwoPhase} are carried out and a $H^{\frac12}$ coercivity norm is obtained for certain classes of parameters. On the other hand, when $\alpha =0$, it was shown in \cite{MazariQuantitative} that the optimal coercivity norm is $L^2$, and a quantitative inequality was then derived.
\end{remark}
\begin{remark}
We believe that our strategy of proof may be used to obtain the same kind of coercivity norm in the three-dimensional case. However, we believe that such a generalization would be non-trivial and need technicalities. Since the main contribution is to introduce a methodology to study the positivity of second-order shape derivative, we simply provide a possible strategy to prove the result in the three dimensional case in the concluding section of the proof of Theorem~\ref{Th:ShapeStability}, see Section~\ref{Se:CclSha}.
\end{remark}
The rest of this article is dedicated to proofs of the results we have just outlined.
\subsection{A biological application of the problem}\label{BiologicalMotivations}
Equation \eqref{Eq:EigenFunction} arises naturally when dealing with simple population dynamics in heterogeneous spaces.
Let ${\varepsilon}\geq 0$ be a parameter of the model. We consider a population density whose flux is given by
$$
\mathcal J_{\varepsilon}=-{\nabla} u+{\varepsilon} u {\nabla} m.
$$
Since ${\nabla} m$ might not make sense if $m$ is assumed to be only measurable, we temporarily omit this difficulty by assuming it smooth enough so that the expression above makes sense. The term $u {\nabla} m$ appears as a drift term and stands for a bias in the population movement, modeling a tendency of the population to disperse along the gradient of resources and hence move to favorable regions. The parameter ${\varepsilon}$ quantifies the influence of the resources distribution on the movement of the species.
The complete associated reaction diffusion equation, called ``logistic diffusive equation'', reads
$$
\frac{\partial u}{\partial t}={\nabla} \cdot \Big({\nabla} u-{\varepsilon} u {\nabla} m\Big)+mu-u^2\quad \text{in }{\Omega},
$$
completed with suitable boundary conditions. In what follows, we will focus on {Dirichlet boundary conditions} meaning that the boundary of ${\Omega}$ is lethal for the population living inside.
Plugging the change of variable $v=e^{-{\varepsilon} m}u$ in this equation leads to
$$
\frac{\partial v}{\partial t}=\Delta v+{\varepsilon} {\nabla} m\cdot {\nabla} u +mv-e^{{\varepsilon} m}v\quad \text{in }{\Omega}.
$$
It is known (see e.g. \cite{BelgacemCosner,Belgacem,Murray}) that the asymptotic behavior of this equation is driven by
the principal eigenvalue of the operator $\tilde {\mathcal L}:u\mapsto -\Delta u-{\varepsilon} {\nabla} m\cdot {\nabla} u -mu$.
The associated principal eigenfunction $\psi$ satisfies
$$
-{\nabla} \cdot \left(e^{{\varepsilon} m}{\nabla} \psi\right)-me^{{\varepsilon} m}\psi=\tilde \lambda_{\varepsilon} \psi e^{{\varepsilon} m}\quad \text{in }{\Omega}.
$$
Following the approach developed in \cite{LamboleyLaurainNadinPrivat}, optimal configurations of resources correspond to the ones ensuring the fastest convergence to the steady-states of the PDE above, which comes to
minimizing $\tilde \lambda_{\varepsilon}(m)$ with respect to $m$.
By using Proposition \ref{Th:BangBang}, which enables us to only deal with {\it bang-bang} densities $m$, one shows easily that minimizing $\tilde \lambda_{\varepsilon}(m)$ over $\mathcal M_{m_0,\kappa}({\Omega})$ is equivalent to minimizing $\lambda_{\varepsilon}(m)$ over $\mathcal M_{m_0,\kappa}({\Omega})$, in other words to
Problem \eqref{Pb:OptimizationEigenvalue} with $\alpha={\varepsilon}$. Theorem \ref{Th:NonExistence} can thus be interpreted as follows in this framework: assuming that the population density moves along the gradient of the resources, it is not possible to lay the resources in an optimal way.
Note that the conclusion is completely different in the case $\alpha=0$ (see \cite{LamboleyLaurainNadinPrivat}) or in the one-dimensional case (i.e. ${\Omega}=(0,1)$) with $\alpha >0$ (see \cite{CaubetDeheuvelsPrivat}), where minimizers exist.
In the last case, optimal configurations for three kinds boundary conditions (Dirichlet, Neumann, Robin) have been obtained, by using a new rearrangement technique.
Finally, let us mention the related result \cite[Theorem 2.1]{Hamel2011}, dealing with Faber-Krahn type inequalities for general operators of the form
$$
\mathcal K:u\mapsto -{\nabla} \cdot(A{\nabla} u)- V\cdot {\nabla} u-mu
$$
where $A$ is a positive symmetric matrix.
Let us denote the first eigenvalue of $\mathcal K$ by $E(A,V,m)$. It is shown, by using new rearrangements, that there exist radially symmetric elements $A^*,V^*,m^*$ such that
$$
0< \inf A\leq A^*\leq \Vert A\Vert _\infty,\quad \Vert A^{-1}\Vert _{L^1}=\Vert (A^*)^{-1}\Vert _{L^1},\quad
\Vert V^*\Vert _{L^\infty}\leq \Vert V\Vert _{L^\infty}
$$
and $E(A,V,m)\geq E(A^*,V^*,m^*)$.
We note that applying this result directly to our problem would not allow us to conclude. Indeed, we would get that for every ${\Omega}$ of volume $V_1$ and every $m\in \mathcal M_{m_0,\kappa}({\Omega})$, if ${\Omega}^*$ is the ball of volume $V_1$, there exist two radially symmetric functions $m_1$ and $m_2$ satisfying $m_1$, $m_2$ in $\mathcal M_{m_0,\kappa}({\Omega})$ such that $\lambda_\alpha(m)\geq \mu_\alpha(m_1,m_2)$, where $\mu_\alpha(m_1,m_2)$ is the first eigenvalue of the operator $-{\nabla} \cdot((1+\alpha m_1){\nabla} )-m_2$.
We note that this result could also be obtained by using the symmetrization techniques of \cite{AlvinoTrombetti}.
Finally, let us mention optimal control problems involving a similar model but a different cost functional, related to:
\begin{itemize}
\item the total size of the population for a logistic diffusive equation in \cite{Mazari2020,MNPChapter,Mazari2021}.
\item optimal harvesting of a marine resource, investigated in the series of articles \cite{MR3035462,MR2476432,MR3628307}.
\end{itemize}
\subsection{Notations and notational conventions, technical properties of the eigenfunctions}\label{sec:notations}
Let us {sum up} the notations used throughout this article.
\begin{itemize}[label=\textbullet]
\item $\textrm{I\kern-0.21emR}_+$ is the set of non-negative real numbers. $\textrm{I\kern-0.21emR}_+^*$ is the set of positive real numbers.
\item $n$ is a fixed positive integer and ${\Omega}$ is a bounded connected domain in $\textrm{I\kern-0.21emR}^n$.
\item if $E$ denotes a subset of ${\Omega}$, the notation $\mathds{1}_E$ stands for the characteristic function of $E$, equal to 1 in $E$ and 0 elsewhere.
\item the notation $\Vert \cdot\Vert $ used without subscript refers to the standard Euclidean norm in $\textrm{I\kern-0.21emR}^n$. When referring to the norm of a Banach space $\mathcal X$, we write it $\Vert \cdot\Vert _{\mathcal X}$.
\item The average of every $f\in L^1({\Omega})$ is denoted by $\fint_{\Omega} f:=\frac1{|{\Omega}|}\int_{\Omega} f$.
\item $\nu$ stands for the outward unit normal vector on $\partial {\Omega}$.
\item if $m(\cdot)$ is a given function in $L^\infty({\Omega})$ and $\alpha$ a positive real number, we will use the notation $\sigma_{\alpha,m}$ to denote the function $1+\alpha m$. {When there is no ambiguity, we sometimes use the notation $\sigma_{\alpha}$ to alleviate notations.}
\item If $E$ denotes a subset of ${\Omega}$ with $\mathscr{C}^2$ boundary, we will use the notations
$$
f|_{int}(y)=\lim_{x\in E,x\to y}f(x)\quad \text{and}\quad f|_{ext}(y):=\lim_{x \in ({\Omega}\backslash E),x\to y}f(y)
$$
so that $\llbracket f\rrbracket =f|_{ext}-f|_{int}$ denotes the jump of $f$ at $x\in \partial E$.
\end{itemize}
\section{Preliminaries}\label{Pr:BangBang}
\def{\dot \lambda_\alpha(m)[h]}{{\dot \lambda_\alpha(m)[h]}}
\def{\ddot \lambda_\alpha(m)[h]}{{\ddot \lambda_\alpha(m)[h]}}
\def{\ddot u_{\alpha,m}[h]}{{\ddot u_{\alpha,m}[h]}}
\subsection{Switching function}
To derive optimality conditions for Problem~\eqref{Pb:OptimizationEigenvalue}, we introduce the tangent cone to $\mathcal{M}_{m_0,\kappa}({\Omega})$ at any point of this set.
\begin{definition} (\cite[chapter 7]{HenrotPierre})\label{def:tgtcone}
For every $m\in \mathcal M_{m_0,\kappa}({\Omega})$, the tangent cone to the set $\mathcal{M}_{m_0,\kappa}({\Omega})$ at $m$, also called the \textit{admissible cone} to the set $\mathcal{M}_{m_0,\kappa}({\Omega})$ at $m$, denoted by $\mathcal{T}_{m}$ is the set of functions $h\in L^\infty({\Omega})$ such that, for any sequence of positive real numbers $\varepsilon_n$ decreasing to $0$, there exists a sequence of functions $h_n\in L^\infty({\Omega})$ converging to $h$ as $n\rightarrow +\infty$, and $m+\varepsilon_nh_n\in\mathcal{M}_{m_0,\kappa}({\Omega})$ for every $n\in\textrm{I\kern-0.21emN}$.\label{footnote:cone}
\end{definition}
Notice that, as a consequence of this definition, any $h \in \mathcal T_m$ satisfies $\fint_{\Omega} h=0$.
\begin{lemma}\label{diff:valPvecP}
Let $m\in \mathcal M_{m_0,\kappa}({\Omega})$ and $h\in \mathcal T_m$. The mapping $ \mathcal M_{m_0,\kappa}({\Omega})\ni m\mapsto u_{\alpha,m}$ is twice differentiable at $m$ in direction $h$ in a strong $L^2({\Omega})$ sense and in a weak $W^{1,2}_0({\Omega})$ sense, and the mapping $ \mathcal M_{m_0,\kappa}({\Omega})\ni m\mapsto \lambda_\alpha$ is twice differentiable in a strong $L^2({\Omega})$ sense.
\end{lemma}
The proof of this lemma is technical and is postponed to Appendix \ref{Ap:Differentiability}.
For $t$ small enough, let us introduce the mapping $g_h:t\mapsto \lambda_\alpha \left([m+th]\right)$. Hence, $g_h$ is twice differentiable.
The first and second order derivatives of $\lambda_\alpha$ at $m$ in direction $h$, denoted by $\dot\lambda_\alpha(m)[h]$ and $\ddot\lambda_\alpha(m)[h]$, are defined by
$$
\dot \lambda_\alpha(m)[h]:=g_h'(0)\quad\text{and}\quad \ddot\lambda_\alpha(m)[h]:=g_h''(0).
$$
\begin{lemma}\label{Le:DeriveeL21}
Let $m\in \mathcal{M}_{m_0,\kappa}({\Omega})$ and $h\in \mathcal{T}_m$. The mapping $m\mapsto \lambda_\alpha(m)$ is differentiable at $m$ in direction $h$ in $L^2$ and its differential reads
\begin{equation}\label{Eq:Derivee21}
\dot \lambda_\alpha(m)[h]=\int_{\Omega} h\psi_{\alpha,m},\quad \text{with}\quad \psi_{\alpha,m}:=\alpha |{\nabla} u_{\alpha,m}|^2-u_{\alpha,m}^2.
\end{equation}
The function $\psi_{\alpha,m}$ is called {\it switching function}.
\end{lemma}
\begin{proof}
According to Lemma~\ref{diff:valPvecP}, we can differentiate the variational formulation associated to \eqref{Eq:EigenFunction} and get that the differential ${\dot u_{\alpha,m}[h]}$ of $m\mapsto u_{\alpha,m}$ at $m$ in direction $h$ satisfies, with $\sigma_\alpha :=1+\alpha m$,
\begin{equation}\label{Eq:L2Derivative1}\left\{\begin{array}{lll}
-{\nabla} \cdot\Big(\sigma_\alpha{\nabla} {\dot u_{\alpha,m}[h]}\Big)-\alpha {\nabla} \cdot\Big(h{\nabla} {u_{\alpha,m}}\Big)=&{\dot \lambda_\alpha(m)[h]} {u_{\alpha,m}} +\lambda_\alpha(m){\dot u_{\alpha,m}[h]} & \\
& +m{\dot u_{\alpha,m}[h]}+h{u_{\alpha,m}} & \text{ in }{\Omega},\\
{\dot u_{\alpha,m}[h]}=0 {\text{ on } \partial {\Omega}},\\
\int_{\Omega} {u_{\alpha,m}} {\dot u_{\alpha,m}[h]}=0.&
\end{array}\right.
\end{equation}
Multiplying {(\ref{Eq:L2Derivative1})} by ${u_{\alpha,m}}$, integrating by parts and using that ${u_{\alpha,m}}$ is normalized in $L^2({\Omega})$ leads to
\begin{align*}
{\dot \lambda_\alpha(m)[h]} =&\underbrace{\int_{\Omega} \sigma_\alpha {\nabla} {\dot u_{\alpha,m}[h]}\cdot {\nabla} {u_{\alpha,m}} -\lambda_{\alpha}(m)\int_{\Omega} {u_{\alpha,m}} {\dot u_{\alpha,m}[h]}-\int_{\Omega} m{u_{\alpha,m}} {\dot u_{\alpha,m}[h]} }_{=0\text{ according to \eqref{Eq:EigenFunction}}}\\
&+\int_{\Omega} \alpha h |{\nabla} {u_{\alpha,m}}|^2-\int_{\Omega} h u_{\alpha,m}^2.
\end{align*}
\end{proof}
\subsection{Proof of Proposition~\ref{Th:BangBang}}
The proof relies on concavity properties of the functional $\lambda_\alpha$. More precisely, let $m_1,m_2\in \mathcal M_{m_0,\kappa}({\Omega})$. We will show that the map $f:[0,1]\ni t\mapsto \lambda_\alpha \left((1-t)m_1+tm_2\right)$ is strictly concave, i.e that $ f''< 0$ on $[0,1]$.
Note that the characterization of the concavity in terms of second order derivatives makes sense, according to Lemma~\ref{diff:valPvecP}, since $\lambda_\alpha$ is twice differentiable. Before showing this concavity property, let us first explain why it implies the conclusion of Proposition~\ref{Th:BangBang} (the weak {\it bang-bang} property).
{Suppose that $m\in \mathcal M_{m_0,\kappa}({\Omega})$ is not { bang-bang}}. The set $\mathcal{I}=\{0<m<\kappa\}$ is then of positive Lebesgue measure and $m$ is therefore not extremal in $ \mathcal M_{m_0,\kappa}({\Omega})$, according to \cite[Prop.~7.2.14]{HenrotPierre}.
We then infer the existence of $t \in (0,1)$ as well as two distinct elements $m_1$ and $m_2$ of $ \mathcal M_{m_0,\kappa}({\Omega})$ such that $m=(1-t) m_1+tm_2$.
Because of the strict concavity of $\lambda_\alpha$, the solution of the optimization problem $\min \{\lambda_\alpha((1-t) m_1+tm_2)\}$ is either $m_1$ or $m_2$, and moreover, $m$ cannot solve this problem.
Assume that $m_1$ solves this problem without loss of generality. One thus has $\lambda_\alpha(m_1)<\lambda_\alpha (m)$. Since the subset of {\it bang-bang} functions of $\mathcal M_{m_0,\kappa}({\Omega})$ is dense in $\mathcal M_{m_0,\kappa}({\Omega})$ for the weak-star topology of $L^\infty({\Omega})$, there exists a sequence of {\it bang-bang} functions $(m^k)_{k\in \textrm{I\kern-0.21emN}}$ of $\mathcal M_{m_0,\kappa}({\Omega})$ converging weakly-star to $m_1$ in $L^\infty({\Omega})$. Furthermore, $\lambda_\alpha$ is upper semicontinuous for the weak-star topology of $L^\infty({\Omega})$, since it reads as the infimum of continuous linear functionals for this topology. Let $\varepsilon>0$. We infer the existence of $k_\varepsilon\in \textrm{I\kern-0.21emN}$ such that $\lambda_\alpha(m^{k_\varepsilon})\leq \lambda_\alpha(m_1)+\varepsilon$. By choosing $\varepsilon$ small enough, we get that $\lambda_\alpha(m^{k_\varepsilon})<\lambda_\alpha (m)$, whence the result.
\medskip
It now remains to prove that $f$ is strictly concave. Let $m\in \mathcal M_{m_0,\kappa}({\Omega})$, and set $m_1=m$, $h=m_2-m_1$, we observe that $f''(t)=\ddot\lambda_{\alpha}((1-t)m_1+tm_2)[h]$ for all $t\in [0,1]$.
The differential ${\dot u_{\alpha,m}[h]}$ of $m\mapsto u_{\alpha,m}$ at $m$ in direction $h$, denoted ${\dot u_{\alpha,m}[h]}$, satisfies \eqref{Eq:L2Derivative1} and the second order {Gateaux derivatives} ${\ddot u_{\alpha,m}[h]}$ and ${\ddot \lambda_\alpha(m)[h]}$ solve, with $\sigma_\alpha:=1+\alpha m$,
\begin{equation}\label{Eq:L2Derivative2}
\left\{\begin{array}{lll}
-{\nabla} \cdot\Big(\sigma_\alpha{\nabla} {\ddot u_{\alpha,m}[h]}\Big)-2\alpha {\nabla} \cdot\Big(h{\nabla} {\dot u_{\alpha,m}[h]}\Big)=&{\ddot \lambda_\alpha(m)[h]} {u_{\alpha,m}} +2{\dot \lambda_\alpha(m)[h]} {\dot u_{\alpha,m}[h]} \\
&+\lambda_\alpha(m){\ddot u_{\alpha,m}[h]}+m{\ddot u_{\alpha,m}[h]}+2h{\dot u_{\alpha,m}[h]} & \text{ in }{\Omega},\\
{\ddot u_{\alpha,m}[h]}=0 \quad {\text{ on } \partial {\Omega}.}&&
\end{array}\right.\end{equation}
Multiplying {(\ref{Eq:L2Derivative2})} by ${u_{\alpha,m}}$, using that ${u_{\alpha,m}}$ is normalized in $L^2({\Omega})$ and integrating by parts yields
\begin{align*}
{\ddot \lambda_\alpha(m)[h]}&=\underbrace{\int_{\Omega} \sigma_\alpha {\nabla} {\ddot u_{\alpha,m}[h]}\cdot {\nabla} {u_{\alpha,m}}-\lambda_{\alpha}(m)\int_{\Omega} {u_{\alpha,m}} {\ddot u_{\alpha,m}[h]}-\int_{\Omega} m{u_{\alpha,m}} {\ddot u_{\alpha,m}[h]} }_{=0\text{ according to \eqref{Eq:EigenFunction}}}
\\&+2\alpha \int_{\Omega} h {\nabla} {\dot u_{\alpha,m}[h]},{\nabla} {u_{\alpha,m}}-2\int_{\Omega} h{u_{\alpha,m}}{\dot u_{\alpha,m}[h]}
\\&=2\left(-\int_{\Omega} \sigma_\alpha |{\nabla} {\dot u_{\alpha,m}[h]}|^2+\int_{\Omega} m{\dot u_{\alpha,m}[h]}^2+\lambda_\alpha(m)\int_{\Omega} {\dot u_{\alpha,m}[h]}^2\right)+2\underbrace{{\dot \lambda_\alpha(m)[h]} \int_{\Omega} u_{\alpha,m} {\dot u_{\alpha,m}[h]}}_{=0\text{ since $\int_{\Omega} {u_{\alpha,m}} {\dot u_{\alpha,m}[h]}=0$}}
\\&=2\int_{\Omega} {\dot u_{\alpha,m}[h]}^2\left(-R_{\alpha,m}[{\dot u_{\alpha,m}[h]}]+\lambda_\alpha(m)\right)< 0,
\end{align*}
where the last inequality comes from the observation that, whenever $h\neq 0$, one has ${\dot u_{\alpha,m}[h]}\neq 0$ and ${\dot u_{\alpha,m}[h]}$ is in the orthogonal space to the first eigenfunction ${u_{\alpha,m}}$ in $L^2({\Omega})$. Since the first eigenvalue is simple, the Rayleigh quotient of ${\dot u_{\alpha,m}[h]}$ is greater than $\lambda_{\alpha}(m)$.
\section{Proof of Theorem \ref{Th:NonExistence}}\label{Pr:NonExistence}
This proof is based on a homogenization argument, inspired from the notions and techniques introduced in \cite{MuratTartar}. In the next section, we gather the preliminary tools and material involved in what follows.\subsection{Background material on homogenization and bibliographical comments}\label{SSE:BackgroundH}
Let us recall several usual definitions and results in homogenization theory we will need hereafter.
\begin{definition}[$H$-convergence]
Let $(m_k)_{k\in \textrm{I\kern-0.21emN}}\in \mathcal M_{m_0,\kappa}({\Omega})^\textrm{I\kern-0.21emN}$ and for every $k\in \textrm{I\kern-0.21emN}$, define respectively $\sigma_k$ and $u_k(f)$ by $\sigma_k=1+\alpha m_k$ and as the unique solution of
$$
\left\{\begin{array}{ll}-{\nabla} \cdot(\sigma_k {\nabla} u_k(f))=f&\text{ in }{\Omega}\\
u_k(f)=0 \text{ on }\partial {\Omega}
\end{array}
\right.
$$
where $f\in L^2({\Omega})$ is given. We say that the sequence $(\sigma_k)_{k\in \textrm{I\kern-0.21emN}}$ $H$-converges to $A:{\Omega}\to M_{n}(\textrm{I\kern-0.21emR})$ if, for every $f\in L^2({\Omega})$, the sequence $(u_k(f))_{k\in \textrm{I\kern-0.21emN}}$ converges weakly to $u_\infty$ in $\wo$ and the sequence $(\sigma_k{\nabla} u_k)_{k\in \textrm{I\kern-0.21emN}}$ converges weakly to $A{\nabla} u_\infty $ in $ L^2({\Omega})$, where $u_\infty$ solves
$$\left\{\begin{array}{ll}-{\nabla} \cdot(A{\nabla} u_\infty)=f&\text{ in }{\Omega}\, ,
\\u_\infty=0 \text{ on }\partial {\Omega}\end{array}
\right.$$
In that case, we will write $\sigma_k \overset{H}{\underset{k\to \infty}\longrightarrow} A$.
\end{definition}
\begin{definition}[arithmetic and geometric means]\label{De:Homo}
Let $m\in \mathcal M_{m_0,\kappa}({\Omega})$ and $\sigma=1+\alpha m$. We define the arithmetic mean of $\sigma$ by $\Lambda_+(m)=\sigma$, and its harmonic mean by $\Lambda_-(m)=\frac{1+\alpha \kappa}{1+\alpha(\kappa-m)}$.
One has $\Lambda_-(m)\leq \Lambda_+(m)$, according to the arithmetic-harmonic inequality, with equality if and only if
$m$ is a {\it bang-bang} function.
\end{definition}
\begin{prnonumbering}\cite[Proposition 10]{MuratTartar}
Let $(m_k)_{k\in \textrm{I\kern-0.21emN}}\in \mathcal M_{m_0,\kappa}({\Omega})^\textrm{I\kern-0.21emN}$ and $(\sigma_k)_{k\in \textrm{I\kern-0.21emN}}$ given by $\sigma_k=1+\alpha m_k$. Up to a subsequence, there exists $m\in \mathcal M_{m_0,\kappa}({\Omega})$ such that $(m_k)_{k\in \textrm{I\kern-0.21emN}}\in \mathcal M_{m_0,\kappa}({\Omega})^\textrm{I\kern-0.21emN}$ converges to $m$ for the weak-star topology of $L^\infty$.
Assume moreover that the sequence $(\sigma_k)_{k\in \textrm{I\kern-0.21emN}}$ $H$-converges to a matrix $A$. Then, $A$ is a symmetric matrix, its spectrum $\Sigma(A)=\{\lambda_1,\dots,\lambda_n\}$ is real, and
\begin{equation} \label{Spec}\tag{$J_1$} \Lambda_-(m)\leq \min \Sigma(A)\leq \max \Sigma(A)\leq \Lambda_+(m).\end{equation}
\begin{equation}\label{Spec1}\tag{$J_2$}\sum_{j=1}^n \frac1{\lambda_j-1}\leq \frac1{\Lambda_-(m)-1}+\frac{n-1}{\Lambda_+(m)-1},\end{equation}
\begin{equation} \label{Spec2}\tag{$J_3$}\sum_{j=1}^n \frac1{1+\alpha\kappa-\lambda_j}\leq \frac1{1+\alpha \kappa-\Lambda_-(m)}+\frac{n-1}{1+\alpha\kappa -\Lambda_+(m)}.\end{equation}
\end{prnonumbering}
For a given $m\in \mathcal M_{m_0,\kappa}({\Omega})$, we introduce
$$
M_m^\alpha=\{A:{\Omega}\rightarrow S_n(\textrm{I\kern-0.21emR})\, , A \text{ satisfies \eqref{Spec}-\eqref{Spec1}-\eqref{Spec2}}\}.
$$
For a matrix-valued application $A\in M_m^\alpha$ for some $m\in \mathcal M_{m_0,\kappa}({\Omega})$, it is possible to define the principal eigenvalue of $A$ via Rayleigh quotients as
\begin{equation}\label{Eq:Zeta}
\zeta_\alpha(m,A):=\inf_{u\in \wo\, , \int_{\Omega} u^2=1}\int_{\Omega} A{\nabla} u\cdot {\nabla} u-\int_{\Omega} mu^2.
\end{equation}
Note that the dependence of $\zeta_\alpha$ on the parameter $\alpha$ is implicitly contained in the condition $A\in M_m^\alpha$. We henceforth focus on the following relaxed version of the optimization problem:
\begin{equation}\label{Pb:OptimizationRelax}
\inf_{m\in \mathcal M_{m_0,\kappa}({\Omega})\, , A\in M_m^\alpha} \zeta_\alpha(m,A).
\end{equation}
for which we have the following result.
\begin{thmnonumbering}\cite[Proposition 10]{MuratTartar}
\begin{itemize}
\item[(i)] For every $m\in \mathcal M_{m_0,\kappa}({\Omega})$ and $A \in M_m^\alpha$, there exists a sequence $(m_k)_{k\in \textrm{I\kern-0.21emN}}\in \mathcal M_{m_0,\kappa}({\Omega})$ such that $(m_k)_{k\in \textrm{I\kern-0.21emN}}$ converges to $m$ for the weak-star topology of $L^\infty$, and the sequence $(\sigma_k)_{k\in \textrm{I\kern-0.21emN}}$ defined by $\sigma_k=1+\alpha m_k$ $H$-converges to $A$, as $k\to +\infty$.
\item[(ii)] The mapping $(m,A)\mapsto \lambda_\alpha(m,A)$ is continuous with respect to the $H$-convergence (see in particular \cite{Oleinik}).
\item[(iii)] The variational problem \eqref{Pb:OptimizationRelax}
has a solution $(\hat{m},\hat{A})$; by definition, $\hat{A}\in M_{\hat{m}}^\alpha$. Furthermore, if $\hat{u}$ is the associated eigenfunction, then $\hat{A} {\nabla} \hat{u} =\Lambda_-(\hat{m}){\nabla} \hat{u}$.
\end{itemize}
\end{thmnonumbering}
This theorem allows us to solve Problem~\eqref{Pb:OptimizationRelax}.
\begin{corollary}\cite{MuratTartar}\label{Co:Equivalence}
If Problem~\eqref{Pb:OptimizationEigenvalue} has a solution $\hat{m}$, then the couple $(\hat{m},1+\alpha \hat{m})$ solves Problem~\eqref{Pb:OptimizationRelax}.
\end{corollary}
\begin{proof}[Proof of Corollary \ref{Co:Equivalence}]
Assume that the solution of \eqref{Pb:OptimizationRelax} is $(\hat{m}, \hat{A})$ and that $\hat{A}\neq 1+\alpha \hat{m}$. Then there exists a sequence $(m_k)_{k\in \textrm{I\kern-0.21emN}}$ converging
weak-star in $L^\infty$ to $\hat{m}$ and such that the sequence $(1+\alpha m_k)_{k\in \textrm{I\kern-0.21emN}}$ $H$-converges to $\hat{A}$. This means that
$$\lambda_\alpha(\hat{m})=\zeta_\alpha(\hat{m},1+\alpha \hat{m})>\zeta_\alpha(\hat{m},\hat{A})=\underset{k\to \infty}\lim \lambda_\alpha(m_k)$$ which immediately yields a contradiction.
\end{proof}
Let us end this section with several bibliographical comments on such problems.
\paragraph{Bibliographical comments on the two-phase conductors problem.}\label{TwoPhase}
Problem \eqref{Pg:TwoPhase} {with $\mathcal M:=\mathcal M_{m_0,\kappa}(\Omega)$} has drawn a lot of attention in the last decades, since the seminal works by Murat and Tartar, \cite{Murat,MuratTartar}
Roughly speaking, this optimal design problem is, in general, ill-posed and one needs to introduce a relaxed formulation to get existence.
We refer to \cite{Allaire,CoxLipton,MuratTartar,Oleinik}.
Let us provide the main lines strategy to investigate existence issues for Problem \eqref{Pg:TwoPhase}, according to \cite{Murat,MuratTartar}.
If the solution $(\hat{m},1+\alpha\hat{m})$ to the relaxed problem \eqref{Pb:OptimizationRelax} is a solution to the original problem \eqref{Pg:TwoPhase}, then there exists a measurable subset $\hat{E}$ of ${\Omega}$ such that $\hat{m}=\kappa \mathds{1}_{\hat{E}}$. If furthermore $\hat{E}$ is assumed to be smooth enough, then, denoting by $\hat{u}$ the principal eigenfunction associated with $(\hat{m},\lambda_\alpha(m))=(\hat{m},\zeta_\alpha(\hat{m}, 1+\alpha \hat{m}))$, we get that $\hat{u}$ and $(1+\alpha \hat{m}) \frac{\partial \hat{u}}{\partial \nu}$ must be constant on $\partial \hat{E}$.
The function $1+\alpha \hat{m}$ being discontinuous across $\partial E$, the optimality condition above has to be understood in the following sense: the function $(1+\alpha \hat{m}) \frac{\partial \hat{u}}{\partial \nu}$, a priori discontinuous, is in fact continuous across $\partial E$ and even constant on it.
Note that these arguments have been generalized in \cite{CoxLipton}. These optimality conditions, combined with Serrin's Theorem \cite{Serrin1971}, suggest that Problem \eqref{Pg:TwoPhase} could have a solution if, and only if ${\Omega}$ is a ball. The best results known to date are the following ones.
\begin{thmnonumbering}
\begin{itemize}
\item[(i)] Let ${\Omega}$ be an open set such that $\partial {\Omega}$ is $\mathscr C^2$ and connected. Problem \eqref{Pg:TwoPhase} has a solution if and only if ${\Omega}$ is a ball \cite{CasadoDiaz3}.
\item[(ii)] If ${\Omega}$ is a ball, then Problem \eqref{Pg:TwoPhase} has a solution which is moreover radially symmetric \cite{ConcaMahadevanSanz}.
\end{itemize}
\end{thmnonumbering}
Regarding the second part of the theorem, the authors used a particular rearrangement coming to replace $1+\alpha m$ by its harmonic mean on each level-set of the eigenfunction. Such a rearrangement has been first introduced by Alvino and Trombetti \cite{AlvinoTrombetti}. This drives the author to reduce the class of admissible functions to radially symmetric ones, which allow them to conclude thanks to a compactness argument \cite{AlvinoTrombettiLions}. These arguments are mimicked to derive the existence part of Theorem~\ref{Th:RadialStability}.
Finally, let us mention \cite{ConcaLaurainMahadevan,Laurain}, where the optimality of annular configurations in the ball is investigated.
A complete picture of the situation is then depicted in the case where $\alpha$ is small, which is often referred to as the "low contrast regime".
We also mention \cite{DambrineKateb} , where a shape derivative approach is undertaken to characterize minimizers when ${\Omega}$ is a ball.
\subsection{Proof of Theorem \ref{Th:NonExistence}}
Let us assume the existence of a solution to Problem \eqref{Pb:OptimizationEigenvalue}, denoted $\hat{m}$. According to Proposition~\ref{Th:BangBang}, there exists a measurable subset $\hat{E}$ of ${\Omega}$ such that
$\hat{m}=\kappa \mathds{1}_{\hat{E}}$. Let us introduce $\hat{\sigma}:=1+\alpha \hat{m}$ and $\hat{u}$, the $L^2$-normalized eigenfunction associated to $\hat{m}$.
Let us now assume that $\partial \hat{E}$ is $\mathscr C^2$.
\paragraph{Step 1: derivation of optimality conditions.}
What follows is an adaptation of \cite{CoxLipton}. For this reason, we only recall the main lines.
Let us write the optimality condition
for the problem
$$\min_{m\in \mathcal M_{m_0,\kappa}({\Omega})}\min_{A\in M_m^\alpha}\zeta_\alpha(m,A)=\lambda_\alpha(\hat{m}),$$
where $\zeta_\alpha$ is given by \eqref{Eq:Zeta}. Let $h$ be an admissible perturbation at $\hat{m}$. In \cite{MuratTartar} it is is proved that for every ${\varepsilon}>0$ small enough, there exists a matrix-valued application $A_{\varepsilon} \in M_{\hat{m}+{\varepsilon} h}$ such that
$$
A_{\varepsilon}{\nabla} \hat{u}=\Lambda_-(\hat{m}+{\varepsilon} h){\nabla} \hat{u}\quad \text{in }{\Omega},
$$
where $\Lambda_-$ has been introduced in Definition \ref{De:Homo}.
Fix ${\varepsilon}$ as above. Since $(\hat{m},1+\alpha \hat{m})$ is a solution of the Problem \eqref{Pb:OptimizationRelax}, one has
\begin{align*}
\int_{\Omega} A_{\varepsilon} {\nabla} \hat{u}\cdot {\nabla} \hat{u}-\int_{\Omega} (m+{\varepsilon} h)\hat{u}^2&\geq \zeta_\alpha(m+{\varepsilon} h,A_{\varepsilon}) \\
&\geq \zeta_\alpha(\hat{m},1+\alpha \hat{m})=\int_{\Omega} \hat{\sigma} |{\nabla} \hat{u}|^2-\int_{\Omega} \hat{m}\, \hat{u}^2.
\end{align*}
where one used the Rayleigh quotient definition of $\zeta_\alpha$ as well as the minimality of $(\hat{m},1+\alpha \hat{m})$.
Dividing the last inequality by ${\varepsilon}$ and passing to the limit yields
$$\int_{\Omega} h\left.\frac{d\Lambda_-(m)}{dm}\right|_{m=\hat{m}}|{\nabla} \hat{u}|^2-h\hat{u}^2\geq 0.$$
Using that $d\Lambda_-/dm=\alpha \Lambda_-(m)^2/(1+\alpha\kappa)$, and that $\hat{m}$ is a {\it bang-bang} function (so that $\Lambda_-(\hat{m})=\hat{\sigma}$),
we infer that the first order optimality conditions read: there exists $\mu\in \textrm{I\kern-0.21emR}$ such that
\begin{equation}\label{Eq:OptimalityH}
\left\{\Psi_\alpha< \mu\right\}\subset \hat{E}\subset \left\{\Psi_\alpha\leq \mu\right\}\quad \text{where}\quad \Psi_\alpha:=\frac{\alpha}{1+\alpha\kappa}{\hat{\sigma}}^2|{\nabla} \hat{u}|^2-{\hat{u}}^2.
\end{equation}
Since the flux $\hat{\sigma} \frac{\partial \hat{u}}{\partial \nu}$ is continuous across $\partial \hat{E}$, one has necessarily $\Psi_\alpha=\mu$ on $\partial \hat{E}\backslash \partial \Omega$.
Now, let us follow the approach used in \cite{MuratTartar} and \cite{CasadoDiaz3} to simplify the writing of the optimality conditions. Notice first that $\hat{u}$ and $\hat{\sigma}^2 \left|\frac{\partial \hat{u}}{\partial \nu_{\hat{E}}}\right|^2$ are continuous across $\partial \hat{E}$. Let ${\nabla}_\tau \hat{u}$ denote the tangential gradient of $\hat{u}$ on $\partial \hat{E}$.
For the sake of clarity, the quantities computed on $\partial \hat{E}$ seen as the boundary of $\hat{E}$ will be denoted with the subscript $int$, whereas the ones computed on $\partial \hat{E}$ seen as part of the boundary of $\hat{E}^c$ will be denoted with the subscript $ext$. According to the optimality conditions \eqref{Eq:OptimalityH}, one has
\begin{align*}
\begin{split}
\left.\frac\alpha{1+\alpha\kappa} \hat{\sigma}^2 \left|{\nabla}_\tau \hat{u}\right|^2+\frac\alpha{1+\alpha\kappa} \hat{\sigma}^2\left(\frac{\partial \hat{u}}{\partial \nu}\right)^2-\hat{u}^2\right|_{int}
\leq \left.\frac\alpha{1+\alpha\kappa} \hat{\sigma}^2 \left|{\nabla}_\tau \hat{u}\right|^2+\frac\alpha{1+\alpha\kappa} \hat{\sigma}^2\left(\frac{\partial \hat{u}}{\partial \nu}\right)^2-\hat{u}^2\right|_{ext}
\end{split}\end{align*}
on $\partial\hat{E}\backslash \partial \Omega$. By continuity of the flux $\hat{\sigma} \frac{\partial \hat{u}}{\partial \nu}$, we infer that
$\alpha \hat{\sigma}^2 \left.\left|{\nabla}_\tau \hat{u}\right|^2\right|_{int}\leq \alpha \hat{\sigma}^2 \left.\left|{\nabla}_\tau \hat{u}\right|^2\right|_{ext}$ which comes to $(1+\alpha \kappa) \left.\left|{\nabla}_\tau \hat{u}\right|^2\right|_{int}\leq \left.\left|{\nabla}_\tau \hat{u}\right|^2\right|_{ext}$. Since $ \left.\left|{\nabla}_\tau \hat{u}\right|^2\right|_{int}= \left.\left|{\nabla}_\tau \hat{u}\right|^2\right|_{ext}$, we have $\left.{\nabla}_\tau \hat{u}\right|_{\partial \hat{E}}=0$. Therefore, $\hat{u}$ is constant on $\partial \hat{E}\backslash \partial \Omega$ and since $\Psi_\alpha$ is constant on $\partial \hat{E}\backslash \partial \Omega$, it follows that $\left|\frac{\partial \hat{u}}{\partial \nu}\right|^2_{int}$ is constant as well on $\partial \hat{E}\backslash \partial \Omega$.
To {sum up}, the first order necessary conditions drive to the following condition:
\begin{equation}\label{Eq:Opt}
\text{The functions }\hat{u}\text{ and } |{\nabla} \hat{u}|\text{ are constant on $\partial \hat{E}\backslash \partial \Omega$.}
\end{equation}
\paragraph{Step 2: proof that ${\Omega}$ is necessarily a ball.}
To prove that ${\Omega}$ is a ball, we will use Serrin's Theorem, that we recall hereafter.
\begin{thmnonumbering}\cite[Theorem 2]{Serrin1971}
Let $\mathscr E$ be a connected domain with a $\mathscr C^2$ boundary, $h$ a $\mathscr C^1(\textrm{I\kern-0.21emR};\textrm{I\kern-0.21emR})$ function and let $f\in \mathscr C^2\left(\overline{\mathscr E}\right)$ be a function satisfying
$$
-\Delta f= h(f)\,, \quad f>0 \text{ in }\mathscr E,\quad f=0\text{ on }\partial \mathscr E,\quad \frac{\partial f}{\partial \nu}\text{ is constant on }\partial \mathscr E.$$
Then $\mathscr E$ is a ball and $f$ is radially symmetric.
\end{thmnonumbering}
According to \eqref{Eq:Opt}, let us introduce $\hat{\mu}=\left.\hat{u}\right|_{\partial \hat{E}}$. One has $\hat{\mu}>0$ by using the maximum principle. Let us set $f=\hat{u}-\hat{\mu}$, $h(z)=\left(\lambda_\alpha(\hat{m})+\kappa\right)z$ and call $\mathscr E$ a given connected component of $\hat{E}$. By assumption, $\hat{E}$ is a $\mathscr C^2$ set, and, according to \eqref{Eq:Opt}, the function $\partial \left(\hat{u}-\hat{\mu}\right)/\partial \nu$ is constant on $\partial \hat{E}$.
The next result allows us to verify the last assumption of Serrin's theorem.
\begin{lemma}\label{Eq:Opt2}
There holds $\hat{u}>\hat{\mu}$ in $\hat{E}$.
\end{lemma}
For the sake of clarity, the proof of this lemma is postponed to the end of this section.\\
Let us now come back to the proof that ${\Omega}$ is necessarily a ball. Take $x\in\partial {\Omega}$. Then $x$ belongs either to the closure of $\Omega \backslash \hat E$ or to the closure of $\hat{E}$. In the first case, there exists a connected component $\Gamma$ of $\partial \big({\Omega}\backslash \hat{E}\big)$ which contains $x$.
Let us assume by contradiction that this connected component also intersects ${\Omega}$, then according to \eqref{Eq:Opt}, one has $\hat{u}$ constant on $\Gamma \cap {\Omega}$ and according to the Dirichlet boundary conditions, one has $\hat{u}=0$ on $\Gamma$, and hence $\hat{u}$ reaches its minimal value in the open set ${\Omega}$. According to the strong maximum principle, one gets that $\hat{u}(\cdot)=0$, and we have reached a contradiction.
Hence, $\Gamma \subset \partial{\Omega}$ and $\Gamma$ is connected. Similarly, if $x$ belongs to the closure of $\hat{E}$, then there exists a connected component of $\partial\hat{E}$ containing $x$, which is included in $\partial{\Omega}$. Hence, $\partial {\Omega}$ is the union of closed connected components of the boundaries of $\hat{E}$ and ${\Omega}\backslash \hat{E}$. As $\partial{\Omega}$ is connected by hypothesis, there only exists one such connected component, that we denote $\Gamma$. This implies in particular that if $\partial \hat E\cap \partial {\Omega}\neq \emptyset$, then $\partial ({\Omega}\backslash \hat E)\cap\partial {\Omega}=\emptyset$, and conversely, if $\partial\hat{E}^c \cap\partial {\Omega}\neq \emptyset$, then $\partial \hat E\cap \partial {\Omega}=\emptyset$.
{Assume first that the closure of ${\Omega}\backslash \hat{E}$ meets $\partial {\Omega}$. As $\hat{E}$ does not intersect $\partial {\Omega}$ in this case, one has $\hat{u}$ and $|\nabla \hat{u}|$ constant over the whole boundary of $\hat{E}$ by (\ref{Eq:Opt}) and thus Serrin's theorem applies: any connected component of $\hat{E}$ is a ball and $\hat{u}$ is radially symmetric over it. We can hence fix a ball $\mathbb B(x_0;a_1)\subset \hat E$ such that $\hat u$ is radially symmetric in it.}
Let us now consider the largest ball $\mathbb{B}(x_{0};a)\subset {\Omega}$ such that $\hat u$ is radially symmetric in $\mathbb B(x_0;a)$. If $\partial \mathbb B(x_0;a)\cap \partial {\Omega}\neq \emptyset$, then $\hat u=0$ on $\partial \mathbb B(x_0;a)$ which, by the maximum principle, implies that ${\Omega}=\mathbb B(x_0;a)$. We can hence assume that $\hat u>0$ on $\partial \mathbb B(x_0;a)$. Let us define
$$\mu_a:=\left.\frac{\alpha}{1+\alpha\kappa}{\hat{\sigma}}^2|{\nabla} \hat{u}|^2-{\hat{u}}^2\right|_{\partial \mathbb B(x_0;a)}.$$ If $\mu_a<\mu$ then by continuity of $\frac{\alpha}{1+\alpha\kappa}{\hat{\sigma}}^2|{\nabla} \hat{u}|^2-{\hat{u}}^2$, it follows that $m=0$ on $\mathbb B(x_0;a+\delta)\backslash \mathbb B(x_0;a-\delta)$. Thus, applying the Cauchy-Kovalevskaya theorem to all the tangential derivatives of $\hat{u}$ yields that $\hat{u}$ is radially symmetric in $\mathbb B(x_0;a+\delta)$ which is a contradiction with the definition of $a$. Indeed, the same arguments as \cite[Proof of Theorem~1, Part~2]{LamboleyLaurainNadinPrivat} would yield that if one writes $\hat{u}(x)=U(r)$, with $r:=|x-x_{0}|$, then there exists $\delta>0$ such that $U'(r)=0$ for all $r\in [a,a+\delta)$. Hence $u$ would remain constant on the annulus $\{a<|x-x_{0}<a+\delta\}$, yielding a contradiction.
The same reasoning yields the same contradiction if $\mu_a>\mu$, in which case $m=\kappa$ on $\mathbb B(x_0;a+\delta)\backslash \mathbb B(x_0;a-\delta)$. Finally, if $\mu_a=\mu$ then if follows from Lemma \ref{Eq:Opt2} that either $m=0$ in $\mathbb B(x_0;a)\backslash \mathbb B(x_0;a-\delta)$ and $m=\kappa$ in $\mathbb B(x_0;a+\delta)\backslash \mathbb B(x_0;a)$ or $m=\kappa$ in $\mathbb B(x_0;a)\backslash \mathbb B(x_0;a-\delta)$ and $m=0$ in $\mathbb B(x_0;a+\delta)\backslash \mathbb B(x_0;a)$. In both case, one can apply Carleman's unique continuation Theorem as done in \cite[Proof of Theorem 2.1, Step 3]{CasadoDiaz3} to conclude that $\hat u$ is radially symmetryc in $\mathbb B(x_0;\delta)$.
{In the case where $\partial\hat{E}$ meets $\partial {\Omega}$, then the boundary of ${\Omega}\backslash \hat{E}$ does not, and we conclude by using the same reasoning on ${\Omega}\backslash \hat{E}$ instead of $\hat{E}$, showing that $\hat{u}<\hat{\mu}$ on ${\Omega}\backslash \hat{E}$ and applying Serrin's theorem to $\hat{\mu}-\hat{u}$. }
\begin{proof}[Proof of Lemma \ref{Eq:Opt2}]
Let us set $v=\hat{u}-\hat{\mu}$, hence $v$ solves
\begin{equation}
\left\{\begin{array}{ll}
-\Delta v=\left(\lambda_\alpha(\hat{m})+\kappa\right)v+\left(\lambda_\alpha(\hat{m})+\kappa\right)\hat{\mu} & \text{ in }\hat{E},\\
v=0 \text{ on }\partial\hat{E}.
\end{array}\right.
\end{equation}
and we are led to show that $v>0$ in $\hat{E}$. Let $\lambda^D({\Omega})$ be the first Dirichlet eigenvalue\footnote{In other words
\begin{equation}\label{Eq:LambdaD}\lambda^D({\Omega})=\inf_{u\in W^{1,2}_0( {\Omega})\,,\int_{{\Omega}}u^2=1}\int_{\Omega}|{\nabla} u|^2>0.\end{equation}} of the Laplace operator in $E$.
By using the Rayleigh quotient \eqref{Eq:Rayleigh} we have
\begin{align*}
\lambda_\alpha(\hat{m})&= \min_{u\in W^{1,2}_0({\Omega})\,,u\neq 0}R_{\alpha,m}(u)
> \min_{u\in W^{1,2}_0({\Omega})\,,u\neq 0}\frac{\frac12\int_{\Omega}|{\nabla} u|^2-\kappa\int_{\Omega} u^2}{\int_{\Omega} u^2}
=\lambda^D({\Omega})-\kappa,
\end{align*}
so that
$\lambda_\alpha(\hat{m})+\kappa> \lambda^D({\Omega})>0$.
Now, since $v=0$ on $\partial \hat{E}$ and that $\hat{E}$ is a $\mathscr C^2$ open subset of ${\Omega}$, the extension $\tilde v$ of $v$ by zero outside $\hat{E}$ belongs to $W^{1,2}_0({\Omega})$. Since $(\lambda_\alpha(\hat{m})+\kappa)$ and $\hat{\mu}$ are non-negative, we get
$$
-\Delta \tilde v\geq (\lambda_\alpha(\hat{m})+\kappa)\tilde v\text{ in }\hat{E}
$$
Splitting $\tilde v$ into its positive and negative parts as $\tilde v=\tilde v_+-\tilde v_-$ and multiplying the equation by $\tilde v_-$ we get after an integration by parts
$$-\int_{{\Omega}} |{\nabla} \tilde v_-|^2=-\int_{\hat{E}} |{\nabla} \tilde v_-|^2\geq -(\lambda_\alpha(\hat{m})+\kappa)\int_{\hat{E}} \tilde v_-^2=-(\lambda_\alpha(\hat{m})+\kappa)\int_{{\Omega}} \tilde v_-^2.$$
Using that $\lambda_\alpha(\hat{m})+\kappa> \lambda^D({\Omega})>0$, we get
$$\int_{\Omega} |{\nabla} \tilde v_-|^2\leq (\lambda_\alpha(\hat{m})+\kappa)\int_{\Omega} \tilde v_-^2<\lambda^D({\Omega})\int_{\Omega} \tilde v_-^2,
$$
which, combined with the Rayleigh quotient formulation of $\lambda^D({\Omega})$ yields $\tilde v_-=0$. Hence $v$ is nonnegative in $\hat{E}$.
Using moreover that $(\lambda_\alpha(\hat{m})+\kappa)\geq 0$ and $\hat{\mu}\geq 0$ yields that $-\Delta v\geq 0$ in $\hat{E}$
Notice that $v$ does not vanish identically in $\hat{E}$. Indeed, $u$ would otherwise be constant in $\hat{E}$ which cannot arise because of \eqref{Eq:EigenFunction}. According to the strong maximum principle, we infer that $v>0$ in $\hat{E}$.
\end{proof}
\begin{remark}
Following the arguments by Casado-Diaz in \cite{CasadoDiaz3}, it would be possible to weaken the regularity assumption on $E$ provided that we assume the stronger hypothesis that $\partial{\Omega}$ is simply connected. Indeed, in that case, assuming that $E$ is only of class $\mathscr C^1$ leads to the same conclusion.
\end{remark}
\section{Proof of Theorem \ref{Th:RadialStability}}
Throughout this section, ${\Omega}$ will denote the ball $\mathbb B(0,R)$, which will also be denoted ${\mathbb B(0,R)}$ for the sake of simplicity. Let $r^*_0\in (0,R)$ be chosen in such a way that
$m^*_0=\kappa \mathds 1_{\mathbb B(0,r^*_0)}$ belongs to $\mathcal M_{m_0,\kappa}({\Omega}).$ Let us introduce the notation $E_0^*={\mathbb B(0,R)}(0,r_0^*)$.
The existence part of { Theorem \ref{Th:RadialStability}} follows from a straightforward adaptation of \cite{ConcaMahadevanSanz}. In what follows, we focus on the second part of this theorem, that is, the stationarity of minimizers provided $\alpha$ is small enough.
\subsection{Steps of the proof for the stationarity } \label{sec:stepproofstationarity}
We argue by contradiction, assuming that, for any $\alpha>0$, there exists a radially symmetric distribution $\tilde m_{\alpha}$ such that $\lambda_\alpha(\tilde m_{\alpha})<\lambda_\alpha(m^*_0)$. Consider the resulting sequence $\{\tilde m_\alpha\}_{\alpha > 0}$.
\begin{itemize}
\item \textbf{Step 1:} we prove that $\{\tilde m_\alpha\}_{\alpha \to 0}$ converges strongly to $m^*_0$ in $L^1$, as $\alpha \to 0$. Regarding the associated eigenfunction, we prove that $\{u_{\alpha,m_\alpha}\}_{\alpha> 0}$ converges strongly to $u_{0,m^*_0}$ in $\mathscr C^0$ and that $\alpha {\nabla} u_{\alpha, m_\alpha}$ converges to 0 in $L^\infty({\mathbb B(0,R)})$, as $\alpha\to 0$.
\item \textbf{Step 2:} by adapting \cite[Theorem 3.7]{Laurain}, we prove that we can {restrict} ourselves to considering {\it bang-bang} radially symmetric distributions of resources $\tilde m_{\alpha}=\kappa \mathds 1_{\tilde E}$ such that the Hausdorff distance $d_H(\tilde E,E_0^*)$ is arbitrarily small.
\item \textbf{Step 3:} this is the main innovation of the proof.
Introduce $h_\alpha=\tilde m_\alpha-m^*_0$, and consider the path $\{m_t\}_{t\in [0,1]}$ from $m_\alpha$ to $m^*_0$ defined by $m_t=m_0^*+th_\alpha$. We then consider the mapping
$$
f_\alpha:t\mapsto \zeta_\alpha(m_t,\Lambda_-(m_t))
$$
where $\zeta_\alpha$ and $\Lambda_-(m_t)$ are respectively given by { equation~\eqref{Eq:Zeta} and definition~\ref{De:Homo}}. Notice that, since $m^*_0$ and $\tilde m_\alpha$ are {\it bang-bang}, $f_\alpha(0)=\lambda_\alpha(m^*_0)$ and $f_\alpha(1)=\lambda_\alpha(m_\alpha)$ according to Def. \ref{De:Homo}. Let $u_t$ be a $L^2$ normalized eigenfunction associated with $(m_t,\Lambda_-(m_t))$, in other words a solution to the equation
\begin{equation}\label{Eq:Ut}
\left\{\begin{array}{ll}
-{\nabla}\cdot\left( \Lambda_-(m_t){\nabla} u_t\right)=\zeta_\alpha(m_t,\Lambda_-(m_t)) u_t+m_tu_t & \text{ in }{\mathbb B(0,R)}\\
u_t=0 & \text{ on }\partial {\mathbb B(0,R)}\\
\int_{\mathbb B(0,R)} u_t^2=1. &
\end{array}\right.
\end{equation}
According to the proof of the optimality conditions \eqref{Eq:OptimalityH}, one has
$$
f_\alpha'(t)=\int_{\mathbb B(0,R)} h_\alpha\left(\frac\alpha{1+\alpha\kappa} \Lambda_-(m_t)^2|{\nabla} u_t|^2-u_t^2 \right).
$$
Applying the mean value theorem yields the existence of $t_1\in (0,1)$ such that $\lambda_\alpha(\tilde m_\alpha)-\lambda_\alpha(m^*_0)={f_{\alpha}'(t_1)}$. This enables us to show that, for $t\in [0,1]$ and $\alpha$ small enough, one has
$$
f_\alpha'(t)\geq C \int_{\mathbb B(0,R)} |h_\alpha| \operatorname{dist}(\cdot, {\mathbb S(0,r^*_0)})
$$ for some $C>0$, giving in turn $\lambda_\alpha(m_\alpha)-\lambda_\alpha(m^*_0)\geq C \int_{\mathbb B(0,R)} |h_\alpha| \operatorname{dist}(\cdot, {\mathbb S(0,r^*_0)}).$ (we note that the same quantity is obtained in \cite{Laurain}. Nevertheless, we obtain it in a more straightforward manner which bypasses the exact decomposition of eigenfunctions and eigenvalues used there.).
\end{itemize}
Let us now provide the details of each step.
\subsection{Step 1: convergence of quasi-minimizers and of sequences of eigenfunctions}
We first investigate the convergence of quasi-minimizers.
\begin{lemma}\label{Le:Convergence}
Let $\{m_\alpha\}_{\alpha>0}$ be a sequence in $\mathcal M_{m_0,\kappa}({\Omega})$ such that,
\begin{equation}\label{eq:lem:lambalphaineg}
\forall \alpha>0, \quad \lambda_\alpha(m_\alpha)\leq \lambda_\alpha(m^*_0).
\end{equation}
Then, $\{m_\alpha\}_{\alpha>0}$ converges strongly to $m^*_0$ in $L^1({\Omega})$.
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Le:Convergence}]
The sequence $(\lambda_\alpha(m_\alpha))_{\alpha>0}$ is bounded from above. Indeed, choosing any test function $\psi\in W^{1,2}_0({\Omega})$ such that $\int_{\Omega} \psi^2=1$, it follows from \eqref{Eq:Rayleigh} that $\lambda_\alpha(m_\alpha)\leq (1+\alpha \kappa)\Vert {\nabla} \psi\Vert_2^2+\kappa \Vert \psi\Vert_2^2.$
Similarly, using once again \eqref{Eq:Rayleigh}, we get that if $\xi_\alpha$ is the first eigenvalue associated to the operator $-(1+\alpha \kappa)\Delta-\kappa $, then $\lambda_\alpha(m_\alpha)\geq \xi_\alpha$.
Since $(\xi_\alpha)_{\alpha>0}$ converges to the first eigenvalue of $-\Delta -\kappa$ as $\alpha \to0$, $(\xi_\alpha)_{\alpha>0}$ is bounded from below whenever $\alpha$ is small enough. Combining these facts yields that the sequence $(\lambda_\alpha(m_\alpha))_{\alpha> 0}$ is bounded by some positive constant $M$ and converges, up to a subfamily, to $\tilde \lambda$.
For any $\alpha>0$, let us denote by $u_\alpha$ the associated $L^2$-normalized eigenfunction associated to $\lambda_\alpha(m_\alpha)$. From the weak formulation of equation \eqref{Eq:EigenFunction} and the normalization condition $\int_{\Omega} u_\alpha^2=1$, we infer that
\begin{align*}
\Vert {\nabla} u_\alpha\Vert _2^2&=\int_{\Omega} |{\nabla} u_\alpha|^2\leq \int_{\Omega} (1+\alpha m)|{\nabla} u_\alpha|^2=\int_{\Omega} m_\alpha u_\alpha^2+\lambda_\alpha(m_\alpha)\int_{\Omega} u_\alpha^2 \leq (M+\kappa).
\end{align*}
According to the Poincar\'e inequality and the Rellich-Kondrachov Theorem, the sequence $(u_\alpha)_{\alpha >0}$ is uniformly bounded in $W^{1,2}_0({\Omega})$ and converges, up to subfamily, to $\tilde u\in W^{1,2}_0({\Omega})$ weakly in $W^{1,2}_0({\Omega})$ and strongly in $L^2({\Omega})$, and moreover $\tilde u$ is also normalized in $L^2({\Omega})$.
Furthermore, since $L^2$ convergence implies pointwise convergence (up to a subfamily), $\tilde u$ is necessarily nonnegative in ${\Omega}$. Let $\tilde m$ be a closure point of $(m_\alpha)_{\alpha >0}$ for the weak-star topology of $L^\infty$. Passing to the weak limit in the weak formulation of the equation solved by $u_\alpha$, namely Eq.~\eqref{Eq:EigenFunction}, one gets
$$
-\Delta \tilde u-\tilde m\tilde u=\tilde \lambda \tilde u\quad \text{in }{\Omega}.
$$
Since $\tilde u\geq 0$ and $\int_{{\mathbb B(0,R)}(0,R)}\tilde u^2=1$, it follows that $\tilde u$ is the principal eigenfunction of $-\Delta -\tilde m$, so that $\tilde \lambda=\lambda_0(m^*_0)$.
Mimicking this reasoning enables us to show in a similar way that, up to a subfamily, $(\lambda_\alpha(m^*_0))_{\alpha > 0}$ converges to $\lambda_0(m^*_0)$ and $(u_{\alpha,m^*_0})_{\alpha > 0}$ converges to $u_{0,m^*_0}$ as $\alpha \to 0$. Passing to the limit in the inequality \eqref{eq:lem:lambalphaineg} and since $m^*_0$ is the only minimizer of $\lambda_0$ in $\mathcal M_{m_0,\kappa}({\Omega})$ according to the Faber-Krahn inequality, we infer that necessarily, $\tilde m=m^*_0$. Moreover, $m^*_0$ being an extreme point of $\mathcal M_{m_0,\kappa}({\Omega})$, the subfamily $(m_\alpha)_{\alpha > 0}$ converges to $\tilde m=m^*_0$ (see \cite[Proposition 2.2.1]{HenrotPierre}), strongly in $L^1({\Omega})$.
\end{proof}
A straightforward adaptation of the proof of Lemma \ref{Le:Convergence} yields that both sets $\{\lambda_\alpha(m)\}_{m\in \mathcal M_{m_0,\kappa}({\Omega})}$ and $\{\Vert u_{\alpha,m}\Vert_{W^{1,2}({\Omega})}\}_{m\in \mathcal M_{m_0,\kappa}({\Omega})}$ are uniformly bounded whenever $\alpha \leq 1$.
Let us hence introduce $M_0>0$ such that
\begin{equation}\label{metz1946}
\forall \alpha\in [0,1], \quad \max \{|\lambda_\alpha(m)|, \Vert {u_{\alpha,m}}\Vert _{W^{1,2}_0({\Omega})}\}\leq M_0.
\end{equation}
The next result is the only ingredient of the proof of Theorem~\ref{Th:RadialStability} where the low dimension assumption on $n$ is needed.
\begin{lemma}\label{Le:Bounds}
Let us assume that $n=1,2,3$. There exists $M_1>0$ such that, for every radially symmetric distribution $m\in \mathcal M_{m_0,\kappa}({\Omega})$ and every $\alpha\in [0,1]$, there holds
$$
\Vert u_{\alpha,m}\Vert _{W^{1,\infty}({\mathbb B(0,R)})}\leq M_1.
$$
Furthermore, define ${\tilde \sigma_{\alpha,m}}$, ${\tilde m}$ and ${\p_{\alpha,m}}:(0,R)\to \textrm{I\kern-0.21emR}$ by
$$\forall x \in {\mathbb B(0,R)},\quad {u_{\alpha,m}}(x)={\p_{\alpha,m}}\left(|x|\right),\quad \sigma_{\alpha,m}(x)={\tilde \sigma_{\alpha,m}}(|x|),\quad m(x)={\tilde m}(|x|),$$
then ${\tilde \sigma_{\alpha,m}}({\p_{\alpha,m}})'$ belongs to $W^{1,\infty}(0,R)$.
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Le:Bounds}]
This proof is inspired by \cite[Proof of Theorem 3.3]{Laurain}. It is standard that for every $\alpha\in [0,1]$ and every radially symmetric distribution $m\in \mathcal M_{m_0,\kappa}({\Omega})$, the eigenfunction $u_{\alpha,m}$ is itself radially symmetric.
By rewriting the equation \eqref{Eq:EigenFunction} on $ {u_{\alpha,m}}$ in {spherical} coordinates, on sees that ${\p_{\alpha,m}}$ solves
\begin{equation}\label{Eq:EFRadial}
\left\{
\begin{array}{ll}
-\frac{d}{dr}\left(r^{n-1}{\tilde \sigma_{\alpha,m}} \frac{d}{dr}{\p_{\alpha,m}}\right)=\left(\lambda_\alpha(m){\p_{\alpha,m}}+{\tilde m} {\p_{\alpha,m}}\right)r^{n-1} & \text{in }(0,R)\\
{\p_{\alpha,m}}(R)=0. &
\end{array}
\right.
\end{equation}
By applying the Hardy Inequality\footnote{This inequality reads (see e.g. \cite[Lemma 1.3]{AloisKufner} or \cite{Hardy}): for any non-negative $f$,
$$
\int_0^\infty f(x)^2dx\leq 4\int_0^\infty x^2f'(x)^2dx.
$$}
to $f={\p_{\alpha,m}}$, we get
{\begin{eqnarray*}
\int_0^R {\varphi}^{2}_{\alpha,m}(r)\, dr & \leq & 4\int_0^R r^2{\varphi}'_{\alpha,m}(r)^2\, dr \\
& \leq & 4R^2\int_0^R \left(\frac{r}{R^{2}}\right)^{n-1}{\varphi}'_{\alpha,m}(r)^2dr=4R^{4-2n}\Vert{\nabla} {u_{\alpha,m}}\Vert_{L^2({\mathbb B(0,R)})}^2\leq M,
\end{eqnarray*}}
since $n\in \{1,2,3\}$. Hence, there exists $C>0$ such that
\begin{equation}\label{Borne1D}
\Vert {\p_{\alpha,m}}\Vert _{L^2(0,R)}^2\leq C.
\end{equation}
We will successively prove that ${\p_{\alpha,m}}$ is uniformly bounded in $W^{1,2}_0(0,R)$, then in $L^\infty(0,R)$ to infer that ${\varphi}'_{\alpha,m}$ is bounded in $L^\infty(0,R)$. This proves in particular that $\sigma_{\alpha,m}{\varphi}_{\alpha,m}'\in L^{\infty}(0,R)$.
We will then conclude that $\sigma_{\alpha,m}{\varphi}_{\alpha,m}'\in W^{1,\infty}(0,R)$ by using that it is a continuous function whose derivative is uniformly bounded in $L^\infty$ by the equation on ${\p_{\alpha,m}}$.\
According to \eqref{metz1946}, {one sees that} $\Vert r^{\frac{n-1}2}{\varphi}'_{\alpha,m}\Vert _{L^2(0,R)}=\Vert{\nabla} {u_{\alpha,m}}\Vert_{L^2({\mathbb B(0,R)})}$ is bounded and therefore, $r^{n-1}{\varphi}'_{\alpha,m}(r)$ converges to 0 as $r\to 0$. {If $n=1$, such a convergence holds since ${\varphi}'_{\alpha,m}(0)=0$ by radial symmetry}. Hence, integrating Eq. \eqref{Eq:EFRadial} between $0$ and $r>0$ yields
$$
{\tilde \sigma_{\alpha,m}}(r){\varphi}'_{\alpha,m}(r)=-\frac1{r^{n-1}}\int_0^r t^{n-1}\left(\lambda_\alpha(m){\p_{\alpha,m}}(t)+{\tilde m}(t) {\p_{\alpha,m}}(t)\right) dt.
$$
By using the Cauchy-Schwarz inequality and \eqref{Borne1D}, we get the existence of $\tilde M>0$ such that
\begin{eqnarray*}
\Vert {\varphi}'_{\alpha,m}\Vert _{L^2(0,R)}^2 &\leq &\int_0^1 \left({\tilde \sigma_{\alpha,m}} {\varphi}'_{\alpha,m}\right)^2(r)\, dr\\
&=&\int_0^1\frac1{r^{2(n-1)}}\left(\int_0^r t^{n-1}\left(\lambda_\alpha(m){\p_{\alpha,m}}(t)+{\tilde m}(t) {\p_{\alpha,m}}(t)\right) dt\right)^2\, dr\\
&\leq &\int_0^1\frac1{r^{2(n-1)}}(\lambda_\alpha(m)+\kappa)^2\Vert {\p_{\alpha,m}}\Vert _{L^2(0,R)}^2\Vert {t\mapsto t^{n-1}}\Vert _{L^2(0,r)}^2\, dr\\
&\leq &\frac{\tilde M}{4n-2}\Vert {\p_{\alpha,m}}\Vert _{L^2(0,R)}^2\leq \tilde M\Vert {\p_{\alpha,m}}\Vert _{L^2(0,R)}^2\leq \tilde M C,
\end{eqnarray*}
Hence, ${\p_{\alpha,m}}$ is uniformly bounded in $W^{1,2}_0(0,R)$.
It follows from standard Sobolev embedding's theorems that there exists a constant $M_2>0$, such that
$\Vert {\p_{\alpha,m}}\Vert _{L^\infty(0,R)}\leq M_2$.
Finally, plugging this estimate in the equality
$$
{\tilde \sigma_{\alpha,m}}(r){\varphi}'_{\alpha,m}(r)=-\frac1{r^{n-1}}\int_0^r t^{n-1}\left(\lambda_\alpha(m){\p_{\alpha,m}}(t)+{\tilde m}(t) {\p_{\alpha,m}}(t)\right) dt
$$
and since $t^{n-1}\leq r^{n-1}$ on $(0,r)$, we get that ${\varphi}'_{\alpha,m}$ is uniformly bounded in $L^\infty(0,R)$.
\end{proof}
The next lemma is a direct corollary of Lemma \ref{Le:Convergence}, Lemma \ref{Le:Bounds} and the Arzela-Ascoli Theorem.
\begin{lemma}\label{Le:ConvergenceFP}
Let $(m_\alpha)_{\alpha>0}$ be a sequence of radially symmetric functions of $\mathcal M_{m_0,\kappa}({\Omega})$ such that, for every $\alpha\in [0,1]$, $\lambda_{\alpha}(m_\alpha)\leq \lambda_{\alpha}(m^*_0)$. Then, up to a subfamily,
$u_{\alpha,m_\alpha}$ converges to $u_{0,m^*_0}$ for the strong topology of $\mathscr C^0(\overline{\Omega})$ as $\alpha\to 0$.
\end{lemma}
\subsection{Step 2: reduction to particular resource distributions close to $m_0^*$}
Let us consider a sequence of radially symmetric distributions $(m_\alpha)_{\alpha > 0}$ such that, for every $\alpha\in [0,1]$, $\lambda_\alpha(m_\alpha)\leq \lambda_\alpha(m^*_0)$.
According to Proposition~\ref{Th:BangBang}, we can assume that each $m_\alpha$ is a {\it bang-bang}, in other words that $m_\alpha =\kappa \mathds{1}_{E_\alpha}$ where $E_\alpha$ is a measurable subset of ${\mathbb B(0,R)}(0,R)$. For every $\alpha \in [0,1]$, one introduces $d_\alpha=d_H(E_\alpha,E^*_0)$, the Hausdorff distance of $E_\alpha$ to $E^*_0$.
\begin{lemma}\label{Le:Hausdorff}
For every ${\varepsilon}>0$ small enough, there exists $\overline \alpha>0$ such that, for every $\alpha\in [0, \overline \alpha]$, there exists a {radially symmetric } measurable subset $\tilde{E}_\alpha$ of ${\Omega}$ such that
$$
\lambda_\alpha(\kappa\mathds{1}_{E_\alpha})\geq \lambda_\alpha(\kappa\mathds{1}_{\tilde{E}_\alpha}), \quad |E_\alpha|=|\tilde{E}_\alpha|\quad \text{and}\quad d_H(\tilde{E}_\alpha,E^*_0) \leq {\varepsilon}.
$$
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Le:Hausdorff}]
Let $\alpha\in [0,1]$. Observe first that
$\lambda_\alpha(m)=\int_{\mathbb B(0,R)}|{\nabla} {u_{\alpha,m}}|^2+\alpha \int_{\mathbb B(0,R)} m|{\nabla} {u_{\alpha,m}}|^2-\int_{\mathbb B(0,R)} mu_{\alpha,m}^2=\int_{\mathbb B(0,R)} |{\nabla} {u_{\alpha,m}}|^2+\int_{\mathbb B(0,R)} m\psi_{\alpha,m}$, where $\psi_{\alpha,m}$ has been introduced in Lemma~\ref{Le:DeriveeL21}.
We will first construct $ \tilde m_\alpha$ in such a way that
$$
\lambda_\alpha(m_\alpha)\geq \int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\int_{\Omega} \psi_{\alpha,m_\alpha} \tilde m_\alpha\geq \lambda_\alpha(\tilde m_\alpha),
$$
and, to this aim, we will define $\tilde m_\alpha$ as a suitable level set of $\psi_{\alpha,m_\alpha}$. Thus, we will evaluate the Hausdorff distance of these level sets to $E^*_0$. The main difficulty here rests upon the lack of regularity of the switching function $\psi_{\alpha,m_\alpha}$, which is { not even continuous (see Figure (\ref{fig:phi})).}
According to Lemmas~\ref{Le:Bounds} and \ref{Le:ConvergenceFP}, $\psi_{\alpha,m_\alpha}$ converges to $-u_{0,m^*_0}^2$ for the strong topology of $L^\infty({\mathbb B(0,R)})$.
Recall that $m^*_0=\kappa \mathds{1}_{\mathbb B(0,r^*_0)}$ and let $V_0$ be defined by $V_0=|\mathbb B(0,r^*_0)|$.
Let us define $\mu_\alpha^*$ by dichotomy, as the only real number such that
$$
|\underline \omega_\alpha|\leq V_0\leq |\overline \omega_\alpha|,
$$
where $\underline \omega_\alpha=\{\psi_{\alpha,m_\alpha}<\mu_\alpha^*\}$ and $\overline \omega_\alpha=\{\psi_{\alpha,m_\alpha}\leq \mu_\alpha^*\}$.
Since $\left|\{\psi_{0,m^*_0}<-\varphi_{0,m_0^*}^2(r_0^*)\}\right|=V_0$, we deduce that $(\mu_\alpha^*)$ converges to $-\varphi_{0,m^*_0}^2(r^*_0)$ as $\alpha\to 0$.
Since $\varphi_{0,m^*_0}$ is decreasing, we infer that for any ${\varepsilon}>0$ small enough, there exists $\overline \alpha>0$ such that: for every $\alpha\in [0, \overline \alpha]$, $\mathbb B(0,r^*_0-{\varepsilon})\subset \underline \omega_\alpha\subset \overline \omega_\alpha\subset \mathbb B(0,r^*_0+{\varepsilon})$.
Therefore, there exists a radially symmetric set $B_{\varepsilon}^\alpha$ such that
$$
\underline \omega_\alpha \subset B_{\varepsilon}^\alpha\subset \overline \omega_\alpha,\quad |B_{\varepsilon}^\alpha|=V_0 ,\quad d_H(B_{\varepsilon}^\alpha,E^*_0)\leq {\varepsilon}.
$$
Since $E_\alpha$ and $B_{\varepsilon}^\alpha$ have the same measure, one has $|(E_\alpha)^c\cap B_{\varepsilon}^\alpha|=|E_\alpha\cap (B_{\varepsilon}^\alpha)^c|$, we introduce $\tilde m_\alpha=\kappa \mathds{1}_{B_{\varepsilon}^\alpha}$ so that $\tilde m_\alpha$ belongs to $\mathcal M_{m_0,\kappa}({\Omega})$.
\begin{figure}[H]\label{fig:phi}
\begin{center}
\includegraphics[width=5.2cm]{graphPsialpha.png}
\caption{Possible graph of the discontinuous function $\psi_{\alpha,m_\alpha}$. The bold intervals on the $x$ axis correspond to $\{m_\alpha=0\}$.
}
\end{center}
\end{figure}
By construction, one has
\begin{align*}
\lambda_\alpha(m_\alpha)&=\int_{\mathbb B(0,R)} (1+\alpha m_\alpha)|{\nabla} u_{\alpha,m_\alpha}|^2-\int_{\mathbb B(0,R)} m_\alpha u_{\alpha,m}^2
=\int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\int_{\mathbb B(0,R)} \psi_{\alpha,m_\alpha}m_\alpha
\\&=\int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\kappa \int_{E_\alpha}\psi_{\alpha,m_\alpha}
=\int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\kappa \int_{E_\alpha\cap (B_{\varepsilon}^\alpha)^c}\psi_{\alpha,m_\alpha}+\kappa \int_{E_\alpha\cap B_{\varepsilon}^\alpha}\psi_{\alpha,m_\alpha}
\\&\geq \int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\kappa \mu_\alpha^* |E_\alpha \cap( B_{\varepsilon}^\alpha)^c|+\kappa \int_{E_\alpha\cap B_{\varepsilon}^\alpha}\psi_{\alpha,m_\alpha}
\\&=\int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\kappa \mu_\alpha^* |(E_\alpha)^c \cap B_{\varepsilon}^\alpha|+\kappa \int_{E_\alpha\cap B_{\varepsilon}^\alpha}\psi_{\alpha,m_\alpha}
\\&\geq \int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\kappa \int_{(E_\alpha)^c\cap B_{\varepsilon}^\alpha}\psi_{\alpha,m_\alpha}+\kappa \int_{E_\alpha\cap B_{\varepsilon}^\alpha}\psi_{\alpha,m_\alpha}
=\int_{\mathbb B(0,R)} |{\nabla} u_{\alpha,m_\alpha}|^2+\int_{\mathbb B(0,R)} \tilde m_\alpha \psi_{\alpha,m_\alpha}
\\&=\int_{\mathbb B(0,R)}\sigma_{\alpha,\tilde m_\alpha} |{\nabla} u_{\alpha,m_\alpha}|^2-\int_{\mathbb B(0,R)} \tilde m_\alpha u_{\alpha,m}^2
\geq \lambda_\alpha(\tilde m_\alpha),
\end{align*}
the last inequality coming from the variational formulation \eqref{Eq:Rayleigh}.
The expected conclusion thus follows by taking $\tilde E_\alpha:=B_{\varepsilon}^\alpha$
\end{proof}
From now on we will replace $m_\alpha$ by $\kappa\mathds{1}_{\tilde{E}_\alpha}$ and still denote this function by $m_\alpha$ with a slight abuse of notation.
\subsection{Step 3: conclusion, by the mean value theorem}
Recall that, according to Section~\ref{sec:stepproofstationarity}, for every $\alpha\in [0,1]$, the mapping $f_\alpha$ is defined by $f_\alpha(t):=\zeta_\alpha(m_t,\Lambda_-(m_t))$ for all $t\in [0,1]$
We claim that $f_\alpha$ belongs to $\mathscr C^1$. This follows from similar arguments to those of the $L^2$ differentiability of $m\mapsto \lambda_\alpha(m)$ in Appendix \ref{Ap:Differentiability}. Following the proof of \eqref{Eq:OptimalityH}, it is also straightforward that for every $t\in [0,1]$, one has
\begin{equation}\label{Eq:DeriveeFAlpha}
f_\alpha'(t)=\int_{\mathbb B(0,R)} \left(\frac\alpha{1+\alpha\kappa} \Lambda_-(m_t)^2 |{\nabla} u_t|^2-u_t^2\right)h_\alpha.\end{equation}
Finally, since $m^*_0$ and $m_\alpha$ are {\it bang-bang}, it follows from Definition~\ref{De:Homo} that
$f_\alpha(0)=\lambda_\alpha(m^*_0)$ and $f_\alpha(1)=\lambda_\alpha(m_\alpha)$.
Since $m_\alpha$ is assumed to be radially symmetric, so is $m_t$ for every $t\in [0,1]$ thanks to a standard reasoning, and, therefore, so is $u_t$. With a slight abuse of notation, we identify $m_t$, $u_t$ and $\Lambda_-(m_t)$ with their radially symmetric part $\tilde m_t$, $\tilde u_t$, $\tilde \Lambda_-(m_t)$ defined on $[0,R]$ by
$$
u_t(x)=\tilde u_t(|x|), \quad m_t(x)=\tilde m_t(|x|),\quad \Lambda_-(m_t)(x)=\tilde \Lambda_-(\tilde m_t)(|x|).
$$
Then the function $u_t$ (defined on $[0,R]$) solves the equation
\begin{equation}\label{Eq:UtRadial}
\left\{\begin{array}{ll}-\frac{d}{dr}\left(r^{n-1}\Lambda_-(m_t)\frac{du_t}{dr}\right)=\left({\zeta_\alpha(m_t,\Lambda_-(m_t))} u_t+m_t u_t\right)r^{n-1}& r\in [0,R]\\
u_t(R)=0 &\\
\int_0^R r^{n-1}u_t(r)^2dr=\frac1{c_n},&
\end{array}\right.
\end{equation}
where $c_n=|\mathbb S(0,1)|$. As a consequence, an immediate adaptation of the proof of Lemma~\ref{Le:Bounds} yields:
\begin{lemma}\label{Le:BoundsUt}
There exists $M>0$ such that
$$
\max \left\{\Vert u_t\Vert _{W^{1,\infty}},\Big\Vert \Lambda_-(m_t)u_t'\Big\Vert_{W^{1,\infty}}\right\}\leq M.
$$
Furthermore, $\Lambda_-(m_t)u_t'$ converges to $u_{0,m^*_0}'$ in $L^\infty(0,R)$ and uniformly with respect to $t\in [0,1]$, as $\alpha\to 0$.
\end{lemma}
According to the mean value theorem, there exists $t_1=t_1(\alpha)\in [0,1]$ such that
$$
\lambda_\alpha(m_\alpha)-\lambda_\alpha(m^*_0)=f_\alpha(1)-f_\alpha(0)=f_\alpha'(t_1)
$$
and by using Eq. \eqref{Eq:DeriveeFAlpha}, one has
$$
f_\alpha'(t_1)=\int_{\mathbb B(0,R)} \left(\frac\alpha{1+\alpha\kappa} \Lambda_-(m_{t_1})^2 |{\nabla} u_{t_1}|^2-u_{t_1}^2\right)h_\alpha,
$$
where $h_\alpha=m_\alpha-m^*_0$.
Let us introduce $I_\alpha^\pm$ as the two subsets of $ [0,R]$ given by $I_\alpha^\pm=\{h_\alpha=\pm1\}$. Let ${\varepsilon}>0$. According to Lemma~\ref{Le:Hausdorff}, we have, for $\alpha$ small enough,
$I_\alpha^+\subset [r^*_0,r^*_0+{\varepsilon}]$ and $ I_\alpha^-\subset [r^*_0-{\varepsilon};r^*_0]$.
Finally, let us introduce
$$
\mathfrak F_1:=\frac\alpha{1+\alpha\kappa} \Lambda_-(m_{t_1})^2 |{\nabla} u_{t_1}|^2-u_{t_1}^2.
$$
According to Lemma~\ref{Le:BoundsUt}, $\mathfrak F_1$ belongs to $W^{1,\infty}$ and $ \mathfrak F_1+u_{\alpha,m^*_0}^2$ converges to $0$ as $\alpha\to 0$, for the strong topology of $W^{1,\infty}(0,R)$.
Moreover, there exists $M>0$ independent of $\alpha$ such that for ${\varepsilon}>0$ small enough,
$$
-M\leq 2 u_{\alpha,m^*_0}\frac{du_{\alpha,m^*_0}}{dr}\leq-M\quad \text{ in } [r^*_0-{\varepsilon};r^*_0+{\varepsilon}]
$$
and it follows that
$$
\frac{M}2\leq \frac{d\mathfrak F_1}{dr}\leq 2M\quad \text{ in }[r^*_0-{\varepsilon};r^*_0+{\varepsilon}]$$ for $\alpha$ small enough.
Hence, since $\mathfrak F_1$ is Lipschitz continuous and thus absolutely continuous, one has for every $y\in [0,{\varepsilon}]$,
\begin{align*}
\mathfrak F_1(r^*_0+y)&=\mathfrak F_1(r^*_0)+\int_{r^*_0}^{r^*_0+y} \mathfrak F_1'(s)\, ds\geq \mathfrak F_1(r^*_0)+\frac{M}2 y \\
\text{and}\quad \mathfrak F_1(r^*_0-y)&=\mathfrak F_1(r^*_0)+\int_{r^*_0-y}^{r^*_0} (-\mathfrak F_1'(s))\, ds\leq \mathfrak F_1(r^*_0)-\frac{M}{2} y.
\end{align*}
Since $h_\alpha \leq 0 $ in $[r^*_0-{\varepsilon};r^*_0]$ and $h_\alpha\geq 0$ in $ [r^*_0,r^*_0+{\varepsilon}]$, we have
\begin{eqnarray*}
h_\alpha(r^*_0+y)\mathfrak F_1(r^*_0+y)&\geq &h_\alpha(r^*_0+y)\mathfrak F_1(r^*)+\frac{|h_\alpha| (r^*_0+y)M}2 y\\
\text{and}\quad h_\alpha(r^*_0-y)\mathfrak F_1(r^*_0-y)&\geq& h_\alpha(r^*_0-y)\mathfrak F_1(r^*)+\frac{|h_\alpha|(r^*_0-y) M}2 y.
\end{eqnarray*}
for every $y\in [0,{\varepsilon}]$. Hence, using that $\int_{\mathbb B(0,R)} h_\alpha=0$, we infer that
\begin{align*}
f_\alpha'(t_1)&=\int_{\mathbb B(0,R)} \left(\frac\alpha{1+\alpha\kappa}\Lambda_-(m_{t_1})^2 |{\nabla} u_{t_1}|^2-u_{t_1}^2\right)h_\alpha =c_{n}\int_0^R h_\alpha(s) \mathfrak F_1(s) s^{n-1}\, ds
\\&=c_n \left(\int_{r^*_0-{\varepsilon}}^{r^*_0}h_\alpha\mathfrak F_1(s)s^{n-1}ds+\int_{r^*_0}^{r^*_0+{\varepsilon}} h_\alpha\mathfrak F_1(s)s^{n-1}\, ds\right)
\\&\geq c_n\left(\int_{r^*_0-{\varepsilon}}^{r^*_0}h_\alpha(s)\mathfrak F_1(r^*)s^{n-1}ds+\int_{r^*_0}^{r^*_0+{\varepsilon}} h_\alpha(s)\mathfrak F_1(r^*)s^{n-1}\, ds\right)&
\\&+\frac{c_nM}2 \left(\int_{r^*_0-{\varepsilon}}^{r^*_0}|h_\alpha|(s) |r^*_0-s|s^{n-1}ds+\int_{r^*_0}^{r^*_0+{\varepsilon}} |h_\alpha|(s)|r^*_0-s|s^{n-1}\, ds\right)
\\&=\frac{c_nM}2\int_{\mathbb B(0,R)} |h_\alpha|\operatorname{dist}(\cdot,{\mathbb S(0,r^*_0)}),
\end{align*}
which concludes Step 3. Theorem \ref{Th:RadialStability} is thus proved.
\begin{remark}\label{Susu:Concluding}
Regarding the proof of Theorem \ref{Th:RadialStability}, it would have been more natural to consider the path $t\mapsto \left(\lambda_\alpha(m_t),m_t\right)$ rather than $t\mapsto \left( {\zeta_\alpha(m_t,\Lambda_-(m_t))},m_t\right)$. However, we would have been led to consider {$\mathfrak G_1=\alpha \kappa |{\nabla} u_{\alpha,m_{t_{1}}}|^2-u_{\alpha,m_{t_{1}}}^2$ instead of $\mathfrak F_1$}. Unfortunately, this would have been more intricate because of the regularity of $\mathfrak G_1$, which is discontinuous and thus, no longer a $W^{1,\infty}$ function, so that a Lemma analogous to Lemma \ref{Le:BoundsUt} would not be true. Adapting step by step the arguments of \cite{Laurain} would nevertheless be possible although much more technical.
\end{remark}
\section{Sketch of the proof of Corollary \ref{Th:Sketch}}\label{Se:Sketch}
We do not give all details since the proof is then very similar to the ones written previously. We only underline the slight differences in every step.
To prove this result, we consider the following relaxation of our problem, which is reminiscent of the problems considered in \cite{Hamel2011}. Let us consider, for any pair $(m_1,m_2)\in \mathcal M_{m_0,\kappa}({\Omega})^2$, the first eigenvalue of the operator $\mathscr N:u\mapsto -{\nabla} \cdot\left((1+\alpha m_1){\nabla} u\right)-m_2 u$, and write it $\eta_\alpha(m_1,m_2)$. Let $m^*_0:=\kappa \mathds 1_{\mathbb B(0,R)}$.
By using the results of \cite{Hamel2011} or alternatively, applying the rearrangement of Alvino and Trombetti, \cite{AlvinoTrombetti} as it has been done in \cite{ConcaMahadevanSanz}, one proves the existence of a radially symmetric function $\tilde m_1$ such that
$$\eta_\alpha(m_1,m_2)\geq \eta_\alpha(\tilde m_1,m^*_0),$$
so that we are done if we can prove that, for any $m\in \mathcal M({\Omega})$ there holds
\begin{equation}\label{Eq:Sk}
\eta_\alpha(m,m^*_0)\geq \eta_\alpha(m^*_0,m^*_0).
\end{equation}
We claim that \eqref{Eq:Sk} holds for any $m\in \mathcal M_{m_0,\kappa}$, provided that $m_0$ and $\alpha$ be small enough. Let us describe the main steps of the proof:
\begin{itemize}
\item {\bf Step 1:} mimicking the compactness argument used in \cite{ConcaMahadevanSanz}, one shows that there exists a solution $m_\alpha$ to the problem
$$\inf_{m\in \mathcal M_{m_0,\kappa}({\Omega})}\eta_\alpha(m,m^*_0),$$
which is radially symmetric and {\it bang-bang}. We write it $m_\alpha=\kappa \mathds 1_{E_\alpha}$.
\item {\bf Step 2:} let $\mu_0$ and $r_0^*$ be the unique real numbers such that
$$
\left|\left\{|{\nabla} u_{0,m^*_0}|^2\leq \mu_0\right\}\right|=V_0=|\mathbb B(0,r^*_0)|.
$$
Introducing $E_0=\left\{|{\nabla} u_{0,m^*_0}|^2\leq \mu_0\right\}$, we prove that
$m_\alpha$ converges in $L^1({\Omega})$ to $\kappa \mathds 1_{E_0}$ as $\alpha \to 0$.
\item {\bf Step 3:} we establish that if $m_0$ is small enough, then $E_0=\mathbb B(0,r^*_0)$. This is done by proving that $u_{0,m^*_0}$ converges in $\mathscr C^1$ to the first Dirichlet eigenfunction of the ball as $r^*_0\to 0$ and by determining the level-sets of this first eigenfunction, as done in \cite[Section 2.2]{ConcaLaurainMahadevan}.
\item {\bf Step 4:} once this limit identified, we mimick the steps of the proof of Theorem \ref{Th:RadialStability} (reduction to a small Hausdorff distance and mean value theorem for a well-chosen auxiliary function) to conclude that one necessarily has $m_\alpha=m^*_0$ for $\alpha$ small enough.
\end{itemize}
\section{Proof of Theorem \ref{Th:ShapeStability}}
Throughout this section, we will denote by $\mathbb B^*$ the ball $\mathbb B(0,r^*_0)$, where $r^*_0$ is chosen so that $m^*_0=\kappa \mathds{1}_{\mathbb B^*}$ belongs to $\mathcal M_{m_0,\kappa}({\Omega})$.
When it makes sense, we will write $f|_{int}(y)=\lim_{x\in {\mathbb B(0,R)}^*,x\to y}f(x)$, $f|_{ext}(y):=\lim_{x \in ({\mathbb B(0,R)}^*)^c,x\to y}f(y)$, so that $\llbracket f\rrbracket =f|_{ext}-f|_{int}$ denotes the jump of $f$ at the boundary ${\mathbb S(0,r^*_0)}$.
\subsection{Preliminaries}
\allowdisplaybreaks
\def{\frac{\partial}{\partial r_i}}{{\frac{\partial}{\partial r_i}}}
\def{u_{1,\alpha}^{(k)}}{{u_{1,\alpha}^{(k)}}}
For ${\varepsilon}>0$, let us introduce $\mathbb B^*_{\varepsilon}:=(\operatorname{Id}+{\varepsilon} V)\mathbb B^*$ and define $u_{\varepsilon}$ as the $L^2$-normalized first eigenfunction associated with $m_{\varepsilon}=\kappa \mathds{1}_{\mathbb B^*_{\varepsilon}}$.
It is well known (see e.g. \cite{Henrot2006,HenrotPierre}) that $u_{\varepsilon}$ expands as
\begin{equation}\label{An:EigenFunction}
u_{\varepsilon}=u_{0,\alpha}+{\varepsilon} u_{1,\alpha}+{\varepsilon}^2 \frac{u_{2,\alpha}}2+\underset{{\varepsilon} \to 0}{\operatorname{o}} ({\varepsilon}^2)\quad \text{in }H^1({\mathbb B(0,R)}^*)\text{ and in }H^1({\Omega} \backslash {\mathbb B(0,R)}^*),
\end{equation}
where, in particular, $u_{0,\alpha}=u_{\alpha,m^*_0}$, whereas $\lambda_\alpha(\mathbb B^*_{\varepsilon})$ expands as
\begin{eqnarray}
\lambda_\alpha(\mathbb B^*_{\varepsilon})&=&\lambda_{0,\alpha}+{\varepsilon}\lambda_\alpha'(\mathbb B^*)[V]+\frac{{\varepsilon}^2}{2}\lambda_\alpha''(\mathbb B^*)[V]+\underset{{\varepsilon} \to 0}{\operatorname{o}}({\varepsilon}^2)\nonumber \\
&=&\lambda_{0,\alpha}+{\varepsilon}\lambda_{1,\alpha}+\frac{{\varepsilon}^2}{2}\lambda_{2,\alpha}+\underset{{\varepsilon} \to 0}{\operatorname{o}}({\varepsilon}^2).\label{An:EigenValue}
\end{eqnarray}
By mimicking the proof of Lemma \ref{Le:Bounds}, one shows the following symmetry result.
\begin{lemma}\label{Le:Rad}
The function $u_{\alpha,m^*_0}$ is radially symmetric. Let ${\varphi}_{\alpha,m^*_0}$, ${\tilde m}$ and $\tilde \sigma_{\alpha,\tilde m}$ be such that $u_{\alpha,m^*_0}={\varphi}_{\alpha,m^*_0}(|\cdot|)$, $m^*_0={\tilde m}(|\cdot|)$ and $\tilde \sigma_{\alpha,\tilde m}=1+\alpha {\tilde m}$. Then ${\varphi}_{\alpha,m^*_0}$ satisfies the ODE
\begin{equation}\label{Eq:RadU0}\left\{
\begin{array}{ll}
-\frac{d}{dr}\left(r^{n-1}\tilde \sigma_{\alpha,\tilde m} \varphi_{\alpha,m_0^*}'\right)=\left(\lambda_\alpha(m_0^*)+{\tilde m} \right)\varphi_{\alpha,m_0^*} r^{n-1} & \text{in }(0,R)\\
\varphi_{\alpha,m_0^*}(R)=0 &
\end{array}
\right.
\end{equation}
complemented by the following jump conditions
\begin{equation}\label{Eq:U0}
\llbracket {\varphi}_{\alpha,m^*_0}\rrbracket (r^*_0)=\llbracket \tilde \sigma_{\alpha,m^*_0}{\varphi}_{\alpha,m^*_0}'\rrbracket (r^*_0)=0,\quad \llbracket \tilde \sigma_{\alpha,\tilde m}{\varphi}_{\alpha,m_0^*}''\rrbracket (r^*_0)=\kappa {\varphi}_{\alpha,m^*_0}(r^*_0).
\end{equation}
Furthermore, ${\varphi}_{\alpha,m^*_0}$ converges to ${\varphi}_{0,m^*_0}$ for the strong topology of $\mathscr C^1$ as $\alpha\to 0$.
\end{lemma}
\subsection{Computation of the first and second order shape derivatives}
\paragraph{A remark on the type of vector fields we consider}
Hadamard's structure theorem (see for instance \cite[Theorem~5.9.2 and the remark below]{HenrotPierre}) ensures that the first order derivative in the direction of a vector field $V$ only depends on the normal trace of $V$. This allows us to work with only normal vector fields $V$ to compute the first order derivative.
Once it is established that $\mathbb B^*$ is a critical shape, we can use Hadamard's structure theorem \cite[Theorem~5.9.2 and the remark below]{HenrotPierre} which states that the second order shape derivative, when computed at a critical shape only depends on the normal trace, hence we will also, for second order shape derivatives, work with normal vector fields.
Since we are working in {two dimensions}, this means that one can deal with vector fields $V$ given in polar coordinates by
$$
V(r^*_0,\theta)=g(\theta)\begin{pmatrix}\cos\theta\\\sin\theta \end{pmatrix}.
$$
The proof of the shape differentiability at the first and second order of $\lambda_\alpha$, based on an implicit function argument according to the method of \cite{MignotMuratPuel}, is exactly similar to \cite[Proof of Theorem 2.2]{DambrineKateb}. For this reason, we admit it. Nevertheless, in what follows, we provide some details on the computation of these derivatives for the sake of completeness, since some steps differ a bit from those done in the references above.
\paragraph{Computation and analysis of the first order shape derivative.}
Let us prove that ${\mathbb B(0,R)}^*$ is a critical shape in the sense of \eqref{Eq:FOO}.
\begin{lemma}\label{Le:Cr}
The first order shape derivative of $\lambda_\alpha$ at ${\mathbb B(0,R)}^*$ in direction $V$ reads
\begin{equation}
\lambda_{1,\alpha}=\lambda_\alpha'({\mathbb B(0,R)}^*)[V]=\int_{{\mathbb S(0,r^*_0)}} V\cdot \nu .
\end{equation}
For all $V \in \mathcal X({\mathbb B(0,R)}^*)$ (defined by \eqref{Eq:X}), one has $\lambda_{1,\alpha}=0$ meaning that ${\mathbb B(0,R)}^*$ satisfies \eqref{Eq:FOO}.
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Le:Cr}]
First, elementary computations show that $u_{1,\alpha}$ solves
\begin{equation}\label{Eq:EigenDerivative1}
\left\{
\begin{array}{ll}
-{\nabla} \cdot \Big(\sigma_\alpha {\nabla} u_{1,\alpha}\Big)=\lambda_{1,\alpha}u_{0,\alpha}+\lambda_{0,\alpha}u_{1,\alpha}+m^*_0u_{1,\alpha}&\text{ in }{\mathbb B(0,R)}(0,R),\\
\left\llbracket \sigma_\alpha\frac{\partial u_{1,\alpha}}{\partial \nu}\right\rrbracket (r^*_0\cos\theta ,r^*_0\sin\theta )=-\kappa g(\theta)u_{0,\alpha},& \\
\left\llbracket u_{1,\alpha}\right\rrbracket (r^*_0\cos\theta ,r^*_0\sin\theta )=-g(\theta)\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket (r^*_0\cos\theta ,r^*_0\sin\theta ),&
\end{array}
\right.
\end{equation}
where $\sigma_\alpha=1+\alpha m_0^*$ and the notation $\llbracket\cdot\rrbracket$ denote the jumps of the functions at ${\mathbb S(0,r^*_0)}$.
The derivation of the main equation of \eqref{Eq:EigenDerivative1} is an adaptation of the computations in \cite{DambrineKateb}.
To derive the jump on $u_{1,\alpha}$, we follow \cite{DambrineKateb} and differentiate the continuity equation
$\llbracket u_{\varepsilon}\rrbracket _{\partial {\mathbb B(0,R)}_{\varepsilon}^*}=0$.
Formally plugging \eqref{An:EigenFunction} in this equation yields
$$u_{1,\alpha}\vert_{int}(r^*_0,\theta)+\left.g(\theta)\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int}=u_{1,\alpha}\vert_{ext}(r^*_0,\theta)+\left.g(\theta)\frac{\partial u_{0,\alpha}}{\partial r}\right|_{ext},$$
and hence
$$
\llbracket u_{1,\alpha}\rrbracket = u_{1,\alpha}|_{ext}-u_{1,\alpha}|_{int}=-g(\theta)\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket.
$$
Note that the same goes for the normal derivative: we differentiate the continuity equation
$$\left\llbracket (1+\alpha m_{\varepsilon})\frac{\partial u_{{\varepsilon},\alpha}}{\partial \nu}\right\rrbracket_{\partial {\mathbb B(0,R)}_{\varepsilon}^*}=0,$$
yielding
$$\left\llbracket \sigma_\alpha \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket=-g(\theta)\left\llbracket\sigma_\alpha \frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right\rrbracket.
$$
According to the equation $-\sigma_\alpha \Delta u_{0,\alpha}=\lambda_\alpha(m^*_0)u_{0,\alpha}+m^*_0u_{0,\alpha}$ in ${\mathbb B(0,R)}^*$, this rewrites
\begin{equation}\label{Eq:Jump1}
\left\llbracket \sigma_\alpha \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket =-\kappa g(\theta)u_{0,\alpha}.
\end{equation}
Now, using $u_{0,\alpha}$ as a test function in \eqref{Eq:EigenDerivative1}, we get
\begin{align*}
\lambda_{1,\alpha}&=-\int_{\mathbb B(0,R)} u_{0,\alpha} \nabla\cdot (\sigma_\alpha\nabla u_{1,\alpha})-\int_{\mathbb B(0,R)} m^*_0 u_{1,\alpha}u_{0,\alpha}\\
&=-\int_{\mathbb B(0,R)} u_{0,\alpha} \nabla\cdot (\sigma_\alpha\nabla u_{1,\alpha})+\int_{\mathbb B(0,R)} u_{1,\alpha} \nabla \cdot (\sigma_\alpha u_{0,\alpha})
\\&=\int_{{\mathbb S(0,r^*_0)}}u_{0,\alpha}\left\llbracket \sigma_\alpha \frac{\partial u_{1,\alpha}}{\partial \nu} \right\rrbracket -\int_{{\mathbb S(0,r^*_0)}}\left\llbracket \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}u_{1,\alpha}\right\rrbracket \\
&=-r^*_0\int_{0}^{2\pi} \kappa g(\theta)u_{0,\alpha}(r^*_0)^2\, d\theta+r^*_0\int_0^{2\pi}g(\theta)\left(\sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\right)\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket\, d\theta\\
&=r^*_0\int_0^{2\pi}g(\theta)\left(-\kappa u_{0,\alpha}(r^*_0)^2+\left\llbracket \sigma_\alpha\left( \frac{\partial u_{0,\alpha}}{\partial r}\right)^2\right\rrbracket \right)\, d\theta.
\end{align*}
by using that $\int_{{\mathbb B(0,R)}(0,R)}u_{{\varepsilon}}^2=1$, so that $\int_{{\mathbb B(0,R)}(0,R)} u_{0,\alpha}u_{1,\alpha}=0$ by differentiation.
Since $u_{0,\alpha}$ is radially symmetric according to Lemma \ref{Le:Rad}, we introduce the two real numbers
{\begin{equation}\label{LM}
\eta_\alpha:=-\kappa u_{0,\alpha}(r^*_0)^2+\left\llbracket \sigma_\alpha\left( \frac{\partial u_{0,\alpha}}{\partial r}\right)^2\right\rrbracket
\quad \text{and}\quad \lambda_{1,\alpha}:=r^*_0\eta_\alpha \int_0^{2\pi}g(\theta)\, d\theta.
\end{equation}}
It is easy to see that $V$ belongs to $\mathcal X({\mathbb B(0,R)}^*)$ if, and only if $\int_0^{2\pi}g=0$ so that we finally have $\lambda_{1,\alpha}=0.$
\end{proof}
\paragraph{Computation of the Lagrange multiplier.}
The existence of a Lagrange multiplier $\Lambda_\alpha\in \textrm{I\kern-0.21emR}$ related to the volume constraint $\operatorname{Vol}(E)=m_0\operatorname{Vol}({\Omega})/\kappa$ is standard, and one has
$$
\forall V \in \mathcal X({\mathbb B(0,R)}^*), \quad \left(\lambda_\alpha'-\Lambda_\alpha\operatorname{Vol}'\right)({\mathbb B(0,R)}^*)[V]=0.
$$
Since
$$\operatorname{Vol}'({\mathbb B(0,R)}^*)[V]=\int_{{\mathbb S(0,r^*_0)}} V\cdot \nu =r^*_0\int_0^{2\pi}g(\theta)d\theta.$$
(see e.g. \cite[chapitre 5]{HenrotPierre}) and since
$$\lambda_\alpha'({\mathbb B(0,R)}^*)[V]=r^*_0\eta_\alpha\int_0^{2\pi}g(\theta)d\theta,$$
where $\eta_\alpha$ is defined by \eqref{LM}, the Lagrange multiplier reads
$$\Lambda_\alpha=\eta_\alpha=-\kappa u_{0,\alpha}(r^*_0)^2+\left\llbracket \sigma_\alpha\left( \frac{\partial u_{0,\alpha}}{\partial r}\right)^2\right\rrbracket .$$
\paragraph{Computation of the second order derivative and second order optimality conditions.}
Let us compute the second order derivative of $\lambda_\alpha$.
By using the Hadamard structure Theorem (see \cite[Theorem~5.9.2 and the following remark]{HenrotPierre}), since ${\mathbb B(0,R)}^*$ is a critical shape in the sense of \eqref{Eq:FOO}, it is not restrictive to deal with vector fields that are normal to the $\partial {\mathbb B(0,R)}^*=\mathbb S^*$, according to the so-called structure theorem which provides the generic structure of second order shape derivatives. This allow us to identify any such $V\in \mathcal X({\mathbb B(0,R)}^*)$ with a periodic function $g:[0,2\pi]\rightarrow \textrm{I\kern-0.21emR}$ such that
$$
V(r^*_0\cos\theta ,r^*_0\sin\theta )=g(\theta)\begin{pmatrix}\cos\theta \\\sin\theta \end{pmatrix}.
$$
\begin{lemma}\label{Le:SecondDerivative}
For every $V\in \mathcal X({\mathbb B(0,R)}^*)$, one has for the coefficient $\lambda_{2,\alpha}= \lambda_\alpha''({\mathbb B(0,R)}^*)[V,V]$ introduced in \eqref{An:EigenValue} the expression
\begin{eqnarray*}
\lambda_{2,\alpha}&=& 2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \partial_r u_{1,\alpha}|_{int}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket V\cdot \nu -2\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu \\
&& +\int_{{\mathbb S(0,r^*_0)}}\left(-\frac1{r^*_0}\left\llbracket \sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right\rrbracket -\frac{\kappa}{r^*_0}u_{0,\alpha}^2\right) (V\cdot \nu)^2
-2\int_{{\mathbb S(0,r^*_0)}}\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2.
\end{eqnarray*}
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Le:SecondDerivative}]
In the computations below, we do not need to make the equation satisfied by $u_{2,\alpha}$ explicit, but we nevertheless will need several times the knowledge of $\llbracket u_{2,\alpha}\rrbracket $ at ${\mathbb S(0,r^*_0)}$.
In the same fashion that we obtained the jump conditions on $u_{1,\alpha}$
Let us differentiate two times the continuity equation $\llbracket u_{\varepsilon}\rrbracket _{\partial {\mathbb B(0,R)}^*_{\varepsilon}}=0$. We obtain
\begin{equation}\label{Eq:Jump2}
\llbracket u_{2,\alpha}\rrbracket _{\partial {\mathbb B(0,R)}_{\varepsilon}^*}=-2g(\theta)\left\llbracket \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket -g(\theta)^2 \left\llbracket \frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right\rrbracket .
\end{equation}
Now, according to Hadamard second variation formula (see \cite[Chapitre 5, page 227]{HenrotPierre} for a proof), if ${\Omega}$ is a $\mathscr C^2$ domain and $f$ is two times differentiable at 0 and taking values in $W^{2,2}({\Omega})$, then one has
\begin{equation}\label{Eq:Hada2}
\left.\frac{d^2}{dt^2}\right|_{t=0}\int_{(\operatorname{Id}+tV){\Omega}} f(t)=\int_{{\Omega}} f''(0)+2\int_{\partial {\Omega}} f'(0) V\cdot \nu +\int_{\partial {\Omega}}\left(H f(0)+\frac{\partial f(0)}{\partial \nu}\right) (V\cdot \nu)^2,
\end{equation}
where $H$ denotes the mean curvature. We apply it to $f({\varepsilon})=\sigma_{\alpha,{\varepsilon}}|{\nabla} u_{\varepsilon}|^2-m_{\varepsilon} u_{\varepsilon}^2$ on $\mathbb B(0,R)$, since $\lambda_\alpha(m_{\varepsilon})=\int_{\mathbb B(0,R)}{ f({\varepsilon})}$.
Let us distinguish between the two subdomains ${\mathbb B(0,R)}_{\varepsilon}^*$ and $({\mathbb B(0,R)}_{\varepsilon}^*)^c$. We introduce
$$
D_1 = \left. \frac{d^2}{d{\varepsilon}^2}\right|_{{\varepsilon}=0}\int_{{\mathbb B(0,R)}_{\varepsilon}^*}\left(\sigma_{\alpha,{\varepsilon}}|{\nabla} u_{\varepsilon}|^2-\kappa u_{\varepsilon}^2\right) \quad \text{and}\quad
D_2 = \left. \frac{d^2}{d{\varepsilon}^2}\right|_{{\varepsilon}=0}\int_{({\mathbb B(0,R)}_{\varepsilon}^*)^c}\left(\sigma_{\alpha,{\varepsilon}}|{\nabla} u_{\varepsilon}|^2\right),
$$
so that $ \lambda_\alpha''({\mathbb B(0,R)}^*)[V,V]=D_1+D_2$.
One has
\begin{eqnarray*}
D_1 &=&\int_{{\mathbb B(0,R)}^*}2(1+\alpha \kappa) {\nabla} u_{2,\alpha}\cdot {\nabla} u_{1,\alpha}+2\int_{{\mathbb B(0,R)}_{\varepsilon}^*}(1+\alpha \kappa)|{\nabla} u_{1,\alpha}|^2
\\&&-2\kappa\int_{{\mathbb B(0,R)}^*}u_{2,\alpha}u_{0,\alpha}-2\kappa\int_{{\mathbb B(0,R)}^*}u_{1,\alpha}u_{0,\alpha}
\\&&+4\int_{{\mathbb S(0,r^*_0)}}(1+\alpha \kappa) ({\nabla} u_{1,\alpha}|_{int}\cdot {\nabla} u_{0,\alpha}|_{int}) V\cdot \nu -4\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu
\\&&+\int_{{\mathbb S(0,r^*_0)}}\left(\frac1{r^*_0}(1+\alpha \kappa)|{\nabla} u_{0,\alpha}|_{int}^2-\frac\kappa{r^*}u_{0,\alpha}^2+2(1+\alpha\kappa)\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int}\left.\frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right|_{int}\right.
\\&&\left.-2\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int}\right) (V\cdot \nu)^2,
\end{eqnarray*}
and taking into account that the mean curvature has a sign on $({\mathbb B(0,R)}_{\varepsilon}^*)^c$, one has
\begin{eqnarray*}
D_2 &=&\int_{({\mathbb B(0,R)}^*)^c}2 {\nabla} u_{2,\alpha}\cdot {\nabla} u_{1,\alpha}+2\int_{({\mathbb B(0,R)}^*)^c}|{\nabla} u_{1,\alpha}|^2
\\&&-4\int_{{\mathbb S(0,r^*_0)}} ({\nabla} u_{1,\alpha}|_{ext}\cdot {\nabla} u_{0,\alpha}|_{ext}) V\cdot \nu
\\&&+\int_{{\mathbb S(0,r^*_0)}}\left(-\frac1{r^*_0}|{\nabla} u_{0,\alpha}|_{ext}^2-2\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{ext}\left.\frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right|_{ext}\right) (V\cdot \nu)^2.
\end{eqnarray*}
Summing these two quantities, we get
\begin{eqnarray*}
\lambda_{2,\alpha}&=&2\int_{{\mathbb B(0,R)}(0,R)}\sigma_\alpha {\nabla} u_{0,\alpha}\cdot {\nabla} u_{2,\alpha}-2\int_{{\mathbb B(0,R)}(0,R)}m^*_0 u_{0,\alpha}u_{2,\alpha}
+2\int_{{\mathbb B(0,R)}(0,R)}\sigma_\alpha|{\nabla} u_{1,\alpha}|^2-2\int_{\mathbb B(0,R)}m^*_0 u_{1,\alpha}^2
\\&&-4\int_{{\mathbb S(0,r^*_0)}}\sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\left\llbracket \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket V\cdot \nu -4\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu
\\&&+\int_{{\mathbb S(0,r^*_0)}}\left(-\frac1{r^*_0}\left\llbracket \sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right\rrbracket -\frac{\kappa}{r^*_0}u_{0,\alpha}^2\right) (V\cdot \nu)^2
\\&&
-2\int_{{\mathbb S(0,r^*_0)}}\left\llbracket \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right\rrbracket (V\cdot \nu)^2-2\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2.
\end{eqnarray*}
To simplify this expression, let us use Eq. \eqref{Eq:U0}. Introducing
$$
D_3= \int_{{\mathbb B(0,R)}(0,R)}\sigma_\alpha {\nabla} u_{0,\alpha}\cdot {\nabla} u_{2,\alpha}-\int_{\mathbb B(0,R)} u_{0,\alpha}u_{2,\alpha}-\lambda_\alpha({\mathbb B(0,R)}^*)\int_{{\mathbb B(0,R)}(0,R)} u_{0,\alpha}u_{2,\alpha},
$$
one has
$$
D_3=\int_{{\mathbb S(0,r^*_0)}}\llbracket u_{2,\alpha}\rrbracket \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r},
$$
and hence, by using Equation \eqref{Eq:Jump2}, one has
\begin{align*}
D_3 &=-2\int_{{\mathbb S(0,r^*_0)}}\llbracket u_{2,\alpha}\rrbracket \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\\
&= 4\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\left\llbracket \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket V\cdot \nu +2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\left\llbracket \frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right\rrbracket (V\cdot \nu)^2.
\end{align*}
Similarly, let
$$
D_4=\int_{{\mathbb B(0,R)}(0,R)}\sigma_\alpha |{\nabla} u_{1,\alpha}|^2-\int_{{\mathbb B(0,R)}(0,R)}m^*_0u_{1,\alpha}^2.
$$
By using Eq. \eqref{Eq:EigenDerivative1} and the fact that $\lambda_{1,\alpha}=0$, one has
\begin{eqnarray*}
D_4 &=& \lambda_\alpha({\mathbb B(0,R)}^*)\int_{\mathbb B(0,R)} u_{1,\alpha}^2-\int_{{\mathbb S(0,r^*_0)}}\left\llbracket u_{1,\alpha}\sigma_\alpha \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket
\\&=&\lambda_\alpha({\mathbb B(0,R)}^*)\int_{\mathbb B(0,R)} u_{1,\alpha}^2-\int_{{\mathbb S(0,r^*_0)}}\left\llbracket u_{1,\alpha}\right\rrbracket\left.\left(\sigma_\alpha \frac{\partial u_{1,\alpha}}{\partial r}\right)\right|_{ext}-\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}\left\llbracket \sigma_\alpha \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket
\\&=&\lambda_\alpha({\mathbb B(0,R)}^*)\int_{\mathbb B(0,R)} u_{1,\alpha}^2+\int_{{\mathbb S(0,r^*_0)}} \left(\sigma_\alpha\partial_r u_{1,\alpha}|_{ext}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket+\kappa u_{1,\alpha}|_{int} u_{0,\alpha}\right) V\cdot \nu.
\end{eqnarray*}
Finally, by differentiating the normalization condition $\int_{{\mathbb B(0,R)}(0,R)}u_{\varepsilon}^2=1$, we get
\begin{equation}\label{normal}
\int_{{\mathbb B(0,R)}(0,R)}u_{0,\alpha}u_{2,\alpha}+\int_{{\mathbb B(0,R)}(0,R)}u_{1,\alpha}^2=0.\end{equation}
Combining the equalities above, one gets
\begin{eqnarray*}
\lambda_{2,\alpha}&=&2\lambda_\alpha({\mathbb B(0,R)}^*)\left(\int_{{\mathbb B(0,R)}(0,R)}u_{0,\alpha}u_{2,\alpha}+\int_{{\mathbb B(0,R)}(0,R)}u_{1,\alpha}^2\right) +4\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\left\llbracket \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket V\cdot \nu
\\&&+2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\left\llbracket \frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right\rrbracket (V\cdot \nu)^2+2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \partial_r u_{1,\alpha}|_{ext}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket V\cdot \nu
\\&&+2\kappa \int_{{\mathbb S(0,r^*_0)}} u_{1,\alpha}|_{int} u_{0,\alpha}V\cdot \nu-4\int_{{\mathbb S(0,r^*_0)}}\sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\left\llbracket \frac{\partial u_{1,\alpha}}{\partial r}\right\rrbracket V\cdot \nu
\\&&-4\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu -\int_{{\mathbb S(0,r^*_0)}}\left(\frac1{r^*_0}\left\llbracket \sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right\rrbracket +\frac{\kappa}{r^*_0}u_{0,\alpha}^2\right) (V\cdot \nu)^2
\\&&-2\int_{{\mathbb S(0,r^*_0)}}\left[\sigma_\alpha \frac{\partial u_{0,\alpha}}{\partial r}\frac{\partial^2 u_{0,\alpha}}{\partial r^2}\right] (V\cdot \nu)^2-2\int_{{\mathbb S(0,r^*_0)}}\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2
\\&=&2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \partial_r u_{1,\alpha}|_{ext}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket V\cdot \nu
-2\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu
\\&& -\int_{{\mathbb S(0,r^*_0)}}\left(\frac1{r^*_0}\left[\sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right]+\frac{\kappa}{r^*_0}u_{0,\alpha}^2\right) (V\cdot \nu)^2-2\int_{{\mathbb S(0,r^*_0)}}\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2.
\end{eqnarray*}
We have then obtained the desired expression.
\end{proof}
\paragraph{Strong stability.}
Recall here that, as mentioned before, since we are dealing with a critical point of the functional $\lambda_\alpha$, it is enough to consider perturbation $V$ normal to the boundary of ${\mathbb B(0,R)}^*$, in other words such that $V=(V\cdot \nu)\nu$.
Under such an assumption, the second derivative of the volume is known to be (see e.g. \cite[Section 5.9.6]{HenrotPierre})
\begin{equation}\label{Eq:VolumeDerivative}
\operatorname{Vol}''({\mathbb B(0,R)}^*)[V,V]=\int_{{\mathbb S(0,r^*_0)}}H (V\cdot \nu)^2.
\end{equation}
Hence, introducing $D_5=(\lambda_\alpha''-\eta_\alpha\operatorname{Vol}'')({\mathbb B(0,R)}^*)[V,V]$ and taking into account Lemma \ref{Le:SecondDerivative}, \eqref{LM} and \eqref{Eq:VolumeDerivative}, we have
\begin{eqnarray*}
D_5&=&2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \partial_r u_{1,\alpha}|_{ext}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket V\cdot \nu-2\int_{{\mathbb S(0,r^*_0)}}\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2
\\ &&-2\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{ext}u_{0,\alpha} V\cdot \nu
+\int_{{\mathbb S(0,r^*_0)}}\left(-\frac1{r^*_0}\left[\sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right]-\frac{\kappa}{r^*_0}u_{0,\alpha}^2\right) (V\cdot \nu)^2
\\&&+\kappa \int_{{\mathbb S(0,r^*_0)}}\frac1{r^*_0}u_{0,\alpha}^2 (V\cdot \nu)^2-\int_{{\mathbb S(0,r^*_0)}}\frac1{r^*_0}\left[\sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right] (V\cdot \nu)^2
\\&=&2\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \partial_r u_{1,\alpha}|_{ext}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket V\cdot \nu-2\int_{{\mathbb S(0,r^*_0)}}\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2
\\ &&-2\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu
-\int_{{\mathbb S(0,r^*_0)}}\frac2{r^*_0}\left[\sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right] (V\cdot \nu)^2.
\end{eqnarray*}
We are then led to determine the signature of the quadratic form
\begin{eqnarray}
\mathcal F_\alpha[V,V]&=&\frac12(\lambda_\alpha''-\Lambda_\alpha\operatorname{Vol}'')({\mathbb B(0,R)}^*)[V,V] \label{Eq:Lambda2}\\
&=&\int_{{\mathbb S(0,r^*_0)}} \sigma_\alpha \partial_r u_{1,\alpha}|_{ext}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket V\cdot \nu -\kappa\int_{{\mathbb S(0,r^*_0)}}u_{1,\alpha}|_{int}u_{0,\alpha} V\cdot \nu \nonumber
\\&&+\int_{{\mathbb S(0,r^*_0)}}\left(-\frac2{r^*}\left\llbracket \sigma_\alpha|{\nabla} u_{0,\alpha}|^2\right\rrbracket \right) (V\cdot \nu)^2-\int_{{\mathbb S(0,r^*_0)}}\kappa u_{0,\alpha}\left.\frac{\partial u_{0,\alpha}}{\partial r}\right|_{int} (V\cdot \nu)^2.\nonumber
\end{eqnarray}
\subsection{Analysis of the quadratic form $\mathcal{F}_\alpha$}
\paragraph{Separation of variables and first simplification.}
Each perturbation $g\in L^2(0,2\pi)$ such that $\int_0^{2\pi} g=0$ expands as
$$
g=\sum_{k=1}^\infty \left( \gamma_k \cos(k\cdot)+\beta_k\sin(k\cdot)\right), \quad \text{with }\gamma_0=0.
$$
For every $k\in \textrm{I\kern-0.21emN}^*$, let us introduce $g_k:=\cos(k\cdot)$ and $\tilde g_k:=\sin(k\cdot)$. For any $k\in \textrm{I\kern-0.21emN}^*$, let $u_{1,\alpha}^{(k)}$ be the solution of Eq.~\eqref{Eq:EigenDerivative1} associated with the perturbation $g_k$.
It is readily checked that there exists a function $\varphi_{k,\alpha}:[0,R]\rightarrow \textrm{I\kern-0.21emR}$ such that
$$
\forall (r,\theta)\in [0,R]\times [0,2\pi], \quad {u_{1,\alpha}^{(k)}}(r,\theta)=g_k(\theta)z_{k,\alpha}(r).
$$
Furthermore, $\varphi_{k,\alpha}$ solves the ODE
\begin{equation}\label{Eq:PsiK}
\left\{
\begin{array}{ll}
-\sigma_\alpha z_{k,\alpha}''-\frac{\sigma_{\alpha}}rz_{k,\alpha}'(r) =\left(\lambda_{0,\alpha}-\frac{k^2}{r^2}\right)z_{k,\alpha}+m^*_0z_{k,\alpha}&\text{ in }(0,r_0^*)\cup (r_0^*,R),\\
\left\llbracket \sigma_\alpha z_{k,\alpha}'\right\rrbracket (r^*_0)=-\kappa u_{0,\alpha}(r^*_0)& \\
\left\llbracket z_{k,\alpha}\right\rrbracket (r^*_0)=-\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket(r^*_0),
\\z_{k,\alpha}'(0)=z_{k,\alpha}(R)=0.&
\end{array}
\right.
\end{equation}
Regarding $\tilde g_k$, if we define $\tilde u_{1,\alpha}^{(k)}$ in a similar fashion, it is readily checked that
$$\forall (r,\theta)\in [0,R]\times [0,2\pi]\, ,\tilde u_{1,\alpha}^{(k)}(r,\theta)=\tilde g_k(\theta) z_{k,\alpha}(r).$$
Therefore, any admissible perturbation $g$ writes
$$
g=\sum_{k=1}^\infty \left\{\gamma_kg_k+\beta_k\tilde g_k\right\}\quad \text{with } \gamma_0=0,
$$
and the solution $u_{1,\alpha}$ associated with $g$ writes
$$u_{1,\alpha}=\sum_{k=1}^\infty \left\{\gamma_k {u_{1,\alpha}^{(k)}}+\beta_k \tilde u_{1,\alpha}^{(k)}\right\}.$$
Using the orthogonality properties of the family $\{g_k\}_{k\in \textrm{I\kern-0.21emN}^*}\cup \{\tilde g_k\}_{k\in \textrm{I\kern-0.21emN}}$, it follows that $\mathcal F_\alpha[V,V]$ given by \eqref{Eq:Lambda2} reads
\begin{eqnarray}
\mathcal F_\alpha[V,V] &=&\frac{r^*_0}2\sum_{k=1}^\infty\left(\sigma_\alpha z_{k,\alpha}'(r^*_0)|_{ext}\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket -\kappa z_{k,\alpha}|_{int}u_{0,\alpha}(r^*_0)\right)\left(\gamma_k^2+\beta_k^2\right)\nonumber
\\&&-\frac{r^*_0}2\sum_{k=0}^\infty \kappa u_{0,\alpha}(r^*_0)\frac{\partial u_{0,\alpha}}{\partial r}(r^*_0)\left(\gamma_k^2+\beta_k^2\right)-\sum_{k=1}^\infty2 \left\llbracket \sigma_\alpha |{\nabla} u_{0,\alpha}|^2\right\rrbracket \left(\gamma_k^2+\beta_k^2\right)\nonumber
\\&=&\frac{r^*_0\kappa u_{0,\alpha}(r^*_0)}2\sum_{k=1}^\infty\left(-\left.\frac{\partial u_{0,\alpha}}{\partial r}(r^*_0)\right|_{int}- z_{k,\alpha}|_{int}\right)\left(\gamma_k^2+\beta_k^2\right)\nonumber
\\&&+\sum_{k=1}^\infty \left(-2\left[\sigma_\alpha |{\nabla} u_{0,\alpha}|^2\right]+\sigma_\alpha z_{k,\alpha}'(r^*_0)|_{ext}\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket \right)\left(\gamma_k^2+\beta_k^2\right).\label{Eq:Ptn}
\end{eqnarray}
Define, for any $k\in \textrm{I\kern-0.21emN}$,
$$
\omega_{k,\alpha}:=\frac{r_0^*\kappa u_{0,\alpha}(r_0^*)}{2}\left(\left.-\frac{\partial u_{0,\alpha}}{\partial r}(r^*_0)\right|_{int}- z_{k,\alpha}|_{int}(r_0^*)\right)$$ and $$ \zeta_{k,\alpha}:=-2\left\llbracket \sigma_\alpha |{\nabla} u_{0,\alpha}|^2\right\rrbracket +\sigma_\alpha z_{k,\alpha}'(r^*_0)|_{ext}\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket .
$$
Thus,
$$
\mathcal F_\alpha[V,V]=\sum_{k=1}^\infty\left(\omega_{k,\alpha}+\zeta_{k,\alpha}\right)\left(\gamma_k^2+\beta_k^2\right).
$$
The end of the proof is devoted to proving the local shape minimality of the centered ball, which relies on an asymptotic analysis of the sequences $\{\omega_{k,\alpha}\}_{k\in \textrm{I\kern-0.21emN}}$ and $\{\zeta_{k,\alpha}\}_{k\in \textrm{I\kern-0.21emN}}$ as $\alpha$ converges to 0.
\begin{proposition}\label{Le:AAO}
There exists $C>0$ and $\overline \alpha>0$, there exists $M\in \textrm{I\kern-0.21emR}$ such that for any $\alpha \leq \overline \alpha$ and any $k\in \textrm{I\kern-0.21emN}$, one has
\begin{equation}\label{Eq:OmegaK}
\omega_{k,\alpha}\geq C>0,
\quad \text{and}\quad
\zeta_{k,\alpha}\geq - M\alpha.
\end{equation}
\end{proposition}
The last claim of Theorem \ref{Th:ShapeStability} is then an easy consequence of this proposition. The rest of the proof is devoted to the proof of Proposition~\ref{Le:AAO}, which follows from the combination of the following series of lemmas.
\begin{lemma}\label{Le:Positivity}
There exists $\overline \alpha>0$ such that, for every $\alpha\in [0, \overline \alpha]$, $z_{1,\alpha}$ is nonnegative on $(0,R)$.
\end{lemma}
\begin{proof}[Proof of Lemma~\ref{Le:Positivity}]
For the sake of notational simplicity, we temporarily drop the dependence on $\alpha$ and denote $z_{1,\alpha}$ by $z_\alpha$. The function $z_\alpha$ solves the ODE
$$
\left\{
\begin{array}{ll}
-\sigma_\alpha z_\alpha''-\frac{\sigma_{\alpha}}rz_\alpha'(r) =\left(\lambda_{0,\alpha}-\frac{1}{r^2}\right)z_\alpha+m^*z_\alpha&\text{ in }(0,r_0^*)\cup (r_0^*,R),
\\\left\llbracket \sigma_\alpha z_\alpha'\right\rrbracket (r^*_0)=-\kappa u_{0,\alpha}(r^*_0)&
\\\left\llbracket z_\alpha\right\rrbracket (r^*_0)=-\left\llbracket \frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket (r^*_0),
\\z_{\alpha}(R)=0.&
\end{array}
\right.
$$
Let us introduce $p_\alpha=z_\alpha/u_{0,\alpha}$. One checks easily that $p_\alpha$ solves the ODE
$$
-\sigma_\alpha p_\alpha''-\frac{\sigma_{\alpha}}rp_\alpha' =-\frac{1}{r^2}p_\alpha-2p_\alpha'\frac{u_{0,\alpha}'}{u_{0,\alpha}}\quad \text{in }(0,R).
$$
Furthermore, $p_\alpha$ satisfies the jump conditions
$$\llbracket p_\alpha\rrbracket (r^*_0)=-\frac{\left\llbracket \partial_r u_{0,\alpha}\right\rrbracket (r^*_0)}{u_{0,\alpha}(r^*_0)} =\frac{-\alpha \kappa \partial_r u_{0,\alpha}|_{int}}{u_{0,\alpha}(r^*_0)}>0
\quad \text{and}\quad
\llbracket \sigma_\alpha p_\alpha'\rrbracket (r^*_0)=-\kappa+\frac{\sigma_\alpha \partial_r u_{0,\alpha}}{u_{0,\alpha}(r^*_0)^2}\left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket.$$
To show that $z_\alpha$ is nonnegative, we argue by contradiction and consider first the case where a negative minimum is reached at an interior point $r_-\neq r^*_0$. Then, $p_\alpha$ is $\mathscr C^2$ in a neighborhood of $r_-$ and we have
$$0\geq -p_\alpha''(r_-)=-\frac{p_\alpha(r_-)}{\sigma_\alpha r_-^2}>0,$$
whence the contradiction.
To exclude the case $r_-=R$, let us notice that, according to L'Hospital's rule, one has $p_\alpha(R)=z_\alpha'(R)/u_{0,\alpha}'(R)$. According to the Hopf lemma applied to $u_{0,\alpha}$, this quotient is well-defined. If $p_\alpha(R)<0$ then it follows that $z_\alpha'(R)>0$. However, one has $p_\alpha'(r)\sim z_\alpha'(r)/(2u_{0,\alpha}(r))>0$ as $r\to R$, which contradicts the fact that a minimum is reached at $R$.
Let us finally exclude the case where $r_-=r^*_0$. Mimicking the elliptic regularity arguments used in the proofs of Lemmas \ref{Le:Bounds} and \ref{Le:ConvergenceFP}, we get that $p_\alpha$ converges to $p_0$ as $\alpha \to 0$ for the strong topologies of $\mathscr C^0([0,r^*_0])$ and $\mathscr C^0([r^*_0,R])$.
To conclude, it suffices hence to prove that $p_0$ is positive in a neighborhood of $r^*_0$. We once again argue by contradiction and assume that $p_0$ reaches a negative minimum at $r_-\in [0,R]$.
Notice that $r_-\neq r_0^*$ since $\llbracket p_0\rrbracket (r^*_0)=0$ and $\llbracket p_0'\rrbracket (r^*_0)=-\kappa<0$.
If $r_-\in (0,R)$, since $r_-\neq r_0^*$, we claim that $p_0$ is $\mathscr C^2$ in a neighborhood of $r_-$ and, if $p_0(r_-)<0$, the contradiction follows from
$$0\geq -p_0''(r_-)=-\frac{p_0(r_-)}{(r_-)^2}>0.$$ For the same reason, a negative minimum cannot be reached at $r=0$.
If $r_-=R$, we observe that $p_0(R)=z_0'(R)/u_{0,0}'(R)$.
According to the Hopf lemma applied to $u_{0,0}$, this quantity is well-defined. If $p_0(R)<0$, then it follows that $z_0'(R)>0$. However, $p_0'(r)\sim z_0'(r)/(2u(r))>0$ as $r\to R$, which contradicts the fact that $R$ is a minimizer.
Therefore $p_0$ is positive in a neighborhood of $r^*_0$ and we infer that $p_\alpha$ is non-negative, so that, in turn, $z_\alpha \geq 0$ in $[0,R]$.
\end{proof}
\begin{lemma}\label{Cl:Comp}
Let $\overline \alpha$ be defined as in Lemma \ref{Le:Positivity}. Then, for every $ \alpha\in [0, \overline \alpha]$ and every $k\in \textrm{I\kern-0.21emN}$,
\begin{equation}\label{Eq:CompPs}z_{k,\alpha}\leq z_{1,\alpha}.\end{equation}
As a consequence, for any $\alpha \leq \overline \alpha$ and any $k\in \textrm{I\kern-0.21emN}$, there holds $\omega_{k,\alpha}\geq \omega_{1,\alpha}$.
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Cl:Comp}]
Since $\omega_{k,\alpha}-\omega_{1,\alpha}=\frac{r_0^*\kappa u_{0,\alpha}(r_0^*)}{2}\left(- z_{k,\alpha}|_{int}(r^*_0)+z_{1,\alpha}|_{int}(r^*_0)\right)$, and since we further have $\frac{r_0^*\kappa u_{0,\alpha}(r_0^*)}{2}>0$,
the fact that $\omega_{k,\alpha}\geq \omega_{1,\alpha}$ will follow from \eqref{Eq:CompPs}, on which we now focus. Let us set $\Psi_k=z_{1,\alpha}-z_{k,\alpha}$. From the jump conditions on $z_{1,\alpha}$ and $z_{k,\alpha}$, one has $\llbracket \Psi_k\rrbracket (r_0)=\llbracket \sigma_\alpha\Psi_k'\rrbracket (r_0)=0$.
The function $\Psi_k$ satisfies
\begin{eqnarray*}
-\sigma_\alpha\Psi_k''-\sigma_\alpha\frac{\Psi_k'}r&=&-\left(\lambda_{0,\alpha}-\frac{k^2}{r^2}\right)z_{k,\alpha}-m^*_0z_{k,\alpha}+\left(\lambda_{0,\alpha}-\frac1{r^2}\right)z_{1,\alpha}+m^*_0\psi_{1,\alpha}
\\&>&\left(\lambda_{0,\alpha}-\frac{k^2}{r^2}\right)z_{k,\alpha}-m^*_0z_{k,\alpha}+\left(\lambda_{0,\alpha}-\frac{k^2}{r^2}\right)z_{1,\alpha}+m^*_0\psi_{1,\alpha}
\\&>&\left(\lambda_{0,\alpha}-\frac{k^2}{r^2}\right)\Psi_{k}+m^*_0\Psi_{k}.
\end{eqnarray*}
since $z_{1,\alpha}\geq0$, according to Lemma~\ref{Le:Positivity}. Since $\Psi_k$ satisfies Dirichlet boundary conditions, $\Psi_k\geq0$ in $(0,R)$.
\end{proof}
\begin{lemma}\label{Cl:Final}
There exists $C>0$ such that, for every $\alpha \in [0, \overline \alpha]$, where $\overline \alpha$ is introduced on Lemma~\eqref{Le:Positivity}, one has $\omega_{1,\alpha}\geq C$.
\end{lemma}
\begin{proof}[Proof of Lemma \ref{Cl:Final}]
Let us introduce $\Psi=-{\partial u_{0,\alpha}}/{\partial r}- z_{1,\alpha}$. Since
$$\omega_{1,\alpha}=\frac{r_0^*\kappa u_{0,\alpha}(r_0^*)}{2}\Psi(r_0^*)$$ and since $\frac{r_0^*\kappa u_{0,\alpha}(r_0^*)}{2}$ converges, as $\alpha \to0$, to $\frac{r_0^*\kappa u_{0,0}(r_0^*)}{2}>0$, it suffices to prove that $\Psi(r_0^*)\geq C>0$ for some $C$ when $\alpha \to 0$.
According to \eqref{Eq:RadU0}, we have $\llbracket \Psi\rrbracket (r^*_0)=\llbracket \Psi'\rrbracket (r^*_0)=0$. Furthermore,
$\Psi(R)=-\frac{\partial u_{0,\alpha}}{\partial r}(R)>0$ according to the Hopf Lemma and $\Psi(0)=0$. Finally, since $\Psi$ solves the ODE, one has
$$
-\frac1r(\sigma_\alpha \Psi')'=\left(\lambda_{0,\alpha}-\frac1{r^2}\right)\Psi+m^*\Psi\quad \text{in }(0,R),
$$
it follows that $\Psi$ is positive in $(0,R]$. Furthermore, $\Psi$ converges to $\Psi_0$ for the strong topology of $\mathscr C^0([0,R])$ and $\Psi_0$ solves the ODE
$$
\left\{\begin{array}{ll}
-\frac1r( \Psi_0')'=\left(\lambda_{0,0}-\frac1{r^2}\right)\Psi_0+m^*_0\Psi_0 & \text{in }(0,R)\\
\Psi_0(R)=-\frac{\partial u_{0,0}}{\partial r}(R)>0. &
\end{array}\right.
$$
Hence there exists $C>0$ such that, for every $\alpha \in [0, \overline \alpha]$, one has $\Psi(r^*_0)\geq C>0$.
\end{proof}
It remains to prove the second inequality of\eqref{Eq:OmegaK}.
As a consequence of the convergence result stated in Lemma~\ref{Le:Rad}, one has
\begin{equation}\label{bus:1904}
\left\llbracket \sigma_\alpha |{\nabla} u_{0,\alpha}|^2\right\rrbracket =\operatorname{O}(\alpha), \quad \left\llbracket\frac{\partial u_{0,\alpha}}{\partial r}\right\rrbracket=\left.\alpha \kappa \frac{\partial u_{0,\alpha}}{\partial r}\right|_{int}<0.
\end{equation}
It follows that we only need to prove that there exists a constant $M>0$ such that, for any $\alpha\in [0, \overline \alpha]$, and any $k\in \textrm{I\kern-0.21emN}^*$,
\begin{equation}\label{Eq:Tain3}
M\geq \sigma_\alpha z_{k,\alpha}'|_{ext}(r^*_0)
\end{equation}
so that
$$
\zeta_{k,\alpha}=\operatorname{O}(\alpha)+\sigma_\alpha z_{k,\alpha}'\vert_{ext}(r^*_0)\left.\alpha \kappa \frac{\partial u_{0,\alpha}}{\partial r}(r^*_0)\right|_{int}\geq \operatorname{O}(\alpha)-M\alpha\kappa\left|\left.\frac{\partial u_{0,\alpha}}{\partial r}(r^*_0)\right|_{int}\right|
$$
To show the estimate \eqref{Eq:Tain3}, let us distinguish between small and large values of $k$.
To this aim, we introduce $N\in \textrm{I\kern-0.21emN}$ as he smallest integer such that
\begin{equation}\label{tain}
\lambda_{0,\alpha}+m^*_0-\frac{k^2}{r^2}<0 \text{ in }(0,R)
\end{equation}
for every $k\geq N$ and $\alpha \in [0, \overline \alpha]$. The existence of such an integer follows immediately from the convergence of $(\lambda_{0,\alpha})_{\alpha>0}$ to $\lambda_0(m^*_0)$ as $\alpha\to 0$.
\medskip
First, we will prove that, for every $k\geq N$,
\begin{equation}\label{Eq:DerPos}
z_{k,\alpha}'(r^*_0)|_{ext}<0
\end{equation}
and that there exists $M>0$ such that, for every $k\leq N$,
\begin{equation}\label{Eq:DerPos2}
|z_{k,\alpha}'(r^*_0)|_{ext}|\leq M
\end{equation}
which will lead to \eqref{Eq:Tain3} and thus yield the desired conclusion.
To show \eqref{Eq:DerPos}, let us argue by contradiction, assuming that $z_{k,\alpha}'(r^*_0)|_{ext}>0$.
Since the jump $\llbracket \sigma_\alpha z_{k,\alpha}'\rrbracket =-\kappa u_{0,\alpha}(r^*_0)$ is negative, it follows that
$$
(1+\alpha \kappa)z_{k,\alpha}'(r^*_0)|_{int}=z_{k,\alpha}'(r^*_0)|_{ext}-\llbracket \sigma_\alpha z_{k,\alpha}'\rrbracket >0.
$$
By mimicking the reasonings in the proof of Lemma~\ref{Le:Positivity}, $z_{k,\alpha}$ cannot reach a negative minimum on $(0,r^*_0)$ since \eqref{tain} holds true. Therefore, since $z_{k,\alpha}(0)=0$ and $z_{k,\alpha}'(r^*_0)|_{int}>0$, one has necessarily $z_{k,\alpha}(r^*_0)|_{int}>0$, which in turn gives
$z_{k,\alpha}(r^*_0)|_{ext}>0$ since $\llbracket z_{k,\alpha}\rrbracket =-\alpha \kappa \frac{\partial u_{0,\alpha}}{\partial r}>0$.
Furthermore, $z_{k,\alpha}(R)=0$. Since $z_{k,\alpha}(r^*_0)|_{ext}>0$ and $z_{k,\alpha}'(r^*_0)|_{ext}>0$, it follows that $z_{k,\alpha}$ reaches a positive maximum at some interior point $r_1$, satisfying hence
$$
0\leq -z_{k,\alpha}''(r_-)=\left(\lambda_{0,\alpha}+m^*-\frac{k^2}{r^2}\right)z_{k,\alpha}(r_-)<0,
$$
leading to a contradiction.
\medskip
Let us now deal with small values of $k$, by assuming $k\leq N$. We will prove that \eqref{Eq:DerPos2} holds true.
To this aim, we will compute $z_{k,\alpha}$. Let $J_k$ (resp. $Y_k$) be the $k$-th Bessel function of the first (resp. the second) kind. One has
$$z_{k,\alpha}(r)=\left\{\begin{array}{ll}
A_{k,\alpha}J_k(\sqrt{\frac{\lambda_{0,\alpha}+\kappa}{1+\alpha\kappa}} \frac{r}R) & \text{ if }r\leq r^*_0, \\
B_{k,\alpha}J_k(\sqrt{\lambda_{0,\alpha}}\frac{r}R)+C_{k,\alpha} Y_k(\sqrt{\lambda_{0,\alpha}}\frac{r}R) & \text{ if }r^*_0\leq r\leq R,
\end{array}\right.$$
where $X_{k,\alpha}=(B_{k,\alpha},C_{k,\alpha},A_{k,\alpha})$ solves the linear system
$$
\mathcal A_{k,\alpha}X_{k,\alpha}=b_\alpha
$$
where
$$
b_\alpha=\begin{pmatrix}0\\-\kappa u_{0,\alpha}(r^*_0)\\-\left[\partial_R u_{0,\alpha}\right]\\\end{pmatrix}
$$
and
$$
\mathcal A_{k,\alpha}=\begin{pmatrix}
J_k\left(\sqrt{\lambda_0}\right)&Y_k\left(\sqrt{\lambda_0}\right)&0
\\\sqrt{\lambda_{0,\alpha}}J_k'(\sqrt{\lambda_{0,\alpha}+\kappa}{\frac{r^*_0}R})&\sqrt{\lambda_{0,\alpha}}Y_k'(\sqrt{\lambda_{0,\alpha}+\kappa}{\frac{r^*_0}R})&-\sqrt{\frac{\lambda_{0,\alpha}+\kappa}{1+\alpha\kappa}}J_k'(\sqrt{\frac{\lambda_{0,\alpha}+\kappa}{1+\alpha\kappa}}{\frac{r^*_0}R})
\\J_k(\sqrt{\lambda_{0,\alpha}}{\frac{r^*_0}R})&Y_k(\sqrt{\lambda_{0,\alpha}}{\frac{r^*_0}R})&-J_k(\sqrt{\frac{\lambda_{0,\alpha}+\kappa}{1+\alpha\kappa}}{\frac{r^*_0}R})
\end{pmatrix}.
$$
It is easy to check that
$$
\Vert \mathcal A_{k,\alpha}-\mathcal A_{k,0}\Vert \leq M\alpha
$$
where $M$ only depends\footnote{Indeed, $\{J_k,Y_k\}_{k\leq N}$ are uniformly bounded in $\mathscr C^2([r_0^*/R-{\varepsilon},R])$ for every ${\varepsilon}>0$ small enough. Since we consider a finite number of indices $k$, there exists $\delta>0$ (depending only on $N$) such that
$$
\forall k\in \{0, \dots,N\}, \quad \det(\mathcal A_{k,\alpha})\geq \delta>0.
$$
Then, since $\Vert X_\alpha-X_0\Vert \leq M\alpha$, it follows from the Cramer formula that there exists $M$ (depending only on $N$) such that
$\Vert
X_{k,\alpha}-X_{k,0}\Vert_{L^\infty}\leq M\alpha.
$
} on $N$.
Hence it is enough to prove that
$|\psi_{k,0}'(r^*_0)|\leq M$ for some $M>0$ depending only on $N$, which is straightforward since the set of indices is finite. The expected conclusion follows.
\subsection{Conclusion}
From Eq. \eqref{Eq:OmegaK} and Lemma \ref{Cl:Final}, there exists $C>0$ and $M>0$ such that $\omega_{k,\alpha}\geq C>0$ and $\zeta_{k,\alpha}\geq -M\alpha$ for every $\alpha\in [0, \overline \alpha]$ and $k\in \textrm{I\kern-0.21emN}$, from which we infer that
\begin{align*}
\mathcal F_\alpha[V,V]&\geq \left(C-M\alpha\right)\sum_{k=1}^\infty \left(\gamma_k^2+\beta_k^2\right)\geq \frac{C}2 \Vert V\cdot \nu \Vert_{L^2}^2.
\end{align*}
according to Eq. \eqref{Eq:Ptn}.
\subsection{Concluding remark: possible extension to higher dimensions}\label{Se:CclSha}
Let us briefly comment on possible extensions of this method to higher dimensions. Indeed, although we do not tackle this issue in this article, we believe that the coercivity norm obtained in Theorem~\ref{Th:ShapeStability} could also be obtained in the three-dimensional case. Nevertheless, we believe that such an extension would need tedious and technical computations.
Since our objective here was to introduce a methodology to investigate stability issues for the shape optimization problems we deal with, we slightly comment on this claim and explain how we believe that our proof can be adapted to the case $d=3$.
Let ${\Omega}$ denote the ball $\mathbb B(0,R)$ in $\textrm{I\kern-0.21emR}^3$ and ${\mathbb B(0,R)}^*$ be the centered three-dimensional ball $\mathbb B(0,R)$ of volume $m_0|{\Omega}|/\kappa$. Let us assume without loss of generality that $R=1$, so that $\partial {\mathbb B(0,R)}^*$ is the euclidean unit sphere $\mathbb S^2$.
As a preliminary result, one first has to show that the principal eigenfunction $u_{\alpha,m^*_0}$ is radially symmetric and that ${\mathbb B(0,R)}^*$ is a critical shape by the same arguments as in the proof of Theorem~\ref{Th:ShapeStability}, which allows us to compute the Lagrange multiplier $\Lambda_\rho$ associated to the volume constraint. Let $\mathcal L_{\Lambda_\rho}$ be the associated shape Lagrangian.
For an integer $k$, we define $H_k$ as the space of spherical harmonics of degree $k$ i.e as the eigenspace associated with the eigenvalue $-k(k+1)$ of the Laplace-Beltrami operator $\Delta_{\mathbb S^2}$. $H_k$ has finite dimension $d_k$, and we furthermore have
$$L^2(\mathbb S^2)=\bigoplus_{k=1}^\infty H_k.$$
Let us consider a Hilbert basis $\{y_{k,\ell}\}_{\ell=1,\dots,d_k}$ of $H_k$.
For an admissible vector field $V,$ one must then expand $ V\cdot \nu$ in the basis of spherical harmonics as
\begin{equation}\label{eq:fin}
V\cdot \nu=\sum_{k=1}^\infty\sum_{\ell=1}^{d_k} \alpha_{k,\ell}(V\cdot \nu) y_{k,\ell}.\end{equation}
Then, one has to diagonalise the second-order shape derivative of $\mathcal L_{\Lambda_\rho}$ and prove that there exists a sequence of coefficients $\{\omega_{k,\ell,\rho}\}_{k\in \textrm{I\kern-0.21emN}\,, 0\leq \ell\leq d_k}$ such that for every $V\cdot \nu$ expanding as \eqref{eq:fin}, the second order derivative of the shape Lagrangian in direction $V$ reads
$$
\mathcal L_{\Lambda_\rho}''=\sum_{k=1}^\infty\sum_{\ell=1}^{d_k}\left(\alpha_{k,\ell}(V\cdot\nu)\right)^2 \omega_{k,\ell}.
$$
We believe this diagonalization can be proved using separation of variables and the orthogonality properties of the family $\{y_{k,\ell}\}_{k\in \textrm{I\kern-0.21emN}^*,\ell=1,\dots,d_k}$.
Using the separation of variables, each coefficient $\omega_{k,\ell}$ can be written in terms of derivatives of a family of solutions of one dimensional differential equations. The main difference with the proof of Theorem~\ref{Th:ShapeStability} comes from the fact that the main part of the ODE is not $-\frac1r\frac{d}{dr}(r(1+\alpha m_0^*)\frac{d}{dr})$ anymore, but $-\frac1{r^2}\frac{d}{dr}(r^2(1+\alpha m_0^*)\frac{d}{dr})$. The important fact is that maximum principle arguments may still be used to analyze the diagonalized expression of $\mathcal L_{\Lambda_\rho}''$ and to obtain a uniform bound from below for the sequence $\{\alpha_{k,\ell}\}_{k\in \textrm{I\kern-0.21emN}^*\,, \ell=1,\dots,d_k}$.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 3,353
|
\subsection{Derivation of nuclear spin equations of motion}
The starting point is
\begin{align}
\dot{\rho}^{(\mathbf{m},\mathbf{n})} & =\mathcal{L}_{\mathbf{m},\mathbf{n
}\hat{\rho}^{(\mathbf{m},\mathbf{n})}-i\frac{\{\hat{\rho}^{(\mathbf{m
,\mathbf{n})},\hat{K}^{(\mathbf{m},\mathbf{n})}\}}{2}-i\langle\mathbf{m
|[\hat{V},\hat{\rho}]|\mathbf{n}\rangle,\label{EOM_RHO}\\
\dot{p}^{(\mathbf{m},\mathbf{n})} & =-i\operatorname*{Tr}\nolimits_{e
\frac{\{\hat{\rho}^{(\mathbf{m},\mathbf{n})},\hat{K}^{(\mathbf{m},\mathbf{n
)}\}}{2}-i\operatorname*{Tr}\nolimits_{e}\langle\mathbf{m}|[\hat{V},\hat{\rho
}]|\mathbf{n}\rangle. \label{EOM_P
\end{align}
On the coarse grained time scale $T_{e}\ll\Delta t\ll T_{\mathrm{coh}
,T_{1},T_{2}$, the trace $p^{(\mathbf{m},\mathbf{n})}\equiv\operatorname*{Tr
_{e}\hat{\rho}^{(\mathbf{m},\mathbf{n})}$ is slowly varying, while other
elements of $\hat{\rho}^{(\mathbf{m},\mathbf{n})}$ almost instantaneously
follows the trace $p^{(\mathbf{m},\mathbf{n})}$. Thus we treat the nuclear
spin density matrix elements $\{p^{(\mathbf{m},\mathbf{n})}\}$ (zeroth-order
quantities) as slow variables and other variables as fast variables and
identify $\hat{K}_{\mathbf{m},\mathbf{n}}$ and $\hat{V}$ as first-order small
quantities. Correspondingly, we decompose $\hat{\rho}=\hat{\rho}_{0}+\hat
{\rho}_{1}+\cdots$, where $\hat{\rho}_{k}$ is a $k$th-order small quantity.
Note that the trace of $\{\hat{\rho}_{0}\}$ is the slow variable $\hat
{p}=\operatorname*{Tr}_{e}\hat{\rho}_{0}$, while other elements of $\hat{\rho
}_{0}$ are fast variables. For consistency, for $k=1,2,\cdots$, we require
$\operatorname*{Tr}\hat{\rho}_{k}=0$, so $\hat{\rho}_{k}$ are fast variables.
Since we always work in the nuclear spin interaction picture, any nuclear spin
evolution is caused by the HFI with the electron. Thus the zeroth-order
dynamics vanishes: $(\dot{p})_{0}=0$.
\subsubsection{First-order dynamics}
The first-order evolution $(\dot{p}^{(\mathbf{m},\mathbf{n}})_{1}$ of the slow
variable $p^{(\mathbf{m},\mathbf{n})}=\langle\mathbf{m}|\hat{p}|\mathbf{n
\rangle$ is obtained from Eq. (\ref{EOM_P}) by replacing $\hat{\rho}(t)$ with
$\hat{\rho}_{0}(t)$, i.e., the solution to the zeroth-order version of Eq.
(\ref{EOM_RHO})
\[
\dot{\rho}_{0}^{(\mathbf{m},\mathbf{n})}=\mathcal{L}_{\mathbf{m},\mathbf{n
}\hat{\rho}_{0}^{(\mathbf{m},\mathbf{n})}.
\]
The trace of this equation gives $\dot{p}^{(\mathbf{m},\mathbf{n})}=0$, i.e.,
we solve for the fast variables contained in $\hat{\rho}_{0}^{(\mathbf{m
,\mathbf{n})}$ as functions of the slow variables $p^{(\mathbf{m},\mathbf{n
)}$, which are regarded as fixed. Coarse-graining for an interval $T_{e
\ll\Delta t\ll T_{\mathrm{coh}},T_{1},T_{2}$ or equivalently calculating the
steady-state solution gives $\hat{\rho}_{0}^{(\mathbf{m},\mathbf{n})
(t)=\hat{P}_{\mathbf{m,n}}p^{(\mathbf{m,n})}(t)$, where $\hat{P
_{\mathbf{m},\mathbf{n}}=\hat{P}_{\mathbf{n},\mathbf{m}}$ is the
time-independent electron steady state as determined by $\mathcal{L
_{\mathbf{m,n}}\hat{P}_{\mathbf{m,n}}=0$ and $\operatorname*{Tr}_{e}\hat
{P}_{\mathbf{m},\mathbf{n}}=1$. By replacing $\hat{\rho}(t)$ with $\hat{\rho
}_{0}(t)=\sum_{\mathbf{m},\mathbf{n}}|\mathbf{m}\rangle\langle\mathbf{n
|\hat{\rho}_{0}^{(\mathbf{m},\mathbf{n})}(t)$ in Eq. (\ref{EOM_P}), we obtain
$(\dot{p}^{(\mathbf{m},\mathbf{n}})_{1}$ as given by Eq. (1) of the main text.
\subsubsection{Second-order dynamics}
The second-order evolution $(\dot{p}^{(\mathbf{m},\mathbf{n}})_{2}$ of the
slow variable $p^{(\mathbf{m},\mathbf{n})}$ is obtained from Eq. (\ref{EOM_P})
by replacing $\hat{\rho}(t)$ with $\hat{\rho}_{1}(t)$, the solution to the
first-order version of Eq. (\ref{EOM_RHO})
\[
(\dot{\rho}_{0}^{(\mathbf{m},\mathbf{n})})_{1}+\dot{\rho}_{1}^{(\mathbf{m
,\mathbf{n})}=\mathcal{L}_{\mathbf{m},\mathbf{n}}\hat{\rho}_{1}^{(\mathbf{m
,\mathbf{n})}-i\frac{\{\hat{\rho}_{0}^{(\mathbf{m},\mathbf{n})},\hat
{K}^{(\mathbf{m},\mathbf{n})}\}}{2}-i\langle\mathbf{m}|[\hat{V},\hat{\rho
_{0}]\mathbf{n}\rangle,
\]
where $(\dot{\rho}_{0}^{(\mathbf{m},\mathbf{n})})_{1}=P_{\mathbf{m
,\mathbf{n}}(\dot{p}^{(\mathbf{m},\mathbf{n})})_{1}$ comes from the
first-order evolution $(\dot{p}^{(\mathbf{m},\mathbf{n})})_{1}$. Substituting
into the above equation and coarse-graining for an interval $1/\gamma_{e
\ll\Delta t\ll T_{\mathrm{coh}},T_{1},T_{2}$ (or equivalently calculating the
steady-state solution) give
\begin{align*}
\hat{\rho}_{1}^{(\mathbf{m},\mathbf{n})}=ip^{(\mathbf{m},\mathbf{n
)}\mathcal{L}_{\mathbf{m},\mathbf{n}}^{-1}\frac{\{P_{\mathbf{m},\mathbf{n
},\tilde{K}^{(\mathbf{m},\mathbf{n})}\}}{2} & +i\sum_{\mathbf{p
}(\mathcal{L}_{\mathbf{m},\mathbf{n}}+i\omega_{\mathbf{m},\mathbf{p}
)^{-1}(\hat{V}^{(\mathbf{m},\mathbf{p})}P_{\mathbf{p},\mathbf{n}}-\langle
\hat{V}^{(\mathbf{m},\mathbf{p})}\rangle_{\mathbf{p},\mathbf{n}
P_{\mathbf{m},\mathbf{n}})p^{(\mathbf{p},\mathbf{n})}\\
& -i\sum_{\mathbf{p}}(\mathcal{L}_{\mathbf{m},\mathbf{n}}+i\omega
_{\mathbf{p},\mathbf{n}})^{-1}(P_{\mathbf{m,p}}\hat{V}^{(\mathbf{p
,\mathbf{n})}-\langle\hat{V}^{(\mathbf{p},\mathbf{n})}\rangle_{\mathbf{m
,\mathbf{p}}P_{\mathbf{m},\mathbf{n}})p^{(\mathbf{m,p})
\end{align*}
as a function of the slow variables $\{p^{(\mathbf{m},\mathbf{n})}\}$. Here
$\mathcal{L}_{\mathbf{m},\mathbf{n}}^{-1}\equiv-\int_{0}^{\infty
}e^{\mathcal{L}_{\mathbf{m},\mathbf{n}}t}dt$ and $(\mathcal{L}_{\mathbf{m
,\mathbf{n}}+i\omega_{\mathbf{m},\mathbf{p}})^{-1}\equiv-\int_{0}^{\infty
}e^{(\mathcal{L}_{\mathbf{m},\mathbf{n}}+i\omega_{\mathbf{m},\mathbf{p}})t
dt$, and we have assumed that $\hat{V}^{(\mathbf{m},\mathbf{n})}(t)=\hat
{V}^{(\mathbf{m},\mathbf{n})}e^{-i\omega_{\mathbf{m},\mathbf{n}}t}$ oscillates
at a single frequency $\omega_{\mathbf{m},\mathbf{n}}$. We can verify the
consistent condition $\operatorname*{Tr}_{e}\hat{\rho}_{1}=0$. Then replacing
$\hat{\rho}(t)$ with $\hat{\rho}_{1}(t)$ in Eq. (\ref{EOM_P}) gives the
desired second-order nuclear spin evolution.
For $\mathbf{m}=\mathbf{n}$, neglecting the coupling of the nuclear spin
population $p^{(\mathbf{m},\mathbf{m})}$ to nuclear spin coherences
$p^{(\mathbf{p},\mathbf{q})}$ $(\mathbf{p}\neq\mathbf{q})$, which amounts to
neglecting the small second-order corrections to the transverse mean field
$\langle\hat{V}(t)\rangle$ and electron-mediated nuclear spin interactions
(they induce nuclear spin diffusion and depolarization, which will be included
phenomenologically at the end of the derivation), we obtain the nuclear spin
relaxation Eq. (2) in the main text. The transition rate from $|\mathbf{m
\rangle$ to $|\mathbf{p\rangle}$ induced by $\hat{V}(t)$ i
\[
W_{\mathbf{p}\leftarrow\mathbf{m}}\equiv-2\operatorname{Re}\operatorname*{Tr
\nolimits_{e}\hat{V}^{(\mathbf{m},\mathbf{p})}(\mathcal{L}_{\mathbf{p,m
}+i\omega_{\mathbf{p,m}})^{-1}(\hat{V}^{(\mathbf{p},\mathbf{m})}\hat
{P}_{\mathbf{m},\mathbf{m}}-\langle\hat{V}^{(\mathbf{p},\mathbf{m})
\rangle_{\mathbf{m},\mathbf{m}}\hat{P}_{\mathbf{p,m}}).
\]
Using $(\mathcal{L}_{\mathbf{p,m}}+i\omega_{\mathbf{p,m}})^{-1}\hat
{P}_{\mathbf{p,m}}=\hat{P}_{\mathbf{p,m}}/(i\omega_{\mathbf{p,m}})$, the
contribution from the second term $\propto\hat{P}_{\mathbf{p,m}}$ is
$(2/\omega_{\mathbf{p,m}})\operatorname{Im}\langle\hat{V}^{(\mathbf{m
,\mathbf{p})}\rangle_{\mathbf{p},\mathbf{m}}\langle\hat{V}^{(\mathbf{p
,\mathbf{m})}\rangle_{\mathbf{m},\mathbf{m}}$. Since $|\mathbf{p}\rangle$ and
$|\mathbf{m}\rangle$ differs only by a single nuclear spin flip, the
difference between $P_{\mathbf{m},\mathbf{m}}$ and $P_{\mathbf{p},\mathbf{m}}$
are negligible and we recover Eq. (3) in the main text.
For $\mathbf{m}\neq\mathbf{n}$, neglecting the coupling of $p^{(\mathbf{m
,\mathbf{n})}$ to other variables, which amounts to neglecting the small
second-order corrections to the transverse mean field and the
electron-mediated nuclear spin interaction, we obtain
\[
(\dot{p}^{(\mathbf{m},\mathbf{n}})_{2}=-i(\delta\epsilon_{\mathbf{m|n}
^{(2)}-\delta\epsilon_{\mathbf{n}|\mathbf{m}}^{(2)})p^{(\mathbf{m
,\mathbf{n})}-\Gamma_{\mathbf{m},\mathbf{n}}^{\varphi}p^{(\mathbf{m
,\mathbf{n})}-\frac{\sum_{\mathbf{p}}(W_{\mathbf{p}\leftarrow\mathbf{n
|\mathbf{m}}+W_{\mathbf{p}\leftarrow\mathbf{m}|\mathbf{n}})}{2
p^{(\mathbf{m,n})},
\]
where $\Gamma_{\mathbf{m},\mathbf{n}}^{\varphi}$ is given by Eq. (5) of the
main text an
\begin{align*}
\delta\epsilon_{\mathbf{n}|\mathbf{m}}^{(2)} & \equiv-\sum_{\mathbf{p
}\operatorname{Im}\operatorname*{Tr}\nolimits_{e}\hat{V}^{(\mathbf{n,p
)}(\mathcal{L}_{\mathbf{p},\mathbf{m}}+i\omega_{\mathbf{p,n}})^{-1}(\hat
{V}^{(\mathbf{p,n})}P_{\mathbf{n,m}}-\langle\hat{V}^{(\mathbf{p,n})
\rangle_{\mathbf{n},\mathbf{m}}P_{\mathbf{m},\mathbf{p}}),\\
W_{\mathbf{p}\leftarrow\mathbf{m|n}} & \equiv-2\operatorname{Re
\operatorname*{Tr}\nolimits_{e}\hat{V}^{(\mathbf{m,p})}(\mathcal{L
_{\mathbf{p,n}}+i\omega_{\mathbf{p,m}})^{-1}(\hat{V}^{(\mathbf{p,m
)}P_{\mathbf{m,n}}-\langle\hat{V}^{(\mathbf{p,m})}\rangle_{\mathbf{m
,\mathbf{n}}P_{\mathbf{p,n}}).
\end{align*}
\subsection{Exact treatment of longitudinal hyperfine interaction $\hat{K}$}
For the nuclear spin relaxation time $T_{1}$ much longer than the electron
damping time $T_{e}$, we can adiabatically single out the dynamics of
$\{p^{(\mathbf{m},\mathbf{m})}\}$ on the coarse-grained time scale
$T_{e},T_{\mathrm{coh}},T_{2}\ll\Delta t\ll T_{1}$ by treating both
$\mathcal{L}_{\mathbf{m,n}}$ and $\hat{K}^{(\mathbf{m},\mathbf{n})}$
\emph{exactly} and $\hat{V}$ as a perturbation. The slow variables are
$\{p^{(\mathbf{m},\mathbf{m})}\}$ and others are fast variables. The
zeroth-order steady state solution $\rho_{0}^{(\mathbf{m},\mathbf{n})}=0$
$(\mathbf{m}\neq\mathbf{n})$, so the first-order slow variable dynamics
$(\dot{p}^{(\mathbf{m},\mathbf{m})})_{1}=0$. The first-order steady-state
solution is determined from
\[
\dot{\rho}_{1}^{(\mathbf{m},\mathbf{n})}=\mathcal{L}_{\mathbf{m,n
}^{\mathrm{tot}}\hat{\rho}_{1}^{(\mathbf{m},\mathbf{n})}-i\langle
\mathbf{m}|[\hat{V},\hat{\rho}_{0}]|\mathbf{n}\rangle
\]
a
\[
\hat{\rho}_{1}^{(\mathbf{m},\mathbf{n})}=i\sum_{\mathbf{p}}[(\mathcal{L
_{\mathbf{m,n}}^{\mathrm{tot}}+i\omega_{\mathbf{m},\mathbf{p}})^{-1}\hat
{V}^{(\mathbf{m},\mathbf{p})}\hat{\rho}_{0}^{(\mathbf{p},\mathbf{n
)}-(\mathcal{L}_{\mathbf{m,n}}^{\mathrm{tot}}+i\omega_{\mathbf{p},\mathbf{n
})^{-1}\hat{\rho}_{0}^{(\mathbf{m},\mathbf{p})}\hat{V}^{(\mathbf{p
,\mathbf{n})}]
\]
where $\mathcal{L}_{\mathbf{m,n}}^{\mathrm{tot}}(\bullet)\equiv\mathcal{L
_{\mathbf{m},\mathbf{n}}(\bullet)-i\{\bullet,\hat{K}^{(\mathbf{m},\mathbf{n
)}\}/2$. Replacing $\hat{\rho}(t)$ in Eq. (\ref{EOM_P}) with $\hat{\rho
_{1}(t)$ and neglecting the coupling of $p^{(\mathbf{m},\mathbf{m})}$ to
nuclear spin coherences gives Eqs. (2) of the main text, with the transition
rat
\[
W_{\mathbf{p}\leftarrow\mathbf{m}}=2\operatorname{Re}\int_{0}^{\infty
}e^{i\omega_{\mathbf{p,m}}t}dt\operatorname*{Tr}\nolimits_{e}\hat
{V}^{(\mathbf{m},\mathbf{p})}e^{\mathcal{L}_{\mathbf{p,m}}^{\mathrm{tot}
t}\hat{V}^{(\mathbf{p},\mathbf{m})}\hat{P}_{\mathbf{m},\mathbf{m}
\]
obtained from Eq. (3) of the main text by replacing $\mathcal{L
_{\mathbf{p,m}}^{\mathrm{tot}}$ with $\mathcal{L}_{\mathbf{p,m}}$. In the
perturbation limit, we have $\mathcal{L}_{\mathbf{p,m}}^{\mathrm{tot}
\approx\mathcal{L}_{\mathbf{p,m}}$ and recovers Eq. (3) of the main text.
\subsection{NV Hamiltonian under coherent population trapping}
The NV Hamiltonian for the coherent population trapping (CPT) experiment has
been discussed in \cite{ToganNature2011}. Here we reproduce it with greater
detail using our own notations. The NV states of relevance include 3 ground
triplet states $|0\rangle$ (energy $0$), $|\pm1\rangle$ (energy
$D_{\mathrm{gs}}$), 6 excited triplet states $|E_{y}\rangle,|E_{x}\rangle$
(energy $\varepsilon_{E_{x}}=\varepsilon_{E_{y}}$)$,|E_{1}\rangle
,|E_{2}\rangle$ (energy $\varepsilon_{E_{1}}=\varepsilon_{E_{2}}$),
$|A_{1}\rangle$ (energy $\varepsilon_{A_{1}}$)$,|A_{2}\rangle$ (energy
$\varepsilon_{A_{2}}$), and one metastable singlet $|S\rangle$ (energy
$\varepsilon_{S}$). A linearly polarized laser with electric field
$\mathbf{E}_{1}e^{-i\omega_{1}t}/2+c.c.$ and frequency $\omega_{1
=\varepsilon_{A_{1}}-D_{\mathrm{gs}}$ resonantly excites the ground states
$|\pm1\rangle$ to the excited state $|A_{1}\rangle$ and, at the same time,
off-resonantly excited $|\pm1\rangle$ to the excited state $|A_{2}\rangle$
with detuning $\Delta=\varepsilon_{A_{2}}-\varepsilon_{A_{1}}$. Another
linearly polarized laser with electric field $\mathbf{E}_{2}e^{-i\omega_{2
t}/2+c.c.$ and frequency $\omega_{2}=\varepsilon_{E_{x}}$ resonantly excites
$|0\rangle$ to $|E_{y}\rangle$. The relevant optical transition matrix element
of the electric dipole operator $\mathbf{d}\equiv-e\mathbf{r}$ are $\langle
A_{1}|\mathbf{d}|\pm1\rangle=\pm d_{a,E}\mathbf{e}_{\pm}/(2\sqrt{2})$, $\langle
A_{2}|\mathbf{d}|\pm1\rangle=id_{a,E}\mathbf{e}_{\pm}/(2\sqrt{2})$, and
$\langle E_{y}|\mathbf{d}|0\rangle=d_{a,E}\mathbf{e}_{x}/\sqrt{2}$, where
$\mathbf{e}_{\pm}\equiv\mathbf{e}_{x}\pm i\mathbf{e}_{y}$ and $d_{a,E}$ is the
reduced matrix element of the electric dipole moment
\cite{Dohertyelectronic2011}. Thus the laser coupling Hamiltonian
\[
\hat{H}_{c}(t)=\frac{\Omega_{A}}{2}(e^{i\phi}\hat{\sigma}_{A_{1},+1
-e^{-i\phi}\hat{\sigma}_{A_{1},-1}+ie^{i\phi}\hat{\sigma}_{A_{2
,+1}+ie^{-i\phi}\hat{\sigma}_{A_{2},-1})e^{-i\omega_{1}t}+\frac{\Omega_{E}
{2}e^{-i\omega_{2}t}\hat{\sigma}_{E_{y},0}+h.c.,
\]
where we have defined $\hat{\sigma}_{ij}\equiv|i\rangle\langle j|$,
$\Omega_{E}=\mathbf{E}_{2}\cdot\langle E_{y}|\mathbf{d}|0\rangle
=E_{2,x}d_{a,E}/\sqrt{2}$, and $\Omega_{A}e^{i\phi}\equiv\mathbf{E}_{1
\cdot\langle A_{1}|\mathbf{d}|+1\rangle=d_{aE}(E_{1,x}+iE_{1,y})/(2\sqrt{2})$.
Defining the bright state $|b\rangle=(e^{-i\phi}|+1\rangle-e^{i\phi
|-1\rangle)/\sqrt{2}$ and dark state $|d\rangle=(e^{-i\phi}|+1\rangle
+e^{i\phi}|-1\rangle)/\sqrt{2}$, the laser coupling Hamiltonian simplifies to
\[
\hat{H}_{c}(t)=\frac{\Omega_{A}}{\sqrt{2}}(\hat{\sigma}_{A_{1},b}+i\hat
{\sigma}_{A_{2},d})e^{-i\omega_{1}t}+\frac{\Omega_{E}}{2}e^{-i\omega_{2}t
\hat{\sigma}_{E_{y},0}+h.c..
\]
In the rotating frame $\vert \psi(t)\rangle$ connected to the laboratory frame
$|\psi_{\mathrm{lab}}(t)\rangle$ via $\vert \psi(t)\rangle=e^{i\hat{H}_{0}t}|\psi
_{\mathrm{lab}}(t)\rangle$ with $\hat{H}_{0}=\varepsilon_{E_{x}}(\hat{\sigma
}_{E_{x},E_{x}}+\hat{\sigma}_{E_{y},E_{y}})+\varepsilon_{E_{1}}(\hat{\sigma
}_{E_{1},E_{1}}+\hat{\sigma}_{E_{2},E_{2}})+\varepsilon_{A_{1}}(\hat{\sigma
}_{A_{1},A_{1}}+\hat{\sigma}_{A_{2},A_{2}})+\varepsilon_{S}\hat{\sigma
_{S,S}+D_{\mathrm{gs}}(\hat{\sigma}_{1,1}+\hat{\sigma}_{-1,-1})$, the NV
Hamiltonian i
\[
\Delta\hat{\sigma}_{A_{2},A_{2}}+\frac{\Omega_{A}}{\sqrt{2}}(\hat{\sigma
}_{A_{1},b}+i\hat{\sigma}_{A_{2},d}+h.c.)+\frac{\Omega_{E}}{2}(\hat{\sigma
}_{E_{y},0}+\hat{\sigma}_{0,E_{y}}).
\]
We further include an external magnetic field $B$ along the $z$ (N-V) axis and
uniaxial strain with effective ground strain fields $\xi_{x},\xi_{y},\xi_{z}$.
The former gives rise to ground state Zeeman term $g_{e}\mu_{B}B\hat{S
_{g}^{z}\equiv\omega_{e}\hat{S}_{g}^{z}$. The latter induces small corrections
to the energy splitting between $|\pm1\rangle$ and couples $|+1\rangle$ and
$|-1\rangle$ through $\hat{H}_{\mathrm{strain}}\equiv-(\xi_{x}+i\xi_{y
)\hat{\sigma}_{1,-1}+h.c=-\xi_{\perp}e^{i\Theta}\hat{\sigma}_{1,-1}+h.c$,
where $\xi_{\perp}\equiv(\xi_{x}^{2}+\xi_{y}^{2})^{1/2}$ and $\Theta$ is
determined by $\xi_{\perp}\cos\Theta=\xi_{x}$ and $\xi_{\perp}\sin\Theta
=\xi_{y}$ (cf. Ref. \cite{ToganNature2011}). The NV Hamiltonian becomes
\[
\hat{H}_{e}=\omega_{e}\hat{S}_{g}^{z}+\hat{H}_{\mathrm{strain}}+\Delta
\hat{\sigma}_{A_{2},A_{2}}+\frac{\Omega_{A}}{\sqrt{2}}(\hat{\sigma}_{A_{1
,b}+i\hat{\sigma}_{A_{2},d}+h.c.)+\frac{\Omega_{E}}{2}(\hat{\sigma}_{E_{y
,0}+\hat{\sigma}_{0,E_{y}}).
\]
In the basis $|b\rangle,|d\rangle,|0\rangle$, we have $\hat{S}_{g}^{z
=\hat{\sigma}_{d,b}+\hat{\sigma}_{b,d}$ and $\hat{\sigma}_{1,-1}=e^{2i\phi
}(\hat{\sigma}_{d,d}-\hat{\sigma}_{b,b}+\hat{\sigma}_{b,d}-\hat{\sigma
_{d,b})/2$. Thus the strain ter
\[
\hat{H}_{\mathrm{strain}}=-\xi_{\perp}(\hat{\sigma}_{d,d}-\hat{\sigma
_{b,b})\cos(\Theta+2\phi)-i\xi_{\perp}(\hat{\sigma}_{b,d}-\hat{\sigma
_{d,b})\sin(\Theta+2\phi)
\]
and the Zeeman term $\omega_{e}\hat{S}_{g}^{z}=\omega_{e}(\hat{\sigma
_{d,b}+\hat{\sigma}_{b,d})$ induces coupling between the dark state and the
bright state. The strain-induced coupling is minimized by choosing
$\Theta+2\phi=\pi$ \cite{ToganNature2011}, such tha
\begin{equation}
\hat{H}_{e}=\xi_{\perp}(\hat{\sigma}_{d,d}-\hat{\sigma}_{b,b})+\omega_{e
(\hat{\sigma}_{b,d}+\hat{\sigma}_{d,b})+\Delta\hat{\sigma}_{A_{2},A_{2}
+\frac{\Omega_{A}}{\sqrt{2}}(\hat{\sigma}_{A_{1},b}+i\hat{\sigma}_{A_{2
,d}+h.c.)+\frac{\Omega_{E}}{2}(\hat{\sigma}_{E_{y},0}+\hat{\sigma}_{0,E_{y}}).
\label{HE
\end{equation}
\subsection{Analytical expressions for steady NV state}
The NV steady state $\hat{P}_{\mathbf{m},\mathbf{m}}$ (denoted by $\hat{P}$
for brevity) is determined by $\mathcal{L}_{\mathbf{m},\mathbf{m}}\hat{P}=0$
and $\operatorname*{Tr}_{e}\hat{P}=1$, where $\mathcal{L}_{\mathbf{m
,\mathbf{m}}(\bullet)\equiv-i[\hat{H}_{e}+\hat{S}_{g}^{z}h_{\mathbf{m
},\bullet]+\mathcal{L}_{e}(\bullet)$ and $\hat{H}_{e}+\hat{S}_{g
^{z}h_{\mathbf{m}}=\hat{H}_{e}|_{\omega_{e}\rightarrow\delta_{\mathbf{m
}\equiv\omega_{e}+h_{\mathbf{m}}}$. Straightforward calculation gives
\begin{align*}
P_{S,S} & =\frac{2\gamma_{ce}}{\gamma_{s}}P_{E_{y},E_{y}},\\
P_{0,0} & =\left( 1+\frac{\gamma+2\gamma_{ce}}{W_{E}}\right) P_{E_{y
,E_{y}},\\
\gamma_{s}P_{S,S} & =\gamma_{s1}P_{A_{1},A_{1}}+\gamma_{s2}P_{A_{2},A_{2}},\\
P_{b,b} & \approx\eta_{1}\left( 1+\frac{\gamma_{\varphi}}{\Gamma_{_{A_{1}}
}+\frac{\gamma+\gamma_{s1}}{W_{A}}\right) P_{E_{y},E_{y}},
\end{align*}
where $W_{E}\equiv\Omega_{E}^{2}/\Gamma_{A_{1}}$ is the resonant optical
transition rate between $|0\rangle$ and $|E_{y}\rangle$, $W_{A}=\Omega_{A
^{2}/\Gamma_{A_{1}}$ is the resonant optical transition rate from
$|\pm1\rangle$ to $|A_{1}\rangle$, and $\eta_{1}\equiv\gamma_{ce}/\gamma_{s1
$. We treat the off-resonant $|\pm1\rangle\rightarrow|A_{2}\rangle$ excitation
perturbatively. Up to zeroth order, near the two-photon dark resonance (small
$\delta_{\mathbf{m}}$), we neglect $O(\delta_{\mathbf{m}}^{4})$ and high order
terms and obtai
\begin{align*}
P_{E_{y},E_{y}}^{(0)} & =P_{0}\frac{\delta_{\mathbf{m}}^{2}}{\delta
_{\mathbf{m}}^{2}+\delta_{0}^{2}},\\
P_{A_{1},A_{1}}^{(0)} & =2\eta_{1}P_{E_{y},E_{y}}^{(0)},\\
P_{d,d}^{(0)} & \approx1-\left( 2+\frac{\gamma+2\gamma_{ce}}{W_{E}}\right)
P_{E_{y},E_{y}}^{(0)},
\end{align*}
wher
\begin{align*}
P_{0} & \approx\frac{1}{\frac{2}{\eta_{2}}+2\eta_{1}\frac{\gamma+\gamma_{s1
}{W_{A}}+\frac{\gamma+2\gamma_{ce}}{W_{E}}},\\
\delta_{0}^{2} & \approx\frac{\eta_{1}\eta_{2}}{4}\frac{W_{A}^{2}-8W_{A
\xi_{\perp}^{2}/\Gamma_{_{A_{1}}}+4\xi^{2}}{\eta_{1}\eta_{2}+\frac{W_{A
}{\gamma+\gamma_{s1}}\left( 1+\frac{\eta_{2}}{2}\frac{\gamma+2\gamma_{ce
}{W_{E}}\right) },
\end{align*}
with $\eta_{2}\equiv\gamma_{s}/(\gamma_{s}+\gamma_{ce})$ and we have used
$\xi_{\perp}\ll\Gamma_{_{A_{1}}}$ and $\eta_{1}\ll1$. Note that only
$|E_{y}\rangle,|0\rangle$, and $|d\rangle$ are significantly populated, while
the populations on $|S\rangle,|b\rangle,|A_{1}\rangle$ are much smaller than
that on $|E_{y}\rangle$ since $\gamma_{ce}\ll\gamma_{s},\gamma_{s1},\gamma$.
The leading order correction from the off-resonant excitation to
$|A_{2}\rangle$ ar
\begin{align*}
P_{E_{y}E_{y}}^{(2)} & \approx\frac{1}{2\eta_{1}}\frac{W_{A_{2}}
{\gamma+\gamma_{s1}}P_{dd}^{(0)},\\
P_{A_{1}A_{1}}^{(2)} & =\frac{W_{A_{2}}}{\gamma+\gamma_{s1}}P_{dd}^{(0)},\\
P_{A_{2}A_{2}}^{(2)} & =\frac{W_{A_{2}}}{\gamma+\gamma_{s2}}P_{d,d}^{(0)},
\end{align*}
where $W_{A_{2}}=(\Omega_{A}^{2}/2)\Gamma_{_{A_{2}}}/\Delta^{2}$ is the
off-resonant transition rate from $|\pm1\rangle$ to $|A_{2}\rangle$ and we
have used $\gamma_{s2}\ll\gamma,\gamma_{s1}$. Now, with the off-resonant
excitation from $|d\rangle$ to $|A_{2}\rangle$, the populations on the excited
states no longer vanish even on two-photon resonance $\delta_{\mathbf{m}}=0$.
As suggested in Ref. \cite{ToganNature2011}, this off-resonant excitation
limits the efficiency of nuclear spin cooling and narrowing. Under saturated
pumping $W_{E}\gg\gamma$, we have $P_{dd}^{(0)}\approx1-2P_{E_{y},E_{y}
^{(0)}$.
\bibliographystyle{apsrev}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 0
|
Q: What's wrong with my .htaccess code? I want to redirect www.example.com/ext_jobs/ to www.example.com/customer/tremoria/ext_jobs/
Options +FollowSymLinks
RewriteEngine on
RewriteRule ^/ext_jobs/(.*)$ customer/tremoria/ext_jobs/$1 [L]
But my solution is not working. What's wrong?
A: You can use the rewritebase instead of the / in ^/ext_jobs/(.*)$
Options +FollowSymLinks
RewriteEngine on
RewriteBase /
RewriteCond %{REQUEST_URI} !^customer/tremoria/ext_jobs
RewriteRule ^ext_jobs/(.*)/?$ customer/tremoria/ext_jobs/$1 [L,NC,QSA]
A: Do not use mod-rewrite for simple things. You have simple solutions with mod-alias like Redirect, RedirectMatch and Alias.
In your case I would try a Redirect if you really need a Redirect. But in your sample you are only doing an internal redirect, transparent to the user browser.
So an Alias is maybe the solution
Alias ext_jobs/ customer/tremoria/ext_jobs/
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 6,516
|
Andreas Puelacher (* 21. April 1964) ist ein österreichischer Alpinskitrainer. Nach einem achtjährigen Engagement als Rennsportleiter der Herrenmannschaft im Österreichischen Skiverband (ÖSV) betreut er seit 2022 die Damenmannschaft des Deutschen Skiverbandes (DSV).
Biografie
Andreas Puelacher stammt aus Oberhofen im Inntal, wo er auch das Skifahren erlernte. Sein Elternhaus ist ein Bauernhof, der heute von seinem Bruder bewirtschaftet wird. Er ist verheiratet und hat mit seiner Frau Andrea einen Sohn und zwei Töchter. Eine Leidenschaft von ihm ist das Drachenfliegen.
Puelacher ist staatlich geprüfter Skilehrer, Skilehrwart, C-Trainer und diplomierter Sportlehrer. Zwischen 1990 und 1995 war er als Trainer erstmals für den Österreichischen Skiverband (ÖSV) tätig. Danach betreute er bis 1999 die Damenmannschaft Liechtensteins. Von 1999 bis 2005 war er Cheftrainer der Schweizer Damen und kehrte danach nach Liechtenstein zurück. Ab 2008 betreute er mit den Schweizer Skiherren erstmals eine männliche Nationalmannschaft. Nach 15 Jahren ging er 2010 zurück zum ÖSV und arbeitete fünf Saisonen lang als Trainer für die Weltcup-Gruppen Riesenslalom und Kombination Herren. Im April 2014 folgte er dem Vorarlberger Mathias Berthold als sportlicher Leiter der ÖSV-Herrenmannschaft nach. In der Saison 2019/20 gewannen Österreichs Herren erstmals seit 28 Jahren nicht die Mannschaftswertung im Weltcup, wofür sich der Cheftrainer rechtfertigen musste. Am 5. März 2022 verkündete Puelacher nach acht Jahren als ÖSV-Rennsportleiter der alpinen Herren, seinen Vertrag nicht mehr über die laufende Saison hinaus verlängern zu wollen. Mitte April gab der deutsche Skiverband bekannt, Puelacher als Cheftrainer seiner alpinen Damenmannschaft verpflichtet zu haben.
Puelacher engagiert sich auf kommunaler Ebene politisch und ist in seiner Heimatgemeinde Ersatz-Gemeinderat für die Liste "Unabhängig für Oberhofen".
Auszeichnungen
2015: Sportehrenzeichen der Gemeinde Oberhofen
Weblinks
Profil auf der Website des ÖSV
Einzelnachweise
Alpinskitrainer (Österreich)
Alpinskitrainer (Schweiz)
Alpinskitrainer (Liechtenstein)
Person (Oberhofen im Inntal)
Österreicher
Geboren 1964
Mann
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 9,910
|
\section{Supplemental Materials}
\twocolumngrid
\pagenumbering{arabic}
\renewcommand{\thepage}{S-\arabic{page}}
\setcounter{equation}{0}
\setcounter{figure}{0}
\setcounter{table}{0}
\setcounter{page}{1}
\makeatletter
\renewcommand{\theequation}{S\arabic{equation}}
\renewcommand{\thefigure}{S\arabic{figure}}
\section{I. SU($2$) Wilson Loop}
In this supplementary material, the equivalence between the two $\mathbb{Z}_2$ invariants defined in the main text is established. The first invariant, $\delta_0$, is the relative sign between the $\SU(2)$ transition matrices at points related by $\hat{\mathcal{T}}$-symmetry \cite{Lee2008},
\begin{equation}
G(\mathbf{q})=\delta_0 G(\mathbf{q}+\pi \hat{\vecg{\phi}}_\mathbf{q}).
\end{equation}
The second is the Fu-Kane formula in Eq. \eqref{eq: Pfaffian} \cite{Fang2016},
\begin{equation}\label{eq: Fu-Kane}
\delta=\Pf \tilde{w}_+(\theta_\mathbf{q}=0)\Pf \tilde{w}_+(\theta_\mathbf{q}=\pi),
\end{equation}
which characterizes the inability to select the $\SU(2)$ part of $\hat{\mathcal{T}}$ to be $i\sigma_y$ globally. As discussed in the main text, the two Fermi surfaces belong to the same topological class, so we will henceforth focus on $\text{FS}_{+}$ and omit the $+$ subscript for notational convenience.
Following Ref. \cite{Lee2008}, the two $\mathbb{Z}_2$ invariants can be expressed as an $\SU(2)$ Wilson loop. In a gauge where the eigenstates are non-singular, $\ket{\chi_{a}(\mathbf{q})}\equiv \hat{\chi}_{a}^\dagger(\mathbf{q})\ket{0}$, it is possible to define the $\Un(2)$ Berry connection $\vec{A}_{ab}(\mathbf{q})=i\bra{\chi_{a}(\mathbf{q})}\nabla_\mathbf{q}\ket{\chi_{b}(\mathbf{q})}$ over the entire FS. It is useful to decompose $\vec{A}$ into $\Un(1)$ and $\SU(2)$ parts: $\vec{A}=\vec{A}_{\Un(1)}+\vec{A}_{\SU(2)}$, where $\vec{A}_{\Un(1)}$ and $\vec{A}_{\SU(2)}$ are the traceful and traceless parts of $\vec{A}$, respectively. The central object connecting the two definitions for the $\mathbb{Z}_2$ invariant is the Wilson loop \cite{Lee2008,Fu2006}
\begin{equation}
W[C]=\frac{1}{2}\Tr \mathcal{P}\exp\left[i\oint_C \d\vec{q}\cdot \vec{A}_{\SU(2)}(\mathbf{q})\right],
\end{equation}
where $\mathcal{P}$ is the path-ordering operator and $C$ is the $\hat{\mathcal{T}}$-invariant loop in Fig. \ref{fig: Wilson}, which is separated into four segments $C_{1-4}$. The Wilson loop is gauge invariant as a result of the trace.
\begin{figure}[htpb]
\centering {\epsfig{file=Wilson.png,width=0.8\linewidth}}
\caption{The Fermi surface $\text{FS}_+$ consists of two gauge patches, $N$ (red) and $S$ (blue), with overlap $N\cap S$ (green). The $\hat{\mathcal{T}}$-invariant Wilson loop $C$, which is separated into four segments $C_{1-4}$, is a great circle that passes through the north and south poles. For concreteness the meridians are taken to be at $\phi_\mathbf{q}=0,\pi$.
}
\label{fig: Wilson}
\end{figure}
To evaluate the Wilson loop, consider the unitary infinitesimal propagator \cite{Ryu2010}
\begin{equation}
K_{ab}(\mathbf{q}_2,\mathbf{q}_1)\equiv\braket{\chi_a(\mathbf{q}_2)}{\chi_b(\mathbf{q}_1)}=[e^{i\d \mathbf{q}\cdot \vec{A}(\mathbf{q}_2)}]_{ab},
\end{equation}
where $\d \mathbf{q}=\mathbf{q}_2-\mathbf{q}_1$. As a result of $\hat{\mathcal{T}}$-symmetry, it satisfies the sewing condition
$K(\mathbf{q}_1+\pi \hat{\vecg{\phi}}_\mathbf{q},\mathbf{q}_2+\pi \hat{\vecg{\phi}}_\mathbf{q})=w(\mathbf{q}_1)K^T(\mathbf{q}_2,\mathbf{q}_1)w^\dagger(\mathbf{q}_2)$. When the propagator is decomposed into $\Un(1)$ and $\SU(2)$ parts, $K(\mathbf{q}_1,\mathbf{q}_2)=e^{i\d \mathbf{q}\cdot \vec{A}_{\Un(1)}}\tilde{K}(\mathbf{q}_1,\mathbf{q}_2)$, where $\tilde{U}(\mathbf{q}_2,\mathbf{q}_1)=e^{i\d\mathbf{q}\cdot \vec{A}_{\SU(2)}(\mathbf{q})}$, the $\SU(2)$ part of the propagator satisfies the corresponding sewing condition
\begin{equation}\label{eq: prop sewing}
\tilde{K}(\mathbf{q}+\pi \hat{\vecg{\phi}}_\mathbf{q},\mathbf{q}_2+\pi \hat{\vecg{\phi}}_\mathbf{q})=\tilde{w}(\mathbf{q}_1)\tilde{K}^T(\mathbf{q}_2,\mathbf{q}_1)\tilde{w}^\dagger(\mathbf{q}_2).
\end{equation}
Here $\tilde{w}(\mathbf{q})$ is the $\SU(2)$ part of the sewing matrix, as defined in the main text. The Wilson loop can be constructed from the infinitesimal propagators by
\begin{equation}
W[C]=\frac{1}{2}\Tr \tilde{K}_4\tilde{K}_3\tilde{K}_2\tilde{K}_1,
\end{equation}
where the propagators for the four segments, $C_{1-4}$, are
\begin{align}
\begin{split}
\tilde{K}_1&\equiv\prod_{n=1}^N \tilde{K}_\pi\left(\frac{\pi}{2}+n\delta\theta_\mathbf{q},\frac{\pi}{2}+(n-1)\delta\theta_\mathbf{q} \right),\\
\tilde{K}_2&\equiv\prod_{n=1}^N \tilde{K}_0\left(\pi-n\delta\theta_\mathbf{q},\pi-(n-1)\delta\theta_\mathbf{q}\right),\\
\tilde{K}_3&\equiv\prod_{n=1}^N \tilde{K}_0\left(\frac{\pi}{2}-n\delta\theta_\mathbf{q},\frac{\pi}{2}-(n-1)\delta\theta_\mathbf{q}\right),\\
\tilde{K}_4&\equiv\prod_{n=1}^N \tilde{K}_\pi\left(n\delta\theta_\mathbf{q},(n-1)\delta\theta_\mathbf{q} \right).
\end{split}
\end{align}
Here $\delta\theta_\mathbf{q}=\frac{\pi}{2N}$ and we use the notation $\tilde{K}_{\phi_\mathbf{q}}(\theta_{\mathbf{q}_2},\theta_{\mathbf{q}_1})=\tilde{K}(\mathbf{q}_2,\mathbf{q}_2)$, where $\mathbf{q}_1=(\theta_{\mathbf{q}_1},\phi_\mathbf{q})$ and $\mathbf{q}_2=(\theta_{\mathbf{q}_2},\phi_\mathbf{q})$. The sewing condition gives the constraints, $\tilde{K}_1=\tilde{w}(\theta_\mathbf{q}=\pi)\tilde{K}_2^T\tilde{w}^\dagger(\theta_\mathbf{q}=\pi/2)$ and $\tilde{K}_4=\tilde{w}(\theta_\mathbf{q}=\pi/2)\tilde{K}_3^T\tilde{w}^\dagger (\theta_\mathbf{q}=0)$. Using $\tilde{w}(\theta_\mathbf{q}=0,\pi)=\Pf \tilde{w}(\theta_\mathbf{q}=0,\pi)i\sigma_y$ and the identity $\sigma_y F^T\sigma_y=F^\dagger$ for any $\SU(2)$ matrix $F$, the Wilson loop reduces to the Fu-Kane invariant, $W[C]=\delta$.
The Wilson loop $W[C]$ is also equal to the invariant $\delta_0$. To arrive at the appropriate form for $W[C]$, perform for the propagators in the segments $C_3$ and $C_4$ a gauge transformation to $\ket{\chi_a(\mathbf{q})}=\sum_b\ket{\alpha_b(\mathbf{q})}U_{ba}(\mathbf{q})$, where $\ket{\alpha_a(\mathbf{q})}\equiv \hat{\alpha}^\dagger_a(\mathbf{q})\ket{0}$ are states smooth over $N$ and satisfy the $\hat{\mathcal{T}}$-symmetry condition, Eq. \eqref{eq: TR FS}. Under this gauge transformation, the propagators transform as $\tilde{K}(\mathbf{q}_2,\mathbf{q}_1)=\tilde{U}^\dagger(\mathbf{q}_2) \tilde{K}^N(\mathbf{q}_2,\mathbf{q}_1) \tilde{U}(\mathbf{q}_1)$, where $\tilde{U}(\mathbf{q})$ and $\tilde{K}^N(\mathbf{q}_2,\mathbf{q}_1)$ are the $\SU(2)$ parts of $U(\mathbf{q})$ and $K^N_{ab}(\mathbf{q}_2,\mathbf{q}_1)\equiv \braket{\alpha_a(\mathbf{q})}{\alpha_b(\mathbf{q})}$, respectively. In this gauge, the $\hat{\mathcal{T}}$-symmetry condition is enforced, hence Eq. \eqref{eq: prop sewing} simplifies to $\tilde{K}^N(\mathbf{q}_1+\pi \hat{\vecg{\phi}}_\mathbf{q},\mathbf{q}_2+\pi \hat{\vecg{\phi}}_\mathbf{q})=[\tilde{K}^N(\mathbf{q}_2,\mathbf{q}_1)]^\dagger$. Consequently, the propagators $\tilde{K}^N$ in the entire segment $C_3\cup C_4$ cancel and $\tilde{K}_4\tilde{K}_3=\tilde{U}^\dagger(\theta_\mathbf{q}=\pi/2,\phi_\mathbf{q}=\pi) \tilde{U}(\theta_\mathbf{q}=\pi/2,\phi_\mathbf{q}=0)$. Similarly, the propagators in the segments $C_1$ and $C_2$ are evaluated in the gauge $S$ by performing the gauge transformation $\ket{\chi_a(\mathbf{q})}=\sum_{b}\ket{\beta_b(\mathbf{q})}V_{ba}(\mathbf{q})$, where $\ket{\beta_b(\mathbf{q})}\equiv \hat{\beta}^\dagger_{b}(\mathbf{q})\ket{0}$. By the same argument, only the gauge transformations at the end points contribute and $\tilde{K}_2\tilde{K}_1=\tilde{V}^\dagger(\theta_\mathbf{q}=\pi/2,\phi_\mathbf{q}=0) \tilde{V}(\theta_\mathbf{q}=\pi/2,\phi_\mathbf{q}=\pi)$. The change of basis matrices $U(\mathbf{q})$ and $V(\mathbf{q})$ are related to the transition matrix by $G(\mathbf{q})=\tilde{V}(\mathbf{q})\tilde{U}^\dagger(\mathbf{q})$. Hence, $W[C]=\frac{1}{2}\Tr[ G(\theta_\mathbf{q}=\pi/2,\phi_\mathbf{q}=\pi)G^\dagger(\theta_\mathbf{q}=\pi/2,\phi_\mathbf{q}=0)]=\delta_0$. This establishes that the two invariants, $\delta_0$ and $\delta$, are equal to the Wilson loop $W[C]$.
\section{II. Monopole harmonics in momentum space}
The monopole harmonics in momentum space are defined as irreducible representations of the gauge-covariant angular momentum operator
\begin{equation}
\vec{L}(\mathbf{k})=-\hbar\mathbf{k}\times (i\nabla_\mathbf{k} -\vec{A}(\mathbf{k}))+\hbar q\hat{\mathbf{k}},
\end{equation}
where the $\Un(1)$ Berry connection of the state $\ket{u(\mathbf{k})}$, $\vec{A}(\mathbf{k})=\bra{u(\mathbf{k})}i\nabla_\mathbf{k} \ket{u(\mathbf{k})}$, describes a magnetic monopole with monopole charge $q$ at the origin in momentum space. More concretely, we take the Berry connections in the regions $N$ and $S$ to be
\begin{align}
\vec{A}^N&=q\frac{1-\cos\theta_\mathbf{k}}{k\sin\theta_\mathbf{k}}\hat{\vecg{\phi}}_\mathbf{k},&\vec{A}^S&=-q\frac{1+\cos\theta_\mathbf{k}}{k\sin\theta_\mathbf{k}}\hat{\vecg{\phi}}_\mathbf{k}.
\end{align}
Note that our definition of the angular momentum operator in momentum space differs from Ref. \onlinecite{Wu1976}, which is defined in real space as $\vec{L}(\vr)=\vr \times (-i\hbar \nabla_\vr - \vec{A}(\vr))-\hbar q \hat{\vr}$, by a minus sign in the monopole charge $q$, and hence our monopole harmonic $Y_{q;lm}(\hat{\mathbf{k}})$ has the same functional form as $Y_{-q;lm}(\hat{\vr})$ in Ref. \onlinecite{Wu1976}.
The monopole harmonics can be written in terms of the Wigner $D$-matrices as
\begin{align}
Y_{q;lm}(\theta_\mathbf{k},\phi_\mathbf{k},\psi_\mathbf{k})&=\sqrt{\frac{2l+1}{4\pi}}D_{m,q}^{(l)*}(\phi_\mathbf{k},\theta_\mathbf{k},\psi_\mathbf{k})\\
&=\sqrt{\frac{2l+1}{4\pi}}d_{m,q}^{(l)*}(\theta_\mathbf{k})e^{im\phi_\mathbf{k}}e^{iq\psi_\mathbf{k}},
\end{align}
where the Wigner matrices are defined as
\begin{align}
\label{eq: Wigner D}
D^{(l)}_{mn}(\phi_\mathbf{k},\theta_\mathbf{k},\psi_\mathbf{k})&=\bra{l,m}e^{-\frac{i}{\hbar} \hat{J}_z \phi_\mathbf{k}}e^{-\frac{i}{\hbar}\hat{J}_y \theta_\mathbf{k}}e^{-\frac{i}{\hbar}\hat{J}_z\psi_\mathbf{k}}\ket{l,n}\nonumber\\
d^{(l)}_{mn}(\theta_\mathbf{k})&=\bra{l,m}e^{-\frac{i}{\hbar}\hat{J}_y\theta_\mathbf{k}}\ket{l,n}.
\end{align}
Here $(\phi_\mathbf{k},\theta_\mathbf{k},\psi_\mathbf{k})$ are the usual Euler angles. The choice of the third Euler angle, $\psi_\mathbf{k}$, corresponds to the choice of gauge. For example, the vector potentials $\vec{A}^N$ and $\vec{A}^S$ correspond to the gauge choices $\psi_\mathbf{q}=-\phi_\mathbf{q}$ and $\psi_\mathbf{q}=\phi_\mathbf{q}$, respectively. Examples of monopole harmonics in the gauge $N$ with $q=0,\pm 1$ and $l=1$, including those used in the main text, are listed in Tab. \ref{tab: monopole harmonics}. Those in the gauge $S$ can be obtained by the gauge transformation $Y^N_{q;lm}(\hat{\mathbf{k}})=Y^S_{q;lm}(\hat{\mathbf{k}})e^{-2iq\phi_\mathbf{k}}$.
\begin{table}[htpb]
\centering
\begin{tabular}{|c|ccc|}
\hline
\diagbox{$m$}{$q$} & $1$&$0$& $-1$\\
\hline
$1$&$\frac{1}{2}(1+\cos\theta_\mathbf{k})$ &$-\frac{1}{\sqrt{2}}\sin\theta_\mathbf{k} e^{i\phi_\mathbf{k}}$&$\frac{1}{2}(1-\cos\theta_\mathbf{k})e^{i2\phi_\mathbf{k}}$\\
$0$&$\frac{1}{\sqrt{2}}\sin\theta_\mathbf{k} e^{-i\phi_\mathbf{k}}$&$\cos\theta_\mathbf{k}$&$-\frac{1}{\sqrt{2}}\sin\theta_\mathbf{k} e^{i\phi_\mathbf{k}}$\\
$-1$&$\frac{1}{2}(1-\cos\theta_\mathbf{k})e^{-i2\phi_\mathbf{k}}$&$\frac{1}{\sqrt{2}}\sin\theta_\mathbf{k} e^{-i\phi_\mathbf{k}}$&$\frac{1}{2}(1+\cos\theta_\mathbf{k})$ \\
\hline
\end{tabular}
\caption{The monopole harmonics $\sqrt{\frac{4\pi}{3}}Y_{q;lm}(\hat{\mathbf{k}})$ with $l=1$ and $q,m=0,\pm 1$ in the gauge $N$.}
\label{tab: monopole harmonics}
\end{table}
\end{document}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 8,574
|
\section{\label{}}
\section*{Introduction}
Synchronous oscillations are found in many brain areas and are
responsible for macroscopic electrical responses of the brain
including field potentials and EEG signals. Within a single brain
area, synchronization of neuronal activity serves to amplify signals
to upstream regions \cite{tiesinga:2004}, while synchronization
across different areas may allow activity to be selectively routed.
Considerable theoretical interest has recently emerged in the
generation of synchrony by correlated ``noisy'' inputs to uncoupled
oscillators \cite{teramae:2004, goldobin:2005, nakao:2005, stroeve:2001},
a phenomenon we will refer to as stochastic synchrony. In the brain,
stochastic synchrony may account for observations such as long-range
synchronization \cite{engel:1991, engel:1991b}, that are
difficult to explain by the presence of synaptic connectivity
alone. Moreover, noisy inputs have been shown to synchronize real
neurons \textit{in vitro} \cite{galan:2006}.
The key component in the study of noisy oscillators is the
phase-resetting curve (PRC). This curve characterizes how inputs to an
oscillator shift its timing, or phase. In the context of neurons, spike
times are believed to play an important role in coding and in the propagation of information across brain regions. Thus, the PRC
provides a quantitative characterization of how inputs to neural
oscillators alter the timing of spikes.
The theory of deterministic oscillators has shown that
the type of bifurcation from steady-state to periodic behavior
determines the shape of the PRC. Weak coupling theory shows that the form of the
interaction between oscillators together with their intrinsic response (the PRC)
provide sufficient information about the ability of the coupling to
synchronize (or desynchronize) the oscillations. For very fast
excitatory synaptic interactions, Type II oscillators characterized
by the Hopf bifurcation synchronize more readily than Type I
oscillators characterized by the
saddle-node-on-an-invariant-circle (SNIC) bifurcation
\cite{hansel:1995, ermentrout:2001, gutkin:2005, netoff:2005}. This difference
in ability to synchronize with excitatory
coupling is a consequence of the shape of the PRC occurring near the
two different bifurcations. A PRC which contains both negative and
positive lobes can allow inputs to both slow down the oscillator which is ahead and
speed up the oscillator which is behind. In contrast, a non-negative
PRC can only speed up the timing of both oscillators, so that
synchronization becomes more difficult. A number of authors
\cite{ermentrout:2001, izhikevich, brown} have shown that the PRC near
a SNIC is non-negative and approximately proportional to $1-\cos t$,
while the PRC near a Hopf is proportional to $\sin(t+\alpha).$ Thus, Type II
PRCs have a large negative lobe, whereas Type I PRCs are strictly positive.
Two recent papers have shown that Type II PRCs
are better than Type I PRCs
at synchronizing uncoupled oscillators with correlated input \cite{galanfp,sashi}. That is,
for a given input correlation of the noisy stimulus, the output
correlation of the oscillators is higher with Type II than with Type I
PRCs. In these two papers, specific functions for PRCs were checked
(namely, $\sin(t)$ and $1-\cos(t)$), and the correlations and degree of
synchrony were analytically and numerically computed. However, it is
not known whether there are other PRC shapes that might produce even stronger
stochastic synchronization.
The easiest way to quantify stochastic synchrony
is to examine the Lyapunov exponent, the rate at which two oscillators
receiving identical inputs converge to synchrony. In this paper we will explore how
this quantity depends on the shape of the PRC. In particular, we find
that Type II PRCs lead to faster convergence than do Type I, and
we use variational principles to determinine the optimal shape of the
PRC to maximize this convergence.
First in Section \ref{sec:ito} we introduce the phase reduction of a
stochastically driven neural oscillator using the It\^o change of
variables, and in Section \ref{sec:lyapunov} we derive the Lyapunov
exponent for two such oscillators receiving common noise. Next we use
the Fokker-Planck equation in Section \ref{sec:phase} to obtain the
probability distribution of the phase of a noise-driven neural
oscillator. The Euler-Lagrange method for constrained optimization
allows us in Section \ref{sec:euler-lagrange} to find the PRC that
minimizes the Lyapunov exponent. This leads to a 4th order system of
nonlinear differential equations, which we approximate to an arbitrary
order of accuracy using regular perturbations in Section
\ref{sec:perturbation}. The resulting approximation shows that a Type
II PRC achieves the minimal Lyapunov exponent, hence producing more
robust convergence to synchrony than a Type I PRC. Several interesting
cases that arise as a function of the constraint parameters are
discussed in Section \ref{sec:constraints}. Finally in Section
\ref{sec:numerics} we show that numerical solution of the 4th order
system agrees with the perturbation-derived approximation.
\section{\label{sec:ito} It\^o Phase Reduction}
Consider a neural oscillator with additive white noise decribed by the stochastic differential equation
\begin{equation}
dX = F(X) dt + \sigma M dW,
\label{eqn:SDE}
\end{equation}
where $F(X)$ represents the deterministic equations of motion,
$\sigma$ is the amplitude of the noise, $M$ is a constant matrix, and
$dW$ is a vector of Gaussian
white noise. Note that for a general limit-cycle oscillator,
there need be no constraints on the entries of $M$. For neural models however,
the noise typically occurs in current felt by the neuron, and this current
appears only in the voltage-component of the deterministic
model. Without loss of generality, we take the voltage to be the first
component. Thus, we will assume here that $M$ has all zero entries except
for the $(1,1)$ element, which is identically 1.
The phase reduction method \cite{teramae:2004} applied to
Eq.(\ref{eqn:SDE}) gives a stochastic differential equation for the
evolution of the oscillator's phase:
\begin{equation}
d\theta = dt + \sigma \Delta(\theta) dW,
\label{eqn:Stratonovich}
\end{equation}
where we have assumed without loss of generality that the intrinsic
frequency of the oscillator is $\omega = 1$, and $dW$ is now a scalar
white noise process. Here $\Delta$ is the
infinitesimal phase response curve defined by
\[
\Delta(\theta) := \nabla_X \theta \Big\vert_{X=X_0(\theta)}
\]
where $X_0(\theta)$ is the unperturbed limit-cycle solution of the
deterministic equation $\dot{X} = F(X)$. See Kuramoto \cite{kuramoto:1984}, pages 26-27.
It is now important to note that the usual phase reduction method uses
the conventional change of variables, so Eq.(\ref{eqn:Stratonovich})
must be regarded as a Stratonovich differential equation
\cite{teramae:2004, horsthemke:1984}. To eliminate the
correlation between $\theta$ and the white noise $\xi=dW$, we must
apply It\^{o}'s Lemma to obtain an equivalent but analytically more
convenient formulation
\begin{equation}
d\theta = \left[ 1 + \frac{\sigma ^2}{2} \Delta'(\theta) \Delta(\theta)\right]dt + \sigma \Delta(\theta) dW,
\label{eqn:Ito}
\end{equation}
where $'$ denotes $\frac{\partial}{\partial \theta}$.
In a recent paper, Yoshimura and Arai \cite{yoshi} show that Eq.(\ref{eqn:Ito}) is incomplete and that another term must be added in
the case where the noise is strictly white. However, more recently (in
preparation) we show that the correct reduction is more subtle, and
under some reasonable circumstances the additional term can be made
arbitrarily small. Thus we will stay with the conventional
phase-reduced model as first proposed by Teramae and Tanaka \cite{teramae:2004}.
\section{\label{sec:lyapunov} Lyapunov Exponent}
As a standard measure of susceptibility to synchrony, we will now
derive the Lyapunov exponent for two identical uncoupled neural
oscillators receiving common additive white noise. The resulting
analysis, however, applies equally well to an arbitrary number of
identical noninteracting oscillators.
Let us define the phase difference $\phi := \theta_2-\theta_1$, where
$\theta_1$ and $\theta_2$ each obey Eq.(\ref{eqn:Ito}). Linearizing
around the synchronous state $\phi=0$, we obtain as in \cite{teramae:2004}:
\[
d\phi = \frac{\sigma ^2}{2} \left[(\Delta'\Delta)'(\theta)\phi\right]dt + \sigma [ \Delta'(\theta)\phi] dW,
\]
where $\theta$ obeys Eq.(\ref{eqn:Ito}) as well. Since the Lyapunov exponent
is defined as $\lambda := \lim_{t\rightarrow \infty}
\frac{\log(\phi(t))}{t}$, let us make the change of variables
$y:=\log(\phi)$. Once again we invoke It\^{o}'s Lemma, and after
simplification we find that $y$ satisfies the stochastic differential
equation
\[
dy = \frac{\sigma^2}{2}[\Delta''\Delta] dt + \sigma \Delta' dW.
\]
Next we integrate, divide by $t$ and take the limit as $t\to\infty$
to obtain an expression for $\lambda$.
\begin{eqnarray*}
\lambda &=& \lim_{t\rightarrow \infty}\frac{y(t)}{t} \\
&=& \lim_{t\rightarrow \infty} \frac{\sigma^2}{2t}\int_0^t \Delta''(\theta(s))\Delta(\theta(s))ds \\
&& + \frac{\sigma}{t} \int_0^t \Delta'(\theta(s)) dW(s)
\end{eqnarray*}
Assuming the system is ergodic, we can replace the long time average
on the right hand side with the spatial or ensemble average. Due to
the It\^{o} change of variables, the last term drops out leaving
\begin{equation}
\lambda = \frac{\sigma^2}{2}\int_0^1 \Delta''(\theta) \Delta(\theta) P(\theta) d\theta,
\label{eqn:lambda}
\end{equation}
where $P(\theta)$ is the steady-state distribution of the phase.
Note that Teramae and Tanaka derive an expression for $\lambda$ in \cite{teramae:2004} by making the approximation $P(\theta)=1$. Substituting this value into Eq.(\ref{eqn:lambda}) and performing integration by parts, they obtain
\[
\lambda \approx - \frac{\sigma^2}{2}\int_0^1(\Delta'(\theta))^2 d\theta.
\]
In this paper, however, we wish to retain the generality of $P(\theta)$ as discussed below.
\section{\label{sec:phase} Steady-State Phase Distribution}
In order to evaluate the Lyapunov exponent, we need to
obtain the stationary density of the phase when perturbed by
noise. Teramae and Tanaka \cite{teramae:2004} have treated the density as uniform,
which is correct for weak noise. However our subsequent perturbation analysis will require higher-order terms, so we will need to derive a more accurate value for the steady-state phase distribution.
By applying the Fokker-Planck equation to (\ref{eqn:Ito}), we obtain
after simplification a partial differential equation for the
probability distribution $P(\theta,t)$:
\[
\frac{\partial P}{\partial t} = -\frac{\partial P}{\partial \theta} + \frac{\sigma^2}{2}\frac{\partial }{\partial \theta} \left[\Delta \frac{\partial (\Delta P)}{\partial \theta}\right].
\]
Now we may set $\frac{\partial P}{\partial t} = 0$ to find the steady state, then integrate once with respect to $\theta$ to obtain:
\begin{equation}
-J = - P + \frac{\sigma^2}{2} \left[\Delta \frac{\partial (\Delta P)}{\partial \theta}\right],
\label{eqn:steadystate}
\end{equation}
where $-J$ is a constant of integration. We require that $P(0)=P(1)$
and that the solution be normalized, namely $\int_0^1P(\theta)\ d\theta=1.$ Note that the equations are singular, since $\Delta(\theta)$ generally vanishes at several
places, in particular at $\theta=0, 1$. In the appendix below, we prove the
existence of the stationary density by directly solving the linear equations
and taking appropriate limits.
In the remainder of this section, we use regular perturbation theory to approximate the
stationary density for small noise, $0<\sigma\ll 1.$
To approximate both $J$ and $P$ we substitute
\begin{eqnarray*}
J &=& 1 + \sigma^2 J_1 + \sigma^4J_2 + \cdots \\
P(\theta) &=& 1 + \sigma^2 P_1(\theta) + \sigma^4 P_2(\theta) + \cdots
\end{eqnarray*}
into equation (\ref{eqn:steadystate}).
Equating like powers of $\sigma$ gives
\[
-J_1 = -P_1(\theta) + \frac{1}{2}\Delta(\theta) \Delta'(\theta).
\]
Integrating both sides over $[0,1]$ leaves the constant on the
left hand side unchanged. For the right hand side,
note that $\int_0^1 P(\theta) d\theta = 1$, and
hence $\int_0^1 P_1(\theta) d\theta = 0$. Furthermore,
$\Delta\Delta' = \frac{1}{2}\frac{d}{d\theta}(\Delta^2)$ so that
\begin{eqnarray*}
J_1 &=& -\frac{1}{4}(\Delta(1)^2-\Delta(0)^2) \\
&=& 0,
\end{eqnarray*}
since $\Delta$ is periodic. Thus we have $P_1(\theta) = \frac{1}{2}\Delta(\theta) \Delta'(\theta)$.
Similarly,
\[
-J_2 = - P_2(\theta) + \frac{1}{2}\Delta(\theta)^2\Delta'(\theta)^2 + \frac{1}{4}\Delta(\theta)^3\Delta''(\theta).
\]
Since $\int_0^1 P_2(\theta) d\theta = 0$ as well, we can integrate both sides as above and use integration by parts to obtain
\begin{eqnarray*}
J_2 &=& \frac{1}{4}\int_0^1(\Delta(\theta)\Delta'(\theta))^2 d\theta \\
P_2(\theta) &=& \frac{1}{2}\Delta(\theta)^2\Delta'(\theta)^2 + \frac{1}{4}\Delta(\theta)^3\Delta''(\theta) \\
&& + \frac{1}{4}\int_0^1(\Delta(\theta)\Delta'(\theta))^2 d\theta.
\end{eqnarray*}
In summary,
\begin{eqnarray}
\nonumber
J &=& 1 + \frac{\sigma^4}{4}\int_0^1(\Delta(\theta)\Delta'(\theta))^2 d\theta \\
\nonumber
P(\theta) &=& 1 + \frac{\sigma^2}{2}\Delta(\theta) \Delta'(\theta) + \frac{\sigma^4}{4} \Bigg[ 2\Delta(\theta)^2\Delta'(\theta)^2 \\
&& + \Delta(\theta)^3\Delta''(\theta) + \int_0^1(\Delta(\theta)\Delta'(\theta))^2 d\theta \Bigg].
\label{eqn:ptosigma^4}
\end{eqnarray}
For the perturbation expansions in the next section, it will suffice to write $J = 1$. We will use Eq.(\ref{eqn:ptosigma^4}) in Section \ref{sec:constraints} and for the numerical verifications in Section \ref{sec:numerics}.
\section{\label{sec:euler-lagrange} Constrained Optimization}
The Euler-Lagrange variational technique provides a method for
determining the phase resetting curve $\Delta$ that minimizes the
Lyapunov exponent, subject to appropriate constraints. To ensure
smooth solutions and to eliminate uninformative harmonics of the
optimal solution, we begin by imposing the general constraint
\begin{equation}
\int_0^1 a(\Delta(\theta))^2 + b(\Delta'(\theta))^2 +c(\Delta''(\theta))^2 d\theta = 1,
\label{eqn:constraint}
\end{equation}
where $a$, $b$ and $c$ are free parameters. A standard normalization
has $a=1,b=0,c=0$, but non-zero values of $b$, $c$ endow solutions with
additional smoothness. Below we will explore the
cases that arise from specific choices of these.
We proceed by placing Eqs.(\ref{eqn:lambda}), (\ref{eqn:steadystate}) and (\ref{eqn:constraint}) together with the approximation $J=1$ into the Euler-Lagrange formula to obtain the functional
\begin{eqnarray}
&\int_0^1 &\Delta'' \Delta P + \nu_1 \left[ a\Delta^2 + b(\Delta')^2 +c(\Delta'')^2 -1\right] \nonumber \\
&& + \nu_2(\theta) \left[1 -P + \frac{\sigma^2}{2} \Delta (\Delta P)' \right] d\theta = 0,
\label{eqn:EL-functional}
\end{eqnarray}
where $\nu_1$ is a free parameter, and $\nu_2(\theta)$ represents a continuum of free parameters.
Define the operator
\begin{eqnarray*}
\mathcal{L}(\Delta) &:=& \Delta'' \Delta P + \nu_1 \left[ a\Delta^2 + b(\Delta')^2 +c(\Delta'')^2-1 \right] \\
&& + \nu_2(\theta) \left[ 1 - P + \frac{\sigma^2}{2} \Delta (\Delta P)' \right].
\end{eqnarray*}
The optimal $\Delta$ we seek will satisfy the two equations
\begin{eqnarray}
\frac{\partial \mathcal{L}}{\partial \Delta} - \frac{d}{d\theta} \frac{\partial \mathcal{L}}{\partial \Delta'} + \frac{d^2}{d\theta^2}\frac{\partial L}{\partial \Delta''} &=& 0 \label{eqn:EL1} \\
\frac{\partial \mathcal{L}}{\partial P} - \frac{d}{d\theta}\frac{\partial \mathcal{L}}{\partial P'} &=& 0. \label{eqn:EL2}
\end{eqnarray}
Note that we can write two more Euler-Lagrange equations, but $\frac{\partial \mathcal{L}}{\partial \nu_1} = 0$ simply restates Eq.(\ref{eqn:constraint}), and $\frac{\partial \mathcal{L}}{\partial \nu_2} = 0$ returns Eq.(\ref{eqn:steadystate}) governing $P$.
Assuming the parameter $c$ is nonzero, we obtain from Eqs.(\ref{eqn:EL1}) and (\ref{eqn:EL2}) a 4th order system of ordinary differential equations:
\begin{widetext}
\begin{eqnarray}
P'' \Delta + 2(P' \Delta' + P \Delta'' + a \Delta \nu_1 - b \Delta'' \nu_1 + c \Delta^{(4)} \nu_1) + \frac{1}{2} \Delta (P' \nu_2 - P \nu_2') \sigma^2 &=& 0 \label{eqn:4th-ode1} \\
\Delta \Delta'' - \nu_2 - \frac{1}{2} \Delta (\Delta' \nu_2 + \Delta \nu_2') \sigma^2 &=& 0.
\label{eqn:4th-ode2}
\end{eqnarray}
\end{widetext}
If $c = 0$, we will have instead the 2nd order system which
obtains by setting $c = 0$ in Eq.(\ref{eqn:4th-ode1}).
When we examine the effects of varying the constraint
parameters in Section \ref{sec:constraints}, we will see that
the main result remains the same in this case as well.
\section{\label{sec:perturbation} Perturbation Approximation}
Let us first consider the 4th order case where the parameter $c$ is
nonzero.
Assuming the noise amplitude $\sigma$ is sufficiently small, we write the following expansions
\begin{eqnarray}
P (\theta) &=& P_0 (\theta) + \sigma^2 P_1 (\theta) + ... \nonumber \\
\Delta (\theta) &=& \Delta_0 (\theta) + \sigma^2 \Delta_1 (\theta) + ...\label{eqn:perturbation} \\
\nu_1 &=& \nu_{1,0} + \sigma^2 \nu_{1,1} + ... \nonumber \\
\nu_2 (\theta) &=& \nu_{2,0} (\theta) + \sigma^2 \nu_{2,1} (\theta)+ ... \nonumber
\end{eqnarray}
Substituting these into (\ref{eqn:4th-ode1}) and (\ref{eqn:4th-ode2}) and equating like powers of $\sigma$ gives to lowest order: $P_0(\theta) = 1$, $\nu_{2,0}(\theta) = \Delta_0(\theta) \Delta_0''(\theta)$ and the fourth order homogeneous equation
\begin{equation}
a\nu_{1,0} \Delta_0 + (1-b\nu_{1,0})\Delta_0''+c\nu_{1,0}\Delta_0^{(4)} = 0.
\label{eqn:perturbation0}
\end{equation}
For convenience let us define the differential operator
\[
\mathcal{J} = a\nu_{1,0} + (1-b\nu_{1,0})\frac{\partial^2}{\partial\theta^2} + c\nu_{1,0}\frac{\partial^4}{\partial\theta^4}.
\]
Thus Eq.(\ref{eqn:perturbation0}) becomes $\mathcal{J}(\Delta_0) = 0$, and the first order correction $\Delta_1$ obeys the inhomogeneous equation
\begin{eqnarray}
\mathcal{J}(\Delta_1)&=& (\Delta_0')^3 - b\nu_{1,1}\Delta_0'' \nonumber \\
&& + \Delta_0(a\nu_{1,1} + 3\Delta_0'\Delta_0'') + c\nu_{1,1}\Delta_0^{(4)}.
\label{eqn:perturbation1}
\end{eqnarray}
Furthermore, substituting the expansions (\ref{eqn:perturbation}) into Eq.(\ref{eqn:constraint}) gives the corresponding constraints:
\begin{eqnarray}
\int_0^1 a\Delta_0^2 + b(\Delta_0')^2+c(\Delta_0'')^2 &=& 1 \label{eqn:constraint0}\\
\int_0^1 a\Delta_0\Delta_1 + b\Delta_0'\Delta_1' + c \Delta_0''\Delta_1'' &=& 0.
\label{eqn:constraint1}
\end{eqnarray}
\begin{figure*}
\includegraphics[width=5in,height=3.75in]{fig1.eps}
\caption{\label{fig:deformation} In the case where the second derivative is left unconstrained, the optimal PRC deviates from a pure cosine function as the noise amplitude $\sigma$ increases. Parameters are a=1, b=1, c=0.}
\end{figure*}
Before solving Eq.(\ref{eqn:perturbation0}), we must first determine
the unknown parameter $\nu_{1,0}$. Since we seek only periodic
solutions, we can impose a condition on the characteristic equation of
(\ref{eqn:perturbation0}):
\begin{equation}
a\nu_{1,0} + (1-b\nu_{1,0})y^2 + c\nu_{1,0} y^4 = 0.
\label{eqn:poly}
\end{equation}
Specifically, by requiring that the roots of this polynomial satisfy $y = 2\pi i$, we determine that
\[
\nu_{1,0} = \frac{4\pi^2}{a+4b\pi^2+16c\pi^4}.
\]
Now we are ready to impose periodic boundary conditions, and we find that the solution of (\ref{eqn:perturbation0}) is just $\Delta_0(\theta) = C_0 \sin(2\pi\theta)$. The constant of integration $C_0$ is determined from the constraint (\ref{eqn:constraint0}) so that
\[
C_0 = \pm\frac{\sqrt{2}}{\sqrt{a+4b\pi^2+16c\pi^4}}.
\]
While both values of $C_0$ will give the same minimal value of the Lyapunov exponent, we choose the negative value for biological plausibility. Hence to lowest order we find the optimal phase resetting curve is Type II:
\begin{equation}
\Delta_0(\theta) = -\frac{\sqrt{2}\sin(2\pi\theta)}{\sqrt{a+4b\pi^2+16c\pi^4}}.
\label{eqn:delta0}
\end{equation}
The next order correction does not appreciably change this result. To obtain the $\sigma^2$ term, we must solve (\ref{eqn:perturbation1}) subject to (\ref{eqn:constraint1}). By the Fredholm Alternative, a solution to the inhomogeneous problem exists if and only if the right-hand side of (\ref{eqn:perturbation1}), call it $r(\theta)$, is orthogonal to the nullspace of $\mathcal{J^*}$. However, since $\mathcal{J}$ is self-adjoint we simply solve for the value of $\nu_{1,1}$ such that
\[
\int_0^1 \sin(2\pi\theta) r(\theta) d\theta = 0,
\]
namely, $\nu_{1,1} = 0$.
Imposing periodic boundary conditions on the resulting equation yields the solution
\[
\Delta_1(\theta) = C_1 \sin(2\pi\theta) + \frac{\sqrt{2}\pi\sin(2\pi\theta)\sin(4\pi\theta)}{(a-144c\pi^4)\sqrt{a+4b\pi^2+16c\pi^4}}.
\]
As before, we use the constraint (\ref{eqn:constraint1}) to obtain $C_1 = 0$. Hence to order $\sigma^2$ the optimal phase resetting curve is given by
\begin{eqnarray}
\Delta(\theta) &=& -\frac{\sqrt{2}\sin(2\pi\theta)}{\sqrt{a+4b\pi^2+16c\pi^4}} \nonumber \\
&& + \frac{\sigma^2}{2} \frac{\sqrt{2}\pi\sin(2\pi\theta)\sin(4\pi\theta)}{(a-144c\pi^4)\sqrt{a+4b\pi^2+16c\pi^4}}.
\label{eqn:optimal}
\end{eqnarray}
\section{\label{sec:constraints} Constraint Parameters}
\begin{figure*}
\centering
\includegraphics[width=5in,height=3.75in]{fig2.eps}
\caption{\label{fig:family} When the first derivative is unconstrained while the second derivative is constrained, Euler-Lagrange optimization produces a family of candidates for the minimizer of the Lyapunov exponent ranging smoothly from Type II to Type I as the parameter $K$ ranges from 0 to 1. For negative $K$ (dashed), the curves do not represent biologically plausible PRCs. Parameters are $a=0,b=1,c=1$.}
\end{figure*}
Let us next explore the influence of the constraint parameters $a$,
$b$ and $c$, which we will allow to take on the values of $0$ or
$1$. Of the seven nontrivial combinations, one has no periodic
solution at all and is thus inadmissible. Four parameter choices give
rise to the same optimum already found in Eq.(\ref{eqn:optimal}), and
two parameter combinations do not produce a unique solution but
instead yield a family of solutions ranging smoothly from Type I to
Type II. In this case, we explicitly find the minimizer of $\lambda$
among the family of solutions.
All of the cases can be analyzed by examining Eq.(\ref{eqn:poly}), the
characteristic equation of $\mathcal{L}(\Delta)=0$. For example, the
case $a=c=0$ and $b=1$ can have no periodic solution, since the
polynomial $(1-\nu_{1,0})y^2 = 0$ has no nontrivial roots.
The four parameter combinations that lead to Eq.(\ref{eqn:optimal})
are those in which $a = 1$. In these cases we have
\[
\nu_{1,0} + (1-b\nu_{1,0})y^2 + c\nu_{1,0} y^4 = 0.
\]
If $c \neq 0$, the polynomial is 4th degree having four distinct roots; if $c = 0$ the polynomial is quadratic with two distinct roots. In each case we can set $y = 2\pi i$ and solve uniquely for $\nu_{1,0}$ as discussed above.
The case $c=0$ (while $a=1$) deserves further attention for
another reason. In this regime, the optimal PRC becomes sensitive to
the noise amplitude $\sigma$ as illustrated in
Fig.(\ref{fig:deformation}). To understand why the curve deforms, let
us focus on the extrema of Eq.(\ref{eqn:optimal}), which are given by
the zeros of the derivative:
\begin{eqnarray*}
\Delta'(\theta) &=& -\frac{2\sqrt{2}\pi}{\sqrt{a+4b\pi^2+16c\pi^4}} \Bigg[\cos(2\pi \theta) \\
&& + \frac{\sigma^2\pi}{a-144c\pi^4}\Bigg(\cos(4\pi\theta)\sin(2\pi\theta) \\
&& + \frac{1}{2}\cos(2\pi\theta)\sin(4\pi\theta)\Bigg)\Bigg]
\end{eqnarray*}
In this form we clearly see that the unperturbed extrema (when $\sigma = 0$) occur at
$\theta=1/4$ and $3/4$, while deformation due to noise is on the order
of $\sigma^2\pi/(a-144c\pi^4)$. More specifically, when $c\neq0$ this
quantity is $\mathcal{O}(\sigma^2 10^{-4})$ so that the weak noise in our model
($\sigma\ll 1$) has negligible effect. However when $c=0$, this
quantity is $\mathcal{O}(\sigma^2)$, so that even relatively small magnitude
noise can have a noticible impact on the shape of the optimal PRC.
Another interesting situation arises in the two cases where $a = 0$,
$c = 1$ and $b$ is arbitrary. Here the characteristic equation has a
double root at $y=0$:
\[
(1-b\nu_{1,0})y^2 + \nu_{1,0} y^4 = 0.
\]
After accounting for the boundary conditions, we have a superposition of two independent solutions
\[
\Delta_0(\theta) = C_3(1-\cos(2\pi \theta)) + C_4\sin(2\pi \theta).
\]
The constraint (\ref{eqn:constraint0}) eliminates only one degree of freedom, leaving a family of solutions as candidates for the optimum:
\begin{eqnarray}
\Delta_0(\theta) &=& K \frac{1-\cos(2\pi \theta)}{\sqrt{2\pi^2 (b+4\pi^2)}} \nonumber \\
&& - \sqrt{1-K^2}\frac{\sin(2\pi \theta)}{\sqrt{2\pi^2 (b+4\pi^2)}},
\label{eqn:family}
\end{eqnarray}
where the remaining degree of freedom $K$ has been normalized to range between $-1$ and $1$. See Fig.(\ref{fig:family}).
Combining Eq.(\ref{eqn:lambda}) for the Lyapunov exponent with Eq.(\ref{eqn:ptosigma^4}) for the steady-state phase distribution, we insert Eq.(\ref{eqn:family}) to obtain the following expression:
\[
\lambda = -\frac{1}{b+4\pi^2} + \frac{\sigma^4}{4} \frac{(4K^4+10K^2+1)}{4\pi^2(b+4\pi^2)^3},
\]
where we have set $a=0,c=1$. Note that we needed to carry out the expansion of $\lambda$ to $\sigma^4$ in order to discover the dependence on $K$.
Since the derivative of $\lambda$ with respect to $K$ has only one real root at $K=0$, where a minimum occurs, the Type II curve remains the optimal PRC even in this case.
\begin{figure*}
\centering
\includegraphics[width=5in,height=3.75in]{fig3.eps}
\caption{\label{fig:abc} The magnitude of the optimal PRC depends on the whether or not the second derivative is constrained. The numerical solution (open circles) and the analytic result (solid lines) coincide. Parameters are $a=1,b=1$ and $\sigma = 0.05$.}
\end{figure*}
\section{\label{sec:numerics} Numerical Verification}
We would like to independently verify the accuracy of the optimal PRC (\ref{eqn:optimal})
derived via perturbation expansion by
numerically solving the Euler-Lagrange equations (\ref{eqn:4th-ode1})
and (\ref{eqn:4th-ode2}) with periodic boundary
conditions. Unfortunately, the resulting system is singular and
therefore very difficult to solve numerically. Instead we substitute the
approximation $P(\theta) = 1 + \frac{\sigma^2}{2}\Delta(\theta)
\Delta'(\theta)$ into the Euler-Lagrange functional
(\ref{eqn:EL-functional}) to obtain a new functional
\begin{eqnarray*}
&\int_0^1& \Delta'' \Delta \left(1 + \frac{\sigma^2}{2}\Delta(\theta) \Delta'(\theta)\right) \\
&& + \nu_1 \left[ a\Delta^2 + b(\Delta')^2 +c(\Delta'')^2 -1\right] d\theta = 0,
\label{eqn:num-functional}
\end{eqnarray*}
which gives rise via Eq.(\ref{eqn:EL1}) to the 4th order boundary value problem
\[
\Delta^{(4)} = \frac{-2\Delta''-2a\Delta\nu_1 + 2b\Delta'' \nu_1 - \Delta'^3 \sigma^2 - 3 \Delta\Delta'\Delta'' \sigma^2}{2 c \nu_1}.
\]
When $c=0$, we similarly obtain a 2nd order boundary value problem.
Using the numerical integration package XPPAUT, we are able to achieve
excellent agreement with our analytical
approximation. Fig.(\ref{fig:abc}) illustrates numerical and analytic
solutions in the case where $c=1$ and where $c=0$. Note that imposing
a constraint on the second derivative of $\Delta$ results in an
optimal PRC of much smaller magnitude.
\begin{figure*}
\centering
\includegraphics[width=5in,height=3.75in]{fig4.eps}
\caption{\label{fig:ab0} When the second derivative is unconstrained, the optimal PRC shape deforms with increasing noise. The numerical solution (open circles) and the analytic result (solid lines) are in good agreement. Parameters are $a=1,b=1,c=0$.}
\end{figure*}
In Fig.(\ref{fig:ab0}) we find good agreement between the analytic
and numerical results even for the regime in which $a=1$, $c=0$ and PRC
shape is sensitive to noise amplitude. The numerical simulation
deforms with increasing $\sigma$ just as the analytic
approximation does.
\section*{Discussion}
In this paper we have used perturbation theory and the calculus of
variations to analyze the rate at which neurons can synchronize when
subjected to common inputs. We treat the inputs as ``noise,'' that
is, as if they are delta-correlated with no structure. Real neuronal
inputs do have correlational structure, however, so that the expression for the rate of
synchronization (the Lyapunov exponent) is more complex.
Indeed, in previous work \cite{galanfp} we have shown that
the temporal characteristics of the noise can also have an effect on
how rapidly neurons synchronize. In that work, we asked the reverse
question: given a particular PRC, what correlation time for the noise
minimizes the Lyapunov exponent?
Suppose that we use some signal that
is not white noise but still has zero mean and is stationary. Then the
phase satisfies
\[
\frac{d\theta}{dt} = 1 + \Delta(\theta)\xi(t)
\]
where $\xi(t)$ is the input. The Lyapunov exponent is
\[
\lambda:= \lim_{T\to\infty} \frac{1}{T}\int_0^T
\Delta'(\theta(t))\xi(t)\ dt.
\]
By using an approximation of $\theta(t)$ as in \cite{galan-optima} we
may be able to obtain a functional for $\lambda$ depending on $\xi(t)$ and
$\Delta$, and from this apply similar methods to estimate the optimal
shape of the PRC given the statistics of the inputs.
Optimization has been applied to other aspects of neural
oscillators. Moehlis, et al. \cite{moehlis} asked the following
question. Consider the scalar oscillator model:
\[
\frac{d\theta}{dt} = f(\theta) + \Delta(\theta)I(t).
\]
(Note that if $f(\theta)=1$, we have Eq.(\ref{eqn:Stratonovich}), the case considered in this paper.) Suppose the neuron fired at $t=0$ and we
desire it to fire again at time $T>0$. What is the minimum stimulus,
$I(t)$ (which, say, minimizes $\int_0^T I(t)^2 dt$) to do this? Moehlis, et al. \cite{moehlis}
write the Euler-Lagrange equations for this optimization problem and
then assume that $I(t)$ is small in order to use perturbation
methods. A related issue is the ``optimal stimulus'' \cite{rieke} for
producing a spike in a neuron, and for neural
oscillators this has been answered in \cite{ermentrout}.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 1,923
|
Q: How can I fix my CLion installation? I can successfully run /opt/clion-2017.2.1/bin/clion.sh from where I installed it but how do I update what is shown in the launcher? In the launcher it still links to the old 2017.1.1.
A: From the Tools menu of the new installation create new desktop and command line launcher and accept to override the old ones when asked by the process. This will override the former configurations and point to the knew configuration for the update.
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 4,193
|
Out in Church: Summer Progress Expected
June 6, 2010 — Terence
Change is coming, of that there can be no doubt. (Sadly, in spite of the suggestion by Cardinal Schonborn, I am not here referring to the Catholic Church, but to others. Just how long Rome can lag behind, is another matter.)
First, consider the progress up to now. The Unitarian Universalist Association and the United Church of Christ accepted full inclusion some time ago, and have many openly gay, lesbian and trans pastors. Last year, in a blaze of publicity, the ELCA voted to approve the recognition of lesbian or gay partners in committed and faithful permanent relationships, without the expectation of celibacy, on exactly the same basis as heterosexual marriages. That decision was accompanied by many dire warnings of doom, and predictions that many congregations would secede in protest. There have been some withdrawals, but as far as I can tell, not too many. Meanwhile, life in the ELCA has continued as normal, and the decision came into practical effect earlier this year.
In the Episcopalian/ Anglican communion, things are more complex, with international ties and allegiances also coming into play. The US Episcopalians have now consecrated two openly gay or lesbian bishops, and have nominated but not approved a few others. In the UK, that decision has been highly controversial, but is closely tied up with controversy over women bishops, which (unlike the US) have hitherto not been permitted. British Quakers resolved last year bless same sex couples in church, and were influential in the recent change in the law to permit civil partnerships in religious premises. Meanwhile, the Swedish Lutheran church has also ordained a lesbian bishop, and has agreed to apply the new law on same sex marriage in church as well. The Danish and Icelandic Lutherans are considering following their Swedish counterparts in applying their own countries' laws on gay marriage when they are approved by parliament.
In the US and Europe, therefore, progress to full inclusion in church is substantial, at least in American mainline Protestant denominations and their European counterparts. Where can we expect the next victories? With the summer assembly season approaching, these are the major things we should be looking for.
For the American Presbyterian Church, progress has been steady, with repeated discussions and votes at the biennial assembly. Four years ago, a motion for inclusion failed. Two years ago, a similar resolution was approved by the assembly, but failed to gain the necessary ratification by local congregations. This year, the motion is up for a vote again. It is likely to pass again, probably with a bigger majority. The test, once again, will come with the push for local ratification. That will take time, but in the meantime, the size of the majority at assembly will give the first indication of the chances of success later this year. (Even if ratification fails, I would think that any defeat would be only temporary. Full inclusion in the Presbyterian Church, following that already approved for the UU Association, United Congregational and ELCA churches, is only a matter of time.) For more detail on the Presbyterians (and ELCA), see the helpful fuller report by Olbie Holman at "Spirit of a Liberal", or read the blow by blow reports at "More Light Presbyterians".
The United Methodists have been following in much the same path as the Presbyterians, with a parallel effort on the ground by the Reconciling Ministries Network to win over local congregations. I have been struggling to find any up to date information on the Methodists, except that the 2008 decision was to affirm the traditional stance to affirm the sacred worth of homosexual people, but to refuse approval of homosexual actions, or to ordain homosexual clergy. A decision last year to enter full communion with the ELCA will put some strain on that, but for now the agreement specifically excludes recognition of the ECLA openly gay clergy.
While the US Episcopalians' debates have been over the controversial ordination of bishops, there has been very little reported on gay marriage in church, although there have been reports of some local jurisdictions going ahead and implementing same sex church weddings where state law allows it. In Canada, the Episcopalian balance is the reverse. There have been no openly gay bishops, but approval for same sex "blessings" (not full church weddings) could be close. I wasn't aware of it previously, but at last national gathering of the Anglican Church of Canada, a majority of delegates agreed to allow individual dioceses to allow church blessing of same sex couples. The decision could not take effect though, as the requirement is for approval by separate majorities of each of the laity, clergy, and bishops. Laity and clergy voted in favour, but the bishops narrowly against. This year, the aim is to "defuse" the issue, which I take as code for allowing some compromise by the bishops after extensive discussion. I would expect approval to come. Later, I would expect local approval to expand nationally, and for acceptance of "blessings" to broaden in time to include full marriage.
"This synod, we're approaching it not so much from the form of a winner-takes-all resolution," he said. "We're approaching it from a kind of conversational route and hoping by the end of synod, after numerous conversations and meetings . . . that the synod may come up with a statement on where the church is on this matter."
At the last national gathering in Winnipeg in 2007, a divided Anglican Church of Canada ultimately decided on a technicality not to give individual dioceses the right to decide for themselves whether priests should be allowed to bless same-sex unions.
The majority of the 300 delegates agreed to the idea, but church law requires separate majorities among priests, laity and bishops.
Priests and laity approved the change. However, Canada's 40 bishops voted it down with a two-person majority, unleashing both outrage and sighs of relief.
Two dioceses have since interpreted the vote as a draw and are now blessing same sex unions, while others are considering the idea.
In the UK, church recognition for openly gay bishops and church blessings will not happen yet, as they are still mired in the wider conflicts over women bishops and international controversies. However, unexpected progress on two other fronts could have a marked impact over the longer term. First, the controversy over women bishops earlier this year ended with clear and unqualified approval for women bishops, without any attempt to placate the misogynists with any kind of special provision for those unwilling to accept women (this had been part of an earlier compromise proposal). It will still take a while for this to come into effect, but as it does, the inevitable shift in the balance of participants in future discussions, as well as the moral example, will surely make it easier for future conferences to consider and approve movement on gay bishops, clergy and blessings. Another development this week will work in the same way. Totally unexpectedly, the decision has just been announced that in the future, divorced clergy will be eligible for nomination as bishops, which has been an absolute no-no.
This has angered some conservatives, but like the decision on women bishops, their defeat here will only weaken further their ability to hold on on LGBT inclusion.
Critics described the change in Church rules as "utterly unacceptable" and warned it would undermine the biblical teaching that marriage is for life.
So: in leading Protestant denominations, in Europe and in the US, progress has already been significant, with further progress confidently expected for this summer and the years immediately ahead. What of the Catholic church? Well, we all know how actively they have lobbied in every legal battle over gay marriage, gay adoption, and especially anything that even appears to suggest abortion. However, for a really stark illustration of the contrast between the Catholic Church and those more firmly rooted in reality, we need to go to the other side of the world, where the bishops are locked in a titanic battle with government to forestall the expansion of access to …………
Contraceptives!
Church leaders in the devoutly Catholic country immediately hit back at the plan, saying the state had no business to be talking to youngsters about sex.
"The students should not be taught … sexuality based on the physical aspect but as a gift from God," said Monsignor Pedro Quitorio, a spokesman for the country's Catholic bishops, adding that sex should be taught by parents.
Archbishop Oscar Cruz said that instead of sex, boys should be taught how to respect girls and women while girls should be taught self defence and how to avoid becoming victims of sexual abuse.
LGBT Inclusion in Church: A Study in Contrasts (queertheology.blogspot.com)
Gay Clergy: God's Spirit at Work (Huff Post) (queertheology.blogspot.com)
Gay Marriage: Coming (Soon?) to a Church Near You. (queering-the-church.com)
In Celebration: Rev Jane Spahr, "Lesbyterian" (queerhistory.blogspot.com)
Texas Baptist Church Takes a Stand FOR Gay Members. (queertheology.blogspot.com)
Posted in Gay Christians. Tags: Catholic Church, Episcopal church, Evangelical Lutheran Church in America, gay inclusion, LGBT, Out in Church, Presbyterianism, Reconciling Ministries Network, Unitarian Universalist Association, United Church of Christ. Leave a Comment »
« The Shower of Stoles Project.
Gay Popes, Papal Sodomites »
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 6,181
|
<div>
Inner Action 2 template
</div>
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 2,968
|
Also by Barrie G. James
The Little Black Book of Pharma Marketing
Trojan Horse: The Ultimate Challenge to Western Industry
Business Wargames
The Marketing of Generic Drugs
The Future of the Multinational Pharmaceutical Industry
## First published in Great Britain in 2009 by
## Pen & Sword Aviation
## an imprint of
## Pen & Sword Books Ltd
## 47 Church Street
## Barnsley
## South Yorkshire
## S70 2AS
## Copyright © Barrie G. James, 2009
## ISBN 978 1 84484 090 4
## eISBN 9781844688227
## The right of Barrie G. James to be identified as Author of this Work has been asserted by him in accordance with the Copyright, Designs and Patents Act 1988.
## A CIP catalogue record for this book is available from the British Library
## All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording or by any information storage and retrieval system, without permission from the Publisher in writing.
## Typeset in Sabon by Phoenix Typesetting, Auldgirth, Dumfriesshire
## Printed and bound in the UK by the MPG Books Group
## Pen & Sword Books Ltd incorporates the Imprints of Pen & Sword Aviation, Pen & Sword Maritime, Pen & Sword Military, Wharncliffe Local History, Pen & Sword Select, Pen & Sword Military Classics and Leo Cooper.
## For a complete list of Pen & Sword titles please contact
## PEN & SWORD BOOKS LIMITED
## 47 Church Street, Barnsley, South Yorkshire, S70 2AS, England
## E-mail: enquiries@pen-and-sword.co.uk
## Website: www.pen-and-sword.co.uk
Previews of Hitler's Gulf War
"What a fantastic and enjoyable book, fast moving, entertaining and exciting. I read this in a single dose and loved it. It reads like a thriller yet is a tale of courage, bluff and daring the like of which is rarely seen anymore. The characters that populate the tale are really from another era of heroes from a desperate yet more noble age and the daring of senior leaders causes me to reflect that we have few like this, if any, nowadays. I found it special as I was in Habbaniya often, the last time being March 2008. I was aware of the incident but not the details. I look forward to returning and telling anyone who will stop and listen – read Hitler's Gulf War!"
Colonel Tim Collins OBE, Gulf War 2003 (the inspirational Eve of Battle speech in Kuwait in 2003, a copy of which reputedly hung on the wall of the Oval Office in the Bush White House.)
"Twice in the last twenty years, British troops have been active in a Gulf war. I was Prime Minister during the first war in 1991. I saw events as they unfolded and marvelled at the skill and courage shown by our armed forces. Barrie James reverts to an earlier conflict against a more deadly enemy. It is a stirring story of courage against the odds that would grace the pages of any work of fiction – yet this is all fact. It tells the story of a forgotten war that deserves to be remembered and hopefully, through these pages, it now will be. I thoroughly enjoyed it."
The Rt. Hon. Sir John Major KG, CH Gulf War 1991
"Intrigue, politics, compelling characters, chaos and courage. Barrie James brings Iraq's history alive, revealing the best and worst of human nature and how individual acts of heroism and bluff combined with tenacious spirit and individual choice determine international events. James is Hercule Poirot in a murder mystery. He deconstructs the whole to make us realise the significance of each clue, each individual foible, each decision demanding us towards the conclusion, compelling us to understand how it all fits together. I read this story in one day and if you want to understand how wars are won and lost, how impossible it is to predict outcomes and gain a personal insight into how timing and one human choice really can tip the balance of history, then read this book. Iraq is still the world's stage and we are all merely players."
Squadron Leader John Peters, Gulf War 1991 (co-author of Tornado Down)
"In the spring of 1941 I was a subaltern in the Royal Horse Guards (The Blues) and a member of 'A' Squadron of the Household Cavalry Regiment which in turn was part of the 4th Cavalry Brigade in Palestine, together with the Wiltshire and Warwickshire Yeomanry. In May 1941 the Brigade, christened 'Habforce', together with the Arab Legion from Jordan and a Battalion of the Essex Regiment, was despatched across the desert between Palestine and Iraq to relieve the siege of the RAF base at Habbaniya. The siege had resulted from a coup by a pro German faction in Iraq under the leadership of Raschid Ali. The story of that siege and that of the British Embassy in Baghdad is recounted in fascinating detail in Dr. James' book. In the course of the relief of both Habbaniya and the British Embassy, two regular divisions of the Iraqi army were defeated by the Brigade group and the valiant efforts of the RAF. I am proud to have taken a small part in the 'Fight for Iraq 1941' and would like to pay tribute to Dr. James for the research he has undertaken to bring a little known but successful conflict to the knowledge of a wider world than that of those who took part in it."
His Grace The Duke of Wellington KG, LVO, OBE, MC, DL, Gulf War 1941
"The British operation in Iraq in 1941 is one of the most interesting and yet least known episodes of the Second World War. It is an extraordinary story of audacity and courage that prompted even Churchill to get excited about 'the arrival of His Majesty's Life Guards and Royal Horse Guards, in armoured cars, across many hundreds of miles of desert'. Barrie James' well-researched account now tells this intriguing story, and he tells it very well, so that this book reads as much like a thriller as it does the very important historical contribution that it undoubtedly is. I enjoyed the book greatly'.
Major General Barney White-Spunner CBE Iraq 2008, GOC HQ, Multinational Division [South East], Basra
Contents
Preface
Acknowledgements
Prologue
Chapter 1 Slippery Slope to War
Chapter 2 Target Habbaniya
Chapter 3 Strike Hard, First
Chapter 4 Shortening the Odds
Chapter 5 The Road to Baghdad
Epilogue
Appendix What Happened to...?
Bibliography
Preface
Hitler's Gulf War is set around the epic story of one of the most remarkable examples of daring in military history and the first real German defeat in the Second World War.
On an isolated, indefensible airfield fifty-five miles from Baghdad a group of poorly armed and outnumbered RAF airmen equipped with obsolete aircraft, together with a few soldiers, outfought the much larger and better-equipped Iraqi forces aided by the Germans and Italians.
It also tells the story of how trucks, taxis, buses and antiquated armoured cars were commandeered to move a small, hastily assembled relief column of cavalry, infantry and Bedouins. They accomplished what no conquering army in history had done. They fought their way across a 500-mile barren, unmapped desert, enduring temperatures approaching 120 degrees to reach the airfield.
In a gigantic game of bluff, fewer than fifteen hundred soldiers, supported by the RAF in their obsolete aircraft, against odds of twenty to one went on to take Baghdad. They foiled a coup, returned a king to his throne and destroyed Axis aspirations in the Middle East.
Following German successes in Greece and the Western Desert in 1941, a defeat in Iraq would have rolled-up Britain's position in the Middle East, lost her the Suez Canal and her independent oil supplies and threatened India. This would have provided the Axis with the opportunity to create a land bridge between Europe and Asia. A mortally wounded Britain would no longer have had the capacity to continue her sole resistance against Germany, making it impossible to free mainland Europe from the Nazis.
A strong German presence in Iraq and Iran would have brought Turkey into the Axis camp, enabling Germany to finish off Russia at her leisure. The spectre of an Axis-dominated region running from the Atlantic to the Pacific would have provided the isolationists in the United States with enough ammunition to persuade Roosevelt to abandon Britain as a lost cause.
History has overlooked the strategic implications of this isolated battle. The victors and the vanquished were buried where they fell. No monuments exist and no awards were made to commemorate the immense personal courage and the ultimate sacrifices made by many of all races, religions and nationalities.
Hitler's Gulf War is based on a wide range of official, academic and public sources. The book uses reported American, British, German, Italian, Iraqi and Saudi dialogue, and personal reminiscences, to add richness to give what is, to the best of my knowledge, the first full account of the events precipitating the thirty-one-day war in Iraq in 1941. Hitler's Gulf War covers the interplay between individuals and countries which reordered the landscape of Iraq, it describes how the war played out and how it set the scene for Iraq's turbulent future which has come to haunt the West.
Acknowledgements
I am breaking with tradition to start by thanking my wife. Mary is my soul mate, best friend and encourager, whose patience, enthusiasm and support have seen me through five previous books, a Master's and a PhD, and have effortlessly handled several very disruptive country-to-country moves.
Mary makes everything possible in my life.
I also need to thank my late parents. My father Ivor was a man with a deep sense of right and personal honour. His tales of travels in the Middle East, including a hilarious trip across the Iraqi desert in a Nairn Brothers bus from Haifa to Baghdad, on his way to work for the Anglo-Persian Oil Company in Abadan in 1943, gave me the wanderlust and a consuming passion to find out why things happened the way that they did. Both of these have driven my life. My mother Dora nursed me back to life on so many occasions, supported me in everything that I wanted to do and gave me the will to succeed.
The genesis of Hitler's Gulf War was a Starbucks in the Barnes & Noble bookshop in the Menlo Park Mall in New Jersey in early September 2000. In between cappuccinos while skimming through a number of military history books, I became fascinated reading a few snippets about the events in Iraq in 1941, and the idea emerged for a book. However, things only began to get serious in November 2005, following my relocation from Switzerland to the UK.
I had chosen by accident to live in East Anglia, which put me among not only a number of RAF stations but, more importantly, residing in my village and the next, a number of retired RAF officers who willingly gave me their time and experience.
My neighbour, Gp Capt Steve Lloyd, is the Deputy Director of the Air Historic Branch of the RAF. Steve kindly started me off with a trawl through their files. I was also very fortunate to have AVM Warwick Pike brief me on combat medicine, Air Cdre David Bywater related his experiences in flying the Shuttleworth Trust's biplane fighters and Gp Capt Brian Lee gave me a run-down on photo-reconnaissance. By the greatest of coincidences I also found Wg Cdr Alan Robertson in the next village, who told me about his fascinating visit to Habbaniya in 1988.
Any researched work depends on assistance and advice to help to fill in the inevitable blank spaces, tease out long-forgotten gems of information and broaden the work by suggesting new directions.
The regimental museums – the Essex Regiment (Ian and Keith Hook), the Green Howards (Sarah Taylor), the Household Cavalry (John Lange) and the King's Own Royal Regiment (Peter Donnelly) – helped to add both depth and balance to what has been largely written about as solely an RAF victory rather than the first true British use of combined operations.
Catherine Morris (The Times), Julia Burgess (RAF Disclosures) Fregatenkapitan Dr Jorg Hillmar (Militargeschichtliches Forschungsamt), Herr Betten (Deutsche Dienstelle), John Rolph (RAF Armoured Car Association) and Arnaldo Minuti (Italian Foreign Ministry) were kind enough to provide me with obituaries from their files, enabling me to complete a further dimension by featuring the post-1941 lives of the key characters involved in those thirty-one days in May.
Wg Cdr Martin Hooker of the RAF Regiment Association was even kind enough to recommend a publisher.
I was also very fortunate to be able to use the formidable resources of three of the world's greatest archives – the British Library, the Imperial War Museum and the National Archives.
However, I would not have been able to source the bulk of my research without the considerable help of Huntingdon Library staff: Sue Booker, Kathryn Bottomley, Michelle Clark, Joy Glazier, Manju Mathew, Elaine Mountfort, Sharon Reed, Lynn Reynolds, Laura Southgate, Moira Storey, Ann Teal and Lesley Wood with great kindness, enthusiasm and understanding, obtained all manner of difficult-to-source books and journal articles; nothing was too much trouble.
Motorbooks in London, probably the best military bookshop in the world, located some obscure books for my research, accompanied by some mean rockabilly from Wanda Jackson.
Two old friends were of immense help. Michael Fenwick made a crucial introduction at Jak's Black Pudding Luncheon Club, and military historian Guy di Carlo in Florida cast his critical eye over the work and made some important suggestions which transformed the work.
I thank you all.
Prologue
The make-believe country
The very name Iraq conjures up notions of intrigue, passion, obsession, war and wanton violence – in almost equal measures.
These perceptions are rooted in the way in which Iraq was created out of the complex past of Mesopotamia, the influence of seminal events – the disintegration of the Ottoman Empire and two world wars – and the forces that they unleashed, nationalism, ethnic feuds and religious intolerance.
Iraq is pure invention, the result of diplomatic gerrymandering.
Iraq only became a country in the aftermath of the Great War with the fateful Ottoman decision to ally itself with Germany and the equally fateful British decision to stay after the Turkish defeat and carve out a subservient nation in the heart of the volatile Middle East. Iraq, based loosely around the ancient land of Mesopotamia, encompassed Basra, Baghdad and later Mosul – three previously quite separate and disparate Vilayets, or provinces, at the eastern edge of the Ottoman Empire.
A continued and controlling British presence in Iraq was seen as crucial to Imperial communications. The central concern was a secure route to India, and Iraq offered an air and land bridge across the Middle East shortening both time and distance. The proximity of Iraq to Britain's growing investments in oil-rich southern Persia and the discover of oil at Kirkuk in 1927 made the justification, retrospectively, even more compelling.
Control was legitimized by having the League of Nations grant Britain a Mandate over Iraq in 1920.
From the very start things went wrong, as the British completely underestimated their task. The Vilayets had never experienced democracy or produced a society not dominated by oppression, autocracy or mayhem. Among Iraq's two million people there was no national unity, and the predominant influences were family, blood, tribe and religion, and where all else failed, and even if it did not, brutal violence.
The British found Iraq virtually ungovernable, and faced a continual round of uprisings, massacres of their troops and officials, widespread brigandage and incursions across Iraq's porous borders by the Turks from the north and Ibn Saud's militant Wahhabi Ikhwan from the south and west.
A widespread and ferocious series of revolts that broke out in 1920 from Shia and Sunni clergy opposed to rule by non-Muslims, Kurds who refused to submit to an Anglo-Arab authority ruling from Baghdad and tribes resistant to paying taxes brought the British face-to-face with 130,000 mobilized and armed Iraquis. It took until late February 1921 to regain full control of what became known as the Ath Thawra al Iraqiyya al Kubra (The Great Iraqi Revolution), at the cost of the lives of 6,000 Iraqis, 500 Imperial troops and £40 million.
The large army needed to police Iraq, as well as simultaneously fighting an insurrection in Ireland, manning a stand-off with Russia over Persia and containing growing problems in India, caused increasing domestic difficulties in Britain. Compulsory military service abroad and the fact that many soldiers were serving when they should have been demobilized at the end of the Great War caused widespread unrest. Britain was going through an industrial class struggle supported by the emerging power of the trades union movement and the growth of the Labour Party, which gave a sharp political edge to government concerns over stability in Britain.
The continued unrest in Iraq was also an unacceptable financial strain on a Britain which was struggling to come to terms with the crippling economic costs of the Great War. Cutting government spending in order to restore an overstretched Britain's declining economic position and to reduce the vast debt accumulated during the First World War became a political necessity.
However, cutting the cost of the large garrison in Iraq called for radical political and military remedies other than direct military occupation.
The solution adopted was indirect rule, using a surrogate ruling through a strong centralized government beholding to the British. This approach had worked successfully in both the Indian and Malay states, and there was every indication that this would apply in Iraq.
The British arranged for Feisal, the third son of Sharif Hussein of Mecca, to be enthroned as King in 1921. This also helped to mollify his father, to whom the British had promised an independent Arab state as payment for his support against the Ottomans, as well as assuage the ambitious and power-hungry Hashemite dynasty. Although acclaimed in a one-question plebiscite as the people's choice, which gave him a ridiculous 96% of the vote, Feisal was elected by ballot-rigging by the British on a monumental scale. Feisal, from the Hejaz and a Sunni, relied on his descent from the Prophet Mohammed's family as his Islamic credentials to appeal to Muslims. While Feisal looked like a king – slim, bearded and aquiline nosed – and acted the part with regal dignity, he, like others uncertain about where their best interests lay, intrigued with his subjects and dissembled his patrons. He was a slippery customer and few believed a word he said, with the result that everyone distrusted him.
Feisal depended on British political and military support, but to cultivate popularity and enhance his legitimacy he needed to appease anti-British sentiment by obtaining concessions from the British. Convinced of the advantages of indirect rule, the British went along with Feisal's demands for the growing transfer of power to the Iraqi government.
In an attempt to reduce the size of their garrison and the crippling costs of Imperialism, the British hit on the novel idea of using air power to control the country. Air power with its speed of application, geographic reach and economy far outweighed the value of ground troops, and, applied by the Royal Air Force, air power quickly became the major instrument for maintaining order.
As part of the policy of indirect rule and to create stability and stem the outflow of money, the British later formed and financed an Iraqi army to take over from Imperial troops, using the RAF in a supporting role. However, this was one more factor in the uneasy relationship between Britain and the regime she had created in Iraq. The British visualized an Iraqi army as a small voluntary force to maintain internal security, while the Iraqi nationalists visualized a large conscript army to unite the disparate communities in a cohesive national institution.
To their credit the British spent millions to vastly improve the lot of Iraqis. They rebuilt roads, canals and bridges, repaired the railways and established clinics and hospitals, and provided Iraq with a revenue department, a judiciary, medical and education systems and a police force, and put in place an elected government.
The British finally gave Iraq its formal independence in 1932, and Iraq joined the League of Nations.
Nevertheless, Britain still maintained a level of indirect control. The Anglo-Iraqi Treaty signed in 1930, which was to run for twenty-five years, guaranteed Iraqi assistance to Britain in time of war and allowed the passage of British troops through Iraq. A British Military Mission remained with the Iraqi Army, and the RAF retained two bases – Habbaniya, fifty-five miles west of Baghdad, and at Shaiba, near Basra.
On the surface Iraq seemed to be moving towards a prosperous future. Beneath the surface, however, the country was in turmoil. This was not surprising, as Iraq, in many ways, was a make-believe kingdom built on false pretence with a monarchy that lacked legitimacy with the majority of its subjects. Iraq had an imposed king, in a country where monarchy was an alien concept. Feisal, a Sunni from the Hejaz, was an outsider in a country with a Shia majority. He spoke an Arab dialect barely intelligible to most of his subjects, and had no sense of connection with Iraq's ancient past. Iraq was kept going by a government designed and sponsored by the British for British purposes, communications and oil, and its elaborate and expensive administration was never asked for nor wanted by Iraqis.
Britain's policy of indirect control carried with it the seeds of its own destruction. Running a country on the cheap inevitably produces massive risks. By ridding themselves of the immediate responsibility of direct government, the British, in the interests of saving money and tying this neatly together with Hashemite ambitions, conceded their influence in the country. Ever greater concessions to the Iraqis to prop-up Feisal's legitimacy, and giving Iraq its own army, contributed to the emergence of a powerful military that was hostile to both the British and civilian political control.
Internal dissent
There were constant problems among Iraq's heterogeneous population.
The Sunni minority, who owed their allegiance to Mecca, were concentrated in the centre and west of Iraq and made up around a quarter of the population. They were exposed to education and urbanization and were inclined to support nationalism, and resented even surrogate British rule. The Shia majority, around half the population, lived mainly in the south, followed tribal and Sharia law and looked towards Persia. The Shia were suspicious of any government control, particularly because of their experience under the Ottomans of Sunni autocracy. The Kurds, although Sunni Muslims, were not Arabs. They represented about a fifth of the population, lived mainly in the north and had more in common with Kurds living in Turkey and Persia than they did with their fellow Iraqis. Then there were scattered communities of Assyrians, Armenians, Turkomans and Yazidis, together with Jewish and a number of Christian minorities in the towns.
Added to this volatile mix were endemic conflicts between the tribes and the cities, the landowners and the peasants, the merchants and the bazaaris, Iraqi nationalists and pan-Arabists, and the urban Effendi, who had served under the Ottomans, and young Army officers. These disparate groups were all fighting for a place in the emerging state structure.
Feisal was forced continuously to maintain a balance between these competing interests in a country lacking both internal cohesion and unity.
Parliamentary democracy under Feisal became a façade, with cabinets formed and forced to resign by direct intervention of the King. Iraq was a parliamentary state where parliament had no effective power and where political opportunism, greed, corruption, blood feuds and personal rivalries flourished, leading to continuous changes of government. In fact, between 1922 and 1936 there was not even a pretence of constitutional rule: military power, not the electorate, determined who had the right to rule.
There was no universal suffrage in Iraq, and few except the wealthy and the powerful played any part in the political process. The political process itself failed to develop procedures for resolving internal conflicts other than rule by decree and the frequent use of repression.
More problematic was the fact that fifty families, largely wealthy Sunni, governed the destiny of Iraq, as parliament primarily represented their interests. The ruling élite consisted of fourteen people who frequently changed cabinet posts and were trying to get into the government if they happened to be out. A popular joke in Baghdad in the 1930s was that the Ins were trying to keep the Outs out while the Outs were trying to get in and get the Ins out.
Shortly before his death Feisal said, rather tellingly, of the people over whom he ruled, that they were 'unimaginable masses of human beings devoid of any patriotic idea, imbued with religious traditions and absurdities, connected by no common tie, prone to anarchy and perpetually ready to rise against any government whatever'.
Feisal had hoped that once independence was achieved the politicians would close ranks and would devote their efforts to internal reform.
Nuri al Said, a former Ottoman officer and the then Prime Minister, with a pro-British stance, resigned, and the King invited Rashid Ali al Gailani, a shrewd, industrious and ambitious lawyer and a political pragmatist, as well as a passionate nationalist, to form a new government.
The Christian Assyrians rebelled in 1933 as the government would not provide them with assurances of security as had been guaranteed by the British and Iraqis under the Mandate. Rashid Ali wanted to impress the public with a tough policy towards the Assyrians to demonstrate that the government could keep the many minorities in Iraq in line. Feisal was in Europe when he heard of the Iraqi Army's brutal massacre of 300 Assyrian women and children. Facing mounting criticism from the League of Nations, he replaced Rashid Ali with a more moderate government.
Feisal died in 1933 and was succeeded by his son. Ghazi, a 25-yearold playboy, was a wayward, immature and self-willed young man, ignorant of state affairs and unequipped by training for his duty of kingship. His main pursuits in life had been motor-car racing, flying, horse riding and the cinema.
Not surprisingly Ghazi was unable to keep the fragile balance his father had struggled, often successfully, to maintain between competing interests. Politicians took advantage of his inexperience and the flimsy political order to compete for power. Between 1932 and 1934 five more changes of government occurred as politicians embarrassed each other with press attacks, palace intrigues and incidents. Politicians also incited the tribes, habitually opposed to any central authority, to rebel. This caused the fall of a further three governments between 1934 and 1935.
From the early 1930s the Iraqi government became utterly dependent on the Army to maintain authority and unity to such an extent that the Iraqi Army became the de facto arbiter of political power. The Army, disillusioned with chronic and unresolved instability due to the failure of the politicians to manage Iraq, eventually took matters into its own hands and moved into the power vacuum. Iraq suffered its first military intervention, the Bakr Sidqi coup, in 1936. Until parliamentary democracy, of a sort, was restored in mid-1941 all governments in Iraq were either run directly by the military or functioned only at the pleasure of the Army, with cabinets heavily staffed by Army officers.
However, even in the military there were differences of opinion and jealousy. The 'old guard' of Sharifian veterans, predominantly Sunnis, who had fought with Feisal and Hussein against the Turks, were increasingly challenged by the ex-Ottoman officers, who were often from urban and lower-middle-class Sunni and Kurdish families. The result was that Iraq suffered a series of coups and counter-coups as the two military factions divided by networks of patronage built around religion, sect, ethnicity, regional affinity or ideology jockeyed for power, with each faction supporting a different set of civilian leaders.
Exacerbating the praetorianization of Iraqi politics was the rise of pan-Arabism and the Palestine Question.
Arab nationalism
The defeat of Turkey in 1918 and the abolition of the Ottoman Empire in 1922 created an ideological tidal wave among its subjugated people of nationalism – the desire of people to run their own affairs.
The failure of the British in Palestine and the French in Syria to fulfil their mandates and to create independent Arab states stoked the fire of nationalism among both Arab intellectuals and Army officers. This intensified as the prestige of Britain and France declined in the late 1930s, resulting from their repeated appeasement in the face of German and Italian aggression. The impression of weakness that this created fanned the flames of pan-Arabism throughout the Middle East.
In Iraq leading nationalists saw an emerging opportunity to create a much larger state centred on Iraq, encompassing Palestine, Syria and Transjordan. From the beginning of the 1930s Iraq became steadily more nationalistic and more anti-British.
German nationalism, with its emphasis on language and history as unifying factors, and the efficiency of Hitler's totalitarian state, which seemed to offer a more effective way of governing in a fragmented society with endemic political instability, was seen as the perfect role model by many Iraqi nationalists and Army officers.
The primary purpose of the Iraqi education policy from 1930, under the influence of Sati al Husri, the Director-General of Education, and his successor Sami Shawkat, was to overcome the difficulties of forging a nation-state out of the diverse and conflict-ridden social structures, as well as transform parochial loyalties – tribal, ethnic and religious – into a national identity. To inculcate a national ideology and strengthen patriotism among the young, government schools quickly degenerated into political seminaries to impart Husriyya – political extremism and xenophobia. Military training became a compulsory subject at the Teachers' Training College, technical schools and all government schools, and a paramilitary youth movement, the Futuwwah, was formed. Following the German and Japanese militaristic models of the 1930s, the Futuwwah had enforced codes of conduct, its members drilled military style and wore uniforms, complete with ranks and insignia.
A vision of Iraq as an autocratic, Prussian-style, highly centralized state which could eventually engulf its neighbours began to capture the imagination of the nationalist intellectuals and the Army officers like Taha al Hashimi. This clashed with ex-Sharifians like Nuri Said, who saw Iraq's future in a Hashemite federation with Transjordan.
Clubs like the al-Muthanna and leading newspapers, such as Al Alam Al Arabi, were used as conduits for promoting a national ideology and shaping anti-British feeling to further develop nationalism.
Under Feisal Iraq was not over-concerned with Zionism. Feisal believed that an accommodation could be reached between the Arabs and the Jews. The situation changed radically under Ghazi. Ghazi listened to neither Iraqis nor the British, and proved to be an ineffective ruler easily manipulated by politicians. Despite his failings be began to be seen as a national hero following his demands for the annexation of Kuwait and his support for the Palestinians Arabs during their 1936–1939 revolt.
On his death in 1939, in a road accident, Ghazi became revered as an Iraqi martyr. Opportunistically the nationalist movement was able to convince a substantial part of the population that Ghazi had been killed by the British, who opposed his policies.
Ghazi's son, King Feisal II, ascended to the throne, but as he was only four years of age his pro-British uncle, Emir Abdul Ilah, was appointed Regent.
Religious zealotry
The catalyst moving forward both Iraqi nationalism and pan-Arabism was the arrival in Baghdad in October 1939 of the Mufti of Jerusalem, Haj Amin al Husaini.
Haj Amin, a former artillery officer in the Turkish Army, was a powerful combination of a Palestinian Arab from a leading family, a fervent nationalist and a Muslim religious leader. At the end of the First World War Haj Amin agitated for the creation of an Arabic state in the Middle East comprising Palestine, Syria, Lebanon, Jordan, Iraq and even Cyprus and parts of Turkey. When this failed he created an anti-Jewish party, El Nadi al-Arabi, and fomented the murderous riots in Jerusalem in 1920. He fled to Damascus but was sentenced to 10 years imprisonment in absentia by the outraged British Colonial Administration. The British, in a mistaken gesture of lenience, pardoned Haj Amin. Fooled by his soft-spoken, mild manner and dignified exterior, which concealed burning Arab nationalism and a violent non-compromising nature, the British committed an even bigger error in 1921. The British agreed to Haj Amin's appointment to the post of Mufti, which his family had held almost unbroken since the seventeenth century, and a role that usually went to jurists who arbitrated disputes by interpreting Koranic law.
As Mufti he began to work behind the scenes, playing a major role in the anti-Jewish violence in Palestine in the late 1920s, and he was a major force behind the Palestine Arab revolt of 1936–1939. Assassinating moderate Arabs in the cities, who supported the British and urged compromise, and attacks orchestrated by the Mufti on Jewish settlements and British installations became the norm.
Haj Amin fled to Beirut in October 1937, one step ahead of the British, who were seeking him for complicity in the murder of Lewis Andrews, the British Commissioner in the Northern District of Galilee. The Mufti eventually fell foul of the French, due to his involvement in nationalist politics in Syria. He again fled, this time to Baghdad, in 1939.
The Iraqi parliament offered sanctuary, welcoming Haj Amin with open arms as a Palestinian hero, and they and the Iraqi government and the public helped finance the Mufti and his cause. Haj Amin was not short of funds, as money was also supplied by other Arab countries, like Egypt and Saudi Arabia, as well as Germany and Italy. All saw the potential for the Mufti to ferment trouble for the British.
The Mufti established his headquarters in Baghdad with his Chief of Staff, Jamal al Hussaini, his Islamic religious adviser Sheik Musa al Alami, as well as his military commander, the ruthless Syrian-born guerrilla leader Fawzi al Qawujki, who had been trained at St Cyr, the French military academy.
The Mufti began to spread a malevolent influence throughout Iraq. He started his own newspaper, Istiqal, which was subsidized by the Germans and Italians. More pointedly, Haj Amin proposed Palestinian and Syrian refugees to fill essential government positions which the state of illiteracy in Iraq prevented its own subjects from assuming. Rapidly key posts in the administration and in the professions – engineers, doctors, lawyers, teachers and, importantly, the police – were filled by Palestinians and Syrians who owed allegiance to the Mufti.
Haj Amin quickly established a broad range of relationships with the leading politicians, the important merchants and the Army leaders, and became the guest of honour at many government functions. Against a background of general unrest created by the war, the Mufti became the focal point of nationalist aspirations among the Army leaders and younger politicians.
In this safe haven, leadership of Arab nationalism throughout the Middle East quickly began to coalesce around Haj Amin. On a pan-Arab level he became the spokesman of a secret committee of leaders from Syria, Iraq, Palestine and Transjordan, as well as the leader of the Palestine nationalists.
However, the Mufti came into his own in Iraq as the leader of the Committee of Seven – a cabal of influential Army officers and politicians – who increasingly made the political decisions in Iraq.
German Aspiration
Germany had had designs on the Middle East, and particularly on Iraq, since the 1890s, driven by the Kaiser's jealousy of Britain's commercial empire in the East. His obsession, the Berlin–Baghdad railway, reached the Turkey-Iraq border in 1913 and was only halted by the outbreak of the Great War
The Arabists at the Auswartiges Amt, the German Foreign Ministry, in Berlin kept the idea of German influence in the Middle East alive through the Weimar period, and made an astute choice in selecting Dr Fritz Grobba as the Chargé d'affaires of the German Mission in Baghdad in 1932.
Grobba was the central figure in Germany's Middle East policy in the 1930s and early 1940s. Although being a member neither of the Nazi Party nor of the aristocracy that traditionally ran foreign policy, he was, nevertheless, a highly able and influential German agent.
Grobba, who spoke fluent Arabic, Persian and Turkish, had served with the German Military Mission to the Turkish Army in Palestine during the Great War, and knew the people and the mentality of the Middle East. While he found an increasing radicalization of Iraqi politics and a country simmering with resentment towards the British, he also found people who were impressed by Germany's strong leadership and militarism and inspired by the resurgence of German power and by its intimidation of Europe.
Grobba was very ambitious, some say unscrupulously so, and many thought that he had 'Lawrentian dreams'. Grobba saw that bringing Iraq into the German camp could provide him with a springboard into a major career at the Auswartiges Amt.
Grobba was egalitarian and a very active and highly personable diplomat, and he and his charming wife worked diligently to create a wide range of relationships embracing leading political, religious, military and economic leaders in Iraq. He was very successful in promoting German trade, but his role changed after 1935 and he became much more political.
Although the Iraqi Army, and later the Mufti, had continuously requested German arms, Grobba, while sympathetic to the nationalist cause and to the trouble that the Mufti wanted to create for the British in Palestine, was caught in a policy trap.
Between 1933 and 1939 Hitler's England Politik was designed to strike an alliance with England, which meant that Grobba was unable to be seen to support radical anti-British nationalists. Complicating the issue was the 1936 Rome-Berlin Axis agreement, where the Eastern Mediterranean was deemed to be within the scope of the Italians, who were already trying to undermine the British funding newspapers like Saut al Shab.
Grobba believed that these factors, together with Hitler's antiSemitic and racial doctrine, the Weltanschauung, which excluded Arabs, and the Auswartiges Amt, with a greater interest in Europe, all contributed to underestimating the value of Arab nationalism. This compromised Germany's Middle East policy, which Grobba had helped to both design and implement, and undermined his mission in Iraq.
Attitudes towards the Middle East policy in Germany were diverse. The Aussenpolitisches Amt, the Office for Foreign Policy of the Nazi Party, and the German military intelligence organization, the Abwehr, were interested in expanding Germany's influence. In the Auswartiges Amt only a few people valued the Arab nationalist movements positively, and the prevailing view was that they were not to be taken seriously. The Wehrmacht, the German armed forces, felt that Iraq was too far away and too difficult to reach to support an armed insurrection.
Grobba also felt that Italian involvement was an embarrassment, as many Arabs viewed Italy as a colonial power with imperial ambitions in the Middle East.
As a consequence a frustrated Grobba could provide only financial support and personal encouragement rather than public displays of support, overt propaganda and arms shipments.
Nevertheless, Grobba decided to continue 'his' policy in a covert fashion. He intrigued with Iraqi Army officers, exploiting their Anglophobic and Germanophilic sentiments through dinners, parties and film shows, and financed pro-fascist groups and cells. From the mid-1930s his house was a central meeting point for Iraqi nationalists. He subsidized newspapers to run pro-German and anti-British propaganda as well as the serialization of Hitler's Mein Kampf in Arabic in Al Alam, Al Arabi in 1933, and negotiated successfully for German to replace French as the second language in Iraqi schools. Grobba also began to subsidize the creation of clubs promoting Iraqi-German friendship in Baghdad, as well as reciprocal visits by politicians of the two countries. In 1937 he arranged for the Hitler Youth leader Baldur von Schirach to visit Baghdad with a delegation, and financed representatives of the paramilitary Futuwwah to attend the Nuremberg Rally in 1938.
After Munich, Germany paid less attention to appearances, and in 1938 followed Italy's lead by broadcasting anti-British propaganda in Arabic from Radio Zessen, near Berlin, using an Iraqi announcer, Yunis al Bahri.
With high levels of illiteracy throughout the Middle East, the radio was an ideal instrument for spreading propaganda. As early as 1934 the Italians had launched an Arabic-language service from Radio Bari, glorifying Italy and its achievements and supporting the Arab nationalist cause against the British and French. The Italians provided radio sets at nominal prices throughout the Middle East, which was extremely popular with the owners of Arab cafés, the centre of social life, who installed the sets for their patrons.
British indifference
Strange as it may seem, the British in Iraq did little to counter the Mufti, nationalism, anti-British propaganda and growing German and Italian intrigue. In particular there was a lack of perception of the symbiotic relationship between nationalism, the Army and public education created by Husriyya. The Military Mission was marginalized by nationalist officers who shared little with their British counterparts, and frequent staff changes and the lack of interest that many, if not most, British diplomats had in Iraq contributed to the problem.
The lack of action had become a hallmark of a British diplomatic service, with a culture steeped in decades of languid appeasement.
The ambassador, Sir Basil Newton, did not speak Arabic and had little direct contact with community leaders except those who were pro-British. He was disinclined to listen to anything which challenged the conventional wisdom that the country was unstable but manageable with 'our' Iraqis, and he dismissed everything else as rumour and speculation. He even obstructed countermeasures by refusing to have intelligence operatives join his staff to conduct covert propaganda and political subversion.
Surprisingly, things did not change when the British declared war on Germany in 1939.
Declaring war on Germany, as the Anglo-Iraqi Treaty required, created controversy among Iraq's leaders. The Army officers opposed, while the pro-British politicians like Nuri supported, a declaration of war. Others, like Rashid Ali, saw opportunities for gaining concessions on Palestine and Syria as the price for Iraq's entry into the war. Major opposition in both September 1939 and again in February 1940 to a declaration of war resulted in merely a break in diplomatic relations between Iraq and Germany.
While Grobba was declared person non grata and had to leave Iraq, and German citizens were deported to India for internment, the nationalists and the Mufti continued to communicate with Germany through the radio of Lugi Gabrelli, the Chargé d'affaires in the Italian Legation.
As the war intensified and the Germans scored victory after victory, Nuri's position worsened and the pro-British group became isolated. Rashid Ali again took office in March 1940.
The second testing time came in June 1940, when Italy declared war on Britain and France. The Iraqi government not only refused to declare war on Italy but also would not break off diplomatic relations, despite intense pressure from the British Ambassador.
Perceptions of weakness
The fall of France in May 1940 changed the Iraqi cabinet's attitude towards Britain. The view prevailed, even among the pro-British faction, that a British defeat was inevitable and that the best course of action was to adopt a strictly neutral stance and limit the fulfilment of treaty obligations towards Britain to a minimum.
When the Ambassador requested Iraqi agreement to land British troops in Basra to proceed to Haifa across Iraq in June, the government responded by limiting the number of troops and imposing a fixed time for their presence on Iraqi soil. Although there were intense negotiations between the British and the Iraqis, particularly over changes in policy in Palestine, the British rejected the Baghdad proposals.
This inevitably led to the Iraqis' increasingly turning towards Germany. Several visits were made by Iraqi government ministers and by the Mufti's private secretary, Osman Kemal Haddad, in June and July 1940 to Fritz von Papen, the German Ambassador in Ankara, to discuss opportunities for collaboration with Germany.
In September 1940 Haddad travelled to Berlin to see Grobba, Otto-Werner von Hentig, the Head of Pol VII, the section of the political department of the Auswartiges Amt responsible for the Middle East, Dr Ernst Woermann, the Under-Secretary of State and Ernst von Weizsacher, the Secretary of State.
Haddad's visit was designed to solicit German support in the form of an immediate formal statement guaranteeing Iraq's independence.
The Germans were cautious, and while agreeing in principle, were unwilling to make a public commitment, and Haddad returned to Baghdad with vague promises rather than firm undertakings. The Germans also requested that the Iraqis should not act militarily against the British without German agreement, or encourage the British to occupy Iraq.
However, the Mufti and Rashid Ali continued to pressure the Germans, and finally, in October 1940, the Germans, together with the Italians, issued a joint communiqué which, while expressing sympathy with Arab aspirations, was more aimed at anti-British propaganda than a full statement of support.
Anglo-Iraqi relations deteriorated further, and despite Nuri's attempt at conciliation Rashid Ali refused to change his stance. Although the majority in the government rejected Nuri's viewpoint that Britain would prevail, and felt that Britain was now isolated and on the brink of ruin, many in parliament were pro-British and favoured cooperation with Britain. When the Italian forces under Maresciallo d'Italia Graziani took a beating from the British in the Western Desert in March 1941, neutrality had an aura of safety about it. The majority feeling in the country was that Iraqi neutrality was essential: if Britain were to win, Iraq would be safe anyway, and if the Germans won, Iraq's best hope was to do nothing to offend them.
Haddad undertook a second visit to Berlin in December 1940 with the specific objective of securing economic and military aid, in particular captured British weapons. As increasing political tension in Baghdad was weakening Rashid Ali's position, he became desperate to clarify the attitude of the Axis towards Iraq and to obtain arms.
This time Rashid Ali's requests fell on sympathetic ears, as Germany needed to put pressure on Britain to counteract her successes against the Italians in Cyrenaica. Creating a new battleground in the rear would distract the British.
Despite the fall of Rashid Ali's government in January 1941 in response to his ongoing dispute with the British and the Regent, the Germans were well advanced in their plan to ship weapons to Iraq. The Germans correctly believed that the new regime of Taha al Hashimi was an extension of previous nationalist governments. al Hashimi was a known admirer of Germany and of the Golden Square, a cadre of four extreme nationalist Iraqi colonels – Salah el Din al Sabbagh, Fahmi Said, Mahmud Salman and Kamil Shabib – whom he regarded as his protégés.
With the promise of German arms on the way and a Britain becoming increasingly weaker, the Mufti, the Golden Square and Rashid Ali and his cabinet of ultra-nationalists began to plan their coup.
Chapter One
Slippery Slope to War
Baghdad, 28 February 1941
Outwardly the white villa with its green window shades in a quiet palm-lined street looked nothing more than the home of a wealthy Iraqi merchant or leading civil servant.
In fact, the large, two-storey structure, with big airy rooms, complete with a garden shaded by nebk trees and surrounded by a 2 m high wall, had been built in the early 1900s as the city residence of a leading pasha in the Ottoman administration.
Since October 1939 the villa in Zahawi Street provided by the Iraqi government was the official residence of Haj Amin, the Mufti of Jerusalem.
The only thing out of the ordinary on that late February day was the succession of unmarked cars that drew up to the entrance to the villa around 10 o'clock that morning.
The first car, a highly polished black Studebaker, arrived at ten minutes to ten and deposited three men outside the villa. All three had a military bearing and looked uncomfortable wearing civilian clothes. A few minutes later, from the direction of central Baghdad, a grey Mercedes drew up to the main gate, and a small man in a white suit and a red fez, carrying a black briefcase, hurried past the guards into the villa. At two minutes past ten o'clock a dark blue, dust-covered Wolseley with a dented wing stopped outside the villa. The two men inside had driven themselves, and a guard slipped behind the wheel and drove off as the two men walked down into the sunken garden, designed like most Baghdad suburban gardens for watering, and mounted the steps to the villa.
Once assembled in one of the larger downstairs rooms, the six guests were joined by their host, the Mufti.
Haj Amin, a tall man, surveyed the gathering. The Mufti's delicate features and gentle manner, accentuated by his unusually bright, deep blue eyes, trimmed and hennaed goatee, which gave him his nickname, 'The Red Fox'. His calm voice belied his zealousness and his malevolent and violent nature. He was ruthless, humourless and of the authentic stuff of which dictators are made. He emitted, as Freya Stark noted, 'a radiance of a just-fallen Lucifer'.
The two young boys in white skull caps and green and white striped robes who had been serving bitter Arabic coffee from long-necked copper pots to the guests were dismissed. Haj Amin then closed the double doors to the room and indicated to the group that they should take their seats at the large table.
The Mufti in his black gold-trimmed robe, wearing a white turban around a scarlet tarbush, took his place at the head of the table on a throne-like chair of the kind peculiar to the Orient – plenty of gold and tinsel, with its back upholstered in a vivid green.
'To bind us to our purpose we must swear our allegiance to our duty on the Koran', said Haj Amin.
Following Haj Amin's lead, Rashid Ali, Naji Shawkat the Minister of Defence, al Sabbagh, Yunis al Sabawi the Minister of Economics, Fahmi Said and Mahmud Salman all swore allegiance
The Mufti, to the irritation of Rashid Ali, was chosen as the leader of the group as much for his closeness to the Golden Square as for his link between Rashid Ali and the nationalist politicians.
'My friends,' the Mufti began, 'Britain is finished. She has been defeated at Dunkirk and is now isolated from mainland Europe. A few days ago the Germans began to attack the British in Libya. Britain is already stretched fighting the Italians in East Africa. Malta is under siege. Palestine is an armed camp and the Vichy French are already working with the Italians in Syria. German and Italian radio propaganda is working in our favour. Arabs all over the Middle East now see the British as the vanquished. The British are surrounded by enemies and they are vulnerable. Now is the time for us to move decisively against them.
'We need to take the initiative in defining our own future. If we allow the Germans or the Italians to clear away the British we will be merely substituting one master for another as we have done with the British for the Turk. First we must show the British that we will not tolerate one more concession which undermines our independence. This means that we must refuse all and any attempts to land more British troops in Basra or have any of their troops transit Iraq. We must also continue to refuse to break off diplomatic relations with Italy.'
'We must be careful', interjected Rashid Ali. 'If we stick rigidly to the terms of the Anglo-Iraqi Agreement of 1930 we will give the British no reason to act against us. This will also give us the breathing space to arm ourselves with more modern weapons from the Germans and prepare for German military help.'
'No!' said al Sabbagh, 'The time has come to show our resolve. Arguing the legalities has brought us nothing!'
'It has prevented the British from moving against us until we are ready', Rashid Ali responded angrily.
'Friends, we must speak with one voice', soothed the Mufti. 'We must not fall into the common trap in Iraqi politics of arguing among ourselves. This only destroys our ability to fulfil our destiny.'
'We must also remember that we need to move decisively against pro-British elements in politics, the armed forces and in the administration', pointed out Salman.
'That means that we need a new government,' added Shawkat, 'and we must neutralize the Regent. Why not appoint Sharif Sharaf, who is a distant relative of Abdul Ilah, so that there is a bloodline? He is a bit gaga, but he is someone who will do what he is told. We also have to make our government appear to be a unifying agent for the country, and at the same time have a semblance of legality about it. Sharif Sharaf offers us a way out of a constitutional crisis.'
'Why don't we call it a Government of National Defence? This would imply that our actions are being taken in the public interest', added Said
'But,' said al Sabbagh, 'it must not appear that the Army is controlling the government, so we will need a civilian prime minister to give the government a semblance of legality.'
'Precisely! I nominate Rashid Ali,' Haj Amin said, 'to head our new revolutionary government.'
While Rashid Ali smiled, acknowledging the agreement of the others, he was under no illusion that the Mufti would be the power behind the throne of any new Iraqi government.
'We should establish a date for our assumption of control', suggested Said.
'We need enough time to get our forces, their equipment and supplies into position without arousing the suspicions of the British', rejoined al Sabbagh.
'But above all, we must not lose the initiative, and we need to move soon', added Haj Amin.
al Sabawi, who had remained quiet throughout the discussion, said, 'The perfect time would be March 31st, the day of the closing session of parliament. All the politicians will be in Baghdad, enabling us to pick up the pro-British elements and capture the Regent with ease.' He added, 'March 31st will give us enough time to complete our military preparations and for the German and Italian aid to reach us.'
'Good, then it is agreed', said the Mufti. 'Now we need to decide upon our individual responsibilities and set up a timetable for action.'
Throughout March 1941, planning for the coup d'état progressed. However, the plotters faced a challenge. The Regent, together with the British Ambassador, pressured al Hashimi to move members of the Golden Square out of Baghdad to provincial garrisons to reduce their influence. While the Regent did order the transfer of the least influential of the Golden Square, Kamel Shabib, to a remote post, the Colonels managed to get the order reversed. Realizing that they could no longer count on al Hashimi's support, and that the Regent was their enemy, they planned to neutralize both problems.
The British, who knew of al Hashimi's friendship with the pro-British Nuri, invited al Hashimi's Foreign Minister, Tawfik al Suwaidi, to meet Anthony Eden, the British Foreign Minister, in Cairo to discuss Anglo-Iraqi relations.
Cairo, 6 March 1941
Anthony Eden, a handsome, patrician figure, motioned al Suwaidi to sit down at the lunch table set for two on the columned verandah of the large, two-storey, colonial-style villa which housed the British Embassy in Cairo.
The view from the veranda of the magnificent lawns and gardens that spilled down almost to the Nile did what it was supposed to do – impress visitors of Britain's Imperial power.
When the lunch was finished, the coffee served and the small talk ended, Eden rose and asked al Suwaidi to follow him down the steps into the garden so that they could walk down to the low wall which bordered on the corniche and be alone for a private discussion.
'Minister,' Eden began, 'Britain's problem is that we have endured a lack of cooperation with Iraq since the war began which is in conflict with our 1930 Treaty. I realize that your government has inherited this situation, and my reason for inviting you to Cairo is to suggest how we may resolve this together.
I have a number of recommendations which I feel your government would be wise to consider. First you must sever diplomatic relations with Italy. Neither you nor we can tolerate an enemy presence which poisons our relationship. Secondly, British troops must occupy a number of areas in Iraq.'
'But, Sir Anthony, that would cause considerable unrest', interjected al Suwaidi for the first time.
'You need to sell this both as a build-up to liberating your brother Arabs, as we will need to move against the Vichy French in Syria, as well as protect the oil wells and pipelines from sabotage to ensure that Iraq receives an uninterrupted flow of revenue from the sale of its oil.
Thirdly, there is a strong anti-British clique centred on the Army and some leading politicians. It is probably wise that you retire these senior officers and replace the politicians with a pro-British cabinet.'
The Foreign Minister considered Eden's remarks carefully before responding. 'Everything that you have mentioned, Sir Anthony, depends on overcoming the considerable opposition of the Army commanders. This is not something that we can achieve quickly. We need time.'
'Minister,' Eden said as he turned and led al Suwaidi back towards the villa, 'we both know that neither Iraq nor Britain has time. We must do this together, now, before events spiral out of control and we have no way to resolve this problem as friends.'
They mounted the steps to the veranda and strolled in silence through the villa to the lobby, with its antique Chinese chests and chairs.
al Suwaidi collected his Homburg and briefcase from the servant waiting in the lobby, and the two men walked together down the short stairway flanked by two stone lions towards the Rolls Royce waiting in the driveway.
al Suwaidi paused, and Eden stopped and turned. 'Perhaps, Sir Anthony, it would be easier for the Iraqi public to accept these concessions if Britain implements the 1939 White Paper on Palestine and let the Palestinian leaders return.'
'Minister,' responded Eden sharply, 'I cannot implement such sweeping changes in the middle of a war. As we discussed, we need to focus on resolving our mutual problem in Iraq before we even begin to think about problems which do not affect your country.'
Eden quickly changed his tone and smiled at the Iraqi minister, who was now looking glum. They shook hands. al Suwaidi entered the car, and as the door closed Eden lent forward and spoke to him through the open window: 'Minister, I am glad that we had this opportunity to discuss our mutual problem. I know that you will present our recommendations to Prime Minister al Hashimi. I look forward to receiving your agreement to resolve the problem and the steps you will take so that we may put this matter behind us and work together for a better future.
As the car pulled away from the villa, al Suwaidi sat back and reflected that Eden had given al Hashimi little more than an ultimatum. Either you fix the problem in Iraq or we shall do it for you. He knew that whatever al Hashimi felt, implementing these recommendations would be virtually impossible, given the certain opposition of the Golden Square and the Mufti.
Cairo in early 1941 was a nest of spies. Within twenty-four hours Eden's 'recommendations' were in front of the Mufti and the Colonels in Baghdad, and on the desks of Joachim von Ribbentrop at the Auswartiges Amt on Wilhelmstrasse, in Berlin, Count Ciano the Italian Foreign Minister at the Palazzo Chigi in Rome and Cordell Hull the Secretary of State at the State Department in Washington.
In Baghdad the conspirators realized that if al Hashimi capitulated to the British they would be sidelined and would lose a golden opportunity to rid themselves of the British. If anything, Eden's 'recommendations' spurred them past the point of no return.
Baghdad, 1 April 1941
On the night of 1 April 1941 the plotters seized power. Truck-loads of troops were ferried into Baghdad to secure the radio station, the telegraph office, the three main railway stations, the bridges across the Tigris, the airport, the palace and the main roads into the city.
The doctor who had signed, under duress, a certificate of the Regent's death from coronary thrombosis, managed to smuggle a copy to the Regent through the palace cook.
The Regent spirited himself away to the house of an aunt, Princess Salha, who lived in an old building on the east bank of the river. They, together with the Queen Mother and Princess Badiya, asked Dr Harry 'Sinbad' Sinderson, the palace physician and royal confidant, to come to the Princess's house to discuss what they should do. Sinderson, after a roundabout journey to avoid the growing military presence on the streets, managed to persuade the Regent that he should dress as a woman and make his way by horse cab to the American Legation rather than to the British Embassy, which was on the other side of the Tigris and bound to be guarded by the plotters.
Dressed in a woman's abba, veil and shoes, the Regent arrived at the American Legation in the early hours of the next morning. As an unknown 'woman' he had difficulty in gaining admittance to see the Minister, Paul Knabenshue. Eventually he managed to see the Minister's wife, Olive, who, realizing it was the Regent, quickly admitted him
Transferring the Regent safely to the protection of the British Embassy across the river would be impossible, and it was decided that he would leave Baghdad with the Knabenshues. They were due to welcome the new British Ambassador, Sir Kinahan Cornwallis, who was flying into RAF Habbaniya, some fifty-five miles due west of the city, later in the day. If they could get the Regent to Habbaniya he could be flown to safety.
Concealed in the Ambassador's white Buick under a rug which covered the legs of the Minister and his wife, the Regent left the American Legation on his way to Habbaniya.
As the car approached the King Feisal bridge across the Tigris, it was flagged down at an Iraqi Army roadblock.
The soldiers raised their rifles, pointing them at the car, and a captain from the 3rd Division strode forward and peered into the vehicle. Knabenshue rolled down the window. 'This is an official car from the American Legation and I am the accredited Minister. Please let this car pass', he said authoritatively.
He rolled up the window quickly and sat back as if to signify that this was all he needed to say to be allowed through the roadblock.
The captain stood back, clearly puzzled, as he had been ordered to detain all Europeans.
Olive Knabenshue looked on in alarm as the Minister, out of sight underneath the window, drew back the slide cocking his Colt 45 automatic, and she saw the end of the Regent's pistol sticking out from under the rug.
The captain conferred with another officer and returned to the car. The flag of the United States on the bonnet of the car seemed to have done the trick, and the captain stood back, smiled, gestured for his troops to raise the barrier and waved the Buick through the roadblock.
Just over an hour and forty minutes later the Buick entered the gates of RAF Habbaniya.
Across Baghdad the hunt was under way by the conspirators for the Regent, Nuri Pasha and other pro-British politicians.
Also forewarned, Nuri Pasha had called Sqn Ldr Patrick Domvile, the British Air Liaison Officer with the Iraqi Air Force. Domvile met Nuri at a mutual friend's house and had him dress in one of his spare RAF uniforms before driving out to Habbaniya. They had a close call at the Falluja bridge, fifteen miles from Habbaniya, when they were stopped by an Iraqi Army roadblock.
With hats pulled down over their eyes and collars turned up to hide Nuri's well-known face, they eventually managed to convince the lieutenant in charge that they were just a couple of RAF officers returning from a weekend leave in Baghdad. After a few minutes' deliberation, the barrier across the road was raised and they were allowed to proceed to Habbaniya.
Later that evening the Regent and Nuri Pasha, as well as three other politicians, a senior member of the Royal bodyguard, Ubaid bin Abdulla, and another member of the Royal family, all of whom were able to make their way to Habbaniya, were flown out to Amman in Transjordan, and then on to Jerusalem.
Habbaniya and Baghdad, 2 April 1941
After inspecting the RAF honour guard at Habbaniya, Sir Kinahan 'Ken' Cornwallis, the incoming British Ambassador, met Paul Knabenshue in one of the Air Headquarters block's small conference rooms.
Knabenshue, who had been the American Minister in Baghdad since 1932, quickly brought Cornwallis up to date with the current situation. Rashid Ali's intrigues, the malevolent influence of the Mufti on Iraqi affairs, the coup engineered by the Golden Square and the flight of the Regent and the pro-British faction led by Nuri Said had combined to create a very unstable situation.
Knabenshue and Cornwallis quickly put together a joint programme of evacuation of both countries' citizens, should the need arise, calling the plan Operation Concentrate, and agreed to meet at the British Embassy the following afternoon to finalize their positions.
'What are you going to do about presenting your credentials to Rashid Ali?', Knabenshue asked as they emerged out of the cool of the admistration block into the blazing sun and walked towards Cornwallis's cars.
'I think that my best move is to withhold the presentation as long as possible. Rashid Ali expects me to present my credentials, which is an endorsement by the British government legitimizing his illegal coup. My strategy is to make Rashid Ali work for our endorsement', Cornwallis replied.
Cornwallis, a giant of a man well over six feet tall, with a distinguishingly large nose, needed time to think. He had his wife travel in his deputy, Adrian Holman's, car, and Cornwallis now sat silently in the back of his Daimler as it took him from Habbaniya towards the Embassy.
An Arabist, Cornwallis had first arrived in Baghdad in 1921 and had worked diligently to help King Feisal create the new state of Iraq. Those had been heady days, he remembered with fondness, building a modern state out a group of warring tribes and religious factions. Although they had had their setbacks they had steadily developed a form of democracy and the infrastructure to support a state.
However, his closeness to King Feisal had been resented by the ambitious Rashid Ali. Following Feisal's death, Rashid Ali, then Minister of the Interior, had conspired to have Cornwallis recalled to the Foreign Office in London. This was Cornwallis's first visit to Iraq since 1935 and he steadfastly resolved to put his animosity towards Rashid Ali behind him in the interests of stabilizing the current state of unrest. He hoped that Rashid Ali would also put aside his resentment.
As the car neared Baghdad Cornwallis could see the tops of palm trees poking over the walls surrounding houses and the sand-coloured minarets hidden in dark green palm groves. Over to his left, some five miles or so in the distance, he could see the sun reflecting off the twin golden domes of the Khadimain mosque with its four minarets looking like soldiers protecting one of the most holy Shia shrines. To his right was a large date orchard and, just beyond, he could see a bund with a line of gum trees which marked a bend in the Tigris.
As the car moved past Baghdad West station nothing much appeared to have changed in the fabled city of 'One Thousand and One Nights'. However romantic the images that Baghdad conjured up, it had not been a true capital city since the seventh century, and in many respects it was just another hot, dusty city on the way to somewhere else.
Little had happened since the time of the caliphs, and most of the old houses with their white or ochre-painted mud, brick or plaster walls had been rebuilt, refaced and reroofed. Cornwallis remembered that when he had left Baghdad in 1935 the city boasted only a few modern houses. A group of wealthy merchants had been roundly criticized when they built several villas in colonial style, complete with balconies, along Haifa Street
As the car entered the city proper he could see the houses in the Salhya district. With grated windows and intricately carved wooden shutters on the first floor, wooden balconies that jutted out into the street and nail-studded wooden doors, the houses gave out an aura of timelessness. Interspersed between the houses were twisted and dusty alleyways so narrow that donkeys with their panniers could fill them from side to side. Even quite wide streets radiating off the main thoroughfares like King Feisal Avenue, along which he was travelling, were still unpaved and became virtually impassable in a rain storm and contained any amount of dust in the summer.
None the less, Baghdad had a certain charm, and Cornwallis looked forward to meeting the many people who had befriended him and with whom he formed close relationships in the fourteen years he had spent among the Iraqis.
April 1941 was a tense time for both Iraqis and the British. Political and military reverses for the British mounted. Bengazi fell to Rommel's Afrika Korps on 4 April, and by the 10th German tanks were on the Egyptian border at Sollum. Within twelve days Rommel had regained the ground that had taken Wavell fifty days to win against the Italians.
Yugoslavia capitulated on the 17th and the Greeks on 24 April, and the remnants of Wavell's battered force, sent to support the Greeks at Churchill's insistence, were digging in on Crete awaiting a German assault.
In Iraq the plotters consolidated their position. They forced al Hashimi's government to formally resign, and they arrested pro-British politicians, including the Governor of Basra. The King, his mother and other members of the Royal family were exiled to Kurdistan.
The new government was quick to manage the flow of news, and jammed the Regent's broadcasts from Jerusalem. The tone of the press changed and became vitriolic. The Arab-language broadcasts of Radio Zessen and the Italian station, Radio Bari, contributed to the propaganda by claiming that the British had poisoned King Feisal and murdered King Ghazi.
But the Germans and the Italians were still a long way from Iraq, and Rashid Ali needed an excuse to keep the British talking until Axis arms arrived and the Germans could intervene directly. Cornwallis correctly assumed that Rashid Ali was playing for time, and withheld presenting his credentials to the new government to indicate Britain's refusal to accept the new regime. With the Regent in Jerusalem preparing to form a new Iraqi government to challenge the legitimacy of his regime, Rashid Ali was forced to solicit recognition from the British and the Americans.
Rashid Ali had publicly declared on 3 April that his government would respect all its international obligations. This made it difficult for the regime to prohibit the landing and transit of British troops, and Rashid Ali sought a compromise with Cornwallis.
However, in Iraq the pressure on the British began to increase. Following the banning of British military personnel from travelling between Baghdad and Habbaniya, and the seizure of the radios from the British Military Mission on 6 April, AVM Harry Smart, the AOC RAF Habbaniya, lobbied Cairo for both air and land reinforcements. AVM Arthur Longmore, Wavell's AOC Middle East, responded that due to the gravity of the situation in North Africa, the German campaign in the Balkans and Greece, the considered view was that Iraq's priority was too low and nothing could be spared.
The next day Wavell advised the Foreign Office that any support for Iraq was beyond his means.
A frustrated Churchill contacted Gen Claude Auchinleck, the C-in-C India, who agreed to spare an infantry brigade and an artillery regiment intended for Malaya. Convoy BP7, supported by the aircraft carrier HMS Hermes, the cruiser HMS Emerald and a few small warships, set sail for Basra from Bombay on 12 April.
Baghdad, 16 April 1941
It had taken Cornwallis less than twenty minutes to drive the two miles from the Embassy, cross the new King Ghazi bridge and arrive at the Serai government building which housed the Council of Ministers and Rashid Ali's office. In 1935 it would have taken him almost two hours to cross the wide river by ferry and then pass through streets thronged with people.
Cornwallis had a fascination for the streets of Baghdad and its cosmopolitan inhabitants, who could be distinguished by their bewildering array of headgear and clothes. All of which seemed to Cornwallis to be an attempt to resist assimilation.
The wealthy, educated and employed Iraqis wore western dress topped with black sidora caps. The Effendis wore red fezzes, the Kurds in their sheepskin coats had black felt pot-caps perched on their heads and the Mujtahids, the elders, wore black turbans. The burnoosed marsh Arabs in their gallabeahs and the Bedouins from the desert in camel-hair cloaks topped by agals and keffiyehs mixed with Sikhs in khaki and white turbans, Baktiaris in their brimless top hats, Bengalis in dhotis and Europeans in tropical suits topped by straw hats or topees. They were only part of the rich pageant of styles, colours and materials. Moving among them were Persian pilgrims on their way to and from the Shia holy places in black and brown abbas, Sunni and Shia mullahs, priests from the Orthodox faiths and rabbis and Arab, Armenian, Christian and Jewish merchants and the odd Parsee in a hat like a cow's foot.
Here and there were story tellers and scribes, money lenders and carpet sellers and the tea and water carriers, and weaving among them would be donkeys, camels, horses, oxen and even barefoot children pulling carts loaded with all sorts of imaginable, and even unimaginable goods. Everywhere were fellahin and the ubiquitous beggars all wearing once-white dishdash. Gliding amost silently, and virtually unnoticed, through the melée were women in black abbas and veils.
Baghdad's miles of covered bazaars were interconnected with vaulted passageways crammed with a bewildering array of stalls piled high with bananas and dates, manna and fish and grain and nuts. Inside was a riot of smells of dung and wood smoke mixed with incense and spices and a cacophony of sound, as people haggled in Arabic, Armenian, Farsi, Greek, Hebrew, Hindi, Turkish, Syriac and English amid the incessant barking of dogs and the braying of donkeys.
Cornwallis could see that since the coup d'état things had changed. The city was now strangely silent. The soldiers on every street in dark khaki topees or fezzes, wearing khaki shirts and shorts, with rifles slung over their shoulders, watching the few people that ventured into the streets, created an air of menace.
As Cornwallis's car drove down Raschid Street he saw that Dellal's department store was closed, as were the banks – The Eastern, The Ottoman and The Rafidan – while shutters covered the store fronts of Bata Shoes, Loya Brothers American Watch Company, Abu Maurice's ice-cream parlour and the cinema that showed Egyptian films. Even those fixtures of Baghdad street life – the outdoor cafes with customers smoking their nargilehs and the hawkers, peddlers and street vendors – were all gone. The wise Baghdadis stayed at home, fearful that they could be part of any further round-up of people who could possibly challenge the new regime.
Cornwallis's journey had not been without incident. Twice when the Daimler, flying the Union pennant, had slowed down to turn a corner people on the pavement had spat at the car, confirming to Cornwallis that if he had arrived sooner things might now have been different.
Cornwallis was responding to an invitation from Rashid Ali to meet to discuss, as it was delicately phrased, 'matters of mutual interest'.
He had no intention of letting the meeting become a negotiation session or of being ambushed by a group of people from the new regime and harangued, and he had stipulated that they meet alone for an hour.
Leaving his companion, Adrian Holman, the Councillor and Cornwallis's deputy, in the anteroom, Cornwallis entered Rashid Ali's office. The Prime Minister rose from his desk smiling and indicated that they should sit at a group of comfortable chairs.
'Thank you for coming, Sir Kinahan,' Rashid Ali, a small bespectacled man, began in Arabic as they shook hands, 'it is nice to see you again despite the circumstances under which we meet. My government is concerned about the poor feelings that the British have about our new regime. I have been instructed to find a way in which we can quickly reach a mutually beneficial agreement so we can normalize relations between our countries.'
Cornwallis was too old a hand to be drawn into an endless discussion in flowery Arabic. Exaggerations, implied meanings and false perceptions of wounded pride were widely used in Arabic as bargaining points throughout the Middle East.
'Rashid Ali', he responded in English in his slow, deep, husky voice, to convey a simple but one meaning message. He was also careful not to use the title of Prime Minister, which could be taken as an indication that the British government accepted the legitimacy of his office and his regime. 'I would like your permission in accordance with the Anglo-Iraqi Treaty of 1930 to land additional troops at Basra and to have them transit the country. You have confirmed on 3 April that your regime intends to honour its obligations, and, given the situation in Iraq, we consider that the presence of British troops can hardy fail to act as a stabilizing factor. Consequently this would be to our mutual benefit.'
He waited a few seconds for the message to sink in. 'Furthermore, when we receive your unconditional approval to establish our communication route through Iraq from Basra, the British government will immediately extend informal relations, and after a while we will fully recognize your government.'
Rashid Ali, unused to such directness, was momentarily taken aback. 'Naturally I will need to discuss this with my cabinet before I can give you a definitive answer,' he responded. 'I have a number of other issues, Sir Kinahan, which we should also discuss, which concern propaganda, gaining public support for the rupture of diplomatic relations with Italy, the White Paper on Palestine, as well as problems of interpretation of the Anglo-Iraqi Treaty. I would like these to be linked to your request for landing additional troops and diplomatic recognition so that we have a comprehensive agreement.'
'No', said Cornwallis, sidestepping Rashid Ali's attempt to broaden the discussions to drag out their resolution, so giving time for help to arrive from the Axis. 'We have to handle each issue separately, and begin with a simple permission to land and transit our troops in accordance with our agreement. As a gesture of good faith we will then immediately resume informal relations with your government. You will now have to excuse me', he said as he stood up. 'I have a prior engagement and I was only just able to fit in our discussion among my other meetings today.'
He shook hands with Rashid Ali, and as they walked together towards the large double doors he added, 'I very much look forward to receiving your agreement at the earliest possible time so that I may initiate the process of resuming informal relations.'
While Cornwallis kept his silence until he was safely back in his office in the Embassy, Rashid Ali was immediately on the telephone, arranging a meeting with his fellow conspirators to discuss the British offer.
Baghdad, 17 April 1941
At ten o'clock in the morning the Mufti, three of the four Colonels of the Golden Square and Yunis al Sabawi met with Rashid Ali in al Sabbagh's office in the Citadel.
The Citadel was where the Ottoman governors of the Vilayet of Baghdad had lived and where Feisal had stayed during the early part of his reign. It now formed part of the complex which housed the headquarters of the Iraqi armed forces.
Rashid Ali quickly outlined Cornwallis's offer and his acceptance, but minimized his failure to broaden the discussions.
'This is treason to the nationalist cause!' shouted Kamil Shabib, a small, obstinate and arrogant colonel who commanded the 1st Division. Salah al Din al Sabbagh, a tall sallow man with a reputation of being a careerist with ambition, who commanded the 3rd Division, added in a quiet yet forceful way, 'Rashid Ali, the Army opposes your undertaking with the British. This is just a tactic to stall for time until they are strong enough to move against us.'
'You have to understand, Rashid Ali, that a war with Britain is necessary – if not immediately, then within a matter of days or weeks', said al Sabawi.
'There can be no compromises', Colonel Mahmud Salman, the head of the Royal Iraqi Air Force, added.
The Mufti contemplated the position before raising his hand to stop the diatribe against Rashid Ali. 'On the one hand we must not do anything to provoke the British to occupy Iraq before the delivery of weapons from our allies. On the other hand if we allow the British to land troops and transit them through Iraq we give them the opportunity to occupy us.
The Germans and Italians will not be pleased if we allow British troops to enter the country in force as it would compromise any military action that they will take to support us. They are unlikely to help us if they are convinced that we are an occupied country. There is also the need to manage the public perception that we are capitulating to the British, which could lose us popular support.
I recommend', said Haj Amin, 'that we accept the British request. However, we need to stipulate that this is acceptable only if it is limited to 3,000 troops in Iraq at any one time, and under the condition that they proceed as rapidly as possible from Basra to Rutba Wells. This way we can limit their size and presence while still fulfilling our legal obligations.'
The Colonels, Rashid Ali and al Sabawi nodded their agreement, and the Mufti suggested that Rashid Ali meet Luigi Gabrielli to advise him of their planned action.
Baghdad, 19 April 1941
Five significant events occurred that day.
Rashid Ali sent a message to Cornwallis at the Embassy. While confirming the Iraqi government's agreement to the landing of British troops in Basra, it had three conditions. Firstly, that all troops must proceed as rapidly as possible to Rutba Wells. Secondly, that the British government must announce well in advance any intention to ship any more detachments of troops to Iraq. Finally, British troops must not exceed at any one time the strength of one mixed brigade, and that no more units should land at Basra as long as there were still some British troops in the country. Cornwallis pointedly sent no reply.
The 20th Indian Infantry Brigade, the 3rd Field Regiment RA and the Headquarters of the 10th Indian Division began their disembarkation from Convoy BP7 at Maqil docks, Basra, and Maj Gen William Fraser assumed command of all British Army forces in Iraq.
Rashid Ali met the Italian Minister at his office. 'This is to advise you of the decisions taken by the National Defence Council yesterday.' He went on to outline that the Iraqi response to the British landing was a way of stalling for time. 'However, Minister, we need to know as quickly as possible from the governments of Italy and Germany whether we can count on support from the air forces of the Axis. Naturally our airfields will be placed at your disposal. We also need to know whether you can send our orders for weapons by air. I would also like to reiterate our request of 25 March that all the arms requested and the financial support promised be sent as soon as possible.'
Iraqi hostility began to show itself. One of two 31 Squadron old biplane Valentias carrying troops from Shaiba to Habbaniya was forced to put down at Pumping Station K-4 on the Tripoli pipeline, and it was immediately attacked by Iraqi troops. The crew and the fifteen soldiers on board were captured and the Valentia was set on fire. The other Valentia attempted to land but was driven off by Iraqi ground fire, which wounded two of the crew.
A change in AOC in Cairo to AVM Arthur Tedder brought reinforcements in the shape of six obsolete and worn-out Gloster Gladiator biplane fighters. They were flown in to Habbaniya from storage in Cairo. A request from AVM Harry Smart commanding RAF Habbaniya to retain the motley collection of ferry pilots – two Free French, one Pole, a Yugoslav and two Britons – to fly the aircraft was overruled by Tedder, and the pilots were flown back to Cairo.
Habbaniya, 20 April 1941
Smart realized that there was little, if anything, that could be sent in terms of reinforcements in the short term. Habbaniya was on its own. Although he had every confidence in Cornwallis, Rashid Ali did not seem to be interested in negotiation.
RAF Habbaniya had been designed from the outset as a peacetime flying school, and was indefensible against a ground attack. While Smart agonized over the outcome of negotiations and what might happen to his command should it be attacked, a meeting was taking place in the nearby office of the Commander of No. 4 Service Flying Training School, Gp Capt William Savile.
Larry Ling, a young wing commander and chief flying instructor, had approached Savile with one of his flying instructors, Sqn Ldr Tony Dudgeon, with ideas about what could be done to protect Habbaniya in the case of an attack.
Savile, a non-flying officer, was regulation and discipline bound. He was more interested in keeping up the flying programme to meet Habbaniya's quota of pilots than in doing anything which could interrupt the programme.
'Sir, I think that you will be interested in Tony's ideas, which have relevance as they come from one of the few pilots at Habbaniya with recent air combat experience', Ling began.
Savile nodded his approval with a frown. He did not like his officers referring to their subordinates by their Christian names, believing this to led to poor discipline.
'Sir,' began Dudgeon, 'all of our aircraft are obsolete and are not designed for modern warfare. However, I have done some checking. I believe that with a little imagination we could make what we have into a potent strike force.
'We have nine Gladiators, our three plus the six which arrived from Cairo yesterday. Although they are pretty clapped-out we can still turn these into useful fighters. Our only problem is belted ammunition for their four Brownings. We do have a slow hand-cranked belt filler in the armoury, together with belts and .303 ammunition. If we run the machine around the clock in shifts we can get sufficient ammunition to arm the Gladiators. We need to run the belt filler around the clock as a ten-second burst from the four Brownings will fire off an hour's work, and we will need as many belts as we can get our hands on.
'Our Fairey Gordon biplane bombers date back to 1931, and we now use them to tow targets. But they still have their bomb-release gear, and if we fit Universal racks they can handle two 250 lb bombs. This will give us seven bombers.
'We have thirty Hawker Audax biplane trainers which are rigged to carry eight 20 lb practice bombs. With the bomb-release gear already in place all we have to do is fit Universal racks, and again we have bombers that can handle a useful load of two 250 lb bombs.'
Savile interrupted. 'But the Audax were Army cooperation aircraft and not bombers like the twenty five Hawker Harts that we have on strength?'
'Yes,' replied Dudgeon, 'but the Harts have been stripped of their mountings, cables and levers for practice bombing so there is little that we can do with them. The Audax is essential a Hart. I have flown them in India with a load of over 1,000 lb, so we should be able to convert the Audax to a bomber, as it has all the equipment excepting the Universal racks.
'We have only nineteen sets of Universal racks in the stores, so with the seven for the Gordons we can create another twelve bombers with 500 lb bomb-loads out of the thirty Audaxes we have on strength. The remaining eighteen Audaxes can carry the eight 20 lb bomb-loads.
'Finally we have twenty-seven Airspeed Oxfords which we use for advanced and photographic training. They are rigged to carry eight 9 lb practice smoke-bombs. The airframe is largely wood and is not strong enough to handle 250 lb bombs. But if we cut some holes in the bottom of the fuselage and bolt two plates onto the plywood fuselage and mount bomb-racks on the plates, we can carry a useful eight 20 lb bombs.
Dudgeon and Ling sat back waiting for Savile's reply.
Savile readily agreed to the Gladiators and that they be commanded by Flt Lt Dicky Cleaver, and to the modifications to the Gordons and to their command by Flt Lt David Evans. However, when the conversation turned to the Audax and the Oxford the air turned purple. Savile turned around and took two weighty volumes off the shelving behind his desk and placed them in front of Ling and Dudgeon. 'There is no standard bomb-load quoted for either the Audax or the Oxford in the manufacturers' manuals. No quotation means no modification.'
Ling argued tenaciously, citing Dudgeon's experience with the Audax, and finally had Savile, more to get them out of his office, agreeing to Ling and Dudgeon talking to the Engineers.
The Engineers were even more straight laced, and would not countenance any modifications. Through his continuing tenacity, Ling got the matter referred to the Operating Air Staff. While they did see the point, it was not authorized. Again Ling's tenacity got authorization for the Engineers to contact the Air Ministry in London.
Habbaniya, 21 April 1941
The Habbaniya Engineers, allowed to draft the signal to London, asked what the bomb-load was for the Audax. The obvious reply was eight 9 lb practice smoke-bombs, which ruled out authorization for the modification.
While the Engineers were happy, Ling and Dudgeon, over several beers in the officers' mess, decided on a new tack. Ling went back alone to Savile. 'Sir, out of courtesy I am advising you that we are going ahead and fitting the Universal racks to an Audax and taking it up for a test flight.'
'Absolutely No!' spluttered Savile.
'I know the problem, but we potentially face a much bigger problem. Unless we defend ourselves we are going to have to surrender. This will be the only RAF station in history to give in to an enemy without firing a shot in its defence. We have the opportunity. All I want is to seize that opportunity. It will not look good in the official enquiry that we did nothing to defend ourselves.'
Savile thought for a moment, realizing the implications. 'All right, I want a written note from you, or Dudgeon, that the test flight is being made on your sole responsibility and in the full knowledge that this is against my wish and my advice.'
'No, Sir,' Ling replied, furious at the buck-passing, 'I will not write such a note. We are all in this together and I plan to take this action on behalf of the station.'
Within one hour the Audax had the racks fitted and had made its successful test flight.
The Oxford proved to be a bigger problem. Dudgeon went to the Station Workshops to use their cutting and drilling machines to make the hundred or so metal strips to fix to the bottom of the Oxford's fuselage as the base on which to bolt the light series bomb-racks.
The Workshops flatly refused to make the strips unless Headquarters accepted all responsibility for the change of aircraft load and to its flying characteristics, to be responsible for whatever might happen and to supply an official set of blueprints for the modifications.
The Engineering staff officers were also against any 'unauthorized modification', which could only be agreed by the Air Ministry in London after discussions with Airspeed, the designers and manufacturers, of the Oxford.
Smart got wind of the situation, and to make sure that there was no repetition of the unauthorized Audax flight issued written orders to Ling, forbidding a test flight.
Ling and Dudgeon sat in the mess. 'Tony. I am not showing you Smart's orders, so you don't know its content. The only way we can do this is for you to do it alone. You have not seen this order and you are not to involve anyone else.'
Dudgeon borrowed a drill, bits, punch, hammer, hacksaw, files and a work-bench, plus some strip steel that someone purloined for him. He set to work in the back of Hangar Six, creating the plates on which to bolt the bomb-racks.
Nothing can be kept quiet on an RAF station, and word soon reached Smart that Dudgeon was hard at work in Hangar Six.
Smart sent for Ling, as Dudgeon's direct superior, together with Savile. 'Ling, by not stopping Dudgeon at the outset you have conspicuously failed in your duty. This is a gross breach of Good Order and of Air Force Discipline. Clear and specific orders in writing have been issued and are being flagrantly disobeyed. If we lose a valuable aircraft or a pilot you will be held fully and personally to blame. Now I am issuing you with a clear order. Find Dudgeon and stop him. Arrest him if necessary!'
By luck Dudgeon had already taken off with eight 20 lb bombs in the racks bolted on to the bottom of the Oxford. Dudgeon had made a successful landing by the time that a very slow-walking Ling arrived at the hangar.
Within a few days of the incident all was forgotten, Ling and Dudgeon were forgiven and the Workshops were instructed to make and fit the plates to each of the remaining Oxfords
Berlin, 27 April 1941
Woermann sat in front of Ribbentrop's massive oak desk in his opulent office at the Auswartiges Amt on the Wilhelmstrasse. 'I thought, Woermann, that you should be aware of my letter to the Fuhrer which endorses action in Iraq.
'I have pointed out that the movement of British troops through Iraqi territory might turn the scales in North Africa and that Britain's real aim is to gain a jumping-off position against the Vichy in Syria and to exert pressure on Turkey. The ultimate success of British policy in Turkey might have a major influence on the situation of German troops should they fight the Red Army.
'I have also added that the scepticism shown by the Luftwaffe and Wehrmacht is unfounded as the information they submit on the number of enemy troops is too high. We know this because we have penetrated the US Embassy in Cairo and we are able to read the dispatches sent by their Legation in Baghdad, which is based on British-supplied information.
'We are also now in a position to pressure the Vichy French so that our planes can refuel in Syria on their way to Iraq. Is there anything I should add?'
'I would impress on the Fuhrer, Herr Reichsaussenminister', said Woermann, 'that we need to act quickly in providing both materiél and a military presence. The longer we delay, the more chance there is for the British to reinforce Iraq and for the Iraqis to lose the initiative and their nerve.'
Baghdad, 28 April 1941
At three o'clock in the afternoon Cornwallis met Rashid Ali at the latter's office at the Seria government building.
'Rashid Ali, the British government requests the Iraqi government's agreement to the landing of a further 3,500 troops at Basra in the next few days.'
'Ambassador, we had an agreement to limit the number of troops, the transitory nature of their stay and no retention of troops to create a base in Basra. Any deviation is a gross breach of the treaty.'
'Rashid Ali, we need to take a much broader interpretation of the treaty. We cannot obviously give you prior notification in wartime for security reasons. We also need to safeguard the Imperial lines of communication which transverse Iraq. Consequently we need more troops, and some will have to stay in the country.'
'No, No, No! We have an agreement which you are now seeking to change. This is something neither I nor my government can countenance.'
'Rashid Ali, the landing of these additional forces will nevertheless take place, and the consequences of opposing these landings will be serious.'
'Ambassador, let me also make myself perfectly clear. If these additional troops disembark before the dispatch of those already in Basra I will broadcast a denunciation of British action. I cannot and will not be responsible for the consequences which might follow an outburst of popular feeling!'
'From your response, Rashid Ali, I have grave concerns about the safety of British women and children in Iraq, and I need to send them outside the country at once. I would appreciate your giving the women and children safe-conduct.'
'Ambassador, the Iraqi government has no quarrel with women and children, and you should feel free to move them outside Iraq should you feel the necessity. As we cannot seem to make progress, I suggest that we call a halt to this meeting to give you and your government time to reflect on your position.'
The meeting with Rashid Ali had been deeply disconcerting for Cornwallis. With the Iraqis' uncompromising position, no doubt bolstered by British losses in Greece and North Africa, it seemed almost impossible to patch up the difficulties between the two countries.
Immediately on his return to the Embassy, Cornwallis sent a cipher telegram to Smart at Habbaniya initiating Operation Concentrate. Thirty RAF buses and trucks, together with a requisitioned Harrington-Marmon 38-seater from the Nairn Brothers' Baghdad-Damascus desert bus service, were to arrive at Baghdad's civil airport starting at 2.30 p.m. the next day. They were to begin to evacuate the women and children to Habbaniya, from where they would be flown to Basra. Next he called in Adrian Holman to initiate the phone calls to the Alwiyah Club, British companies with offices in Iraq and to the group of Embassy wives who would begin to call their lists of expatriates to have their wives and children assemble at the airport late the next morning.
Warning notices were also prepared and delivered to the district wardens and to the Vice-Consul for delivery on the other side of the Tigris, where the bulk of the British community lived. Finally he called Paul Knabenshue at the American Legation on the east bank of the Tigris. 'Paul, this is Ken. I have had a very unsatisfactory time today which has left me no option but to initiate our agreement. Operation Concentrate starts at eight tomorrow morning.'
Knabenshue called in his secretary and asked her to begin to call the small number of Americans in Baghdad, to have them assemble at the Legation the following morning. Cornwallis and Knabenshue had agreed on an Anglo-American plan where the citizens of either country, and any other friendly neutrals or allies, could seek refuge in the British Embassy or American Legation, whichever was closer, with the RAF transporting all British and American wives and their children to Habbaniya.
Luigi Gabrielli, the Minister at the Italian Legation, looked through the list which had been given to him by Rashid Ali. The Iraqi requests were becoming more detailed, and the tone more urgent. He wondered how long he could continue to stall Rashid Ali until Rome and Berlin agreed between themselves what they were going to do.
The Iraqis wanted large sums of money to offset the costs of their war – not only for military operations, but also to bribe the tribes for support and to compensate for the loss of Iraq's income from customs revenues and oil royalties. They also wanted Axis aircraft to bomb British naval units in the Gulf and the bases at Habbaniya and Shaiba, and offered the use of airfields at Baghdad, Mosul and Mikdadia, as well as Iraqi petrol stocks.
Rashid Ali had also requested the Axis to supply large quantities of aircraft, armoured cars, armour-piercing guns and ammunition, grenades and machine-guns. The Iraqis were particularly anxious, he noted, to obtain captured British infantry weapons with which they were familiar.
While the Iraqis were no longer insisting on a declaration from the Axis on the future independence of Arab countries, the list was long and extensive.
Gabrielli feared that Rome and Berlin would again ask him to stall, at a time when it looked as if war between Iraq and the British was inevitable, compromising his position with the Iraqi leaders.
Back in his office in the Serai building, Rashid Ali contemplated his future as Iraq drifted towards war with Britain. Throughout his political life he had changed cliques with the tide, always one step ahead in picking the winning side. His own personal pique had forced him to turn to the Mufti and the Colonels and the Axis. He had played both the Germans and the British in an attempt to maintain neutrality in a war which did not concern Iraq, while veering to support the Axis as insurance in case they, rather than the British, won.
Now the Colonels and the Mufti had made the decision to go to war, shattering the delicate balance he had created and striven to maintain. Iraq's only hope was urgent support from the Germans and Italians, not only in terms of money and equipment but also in the form of a military mission, complete with ground troops and airmen to stiffen the Iraqi armed forces.
Ling and Dudgeon's request to Savile to prepare the polo field and the golf course as an additional landing and parking area was finally approved. Both had argued that the airfield alone was not large enough to support the Habbaniya Striking Force, and they needed an alternative. Although opposed by some of the senior officers, Savile, much to the surprise of Larry and Tony, had fought tenaciously and won permission from Smart. Two steamrollers had been pushed into service, and with floodlights would finish the make-shift runway before midnight that day.
Baghdad, 29 April 1941
Just after ten that morning, buses and cars began to arrive at the British Embassy and the American Legation, and started to ferry the women and children to Baghdad's civil airport. Only a few women chose to stay behind. The wives of the British Ambassador and the US Minister remained, together with two Consulate wives, one typist and the Matron and nursing sisters from the hospital.
Masked by the loading of the women and children, two covered RAF trucks pulled up outside the Chancellery of the British Embassy. Airmen began to unload sandbags, rolls of barbed wire, several cases of rifles, ammunition and a few tear-gas bombs, which were labelled variously 'stationery' and 'garden tools', together with cardboard boxes full of tinned goods. All of these were hurried out of sight into the ballroom for storage.
By the end of the day 132 women and ninety-nine children had been registered and packed aboard a convoy of buses and lorries at the airport. Twenty minutes later the convoy completing Operation Concentrate moved off in the direction of Habbaniya.
An emergency cabinet meeting was held at al Sabbagh's office in the Citadel at eleven in the morning with Rashid Ali, the four Colonels and the Mufti.
Rashid Ali summed up the situation. 'The British action clearly signals that the British have decided to pay no attention to their agreement and that they plan not only to station more troops on our soil, but to occupy the whole of Iraq. I believe that we have no choice now but to resist the British with the full force of our Army.'
'Are we all in agreement?' asked Rashid Ali.
'I think that this is an opportune time', said al Sabbagh. 'The British are isolated and in an indefensible position at Habbaniya. If we breach the bunds in the south and cut the railway and telegraph links between Basra and Baghdad, we prevent their troops, three hundred miles away, from coming to their aid. Habbaniya gives us a bargaining chip should the British try to force the issue.'
'I agree', said Salman. 'Our Air Force is numerically larger and our planes are more modern than those the British have in Iraq. With all their other commitments I do not see them reinforcing Habbaniya before we overwhelm them.'
'They have no armour and no artillery and are not a match for my Mechanized Force', added Fahmi Said.
'It is time we showed the British that we are not a vassal state', added Kamil Shabib.
'Naturally I will send a note of protest to the British Embassy rejecting their demand to land more troops, and our objections against any further landings.'
'We also need to send a further note to Gabrielli at the Italian Legation, advising him of our plans and requesting immediate military and financial support from Italy and Germany', suggested the Mufti.
'That is under way, Haj Amin', confirmed Rashid Ali.
'Well, gentlemen,' said the Mufti, rising from his seat, 'the situation has developed much better than I could have hoped for. If the British stand and fight we will win. If they do not fight we will starve them out of their airfield. Whatever the outcome we will humiliate the British.
They will be forced to seek an agreement on our terms. This will mean the end of British involvement in Iraq and a just settlement for Palestine. Britain's loss of face will have profound effects throughout the Middle East and will make their position in both Egypt and Palestine impossible. Our victory will also send shock waves through their other possessions, like India and Burma, where they also face unrest.
I will leave you to plan for our victory!'
At the time of the meeting of the Golden Square, men of the King's Own Royal Regiment began to arrive at Habbaniya from Shaiba in an airlift by 31 Squadron's old Vickers Valentia troop carriers. By the end of the day 364 officers and men of the 1st Battalion, with their twelve light and six heavy Vickers machine-guns and two Boys anti-tank rifles, were reinforcing RAF Habbaniya.
In the late afternoon additional elements of 20th Indian Brigade landed at Basra from Convoy BP1 without incident, despite the Iraqi government's refusal of permission to land.
Both the British and the Germans failed to exploit their opportunities, and as a result a war started which could either have been easily averted or quickly won.
Inaction by the British in nipping the problem 'in the bud', and the snail-like pace of the Germans in taking advantage of Britain's vulnerability, appears strange in hindsight. However, this was due to competing sets of interests on both sides, which coloured their perceptions of the value of Iraq and the urgency of their responses.
In the British camp there were vastly different views held about the evolving problem. The men on the spot, Wavell in Cairo and Cornwallis in Baghdad, understood the issues of Arab nationalism, and with the lack of British resources saw armed intervention as both provocative and risky, believing diplomacy and conciliation were sound roads to take. This position was supported by AVM Smart, who commanded RAF Habbaniya, who saw conciliation as a way to avoid armed confrontation on the ground which he would lose.
Wavell's position was understandable. He was already fighting for his life in North Africa against Rommel and in East Africa against the Italians. He was awaiting an air assault on Crete on his battered forces that had been evacuated from Greece. He was also unsure of the intentions of the well-armed Vichy French in Syria, who could invade Palestine at their choosing. With too many demands on his few troops and limited equipment, Wavell was reluctant to assume more commitments.
In contrast, Churchill in London and Claude Auchinleck, the C-in-C India, favoured rapid armed intervention, not only to safeguard oil supplies and the quickest way to India by air and land, but to avoid undermining the British position in the Middle East as a whole. Both men feared that conciliation, together with Britain's lackustre performance against the Germans, could signal weakness and incite unrest throughout the region.
Churchill, the strategist, saw the British position in the Middle East as a house of cards held in place by Egypt. Already pressured in the Western Desert by Rommel and in Abyssinia by the Duke of Aosta, and the possible loss of Crete, would make it easy for Axis forces to take Cyprus and to then threaten Egypt from the north, west and south. A German bridgehead in Iraq, together with the Vichy French in Syria, would enable the Axis to complete their encirclement of Egypt, forcing the collapse of Britain's position in the Middle East and cutting Britain off from India, Australia and New Zealand.
Both Churchill and Wavell realized that they were fighting a global conflict which demanded difficult strategic choices, each had a powerful intellect and was an aristocrat and scholar, and each was attracted to unorthodox solutions. However, there the similarities ended.
Wavell was one of the British Army's great trainers of men, a meticulous planner with a talent for administrative detail who was attuned to the complexities of operation rather than the strategic implications. Wavell was an altogether too simple, too straight and too decent a person to survive in the muddy waters of politics where the intriguers laid traps and the sharks hunted. However, Wavell lacked the imagination, the decisiveness and the speed of action which are the hallmarks of successful military leadership. In contrast Churchill had strategic vision and demanded from his commanders an enthusiasm that bordered on zealotry, and took it upon himself to meddle in campaigns from a distance, often in excruciating detail.
Wavell's resentment of Churchill's constant meddling led him to shield information from Churchill. This lack of transparency and Wavell's caution, bordering on inaction, only increased Churchill's distrust of his C-in-C Middle East.
Things were no better in Berlin. Advocates of support for Iraq were primarily driven by two different factions in the Auswartiges Amt. Grobba and Woermann were concerned with exploiting their carefully built relationships in Iraq to strengthen Germany's position while weakening that of a vulnerable Britain. However, Von Hentig and Von Weiszacher were more interested in using Iraq to deflect Hitler from prosecuting Operation Barbarossa, the attack on Russia, which they felt was the wrong fight for Germany at the wrong time.
Hitler was totally focused on Barbarossa. believing that Iraq was a side show. His racial theories about Semites, which equated Arab with Jew, provided a stumbling block to overcome before support could be given to Iraq. Adding to these problems was the Luftwaffe, which felt that Iraq was too far away to provide meaningful help. The Wehrmacht, with its hands full in Greece and Crete and with Rommel in Africa, did not want to get involved in another 'adventure', given the pending attack on Russia.
Complicating the matter further was an agreement with Mussolini which gave Italy a free hand in the Middle East to create a new colonial empire. With the need to keep Italy firmly in the Axis fold, both Hitler and Ribbentrop were reluctant to take any action which would compromise their relationship with Mussolini.
These perceptions and the delayed responses by both sides not only caused the crisis but dictated how it would play out and when and how it would end.
Chapter Two
Target Habbaniya
Baghdad, 30 April 1941
A little after three in the morning Pat Domvile, the Air Liason Officer, gently shook awake Maj Gen George Waterhouse, Head of the British Military Mission, who was asleep on a camp bed in his office in the Embassy.
'Sir, we have company', he said to a sleepy-eyed Waterhouse. 'The guards woke me about half an hour ago. They thought that we were being surrounded by the whole Iraqi Army. I took a shufti along the river bank. There are trucks filled with troops, towed artillery, tanks and armoured cars heading west across the King Feisal Bridge. It looks to me that they are moving out of the Raschid Camp. They have not come our way, and as the only possible target to the west is Habbaniya my chaps could be in for a spot of trouble.'
'You had better let them know', said Waterhouse.
'I have already sent a cipher telegram to the Air Vice-Marshal,' Domvile said, 'but I wanted you to be aware of events.'
'Good thinking, Domvile. Now that I'm up I need a cup of tea!'
Habbaniya
It took twenty-eight minutes for the cipher telegram to reach Smart. Sending the cipher took seconds. The rest of the time was spent in decryption by the Duty Officer, verification by the Chief Signals Officer, a dispatch rider to take the message to Smart's quarters and for his batman to wake him.
Quickly reading the message, Smart had his batman contact his ADC to arrange an emergency meeting in the conference room adjacent to his office on the top floor of the Air Headquarters block at 04.00. The summons went to Savile, Wg Cdr Paul Holder, the Senior Administration Officer, Lt Col Alastair Brawn commanding the RAF Levies, and Lt Col Edward Everett of the King's Own Royal Regiment.
At precisely 04.00 Smart, a small, dapper man, walked into his conference room where those summoned to the meeting waited.
'As you all know, the situation over the last month in Iraq has been very tense. Domvile at the Embassy sent me a cipher at 03.00 which suggests that Iraqi troops, guns and armour are headed in our direction. Whether this is fact or surmise by Domvile I just don't know. However, we need to be prepared for any outcome. Bill,' to Savile, 'at first light, which should be in about an hour, I want you to get an Audax up and take a shufti to see what, if anything, the Iraqis are up to.
'I am going to sound a General Alarm and I want you all back here at 07.00, together with Page who runs the Armoured Car Company and the Chief Flying Instructor. I would like the Levies, the Army and the Armoured Car unit to give me a ground defence plan. Bill, I want you and Ling to brief us on the air striking component, and Paul, I would like you to let us know our supplies situation.'
After sounding the General Alarm, Smart went to his office and sat quietly contemplating his two predicaments.
Firstly, RAF Habbaniya, while in many ways a perfect posting, was almost defenceless. The station, situated on a bend on the Euphrates, embraced RAF Headquarters Iraq, No. 4 Service Flying Training School, six large hangars, an aircraft depot with repair and workshops, a hospital, a meteorological station, administration and barrack blocks and an airfield with two intersecting runways of packed sand. Complicating things was the fact that Habbaniya was home to some 9,000 people – Arabs, Armenians, Assyrians, Indians and Kurds and their families – who worked at the station.
The base, built in 1937, contained all the amenities of a peacetime station. RAF Habbaniya boasted a social club, a swimming pool, a polo ground, riding stables, thirty-six tennis courts, rugby, soccer, and hockey fields, a cricket ground, a gymnasium, a golf course and the Astra cinema with its outdoor and indoor facilities. Incongruously it also had horses and a pack of hounds known as the Royal Exodus Hunt managed by Lt Alistair Graham of the Levies, and a Yacht Club, based on the other side of the plateau, on Lake Habbaniya.
Lines of eucalyptus trees were planted along all twenty-eight miles of the camp's paved roads to give shade. Hibiscus and oleander shrubs, green lawns and carefully tended vegetable gardens were dotted around the station. The camp had its own water supplies contained in a single, high tower, and an electric power station provided all the camp's power, while three churches catered for the Anglican, the Armenian and the Assyrian communities.
RAF Habbaniya was not built with war in mind, and had none of the usual defences of dispersal areas, bomb-proof shelters, redoubts and antiaircraft positions that an operational station would contain in a war zone.
The only protection for the camp's five hundred acres was a seven-mile-long, ten-foot-high steel perimeter fence, curved outwards to keep the kleptiwallahs at bay. The fence was interspersed, every 300 yards or so, by one of fourteen twenty-foot-high brick blockhouses. However, the airfield was outside the perimeter fence. The Air Ministry had considered that expanding the perimeter fence was too expensive to protect an airfield of packed sand which sat in a boiling desert miles from anywhere in a friendly country.
On the camp's southern and western flanks the station was dominated by a plateau some one hundred and fifty feet high, while to the north and east the Euphrates contained the station. Beyond the plateau to the south was Lake Habbaniya used by RAF and BOAC flying-boats. RAF Habbaniya was a very soft target for both ground and air attack.
Secondly, the senior officers, including Smart, reflected the ambience of Habbaniya. They had been selected on the basis of suitability for managing training. Their rules, regulations, targets, timetables and patterns of work were designed solely for the production of an efficient, effective and regular flow of trained aircrew. Anything that interfered with this process could not, and would not, be tolerated.
AVM Harry 'Reggie' Smart was a very brave man who had served his country with distinction in both war and peace, in the air and on the ground. But he knew that both he and his senior officers lacked the leadership qualities, the flexibility in thinking, the sense of urgency and the combat experience demanded in a modern war to turn the adversity they faced into success. Smart dreaded the coming meeting.
Half an hour after first light, Colonel Fahmi Said, commander of the Mechanized Force and a key member of the Golden Square, stood at the edge of the escarpment looking down on RAF Habbaniya.
From his vantage point the camp and its airfield were spread out before him, like a toy village, with the Euphrates like a silver snake coiled at its back. The hangars, the barrack blocks and the playing fields all dominated by the water tower seemed so close in the sharp early morning sunlight that Said felt he could reach out and touch them.
As the single Audax landed back at the camp, Said smiled, knowing that the British would be aware that his armour and guns now totally dominated their base and that more troops were on the way to support him.
Now was the time to challenge the British. He signalled to his ADC, Major Hussain Najib, that it was time to take the message to Habbaniya.
The Ford staff car carrying a white flag passed by the signpost with RAF Habbaniya in English and Arabic topped by the little direction board pointing to London, a long 3,287 miles away, and Baghdad, now an uncomfortable fifty-five miles near.
To Najib the sign seemed to amplify Habbaniya's isolation and vulnerability, and to reinforce the Iraqi Army's dominance over the British.
At precisely 6 am Najib's car pulled up at the guardhouse at the main gate. Najib, immaculately dressed in his cavalry uniform complete with gleaming riding-boots, got out carrying a white flag and walked over to the duty officer who was just emerging from the guardroom.
'I have an urgent message for your commanding officer', Najib stated, dispensing with a salute and looking contemptuously at the young pilot officer wearing a khaki shirt and shorts and an RAF Police arm band.
The pilot officer snapped off a salute and the guard detail came to attention. 'I will have you taken to his office, sir. But first you must wear a blindfold.'
The pilot officer motioned to a burly flight sergeant to fix the blindfold, and Najib was then led back to the car. A grinning sergeant from the Levies, immaculate in a pale blue open-necked shirt, khaki shorts and puttees and a dark brown felt hat turned up at the side, Australian style, climbed onto the running board to direct the driver to the two-storey Air Headquarters block a mile away.
Twenty minutes later Smart contemplated Said's message:
For the purposes of training we have occupied the Habbaniya hills. Please make no flying or the going out of any force or persons from the cantonment. If any aircraft or armoured car attempts to go out it will be shelled by our batteries, and we will not be responsible for it.
Smart realized that he could not comply, yet had no orders to take things further, and decided to bluff it out replying:
Any interference with training flights will be considered an 'act of war' and will be met by immediate counter-offensive action. We demand the withdrawal of the Iraqi forces from positions which are clearly hostile and must place my camp at their mercy.
When Najib had left with the reply to Said, Smart immediately signalled Cornwallis in Baghdad, Fraser in Basra and Tedder in Cairo with Said's message and his reply. He asked Cornwallis to arrange for the immediate withdrawal of the Iraqi forces from Habbaniya, adding that any increase in Iraqi forces would compel him to take air action.
Punctually at seven, Smart entered the conference room adjacent to his office and gestured to the participants to bring their coffee and take their places at the table.
'As you probably know, I received an envoy from the Iraqis, who told us to stop flying. I told him that we will continue to fly and any action that they may take will be considered as an act of war. I regret having to force the issue but I had no choice. Right. Bill, why don't you bring us up to date on the Iraqis' deployment?'
'Ling here', said Savile, 'conducted the air reconnaissance.'
'Things do not look good. I flew across the plateau and then carried on to Khan Nuqta, half way between Falluja and Baghdad, before looping back across Lake Habbaniya, and then flew at low level over the plateau for a closer look. There are about 2,000 Iraqi troops on the plateau busy digging-in. There are a number of gun batteries and machine-gun positions already in place, plus some tanks and armoured cars. All of which are protected by anti-aircraft guns including some 20 mm cannon.
'If you look out of the window you can see them for yourself. Some of the heavy guns are sited less than 2,000 feet from our perimeter, and cover the camp at point-blank range. The road from the plateau back past Falluja towards Baghdad is jam-packed with troops in trucks, towed artillery, armoured cars and some of their new light tanks. I estimate that by the middle of tomorrow there will be between 8,000 and 9,000 Iraqi troops up there on the plateau.
'Coming back over Lake Habbaniya, I saw two of the big Empire flying-boats sitting on the water. There were no Iraqi military formations on their way, but I suspect that a visit to BOAC is only a matter of time. I took photos of the Iraqi positions on the plateau, as well as along the Falluja road, which we can use to both verify the number of troops we face and map out their positions on the plateau.'
Smart was clearly shocked. 'This is a gigantic bluff! They are just throwing their weight about to see if we will crack. Bill, I want you to get every airman to start digging trenches. Commandeer all the airmen, all the pupils and junior officers and get them to start digging defence trenches around the camp.'
'Why don't we hear from the land forces to ensure that we work in concert on our defences?' responded Savile, taken aback by Smart's outburst.
'A good idea, Bill. Why don't you start, Col Brawn, and let me know how you plan to deploy the land forces?'
'The overall plan for ground defence has been a joint undertaking with the King's Own and the Armoured Car unit, so I will just elaborate on the deployment of the Levies. We have a strength of 1,199 officers and other ranks composed of five companies of Assyrians, one company of Kurds, together totalling 750 men, plus a headquarters company. Each of the blockhouses along the perimeter fence will contain two NCOs and six troopers. I have divided the camp into three sectors, each covered by one company with two companies in a mobile reserve.'
'The Levies are Iraqis and have never been in combat, Colonel. What is their reliability and capability?' questioned Smart.
'The Assyrians are Christians and the Kurds, although Muslim, are not Arabs. Both have recently suffered very badly at the hands of the Iraqis, so I am sure that we can count on their support. They have been well trained and are very enthusiastic, and I am sure that they will acquit themselves well. However, the Levies are lightly armed. We have one machine-gun section armed with some ancient Hotchkiss and Lewis guns, a 3-inch mortar section and one Boys anti-tank rifle.'
'Col Everett, what can the King's Own do to help out?' asked Smart.
'We have 364 officers and men. Given the size of the base, it would be best to split my force into two. Half of the King's Own will be involved in point defence around the key installations like the Air Headquarters block, the power station, the hospital, the fuel store and the water tower. The other half will join the Levies' mobile reserve. We have a few Vickers heavy machine-guns and some Boys anti-tank rifles.'
Smart turned to the camp's youngest 'military man', Sqn Ldr John Page, who commanded No. 1 Armoured Car Company RAF.
'How are you fixed, Page?' questioned Smart.
'I have eighteen armoured cars. These were originally built for the Royal Navy in 1915 on Silver Spirit chassis. They are thin skinned and no match in a confrontation with the modern Crossley armoured cars or the FIAT light tanks operated by the Iraqis. The Colonels and I agree that our best bet is to use these with their heavy Vickers machine-guns as part of the mobile reserve to break up infantry ground attacks and to carry the fight to the enemy at night. We have also given some of the cars a Boys anti-tank rifle to give us a bit more punch if we come up against Iraqi armour.'
'Bill, you have command of the Air Striking Force. What do you plan?'
'We have divided our forces in two. The Audaxes will be in two squadrons under the overall command of Ling, here. One squadron of twelve aircraft with two 250 lb bombs will be run by Wg Cdr John Hawtrey. The second squadron of nine Audaxes carrying eight 20 lb bombs will be headed by Wg Cdr Glynn Silyn-Roberts. Paul Holder, here, will have a roving commission to fly Audaxes for anyone. Ling's force will fly out of the converted polo field, which is out of sight of the gunners on the plateau. The rest of the aircraft, under the overall command of Sqn Ldr Tony Dudgeon, will fly out of the main airfield. Dudgeon will have the nine Gladiators for airfield defence under Flt Lt Dicky Cleaver. The seven Gordons with their two 250 lb bomb-loads will be run by Flt Lt David Evans, while Dudgeon will command all twenty-seven Oxfords with their eight 20 lb bomb-loads. While we can field sixty-four aircraft out of the eighty-nine airframes we have on station, we have a shortage of pilots. We have no pilots classed as both medically fit and operationally experienced for combat. Most are instructors who have never fired a bullet or dropped a bomb in anger. Some are pilots who have been classed as 'unsuitable' for operational flying, others are woefully short of productive flying training or recent practice. Three, including Dudgeon, have been sent here for a rest following extended combat tours. We also have some pilots from the Royal Hellenic Air Force, whose command of the English language is limited, to say the least.
'All told, we have nineteen pilots for the twenty-one Audaxes and twenty pilots for Dudgeon's forty-three aircraft. So our true operating strength is thirty-nine aircraft. On the aircrew side the situation is worse. We have split the two experienced bomb aimers and four part-time air gunners equally between Ling and Dudgeon. So we have had to revert to asking pupils and ground-crew to volunteer to fill the vacant aircrew seats.'
Smart, growing more despondent, asked his Senior Administration Officer, Paul Holder, to bring the meeting up to speed on the supply situation.
'We have food and supplies for the RAF, the Levies and the King's Own for twelve days and four days for the 9,000 civilians in the cantonment. I plan to fly down to Shaiba and buy supplies to supplement our food stocks, so we should be able to keep up the level of our stores. We have three months' supply of 100-octane aviation-grade petrol stored in "A" and "B" dumps, so we will have adequate supplies through the short to mid-term.'
'Good, Paul. Now does anyone else have anything to add?' asked Smart, looking around the room. Two hands were raised, Everett and Ling. Smart gestured to Everett.
'Sir, this camp is vulnerable not only to Iraqi guns. We have no defences against their armour. The temperatures are rising in the day. If the water tower or the electric power station are put out of action our capability to fight stops. The camp is also situated on low ground and the Euphrates runs for about ten miles behind us at ten feet or so above the surrounding land protected by artificial embankments, bunds. If these are breached by Iraqi shellfire Habbaniya could be flooded out. Again this would curtail our capability to resist.'
'Your point, Colonel? a clearly annoyed Smart asked.
'We have no military option other than to take the fight to the enemy, and to do it first', said Everett.
Smart cut Everett off and gestured to Ling.
'The Iraqi Air Force is larger and better equipped than our own here at Habbaniya. They have over seventy operational aircraft, including Savoia Marchetti bombers and Northrop and Breda fighters that are all faster than anything we have. Even their version of the Audax, the Nisr, has a more powerful engine. We have also trained many of their pilots. While I agree that the clapped-out Gladiators are better than nothing, we need to take the fight to the Iraqis. We need to hit them on their airfields before they come here and take us out.'
Smart went apoplectic as the murmurs of general assent followed Everett's and Ling's remarks. 'What you two are advocating is for me to start my own private war against the Iraqis! We have a treaty with the Iraqis, and while things do look a little dicey I am sure that these problems things can be resolved diplomatically. In the absence of any directives from higher authority no immediate defensive or offensive action will be taken in spite of the risk of not taking such a course involves. I have already contacted Basra, Cairo and New Delhi, so we will just have to sit tight until we hear from them. In the meantime we need to get cracking with our defences. Bill, I would like you to keep an Oxford up, performing continuous reconnaissance, so that we are up to date with developments. You also need to get started on trench building. Thank you all. I will advise you as and when the situation develops.
Smart went back to his office, and after shutting the door sat pensively at his desk, staring out at the plateau, where he could see, in the distance, figures moving about. He did not want to make a decision to attack the Iraqis without provocation, and even if he did there was no guarantee of success. However, if the Iraqis struck first, Habbaniya could be lost. Either way his decision would bring death and destruction to the people and equipment under his charge. Habbaniya was completely vulnerable to attack and there were 9,000 civilians in the cantonment at the mercy of the Iraqis. The longer he waited, the more vulnerable the camp became as the initiative shifted in favour of the Iraqis.
Alex Thompson, the BOAC Manager at Lake Habbaniya, was not expecting a call before eight o'clock in the morning, and just managed to pick up the telephone at the third ring.
'Hello Alex, it's Paul Holder at the camp.'
'I hope that you got some personal mail in the sacks I sent up on Wednesday', Thompson said.
'Alex, we don't have much time,' Paul replied, cutting through the pleasantries, 'and I want you to listen very carefully as time is of the essence. The Iraqis have turned funny and they have invested the plateau looking down on the station. I flew over earlier and saw the two Empire boats on the lake. How quickly can you get them off?'
'We fuel at dawn in the cool to minimize fuel evaporation. The passengers and crew are just sitting down for breakfast. By the time they have finished it will be a couple of hours, so we will meet our 9.30 takeoff time. One 'boat is on its way to Bahrain and the other is going to Tiberius', answered Thompson.
'Alex, forget breakfast. Have them load up now, and get Doris and the children off as well as anyone else who is non-essential. There is nothing we can do to help you out. Your best bet is to get out now, while you can, as the Iraqis will be coming around to see you later today. Best of luck', said Paul, breaking the connection.
Thompson, still holding the phone, dialled his bungalow. 'Doris. Listen to me. Pack two suitcases for you and the children and get down to the jetty in half an hour. The Iraqis are being very difficult and the RAF suggests women and children leave as a precautionary measure. There is no real danger, it's more a move to avoid provoking the Iraqis. I will tell Stan what is happening, and you tell Maggie to pack right away.'
He walked out of the BOAC Manager's office and gazed out at the two big silver Short S23C Empire-class flying-boats gently bobbing at their moorings, looking like two contented whales. BOAC, and its predecessor Imperial Airways, had not seen fit to invest in pontoons at Habbaniya, so everything had to be ferried out by launch. Luckily the time-consuming and messy process of refuelling was completed, and all that remained was to load the crew, passengers and their baggage and the aircraft could be on their way. A quick word to his radio operator and deputy, Stan Freeman, and he would walk over to the Airways Guest House and break the bad news to the two captains, and give them the job of telling the passengers.
Thompson knew that he should not enjoy the passengers' discomfort, but wealth and privilege often seemed to bring out the worst in people. He could just imagine the scene as the two captains had to face the angry reactions of the passengers. That insufferable countess, the arrogant politician who owed his junior cabinet position to the fact that his betters had gone off to fight, the pushy self-made millionaire with those awful teeth who had found that money was not enough and wanted power and public recognition, and the fading actor given a new lease of life entertaining the troops, who was deeply in love with himself and wanted everyone to share his fascination.
Just after nine in the morning, Smart sent off another signal to Cornwallis, Fraser and Tedder, adding the Air Ministry in London. This time Smart reported on the steady increase in Iraqi strength and that heavy artillery was now trained on the camp. Smart added that air action might have to be taken that day if the Iraqi forces did not withdraw, as the risk of night attack would be unacceptable. Again, Smart requested assistance to be sent by air that day. A few minutes later Cornwallis replied, endorsing Smart's action in reply to the Iraqi message.
The very air seemed to sweat. As the sun rose higher in the clear sky the temperature began to move past 104ºF, and perspiration began to roll down the bodies of the airmen digging trenches all over the camp.
When the temperature rose above blood heat, prolonged human activity became difficult, and to avoid heat stroke the men were forced to wear their shirts and topees and to stop frequently to drink water to avoid dehydration.
The airmen were not used to being out in the sun. At Habbaniya flying began at dawn and ended at eleven in the morning. After midday, with the mercury rising, you either stayed in the shade or indoors until nightfall, when the temperature fell sharply.
Major Najib returned to RAF Habbaniya half an hour before noon with a second message, and again demanded to see Smart. This time the pilot officer asked Najib if he would care to wait in the guardhouse or in his car, as it would take a little time to contact the Air Vice-Marshal.
Najib elected to sit in his car for twenty minutes. By the time that Najib reached Smart's office, his once immaculate uniform was drenched in sweat and he looked less the martinet and more the drowned rat.
Smart looked at Said's second message:
The British have broken the Anglo-Iraqi Treaty. The Iraqi Commander will not allow any training or anything else to be carried out as long as the Treaty is not respected.
This placed Smart in a precarious position, but he decided to stall for time, sending Said an ambiguous reply:
You have raised a political question which neither you nor I can answer. I suggest that you and I refer this to our respective Ministers to come to an agreement.
In the meantime I would recommend that you remove your forces from the plateau to avoid the occurrence of any accidental problems.
Smart immediately sent the gist of this second Iraqi demand and his reply by signal to Cornwallis, Fraser and Tedder, as well as the Air Ministry in London. Smart also stated that he would continue to avoid offensive action until the Iraqis opened fire. He asked for an immediate directive as well as the status on reinforcements.
Ling and Dudgeon left Savile's office in a state of shock. While Savile had agreed to Ling's suggestion of Savile setting up his 'operations room' in the back of Hangar Six, the furthest away from the plateau, and to Dudgeon's recommendation to take up an Oxford and photograph the whole plateau and the surrounding areas to create photo-mosaic maps on which to pinpoint Iraqi positions for targeting, he would not countenance relieving the ground-crews from trench digging.
'Sir, if the Iraqis start something, the trenches and our few machine-guns, with the best will in the world, are not going to stop their tanks. However, a determined attack from the air with everything that we could put up might hold them off until such time as we can be reinforced. The problem is that all the essential people – the armourers, the electricians, the mechanics and the riggers – have been spread all over the camp. We don't know where they are so we cannot find them in a hurry if the balloon goes up.'
'I understand your position but the AOC has given explicit orders', replied Savile.
'Sir, you cannot expect the few pilots that we have to fuel and service the aircraft and find and load the bombs and ammunition, get briefed and start the aircraft all on their own. Just as an example, the Audax, Gladiator and the Gordon all need a second person to hand-crank to get the engine going!'
'Dudgeon, until the AOC countermands the order it stands', Savile replied frostily.
Later in the officers' mess Ling turned to Dudgeon. 'We cannot get caught again disobeying orders, so we have to find away around this absurdity.'
'We will have to get the pilots to walk around the trenches now and tell the individual ground-crews their assembly points – behind the hangers or on the polo field – when we get the call to go into action. While this is not perfect, at least it will give them a rallying point and stop them running around like headless chickens', suggested Dudgeon.
'You're right. We have to avoid a repetition of Smart's General Alarm this morning. No one knew what was up, where to go or what to do, and it took hours to get some sort of order in place. The last thing we need when we have to go into action is confusion', agreed Ling.
Smart looked unhappily at the copy of a fence-sitting signal he had just received from Cornwallis to the Foreign Office in London. Cornwallis agreed that Smart should use force to restore the situation, indicating that the Iraqi threats were acts of war justifying immediate air action. While he awaited a Foreign Office directive to force the issue, Cornwallis was trying to get the Iraqi government to withdraw their forces
Smart, watching the continual build-up of Iraqi forces on the plateau, sent a signal in frustration to Cornwallis. Smart wanted to know if Cornwallis agreed with his replies to the Iraqi commander and whether Cornwallis would have preferred that Smart had sent an ultimatum and, if necessary, an air attack against the steadily increasing forces threatening Habbaniya.
A strict system of rationing was introduced that morning to preserve food supplies. It became increasingly common to see both RAF and Army officers sit down in the mess to a table complete with a peacetime setting of silver and crockery and staff, but precious little food.
During the late afternoon Alistair Thompson and five other BOAC employees were rounded up by the Iraqi police and taken to a civilian jail in Baghdad, and remained there for the duration.
That evening Smart received a signal from AM Tedder in Cairo. Ten Vickers Wellington bombers would be sent to Basra immediately, and another ten would be on stand-by in Palestine. If the Ambassador agreed, they could be used immediately should the Iraqis open fire on Habbaniya.
To Smart these 'reinforcements' were virtually useless. To get the Wellingtons, from either Shaiba three hundred miles south or Aqir in Palestine five hundred and fifty miles to the west, bombed-up and fuelled and over Habbaniya would take at least five hours, by which time the Iraqis could have completely overrun the camp.
Habbaniya, 1 May 1941
Smart had taken to sleeping in a camp bed in his office – partly to be available for any eventuality and partly to avoid waking his wife.
A flurry of messages arrived for him after midnight, which all gave moral support but were short on both directive and reinforcements. Both Cornwallis and Auchinleck supported immediate action. Fraser in Basra was unable to support Habbaniya due to the extensive flooding, and suggested that he seek help from Middle East Command in Cairo. The breakfast-time signal from the Foreign Office in London received by Cornwallis was also copied to Smart. It stated that the position must be restored and Iraqi troops withdrawn without delay, giving Cornwallis, and by implication Smart, full authority for taking any steps necessary to ensure this, including air attack. For Smart it was no longer if but when.
Smart contemplated issuing an immediate ultimatum to the Iraqi commander to withdraw his troops. However, he knew that the commander on the spot would have to consult Baghdad before making a move. Reasonably this would take a minimum of three hours, bringing them up to noon. If there was no withdrawal, Smart would then have to attack.
However, a noon attack would leave only half a day for air operations before nightfall. Smart would need the maximum amount of daylight to mount a heavy first strike, which would hopefully dissuade the Iraqis from a night attack. Smart signalled Cornwallis to issue an ultimatum to the Iraqi government at 05.00 hours on 2 May, which would run to 08.00, giving Habbaniya the whole day from 08.00 to attack if the ultimatum was unsuccessful.
Smart was incensed when he saw his copy of Cornwallis's lunchtime signal to the Foreign Office in which the Ambassador stated that he had asked Smart to attack that very day. Smart quickly fired back that he would not attack that day given the few remaining hours of daylight, and that he would attack the next morning. However, Smart also changed the delivery of the ultimatum. It would now be given to the Iraqi commander at 05.45, with thirty minutes' grace to observe movement, otherwise air action would be taken.
Tedder's advice, which followed at the end of the afternoon, added nothing, and if followed could have had a disastrous effect on the situation. Tedder gave Smart the choice of telling the Iraqis either that he was continuing his flying training programme or that he was giving them an ultimatum to remove their forces surrounding Habbaniya. If they opened fire or refused to accept the ultimatum he should attack the Iraqi government offices and their Raschid barracks. Either course would have left Habbaniya wide open for an unstoppable ground assault on the camp
Smart knew that the time had come for him make a decision. He could no longer wait for reinforcements that might or might not come, nor could he wait for the Iraqis to make the first strike.
Attacking well-dug-in armour, artillery and troops exclusively from the air had never been achieved in warfare. If he failed he would personally carry the blame both for wrecking any chance to negotiate their way out of the problem and for the civilian and military casualties. If he succeeded he could still face censure for making an attack on a friendly nation where a peaceful solution could have been negotiated.
Smart now knew what the loneliness of command really meant. The chips were down. He was out on a limb with no directives and no help. He called Savile and asked him to assemble the Air Striking Force commanders and the ground force commanders for a meeting in his conference room at 20.00.
Arriving at the designated time, Smart found Savile waiting to introduce him to Col Ouvry Roberts, chief of staff of 10th Indian Division. Roberts had flown in that morning from Shaiba on a 31 Squadron DC-2 to take a look at the situation. As Roberts was the senior Army officer, Savile had requested that Roberts stay and take over command of the land forces at Habbaniya.
Smart, Savile and Roberts moved over to where Dudgeon had laid a photo-mosaic map across the conference table, where the other participants were crowding around. They all looked up as Smart walked over. He turned to Dudgeon and said, in a subdued voice, 'Why don't you bring us up to date with the Iraqi investment?'
'The photos, taken just before dusk, show that the Iraqis now have at least twenty-eight field guns and howitzers and up to 9,000 troops dug in on the escarpment. Vehicles are dispersed here, and here, and armoured cars there, there and near our main gate,. Iraqi troops have also occupied Humfriya and Sin el Dhibban, the little villages on the outskirts of the cantonment. The Iraqis have also ferried two howitzers and some machine-guns across the Euphrates so that they are now in a position to shell us from both sides. Troops and supplies are still coming down the Falluja road to reinforce their positions. I also took a shufti at Ramadi. I estimate that the Iraqis are there in brigade strength. This is probably insurance in case someone comes to our aid from Palestine', Dudgeon said, looking up from the map table.
Turning to Roberts, Brawn, Everett and Page, Smart asked whether they had finished their preparations. 'Yes, speaking on behalf of the ground forces', said Roberts. 'We are all in place and have already initiated fighting patrols along the perimeter fence this evening to deter the Iraqis from mounting night attacks.'
'Bill, how are the trench defences coming along?'
'They were completed late this afternoon', he replied. 'We also found a half-a-dozen more Lewis guns with anti-aircraft pedestal mountings tucked away in the stores. We have set these up, together with the other 34 Lewis guns, in sandbagged emplacements around the vulnerable targets to give us some protection from attacks by the Iraqi Air Force.'
'Gentlemen,' began Smart, 'I think that we no longer have the option to wait, and if the Iraqis are still up on the plateau I am ordering an air attack at first light tomorrow morning.'
One could almost sense the feeling of relief which swept through the conference room.
'I want you to get every aircraft we can get into the air before dawn so that we can commence bombing and strafing the Iraqis as soon as you can distinguish ground targets', Smart went on. 'Our strategy is concentrated, continuous bombing without warning to demoralize the Iraqis. We should have them in full flight within about three hours. Air Marshall Tedder has sent ten Wellingtons of 37 and 70 Squadrons from Cairo to Shaiba to help out. They carry a useful bomb-load of four and a half tons each, so that should augment our striking power.
'Just before the meeting I received a signal from the Prime Minister which I think you all should hear, as it sums up what we are about to do.
If you have to strike, strike hard. Use all necessary force.
'I think that you have all a lot of things to do and you face a busy night ahead. But before I let you go are there any things that we need to finalize?'
'Yes, sir', said Ling, raising his hand. 'There are a number of practical issues that we need to resolve. For example, the lack of facilities at the polo ground for rapid resupply of fuel, ammunition and bombs. We also need to release ground-crew from the trenches so that we can get ready for our pre-dawn take-off.'
'Wing Commander, we cannot let any one leave the trenches: they are our only form of ground defence, I repeat our only defence, so they must stay! Gentlemen, do the best you can', Smart said as he concluded the meeting.
Smart went back to his office to draft a signal to Cornwallis, Tedder, Fraser and the Air Ministry in London, explaining the deterioration in the situation, which now made it necessary to attack the next morning without warning
Leaving the meeting, Larry motioned to Dudgeon, Hawtrey, Holder and Silyn-Roberts, as well as the two flight lieutenants, Cleaver and Evans, to follow him into an empty office on the ground floor of the administration block.
'You all know where your ground-crews are, so get them out of their trenches and round up your pilots. You need to name your aircrews and allocate their tasks for tomorrow morning. We all need to be fuelled and bombed-up ready to go by 04.30 latest to get us over the plateau to start action at 05.00. See you at your aircraft at 03.00 hours.'
'Just one thing,' Ling continued as they moved towards the door, 'for God's sake keep this quiet from Headquarters and from Savile. What they don't know about, they can't stop!'
None of the 'Habbaniya Striking Force' commanders got to bed much before midnight, and the last pilot was back at his aircraft by 03.15 the next morning.
Chapter Three
Strike Hard, First
Habbaniya and Baghdad, 2 May 1941
As the first molten rays of the coming dawn spread like a halo around the plateau, all thirty-nine aircraft started their engines. They moved into position on both the main airfield and on the converted polo ground for take-off, guided by airmen with hooded torches.
In a combat situation no flare-path or navigation lights could be used, and it had been reluctantly decided that those pupil pilots with no night-flying experience would be excluded.
Dudgeon in an Oxford led the rest of the aircraft allocated to the main airfield out of the gate. He pointed the nose at the black mass of the plateau framed against the lightening sky in front of him and opened up the throttles wide, remembering to pull back the stick as soon as he possibly could to miss the casuarina and pepper trees at the far end of the runway.
Dudgeon's aircraft led the Oxfords and the Gordons, followed by the Audaxes, into a climbing turn to form a huge circle which wheeled over Habbaniya.
The air above the plateau seemed to be filled with aircraft as the ten Wellingtons from RAF Shaiba joined up with the Habbaniya fleet. Now aircraft of five different types, sizes and speeds were all jockeying for space in an area not much bigger than a medium-size golf course.
At 05.00 hours, when it became possible to distinguish objects on the ground, Dudgeon led the attack on the plateau, dropping the first bomb-load at 1,000 feet from his 'training' aircraft.
As Dudgeon banked away from the plateau he saw the rest of the Oxfords painted in 'trainer' yellow follow in at 1,000 feet, the camouflaged Wellingtons start to bomb from 5,000 feet, while the silver Audaxes and Gordons began to dive-bomb the Iraqi positions. Patrolling at 10,000 feet like protective eagles were the silver-painted Gladiators.
Quickly overcoming their initial surprise, the Iraqis began to retaliate immediately. They opened up with anti-aircraft fire from machine-guns and 20 mm cannon, while their batteries of 18-pounder field guns and 4.5-inch howitzers began to shell RAF Habbaniya.
One of the first Iraqi shells fell on the square, killing Privates Adshead and Rooney of B Company of the King's Own, and badly wounding several others, while other salvoes fell all around D Company, with near-misses throwing up fountains of earth covering the troops in their slit-trenches.
Baghdad
Freya Stark arrived at the British Embassy just after seven in the morning in style, sitting in a horse-drawn ghary, holding a parasol and wearing one of her eccentric hats. She was a small woman who wore her hair looped over one ear to hide a deformity caused by a childhood accident. Her dazzling career as an explorer, writer, photographer and Arabist had been achieved through a combination of audacity and charisma. However, her driven personality and extravagance could exasperate both her friends and her detractors.
Her command of Arabic and her personality helped her to successfully create the anti-fascist Brotherhood of Freedom in Aden and Egypt for the Ministry of Information, and this had led to her secondment to Baghdad to help counter Axis propaganda.
Freya was on a train on the one-metre-gauge line from the Persian border at Khanikin to Baghdad when the RAF bombs began to fall on the plateau.
It had taken some time to get transport from Baghdad North station to the Embassy.
The early hour, the general surliness of the people at the station and the quiet streets with their windows tightly shuttered in a normally vibrant city created a sense of unease in Freya.
Past the Munahiya and Haidar-Khana Mosques, along al Raschid Street lined with the shops of merchants selling brocade, velvet and silks, past Armenian dentists, Muslim photographers, Jewish jewellers, Christian booksellers, past the Bank of Cairo, the three-storey Orozdibak department store and the Lowel Brothers' Buick and Chrysler agency and over the King Ghazi Bridge, the story was the same: a few, sullen people and a large number of troops on the streets.
Arriving at the Embassy, Freya was confronted by a police cordon and had to wait outside the small wooden postern gate until Leslie Potts, the Consul, identified her before the police let her and her two suitcases through the gate.
The sight inside the Embassy confirmed Freya's worst fears: they were under siege. Three large incinerators in the Chancery courtyard were being fed by staff, while others carried out boxes full of secret files and archives and most of the cyphers to feed the fires. Two Bedford light trucks were parked on the drive and a few cars were on the lawn, petrol tins full of sand were strategically placed to put out fires and a large 'V' of white sheets was spread out on the lawn to warn RAF planes of their position.
The faces of the people Freya passed were pale from sleepless nights, and bore out the general air of uncertainty and even fear. The British looked hot and pensive, the Middle Easterners who had taken refuge in the Embassy looked gloomy, the Indians deeply unsettled, and the Iraqis who had stayed on looked resigned to their fate.
Freya walked into the large hall at the main entrance, where the blue-tiled fountain still played, up the circular staircase and along the passage to the Ambassador's office. Freya knocked on the open door, and Cornwallis, sitting at his desk, looked up and smiled, waving Freya to come in. 'Freya! What a pleasant surprise. I thought that you might have heard and would keep a discreet distance until all this unpleasantness has been resolved', he said, walking around the desk to embrace her.
'How are things, Ken?', asked Freya.
'Not good. Rashid Ali's blind ambition has trapped him between the Mufti, the Golden Square and the Axis on one side and us on the other – and he had no way out. The RAF started bombing at five this morning and I am worried that the Iraqis will take revenge on the people here at the Embassy', he said as he walked out onto the balcony overlooking the courtyard and the lawn.
'The RAF buses and trucks which took the women and children out to Habbaniya on the 29th also brought in some Lee Enfield rifles, sandbags and barbed wire. George Waterhouse is having the time of his life drilling men and siting the sandbags and barbed wire', he said, pointing to activity in front of the main gate; 'I just hope to God that it does not come to that.'
'We also have to contend with overcrowding. We now have three hundred and sixty-six men as well as nineteen women – Armenians, Indians, Jews, Poles and Yugoslavs – with no papers which disqualified them from the evacuation. Unfortunately we have toilet facilities for only ten, so Pat Domvile has got some men over there by the west wall,' he gestured, 'digging latrines and wastewater sumps. So far the telephones still work and the electricity and water are running, and we do have enough food stores for about a week. At the moment there is little more we can do.'
Freya tried to look out at the city beyond the walls of the Embassy, but the pall of smoke now hanging over the garden, and the black ash swirling around from the bonfire, which made her eyes smart, seemed to be imprisoning her.
Habbaniya
The RAF aircraft attacking the plateau soon established a routine to maintain their strategy of continuous bombing. After dropping their bomb-loads they would return to their landing grounds to rearm and refuel and then take off again for another bombing run.
Shells from the Iraqi guns were falling all over the camp, making takeoffs and landings a hazardous operation. The Audaxes based on the polo field were fortunate as they were screened by a row of trees, which hid not only the parked aircraft but much of the landing and take-off run from the Iraqi gunners. However, the Gladiators, Gordons and Oxfords based on the main airfield were fully exposed to the guns on the plateau.
Dudgeon's flights devised both take-off and landing patterns to reduce their vulnerability to ground fire. Engines were started in the lee of hangars, and after an airman based at the corner of a hangar gave the all-clear if there were no aircraft landing, the pilot would juggle with the throttles and brakes as the aircraft swept out from behind the hangar and opened-up the engine. As soon as the pilot had achieved flying speed and sufficient height he would drop a wing and pull the aircraft to starboard away from the plateau, and then climb to his bombing height before turning and heading back towards the plateau for his next run over the target.
On landing, the aircraft would swing away over the Euphrates on the far side of the camp and then fly in very low, getting the best cover from buildings and trees, and then turn in between the hangars, landing on the taxi way by the airfield fence before swinging through the gate and back between the hangars.
As soon as the aircraft parked, one of the crew would report to Savile in the operations centre on the result of the mission. Savile would mark this on the photo-mosaic map and then assign a new target for the next mission. While this was going on, the other crew member would help the ground-crew to rearm and refuel the aircraft and check for battle damage.
David Evans, commanding the Gordon flight, had picked up several wounds in the first day but kept flying. He experimented and came up with a way of destroying pin-point targets on the plateau. The 250 lb bombs were fitted with safety devices to ensure that the bombs fell at least two hundred feet before exploding, to avoid blowing up the plane from below. Before take-off Evans removed the safety device and had the bombs fitted with fuses to give a 7-second delay between impact and explosion. Climbing to 3,000 feet, he would pick out a target on the plateau and go into a near-vertical dive at a lunatic speed for a Gordon of 200 mph. Pulling out, Evans would lay his bomb at between eight and ten feet above the target, and with his speed and the 7-second delay would get just far enough away before the bomb exploded. The Iraqis soon got the point, and refused to come out and fire at Evans when he started his dive-bombing attacks.
Baghdad
Rashid Ali and Haj Amin sat opposite Luigi Gabrielli, the head of the Italian Legation, in the library in the Mufti's house in Zahawi Street.
'Minister,' Rashid Ali began, 'as you know, we have been attacked by the British without warning early this morning and are fighting back. We have breached the dams on the lower Euphrates and Tigris and have destroyed railway tracks and the telegraph to the north of Basra, preventing the British from moving to support their airfield at Habbaniya. I would like you to convey to Rome and Berlin that we are counting on military aid, and we are relying on your promises as we have been forced to make a stand on Britain's unacceptable demands.
'Mr Gabrielli,' repeating the Minister's name to reflect the gravity of his message, 'we want a prompt reply and emphasize that we need immediate shipments of aircraft, and arms, a military mission and financial aid. If we do not receive this assistance without delay our only hope is to gain time through negotiation. However, I see little chance that this would work, as the British seem bent on occupying Iraq.'
'Minister,' the Mufti added in a quiet, measured tone, 'as soon as we receive aid from your governments I will direct my supporters to begin subversion in Palestine and Transjordan which will help to pin down the British.'
Gabrielli promised to convey the Iraqi message urgently to Rome and Berlin, and left hurriedly for his Legation.
Haj Amin turned to Rashid Ali. 'al Sabbagh made a serious miscalculation by expecting the British to negotiate and not fight. Everything now depends upon Iraq's ability to hold on – the longer that we resist the greater the chance of serious German and Italian support. We need to stall for time. Contact the Turkish Embassy and ask for an urgent meeting with their Ambassador. If we can get the Turks to mediate with the British on our behalf it may weaken British determination and delay further armed clashes.'
Rashid Ali bitterly resented being ordered around by the Mufti. However, Haj Amin's instruction made sense, and this was not the time or the place to risk a confrontation with the Mufti.
Habbaniya
By mid-morning the attackers were experiencing intense, accurate antiaircraft fire from the plateau.
With the Oxfords cruising at 1,000 feet and the Audaxes and Gordons sweeping even lower as they dive-bombed, they were hard targets to miss, even for the most inexperienced Iraqi gunner. Flg Off Walsh's Oxford was shot down in flames and Plt Off Gillespie's Audax failed to return.
For most of the pilots the tracers floating up gently from below seemed at first to be just pretty lights. When they saw bullet holes being punched through their fabric-covered wings, and bullet strikes through their cockpits and windscreens, the realization set in that this was a fight to the death, and that their flimsy machines with a lack of protective armour plate were virtual death traps.
After a sortie each aircraft was checked for additional battle damage. If it was superficial – it seemed to have passed through the aircraft without hitting anything vital like a fuel line, a spar or control wires – it was patched up with the engine running.
However, within a few hours every aircraft was damaged. The record was held by Flt Lt Dan Cremin, who brought his Audax back after one ten-minute sortie with fifty-two new bullet holes.
As the day grew longer, the human casualties began to grow. Aircrew with minor wounds which were not incapacitating were returned to the attack as soon as the planes could be rearmed. Those with flesh wounds were bundled off to the hospital, often to return in a few hours or the next day, stitched, bandaged and plastered for another sortie.
The example of Wad Taylor, a part-time air gunner, was typical. Wad, because he always took a large sandwich aloft, was in a Gordon when a bullet came up through the floor, went through his eyebrow, scarred his forehead, entered his leather helmet and went out through the top. He covered his blood-filled eye with one hand and used the other hand to fire his Lewis gun. He was rushed to hospital when the Gordon landed, and an hour and a half later, stitched and wrapped in bandages, he took his place in another Gordon and went back into action.
However, the number of seriously injured aircrew began to mount. Flt Lt Jimmy Page, an Audax pilot, was shot through the jaw but managed to land safely on the polo field despite being in great pain and barely able to see. Another Audax pilot hit by three bullets through the lung and shoulder collapsed over his stick, causing his aircraft to go into a steep dive. His pupil gunner, also wounded, was able to pull the pilot back and hold him there until he recovered enough to pull the aircraft out of the drive and land it semi-consciously on the polo field with his one good arm. A Gordon pilot shot in the thigh by a burst of ground fire was able to land. The same burst had hit his oil cooler, and as the oil pressure fell the plane limped to a halt just outside the airfield fence as the engine seized. The pilot and his gunner only just made the safety of the hangar before Iraqi gunfire bracketed the Gordon and then scored a direct hit.
Cairo
GHQ Middle East Command was located in a sprawling warren of offices in a modern five-storey Art Nouveau apartment block dubbed 'Grey Pillars' in Tonbalat Street. This was located at the southern end of Garden City, one of the better residential suburbs of Cairo.
Lt-Gen Henry 'Jumbo' Wilson, an enormous bald man, surprisingly light on his feet for such bulk, alighted from the Humber staff car as it drew to a halt outside GHQ, After showing his pass to the Military Police sergeant at the guardhouse, he quickly passed through the barbed-wire fence.
Wavell, deep in thought, did not hear the first knock of his ADC on the main door to his office. The second, louder and more insistent knock, penetrated Wavell's thoughts and he looked up from his desk.
'Sir, General Wilson is here to see you.'
'Thank you', replied Wavell to his senior ADC, Peter Coates. 'Give me a few minutes to clear my thoughts and then send him in.'
Wavell's responsibility was vast. Geographically, Middle East Command stretched from the Tunisian border with Libya in the west, to Transjordan's border with Iraq in the east and from Greece in the north to Madagascar in the south. No commander in the history of warfare had to deal with an area covering three million square miles, over one million people or with so many countries, cultures, religions, climates, or allies or enemies.
In early May he faced an array of problems. ULTRA intercepts indicated that an airborne assault by Generalleutnant Kurt Student on Crete was imminent. Rommel had Tobruk under siege, and advance elements of the Afrika Korps were on the Egyptian border at Sollum. The sea lanes across the Mediterranean were in danger of being cut, and Malta was under continuous aerial bombardment. In East Africa his forces engaged in a major battle at Amba Alagi in Eritrea were outnumbered five to one, and intelligence reports indicated that Axis agents were at work throughout the Middle East fomenting unrest.
Worse, Wavell was at war with Churchill. Churchill's distrust of generals was matched by Wavell's doubts about politicians.
Wilson strode into the office and sat down when Wavell indicated a chair. Jumbo was Wavell's right-hand man, 'a man who got things done' – a situation that worked well probably because they were so different.
Wavell, a muscular, bull-necked man with an eye blinded in the First World War, was possibly the best-educated soldier of his generation. He appeared distant, detached and even remote. Celebrated for his unexpected silences and his prolific memory, his reputation was built on hard work and integrity. Wavell attracted loyalty and affection as few others, not least for his personal bravery.
In contrast, Jumbo was a solid and dependable but aggressive man, some said a bully, who was not universally liked by either peers or subordinates. He resembled a benevolent uncle, but looks were deceptive, as few experienced the benevolence he appeared to exude.
'Jumbo,' Wavell began, 'I want you to put together a plan with George Clark in Jerusalem to relieve Habbaniya.'
Discussions with Wavell always contained long silences and Wilson waited patiently for Wavell to continue.
'As you know, I have fought this, but it is now a directive from the Chiefs of Staff, so we have to get on with it', Wavell finally added.
'I know the position, sir, but we face enemy pressures throughout the Middle East and we have serious manpower and equipment shortages. We lost half of our equipment when we had to evacuate Greece, and some of our best troops, the Australians and New Zealanders, are tied up in Crete waiting for Student to attack. Where do I get the resources?' asked Wilson.
Wavell responded enigmatically with, 'That is your affair', and rose to indicate the end of the meeting.
Habbaniya
By noon three of the four flight commanders flying Audaxes were out of the fight. Ling was wounded in the air by ground fire before lunch and out of commission for the rest of the campaign, and Silyn-Roberts was rushed to the hospital for an emergency appendectomy.
Paul Holder's Audax was hit, wounding his air gunner Aircraftman Wad Taylor, and his engine stopped. Holder was able to glide in over the perimeter fence and land. As he and Taylor ran for the nearest trench an Iraqi shell blew up the Audax
All ten Wellingtons from 37 and 70 Squadrons were also out of the fight. The big Wellingtons were nice juicy targets, and on return to Shaiba nine were found to be unserviceable due to battle damage The tenth Wellington, piloted by Flg Off Anstey, made a forced landing at Habbaniya after sustaining serious flak damage to its engines before it had even dropped its bombs.
The large aircraft sitting on the main airfield acted as a magnate for the Iraqi guns, and the aircrew made for the hangars with machine-gun bullets kicking up the sand around their feet as they sprinted to safety.
The Habbaniya ground-crews, although lacking any Wellington equipment, tried desperately to save the aircraft, and a tractor was driven out boxed in by RAF armoured cars on all sides to screen the unprotected driver from the Iraqi bullets. Before the tailwheel of the Wellington could be attached to the tractor the Iraqi howitzers had bracketed the aircraft and scored a direct hit, which also put the tractor out of action.
The tractor driver, Aircraftman Kevin O'Reilly, was unhurt but left wandering around in the open, dazed by the exploding shells. One of the armoured-car commanders, Flt Lt Pyne, saw O'Reilly and had the driver, LAC Tait, drive back, and Pyne and LAC McLennan hauled him into the vehicle in the face of intensive fire before heading back to the comparative safety of the lee of the hangars. Halfway across the airfield, Pyne's armoured car was almost blown over by a huge blast wave as the full 4,500 lb bomb-load and fuel in the burning Wellington detonated.
Just before noon the Iraqi Air Force began a series of attacks on Habbaniya. The Bredas, Gladiators, Nisrs, Northrops and Savoias of the numerically stronger and more modern Iraqi Air Force made a number of sorties, bombing and strafing the camp. As most of these attacks were made with low-level, high-speed runs, they were fairly inaccurate, although the RAF did lose three aircraft on the ground.
While Habbaniya did mount combat air patrols with the Gladiators to counter Iraqi air raids, they had no radar or radio control. This, together with operating worn-out aircraft which could not match the speed of the enemy aircraft, allowed the Iraqis to bomb and strafe the camp with impunity. Probably the only value in the Gladiator standing patrols were that they deterred the Iraqi Air Force from pressing home more determined attacks.
The oxygen-equipped Savoias flew over at high level, well above the operating ceiling of the RAF Gladiators. Dicky Cleaver saw one at around 20,000 feet, and reckoned that he could reach it before he lost consciousness from oxygen starvation or the Mercury engine's carburettor iced up. The Gladiators in the Middle East were usually flown with the canopy slid back to give a better view and fresh air. With no heated suit, Dicky wound the sliding canopy closed and used full throttle to provide the boost to get a full-power ascent. It took almost ten minutes for Dicky to climb to 20,000 feet and get behind the Savoia. Perfectly placed, Dicky took careful aim and pressed the gun button on the top of the control column. However, he did not reckon on the cold at that altitude or the lack of anti-freeze oil at Habbaniya, and all four Browning .303 machine-guns iced up. Helplessly Dicky watched the Savoia fly safely away.
Getting food to the defenders was a difficult and even hazardous job. With the King's Own and the Levies dispersed around the camp, a large part of the RAF personnel manning slit-trenches and the aircrews in continuous combat, the only feasible way to feed the defenders was for the mess staff to bring sandwiches and dixies of tea to them. Many of those shuttling food, tea and water to the troops, airmen and aircrews were locally recruited Iraqi cooks, kitchen staff and barmen who loyally remained on duty throughout the siege.
However, the officers' mess was still functioning, and the senior officers, none of whom had volunteered to fly or fight, were continuing to use the mess, oblivious to the battle ranging around them, much to the annoyance of their younger, fighting subordinates. When disturbed by Iraqi machine-gun bullets through the corrugated asbestos roof and splinters from exploding bombs through the walls, they made vigorous complaints in the Mess Suggestion Book.
To the delight of the defenders, six Iraqi Gladiators strafed the officers' mess, and a number of elderly and portly senior officers were seen to be hastily diving for cover under tables and chairs as the Iraqis roared a few feet over the building.
Early in the afternoon LAC Arthur Briggs, with a nose for business, opened a book and began taking bets on how long it would take for the water tower to catch a direct hit from an Iraqi shell. With his position in a slit-trench directly outside the Air Headquarters block, the large amount of traffic to and from the command centre provided him with the perfect cover to receive bets and pass betting slips across the camp.
Throughout the day the blockhouses along the perimeter fence, and especially those facing the plateau, had been taking sustained fire from the well-dug-in Iraqi positions. Although none of the blockhouses were put out of action, No. 8 was under continuous heavy fire and an Iraqi shell knocked in the front of No. 9.
However, the blockhouses kept up constant return fire on the Iraqi positions, and No. 10, with a field of fire across the Falluja–Ramadi road, shot up Iraqi traffic throughout the day. Four Iraqi officers in a staff car were believed wounded, and an Iraqi dispatch rider was shot off his motor-bike and killed.
Roberts from the outset believed that their best ground defence was in the attack. While Brawn and Everett had planned their defences by meticulously dividing the perimeters into sectors of responsibility, with the blockhouses being supported by communications trenches, Roberts believed that their key objective was to seek out and destroy the Iraqis.
Both the King's Own and the Levies began ground offensives against the Iraqis.
An Assyrian company, using its Boys anti-tank rifles, successfully repelled the only armoured foray towards the camp from the south by eight Crossley armoured cars supported by three FIAT light tanks from the Iraqi Mechanized Brigade.
The two Iraqi 3.7-inch howitzers across the Euphrates to the north began to pour accurate fire into the camp. A Levy platoon from the Kurdish company crossed the wide, fast-flowing river in an old motorboat supported by small-arms fire from the nearest blockhouse. They were forced to advance across open and partly flooded ground, but were able to inflict over thirty casualties among the Iraqi gun crews before heavy machine-gun fire and a lack of ammunition drove them back.
Later in the day Audax dive-bombing was used to silence the guns.
Under cover of night Roberts planned to use fighting patrols of Brawn's Levies to harass the Iraqi forward positions and to deter enemy patrols from penetrating the camp.
By the end of the day the growing list of casualties from Iraqi shell and sniper fire and bombing strafing raids had taught everyone that digging slit-trenches was more than a necessary fatigue.
Baghdad
As the day began to draw to a close, Adrian Holman found Cornwallis in conversation with Edmund Loftman, one of the leading lights of the British community in Baghdad, in the courtyard of the Embassy, and coughed discreetly to get his attention.
'Adrian, you need me?' said Cornwallis.
'Sorry to disturb you, sir, but I have a pressing problem.'
Cornwallis, sensing Adrian's discomfort in speaking in front of Loftman, excused himself and drew Adrian away. 'My thanks to you, Adrian, in getting me away from that bore. Now what can I do for you?'
'I have an Iraqi captain at the postern gate with about a dozen armed soldiers. He has cut off our telephones and wants us to hand over all our radio transmitters and wants to search the Chancery buildings. What do you want me to do?' asked Adrian.
'I will protest in the strongest possible terms about the violation of our diplomatic immunity, but in reality there is little we can do. I will use this opportunity to press for the restoration of our electicity, as its loss endangers our health. The septic tank for the Embassy's sewage depends entirely on an electically operated pump. While I am protesting and negotiating, Smy in the radio room will have enough time to send a final message to London and Habbaniya, and then destroy our transmitter before we turn it over. Also ask Smy to burn the remaining cipher books and have General Waterhouse present so that he can authenticate their destruction.'
'That means that we will be totally cut off from the outside world, doesn't it?' asked Holman as more of a statement than a question.
'Not quite', said Cornwallis. 'I will hide my Philips radio so that we can get some idea what is happening, and we do have a hidden reserve radio transmitter.'
At the American Legation Paul Knabenshue took stock. He had given shelter to 170 refugees – fifteen Americans, 140 Britons and fifteen people of other nationalities.
Col Page, a retired Indian Army officer, was drilling some of the men, armed with a motley collection of hunting-rifles, shotguns and pistols, while Betty Sulman, the young King's British governess, was organizing food and shelter. It came as a great shock to Betty to encounter lying, cheating and stealing which people who were not normally dishonest or greedy did to increase their rations.
Knabenshue had also lost his radios and telephones to the Iraqis, and was isolated like the British across the river.
Habbaniya
As dusk began to filter in from the east, five Wellingtons of 70 Squadron returned for a second attack on the plateau. Although the Iraqi Air Force intercepted the Wellingtons with two Northrops and two Gladiators and attacked for twenty minutes, there was no damage. However, ground fire did succeed in hitting one of the Wellingtons, which had force-landed at Habbaniya.
During the early evening the first issue of a daily sheet giving news of the progress of the war, with a note or two about events at Habbaniya, began to roll off the camp printing-press.
With the fall of darkness, combat flying was halted for the day.
Smart called for a briefing at 20.00 hours in his first-floor conference room in the Air Headquarters block.
The number attending this meeting was much smaller than the day before. Smart now presided over the Army officers, Roberts and Everett from the King's Own and Brawn from the Levies, Page from the Armoured Car Company and Savile, Hawtrey and Dudgeon from the Air Striking Force.
The results of just one day's sustained combat were not good.
The Air Striking Force had flown 193 recorded sorties, although there were probably more which went unrecorded in the heat of the engagement. Of the sixty-four aircraft, over one-third, twenty-two, had been shot down, strafed by the Iraqi Air Force, blown up by the Iraqi guns or damaged beyond repair. Many of the remaining aircraft were unfit for service under normal RAF standards, but fitters and riggers working with light from hooded torches behind the hangars and on the polo field were trying desperately to coax them back into the air for tomorrow's battle.
More seriously, ten, more than a quarter of the thirty-nine pilots were dead, missing or hospitalized, and most of the rest were wearing bandages and plasters.
The British casualties were thirteen dead and twenty-nine wounded, including four Levies and nine civilians.
Despite the continuous bombing and strafing campaign, the Iraqis were still firmly ensconced on the plateau, looking down on the camp with the undiminished capacity to shell Habbaniya into submission or to mount an unstoppable ground attack. Iraqi morale was higher than Smart had been expecting, and, more pointedly, showed no signs of weakening.
Smart looked visible shaken by the news, and began to blame Wavell in Cairo, Auchinleck in New Delhi and Bill Fraser in Basra for failing to come to his aid, and the politicians in general for getting Habbaniya into this mess in the first place. 'What else could I have done?' he said. 'We have attacked and it's not working. My station is being destroyed around me and my people are dying and no one is helping us.'
Smart called an end to the meeting, to the relief of the participants embarrassed by his outburst. 'Well,' he concluded as the participants were filing out of the door, 'tomorrow is another day and I will be talking to you about some better ideas after I have spent some time thinking about them.' No one commented on Smart's behaviour as they went their separate ways.
After the disastrous briefing, Dudgeon, Savile and Hawtrey met at the operations room at the back of Hangar Six to discuss tactics for the next day.
'In the face of our experience today and particularly our losses, I think that we should divide our resources differently', Dudgeon said. 'By switching the remaining Audaxes and the Gladiator flight to the polo field, where they are well screened from the Iraqi gunners, they can bomb-up, take-off and land back quicker and safer. Cutting back on my command to just the Oxfords and the Gordons means that I will be able to get them all behind the hangars out of direct sight of the gunners. The corrugated iron of the roofs and walls will also explode most of the incoming shells. This will give my crews their best measure of protection.'
Hawtrey confirmed his agreement, adding, 'It would simplify our tasks as we would each only have to crew two different types of aircraft.'
'We have a problem with the women and children still in the camp, as well as the wounded at the hospital. All require evacuation. Obviously we cannot get them through the Iraqi lines, so the only way out is by air. What about the Oxfords, Dudgeon?', asked Savile.
'The Oxfords can only carry a few at a time and have no passenger comforts', replied Dudgeon. 'However, 31 Squadron at Shaiba has some nice shiny new DC-2s which they commandeered from Indian National Airways and fitted out for troop transport. We could get them to shuttle the women, children and wounded out to Shaiba.'
'The DC-2s are big targets for the Iraqis and we could lose not only the planes but also the women and children', replied Savile.
'As long as we know when they are coming we could have the Audaxes dive-bomb and strafe the Iraqis to keep their heads down', said Hawtrey.
'I'll get in touch with Richard Burberry, the CO of 31 Squadron, and get them to arrive a few minutes after five, just as we start our attack on the plateau, which should keep the Iraqis tied up', agreed Savile.
Savile confirmed the operational plan for the second day and to a 04.00 hours briefing for the crews, with a take-off time scheduled for 05.00 hours the next morning.
After the meeting Hawtrey and Dudgeon toured the polo field and behind the hangars to assess the number of aircraft that would be available for combat the next day, and spoke individually to the fitters, riggers and armourers who were working in virtual darkness, to keep up their morale.
Smart signalled Tedder in Cairo, urgently requesting reinforcements due to the damage suffered by so many of Habbaniya's aircraft and pilots in the first day of combat.
Dudgeon got back to his room shortly before midnight, to be greeted by his long-suffering dachshund, Frankie, who had been frightened out of his wits by the non-stop shelling and bombing. Just as Frankie curled up in Dudgeon's arms, the Iraqis, free from instant retribution by the Air Striking Force, began a sustained bombardment which went on from midnight through to 03.00, pumping over two hundred shells into the camp.
Dudgeon and a shivering and whimpering Frankie crawled underneath the bed just before a near-miss blew in the windows, scattering jagged shards of broken glass, part of the window frame and bits of curtain all over the room and bed. After a few minutes, desperate for sleep, Dudgeon, complete with Frankie, brushed the debris off the bed and climbed in, with Frankie burrowing down under the sheets to curl up at his feet, apparently safe in the belief that his master would protect him.
Cairo
Smart's initial handling of the attack, his dispatch of somewhat hysterical messages, including his failure to press home ground attacks in concert with the bombing and strafing, suggested to Tedder that Smart's judgement might be affected. He began to consider replacing him.
Habbaniya, 3 May 1941
Crew briefing took place at 04.00, and Habbaniya's little air force took off at 05.00, joined by a single serviceable Wellington from Shaiba. The plan was to continue the relentless programme of bombing and machine-gunning the Iraqi positions on the plateau overlooking the camp. As the bombing began the Iraqis resumed their intense shelling of the camp.
When the first of 31 Squadron's DC-2s appeared overhead, still sporting their Indian National Airways livery, John Hawtrey's Audaxes began an intense dive-bombing and strafing attack, which enabled the DC-2 to slip unscathed into Habbaniya. Throughout the airlift only a few DC-2s were damaged, none seriously, and no women or children were wounded or injured.
Paul Holder flew down to Shaiba on the first Douglas flight out of Habbaniya to organize the purchase of meat, vegetables and other foodstuffs to fill the aircraft on their flights up to Habbaniya to restock the camp's food stores.
On the second day there were some tactical changes. Three Wellingtons operating out of Shaiba, together with some dive-bombing Audaxes and Gordons from Habbaniya, took the fight to the enemy. Several Iraqi fighters intercepted the RAF aircraft over Raschid air base, and gunners aboard a Wellington shot down one Nisr and damaged a second. Photographic evidence from Dudgeon's Oxford later showed that the surprise attack on Raschid air base, a little to the south of Baghdad, where the Wellingtons had unloaded 7,100 lb of bombs, damaged twenty-nine of the Iraqi Air Force's Savoia Marchetti bombers and Breda fighters on the ground.
Back at Habbaniya, SM79s and Northrops again attacked the camp. Dicky Cleaver made three passes firing 1,200 rounds, and a Savioa Marchetti left the area trailing black smoke, while Flt Lt May attacked a formation of Northrops and damaged one.
On the way back from the Raschid attack, Plt Off Michael Strange, flying an Audax, saw a lone Iraqi Nisr flying north-north-east away from Habbaniya. Following the same course, he came across a Northrop and a Savoia, both force-landed in the desert, signifying that Cleaver and May had made confirmed kills.
The Nisr disappeared, but could only have come from Baquba air base some seventy miles away on the same course. The Air Striking Force decided to pay Baquba a visit later that day.
A group of Audaxes supported by Gladiators found twenty-one parked Iraqi aircraft at Baquba. The dive-bombing Audaxes took ten on the ground and the Gladiators got three on the ground and one in the air.
While the Raschid attack was accomplished without loss, an Audax was shot down over Baquba with a bullet through its Kestrel engine. The pilot managed to glide away and landed safely on packed sand. Iraqi troops quickly arrived on the scene, none too happy at having their airfield soundly beaten up.
They stripped the two RAF aircrew, tied their hands behind their backs, badly beating them with rifle butts and staves as they made them run barefoot to the air base. Luckily the troops did not resort to the common Iraqi custom of slashing off their prisoners' genitals.
It had been decided to avoid bombing civilians in Iraq since no one wanted to inflict suffering on the population, with whom the British had no quarrel. However, Gordons began dropping leaflets which Cornwallis had composed in Arabic a few days before.
These were designed to assure the population that the British would drive out their enemies and that Rashid Ali and the Golden Square had betrayed them for German gold. Not all the leaflet-dropping sorties were received without retaliation and many took ground fire from the Iraqis. Several pilots responded by dropping the wrapped parcels so that the tightly packed paper fell on the heads of their assailants.
Baghdad
At 07.30 Rashid Ali made a radio broadcast which was repeated hourly through the day:
To the people of Iraq
We were forced to take defensive measures, and the military operations which have begun are continuing with successes to our Army. The noble Iraqi nation is requested to remain quiet, proving its political maturity, and confidence in our national forces. The people are requested never to attack foreigners among us who will be regarded as our guests.
The early part of the day at the Embassy was marred by several incidents. The reserve wireless set was up and working by 8 a.m., and by 10 a.m. contact was re-established with Habbaniya. By 11 a.m. an official from the Iraqi Ministry of Foreign Affairs turned up at the main gate with some army officers, demanding the surrender of the reserve wireless set. The Embassy gave the set up and reluctantly allowed a further search of the buildings under protest at the implied threat of force. Around noon the Iraqis requested that the British managers of the Imperial Bank of Iran and the Eastern Bank, who were at the Embassy, and the manager of the Ottoman Bank, who was at the American Legation, hand over their keys. If the managers refused, the Iraqi Ministry of the Interior would open the banks by force. The managers eventually agreed under protest in the face of force majeure.
Freya stood on the long, low terrace wall at the foot of the Embassy lawn looking out over the river. It was strange to see the Tigris so silent. Except for the police launch riding quietly off the Embassy wall, the river was devoid of traffic. Normally the 400-metre-wide dirty brown river would still be swollen from the winter run-off of snow in the mountains, and would be alive with river traffic almost as varied as the streets of Baghdad.
Gufas, large round barrel-like coracles made of reeds filled with watermelons would josh with Kalaks made up of planks and goatskins from Mosul transporting brushwood and grain. The beautiful slender Maheiles with pointed, overhanging prows would slid silently past the long, rectangular Shahturs, lateen-sailed booms or Dhows up from the Gulf with their sharp sterns and nose like a swordfish, Arab buggalows with intricately carved sterns and the gondola-like Bellums. All the while the passenger ferries operated by the Lynch Brothers would crisscross the Tigris, darting around the big flat-bottomed British-built river steamers from Basra downstream, who presided over the noise and chaos like disapproving parents looking down on unruly children.
However, Baghdad was a town under curfew, and nothing moved on land or river. Iraqi river commerce was dead. The high-sided ships bringing Welsh coal, Kenyan coffee, Egyptian cotton and Indian spices and tea, and carrying out Arabian horses, dates, gums, hides, liquorice, nuts and wheat, had all gone. The only thing in the river that remained the same was the odd dead dog or donkey floating downstream with the current.
Later in the day the Iraqis had demanded that the Embassy hand over the seven cars and two trucks inside the walls. They went away for a while when this was refused, but were back later with a request for the British to lower the Union Jack on the roof of the Embassy as this could incite hatred
While Loftman and some of the permanent residents became angry, Freya was pleased that Cornwallis overruled them and agreed to lower the flag, since the Iraqi request made sense under the circumstances. A little later he had the Union Flag raised on a flag pole in front of the main Embassy building, which could not be seen from outside the walls, thus pacifying Loftman.
Although Rashid Ali's radio broadcast was designed to keep the lid on popular opinion, this was being inflamed by the broadcasts from Radio Zessen and Radio Bari, which Freya could hear blaring out from nearby houses. Most irritating was the radio in a coffee shop a few yards from the southern boundary wall of the Embassy, which seemed to be amplified every time that Yunis al Bahri ranted and raved about the British. Radio Zessen and Radio Bari transmitted programmes five times a day, including readings from the Koran, anti-British news and German and Italian military successes. Since the outbreak of hostilities the radio broadcasts had begun to add historical descriptions of the glories of the Arab past to strengthen Iraqi self-assurance.
Heavy doses of martial music and a focus on nationalism, rather than religion, constantly calling on Arabs to revolt against the British, was heady stuff in a country with a low level of literacy. A demonstration by the Futuwwah, disparagingly known to the British as the 'Footwear', outside earlier in the day had also caused concern among those incarcerated in the Embassy.
Freya looked back and saw that in some windows of the houses overlooking the Embassy Iraqi soldiers had mounted machine-guns, and shivered, hoping that things would not spiral so far out of control that the Iraqi armed forces would either participate in, or turn a blind eye to, a massacre at the Embassy.
Freya could not help but think that the leaflets that the RAF had dropped earlier in the day had been of little use. Although written in Arabic they were not widely read by a largely illiterate population, and they probably helped to inflame rather than convert those who could read the content.
Berlin
Ribbentrop wrote to Hitler in an attempt to force support for Iraq:
If the available reports are correct regarding the relatively small forces the English have landed in Iraq so far, there would seem to be a great opportunity for establishing a base for warfare against England through an armed Iraq. A constantly expanding insurrection of the Arab world would be of the greatest help in our decisive advance toward Egypt. The figures regarding the British in Iraq show how weak England still is today at the Suez Canal.
The Iraqis have requested that Fritz Grobba be sent to Baghdad at once to help their efforts. Grobba had been our Minister in Baghdad until the war started and relations were broken off. He is our best Iraq expert. He should be flown at once to Iraq to direct and further expand our network of agents in the Near East.
The whole Arab world shall then be aroused into rebellion against England from our centre in Iraq.
I await your agreement on sending immediately a fighter and bomber squadron to aid the Iraqis, as well as air delivery of arms and ammunition.
Later in the day Hitler responded favourably to Ribbentrop's request, and the Reichsaussenminister sought an immediate meeting with Goering to convince him of the strategic value of the Iraqi operation.
Habbaniya
Dudgeon had a problem with Frankie. He could not leave the Dachshund alone to endure the continuous shelling, which was driving the dog mad. With nowhere to leave him, he decided, against all known RAF rules and regulations, to take Frankie with him on operational sorties.
He figured that as no one of any seniority seemed to be risking his neck if he didn't have to, there was no one close enough to cause him trouble.
After a couple of sorties Frankie had got the hang of flying. As soon as Dudgeon picked up his flying helmet Frankie would trot a few feet in front of him trying to divine which aircraft they would take. When they got to an Oxford Frankie would stand near the door at the rear of the plane waiting to be lifted in, as he was not tall enough to jump in. Once in the Oxford he would scamper up front and look out of the bomb aimer's window in the nose. As soon as they were airborne Frankie would come back and curl up on the floor next to Dudgeon's seat and close his eyes, and sit there undisturbed by all the manoeuvring until Dudgeon throttled back his engines for the approach. Then Frankie would run back to the bomb aimer's window to supervise the landing, only leaving his position to run aft when the aircraft stopped, to wait for the door to open.
Flying a Gladiator or Gordon, both with no floor, where Frankie could not lie beside Dudgeon, almost drove Frankie to distraction, but anything was better for him than leaving him behind among the Iraqi shells and bombs which fell continuously on the camp.
Holder's second Audax was hit in the fuel tank, drenching him in petrol. He crash-landed upside-down on the station's golf course and was trapped with his head in the sand of a bunker. Wad Taylor, whose wounds had been patched up, scrambled clear and with considerable effort raised the aircraft's tail, freeing Holder. They ran for their lives and had made some hundred yards when the Audax blew up.
Roberts, standing outside the Air Headquarters block, made his way between the slit-trenches to two old 18-pounder field guns that graced the entrance to the two-storey building. He had asked around and found out that they had been used against the Turks during the 1916–18 Mesopotamia campaign, and had been sitting around for years until brought to Habbaniya in 1937.
Although they were covered in paint in typical armed services fashion of 'if it doesn't move, paint it', with a coat for every year in place, they had not been disabled.
Roberts decided to fly a Royal Artillery artificer up from Shaiba on the next 31 Squadron flight, plus some 3.3-inch shells for the guns, to see if they could be pressed into service.
London
Churchill, increasingly concerned about the situation, and annoyed with Wavell's continuing reluctance to assume a greater burden, had the Chiefs of Staff send a directive to Cairo:
A commitment in Iraq is inevitable. We have to establish a base at Basra and control that port to safeguard Persian oil in case of need.
The line of communication to Turkey through Iraq has also assumed greater importance owing to German air superiority in the Aegean Sea. Had we sent no forces to Basra the present situation at Habbaniya might still have arisen under Axis direction, and we should also have had to face an opposed landing at Basra later on instead of being able to secure a bridgehead there without opposition.
There can be no question of accepting the Turkish offer of mediation. We can make no concessions. The security of Egypt remains paramount. But it is essential to do all in our power to save Habbaniya and to control the pipeline to the Mediterranean.
Habbaniya
The meeting with Savile was short and perfunctory. The Iraqis were still on the plateau, seemingly impervious to the death and destruction rained down upon them by the Air Striking Force while sending shell after shell into the camp.
The damage in the camp was beginning to mount. A number of barracks had lost their roofs, some houses and all of the hangars had been hit, slight damage was recorded to the officers' mess, the Assyrian church had been seriously damaged and a number of vehicles, including Smart's car, were write-offs.
As night fell and combat flying ceased, the Iraqi artillery began to use the freedom of darkness to step up their fire on the camp, throwing in about a shell a minute.
The only possible way to reduce the barrage was to bomb at night. Dudgeon and Hawtrey decided on a bold new plan. While they could not, for obvious reasons, use a flare-path which would illuminate them for the Iraqi gunners, they could use hand-held shielded torches among the trees along the polo ground to take-off, and use moonlight to land. While this took care of the Audaxes, the Oxfords faced a more difficult problem.
The polo ground was too short for the Oxfords to land, so they were confined to the main airfield. The final approach run to the main airfield was directly over the plateau, which would expose them to the Iraqi gunners. However, if the run was made without moonlight they could use their landing lights for a few seconds to touch down.
With only three Oxford pilots with night-flying experience, the decision taken was to fly three two-hour patrols, with the last pilot landing at daybreak. The Iraqi gun flashes would give away their positions on the plateau, allowing the Oxfords to target the guns. If the Iraqis did not fire, the Oxfords could drop one of their eight 20 lb bombs every fifteen minutes to encourage the Iraqis to keep their heads down.
Dudgeon took the first sortie shortly before midnight and ten minutes after moonset.
He started up the Oxford's twin Cheetah engines and was guided by shielded torches through the gates out onto the main airfield and into pitch darkness.
Lining up on a set compass course, and guided by his bomb aimer, Sgt Arthur Prickett, Dudgeon set the throttles and taxied forward for 4½ minutes, which brought them close to the plateau, and then turned onto a new compass heading and opened the throttles. The Oxford finally lifted off and Dudgeon brought the aircraft up to 1,000 feet.
Once they were airborne, the darkness was not so intense, and Dudgeon was able to pick out the winding Euphrates in the background, and the canopy of stars above gave him just enough ambient light with which to distinguish trees and roads and water from the desert.
For the next two hours Dudgeon criss-crossed the plateau, dropping bombs every now and again to keep the Iraqis from shelling the camp and from enjoying a good night's sleep.
The tricky bit was landing in pitch darkness. Switching off his cabin lights to regain his night vision, Dudgeon steadily banked to bring the Oxford over the lip of the plateau, and lost height down to 250 feet, lining up on a compass heading that should put the airfield dead-ahead. Counting ten seconds, he descended to just above the airfield, and when the altimeter read 50 feet he switched on his landing-lights. Seconds later the road around the airfield and then the ten-foot dyke flashed beneath his wings.
Dudgeon switched off the landing-lights and snapped the throttles shut at the instant the wheels touched the ground, using the light from the instruments as he braked to a standstill.
The emotional stress of the landing hit Dudgeon hard. He began to shake, hyperventilate and sweat profusely. After several minutes of silence, Arthur Prickett asked over the intercom whether he was all right. He shook his head to clear it and then looked over towards the hangars, and saw shielded torches signalling to him to taxi in.
The other two Oxford pilots did not fare so well. The first, a flying officer, managed to get down in one piece, but got out of the aircraft shaking, uttering, 'Don't ask that again, that's enough.' Dudgeon did not ask him again, nor did he think badly of him for it. The second Oxford, piloted by a WO, turned too early, leaving too short a take-off run. The wheels clipped the edge of the ten-foot dyke, and the Oxford somersaulted into the marshes beyond, the wooden aircraft exploding in a ball of fire: another Oxford lost, as well as a vital crew.
On the plus side, the night shelling of Habbaniya had been significantly reduced as a result of the night combat missions flown by the Audaxes and the Oxfords.
Brawn's fighting Levy patrols at night began to range far and wide outside the camp as they searched for Iraqis in outlying posts. So powerful was the impact of these patrols that the Iraqis began to vacate their forward posts at dusk, and never seriously attempted to penetrate Habbaniya at night.
Habbaniya, 4 May 1941
By the third day of the siege the defenders were getting into a pattern of 04.30 take-offs, strafing and bombing the Iraqis on the plateau at first light in conjunction with the Wellingtons of 37 and 70 Squadrons from Shaiba.
Eight Wellingtons again attacked Raschid Airfield, bombing buildings and strafing aircraft and scoring a direct hit on an Iraqi Nisr.
The Wellingtons attracted significant ground fire as well as interception by the Iraqi Air Force. The rear gunners on two Wellingtons put a Northrop and a Breda fighter out of action. However, one Wellington flew low over a 20 mm anti-aircraft gun while being pursued by an Iraqi Gladiator. Both fighter and guns hit the Wellington, forcing it to land, with the crew being taken prisoner.
The loss of irreplaceable pilots was a growing concern, and Dudgeon decided to push his luck with Savile. While Savile had been doing a sterling job in keeping supplies of fuel, ammunition and bombs going in spite of problems with transportation and shell damage, he disappointed the younger pilots. Dudgeon became more and more frustrated as he saw Savile pouring over the photo-mosaic maps plotting targets and assigning them to the pilots. Although Savile was an instructor of the highest quality and graded fully medically fit for any duty, he rarely flew, in common with almost all the senior officers, and he had not taken part in the strikes against the Iraqis.
Dudgeon, young, fatigued with flying around the clock and intolerant, saw a man dodging his responsibility. Finally, after Savile continuously turned down opportunities to take a sortie as he was too busy, he was confronted by Dudgeon. 'Sir, you have not yet seen the top of the plateau. I and your pilots believe that you should. My aircraft is outside, engines running, bombed-up and ready to hit the targets you just gave me. Come with me and we will be back in twenty minutes.'
'Dudgeon,' replied Savile, 'I just cannot leave the operations centre.'
'Sir, with all due respect I think that your presence in the air will encourage the younger pilots and will boost morale among the hard-pressed ground-crews. A personal example will go a long way to supporting their efforts.'
Savile looked very unhappy, but at last gave in and agreed to go.
After Dudgeon had completed several bombing runs, Savile, sitting next to Tony, asked, 'What are those zipping noises?'
'They are machine-gun and rifle bullets passing close by', responded Dudgeon.
At that instant Savile looked out of the cockpit and saw a row of holes being stitched through the plywood wing. He yelled, 'My God! We've been hit! Go back and land at once.'
Dudgeon as captain ignored the demand and completed the remaining bombing runs of his sortie. Savile sat quietly next to Dudgeon with his hands on his knees, staring straight ahead.
This was the only time a senior RAF officer ventured into combat at Habbaniya.
Dudgeon later regretted what he had done to Savile, whose medal ribbons clearly showed that he had contributed his full share during the Great War.
Smart and his family were evacuated on the first 31 Squadron DC-2 early that morning. The reason for Smart's departure has never been satisfactorily explained. The official view is that he suffered severe injuries including concussion, a broken jaw and multiple fractures in a car crash in the blackout on the night of 3 May. Those on the spot believed that it was nervous exhaustion.
Unexpectedly Sqn Ldr Pike of 203 Squadron arrived with four Blenheim fighters from Cairo. No one had told the defenders that they were coming, and no one had briefed the Blenheims that the plateau was thick with Iraqi anti-aircraft guns.
Coming in to land at Habbaniya with a normal approach over the plateau, 203 Squadron got a rough reception from the Iraqi gunners. Luckily some Audax pilots saw the approach of the Blenheims and did their best to keep the gunners' heads down by dive-bombing ahead of the incoming Blenheims. The arriving pilots were staggered to see airmen waving them in behind the hangars, and hesitated, but bursting Iraqi shells got them to make their minds up. Three of the Blenheims were subsequently found to have bullet holes, while one of the pilots swore that he had flown through the dust of an exploding 250 lb bomb on his approach run.
London
With no response from Wavell, Churchill sent another, more strongly worded, personal directive to Cairo:
Our information is that Rashid Ali and his partisans are in desperate straits. However this may be, you are to fight hard against them. The mobile column being prepared in Palestine should advance as you propose, or earlier if possible, and actively engage the enemy, whether at Rutba or Habbaniya. Having joined the Habbaniya forces, you should exploit the situation to the utmost, not hesitating to try to break into Baghdad, even with quite small forces, and running the same kind of risks the Germans are accustomed to run and profit by.
There can be no question of negotiation with Rashid Ali. Such negotiations would only lead to delays, during which time the German Air Force will arrive.
You do not need to bother much about the long future in Iraq. Your immediate task is to get a friendly government set up in Baghdad, and to beat down Rashid Ali with the utmost vigour.
Baghdad
Paul Knabenshue at the American Legation was incensed by the Iraqi warning he had received that morning. The Iraqis had threatened to bomb the British refugees in the American Legation in retaliation for an alleged British threat to bomb public buildings in Baghdad if Iraqi forces did not evacuate the Habbaniya area.
Knabenshue agreed to surrender the British refugees only if an Iraqi cabinet minister would sign a statement guaranteeing adequate and safe internment facilities. Since no official would comply, Knabenshue offered the American Legation as an internment camp and accepted responsibility.
The Iraqis reluctantly accepted Knabenshue's offer.
Habbaniya
Lt Kit Wilson of the 3rd Field Regiment, Royal Artillery, beamed happily at the two 18-pounder field guns sitting outside the Air Headquarters block.
He had flown up from Shaiba late on the 3rd with a set of tools, an engineer and an artificer, sitting on three crates of 3.3-inch shells amid a jumble of boxes of tinned foodstuffs and bags of vegetables and meat from Paul Holder's first 'shopping' expedition.
It had taken several hours for the working party to strip off the accumulated paint and get the breech, rifling and firing mechanism cleaned.
Wilson walked over to Ouvry Roberts. 'Well, sir, I think you have got yourself two heavy guns', pointing at the two now gleaming howitzers. 'Would you like me to give a demonstration?' he asked.
Wilson got help from several airmen to turn the guns to point at the plateau. His engineer, together with the artificer, loaded both howitzers with their heavy 23 lb shells while he adjusted the sights for an airburst at the edge of the escarpment.
He walked back to Ouvry Roberts. 'Would you like to join me, sir?' he said, offering the lanyard of one of the guns to the Colonel, who happily accepted the offer.
'On my count of three, pull the lanyard. One, two, three... '
Both guns thundered simultaneously, which brought airmen and soldiers running out of the barracks and blocks nearby to see what was happening. They, together with those manning the slit-trenches and the working party, stopped to watch with awe as one shell hit the lip of the plateau and the other exploded about fifty feet further along and twenty feet above the escarpment.
As a loud cheer went up from the Habbaniya defenders, Wilson said to Roberts, 'I think it just needs an adjustment or two.'
Baghdad
Across the river from Knabenshue's 'White House on the Tigris', life in the British Embassy had its trials and tribulations.
Freya had spent part of the previous evening trying to suppress false rumours that Rutba Wells had been taken, and had finally ended the evening with a welcome hand of bridge.
In the morning, up on the Embassy roof Freya could just see the Wellingtons bombing Raschid air base to the south, and hear the dull crump as the bombs exploded and the sharp crackle of returning antiaircraft fire.
The police in the motor launches retaliated by shooting their rifles every time that they glided past the Embassy, and Freya spent a lot of her time soothing the nineteen female refugees – Iraqi Armenians and Jews, Greeks, an Indian family and a Yugoslav.
Later Freya, together with some Embassy staff and Ernest Main of the Baghdad Times, helped to compose, print and distribute a daily news bulletin compiled from the BBC broadcasts.
A drama occurred when a superintendent of the Iraqi police arrived at the main gate. However, this was a false alarm, as he was only there to provide a friendly escort for the Embassy lorry to buy food. Unfortunately all the shops had been closed by their owners, terrified of being bombed by the British. This precipitated strict rationing in the Embassy.
Freya managed to get on friendly terms with many of the police guarding the Embassy, and found them happy to help in any way provided that they were given a little bakhseesh.
Late in the day the Iraqi Ministry of Foreign Affairs asked for a list of the names of all people in the Embassy of whatever nationality, accompanied by a certificate giving an assurance that no one would leave the Embassy before the Ministry had been advised. Since no one at the Embassy had any desire to do so, and this contravened diplomatic protocols, the Embassy refused to respond.
Habbaniya
At the end of the day the number of recorded sorties by Habbaniya's Air Striking Force had fallen to fifty-three, added to which were the night sorties flown by the Audaxes in the moonlight and the Oxfords in pitch darkness.
However, the Iraqis were still on the plateau and raining shell, machine-gun and sniper fire down on the camp.
Habbaniya, 5 May 1941
Overnight the ground-crews had been able to repair the bullet holes in the three 203 Squadron Blenheims, and one took off shortly after dawn before the departure of the Air Striking Force to maintain a standing combat patrol over Habbaniya.
The Iraqi Air Force later sent up two of their Pegasus-engined Nisrs to strafe and bomb Habbaniya.
They saw the Blenheim and fled, although one was not fast enough. The Blenheim overhauled the Nisr flying flat-out, low down over the Euphrates, and shot it down into the river with a single burst of its four .303 machine-guns.
As the Blenheim sedately waggled its wings and returned to its combat patrol, a tremendous cheer came from the camp, which must have been heard by the Iraqis on the plateau. From then on the frequency and strength of the Iraqi Air Force raids on Habbaniya declined dramatically.
Two of the Blenheims hit Raschid air base and the civil airport at Baghdad. At Baghdad, a Gladiator and three Nisrs were strafed, while at Raschid a Savoia, a Breda and a Nisr were attacked, with the Nisr being left in flames.
The fourth Blenheim, piloted by Sgt Hemsted, flew a photo-reconnaissance mission to Kirkuk, Baghdad, Solomon Doh and Mosul. At the latter he strafed two Nisrs, but was hit in the tail seven times by anti-aircraft fire.
Cairo
Wavell to Chiefs of Staff:
Your message takes little account of realities. You must face facts. I feel it my duty to warn you in the strongest possible terms that I consider that the prolongation of fighting in Iraq will seriously endanger the defence of Palestine and Egypt.
The political repercussions will be incalculable and may result in what I have spent the last two years trying to avoid, namely, serious internal troubles in our bases. We should seek to accept a negotiated settlement via Turkish good offices.
Later that day the operational command for northern Iraq passed from Auchinleck in New Delhi to Wavell in Cairo.
Baghdad
Day after day the temperature soared, reaching 115ºF on 5 May. The heat in Baghdad was moist and tropical and oppressive, and added to the air of depression. Men at the Embassy stripped to their shorts, and the women did what they could to stay cool.
To combat depression, life in the beleagured Embassy was taken in hand. The defence of the premises grew daily more complicated as more and more barbed wire was wound around trees and hidden in flower-beds, and more and more sandbags were filled and barricades built – only to be resited the next day. The largest organization in the Embassy was the guard, which, with few exceptions, embraced all the able-bodied European men. Divided into four main and one subsidiary squads, these volunteers watched the perimeter by day and patrolled it by night.
As George Waterhouse was liable to conduct an inspection at any time, all the watchers, wisely, were at their posts and awake – all except an elderly Lancastrian who was provided with a chair for his short post-lunch duty of keeping an eye on the main gate. One afternoon Waterhouse, on a tour of the defences, found him asleep, and severely reprimanded the man. The following day his chair was seen to be facing the Embassy drive, with his back to the gate. The defaulter's superior demanded an explanation, and was told, 'I can see the old bastard better comin' from 'ere!'
There was also a group of roof spotters who did duty from dawn to dusk in the increasingly hot weather to follow the progress of the war in the air. Their reward came from seeing the RAF shoot down Iraqi planes and drop bombs on Raschid air base, the Washash camp and the airport.
Except for his bedroom and a small dining room, Cornwallis had given over the Embassy to his 'guests'.
Catering had been divided into five sections according to dietary habits – European, carnivorous Indian, vegetarian Indian, Arab and Orthodox Jew. The European canteen, catering for about 150, nicknamed the 'Corner House', was run by the Matron of Faisal College and one of the nurses from the Royal Hospital, under the supervision of the Director-General of Iraqi State Railways.
Many of the women and the older men sat around in the shade of trees in the garden, and even the police in their motor-boats on the Tigris opposite the Embassy lay stretched out asleep on the thwarts in a square of shade which fell on the water.
The lack of police vigilance made it easier for people to enter and leave the Embassy, and a man was seen hawking a bundle of fish on a pole as he walked in and around the premises.
One of the biggest problems for the Iraqi refugees in the Embassy was an absence of their own bread. Freya, using her Arabic and some baksheesh, was able to arrange for supplies of tannura, which was gratefully received by the Arabs, Persians and Kurds.
The six horses stabled in the Embassy were getting rather bored by the lack of exercise, and the Kurdish grooms walked them around the grounds and Freya fed them dates.
She found that the leaflets dropped by the RAF were causing more trouble than they were worth. Empty threats and the promise to bomb government buildings had been counter-productive, and Freya advised Ken to telegraph Habbaniya to stop the leaflet drops.
Habbaniya
By the evening of the fourth day Savile, Hawtrey and Dudgeon took stock.
Although the Air Striking Force aided by the Wellingtons out of Shaiba had flown some eighty recorded sorties, only four of the original twenty-seven Oxfords were now 'flyable', and the remaining Audaxes and Gordons were literally held together with fabric patches pasted over the bullet holes. Even more problematical was the fact that in addition to the wounded, evacuated and dead pilots, an additional four, from Dudgeon's original nineteen, had to be relieved of flying duties due to nervous exhaustion.
During the night a fighting patrol of the King's Own and the Levies had tried to drive off the Iraqi troops which had invested Sin el Dhibban. The village lay below the northern end of the escarpment some four miles to the south of Habbaniya and was a vital ferry point across the Euphrates, providing a route to reinforce the Iraqis on the plateau. For the first time they failed to shift 200 Iraqis entrenched in well-defended positions.
Washington DC
'Wild Bill' Donovan had spent the morning with the President, Franklin D. Roosevelt, in the Oval Office finalizing proposals for a new government agency which would combine intelligence gathering with clandestine warfare.
On several trips to Britain as FDR's personal emissary, he had been impressed by SOE and the Commando units, and by Churchill's doctrine of taking the war to the enemy by 'setting Europe alight'.
Although both FRD and Donovan had little doubt that America would be fighting the Germans and the Japanese, sooner rather than later, they had to overcome not only strong domestic political pressures to remain neutral but also the Army, the Navy and the State Department, who would fight to the death to retain their control over intelligence.
FDR had proposed that Donovan be named first as Coordinator of Information in June, followed a year later by the formation of the Office of Strategic Services.
This, believed FDR, would give the military time to adjust to a new agency and Donovan the space to begin recruitment. FDR had also decided to promote Donovan from his current rank of lieutenant-colonel to major-general, to give him more political clout, recognizing that Washington DC was a city where the trappings of power and position were everything.
As Donovan shut the Oval Office door behind him, Marguerite 'Missy' Le Hand, FDR's personal secretary and close confidante, asked Donovan if he could spare a few minutes to talk to the Secretary of State. As Donovan nodded his agreement, Missy called through to Cordell Hull's secretary to advise her that Donovan was on his way over.
It took Donovan a good eight minutes to cross from the White House to the Old State Building, climb the ornate staircase to the second floor, and walk down the long corridor to the Secretary of State's office in the south wing.
Donovan did not have too much time for politicians, especially the brand in Washington, who seemed to be doing their best to line their own pockets or those of their sponsors. But Cordell Hull was different. Genuinely born in a log cabin in Tennessee, Hull, a tall lean man, was earnest and sincere in a way that comes to someone thoroughly convinced of the righteousness of his political and economic policies for peace and justice, as well as the moral and commercial supremacy of the United States.
Hull's secretary was waiting for Donovan and showed him into Hull's office, where Hull, already on his feet, gestured to a dark-green leather-buttoned chesterfield and matching easy chairs that seemed lost in a corner of the large office.
Hull was a no-nonsense politician, and the ornate stencilled wallpaper and the decorative furniture had been panelled over or removed, leaving only the massive fireplace, with its carved mantel and Doric columns as a reminder of Room 208's former richness. For Donovan the nicest feature of the room was the three windows opening onto the south portico, which gave a magnificent view of the Potomac.
'Thank you for coming over to see me, Bill. I heard that you were with the President, and hoped that we could talk for a few minutes.'
'Of course, Cordell, anything for a fellow lawyer, how can I help?'
'The situation in Iraq does not look good for the British. The Iraqis have besieged a British air base, the Regent has fled the country and our Legation is surrounded, and the Iraqis have taken away our radio. All our information now comes from our Ambassador in Cairo, who is briefed by British Headquarters. Adding to the problem is news from our highly placed source in the Vichy government that the Germans want to use Syria as a staging-post to supply the Iraqis with arms and equipment and possibly troops.
'In my eyes the Near East offers the greatest danger of the war, the possible juncture of German and Japanese forces, effectively cutting the world in two. Not only do we have to protect our own economic interests in the area, but the British cannot be seen to lose. It is not the oil that is at stake – we supply 90% of Britain's gasoline – but a British loss of face that would have a catastrophic effect on their position in Egypt, as well as Palestine fanning the flames for revolution. With revolts breaking out all over the Middle East the British would have little option but to disengage, leaving the Germans with North Africa, the Middle East and the Suez Canal.
'However, that is only my opinion. As Secretary of State I have to deal with a set of competing interests which has led to muddled thinking in US policy over Iraq, as well as the whole of the Middle East. Congress will not stomach war, so FDR has gone as far as he can go by virtually declaring economic warfare.
'Added to FDR's problems are the isolationist lobby led by the America First Committee supported by Lindbergh, a bunch of Ivy League intellectuals, a number of influential senators and congressmen and some of the Hollywood set. All of which are financed by a motley collection of industrialists and newspaper magnates.
'FDR is committed to keeping Britain afloat because sooner or later he will have to face the Nazis. FDR puts the defence of the Middle East in fourth place behind that of Britain, Singapore and the Ocean trade routes. FDR is not interested in helping the British to preserve British Imperialism nor their political or commercial pre-eminence in the Middle East after the war. He sees this as an opportunity for the United States.
'This is to do with the interests of US oil companies who want to be able to muscle in on the British monopolies in Iraq and Iran, but also the possibility that there are substantial quantities of oil in Saudi Arabia. It is also to do with the policy we adopted at the end of the Great War, which commits us to the creation of a Jewish national home in Palestine. Those are the interests that I have to contend with.
'Bill, you have been to Baghdad, how do you read the situation?'
'I share your concerns. I was there for a few days this January during my Mediterranean fact-finding trip. In my view the British had done little to quench the flames of dissent. Their then Ambassador spoke only to pro-British Iraqis, people with position and power to maintain, so he had no real idea of what was going on. On the other hand our Ambassador, Knabenshue, was a lot more switched on and had already made strong representations to the Iraqi government to cooperate with the British. Knabenshue fixed up meetings for me with a number of leading anti-British and neutral politicians for me to outline America's firm support and all possible aid short of war to the British.
'I also met with Haj Amin, the Mufti of Jerusalem. He very reluctantly came to see me at the Legation. I expected him to arrive with his retinue, but I think that he concealed his visit as he came alone. He is a malevolent influence and is behind the current problems in Iraq. I explained to the Mufti, in no uncertain terms, that America, though we have not yet declared war, is behind Britain and would resent the activities of people working against her. I told him to look ahead and understand the consequences of actions which, being against the interests of Britain, were against the interests of America. I asked him to explain our position to members of the Golden Square.'
'What was his response?'
'He agreed, but I am sure that he did not relay my message to them, as he continued his intrigues. The result has been this revolt.'
'What do you think will happen?'
'It is difficult to judge, as it could go either way. The British are heavily extended and have very little, if anything, left to give. It largely depends on whether the Germans get to Baghdad in strength before the British. If that happens it's the beginning of the end for the British in the Middle East, and it may well signal the end of British resistance to Germany.'
'Is there anything that we could do?'
'We are already doing everything we can short of war', replied Donovan.
Habbaniya, 6 May 1941
The sharp decline in airworthy aircraft and the loss of pilots and aircrew now meant that what was left of the Air Striking Force's resources were at a premium, and they had to husband their ever-diminishing planes and crews.
Savile, Hawtrey and Dugdeon had waited until dawn to send up an Audax on a photo-reconnaissance run over the plateau and up the Baghdad road as far as Falluja.
The Audax pilot's news was good. By dawn the plateau was beginning to empty of Iraqi troops, who were withdrawing in the direction of Falluja.
Roberts immediately dispatched his armoured cars and infantry patrols in pursuit, clashing with the Iraqis on the Falluja road and in the village of Sin el Dhibban.
The lack of supplies, the continual air attacks and the use of the camp's 18-pounder 'artillery', which suggested that the camp was being reinforced by air, they later found out all contributed to shaky Iraqi morale.
Eight Iraqi Crossley armoured cars out of a column of fourteen were able to withdraw safely to Falluja, but the other six were destroyed as a counter-attack was being prepared to recapture lost Iraqi positions.
The 4th Iraqi Infantry Brigade which had initially been holding the plateau had been withdrawn after two days, being replaced by the 11th Iraqi Infantry Brigade and the 7th Field Artillery Brigade. These were now withdrawing due to the psychological pressure of continuous bombing and strafing and high casualties.
As the Iraqis from the plateau streamed onto the road to Falluja, an Iraqi column was leaving Falluja bound for the plateau to follow Colonel Fahmi Said's orders to recapture the lost Iraqi positions at whatever cost.
A patrolling Audax saw the column of four Crossley armoured cars, the 2nd Infantry Battalion of the 1st Infantry Brigade transported in civilian cars and the 5th Desert Artillery Regiment coming from Falluja meet up with the 11th Infantry Brigade and the 7th Field Artillery Brigade and their equipment, staff cars and trucks retreating from the plateau. This caused a huge nose-to-tail jam of men, vehicles, equipment and horses.
The Audax pilot wasted no time in landing and relaying the situation to the operations room in Hangar Six. The Air Striking Force would never get another opportunity like this, and Savile got every Audax, Gladiator, Gordon and Oxford that could be possibly coaxed into getting off the ground into the air to bomb and machine-gun the Iraqis on the ground. Forty aircraft from Habbaniya went into a pattern of bombing and firing on everything on the Falluja road, racing back to the airfield and the polo ground, rearming with their engines running and taking off again and heading back to the Falluja road to repeat the process.
The Air Striking Force took full advantage to cause mayhem by shooting and bombing everything along the only available road. They first attacked vehicles at the front and rear of the column to wreck their vehicles to trap those in the centre. With marsh and flooded ground on either side of the road, the vehicles had nowhere to go and were destroyed where they had stopped.
In a devastating two-hour attack on the road the Air Striking Force made 139 sorties When the last aircraft left, the pilot reported that the road was a strip of flame hundreds of yards long with ammunition limbers exploding and civilian and Army staff cars and troop transport trucks burning by the dozen. The charred and battered remnants of the two convoys littered the Falluja road for several weeks afterwards.
The Iraqis had lost about seventy vehicles and over five hundred troops killed or wounded in this single engagement, while the Air Striking Force had lost one Audax shot down.
Six Wellingtons of 37 Squadron from Shaiba attacked Raschid air base and Baghdad airport, setting fire to a hangar and inflicting damage on parked aircraft. They then dropped thousands of leaflets over Baghdad and bombed the Washash barracks in the western part of the city. However, the transport Valentias had a bad day. One crashed at Pumping Station K-3, while three more had to land in the open desert for a variety of reasons.
Sin el Dhibban
Unfortunately the Iraqi troops dug in at Sin el Dhibban, on the high ground beyond and in trenches by the New Lake Road did not follow the example of their compatriots and withdraw, but stubbornly held the village, forcing Habbaniya's ground troops into a major offensive.
Roberts's plan was for B Company of the King's Own under Capt David Clayton to attack the village at the eastern end of the plateau from the flank, while D Company under Maj Nigel Gribbon moved around to the west and made a frontal attack.
B Company met stiff resistance from intense Iraqi automatic weapons fire, and started to take heavy casualties. Lt Thompson and Private Owen pressed home the attack, firing bren-guns from the hip and lobbing hand grenades. Thompson came upon an Iraqi heavy machine-gun post, and called on them to surrender. As they raised their arms Thompson lowered his bren-gun and an Iraqi shot him dead. Owen and the last of the platoon finished off the rest of the post. B Company started to extricate themselves, with a wounded Pat Weir giving covering fire from his bren-gun.
D Company had better luck following behind the RAF armoured cars, and mounted three successive attacks supported by the Levies, the RAF armoured cars and the two 18-pounder field guns to drive the Iraqis out of the village. At one point the Levies had to intervene. Capt Cottingham, RAB100 Stephan Nessan and Sgt Lazar Adam, all from the Levies, loaded up a Commer truck with a Vickers heavy machine-gun and a 3-inch mortar, and were able to get into a position by the sewage farm to enfilade the Iraqi positions. Together with support from a Levie company under RAB100 Khamsi Putros Odisho, their withering fire broke up the dangerous possibility of an Iraqi counter-attack.
D company began to take large numbers of prisoners, which required guards, reducing the strength of the already depleted company. At that moment a Valentia landed at Habbaniya with the band of the King's Own, who were rapidly trucked over to the village and pressed into service to reinforce D Company in mopping-up Sin el Dhibban. A young lance-corporal with a Webley who was seen chasing a terrified Iraqi stopped only to shout to the nearest officer, 'Do we take prisoners, Sir?'
The dust and heat haze, together with the fluid lines and the lack of radio communications, made air-to-ground and artillery cooperation extremely difficult, and much of the action was in hand-to-hand combat.
At the end of the day when the action had finished and the British were in possession of Sin el Dhibban and the plateau, they had lost seven killed and fifteen wounded in action. Together with the estimated five hundred killed on the Falluja road, the Iraqis had lost anything up to one thousand men killed in the battle for Habbaniya.
Basra
Convoy BP1 with the 21st Indian Infantry Brigade, comprising the Frontier Force Rifles and two battalions of the Gurkha Rifles, with two troops of the 13th Lancers equipped with armoured cars and a detachment of Engineers landed at Basra from Bombay. This gave enough strength for the 10th Indian Division to take over more of Basra, including the main telegraph office and the wireless station, albeit at the loss of a few Gurkhas to snipers.
Although Iraqi civil servants, dock workers and the crews of the dredgers keeping the port free from silt still refused to work, the situation in Basra was now stable enough to begin to re-establish communications inside Iraq.
Berlin
The drive from Templehof airfield to the Reichsluftfahrtministerium building took only twenty-five minutes, but Oberst Werner Junck was late. His flight from Calais, where he was stationed with Jagdfliegerfuhrer 3 at Wissant, had been delayed by almost an hour due to a violent spring storm over the Ruhr.
It was well known that General Hans Jeschonnek, the Luftwaffe's Chief of Staff, detested tardiness. Junck's sense of foreboding increased as he entered the sober façade of the ministry building – another monument on the Wilhelmstrasse to Ernst Sagebiel's National Socialist intimidation style of architecture.
Jeschonnek, a hardworking and intelligent officer, was not best pleased with the mission to Iraq when he was stretched for men and machines with both the up-coming attack on Crete and Barbarossa. Nor was he ecstatic with his visitor's tardiness, which he put down to disrespect. None of this was the fault of Junck. But to show his displeasure he had Junck stand in front of his desk for a few moments before looking up.
Jeschonnek was a complex man, and although highly professional he was out of place in the higher echelons of the Luftwaffe. He was complex and often abrasive, brusque and sarcastic to his subordinates, while his youth and immaturity made life difficult for him with more experienced leaders.
'Sit down, Junck, this is not a parade ground', Jeschonnek snapped. 'This is a difficult operation. You are to organize and lead a special force, which we have codenamed Sonderkommando Junck, to go to Iraq to help them to fight the British. The Fuhrer desires an heroic action', he said, pursing his lips.
'What does that mean precisely, Sir?' questioned Junck.
Jeschonnek, realizing that he had treated Junck poorly, replied, in a less aggressive tone, 'This is an operation which would have significant effect, possibly leading to an Arab uprising, in order to start a Jihad, or Holy war, against the British.
I want you, as Fliegerfuhrer Iraq, to establish your staff headquarters on Rhodes. Reporting to you will be Oberst Rodiger von Manteuffel and his team with headquarters at Aleppo airfield in Syria. He is responsible to you for organizing ground support in Syria, including refuelling, flight planning and the transportation of your equipment. You will also have Hauptmann Harry Rother who will be in overall command of the transport element, and Major Axel von Blomberg who will be your liaison to their Prime Minister, Rashid Ali.
Our operations staff working with the Oberkommando der Wehrmacht (OKW) have found you a Staffel each of Messerschmitt Bf110 fighters and Heinkel He111 medium bombers, together with a transport element – some Junkers 52s and a few Junkers 90s. We have added the big Junkers to carry bombs, anti-aircraft guns, lubricants and even petrol, as we cannot guarantee the quality of Iraqi-produced aviation spirit. When we finish I suggest that you go down with Blomberg, who is waiting for you outside, to the operations staff office on the third floor, and they will fill you both in on all the details.
We have started negotiating with the Vichy French to provide refuelling and transit rights for Syria. They have been stalling, but the Fuhrer has a meeting with Petain in a few days and he will be put under extreme pressure to cooperate.'
'Is there anything else I should know, sir?' asked Junck.
'Yes, General Hellmuth Felmy, who is in Athens, has been earmarked as the head of the German military mission to Iraq. You may have heard of Felmy. He is extremely ambitious. Stay out of his way. Your chain of command is through me. Oh, there will also be a civilian involved. His name is Grobba and he is from the Auswartiges Amt. Grobba will be based in Baghdad and is the political link with Rashid Ali. He also has the reputation of being a prima donna, so you will need to step carefully on the ground. You should also know that Felmy is Grobba's brother-in-law. Good luck!'
With that he dismissed the Oberst. Junck was left with the distinct feeling that he was stepping into a snake pit.
London
Chiefs of Staff to General Wavell
Your telegram of yesterday has been considered by Defence Committee. Settlement by negotiation cannot be entertained except on the basis of a climb-down by the Iraqis, on Iraq. Realities of the situation are that Rashid Ali has all along been hand-in-glove with Axis powers, and was merely waiting until they could support him before exposing his hand. Our arrival at Basra forced him to go off at half-cock before the Axis was ready. Thus there is an excellent chance of restoring the situation by bold action, if it is not delayed.
Chiefs of Staff have therefore advised Defence Committee that they are prepared to accept responsibility for dispatch of the force specified in your telegram at the earliest moment. Defence Committee direct that the AOC Habbaniya should be informed that he will be given assistance, and in the meanwhile it is his duty to defend Habbaniya to the last. Subject to the security of Egypt being maintained, maximum air support possible should be given to operations in Iraq.
Habbaniya
The Iraqi Air Force hit Habbaniya twice that day. While a Gladiator did intercept and damage a Northrop, Iraqi bombs and machine-gun fire destroyed two Oxfords, one Gladiator and an Audax on the ground, as well as killing seven people and wounding eight more.
The second raid was particularly upsetting for Dudgeon. He was away from his Oxford making his sortie report to Savile and getting details of his next target when an Iraqi plane dropped a stick of bombs nearby. When he returned to the Oxford the engines were still running but his bomb aimer, Arthur Prickett, was lying on his side underneath the aircraft where he had been loading the bomb-racks for the next sortie.
He felt no pulse and, rolling Prickett over, saw a small wound just over his heart with only a little blood. It had been Dudgeon's turn to load the bombs while Prickett reported in to Savile, and by rights Dudgeon should be on the ground
With a heavy heart Dudgeon called for an ambulance to take Prickett's body to the hospital, and told one of the ground-crew to inform Savile. He checked the Oxford for other damage and loaded the remaining three bombs, calling out to the nearest pupil-observer, Stuart Smith, that he was Dudgeon's new bomb aimer. Once aboard, he and Smith, together with Frankie, took off for the next raid on the Falluja road.
Later that night Savile, Hawtrey and Dudgeon met at the operations room at the back of Hangar six for their review of the day's actions and plans for the following day together with Roberts, Everett and Brawn.
'Just before nightfall I sent up an Audax to take a shufti over the plateau,' said Savile, 'to see if there were any Iraqis still there. He dropped some bombs and machine-gunned his way around the plateau, but there was no return fire. He took photos from height and from some low-level passes. But as you can see,' Savile said, and pointed at the new photo-mosaic covering the map table, 'there is no sign of the Iraqis whatsoever.'
With the immediate threat ended, they agreed to wait until the following morning and a second photo-reconnaissance flight to decide upon their next moves.
Dudgeon and Hawtrey retired to the bar at the officers' mess to celebrate their victory over the Iraqi ground forces investing Habbaniya, and to recount their operations.
During the five days of combat in their poorly armed and obsolete training aircraft they had managed to drop well over 3,000 bombs, about fifty tons of ordnance, and had fired 116,000 rounds of ammunition. Savile had recorded 647 sorties, but they both knew that there were many more. Many of the Audax pilots were too tired to walk the half-mile from the polo field to report to the operations room after each sortie, and both Dudgeon and Hawtrey had also sent up the occasional aircraft without reference to Savile.
However, the price had been high. From the school alone they had lost thirteen killed and twenty-one too badly wounded to fly, as well as four grounded by nervous exhaustion. Those who joined from other formations in the camp, serving as both pilots and aircrew, were not included as they were never recorded as casualties by Savile.
It terms of sheer losses sustained in relation to the resources employed, the RAF had lost more men in air combat over Habbaniya than in the Battle of Britain.
In between whiskies, Hawtrey said, 'Do you realize that not one senior officer from Air Headquarters or the padres ever came onto my polo pitch, or down to your airfield, to give us an encouraging word or to comfort the wounded?'
It was a sad but true reflection of the support given by senior noncombatants. Many found themselves in an unexpected situation for which they were unprepared and inexperienced, and they just ground to a halt.
Habbaniya, 7 May 1941
For the first time in more than a week a more relaxed Savile, Hawtrey, Dudgeon and Holder met with Roberts, Everett and Brawn at the operations office at 06.00.
The night had been strangely silent with no Iraqi shelling, and they had all benefited from a night's uninterrupted sleep. The only one put out by the changing fortunes of Habbaniya was Frankie, who no longer slept in Dudgeon's bed.
The reconnaissance Audax had reported that there was no large-scale Iraqi presence in the area surrounding the camp, but only a few small, isolated units, while Page's armoured cars had confirmed that the plateau was completely deserted.
Without the threat of constant shelling, the wreckage of the Wellington and an Oxford, as well as the tractor, could now be cleared away from the main airfield, and aircraft could be repaired and serviced in the hangars in the shade out of the searing heat.
The King's Own and the Levies had begun to collect equipment abandoned by the Iraqis on the plateau, at Sin el Dhibban and along the Falluja road. The haul was impressive – field guns, gun tractors, cannon, machine-guns, a light tank and ten armoured cars, trucks and civilian cars and over three hundred rifles and half a million rounds of ammunition.
All of this equipment was far superior to the arms of both the British and Indian Armies, enabling Roberts to rearm the King's Own and the Levies with more and better weapons.
However, the war was far from finished. The decision taken was that the Air Striking Force's main role would now be in finishing off the Iraqi Air Force as a fighting force, while the ground forces would concentrate on mopping-up the few Iraqi units still in the area and pushing the Iraqis back to Falluja.
A lone Iraqi Nisr raided Habbaniya in mid-morning. A fighter Blenheim was scrambled, piloted by Sgt Hemsted. Giving chase, Hemsted intercepted the Nisr, shooting it down. In the early afternoon two Blenheims flew a reconnaissance over Baquba and found a collection of twenty-one Bredas, Nisrs and Northrops on the ground and strafed them. Later in the afternoon they returned with some of the Air Striking Force's Audaxes and Oxfords. The Blenheims set three Iraqi aircraft on fire while the Audaxes and Oxfords destroyed three more and damaged others.
An Iraqi Gladiator attempted an interception but was engaged by Plt Off Watson in an RAF Gladiator. Watson fired a long burst from astern at point-blank range, and the Nisr rolled over onto its back and was last seen diving steeply away, trailing smoke.
Baghdad
In contrast to previous days it was cool and pleasant at the Embassy. However, an air of depression began to take root as the idea that the problems would not be solved soon began to set in among the residents.
In an effort to fight the growing morale problem, Cecil Hope-Gill turned his public affairs section into an Amusement Committee to keep the enforced inmates happy. He collected all available talent, drew up programmes for the afternoon and evening, and lectures were provided on the most unlikely subjects, as well as producing entertaining diversions like Tea-Time Topics, Siesta Spotlights and the like on widely differing topics.
Every evening there was a least one dance, concert, lecture or cinema show, as well as bridge drives, a treasure hunt, crosswords and puzzles. One of the big moments of the day was the Embassy 'Broadcast' at
7.30 p.m., when Dr Harry Sinderson, doyen of the British medical community in Baghdad, produced a fifteen-minute roundup of news, anecdotes, jokes and the inevitable limerick under the guise of the Embassy Broadcasting Corporation.
Several RAF air raids occurred that day, and an Iraqi Petroleum Company oil tank was hit to the north on the eastern side of the river. A great plume of black smoke billowed out of the oil tank and rolled over the houses. The fire burned through the night, greeting the dawn with a pall of smoke which hung over Baghdad and the river.
An RAF Audax dropped a message bag into the Embassy during the late afternoon but was driven away by machine-gun fire.
Habbaniya
During the night the AOC's office at Habbaniya received a telegram from Churchill:
Your vigorous and splendid action has largely restored the situation. We are all watching the fight you are making. All possible aid will be sent. Keep it up.
Savile had copies made and posted on notice boards throughout the camp.
Berlin
Grobba, accompanied by Dr Hans-Ulrich Granow, Willi Steffan, two radio technicians and Ahmed Jabr, a Palestinian German from the Brandenberger Regiment, left Staaken airfield in Berlin at 16.10 on a Luftwaffe Junkers 52/3m en route to Baghdad.
It was going to be a long, circuitous trip. They were scheduled for refuelling stops at Foggia on the Adriatic late that night, which would put them into Rhodes in the Dodecanese by late morning the next day. They would fly on to night-stop at Aleppo in Syria that evening before arriving via Mosul in Baghdad on 11 May.
Habbaniya, 8 May 1941
The Air Striking Force swung into action with renewed vigour, and raided the Iraqi air bases at Baquba, Shahraban, Khaniqin and Kirkuk, destroying twelve aircraft on the ground in fifty sorties.
At first the Air Striking Force raided in penny-packets as so few aircraft were in any fit state to fly. With the lack of shelling and working out of the sun in the hangars the mechanics began to get more and more aircraft back into flyable condition. This enabled the Air Striking Force to increase the size and frequency of the raids.
With little or no Iraqi reaction to the raids, the fighter escorts of Gladiators used up their ammunition strafing ground targets. Within a few days Iraqi air attacks on Habbaniya had fallen to two to three aircraft a day, signalling that Iraqi air power was a spent force.
However, the Air Striking Force was still taking casualties.
An Oxford reconnaissance sortie flown by Flt Sgt Harry Brattan, one of Dudgeon's few remaining qualified pilots, resulted in his death. Iraqi forces in brigade strength were known to be in Ramadi some twenty miles west of Habbaniya on the Euphrates, and a sortie was flown to see if anything had changed and whether they were preparing for an attack against Habbaniya
Even though the Iraqis were quiet, there was no point in taking risks, and Dudgeon ordered Brattan to fly at 3,000 feet, out of range of rifle and machine-gun fire. Brattan disobeyed Dudgeon's orders and came in over Ramadi at 1,500 feet. He learned the hard way when a single bullet came up through the plywood floor of the cockpit and through his left side, severing a main artery. He died within a matter of minutes.
AC Kenneth Clifton, being used as an air gunner, a big step up from operating a wind-driven winch in a target-towing Gordon, left his gun and got Brattan out of the seat. Clifton then took over the controls to bring the Oxford back.
The only problem was that Clifton had only held the controls once before, and his efforts in trying to get the Oxford down in one piece, viewed from the ground, were hair raising.
Finally landing, he attempted to taxi the aircraft in, and was saved by one of the flying instructors, Kenneth Osbourne-Young, who got aboard and had to sit in the pilot's seat full of blood in a cockpit awash with it. Although there was a pupil-pilot on board who was fully capable of flying and landing the aircraft, he was outranked by Clifton until he qualified. Clifton almost got court-martialled until sanity prevailed and he was promoted to leading aircraftman and immediately awarded the Distinguished Flying Medal. However, the DFM caused him immense trouble. The medal was only awarded to aircrew, and he was constantly stopped by officers, NCOs and the RAF police demanding to know why he was wearing the medal.
At the end of the day welcome reinforcements in the shape of five Gladiators from the Storage Unit at Ismalia in Egypt flew in with Sqn Ldr Freddie Wightman, Flt Lt Sir Robert MacRobert, Flg Off Gerard Herrtage and Sgts Bill Dunwoodie and Len Smith.
Baghdad
Starting that day, the Iraqi National Defence Council began to repeat the following statement over Radio Baghdad:
We wish to inform the public that in their joy over the victory they are spending their ammunition in vain. We wish peace to prevail in every place, and after the victory over the British, revenge shall be taken on the internal enemy, and we shall hand him over to your hands for destruction.
Citizens in Baghdad began to respond, often vengefully. A woman whose gold button inadvertently appeared through her black abba was detained for signalling the British, as was a French violin teacher accused of carrying a radio in his violin case. Patients in the Meir Eliyahu hospital were accused of signalling the British, and a mob entered the hospital, attacking staff and patients.
Despite a total blackout, crime was down. Schools were closed, businesses paralysed as banks were ordered not to make payments and collections were taken up for the Army and the wounded.
Many hostile and over-enthusiastic members of the paramilitary Futuwwah policed Baghdad and used the opportunity to exact personal revenge, and to break into private houses and beat and terrorize people.
News of the Iraqi retreat from Habbaniya reached the Embassy through the BBC broadcasts picked up by their clandestine radio, and Dr Sinderson read out the news to the refugees sitting under the palm trees on the lawn.
Food was still coming in to the Embassy, and Freya went to talk to the police at the main gate to see if she could get any local news. The Iraqis claimed to have shot down forty-five RAF aircraft for the loss of one of their own, and though friendly they refused to give any further news. Freya hoped that London was making political capital out of the fact that the Germans, who had promised aid within two days of hostilities, had still to deliver on their promise.
Later in the day an RAF Gladiator swooped down low over the Embassy and dropped a letter, confirming that all except thirty-two women and children had been safely evacuated from Habbaniya to Basra. However, an attempt later by an Audax to pick up a message from the Embassy using a hook and cable was abandoned due to intense ground fire.
Jeddah
King Abdul-Aziz Ibn Saud, a tall thin man with a great hook of a nose and drooping eyelids that belied his sharp intellect, sat on a raised chair as Naji al Suwaidi, Rashid Ali's Finance Minister, was ushered into his office in the Summer Palace at Jeddah.
Ibn Saud had little time for town Arabs who donned keffiyehs when they came to see him, and had even less time for so-called Arab nationalists, a small band of intellectuals who fought to secure their own power for their own ends rather that that of all Arabs. Although the Hashemites were no friends of the Sauds, deposing a king was a serious business. It set a bad precedent and could not be encouraged in case it spread throughout the Middle East.
'Your Majesty,' began al Suwaidi, 'I have come at the behest of Rashid Ali to seek your support to repel British aggression under the terms of the Treaty, specifically Article 4, signed between our countries in April 1931. Rashid Ali requests your help to mediate with Great Britain to call off the fighting and to send your forces to threaten Transjordan to prevent the Arab Legion being used against the Iraqi Army. Rashid Ali also is proposing a conference be held in Baghdad to make a pact between Iraq, Iran, Turkey, Russia, Germany and Italy and to consider the present situation and the future of the Arab world. He would appreciate your sending a representative to participate in the conference.'
'I have heard the request from Rashid Ali for support from Saudi Arabia', Ibn Saud began. 'As I told you during your visit in April, you should have reached a decisive agreement with Britain, avoided sedition and have no dealings with governments or individuals who seek personal advantage to the detriment of Iraqi interests', he said, taking a swipe at the Mufti, whom he viewed as a vicious zealot, and the Germans and Italians whom he mistrusted. 'Your Prime Minister chose to ignore my advice. Rashid Ali has made a big mistake in fighting Great Britain at such a critical time. If one puts one's hand in a nest of scorpions one will be stung. Any differences of opinion between Iraq and Great Britain should have been solved by peaceful means. You seem to forget that Iraq owes both its nationhood and its independence to Great Britain and, as such, is heavily indebted.
'I am a staunch friend of Great Britain and would have gone to her aid if I had sufficient arms. With the exception of Palestine, Great Britain has done nothing against Arab interests, and the present war is one of life or death for her. Our duty, if not to help Great Britain, is to be neutral. To the people of Iraq I say if you wish us to mediate between you and Great Britain we are willing to do what we can, but do not imagine for a moment that we shall take any steps which could cause a breach between us and the British. Our firm and fixed belief remains that interests of all Arabs lies in friendly cooperation with the British.
'Thank you for your visit. My warmest greetings to Rashid Ali.'
As al Suwaidi turned to leave, Ibn Saud said as an aside to his chamberlain, 'See that he gets one of our new Buick cars and make sure that he leaves quickly.'
Berlin
Ribbentrop waved a paper at Woermann when the latter entered the Reichsaussenminister's office in the Auswartiges Amt.
'Abetz, our Ambassador in Paris, has been able to get the following concessions out of Darlan. The stock of French arms under Italian control in Syria can be transported to Iraq, they will help us forward all shipments destined for Iraq, give us permission for our planes to land and take on fuel, make an airfield available specially for Iraq operations and until this is available allow us to use all airfields in Syria.'
'What is the price we have to pay', asked Woermann, always cautious of concessions.
'Surprisingly little', responded Ribbentrop. 'We allow them to rearm six destroyers and seven torpedo-boats, relax travel regulations between Vichy and the rest of France and reduce their payments covering our occupation costs. It's really a few Reichsmarks compared to the gains that we could make with Iraq in our hands. I am sending Rahn to supervise the implementation of the details.'
'Do you want me to do anything?'
'No. I have sent my personal aircraft to Paris to pick up Rahn and take him to Syria.'
As Woermann left Ribbentrop's office he wondered what was driving the Foreign Minister. Was it to get back at the British? They had ridiculed his clumsy efforts at diplomacy when he was the Ambassador in London in the mid-1930s. His attempts at mixing with the aristocracy and the élite were rebuffed as they saw nothing but a social climbing 'Champagne salesman', which had made him bitter towards the British.
Or was it his constant need to impress Hitler? Ribbentrop's shameless flattery and sycophancy and his way of always telling Hitler what he wanted to hear were the stuff of legend in the Auswartiges Amt. It was probably a combination of both, mused Woermann.
Cairo
Wavell officially assumed operational control of South-Eastern Iraq from Auchinleck, the C-in-C India. Wavell informed the new commander of the 10th Indian Division, General Edward Quinlan, that his task was to secure the Basra–Shaiba area and organize a base to receive further reinforcements.
This contrasted with Auchinleck's policy, which was more aggressive, being designed to move the Army north as quickly as possible to set up a friendly government.
Wavell's strategy was to first ensure the cooperation of local tribes before moving north from Basra, while a force from Palestine, Habforce, would advance on Baghdad from the east.
In 1941 British Intelligence in the Middle East was locked into a turf war marked by failures to communicate and coordinate. At the heart of this was a fight for influence between the War Office and the Foreign Office. The Military wanted to centralize information from the many secret and semi-secret groups operating out of Cairo and to ensure that action was taken. In contrast, the Foreign Office wanted exclusive control over all actions with foreign powers, including those at war with Britain.
Rising above the bitterness and petty squabbles were two military operations that had proved their worth. Security Intelligence Middle East (SIME), formed by Lt Col Raymund Maunsell in Cairo in 1939, was primarily involved in counter-espionage. SIME had captured virtually every Axis agent in the Middle East, and had turned many of them. Modelled after MI5, SIME behaved more like MI6, one of the cover names for the Foreign Office's Secret Intelligence Service... and did it better. Maunsell ran SIME on loose and informal lines, with little attention to military proprieties. First names were encouraged, and his senior staff called him R.J.
The other operation was Brig Dudley Clarke's 'A' Force, whose mission was deception and disinformation. 'A' Force was the real reason behind the defeat of 250,000 well-armed and supplied Italians in Cyrenaica and East Africa by Wavell's poorly equipped army of 55,000. A small neat man with sparkling eyes, his laconic, self-depracating manner, cigarette holder and love of the good life masked Clarke's brilliantly clever and imaginative mind and his photographic memory. Maunsell and Clarke had formed an alliance to share information and give assistance in the interests of 'getting the job done'.
The headquarters of British Forces Egypt, where SIME was based, was too close to the intrigue, so Maunsell met Clarke weekly in 'A' Force's headquarters in a four-storey building at 6 Shari Qasr el Nil, a few blocks from Groppi's, the famous Cairo coffee house.
The one thing that Maunsell had not expected was the fact that Clarke's headquarters was on the floor below a fashionable brothel, which Clarke had gallantly permitted to continue. Maunsell was always embarrassed when he entered the building and was greeted cordially by the brothel's ladies, but Clarke reckoned that it was perfect cover, shielding his staff and agents from prying eyes as the constant flow of people in and out of the building was easily explicable.
Today's meeting was a little different. Both had been appalled by the lack of Foreign Office interest in neutralizing the Mufti and in countering Axis and nationalist subversion in Iraq. Clarke was fully aware of the Mufti's objective. He had found an Arabic translation of an IRA booklet on fighting the British when they raided the Mufti's office in Jerusalem, following his flight to Beirut. The Foreign Office's intelligence arm, MI6, and its various cover organizations, like the Allied Services Liaison Unit (ASLU) and the Intelligence Security Liaison Department (ISLD), had done little, if anything, and the result was Rashid Ali's revolt.
Both men had guessed that Wavell had access to high-grade German intelligence, although neither had any idea that the source was ULTRA intercepts or that the Italian diplomatic cipher had been broken, enabling the British to read messages to and from Gabrielli. MI6 kept its information away from all in the Middle East save for Wavell. Wavell had told Maunsell and Clarke that he was expecting German intervention in Iraq and wanted an operation mounted to counter this.
They had invited Maj David Collins from Special Projects Operating Centre (SPOC), the sabotage section of SOE (Special Operations Executive), to discuss how they could intervene to block the Germans supporting the Iraqis.
Collins, a former oil company executive based in the Middle East and a straight-talking Londoner with a reputation for making 'things happen', said nothing for a few minutes before speaking. 'If the Germans come to the aid of Iraq their vulnerability will be their supply lines. They hold no territory adjacent to Iraq, and their nearest Axis base is Rhodes, which is six to seven hours' flying time to northern Iraq. No plane can carry a heavy load for that distance non-stop, which means that they have to refuel in either Turkey or Syria. This will require that they persuade or coerce the Turks or the Vichy French to help them. From what I gather, the Turks show little interest in getting involved either with the Germans or with us, so that leaves the Vichy French. The Vichy do not like us, and with the right pressure they will relent and allow the Germans use of their airfields, roads and railways.'
'What can we do about it?' asked Clarke.
'I do not have enough resources to sabotage every airfield in Syria, but I could take out the railway line between Aleppo and Mosul at a viaduct a few miles before Tel Kotchek, just inside Syria. If they need to ship anything heavy it will have to go by rail, and this is the only railway line.'
'Apart from sabotaging the railway line, what else could be done?' probed Maunsell.
'The one thing that the Germans cannot do is fly in enough aviation spirit to support their aircraft. Petrol is heavy, messy and dangerous to ship by air, and it would use up space that is necessary to support their aircraft, like ammunition, bombs and spare parts.'
'Yes, but Iraq is an oil producer and is awash with the stuff', rejoined Clarke.
'That may be so, but it exports crude oil and has only a small refinery in Baghdad for domestic supplies. If we could sabotage the 90-octane supplies held at the Iraqi Oil Company's storage tanks for their Air Force, we force the Germans to either fly it in or to refine Iraqi oil on the spot.'
'Why?'
'Local Iraqi oil is low-grade 80-octane, and the Germans need a minimum of 87-octane. It would be essential to improve the grade as this has an important effect on aero engine efficiency, particularly in terms of speed, manoeuvrability and climb, as well as engine life.'
'David, can you sabotage these storage tanks?' asked Clarke.
'Yes, but I will need help. Unlike Tel Kotchek, I cannot fly in, blow them up and fly out. We will have to infiltrate people into the middle of an enemy city of almost half a million, which means I will need people who can pass for Iraqis – look, talk and dress.'
Maunsell thought for a moment. 'I think I know where we can get some help. Do you have any problems in working with the Jews, David?'
'None at all, sir, I will work with anyone who is willing to fight our enemies.'
'Good. I will have to pass this by Wavell, but I expect that you will be on a flight to Lydda in the next few days.'
'What about SOE and Thornhill?' questioned Collins.
There was a general feeling among senior officers at GHQ that SOE Cairo was out of control. Although SOE had been operating in Cairo for almost a year at great expense, it had scored few successes. Slack security, inexperience, the cavalier attitude of its operatives, who swaggered around Cairo like 'Good-time Charlies' who could be found at all the up-market restaurants, bars and parties had created suspicion and hostility.
Col Thornhill, one of the leading lights of SOE, was an amiable but very indiscreet man with an insatiable need to know what was going on. Unfortunately he was often to be found propping up the Long Bar at Shepheard's – the centre of the Cairo rumour mill. This was problematic for SIME, who believed, but could never prove, that the bar-man, 'Joe the Swiss', was a German agent.
'David, SOE will receive a change-of-duties order later this afternoon assigning you to Jumbo Wilson's staff ostensibly as an engineer working on the defences in Crete. As from now you will operate out of sight here in Dudley's offices. That should take care of SOE and keep Thornhill away', replied Maunsell.
Transjordanian-Iraqi border
Wavell ordered a mechanized squadron of the Transjordan Frontier Force (TJFF), an Imperial unit with local troops and senior British officers, to link up with Habforce – the column being mobilized in Palestine
King Abdullah of Transjordan, incensed by the treatment of his brother, Abdul Ilah, by Rashid Ali, committed part of his own private army, the Desert Mechanized Regiment of the Arab Legion and its commander, Maj John Glubb, who would act as the political officer, to Habforce.
As the combined Arab force advanced on Rutba Wells from H-3, a number of Arab officers and NCOs of the TJFF made it clear that they would not cross the border into Iraq and fight their Arab brothers. They argued that they had no quarrel with Iraq, but with the British, who made others fight for them.
They were surprised, when they approached the Arab Legion to join them, to be disarmed and led back to the border, where they were dismissed.
The conspirators had not realized the big difference between the two units. The TJFF were mostly Palestinians and Syrians who were promoted on their length of service. In contrast the Arab Legion was picked from the leading families of Transjordan, largely Bedouin, with personal allegiance to Abdullah and where promotion was patriarchal.
After leaving the TJFF stranded over the border, Glubb's force of 350 troopers in Chevrolet and Ford trucks and a few odd-looking armoured cars made by a German company in Haifa, turned around and headed back towards Rutba Wells.
Habbaniya, 9 May 1941
An 05.30 take-off from Habbaniya put two 203 Squadron Blenheims over the Iraqi air base at Mosul two hours later. One of the Blenheims, flown by Flt Lt Gordon-Hall, climbed away after dropping its bomb-load and made for Habbaniya, while Sqn Ldr Pike's aircraft strafed the airfield, destroying four aircraft.
On the way back to Habbaniya, Pike saw Gordon-Hall's crashed aircraft on the ground some thirty miles south of Mosul. Radio Baghdad announced later in the day that the pilot and copilot had been captured, but the gunner had died in the crash
At the same time as the Mosul raid, two ancient Vincents of 244 Squadron carried out a dive-bombing attack on Iraqi Army barracks at Amara in the south, while Wellingtons of 37 Squadron at Shaiba attacked Iraqi airfields at Kirkuk and Mosul.
A Blenheim of 203 Squadron on a reconnaissance mission over Raschid air base saw a Savoia, three Breda 65s and a Nisr on the ground, which he strafed, claiming to have destroyed the SM79.
The Air Striking Force was in action throughout the day, making fifty-three sorties against Iraqi ground forces in the Falluja area.
Baghdad
During the morning the Mufti broadcast over Radio Baghdad, as well as those of Radio Bari and Radio Zessen, a highly inflammatory fatwa announcing a jihad against Britain, urging every Muslim to join in the struggle against the greatest foe of Islam.
In the name of Merciful and Almighty God
I invite all my Muslim brothers throughout the whole world to join in the Holy War for God, for the defence of Islam and her lands against her enemy. Faithful, obey and respond to my call.
O Muslims!
Proud Iraq has placed herself in the vanguard of this Holy Struggle, and has thrown herself against the strongest enemy of Islam, certain that God will grant her victory.
The English have tried to seize this Arab-Muslim land, but she has risen, full of dignity and pride to defend her safety, to fight for her honour and to safeguard her integrity. Iraq fights the tyranny which has always had as its aim the destruction of Islam in every land. It is the duty of all Muslims to aid Iraq in her struggle and to seek every means to fight the enemy, the traditional traitor in every age and every situation.
Whoever knows the history of the East has everywhere seen the hand of the English working to destroy the Ottoman Empire and to divide the Arab countries. British politics toward the Arab people is masked under a veil of hypocrisy. The minute she sees her chance, England squeezes the prostrate country in her imperialistic grasp, adding futile justifications. She creates discord and division within a country and while feeding it in secret openly she assumes the role of advisor and trusted friend.
The time when England could deceive the peoples of the East is passed. The Arab Nation and the Muslim people have awakened to fight British domination. The English have overthrown the Ottoman Empire, have destroyed Muslim rule in India, inciting one community against another; they stifled the Egyptian awakening, the dream of Mohammed Ali, colonizing Egypt for half a century. They took advantage of the weakening of the Ottoman Empire to stretch out their hands and use every sort of trick to take possession of many Arab countries as happened to Aden, the nine districts, the Hadramut, Oman, Muscat and the Emirates of the Persian Gulf and Transjordania.
The vivid proof of the imperialistic designs of the British is to be found in Muslim Palestine which, although promised by England to Sheriff Hussein, has had to submit to the outrageous infiltration of Jews, shameful politics designed to divide Arab-Muslim countries of Asia from those of Africa. In Palestine the English have committed unheard-of barbarisms; among others, they have profaned the el-Aqsa Mosque and have declared the most unyielding war against Islam, both in deed and in word. The Prime Minister at that time told Parliament that the world would never see peace as long as the Koran existed. What hatred against Islam is stronger than that which publicly declares the Sacred Koran an enemy of human kind? Should such sacrilege go unpunished?
After the dissolution of the Muslim Empire in India and the Ottoman Caliphate, England, adhering to the policy of Gladstone, pursued her work of destruction to Islam, depriving many Islamic states both in the East and in the West of their freedom and independence. The number of Muslims who today live under the rule of England and invoke liberation from their terrible yoke exceeds 220,000,000.
Therefore I invite you, O Brothers, to join in the War for God to preserve Islam. your independence and your lands from English aggression. I invite you to bring all your weight to bear in helping Iraq that she may throw off the shame that torments her.
O Heroic Iraq, God is with Thee, the Arab Nation and the Muslim World are solidly with Thee in Thy Holy Struggle!
Habbaniya
Roberts had a meeting with Everett and Brawn during the early afternoon to finalize the next phase of their ground strategy.
'The further we push the Iraqis back,' Roberts began, 'the safer we will be. We now hold ground as far west as Ramadi, twenty miles away, and to the north of the river a wide arc is free of Iraqi troops. Soon we will have to go forward to Baghdad to release the British and other Westerners in the British Embassy and the American Legation and get a pro-British government in place.
'The choke point is here at Falluja,' he indicated on the map spread across the trestle table, 'which is the only bridge crossing between here and Baghdad. If we hold the bridge we deny the Iraqis the ability to cross the Euphrates and we make it easier to advance on Baghdad. If the Iraqis hold the bridge they are positioned to retake Habbaniya with a counter-offensive. The RAF is already systematically bombing and machine-gunning Iraqi reinforcements, making it difficult for them to build up a strong counter-force. We are now better armed and even have our own artillery', he said, to laughter from Everett and Brawn.
'We need to get the ferry back into operation at Sin el Dhibban. This will help us to put a flying column across the Euphrates which can take Falluja from the rear. Until we are ready to attack Falluja we can start small night reconnaissance patrols across the river to check on both Iraqi strength at Saqlawiya and Falluja and suitable routes for our flying column.'
'What do you think the chances are of capturing Falluja?' asked Everett.
'I don't think that the Iraqis have a hope in Hades against us, and Falluja is a piece of cake', responded an upbeat Roberts.
Baghdad
Freya had kept herself busy at the Embassy by trying to gain news on the evolving situation with the police guards as well as helping out with tasks like a lecture on the lawn the previous evening on Aden in Wartime.
The police had been buoyed by the Mufti's call for a Holy War, and Freya could not understand why the British, seeing the effect other people managed to obtain through propaganda, persisted in thinking it quite useless for themselves.
A long conference had been held that day full of a spirit of vengeance, with plans not only to humiliate Rashid Ali's government but for the one that would replace it to be filled with pro-British politicians. Freya voiced the opinion that they should only deal with the five top people and the Mufti and any found to be working with the Germans, and then work at discrediting the Germans with the Army. Any other approach would be counter-productive, she argued, as the majority of Army officers, like the rank and file, were honest dupes of a small number of ambitious staff officers.
Rutba Wells
The fort at Rutba Wells was two hundred empty miles from anywhere. The fort, with its high windowed walls and four round towers, sat at the convergence of tracks from Amman, Damascus and Baghdad. There were no roads in 1941 except those made by the tracks of previous vehicles, which were quickly blotted out by the sand blown by the wind over the hard, red earth and the beds of flint.
The old fort had been turned into an overnight rest-house in 1927 to accommodate passengers at the Imperial Airways refuelling stop on the route to and from Haifa, Baghdad, the Gulf and India. When Imperial Airways switched to flying-boats, the mud-walled fort, with its bedrooms, lounge and restaurant, became a rest-stop for desert travellers and for the Nairn Brothers' cross-desert bus service from Damascus to Baghdad.
The fort had been seized at the beginning of May by the Iraqi Desert Police. They had been joined by Fawzi al Qawujki, the Syrian guerrilla leader. al Qawujki, a large, red-haired, scar-faced man who had been made a colonel in the Iraqi Army, responsible for irregular forces, brought forty trucks full of irregulars armed with machine-guns to support the Desert Police.
Following the attack on the British road survey party on 1 May, the Europeans working at the H-3 pumping station on the Kirkuk–Haifa pipeline had fled.
With its water source and airfield, the fort could be used as a base by German airborne forces to seize the oil pipeline. The Mechanized Regiment of the Arab Legion arriving from H-4 attempted to secure Rutba Wells. However, the Iraqis put up stiff resistance, and, armed only with Lewis and Hotchkiss machine-guns dating back to the Great War, and without artillery or mortars, the Legion retired to wait for the RAF.
The Iraqis triumphantly announced Glubb's death on Radio Baghdad that evening, which led to the premature publishing of his obituary in The Times in London the next day.
Four Blenheims from 84 Squadron operating out of Aqir in Palestine flew in to support the Legion. Each made two passes over Rutba Wells, attracting severe ground fire that damaged all four aircraft, forcing Flg Off Goudge to make a crash-landing near H-3.
Chapter Four
Shortening the Odds
Rhodes, 9 May 1941
Two weeks earlier, Dr Rudolf Rahn had been on holiday in the Obersdorff on a short break from his post at the German Embassy in Paris. He had been summoned urgently back to Paris and told by the Ambassador, Otto Abetz, that he had been selected to go immediately to Syria to help to save the Iraqi revolt.
Rahn, a young man going places, was not surprised to have been selected for the mission. He had had trouble-shooting assignments in Iran, Turkey and South America, as well as stints in the legal, oriental and press departments at the Auswartiges Amt under his belt. While he liked the ambience of war-time Paris, where he could enjoy the culture and, as a German, still get a gourmet meal, he needed a new challenge – particularly one that would advance his career.
Abetz had hinted privately that some of the Arabists at the Wilhelmstrasse had encouraged an Iraqi coup d'état, followed by a revolt, without thinking of the consequences. Iraq's military action had failed to stop the British, and the revolt was now in serious trouble
The Luftwaffe was en route with two Staffeln, and Rahn's job would be to ensure that the Vichy French in Syria complied with Admiral Darlan's directive to provide refuelling facilities and release weapons and ammunition stored by the Italian Control Commission, and transfer these to Iraq.
He had arrived just before noon at Rhodes' Gaddura airfield in Ribbentrop's personal Heinkel 111 transport to refuel on his way to Aleppo. He noticed a number of two- and three-engined Luftwaffe aircraft among a collection of Regia Aeronautica's Savoia bombers and FIAT biplane fighters crowding the grass along the single paved runway.
All the Luftwaffe aircraft seemed to be in various stages of refuelling, and there were Italian ground-crews, supervised by the aircrews, busy painting out the Balkenkreuz and the Swastika national markings, and overpainting them with what seemed to be, at a distance, green triangles outlined in red.
As they left the plane, an Italian staff car drew up alongside the Heinkel, and a young tenente stepped out and greeted them. 'Would you please join me,' he said, pointing at the Lancia, 'and I will take you to our mess, where you can wait in comfort.'
'Tenente, what are you painting on the aircraft?' asked Rahn.
'We have been requested to repaint the aircraft with Royal Iraqi Air Force markings', he replied.
Rahn and his translator, Eital Malhausen, an Austrian born in Smyrna of a French mother, were driven across to Gaddura's administration block by the helpful tenente, who ushered them into the small mess hall.
The room was empty with the exception of one table by the window, where six men were seated. One detached himself and walked over to Rahn and Malhausen. 'You must be Rahn', he said, ignoring Malhausen. 'My name is Grobba. We should talk', and he guided Rahn by his elbow to a table by the far wall, leaving Malhausen to join the others.
They both ordered coffee from a mess waiter, and sat quietly sizing each-other up while they waited.
Rahn knew that Grobba was the ranking Iraqi specialist at the Auswartiges Amt. Although he had the reputation of being obstinate and high handed, he had been placed in charge of supervising the revolt on the spot by Ribbentrop.
Conversely Grobba was well aware that Abetz was part of Ribbentrop's inner circle, a personal Foreign Office nicknamed Dienstelle Ribbentrop, which met in a private office across the Wilhemstrasse from the 'official' Auswartiges Amt. Rahn was supported by Abetz and Ribbentrop, who at Abetz's suggestion, had given Rahn plenipotentiary powers, as well as use of his personal VIP Heinkel transport.
Both realized that they had more to gain by cooperating than competing: if the Iraqis got the weapons and equipment both men would benefit. However, what the Iraqis did with those arms was out of both men's control.
'I am travelling under the name Gehreke,' began Grobba rather self-importantly, 'to confuse British spies.'
'Is that why you have repainted the aircraft in Iraqi colours?' responded Rahn.
'Oh, no, that is political. We do not want to be identified as German until our success is virtually guaranteed.'
They had just finished agreeing on how Rahn would notify Grobba of where and when the arms were to be delivered, and Rahn's acceptance of one of Grobba's radio technicians, Hontsch, when the tenente returned to announce that their planes had been refuelled.
Forty minutes later Grobba and Rahn, together with a third Heinkel, left Gaddura for the 400-mile flight to Aleppo, where they landed later that evening.
Aleppo
At Aleppo Grobba received bad news. The Iraqi Consul-General met the aircraft and advised him that the British had broken the siege of Habbaniya and seemed to be set to march on Baghdad.
Grobba decided to fly on that night to Mosul, to be in Baghdad early the next morning.
Rutba Wells, 10 May 1941
Just after noon, two Blenheims of 203 Squadron operating out of H-4's airfield, flown by Sqn Ldr Gethin and Flg Off Watson, attacked the fort at Rutba Wells, making several strafing and bombing attacks. Two columns of vehicles were seen to be engaged in a fight, but Gethin and Watson refrained from attacking as it was uncertain which of the columns was British.
Both aircraft took heavy rifle and machine-gun fire, and Gethin flew off in the direction of the column, but the Blenheim lost height and crash-landed. Watson landed alongside, and together with Gethin tried to get the other members of the aircrew out of the crashed and burning aircraft. The intense heat and the exploding ammunition and fire from some Iraqi armoured cars drove them back. They were lucky to get off the ground with the surviving Blenheim, but Gethin died of his injuries a few hours after they landed back at H-4.
One of the columns of vehicles that Gethin and Watson had seen fighting al Qawujki's machine-gun-armed trucks was a forward detachment of No. 2 Armoured Car Company RAF. On Thursday, three days before, Sqn Ldr Casano's Rolls Royce armoured cars – similar in type and vintage to those at Habbaniya – were guarding an RAF airfield one thousand miles away at Sidi Barrani in the Western Desert. Casano made an epic journey in record time to get to Iraq, arriving at Rutba Wells on Saturday.
During the night the Iraqis, shaken by the bombing and the appearance of the British armoured cars, evacuated the fort and retired to the east.
Baghdad
There was relief at the Embassy that the RAF made no low-level passes that day. The Iraqi machine-guns were now quite visible around the Embassy, and there was concern that the Iraqis would shoot down an RAF plane.
The BBC reported that the Iraqi Army was dispersed and their Air Force destroyed. But there was no confirmation that the Iraqis were in trouble. Things had settled down, and life in the Embassy was one of ease and boredom.
Freya had managed to get cigarettes for the servants, which helped to ease the problem of supplying them with a special diet. Pat Domvile told Freya that he wanted to become a servant so that she would look after him. Freya had now collected a 'family' of forty-five, including Embassy and personal servants, grooms and gardeners to look after.
Freya was invited to dine with Ken and enjoyed the opportunity to wear her evening gown and sit at a well-set table. But the conversation was stilted – not because of the situation, but because the guests were chosen in order of merit, which had nothing to do with the art of conversation. 'Keeping up appearances' in the face of adversity was maybe good for morale, mused Freya, but it was hard going.
Habbaniya, 11 May 1941
At 05.10 the Air Striking Force's Audaxes and Oxfords, escorted by two Blenhiems of 203 Squadron and the Gladiators of A Squadron from Habbaniya, attacked Raschid air base. The Audaxes, Oxfords and Blenheims bombed, while the Gladiators strafed. Several more attacks were made on Raschid that day, and formed the pattern of Habbaniya air strikes over the next three days.
Baghdad
As the three-engined Junkers taxied to the terminal at Baghdad's civil airport, Grobba could smell the city through the partly opened sliding window. For good reason the city of 1001 nights was referred to by its foreign residents as the city of 1001 smells. He remembered every city by its distinctive odour, and the smell of Baghdad, which would require an entire perfume industry to mask, brought memories of the eight years that he had spent there flooding back.
A big, fair, rugged-faced man, Grobba had arrived in Baghdad in 1932 to set up the German Legation. Gregarious by nature, and a Mason and a democrat by conviction, he and his wife had worked diligently to create a wide circle of important local dignitaries and to fit into the expatriate community, becoming friendly with many British families. Unlike most Germans, he maintained close friendships with foreign and Iraqi Jews long after Hitler's time. Although he left Masonry after Hitler's accession to power, he still gave the Masonic grip when he met a fellow Mason. His success was also a factor of his staying power and his command of Arabic. While non-Arabic-speaking British Ambassadors came and left year after year, Grobba remained solidly in place from 1932.
Grobba's forte was intrigue, and in manipulating relationships and fixing deals, leaving politics and propaganda to Dr Julius Jordan, a German archaeologist and a rabid Nazi, who worked for the Iraqi National Museum.
Grobba's career was in the ascent until he was declared person non grata in September 1939, when Iraq, under Nuri al Said's government, broke off diplomatic relations following the British declaration of war against Germany.
Twenty months at the Wilmhelmstrasse among the Junker aristocrats, the Nazis and the bureaucrats had taken its toll, and he was happy to be back in the field with an opportunity to kick-start his stalled career.
He was highly ambitious, and was determined that nothing would stand in the way of his achieving a promotion to the top ranks of the Auswartiges Amt. A victory in Iraq, he believed, would catapult him to his rightful place over the heads of the old-school network, the political appointees and the time servers.
He had survived the last two years in Baghdad walking on a very thin tightrope between Iraqi demands for arms, money and recognition and Germany's reluctance to get involved. It was a tribute to his ability to influence people that the Iraqis had requested his presence in Baghdad as the German liaison.
His assignment now was more difficult. Rashid Ali and the Golden Square needed to be motivated to continue to resist the British for the next few weeks until Germany could rearm the Iraqis and the Luftwaffe could destroy the RAF.
Jerusalem
Wavell and Jumbo Wilson flew into Lydda Airport early in the morning in a 267 Squadron Lodestar and took the waiting Humber staff car to their appointment with George Clark at Force Headquarters in the King David Hotel in Jerusalem.
Although the King David was a massive building of some seven storeys with two hundred rooms, which towered above its surroundings, it fitted in. Municipal regulations in 1918 decreed that the front of every building be at least partially made of local 'Jerusalem' limestone with its gold or ochre hue. This meant that even the largest, highest and even ugliest of buildings seemed to blend with the city, giving a semblance of conformity.
In 1939 the British had taken over the entire southern wing of the hotel, housing their communications gear in the basement and building a side entrance which linked the hotel with the Army camp at its side.
The barbed wire and the armed guards at the airport, in the escorts that had accompanied Wavell and Wilson, and at the entrance to the hotel were not there as a precaution against the Axis. They served as a reminder of the fragile peace in Palestine between the Arabs and the Jews, and the fact that the British were caught in the cross-fire between the warring sides.
Maj Gen George Clark, together with AVM John D'Albiac, who was to replace Smart at Habbaniya, were waiting for Wavell and Wilson in the large top-floor conference room which overlooked the old city and gave a magnificent view of Jerusalem and the Dome of the Rock.
Wavell, a man of few words, quickly got to the point, after the introductions had been made and the coffee served. 'Although the encirclement of Habbaniya has been broken, we still need to send a force to pre-empt Rashid Ali joining with the Germans to oust us from Iraq. As you all know,' he explained, 'we have elements of the 10th Indian Division here,' he pointed at the map, 'south at Basra. They have encountered great problems in moving north to Baghdad due to the seasonal floods. The Iraqis have added to the problem by cutting the roads and destroying rail and telegraph communications, making movement out of Basra impossible. Our only option for timely intervention is from Palestine. Time is of the essence, gentlemen. George, what is happening?'
'We have put together a relief column, Habforce, consisting of 6,000 men and almost 1,000 vehicles', Clark responded. 'Obviously this quantity of men and equipment will take time to get to Habbaniya, so we have divided up the force into two components. Joe Kingstone, the Commander of the 4th Cavalry Brigade, has formed a small advance striking-force, Kingcol, with 2,000 men, together with the Arab Legion, who are at Rutba Wells. They will make a dash for Habbaniya. The bulk of Habforce will follow on Joe's tail.'
'When will you move?' interjected Wilson.
'Joe moved out from his camp at Nathanya at first light this morning. The rest of Habforce is moving into position and will rendezvous at the Beit Lied crossroads late today.'
Wavell smiled at Clark. 'You have done a first-class job! When do you think that you will be able to relieve the siege of the Embassy in Baghdad?'
'Baghdad!' exploded Clark, 'No one said anything about Baghdad!'
'We need to get to Baghdad and the Embassy and get rid of Rashid Ali', said Wilson, raising his voice. 'Your job is to do that. Do I make myself clear?'
Visibly shaken by Wilson's outburst, Clark responded, 'Of course I will follow orders, but you all need to be aware of the risks that we face. First, it's a long way: it's four hundred and seventy miles to Habbaniya and a further fifty-five miles on to Baghdad. Most of it is across unmapped desert and we will be facing temperatures of over 100 degrees. Water and heat exhaustion will be our greatest problems. The Met Office predicts the highest temperatures in the desert for the last twenty-five years. Second it is hostile territory the whole way. John here', he indicated D'Albiac, 'cannot guarantee me fighter cover over Habforce. Our column will be thirty to fifty miles long, which is a difficult target for any aircraft to miss. There are at least two untouched Iraqi divisions and remnants of a third, plus an armoured brigade, and we have no armour.'
Wilson, now furious, began to get out of his chair. Wavell leaned across, placing his hand on Wilson's arm, and said, 'Jumbo, we need to hear from George.'
Determined to make his point, Clark went on, 'And we have men and equipment problems. The Household Cavalry, the Life Guards and the Blues, only made the move from horses to vehicles last month. They have never fought as a mechanized unit, they have had little training and many of the other ranks cannot drive. The Royal Artillery battery was unable to calibrate its 25-pounders before moving off, and the RAF Armoured Car unit with its museum-pieces, got on station only yesterday. To make ourselves mobile we have had to hire scores of resentful Jewish and Arab drivers and their trucks and buses from Jerusalem and Haifa, as well as borrowing twenty trucks and drivers from the Palestine Police and commandeering some taxis from Haifa. We have little or no modern equipment, and even our machine-guns date back to the First World War.'
Before anyone could speak, Clark held up his hand. 'That being said, we are all as keen as mustard, and morale is very high throughout Habforce.'
'Well, George, you seem to have things well under control. Keep us briefed', Wavell said, rising from his chair, so signalling that the meeting was over. 'Jumbo and I have to get back to Cairo, as that Rommel fellow is giving us a bit of trouble at Tobruk', he said, making his way to the door. As he turned the handle he paused and looked back at Clark. 'It is long odds, but I think that you will make it,'
Nathanya
At first light Kingcol had begun to move out from Brigade Headquarters of the 4th Cavalry at Nathanya, by the Mediterranean.
The previous days had been a blur of frantic activity as the various units assigned to Kingcol hunted for spare parts, traded equipment and loaded up on ordnance. No one had been briefed on the mission, but cleaning two-gallon petrol cans, marking them 'water' and filling them, as well as numerous canvas bags reminiscent of the traditional seven-pint goatskin chargules, with water suggested an imminent desert crossing.
The previous evening Brigadier John Joseph 'Joe' Kingstone, who had been acting as commander of the 1st Cavalry Division at Haifa, resumed control of the 4th Cavalry Brigade. He held a brief meeting in the wooden hut which served as the Brigade mess. to outline to his senior officers the formation of Kingcol and its mission to relieve Habbaniya. Peter Grant-Lawson, the Brigade Major, impressed on the officers that this was a fighting column crossing a desert that no conquering army had ever crossed, and Kingcol needed to be self-supporting in water, fuel, oil and ammunition. They were to take the minimum of personal equipment in their vehicles, as every nook and cranny would be needed for Kingcol's supplies.
To ensure that every one got the message, Grant-Lawson concluded with a sobering comment: 'Gentlemen, we will be travelling with twelve days of rations, five days of water and five days of fuel for the column, There are 2,000 of us in two hundred vehicles, so that is a lot of food, water and fuel to carry. This means that we will not be able to take anything along with us under any circumstances that we cannot fight with, eat, drink or drive. That means, Bobo,' to the Duke of Roxburghe, 'you can't bring your dinner jacket,' to peals of laughter, 'but it also means that you and your men can't take along the exotic pets you have collected in our stay here. You have to get rid of them tonight. Our most precious commodity is water. Officers are personally responsible for water discipline. It's one gallon per man per day and one gallon per vehicle per day. There are no exceptions. You will allow drinking only at scheduled halts.'
Kingcol in its ageing Morris 15 cwt trucks borrowed from the Scots Greys at Rehovot, together with a motley collection of RASC 3-ton trucks, commandeered taxis, buses and even some decrepit flat-bed trucks with solid tyres and little or no spare parts, driven by a bunch of rebellious civilians, together with a sprinkling of new Chevrolet staff cars, streamed out of Nathanya at dawn.
As the leading elements of Kingcol pulled out of the camp gates, Sgt Major Hugh Seekings of the Royal Military Police turned to Staff Sgt Jack Holmes, the Field Security Officer. 'Kingstone's got more nobles with him than Henry V. There's the Duke of Roxburghe, a lord, an earl, a marquesss, two viscounts, a couple of baronets and at least half a dozen knights. Oh, yes, there are also two Members of Parliament.'
'He will need all the help he can get. Where he's going the odds are far greater than they ever were at Agincourt', replied Holmes.
The road skirted the lower, northern slopes of Mount Carmel and then plunged down to the wide clay-covered plain of Esdraelon, until it crossed the eastern part of the plain to reach Beisan, where the Jordan Valley met the Plain of Jezreel. At Beisan Kingcol halted for the first of the 'twenty minutes every two hours' halts designed to minimize heat exhaustion and dehydration. The men got out of their vehicles for their first water stop, some taking advantage of the time to stretch their legs, others to buy some oranges, while a few looked out across the Jordan valley to the desert hills of Transjordan shimmering in the heat haze.
After the halt Kingcol crossed the narrow-gauge railway line of the old Hejaz Railway and turned north down the Jordan Valley, which contained the heat like a furnace. It passed over the three bridges over the swift-flowing Jordan, passing a sign which indicated that it was at sea level, and, reaching Josr Majami, began the ascent up the long, winding and steep gorge, then on the dusty road through rugged, wild and inhospitable country to Irbid. Beyond Irbid Kingcol reached a stretch of tarmac road which took it out onto the high desert plateau of Transjordan to the column's night stop at Mafraq.
An hour or so later Kingcol was joined by Maj Keith May, who brought A and D companies of the 1st Essex in requisitioned Haifa buses and two bren-gun carriers – christened Chelmsford and Colchester – carried on RASC lifting transport.
The column was dispersed over a large area, and dug in as a precautionary measure. Mafraq, a desolate place consisting of nothing but a few old huts, was just twenty-five miles south of Deraa in Vichy-controlled Syria. Deraa had a rail junction and an airfield which could be used by either the Vichy or the Germans as a base from which to attack Kingcol.
But Kingcol's luck held. At that point the Luftwaffe had not yet arrived in strength in Syria, and the Vichy French were not interested in attacking the British. Kingstone was also pleased with Kingcol's performance – they had covered 140 miles in the first day.
Beirut
Shortly before nine in the morning, Rahn and Malhausen arrived in their Heinkel from Aleppo at Beirut airport, where they were met by Jacques Guerard, Admiral Darlan's (the Vichy military leader) special representative, who was to take them to General Henri Dentz, the Vichy High Commissioner for Syria.
After the introductions Guerard took Rahn aside. 'Dentz argues about everything with everyone. His problem at the moment is the loyalty of his troops to Vichy, as the Free French in Palestine have been making overtures across the border. He is also angry at the Italian Control Commission, who have secured a large number of his weapons and who generally strut around as if they own Syria. My job is to smooth the way for you to accomplish your mission. It would help us both if you were very diplomatic with Dentz. He does not like to be reminded of the surrender, added to which he is an Alsatian. He is only pro-German in the sense that he believes that the British let down the French.'
Rahn smiled and nodded in agreement.
The Vichy administrative and military headquarters for Syria was based at the Grand Serail. The striking 430-room, three-storey building with its red roof, yellowish walls and windows encased in white limestone was built by the Ottomans in the French Romantic style in the 1820s. It was imposing, not only because of its size, but by its location, crowning the Serail hill, which gave it a sweeping view across the city to the sea, being framed against the Lebanon Mountains which rose in the background.
Guerard ushered them through the entrance and up the wide marble staircase, to Dentz's second-floor office suite located above the portico.
The meeting started off badly, with Dentz in turn being aggressive, resentful and sarcastic. However, Rahn, a skilful and accomplished diplomat, managed to get Dentz on side by impressing and charming him with his good humour and modesty.
'Whatever the recent differences between our countries, General, the supply of weapons to Iraq, for which we will pay in both gold and French francs, is a great opportunity for a new chapter in Franco-German relations, and offers the prospect of seeing the British defeated in its colony', Rahn began. 'The victory in Iraq will also clear away the plan to invade Syria by the British and Free French from Palestine It will also give us the chance to work together to throw the British out of the Middle East and North Africa for good.'
Proffering the papers he had taken out of his case, Guerard added, 'This has been thoroughly discussed with and agreed by Admiral Darlan, who has given me these written instructions to enable you to comply urgently with this request.'
'Is this a request or an order?' responded Dentz.
'The Admiral prefers to see this as a request. However, he has asked me to impress upon you the importance of the successful accomplishment of our mission to Vichy and to the Admiral in particular', answered Guerard.
'What will you do about General di Giorgis, who is responsible for the arms which are stored under his supervision by the Italian Control Commission? I have no authority over the Italians', Dentz said.
'This has already been agreed, and di Giorgis has been informed', replied Rahn.
'What weapons do you want from the arms stored under the Control Commission's supervision?' asked Dentz
'For the first shipment,' responded Rahn, 'I require the following items: 15,500 rifles and 200 machine-guns with 6 million cartridges and 900 filled belts of ammunition, four 7.5 cm field guns and 10,000 shells with all the necessary equipment and spare parts'
'I am not sure if we have that much ordnance on hand', began Dentz.
'I have the latest inventory here from the Control Commission, and this is just a fraction of what they have', Rahn said, smiling at Dentz.
Dentz sidestepped Rahn's response and asked, 'How do you plan to get these weapons to Iraq?'
'These are too heavy to send by air, so we will use the railway from Aleppo to ship the arms to Mosul', said Rahn.
'North of Aleppo the railway passes through Turkish territory parallel to the border up to the town of Kamishli, where it re-enters Syrian territory. I am sure that the Turks will not give you permission to ship the arms to Iraq through Turkey', opined Dentz.
'If you say that they are urgently needed to strengthen French forces along your northern border with Iraq, I am sure that the Turks will agree', Rahns said, again smiling at Dentz. 'The Turks do not need to know the final destination. The railway passes from Kamishli to Tel Kochek, where it intersects with the Syrian-Iraqi border and from there runs directly to Mosul and Baghdad.'
'The Franco-Turkish Convention allows the Turks five days' advance notice for military shipments on the railway, and I see no way around this', said Dentz
'I am sure that you will find a way, General', Rahn responded.
Guerard turned to Rahn, 'Would you kindly give us a few moments, M. Rahn?'
After Rahn had left the room, Guerard looked at Dentz for a few moments before speaking. 'Please do not keep trying to make this more difficult than it is. Yesterday Marshal Petain meet with Hitler and Ribbentrop at Berchtesgaden. A part of their discussion concerned Syria and Iraq. Although I do not have all the details, the Marshal agreed to four main points. Firstly to turn over three-quarters of the war materials stored in Syria. Secondly, to allow the landing of German and Italian aircraft, to provide them with fuel and make available to the Luftwaffe a special base at Aleppo. Thirdly, permitting the use of roads, ports and railways for transporting arms to Iraq, and fourthly to train in Syria Iraqi soldiers equipped with French weapons. I have here General Huntzinger's written authorization. Your Air Commander will receive today a telegram from General Bergeret confirming that he also must give every assistance to the Germans and Italians. You should think very carefully. Admiral Darlan asked me to convey to you that you personally bear the full responsibility for the suffering which non-compliance would certainly lead the Germans to inflict on France.'
They rejoined Rahn, and a now mischievous Dentz said, 'As this is a clandestine operation what name do you want to be known by?'
'Renoir', responded Rahn, to which Dentz replied sarcastically, as he wrote out his orders, 'Renouard it will be.'
'You are an unusual German, Renouard. The Iraqi coup d'état has been badly handled and the fighting even more so, and I fear that these weapon shipments will fall into the hands of the British. Nevertheless I wish you good luck.'
Rahn reluctantly declined Dentz's invitation to dinner that evening at his official residence at Les Pins, near the racecourse, as he needed to get back to Aleppo to advise Grobba of his progress over a secure radio link.
Dentz later spoke directly with Fedzi Pasha, the Turkish Military Commissioner, who, having no grounds for suspicion, immediately granted the request for trains with military stores destined for the protection of the northern Syrian border to pass through Turkish territory.
Baghdad
The Embassy was like an oven, 111ºF in the shade, with the hot Khamsin wind from the north made humid by the floodwaters.
Freya could not face a siesta in the garden with heaps of sleeping people, or in her dormitory with the seventeen women, or on her balcony, but was able to sleep on a borrowed mattress on the floor in a friend's office. It was surprising to Freya how small things like sleeping and waking, a cool breeze from the river, and the scent of flowers could give such pleasure.
Church was down to just thirty people that morning, which together with a poor service suggested that people wanted something more profound.
The police were less amiable, and the servants confirmed to Freya that this was a response to the Mufti's fatwa; they said that the police promised to massacre the Iraqi servants who stayed in the Embassy.
Later that evening, as Ken walked around the garden with Freya, he asked, 'Who do you think that we should have to deal with Anglo-Arab relations after this fiasco, Freya?'
'Ken, it would be disastrous to have the wrong man. This may well be our last chance, and our very last chance, to set good relations with Iraq on their feet once more. There is bound to be resentment, but it should not be beyond us to switch this against the Germans', Freya replied.
Aleppo, 12 May 1941
Rahn was aboard the first train which left Aleppo central station at
10.00 a.m. with 300 tons of French weapons and ammunition destined for Tel Kochek on the Syrian-Iraqi border. At noon a second train followed, carrying 160 tons of weapons and ammunition. Each train was accompanied by half a company of the 16th Tunisian Tirailleurs, to give a semblance of reality to Dentz's request to Fedzi Pasha.
Baghdad
At nine thirty in the morning, Grobba was ushered into Rashid Ali's office in the Serai Building, and found that the Mufti had already arrived and was deep in conversation with the Prime Minister.
'Dr Grobba, welcome again to my office. It seems ages since you were last here', Rashid Ali said as he stood up to greet Grobba.
'I hope that my stay this time in Iraq will be longer', the German answered as they shook hands.
The Mufti also rose from his seat and took Grobba's hand between his own two hands and said, 'It is nice to see our old friends once more.'
'It is good to be back among my friends', replied Grobba.
'Gentlemen', he began, 'I am pleased that we could meet privately before the others join us. We have a number of delicate issues to discuss. I have brought financial assistance as we had agreed', he said, walking to the open door, where his private secretary, Ahmed Jabr, and his radio operator, Emde, were waiting with two wooden boxes which he had them bring into Rashid Ali's office. He had them open the boxes on Rashid Ali's desk and then retire, closing the door behind them.
He reached into one of the boxes, handing Rashid Ali a 250 g gold bar, and then turned to Haj Amin and reached into the second box, bringing out a small stack of $50 bills. 'The boxes', he said, 'contain £10,000 in gold and $15, 000 in notes. 'This', Grobba said, pointing at the money, 'is the first instalment, and more is to follow when Dr Granow arrives from Berlin next week.
'We propose to extend to you a credit agreement of £1 million, and in addition the Italian government will assign 10 million lira to Iraq. Within the next few days advance elements of Sonderkommando Junck with Luftwaffe bombers and fighters will be arriving, followed by aircraft from Italy's Regia Aeronautica. We have also arranged to have weapons and ammunition shipped to you through Syria. The first consignment will arrive by train in Mosul in a day or two.
'However, you must realize that in exchange for this considerable assistance we require something from Iraq. I have a document here which lays out the agreement between Iraq and Germany in respect of our support during these difficult times. Let me explain what these conditions are. Firstly that Iraq transfers to Germany the concessions for the Iraqi oilfields with royalties to be paid in German commodities. Secondly that Germany has control of all Iraqi airfields for the duration of the war, and finally that Iraq recognizes the original concession granted to Germany by the Ottomans for the Baghdad–Istanbul railway.'
The growing bonhomie which had filled Rashid Ali and Haj Amin as they had listened to Grobba explain Germany's largesse gave way to silence as they took in the implications of Germany's quid pro quo.
The only noise for several minutes in the room was the two slowly revolving fans in the high-ceilinged room, as they tried in vain to create a cooling breeze from the humid air which seemed to just hang there.
Finally Haj Amin spoke. 'From my perspective I think that this is something that we must accept to achieve a greater good. However, you must realize that anything which affects our hard-won independence or places us in thrall to another country would be totally unacceptable to all Iraqis. Until we have had time to digest your document. this should remain between us. The Golden Square will see this as merely replacing the British with the Germans. While we know that not to be the case, it would serve all our interests if this were not to be discussed during our forthcoming meeting. Rashid Ali and I need to develop an approach that will not antagonize the Colonels.'
Rashid Ali, Haj Amin and Grobba moved to the adjacent conference room just before 10.00 to await the arrival of the four Colonels and the Defence Minister, Naji Shawkat.
Shawkat, together with Rashid Ali and Haj Amin, had formed the party welcoming Grobba the day before at Baghdad's civil airport. However, this was the first time that Grobba had seen al Sabbagh, Fahmi Said, Mahmud Salman and Kamil Shabib since he had been deported in September 1939.
All four Colonels looked tired and drawn, highlighting the self-imposed stresses and strains of a war which they had unleashed. Fahmi Said in particular seemed very agitated, and spent most of the meeting trying unsuccessfully to sit still.
Rashid Ali opened the meeting, welcoming Grobba profusely, outlined the course of events since the coup of 2 April, and then he suggested that each of the Colonels bring him up to date with the current battlefield situation.
Grobba had lived sufficiently long in the Middle East, and knew each of the participants well enough, to be able to pick out reality from banter and fact from enthusiastic exaggeration. While the Colonels each tried to put forward a positive position, the reality was quite different. The British from Habbaniya were still contained by the Iraqis holding the bridge at Falluja, and the floods and broken rail links in the south stopped the British from moving up to Baghdad. However, it was only a matter of time before a relief column from Palestine arrived, which would dramatically change the situation. The Royal Iraqi Air Force was no longer in any position to support the Iraqi Army, and everything depended on the Iraqis holding out for a few weeks. Grobba began to have serious doubts whether the Iraqis could hold things together long enough for the Luftwaffe and the arms from Syria to make a material difference to the outcome of the revolt.
When his turn came, he outlined both the financial and military commitments that the Germans and Italians were making, and the necessity that the Iraqis hold firm for a week to ten days until the Luftwaffe and the arms from Syria could be brought to bear.
'My feeling is that the Luftwaffe should send General Felmy as military adviser to Iraq', proposed Grobba, sensing that the Iraqis badly needed both some seasoned military advice as well as a stiffening of their resolve.
'That is precisely what the British wanted to do in connexion with the Anglo-Iraqi Treaty of 1930, which is why we are sitting here today', an annoyed al Sabbagh responded, sensitive about his own position as the de facto commander of all Iraqi forces.
'Felmy would only have the Luftwaffe under his command', Grobba countered quickly to defuse the hostility. 'Felmy is one of our ablest and youngest generals, and would be an ideal person to act as a liaison between the Iraqi and German forces.'
'Gentlemen,' Haj Amin interrupted, 'I think that this is an excellent suggestion. We get a first-class military adviser who is not involved in the day-to-day operation of the Iraqi Army so that we have all the benefits and none of the drawbacks that we experienced under the British.'
'What do we need to do to get your General Felmy here?' asked the Mufti.
'Rashid Ali should make a formal request, as postings at this level are strictly approved by the Fuhrer', Grobba said, pleased that they were making practical progress.
The meeting lasted another hour as Grobba outlined the fuel and accommodation needs of the German forces which were en route to Mosul.
The constant heat, the anxiety of not knowing what was going to happen, and the wear and tear of having to listen to Radio Baghdad blaring out depressing news for six hours up to ten o'clock in the evening, accompanied by the rantings of Yunis al Bahri on Radio Zessen, began to tell. The BBC News was no better, reporting that things were 'all quiet in Iraq', with complete forgetfulness of the Embassy's precarious situation. People became increasingly short tempered, and arguments seemed to erupt over the slightest thing.
Freya was shocked to see a rifle and three boxes of cartridges outside the Secretariat, and both of the large doors leading into the Embassy building shut and locked, with their handles tied with rope. It seemed to reinforce their desperate situation.
Freya's overriding goal today was to try to get face-powder out of the Iraqis.
Within the Embassy a number of things beside defence and food required organization. The doctors set up a surgery and dispensary capable of dealing with anything from a headache to a major operation
– although there was fortunately no need for the latter. The doctors also presided over the 'decontamination squad' which paraded every morning to wash out the bathrooms and 'Flit' all the rooms, while the 'water controller' had to cope with the fact that the Embassy drainage system was only designed to cater for fifteen people.
Mosul
During the day, advance elements of Sonderkommando Junck began landing at Mosul.
Hauptmann Harry Rother's Junkers 90 had left Gaddura at 22.00 the night before and had overflown Syria during the night as the Luftwaffe had still to receive landing permission. Rother was well aware of the need to support the incoming aircraft, and had ensured that they began by first delivering the anti-aircraft defence and the ground-crews.
Rother's aircraft carried the Flakzug's three Flak 38 mountings which made up the battery, together with their twelve-man crew and ammunition, while the accompanying Junkers 90 had brought in the thirty-man Zug of aircraft, engine, instrument and wireless mechanics and fitters, and their equipment for the Messerschmitt fighters.
Rother was appalled to see that the Royal Iraqi Air Force, despite being at war, had made no preparations for air defence. This shortcoming was highlighted by the wrecks of several of their aircraft which had been destroyed in RAF air raids.
Before leaving on his return trip to Gaddura, Rother had had the single-barrelled 20 mm Flak 38s mounted to give defensive fire across the airfield, while the ground-crews dug slit-trenches and constructed fire-control points. He had improvised around the lack of revetments, and had the ground-crews mark out random stands for the incoming aircraft, which would help to make destroying more than one aircraft at a time from the air more difficult.
During the morning all twelve Messerschmitt 110Cs of the 4th Staffel of Zerstorergeschwader 76 (4/ZG 76), the 'Haifischgruppen', under Oberleutnant Max Hobein, together with three Heinkel 111H6s of the 4th Staffel of Kampfgeschwader 4 'General Wever' (4/KG 4) led by Hauptman Schwanhauser, landed, having flown through Aleppo. The Bf 110s were weighted down. On the inside they carried a third 'crew member', one of the ground-crew, cramped up inside the long cockpit 'greenhouse', and on the outside they sported two 900-litre external fuel tanks to extend their range.
The pressure to get Sonderkommando Junck in place as quickly as possible to bolster their flagging ally meant that units had been commandeered from all over Europe. For example, 4/ZG 76 had been transferred virtually overnight from its base in Belgrade.
The transport element was composed of twenty Junkers 52/3ms and a few large four-engined Junkers 90s. All of the transports had been snatched away from the build-up at Athens Tatoi airfield in readiness for Student's airborne invasion of Crete. The transport aircraft began an immediate shuttle between Aleppo and Mosul, bringing in spare parts, ammunition and bombs.
Three of the Junkers 52s stayed at Mosul. One was completely equipped as a powerful mobile radio station to keep Junck in contact with Aleppo, Athens and Jeschonnek in Berlin. Of the other two, one was fitted out as a chemical laboratory to analyse and determine the additives necessary to bring Iraqi aviation spirit up to German standards, while the other was designed to mix the additives with the fuel.
The logistics of assembling the force at such short notice, together with delays by some of the Vichy French in Syria, either due to their reluctance to help their recent conquerors or as a result of German condescension to the defeated French, meant that Junck's forces were flying in to Mosul piecemeal.
The final three Heinkel 111s of 4/KG 4 flew into Mosul on 13 May, while two Bf 110Cs of Zerstorergeschwader 26 (ZG 26) under Lieutnant Woerner arrived on 14 May.
Wadi Tarifa
With an early start, Kingcol moved into the Great Black Lava Belt on its way to H-4. For mile after mile the column passed through a wilderness of black basalt rocks polished smooth by time and the sun's heat, and so densely packed off-road that it was almost impossible to step between them. They passed A25, a wide patch of sand which had been cleared of the black rocks to provide a landing-ground, to a point where the road bent towards the north-east and the Iraqi border.
Emerging from the Lava Belt, Kingcol made good time on the tarmac road to H-4 near the Iraqi frontier, although making 122 miles in eight hours was an exhausting journey. H-4, like all the pumping stations, was a section of some ten acres of desert fenced-in by a high perimeter wire to keep out the kleptiwallahs. Inside were the large pumping units, used to boost the flow of oil along the pipeline, as well as sleeping and administrative quarters for the crew.
Kingstone decided against a bivouac inside the H-4 compound in view of the excellent target the pumping station presented to the Luftwaffe, and stopped seven miles beyond at Wadi Tarifa to camp for the night. The next day, Kingcol would cross the frontier into Iraq and face the Iraqi Army on its run down to Habbaniya and the beleaguered Embassy in Baghdad.
Tel Kotchek, 13 May 1941
The two trains carrying the Vichy French arms and ammunition arrived at the Syrian-Iraqi border at 04.00 that morning.
To Rahn's chagrin, Grobba was not there to meet him. Even worse, the Iraqis had failed to send a locomotive to take over from the Syrian engine. Finally Dentz's right-hand man, Commandant Teze, ordered the station-master to form the two trains into one and have them drawn to Mosul by a Syrian train. The station-master, confused, called Aleppo for confirmation, and was ordered to do as he was told. The train arrived just after noon that day in Mosul.
Grobba, back from his meeting in Baghdad, met Rahn at the station. They arranged for the Iraqi officer commanding the Mosul garrison, Colonel Qasim Maqsud, to sign a three-year agreement with Rahn to pay for the arms with shipments of Iraqi grain, rice, sugar and oil to Syria. The train was scheduled to return to Syria the next day filled with grain. This was Rahn's gesture to Dentz to help him to overcome the serious shortage of bread which had created riots across Syria
Mosul
The first Luftwaffe interception took place when a Messerschitt Bf 110 of 4/ZG 76 attacked a Blenheim of 203 Squadron on a reconnaissance flight over Mosul. Although the Blenheim was attacked four times, it suffered no damage. On a bombing mission on the Mosul–Tel Awainot railway later that day, Flt Lt Plinston of 84 Squadron saw a Bf 110 in the air over Mosul, confirming that the Luftwaffe had arrived in some force.
Habbaniya
A section of the Madras Sappers and Miners, flown in from Shaiba, and guarded by a company of Levies, built a flying ferry across the Euphrates at Sin el Dhibban
Here the river was over 750 feet wide, and very fast and swollen by the winter rains in the Turkish mountains. It required 1,500 feet of wire hawser to secure the raft. A platoon of the 2nd Levy Company took up positions on the northern shore to guard the bridgehead against an Iraqi attack.
Baghdad
The previous day Freya had noticed an aircraft with strange markings flying over the city making for the civil airport. No one recognized the aircraft, although someone had suggested it might be the Persian weekly mail flight.
The BBC News depressed the Embassy. Soviet recognition of Iraq was a blow, as was the return of von Papen to Ankara. The third day of hearing that Rutba Wells was in British hands seemed to irritate everyone. Freya firmly expected to remain at the Embassy for a week to ten days, although it seemed to her to be wishful thinking.
Freya kept herself busy making out a list of cosmetics and soaps for the ladies and giving it to Vyvyan Holt, the Oriental Secretary, who negotiated with the Iraqi Ministry of Foreign Affairs for all their food and supplies.
Cornwallis had the Embassy spread Union Flags flat on the roof to warn the RAF away from mistakenly bombing their refuge.
A later institution in the Embassy was the keep-fit class held on the temporary 'bowling green', the bowls being small cannon balls for which General Waterhouse could find no other use. The keep-fit, or 'Physical Jerks' class, was held at 06.30 every morning. While the attendance was very good, their gymnastics were put in the shade by an elderly Indian, who, at the same time, used to stand on his head, without using his hands to keep him there, for a full five minutes.
Rome
Mussolini entered the war gambling that the early German victories would continue, enabling Italy to profit handsomely and cheaply. Within a year Il Duce was chaffing at his relegation to a junior partner while Hitler dictated events. To redress the balance he made the fateful decision to invade Greece. His easy victory turned into a catastrophe as the small Greek Army drove the Italians back into Albania. Hitler was forced to bail out Italy by diverting troops and aircraft, which delayed the launch of Barbarossa.
Nevertheless, Mussolini still harboured grandiose aims of making Italy a world power, and saw the opportunity of creating an Italian Imperium in Africa and the Middle East centred on oil-rich Iraq. With Rommel in Africa recapturing the whole of Cyrenaica, which Graziani had lost to a very much smaller British force, Mussolini's power of independent decision making and of acting separately from Hitler was fast eroding. Iraq was Mussolini's last throw of the dice. Mussolini was a firm believer in his own infallibility, and was buoyed by subordinates who pandered to his ego by telling him what he wanted to hear. Galeazzo Ciano, who had married Il Duce's daughter Edda, owed his position as Foreign Minister to Mussolini. Ciano, a young, inexperienced dilettante, who nursed ambitions of succeeding Mussolini, went along with Il Duce's delusions of conquest, giving personal opinion which suited the mood rather than the professional advice of his experienced diplomats.
Publically Ciano seemed to be a poor copy, slavishly imitating Mussolini's pugnacious gestures and scowling facial poses. In private Ciano was a charming, educated and intelligent man undecieved by Mussolini's pretensions of reviving the Roman empire.
Mussolini sat back in his chair in his large and ostentatious office in the Palazzo Venezia, and turned to Ribbentrop. 'Well, that covers all but one item on our agenda. What is your thinking on Iraq, Minister?' he enquired.
'Excellency,' he replied, 'the Count', gesturing to Ciano, the Italian Foreign Minister, who sat opposite, 'and I have discussed in broad terms the need to help Iraq. If a sizeable shipment of arms could reach Iraq, airborne troops could then be brought into the area. With the material on hand they could advance against the British and, in certain circumstances, attack Egypt from the East', voicing an opinion quite independently from that of Hitler. 'Iraq has to be helped. As you suggest, arms shipments and airborne forces could open up a new front against the British and start a revolt, not only of Arabs but of a great number of Muslims. Already the Mufti has summoned the Arabs of the world to a holy war against Britain.'
'The possession of Iraq and its oil wells might well have a more profound effect upon the British world position than a landing in the British Isles themselves', responded Mussolini. 'It would be necessary to get possession of Crete and Cyprus, the anterooms, if you like, to Syria. If we could get French permission to land troops and planes in Syria, our help to Iraq could be substantial. We have already prepared five planes which could proceed to Baghdad via Rhodes. They can transport 400 machine-guns as well as twenty anti-tank guns. We also have a squadron of fighters ready for action. Should the passage of arms through Turkey prove impossible, we would have to march against Britain through Syria. The great advantage here is that an attack on Egypt from Syria would require crossing a desert of only 100 km compared with the 500 km stretch of desert if we attack Egypt from the west.'
'How long, Excellency, do you think that Iraq would be able to hold out against the British?' asked Ribbentrop.
'The Prime Minister has assured us that he can hold his own against the British provided that he receives war material. If he receives no aid the British would overwhelm the Iraqis in three to four weeks', Il Duce replied. 'But will Darlan let us send men and material across Syria?'
'That has all been taken care of, Excellency. German weapons, ammunition and aircraft are arriving in Mosul as we speak.'
Mussolini, not wanting to be left out of any spoils that could be gained, turned to Ciano and said to his son-in-law, 'Please arrange for our aircraft and our weapons to be sent to Iraq immediately.'
When Ribbentrop had left, Mussolini sat with his Foreign Minister and reviewed their meeting. 'Galeazzo, we have to make sure that we have Italian soldiers and airmen in Iraq when it is freed from the British. If not, we shall give the Germans complete access to the Middle East, the oil and the Persian Gulf, and we shall become minor partners in the Axis. I do not want to have to beg for access to the oil. Make sure that you talk to Darlan immediately to secure transit and refuelling rights for the Regia Aeronautica.'
Ciano agreed and asked, 'Duce, what do you make of von Ribbentrop?'
Mussolini replied, dismissively, 'He belongs to the category of Germans who are a disaster for their country. He talks about making war right and left without naming an enemy or defining an objective.'
'The King hosted Ribbentrop at the Quirinale Palace, and could not wait until he was gone. The King regards him as a cross between a head butler and a pimp', Ciano added mischievously.
Ankara
'My dear Minister,' Franz von Papen, the German Ambassador, said, turning as Naji Shawkat, Rashid Ali's Defence Minister, entered his office, 'it is so nice to see you again.'
Von Papen had just returned from a difficult week in Berlin with Ribbentrop at a high-level Auswartiges Amt briefing on Barbarossa. The very last thing he wanted was a session of hand-holding with the Iraqi.
'I am leaving for Baghdad this afternoon, and I needed to see you before I leave. I have had several meetings with Sukru Saracoglu, the Foreign Minister, concerning Turkey's mediation on our behalf with the British.'
'I hope, Minister, that your government has not decided to capitulate', a concerned Von Papen replied.
'No, no. We see this merely as a truce, a ploy, to give us time to organize political, economic and military aid so that we can renew our fight. We are only pursuing mediation for a postponement of hostilities. We are not changing our political orientation.'
'As I have said before, Minister, I and my government feel that the Iraqi revolt was premature. You must keep pressure on the British for another two weeks until we can send to you the aid that you have been promised. You must hold the airfields and the railways, as they are the keys to German aid. According to a message that I received earlier today, two train loads of weapons and ammunition which we have arranged for you have crossed the Syrian border this morning on their way to Mosul.'
'That is most pleasing, Ambassador.'
'What do you think are the chances for a successful mediation?' questioned Von Papen.
'I am afraid that they are slim. My government will not cede any ground on the number of British troops in Iraq, and Saracoglu does not see the British accepting our position.'
H-3
Kingstone, in his staff car, a low-slung Plymouth taxi from Haifa, painted a vivid green, accompanied by an escort of two trucks full of motorized troops from the Household Cavalry and Lt Somerset de Chair, his intelligence officer, set off across the Iraqi border. They arrived at the next pumping station, H-3, ten miles over the Iraqi border, just after six o'clock in the evening.
Waiting for Kingstone at H-3 was Sqn Ldr Michael Casano, the commander of No. 2 RAF Armoured Car Company, with six of his Fordson vehicles. A tall, dashing and colourful character with a big black moustache, red-spotted scarf and hat set at a rakish angle, he was waiting for Joe at the lead armoured car. As Joe walked up he was met with a raffish salute from Casano and a growl from Casano's black and white mongrel, Butch, who had been wisely sitting in the shade.
'What happened at Rutba? Kingstone asked Casano.
'The Arab Legion got there first and was involved in a fire-fight with the Iraqi Desert Police who were holding the Fort. The Legion had to make a strategic withdrawal when Fawzi al Qawujki turned up, reinforcing Rutba with a motorized column. The RAF attacked with Blenheims, but they lost two aircraft to Iraqi ground fire. At that point we arrived and evened up the score. After an hour or so of intense fighting, Fawzi and the Desert Police withdrew to the east. '
'Fawzi's a nasty piece of work. He is a competent guerrilla leader and gave us a lot of trouble in Palestine in the 1936–39 revolt. Did you leave any troops to guard Rutba?' probed Kingstone.
'Glubb Pasha, the commander of the Arab Legion, is there with about three hundred and fifty legionnaires. I left two cars in case Fawzi tried to retake Rutba', replied Casano.
'How long will it take to get to Rutba?' Kingstone asked.
'About three hours', replied Casano. 'We have five miles of tarmac highway, then it's another forty-five miles across the desert using astral navigation.'
'Let's go to Rutba and meet up with this Glubb Pasha', said Kingstone, turning back towards his staff car.
Rutba Wells
They made good time across the desert in the cool of the night, and rolled into Rutba Wells at nine in the evening.
Joe left the escort and Casano's cars on the outskirts. He drove on through the small town with de Chair past the single-storey mud brick dwellings until they reached the eastern end, dominated by the large fort silhouetted against the starry night.
Dismounting among a jumble of Chevrolet and Ford trucks, with Lewis guns on plinths, they walked through the massive open wooden gates of the fort. Inside they were greeted by a mass of khaki-robed Bedouins with red-and-white check keffiyehs and red sashes, wearing bandoliers full of ammunition, long silver-handled daggers and .38 revolvers in their belts, and Lee-Enfield rifles slung over their shoulders. The legionnaires quickly became affectionately known as 'Glubb's Girls' by cavalry and infantrymen alike.
As Kingstone and de Chair moved through the throng, they met a short man in a khaki-drill uniform and a yellow keffiyeh with double ropes of a black silk aqal.
The man beckoned to them and led them through the courtyard, which still displayed tangled and charred wreckage from the fire-fight, down a corridor into a large hall. Judging by the refectory-style tables and a blackboard on one of the walls, with 'Roast Beef & Yorkshire pudding and custard tart' written in faded white chalk, this was the restaurant.
He sat down at one of the tables, putting his riding-boots up on the table. The hall was illuminated by several candles thrust into empty beer bottles. The flickering light showed the man to have a dent in the side of his jaw and a sandy moustache.
'Please sit down', said the man. 'My name is John Glubb and I command the Arab Legion.'
Kingstone, a tall, red-faced man, stood towering above Glubb. He did not know what to make of Glubb, with the many ribbons on his chest and the crossed swords and stars of an unknown design on his shoulder tabs.
Establishing that the wells were working, Kingstone went outside with de Chair. 'He thinks he is the King of Saudi Arabia. I am going to get him out of the way. The trouble is, I don't know if he is senior to me or not.'
de Chair replied, unhelpfully, 'With all those insignia I thought he looked like a lieutenant-general.'
'You go ahead', said Kingstone. 'I am going to have a word with Glubb.'
As Glubb reached the main gateway, Kingstone stepped back to talk to him. They looked like two lions sizing each other up to find out who would win the fight to lead the pride.
'Are you senior to me?' Kingstone asked Glubb.
'That does not arise, because I am a civilian', answered Glubb.
'Yes, but aren't you seconded from the Army', he persisted, determined to get to the bottom of this.
'No. I resigned my commission in 1925 and I am now a colonial civil servant', replied Glubb.
Both looked a little nonplussed as they parted company. Kingstone drove back to the escort camped on the outskirts, where he bedded down for the night under the desert sky, lit by a large, bright moon.
The last thing he remembered was the conversation of two troopers on guard, as they patrolled noisily around the camp. 'I have absolutely no idea why they brought us to this sodding place, do you?' said the first.
'Well I wouldn't give you tuppence for Transjordan, and as for this place I wouldn't give you a bloody penny for it', replied the second.
Rutba Wells, 14 May 1941
Starting just after 04.00, the rest of Kingcol began to trickle into Rutba Wells, with the final vehicles straggling in at around 08.00. All the vehicles were dispersed with the armoured cars on point, backed up by the 25-pounder battery with their barrels pointing outwards.
The Brigade was to move out at first light on the 15th, and the day was spent filling up with water and fuel, repairing vehicles and equipment, and checking ammunition and weapons.
In the early afternoon Kingstone held a meeting around the Brigade Office truck. As there were too many people and it was too hot to sit inside the truck, the awnings had been opened up and chairs brought outside to accommodate the participants in the shade, out of the 122ºF heat.
Kingstone, regarded as one of the best fighting brigadiers in the British Army, was an inclusive officer, and in addition to his staff of Grant-Lawson and de Chair, and the outsiders Casano and Glubb, who were now included in Kingcol, another twelve officers gathered at the Brigade Office truck for the briefing.
It was a tight squeeze under the awning. There were Lt Col Andrew Ferguson, who commanded the Life Guards and the Royal Horse Guards, Maj Keith 'Steve' May who had brought A and D Companies of 1st Battalion Essex Regiment and the two bren-gun carriers, Maj Jack 'Daddy' Wright with 237 Battery 60th Field Regiment Royal Artillery and his useful punch of 25-pounder field guns, Lt Barraclough with No. 1 Anti-Tank Troop, Royal Artillery, with 2-pounder guns to take care of enemy armour, Lt 'Fish' Ackroyd of the Light Aid Detachment to repair vehicles, Maj 'Doc' Arundel of 166 Field Ambulance to patch up the wounded, Capt Dick Shuster of the Middlesex Yeomanry's Signals Troop, Maj Newmarsh of the RASC, commanding 552 Transport Company providing heavy trucks, and Lt 'Cheeky Chappie' Oldhan with a troop of Royal Engineers from the 2nd (Cheshire) Field Squadron. Together with the non-combatants Capt Corpe, Lt John Hampton the transport officer and Maj Victor Toler-Aylward the camp commander, this represented all the combat and support elements in Kingcol.
'Tomorrow at first light we are making a dash for RAF Habbaniya, which is fifty-five miles west of Baghdad. Habbaniya has been under siege, which was lifted by the RAF a few days ago', Kingstone began. 'However, there are substantial Iraqi forces in the area which still pose a threat and have to be neutralized. We will join up with Habforce, which is following closely in our wake, and move on Baghdad to relieve our Embassy, which is under siege. We will first go east for 220 miles to Kilo 25 here,' he indicated on the map, 'which is about fourteen miles west of Ramadi. We will bivouac there overnight and then turn south. The RAF has established that there is an Iraqi Brigade well dug in at Ramadi. We are going to leave that for Habforce to finish off while we strike south to Wadi Abu Faruk. We will then turn east around the south of Lake Habbaniya, cross the Mujara bridge, here, and enter Habbaniya from the south. RAF reconnaissance shows that the land between Ramadi and the north shore of Lake Habbaniya has been deliberately flooded. The Iraqis have destroyed a long culvert west of Ramadi on the road to Habbaniya, which adds to the problem. This will necessitate Kingcol striking south. rather than east, and puts perhaps half a day on our journey.
'Now to formations. Glubb Pasha of the Arab Legion is splitting his force into three. He is leaving fifty men each to occupy H-3 and Rutba, and the remaining 250 Legionnaires in their trucks will scout ahead and form a rearguard for the column. Casano's armoured cars will be strategically placed in pairs throughout the column. The trucked infantry, the Cavalry and the Essexes will be in the vanguard, followed by the guns, with the RASC heavy trucks and the support units bringing up the rear. Kingcol will move at fifteen miles an hour with twenty vehicles to the mile.
'There are some important things that you all need to note. Fish's Light Aid Detachment will be at the end of the column. If your vehicle breaks down stay where you are and Fish's chaps will fix it. We are not going to stop or wait for anyone. We made the mistake of hiring the trucks and buses from Haifa and Jerusalem by the day, and a number of the owners are having frequent breakdowns, as this increases their charges. John,' to the Transport Manager, 'we need to stop this. Tell them that if they break down we will offload what we can and will leave them and their vehicle behind. That should put a stop to their game.
'Gentlemen, this is the desert. It is very inhospitable and waterless. It's sand and stone all the way. There is a main track but there are no reference points and no maps, so we will be using compasses to guide us. Do not get lost. We are not going to stop and go looking for you. It is now 15.00 and the temperature under our awning is 124ºF. Outside there is no shade. You are all responsible to see that you and your men wear topees or sun hats as well as shirts. I do not want to hear that anyone in Kingcol has heat stroke or has been burned by the sun. Finally the RAF has notified us that German aircraft are passing through Syria and some are already in Mosul in the north of Iraq. Our column is a big and tempting target spread out across the desert. A number of the trucks have anti-aircraft mountings, and you should attach your Lewis guns. We need to be vigilant and be able to hit back if we are attacked. I know that you all have lots to do, so I suggest that you get cracking. Corporal of Horse Barnes will be setting up mess tables here tonight, and dinner will be at 19.00 sharp.'
Although forbidden, two pets had been smuggled aboard Kingcol – Trooper Dan's rabbit and the unofficial mascot, a small monkey. The intense heat that day was too much for the rabbit. Well out of sight of Kingstone and Grant-Lawson, 'A' Squadron buried the rabbit with full military honours. The monkey, however, survived, but was not best pleased with the heat. Put down onto the sand, he would run quickly, hopping from one foot to the other, to the nearest patch of shade. Kingstone had wisely prevented Grant-Lawson from telling Casano to get rid of Butch, deciding to let 'sleeping dogs lie'. Butch and Casano were inseparable. Casano displayed that rare trait of respect, and engendered an esprit de corps that officers earn only by leading by example. 'Cass's Men', as they liked to be called, and their armoured cars, museum pieces maybe, were critical to Kingcol.
Baghdad
The record heat wave brought an announcement at the Embassy that all people should wear topees to guard against heat stroke.
Freya could feel and see that the enforced incarceration was beginning to tell, and many people, particularly those who worked outside the Embassy, were becoming impatient. Freya thought it sad that all the traits that people thought of as Nazi were surfacing each time that there was a problem and people gave vent to their anger. Luckily the Embassy seemed to have kept its 'Nazis' under control, but there was a limit to how long a spirit of amiability could be maintained.
At least the women were very pleased with the face-powder that Freya had organized.
Bet Yar
Lt Adrian Mumford sat nervously with his back against the wall, with a Smith & Wesson .38 Police Special resting on his knee under the table.
'For God's sake put that bloody gun away! This is not the Wild West', said Maj David Collins.
'Yes, but you are not armed. We are meeting against my better judgement with the Irgun, a gang of terrorists. This could be a trap.'
Collins was fed up with Mumford's petulant whining, which he had had to endure from the moment the man picked him up at Lydda airport's small terminal an hour and twenty minutes ago. 'Stop thinking of your own safety and what it would do to your career if the Irgun killed me. No one is going to do anything, although you could drop the gun and injure someone. So put that bloody gun away and look as if this meeting is important to you. If you can't do this, go and sit in the car.'
Collins regretted that he had to involve MI6, but Wavell, a stickler for form, had insisted that Maunsell go through channels. However, he had not bargained for this popinjay from the ASLU, the MI6 cover organization, in Jerusalem. It was all he could do to keep his hands off Mumford's throat.
Collins had flown in on Misr's DH 86 flight that morning from Cairo to be collected by Mumford and driven the sixteen miles to the village of Bet Yar on the coast just south of Tel Aviv. Maunsell had arranged a meeting with David Raziel, the head of the Irgun Zvai Leumi, through Ralph Harari, whose family was part of Cairo's haute Juiverie. Harari had contacted Arey Posek, the Irgun's liaison with British Intelligence, to arrange a meeting for Collins. The rendezvous had been fixed for
12.30 at the Neptune restaurant, where they were to sit inside and wait for Raziel.
Collins was not worried what Raziel, a Lithuanian-born Jew and a naturalized Palestinian, would do. Although Raziel had had some brushes with the Palestine Police, he was not anti-British, and wanted only to create a state for the Jews – by force if necessary. At the beginning of the war Raziel had argued successfully, on the principle of 'my enemy's enemy is my friend', for a suspension of all offensive action against the British in Palestine that could assist the greater enemy of the Jews – the Nazis.
Not all members of the Irgun agreed. Avraham Stern had ceded from the Irgun to found his own organization, Lehi, better known as the Stern Gang. Stern's extremism, especially the robberies of wealthy Jews, had already earned the emnity not only of the British but of many Jews.
A few minutes before the appointed time, Posek, together with a youthful, handsome man in a white, open-necked shirt and dark trousers entered the room.
'I am David Raziel', the young man said, shaking hands with both men and sitting down with his back to the door. 'There is no need for weapons', he went on. 'No one followed you from the airport and my men have secured the area. To save time, as I know that the Major will need to be back at Lydda to catch the return Misr flight at 4.30 p.m., I have ordered sea bass, which I think that you will like.'
Mumford was taken aback by Raziel's effrontery, his presumptions and his knowledge. The lieutenant was a career Foreign Office man who considered, like many of the Arabists in the FO, that the Jews were troublemakers who were upsetting the apple-cart, not only in Palestine, but throughout the Middle East. A meeting with Raziel was out of order. It looked as if the British were coming cap in hand to the terrorists. He had already voiced his concerns to Collins, who had said that he would go alone. He was also deeply concerned about the lack of an escort and Collins's veto of prior surveillance. All this just confirmed his misgivings.
After the meal Raziel sat back, opened a pack of Capstans and lit up a cigarette.
'So, Major, what can I do to help you?'
'Mr Raziel, we have a problem in Iraq, and I think that it may be to our mutual benefit to cooperate. The Luftwaffe has already flown into Mosul to reinforce Rashid Ali's revolt. Intelligence indicates that they have problems in getting adequate supplies of high-octane aviation fuel. The Iraqis have stocks of 90-octane fuel in storage tanks at the Iraqi Oil Company's facility in Baghdad. If we could sabotage these tanks it would stop the Germans from gaining a hold in Iraq. I don't need to tell you what that would mean for the Jews throughout the Middle East. We need a small team of saboteurs who can pass as Iraqis to cross our lines at Falluja, enter Baghdad and blow up these storage tanks.'
Raziel sat back, slowly sipping his black coffee to give him time to think. 'If you can fly us into Habbaniya, give us explosives and defensive weapons, we can take out these storage tanks for you. We have some Iraqi Jews who can pass as Arabs, and many of us have experience with explosives. There is, however, one thing that we want:' Raziel waited a few seconds to get everyone's attention before adding, 'a free hand with the Mufti.'
Mumford spluttered and was about to open his mouth when he saw the hostile glance from Collins signalling him to shut up.
'Obviously my government cannot and will not give you any encouragement to take the law into your own hands. However,' he added with a wry smile, 'in the confusion following your destruction of the storage tanks you might well find yourself in Zahawi Street. What happens then is, of course, none of our concern.'
'Give me two days to assemble my team and we can be at your military airfield at Tel Nor at 07.00 for a flight to Habbaniya on 17 May', responded Raziel.
All four men got up to leave the restaurant. Outside they shook hands and Collins walked across to the driver's side on Mumford's Morris 10 and asked for the key.
As Collins and Mumford's car disappeared into the distance, Posek, who had been quiet throughout the meeting, turned to Raziel. 'David. Doing a deal with the British is one thing, but going on the mission yourself is foolish. We need you here.'
'Arey, I have no choice. Stern has made me vulnerable. I have to be seen as a man of action, not merely a committee member going to endless meetings like those in the Haganah. I also have to be seen to be getting something from the British in exchange for our cooperation. By ridding ourselves of the Mufti we not only destroy a rallying-point for all Arabs but we show all Jews that the link with the British is to our advantage. If we demonstrate to the British that we can be trusted, our next move is to have them train and arm a Jewish brigade to fight the Germans, This will give us a cadre of trained, disciplined and experienced troops to form the nucleus of a Jewish Army.'
A very different conversation was taking place in the car driven by Collins.
'I just cannot believe that you will be using Jews on a sensitive mission and have given them the green light to commit murder. This is unconscionable. The British do not do this sort of thing. Where will it stop. The FO is not going to like this one little bit', whined Mumford.
With that Collins wrenched the steering wheel to the right and slammed on the brakes, stopping the car in its tracks. As Mumford picked himself off the dashboard Collins grabbed the younger man's shirt-front. 'Right, you little shit. Let's get things straight. Your behaviour in the restaurant almost ruined this mission. The Colonial Office buggered about in the 20s and 30s and even made Haj Amin Mufti. They compounded the error by not stopping him when he led the Arab revolt in the mid-1930s. Then the Foreign Office allowed the Mufti to set up shop in Baghdad and poison the Iraqis. The Foreign Office did nothing about the Mufti and his gang. We would not be in this position now if your masters in Whitehall had been doing their job. The issue is that the Foreign Office still does not want to do anything in case it upsets the Arabs. We have a problem where time is of the essence. I don't care what he is or where he comes from, but Raziel can help us to solve this problem. What they do in their spare time is none of our business. Frankly if they take out the Mufti it will save us all a lot of bother.
'Forget what you have heard about SOE. What little time I spend in Cairo does not involve propping up the bar at the Continental, eating roast pigeon at the Roof Garden, dancing the night away at the Kit Kat or playing tennis at the Gezira Club. My job is to take the fight to the enemy, and I take on the unpleasant but necessary jobs to help to win this war. This involves using and killing people. The Chiefs of the Imperial General Staff outrank the Foreign Office in a war zone and I outrank you. If you try to thwart me I will have your guts for garters. I will make it my personal business to have you transferred to the Regular Army and push a desk in the remotest posting there is in the Empire.'
With that, he let go of Mumford's shirt-front, switched the engine on, put the car into gear and let the clutch out, and swung the car back onto the road for Lydda. 'Let's see if we can get a cup of coffee at the airport before you see me off', added Collins casually, knowing that he had got his point across to a very confused and dazed Mumford.
H-4
To find out where the Luftwaffe was refuelling its aircraft en route to Mosul, 203 Squadron flew two reconnaisance sorties. The first sortie showed a Junkers 90 taking off from Palmyra, while the second identified three large and four small aircraft being refuelled at Aleppo.
Plt Off Watson, who had flown the sorties, asked Gp Capt Brown if he could go back with a fighter Blenheim and attack the aircraft. Brown referred him to the newly arrived Habforce commander, George Clark. Clark turned on Watson and barked, 'Do you want to declare war on Syria?'
Watson, unfazed by Clark's rank, responded, 'I think that would be a bloody good idea, sir', much to the delight of the General's staff.
Sense prevailed and an attack on Palmyra was ordered and timed for 17.40.
Two 84 Squadron Blenheims and two 250 Squadron Tomahawks, which had flown in for the raid from Aqir in Palestine, and the 203 Squadron fighter Blenheim took off to attack all German aircraft found at Palmyra. The British raid missed three Luftwaffe Heinkel 111s which had taken off for Mosul an hour earlier, and they were able to destroy only two German aircraft on the ground.
Baghdad, 15 May 1941
Major Axel von Blomberg was grateful to Jeschonneck for selecting him for Sonderkommando Junck. For the first time he hoped to be able to prove his own worth and come out of the shadow that is inevitably cast over your career when your father, in this case a field marshal, is a senior serving officer.
Von Blomberg was flying as a passenger on a Heinkel 111 from Mosul to Baghdad to take on the task of air liaison with Rashid Ali.
His first assignment was to set up a council-of-war meeting between Junck and the Iraqi government on 16 May.
The Heinkel pilot had deliberately flown low hoping to avoid British fighters on its way to Baghdad. The pilot brought the plane down low over the Tigris so that he could make a sweeping starboard turn just after the King Feisal Bridge to line the Heinkel up with the runway to get the aircraft as quickly as possible down into the civil airport in the western suburbs of Baghdad.
As the Heinkel approached the King Feisal Bridge, Hassan al Bassam, a corporal in the 6th Infantry Brigade, saw an unfamiliar aircraft coming head-on to the bridge he was guarding. Unable to see the Iraqi markings newly painted on the side of the Heinkel, al Bassam unslung his Lee-Enfield rifle, released the safety catch and fired three rounds into the belly of the low-flying plane as it passed directly over his position.
Only after the Heinkel had landed did the crew find Von Blomberg slumped forward in his seat, dead. One of al Bassam's rounds had passed vertically through the floor, entering Von Blomberg's throat and travelling on up into his brain. If the Major had not sat forward to see Baghdad through the small side-window, the bullet would have completely missed him.
Rutba Wells
Kingcol column left Rutba Wells at daybreak. As the vehicles moved out onto the desert they began to kick up a yellow dust cloud which created a huge funnel in their wake, stretching back almost thirty miles.
The enormity of crossing the desert to Habbaniya began to strike home to Kingstone on the road to Ramadi. It all depended on de Chair – one small man with a simple compass riding in the back of a pale yellow Chevrolet staff car at the head of the column.
The heat of the desert was staggering: a fierce, parched heat which made you feel as if you were standing on the edge of a huge fire in a high wind, being licked by gusts of flame. Those that would leave Iraq would bear the permanent impression of the country burnt into them – a cicatrix for life. Then there were the almost daily dust storms – a sudden darkness, a rustle and a shadow, a pillar of dust swirling around its base tearing across the desert, which then almost as quickly stopped. Then there were groups of 'Dust Devils' – near-vertical whirlwinds drawn up from the desert in swirling pillars that could reach up several thousand feet and then wander across the skies, and seemed to leap up and spiral around, almost at will.
No stretch of the desert excluded the fly. Iraq in May was plagued with dust, heat and flies, and the greatest scourge of all was the fly. Flies in the tents, in the hastily dug anti-aircraft trenches and in the desert were unbelievable and hard to describe. Flies settled in clouds on everything from equipment and uniforms to food. The troops never seemed to be free of flies; they were their constant companions whether they moved, sat, slept or ate.
Then there was the problem of navigation. de Chair, who was leading the brigade, was faced with a dilemma. Should he follow the camel tracks of the Old Damascus Road marked on all the old maps, or follow the cairns which marked the rough course of the beginning of the proposed Haifa–Baghdad road, which was being surveyed by Brooks & Murdoch, a civil engineering firm in Baghdad? He decided to follow the cairns.
It was a few minutes before 08.00 when de Chair climbed a rise in the ground and looked back at the column, stretching back into the distance like a long, black snake. At that point several explosions, followed by the eruption of clouds of black smoke, came from somewhere to the rear. This was followed by several streams of tracer arcing up towards three aircraft high above the column.
It was not until the stop for water at noon that de Chair heard that Kingcol had been attacked by Heinkels. One of the trucks carrying the Essexes had been hit, and there were several casualties.
The noonday halt was for two hours, to help to limit dehydration in the appalling heat. Following the military manual, the troops began trying to dig slit-trenches in case of another air raid. However, the desert at the halt contained only a thin layer of sand over rock, so most had to do with six-inch-deep trenches – useless against any air raid.
Kingcol was lucky: there were no more air raids that day.
de Chair became concerned about Kingcol's position. Driving back down the column, he found Peter Brooks, a partner in the road surveyors, who had offered to help out with his knowledge of the area.
He motioned for Brooks, who was driving sedately in his blue Packard, to pull his car out of the column and join him. Brooks, a short, quiet, grey-haired man, sauntered over to de Chair's car.
'Are we on the right route?' de Chair asked.
'Yes. You can strike left up ahead until you hit the Old Damascus Road, which is about ten miles away. The alternative is to keep going in this direction along the course of the new Haifa–Baghdad Road. You will meet an area of treacherous sand in about fifteen miles which stretches between the two roads. You should be able to cross the sand trap provided you rush it at speed. Its only sixty to a hundred feet wide, and then you connect with the foundation of the new road', said Brooks. 'Later you will come across the beginning of a seventy-mile stretch of tarmac road which leads out into the desert west of Ramadi.'
They reached the sand trap an hour later. Despite a few heart-stopping incidents, all the Kingcol vehicles managed to cross the barrier and link up with the metalled road. Picking up speed from its customary fifteen miles an hour, the column made up time along the tarmac, and finally pulled off the road twenty-five miles short of Ramadi at 21.00, with every man and vehicle up at the finish.
After dispersing the vehicles Kingstone held a conference at the Brigade truck at 21.30. 'We strike out tomorrow at 07.00 and turn south fourteen miles before Ramadi, heading towards Abu Farouk. Two companies of the Essexes will lead off under Maj May, followed by Maj Newmarsh's RASC trucks and the rest of the support units. The Household Cavalry will act as our rearguard. At Abu Farouk we turn west, skirting Lake Habbaniya and passing over the bridge at Mujara, before we turn north into the RAF station. Glubb Pasha and the Arab Legion will scout north of the column and will protect our flank.'
After almost 160 tortuous miles in intense heat, the officers and men of Kingcol were tired, dirty and impatient for a meal and a night's rest before their last day in the desert.
Baghdad
A small sensation swept through the Embassy. One of the women, nicknamed the Prima Donna, caused a scandal by strolling about in just a camisole and sleeping in the women's dormitory in the nude. One of the senior Embassy wives was forced to confront the woman as a result of the complaints from the other women. Freya was amused that there were no complaints from the men.
Cecil Hope-Gill let it slip to Freya that he was beginning to realize the full import of the phrase 'stark naked'. Freya was furious, and refused to talk to him again. Private relations in the Amusement Committee were not as happy as their public achievements. Hope-Gill usually found himself at loggerheads with the majority. The limerick from one of his colleagues sealed his fate with Freya:
Beware of Miss Poison-pen Stark,
Whose bite is far worse than her bark.
She has glutted her fill,
With the blood of Hope-Gill,
And thinks it is only a lark.
Hope-Gill later remarked that he thought that was the reason that Freya had left him out of her account of the siege.
A rumour in the market-place was that Fritz Grobba was back in Baghdad and German planes were in Mosul. The BBC News later confirmed that the RAF had bombed German planes in Syria and that Germans were indeed in Iraq.
Freya managed to buy two fish from one of the policemen in the launch guarding the Embassy from the river.
Baghdad, 16 May 1941
Despite the unfortunate death of the valuable Von Blomberg, Junck had hastily arranged a meeting with Rashid Ali and a number of key Iraqi officers to set priorities for Sonderkommando Junck and to stiffen Iraqi resolve.
Junck flew in to Baghdad's civil airport from Mosul, landing at first light in a Messerschmitt, and was met by Hauptman Erik Kohlhaas, the Abwehr's head of station in Iraq. Kohlhaas took Junck to the Maude Hotel for breakfast and to give Junck a briefing.
Junck and Kohlhaas sat under the awning in the restaurant on the first-floor balcony of the Maude. They looked out across a tranquil Tigris and a Baghdad slowly coming to life shimmering in an early-morning heat haze. Junck found it strange, sitting on the balcony, that there was a war going on and men dying less than a hundred kilometres away to the west. Kohlhass, sensing Junck's interest, gave him a quick thumb-nail sketch of Baghdad and Iraq
'To our right, across the river, is the British Embassy,' he said, pointing at a large villa-style building with tall windows and decorative pilasters, 'which is under siege by the Iraqis. The Iraqis have also laid siege to the American Legation, which is a little further down on this side of the river. The Ghazi and the Feisal Bridges were built in the last five years. Before that you could cross the river only by ferry or on a rickety wooden bridge built across boats lashed together. All the government offices, including the Ministry of Defence, where you will have your meeting, are on this side of the Tigris, about a mile upstream. Iraq is a place with a lot of history, the present is a bit of a quandary and the future is clouded.'
'Why do you say that?'
'The politics in Baghdad are as labyrinthine as its streets. Beneath the surface are all sorts of tribal and religious factions attempting to upset the status quo. To me, the only way power is exercised in Iraq is by strong military force. Brutalizing Iraqis appears to be the way that power is maintained. It seems that all that they respect is force. The more brutal, the more respect. This is not just the British and the Turks; before them it has a long history, going back beyond Gengis Khan and even Alexander the Great. Without this the minority élite, the Sunni, would lose their power to tribes and religious factions, and the country would then fall apart. Unity is not one of Iraq's strong points.'
Kohlhaas then provided enough insight on the poor showing of the Iraqi military against a much smaller and weakly armed foe, and on the lack of military skills and morale fibre among the Iraqi officers, to make Junck realize that helping the Iraqis to win would be an up-hill battle. A lack of coordination, poor command skills and little enemy intelligence would be difficult to overcome in the short term.
Kohlhaas, however, was damning in his portrayal of Grobba. 'Grobba is arrogant and he has an insulting personal style. He has deliberately distorted the facts with self-justifying and exaggerated reports to build his importance here in Iraq with the Wilhelmstrasse. He is an obsessive with all things Arab and likes to present himself as the 'German Lawrence'. He is unscrupulously ambitious, and we at the Abwehr, both here and on the Tirpitzüfer in Berlin, have found to our cost that we cannot trust him. Oberst, I suggest that you are very careful in your dealings with Grobba. He is neither a Nazi nor a Junker, but he is protected by von Ribbentrop, and it would do you no good to confront him.'
Junck entered the second-floor conference room in the Citadel at 10.00 a.m., hoping that Kohlhaas's analysis of the situation and the people was exaggerated due to some personal or professional animosity. He was to find, to his chagrin, that the Abwehr head of station's portrayal was all too accurate.
Grobba welcomed him, rather perfunctorily, and introduced him to Rashid Ali, to the Chief of Staff, General Amin Zaki, and his Chief of Operations, Colonel Nur ed Din Mahmud, and the head of what was left of the Royal Iraqi Air Force, Mahmud Salman.
To Junck's surprise, Grobba ran the meeting, building up each of the Iraqis before their presentations and commenting favourably, even when it was clear that the decisions which had been taken and the resulting performance of the Iraqi armed forces were deeply flawed.
Junck realized that the Luftwaffe would have to take the initiative, and soon, if the Iraqis were to overcome their fear of British air attacks and regain their self-confidence.
'Oberst Junck. You have heard from our Iraqi colleagues in detail. What do plan to do to help them at this critical point', Grobba asked pointedly, as if Junck was his subordinate.
Junck did not rise to the bait, and directly addressed the Iraqis. 'Gentlemen, thank you for your candour. With the British reinforcements isolated in the Basra area I believe that we have two important priorities. Our first order of business is to stop their relief column from Palestine, which is now in Iraq on its way to Habbaniya. Our second priority is to eliminate the British presence by capturing the British base at Habbaniya.
'Sonderkommando Junck will provide you with air cover, while the considerable forces at your disposal, which have not yet been committed, the 2nd Division in Kirkuk and the 3rd Division, here in Baghdad, will retake Habbaniya and stop the remnants of the Palestine column. We need to take the initiative, and I have already authorized a substantial raid on Habbaniya which is taking place as we speak. We found and bombed the British relief column in the desert near Ramadi yesterday.
'Gentlemen, to help you I need to know your positions and the situation on a daily basis. In that way we can be sure that the Luftwaffe brings you the right support, at the right place and at the right time.'
The Iraqis seemed delighted with Junck's direct approach, and agreed to the programme that they subsequently elaborated with him that morning.
An Iraqi staff car was to take Junck back to the airport, but as Junck walked through the main entrance and down the steps of the Citadel to the waiting car, Grobba caught up with him and pulled him aside.
'What do you think you are trying to do?' he said accusingly to Junck. 'These are very difficult times and we need to handle these people very carefully. Everything must be cleared by me first. You must not make commitments to them or expect them to make commitments to you. I have to coordinate everything. Do you understand, Oberst?'
Junck pointedly removed Grobba's hand from his arm. 'Firstly, you asked me what I planned to do for the Iraqis and I told them. Secondly, I am happy to work with you but I do not work for you. I report directly to General Jeshonneck. Thirdly, until General Felmy and the Military Mission arrive I make the decisions on military issues. I will be happy to discuss these in advance of meetings with the Iraqis, but my decision must be final in these matters.'
As Junck turned to go, Grobba yelled after him, 'I warn you, do not try to sabotage my mission here in Iraq. I have the Fuhrer's support!'
However, earlier that day the RAF and the Luftwaffe had been in combat with the Germans, drawing first blood. The first to attack was Flg Off Lane-Sansome of 203 Squadron, who arrived over the main airfield at Mosul at first light in a Blenheim.
Lane-Sansome saw a number of Messerschmitt fighters and Heinkel bombers, as well as some single-engined monoplanes on the ground – most probably some of the surviving Iraqi Bredas and Northrops. He dropped a few 20 lb bombs, and as he prepared to drop down to strafe the airfield at low level he found a Bf 110 trying to line up on his tail. Lane-Sansome was lucky and had enough height to pull some violent manoeuvres, managing to give the Messerschmitt the slip before flying the ninety minutes back to Habbaniya.
At 09.35 a Kette of Heinkels of 4/KG 4 led by their commander, Hauptmann Schwanhauser, raided Habbaniya. Schwanhauser split his force into three, leading one of the Heinkels into a low-level attack on the camp while the remaining aircraft swung to port to bomb aircraft parked on the ground. All three aircraft then made a second pass, bombing the hangars and causing more damage than that of the past Iraqi air raids combined, and killing a number of RAF personnel.
During the attack a lone Gladiator of 94 Squadron, piloted by Flg Off Gerald Herrtage of 94 Squadron, managed to get airborne and closed with the Heinkels as they began their run back to Mosul. Herrtage lined up on the middle Heinkel and began to pump .303 rounds at point-blank range into its port wing, causing white vapour from a ruptured fuel line to stream out behind it. However, the upper and lower gunners in all three Heinkels opened up on the Gladiator, bracketing Herrtage in a withering cross-fire. Herrtage died at the controls of the Gladiator, which then slowly banked away from the Heinkels, tipped over on its starboard wing and then rolled over and drove straight into the ground.
LAC Ernest 'Mel 'Melluish, an armourer from the Armoured Car Company, was on guard at the powerhouse. The three large diesel generators powered the whole camp and, importantly, brought water from the Euphrates, purified it and then pumped it sixty feet up to the 250,000-gallon assembly of storage tanks which dominated the skyline.
At 10.30, when the air raid sirens began to wail, Mel decided to stay out of the slit-trench which ran alongside the diesel storage tanks. As bullets began to fly and holes appeared in the blast-wall of sandbags, Mel's interpretation of guard duty changed, and he quickly jumped into the trench. He saw the Heinkels cross the sky overhead, pursued by a lone Gladiator, and as the gunfire erupted he hugged the ground face down at the bottom of the slit-trench. A slight moaning sound, followed by a tremendous blow in the small of his back, convinced Mel that he had sustained a serious injury. He decided to get out of the trench and look for medical aid. As he stood up, several things dropped off his back, and he reached down into the trench, where he found some empty cartridge cases. Feeling rather sheepish, he put the cartridge cases in his pocket and decided not to tell anyone about the 'incident'.
KG 4's Heinkels were not unscathed. Oberleutnant Karl-Heinz Graubner, flying the aircraft attacked by Herrtage, began to lose oil pressure, the temperature rose and his port engine threatened to shake itself off its mountings. After ten minutes the engine died and Graubner's crew began to throw everything out of the aircraft in a vain attempt to maintain height. After several failed attempts to restart the engine, the Heinkel had lost so much height that a crash-landing in the desert was unavoidable. The crew strapped themselves and braced themselves for the crash, while Graubner turned off the fuel cocks. The aircraft was later found wheels-up alongside the Haditha road, pointing towards Mosul. This was the one and only raid flown by KG 4's Heinkels in Iraq.
In retaliation for the raid on Habbaniya, two Blenheims of 203 Squadron, flown by Sqn Ldr Pike and Flg Off Lane-Sansome, hit the airfield at Mosul with a low-level strafing attack at dusk. Both reported to have damaged aircraft on the ground.
Later in the day four more Blenheims arrived at Habbaniya, three from 84 Squadron, and the remaining 203 Squadron aircraft from H-4.
South of Kilo 25
Thursday had been a bad night for Kingcol. After the dust, heat and flies of the day, mosquitoes and sandflies in the night completed the vicious circle. The mosquito nets were not fine enough to exclude the sandflies, which had free rein to bite their hosts unmercifully throughout the night, denying the tired troops any rest.
Friday was an equally bad day for Kingcol. At 07.00 de Chair, joined by Maj Henry Abel-Smith, took up position by a rusty iron plate on a pole which pointed west, with the words 'Ramadi 25 kilos'.
With the Arab Legion scouting ahead, de Chair estimated a compass course of 140, and added three minutes to the bearing to compensate for the magnetic deviation. He passed the bearing to Maj Newmarsh and the RASC drivers, with instructions to follow compass bearing 143 until they came to Abu Farouk, and then to head west to Mujara.
Just over two hours later the column began to flounder, and took to driving along the wadi beds, which were firmer than the ridges. However, the relief was short lived, and less than an hour later the whole column ground to a halt. The heat was now a punishing 122ºF and water was becoming scarce. They had not been able to replenish since Rutba Wells, and would have to reach Mujara before they could top up their water supply.
Men were exhausted by the heat. Adding to their discomfort was sweat, which began to soak through their clothing, making it uncomfortable. Everything not covered – metal gun-barrels, plastic steering-wheels, even leather map-cases – became too hot to touch, while those wearing spectacles and sunglasses and goggles had them steam-up.
Moving to the head of the column, Kingstone saw the problem. The big three-ton RASC supply trucks littered the desert. They had broken through the hard crust into the soft sand beneath and floundered up to their axles. A frustrated Newmarsh watched his crews digging out the sand from in front of the wheels, insert metal sand channels under them to give grip and then try to drive themselves out. They would heave themselves out, like a horse getting up, lurching from side to side in a cloud of steam and sand, only to move a few feet forward before getting embedded again.
One of the vital water-trucks had been emptied out onto the sand. Several of the trucks carrying petrol had been unloaded, and the bare, thin sheet-metal of the piles of petrol tins reflected the sun, acting like a gigantic mirror and providing a beacon for any passing enemy aircraft.
By mid-afternoon the temperature was up to 130ºF and it became impossible to work, and the crews lay exhausted under their trucks. Kingstone reluctantly admitted defeat, ordering all disabled vehicles to be left where they stood and directing the column to turn around and head back to Kilo 25.
At the Kilo 25 marker Kingstone held a brief conference early that evening when the temperature began to fall and the men were somewhat refreshed. 'I am not going to go over the events of today – we all know what happened. Luckily Maj Newmarsh has had to abandon only two of the RASC trucks, so we are still pretty much intact. The bad news is that we lost one of our precious water cargoes. We have enough supplies of water, on reduced rations, to stay here one more day. After that we have to go on to Habbaniya, or we go back to Rutba. Gentlemen, let's talk about this tomorrow after a good night's sleep.'
As the officers dispersed back to their units, Kingstone turned to Glubb, who had arrived back from shielding the column to the north. 'Tomorrow I need the Legion's help in finding us an alternative route. I know that we can get one or two trucks through, but I have to get all two hundred to Habbaniya.'
Kingcol suffered its first Turab as the sun went down, The darkening sky quickly turned ochre, blocking out the remaining sun as the dust storm hit the camp. The wind pushed dust into the eyes, ears and mouths of the troops and into the tiniest cracks and crevices in their clothes, vehicles, weapons and supplies.
Baghdad
Freya was amused by the reaction of the Iraqi police at the gate to the amount of cosmetics that had been sent for. The sergeant said that he could not think why the women needed such things as face-powder when they were all to be murdered in a few days. One of the women later said to Freya that she wondered if they should stop using lipstick, but Freya responded that at least her face was going to be in order if she was going to die.
The Embassy was beginning to look like a camp of immigrants. In the ballroom people were defending their little islands of privacy with pathetic barricades of boxes and suitcases, and two of the offices had been converted into dining areas, instead of the outdoor trestle tables, to get people indoors and out of the ferocious heat. Ominously Dr Sinderson began to vaccinate everyone in the Embassy to guard against typhoid.
Habbaniya, 17 May 1941
Just after 07.00 two 94 Squadron Gladiators looking for 'trade' over Raschid airfield saw two Bf 110s of ZG26 taking off below them.
Sgts Bill Dunwoodie and Len Smith dived behind the Messerschmitts and lined up for a quarter attack. Dunwoodie got in a few short burts, then moved into an excellent position astern and put a sustained stream of fire into the 110. The first that Unteroffizier Werner Fischer knew of the danger was a line of Dunwoodie's .303 bullets stitching their way through his wing from outboard of the starboard engine across the aircraft to the front of the long 'greenhouse' canopy. Before he could bring his rear-mounted machine-gun to bear to return fire, his pilot, Lieutnant Kai Woerner, leader of the ZG26 Rotte, slumped over, pushing the control column forward. A terrible flash followed, and the 110 became a mass of flame before disintegrating in a mid-air explosion. Almost simultaneously, Smith's target slammed into the desert two hundred feet outside the Raschid airfield perimeter.
Savile, conscious of the need to thwart a build-up of Luftwaffe aircraft so near to Habbaniya, ordered a strike on Raschid. At 09.45 six Blenheims of 84 Squadron, escorted by six Gladiators, bombed and strafed the Iraqi Air Force's main base.
To prepare for the arrival of Kingcol the bridge at Mujara had to be secured. During the morning two companies of the King's Own, some Sappers and three armoured cars, accompanied by Roberts and civilian guides, set off for a twenty-five-mile drive across the desert to the south of Lake Habbaniya.
There was no Iraqi opposition and Mujara was deserted, but there was damage to the irrigation plant and to the wooden trestle bridge. These could be repaired, and D Company and two of the armoured cars remained to provide protection for the Sappers. It was extremely hot, 125ºF, with no shade, and D Company and the Sappers had a tough time. The company was under strength, and desert sores and boils began to take their toll. The flies at Mujara were the largest and most blood-thirsty that the King's Own had encountered, and they became a byword for the hostile environment.
At noon four Hurricane fighters flew in from Aboukir in Egypt and were handed over to No. 1 Flight of A squadron. Two of the Hurricanes were IIcs. With extra tankage and four 20 mm cannon, they would give the Air Striking Force the range and punch to do some serious damage to the Luftwaffe in Mosul 220 miles away. The Gladiators had barely enough fuel to reach Mosul, and had no reserves for a dogfight, which put the Blenheim bombers at risk to the cannon-firing Bf 110s.
Savile decided to hit quickly with his new resources, and immediately arranged a long-range strike at Mosul with a mixed force of Hurricanes and Blenheims.
During the late afternoon two further companies of the Essex Regiment with their CO, Lt Col 'Crasher' Nichols, arrived by air and took over some of the duties of the King's Own, who had been in combat for over two weeks.
Mosul
Flak Oberfeldwebel Oscar Unternahrer sat smoking a pipe under the canvas awning that had been erected to keep the fierce sun off the Flak 38 anti-aircraft cannon. The afternoon heat had been intense, and all the aircraft dispersed around the airfield had their canopies and their main wheels draped in canvas to keep off the sun. On the nearest Bf 110 of ZG76, still sporting its distinctive shark's-mouth Haifischmaul nose insignia, one of the ground-crew, wearing an unwieldy canvas-covered cork tropenheim, was adding an Iraqi flag to the line of flags, which ran beneath the cockpit window, of the nations that the Geschwader had fought over or been based in – Poland, Denmark, Norway, Belgium, Holland, France, Great Britain and Yugoslavia.
On his other side Unternahrer could see a ground-crew being urged on by an energetic Erst Wart struggling to repair a Jumo engine from one of the 4/KG 4 Heinkel 111s which hung from an 'A'-frame trestle.
He had heard from one of the ground-crew that operating in this intense heat, in the open, with wind-blown sand and a lack of tropicalization, such as special oil and dust filters, was beginning to tax the mechanics, who had limited access to spare parts. Adding to the problem was the fact that the Iraqis had not supplied them with aviation spirit and modern lubricants – just rather poor-quality fuel and heavy-grade oil. This meant that the Luftwaffe was dependent on the mobile laboratory and the mixing unit aboard the Ju 52/3ms to refine the Iraqi fuel for use by Sonderkommando Junck.
The RAF raid on Mosul caught the Germans completely by surprise. The six Blenheims, preceded by two of the cannon-armed Hurricane IICs, came in fast and low from the west, putting the sun directly in the eyes of the defenders. A lucky burst from a Hurricane's cannon caught the ammunition of a Flak 38, which exploded in a fire-ball, sending 20 mm shells in all directions across the airfield and making the defenders dive for cover. The Blenheims came next, and each unloaded four 250 lb bombs on and around aircraft and buildings. By the time the two remaining Flak 38 crews had lined up on the attackers and started pumping out shells, the raiders had cleared the airfield and were disappearing to the south.
Junck came out of the slit-trench nearest the administration block, which he had commandeered, to survey the damage. While Unternahrer's gun crew had managed to take out one of the .303-armed Hurricanes which followed in the wake of the Blenheims, this was poor recompense. Sonderkommando Junck had just lost two more planes destroyed, and the ground-crews now had to contend with trying to repair four more damaged aircraft.
In just two days of fighting Junck had lost a third of his combat forces on the ground and in the air in Iraq and in Syria, and he now had only four Heinkel 111s and eight Bf 110s and two Ju 52/3ms transports left in his command. That evening he told his pilots that new tactics were to conserve aircraft, aircrew, and fuel by bombing and strafing at frequent but irregular intervals, and to come in fast from a shallow dive, make one pass and then fly away.
Kilo 25
Dawn brought a camp covered in a heavy layer of sand and dust. The troops stayed put all day at Kilo 25, brushing away the dust and digging out the wind-driven sand, which had drifted as high as the tops of the wheels of Kingcol's cars and trucks.
This delay gave the Arab Legion time to send three reconnaissance parties out into the desert, looking for ways around the soft sand to Mujara.
de Chair decided to look for water, and with Glubb and Brook's advice he back-tracked along the tarmac road to the Kilo 40 signpost, and turned south into the desert along a track which descended sharply through a small gorge and opened up to a vast lake of dried asphalt. They found pools of sulphurous water at Abu Jir which would be useful for the radiators of the column's vehicles. After a welcome swim, they headed west until they reached an abandoned oil boring site at Awasil, and then went on to reach the Euphrates at Abu Tibir.
de Chair got back to Kilo 25 to see a lone Heinkel circling the camp and dropping bombs.
Later, as he was arranging for the RASC to send out light trucks to collect the water from Abu Jir in old petrol tins, one of the Legion's reconnaissance parties returned with the news that they had discovered a route through to Habbaniya. Although it required a long detour, the whole column could follow the route in single file.
Kingstone and Glubb had gone on by car from Mujara to Habbaniya, and were flown back in a Blenheim, which landed near Kilo 25 a few minutes before 20.00, just as the light was fading. In that night's conference Kingstone announced that the next day the column would strike out for Mujara and Habbaniya.
Baghdad
From the Embassy they saw the two Messerschmitts fall over Raschid in the early morning, and at noon heard over their clandestine radio from a BBC news broadcast that the Vichy French had by 'accident' allowed some thirty German aircraft to force-land in Syria. The former evoked no cries of joy, and the latter just helped to feed the despondency of the people interned in the Embassy.
Freya refused to share the feeling of gloom, and tried to remain positive. The daily list prepared for the Iraqi Chef de Protocol for the Embassy supplies always gave Freya a lift. This time it included ten tins of Brylcream, Kotex and mothballs – hardly the order of pessimists.
Nevertheless, everything in the Embassy was now strictly rationed.
Hammond's Bund, 18 May 1941
It was a tight fit in the four-door Ford Prefect. David Raziel sat up front with Capt Tim Arnold, SOE's man at Habbaniya. Raziel's second-incommand, Yaakov Meridor, sat squashed in the back with the two Iraqis, Yaakov Aharoni and Yaakov Harazi.
They had flown in to Habbaniya from Tel Nor in a Bombay transport before noon the day before. Arnold had arranged their overnight accommodation and meals away from prying eyes in one of the bungalows recently vacated by an Assyrian family who had been evacuated to Basra. He briefed them with a new objective to reflect changing circumstances. The mission was no longer about sabotaging the Iraqi Oil Company's 90-octane fuel tanks in Baghdad It was now about gathering intelligence on Iraqi troop dispositions – numbers, units and weapons – in and around Baghdad. The British were still outnumbered by the Iraqi Army by around twenty to one, and it was vital to know who, where and what they would be facing on their way to Baghdad.
Arnold purposely did not mention Raziel's arrangement with Collins, but ensured that the group was armed with pistols and hand grenades, as well as two limpet mines. Only Mirador knew Raziel's true identity, and that the primary target would be the Mufti.
Early on the 18th they had changed into Iraqi-made clothing which RAF Intelligence had acquired from some of the Iraqi employees on the base. Raziel and Mirador, who were Ashkenazi, dressed in suits and wore fezzes to pass as Effendis, while Aharoni and Harazi, who were both Hasidim, with more swarthy complexions, dressed in robes and keffiyehs like fellaheen.
They left Habbaniya at 13.00, reaching Hammond's Bund just after two. They had driven along the top of the Bund for a mile or so until they found a small rowing-boat. Arnold stopped the Prefect, and after Raziel had pronounced the boat 'river worthy' they had all followed him down the bank to the river.
The frail boat was too small to accommodate all four of Raziel's party and their weapons, so Mirador had rowed Aharoni and Harazi across first. He then turned the small boat around to return to pick up Raziel and the weapons.
Raziel and Arnold, who were standing at the water's edge, simultaneously heard the noise of an aircraft. The engine note changed, taking on a more urgent note as it began to drop lower and head directly towards them.
Hauptman Erich Vogel and his rear gunner, Unteroffizier Joachim Elmann, had been looking for British troops and vehicles across the Euphrates opposite Falluja. Vogel had found a small troop concentration which had looked tempting. As they flew lower they had been greeted by bursts of concentrated anti-aircraft fire from guns on plinths, as well as a truck mounting a 20 mm cannon. Vogel and Ellmann had no idea that they were on the receiving end of some Iraqi ordnance which Roberts had pressed into service. The wall of shells and bullets reaching up to the low-flying Bf 110 made Vogel quickly decide that finding easier prey was the most sensible course of action, and they had moved on.
'Erich, we have something at your 9.00 o'clock.' Normally Vogel would not have bothered with a single car and a few people. However, there was no flak, and this was a quick and easy kill. He banked to the east and swung around until he faced west, and then lined up with the Bund and eased the Bf 110 lower so that the nose-mounted package of two 20 mm cannon and four 7. 9 mm machine-guns were lined up with the target.
As Vogel closed on the target, he switched on the reflector gunsight mounted on top of the main instrument coaming. As the target began to fill the sight he saw two figures scrambling up the bank towards the car and into his field of fire. He flicked off the hinged cover and pressed the red gun button. Raziel and Arnold were, perhaps, ten feet below the Prefect, which was parked on top of the Bund. Vogel's fusillade of shells and bullets 'walked' up the bank from the water's edge, slammed into both men and continued up the bank to shred the car. As Mirador reached the shore and ran up the slope towards the two mangled bodies, Vogel had already gained height and turned east looking for more 'trade'
Mirador knelt over Raziel's shattered remains and began to cry over the body of a man whom he had admired and who had held so much promise, who had died so needlessly.
Kilo 25
A patrol led by Lt the Hon. Charles Wood carried out a daybreak reconnaissance towards Ramadi, and was fired upon but suffered no casualties.This was Kincol's first contact with the enemy on the ground.
Kingcol's reveille was at 07.00, and breakfast of tinned bully-beef and biscuits, all that was left of their dwindling food rations followed half an hour later. The column got on the move just before nine, after loading the vehicles and filling the radiators and petrol tanks.
They were now down to less than half a gallon of water per man per day, if you could call it water. Sometimes it was black and at other times purple; sometimes it had to be left for ten minutes until the sediment settled. Some of the water came all the way from Egypt.
First they would go south-west, away from Mujara, to bypass the soft sand before swinging west, pointing towards Mujara. There they would cross the wooden trestle bridge before turning north to drive into the RAF cantonment.
In the lead was Glubb in a Ford 8 cwt of the Arab Legion, which had reconnoiterd the route, followed by a Morris 15 cwt full of troopers of the Blues, with Kingstone's commandeered taxi bringing up the rear of the advance party. The heavy 3-ton Bedford trucks of the RASC moved off directly behind Kingstone, followed by the towed guns, while the fighting-troops brought up the rear. All vehicles were to follow directly in the tracks of the truck in front. If a vehicle broke down for any reason it was to be left where it stood, with those in the rear passing as close to it as possible.
The word had been passed from Kingstone that there would be no stopping, even for the twenty-minute water and rest halts. This was a race to get to Habbaniya. Yesterday's visit by a Heinkel meant that the Luftwaffe had Kingcol's rough location, and today's forty-mile-long column of vehicles in single file was a very big and tempting target. All the anti-aircraft plinths were to be manned, and everyone not driving had to keep his eyes peeled for aircraft.
While Kingcol was safe from the sand barrier on the new route, travelling in single file had quickly become uncomfortable. The heavy lead vehicles threw up great clouds of dusty sand, forcing those following to use sand-goggles, sun-glasses and even spectacles to see, and handkerchiefs and scarves over their faces to breathe. Worse still, the heavier RASC trucks and the towed 25-pounder guns had scored deep ruts into the hard, packed soil, which forced the troops following in their lighter trucks and buses to lurch from side to side and porpoise up and down, straining the chassis and destroying the springs.
Leutnant Kurt Eisenach leading a Schwarm of ZG76's Messerschmitts looking for Kingcol was frustrated. Eisenach was unaware of the soft sand that Kingcol had encountered, and of the waterlogged ground along the Euphrates. He had assumed, from the Heinkel pilot's debriefing, that Kingcol had moved due south from Ramadi along the Euphrates, as this was the quickest route to Habbaniya.
The Schwarm had cruised up and down the west bank of the Euphrates between Ramadi and the top of Lake Habbaniya several times. Fuel was getting low, and Eisenach was wary of getting into dogfights with the RAF's faster Hurricanes, or even the ageing Gladiators, which were much more manoeuvrable than the 110s.
As Eisenarch neared Ramadi on what was to be their final sweep before heading back to Mosul, a shout from the rear gunner in his wingman's aircraft changed the picture. 'A dust cloud at 11.00 o'clock', yelled Obergreifreiter Weber.
Eisenach could see a large cloud of dust in the far distance, some way to the west of Ramadi. He looked quickly at his instruments before making a decision. 'We have fuel for just one pass. Circle to the east to put the sun in their eyes, then drop down low and follow me in. Kai, I want you to fly top cover in case we attract any Hurricanes or Gladiators.'
Just after 09.15, as C Squadron of the Household Cavalry left Kilo 25 to form the rearguard of Kingcol, the three 110s of ZG76 came in low, at about fifty feet, out of the sun. They began to strafe the column with their cannon and machine-guns even before they were in range. C Squadron's troopers were caught by surprise, quickly sustaining four casualties and two destroyed Morris trucks.
Ancient Hotchkiss machine-guns, 1915-vintage Lewis guns and even troopers with Lee-Enfield rifles and officers with Webley .45 pistols from all along the column opened up on the Schwarm.
In a lone Arab Legion Chevrolet truck parked directly in the path of the attackers, two Legionnaires calmly stood their ground and brought their plinth-mounted Lewis gun unhurriedly to bear on Eisenbach. As the nose with its white upper coaming began to fill the sight, the gunner, Mutr Fuqaan, began to fire controlled bursts in text-book fashion at the oncoming lead aircraft. As Eisenbach, who had been firing continuously, lined up on this new threat he ran out of 20 mm shells and fired the last remaining rounds from his 7.9 mm machine-guns. Those few rounds were enough to kill Mutr Fuqaan and seriously wound his loader, Mibrad Mohamad, in the face when he bent down to pick up a new 47-round drum of .303 ammunition for the Lewis gun.
As the rest of the column ground on towards Mujara, Doc Arundel arrived with one of 166 Field Ambulance's vehicles. He dressed the wounds of the injured Legionnaire and Cavalry troopers and, together with the dead bedou, put them on stretchers and loaded them into the ambulance. The wall of fire from the ground was so intense that the other two Messerschmitts that had followed Eisenach down and had hit the column near its centre, wounding several men and disabling two trucks and a bus, broke off their attack.
There were no further attacks, and Kingcol rolled into Mujara just after 16.00 and crossed over the wooden bridge, beneath which raced the run-off from Lake Habbaniya, and turned north for the run into the cantonment.
'Well, isn't this a pretty sight. It looks as if the Cavalry toffs have had a hard day in the desert', said Sgt Alfred Humphreys of the King's Own, as Kingcol turned onto the tarmac road which his platoon was guarding.
Lance Corporal 'Stinky' Evans added, 'It's like something out of that film, Beau Geste, we saw at the NAAFI in Karachi a few months ago.'
Humphreys walked slowly down from the ridge into the road in front of the column and raised his arm for the lead vehicles to stop. To his right, behind him on the small ridge, a group of the King's Own with Lewis guns on tripods pointing their way, backed up by a pair of RAF armoured cars on his left, guarded the southern route into Habbaniya.
An immaculate Humphreys walked slowly over to the lead vehicle, the dust-covered taxi and its grimy occupants, looked at Kingstone and snapped off a prarade-ground salute. 'Good afternoon, sir. Our dispatch rider will take you about twenty miles up this road to your camp site by the lake. It's tarmac for another eight miles, then gravel desert for twelve miles to where you camp. Then it's another fifteen miles or so on tarmac road into the cantonment.'
'Stinky' Evans, riding the BSA they used for dispatch duty, drove down the slope from the ridge, and with a rather dramatic flourish of his arm to Kingcol rode off down the road.
Humphreys saluted, then walked back up the ridge to get out of the dust while Kingcol began to file past. The men of the King's Own showed mild interest as the long column snaked past their ridge, but there was no waving or shouting. An hour later Kingcol reached its bivouac area on the shores of the lake near to the Airways Rest House, and began to form up in their defensive positions. The rough undulating gravel gave way to soft sand in some areas of the camp, and the RASC was again busy digging out their 3-ton trucks.
After the tribulations of the desert, and particularly the lack of water for washing, everyone found the bathing offered by the lake a great relief.
While the rest of the column seemed to be intent on getting into the water, Kingstone and de Chair met briefly before they each took their staff cars and set out for Habbaniya – Kingstone to talk to Roberts, and de Chair to meet with RAF Intelligence.
'The King's Own certainly did not shown any reverence at our achievement in crossing the desert to get to them. It's almost as if we are spare grooms at a wedding', said a slightly miffed de Chair.
'Somerset, you need to be more charitable. No one came to their aid, and their 1,500 or so men had to beat back over 9,000 Iraqis, which is a tremendous military feat. They have held out for another twelve days without any support. We just look like tourists to these men. So when you are in Habbaniya say little. Remember, our job is to take Baghdad, not to talk about yesterday's battle.'
Habbaniya
Arriving at the Air Headquarters block, Kingstone was shown upstairs to the large conference room. He was surprised to find George Clark, the Habforce commander, and John D'Albiac, Smart's replacement, who had flown in earlier that day, as well as a room full of Army and RAF officers.
'Hello, Joe, good to see you. You are just in time for our final planning meeting for the attack on Falluja in the morning. I'll introduce you around while we organize a beer for you. Sandwiches, I am afraid, are out of the question. Everyone here is on hard rations.'
The introductions made, and a bottle of beer on the table in front of Kingstone, Clark gestured for Roberts to begin.
'Falluja occupies a strategic position. It is on the Euphrates midway between Habbaniya and Baghdad, and home to the only bridge across the river for fifty miles in either direction. Complicating any operation against Falluja is the fact that the Iraqis have flooded the approaches to the town, limiting our approach to a single causeway. They have breached the banks of the Euphrates and they have opened the regulator of the Saqlawiya Canal, flooding the ground to their north. We have no landing craft, specialized equipment or heavy artillery support, nevertheless our job is to secure the bridge before the Iraqis blow it up and isolate us. Securing the bridge intact is vital. Falluja is less than thirty miles from Baghdad, and with the bridge in our hands we can dictate the course of events. The objective of this operation is to cut off, isolate and bombard Falluja into submission.
'Phase One will be a massive air raid on the town by Bill Savile's Air Striking Force. This is timed to start at dawn. We will call up dive-bombing attacks on specific targets as and when necessary.
'Phase Two covers the land operations. Also at daybreak, Reese, with V Column, will be air-landed with C Company of the King's Own, a machine-gun section and two Boys rifles here,' he gestured at the large wall map, 'about two miles north of Falluja. Reese will take up position covering the Falluja–Baghdad road to prevent Iraqi troops from leaving Falluja, as well as stop any forces sent from Baghdad to relieve the town.
'Tonight, A Column, under Anderson, will cross the Sin el Dhibban ferry. His company of Levies with a detachment of Sappers and Miners will capture the Police Post and the Regulator Bridge a mile west of Saqlawiya. Anderson will then move south to the high ground just north of Falluja. Going across the ferry after Anderson will be Strickland with a detachment of the 2/4 Gurka Rifles plus four of Habbaniya's captured 3.7-inch howitzers. This is S Column. Strickland's job is to protect the rear and flanks of A Column. Following on Strickland's heels will be L Column under Hilliard, with an Armoured Car section, a platoon of the King's Own, some Sappers and Miners and a lorry-borne raft. They will strike north and then west, and put a ferry across the Notch Fall Regulator.
'Graham with G Column, consisting of a platoon of Levies, six of Joe Kingstone's 25-pounders and some Sappers and Miners, are to get in front of the bridge and then take it.'
Clark, well pleased with the plan, wanted to control the operation himself, using the argument that he was now the senior land officer. But Roberts had devised the plan, had fought with most of the participants during the Habbaniya crisis and had their loyalty and infected them with his enthusiasm. Clark eventually, and wisely, relented and agreed that Roberts could continue without interference.
After the meeting over another beer, Clark got Kingstone to bring him up to date with Kingcol's dash across the desert. Just before he left, Kingstone asked Clark what he thought of Roberts.
'Although he is forceful and intolerant of inefficiency, everyone who has fought with him – The King's Own, the Levies and the RAF armoured cars – will follow him anywhere. The other thing is that he is unfailingly confident, and those under his command say that they always seem to get a laugh out of him, even under the worst conditions.'
'Not a bad man to have around', mused Kingstone.
However, D'Albiac was extremely unhappy with Roberts's plan, which involved integrated air-land operations. D'Albiac believed, as did many senior RAF officers, that the services were designed to operate independently, and Roberts's plan set a very dangerous precedent. D'Albiac had not been present during the early days of the Habbaniya siege, and was unaware of the tremendous psychological effect of continuous heavy bombing, backed up by aggressive ground action, which enabled a far smaller force to prevail. Combined Operations, which were to become the norm in the Western Desert within eighteen months, was still in its infancy in early 1941. While Roberts had seen the opportunity and witnessed the result, D'Albiac was furious that Army commanders could call on his resources seemingly at will, deflecting them away from what he believed was the RAF's real job – defeating the Luftwaffe. Although D'Albiac did not fight Roberts's plan, he made sure that Tedder in Cairo was aware of his misgivings.
While Kingcol bivouacked at the lake, the Levies, the King's Own and the Gurkhas began to move out of Habbaniya to take up their positions for the attack on Felluja the following day. Leaving at 17.00, it took some two and a half hours to get A, S and L Columns across the Euphrates on the Sin el Dhibban ferry with their mules, armoured cars and captured Iraqi artillery and a few trucks.
At 19.00 G Column under Alistair Graham left Habbaniya to set up positions on a bluff overlooking Canal Turn, while Advanced Brigade and Battalion Headquarters were set up in a well-camouflaged position in direct contact with both companies of the Levies and with Habbaniya. However, this took all night, as the Levies had first to cross the gap blown in Hammond's Bund by the Iraqis, using a combination of rafts and sailing-boats manhandled from the RAF Yacht Club on Lake Habbaniya to the river.
Only half of the Levy column arrived, and Graham had to return and extricate the remaining vehicles, which had been driven into the ditch in the darkness. The river crossing was made in absolute silence. Despite several water-logged boats, the Levies managed to get all their weapons and ammunition across the flooded ground and onto dry land. Only two incidents marred the crossing. One was a Gunner officer and his signaller who upset their boat when trying to lay a cable across the breach in Hammond's Bund, and had to be saved from drowning. The other was the ten mules who lived up to their reputation: having gone halfway across the water they panicked and returned to the shore, despite furious efforts by their handlers to stop them.
Shortly before dawn the whole company was in position at Palm Grove, overlooking Falluja, which lay some two thousand yards to the east.
Chapter Five
The Road to Baghdad
Falluja, 19 May 1941
From first light, the Air Striking Force began to soften up Falluja, some fifteen miles east of Habbaniya, with a fifty-seven-aircraft strike against Iraqi ground forces in the town. This time 4 Service Flying Training School's Audaxes, Gordons and Oxfords were supplemented by the heavier bomb-loads of 84 and 203 Squadrons' Blenheims.
Simultaneously Lt David Reese's C Company of the King's Own was flown in four of the Habbaniya Communications Flight's ancient Valentias, and landed two miles to the north-east of Falluja to cut the Baghdad–Falluja road. The airborne operation went well, with each aircraft remaining on the ground for less than sixty seconds, speaking well for Reese, who had drilled his company assiduously in getting themselves, weapons and equipment out of the Valentias in the minimum possible time.
To prevent the Iraqis summoning reinforcements, the Air Striking Force had been tasked to cut the telephone lines from Falluja. There were two telephone lines. One, with a few wires which led across the cultivated ground, was easily destroyed by an Audax that flew backwards and forwards just above the ground between the telephone poles for about a mile, using its propeller blades to cut the wire. The other telephone line, with a large number of wires, ran across the desert behind Falluja. There were too many wires to use the propeller, and the Audaxes had little success in bombing the poles.
A frustrated pilot officer, Martin Bennett, flew back to Habbaniya. While his Audax was being rearmed and refuelled, he walked over to the workshop at the back of Hangar Two and borrowed a pair of metal shears and an axe. Bennett flew the Audax back to the line of telephone poles which snaked from Falluja towards Baghdad, landed and taxied in between two of the poles. While his gunner gave him covering fire, Bennett calmly stood up on the main plane and cut the telephone wires with the pair of metal shears. To make certain the Iraqis could not repair their landline communications, he then took an axe and unhurriedly cut the telegraph poles on either side of the Audax before returning to the plane to take off and fly back to Habbaniya.
However, Falluja was holding out, and Clark began to worry. By mid-morning he started to pester Roberts to throw more troops into the battle. Roberts steadfastly resisted, as he wanted at all costs to avoid a pitched house-to-house battle with his limited number of troops and the large number of unnecessary casualties that he would take.
Following a lull, Roberts ordered an intensive dive-bombing attack concentrated on the eastern edge of the bridge just before three in the afternoon. After the ten-minute air strike, the same area was subjected to a thirty-five-minute bombardment by Kingcol's 25-pounder guns which had been manhandled across Hammond's Bund the previous night.
Crouched in a filthy ditch fifty feet from the western edge of the bridge, Capt Alistair Graham scanned the eastern edge of the bridge through his binoculars. Fires from the bombing were burning all over the town, and a pall of grey smoke drifted across the face of the buildings opposite, obscuring them and the far end of the bridge.
Graham turned to his Levy lieutenants, RAB100 Baijan Peeku and Khamsi Lazar Oshana: 'If we can't see them, they can't see us. We will use the armoured cars as shields and rush the bridge as they give us covering fire. Pass the word to fix bayonets, and after the count of three follow me. One, Two, Three... '
As Graham pounded across the bridge with the Levies at his heels and Iraqi bullets buzzing past his head, the other end of the 177-foot bridge seemed an awfully long way away. When the Levies reached the end of the bridge they were met by a party of Iraqis with a white flag, who surrendered the whole of the Iraqi force in Falluja.
The bridge had been captured without incurring a single casualty, and a large quantity of rifles, machine-guns and ammunition which had been abandoned by the Iraqis was shared out among the Levies and the King's Own. After establishing a bridgehead, the Levies were reinforced by Capt Cottingham, who brought two mortars and a machine-gun section. A little later they were all amused to see Alastair Brawn riding into view on a large white mule which he had used to cross the water.
By 17.00 the town was secure. Although the three northern columns, A, S and L, took no part in the fighting due to numerous delays and difficulties in crossing canals, Roberts's plan had worked well beyond everyone's expectations. The softening-up process, with 134 sorties dropping over ten tons of bombs, had so demoralized the Iraqis that three hundred prisoners, including twenty-seven officers, were captured with a remarkable lack of British casualties.
Col Everett of the King's Own arrived with D Company from Mujara to take command, and the air-landed troops of C Company to the north moved into the town to take up defensive positions, while a Levy platoon under Capt Armstrong returned to Habbaniya with the Iraqi prisoners.
Falluja was badly damaged. There was no water or power and no glass in the windows. Most of the inhabitants had fled, and those who remained were formed into burial-parties and working-parties to help to clean up the town. The intense heat, mosquito infestation and a lack of mosquito nets, coupled with battle rations and continuing offensive actions against pockets of local resistance, made life difficult for the King's Own.
Mosul
Cutting the telephone wires had prevented any news of the battle reaching Baghdad, and it was not until early evening that an Iraqi captain managed to phone the Citadel with the news from Khan Nuqta.
Junck, who had already been forewarned of the Iraqi defeat at Falluja by Kohlhaas, was more than ready for Grobba's call later that evening.
'Junck, where was the Luftwaffe? The Iraqis have sustained a major set-back around Falluja and they need immediate support.'
'Get one thing straight, Grobba. At our meeting with the Iraqis I told you that you must keep me informed of the Iraqis' military plans. This you have conspicuously failed to do, and Falluja is the result of your lack of communication. We have no radio messages in our log from you about the Iraqi attack on Falluja, and I am not a mind reader. Let us finish with arguing who is at fault and get down to how we can help them. I will have my fighters and bombers all over Falluja and Habbaniya tomorrow from daybreak. While we take on the British in the air and on the ground, you have the Iraqis mount a counter-attack. We must resecure that bridge at all costs, as it is the only way to and from Habbaniya and Baghdad. Whoever controls that bridge controls the outcome of the revolt. Remember, I am counting on you to convince Rashid Ali and the Colonels that they must give this counter-attack all they have.'
With that, Junck switched off the radio before a stunned Grobba could reply, and got back to the serious business of planning his operations for the following day. Truth be told, Junck was a very worried commander. Not only had he begun to take significant losses from enemy action and suffer from a lack of spare parts, but he had received a signal from Athens, warning of the recall of all his transports. Operation Merkur, Student's airborne invasion of Crete, was to begin the next day, and Student had commandeered all transport aircraft in the region.
Baghdad
News reaching the Embassy indicates that no one really knows what is happening here, mused Freya. The water supply was turned off, and although it was said to be so throughout Baghdad, Freya thought that it might be just a ruse to pressure the Embassy. She made sure that the cawasses were fed, although she began to worry that with the war growing closer to Baghdad they might soon run out of food. There was still no answer from the Iraqi government to their protests that the cawasses were being threatened with death.
Although Rashid Ali had gone to von Blomberg's funeral, there was still no official admission by the Iraqis of the presence of Germans in Iraq.
Falluja, 20 May 1941
Just after first light, three Gladiators of A Flight encountered a Messerschmitt ground-attacking British troops around Falluja, and drove it off. A little later, two Hurricanes on patrol west of Ramadi chased away four 110s, trying to sneak into Habbaniya from the west. Both engagements resulted in no wins or losses for either side.
Flt Sgt Smith was a lucky man. He was one of four Gladiators flying top cover over an Audax which had force-landed in the desert. The Gladiators were attacked by five 110s of ZG76. The Messerschmitts concentrated their fire on Smith's aircraft, and he only just managed to escape after his Gladiator had suffered severe damage. Leutnant Martin Drewes mistakenly claimed Smith as his first victory.
The Hurricanes scrambled seven more times that day. Two Hurricanes chased another 110 strafing troops around Falluja and Hammond's Bund. The latter attack almost killed Everett, RAB200 Yacob Khoshaba Aboona and Maj David Rose, who were machine-gunned by the 110, which wounded their driver in the head and riddled their car. Two other Hurricanes found a Kette of Heinkels of KG4. The Hurricanes all reported damaging their opponents
By the end of the day A and B Companies of the King's Own, who had been in or near Habbaniya, began to drift into Falluja in penny packets by launch, truck, swimming and walking. One of the immediate tasks should have been a thorough house-to-house search for hidden Iraqi soldiers, weapons and potential snipers. However, with the very small number of troops available, this was out of the question. It was not even possible to prevent refugees from returning to Falluja at night – some of whom were undoubtably Iraqi soldiers – a good many of whom were dressed as women wearing abbas.
Communications were tenuous. Everett had one radio, an ancient heliograph and a mirror.
Habbaniya
While the Hurricanes were away, the Luftwaffe played. Six 110s strafed Habbaniya, destroying a Blenheim of 84 Squadron, a DC-2 and two Valentias of 31 Squadron, and damaging two more Blenheims. They left several fires belching columns of black smoke and several dead and wounded in their wake
In the late afternoon six Blenheims with four Gladiators attacked Raschid air base, with two Hurricanes flying top cover, in a bid to destroy all the airfield's large hangars to deny the Germans their use.
Baghdad
The largest meeting room in the Citadel was on the ground floor. It was windowless and in the centre of the building to provide a cool, safe working environment. However, the notable chill in the air had nothing to do with the location of the room.
Ostensibly General Amin Zaki, the Chief of Staff of the Iraqi Army, presided over the meeting. However, Grobba's earlier stormy meeting with the General had set the tone of the conference. Grobba was incensed that the 3rd Infantry Brigade defending Falluja had failed to blow the bridge as planned. Privately he had told Zaki that if there was not an immediate Iraqi counter-attack the Luftwaffe would withdraw.
al Sabbagh, as C-in-C of Western Command, introduced Colonel Nur al Din Mahmud, Director of Military Operations at the Ministry of Defence, who was to present the plan.
'The best plan is a tried and tested plan. The British Military Mission devised a number of operational training plans for the Iraqi Army, which included a counter-attack. The last time we practised this plan was just over a year ago, and Colonel Yahin Hasan', he indicated an officer sitting to his right, 'commanded the exercise with the 6th Infantry Brigade. We have a tested plan, the officer commanding and the troops who have practised the exercise. Our best course of action is to use them as the blueprint for our counter-attack using the directing staff solutions.'
Observing the nods of assent from around the table, Mahmud passed the presentation over to Hasan, who at 35 was young to hold the rank of colonel.
Hasan described in detail the plan and his intention to support the attack with two light tanks and two mechanized machine-gun companies which would help to cover the attack, capture and demolish the bridge, and block the escape route of the British troops.
'When do you plan to attack?' asked Grobba.
'I plan to get into place tonight and begin the attack in the early hours of tomorrow morning.'
'So quickly and at night?' countered Zaki, more to show that he was still in charge of the meeting than concern for the operation.
'Yes. I want to catch the British off guard and give them little time to fortify Falluja. By daybreak we should be in possession of the town. This will limit the effectiveness of their bombers.'
With the blessing of the participants, Hasan left the meeting to prepare for the operation.
Habbaniya
An hour later, Clark held a meeting at D'Albiac's conference room at the Air Headquarters block to hammer out the plan for the advance on Baghdad.
In addition to Clark, the Army was represented by Kingstone, Ferguson and de Chair from Kingcol, Everett of the King's Own, Nichols of the Essexes, Strickland of the Gurkhas and Glubb Pasha.
With the arrival of Habforce, Roberts had relinquished command of Habbaniya ground forces. He had flown back to Shaiba late that afternoon to rejoin the 10th Indian Division as Chief of Staff to Maj Gen Bill Slim, who had replaced William Fraser.
The RAF contingent included D'Albiac, Savile, Hawtrey, Casey and Dudgeon, the Blenheim squadron commanders, Wightman of 94 Squadron and Pike of 203 Squadron, the two armoured-car commanders, Casano and Page and Brawn and Graham of the Levies.
The room was crowded. Clark sat on the conference table with his legs dangling, Kingstone and de Chair sat on the floor behind D'Albiac with their backs to the wall, while the rest of the participants sat at tables, on chairs and on the floor.
After some two hours of debate and discussion around Dudgeon's new photo-mosaic map, which covered the whole Falluja–Baghdad sector, Clark, supported by D'Albiac, summarized the strategy to take Baghdad. 'We are going to play a gigantic game of bluff. It worked at Habbaniya and it worked at Falluja, and there is every reason to believe that we can pull it off again. However, everything depends on our maintaining our momentum, giving the Iraqis no time to recover and regroup. There are two things in our favour. The Iraqis clearly have no intelligence on our strength. We have to continue to convince them that we are the vanguard of many. The Iraqi Air Force has ceased to exist for all intents and purposes, and the Luftwaffe seems to be attacking sporadically. We can rely on the RAF to provide us with ground-attack support, which has proved to be a very effective weapon in softening-up the Iraqis, as well as protecting us from attack by the Germans. We are all agreed that we need two days to finish our preparations for the ground assault.
'We will split our force into two columns to form a pincer movement. On 25 May one column will leave Habbaniya and head north from Falluja and advance to the north to Khadimain to cut the road and rail link between Mosul and Baghdad. This is designed to stop the Iraqis from moving their 2nd Division down from Mosul and Kirkuk to reinforce their 1st, 5th, and what's left of their 3rd Divisions around Baghdad.
'The northern column under Andrew Ferguson will comprise A and B Squadrons of the Household Cavalry under Eion Merry and Eric Gooch, three armoured cars, one troop of 237 Battery's 25-pounder guns, a detachment of the Field Troop of Royal Engineers and one of the 166th's ambulances.
'The southern column leaving on the 26th will take the shorter, more direct, Falluja–Baghdad route under the command of Joe Kingstone. Joe will have C squadron of the Household Cavalry under Rupert Hardy, Steve May with A and D Companies of the Essexes, 237 Battery's 25-pounders less one troop, the anti-tank troop with their 2-pounder guns, three armoured cars and one of the 166th's ambulances. There will be around 700 men in each column.
'Supplementing Andrew's and Joe's troops will be Glubb's Legionnaires, who will run guerrilla operations in the Jezirah – the area between the Euphrates and the Tigris from Falluja to Mosul. The Legion will hamper the passage of reinforcements and ammunition, deny road access and generally create alarm and despondency. They will be tasked with riding around the countryside ahead and on the flanks of both columns, distributing baksheesh to the local headmen and sheikhs and advising them to keep out of the way of the mighty British force on its way from Habbaniya to Baghdad. This is designed to stop the tribes from harassing our columns, as well as getting the word through to Baghdad that the Iraqi Army faces impossible odds.
'I am afraid that the food situation is still critical, so you will be carrying seven days of hard rations and water for five days. Each column will be carrying enough petrol for 460 miles.'
'Anything to add, John?' Clark asked D'Albiac.
'Yes, George. The RAF will be flying fighter cover over both columns to keep the Germans and any stray Iraqis off your backs. The Air Striking Force will be kept ready bombed-up at Habbaniya to give you air support. Casey and his man with Ferguson will be able to call up support over their radios. Dudgeon will continue to fly reconnaissance missions so that you are kept in the picture.'
After they wound up the meeting, Clark sat across the desk from D'Albiac in Smart's old office, nursing a whisky. 'What are the odds of pulling this off, George?'
'To be truthful it's touch and go. The bulk of Habforce is tied up around Ramadi, engaged in fighting a very aggressive Iraqi brigade. We have some 1,450 lightly armed troops facing at least 30,000 heavily armed Iraqis supported by armour and artillery. The Iraqis have some good, fresh troops, some professional commanders, and their formations are largely intact. We have to pray that the Iraqis do not stand and fight or take the initiative and attack.'
'What happens if we cannot take Baghdad?'
'Then we have to wait until Slim's 10th Indian Division arrives from the south and the rest of Habforce from the west. Then it depends upon the Iraqis. We certainly do not want to have to take Baghdad house by house and street by street.'
'And the Embassy?'
'They will just have to take their chance', he shrugged.
'It's been a long journey from Jerusalem to Baghdad', remarked D'Albiac.
'It's only the last fifty-five miles that really count', replied Clark.
Baghdad
Shortly before midnight Hasan took the decision to postpone the counter-attack until the night of 21/22 May. He had encountered serious delays in getting fuel for his tanks and vehicles from Baghdad, and in fixing the ground navigation lights to guide his troops into position for the attack. Everything depended on a successful counterattack, and Hasan decided to err on the side of caution.
Habbaniya, 21 May 1941
Out by the lake, Kingcol refitted, rearmed, refuelled and reprovisioned ready for its first combat. The advance Brigade headquarters was withdrawn from Falluja to Habbaniya, leaving a small force of Levies and King's Own to patrol the town.
Kingstone had the Household Cavalry supplement the Madras Miners and Sappers and the Royal Engineers shovelling earth and sand to plug the gap in Hammond's Bund in preparation for the movement of the two columns across the river.
In early afternoon D'Albiac held a strategy meeting with the Air Striking Force and the rest of his operational commanders. 'We have to eliminate the Luftwaffe as a threat. For the last seven days we have been conducting raids on Vichy airfields at Damascus, Mezzo, Palmyra and Kayak to choke off their supply route. While we have caught a number of their aircraft on the ground, we have also destroyed Vichy aircraft. We are not far off an invasion of Syria, but that is not our concern. Habbaniya's Order of Battle has changed substantially. A Squadron now consists of thirteen Gladiators and five Hurricanes. Together with B, C and D Squadrons of the Air Striking Wing, with its serviceable Audaxes, Gordons and Oxfords, and 84 and 203 squadrons, each with seven Blenheims, we now have a lot of firepower. The Germans are still striking at Habbaniya and destroying aircraft and killing people, and they continue to strafe our ground troops, and this has to stop. Without radar we have been forced to mount standing patrols, which are wearing out our pilots and aircraft. As the Luftwaffe is still getting through, the only thing we can do is to hit them hard where it hurts most – at their airfields. I have ordered a dusk strike at Raschid, and tomorrow morning I want reconnaissance over Mosul and their other airfields so that we can plan our strikes.'
At dusk the Air Striking Force with fourteen Audaxes, Four Gordons and nine Oxfords and four Blenheims of 84 Squadron, escorted by no fewer than thirteen Gladiators, mounted a large raid on Raschid and systematically destroyed the Iraqi Air Force base.
Falluja, 22 May 1941
The night of 21 and 22 May was extremely hot and there was no moon.
The 2nd Battalion of the 6th Infantry Brigade failed to reach its rendezvous, and Hasan, after waiting for a fruitless hour and a half, decided to attack at 02.30 with only the 1st Battalion.
Although alerted by the glare of trucks' headlights and a growing rumble from the direction of Baghdad, the defenders were still unprepared for the ferocious mortar barrage that preceded the fast and furious Iraqi assault on the southern and eastern defences. Simultaneously sniping broke out in the town itself
The British forces had been thinned out in preparation for the march on Baghdad, and only two companies of the King's Own and a platoon of Levies manned the town defences.
Lt Chris Hodgson of the King's Own, commanding the first outpost to be reached by the Iraqis, led a counter-attack with a haversack full of grenades, and was killed within minutes, and the outer defences were quickly overrun.
By three o'clock in the morning the Iraqis had penetrated the northeastern edge of the town and forced C Company, which was suffering heavy casualties, to withdraw from their trenches into buildings astride the main street. David Reese and Lt Ben King were seriously wounded and CSM Maguire was killed. D Company felt the full force of the attack, and was overwhelmed and forced to fall back to the river bank towards the vital bridge, pursued by two of the FIAT light tanks. At the bridge, the Levy platoon's anti-tank gunner, Corporal Marcus Odisho, stood up and fired his Boys rifle from the shoulder, stopping one of the tanks at seventy-five yards and forcing the other to retreat. The rifle was not designed to be fired standing up from the shoulder, and the gunner was knocked flat on his back by the recoil, suffering heavy bruising.
The Iraqis were now very close to the eastern edge of Falluja, and the situation was becoming serious. Graham counter-attacked with one platoon from the bridgehead. One of the platoon commanders, RAB100 Baijan Peeku, led the attack, coming under heavy and he and half of his platoon became casualties, although they did succeed in driving back the Iraqis.
Kingstone received an urgent order from Clark an hour later to go down to Falluja and take personal command of the battle.
At dawn Everett led a platoon of C Company with fixed bayonets against dug-in Iraqi positions, managing to push the Iraqis temporarily out of the town. They came under heavy sniper fire that wounded several men, including Everett, who continued to direct the King's Own from a stretcher.
A new Iraqi attack was mounted at the south-eastern outskirts of the town with infantry and two additional tanks against a heavily outnumbered C Company, which was holding on by the skin of its teeth.
Just after nine Kingstone arrived and found himself in the middle of a ferocious battle with hand-to-hand, house-to-house combat, with an outnumbered King's Own and Levies, who were running out of both ammunition and water, and facing the arrival of fresh enemy troops in the form of the 2nd Battalion of the Iraqi 6th Infantry Brigade.
Miraculously the sole radio still worked. Kingstone first called for reinforcements from Kingcol, and C Squadron of the Household Cavalry and A and D Companies of the 1st Essex camped at Lake Habbaniya left for Falluja. The second call was to Savile to get the Air Striking Force into the air to ground-attack both the Iraqi reinforcements and the enemy strongpoints in Falluja.
Hammond's Bund was still breached, and the reinforcements, having arrived in trucks, had to dismount, strip and wade the six hundred yards across the bund, piling their arms, ammunition and clothing, and those who could not swim, on top of rafts of old oil drums and planks made up by the Madras Sappers.
In parties of ten to fifteen, it took over two hours to get C Squadron of the Household Cavalry and the 1st Essexes across. They then had to hump their equipment, including their heavy Boys anti-tank rifles, their Hotchkiss machine-guns and their rifles and ammunition, another seven miles on foot into Falluja, wading through floodwaters, sometimes chest deep, holding their arms and ammunition over their heads. Capt Morgan Jones of the Cavalry moved up and down the column with words of encouragement, and relieved men of their loads.
A troop of 25-pounders from 237 Battery arrived at the bund and began shelling the Iraqi positions from across the river to give support to the King's Own and the Levies fighting in Falluja.
The reinforcements were thrown into the battle as soon as they arrived. Sniping had become a problem, and the British were taking heavy casualties. The Iraqis developed a habit of wearing gallabeahs over their uniforms, making it difficult to distinguish soldiers from civilians, which resulted in a number of ambushes. Kingstone called in the mayor and told him to get the entire population of the town into the Mosque within half an hour. Any civilian seen outside would then be shot. This had a major impact in reducing the level of civilian involvement in the battle.
Despite the additional reinforcements, retaking the town was a long and difficult job, as the Iraqis had dug themselves into trenches and had placed machine-guns on the flat roof-tops commanding the main streets and their approaches.
A Company, under Maj Robert James, was ordered into the attack. A platoon under Lt Terry Muirhead was sent to clear the right-hand side of the main road, and a further platoon was ordered to contain the Iraqis south of the main road, while the remaining platoon was to guard Iraqi prisoners, conduct anti-sniping patrols and keep the road clear for the evacuation of British wounded and for replenishment of ammunition.
Muirhead soon ran into trouble at the eastern end of Falluja from strong Iraqi fire from trenches and two houses on the north side of the main road, where machine-guns had been placed on flat roof-tops giving an excellent field of fire. James, using his last remaining reserves, his company headquarters platoon, had to destroy the Iraqi machine-gun positions to get to Muirhead. Leading the platoon across the main road under heavy fire, James moved them into an alley leading towards the machine-guns. However, progress was slow as they had to clear each house as they went forward. About fifty yards short of their objective, a light machine-gun opened fire from a window in the alley, severely wounding Private Sibbold. After firing back into the window, the platoon took cover in the house opposite, only to find it full of Iraqi soldiers, but after some fierce hand-to-hand combat they managed to clear the premises. From the roof of the house they engaged the Iraqi machine-guns that were pinning Muirhead down with sustained rifle fire, until the Iraqis retreated, leaving their dead and wounded. After they had poured more fire into the windows of houses occupied by the Iraqis, a white flag was put out of one of the windows and the whole Iraqi force in those houses surrendered.
With daylight the Air Striking Force mounted fifty-six sorties over Falluja, supporting the Army. They found and attacked an Iraqi convoy of some forty vehicles in the open on its way to Falluja, and damaged over twenty. One of the trucks, loaded with gun cotton to demolish the bridge, exploded in a spectacular fireball after a direct hit from an Oxford. RAB100 Peeku returned to the fray with six recently recruited Levies, and captured a Bren gun after a fire-fight with some Iraqis. Setting up in a house, they engaged a large group of Iraqi soldiers and their vehicles. Peeku continued firing the Bren gun until it became too hot to continue. After borrowing a replacement barrel from the King's Own, and supported by a Levy mortar section, Peeku took more than a hundred Iraqi prisoners. The wounded Peeku was awarded the Military Cross for his bravery at Falluja.
An Audax, piloted by Flg Off Dremas of the Royal Hellenic Air Force, was shot down but managed to land between the Iraqi and British positions. His air gunner, AC Peter Pike, dismounted his Lewis gun, and with two drums of ammunition the pair fought their way back to the British lines. Several Levies joined them, and together they made their way back to the rear under fire from snipers. Dremas spoke no English, so Pike took charge of the party, which now included a number of Levies wounded by snipers.
Two Audaxes on reconnaissance were intercepted and attacked by Iraqi Nisrs, but, although badly damaged, were able to glide down and land. The two pilots and one gunner left the dead gunner and managed to get clear of the aircraft before they were strafed and destroyed on the ground. They were befriended by a local sheikh, but while they were having coffee, an Iraqi policeman burst in and shot them out of hand, killing one pilot. The gunner feigned dead and was later taken to hospital. The other pilot was surrounded by Iraqi fellahs who were going to take him prisoner. However, someone in the crowd hit the pilot on the head with a wooden stave, killing him.
Following a prolonged house-to-house battle, the two Essex companies and the Levies finally managed to clear Falluja, and by six o'clock that evening all the Iraqis in the town had surrendered, been wounded or killed, or had fled.
The Iraqis had come very near to retaking Falluja and blowing the bridge, and only the tough defence, the rapid reinforcements and the Air Striking Force had averted a major disaster. Nevertheless, the British had suffered fifty casualties, including seventeen dead and twenty-nine wounded. All but three of the King's Own officers were either killed or wounded.
The Iraqi 6th Infantry Brigade suffered considerable casualties, including eleven officers and 237 men killed or wounded, and around one hundred taken prisoner.
The Iraqis had found that urban cover was an antidote to the striking power of the RAF. A number of Iraqi soldiers had changed their khaki uniforms for civilian clothes, and those caught were shot. As a precaution 1,300 civilians were expelled from Falluja to prevent further sniping.
Everett was evacuated to Habbaniya by launch and ambulance, and after an operation at the base hospital was flown to Shaiba later that day to await a hospital ship for India. Keith May of the Essexes was given temporary command of the King's Own in Falluja.
Mosul
In keeping with the strategy to attack the Luftwaffe's bases, two hit-and-run raids were mounted on Mosul. Two long-range Hurricanes strafed the airfield, claiming two aircraft destroyed, but one, piloted by Flt Lt Sir Robert MacRobert, failed to return. At dusk Flg Off Lane-Sansome in a 203 Squadron Blenheim bombed the airfield in near-darkness, setting one aircraft on fire.
While two He 111s did attack Habbaniya, no support was given to the 6th Infantry Brigade at Falluja. Effective air support for the Iraqis at Falluja could well have turned the counter-attack into a success.
Habbaniya
Dudgeon had suggested that, in the absence of accurate maps, he would fly along the Falluja–Baghdad route and take a series of line-overlap photos of the road and up to a mile on either side. By drawing an arbitrary grid on the photo strip with numbers and letters it would suffice for the advance.
At 10,000 feet in an Oxford, necessary for the camera and lenses to give the correct footprint, Dudgeon began to fly dead straight and level for the next twenty-five minutes.
In a clear blue sky in a flimsy plywood aircraft painted bright 'trainer' yellow, with light being reflected off its curved Plexiglass windscreen and rear gun cupola, Dudgeon began to feel extremely visible and very vulnerable.
The one good thing was that Frankie was safe back at Habbaniya. With no Iraqi shelling to frighten the dog, Dudgeon had started to leave Frankie behind. However, the dachshund had become used to flying and was beginning to resent being taken off his 'flying duties'.
As soon as he was told by his observer that the last photo had been taken, Dudgeon rolled the Oxford onto its back and pulled the stick back to drag the nose into a near-vertical dive. At about 4,000 feet there was a tremendous explosion, which he initially thought was cannon fire, but he soon realized it was the implosion of the Plexiglass bomb aimer's nose-window, which had fragmented when the Oxford exceeded its design limits. Pulling out just above the desert, Dudgeon got back to Habbaniya as fast as he could.
Baghdad, 23 May 1941
With the failure of the Falluja counter-attack, the road to Baghdad was now wide open to the British. al Sabbagh lost no time in demanding that Zaki form a military council to arrange for the defence of the capital. The meeting was held in the early hours in the same windowless conference room in the Citadel that had served for planning the Falluja counter-attack. Instead of a chill, an air of despondency hung over the room, as many of the participants perceived that this was the final phase of the revolt.
al Sabbagh, as C-in-C Western Command, took charge of the preparations for the defence of Baghdad. 'I have recalled Colonel Nur el Din Mahmud from Basra, and he will oversee the construction of our civil defences. We will form two defensive positions,' he indicated on the large wall map, 'one in the north and one in the south. These cover the only access routes to the city. From the west Lake Aqargouf forms an effective barrier between our two positions.
'Our defences will be built in depth behind anti-tank ditches. I want you', he gestured to the two leaders of the paramilitary Futuwwah and the Kataib al Shabab youth movements, 'to mobilize the youth, the workers and the religious organizations to start immediately digging these ditches. They must be fifteen feet wide and six feet deep. In the northern sector the ditch is to be four miles long, stretching from the right bank of the Tigris north of Khadimain here,' he pointed to the map, 'to the Aqargouf marshes. In the south the ditch is to be six miles long, extending from where the Tigris bends, near al Dorah, to the Aqargouf marshes. This needs to be started immediately, and you should go now to put this into operation.'
After they had left, al Sabbagh turned to the remaining participants. 'Colonel Raghib, I want you to take your 9th Infantry Battalion, the 1st and 9th Desert Artillery Brigades and the remnants of the Mechanized Force to defend our northern sector. Colonel Jahid, take your 4th Infantry Brigade, the reorganized 3rd Infantry Brigade under Lieutenant-Colonel Bapo and Colonel Yamani's 5th Desert Artillery to the southern sector.
'We will have a mobile reserve comprising the Cavalry Brigade and Colonel Hasan's reorganized 6th Infantry Brigade. One of the Cavalry regiments I want on the east bank of the Tigris, opposite Khadimain, to prevent any move towards Baghdad from this direction. We have little time and much to do. Please prepare your positions.'
Istanbulat
A detachment of the Arab Legion plus two RAF armoured cars and a detachment of Royal Engineers were ferried across the Euphrates at Sin El Dhibban and drove north to cut the Baghdad–Mosul railway at Istanbulat, some six miles south of Samara.
Legion scouts found that the Iraqis were sending reinforcements and ammunition to the Iraqi battalion holding up Habforce at Ramadi by first going north up to Meshahida and then, a short distance north of the station, turning west across the desert.
On their return from their successful raid, the Legion lay up on the track to Ramadi for some time, and was able to ambush a party of Iraqis, capturing two cars, a captain and six other ranks, and gaining some useful intelligence.
Rhodes
Eleven CR42 fighters of the 155 Squadriglia of the 3 Gruppo, accompanied by a Savoia Machetti SM 79 under the command of Capitano Francesco Sforza, landed at Gaddura a little after noon. Flying from Valona and Athens, they were on their way to Kirkuk via Aleppo to support the Iraqis.
Although the Regia Aeronautica had been ready for a number of days to send reinforcements, the Vichy French were extremely hostile to having the Italians land in Syria. It had finally taken a phone call from Ciano to Ribbentrop to put pressure on the Vichy French to help the Italians stage through Aleppo.
Berlin
Hitler finally gave in to Ribbentrop, agreeing that an operation in Iraq would tie down British forces, weakening their position on the Western front in Europe, and issued Fuhrer Directive 30:
The Arab liberation movement in the Middle East is our natural ally against England. In this connection, the rising in Iraq has special importance. It strengthens beyond the boundary of Iraq forces hostile to England in the Middle East, disrupts English communications and ties down English troops and shipping space at the expense of other theatres of war. I have therefore decided to advance developments in the Middle East by giving assistance to Iraq. Whether and how the English position between the Mediterranean and Persian Gulf, in connection with an offensive against the Suez Canal, shall later be definitely defeated is to be decided only after Barbarossa.
While the Directive placed only limited tasks on the OKW and postponed a final decision on Middle East strategy, it did not curtail German activity. In a separate act, Hitler created Sonderstab F under Luftwaffe General der Flieger Hellmuth Felmy, based at Athens-Tatoi airport, to lead German support for Iraq.
Campaniya, 24 May 1941
British Intelligence had established that the Germans were supplying the Iraqis with weapons using the railway line from Aleppo to Mosul and decided to cut the link. Early in the morning a 216 Squadron Valentia, piloted by Flt Lt Bartlett flew in from Heliopolis in Egypt with Maj David Collins and picked up a party of Sappers and explosives at Habbaniya. The Valentia flew to a point in northern Syria and landed near Campaniya, where the Aleppo–Mosul line crossed a high viaduct. The Sappers, under the direction of Collins, placed charges and demolished the viaduct, cutting the railway line. As Bartlett swung the Valentia round to take off, a Vichy armoured car arrived. Although it fired at the plane, Bartlett got into the air and made a safe return to Habbaniya.
Habbaniya
The Iraqi Army's almost successful counter-attack had delayed the move on Baghdad and raised concern in Clark's mind about the adequacy of the force at his disposal, compared with the strength of the enemy.
He was further unnerved by a patrol action that day which underlined just how weak his forces really were. D Company of the Essexes sent out a fighting patrol from Falluja, and less than a mile away from the town had been vigorously attacked by a larger body of Iraqi troops. Although the Essexes did manage to inflict some casualties, the enemy was too strong and they were forced to withdraw.
Many Iraqi units were still intact, with their troops showing fight, and water obstacles and new anti-tank ditches had been revealed by air reconnaissance. The continued presence of the Luftwaffe and the continued heavy fighting between Habforce and the Iraqi 1st Infantry Brigade at Ramadi in his rear made Clark wonder privately if he had gone too far with too little.
Baghdad
The Mufti sat forward in his chair to emphasize to Gabrielli his dissatisfaction with the situation. 'In spite of the good prospects for a revolt in the rest of the Arab world, this is only of limited value if our uprising here in Iraq, which is the key to the whole situation, should fail. What worries me most is not so much the military situation but the political situation. There is no lack of pessimists, many of whom are endeavouring covertly and in the open to sabotage the efforts of myself and Rashid Ali. Even the cabinet includes some individuals who hold that if the Axis powers do not render immediate and substantial assistance it might be better to negotiate with the British. I believe that the next fourteen days will be decisive. If we are unable to last through this period it will be necessary to give in. The aircraft that have arrived were insufficient and we need more, as well as war matériel as proof of Axis solidarity.
'If Iraq should fall in the coming days, the anti-British movement throughout the Middle East would succumb to British arms or British gold and intrigues. I realize that you can cope with this, but it would then be necessary to reconquer each step you have lost by force of arms without being able to rely on Arab goodwill.'
Gabrielli signalled his conversation with the Mufti to Rome later that evening, causing Mussolini to call in the German Military Attaché to clarify whether the Axis should furnish effective or merely symbolic assistance.
Early in the evening the families of Rashid Ali and Naji Shawkat left Baghdad by train for Kirkuk.
25 May 1941
B Squadron of the Household Cavalry and two RAF armoured cars under Lt Gerard Leigh had moved into position west of Ramadi across the Ramadi–Falluja road late the previous evening to prevent Iraqi troops from escaping from RAF bombing.
Early that morning, over a hundred of Fawzi al Qawujki's truck-borne irregulars attacked the squadron. The two armoured cars became bogged down in the sand and had to be abandoned to the Iraqis when B Squadron was forced to withdraw in the face of overwhelming odds. One of the armoured cars was later recovered by a troop of A Squadron with an RAF tender under Lt Valerian Wellesley.
Three Heinkels raided Habbaniya using delayed-action bombs which cut water mains leading to the civil cantonment. This was followed by a strafing attack by five of ZG76's 110s.
Three Blenheims of 84 Squadron bombed Mosul, while three more dropped leaflets on Iraqi troops to the north and west of Falluja.
The Miners and Sappers, aided by Iraqi prisoners, had finally realized the futility of trying to bridge the gap in Hammond's Bund with shovelfuls of earth and sand. Calmer heads had prevailed, and they eventually solved the problem of the seventy-five-yard gap in the bund by placing pontoons at either end and constructing a raft which could be pulled to back and forth between them with a wire hawser.
Aleppo
The Regia Aeronautica's CR42 Falco fighters and the supporting SM 79 Sparviero rigged-out as a transport for the ground-crews arrived at Aleppo from Gaddura in the late afternoon, and prepared to overnight at the airfield before flying west out to Kirkuk via Mosul early the next morning.
Iraq, 26 May 1941
Two Blenheims, one piloted by Flg Off Lane-Sansome of 203 Squadron and the other flown by Flg Off Goudge of 84 Squadron, made a low-level strafing attack on Kirkuk airfield at dawn, but found no German aircraft on the ground.
In preparation for the assault on Baghdad, Dudgeon flew a photo-reconnaissance mission between Baghdad and Khan Nuqta, some seventeen miles east of Falluja, and then constructed a photo-mosaic map for use by de Chair.
The Arab Legion again sallied forth in the night to the north. Ten Legion trucks and a detachment of Royal Engineers set out from Habbaniya across the Sin el Dhibban ferry at dusk and headed north to Samarra to blow the railway lines. This time they encountered an Iraqi mechanized patrol in seven cars. In a short, sharp skirmish the Iraqis lost three killed, a large number of wounded and one taken prisoner. The British suffered no losses, and pursued the Iraqis to the banks of the Tigris.
Intelligence gained from the prisoners indicated that great consternation had been caused by the raid on Istanbulat, the Iraqis previously believing that the British forces were concentrated to the west of the Euphrates. The Iraqi patrol had been hastily formed to defend the sector, and was made up of a mixed Army and police force in two armed trucks and five civilian vehicles. The Iraqis seemed to have limited motor transport, relying on commandeered vehicles, which created a lot of ill will among the car owners.
Kirkuk, 27 May 1941
The Regia Aeronautica flew into Kirkuk in the late afternoon after a brief stop in Mosul, and began to prepare for immediate offensive operations. Fortuitously the Italians had landed half an hour after the departure of two 84 Squadron Blenheims which had bombed and strafed Kirkuk airfield.
North of Baghdad
Leaving at dusk, Ferguson's northern column crossed the Euphrates by the Sin el Dhibban ferry and, guided across the desert by the Arab Legion, circumvented floods, broken bridges and bunds, to arrive some six miles north of Baghdad by midnight, where they bivouacked for the night.
Much was made later of the fact that the Iraq defences guarding the city were not yet in place in the north and that Ferguson could have driven right into the centre of Baghdad unopposed and seized the city by surprise. However, it is difficult to see how Ferguson could have held a city of over 400,000 with just 700 men.
Falluja
At Hammond's Bund Kingstone and the southern column's vehicles and equipment began to arrive at dusk for the crossing. The 25pounders and their ammunition limbers were the first to arrive. However, the guns had to be manhandled across, and the gun troop alone consumed three and a half hours. It had been decided to complete the crossings by 04.00 and disperse the vehicles which had not been able to get across, to avoid providing daylight targets for the Luftwaffe. All the supply lorries had to be driven away from the bridgehead and wait to cross the following evening.
Baghdad
When Gabrielli met Rashid Ali early that evening he found a very embittered man.
'I do not think that we can do anything to save the situation in Iraq. German aid has been rapid but meagre, and I fear that Iraq has been used as a pawn by a Germany that does not wish to commit herself wholeheartedly. I also expected more from Italy.'
Falluja, 28 May 1941
After reveille at three in the morning, followed by breakfast half an hour later, Kingstone's southern column loaded up to leave Falluja. It was chilly, and men everywhere were walking around in greatcoats, which they only discarded once the sun came up.
The armoured cars left first, followed by the trucks of C Squadron of the Household Cavalry, and then A and D Companies of the 1st Essex with the brigade support vehicles bringing up the rear. Although there was a metalled road marked by a line of telephone poles, the column spread out across a wide front to minimize targets in case of an air attack.
Twenty-five miles later, just after five o'clock, the southern column topped a rise and was confronted by the fort at Khan Nuqta, a large, square, cream-washed, plaster building. Air reconnaissance had shown a large number of one-man air raid shelters on both flanks of the fort, indicating the possibility that the column could be facing Iraqi strength of up to a battalion.
C Squadron dismounted from its trucks and worked its way around to the north of the fort. By seven in the morning the fort, together with an Iraqi battalion headquarters and around fifty prisoners, was captured. Interrogating the officers, they found that the Iraqis had dug wide trenches along the Baghdad road at irregular intervals; they had then opened the Abu Ghraib Regulator, flooding the road for about a mile; they were relying on these hidden obstacles to impede the column's advance. Armed with this information, 'Cheeky Chappie' the engineer and his men set to work to bridge the underwater trenches with ten-foot steel ramps.
Brigade Headquarters had moved a little way up the road to a set of low, white buildings set in a small grove of oleander trees, complete with a rudimentary garden. Rummaging around, de Chair discovered a working telephone switchboard which was still connected by circuit to the telephone system of the whole area. When the phone rang de Chair called over his driver and interpreter, Reading, a Christian Palestinian, to answer the call.
'Where have you been, I have been trying to raise you?' Reading told de Chair, covering the mouthpiece.
'Tell them that we are surrounded by the British: that they have tanks and that the tanks are already across the flooding.'
Reading duly spoke to the operator in an excited voice, causing consternation at the other end.
de Chair decided not to spread any more deception. Finding Kingstone, Maj Ian Spence, Kingstone's special liaison officer, Capt Dick Shuster, the signals officer, and Grant-Lawson, de Chair gave them the good news.
The rest of the morning was spent intercepting Iraqi messages.
The battalion commander of the 3rd Division came on the line to tell his headquarters that he had 'just escaped from Khan Nuqta with twelve men and I am hiding at Qasr Milh.' A patrol was ordered out by the 3rd Division Headquarters to report on the presence of British tanks, and later in the morning, with de Chair scarcely believing what he heard, it replied, 'The British have at least fifty tanks, and fifteen are already across the floods.' A distraught artillery officer came on the line crying that he was 'shelling the British and I am engaging a formation of five British tanks.' At this point a stern voice from Headquarters admonished him to 'Stand firm!' to which the agonized gunner replied, 'Can't you hear the firing?'
The battalion commander ejected from Khan Nuqta came back on the line asking for reinforcements to stabilize the retreat. 'I don't mind about Khan Nuqta, what about Baghdad if the British have tanks?'
'Don't worry about Baghdad,' came the Headquarters reply, 'we have dug a ditch around Baghdad that will stop any tank.'
Later the 3rd Division Headquarters and the 3rd Battalion at Khadimain asked where the two companies of the 1st Battalion requested to join the Brigade Colonel at Qasr Milh were.
'They started out all right,' came the reply, 'but when the British aircraft approached they dispersed to the villages.'
Around two o'clock in the afternoon the telephone line went dead.
The intercepted telephone messages had provided the southern column with an almost complete picture of the Iraqi order of battle in its sector. By mounting a highly effective deception operation with a non-existent British tank force it had compelled the Iraqis to retreat to their second line of defence on the Abu Ghraib Canal.
Later in the afternoon the column moved on, cautiously entering the water, which now covered the tarmac road, and carefully searching out the pair of steel ramps that Cheeky Chappie's men had laid, bridging each of the trenches.
The column passed through the water barrier, but on reaching a line of sand hills the leading troop came under heavy fire, two men were wounded and four trucks full of troops were stranded on a straight road with ditches on either side, taking fire. Unable to manoeuvre the trucks, Maj Rupert Hardy dismounted the troopers and advanced across open desert without cover. The three troops who carried out the action moved by section rushes under continuous long-range machine-gun fire, and eventually reached the sand hills. Using their Vickers guns, they forced the Iraqis to retreat to a small village. The column, with its exhausted men, ended the day scattered along a low ridge facing the Abu Ghraib Canal, some twelve miles from the centre of Baghdad.
Things had not gone as well for the northern column.
At first Ferguson advanced without incident to a railway crossing about three miles north of Taji station. The RAF armoured cars met enemy armoured cars and some light tanks near the station, but after an exchange of fire the Iraqis retired.
After a brief fire fight the northern column took Taji station and exchanged fire. When Ferguson was about 200 yards short of the station, the Iraqis stopped firing, and as the troopers rushed forward to the station four car-loads of Iraqi troops and police drew rapidly away from the back of the fort protecting the station, enabling the northern column to take Taji station without casualties.
However, the advance was held up three miles further on as they took heavy fire from a line of sand hills. When they turned south along the railway they began to encounter heavy shell and machine-gun fire from the village of Khadimain which killed Trooper Shone and wounded five other cavalrymen. Ordered to recover Shone's body, Valerian Wellesley's troop was pinned down by accurate Iraqi fire, and Wellesley ordered Corporal of Horse Maxted to lead the fall-back. Maxted, all six foot three, stood up, brushed the sand off his uniform and marched erect with his rifle at the short tail across a hundred yards of open desert to the safety of a ridge while bullets kicked up the sand around him.
Fearful of an attack from the rear, the Arab Legion provided a protective screen against reinforcements from Mosul which could pin the northern column into the narrow strip of land between the Aqargouf marshes and the Tigris.
The Arab Legion had fared well, cutting the railway track and the telephone lines on both sides of Meshahida station. The massive two-storey loop-holed fort protecting the station was garrisoned by around thirty Iraqi soldiers and eight police. Although the Iraqis were armed with Vickers machine-guns and put down heavy fire, the Legion suffered no casualties. During the evening a Legion patrol north of Taji station wounded and captured the Mutasserif of Baghdad, Jellal Beg Khalid, who had been on a visit to Samarra. He confirmed that the advance of the northern column via Meshahida was a complete surprise, as all the defences had focused on the British advance on Baghdad from Falluja.
As darkness fell Ferguson decided to bivouac in a square leaguer with fighting patrols out in front throughout the night. A very tired and weary northern column slept in its trucks.
The intensity of fire from Khadimain made Ferguson think that his column might be facing a full Iraqi brigade. He began to wonder if it would be possible, even with all the luck in the world, for his few trucks of cavalry troopers, a machine-gun troop and three 25-pounders to break through the strong enemy position and take Baghdad. Adding to his problem was the arrival of a regiment of Iraqi cavalry on the east bank of the Tigris opposite Taji. Desultory bursts of fire were exchanged across the river throughout the afternoon between the Iraqis and a picket of the Arab Legion on the west bank.
Despite the progress of the southern column, there were still strong pockets of Iraqi resistance to the north and east of Falluja, and consolidating the British position against continuous sniping and skirmish and an extremely mobile enemy, more 'hide and seek' than regular combat, was proving to be very difficult.
Two platoons of the Essexes, together with a Levy machine-gun section under Keith May, fought a major engagement against an entrenched Iraqi force with well-protected flanks which held an oasis by the Abu Ghraib Regulator. The objective was to close the regulator and control the floodwater over the Baghdad road. A heavy air attack and a barrage from a 25-pounder section failed to shift the well-dug-in Iraqis. The British attack was countered by heavy machine-gun and artillery fire from Iraqi reinforcements, which caused a number of British casualties. A second British attack, led by Major Kelly, tried a limited advance under cover of an artillery barrage, hoping that the Iraqis would retreat. But the Iraqis, reinforced by additional machine-guns and light artillery, caused a number of British casualties, including Kelly, who was seriously wounded. During the evening the Essexes and the Levies withdrew.
Mosul
Junck surveyed the battered airfield and the hulks of his aircraft dotted around the perimeter. The crew chiefs had done a magnificent job in keeping his command airworthy. But there was a limit to how much you can cannibalize damaged aircraft to keep others in the air. The lack of spare parts and supplies, the hostile climate, poor-quality fuel and lubricants, together with the relentless British air raids on his base, meant that there was now little left to fight with.
Sonderkommando Junck was reduced to just two Heinkel bombers and four 50 kg bombs and one Ju 52/3m transport. Junck had lost fourteen Messerchmitt 110s and five Heinkel 111s – an unacceptable loss rate of 95%. Junck had reluctantly signalled Felmy at Athens-Tatoi that his command was no longer operational, and requested immediate air evacuation.
Junck also managed to reach Harry Rother, and asked him to fly in to Mosul at daybreak the next day with the two Junkers 90s and the two Junkers 52/3ms which were sitting on the ground at Aleppo, to pick up the rest of the ground-crews, the Flak section and his command group.
He then called a meeting with the flying and ground officers under his command in the only part of the administration building still intact. The tired and lined faces, mostly sporting beards, looked back at him in their stained and dirty mix of tropical and European summer uniforms. They showed both the immense strain that Sonderkommando Junck had been forced to operate under and the poor way in which they had been equipped and supported.
'The situation is as follows. We only have three operational aircraft and virtually no supplies. There are vague promises to send reinforcements, but Student's operation in Crete has taken precedence. I have no choice but to evacuate to avoid our capture. The situation is no reflection on your efforts. You have all performed above and beyond the call of duty under the most difficult of circumstances, and I have been privileged to serve with you. I want you to convey that message to your men. I have notified Berlin and Athens, and I have asked Rother, who is in Aleppo, to get here at first light to help ship us out.
'Miki [to Hauptmann Michael Hull, his Administrative Officer], I want you to have the ground-crews rip out the radio station in our remaining Tante Ju. We will need this to get our people out of here. We also need to strip the two Heinkels of all non-essential equipment as these are also needed as transports. We have little fuel. You will have to get the ground-crews to drain the tanks of all the damaged aircraft so that we have enough aviation spirit in our three flyable aircraft to get us to Aleppo.
'Finally, you and your men will take only what you can wear. Everything else, including our equipment and all papers, are to be burnt before dawn so that we can evacuate as quickly as possible.'
In late afternoon the third and final shipment of arms and ammunition from Syria arrived in Mosul. Like the second shipment of 26 May, this delivery had been delayed by the destruction of the viaduct outside Tel Kotchek. This had necessitated laboriously unloading the equipment and munitions onto a motley collection of trucks and then driving them across the border to be reloaded onto trains bound for Mosul. Both shipments joined the first delivery in warehouses in Mosul and remained there throughout the conflict.
Baghdad
The Supreme Defence Council met in the windowless, chilly, conference room in the Citadel which was now becoming depressingly familiar to many of the participants.
In addition to the Golden Square – al Shabbagh, Salman, Said and Shabib – and senior military figures like Zaki, they were joined by the replacement Regent, Sharif Sharaf. The Mufti, Rashid Ali, also attended, with the Minister of Defence Naji Shawkat, the Finance Minister Naji al Suwaidi and three other cabinet ministers –Yunis Sabawi, Ali Mahmud and Musa Shabandar.
A downcast al Sabbagh took the floor. 'Following the failure of the siege of Habbaniya and the collapse of the Falluja front, our situation is very grave. The Army is rapidly running out of ammunition. The number of British troops facing us is growing by the hour. They are being reinforced by air and by ground from Palestine, and their Army has finally broken out of Basra and is moving north to support their tanks facing Baghdad. They have also enlisted the help of the Arab Legion, and the Jordanians are in Iraq in force.'
'What about the Germans?' asked Shawkat.
'They sent a small, inadequate air force which has now largely been destroyed by the British. The British have also attacked them in Syria, cutting off the German lifeline. The Germans have begun a major operation in Crete and have no more planes to share with us.'
'What happened to the arms that they sent to Mosul from Syria?' asked the replacement Regent, Sharif Sharaf.
'Those arms were French. Their equipment and even their calibres are different from those of the British arms that we use. Their ammunition does not fit our weapons and we have no time to train our soldiers to use different types of weapons.'
'And the Italians?' questioned Said.
'Fewer than twenty planes arrived in Kirkuk in the last few days. Too little, too late.'
'What do you suggest?' the Mufti asked.
'I propose that we withdraw to Kirkuk and establish a new front, using the Diyalah river as a defence line. By withdrawing we can make Baghdad an open city and save the capital from destruction. Kirkuk not only offers us a number of natural defence lines but from there we control Iraq's oil and we can move easily into Turkey and Persia, both neutral, as is necessary.'
'How is Baghdad to be administered?' Yunis Sabawi enquired.
'A Committee for Internal Security is to be formed, consisting of the Mayor of Baghdad, the Director-General of Police and some Army officers who will remain', replied al Sabbagh.
The proposals were endorsed by the replacement Regent, the Defence Council, the military leaders and the cabinet, who lost little time in leaving the meeting to prepare for the move to Kirkuk.
Later that evening Rashid Ali met Suwaidi and Zaki, and they agreed to move the Treasury to Kirkuk in an armed police vehicle, accompanied by the Finance Minister and the Chief-of-Staff.
Grobba had fallen victim to the rumours that were now flying around Baghdad concerning the vast strength of the British forces and their imminent capture of Baghdad. In his daily radio message to the Auswartiges Amt, Grobba reported that the British forces were now only a few miles from Baghdad with more than a hundred tanks. He also added, maliciously, that the air support had been completely mishandled by Junck, who had provided almost no support for the Iraqi Army and had been virtually destroyed by a poorly armed British Air Force flying biplanes. Sonderkommando Junck, which now had only two bombers in flyable condition, was a spent force. Concerned that the Iraqis would capitulate, he advised the Auswartiges Amt that he was preparing to transfer his mission from the capital to Mosul.
Baghdad, 29 May 1941
At two o'clock in the morning Ferguson was woken by a hail of machine-gun bullets from four Iraqi trucks which had come down the main road and opened fire at the northern column's camp at point-blank range. However, accurate counter-fire from an A Squadron machine-gun emplacement knocked out the trucks. After the engagement Ferguson found that he had become the proud owner of a slightly damaged Ford V8 truck, which he used to replace his Haifa taxi, now on its last legs.
Just after dawn, A Squadron took the line of sand hills which had held up its advance on the previous day, and the column again moved forward to the Khadimain brick works. The sand hills had been a covering position of the main Iraqi defence system, and while B Squadron carried on the advance it came up against the main position at Khadimain station, and was held up for five hours by heavy shell fire, which killed one man and wounded five others.
Al Khadimain was one of the most important mosques in Iraq, a factor which seemed to have contributed to the Iraqis' fierce defence – in contrast to the weak defences encountered by the southern column. The flanks of Khadimain were protected by the Tigris on one side and by the virtually impassable Aqargouf marshes on the other. With an Iraqi brigade in place in well-defended positions supported by heavy artillery, it was beyond the power of two dismounted cavalry squadrons, supported by one troop of machine-guns and one troop of 25-pounders, to take. The possibility of collateral damage to the holy shrine ruled out bombing by the Air Striking Force. That evening the northern column returned to its bivouac at Taji station, and although probing patrols were sent out which revealed the strength of the Iraqis, and a further attempt was made the next morning to capture the station, the northern column made no further progress until the armistice was signed.
The southern column moved on towards Baghdad, reaching a twenty-yard-wide canal with a breached road and a bridge which had been blown by the retreating Iraqis. Still entrenched across the canal, the Iraqi heavy machine-gun fire made an immediate crossing impossible. After several hours of sustained fire from the 25-pounders, 2/Lt Grimley led two sections across the damaged bridge, followed by Muirhead and his platoon, to secure the bridgehead
Kingstone sent for 'Cheeky Chappie', who, after surveying the damage, requested a signal to Habbaniya for all the forward bridging equipment available. There was nothing to do except sit and wait.
The Air Striking Force actively sought out and then bombed and strafed Iraqi positions on the outskirts of Baghdad to support the southern column, not only against the Iraqi Army, but also to harry the Zoba tribe, which had thrown in its lot with the revolutionaries, and some Futuwwah irregulars who were mobilizing against the British.
At 10.00 three Audaxes escorted by two Gladiators were bombing across the eastern bank of the Abu Ghraib Canal where the Iraqis were strongly entrenched, when they were intercepted by two Italian Falcos from Kirkuk flown by Tenente De Merich and Sottotenente Valentini.
After an attacking pass at one of the Gladiators, both Falcos lined up on an Audax flown by Flt Lt Ian Webster. The attack left his gunner, Sgt Alex Payne, slightly wounded, and with smoke pouring out of the engine Webster crash-landed as close to the British lines at Falluja as he could.
Freddie Wightman, flying one of the Gladiators, dived from above, and half-rolled, putting himself behind one of the CR42s. As the Falco turned, Wightman kept turning inside, keeping his sights on the CR42 until the range closed to 100 yards. After three short bursts, black smoke poured from Valentini's engine and the Falco began to disintegrate. Valentini pulled the nose up to roll inverted, and simply dropped out of the cockpit after popping his seat-belt and harness. De Merich, now facing two Gladiators, decided to fight another day, and headed back to Kirkuk, while Valentini's parachute opened up and he drifted down to the Abu Ghraib Canal.
de Chair was sitting in his car writing up Kingcol's war diary when his driver shouted to him. Scrambling out of the car, de Chair saw one of the planes begin to fall and a white parachute detaching itself. A few minutes later the aircraft hit the ground about two miles away. Almost in slow motion, the parachute with a man underneath followed the plane down and then landed close to the crashed aircraft. de Chair, followed by his driver and Reading, ran towards the wreck until they were brought up short by the Abu Ghraib Canal. Undeterred, de Chair plunged in and swam across to the eastern bank, clambering up to the slope to the wreckage.
There were Iraqi insignia everywhere, but no interesting papers. Looking around he saw a man a few hundred yards away in long trousers, dragging a parachute along the ground. de Chair, drawing his Webley, ran over to confront the man and was startled to receive a Fascist salute and to be addressed in Italian. This was the first inkling the British had of Italian participation in the revolt.
Later that day the Air Striking Force organized a heavy attack on the Italian base at Kirkuk and a raid on the airfield at Baquba. While no damage was done at Kirkuk, two Italian and one Iraqi Air Force aircraft were damaged at Baquba.
The Arab Legion moved towards Meshahida early in the morning to cover the rear of the northern column. The Legion engaged an Iraqi police patrol from Smaicha, which was reconnoitring the rear of the northern column, killing two and capturing five. At this point a train came into view on its way from Meshahida to Taji, carrying Iraqi reinforcements to attack the rear of the British column. Seeing the fire-fight with the Legion, the train went into reverse, heading back towards Meshahida.
Near Falluja the Essexes got into position for a third assault on the Iraqis, who still held their position at the oasis controlling the Abu Ghraib Regulator. A further Air Striking Force raid, bombing the Iraqi positions for an hour, was supported by a sustained bombardment by the 25-pounders. Most of the bombs and much of the artillery fire missed the target through poor spotting. With no impact on the strong Iraqi force, which continued to concentrate heavy, accurate fire on the British positions, it was concluded that they faced at least a full Iraqi battalion, and the attack was called off. Adding insult to injury, the 1st Essexes were then showered with leaflets in Arabic calling on them to surrender.
Rashid Ali called Grobba in the early evening. 'I wanted to advise you of the decisions made by the Supreme Defence Council last night. We intend to defend Baghdad and we will need to have Luftwaffe support. Conscious of our situation, we may lose the city. However, we will continue to defend the country step by step. We have decided that we would then concentrate our forces around Kirkuk. Please ask Colonel Junck for his continued support, and I would like you to impress upon your Foreign Minister our need for more aircraft and supplies to continue the fight against our mutual enemy.'
Gabrielli managed to reach Sforza by radio just after four in the afternoon, and advised him to leave Kirkuk in a hurry as the Iraqis were about to capitulate. Sforza's men immediately began to pack up to be able to leave Mosul at daybreak the next morning.
Rashid Ali was just finishing the packing of his papers, and was surprised to receive a call in his office from an agitated Mahmud Salman.
'You should know that the Army leaders decided early this morning to abandon their command posts, and have already left in their cars for Persia.'
'But how can they do this when the Supreme Defence Council has decided to continue the resistance and to establish a new front at Kirkuk?'
'Rashid Ali, I am going to follow my colleagues, and I suggest that you do the same, otherwise you will be left here on your own', replied Salman before he cut the line.
Rashid Ali informed Sharif Sharaf of the situation, and then called Suwaidi and Zaki, who were still at Baquba, to hand over the Treasury to the Governor of Diyalah and then strike out for Persia. He then, together with the Mufti, Gabrielli and the rest of his civilian cabinet, left Baghdad for Persia. Later that day they crossed the border a mile or so south of Quasir e Shirin.
Facing delays in repairing the bridge, the 1st Essexes crossed on foot and advanced for six miles until, with the pontoons finally in place by the late afternoon, the southern column caught up with A and D Companies. The land now changed from desert to a region irrigated by dykes and canals, where the hard, sandy ground was broken by ridges covered with thick green scrub, and lines of gum trees followed the banks of the canals.
Early in the afternoon a further canal bridge was reached, this time unblown, and Robert James quickly cut the wires leading to demolition charges. With Capt Greene left commanding the bridgehead, the advance on Baghdad continued. During mid-afternoon an Iraqi Gladiator strafed the bridge, but was finally driven off by sustained rifle fire from A Company Headquarters.
A and B Squadrons of the Cavalry saw action as their advance guard of two armoured cars came under machine-gun and shell fire. They quickly dismounted from their trucks, and while one troop began to work its way around the Iraqi position in a wide sweep on the right flank, the rest of the squadron advanced up a narrow ditch running alongside the road. They occupied this position throughout the night.
When the light began to fade, the main column turned off the Baghdad road at the Government Agricultural Station and drove along a track to the Experimental Farm which nestled in among some eucalyptus trees. The main part of the southern column, which had bivouacked for the night at the Experimental Farm, found that they had acquired a large number of abandoned chickens, which quickly supplemented their hard rations.
Baghdad, 30 May 1941
This was a confusing day for Grobba. His first signal to the Auswartiges Amt at daybreak advised Ribbentrop that the military situation had improved as a result of both Iraqi Army action against the British and their flooding of the approaches to Baghdad. He also confirmed that command of the Iraqi Army and the cabinet were on their way to Kirkuk and were determined to continue the fight, provided that Germany immediately sent military aid.
Kirkuk
At daybreak Sforza's men poured petrol on two damaged Falco biplane fighters and on stores and equipment which were no longer needed. As soon as the fires were well alight, some of the ground-crew left in the unit's Sparviero transport, while the remaining staff left in two small commandeered buses, heading towards Mosul and the Syrian border, some 150 miles away.
The nine remaining CR42s flew cover for both the Sparviero and the buses until they were safely into Syrian territory. This ended the Regia Aeronautica's participation in the Anglo-Iraqi war.
Mosul
The air evacuation of Sonderkommando Junck had been delayed by a day. One of the Junkers 90s had been destroyed on the ground at Aleppo by a British raid, and one of the Junkers 52/3ms had become unserviceable.
Rother, flying the remaining Junkers 90, had already made one trip that day from Aleppo. He had flown in at dawn in company with the three-engined Junkers. He had agreed with Junck that the Command Unit and the remaining officers fly out in the Tante Ju to Aleppo while he took the remaining ground-crew members and the Flak team. This still left a dozen men at Mosul. Rother, in the faster plane, agreed to return from Aleppo to pick up the remaining Luftwaffe personnel, who during his absence would systematically destroy all remaining equipment and stores.
After loading up, Rother took off at 05.30, arriving back at Aleppo by 09.00. Unloading and refuelling took an hour, and a shade after ten in the morning he was back in the air heading towards Mosul to pick up the remaining airmen. Before landing, his radio operator, Klaus Meier, advised him that following another heavy RAF raid on Aleppo, Rother should now route himself through Palmyra before flying on to Gaddura.
Due to a slight headwind it was almost 14.00 when Rother touched down at Mosul airport and taxied over to the damaged administration block and swung the tail around so that he could make a fast take-off.
Leaving his co-pilot, Ernst Hauser, at the controls, Rother walked down the sloping floor of the aircraft, opened the door and jumped out. Rather than a dozen or so men waiting for him, there were just two.
'Where the hell are your comrades?' he asked the Gefreiter.
'They did not believe that you would return for us, Herr Hauptmann. They commandeered two cars and set out for Syria', the tall one said.
'When was this?'
'They argued for a bit after you and Oberst Junck flew out, and then they left about 08.30, just after the last British raid.'
'Why did you stay?'
'You said you were coming back for us, Herr Hauptman', said the shorter one.
'Well, we are not going to hang around here with the RAF looking for us, so you had better get aboard quickly', Rother said, and climbed up the short ladder into the Junkers, followed by the two young airmen.
At 14.12 the last aircraft of Sonderkommando Junck took off from Mosul airfield. The Junkers 90 in its drab camouflage and peeling Royal Iraqi Air Force markings struggled into the hot, overcast sky. The plane made a slow, lazy turn over the airfield and slowly gained height as Rother set course south-west for Palmyra in Syria.
Berlin
Ribbentrop's reply to Grobba advised him that new air units were on the way and would land at Mosul on or around 1 June, and he should brief Rashid Ali and the Iraqis. In the meantime Grobba learned that Rashid Ali, his cabinet and the senior Iraqi military commanders, including the Golden Square, had all left for Persia. Given the circumstances, Grobba advised Ribbentrop that he was evacuating his team to Mosul and would resume contact later that day.
Meshahida
With Meshahida station still in the hands of the Iraqis, Ferguson's column was dangerously penned-in by the Iraqi brigade to the south at Khadimain, an unknown number of Iraqis who could be quickly reinforced by troops by rail from the north at Meshahida, and the natural barriers of the Tigris to the east and the Marsh to the west.
The Arab Legion was tasked with taking the fort at Meshahida, backed up by one 25-pounder and one squadron of the Household Cavalry.
After a heavy bombardment of the fort by the 25-pounder, the garrison fled, dispersing among the neighbouring tribes, where they changed into civilian clothes. While the 25-pounder and the Household Cavalry squadron rejoined the column in front of Khadimain, a detachment of the Legion under Major Lash occupied the fort to cover the rear of the northern column from attack. Lash began to patrol aggressively in the Ramadi–Meshahida–Habbaniya area, keeping it clear of Iraqi military and police forces, intensified subversive activities among the Shammar and other tribes, and maintained the railway-line break near Samarra.
Baghdad
The southern column was now three miles from Baghdad, by the Khir Iron Bridge, facing the Iraqi forward defence line at the Washash Canal, while the northern column was at Khadimain opposite the Iraqi northern defence line, which was some five miles north of the city.
With a combined force of fewer than 1,400 men, eight 25-pounders and four old armoured cars, the British faced an Iraqi Army of well over a division in size which was resolved to defend the sprawling urban capital to the bitter end.
In both sectors the British came under increasingly heavy fire from the Iraqi positions, and replied with strafing and bombing attacks by the Audaxes, Blenheims, Gordons and Oxfords of the Air Striking Force on the northern and southern defence lines, followed by barrages from the 25-pounders.
As dusk began to fall the Audaxes conducted a twelve-minute dive-bombing attack on the Washash Army barracks about a mile behind the Iraqi lines facing Kingstone's southern column.
This last attack, which devastated the barracks, together with the banshee wails from the tails of the 250 lb bombs which had been modified by the engineers at Habbaniya, had a tremendous impact on Iraqi troop morale. This was instrumental in deterring the Iraqi commanders from prolonging the conflict.
Mosul
Arriving at Mosul air base in the early evening, Grobba found a deserted airfield littered with the still-burning hulks of unserviceable and damaged Luftwaffe Messerschmitt 110s, Heinkel 111s and Junkers 52s that marked the end of Sonderkommando Junck. He signalled Ribbentrop that it was now impossible to land at Mosul due to British bombing, and that as a result of the shortage of high-octane aviation spirit reinforcements would now have to operate out of Syria. With the Syrian border only sixty miles away, and with the capture of Mosul by the British imminent, he and his party would head towards Syria.
Grobba and his small group were on their way, and did not receive Ribbentrop's signal to stay in Mosul until the air reinforcements arrived.
Baghdad
Earlier that day al Sabawi, who had organized an armed force of irregulars from his paramilitary Futuwwah and the Kataib al Shabab youth, heard of the flight of Rashid Ali, the Mufti and the Golden Square to Tehran. He immediately issued a statement that Rashid Ali had named him Military Governor of Baghdad, and that he would prepare for popular resistance against the British.
The Committee for Internal Security quickly denounced his move, divested al Sabawi of all powers and dispatched a joint Army and police force to surround his headquarters at the Council of Ministers, where he was arrested.
Brought in front of the Council, al Sabawi conducted a heated discussion about continuing the resistance. 'We must continue to fight the British. We must win. Only then can we be masters of our own destiny.'
'We have lost against superior forces. There is no point in encouraging the destruction of Baghdad nor in sacrificing more Iraqis', said the Mayor of Baghdad, Arshad al Umari.
'You are all defeatists! All we need is a few more weeks until the Germans and the Italians come in strength. Then we can rout the British and together with our Arab brothers throw them out of the Middle East for good.'
Colonel Nur el din Mahmud looked aross the table at al Sabawi and said, pointedly, 'Your Axis friends have gone and none of your brother Arabs bothered to lift a finger to help us. All we will do is to destroy ourselves. This is the wrong time and the wrong place to continue our fight. If you will not agree then we must consult with all the Army commanders and convene an emergency meeting of the Council of Ministers. Let them see this treachery and decide for themselves!'
'Enough!. We have decided to seek a peace treaty to stop the fighting. You would be better to leave immediately for Tehran.'
A little after eight that evening, the Mayor of Baghdad, together with Colonel Nur el din Mahmud, arrived at the British Embassy to meet Cornwallis. They were kept waiting some twenty minutes while Cornwallis quickly shaved and changed into a formal suit, and arranged to have the last of the Embassy's coffee served. Appearances were everything in the Arab world.
As they were shown into his office, Cornwallis rose and walked over to the two men and gestured that they should sit in the comfortable chairs by the coffee table, and help themselves to coffee.
'How can I help you?' asked Cornwallis, as if the siege of the Embassy had never occurred.
'Ambassador,' began al Umari, 'Rashid Ali, the Mufti and the members of the Golden Square have all fled to Persia, and Grobba and Gabrielli and the rest of the Germans and Italians have also left the city. Control of the city has now been assumed by a Committee of Internal Security under my command.'
'Under these conditions the military have elected me to request a cessation of hostilities and to arrange a meeting to negotiate terms', added Colonel Nur el din Mahmud.
'Naturally I will have to discuss this with my colleagues at Habbaniya, and I need access to a telegraph line to communicate with them.'
'I will arrange this immediately', replied al Umari. 'In the meantime could I suggest that our envoys from Baghdad meet with your envoys to negotiate terms?'
'Where do you suggest, and when?'
'At the Khir Iron Bridge over the Washash Canal, where our troops face each other', suggested Mahmud, 'at, say, two o'clock in the morning?'
Kingstone received a message from Clark informing him that Iraqi envoys would arrive at the Iron Bridge at 02.00 to negotiate armistice terms.
Baghdad, 31 May 1941
The meeting had been postponed at the Iraqis' request to four o'clock in the morning.
Even at the height of the summer the desert at night can be a bitterly cold place. On the western bank of the Washash Canal, masked by a ridge, Clark, D'Albiac and Glubb, wearing their greatcoats with upturned collars, stamped their feet to keep warm as they waited for the Iraqi delegation.
The Iraqis had cut all the canal banks in the vicinity, and water was lapping the road in the grey dawn. With waterfowl flapping overhead, the scene, to Glubb, was more reminiscent of the Norfolk Broads than of the City of the Caliphs.
'Cornwallis has recommended that as part of the negotiations we allow the Iraqi Army to keep all their weapons and equipment and let them return to their peacetime locations. He believes that this will help them to keep face by suggesting that the war was against Rashid Ali and his friends, and not the Iraqi people. That would make for simple terms which the Iraqis can agree to quickly. What do you think?' Clark said, turning to Glubb.
'A quick agreement would cover up the fact that we have just carried out a colossal bluff and do not have enough troops to finish off the Iraqi Army. It would also put the Regent in power and get things back to normal as quickly as possible.'
'And you, John?'
'It makes sense. However, I don't think that your troops and my airmen are going to be very happy. The Iraqis will seem to have got off scot-free.'
'Well, that seems to be the word from the Foreign Office, so we will have to go along with their plan, and Kingstone and his men and your pilots will just have to lump it.'
'That's true, but all of this would never have happened if the Foreign Office had done its job in the first place', replied D'Albiac.
Two Iraqi officers, Colonel Ahmed Aquid and Captain Ghazi Daghistani, arrived promptly at four o'clock in a civilian car and walked across the bridge to de Chair and Gerald de Gaury. de Gaury, with a long history of the Middle East, had been appointed Chargé d'Affaires to the Regent by the Foreign Office and had flown in to Habbaniya the day before with the Regent and his retinue.
As the two groups came together, de Gaury and Daghistani, who had met on a course at Sandhurst, incongruously greeted each other as old friends.
Clark asked de Chair for some paper, sat down on the grass by the side of the road and wrote on the back of de Chair's telegraph form:
Your Excellency,
AOC and self have met the delegates. I have ordered the advance to stand fast at 04.00 hrs. I have here in my possession the instructions of my C-in-C General Wavell regarding terms. AOC and I would be grateful if you could see your way to come out here for discussion. We will await your instructions.
G. Clark, Major General
Clark looked at Daghestani and said, 'I will keep Colonel Aquid here as a hostage while you take Lt de Chair, Capt Spence and Mr de Gaury to the Embassy. with my message.'
They boarded Daghestani's car and he asked them to blindfold themselves as they passed through the Iraqi defence lines. As they entered the city he told them to remove their blindfolds. They drove quickly through the quiet streets of Baghdad, almost deserted save for groups of patrolling Iraqi soldiers, to the Embassy.
It took some time before Vyvyan Holt, the Oriental Secretary, wearing a plaid dressing-gown, let them in through the wooden postern gate. They found Cornwallis in bed and gave him Clark's message.
'Right, give me half an hour to dress and I will be with you. In the meantime I expect that they will find you some tea at the canteen, which we have here for the people on night guard.'
They left the Ambassador to dress, and Holt showed them down some stairs into the courtyard, through the garden and into the canteen. A buxom lady poured them tea and with tears in her eyes said, 'We have been waiting a long time for this.'
People, who seemed to be sleeping all over the Embassy grounds, began to get up and to crowd around them, sensing that their ordeal was coming to an end.
Cornwallis emerged from the main entrance, wearing an immaculate white tropical-drill suit topped-off by a white topee, and beckoned to de Chair, de Gaury, Spence and Daghestani to join him.
With Cornwallis and Daghistani in the front, and de Chair, de Gaury and Spence squashed in the back of the car, they drove back to the Iron Bridge. This time it took longer to pass through the Iraqi defence lines. After some argument and clicking of rifle bolts, which seemed very ominous to the three blindfolded men in the back, Daghistani was cleared to take them back to the bridge.
On their arrival, a shade before six o'clock in the morning, Cornwallis hurried across to the British side, joining Clark and D'Albiac in the latter's Air Force blue Humber saloon, while the others were left to wander about.
An hour or so later, Clark emerged and called de Chair over to dictate the armistice agreement which the three had worked out in D'Albiac's car. When the draft was shown to Glubb, he embarrassingly pointed out that it did not provide for a cessation of hostilities. After a further session in the back of D'Albiac's car, a final text was dictated on the back of another of de Chair's telegram forms:
Whereas representatives of the Iraq Army have sought for an Armistice and in view of the fact that His Highness the Regent Abdul Ilah is on his way to reign over the country, the following terms must be observed by the Iraq Army. The mildness of the terms will be noticed because the British Army has, in fact, no quarrel with the Iraqi people or their Army, but only with the pro-Axis adventurers who were determined to use the Army for their own end:
All hostilities between the two armies cease forthwith.
The Iraq Army will be permitted to retain all its arms.
All British prisoners of war (military, RAF and civilian) to be released forthwith.
All enemy (German and Italian) service personnel to be interned and their war material retained by the Iraqi Government pending further instructions.
The town and vicinity of Ramadi to be evacuated by the Iraq Army by 12.00 noon on 1st June, 1941.
All Iraqi prisoners of war now in British hands to be handed over to the Regent as soon as the terms in paragraphs 3, 4 and 5 are complied with.
Cornwallis took one copy and returned across the bridge for Daghestani and Aquid to take him to the Citadel to lay the terms before the Committee of Internal Security.
Early in the afternoon, al Umari, together with Brigadier Ismail Namiq, Colonel Hamd Nasrat and Colonel Nur el din Mahmud, came to the Embassy to discuss terms. At first they wished to make some textual alterations, with the intention of making the armistice appear to have been the result of mutual agreement, and not the terms offered by the victor to the vanquished. Cornwallis refused to debate the issue, indicating that he had no authority to amend the terms offered by Clark and D'Albiac. After some discussion, in which Cornwallis was supported by al Umari, the Iraqi officers signed the armistice terms.
As the sun began to go down, the Air Striking Force dropped hurriedly printed leaflets on Iraqi positions at Amara, Diwaniya, Nasiray and al Qurna, setting out the terms of the armistice.
The news of the armistice reached Paul Knabenshue at the United States Legation a few minutes after four in the afternoon.
He looked back on the siege with a mixture of pride and relief.
Pride because he had refused to submit to intimidation. Midway through the siege, the American University in Beirut had expelled all Iraqi students, and in retaliation the Iraqis threatened to arrest and intern him. With no radio he had asked the Iraqis to send a telegram to Washington requesting his replacement. Knowing that the Iraqis would read his message, he banked on Arab pride by informing Cordell Hull not to judge the Iraqis too harshly as they had only temporarily forgotten their well-known manners. The telegram was never sent and the matter was quietly dropped.
Relief because all of his wards, American and British, had come through the ordeal unharmed. The ten women inside had encountered no problems. He was amused when he recalled that the King's attractive English governess, Betty Sulman, came to ask his advice on how to reject four proposals of marriage without offending her suitors.
For Freya at the British Embassy, the end of the of the siege was a time for reflection.
The Colonel, Aquid, who had come the Embassy with Daghestani, was aggressive and quite rude, reaffirming her belief that the Iraqis have the worst manners of all the Arabs.
People like Bill Bailey, who had been seized in the street, the BOAC staff from Lake Habbaniya and British prisoners of war who were brought to the Embassy, came with tales of vicious treatment. Some had been chained up by both hands for the first six days. None had received medical attention, and all shared cells with convicted criminals. Most were thin, looked strained and carried the marks of severe beatings. Reports began to come in of homes systematically looted, then defiled with excrement, and what was not stolen was destroyed. Dr Sinderson, for example, whose house was thoroughly looted, was amazed to see after an interval his wife's dresses appearing at public balls and concerts.
The terms of the agreement seemed to be very lenient, and Freya wondered if not bringing in the Regent with a great show of force seemed to be placating one's enemies rather than encouraging one's friends. The talk around the Embassy was that it was all over. Just seeing the faces of Iraqis, however, brought home the fact to Freya that the revolt had not changed their attitudes. She knew that it was just a question of time before there was another revolt.
Habbaniya
Around noon Savile authorized Paul Holder to secure replacement stocks to resupply Habbaniya with over 100 tons of bombs and 250,000 rounds of ammunition which the RAF had expended during the revolt. He also requested a full set of checks on the remaining Air Striking Force aircraft, most of which were barely flyable after some 1,600 sorties within a month.
Tony Dudgeon's last photo-reconnaissance mission ended just after three. He had been tasked to photograph the Iraqi Army units around Falluja, Ramadi and in front of Baghdad at Khadimain, at the Abu Ghraib Regulator and at the Washash Canal. These were needed to check that the Iraqis were indeed observing the cessation of hostilities and were evacuating their positions and begining to return to barracks. This was also Frankie's last 'operational' mission.
Berlin
To General Hellmuth Felmy, Athens-Tatoi:
All forces are to gather at Aleppo, They are to stand down until further intentions have been clarified with the French Government in Vichy. Remove all Iraqi insignia. Further orders will follow.
Feldmarschall Wilhelm Keitel
No further orders were ever sent from Berlin.
Epilogue
The war in Iraq in 1941 was lost by those who hesitated, miscalculated and were misled and won by those who had nothing to lose and who were prepared to take massive gambles.
The Iraqis hesitated at Habbaniya, which they could easily have overrun, passing the initiative to the British. The Germans and Italians hesitated in sending adequate support to Rashid Ali, which allowed an inferior British force to prevail.
The Iraqis miscalculated that the British would resort to negotiation to free their hostages surrounded at Habbaniya and besieged at the Embassy and would not fight. The lack of international political experience of the Iraqi leadership led to unfounded optimism concerning the foreign support that they could expect from other states – particularly Saudi Arabia and Turkey – and they conspicuously failed to mobilize pan-Arab support throughout the Middle East to their cause.
Rashid Ali miscalculated that he could balance the demands of the Mufti, a malevolent zealot, and the nationalists in the Golden Square with British needs.
The Germans wrongly assumed that the strength, morale and fighting spirit of the Iraqis was enough to contain the British until their forces could intervene in strength. They sent only a token force, which was poorly prepared and inadequately supported. Hitler made a crucial mistake in thinking that Crete was more valuable strategically than Iraq. For far less than the huge casualties and loss of equipment sustained in Crete the Germans could have comfortably taken Iraq. A great prize which was theirs for the taking at little cost slipped from their grasp.
The Mufti misled the Iraqis by urging them to confront the British for his own ends. A climb-down by the British over Iraq and negotiated changes to the British policy on Palestine would have elevated the Mufti to unassailable leadership of the pan-Arab world. The Iraqi leadership also misled the population. True, many would have gone along with a victory over the British, but the leadership did not have a deep under standing of Iraq's religious and political factions. Many of the Shia tribes and the Kurds were either neutral or against fighting a war which was not their concern or in their interest.
Grobba misled the Wehrmacht, the Abwehr and the Auswartiges Amt on the true strength and capabilities of the Iraqi armed forces. The Germans were economical with the truth about the extent of their commitment, which encouraged the Iraqi leadership to begin hostilities before support was in place, while the Iraqis misled the Germans on their ability to supply high-octane fuel and quality lubricants.
The airmen and soldiers who fought at Habbaniya and went on to take Bahgdad won because they had nothing to lose and everything to gain.
The guiding forces were Churchill and Auchinleck, who saw the strategic implications. Losing Iraq would have had untold consequences. It would have inspired other ultra-nationalists in the region – Zahir Shah and Abdul Majid Khan in Afghanistan, King Farouk and Azil el Misri in Egypt, Subhas Chandra Bose in India and Reza Shah Pahlevi and Bandar Shah in Persia – all of whom had been courting or were being courted by the Germans to revolt.
Losing Iraq would have meant losing the Middle East, Britain's oil supplies in the sterling area and the fight for Roosevelt's support, which in turn meant the battle for American public opinion against the isolationists.
The real heroes at Habbaniya and on the march on Baghdad were the airmen – like Dudgeon, Ling, Hawtrey, Savile, Cleaver, Evans, Casano and Page – and the soldiers, Clark, Kingstone, Roberts, Brawn, Graham, Everett, Ferguson, May and Glubb – who were all prepared to take massive gambles.
While 31 May 1941 was the end of the revolt, in many ways it was the beginning of Iraq's troubles. That night Baghdadis came out onto the streets and, egged on by the Futuwwah and other nationalists, killed a large number of Jews in the streets and in their homes. As far as anyone knows, between two and four hundred Jews died, and up to a thousand were injured that night and over the next two days in what came to be known as the Farhud.
The streets were full of hatred. Rioting and looting erupted throughout the city by a populace poisoned for years by a fascist education system, radio propaganda and the problems in Palestine. Humiliated by their defeat, they went looking for scapegoats, and selected the Jews and anyone who tried to help them.
The Iraqi police and the Army were either part of the Farhud or turned a blind eye, and the 1,500 British soldiers camped outside the city were powerless to protect the Jews from the rage of a city of 400,000.
The Regent arrived in the city the next day and put together a cabinet that would support him. His overriding objective was to hunt down and execute or dismiss and humiliate any officers even vaguely connected with the revolt, in order to protect the royal family. However, the Regent had lost the loyalty of many Iraqis, and his actions merely created hatred for the monarchy and pushed nationalism in the Army underground. The wave of executions, arrests, trials and dismissals had a deep and lasting effect on the Iraqis. The atmosphere of fear that this created bred a deep resentment against the Regent, the royal family, Nuri al Said and the British. al Sabbagh's memoirs, written during his imprisonment in Turkey and published posthumously, inspired a generation of nationalists within the Iraqi armed forces.
This poisonous environment provided a breeding-ground for revenge. Many of the young Army officers dismissed by the Regent came back to haunt him – Ahmad Hassan al Bakr, for example, later became President of Iraq. Young officers who were cadets in the Iraqi Military Academy in 1941 were involved in the bloody coups of Abdul Karem Kassem in 1958 and Abdul Salam Arif in 1966, and they supported the Baath Party coup in 1968 and Saddam Hussein's rise to power in 1979.
The Farhud was the beginning of the end for the 2,600-year-old Baghdad Jewish community, one of the oldest and richest in the world, and one which had brought great cultural achievements and wealth to Iraq. The final nail in the coffin was the Arab-Jewish war of 1948 and the humbling performance of the Iraq Army. From then on the Jewish community was systematically humiliated and their property confiscated. Arrests, trials and even executions became commonplace. By 1951 virtually all the remaining Jews, some 120,000, had been airlifted, or had been smuggled, out of Iraq carrying a few possessions. They, and the great families like the Khadhouries and the Sassoons, made countries like Britain and America wealthier with their talents.
Despite the huge strategic significance of the war in Iraq in 1941, very little is known about what actually happened. In some respects this is understandable. From the British perspective the war being fought in the North Atlantic to protect Britain's food, oil and munitions from Doenitz's Wolfpacks and the Condors, Rommel's successes in the Western Desert, the loss of Crete and the follow-on campaign in Syria dominated the public consciousness.
Also helping to deflect attention was inter-service rivalry. D'Albiac's 'definitive' report, written second-hand after the battle for Habbaniya and the capture of Baghdad, rightly praises the RAF but contains little mention of the role of the Army. With a split command, no soldier of any stature and a diverse set of units, the role of the Army remains largely in the historical archives of the participating regiments.
From the German and Italian perspectives the lack of attention is also understandable, as few wish to dwell on their misfortunes.
Using any yardstick but the war in Europe, the strategic issues were exceptional. Those thirty-one days in Iraq in May 1941 produced extraordinary acts of courage and incredible improvisations in the air and on land, which rivalled anything in Europe and later in the Far East. The tactical lessons learned by the British were to become immensely valuable. Air resupply and air-landing troops behind enemy lines under fire soon became the norm. Montgomery's use of the Desert Air Force from late 1942 and the 2nd Tactical Air Force in 1944–45 to support ground troops from Normandy through to the North German Plain vindicated the use of combined operations.
A faraway place with unfamiliar people and unpronounceable names, and a steady stream of fresh disasters and setbacks, drove the victory in Iraq in 1941 from the news within a matter of days. There has been no official recognition or honours given to anyone on either side, and few, if any, who participated in this epic struggle are alive today.
Sadly the memory of a battle that, if lost, could quite possibly have changed the course of the war and the world has just faded away, like the participants, and has become just another very small, overlooked footnote in history.
Appendix
What Happened to...?
People follow very different paths to their future, and their pasts are largely poor guides to their later lives.
The characters in Hitler's Gulf War are no exception. A few went on to bigger and better things – for some this was the crowning point of their lives – while many others just faded into obscurity. Some had long and productive lives into their 90s while others had their lives cruelly taken from them.
Here are the post-1941 biographies of the key participants that I have been able to trace.
Otto Abetz, the German Ambassador to Vichy France, left Paris ahead of the Allied invasion in 1944. He was captured in the Black Forest in 1945 by the US Army and was sentenced in 1949 by a French court to twenty years′ hard labour for his part in the deportation of French Jews to the death camps. He was released in 1954 from Loos prison, and died in a car accident near Langenfeld on the Cologne-Ruhr Autobahn in May 1958. He was 55 years old.
Prince Abdul Ilah pursued and punished all those officers who had mutinied against the Kingdom in 1941. Death, imprisonment, and loss of status and pensions created a great hatred for the Regent. Saddam Hussein, although a child at the time, was touched personally by the events of 1941. His uncle, Khayrallah Tulfah, was an officer who had participated in the war and whose career was ruined by the Regent. Following the Kassem coup in 1958, the Regent's mutilated body was symbolically hung up at the gate of the Ministry of Defence at the spot where the Regent had had Salah ed Din al Sabbagh hanged in 1945. The Regent was 46 years old.
Yacob Khoshaba Aboon of the RAF Levies went on to serve in Basra and Palestine, and was awarded the MBE. At the disbandment of the Levies in 1955 and the handover to the Iraqi government, Aboona was the senior serving Levy officer. Following his release from military service, he moved to Baghdad and started his own business. He emigrated to the United States in 1981 and died in Skokie, Illinois, in 1984, aged 84.
John D'Albiac was appointed to command 2 Group at the end of 1942, followed by the 2nd Tactical Air Force in Europe. At the end of the war as an air marshal he was knighted. He became the first Commandant of London's Heathrow airport. He retired in 1961 after serving for ten years as Deputy Chairman of the Air Transport Advisory Council. D'Albiac died aged 69 in Beaconsfield in August 1969.
Butch, Casano's devoted black-and-white mongrel, was by his master's side everywhere, from action in Vichy-controlled Syria into the hallowed ground of the Long Bar at Shepheard's Hotel in Cairo, where dogs, except Butch, were not allowed. Butch was killed by the same burst of enemy fire that wounded Casano in the Western Desert. Such was his fame that Butch's demise was reported in the Egyptian Gazette.
John George Walters Clark went on to command in the Syrian Campaign which followed the fall of Baghdad. He briefly commanded the 10th Armoured Division in North Africa and then held administrative posts in the Mediterranean and led missions to the Netherlands and Greece. Clark retired from the Army with the honorary title of Lieutenant-General in 1946, and died after a brief illness in 1948 at the age of 56.
Dudley Wrangel Clarke remained in the Mediterranean theatre with A Force for the rest of the war. Clarke was one of the most influential soldiers of the Second World War. He thought up the name SAS for a phantom army in North Africa which became a reality, and gave Wild Bill Donovan the name ′Rangers′ for US Commando forces. Clarke was the ideas man behind a number of other successful deceptions, such as ′The Man that Never Was′ and ′I was Monty′s Double′, both of which were made into successful films. Clarke was arrested in Madrid in full drag in October 1941, in circumstances which have never been explained, while on a secret mission to Spain. He retired as a brigadier, and was appointed Head of Public Opinion Research for the Conservative Central Office, where he served from 1948 until his retirement in 1952. In the early 1950s the Australian Secret Intelligence Service used Clarke to produce a report on methods to combat communist subversion in South-East Asia using propaganda, economic warfare and covert operations. He later became a director of Securicor. Clarke died at the age of 75 in 1974.
Michael Peter 'Cass' Casano continued to command No. 2 Armoured Car Company RAF, and played a very active role with Habforce in the invasion of Vichy-controlled Syria, for which he was awarded the Military Cross for his inspiring and courageous leadership. He was badly wounded by the same burst of enemy fire in North Africa that killed Butch. At Glubb Pasha's request, Casano was transferred to command the Arab Legion's armoured-car force, and he remained in Transjordan for the rest of the war, marrying the daughter of a White Russian general. Casano was granted a permanent commission in 1945 in administration, serving in Germany, Malta and the UK. Retiring from the RAF in 1958, he became a driving instructor. He died in August 2006 at the age of 93.
Somerset de Chair was wounded in the Syrian campaign in late 1941 and invalided out of the Army. He was in and out of parliament several times, became a shrewd collector of antiques and paintings, bought a succession of stately homes and was a prolific writer. Married four times, de Chair died in Antigua at 83 in January 1995.
Gian Galeazzo Ciano, Mussolini's Foreign Minister and son-in-law, began to lobby for Italy's withdrawal from the conflict following the Axis defeat in North Africa, the Allied landings in Sicily and the Wehrmacht's setbacks in the East.
He was arrested by Mussolini's Italian Social Republic police for treason. Under intense German pressure he was tried and executed by firing squad in Verona in January 1944.
Richard 'Dicky' Cleaver went on to fly Hurricanes with 216 Squadon in Ceylon, achieving a second victory – this time against the Japanese. As a supernumerary squadron leader, Dicky joined 611 Squadron in the summer of 1944. He died when his Spitfire inexplicably crashed into the sea in June that year off the Normandy coast. He was 25 years old.
Kinahan Cornwallis remained British Ambassador in Baghdad until his retirement in 1945. He was then appointed Chairman of the influential Middle East Committee of the Foreign Office and became a Director of the British Bank of the Middle East. Cornwallis died in June 1959 in Basingstoke at the age of 76.
Ghazi Muhammud Fadhil Daghistani, the captain escorting de Gaury and de Chair to the Embassy, remained in the Iraqi Army, and in 1952 was appointed Military Attaché in London. In 1957 he was promoted major-general, but when the monarchy was overthrown in 1958 Daghistani was arrested and sentenced to death by a military court. His sentence was subsequently commuted to three years' imprisonment, and in 1960 he was released and came to live in London, where he died in 1966, aged 56.
Henri-Ferdinand Dentz, the Military Governor of Vichy Syria, was evacuated to France at the end of the Syrian campaign in late 1941 as part of an agreement with the Vichy French. At the end of the war he was arrested, tried and convicted of treason by the French government and sentenced to death. De Gaulle commuted the sentence to life imprisonment. Dentz died in Fresnes jail in December 1945 at the age of 64.
John Patrick Domvile, the Air Liaison Officer in the British Military Mission to Iraq, who was caught at the Embassy siege, was promoted to wing commander in 1942. He resigned his commission in September 1945, retiring as a group captain. He later held a teaching post at the Royal Military Academy, Sandhurst. On a visit to Baghdad in 1958, he was caught up in the Revolution, and was forced to flee for his life. His possessions and money due to him were confiscated, and no repayment was ever made.
William Joseph Donovan went on to found the Office of Strategic Service (OSS) in 1942. After Roosevelt's death in 1945 his personal connexion with the President was broken and his position weakened as a Republican in a Democratic administration; he returned to his Wall Street law firm – Donovan, Leisure, Newton & Irvine. His last official duty was serving as special assistant to Chief Prosecutor Telford Taylor at the Nuremberg War Crimes Trials.
Donovan was the only American to have received all four of the United States' highest awards – the Medal of Honor, the Distinguished Service Cross, the Distinguished Service Medal and the National Security Medal.
Donovan died in early 1959 at the age of 76, and is buried in Arlington National Cemetery. When he heard the news, President Eisenhower remarked, 'What a man! We have lost the last hero.'
Martin Drewes, the leading fighter pilot in 4/ZG 76, later became an 'Experte' in the night-fighter role over Germany. He ran up an impressive score of forty-nine victories, mainly against RAF Lancasters, and ended the war as Gruppen Kommander 111/NJG 3, being awarded the Knight's Cross with Oak Leaves. At the end of the war he emigrated to Brazil, first working as a pilot and then as a businessman.
Tony Dudgeon, despite his unconventional approach to authority, retired from the RAF as an air vice-marshal in 1968. He spent the next ten years as head of McKinsey in Paris. He fought tenaciously to have battle honours awarded to No. 4 Flying Training School for its contribution to the war. This was continuously rejected with the argument that the Flying School was a non-operational unit. Dudgeon died in Paris in July 2004 aged 89. His son is a wing commander in the RAF.
William Henry Dunwoodie left Habbaniya with 94 Squadron on 31 May and flew to Amman, forming part of X Flight, which took part in the conflict with the Vichy French in Syria. He was awarded a DFM in 1941, was commissioned as a pilot officer in 1943 and was a flight lieutenant on his discharge from the RAF in 1946.
Robert Anthony Eden, who tried unsuccessfully to persuade al Suwaidi, the Iraqi Foreign Minister, to have al Hashimi replace the officers of the Golden Square, remained in politics after the war. He was knighted in 1954, and succeeded Winston Churchill as Prime Minister in 1955. However, his famous diplomatic flair failed him. His conduct of the 1956 Suez Crisis proved disastrous, leading to the end of a Britain as the most important player in the Middle East, to the country′s demise as a superpower, and to his own continuous health problems. Eden retired in 1957, ostensibly for health reasons. He was ennobled as the Earl Of Avon in 1961, and died in Salisbury in 1977, at the age of 79.
Edward Newbey Everett (-Heath) spent a year recovering from wounds he sustained at Falluja. He rejoined the King's Own Royal Regiment in the Western Desert and was captured at Sollum in 1942. A POW for the rest of the war, he had a number of attempted escapes and a considerable period of 'freedom' to his credit. He was promoted colonel in 1950 and retired from the Army in 1952. He died in Camberley aged 74 in November 1976.
Hellmuth Felmy remained in Athens following the disbandment of Sonderstab F, and became Commandant of German forces in southern Greece. He was captured by the US Army in May 1945 in Germany, and was tried in 1948 at Nuremberg for war crimes involving reprisal killings in the Balkans while Commandant of southern Greece. He received a sentence of fifteen years, but was released in 1951 after serving only thirty-four months. Felmy died in December 1965, aged 80. His son, Hansjoerg, became a leading German actor.
Andrew Henry Ferguson, who led Kingcol's northern column, remaind in the Life Guards until his retirement. His son Ronald followed the long family tradition of service in the regiment. Ferguson's granddaughter, Sarah, was briefly married to Prince Andrew, Duke of York. Andrew Ferguson died aged 67 in 1966.
Frankie, the dachshund, the veteran of over thirty 'unofficial' bombing missions, accompanied Tony Dudgeon to Cairo in June 1941. One evening at around midnight at the southern end of Gezira Island, Frankie was hit by a passing car, which did not stop. Frankie died of his internal injuries two days later.
Rashid Ali al Gailani, the Iraqi Prime Minister, fled first to Persia during the evening of 29 May 1941, and then to Berlin, where he cooperated with the Germans.
At the end of the war he lived in exile in Saudi Arabia, returning to Iraq in 1958 following the revolution overthrowing the monarchy. In December 1958 he was implicated in a plot against Abdul Karem Kassem, and was sentenced to death for treason and plotting against the regime. His sentence was commuted to life in prison, and he was released by special amnesty in 1961.
Rashid Ali settled in Beirut, where he died in August 1965 at the age of 73. Even though he lived in the same city as Haj Amin for a number of years, it is not believed that they met.
Gerald de Gaury, Chargé d'Affaires at the Regent Abdul Ilah's court in exile, rejoined the Army, and as a lieutenant-colonel raised and commanded the Druze Cavalry, which took part in the Syrian campaign. Further assignments in the Middle East continued, and he travelled extensively into the early 1960s, maintaining his friendship with the ruling families of Saudi Arabia and Kuwait. de Gaury published twelve books on his journeys in the Middle East and biographies of its rulers. He died in Brighton in mid-January 1984.
John Bagot Glubb, commanding the Arab Legion, was devoted to the Hashemite dynasty, the people of Jordan and the Arab cause throughout the rest of his life. Despite his loyalty, he was discourteously sacked as head of the Arab Legion in 1956, and given two hours to leave Jordan, arriving in England with £5 in his pocket and little more in the bank. Although he was knighted, the British government washed its hands of Glubb, offering no employment or asking for advice from a man many believed to be far greater than T.E. Lawrence. He supported his family through lecturing and writing nineteen books. The Pasha died in March 1986 at the age of 88.
Alistair Graham, of the Levies, was awarded the Military Cross for his part in the battle for Falluja. He returned to the Green Howards and took part in the Normandy landings, and was in action in north-western Europe. He was promoted major in 1948 and retired in 1961. Graham died after a long illness at the age of 60 in 1975.
Hans-Ulrich Granow, who flew to Baghdad with Grobba, re-entered the German Foreign Service after the war and served as Ambassador in Malaya, South Africa and Sweden. He died at his post in Stockholm in March 1963 of a heart attack. He was 62 years of age.
Nigel St George Gribbon was wounded at Felluja and remained in Iraq through 1942. His subsequent career took him through staff colleges in Canada, India and with the RAF, operational tours with the Parachute Brigade, the King′s Own and 161 Brigade, and in intelligence roles. He retired from the Army in 1972 as a major-general, with a last posting as Assistant Chief of Staff (Intelligence) with SHAPE. His post-Army career combined high-profile private sector and public service appointments. He died in Sussex aged 91 in early 2009.
Fritz Konrad Ferdinand Grobba, after making his way back to Berlin, became embroiled in a feud between Rashid Ali and the Mufti for leadership of the 'Arab Cause'. Unwisely Grobba chose the 'politician' over the 'fanatical idealist' whom the SS supported, believing that the Mufti had greater influence across the Arab world. Grobba was outmanoeuvred by the far more influential and ruthless SS, who had him dismissed in 1943 and sent to France to analyse documents. Grobba was pensioned in 1944 and joined the Saxon government in Dresden before being appointed Public Prosecutor for Meiningen. Following his capture by the Russian Occupying Forces, who believed him to be a spy, he spent ten years in captivity.
On his return to Germany, Grobba cooperated with the CIA, and, it is claimed, during his capitivity in Russia with the KGB, providing his network of contacts in the Middle East. He later made several trips to Iraq, Lebanon, Saudi Arabia and Syria, promoting a German-Arab Society based in Bonn. Grobba published his memoirs in 1967 and died at the age of 87 in Bad Godesburg in September 1973.
Jacques Guerard returned to Vichy and became Secretary-General to Prime Minister Pierre Laval. He fled to Spain ahead of the Allies in 1944 and was condemned to death in absentia. He returned voluntarily in 1955, and by the time of his trial in 1958 the French government had lost its appetitie for revenge and he was handed out a five-year suspended sentence. He died in 1977.
Habbaniya. Control of the air base reverted to Iraq in May 1955, but it was still used extensively by the RAF. Kassem's bloody revolution in July 1958 made a continued RAF presence untenable. The RAF ensign was finally lowered on 31 May 1959. The Iraqi Air Force took over Habbaniya, but abandoned the air base after the 1991 Gulf War.
Today Habbaniya lies derelict. The cast-iron road signs are still there, but the English text on top has been covered in black paint and the Arab script on the bottom is highlighted in white. In all RAF stations the curbs along the streets are painted white: in Habbaniya the colour has long since gone. Now they are all faded, chipped and scuffed.
The officers' mess, where Ling and Dudgeon met to plan the rebellion against stupidity, is rather forlorn, and deserves a coat of paint, like most of the buildings. The two-storey administration block with its inner balcony encloses a grass quadrangle. The grass has long since gone, and through neglect the roots of the trees at each coner of the quadrangle have been allowed to damage the structure. The Iraqis had no interest in tennis: weeds and bushes now sprout up through the courts. Incongruously there are a number of wooden VHF aerials dotted around the camp, all of which seem to have storks' nests perched on their tops.
When the US Marines took over Habbaniya in 2003 they found a terrible mess. The building that once housed the St George′s Anglican church had been turned into a mosque, the Astra cinema sported a mural of Saddam Hussein, and the 290 graves where those who had sacrificed their lives in the defence of Habbaniya, at Falluja and in the march on Baghdad lie buried were neglected and vandalized.
A group of servicemen from the US Army, Air Force, Marines and Navy, together with some British, Nepalese and Philippino civilians, formed a volunteer work-party to clean up the graveyard.
Remembrance Day services have been held at Habbaniya since 2005, and in 2008, at a ceremony attended by the British Ambassador to Iraq and the top US Marine general in Iraq, the cemetery was rededicated, a ceremonial wreath was laid and two minutes of silence were observed.
John Gosset Hawtrey remained in the RAF and took a number of staff postings, including command of RAF stations in the UK and overseas. Early in 1952, following a promotion to air vice-marshal, he was appointed AOC Air Headquarters Iraq. Hawtrey died on holiday in San Remo, Italy, at the age of 53 in 1954.
Otto-Werner von Hentig, Head of the Political Department responsible for the Middle East at the German Foreign Office, joined the Foreign Service of the post-war German government and served as Ambassador to Indonesia in 1952–53. On his retirement from the Foreign Service, he was a personal adviser to the Saudi Royal Family for almost two years. Von Hentig died in August 1984 in Norway.
Paul Davie Holder had a successful career in the RAF which took him in and out of operations with Bomber, Coastal and Transport Commands and staff postings in the UK and overseas. He was promoted to air marshal and knighted in 1965, retiring from the RAF in 1968. He died aged 89 in November 2001.
Adrian Holman, Cornwallis's deputy, later served in Tehran, Algiers, Paris and Bucharest, becoming Ambassador in Cuba in 1952. He retired from the Foreign Service with a knighthood in 1954, and died in Hampshire in 1974 at the age of 79.
Vyvyan Holt, the Oriental Secretary at the Embassy, remained in post until 1944, when he transferred to the Embassy in Tehran as Oriental Counsellor. He was posted as Minister to the British Legation in Seoul in 1949. Believing in Diplomatic Immunity, he stayed at his post when the North Koreans invaded in 1950. He was detained and taken on a death march to Chung-Gang-jin. Holt was harshly treated, contracted pneumonia and was probably saved by the intervention of George Blake, the SIS agent who converted to Marxism and subsequently became a double agent for the KGB.
He was released in 1953, and took up his final post as Minister in El Salvador in 1954, retiring from the Foreign Service in 1956, when he was knighted. Holt died in 1960, aged 64, his harsh treatment by the North Koreans probably having precipitated his early death.
Cordell Hull, the longest-serving Secretary of State, retained his position to late 1944, eleven years after his appointment, until he resigned because of failing health.
He was the underlying force and architect behind the creation of the United Nations, and was awarded the Nobel Prize for 'cofounding the United Nations' in 1945.
Hull died in 1955 and is buried in the Washington National Cathedral.
Haj Amin al Husayni escaped on the evening of 29 May 1941 first to Persia, and then to Italy and Germany. He helped to organize a Bosnian Muslim Waffen SS unit in the Balkans and Arab saboteurs for action in the Middle East, and he broadcast radio propaganda on behalf of the Germans. Haj Amin was denied residence in Switzerland at the end of the war and fled to Paris. He escaped house arrest in Paris in 1946, and using false papers to travel to Cairo on an American aircraft he obtained asylum in Egypt, where he became one of the sponsors of the 1948 Arab–Israeli war. Haj-Amin set up a Palestinian government for eight days in Gaza in 1948, before the final collapse of the Arab armies and the annexation of the West Bank by Jordan. The Mufti was reputed to have been involved in the assassination of King Abdullah of Jordan in 1951. He was made a scapegoat for the Arab losses in 1948 and was marginalized by moderate Arab groups as well as extreme Palestinian organizations.
Haj Amin settled in Beirut in 1959 and died there, an embittered man, in July 1974, at the age of 77. His wish to be buried in Jerusalem was refused by the Israelis, and his tomb lies in a corner of the Palestinian 'Martyrs' Cemetery' in West Beirut.
Robert James, of the 1st Batallion Essex Regiment, was awarded the DSO for his actions at Falluja. Promoted lieutenant-colonel, he transferred to command the 5th Batallion East Yorkshire Regiment, and fought in Tunisia and Sicily, winning two more DSOs. He was given a staff appointment, but volunteered to rejoin the 5th Battalion to take part in the Normandy landings. James was killed at Amayre by a German shell in early August 1944.
Hans Jeschonnek was promoted to Generaloberst in 1942. Hitler grew increasingly critical of the Luftwaffe, and Goering abandoned Jeschonnek to his fate. The last straws were the massive RAF raid on Hamburg and, less than a month later, another RAF raid which devastated Peenemunde. Facing Hitler's wrath and his own recent personal tragedy with the deaths of his father, brother and brotherin-law, he committed suicide at the Fuhrer's Wolfschanze Headquarters at Rastenburg in East Prussia in August 1943. Jeschonnek was 53 years old.
Werner Junck. Misled by Grobba, Jeschonnek had Junck courtmartialled on his return to Germany. Junck was cleared, and his offers to resign his commission were rejected. Junck later distinguished himself as commander of Jagdkorps II, and he retired as a Generalleutnant in 1944. Junck died aged 81 in July 1976 in Munich.
James Joseph Kingstone, who commanded Kingcol, went on to command 30th Armoured Corps. He became ADC to King George VI in 1944, and although he retired from the Army in 1945 he remained Colonel of the Queen's Bays until 1954. He was made Deputy Lord Lieutenant of Wiltshire in 1953 and died in Warminster in 1966 at the age of 74.
Paul Knabenshue, the American Consul-General, contracted tetanus in Baghdad and died at his post in February 1942 at the age of 56. His Irish-born wife, Olive, wanted to return to Ireland but was refused permission by the Irish government, which said that she was an American citizen, despite never having lived in or visited the United States.
The Regent, Abdul Ilah, remembering how the Knabenshues had saved his life, granted Olive a pension large enough for her to live comfortably in Baghdad until the end of the war, when she could travel safely home.
RAF Levies remained, guarding RAF installations in the Middle East throughout the Second World War. Some joined a Levy parachute company which later served with distinction in Albania, Greece, Italy and Yugoslavia. The Levies were disbanded in Iraq in May 1955. A few decided to stay in Iraq, joined the Iraqi Army and became Iraqi nationals, but most Assyrians emigrated to Britain, Australia, Syria and the United States, which all host Iraqi Christians of different persuasions.
Christopher 'Larry' William Mitchell Ling was awarded the DFC in November 1945 and retired from the RAF as a group captain in June 1946.
Eitel Frederich Rodiger von Manteuffel, Junck's staff officer on Rhodes, later served on the staff of the OKL, then commanded Airbase 4/XIII and, on promotion to Generalmajor in 1944, led the German Mission to Rumania.He was captured by the Russians in Rumania in late 1944, and liberated in 1950. Von Manteuffel died in Germany in 1984, aged 88.
Raymund John Maunsell remained in Cairo until 1944 when he moved to a counter-intelligence post at SHAEF. On leaving the Army in 1948 as a brigadier, he directed Unilever's Information Department until his retirement in 1963. Maunsell died in 1976 at the age of 73.
Keith 'Steve' May of the 1st Battalion Essex Regiment was promoted lieutenant-colonel and captured at Deir el Shein in Cyrenaica in May 1942. He spent the rest of the war as a POW in Italy and Germany, with a number of escape attempts to his credit
Ernest ′Mel′ Melluish left Habbaniya in 1942 as a corporal, taking up duties as an armourer with Bomber Command at Waterbeech. He was scheduled for night duty, and swapped this with his best friend to take out his future wife. While a Lancaster was being bombed-up one of the bombs detonated, killing his friend. This stayed with Mel for the rest of his life. He demobilized in 1945 and worked as an agricultural, and later a heavy goods, mechanic. Mel died aged 81 in Norfolk in 2002.
Benito Mussolini was imprisoned in Gran Sasso in northen Italy following the overthrow of his Fascist regime in September 1943 and a switch in the new administration's alliegiance to the Allies. Spectacularly rescued by Otto Skorzeny on Hitler's orders, he became head of a German puppet government – the Italian Social Republic – in northern Italy.
Mussolini's past eventually caught up with him, and he was captured and executed by partisans on the banks of Lake Como with his mistress Claretta Petacci. Subsequently their bodies were hung upside-down from a girder in the forecourt of a Milan garage in April 1945.
John Joseph Joffre Page, the commander of No. 1 Armoured Car Company RAF at Habbaniya, became a liaison officer with Montgomery's staff, and rose to the rank of group captain, commanding RAF Halton before being demobilized in 1946.
He rejoined the oil industry, working with Iraq Petroleum in Palestine, Syria and the Gulf until his retirement in 1972. He then held a number of chairmanships and was knighted in 1979. He died in February 2006 aged 91.
Franz von Papen, the German Ambassador to Turkey, was acquitted at the Nuremberg War Crimes Trials of crimes against humanity, but was found guilty of political immorality, which was not a punishable offence. On release he was immediately rearrested by the German authorities in their de-Nazification programme, and was released in the late 1940s. Von Papen tried, unsuccessfully, to restart his political career in the 1950s. He wrote a number of books and retired to a castle in Upper Swabia. Von Papen died in Germany in May 1969 at the age of 89.
Baijan Peeku, who was wounded and won the Military Cross at Falluja, together with fellow Military Cross winner Stephan Nessan, represented the RAF Levies in the June 1946 Victory Parade in London. He died aged 88 in Sydney, Australia, in 1984.
James Maitland Nicholson Pike, Flight Commander of 203 Squadron, went on to command 185, 220 and 248 Squadrons. He joined the RAF Staff College as a wing commander in 1945, went on to command RAF bases at St Mawgan, Kinloss, Malta and Gibraltar as a group captain, and finished his service career as an air commodore in Intelligence in 1969. After his retirement he worked for the Ministry of Defence until 1978. Pike died in March 1999 at the age of 83.
Fawzi al Qawujki, Haj Amin's military aide and a colonel in the Iraqi Army in charge of irregular warfare, made a fighting retreat across the border to Syria. He was later wounded in the Syrian campaign and airlifted out by the Germans He married a German whom he met while hospitalized in 1942 in Berlin, and earned the enmity of the Mufti, who denounced him as a British spy when he supported Rashid Ali. In 1947 he returned to Tripoli via Paris after six years in exile in Germany. He served as field commander of the Arab Liberation Army during the Arab–Israeli war of 1948, fighting alongside his old adversary, Glubb Pasha, and emerged as a rival to Haj Amin. His experience as a successful guerrilla leader was ineffective on the battlefield, and his forces were driven out of Palestine into Lebanon in October 1948. al Qawujki died in Beirut in 1977 at the age of 87.
Rudolf Rahn, Ribbentrop's emissary, served later in the war as Reich plenipotentiary in Tunisia, Hungary and Italy. He conducted himself like a relatively discreet viceroy, focusing more on collaborating than on coercing, believing that a more subtle approach would provide greater benefits. Rahn died in Germany aged 75 in 1975.
David Raziel, of the Haganah, was buried in the military cemetery at Habbaniya on 19 May 1941. In 1952 his sister, Esther, contacted the Military Attaché at the British Embassy in Tel Aviv to find out what had happened to her brother. SIME records in Cairo had been destroyed, and it took some time to find his place of burial. His sister requested that the body be returned to Israel. In 1955 David Raziel's remains were exhumed from the Habbaniya cemetery and transferred to Cyprus. In 1961 he was reburied in Jerusalem at the Mount Herzle Military Cemetery.
Joachim von Ribbentrop, the German Foreign Minister, lost Hitler's favour following the 20 July 1944 bomb plot, when it was discovered that a number of those involved came from the Auswartiges Amt. The end of the war came just as Hitler was about to replace him with Artur Seyss-Inquart. Nevertheless, Ribbentrop remained a fanatical Nazi and a vociferous anti-Semite. He was tried and convicted by the Nuremberg Military Tribunal of crimes against humanity, and executed in October 1946.
Ouvry Linfield Roberts served in a number of staff positions in the Middle East and Ceylon until August 1943, when he was given command of the 23rd Indian Division. The 23rd was one of the divisions of 4th Corps which defended Imphal against the Japanese. Encircled, the 23rd fought a bitter battle for three months until the Japanese were forced to abandon their attack on India and fell back into Burma with heavy casualties. Roberts ended the war as Quartermaster-General, was knighted and later served as ADC-General to the Queen. Following a successful business career in Canada, he died at 87 in March 1986 at his home in Buckinghamshire.
Yunis al Sabawi, the most pro-Axis member of Rashid Ali's cabinet, and founder of the fascist Futuwwah movement, fled to Persia. He was brought back to Baghdad, tried and executed in 1942.
Salah ed Din al Sabbagh, a leading member of the Golden Square, who commanded the 3rd Iraqi Division, escaped to Persia on 29 May 1941. He later crossed the border into Turkey, dressed as a Dervish, and was arrested. He was extradited to Iraq and executed in 1945. During his imprisonment in Turkey he wrote his memoirs and thoughts about the thirty-one-day war, which were published posthumously. al Sabbagh's work inspired future nationalists like Abdul Karim Kassem, who was a young officer in 1941, and Saddam Hussein.
Fahmi Said, the member of the Golden Square who commanded the mechanized forces besieging Habbaniya, fled to Persia on 29 May. He was later extradited to Iraq and executed in 1942.
Nuri al Said served as Prime Minister of Iraq fourteen times. His support for Britain over the Suez crisis in 1956 severely weakened his position with the nationalists. A military junta led by Abdul Karem Kassem, who was a junior officer in the Iraqi Army in 1941, seized power in August 1958. The King and the Regent were assassinated at the palace, ending the Hashemite monarchy in Iraq, and Nuri was caught trying to escape in a Baghdad street the next day, disguised as a woman. He was killed on the spot. The mob mutilated his body.
Mahmud Salman, one of the four Colonels of the Golden Square who commanded the Royal Iraqi Air Force, fled to Persia but was extradited and executed with his co-conspirators in 1942.
Walter Archer Bourchier Savile retired from the RAF as a group captain in May 1944 and died aged 90 in June 1988.
Kamil Shabib, Commander of the 1st Iraqi Division and a member of the Golden Square, fled to Persia on 29 May 1941 with his colleagues. He was extradited from Persia and tried and executed in Iraq in 1944.
Naji Shawkat, Minister of Defence in Rashid Ali's cabinet in 1941, was captured by the British in Italy in July 1945 and handed over to the Iraqis to serve a five-year term of imprisonment. Shawkat died in Baghdad in 1979.
Glynn Silyn-Roberts retired from the RAF in November 1961 as an air vice-marshal, and died at the age of 77 in September 1983.
Harry Chapman 'Sinbad' Sinderson remained in Baghdad, continuing his many roles as physician to the Royal Family and Household, adviser to the Iraqi Ministry of Social Affairs, and as Dean and Emeritus Professor of Medicine in the Royal College of Medicine, which he had founded. In 1943 he and his wife created Noah's Ark, a welfare centre for Allied servicemen in Baghdad, of whom a million received a welcome in three years. He also introduced the Red Crescent and the Save the Children Fund to Iraq.
Sinderson was knighted on his retirement in 1946, and he continued to serve people through the Order of St John and local and national charities. He died aged 83 in November 1974.
Harry George Smart, evacuated from his command at RAF Habbaniya, was subsequently appointed AOC No. 17 Operational Training Group, and his promotion to air vice-marshal was confirmed in 1943. He retired from the RAF in September 1945. He emigrated to South Africa, and died at Simonstown in Cape Province in 1966, at the age of 75.
Leslie Ernest Smith left Habbaniya at the end of May with 94 Squadron, and was commissioned in August 1941 as a pilot officer. In 1943, as a flight commander in 609 Squadron, flying Hawker Typhoons on ′Rhubarbs′ across the Channel, he was forced to ditch off Deal in Kent after his aircraft sustained heavy flak damage over Brussels. Smith was awarded a DFC in 1944 and was a flight lieutenant on his discharge from the RAF in 1947.
Freya Stark continued her work with the Ministry of Information until the end of the War. Sent to the United States in 1943 to lecture on the dangers of imposing a Zionist state in Palestine, she received a bruising reception which stayed with her for the rest of her life. In 1945 Freya left public service, and in 1947 married Stewart Perowne, the Public Relations Attaché at the British Embassy in Baghdad during the siege. She accompanied Perowne on postings to Barbados and Libya, but they separated in 1952.
Freya continued to travel, with memorable visits to Afghanistan, India and Nepal. She was a true eccentric, who wore Dior in the wilder reaches of Asia and Arabic dress in London. Her will of iron, infinite patience and remarkable powers of persuasion helped her to overcome her lack of money and worldly advantage to become a widely read and admired explorer, photographer and author, who was courted by generals, intellectuals, politicians and royalty. She was made a Dame Commander by Queen Elizabeth II in 1972. Freya died at the ripe old age of 100 at her home in Asolo, Italy, in May 1993.
Betty Sulman (Morrison) remained the King's Governess until 1942, when she married John Morrison, a British intelligence officer, in Baghdad. After the war Morrison, a classicist, returned to a successful academic career. Betty championed the phonic method of reading, helping many children and adults with learning disabilities to read. Betty had five children and remained very active until her death in Cambridge in 2003 at the age of 94.
Naji al Suwaidi died in South Africa in 1942 during his deportation to Baghdad to face trial for treason.
Tawfiq al Suwaidi became Prime Minister of Iraq in 1946 and again in 1950. He died in Beirut in 1968.
George Guy Waterhouse, head of the British Military Mission in Iraq, retired in 1941, only to be recalled later that year. He served until 1945 as OC North-West District, finally retiring as a major-general. He died in 1975 aged 89.
Archibald Wavell was dismissed by Churchill in June 1941 and sent as C-in-C India, a relative backwater. His appointment coincided with the Japanese declaration of war, and he was made Supreme Commander of Allied Forces in South-East Asia just in time to oversee the British defeats in Burma, Malaya and Singapore, which Churchill saw as failures by Wavell, Now a field marshal, Wavell was offered another 'backwater' post, Viceroy of India, which he held from 1943 to 1947. He died in London in May 1950.
Ernst von Weizsacher, Secretary of State at the German Foreign Office, was appointed Ambassador to the Vatican in 1943. He was convicted by the Nuremberg Military Tribunal of crimes against humanity and sentenced to seven years. He was released in 1950 after serving eighteen months. He died in Germany in 1951 shortly after completing his memoirs. His son, Richard, became President of the Federal Republic of Germany in 1984.
Arthur Valerian Wellesley served throughout the war, winning a Military Cross in 1941. Promoted lieutenant-colonel after the war, he commanded, in turn, the Royal Horse Guards, the Household Cavalry, 22nd Armoured Brigade and the RAC 1st (British) Corps. Prior to his retirement in 1967 he served for three years as Defence Attache in Madrid. He succeeded his father as the 8th Duke of Wellington in 1972, was on the board of several companies and has remained active taking a leading role in many local and national associations and charities, and in animal welfare.
William Taylor Forest Wightman, with three victories flying Gladiators, was taken off flying duties after Habbaniya. He was promoted to wing commander in 1945 and to group captain in 1951. He retired from the RAF in late 1958 and died at the age of 80 in November 1989.
Henry Maitland Wilson led the successful Syrian Campaign against the Vichy French later in 1941, and in 1943 was appointed C-in-C Middle East. Despite a disastrous campaign in the Dodecanese, he became Supreme Allied Commander Mediterranean after Eisenhower. In 1945 Wilson was appointed a field marshal and was ennobled, taking the title of Baron Wilson of Libya. He finished his Army career as the Senior British Military Representative in Washington DC. Wilson died in Oxfordshire at 83 at the end of December 1964.
Ernst Woermann, Under-Secretary of State at the German Foreign Office, was convicted at the Nuremberg Military Tribunal of crimes against humanity, and was sentenced to seven years' imprisonment, which was later reduced to five years. He was released in 1951. Woermann died in Heidelberg in 1979 at the age of 91.
Amin Zaki (Sulaiman), the Iraqi Chief-of-Staff, fled to Persia on the night of 30 May 1941. He was returned to Iraq and sentenced to death in 1942, which was later commuted to life imprisonment.
Bibliography
Hitler's Gulf War has been based on a wide selection of books, articles, official documents and academic work.
The large number of works used to research this book are based on two issues. First I wanted to go beyond the purely physical aspects of the conflict to look at the people and the decisions that they made. I also drew widely on personal reminiscences in the air, on the battlefield, at the Embassy and at the various locations featured in Hitler's Gulf War. This was to help to explore, explain and put into context the factors driving the events which led up to those thirty-one days in May 1941, and the actions of Germany, Italy, Iraq and the United States, in addition to those of Great Britain, as well as the attitudes and experiences of those personally involved. All of these influenced the conflict and its outcome.
Secondly, virtually everyone involved is no longer alive, which means relying on published information. However, anything written about events which occurred almost seventy years ago comes with a set of in-built problems. Time has a subjective influence on the accuracy and the context of information. Works written at the time or shortly afterwards capture the detail and are sharply focused, although they tend to miss out on the wider perspective and are often written for purpose. Those written sometime later have greater perspective and see things in context, although they suffer from self-justification and important lapses of memory. Then there are those that reinterpret the information and those that confuse people, places and fact.
Nothing is black and white, it is merely shades of grey. I have therefore used as wide a variety of sources as possible in an attempt to avoid the most obvious pitfalls and to iron out the inconsistencies.
I have also been very fortunate to tap into the rich vein of Iraqi academic material. This not only gives more dimension but provides both the context in which decisions were made and a new level of detail of combat from the Iraqi perspective. Here are some of the numerous books and other literature worth reading on the subject, which I have used as references for this book:
Books
Aaronson, S., Hitler, the Allies & the Jews, Cambridge University Press, Cambridge, 2004
Allen, M., Arabs, Continuum, London, 2006
Al-Marashi, A. and Salama, S., Iraq's Armed Forces: An Analytical History, Routledge, London, 2008
Antonius, G., The Arab Awakening, Hamish Hamilton, London, 1938
Arthur, M., Lost Voices of the Royal Air Force, Hodder, London, 1993
Avon, The Earl of, The Eden Memoirs: The Reckoning, Cassell, London, 1965
Benjamin, M., Last days of Babylon, Bloomsbury, London, 2007
Birdwood, B., Nuri al Said: A Study in Arab Leadership, Cassell, London, 1959
Black, E., Banking on Baghdad, Wiley, London, 2004
Bloch, M., Ribbentrop, Crown, New York, 1992
Buckley, C., Five Ventures, HMSO, London, 1977
Bullock, A. (Introd.), The Ribbentrop Memoirs, Weidenfeld & Nicholoson,
London, 1954 Burnett, A. (Ed.), Iraq Defence Intelligence, Vol. 3 (1933–1941), Cambridge
Archive Editions, Slough, 2005
Candler, E., The Long Road to Baghdad, Vol. 1, Cassell, London, 1919
Catherwood, C., Winston's Folly: Imperialsm and the Creation of Modern Iraq, Constable, London, 2004
Central Office of Information, Paiforce: The Story of the Persia and Iraq Command, 1941–1946, London, 1948
Chaillot, C., The Syrian Orthodox Church of Antioch and all the East, Inter-Orthodox Dialogue, Geneva, 1995
de Chair, S., The Golden Carpet, Faber & Faber, London, 1943 Morning Glory, Cassell, London, 1994
Churchill, W.S., The Second World War, Vol. III, 'The Grand Alliance', Houghton Mifflin, Boston, 1950
Ciano, G., Diary 1937–1943, Enigma Books, New York, 2002
Cole, W.S., Roosevelt and the Isolationists 1939–1945, University of Nebraska Press, Lincoln, 1983
Cooper, A., Cairo in the War 1939–1945, Hamish Hamilton, London, 1989
Cowper, J.M., The King's Own, Vol. 3, 1914–1950, Gale & Polden, Aldershot, 1957
Dahl, Roald, Going Solo, Puffin, London, 1988
Dann, U. (Ed.), The Great Powers in the Middle East 1919–1939, Holmes & Meier, Teaneck, 1988
Doenecke, J.D., The Battle Against Intervention 1939–1941, Krieger, Malabar, 1997
Dodge, T., Inventing Iraq, Hurst & Co., London, 2003
Dudgeon, A.G., Hidden Victory, Airlife, Shrewsbury, 1991 The War That Never Was, Tempus, Stroud, 2000
El-Sohl, R., Britain's Two Iraq Wars, Ithaca Press, Reading, 1996
Epelag, Z., The Grand Mufti, Taylor & Francis, London, 1963
Everett-Heath, E.N. and Westworth, J.R., The Flying Fourth, KORR Museum, Lancaster, 2005
Felmy, H. and Warlimont, W., German Exploitation of Arab Nationalist Movements in World War II, Hailer, St Peterburg, 2007
Fisk, R., The Great War for Civilization: The Conquest of the Middle East,
Fourth Estate, London, 2005
de Gaury, G., Three Kings in Baghdad, Hutchinson, London, 1961
Geniese, J.F., Freya Stark, Chatto & Windus, London, 1999
Glubb, J., The Story of the Arab Legion, Hodder & Stoughton, London, 1948
Green, H., King's Own Royal Regiment, Leo Cooper, London, 1972
Grobba, F., Maenner und Machte im Orient, Musterschmidt, Gottingen, 1967
Gundelach, K., Kampgeschwader 'General Wever' 4, Motorbuch Verlag,
Stuttgart, 1978 Die Deutsche Luftwaffe im Mittelmeer 1940–45, Band 1: 1940–1942, Verlag
Peter D. Lang, Frankfurt-am-Main, 1981
Gunter, J., Inside Asia, Harper, London, 1942
Haddad, G., Revolutions and Military Rule in the Middle East, Vol. 1, Speller,
New York, 1971
Halbert, C., Il Duce, Little, Brown, Boston, 1962
Halpern, M., The Politics of Social Change in the Middle East & North Africa,
Princeton University Press, Princeton, 1963
von Hentig, W.O., Mein Leben eine Dienstreise, Vandenhoeck & Ruprecht, Gottingen, 1962
Hildebrand, K., The Foreign Policy of the Third Reich, University of California
Press, Berkeley, 1970
Hill, Roderic, The Baghdad Airmail, Nonsuch, 2005
Hinsley, F.H. et al. British Intelligence in the Second World War, Vol. 1,
HMSO, London, 1979
Hirszowicz, L., The Third Reich and the Arab East, Routledge & Keegan Paul, London, 1966
Holt, T., The Deceivers, Weidenfeld & Nicolson, London, 2004
Hooton, E.R., Eagle in Flames, Arms & Armour Press, London, 1997
Hopwood, D., Tales of Empire, I.B. Taurus, London, 1989
Horne, E., A Job Well Done, The Anchor Press, London, 1982
Ingram, E. (Ed.), National & International Politics in the Middle East, Frank
Cass, London, 1986
Jbara, T., Palestinian Leader Hajj Amin Al-Husayni, Mufti of Jerusalem,
Kingston Press, Princeton, 1985 Kadduri, M., Independent Iraq, Oxford University Press, London, 1960 Kattan, N., Farewell Babylon, Souvenir Press, London, 2007 Karsh, E., The Arab-Israeli War 1948, Osprey, Oxford, 2002 Kedourie, E., The Chatham House Version and Other Middle Eastern
Studies, Frank Cass, London, 1970 Arab Political Memoirs and Other Studies, Frank Cass, London, 1974 Kelidar, A. (Ed.), The Integration of Modern Iraq, Croom Helm, London,
1979 Kershaw, I., Fateful Choices, Allen Lane, London, 2007 Kirk, G., The Middle East in the War, Oxford University Press, Oxford, 1952 Kolinsky, M., Britain's War in the Middle East, MacMillan, London, 1999 Langer. W. and Gleason, S.E., The Undeclared War 1940–1941, Little Brown,
London, 1953 Lukitz, L., A Quest in the Middle East: Gertrude Bell and the making of Modern Iraq, I.B. Taurus, London, 2006 Lunt, J.D., Imperial Sunset: Frontier Soldiering in the 20th Century, Madonald Futura, London, 1981 Lyman, R., Iraq 1941, Osprey, Oxford, 2006
First Victory, Constable & Robinson, London, 2006 Mackenzie, C., Eastern Epic Vo.l 1, Chatto & Windus, London, 1951 Martin, T.A., The Essex Regiment, Essex Regiment Association, Brentwood,
1951 Matter, P., The Mufti of Jerusalem, Cambridge University Press, Cambridge, 1988 Meir-Glitzenstein, E., Zionism in an Arab Country: Jews in Iraq in the 1940s,
Routledge, London, 2004 Mockler, A., Our Enemies The French, Leo Cooper, London, 1976 Muggeridge, M. (Ed.), Ciano's Diplomatic Papers, Odhams, London, 1948 Mure, D., Practice to Deceive, William Kimber, London, 1977 Nettleton, S. (Ed.), Social Forces in Middle East, Fisher, New York, 1955 Nightingale, P.R., The East Yorkshire Regiment in the War of 1939–45, Mr
Pye (Books), Goole,1998 Omissi, D., Air Power and Colonial Control, University of Manchester Press,
1990 Pearce, F., Under the Red Eagle, Woodfield, Bognor Regis, 2004 Pearlman, M., The Mufti of Jerusalem, Gollanz, London, 1947 Pelly-Fry, J., Heavenly Days, Crecy Books, 1994 Playfair, I.S.O., History of the Second World War, Vol. 2, HMSO, London,
1956 Priestland, J. (Ed.), Records of Iraq, Vol. 8, Politics and the Palace (1936–1941) and Vol. 9, The Second World War (1941–1945), Cambridge Archive Editions, Slough,2001
Rahn, R., Ruheloses Leben: Aufzeichnungen und Erinnerungen, Diederichs Verlag, Dusseldorf, 1949
Ranfurly, H., To War with Whitaker, Mandarin, 1995
Rankin. N., Churchill's Wizards, Faber & Faber, London, 2008
Rassam, S., Christianity in Iraq, Gracewings, London, 2004
Raugh, H.E., Wavell in the Middle East, 1939–41, Brassey, 1993
Ross, A.E. (Ed.), Through Eyes of Blue, Airlife, Shrewsbury, 2002
Schechtman, J.B., The Mufti and the Fuhrer, Thomas Yoseloff, New York, 1965
Schofield, J., Wavell, John Murray, London, 2006
Schreiber, G., Stegman, B., and Vogel, D., Germany and the Second World War, Clarendon Press, Oxford, 1991
Schroder, B., Deutschland und der Mittlere Ostern in Zweiter Weltkreig, Musterschmidt Verlag, Gottingen, 1975
Schwanitz, W.G., Germany and the Middle East, Marcus Weiner, Princeton, 2004
Shams, I. (Ed.), Iraq: Its History, People & Politics, Humanity Books, London, 2003
Simpson, C., Blowback, Weidenfeld & Nicholson, London, 1988
Smith, D.M., Mussolini's Roman Empire, Longman, London, 1976
Shores, C., Dust Clouds in the Middle East, Grub Street, London, 1996
Silverfarb, D., Britain's Informal Empire in the Middle East: A Case Study of Iraq 1929–1941, Oxford University Press, 1997
Simon, R.S., Iraq between the Two World Wars, Columbia University Press, New York, 1986
Sinderson, H.C., Ten Thousand and One Nights: Memories of the Sherifian Dynasty, Hodder & Stoughton, London, 1973
Sluglett P.J. and Steer, I., British in Iraq 1914–1932: Contriving King & Country, Columbia University Press, New York, 2007
Spears, E., Fulfilment of a Mission, Leo Cooper, London, 1977
Stark, F., Baghdad Sketches, John Murray, London, 1937 Dust in the Lion's Paw, Century, London, 1985
Sweet-Escott, B., Baker Street Irregular, Methuen, London, 1976
Tarbush, M.A., The Role of the Military in Politics: A Case Study of Iraq to 1941, K. Paul International, London, 1982
Tedder, Lord, With Prejudice, Cassell, London, 1966
Tillman, H., Deutschlands Araberpolitik in Zweiten Weltkreig, Deutsche Verlag fur Wissenschaft, Berlin, 1965
Trigg, J., Hitler's Jihadis, The History Press, Stroud, 2008
Tripp, C., A History of Iraq, Cambridge University Press, Cambridge, 2000 Warner, G., Iraq & Syria, University of Delaware Press, Newark, 1974
von Weizsacher, E., Memoirs, Victor Gollanz, London, 1951
White-Spunner, B. Horse Guards, Macmillan, London, 2006
Wilson, Lord, Eight Years Overseas, Hutchinson, London, 1950
Wyndham, H., The Household Cavalry at War, Gale & Polden, Aldershot, 1952
Magazines and Journals
(A selection of those articles found to be of most use)
Arsenian, S., Wartime Propaganda in the Middle East, Middle East Journal, Vol. II, 1948
Atherton, L., SOE: Operations in the Middle East, PRO Publications, 1997
Cohen, H.J., The Anti-Jewish Farhud in Baghdad 1941, Middle Eastern Studies, Vol. 3, No. 1, October 1966
Dawn, C.E., The Formation of Pan-Arab Ideology in the Interwar Years, Int. Journal of Middle East Studies, Vol. 20, No. 1, February 1988
Deringil, S., The Preservation of Turkey's Neutrality During the Second World War, Middle Eastern Studies, Vol. 18, No. 1, February 1988
Dietrich, R., German relations with Iraq, Middle Eastern Studies, Vol. 41, No. 4, July 2005
Documents – German Ideas on Iraq 1937–1938, The Middle East Journal, Vol. 12, No. 2, Spring 1958
Dovey, H.O., The Middle East Intelligence Centre, Intelligence & National Security, Vol. 4, No. 4, October 1989
The Eighth Assignment 1941–1942, Intelligence & National Security, Vol. 11, No. 4, October 1996
Graham, A., The Iraq Levies at Habbaniya, The Army Quarterly, Vol. XLIV, 1942
Heller, M., Politics and the Military in Iraq and Jordan, Armed Forces & Society, Vol. 4, No. 1, November 1977
Hilgruber, A., England's Place in Hitler's Plans for World Domination, Journal of Contemporary History, Vol. 9, No. 9, January 1974
Isaacs, H.D., Sir Harry Sinderson Pasha (1891–1974), British Society For Middle Eastern Studies Bulletin, Vol. 2, No. 1, 1975
Kedourie E., Wavell and Iraq, Middle Eastern Studies, Vol. 11, July 1966
Kelidar, A., Iraq: The Search for Stability, Middle East Review, Vol. 11, No. 4, Summer 1979
Kelly, S., A Succession of Crises: SOE in the Middle East, 1940–45, Intelligence & National Security, Vol. 20, No. 1, March 2005
Khadduri, M., General Nuri's flirtations with the Axis Powers, Middle East Journal, Vol. XVI, 1962
De Luca, A.R., Der Grossmufti in Berlin: The Politics of Collaboration, Int. Journal of Middle East Studies, Vol. 10, No. 1, 1988
MacDonald, C.A., Radio Bari, Middle Eastern Studies, Vol. 13, 1977
Marmorstein, M., Review of Fritz Grobba's Memoirs, Middle Eastern Studies, Vol. 5, No. 3, October 1969
Mattar, P., Al-Husayni and Iraq's Quest for Independence 1939–1941, Arab Studies Quarterly, Vol. 6, No. 4, 1984
Maunsell and Mure, Intelligence & National Security, Vol. 8, No. 1, January 1993
Murphy, J.F., The March on Baghdad, Military Heritage, August 2000
Nafi, B.M., The Arabs and the Axis 1933–1940, Arab Studies Quarterly, Spring 1997
Nevo, J., Al-Hajj Amin and the British in World War II, Middle Eastern Studies, Vol. 20, No. 1, January 1984
Nicosia, F., Arab Nationalism & National Socialist Germany 1933–1939, Int. Journal of Middle East Studies, Vol. 12, 1980
Ogilvy, D., Airspeed Oxford, General Aviation, April, 2007 The Last of a Line, The Aeroplane, January 2008
Parsons, L., Soldiering for Arab Nationalism: Fawzi al-Qawuqji in Palestine, Journal of Palestinian Studies, Vol. 36, No. 4, 2007
Pool, D.C., Light on Nazi Foreign Policy, Foreign Affairs, Vol. 25, 1946
Porch. D., The Other Gulf War, British Intervention in Iraq 1941, JFQ, Issue 35
Price, A., Messerschmitt Bf 110, The Aeroplane, August 2007
Al-Qazzaz, A., The Iraq-British War of 1941: A Review Article, Int. Journal of Middle East Studies, Vol. 7, No. 4, 1976
Schwanitz, W.G., The Jinee and the Magic Bottle, Interdisciplinary Journal of Middle Eastern Studies, Vols X & XI
Stratton, M.B., British Railways & Motor Roads in the Middle East 1930–1940, Economic Geography, Vol. 20, No. 2, April 1944
Sulzberger, D.L., German Preparations in the Middle East, Foreign Affairs, Vol. 20, 1942
Thorpe, J.A., The US and the 1940–1941 Iraqi Crisis, Middle East Journal, Vol. 25, 1979
Williams, M., Mussolini's Secret War in the Mediterranean and the Middle East, Intelligence & National Security, Vol. 22, No. 6, December 2007 Yisraeli, D., The Third Reich and Palestine, Middle East Studies, Vol. 7, 1971
Archival Sources
Cabinet, CAB 106
Air Ministry, AIR 23, 40 & 41
Foreign Office, FO 370, 371 & 406
Security Service, KV 5/34, 5/35 & 5/40
War Office, WO 208
SOE, HS 3/198
PhD Theses
Al-Jamail, K., Nationalism in Iraq 1936–1941: Rashid Ali and Foreign Involvement, University of Keele, 1978
Flacker, E., Fritz Grobba and Germany's Middle East Policy 1933–1942, University of London, 1998
Hamdi, W.M.S., Rashid Ali Al-Gailani and the Nationalist Movement in Iraq 1939–1941: A Political and Military Study of the British campaign in Iraq and the National Revolution of May 1941, University of Birmingham, 1986
Tarbush, M., The role of the Military in Politics: A Case Study of Iraq from 1936–1941, University of Oxford, 1977
Internet Sources
www.assyrianlevies.com
Assyrian RAF Levies: The Battle for Habbaniya
www.csus.edu/indiv/s/scottjc/introduction.htm
Iraqi Coup, 1941
www.generals.dk/
Generals of World War II
www.geocities.com/acrawfordO/revo1t.html
The Iraqi Revolt
www.Habbaniya.org
RAF Habbaniya Association
www.latimesblogs.latimes.com/babylonbeyond/2008/12/fifty-years-aft.html
Rededication of the Habbaniya Cemetery
www.lasecondaguerramondiale.it/iraq.html
La Rivolta dell′Iraq
www.michael-reimer.com/CFS2/CFS2 Profiles/MTO AXIS Iraq.html
Iraqi and Axis aircraft in the fight for Habbaniya
www.mogggy.org/iraq/iraq.htm
Allied, Iraqi and Axis aircraft used in Iraq, 1941
www.mideastweb.org/iraqaxiscoup.htm
Iraq Coup Attempt of 1941: the Mufti and the Farhud
www.piltonenbunker.de/nachtjaeger/Luftwaffe/Drewes Martin.htm
Martin Drewes
www.rafarmouredcarsassociation.com
RAF Armoured Car Company Association
www.surfcity.kund.dalnet.se
Biplane fighter aces of World War II
|
{
"redpajama_set_name": "RedPajamaBook"
}
| 7,701
|
{"url":"https:\/\/mathematica.stackexchange.com\/questions\/193003\/how-can-i-manipulate-the-output-of-information","text":"# How can I manipulate the output of Information?\n\nIf I type\n\nInformation[ProductLog]\n\n\nI see\n\nIf I type\n\nFramed[Information[ProductLog]]\n\n\nI see\n\nThis is expected, though not desired, because Information prints its output, rater than returning it. However, I want to get access to the text of the description, or the boxes of the description, or something, to, e.g., put the descriptions of functions in columns or rows. How do I do things like this?\n\n\u2022 The output cells are created by CellPrint. You can capture the cell expressions with Block[{CellPrint = Sow}, Reap@Information[ProductLog] ], but I'm not sure that's what you want. Mar 10, 2019 at 22:52\n\nYou may use the \"PlaintextUsage\" or \"TypesetUsage\"property of WolframLanguageData.\n\nWolframLanguageData[\"ProductLog\", \"PlaintextUsage\"]\n\n\nor\n\nWolframLanguageData[ToString@ProductLog, \"PlaintextUsage\"]\n\n\ngives\n\n\"ProductLog[z] gives the principal solution for w in z\\[LongEqual]we^w. ProductLog[k, z] gives the k^th solution.\"\n\n\nNote the escapes like \"\\[LongEqual]\" will evaluate to their symbol in strings in Mathematica\/Wolfram Language.\n\nFor typeset lines then\n\nWolframLanguageData[\"ProductLog\", \"TypesetUsage\"][[1, 1]]\n\n\nHope this helps.\n\n\u2022 Thanks! This is exactly what I was looking for. Mar 11, 2019 at 0:30\n\nA simpler method to get the usage messages for further manipulation is to use MessageName[symbol, \"usage\"]:\n\nMessageName[ProductLog, \"usage\"]\n\n\nwhich already yields a typeset version, similar to what is produced by Information[].\n\nAnd of course, Attributes[symbol] will return symbol's associated attributes.","date":"2022-08-19 02:11:58","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 1, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.5584707260131836, \"perplexity\": 3327.0054485436926}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 20, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-33\/segments\/1659882573540.20\/warc\/CC-MAIN-20220819005802-20220819035802-00363.warc.gz\"}"}
| null | null |
/*
* To change this license header, choose License Headers in Project Properties.
* To change this template file, choose Tools | Templates
* and open the template in the editor.
*/
package GoHost;
import java.io.Serializable;
import javax.persistence.Basic;
import javax.persistence.Column;
import javax.persistence.Entity;
import javax.persistence.GeneratedValue;
import javax.persistence.GenerationType;
import javax.persistence.Id;
import javax.persistence.NamedQueries;
import javax.persistence.NamedQuery;
import javax.persistence.Table;
import javax.validation.constraints.NotNull;
import javax.xml.bind.annotation.XmlRootElement;
/**
*
* @author Jerry
*/
@Entity
@Table(name = "attendee")
@XmlRootElement
@NamedQueries({
@NamedQuery(name = "Attendee.findAll", query = "SELECT a FROM Attendee a")
, @NamedQuery(name = "Attendee.findByIdattendee", query = "SELECT a FROM Attendee a WHERE a.idattendee = :idattendee")
, @NamedQuery(name = "Attendee.findByIduser", query = "SELECT a FROM Attendee a WHERE a.iduser = :iduser")
, @NamedQuery(name = "Attendee.findByIdevent", query = "SELECT a FROM Attendee a WHERE a.idevent = :idevent")})
public class Attendee implements Serializable {
private static final long serialVersionUID = 1L;
@Id
@Basic(optional = true)
@GeneratedValue(strategy = GenerationType.IDENTITY)
@Column(name = "idattendee")
private Integer idattendee;
@Column(name = "iduser")
private Integer iduser;
@Column(name = "idevent")
private Integer idevent;
public Attendee() {
}
public Attendee(Integer idattendee) {
this.idattendee = idattendee;
}
public Integer getIdattendee() {
return idattendee;
}
public void setIdattendee(Integer idattendee) {
this.idattendee = idattendee;
}
public Integer getIduser() {
return iduser;
}
public void setIduser(Integer iduser) {
this.iduser = iduser;
}
public Integer getIdevent() {
return idevent;
}
public void setIdevent(Integer idevent) {
this.idevent = idevent;
}
@Override
public int hashCode() {
int hash = 0;
hash += (idattendee != null ? idattendee.hashCode() : 0);
return hash;
}
@Override
public boolean equals(Object object) {
// TODO: Warning - this method won't work in the case the id fields are not set
if (!(object instanceof Attendee)) {
return false;
}
Attendee other = (Attendee) object;
if ((this.idattendee == null && other.idattendee != null) || (this.idattendee != null && !this.idattendee.equals(other.idattendee))) {
return false;
}
return true;
}
@Override
public String toString() {
return "GoHost.Attendee[ idattendee=" + idattendee + " ]";
}
}
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 7,238
|
Core: Suppressed reports about broken digital signature on MSI and CAB files.
Kernel: Fixed issue with Outlook crash due to AVG plugin.
Setup: Fixed installation failure when Toolbar is not installed, and change of search provider is enabled.
Setup: During unistallation no information on specific AVG processes being closed is provided.
Setup: Fixed issue with AVG being repaired into different folder if non-default folder was used for installation.
Setup: Installation works correctly even on computers with Windows on other than C: drive.
Setup: Repair installation on corrupted files runs correctly.
Setup: Semicolons in USER and COMPANY properties are replaced with commas, and do not interfere with the installation process.
Update: Fixed issue with repeated restart requests during update.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 2,447
|
Pałac myśliwski książąt Radziwiłłów w Antoninie – pałac myśliwski zbudowany w latach 1822–1824 przez Karla Friedricha Schinkla dla księcia Antoniego Radziwiłła.
Pałac jest drewniany, oszalowany, na podmurówce, zbudowany na planie krzyża greckiego – korpus główny o kształcie oktagonalnym, do którego dobudowano cztery niższe skrzydła. Elewacja pomalowana jest na kolor tzw. żółci cesarskiej (niem. Kaisergelb). Całe wnętrze korpusu głównego zajmuje wielka, trójkondygnacyjna sala o ozdobnym stropie, wsparta na filarze mieszczącym przewody kominowe z kominków, przyozdobiona trofeami myśliwskimi. Dawniej odbywały się w niej koncerty. Jedno ze skrzydeł mieści klatkę schodową wiodącą na wewnętrzne balkony, prowadzące do pokoi mieszkalnych w pozostałych trzech skrzydłach.
Gościem Antoniego Radziwiłła w Antoninie był Fryderyk Chopin (dwukrotnie, w 1827 i 1829). Na elewacji pałacu znajduje się tablica upamiętniająca te wizyty, a przy wejściu na teren pałacowy umiejscowiony jest pomnik kompozytora. W latach 1939–1944 pałac był własnością Adolfa Hitlera po podarowaniu go w 1939 roku przez ówczesnego właściciela, księcia Michała Radziwiłła
Pałac remontowany w latach 1975–1978 staraniem Jerzego Waldorffa, obecnie hotel i restauracja. Nagrodzony za renowację nagrodą Europa Nostra' 94.
Wokół rozległy park w stylu angielskim ze stawem i pomnikowymi drzewami. Na środku stawu, na wysepce, znajduje się wykonany z białego marmuru nagrobek córek Antoniego Radziwiłła – Fryderyki Heleny i Luizy.
W pobliżu także zabytkowe zabudowania utrzymane w stylu tyrolskim.
Miejsce koncertów Międzynarodowego Festiwalu "Chopin w barwach jesieni".
Bibliografia
Skuratowicz J., Dwory i pałace w Wielkim Księstwie Poznańskim, Międzychód 1992
Antonin (wieś w powiecie ostrowskim)
Antonin
Zamki i pałace Radziwiłłów herbu Trąby
Zabytki w powiecie ostrowskim (województwo wielkopolskie)
Dzieła Karla Friedricha Schinkla
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 8,619
|
Segro to develop last-mile logistics beneath Paris streets
last-mile logisticsParis
UK REIT Segro, a logistics and industrial property specialist, is to develop a 75,000 sq m underground logistics centre as part of the redevelopment of the Gobelins rail station in central Paris.
The development will be a joint project with Paris-based developer Icade and also create two office buildings totalling around 14,000 sq m.
While the Segro portion of the scheme is aimed at the urban distribution and last-mile delivery market, the offices will create workspace both for local start-ups and more established companies. The project will also provide sports facilities, for climbing and basketball. Urban greenhouses and a 1.3-hectare garden on the Olympiades esplanade are also planned.
The project is scheduled to be complete in 2025.
Segro said in a statement that the logistics hub will be developed to high environmental and wellness standards. It will include charging infrastructure for electric vehicles and delivery tricycles as well as recycling points and a bike repair centre.
Project won as part of Reinventing Paris competition
SEGRO and Icade have signed preliminary agreements to acquire land for the redevelopment of the site, which is in the 13th district of Paris. Redevelopment of Gobelins rail station is part of the Reprendre Racines (Back to our Roots) project which is a masterplan devised by SEGRO and Icade and selected by France's national rail company SNCF and the City of Paris as part of the Reinventing Paris 2 competition.
The masterplan team included architectural and urban design firms Data Architectes and Anyoji Beltrando and landscape design firm Wagon Landscaping.
"The growth in ecommerce is driving demand for logistics space in major urban centres, where land for development is a scarce resource," Segro general manager for France Laurence Giard said.
"The underground logistics hub proposed at Gobelins railway station is a new and innovative model for delivering logistics space at scale in one of Europe's most densely populated cities, within a brand new world-class mixed-use destination," he added.
Vaccine gives hope of a 2H 2021 global recovery: Colliers
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 3,325
|
\section{Introduction}
The general Erd\H os distance problem asks to determine the number of distinct distances spanned by a finite set of points. In the Euclidean space, it is conjectured that for any finite set $E \subset \mathbb{R}^d$, $d\ge 2$, we have $|\Delta(E)| \gtrapprox |E|^{\frac{2}{d}}$, where $\Delta(E)=\{\|x-y\| : x,y \in E\}$. Here and throughout, $X\ll Y$ means that there exists $C>0$ such that $X\le CY$, and $X\lessapprox Y$ with the parameter $N$ means that for any $\varepsilon>0$, there exists $C_{\varepsilon}>0$ such that $X \le C_{\varepsilon}N^{\varepsilon}Y$.
The finite field analogue of the distance problem was first studied by Bourgrain, Katz, and Tao \cite{BKT04} over prime fields. In this setting, the Euclidean distance among any two points $\boldsymbol{x}=(x_1,\ldots, x_d),\boldsymbol{y}=(y_1,\ldots,y_d) \in \mathbb{F}_q^d$, the $d$-dimensional vector space over the finite field of order $q$, is defined as $\|\boldsymbol{x}-\boldsymbol{y} \|=\displaystyle\sum_{i=1}^d(x_i-y_i)^2\in\mathbb F_q$.
For prime fields $\mathbb{F}_p$ with $p\equiv -1 \pmod 4$, they showed that if
$E \subset \mathbb{F}_p^2$ with $|E|=p^{\delta}$ for some $0<\delta<2$, then the distance set satisfies $|\Delta(E)| \gg |E|^{\frac{1}{2}+\varepsilon}$, for some $\varepsilon>0$ depending only on $\delta$.
This bound does not hold in general for arbitrary finite fields $\mathbb{F}_q$ as shown by Iosevich and Rudnev \cite{IR07}. In this general setting, they considered the Erd\H os-Falconer distance problem to determine how large $E \subset \mathbb{F}_q^d$ needs to be so that $\Delta(E)$ spans all possible distances or at least a positive proportion of them. More precisely, they proved that $\Delta(E)=\mathbb{F}_q$ if $|E| > 2q^{\frac{d+1}{2}}$, where the exponent is sharp for odd $d$. It is conjectured that in even dimensions, the optimal exponent will be $\frac{d}{2}$. As a relaxed fractional variant for $d=2$, it was shown in \cite{CEHIK} that if $E \subseteq \mathbb{F}_q^2$ satisfies $|E| \gg q^{\frac{4}{3}}$, then $|\Delta(E)| \gg q$. A recent series of other improvements and generalizations on the Erd\H os-Falconer distance problem can be found in \cite{hieu, koh, claudiu, suk, thang}.
Using Fourier analytic techniques, a two-parameter variant of the Erd\H os-Falconer distance problem for the Euclidean distance was studied by Birklbauer and Iosevich in \cite{BI17}. More precisely, given $E \subseteq \mathbb{F}_q^d \times \mathbb{F}_q^d$, where $d\ge 2$, define the two-parameter distance set as
\[
\Delta_{d, d}(E)=\left\{\left(\|x_1-y_1\|, \|x_2-y_2\|\right) : (x_1,x_2), (y_1,y_2) \in E \right\}
\subseteq \mathbb{F}_q \times \mathbb{F}_q. \]
They proved the following results.
\begin{theorem}\label{res11}
Let $E$ be a subset in $\mathbb{F}_q^d \times \mathbb{F}_q^d$. If $|E| \gg q^{\frac{3d+1}{2}}$, then $ |\Delta_{d, d}(E)| = q^2$.
\end{theorem}
\begin{theorem}\label{res22}
Let $E$ be a subset in $\mathbb{F}_q^2 \times \mathbb{F}_q^2$. If $|E| \gg q^{\frac{10}{3}}$, then $ |\Delta_{2, 2}(E)| \gg q^2$.
\end{theorem}
In this short note, we provide an extension and an improvement of these results. Compared to the method in \cite{BI17}, our results are much elementary.
For $\boldsymbol{x}=(x_1,\ldots,x_d),\boldsymbol{y}=(y_1,\ldots,y_d) \in \mathbb{F}_q^d$ and for an integer $s\ge 2$, we introduce
$$\|\boldsymbol{x}-\boldsymbol{y} \|_s:=\sum_{i=1}^d a_i(x_i-y_i)^s,$$
where $a_i \in \mathbb{F}_q$ with $a_i \neq 0$ for $i=1,\ldots, d$. For any set $E\subset \mathbb{F}_q^d\times \mathbb{F}_q^d$, define
\[\Delta_{d, d}^s(E)=\left\{\left(\|x_1-y_1\|_s, \|x_2-y_2\|_s\right) : (x_1,x_2), (y_1,y_2) \in E \right\}.\]
Our first result reads as follows.
\begin{theorem}\label{res1}
Let $E$ be a subset in $\mathbb{F}_q^d \times \mathbb{F}_q^d$. If $|E| \gg q^{\frac{3d+1}{2}}$, then $ |\Delta_{d, d}^s(E)| \gg q^2$.
\end{theorem}
It is worth mentioning that our method also works for the multi-parameter distance set defined for $E \subseteq \mathbb{F}_q^{d_1+\dots+d_k}$, but we do not discuss such extensions herein.
For the case of $d=2$, we get an improved version of Theorem \ref{res22} for the usual distance function over prime fields.
\begin{theorem}\label{res3}
Let $E \subseteq \mathbb{F}_p^2 \times \mathbb{F}_p^2$. If $|E| \gg p^{\frac{13}{4}}$, then $ |\Delta_{2}(E)| \gg p^2$.
\end{theorem}
We note that the continuous version of Theorems \ref{res1} and \ref{res3} have been studied in \cite{ioo0, ioo1}. However, the authors do not know whether the method in this paper can be extended to that setting. Moreover, it follows from our approach that the conjecture exponent $\frac{d}{2}$ of the (one-parameter) distance problem would imply the sharp exponent for two-parameter analogue, namely, $\frac{3d}{2}$ for even dimensions. We refer the reader to \cite{BI17} for constructions and more discussions.
\section{Proof of Theorem \ref{res1}}
The following lemma plays a key role in our proof for Theorem \ref{res1}.
\begin{lemma}[Theorem 2.3, \cite{Vinh}]\label{lmm1}
Let $X, Y \subseteq \mathbb{F}_q^d$. Define $\Delta^s(X, Y)=\{\|x-y\|_s\colon x\in X, y\in Y\}$. If $|X||Y|\gg q^{d+1}$, then $|\Delta^s(X, Y)|\gg q$.
\end{lemma}
\begin{proof}[Proof of Theorem \ref{res1}]
By assumption, we have $|E|\ge Cq^{d+\frac{d+1}{2}}$ for some constant $C>0$. For $y\in \mathbb{F}_q^d$, let $E_y:=\left\{x\in \mathbb F_q^d : (x, y)\in E\right\}$, and define
\[
Y:=\left\{ y\in \mathbb{F}_q^d : ~ |E_y|> \frac{C}{2} q^{\frac{d+1}{2}} \right\}.
\]
We first show that $ |Y| \ge \frac{C}{2} q^{\frac{d+1}{2}}$. Note that
\[
|E|=\sum_{y \in Y} |E_y| + \sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y| ~ \le ~
q^d|Y| + \sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y|,
\]
where the last inequality holds since $|E_y| \le q^d$ for $y\in \mathbb{F}_q^d$. Combining it with the assumption on $|E|$ gives the lower bound
$
\sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y| \ge Cq^{d+\frac{d+1}{2}} - q^d|Y|.
$
On the other hand, by definition, we have $|E_y| \le \frac{C}{2} q^{\frac{d+1}{2}}$ for $y \in \mathbb{F}^d_q\setminus Y$ yielding the upper bound
$
\sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y| \le \frac{C}{2} q^{d+\frac{d+1}{2}}
$.
Thus, these two bounds altogether give $Cq^{d+\frac{d+1}{2}} - q^d|Y| \le \frac{C}{2} q^{d+\frac{d+1}{2}}$, proving the claimed bound $|Y| \ge \frac{C}{2} q^{\frac{d+1}{2}}$.
In particular, Lemma \ref{lmm1} implies $|\Delta^s(Y,Y)|\gg q$, as $|Y||Y| \gg q^{d+1}$. On the other hand, for each $u\in \Delta^s(Y,Y)$, there are $z, t\in Y$ such that $\|z-t\|=u$. One has $|E_z|, |E_t|\gg q^{\frac{d+1}{2}}$, therefore, again by Lemma \ref{lmm1}, $|\Delta^s(E_z, E_t)|\gg q$. Furthermore, for $v\in \Delta^s(E_z, E_t)$, there are $x\in E_z$ and $y\in E_t$ satisfying $\|x-y\|_s=v$. Note that $x\in E_z$ and $y\in E_t$ mean that $(x,z), (y,t)\in E$. Thus, $(v,u)=(\|x-y\|_s, \|z-t\|_s)\in \Delta_{d, d}^s(E)$. From this, we conclude that $|\Delta_{d, d}^s(E)|\gg q|\Delta^s(Y,Y)|\gg q^2$, which completes the proof.
\end{proof}
\section{Proof of Theorem \ref{res3}}
To improve the exponent over prime fields $\mathbb{F}_p$, we strengthen Lemma \ref{lmm1} as follows. Following the proof of Theorem \ref{res1} with Lemma \ref{lmm2} below proves Theorem \ref{res3} then.
\begin{lemma}\label{lmm2}
Let $X, Y \subseteq \mathbb{F}_p^2$. If $|X|, |Y|\gg p^{\frac 5 4}$, then $|\Delta(X, Y)|\gg p$.
\end{lemma}
\begin{proof}
It is clear that if $X' \subseteq X$ and $Y' \subseteq Y$, then $\Delta(X',Y')\subseteq \Delta(X,Y)$. Thus, without loss of generality, we may assume that $|X|=|Y|=N$ with $N\gg p^{\frac 5 4}$.
Let $Q$ be the number of quadruples $(x, y, x', y')\in X\times Y\times X\times Y$ such that $\|x-y\|=\|x'-y'\|$. It follows easily from the Cauchy-Schwarz inequality that
\[|\Delta(X, Y)|\gg \frac{|X|^2|Y|^2}{Q}.\]
Let $T$ be the number of triples $(x, y, y')\in X\times Y\times Y $ such that $\|x-y\|=\|x-y'\|$. By the Cauchy-Schwarz inequality again, one gets $Q\ll |X|\cdot T$. Next, we need to bound $T$. For this, denote $Z=X\cup Y$, then $N\le |Z|\le 2N$. Let $T'$ be the number of triples $(a, b, c)\in Z\times Z\times Z$ such that $\|a-b\|=\|a-c\|$. Obviously, one gets $T\le T'$. On the other hand, it was recently proved (see \cite[Theorem 4]{pham}) that
\[T'\ll \frac{|Z|^3}{p}+p^{2/3}|Z|^{5/3}+p^{1/4}|Z|^2,\]
which gives
\[T\ll \frac{N^3}{p}+p^{2/3}N^{5/3}+p^{1/4}N^2,\]
and then $T\ll \dfrac {N^3}p$ (since $N\gg p^{\frac 5 4}$).
Putting all bounds together we obtain
$$\dfrac{N^3}{|\Delta(X,Y)|}=\dfrac{|X||Y|^2}{|\Delta(X,Y)|}\ll \dfrac Q{|X|}\ll T\ll \dfrac {N^3}p,$$
or equivalently, $|\Delta(X,Y)|\gg p$, as required.
\end{proof}
\section*{Acknowledgment}
The authors would like to thank Thang Pham for sharing insights and new ideas.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 252
|
Q: Visual Studio Setup Project - Uninstall is not being called I am preparing setup project for my winform application. I have created an Installer Class and attached into custom action window. I have tried many ways to get uninstall method called but all attempts have failed.
I have not tried to create a separate class installer as I need to read the savedState object during uninstallation
I have run uninstall msi:
MSI (s) (04:F4) [17:28:46:819]: Windows Installer removed the product. Product Name: Library Manager. Product Version: 1.5.0. Product Language: 1033. Manufacturer: KANAAN. Removal success or error status: 0.
I don't see any error in the log file
Anyone has faced a similar error?
Note: I have created an empty project for demo purposes, and uninstall was called!
A: I have solved the issue by creating a new setup project. Now Uninstall method is called.
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 5,060
|
Padshahan is a village and union council, an administrative subdivision, of Chakwal District in the Punjab Province of Pakistan, it is part of Chakwal Tehsil.
References
Chairman CH SARFRAZ Khan
Union councils of Chakwal District
Populated places in Chakwal District
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 8,231
|
Kateřina Siniaková and Renata Voráčová were the defending champions, but they chose not to participate this year.
Barbora Krejčíková and Mandy Minella won the title, defeating Margarita Gasparyan and Oksana Kalashnikova in the final 1–6, 7–5, [10–6].
Seeds
Draw
Draw
External Links
Draw
Open de Limoges - Doubles
Open de Limoges
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 9,591
|
Conspiracy of Historical Proportions?
How deep does the Conspiracy Swamp Go?
In commemoration of all those who have sacrificed their lives for
the American way of freedom, and defiance of oppressive governments,
we are posting House Resolution 907 which precisely documents the
bulk of illegal and Treasonous actions committed by the Obama
Administration during which he weaponized the US Intelligence
Community against his political opponents,
and against American Citizens.
We have highlighted key passages and added links
to assist you in understanding the gravity of this resolution
House Resolution 907 appears in its entirety below:
115TH CONGRESS 2D SESSION H. RES. 907
Expressing the sense of Congress that the Attorney General of the United States should appoint a Special Counsel to investigate misconduct at the Department of Justice and Federal Bureau of Investigation, including an investigation of abuse of the FISA warrant process, how and why the Hillary Clinton probe ended, and how and why the Donald TrumpRussia probe began.
IN THE HOUSE OF REPRESENTATIVES
Mr. ZELDIN (for himself, Mr. GOSAR, Mr. GAETZ, Mr. PERRY, Mr. DESANTIS, Mr. MOONEY of West Virginia,
Mr. DESJARLAIS, Ms. TENNEY, Mr. MEADOWS, Mr. JODY B. HICE of Georgia, Mr. GOHMERT, Mr. ROTHFUS,
Mr. JORDAN, Mr. BUDD, Mr. ROUZER, Mr. YOHO, Mr. BRAT, Mr. ROKITA, Mr. BIGGS, Mr. POE of Texas,
Mr. WILLIAMS, Mr. DUNCAN of South Carolina, Mr. GIBBS, Mrs. BLACK, and Mr. ISSA) submitted the following resolution; which was referred to the Committee on the Judiciary
Expressing the sense of Congress that the
should appoint a Special Counsel to investigate misconduct at
the Department of Justice and Federal Bureau of Investigation,
including an investigation of abuse of the FISA Warrant process
how and why the Hillary Clinton email probe ended,
and how and why the Donald Trump-Russia probe began;
There is an urgent need for the
appointment of a second Special Counsel
in light of evidence that raises critical concerns about
decisions, activities, and inherent bias displayed at the highest levels of the
Department of Justice (DOJ) and Federal Bureau of Investigation (FBI)
regarding FISA abuse,
how and why the HillaryClinton email probe ended,
how and why the Donald Trump-Russia probe began;
the concerns of the American people are serious and the issues requiring an
immediate, unbiased, independent, and thorough investigation are broad;
misconduct during the 2016 presidential election by high-ranking individuals within the FBI and DOJ may have led to the premature conclusion of the FBI's 2016 probe into then-presidential candidate and former Secretary of State Hillary Clinton;
Mess with me and I'll Clinticide you!
during her tenure as Secretary of State, Hillary Clinton violated Federal law, and Department of State rules, regulations, and protocol, by using a private email server in her Chappaqua, New York, home;
official communications were transmitted on an unsecured server and included emails that contained classified information when they were sent, in addition to additional emails which were retroactively deemed classified by the Department of State;
FBI Director James Comey acknowledged that 65 of these illicit emails were classified as ''Secret'' and 22 were classified as ''Top Secret'';
there is significant evidence that the use of this private server by Secretary Clinton was meant to avoid compliance with the Freedom of Information Act (5 U.S.C. 552) and done to obstruct justice by not having to turn over incriminating emails in the case of a subpoena;
various sensitive emails subject to grand jury and congressional subpoenas
were destroyed on Secretary Clinton's private server through the use of ''BleachBit'' software and the destruction of hardware before they could be obtained by investigators in March 2015;
in a September 2015 meeting between then-Attorney General Loretta Lynch and then-Director Comey, the Attorney General instructed Director Comey to refer to the Clinton email investigation as a ''matter'', thus watering down the severity of the investigation and aligning the FBI's rhetoric with the messaging of the Clinton campaign;
Cheryl has Immunity? Time for a Clinticide?
Cheryl Mills, who served as Counselor and Chief of Staff to Hillary Clinton during her entire tenure as United States Secretary of State, was offered immunityfrom prosecution by the FBI during this investigation in exchange for access to her laptop that contained many of the questionable emails;
I don't need no stinking witnesses!
according to transcripts obtained by the Senate Judiciary Committee, former Director Comey was prepared to exonerate Hillary Clinton as early as April or May of 2016 when he began to draft a statement announcing the end of his investigation, before up to 17 key witnesses, including former Secretary Clinton and several of her closest aides, were interviewed;
former Director Comey contradicted these transcripts when he stated during sworn testimony before the House Judiciary Committee on September 28, 2016, that he made the decision not to recommend criminal charges for Secretary Clinton ''after'' she was interviewed by the FBI on July 2, 2016;
Director Comey, in the final draft of his statement, allowed FBI Agent Peter Strzok to replace ''grossly negligent'', which is legally punishable under Federal law, with ''extremely careless'', which is not legally punishable under Federal law;
Federal law states gross negligence in handling the Nation's intelligence can be punished criminally with prison time or fines (18 USC § 793 and 798);
on June 27, 2016, Attorney General Lynch had a covert meeting with former President Bill Clinton aboard her plane on the tarmac in Phoenix, Arizona;
on July 5, 2016,
FBI Director Comey violated
DOJ rules and unilaterally exonerated then-presidential candidate Hillary Clinton in a public statement to the media;
one day later, on July 6, 2016, an announcement followed from Attorney General Lynch that the DOJ investigation into then-presidential candidate Hillary Clinton would be formally closed with no criminal charges;
Anthony and Huma
in September 2016, the FBI, during an examination of the personal laptop of former
Congressman Anthony Weiner
as part of an unrelated investigation into him
sending sexually explicit messages to a teenage girl,
discovered previously unexamined
Department of State classified emails
belonging to his spouse,
top Clinton aide Huma Abedin;
McAuliff and the McCabes
FBI Deputy Director Andrew McCabe's wife, Dr. Jill McCabe,
was running for Virginia State Senate
at the time and, as of October 26th,
had received $675,000 in donations from the Virginia Democratic Party and
Common Good VA,
the Leadership PAC controlled by Democratic Virginia Governor Terry McAuliffe,
a longtime Clinton associate;
an investigation conducted by the Office of the DOJ Inspector General noted that on October 27, 2016, Director Comey required that Deputy Director McCabe remove himself from a conference call regarding the Clinton emails discovered on Anthony Weiner's laptop to avoid the appearance of a conflict of interest after media reports surfaced noting these questionable political donations;
it took until October 28, 2016, for Director Comey to announce via a letter to the chairs of the relevant congressional committees that he was reopening the investigation into Hillary Clinton, an additional, egregious delay after the FBI failed to even examine the illicit emails after the FBI discovered them on Anthony Weiner's computer;
further investigation into whether then-FBI Deputy Director McCabe and other FBI officials sought to purposely delay the release of these illicit emails for politically motivated purposes is warranted;
throughout the Obama Administration, the DOJ failed to fully investigate serious concerns surrounding former President Clinton, then-Secretary of State Clinton and the Clinton Foundation's connection to Russiancompany Uranium One, which received Department of State approval to purchase U.S. uranium mines in 2010;
Another Mass Clinticide, compliments of Putin?
throughout Hillary Clinton's tenure as Secretary of State, a family foundation controlled by the Chairman of Uranium One made $2,350,000 in contributions to the Clinton Foundation which were not publicly disclosed in violation of an agreement Secretary Clinton had with the Obama White House to publicly identify all donors;
[NOTE: Witnesses who could testify against Hillary died in a mysterious plane crash]
in 2010, while Russian State interests were working to both acquire a majority stake in Uranium One and to purchase American mines, Bill Clinton was paid $500,000 for a speech in Moscow by a Kremlin-linked Russian investment bank that was underwriting Uranium One stock;
a confidential informant who worked with the FBI to uncover bribery and other corruption related to the Uranium One matter was threatened with reprisal by the Justice Department under Attorney General Lynch when he tried to come forward in 2016;
the Senate Judiciary Committee launched a probe in October 2017 to investigate the Uranium One matter, including whether Federal departments and agencies such as the Department of State knew the FBI was looking into possible corruption before the deal was approved;
an investigation conducted by the Office of the DOJ Inspector General noted that a multi-State investigation into the questionable dealings of the Clinton Foundation with corrupt donors was shut down in August 2016, when pressure was asserted on the FBI by senior officials within the Obama Justice Department;
the same Inspector General's report also noted that shutting down this investigation into Clinton Foundation impropriety and influence peddling was connected to high ranking officials in the DOJ and FBI,
including Attorney General Lynch, Director Comey, and Deputy Director McCabe;
the same Inspector General's report also found that Deputy Director McCabe, after consenting to the political pressure to shut down the Clinton Foundation multiState investigation, attempted to later use unauthorized leaks to the press to create a false narrative that he was opposed to the closure of the investigation and that he did this in an attempt
to salvage his reputation following revelations of questionable Clinton-connected money being donated to his wife's Virginia State Senate campaign;
in October 2016, the FBI and DOJ used politically biased, unverified sources to obtain warrants issued by the United States Foreign Intelligence Surveillance Court of Review (FISA Court) that aided in the surveillance of U.S. citizens, including Carter Page;
[NOTE: Chief Justice Roberts, on 03MAY2018, amended the FISA procedural rules]
these warrants grant U.S. intelligence and law enforcement agencies sweeping power to collect bulk information and conduct ''about collection'', which results in surveillance of a broad array of private communications from the past, present, and future, including those of U.S. citizens not specifically targeted in the FISA authorized warrant;
Steele - MI-6 Case Officer
to obtain these warrants, FBI and DOJ officials submitted an unverified dossier prepared by ChristopherSteele
to the FISA Court,
failing to disclose that Christopher Steele was hired by the firm Fusion GPS,
which was hired by the Democratic National Committee and Hillary Clinton
campaign to prepare this dossier and that the source was unreliable and was soon thereafter going to be terminated as a source;
the FISA Court was not informed
that Christopher Steele was actively opposed to the election of Donald Trump,
that he was the unnamed source cited in the media reports that the FBI used to corroborate his dossier, and
that Fusion GPS had been hired to perform previous anti-Trump research efforts in 2015;
the Woods Procedures, which are the FBI's mandatory vetting process required for all FISA warrant applications instituted to ensure that all the facts contained in an application are accurate and verified to clearly support probable cause for a warrant, were not followed;
former Director Comey admitted in sworn testimony to the Senate Judiciary Committee on June 8, 2017, that material contained in the Steele dossier was known to be both ''salacious'' and ''unverified'';
since FISA warrant applications
- are rarely turned down,
- are almost never subject to appeal, and
- are presented in closed court with no public record where the Government is not challenged by
any defense,
it is imperative that the Government take extra care to validate the information being utilized to build their case before they take the extraordinary step of waiving rights of a U.S. citizen without his or her knowledge or the opportunity to present a defense;
at the FISA Court, the Government has a responsibility not only to provide its best evidence in support of its case, but also to provide the best evidence against its case;
these deeply flawed and questionable FISA warrant applications utilizing illicit sources and politically biased intelligence were approved by DOJ and FBI officials at the highest levels before being submitted to the FISA Court;
Bruce Ohr and his wife
it was further not disclosed to the FISA court that
the wife of fourth-ranking DOJ official Bruce Ohr
worked for Fusion GPS and that
Christopher Steele directly transmitted the dossier and other information through Bruce Ohr for submission to the FISA court;
to this day there does not appear to be any evidence
that President Trump colluded with the Russians
to win the 2016 election;
the initial FBI probe into the Trump Campaign and alleged collusion with Russia was launched in July 2016, based on questionable and insufficient intelligence and biased motivations;
former Director Comey prepared a series of seven memoranda containing classified information, including notes on his conversations with President Trump;
Comey's "friend" - now Legal Counsel
former Director Comey admitted in sworn testimony to the Senate Committee on Intelligence on June 8, 2017, that he had leaked this content to a personal friend and encouraged that friend to share the material with the press in order to trigger a Special Counsel investigation;
an investigation conducted by the Senate Judiciary Committee
later revealed that the personal friend of Director Comey was
Professor Daniel Richman of Columbia Law School and that Director Comey provided him with four of the seven memoranda;
Director Comey's actions are a clear violation of non-disclosure agreements he signed as a condition of his appointment and a clear violation of FBI protocols regarding the dissemination of sensitive information outside of the Bureau which are based on provisions of the PrivacyAct of 1974 (5 U.S.C. 552a);
FBI Agent Strzok and FBI Counsel Page
text messages exchanged between FBI Agent Strzok and FBI Counsel Lisa Page, during the period of August 16, 2015, to May 17, 2017, contain serious evidence of political bias and the improper handling of investigations within the agency;
the texts contain egregious evidence of bias against President Trump, including Lisa Page stating ''Trump should go f himself'' and Peter Strzok stating ''F TRUMP'';
those text messages were not stored within the FBI archive system, an egregious oversight blamed on a technical glitch, and even after these messages were partially recovered by the Bureau's Inspector General in January 2018, many unanswered questions remain regarding impropriety and bias;
Possible Political Bias?
in March 2018, former FBI Deputy Director McCabe was fired by Attorney General Jeff Sessions who noted that Deputy Director McCabe
''lacked candor—including under oath—on multiple occasions'' and
had partaken in ''unauthorized disclosure to the news media'',
among other violations noted in a report issued by the Office of the DOJ Inspector General after a wide-reaching investigation into Deputy Director McCabe's conduct;
a myriad of DOJ and FBI personnel have been fired or demoted, or have resigned, including
FBI Director Comey,
Deputy Director McCabe,
Chief of Staff to the Director James Rybicki,
FBI General Counsel James Baker,
FBI Agent Strzok, FBI Counsel Page,
FBI Special Agent Josh Campbell,
DOJ Senior Official Ohr,
FBI Assistant Director Michael Kortan, and
Assistant Attorney General Peter Kadzik;
evidence has come to light that raises serious concerns about egregious misconduct within the DOJ and FBI rooted in political bias;
the DOJ, FBI, or both appear to have planted at least one person into Donald Trump's Presidential campaign to infiltrate and surveil the campaign;
the DOJ has failed to timely comply with several related document requests by Congress;
FBI's Redacted Document
providing Members of Congress with
heavily redacted versions of some but not all of the documents demanded and offering Members limited in-person viewing of these documents is an inadequate response to repeated requests after months of delay by the DOJ;
the mission of the Office of the DOJ Inspector General is limited to detecting and deterring waste, fraud, abuse, and misconduct in DOJ programs and personnel and promoting economy and efficiency in those programs, and a fully independent Special Counsel has greater autonomy than an Inspector General or Federal prosecutors to run a non-biased investigation and if necessary bring forth criminal charges; and
the DOJ and FBI cannot be expected to fully investigate themselves regarding this matter:
Now, therefore, be it
That it is the sense of Congress that—
(1) DOJ, FBI, and all Federal law enforcement agencies have a sacred duty to uphold
our Constitution and to protect our country without any partisan or ideological
inclination affecting their important work;
(2) Congress acknowledges with gratitude that the vast majority of the men and women
who serve within these critical agencies do so with the utmost integrity, independence,
patriotism, and commitment to the rule of law;
(3) Misconduct regarding FISA abuse, how and why the Hillary Clinton probe ended, and
how and why the Trump-Russia probe began should immediately be investigated by a
Special Counsel who can act independently; and
(4) The Attorney General of the United States should immediately appoint a Special
Counsel to conduct a thorough and independent investigation of these grave concerns.
Editorial Comment:
We have posted this House Resolution, with highlights and links to related stories, to educate those who are unable to follow this Byzantine plot line by Obama conspirators. It is a highly convoluted scheme, acted out by some of the most corrupt persons in the history of they US -- who, coincidentally were attorneys who should face disbarment proceedings, and prison sentences.
Most stunning was the fact that these attorneys [officers of the Court] held key positions in the Department of Justice or the Federal Bureau of Investigation -- previously the pinnacle of the
US Justice system.
A series of Attorneys General committed the most egregious of crimes, beginning with Eric Holder's Fast&Furious debacle, and then Loretta Lynch's meeting with Bill Clinton on the tarmac, and later, [Acting AG] Sally Yates who, rather than expressing her concerns to President Trump regarding his Executive Order, publicly defied Trump and condemned his Executive Order -- which had been vetted by DOJ's Legal Counsel -- and so Trump fired her.
The trail is now so convoluted that it is difficult to follow; we hope this post has made this coup plot a bit more clear. There is far more to this scheme than this definitive House Resolution, but, there is still more to come, and these revelations will hopefully lead to the arrest and prosecution of all the conspirators -- beginning with Hillary Clinton.
Labels: CIA, Clapper, Clinton, Comey, Corsi, FBI, Fusion, Giustra, Huma, Lynch, McCabe, Mills, Obama, Ohr, Page, Richman, Steele, Strzok, Trump, Weiner
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 3,053
|
{"url":"https:\/\/conferences.lbl.gov\/event\/212\/contributions\/6077\/","text":"# Nuclear Structure 2022\n\n13-17 June 2022\nBerkeley, CA\nUS\/Pacific timezone\n\n## Sub-keV Decay-Recoil Spectroscopy with Superconducting Quantum Sensors\n\n17 Jun 2022, 09:00\n30m\nBerkeley, CA\n\n#### Berkeley, CA\n\nLawrence Berkeley National Laboratory\nInvited Abstract Oral Presentations\n\n### Speaker\n\nDr Kyle Leach (Colorado School of Mines)\n\n### Description\n\nThe recoiling daughter nucleus in weak nuclear decay processes ($\\beta$ or EC decay) is a unique probe for a wide range of nuclear structure properties and BSM physics scenarios. In particular, precision measurements of the recoil can be used to access model-independent information on the neutrino mass, the $\\beta$-$\\nu$ angular correlation, precise excited- and ground-state energies, and decay branching fractions. Despite the unique information contained in these recoiling nuclei, few cases have been studied due to the technical challenges associated with measuring heavy ions with eV-scale kinetic energies with high resolution. To address this, we have developed a program to perform these low-energy measurements using implanted rare isotopes in thin-film superconducting tunnel junctions (STJs). STJs are a cryogenic-charge sensitive quantum sensor technology that allow for count rate of up to 10 kHz\/pixel, provide energy resolutions of 1-5 eV, and have low-energy thresholds of a few eV [1]. In this talk I will discuss this novel experimental concept and present data on the eV-scale measurements we have performed with STJs using $^7$Be (T$_{1\/2}$ = 53 days) EC decay relevant to structure and astrophysics [2] and BSM neutrino physics [3,4]. I will also present plans on extending this concept to perform measurements on short-lived species on-line at RIB facilities, and how our community can better leverage these technologies for high-impact structure measurements.\n[1] S. Friedrich et al., J. of Low Temp. Phys. 200, 200 (2020)\n[2] S. Fretwell et al., Phys. Rev. Lett. 125, 032701 (2020)\n[3] K.G. Leach and S. Friedrich, arXiv: 2112.02029 (2021)\n[4] S. Friedrich et al., Phys Rev. Lett. 126, 021803 (2021)\n\n### Primary author\n\nDr Kyle Leach (Colorado School of Mines)\n\n### Presentation Materials\n\nThere are no materials yet.","date":"2022-08-13 18:24:26","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.3826017379760742, \"perplexity\": 5723.184116420908}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": false}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-33\/segments\/1659882571982.99\/warc\/CC-MAIN-20220813172349-20220813202349-00120.warc.gz\"}"}
| null | null |
\section{Introduction}
Optical clocks using strontium $^{87}\mathrm{Sr}$ are among the most accurate in the world. The time elapsed between two of their ticks is about $10^{-15}\mathrm{s}$ (the inverse of strontium frequency) with a precision of $10^{-19}$ \cite{clocks}. Physical phenomena that probe much smaller characteristic timescales have also been measured. For instance, the lifetime of the top quark is $10^{-25}\mathrm{s}$. Such a result is obtained experimentally from a statistical analysis, where the short duration of the lifetime is compensated by a large number of events. At the theoretical level, physicists consider even shorter scales: in primordial cosmology, the inflation epoch is believed to have lasted $10^{-32}\mathrm{s}$. Based on a cosmological model, the recent paper \cite{Wendel} even argues that the precision of recent atomic clocks already sets an upper bound of $10^{-33} ~\mathrm{s}$ for a fundamental period of time.
Planck time is a far smaller timescale. We recall that the Planck time is defined as
\begin{equation}
t_P \overset{\text{def}}= \sqrt{\frac{\hbar G}{c^5}} \approx 10^{-44}~\mathrm{s},
\end{equation}
where $G$ is Newton's constant, $\hbar$ the reduced Planck's constant and $c$ the speed of light. It can seem an impossible task to probe time at the Planck scale. However, the example of the lifetime of the top quark shows that it is possible to overtake clock accuracy limitations by several orders of magnitude using statistics. Here, we examine the following question: if time behaves differently than a continuous variable at the planckian scale, how could the departure from this behaviour be inferred experimentally? To answer this question, we assume that proper time differences take discrete values in multiple steps of Planck time, and devise a low energy experiment that would detect this effect.
This work is motivated by recent experimental proposals to detect the non-classicality of the gravitational field by detecting gravity mediated entanglement (GME) \cite{BoseSpinEntanglementWitness2017, MarlettoGravitationallyInducedEntanglement2017, Marshman:2019sne, Krisnanda_2020,QISSsougato} and the production of non-gaussianity \cite{Howl:2020isj}.
Since the quantum gravity regime of particle physics may be practically impossible to probe, it is intriguing that these low energy experiments are not too far removed from current capabilities. Instead of accelerators, the suggestion in these proposals is to quantum control slow moving nanoparticles or use a Bose-Einstein condensate.
To understand how this is possible we remark that the Planck mass is a mesoscopic quantity
\begin{equation}
m_P \overset{\text{def}}= \sqrt{\frac{\hbar c}{G}} \approx 2 \times 10^{-8} ~\mathrm{kg}.
\end{equation}
For both of the above experiments, the formula that measures the quantum gravity effect can be cast in the form \cite{Christodouloupossibilitylaboratoryevidence2019, Christodouloupossibilityexperimentaldetection2018b,Howl:2020isj}
\begin{equation} \label{eq:Q}
Q = \frac{m}{m_P} \frac{\delta \tau}{t_P},
\end{equation}
with $m$ the mass probing the gravitational field and $\delta \tau$ a time dilation. In the case of GME, $Q$ is a quantum mechanical phase. In the case of non--gaussianity growth, $Q$ is the signal-to-noise ratio. The effects become most pronounced when $Q$ approaches order unit. We thus see an interesting interplay between the Planck mass and the Planck time: if $Q\sim1$ and $m\sim m_P$, then $\delta \tau \sim t_P$. In the GME experiment \cite{BoseSpinEntanglementWitness2017}, the mass is $m \sim 10^{-6} m_P$, so that it already probes a proper time difference of $\delta \tau \sim 10
^6 t_P$, as was first noticed in \cite{Christodouloupossibilityexperimentaldetection2018b}. Thus, quantum gravity phenomenology provides a further motivation to the current push to develop technologies for setting mesoscopic masses in path superposition \cite{ Arndt:2014bp, RomeroIsart:2009iv, Eibenberger:2013gm}.
The aforementioned proposals aim at demonstrating $qualitative$ results: to witness the production of entanglement, or of non-gaussianity, i.e. $Q \neq 0$, in order to check whether the gravitational field obeys quantum mechanics. Here, we propose a $quantitative$ method for experimentally measuring a set of values for $Q$, in order to test a hypothetical discreteness of time. We see that this requires a careful consideration of the \emph{uncertainty} on $Q$. $Q$ is estimated through a probability $p_+$ of an event occurring, the uncertainty on which must satisfy
\begin{equation}
\label{eq:uncertainty}
\Delta p_+ < \frac{m}{m_P}.
\end{equation}
We see again that the Planck mass acts as a natural scale for the effect to become prominent: smaller masses would require higher precision in estimating the probability $p_+$. The global analysis of the experimental constraints performed in this work shows that the detection of a fundamental time discreteness may not be too far removed from current technological capabilities.
\medskip
The plan is the following:
\begin{enumerate}[itemsep=0 mm]
\item [\ref{sec:experiment}] we present the experimental setup;
\item [\ref{sec:time}] we introduce the hypothesis that proper time differences are discrete at the Planck level;
\item [\ref{sec:visibility}] we deduce the constraints on the experimental parameters to make this discreteness detectable;
\item [\ref{sec:balAct}] we suggest a set of reasonable parameters that fulfil the constraints;
\item [\ref{sec:decoherence}] we complete the analysis by considering decoherence;
\item [\ref{sec:discussion-hypothesis}] we discuss the hypothesis.
\end{enumerate}
\section{Experimental setup}
\label{sec:experiment}
\begin{figure}[h]
\begin{overpic}[width = 0.9\columnwidth]{setup.png}
\end{overpic}
\caption{\textbf{Spacetime view of the experiment.} For a time $t_{acc}$, an inhomogeneous magnetic field is applied that sets a mass $m$ with embedded spin in a superposition of two paths, at a distance $d$ and $d+l$, respectively, from another mass $M$. The masses are in free fall for a time $t$, as measured in the laboratory, after which the procedure is reversed and the superposition undone. During this time $t$, the two trajectories accumulate a different phase due to the gravitational interaction with $M$.
}
\label{fig:setup}
\end{figure}
The proposed experimental setup is depicted in {figure \ref{fig:setup}}. A spherical nanoparticle of mass $m$ with embedded magnetic spin is dropped simultaneously with a second mass $M$. The mass $m$ is then put into a spin-dependent superposition of paths by the application of a series of electromagnetic pulses. This technique was proposed in \cite{BoseSpinEntanglementWitness2017, bose2018matter}. In the branch of closest approach, $m$ and $M$ are at a distance $d$, in the other, they are at a distance $d+l$.
The superposition is held at these distances for a time $t$ as measured in the laboratory frame. While the two masses free fall, they interact gravitationally. If linearised quantum gravity holds, then the two quantum branches in the total state evolve differently, accumulating a relative phase. After the superposition has been undone, this phase is visible in the state of the spin of the mass $m$.
Let us see this in detail. The quantum state of the mass $m$ is given by its position in the apparatus and the orientation of its embedded spin. There will be three relevant position states\footnote{It has recently been shown \cite{chevalier2020witnessing} that treating the position states as eigenstates is a valid approximation in this setup.} $\ket L$, $\ket C$ and $\ket R$, respectively left, centre and right. For the spin, we use the canonical basis, $\ket \uparrow$ and $\ket \downarrow$, in the $z$-direction. The mass $m$ is prepared at $t_0$ in the central position with the spin in the positive $x$-direction:
\begin{equation}
\ket{\psi_0}= \frac{1}{\sqrt2}|C\rangle\left(\ket\uparrow + \ket\downarrow\right).
\end{equation}
An inhomogeneous magnetic field is then applied to the mass $m$, entangling its position with its spin so that at time $t_1$ the state is
\begin{equation}
\ket{\psi_1} = \frac{1}{\sqrt2}\left(\ket{L\uparrow} + \ket{R\downarrow}\right).
\end{equation}
The particle is then allowed to free-fall for a time $t$. During this time, it interacts gravitationally with the mass $M$. The displacement of the masses due to their gravitational attraction is negligible. The two states $\ket L$ and $\ket R$ are eigenstates of the hamiltonian and each acquires a phase proportional to the newtonian potential induced by $M$. So at time $t_2$ the state is
\begin{equation}
|\psi_2\rangle = \frac{1}{\sqrt{2}}(e^{i\phi_L}|L\uparrow\rangle + e^{i\phi_R} |R\downarrow\rangle),
\end{equation}
where
\begin{equation}
\phi_L = \frac{GMm}{\hbar}\frac{t}{d+l} ~~~\mbox{and}~~~
\phi_R = \frac{GMm}{\hbar}\frac{t}{d}.
\end{equation}
At this point, another inhomogeneous magnetic field is applied to undo the superposition. The final state of the particle is, up to a global phase,
\begin{equation}\label{eq:psi_final}
|\psi_3\rangle = \frac{1}{\sqrt2}\ket C \left(\ket\uparrow+e^{i\delta \phi}\ket\downarrow\right),
\end{equation}
where the relative phase $\delta \phi$ is given by
\begin{equation}\label{eq:phi}
\delta \phi = \frac{GMmt}{\hbar}\frac{l}{d(d+l)}.
\end{equation}
Information about the gravitational field is now contained in the state of the spin, which in turn can be estimated from the statistics of spin measurements.
Concretely, we consider a measurement on the spin of the particle along the $y$-direction
\begin{equation}
\ket{\pm i} \overset{\text{def}}= \frac{1}{\sqrt2}\left(\ket\uparrow\pm i\ket\downarrow\right).
\end{equation} Born's rule gives the probability $P_+$ of finding the spin in the state $\ket{+i}$:
\begin{equation}\label{eq:p+}
P_+(m,M,d,l,t) = \frac12+\frac12\sin \delta \phi,
\end{equation}
where we compute $\delta \phi$ as a function of $m,M, d, l$ and $t$ through equation \eqref{eq:phi}. This equation for the probability is a theoretical prediction of linearised quantum gravity.
Experimentally, the probability can be measured by the relative frequencies in collected statistics. The experiment is repeated $N$ times keeping the experimental parameters fixed. If the outcome $\ket{+i}$ is recorded $N_+$ times, the frequency
\begin{equation}\label{eq:p+exp}
p_+(m,M,d,l,t) \overset{\text{def}}= \frac{~N_+}{N}
\end{equation}
is then the experimentally measured value of the probability.
This procedure can be repeated for different sets of experimental parameters to verify the functional dependence of $p_+$ to these. In what follows, we propose an experiment that can detect a statistically significant discrepancy between $P_+$ and $p_+$. This discrepancy would signal a departure from linearised quantum gravity.
The above experimental setup is similar to that proposed to detect GME in \cite{BoseSpinEntanglementWitness2017}, with the main difference that for our purpose we only require one mass, not two, in a superposition of paths. It is thus conceptually more similar to the celebrated Colella-Overhauser-Werner (COW) experiment \cite{Colella1975, abele2012gravitation}.
However, the task we have set ourselves here and the method to achieve it, goes much beyond showing that gravity can affect a quantum mechanical phase and induce an interference pattern. To detect a potential discreteness of time, we need a more sensitive apparatus, and so the gravitational source $M$ will need to be much \emph{weaker}. In our case, $M$ is not the Earth, but a mesoscopic particle, essentially a speck of dust.
\section{Hypothesis: Time Discreteness}
\label{sec:time}
While the newtonian limit of linearised quantum gravity is sufficient to compute the phase difference $\delta \phi$, it is can also be understood in general relativistic terms \cite{Christodouloupossibilitylaboratoryevidence2019, Christodouloupossibilityexperimentaldetection2018b}. The mass $M$ induces a Schwarzschild metric which dilates time differently along each of the two possible trajectories of $m$. Then, equation \eqref{eq:phi} can be recast as
\begin{equation}\label{eq:main2}
\delta \phi = \frac{m}{m_P}\frac{\delta\tau}{t_P},
\end{equation}
where $\delta \tau$ is the difference of proper time between the two trajectories, given by
\begin{equation}\label{eq:delta_tau}
\delta\tau = \frac{G M}{c^2} \frac{l }{ d(d+l)} t.
\end{equation}
Now, it is widely believed that the smooth geometry of general relativity should be replaced, once quantised, by some discrete structure. In particular, we may expect time to be granular in some sense. In which sense precisely, we do not know. However, since $\delta \tau$ admits a straightforward interpretation of a covariant quantum clock, it makes a good candidate to reveal discrete features of time. Thus we make the following hypothesis: $\delta \tau$ can only take values which are integer multiples of Planck time $t_P$. That is, \eqref{eq:delta_tau} is modified to:
\begin{equation}\label{eq:discrete}
\delta\tau = n \, t_P, \quad n \in \mathbb{N}.
\end{equation}
Additional motivation for the hypothesis and possible alternatives are discussed in section \ref{sec:discussion-hypothesis}. For now, it can be taken just as the simplest implementation of the idea that time is discrete at a fundamental level, similar in philosophy to the idea that everyday-life matter is not continuous, but instead made of atoms. Devising an experiment to detect this discreteness and examining its feasibility is the task we have set ourselves in this work.
Equation \eqref{eq:discrete} is still incomplete and we need to posit a functional relation between the level $n$ and the parameters $M,d,l,t$. We rewrite equation \eqref{eq:delta_tau} as
\begin{equation}\label{eq:smooth}
\delta\tau = \frac{t}{\beta} \; t_P,
\end{equation}
where
\begin{equation} \label{eq:beta}
\beta \overset{\text{def}}= \frac{d(d+l)c^2}{GMl} t_P,
\end{equation}
and we take $n$ to be given by the floor function
\begin{equation}\label{eq:floor}
n = \floor{\frac{t}{\beta}}.
\end{equation}
That is, $n$ is the integer part of the dimensionless quantity $t/\beta$. The main lessons of our results do not depend on the specific choice \eqref{eq:floor} for the functional dependence between $t/\beta$ and $n$. Other modifications of the continuous behaviour in \eqref{eq:delta_tau}, so long as they display features of planckian size, could be probed by the experiment.
We have
\begin{equation}\label{eq:hypothesis}
\delta \tau = \floor{\frac{t}{\beta}} t_P.
\end{equation}
The consequences of this hypothesis are revealed in the measured probability $p_+$
of equation \eqref{eq:p+exp}.
If time behaves continuously, $p_+$, as a function of time $t/\beta$ will fit the smooth (blue) curve in figure \ref{fig:preliminary}, given by
\begin{equation}\label{eq:p_smooth}
P_+ = \frac12+\frac12\sin \left( \frac{m}{m_P} \frac{t}{\beta} \right).
\end{equation}
If the hypothesis holds, the observed profile for the probability will follow that of the red step function in figure \ref{fig:preliminary}, given by
\begin{equation}\label{eq:p_discrete}
P^h_+ = \frac12+\frac12\sin \left( \frac{m}{m_P} \floor{ \frac{t}{\beta}} \right).
\end{equation}
To test the hypothesis, the strategy is thus to plot experimentally the curve $p_+(t/\beta)$. Observing plateaux would be the signature of time-discreteness.
\begin{center}
\begin{figure}[h]
\includegraphics[width = \columnwidth]{Probability.png}
\caption{{\bf Probability of measuring spin $\ket{+i}$ as a function of $t/\beta$} under the continuous and discrete time hypotheses. Blue line: $\delta \tau$ is smooth as in equation \eqref{eq:smooth}. Red line: $\delta \tau$ is discrete as in equation \eqref{eq:hypothesis}. We have taken the value of $m=10^{-2}m_P$. The experimental parameters shown in table \ref{tab:tableB} would produce 100 data points scanning the range of $t/\beta$ depicted here, with a sufficient resolution to decide which of the two curves is realised in nature. }
\label{fig:preliminary}
\end{figure}
\end{center}
\section{Ensuring Visibility of the Effect}
\label{sec:visibility}
Each experimental data point for $p_+(t/\beta)$ is obtained from computing the statistical frequency of the outcome $\ket{+i}$. Point by point, a scatter plot of $p_+$ against $t/\beta$ will be obtained. We must choose the experimental parameters so that the difference between $P_+$ and $P^h_+$ can be resolved. This imposes requirements on the minimal precision of the experimental apparatus and on the maximal permissible gravitational noise in the environment.
\subsection{Visibility of the Vertical Axis}
\label{sec:vertical}
The uncertainty $\Delta p_+$ for the probability $p_+$ after $N$ runs results from using finite statistics and is of the order
\begin{equation} \label{eq:visX}
\Delta p_+ \sim \frac1{ \sqrt N}.
\end{equation}
The vertical step $\alpha$ between the plateaux is given by
\begin{equation}
\alpha =\left|\sin\left(\left(\floor{ \frac{t}{\beta}}+1\right)\frac{m}{m_P}\right)-\sin\left(\floor{ \frac{t}{\beta}}\frac{m}{m_P}\right)\right|.
\end{equation}
We assume that $m\ll m_P$, consistent with the fact that it is hard to put a large mass in a superposition. The above expression simplifies to
\begin{equation}
\alpha(t) \approx \frac{m}{m_P} \cos \left( \floor{ \frac{t}{\beta}}\frac{m}{m_P} \right).
\end{equation}
So the steps are most visible when
\begin{equation}\label{eq:alpha-zero}
\left|\frac{t}{\beta}\frac{m}{m_P}\right| \ll 1 .
\end{equation}
Then the expression simplifies to
\begin{equation}
\alpha(t) \approx \frac{m}{m_P}.
\end{equation}
Requiring that the probability uncertainty is an order of magnitude smaller than the vertical step, ${\Delta p_+ < 10^{-1}\alpha}$, we find the constraint
\begin{equation}
N >10^{2} \left( \frac{m_P}{m} \right)^2.
\label{eq:number}
\end{equation}
We see that a larger mass $m$ means that fewer runs $N$ per data point are required, which implies a shorter total duration $T_{tot}$ of the experiment.
Indeed, since plotting $p_+(t/\beta)$ requires $N$ runs per data point, each run requiring at least a time $t$, a lower bound for the total duration of the experiment is
\begin{equation}
T_{tot} \sim N_{dp} N t,
\end{equation}
where $N_{dp}$ is the number of data points. Thus, the constraint \eqref{eq:number} can be restated as
\begin{equation} \label{eq:con1}
\frac{T_{tot}}{ N_{dp} \; t} >10^{2} \left( \frac{m_P}{m} \right)^2.
\end{equation}
This constraint imposes a trade-off between the time required to resolve the discreteness and the mass that has to be in superposition. It counter-balances the fact that it is harder to achieve quantum control of a large mass.
\subsection{Visibility of the Horizontal Axis}
\label{sec:horizontal}
The uncertainty in $t/\beta$ is found via the the standard formula for the propagation of uncertainty and can be expressed as
\begin{equation}
\Delta(t/\beta) = U\frac{t}{\beta},
\end{equation}
where
\begin{multline}
U \overset{\mathrm{def}}{=}
\left[ \left(\frac{\Delta t}{t}\right)^2
+\left(\frac{d}{d+l}+1\right)^2\left(\frac{\Delta d}{d}\right)^2 \right.
\\
\left. + \left(\frac{\Delta M}{M}\right)^2
+\left(\frac{d}{d+l}\right)^2\left(\frac{\Delta l}{l}\right)^2 \right]^{\frac{1}{2}}.
\label{eq:defU}
\end{multline}
By assumption \eqref{eq:floor}, the width of the plateaux is $1$.
To place several data points on each plateau, we require the typical uncertainty to be an order of magnitude smaller, i.e. $\Delta(t/\beta) < 10^{-1}$. We thus impose the constraint
\begin{equation}
U < 10^{-1}~ \frac{\beta}{t}
\label{eq:time}
\end{equation}
on the experimental parameters. Note that a given $U$ determines the highest value of $n = \floor{t/\beta}$ for which the discontinuities can be resolved.
\subsection{Gravitational Noise }
\label{sec:gravitational}
There is no analog of a Faraday cage for gravitational interactions, so influences by other masses will also contribute to the accumulated phase $\delta \phi$. Since the experiment we are considering is in a sense an extremely sensitive gravimeter, these would need to be taken carefully into account.
We distinguish between `predictable' gravitational influences and `unpredictable' gravitational influences, i.e. gravitational noise.
The latter type will dictate the degree of isolation required for a successful realisation of the experiment, adding another visibility constraint, while the former type can be dealt with by calibration.
The presence of unexpected masses in the vicinity of the apparatus may disturb the measurement. It will contribute to the proper time dilation by an amount $\eta$, modifying \eqref{eq:p_discrete} to
\begin{equation}
P^h_+(\eta) = \frac12+\frac12\sin\left(\frac{m}{m_P}\floor{\frac{t}{\beta}+\frac{\eta}{t_P}}\right).
\end{equation}
Getting a single data-point requires $N$ drops, and for each drop, the perturbation $\eta$ may be a priori different. However, it should be small enough so that it does not make the probability $P^h_+$ jump to another step, i.e. $\eta$ is a negligible noise if
\begin{equation}\label{eq:eta}
\floor{\frac{t}{\beta}+\frac{\eta}{t_P}} = \floor{\frac{t}{\beta}}.
\end{equation}
Of course, $\eta$ is a random variable over which the control is limited. To a first approximation, the condition \eqref{eq:eta} can be implemented over the $N$ drops by requiring
\begin{equation}\label{eq:grav_eta}
\Delta \eta < 10^{-1} t_P.
\end{equation}
For instance, the gravitational noise induced by the presence of a mass $\mu$ at a distance $D\gg l,d$ is at most
\begin{equation}
\eta_{max} = \pm\frac{G\mu l}{D^2}\frac{t}{\hbar}.
\end{equation}
Thus, we get a fair idea of how isolated the apparatus should be with the condition
\begin{equation}
2 G l \frac{\mu }{D^2} \frac{t}{\hbar} < 10^{-1} t_P.
\end{equation}
The ratio
\begin{equation}
A \overset{\mathrm{def}}{=} \frac{\mu}{D^2}
\end{equation}
is a measure of the impact that a mass $\mu$ has on the visibility of the discontinuities if it is allowed to move uncontrollably as close as a distance $D$ away from the experiment. Thus, we end up with the following constraint
\begin{equation}\label{eq:gravinf}
A \, l \, t < 5 \times 10^{-2} \, \frac{t_Pm_P}{l_P}.
\end{equation}
This equation is a requirement on the control of the environment necessary to resolve the discontinuities. Shorter superpositions are less sensitive to the gravitational noise.
Above, we took into account the effect of a single mass $\mu$. This not sufficient to guarantee that there will not be a cumulative effect from several masses around. However, note that if these masses are homogeneously distributed, their contributions average out.
The `predictable' type of gravitational influences are systematic errors arising for example from the gravitational field of the Earth, the Moon, and the motion of other large bodies, such as tectonic activity or sea tides, but also from small masses that will unavoidably be present in the immediate vicinity of the mass $m$, such as the experimental apparatus itself and the surrounding laboratory. Given the extreme sensitivity of the apparatus, it will likely not be possible to make all these gravitational influences satisfy \eqref{eq:gravinf}.
However one can calibrate for the contribution of a mass $\mu$ at distance $D$ if it moves slowly with respect to the time $N t$ that it takes to collect a data point\footnote{A simple method to calibrate when the different values of $t/\beta$ are obtained by changing $d$ only while keeping $M$, $l$ and $t$ fixed, as considered in the following section, is the following. The mass $\mu$ will contribute a constant phase $\phi_B$. The state of the mass $m$ when the experiment is performed \textit{without} $M$ present is $(\ket0+e^{i\phi_B}\ket1)/\sqrt2$. We can estimate the phase $\phi_B$ by running the experiment without $M$. So long as the masses are slow moving, it suffices to rotate the measurement basis to $(\ket0+e^{i\phi_B}\ket1)/\sqrt2$ rather than $\{\ket{\pm i}\}$.}, i.e. if
\begin{equation}
Nt v \ll D,
\end{equation}
with $v$ the speed of the mass. Another possibility that can be calibrated for is if the mass is not moving slowly but the uncertainty in its position is small with respect to $D$ (for instance, a moving mechanical part or the Moon).
\section{Balancing act}
\label{sec:balAct}
The three experimental constraints identified in the previous subsection are repeated below.
\begin{equation}
\left\{~~
\begin{aligned}
10^2 \, \frac{ N_{dp} \; t}{T_{tot}} < \left(\frac{m}{m_P}\right)^2 & \quad \text{[Vertical]}\\
U\frac{t}{\beta} < 10^{-1} & \quad \text{[Horizontal]} \\
A \, l \, t < 5 \times 10^{-2} \, \frac{t_P \; m_P}{l_P} & \quad \text{[Noise]},
\end{aligned}
\right.
\label{eq:ineqs}
\end{equation}
with
\begin{equation}
\frac{t}{\beta} = \frac{M}{m_P}\frac{ctl}{d(d+l)}.
\end{equation}
We now proceed to identify a set of reasonable parameters that satisfy the constraints. Our series of assumptions is an educated guess based on our understanding of current technological trends.
\begin{enumerate}
\item Any of the parameters $M$, $d$, $l$ and $t$ could be modulated to scan a range of $t/\beta$. Since $t/\beta$ is most sensitive to changes in $d$ (quadratic dependence), we assume the modulation of $d$, keeping $M$, $l$ and $t$ fixed.
\item The total duration of the experiment is about a year
\begin{equation}
T_{tot} \sim 10^7 \mathrm{s}.
\end{equation}
\item The plot requires about a hundred of data points
\begin{equation}
N_{dp} \sim 10^{2},
\end{equation}
to be distributed over ten plateaux
\begin{equation}
t/\beta \leq 10.
\end{equation}
\item
Experimentally, the maximal distance between the two branches of the superposition cannot be very large, and so we assume
\begin{equation}
d \gg l.
\end{equation}
\end{enumerate}
From these first assumptions, the system of inequalities simplifies to
\begin{equation}
\left\{~~
\begin{aligned}
t < 10^3 \left(\frac{m_P}{m}\right)^2 ~\mathrm{s} & \quad \text{[Vertical]}\\
U < 10^{-2} & \quad \text{[Horizontal]} \\
A \, l \, t < 5 \times 10^{-2} \, \frac{t_P \; m_P}{l_P} & \quad \text{[Noise]} \\
t/\beta \leq 10 & \quad \text{[Range]},
\end{aligned}
\right.
\end{equation}
with
\begin{equation}
\frac{t}{\beta} = \frac{M}{m_P}\frac{ctl}{\phantom{{}^2}d^2}.
\end{equation}
The uncertainty $U$, defined by equation \eqref{eq:defU}, depends on the precision in $t$, $M$, $d$ and $l$.
With the assumption $l\gg d$ its expression simplifies to
\begin{equation}
U = \sqrt{\left(\frac{\Delta t}{t}\right)^2
+ \left(\frac{\Delta M}{M}\right)^2
+\left(\frac{\Delta d}{d}\right)^2+\left(\frac{\Delta l}{l}\right)^2}.
\end{equation}
Then, the [Horizontal] inequality implies that $t$, $M$, $d$ and $l$ will have to be controlled better than 1 part in 100.
\begin{enumerate}\setcounter{enumi}{4}
\item It is reasonable to expect that the uncertainty $U$ will be dominated by the uncertainty in the superposition size $l$, thus,
\begin{equation}
U \approx \frac{\Delta l}{l}.
\end{equation}
\item We assume possible to control the size of the superposition to the scale of a few atoms, i.e.
\begin{equation}
\Delta l = 10^{-9}\mathrm{m}.
\end{equation}
\item From the above two points we have a lower bound for the value of $l$. Taking $l$ larger, would only make the experiment harder because of decoherence and gravitational noise. We thus take
\begin{equation}
l \sim 10^{-7}\mathrm{m}
\end{equation}
which satisfies the horizontal constraint, allowing to resolve the first 10 steps.
\end{enumerate}
We have now solved the horizontal constraint and fixed $l$. The remaining constraints evaluate to
\begin{equation}
\label{eq:VPR}
\left\{~~
\begin{aligned}
t < 10^3\left(\frac{m}{m_P}\right)^2 \mathrm{s} & \quad \text{[Vertical]}\\
A \, t < 4 \times 10^{-11}~\mathrm{kg~s~ m}^{-2} & \quad \text{[Noise]} \\
\frac{Mt}{\phantom{{}^2}d^2} < 7\times10^{-9}~\mathrm{kg~s~m}^{-2} & \quad \text{[Range]}.
\end{aligned}
\right.
\end{equation}
All three equations suggest to take $t$ as small as possible. Nonetheless, this cannot be too short because the superposition is created by a magnetic field $B$ that separates the branches at a distance $l$. This process requires some time $t_{acc}$, which is bounded from below by the highest magnetic field $B_{max}$ that can be created in the lab. Concretely\footnote{We assume the masses are made of a material that allows neglecting diamagnetic effects. If diamagnetism cannot be ignored, one has to resort to a more complicated scheme of pulses, inverting the direction of the magnetic field gradient at specific intervals as detailed in \cite{pedernales2020motional}, or inverting both the direction of the gradient and the spins as proposed in \cite{vandekamp2020quantum}. Alternatively, one can use a different method of wavepacket separation, like that detailed in \cite{Pino:2016we}.
}
\begin{equation}
\label{eq:B1}
\mu_B \frac{B_{max}}{l} > \frac{m l }{t_{acc}^2},
\end{equation}
where $\mu_B$ is the Bohr magneton ($\mu_B \approx 10^{-23} ~\mathrm{J.T^{-1}}$).
\begin{enumerate}\setcounter{enumi}{7}
\item $t$ should be at least as long as $t_{acc}$, say
\begin{equation}
\label{eq:B2}
t \sim 3 \, t_{acc}.
\end{equation}
\end{enumerate}
\begin{enumerate}\setcounter{enumi}{8}
\item Taking $B_{max} \sim 10^2 ~\mathrm{T}$, which is the value of the strongest pulsed non-destructive magnetic field regularly used in research \cite{magnet}, we get in SI units
\begin{equation}
\label{eq:t2>m}
10^{-8} t^2 > m.
\end{equation}
\item Considering the difficulty to put a heavy mass in superposition, we can minimise both $t$ and $m$ under the vertical constraint of \eqref{eq:VPR} and equation \eqref{eq:t2>m}. We find
\begin{equation}
\begin{split}
&m = 3 \cdot 10^{-10} ~\mathrm{kg} \sim 10^{-2} m_P\\
&t = 10^{-1} ~\mathrm{s}.
\end{split}
\end{equation}
\end{enumerate}
These values are consistent with the assumptions made above that $m \ll m_P$ and $\Delta t / t \ll 10^{-2}$.
We have thus solved the Vertical constraint too. We are left with
\begin{equation}
\left\{~~
\begin{aligned}
A < 4 \times 10^{-10}~\mathrm{kg~ m}^{-2} & \quad \text{[Noise]} \\
\frac{M}{d^2} < 7 \times10^{-8}~\mathrm{kg~m}^{-2} & \quad \text{[Range]}.
\end{aligned}
\right.
\end{equation}
\begin{enumerate}
\setcounter{enumi}{10}
\item Considering a priori the difficulty to isolate the system from external perturbations, the noise inequality fixes the minimal upper bound for $A$, i.e. we want to tolerate perturbations as high as
\begin{equation}
A = 4 \times10^{-10}~\mathrm{kg~m}^{-2}.
\end{equation}
\end{enumerate}
This threshold is very sensitive. To give an example, it corresponds to the gravity induced by a bee flying $230\mathrm{m}$ away. Such a high control might only be attainable in space, where cosmic dust particles, with typical mass of $5\mu\mathrm{g}$ \cite{carrillo2015size}, would need to be kept $4\mathrm{m}$ away from the masses.
We are thus left with one last inequality which reads, in SI units,
\begin{equation}
d> 4 \times10^3 \sqrt{M}. \quad \text{[Range]}
\end{equation}
\begin{enumerate}
\setcounter{enumi}{11}
\item We have implicitly assumed that $m$ is a test mass moving in the geometry defined by $M$, so we require $M \gtrsim 10 ~m$ for consistency. Choosing the minimal value
\begin{equation}
M = 10~m,
\end{equation}
leads to
\begin{equation}
d \geq 0.17~\mathrm{m}.
\end{equation}
\end{enumerate}
This corresponds to the lower bound for the range that $d$ will scan, corresponding to $t/\beta = 10$. The value $t/\beta=1$ provides an upper bound of $d \approx 54 ~\mathrm{cm}$. Note that the assumption made above that $\Delta d/d,\, \Delta M /M \ll 10^{-2}$ is indeed reasonable.
\paragraph*{Casimir-Polder.} So far, we have not taken into account the Casimir-Polder (CP) force between the two masses. The modification of the vacuum energy between two perfectly conducting, parallel discs of area $a$ a distance $d$ apart \cite{schwartz2014quantum} results in a force $F_{\mathrm{CP}}=\frac{\hbar c \pi^2}{240d^4}a$. Taking this force as an overestimate of that between two spherical dielectric particles of cross-sectional area $a$ a distance $d$ apart, we see that the CP force is at most a million times weaker than the gravitational force and can thus be neglected.
\paragraph*{Uncertainty on $m$.}
A small shift $\delta m$ on the mass $m$ adds a phase difference $\epsilon = \delta m/m_P\cdot\floor{t/\beta}$,
which in turn causes a shift $\delta P$ in the probability. Since $m\ll m_P$ and $t/\beta<10$, then $\epsilon\ll 1$ and the shift is to first order $\delta P \approx \frac{1}{2}\epsilon$.
The uncertainty in $m$ does not affect the visibility of the probability axis if $\delta P \ll \alpha$, i.e. if $\delta m/m \ll 2/\floor{t/\beta}$.
This last condition on $m$ means that the mass $m$ should be known to one part in 100, which is easily reachable.
\medskip
This concludes our derivation of a set of parameters that satisfy the constraints of the previous section and, thus, allow to probe planckian features of time.
The values are summarised in table \ref{tab:tableB}.
As a corroboration of the analysis, the experimental plot is simulated for these parameters in figures \ref{fig:simulated} and \ref{fig:simulated_grav}. There, we see how the effect becomes visible when the gravitational noise and the uncertainty on the experimental parameters satisfy the constraints derived above.
\begin{table}
\begin{center}
\begin{tabular}{c c c}
\hline
Parameter & Value & Uncertainty \\
\hline
$m$ & $3 \times 10^{-10}~\mathrm{kg}$ & $10^{-12}~\mathrm{kg}$\\
$M$ & $3 \times 10^{-9}~\mathrm{kg}$ & $10^{-11}~\mathrm{kg}$\\
$t $ & $10^{-1}~\mathrm{s}$ & $10^{-4}~\mathrm{s}$ \\
$l$ & $10^{-7}~\mathrm{m}$ & $10^{-9}~\mathrm{m}$ \\
$d$ & $ [17,54]~\mathrm{cm}$ & $10^{-2}~\mathrm{cm}$ \\
\hline
$A$ & $\leq 4 \times10^{-10}~\mathrm{kg~m}^{-2}$\\
\hline
$N_{dp}$ & $100$ &\\
$N$ & $10^6$ &\\
$T_{tot}$ & $1 \;\text{year}$ & \\
$n$ & $ [0,10]$
\end{tabular}
\end{center}
\caption{The experimental parameters identified in section \ref{sec:balAct}}
\label{tab:tableB}
\end{table}
\begin{figure*}[ht]
\includegraphics[width=\textwidth]{simulations.png}
\caption[width=2\columnwidth]{\textbf{Simulated data points with decreasing values of $\Delta l$.} The value of the parameters is set as in table \ref{tab:tableB}, assuming no gravitational noise.
Each point point is obtained by sampling $N$ times the probability distribution $P^h_+$ in \eqref{eq:p_discrete}, where the parameters $t$ , $l$ and $d$ are themselves each time sampled from a normal distribution with the corresponding uncertainty.
From left to right, the uncertainty in $l$ takes the values $10^{-8}\mathrm{m}, 5\times10^{-9}\mathrm{m}$ and $10^{-9}\mathrm{m}$, demonstrating that the effect becomes visible when the experimental parameters have little uncertainty, see section \ref{sec:horizontal}. Note how the discontinuities on higher values of $t/\beta$ require higher precision to be resolved.
}
\label{fig:simulated}
\end{figure*}
\begin{figure*}[ht]
\includegraphics[width=\textwidth]{simulations_grav.png}
\caption[width=2\columnwidth]{\textbf{Simulated data points with decreasing gravitational noise.} The data points are obtained in the same manner as those in figure \ref{fig:simulated}, with the following difference. At each run, a value of $A$ is picked uniformly at random from $[-A_{max},A_{max}]$ and the quantity $Alt$ is added to $t/\beta$ before sampling the distribution. This procedure simulates the influence of a single mass moving uncontrollably while statistics are collected, see section \ref{sec:gravitational}. The value of the parameters is as set in table \ref{tab:tableB}, while $A_{max}$ is, from left to right, $1/(2tl), 1/(5tl)$ and $1/(20tl)$ in natural units. The discontinuities become visible only if the gravitational noise is reduced.
}
\label{fig:simulated_grav}
\end{figure*}
\section{Maintaining Coherence}
\label{sec:decoherence}
A mass in superposition of paths will interact with the ambient black body radiation and stray gas molecules in the imperfect vacuum of the device. As the photons and molecules get entangled with the position degrees of freedom of the mass, the coherence of the superposition is lost and the phase cannot be recovered by observing interference between the two paths.
These unavoidable environmental sources of decoherence are well studied both theoretically and experimentally \cite{Pino:2016we, RomeroIsart:2009iv, RomeroIsart:2011es}. Gravitational time dilation can also be a source of decoherence for thermal systems \cite{pikovski2015universal}, but requires much stronger gravitational fields than considered in this experiment.
We assume the experiment will be performed with a nanoparticle of mass $m=3\times10^{-10}~\mathrm{kg}$, radius ${R=30~\mu\mathrm{m}}$. For the formulae appearing in this section we refer the reader to \cite{Pino:2016we}.
\subsubsection{Black-Body Radiation}
\label{sec:decoherence_bb}
The typical wavelength of thermal photons ($\approx10^{-5}\mathrm{m}$ at room temperature) is much larger than $l$, thus spatial superpositions decohere exponentially in time with a characteristic time
\begin{equation}
\tau_{bb} = \frac{1}{\Lambda_{bb} l^{2}},
\end{equation}
which is sensitive to the superposition size $l$.
The factor $\Lambda_{bb}$ depends on the material properties of the mass as well as its temperature and that of the environment. If the environment and the mass are at the same temperature $T$ then the factor is
\begin{equation}
\begin{aligned}
\Lambda_{bb} =&
\frac{8!8\zeta(9)}{9\pi}
cR^6\left(\frac{k_BT}{\hbar c}\right)^9\mathrm{Re}\left[\frac{\epsilon-1}{\epsilon+2}\right]^2 \\
&+\frac{32\pi^5}{189}cR^3\left(\frac{k_BT}{\hbar c}\right)^6\mathrm{Im}\left[\frac{\epsilon-1}{\epsilon+2}\right],
\end{aligned}
\end{equation}
where $\zeta$ denotes the Riemann zeta function and $\epsilon$ is the dielectric constant of the nanoparticle's material at the thermal frequency. We take $\epsilon = 5.3$ like that of diamond \cite{bhagavantam1948dielectric} for the purposes of this estimation. Plugging in the the radius of $30~\mu\mathrm m$ of the masses under consideration and the superposition size $10^{-1}\mu\mathrm{m}$, we have
\begin{equation}
\tau_{bb} \approx \frac{2\times10^{5}~\mathrm{s}}{\left(T/\mathrm{K}\right)^9}.
\end{equation}
A coherence time of about $1~\mathrm{s}$, one order of magnitude above $t$ of table \ref{tab:tableB}, will require the temperature to be below $4~\mathrm{K}$.
\subsubsection{Imperfect vacuum}
\label{sec:decoherence_vacuum}
The thermal de Broglie wavelength of a typical gas molecule ($\approx10^{-10}\mathrm{m}$ for He at $4\mathrm{K}$) is many orders of magnitude below the superposition size $l$ considered here, thus a single collision can acquire full which-path information and entail full loss of coherence. The exponential decay rate of the superposition is in this case independent on the size $l$ of the superposition, with a characteristic time
\begin{equation}
\tau_\mathrm{gas}=\frac{\sqrt{3}}{16\pi\sqrt{2\pi}}\frac{\sqrt{2m_gk_BT}}{PR^2}
\end{equation}
in a gas at temperature $T$, pressure $P$ of molecules of mass $m_g$. Assuming the gas is entirely made of helium, and setting the highest possible value for the temperature according to the previous section, we get
\begin{equation}
\tau_\mathrm{gas}\approx\frac{10^{-17}\mathrm{s}}{P/\mathrm{Pa}}.
\end{equation}
Thus a coherence time of $10\, t = 1\mathrm{s}$ requires a pressure of $10^{-17}\mathrm{Pa}$. This is a regime of extremely low pressure and may present the most serious challenge for any experiment that involves setting masses of this scale in path superposition. To put things in perspective, pressures of the order $10^{-18}\mathrm{Pa}$ are found in nature in the warm-hot intergalactic medium \cite{missingbaryons}, while the interstellar medium pressure is at the range of $10^{-14}\mathrm{Pa}$ \cite{RevModPhys.73.1031}.
On the other hand, pressures as low as $10^{-15}\mathrm{Pa}$ at $4\,\mathrm{K}$ have been reported since the 1990's in experiments employing cooling magnetic traps \cite{gabrielse1990thousandfold,gabrielse2001comparing}.
In a similar context to ours, the contemporary GME detection proposals quoted above require pressures of $10^{-15}\mathrm{Pa}$ at $0.15~\mathrm{K}$ \cite{BoseSpinEntanglementWitness2017}. Finally, the cryogenic requirements found in this section can be relaxed if the path superposition can be achieved faster. From equations \eqref{eq:B1} and \eqref{eq:B2}, if a stronger magnetic field can be used this will require shorter coherence times.
\section{Discussion of the hypothesis}
\label{sec:discussion-hypothesis}
At first sight, the hypothesis \begin{equation}
\delta \tau = n \, t_P
\tag{\ref{eq:discrete}}
\end{equation}
mimics the na\"{i}ve picture of a tiny clock ticking at a constant rate, with a lapse $t_P$. This simple physical picture of the quantum mechanical phase as a sort of intrinsic "clock" ticking at planckian time intervals is appealing in its simplicity and does not depend on any particular model of quantum gravity. Thus, in our opinion, it is on its own right worth being looked at.
Whether this hypothesis is backed by a physical theory of time is unclear. In the well corroborated fundamental paradigms of general relativity and quantum mechanics, time is modelled as a continuous variable. However, in a more fundamental theory like quantum gravity, yet to be established, one can reasonably expect a modification of the notion of time at planckian scale. We discuss two main avenues by which the continuous time can become discrete:
\begin{enumerate}
\item [A.] Instead of a smooth spacetime, consider it instead an effective description on large scales, that emerges from an underlying discrete lattice.
\item [B.] Promote time to a quantum observable with a discrete spectrum.
\end{enumerate}
\medskip
{\bf A.} Most straightforwardly, \eqref{eq:discrete} can be taken prima facie to arise from a kind of classical time discreteness. Assuming that the notion of proper time $\tau$ of general relativity becomes discrete in a linear sense, with regular spaced planckian time intervals, then also \textit{differences} of proper time $\delta \tau$ will display a similar behaviour, from which \eqref{eq:discrete} follows. This assumption is made for instance in the program of Digital Physics \cite{Zuse}, which advocates that space may be nothing but a grid.
Of course, such a `classical' discreteness would manifestly break Lorentz invariance. It might be already possible to set upper bounds on the discreteness of time from the limits set on Lorentz invariance violations by the study of the dispersion relations of light \cite{Jacobson:2005bg, Ackermann:2009aa, amelino, Nemiroff:2011fk}.
Before discussing possible implications of quantum theory, a comment on the intermediate case of a classical but stochastic spacetime. For instance, if spacetime can be described by a single causal set, stochastic fluctuations of planckian size in proper times are to be expected \cite{Rideout_1999, Dowker_2005, sorkin2003causal}. Because of the statistical nature of the time measurement proposed here, finding a continuous behaviour for $\delta \tau$ would not necessarily exclude the possibility of a classical discreteness. It could just be masked by stochastic fluctuations.
\medskip
{\bf B.} Turning to the quantum theory, the discreteness of time may appear as the discreteness of the spectrum of some time operator. Contrary to general belief, Pauli's argument \cite{Pauli} has not ruled out the possibility of a time-operator but rather stressed the subtlety of its definition \cite{Galapon}.
There are two main candidates for being the relevant time observable here: the proper time interval $\tau$ in each branch and the difference of proper time $\delta \tau$ between the branches. Then in both cases the question of which spectrum is to be expected should be answered.
Equation \eqref{eq:discrete} can be regarded as the assumption of the linearity of the spectrum. For comparison, this is very different from the energy spectrum of the hydrogen atom $E_n \propto - {1}/{n^2}$
but it is very similar to that of the harmonic oscillator $ E_n \propto n $. If the spectrum of $\tau$ is linear, then so is the spectrum of $\delta \tau$, which is what we assumed in the main analysis with equation \eqref{eq:discrete}. Thus, it does not really matter in this case, whether it is $\tau$ or $\delta \tau$ which is taken as the relevant quantum observable. On the contrary, for a non-linear spectrum, this question is crucial. As said earlier, the assumption of linearity is natural in the sense that it mimics the ticking of a clock, but it is not really backed so far by any theory of quantum gravity.
In Loop Quantum Gravity (LQG) the spectrum of the length, area and volume operators are famously discrete \cite{rovelli1995discreteness}. Discreteness of time may arise in a similar fashion from this theory, although nothing has been proven yet.\footnote{There is also a debate on whether discreteness in the spectrum of observables survives the implementation of the hamiltonian constraint \cite{Dittrich:2007th,rovelli2007comment}.}
The hypothesized linear behaviour is similar to the spectrum of the area operator in LQG \cite{CLQG}
\begin{equation}\label{eq:LQG_area}
A_j = 8 \pi \gamma l_P^2 \sqrt{j(j+1)}, \quad j \in \mathbb{N}/2,
\end{equation}
where $\gamma$ is a fundamental constant called the Immirzi parameter. There are indications that length has a spectrum that goes as a square root progression in $j$ \cite{Bianchi:2008es}. Geometrically, we would expect time to behave similarly to a length.
In such a case, it will make all the difference whether the square-root behaviour applies to the proper time itself
\begin{equation}\label{eq:tau-sqrt}
\tau = \sqrt{n} \, t_P,
\end{equation}
or the difference of proper time
\begin{equation}\label{eq:delta-sqrt}
\delta \tau = \sqrt{n} \, t_P.
\end{equation}
We first analyse the consequences of equation \eqref{eq:tau-sqrt} on the visibility of the plateaux. We work in Planck units and take $l\ll d$ as in the main text, although the same result can be obtained without this assumption. The proper times $\tau_\mathrm{far}$ and $\tau_\mathrm{close}$ of the branch in which $M$ and $m$ are a distance $d+l$ and $d$ apart are given in terms of laboratory time according to general relativity by
\begin{align}
\tau_\mathrm{far} = t\sqrt{1-\frac{2M}{d+l}} ~~~~~
\tau_\mathrm{close} = t\sqrt{1-\frac{2M}{d}}.
\end{align}
These are very large compared to the Planck time, as we are in the weak field regime and $t$ cannot be smaller than the period of the sharpest atomic clock.
Let's now impose the discretisation \eqref{eq:tau-sqrt}
\begin{align}
\tau_\mathrm{far} = \sqrt{n+k}, ~~~~~
\tau_\mathrm{close} = \sqrt{n}
\end{align}
where
\begin{align}
n+k = \floor{\left(1-\frac{2M}{d+l}\right)t^2}, ~~
n = \floor{\left(1-\frac{2M}{d}\right)t^2}.
\end{align}
Equation \eqref{eq:discrete} is thus replaced by
\begin{equation}
\delta\tau = \left( \sqrt {n+k} - \sqrt{n} \right) \, t_P.
\end{equation}
The condition $l\ll d$ implies that $k \ll n$, so that the equation above simplifies to
\begin{equation}
\delta\tau \approx \frac{k}{2\sqrt{n}}.
\end{equation}
So in this case, a square-root behaviour for the spectrum of $\tau$ leads to a linear behaviour for $\delta \tau$. Unfortunately, the factor of $\sqrt{n}$ in the denominator means that different values of $\delta \tau$ are exceedingly close to each other, making the experiment impossible in our proposed setup.
We now consider the case \eqref{eq:delta-sqrt}. We have
\begin{equation}
n = \floor{\left(\frac{t}{\beta}\right)^2},
\end{equation}
so that
\begin{equation}
P^{h'}_+ = \frac12+\frac12\sin \left( \frac{m}{m_P} \sqrt{\floor{ \left(\frac{t}{\beta}\right)^2 }} \right).
\end{equation}
For small values of $t / \beta$, the plot of $P^{h'}_+$ is the same as the one of $P^{h}_+$, studied in the main text. For larger values of $t/\beta$, both the width of the plateaus and the steps between them are smaller. Thus, the detection of such a discreteness is of similar difficulty so long as $t/\beta < 10$ (see figure \ref{fig:hypothesis2}).
\begin{figure}[H]
\includegraphics[width = 1\columnwidth]{hypothesis2.png}
\caption{{\bf Plot of $P_+$ as a function of $t/\beta$ with an alternative hypothesis.} We take $m=10^{-2}m_P$. Blue curve: $\delta \tau$ takes continuous values. Red curve: $\delta \tau = n \, t_P$ as considered in the main text. Green curve: $\delta \tau = \sqrt{n} \, t_P$, as motivated from LQG in this section.}
\label{fig:hypothesis2}
\end{figure}
\section{Conclusion}
In this article, we have devised an experiment that would probe a hypothetical granularity of time at the Planck scale. We have also carried out a feasibility analysis. First, we have determined a set of constraints that would ensure the visibility of the plateaux in the plot of the probability $p_+(t/\beta)$. These constraints are expressed as a set of inequalities on the experimental parameters. Second, based on current claims in the experimental physics literature, we have shown that there exists a reasonable range of parameters that satisfy the constraints. The obtained values are gathered in table \ref{tab:tableB}. Finally, we have determined the temperature and pressure conditions required to avoid too fast decoherence.
We surprisingly conclude that the proposed experiment is a feasible task for the foreseeable future. In particular, it is of comparable difficulty to contemporary experimental proposals for testing the non-classicality of the gravitational field. Nevertheless it remains difficult, and will require pooling expertise in adjacent experimental fields.
The possibility of probing planckian time without involving extremely high energies may be a disturbing idea to many physicists. However, the history of physics shows examples where scientists have gained knowledge at a physical scale that was widely believed to be unreachable with the available technology at the time. The first example is when Einstein proposes a way to measure the size of atoms by observing the brownian motion of mesoscopic pollen grains \cite{einstein1905uber}. Another example is when Millikan shows that the electric charge comes in discrete packets, and measures the charge of the smallest packet (the electron) \cite{millikan1909modification, millikan1913elementary}. Again, such a feat was realised through the observation of the mesoscopic motion of charged drops of oil. In both cases, as in our proposal, the scale of discreteness was reached through mesoscopic observables thanks to two leverage effects: an algebraic game involving very small or very big constants and a statistical game involving the collection of many events.
The importance of realising the proposed experiment lies primarily in the groundbreaking implications of potentially discovering a granularity of time at the Planck scale. A negative result would also have significant implications, guiding fundamental theory. Finally, an easier version of the experiment with relaxed constraints would remain of profound interest, setting new bounds on the continuous behaviour of time.
\begin{acknowledgments}
The authors thank \v{C}aslav Brukner, Giulio Chiribella, Giulia Rubino, Eugenio Bianchi, Philipp H\"{o}hn, Alejandro Perez, Sougato Bose, Vlatko Vedral, Chiara Marletto and Tristan Farrow for insights and constructive criticisms during the course of this work. We also thank Martin Plenio and Anupam Mazumdar for comments on the first version of the article. Last but not least Carlo Rovelli for reading and commenting early drafts of this work.
This publication was made possible through the support of the ID\# 61466 grant from the John Templeton Foundation, as part of the "Quantum Information Structure of Spacetime (QISS)" project (\hyperlink{http://www.qiss.fr}{qiss.fr}). The opinions expressed in this publication are those of the authors and do not necessarily reflect the views of the John Templeton Foundation.
\end{acknowledgments}
\bibliographystyle{scipost}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 2,769
|
\section{Introduction}
The viscoelastic properties of dilute polymer solutions are central to
several physical and engineering applications~\citep{l99}.
When the pure solvent is turbulent the most remarkable effect of
the addition of polymers is a significant
reduction of the turbulent drag below that of
the solvent \citep*{plb08,wm08,b10,g14}.
This phenomenon, also
known as Toms effect \citep{Toms1,Toms2},
is commonly utilized to reduce the energy losses in pipelines
and hence the costs associated with the transport of crude oil.
However, in a turbulent flow,
polymers are subject to mechanical degradation due to
the fluctuating strain-rate, which stretches polymers and thus causes
their scission.
Since turbulent drag reduction decreases with the molecular weight of the
dissolved polymers \citep{v75},
the efficacy of polymers as drag reducing agents
diminishes in time, with
a strong impact on both industrial applications and laboratory experiments
\citep*{PA70,dTDKN95,cllc02,vis05,vcs06,ewsc09,ps12,odp17}.
An analogous mechanical
degradation of polymers and, by association, practical limitations
are observed in experiments of
homogeneous and isotropic turbulence with polymer additives \citep{cmxb08}
as well as in the regime of elastic turbulence \citep{gs04}. In the
latter case, even though the Reynolds number of the solution is low, elastic
instabilities generate a chaotic flow with highly fluctuating
velocity gradients \citep*{bss07}.
A detailed knowledge of the statistics of
polymer scission in turbulent flows
is therefore important for the design of experiments
and the performance of realistic simulations of both
turbulent drag reduction
and elastic turbulence.
However, unlike the fragmentation of liquid sheets and drops \citep{villermaux},
the modelling of flow-driven scission of polymers has received scant attention and, with very few exceptions \citep*[e.g.][]{pms18},
polymer scission is disregarded completely by the available constitutive models of polymer solutions.
We therefore investigate the dynamics of polymers in
a turbulent flow focusing on the scission statistics.
A polymer is described as a bead-spring chain in a time-dependent,
linear velocity field.
This polymer model is known as the \citet{r53} model
and represents one of the most common descriptions
of a polymer molecule in a flow
(in the case in which
only two beads are considered, the Rouse model
reduces to the elastic dumbbell model; see \citealt{bcah77}).
Even when the flow is turbulent,
the assumption of a linear velocity field is justified,
since the size of polymers is generally smaller than the
Kolmogorov dissipation scale $\eta_K$, below which viscosity strongly damps
the spatial fluctuations of the velocity.
We introduce scission into the Rouse model by assuming that
the bead-spring chain breaks into two shorter chains
as soon as the tension in one of the springs
exceeds a critical threshold.
After introducing a modification of the Rouse model that takes polymer scission into account,
we begin by considering the statistics of the first scission.
For {\it passively} transported polymers,
for which the motion of the polymers does not modify the carrier velocity field,
we derive qualitative analytical predictions
by restricting ourselves to the Hookean dumbbell model and by
using a decorrelated-in-time
Gaussian stochastic velocity field (the approach is adapted from
a study of droplet breakup conducted in \citealt{RV18}).
The theoretical predictions are compared with Lagrangian direct numerical
simulations (DNSs) of the Rouse model in three-dimensional homogeneous isotropic turbulence.
For {\em active} polymers, the statistics of the first scission is studied via
hybrid Eulerian--Lagrangian simulations \citep{wg13jfm,wg13b,wg14},
in which the feedback of (dumbbell-like) polymers on the velocity field is taken into account.
These simulations shed light on the transient nature of the dissipation-reduction effect, which owes its origin to polymer stretching and its demise to polymer scission.
Finally, we analyse multiple scissions of passive polymers, via DNSs in which a hierarchy of daughter polymers arise from successive breakups, each of
which with their own statistics.
\section{The Rouse chain}
The Rouse model describes a polymer as a chain of $\mathscr{N}$ inertialess beads connected
to their nearest neighbors by elastic springs.
We consider FENE (finitely extensible nonlinear elastic)
springs with spring constant $H$ and maximum length $Q_m$.
The fluid in which the chain is immersed is Newtonian and its motion is described
by an incompressible velocity field $\bm u(\bm x,t)$.
The drag force of the fluid on each bead is given by the Stokes law
with drag coefficient $\zeta$; the collisions of the molecules of the fluid
with a bead are described by Brownian motion.
Finally, hydrodynamic and excluded-volume interactions between different segments of
the chain are disregarded.
The motion of the chain is described
in terms of the position of its center of mass, $\bm X_c$, and the separation
vectors between the beads, $\bm Q_i$ ($i=1,\dots,\mathscr{N}-1$).
This set of coordinates evolves according to the equations \citep{bcah77,o96}%
\begin{subequations}%
\begin{eqnarray}%
\label{eq:cm}
\dot{\bm X}_c &=& \bm u(\bm X_c(t),t)+\dfrac{1}{\mathscr{N}}\sqrt{\frac{Q_{\rm eq}^2}{6\tau}}
\sum_{i=1}^{\mathscr{N}}\bm\xi_i(t),
\\[2mm]
\label{eq:q}
\dot{\bm Q_i} &=&
\bm\kappa(t)\bcdot\bm Q_i(t)-\dfrac{1}{4\tau}
[2f_i\bm Q_i(t)-f_{i+1}\bm Q_{i+1}(t)-f_{i-1}\bm Q_{i-1}(t)]
\\
\nonumber
&&+\sqrt{\dfrac{Q_{\rm eq}^2}{6\tau}}
[\bm\xi_{i+1}(t)-\bm\xi_i(t)],\qquad i=1,\dots,\mathscr{N}-1,%
\end{eqnarray}%
\label{eq:chain}%
\end{subequations}%
where $\kappa^{\alpha\beta}(t)=\nabla^\beta u^\alpha(\bm X_c(t),t)$
is the velocity gradient evaluated at the position of the center of mass,
$\tau=\zeta/4H$
is the characteristic time scale of the springs, $Q_{\rm eq}=\sqrt{3k_BT/H}$ is their
equilibrium root-mean-square (r.m.s.)
extension ($k_B$ denotes the Boltzmann constant
and $T$ is temperature),
and $\bm\xi_i(t)$
($i=1,\dots,\mathscr{N}$) are independent, vectorial, white noises.
The coefficients
\begin{equation}
f_i=\dfrac{1}{1-\vert\bm Q_i\vert^2/Q_m^2}
\end{equation}
characterize the FENE interactions and
ensure that the extension of each spring
does not exceed its maximum length $Q_m$.
Obviously,
in the equations for $\bm Q_1$ and $\bm Q_{\mathscr{N}-1}$ it is assumed that
$\bm Q_0=\bm Q_{\mathscr{N}}=0$.
The end-to-end separation vector of the polymer is
defined as $\bm R=\sum_{i=1}^{\mathscr{N}-1}\bm Q_i$.
In a still fluid, the equilibrim r.m.s. value of $\vert\bm R\vert$ is
$r_{\rm eq}=Q_{\rm eq}\sqrt{\mathscr{N}-1}$ \citep{bcah77}.
We modify the Rouse model in order to account for the scission of the polymer
when the tension in any of the springs exceeds a critical value.
Since the relation between the tension and the extension of a spring
can be easily inverted~\citep{thiffeault}, we can, equivalently, assume
that for each spring of the chain there exists a critical scission length
$\ell_{\mathrm{sc}}$ such that the spring breaks
if the length of the corresponding separation vector
exceeds $\ell_{\mathrm{sc}}$ (i.e. the chain breaks if
$\vert\bm Q_i\vert\geqslant\ell_{\mathrm{sc}}$ for any $1\leqslant i\leqslant \mathscr{N}-1$).
The scission process is non-stationary; the dynamics of the chain
therefore depends on its initial configuration and, in particular, on
its initial end-to-end separation $r_0=\vert\bm R(0)\vert$.
In the following we shall assume
that $\ell_{\mathrm{sc}}$ is much greater than $r_0/(\mathscr{N}-1)$ and that
$r_0$ is equal to $r_{\rm eq}$
or greater than it.
Finally, the size of the chain always remains
smaller than $\eta_{\rm K}$, so that the velocity field at the scale of the
chain can be considered as linear and the dynamics of the polymer is
entirely determined by the velocity gradient at the location of the
center of mass, consistent with the Rouse model.
To summarize, the spatial scales that characterize
the system are arranged as follows:
$r_{\rm eq}\leqslant r_0\ll\ell_{\mathrm{sc}}(\mathscr{N}-1)<Q_m(\mathscr{N}-1)<\eta_{\rm K}$.
\section{First-scission statistics}
\subsection{Passive polymers}
\label{sect:first-scission}
\subsubsection{Analytical predictions}
\label{sect:analytical}
Here we make some simplifying assumptions on both the polymer model and the
carrier flow in
order to derive analytical predictions for the statistics of polymer scission.
First of all, we only consider the statistics of the first scission.
We then restrict ourselves
to the $\mathscr{N}=2$ case, also known as the dumbbell
model \citep{bcah77,o96},
i.e., we focus on the slowest deformation mode of the polymer.
Many results on single-polymer dynamics in random or turbulent flows
have been obtained by using the
dumbbell model (see \citealt{VPBT15} and references therein)
and the most common constitutive models of polymer
solutions, namely the Oldroyd-B \citep{o50} and the FENE-P \citep{bdj80} models, are based on it.
The legitimacy of this approach is supported by the numerical simulations
in \cite{jc08} and~\cite{wg10}, where it is shown that, in isotropic turbulence
and in the absence of scission,
the statistics of the end-to-end separation of a dumbbell and that of a $\mathscr{N}=20$ chain
coincide (provided, of course,
a proper mapping between the parameters of the two systems is applied).
Finally, we replace the nonlinear spring with a Hookean one ($f_i=1$);
this is because the nonlinearity
of the elastic force enters into play only at extensions
close to the scission length,
and we shall see from our simulations in \S~\ref{sect:sim}
that it does not affect
the qualitative properties of the scission process.
For $\mathscr{N}=2$, \eqref{eq:chain} reduce to
\begin{subequations}
\begin{eqnarray}
\label{eq:cm-dumbbell}
\dot{\bm X}_c
&=& \bm u(\bm X_c(t),t)+\dfrac{1}{2}\sqrt{\frac{r_{\rm eq}^2}{3\tau}}
\,\bm \zeta_1(t),
\\
\label{eq:q-dumbbell}
\dot{\bm R} &=& \bm\kappa(t)\bcdot\bm R(t)-\dfrac{\bm R(t)}{2\tau}
+\sqrt{\dfrac{r_{\rm eq}^2}{3\tau}}\,
\bm\zeta_2(t),
\end{eqnarray}
\label{eq:dumbbell}
\end{subequations}
where $\bm\zeta_1(t)$ and $\bm\zeta_2(t)$ are (non-independent)
vectorial white noises.
We model the flow via the smooth (also known as Batchelor) regime
of the \cite{k68} model.
This model
has been widely employed in the study of turbulent transport \citep*{fgv01}
and has yielded several theoretical results
on the coil-stretch transition in random or turbulent flows
(see \citealt*{pav16} and references therein).
The velocity is a divergence-less and spatially smooth
Gaussian vector field.
It is statistically stationary in time and
homogeneous, isotropic, and parity invariant in space;
it has zero mean and zero correlation time.
Under these assumptions,
$\bm \kappa(t)$ is a tensorial white noise with two-time correlation
\citep{fgv01}
\begin{equation}
\langle \kappa^{ij}(t)\kappa^{mn}(t')\rangle=
\mathscr{K}^{ijmn}\delta(t-t'),
\end{equation}
where
\begin{equation}
\mathscr{K}^{ijmn}=\frac{\lambda}{3}
[4\delta^{im}\delta^{jn}-\delta^{ij}
\delta^{mn}-\delta^{in}\delta^{jm}]
\end{equation}
and $\lambda$ is the maximum Lyapunov exponent of the flow.
Obviously, the assumption of temporal decorrelation is a strong
approximation, since an isotropic turbulent flow has Kubo number
$\mathit{Ku}=\lambda t_{\mathrm{corr}}\approx 0.6$,
where $t_{\mathrm{corr}}$ is the correlation time of the flow \citep{gp90,bec06,wg10}.
However, it was shown in \cite{mv11} that, for a
fluctuating flow with comparable $\mathit{Ku}$,
the statistics of polymer extension
is correctly captured by a time decorrelated velocity field.
The Weissenberg number $\mathit{Wi}=\lambda\tau$ determines
to what extent polymers are stretched by the flow. In particular,
the coil-stretch transition occurs when $\mathit{Wi}$ exceeds the critical value $\mathit{Wi}_{\rm cr}=1/2$
(\citealt*{lumley,bfl00}; note that our definition of $\tau$
and hence that of $\mathit{Wi}$ differ from that of \citealt{bfl00} by a factor of~2).
As the velocity field is homogeneous and isotropic in space, the statistics of
$R=\vert\bm R\vert$
is independent of the position of the center of mass and of the
direction of $\bm R$.
To study polymer scission,
it is therefore sufficient to focus on
the probability density function (p.d.f.) of $R$, which will be denoted as $P(R,t)$.
When $\bm\kappa(t)$ has the properties described above,
$P(R,t)$ satisfies the Fokker--Planck equation \citep*{c00,cmv05}:
\begin{equation}
\label{eq:FPE}
\partial_{t'} P=\mathbb{L} P,
\qquad \mathbb{L} P=- \partial_R (D_1 P)+\partial_R^2(D_2 P),
\end{equation}
with rescaled time $t'=t/2\tau$ and coefficients
\begin{equation}
\label{eq:coeff}
D_1 = \Big(\frac{8}{3}\mathit{Wi}-1\Big)R+\dfrac{2r_{\rm eq}^2}{3R}, \qquad
D_2 = \frac{2\mathit{Wi}}{3}R^2+\frac{r_{\rm eq}^2}{3}
\end{equation}
(once again our definition of $\mathit{Wi}$
differs from that used in \citealt{c00} and \citealt{cmv05}
by a factor of 2).
The appropriate boundary conditions are
reflecting at $R=0$ and absorbing at $R=\ell_{\mathrm{sc}}$, i.e.
\begin{equation}
\label{eq:bc}
D_1 P-\partial_R(D_2 P)=0 \ \text{at $R=0$}
\qquad \text{and} \qquad
P(\ell_{\mathrm{sc}},t)=0
\end{equation}
for all $t$.
The former condition ensures that the extension of the polymer
stays positive, while the latter describes scission at $R=\ell_{\mathrm{sc}}$.
The analysis of \eqref{eq:FPE} to~\eqref{eq:bc}
closely follows that in \citet{RV18}
for the breakup of sub-Kolmogorov droplets in isotropic turbulence.
(The results presented here are deduced directly from those in \S~3 of
\citealt{RV18} by setting $\mathit{Ca}=\mathit{Wi}$, $\mu=1$, $r_{\mathrm{eq}}^2=
Q_{\rm eq}^2/3$, and
$f_1(\mu)=f_2(\mu)=\gamma(\mu)=1$.)
We therefore skip the details of the derivations and
directly present the predictions of scission statistics.
The number of unbroken polymers that survive at time $t$,
$N_p(t)$,
is related to $P(R,t)$ as follows:
\begin{equation}
N_p(t)/N_p(0)=\int_0^{\ell_{\mathrm{sc}}} P(R,t)\mathrm{d}R.
\label{eq:N}
\end{equation}
At times $t\gg \tau$,
$N_p(t)$ therefore decays exponentially as
\begin{equation}
N_p(t)\sim N_p(0)\,e^{-t/T_d},
\label{eq:N-decay}
\end{equation}
where the decay time $T_d$ is the reciprocal of the
lowest eigenvalue of the operator $\mathbb{L}$ with boundary conditions
\eqref{eq:bc}.
The eigenfunctions of $\mathbb{L}$ are hypergeometric functions with
parameters depending on $r_{\rm eq}$, $r_0$, $\mathit{Wi}$ and form a discrete set
selected by the boundary condition at $R=\ell_{\mathrm{sc}}$.
A calculation of the lowest eigenvalue shows that
$T_d$ depends weakly on $\mathit{Wi}$ for small $\mathit{Wi}$,
decreases rapidly as $\mathit{Wi}$ exceeds $\mathit{Wi}_{\rm cr}$, and saturates at large $\mathit{Wi}$.
As we shall see below, the p.d.f. of $R$ integrated over time,
\begin{equation}
\label{eq:Phat}
\widehat{P}(R)=\int_0^\infty P(R,t)\mathrm{d}t,
\end{equation}
allows us to estimate the mean lifetime of a polymer before its first
scission.
If the initial distribution of polymer sizes is `monodisperse',
i.e. $P(R,0)=\delta(R-r_0)$, then $\widehat{P}(R)$ takes the form
\begin{equation}
\widehat{P}(R)\propto
\begin{cases}
e^{-\Phi(R)}[\phi(\ell_{\mathrm{sc}})-\phi(r_0)] & \text{if $0\leqslant R\leqslant r_0$},
\\
e^{-\Phi(R)}[\phi(\ell_{\mathrm{sc}})-\phi(R)] & \text{if $r_0< R\leqslant\ell_{\mathrm{sc}}$}
\end{cases}
\end{equation}
with
\begin{equation}
\Phi(R)=\ln D_2(R)-\int^R \dfrac{D_1(\zeta)}{D_2(\zeta)}\,\mathrm{d}\zeta,
\qquad
\phi(R)=\int^R \dfrac{e^\Phi(\zeta)}{D_2(\zeta)}\,\mathrm{d}\zeta
\end{equation}
and hence
\begin{equation}
\label{eq:pdf-1}
\widehat{P}(R)\sim
\begin{cases}
r_{\rm eq}^{-2}r_0^ {-1}R^2 & \text{if $0\leqslant R\ll r_{\rm eq}$,}
\\[1mm]
\vert r_0^\alpha-\ell_{\mathrm{sc}}^\alpha\vert
R^{-1-\alpha} & \text{if $r_{\rm eq}\ll R\ll r_0$,}
\\[1mm]
\ell_{\mathrm{sc}}^\beta R^{-1-\beta} & \text{if $r_0\ll R\ll\ell_{\mathrm{sc}}$},
\end{cases}
\end{equation}
where $\alpha=3(\mathit{Wi}^{-1}-2)/2$ and
\begin{equation}
\beta = \begin{cases}
\alpha & \text{if $\mathit{Wi}<\mathit{Wi}_{\rm cr}$,}
\\
0 & \text{if $\mathit{Wi}>\mathit{Wi}_{\rm cr}$}.
\end{cases}
\end{equation}
Therefore, above the coil-stretch transition,
the right tail of $\widehat{P}(R)$ saturates to the power law $R^{-1}$.
An analogous behaviour was found previously for the size distribution of
sub-Kolmogorov droplets in isotropic turbulence \citep*{bmv14,RV18}.
Also note that the exponent
$\alpha$ coincides with the one obtained by \cite{bfl00}
for the p.d.f. of intermediate extensions
in the absence of scission.
If the initial distribution of polymer sizes is broad but nonetheless
admits a maximum size $r_0$, then the left ($R\llr_{\rm eq}$)
and right ($r_0\ll R\ll\ell_{\mathrm{sc}}$) power-law tails continue to exist, but
$\widehat{P}(R)$ no longer behaves as a power law
for extensions $r_{\rm eq}\ll R\ll r_0$.
The mean time $\langleT_{\mathrm{sc}}\rangle$ it takes for a polymer
to undergo its first scission can be deduced from the behaviour of
$\widehat{P}(R)$ via the relation:
\begin{equation}
\langleT_{\mathrm{sc}}\rangle=\int_0^{\ell_{\mathrm{sc}}}\widehat{P}(R)\mathrm{d}R.
\end{equation}
Equation \eqref{eq:pdf-1} then yields
two different behaviours below and above the coil-stretch transition:
\begin{equation}
\label{eq:tsc}
\lambda\langleT_{\mathrm{sc}}\rangle\sim
\begin{cases}
(\ell_{\mathrm{sc}}/r_0)^\beta & \text{if $\mathit{Wi}<\mathit{Wi}_{\rm cr}$},
\\
\ln (\ell_{\mathrm{sc}}/r_0) & \text{if $\mathit{Wi}>\mathit{Wi}_{\rm cr}$}.
\end{cases}
\end{equation}
\subsubsection{Direct numerical simulations}
\label{sect:sim}
In this Section,
we present numerical simulations of the Rouse model \eqref{eq:chain}
in homogeneous and isotropic turbulence and compare them with
the analytical predictions of \S~\ref{sect:analytical}.
The velocity field $\bm u(\bm x,t)$
is the solution of the incompressible Navier--Stokes equations,
\begin{equation}
\partial_t \bm u+\bm u\bcdot\bnabla\bm u=-\bnabla p+\nu\Delta\bm u+\bm F,
\quad \bnabla\bcdot\bm u=0,
\label{eq:NS}
\end{equation}
over the periodic cube $[0,2\upi]^3$. Here $p$ is pressure, $\nu$ is the
kinematic viscosity, and $\bm F(\bm x,t)$ a body force that mantains a constant
kinetic-energy input $\epsilon$.
The numerical integration uses a standard, fully de-aliased pseudo-spectral
method with $512^3$ collocation points and, for the time evolution,
a second-order slaved Adams--Bashforth scheme
with time step ${\rm d}t=4\times 10^{-4}$.
The values of $\nu$ and $\epsilon$ are such
that the Taylor-microscale Reynolds number is $\mathit{Re}_\lambda=111$.
The Lyapunov exponent of the flow is $\lambda=0.15\tau_\eta^{-1}$, where
$\tau_\eta$ denotes the Kolmogorov time scale, consistent with the
value found earlier in \citet{bec06} and \citet{wg10}.
The position of the centre of mass of the polymer is obtained by
integrating~\eqref{eq:cm} via a second-order Adam--Bashforth method with same
${\rm d}t$ as for the Navier--Stokes equations.
The noise term in~\eqref{eq:cm}
is disregarded, because it has a negligible effect when $\bm u(\bm x,t)$
is turbulent. Moreover, its amplitude
is smaller than that of the noise terms in the
equations for the separation vectors by a factor of $\mathscr{N}$.
As $\bm u(\bm x,t)$ is only known
over a discrete grid, the integration of~\eqref{eq:cm} requires
interpolation to reconstruct the velocity field at
$\bm X_c(t)$; a trilinear scheme is used for this purpose.
The same approach allows the calculation of the velocity gradient
along the trajectory of the center of mass, $\bm \kappa(t)$, and hence
the integration of~\eqref{eq:q}
by means of the Euler--Maruyama method with time step ${\rm d}t$.
Since we focus on the statistics of the first scission, a chain is removed
from the simulation as soon as it breaks according to the criterion
discussed above.
The time origin for \eqref{eq:q} ($t=0$) is taken in the statistically steady state
of the carrier turbulent flow, so that the temporal dynamics of polymers
is not influenced by the initial transient evolution of $\bm u(\bm x,t)$.
Note that, in the present context, it is not necessary to use
integration schemes specifically designed to
prevent the extension of the links from exceeding $Q_m$,
since the springs, by construction, break well before their extension approaches $Q_m$.
\begin{figure}
\includegraphics[width=0.49\textwidth]{exponential-decay}\hfill%
\includegraphics[width=0.49\textwidth]{t_d}%
\caption{Passive polymers:
(\textit{a}) exponential decay of the fraction of unbroken polymers for different values of $\mathit{Wi}$;
(\textit{b}) decay time of the fraction of unbroken polymers rescaled by the
Kolmogorov time $\tau_\eta$ as a function
of $\mathit{Wi}/\mathit{Wi}_{\rm cr}$.}
\label{fig:decay}
\end{figure}
In our simulations, we consider
$N_p(0)=9\times 10^5$ polymers, whose positions at time $t=0$ are
uniformly distributed in space.
Since the statistics of polymer extension depends on the initial
size of polymers but not on their orientation, for simplicity the
initial condition for the separation vectors is taken to be
$\bm Q_i(0)=Q_0 (1,1,1)/\sqrt{3}$ with $Q_0>0$ for all polymers, i.e.
the polymers
are in a straight configuration and
$P(R,0)=\delta(R-r_0)$ with $r_0=(\mathscr{N}-1)Q_0$.
In order to compare chains with different numbers of beads,
an appropriate mapping of the chain parameters is needed.
Here we use the mapping proposed by \cite{jc08} and also used
by \cite{wg10}.
If the parameters of the individual links of a $\mathscr{N}$-bead chain are
$\tau$, $Q_{\rm eq}$, $Q_m$, $\ell_{\mathrm{sc}}$, then the statistics of the end-to-end
separation of the chain is equivalent to that of a dumbbell with
following parameters:
\begin{equation}
\label{eq:mapping}
\tau^{D}=\frac{\mathscr{N}(\mathscr{N}+1)\tau}{6}, \quad
Q_{\rm eq}^D=Q_{\rm eq}, \quad
Q_m^{D}=Q_m{\sqrt{\mathscr{N}-1}}, \quad
\ell_{\mathrm{sc}}^{D}=\ell_{\mathrm{sc}}{\sqrt{\mathscr{N}-1}},
\end{equation}
where the last relation is introduced for
compatibility with the expression of $Q_m^D$.
This mapping allows us to compare chains with different numbers of beads
by using the dumbbell model as a reference.
Following \cite{wg10}, we
define the Weissenberg number for a Rouse chain
as $\mathit{Wi}=\lambda\tau^{D}$.
In our simulations, $0.4\leqslant \mathit{Wi}
\leqslant 8$.
(Note that small-$\mathit{Wi}$ simulations are computationally more demanding, because
the calculation of quantities like $\widehat{P}(R)$ and
$\langleT_{\mathrm{sc}}\rangle$ requires that the time evolution is
long enough for all polymers to break, and the time at which the scission
process is complete becomes longer
and longer as $\mathit{Wi}$ decreases.)
\begin{figure*}
\setlength{\unitlength}{\textwidth}
\includegraphics[width=0.49\textwidth]{pdf-integrated}\hfill%
\includegraphics[width=0.49\textwidth]{three-powers}%
\caption{Passive polymers:
(\textit{a}) time-integrated p.d.f. of the end-to-end separation of unbroken polymers
for $r_0=r_{\rm eq}$ and for different values of $\mathit{Wi}$. The inset shows the
value of $\beta$, which determines the exponent of the right tail
of the p.d.f. (i.e. $\widehat{P}(R)\propto R^{-1-\beta}$
for $r_0\ll R\ll\ell_{\mathrm{sc}}$),
as a function of $\mathit{Wi}/\mathit{Wi}_{\rm cr}$;
(\textit{b}) time-integrated p.d.f. of the end-to-end separation of unbroken polymers
for $r_0=40r_{\rm eq}$ and for different values of $\mathit{Wi}$.
The inset shows the
value of $\alpha$,
which determines the power-law behaviour
of the p.d.f. for intermediate extensions (i.e.
$\widehat{P}(R)\propto R^{-1-\alpha}$ for
$r_{\rm eq}\ll R\ll r_0$),
as a function of $\mathit{Wi}/\mathit{Wi}_{\rm cr}$.
In both panels (\textit{a}) and (\textit{b}), the p.d.f.s are
normalized to unity for the sake of comparison.
}
\label{fig:pdfs}
\end{figure*}
As for the choice of the other parameters,
we take $r_{\rm eq}=1$, $Q_m^{D}=\sqrt{3000}$
(also following \citealt{jc08} and \citealt{wg10}), and, unless otherwise
specified, $r_0=r_{\rm eq}$. In addition,
it is assumed that a spring breaks as soon as
its extension exceeds $\ell_{\mathrm{sc}}=0.8Q_m$.
The number of beads is set to $\mathscr{N}=10$.
We have also performed simulations with different sets of parameters,
which support the generality of the results presented below.
Figure~\ref{fig:decay} shows the temporal evolution of
the fraction of unbroken polymers.
The decay is exponential
with a time scale $T_d$ that decreases rapidly as $\mathit{Wi}$
exceeds its critical value, in agreement with the predictions of
\S~\ref{sect:analytical}. We shall see in \S~\ref{sect:active}
that, for a dumbbell, it is possible to write an explicit expression
for $T_d$ as a function of $\mathit{Wi}$.
The time-integrated p.d.f. of the end-to-end separation
{\it of unbroken polymers}
is shown in figure~\ref{fig:pdfs}\textit{a}
for an initial polymer size $r_0=r_{\rm eq}$ and different values of $\mathit{Wi}$.
The p.d.f. displays a power-law behaviour
for both $R\ll r_{\rm eq}$ and $r_0=r_{\rm eq}\ll R\ll \ell_{\mathrm{sc}}(\mathscr{N}-1)$. The left tail is
proportional to $R^2$, because
the small separations are dominated by thermal fluctuations.
The right tail rises as a function
of $\mathit{Wi}$, until the power law saturates to $R^{-1}$ for $\mathit{Wi}>\mathit{Wi}_{\rm cr}$.
A third power law emerges for intermediate extensions
if $r_0>r_{\rm eq}$ (figure~\ref{fig:pdfs}\textit{b}).
In this case, the exponent $-1-\alpha$ changes from negative to positive
as $\mathit{Wi}$ increases and saturates to $2$ at large $\mathit{Wi}$.
To appreciate the coexistence of these three power laws more clearly,
in figure~\ref{fig:ld_1e4}\textit{a} we also consider $\widehat{P}(R)$ for
a much larger value of $Q_m^D$ and
a larger separation between $r_{\rm eq}$, $r_0$, and $(\mathscr{N}-1)\ell_{\mathrm{sc}}$.
All these results confirm the predictions reported in \S~\ref{sect:analytical}.
The p.d.f.s presented so far correspond to a ``monodisperse'' initial
state $P(R)=\delta(R-r_0)$ in which all polymers have the same
end-to-end distance. However,
as mentioned in \S~\ref{sect:analytical}, the behaviour of
$\widehat{P}(R)$ for intermediate extensions
is expected to change if the initial distribution of polymer
extensions is broad. To confirm this prediction, we have considered an initial
state in which the end-to-end distance of polymers is
distributed uniformly between $r_{\rm eq}$ and a maximum initial extension
$r_0>r_{\rm eq}$. The time-integrated
p.d.f.s given in figure~\ref{fig:ld_1e4}\textit{b} show that
only the left and right power-law tails persist in this case,
while $\widehat{P}(R)$
does not behave as a power law for intermediate extensions.
\begin{figure*}
\centering
\includegraphics[width=0.483\textwidth]{pdf-ld_1e4}\hfill
\includegraphics[width=0.5\textwidth]{pdf-ld_1e4_r500}%
\caption{Passive polymers:
(\textit{a}) time-integrated p.d.f. of the end-to-end separation of unbroken polymers
for $Q^D_m=10^4$, $r_{\rm eq}=1$, $r_0=10^2$, and different values of $\mathit{Wi}$;
(\textit{b}) time-integrated p.d.f. of the end-to-end separation of unbroken
polymers
for $Q^{D}_m=10^4$, $r_{\rm eq}=1$, $r_0=5\times 10^2$, $\mathit{Wi}=1$, and
a uniform (solid gray line) or a monodisperse (dashed magenta line)
initial distribution of polymer sizes. In both panels,
the p.d.f.s are
normalized to unity for the sake of comparison.}
\label{fig:ld_1e4}
\end{figure*}
\begin{figure}
\includegraphics[width=0.48\textwidth]{pdf-lifetime}\hfill%
\includegraphics[width=0.48\textwidth]{mean-t_sc}%
\caption{Passive polymers:
(\textit{a}) p.d.f. of the lifetime of a polymer for different values of
$\mathit{Wi}$. The inset compares the decay time of the fraction of unbroken polymers,
the mean lifetime of a polymer, and the time scale $\gamma^{-1}$ in the exponential
tail of the p.d.f. ($P(T_{\rm sc})\sim e^{-\gamma T_{\rm sc}}$ for $\gamma T_{\rm sc}\gg 1$) as a function of $\mathit{Wi}/\mathit{Wi}_{\rm cr}$;
(\textit{b}) mean lifetime as a function of $\ell_{\mathrm{sc}}/Q_0$ below the coil-stretch transition.
$T_{\rm sc}'$ is a fitting parameter for the dashed curves.
The data for $\mathit{Wi}=0.4$
(resp. $\mathit{Wi}=0.45$) are multiplied by a factor of $10^2$ (resp.~10) in order to make the three curves more easily
distinguishable;
(\textit{c}) the same as in (\textit{b}) above the coil-stretch transition.}
\label{fig:lifetime}
\end{figure}
We now turn to the statistics of the lifetime $T_{\mathrm{sc}}$ of a polymer.
The DNSs suggest that the
p.d.f. of $T_{\mathrm{sc}}$ has an exponential tail
with a time scale $\gamma^{-1}$ that, beyond $\mathit{Wi}_{\rm cr}$,
decreases rapidly as a function of $\mathit{Wi}$ (figure~\ref{fig:lifetime}\textit{a}).
For all values of the Weissenberg number,
$\gamma^{-1}$ is approximately the same as the
decay time of the fraction of unbroken polymers owing to
the exponential decay of the latter at long times (see \S~\ref{sect:analytical}).
However, $\gamma^{-1}$ differs
from $\langle T_{\rm sc}\rangle$, because the exponential behaviour of the
p.d.f. of $T_{\mathrm{sc}}$ sets in only at relatively large values of $T_{\mathrm{sc}}$.
For a fixed $\mathit{Wi}$, the mean lifetime $\langle T_{\rm sc}\rangle$ behaves as a power
of $\ell_{\mathrm{sc}}/Q_0$ below the coil-stretch transition and as the
logarithm of $\ell_{\mathrm{sc}}/Q_0$ beyond that (see
figures~\ref{fig:lifetime}\textit{b} and~\ref{fig:lifetime}\textit{c});
we remind the reader that $Q_0$ is the initial length of any link of the
chain.
Small deviations
are only observed for $\ell_{\mathrm{sc}}\gg Q_0$.
Moreover, we have checked that, for $\mathit{Wi}<\mathit{Wi}_{\rm cr}$, the exponent $\beta$
that gives the dependence of $\langleT_{\mathrm{sc}}\rangle$ on $\ell_{\mathrm{sc}}/Q_0$
is the same as the exponent that describes the right tail of $\widehat{P}(R)$,
i.e. $\widehat{P}(R)\propto R^{-1-\beta}$ for $r_0\ll R\ll(\mathscr{N}-1)
\ell_{\mathrm{sc}}$,
in agreement with \eqref{eq:tsc}.
Thus, the statistics of $T_{\mathrm{sc}}$ in a turbulent flow is correctly described
by the predictions of \S~\ref{sect:analytical}.
We also note that the exponential tail of the distribution of $T_{\rm
sc}$ originates from the fact that scission is caused by the cumulative action
of the fluctuating strain-rate. This is in contrast to the
fragmentation of sub-Kolmogorov inextensible fibres,
for which the internal tension depends on
the {\it instantaneous} velocity gradient projected along the fibre.
The p.d.f. of the scission time for fibres, therefore, reflects
the intermittent statistics of the velocity gradient and is strongly non-exponential
\citep*{sofia-2}.
Figure~\ref{fig:histograms} presents further results on
the statistics of the scission process.
As previously observed in experiments~\citep{hm84},
scission preferentially happens at
the midpoint of the polymer. However, the probability of scission happening
at the middle link decreases with $\mathit{Wi}$. The reason for this is that, for small $\mathit{Wi}$,
the chain is most of the time in a coiled state and scission occurs because of a sequence of
very strong fluctuations of $\bm\nabla\bm u$, whereas for large $\mathit{Wi}$ all links
are consistently stretched near to the scission length.
The insets of figure~\ref{fig:histograms} show that
the probability of more than one link breaking simultaneously
is generally very small.
\begin{figure}
\includegraphics[width=0.33\textwidth]{histogram_Wi_12}\hfill%
\includegraphics[width=0.33\textwidth]{histogram_Wi_4}\hfill%
\includegraphics[width=0.33\textwidth]{histogram_Wi_3}
\caption{Passive polymers:
probability of polymer scission occuring at the $j_{\rm sc}$-th link for different values of $\mathit{Wi}$.
The insets shows the probability of $n_L$ links breaking simultaneously.}
\label{fig:histograms}
\end{figure}
Finally, it was mentioned in \S~\ref{sect:analytical} that, in the absence of scission,
the dumbbell model ($\mathscr{N}=2$) captures the statistics of the
end-to-end separation of a full chain remarkably well,
provided the mapping in~\eqref{eq:mapping} is applied \citep[see][]{wg10}.
We have performed an analogous comparison in the presence of scission.
The time-integrated p.d.f.s
in figure \ref{fig:comparison}\textit{a} show that, after the parameters of the chain
are suitably rescaled, the time-independent
statistics of the end-to-end distance is independent of
$\mathscr{N}$, except for small deviations close to the maximum extension.
Indeed, for a dumbbell, scission is defined in terms of the end-to-end
separation, whereas chains
with larger $\mathscr{N}$ break before all the links can
stretch up to $\ell_{\mathrm{sc}}$.
These small differences have however a significant impact
on time-dependent quantities, such as
the fraction of unbroken polymers:
small-$\mathscr{N}$ chains capture the temporal decay qualitatively,
but underestimate the scission rate
(see figure~\ref{fig:comparison}\textit{b}).
The results also suggest that the discrepancies between
chains with $\mathscr{N}-1$ and $\mathscr{N}$ beads diminish
as $\mathscr{N}$ increases (figure~\ref{fig:comparison}\textit{b})
as well as when $\mathit{Wi}$ increases (not shown).
We conclude that it is essential to consider the dynamics
of a full bead-spring chain in order to describe the scission
process correctly and achieve quantitative agreement between
experiments and constitutive models of polymer solutions.
\subsection{Active polymers}
\label{sect:active}
We now investigate the implications of the results obtained so far for the two-way-coupling
regime in which polymers perturb the surrounding flow.
When polymers break, their effective relaxation time $\tau^D$ decreases according to
\eqref{eq:mapping} and the solution,
at any point in time, consists of polymers with different $\tau^D$.
We can then introduce a mean Weissenberg number $\langle\mathit{Wi}\rangle (t)$,
which is defined as the average of $\lambda\tau^D$ over all polymers that
compose the solution at time $t$.
Studying the evolution of $\langle\mathit{Wi}\rangle(t)$, in a one-way-coupling simulation, helps us forsee how the effect of polymer-feedback on the flow would decay due to scission.
Let us, for the sake of simplicity,
consider the case of dumbbells ($\mathscr{N}=2$) and take
an initial ensemble of $N_p(0)$ dumbbells with
Weissenberg number $\mathit{Wi}_0$.
When a dumbbell breaks it forms two beads which formally have zero
Weissenberg number. Thus, at time $t$, the system consists of
$N_p(t)$ dumbbells with $\mathit{Wi}=Wi_0$ and $2[N_p(0)-N_p(t)]$ single
beads with $\mathit{Wi}=0$. Hence, for an ensemble of dumbbells,
\begin{equation}
\label{eq:mean-Wi}
\langle\mathit{Wi}\rangle(t)= \dfrac{N_p(t)}{2N_p(0)-N_p(t)}\mathit{Wi}_0.
\end{equation}
The temporal evolution of $\langle\mathit{Wi}\rangle$ is obtained by calculating $N_p(t)$ from the Lagrangian
database used in \S~\ref{sect:sim}
and is shown in figure~\ref{fig:active-3}{\it a}
for different values of $Wi_0>\mathit{Wi}_{\rm cr}$. Dumbbells with larger $\mathit{Wi}_0$ have a larger
scission rate, and therefore $\langle\mathit{Wi}\rangle$ vanishes rapidly.
In contrast, dumbbells with smaller $\mathit{Wi}_0$ break relatively slowly and
the mean $\mathit{Wi}$ of the solution remains nonzero for a longer time.
We note, en passant, that $\langle\mathit{Wi}\rangle$ becomes approximately equal to $\mathit{Wi}_{\rm cr}$
at $t\approx 50\tau_\eta$ for all $\mathit{Wi}_0$. By substituting \eqref{eq:N-decay} into
\eqref{eq:mean-Wi}, we thus deduce the following empirical expression for the
scission rate of dumbbells (see figure \ref{fig:active-3}{\it b}):
\begin{equation}
\frac{\tau_\eta}{T_d}\propto \ln\left[\frac{1}{2} \left( 1+ \frac{\mathit{Wi}}{\mathit{Wi}_{\rm cr}} \right)\right].
\end{equation}
The behaviour of $\langle\mathit{Wi}\rangle$ shown in figure~\ref{fig:active-3}{\it a}
suggests that a large value of $\mathit{Wi}_0$
will produce a polymer feedback that is initially strong but short-lived, decaying rapidly due to scission.
In contrast, a moderate value of $\mathit{Wi}_0$
yields a feedback that, albeit weaker, should last for a longer time and may therefore be more effective.
\begin{figure}
\centering
\includegraphics[width=0.485\textwidth]{pdf-comparison}\hfill%
\includegraphics[width=0.5\textwidth]{decay-comparison}
\caption{Passive polymers: (\textit{a})
time-integrated p.d.f. of the end-to-end separation of unbroken polymers
for $\mathit{Wi}=0.6$ and different numbers of beads $\mathscr{N}$;
(\textit{b}) fraction of unbroken polymers as a function of time
for $\mathit{Wi}=0.6$ and different numbers of beads $\mathscr{N}$.}
\label{fig:comparison}
\end{figure}
To investigate this point in a two-way-coupling simulation,
we take the hybrid Eulerian--Lagrangian approach
proposed by \citet{wg13jfm,wg13b,wg14}, which consists
in seeding the fluid
with a large number of FENE dumbbells and calculating the reaction force
exterted by the dumbbells upon the fluid.
This amounts to adding the term $\bnabla\bcdot \mathsfbi{T}_p$
to the right-hand side of the Navier--Stokes
equations \eqref{eq:NS}, where $\mathsfbi{T}_p$ is
the polymeric contribution to the stress tensor:
\begin{equation}
\mathsfbi{T}_p=\dfrac{\nu\eta L^3}{N_p(0)}\sum_{n=1}^{N_p(0)}
\dfrac{1}{\tau^{(n)}}
\left[
f\left(\dfrac{R^{(n)}}{Q_m^{D}}\right)
\dfrac{3\bm R^{(n)}\otimes\bm R^{(n)}}{r_{\rm eq}^2}
-\mathsfbi{I}
\right]
\delta\left(\bm x-\bm X^{(n)}_c\right).
\end{equation}
In the expression of $\mathsfbi{T}_p$, $L$ is the linear size of the
domain and $\eta$ is the ratio of the polymer to the solvent contribution
to the total viscosity of the solution ($\eta$ is proportional to the
volume fraction of dumbbells). The vectors $\bm X^{(n)}_c$ and
$\bm R^{(n)}$
are the positions of the centre of mass and the end-to-end
separation of the $n$-th dumbbell, respectively,
$\mathsfbi{I}$ is the identity matrix, and
\begin{equation}
\frac{1}{\tau^{(n)}}=
\begin{cases}
\tau^{-1} & \text{for $t < t^{(n)}_{\rm sc}$}
\\
0 & \text{for $t \geqslant t^{(n)}_{\rm sc}$,}
\end{cases}
\end{equation}
where $t^{(n)}_{\rm sc}$ is the smallest time such that $R^{(n)}=\ell_{\mathrm{sc}}$.
Thus, the dumbbells stop affecting the velocity field after breaking.
The evolution of the position and the configuration of each dumbbell is
given by~\eqref{eq:dumbbell}.
\begin{figure}
\centering
\includegraphics[width=0.485\textwidth]{mean-Wi-dumbbell.pdf}%
\hfill%
\includegraphics[width=0.5\textwidth]{t_d-dumbbell.pdf}%
\caption{Passive polymers: ({\it a}) mean Weissenberg number as a function of time
for $\mathscr{N}=2$ and different values of $\mathit{Wi}_0$; ({\it b}) reciprocal of $T_d$
(multiplied by the Kolmogorov time scale) as a function of $1+\mathit{Wi}/\mathit{Wi}_{\rm cr}$.
The dashed line is
proportional to $\ln[( 1+ \mathit{Wi}/\mathit{Wi}_{\rm cr})/2]$.}
\label{fig:active-3}
\end{figure}
The computational domain is a three-dimensional periodic box
with $L=2\upi$. A pseudo-spectral method with $128^3$ grid points
and the second-order Runge--Kutta algorithm
are used for the integration of the Navier--Stokes equations
in space and in time, respectively.
The equation for the dumbbells are solved by using the Euler--Maruyama
scheme. The turbulent flow is maintained by a forcing with Fourier
transform $\widehat{\bm F}(\bm k,t)=A(t)\hat{\bm u}(\bm k,t)$
for $1\leqslant \vert\bm k\vert \leqslant 2$ and zero otherwise,
where $A(t)$ is such that the kinetic-energy input rate
$\epsilon_{\rm in}=\langle \bm u\bcdot\bm F\rangle$ is constant.
We take $\epsilon_{\rm in}=0.5$ and
$\nu = 0.015$, which yields $\mathit{Re}_\lambda
=51$ in the absence of polymer feedback
(for more details on the simulation, the reader is referred
to \citealt{wg13jfm}).
We first evolve the velocity field alone with $\eta=0$ and, once
a statistically steady flow is obtained, $N_p(0)=5\times 10^7$
polymers with $\eta=4\times 10^{-2}$ are dispersed
randomly into the fluid. In the following, the time $t=0$ corresponds to the
time at which polymers are added to the flow.
As in \S~\ref{sect:first-scission}, $Q^{D}_m/r_{\rm eq}=\sqrt{3000}$,
$\ell_{\mathrm{sc}}=0.8Q_m^{D}$ and $r_0/r_{\rm eq}=1$.
We shall consider
three values of the Weissenberg number $\mathit{Wi}=\lambda\tau=0.6,0.8,1.0$,
where $\lambda$ is the Lyapunov exponent of the Newtonian flow ($\eta=0$).
We begin by examining the time evolution of $N_p(t)/N_p(0)$.
Figure~\ref{fig:active-1}{\em a}
shows that initially the fraction of unbroken polymers
decreases slowly. However, at longer times and for large enough $\mathit{Wi}$,
the decay becomes exponential.
The time at which the exponential decay sets in decreases rapidly with $\mathit{Wi}$, while the corresponding decay rate increases.
This behaviour is the consequence of the way
the feedback of an ensemble of polymers evolves as a result of scission.
In isotropic turbulence, the dispersion of polymers into a Newtonian
solvent reduces the kinetic-energy dissipation rate by a factor proportional
to the polymer concentration \citep*{chirag,de_angelis,pmr06,pmr10,wg13jfm}.
The time series of the
instantaneous kinetic-energy dissipation rate
\begin{equation}
\epsilon(t)=\frac{\nu}{L^3}\int_{[0,L]^3} \vert\nabla\bm u(\bm x,t)\vert^2
\mathrm{d}\bm x
\end{equation}
is shown in figure~\ref{fig:active-1}{\em b}
for different values of $\mathit{Wi}$.
At short times, the polymer feedback is fairly strong and
causes a significant reduction
of $\epsilon(t)$ compared to the value of the Newtonian flow.
The reduction of $\epsilon(t)$ is associated with a corresponding
reduction of the
amplitude of the velocity gradient and hence entails a weaker polymer
stretching and a lower probability of scission. For this reason,
the scission rate is initially small.
As time progresses, however, the concentration of unbroken polymers
keeps decreasing, until the effect of polymers becomes negligible
and $\epsilon(t)$ returns to its Newtonian value.
This is accompanied by a growth of the velocity gradient and thus a faster
decay of the fraction of unbroken polymers.
Since for higher values of $\mathit{Wi}$ polymers break more easily,
the time needed for the polymer feedback to vanish decreases
with increasing $\mathit{Wi}$. So for $\mathit{Wi}=1$ the reduction of energy
dissipation is initially stronger, but the return to the
Newtonian regime is fast; for $\mathit{Wi}=0.6$ the same process is
extremely slow and the dissipation reduction, albeit smaller,
lasts for a much longer time.
\begin{figure}
\centering
\includegraphics[width=0.5\textwidth]{fig8a.pdf}%
\includegraphics[width=0.5\textwidth]{fig8b.pdf}%
\caption{Active polymers: ({\em a}) fraction of unbroken polymers
as a function of time;
({\em b}) time series of the kinetic-energy dissipation rate.
}
\label{fig:active-1}
\end{figure}
One important consequence of this analysis is that, for a given experiment, an optimal, not necessarily very large, $\mathit{Wi}$
should maximize the energy-dissipation reduction integrated over the duration of the experiment.
\section{Multiple scissions}
We now return to passive polymers and examine how a population evolves when polymers can undergo multiple, repeated
scissions.
When a polymer composed of several beads undergoes its first scission (whose statistics was analysed in \S~\ref{sect:first-scission}), it results in two daughter polymers, each containing a smaller number of beads. These daughter polymers can themselves undergo further scissions to produce a tertiary generation, and so on, until, in case of complete breakage, we are left with only individual beads which represent small, inextensible polymer fragments. In this section, we study how this hierarchy of daughter polymers evolves, owing to multiple breakups of the parent polymers. After every scission event, we discard the broken link, form two new daughter polymers, and then follow their evolution along the trajectories of their respective centres of mass, which typically separate exponentially in time due to the positive Lyapunov exponent of the turbulent flow.
In the simulations for first-scission statistics, described earlier in \S~\ref{sect:sim}, we treated the centre of mass and the separation vectors of the polymer chains as the dynamical variables. This approach becomes inconvenient when dealing with multiple scissions, because new trajectories would have to be spawned after each scission event. So, we instead adopt the following approach. Consider $N_p(0)$ parent polymer chains, each of which is composed of $\mathscr{N}_0$ beads. Thus, the total number of beads in the simulation, which remains constant in time, is $N_b=\mathscr{N}_0 N_p(0)$. We assign to the beads distinct labels from 1 to $N_b$ and follow the time evolution of their positions. In addition, we maintain an array which records the labels of the first and last beads of every polymer chain in the simulation. So when a polymer chain undergoes a scission, we simply update this array: the broken chain is assigned all the beads preceding the broken link (and thereby becomes the first daughter polymer), while the remaining beads are assigned to a new entry in the array (to form the second daughter polymer). This procedure allows us to simulate the growing population of polymer chains without increasing the number of dynamical variables. For a given chain with $\mathscr{N}$ beads,
we still use \eqref{eq:chain} to evolve its dynamics,
by calculating $\bm X_c$ and $\bm Q_i$ from the position vectors of the beads, $\bm X_1,\dots,\bm X_{\mathscr{N}}$:
\begin{equation}
{\bm X}_c = \sum_{i=1}^{\mathscr{N}}\bm X_i \quad {\rm and} \quad {\bm Q}_i = {\bm X}_{i+1} - {\bm X}_{i} \;\; (1 \le i \le \mathscr{N}-1).
\end{equation}
All other aspects of these simulations follow the description given for the first-scission simulations in \S~\ref{sect:sim}, except that we now simulate the turbulent flow with a more moderate value of $Re_{\lambda} = 90$, by using $256^3$ collocation points and a time step of ${\rm d}t = 10^{-3}$. The reason for this is that multiple scission simulations require the polymers to be evolved simultaneously along with the flow (whereas a precomputed set of tracer trajectories were used to obtain all first-scission statistics), thus necessitating a new flow computation for each variation of the polymer parameters.
\begin{figure}
\begin{center}
\includegraphics[width=0.95\textwidth]{Nevol_4}
\end{center}
\caption{Passive polymers undergoing multiple scissions: evolution of the number of polymers, $N_P (t,\mathscr{N})$, categorized according to the number of constituent beads $\mathscr{N}$, due to the repeated scission of $N_P (0,10)$ 10-bead parent polymers with (\textit{a}) $\mathit{Wi}_0=0.8$, (\textit{b}) $\mathit{Wi}_0=1.0$, (\textit{c}) $\mathit{Wi}_0=2.0$ and (\textit{d}) $\mathit{Wi}_0=4.0$.}
\label{fig:decay_mult}
\end{figure}
At time $t=0$, the parents polymers are composed of $\mathscr{N}_0$ beads.
As a result of scissions, at later times polymers with different numbers of beads will be found in the flow.
We thus denote by $N_P(t,\mathscr{N})$ the number of polymers that, at time $t$, are composed of $\mathscr{N}$ beads.
We begin by examining how $N_P(t,\mathscr{N})$ evolves due to the repeated scission of an initial population of $N_P(0,\mathscr{N}_0)$ parent polymer chains, each with $\mathscr{N}_0=10$ beads. Figure~\ref{fig:decay_mult} presents results for various values of the parent-polymer Weissenberg number, $\mathit{Wi}_0$ = (\textit{a}) 0.8, (\textit{b}) 1.0, (\textit{c}) 2.0, and (\textit{d}) 4.0. Each $\mathscr{N}$-bead parent/daughter polymer is represented by a different curve, as indicated in the legend. We observe that for the small values of $\mathit{Wi}_0$ in panels (\textit{a}) and (\textit{b}), only 10-bead polymers break.
As shown in figure~\ref{fig:histograms}, most of the daughter polymers have 5 beads (along with a few non-5-bead polymers due to rare off-centre scissions),
because scission typically occurs at the central link for small $Wi_0$.
For the larger values of $\mathit{Wi}_0$ in panels (\textit{c}) and (\textit{d}), the daughter polymers also undergo scissions and the population is eventually dominated by small polymers and individual beads (inextensible polymer fragments).
Clearly, the extent to which polymers can undergo repeated scissions depends on the $\mathit{Wi}_0$ of the parent polymers. This is because each subsequent scission produces daughter polymers with a smaller relaxation time. This can be quantified by using \eqref{eq:mapping} to calculate the effective Weissenberg number of the $\mathscr{N}$-bead daughter polymer:
\begin{equation}
\mathit{Wi} = \frac{\mathscr{N} (\mathscr{N}+1)}{\mathscr{N}_0 (\mathscr{N}_0+1)} \mathit{Wi}_0,
\label{eq:WiN}
\end{equation}
where $2 \leq \mathscr{N} < \mathscr{N}_0$, and $\mathit{Wi}_0$ is the Weissenberg number of the parent polymer with $\mathscr{N}_0$ beads. Note that $\mathit{Wi} \equiv 0$ for individual beads ($\mathscr{N} = 1$). An approximate condition for a significant fraction of any given generation of daughter polymers to breakup is $Wi > Wi_{c} = 1/2$. So, if $\mathscr{N}_0 = 10$, \eqref{eq:WiN} implies that 5-bead daughter chains will not breakup, unless $\mathit{Wi}_0 > 1.83$. The results in figure~\ref{fig:decay_mult} agree with this estimate: 5-bead daughters experience a slow rate of scission for $\mathit{Wi}_0 = 2.0$ (panel ({\em c})), but no scission at all for $\mathit{Wi}_0$ = 0.8 and 1.0 (panels ({\em a}) and ({\em b}) respectively). Furthermore, for $\mathit{Wi}_0 = 4.0$, we expect scissions of daughter polymers that have $\mathscr{N} = 4$ beads or more, but certainly not if they are composed of only 2 beads. This is exactly what we observe in figure~\ref{fig:decay_mult} (panel ({\em d})). Thus, the condition $\mathit{Wi} > Wi_{\rm cr}$, in conjunction with \eqref{eq:WiN}, provides a simple way of estimating the number of beads of the smallest polymer that can be formed by repeated scission, given $\mathit{Wi}_0$.
Panels ({\em c}) and ({\em d}) of figure~\ref{fig:decay_mult} show that the evolution of the number of daughter polymers typically have two regimes. First, there is an enrichment phase, during which daughter polymers are formed due to the rapid breakup of the 10-bead parent polymer. Daughter polymers with various number of beads can be formed, especially at large $Wi$ for which off-centre breakups are quite frequent (figure~\ref{fig:histograms}). An exponential decay phase then appears for the daughter polymers that have enough beads to undergo further scission, e.g., $\mathscr{N} = 7, 6, 5, 4$ in panel ({\em d}). These secondary breakups produce a second phase of mild enrichment for the smallest polymers that do not breakup, e.g., $\mathscr{N} = 2$ in panels ({\em c}) and ({\em d}).
\begin{figure}
\begin{center}
\includegraphics[width=\textwidth]{Tsc_tauN_Ravg}
\end{center}
\caption{Passive polymers undergoing multiple scissions: (\textit{a}) probability distributions of the lifetime of a parent 10-bead polymer, a daughter 5-bead polymer formed because of the scission of a 10-bead polymer ($\mathscr{N}_0 = 10$) and a parent 5-bead polymer ($\mathscr{N}_0 = 5$) with same $\mathit{Wi}_0=4$; (\textit{b}) decay of $\mathit{Wi}$, averaged over the entire population of polymers, for various values of the initial $\mathit{Wi}_0$ of the 10-bead parent polymers; (\textit{c}) evolution of the averaged end-to-end extension $\langle R \rangle$ of the polymers for various $\mathit{Wi}_0$ of the parent polymers. }
\label{fig:Tsc_mult}
\end{figure}
The decay time $T_d$ of the number of daughter polymers, in the exponential decay regime, is larger for smaller $\mathscr{N}$, as evidenced by the shallower slopes of the corresponding curves (e.g., $\mathscr{N} = 7, 6, 5, 4$ in panel ({\em d})). This variation follows from figure~\ref{fig:decay}\textit{b}, provided that the decay time of each daughter polymer can be estimated from figure~\ref{fig:decay}\textit{b} by
using \eqref{eq:WiN} to calculate $\mathit{Wi}$. However, the use of figure~\ref{fig:decay}\textit{b} for daughter polymers is permissible only if the scission kinetics of a $\mathscr{N}$-bead daughter polymer is the same as the first-scission kinetics of a parent polymer that starts with $\mathscr{N}$ beads. To check if this is the case, we ran a separate simulation to calculate the distribution of survival times of 5-bead parent polymers ($\mathscr{N}_0 = 5$, $\mathscr{N} = 5$). In figure~\ref{fig:Tsc_mult}\textit{a}, these results are compared with the survival time distribution of 5-bead daughters, which form from the breakup of 10-bead parent polymers. The corresponding distribution for the 10-bead parent polymers is also shown for comparison. Remarkably, we see that daughter polymers have a much higher probability of breaking at early times than parent polymers with the same number of beads. This is because daughter polymers are formed from scission events:
they typically start out in a stretched configuration and so have a much higher probability of breaking quickly than a randomly initialized, parent polymer with the same number of beads. The effect of this fades quite quickly, though, and for times larger than the Lagrangian decorrelation time $T_L \approx 10 \tau_\eta$
(for $Re_{\lambda} \sim 10^2$; see \citealt{Yeung2002}) the survival time p.d.f of the daughter polymers begins to resemble that of the parent polymers with the same number of beads. The time scales associated with the two exponential tails is therefore approximately the same, and one can indeed estimate the long-term decay time scale (or equivalently the decay rate) of an $\mathscr{N}$-bead daughter polymer, formed due to multiple scissions, by using just the first-scission statistics of a parent polymer with the same number of beads.
Can the early-time, high scission probability of daughter polymers be ignored? This depends on the $\mathit{Wi}$ of the daughter polymers. From figure~\ref{fig:decay}\textit{b}, we expect $T_d \gg T_L$ for small $\mathit{Wi} \lesssim 1$; the number of polymers that would breakup during the initial time $T_L$ will then be relatively small, and may be ignored. However, for larger
$\mathit{Wi} \gtrsim 4$, we have $T_d \approx T_L$ and a significant fraction of daughter polymers would break before they forget their stretched initial conditions. Therefore, for large $\mathit{Wi}_0$ (which would produce daughter polymers with large $\mathit{Wi}$), it becomes important to simulate multiple scissions in order to faithfully describe the decay of the polymer population. Such simulations would have to be repeated if the number of beads of the parent polymer changes. The situation simplifies considerably, however, for small $\mathit{Wi}_0$ as one can then use first-scission statistics, calculated for a range of $\mathscr{N}$, to describe the decay of daughter polymers, regardless of the number of beads of the parent polymer.
The evolution of $\mathit{Wi}$ averaged over all polymers, which represents the mean effective $\mathit{Wi}$ of the entire population, is shown in figure~\ref{fig:Tsc_mult}\textit{b}, for various values of $\mathit{Wi}_0$ of the parent polymers. This is the analogue of figure~\ref{fig:active-3},
but for $\mathscr{N}_0 = 10$ rather than $2$. We see that $\langle \mathit{Wi} \rangle$ decreases rapidly while the parent polymers are breaking. After this, the slower
scission rates of the daughter polymers lead to a more gradual decrease in $\langle \mathit{Wi} \rangle$, which eventually will saturate to a value less than $Wi_{\rm cr}$, corresponding to daughter polymers that cannot be broken up further by the flow.
This effect cannot be captured by dumbbells, which directly breaks into
beads (figure~\ref{fig:active-3}).
The reduction of $\langle \mathit{Wi} \rangle$ with time has a strong influence on the evolution of the mean end-to-end extension of the polymers $\langle R \rangle$, as shown in figure~\ref{fig:Tsc_mult}\textit{c}. After an initial stretching phase, $\langle R \rangle$ begins to decrease, owing to both the scission of highly
stretched polymers and the smaller $\mathit{Wi}$ of the resulting daughter polymers. After all scissions cease, we are left with relatively in-extensible daughter polymers whose extensions fluctuate near $r_{eq}$. Interestingly, the curve for $\mathit{Wi}_0 = 2.0$ shows two regimes in the decay of $\langle R \rangle$. The first fast-decay regime is due to the rapid scission of 10-bead parents (see figure~\ref{fig:decay_mult}\textit{c}); the second slow-decay regime results from the much slower breakup of 5-bead daughter polymers, for which $\mathit{Wi} = 0.55$ only {\it just}
satisfies the condition for secondary scission ($\mathit{Wi} > Wi_{\rm cr} = 1/2$). This second slow regime is not seen for either smaller or larger $Wi_0$: in the former case, secondary breakups do not occur, whereas in the latter case secondary breakups occur very quickly (see figure~\ref{fig:decay_mult}).
\section{Concluding remarks}
Polymers, even in small quantities, have a dramatic impact on a turbulent flow, reducing drag or dissipation and suppressing small-scale motion. However, because these effects
originate from polymer stretching (and the resultant feedback forces), the polymers which exert the strongest influence on the flow are also the most susceptible to strain-induced
scission. Therefore, to achieve effective flow modification, one must strike a balance between these opposing tendencies, which in turn demands a detailed understanding of the scission
process and the factors that influence the rate of scission.
In this work, we have analysed the scission of polymers in homogeneous isotropic turbulence, with a focus on the temporal decay of unbroken polymers, and the statistics of their survival
times. By using direct numerical simulations, we have quantified the decay time (or scission rate) as a function of $Wi$, which can serve as inputs for coarse-grained models. Importantly,
all the key qualitative features of the numerical results can be predicted analytically by replacing the fluctuating, turbulent velocity gradient by a time-decorrelated Gaussian random
flow.
This is possible
because scission is caused by the cumulative action of fluctuating strain, and \textit{not} by sudden stretching in high-strain regions of the flow.
Another finding, relevant for future
computations, is that a multi-bead polymer chain cannot be replaced by a dumbbell model without incurring quantitative
errors in the prediction of breakup rates. However, the results appear to
converge as $N_b$ increases, so we expect $N_b \sim O(10)$ to be sufficient even if a polymer model may strictly demand many more beads.
Our study of the scission of active polymers has shown that there is an intermediate value of $\mathit{Wi}$ for which the overall, time-integrated, reduction
of the kinetic-energy dissipation rate is maximum: for small $\mathit{Wi}$
the polymers do not stretch and the feedback is weak, whereas for large $Wi$ the stretching and feedback is
initially strong, but the resultant dissipation-reduction is lost rapidly as the polymers breakup very quickly.
We have shown that in a sufficiently strong turbulent flow (large $Wi$) polymers can breakup repeatedly. However, because the fragments in each successive generation have a smaller relaxation time, the breakup process eventually ceases once the effective $Wi$ of the surviving polymer fragments becomes less that $Wi_{cr}$. From this condition, we can estimate the number of beads in the largest surviving chains as $\mathscr{N} (\mathscr{N}+1) \le 6 Wi_{\rm cr} \tau_{\eta}/\tau$ [using ~\eqref{eq:mapping}]. Now, as $\mathscr{N}$ is linearly related to the mass of a polymer, this condition allows us to estimate how the weight-averaged molar mass (biased towards the largest chains) of the surviving polymers $M_{\rm ws}$ scales with the Reynolds number $Re$. For large $\mathscr{N}$, we have $\mathscr{N}^2 \sim \tau_{\eta}/\tau \sim Re^{-3/2} \nu/D^2 \tau$, where $D$ is the large length-scale of the flow system. Therefore, for a specified polymer, solvent and system geometry we obtain $M_{\rm ws}^2 \sim Re^{-3/2}$. This scaling is consistent with the experimental data of \cite{vcs06}, who obtain a power-law exponent close to $-3/2$ for a variety of polymers and system geometries (these results are reported in terms of the squared, weight-averaged polymer length which is linearly related to $M_{\rm ws}^2$, as described in the Data Analysis section of \citealt{vcs06}).
The multiple-scission statistics of a polymer also show a non-monotonic dependence on $\mathit{Wi}$. Small $\mathit{Wi}$ polymers break only once, if at all, whereas large $\mathit{Wi}$ polymers undergo a
rapid succession of breakups and quickly reach their limiting generation (fragments which are no longer stretched by the flow). However, for intermediate $\mathit{Wi}$, the first-scission
occurs quickly, but then the daughter polymers breakup much more slowly. This introduces multiple time-scales into the decay of $\langle Wi \rangle$, the average effective $Wi$ of the
polymer population. This average quantity and its evolution are relevant to coarse-grained continuum models of polymer solutions which typically contain a single mean
polymer-relaxation-time parameter. Indeed, the development of continuum models that incorporate scission is essential for predicting the long-time dynamics of turbulent polymer
solutions in complex applications. The quantitative results as well as physical insights gained from this study should aid in the development of such models.
\begin{acknowledgments}
JRP, SSR
and DV thank the Indo--French Centre for Applied Mathematics (IFCAM)
for financial support.
DV acknowledges his Associateship with the International Centre for Theoretical Sciences, Tata Institute of Fundamental Research, Bangalore, India.
The computations were perfomed at Centre de Calculs Interactifs of
Universit\'e C\^ote d'Azur, at
Information Technology Centre in Nagoya University and National Institute for Fusion Science (NIFS),
and at the cluster Mowgli and the work station Goopy at the ICTS--TIFR.
TW thanks the financial support by MEXT KAKENHI through Grant No.~20H00225 and JSPS KAKENHI Grants No.~18K03925, and the computational supports provided by JHPCN (jh190018-NAH, jh200006) and by NIFS (NIFS18KNXN366, NIFS18KNSS105).
SSR acknowledges support of the DAE,
Govt. of India, under project no.~12-R\&D-TFR-5.10-1100 and DST (India) project MTR/2019/001553 for support.
JRP acknowledges funding from the IIT Bombay IRCC Seed
grant.
\end{acknowledgments}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 9,299
|
Талгортъёган (устар. Тал-Горт-Юган) — река в Шурышкарском районе Ямало-Ненецкого АО. До слияния с левым притоком Ай-Талгортъёган называется Ун-Талгортъёган (или Ун-Тал-Горт-Юган). Устье реки находится в 9 км по левому берегу протока Нярхулпосл. Длина реки составляет 23 км.
Данные водного реестра
По данным государственного водного реестра России относится к Нижнеобскому бассейновому округу, водохозяйственный участок реки — Обь от впадения реки Северная Сосьва до города Салехард, речной подбассейн реки — бассейны притоков Оби ниже впадения Северной Сосьвы. Речной бассейн реки — (Нижняя) Обь от впадения Иртыша.
Примечания
Реки Шурышкарского района
Бассейн Малой Оби
Притоки проток и рукавов Оби
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 248
|
Mount Siyeh, with a height of , is the fifth tallest and one of six peaks over in Glacier National Park, Montana, United States. Mount Siyeh was named after a Blackfeet Indian, Sai-yeh, whose name means "Crazy Dog" or "Mad Wolf."
Mount Siyeh stands about two miles east of the Continental Divide, within the watershed of the Saint Mary River, which drains into the Saskatchewan River, and ultimately into Hudson Bay.
Mount Siyeh is rather easily accessible from a variety of different routes. It can be ascended via a long scree scramble from the Preston Park area, or from Piegan Pass via Cataract Mountain. The summit can also be reached via a long hiking/mountaineering route dubbed the "Skyline Experience"; this route starts from the Many Glacier Hotel and involves ridge walk and includes the summits of Wynn Mountain () and Cracker Peak () before reaching the top of Siyeh.
Route Descriptions can be found on Summitpost or in A Climbers Guide to Glacier National Park.
Geology
Like other mountains in Glacier National Park, it is composed of sedimentary rock laid down during the Precambrian to Jurassic periods. Formed in shallow seas, this sedimentary rock was initially uplifted beginning 170 million years ago when the Lewis Overthrust fault pushed an enormous slab of precambrian rocks thick, wide and long over younger rock of the cretaceous period.
Climate
Based on the Köppen climate classification, it is located in an alpine subarctic climate zone with long, cold, snowy winters, and cool to warm summers. Winter temperatures can drop below −10 °F with wind chill factors below −30 °F.
Gallery
See also
List of mountains and mountain ranges of Glacier National Park (U.S.)
References
Mountains of Glacier County, Montana
Mountains of Glacier National Park (U.S.)
Lewis Range
Mountains of Montana
North American 3000 m summits
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 5,700
|
{"url":"http:\/\/www.bot-thoughts.com\/2009\/07\/spice-part-5.html","text":"## Tuesday, July 7, 2009\n\n### SPICE: Part 5\n\n<< [Part 1 2 3 4 5]\n\nTo continue with the SPICE experimentation, let's take a look at doing a better job of simulating a motor while using the SPICE subcircuit (.SUBCKT) feature. Here's a fantastic article from eCircuit Center on modeling not only the electronic but the mechanical properties of a DC motor in SPICE. I'll let you read the article yourself to get all the nuances.\n\nSubcircuits Overview\n\nMeanwhile, I'd like to take the motor definition and turn it into a subcircuit in SPICE. Subcircuits provide a way to encapsulate a lot of electronics detail, like that of an op amp, logic chips, or other integrated circuits, while presenting a simple interface to the main model in the form of external nodes, aka input\/output pins.\n\nOur motor subcircuit will have two external nodes: A, and B. These represent the two terminals you'd see on this on any DC brush motor. This subcircuit will plug into our earlier circuit in place of the resistor RM that we were using to simulate a motor.\n\nHere's what the motor simulation circuit looks like. For a full explanation of the circuit see the article referenced above. This diagram was drafted in LTspice, a Windows-based implementaiton of SPICE that makes it extremely easy to draft and simulate circuits. I sort of spontaneously decided to give LTspice a try.\n\nA side note: if you work this out yourself, I ran into some glitches trying to build a current controlled voltage source in LTspice, the E_EMF above. If you use the drawing above, you get a different motor simulation than if you use the file provided on the DC motor simulation website.\n\nBefore we go further, I want to talk a little about LTspice.\n\nQuick LTspice Overview\n\nLinear Technology's LTspice is a Windows program that combines circuit CAD and SPICE simulation. I've been hearing about it over on Electro-Tech-Online forums and with the recent series on SPICE using MacSpice I figured why not give this little gem a try. Having just acquired a \\$5 1.5GHz Pentium 4 tower with Windows XP at a nearby yard sale, there's no time like the present.\n\nNot to make this into an LTspice user guide or extensive tutorial, here is a quick overview of the software. In short, the process for using the program is:\n2. Define component models,\n3. Wire components, and\n4. Set up and run the simulation.\nOn the right side of the toolbar is the set of icons for adding components to the schematic.\n\nResistors, capacitors, inductors, ground, and diodes each have their own symbol.\n\nTransistors and voltage supplies are added via the add component button which looks like an AND gate. From a dialog box, you select the type of component such as, LED, MOSFET, JFET, Comparator, Voltage Regulator, and more, even custom components (more on this later). In our case above, an NPN transistor, and voltage (supply).\n\nOnce you have components in place, you specify details about them, like passive component values, supply voltages, supply types (PULSE, SINE, etc), transistor models (e.g., 2N3055 or 2N2222), and so forth.\n\nWire the components with the pencil icon which features a handy vertical & horizontal alignment rules. You can easily move or drag components or entire sections of the circuit, rotate them, mirror them. It's really simple to quickly create your circuit diagram and lay it out just how you want.\n\nOnce you have your circuit defined you can run a simulation. In this case I chose to run an operating point simulation. Just select Edit simulation command from the Simulate menu. Select the DC op pnt tab and click Ok. Place the .OP directive onto the sheet. A similar procedure would be used for transient, AC or DC analysis.\n\nNow select Run from the Simulate menu or click the run icon in the toolbar and you get very readable outputs (unlike the raw SPICE or MacSpice output). Now that we've briefly reviewed LTspice, let's get back to improving the motor simulation.\n\nSubircuits\n\nTo encapsulate all the motor simulation circuitry above into a subcircuit, the circuit description is wrapped in SUBCKT \/ ENDS cards. The SUBCKT line specifies the name of the subcircuit and the number of externally accessible nodes ala:\n\n.SUBCKT SUBNAM N1 [ N2 N3 ... ]\n\n.ENDS\n\nSubcircuits are referenced (called) by the pseudo-element identifier X followed by whatever name assigned to the instance of the subcircuit:\n\nXYYYYYYY N1 [ N2 N3 ... ] SUBNAM\n\nFor our motor, we'll use the following sort of template.\n.SUBCKT motor A B** Motor description cards go here*.ENDS\nNode A will be the + voltage supply terminal, and B will be the - terminal.\n\nMotor Model\n\nHere's the SPICE deck for the motor from the website above, less the voltage supply and analysis\/print directives and with the .SUBCKT \/ .ENDS cards added. I also changed the series resistance, RA, to 16\u03a9. While I was toying with all this, it seemed that 16\u03a9 resulted in a motor with current requirements compatible with our original circuit, about 490mA, steady-state:\n.SUBCKT MOTOR A B* MOTOR VOLTAGERA A 2 16LA 2 3 0.0015H_EMF 3 4 VSENSE2 0.05VSENSE1 4 B DC 0V** MOTOR TORQUE BASED ON INERTIA AND FRICTIONH_TORQ 6 0 VSENSE1 0.05LJ 6 7 0.00025RB 7 8 0.0001VSENSE2 8 0 DC 0V** MOTOR POSITIONFPOS 0 11 VSENSE2 1CPOS 11 0 1RPOS 11 0 1MEG.ENDS\nAs mentioned before, the subcircuit generated by LTspice looks a bit different:\n.subckt motor a bLM a N002 0.0015LJ N001 N004 0.00025RM N002 N005 16RB N004 N006 0.0001E_EMF N005 N007 N006 0 0.05VSENSE1 N007 b 0VSENSE2 N006 0 0H_TORQ N001 0 VSENSE1 0.05FPOS 0 N003 VSENSE2 1CPOS N003 0 1RPOS N003 0 1m.ends motor\nIncluding Files\n\nSPICE also lets you include entire files via the .INCLUDE card. I used this when simulating the motor in MacSpice.\n\n.INCLUDE filename\n\nSo let's put the motor subcircuit in a separate file, motor.cir, and we can include it into our earlier circuit description.\n\nCustom Components in LTspice\n\nWhereas in MacSpice one has to manually set up files with subcircuits to be included, LTspice front ends these SPICE features by enabling the user to create new components. First, you draft a new symbol, in this case, I drew a typical motor symbol (right), a circle with an M in it, and two pins (ports) shown as squares.\n\nNext you set up a schematic in the same directory and with the same base name as the symbol (e.g., motor.asy and motor.asc). You associate the pins\/ports on the drawing with those in the schematic and now you can just plug in this component by adding it out of your directory just like you'd add any other LTspice component. This is a pretty clever approach that gives LTspice a lot of flexibility and power.\n\nBehind the scenes, LTspice is using subckt cards to implement this hierarchy. It puts the subcircuit description inline in your circuit diagram when you load a component. You can see how this is implemented by selecting View>SPICE Netlist.\n\nHooking up the Motor\n\nIn MacSpice you can include the subcircuit file, so after adding the appropriate file include card, we replace RM, connected to nodes 1 and 3, with a call to the motor subcircuit (XMOTOR 1 3 MOTOR). So here's what we end up with:\nMotor driver** Sources*V9 1 0 9.0V5 2 0 5.0** Simulate motor as resistor*.INCLUDE Macintosh\\ HD:Users:mes:Desktop:motor.cir*RM 1 3 18XMOTOR 1 3 MOTOR* Measure motor currentVM 3 4 0...\nThe complete MacSpice circuit file is here: ex5.cir\n\nMeanwhile, using LTspice, I created the motor driver circuit, replacing RM with the motor custom component and then added a few extras. To simulate the 5V logic signal from the AVR microcontroller, I installed a voltage controlled switch, S1, which is repeatedly turned on and off by a 5V PULSE voltage supply.\n\nWhen I simulated the circuit at first, as expected, I was seeing massive voltage spikes across Q1 because of the motor's inductance. The cure for this is to place a schottky diode across the motor itself, D1 above.\n\nNote that when using a MOSFET driver instead of a BJT, a diode is present within the transistor package. However, I've seen advice to use an external diode just in case the current or voltage is too high for the internal diode to handle.\n\nFinally, I added a pull up resistor, R1, to ensure that the transistor is off when the AVR MCU's output pin is in a high impedance state (during reboot, MCU off, etc.). The only way to turn on the motor is if MCU sets the pin low to sink current.\n\nSimulation Results\n\nIn LTspice, to do operating point analysis, I bypassed the S1 switch and hardwired the base resistor to ground. Here are the results. The interesting results are in blue.\nV(m2): 0.111133 voltageV(n005): 0.496462 voltageV(n004): 4.95934 voltageV(m1): 9 voltageV(n002): 4.04701 voltageV(n001): 5 voltageV(n003): 0 voltageIc(Q3): -0.0495875 device_currentIb(Q3): -0.00861067 device_currentIe(Q3): 0.0581982 device_currentIc(Q1): 0.555551 device_currentIb(Q1): 0.0495875 device_currentIe(Q1): -0.605139 device_currentI(D1): -3.17e-005 device_currentI(R1): 0.005 device_currentI(R3): 0.00861067 device_currentI(R2): -0.0495875 device_currentI(S1): 0 device_currentI(Vsw): 0 device_currentI(V2): -0.0631982 device_currentI(V1): -0.555551 device_currentIx(1:A): 0.555519 subckt_currentIx(1:B): -0.555519 subckt_current\nNote that LTspice displays all currents without having to insert 0v supplies -- it does this behind the scenes. A nice feature!\n\nAlso, it usually displays voltages by referencing the device name, rather than the associated node names, making it a bit easier to understand the results.\n\nFinally it displays the values as decimals rather than scientific notation so there's no mental calculations (e.g., let's see 0.1E-2 is, uhh... 1mA...)\n\nSo we're seeing about 556mA through the motor. When I used the subcircuit file given by the DC Motor simulation website, apparently the back emf behaves a bit differently than what I drew up in LTspice because current is quite a bit lower, about 217mA. So, time permitting I'll play around with different values of the components in the motor simulation. At some point it would be handy to be able to model any motor based on datasheet specs. But that's best saved for a whole 'nother article.\n\nNext, I performed transient analysis in LTspice. The software makes it easy to graph the results; run the transient analysis, then click on areas of the circuit with your virtual 'probe' to read voltage\/current.\n\nMacSpice plots looked the same, in case you were wondering. Notice the voltage clamping effect of the schottky diode. The motor is still generating small voltage spikes at turn on and shut off, but they are orders of magnitude smaller than before. One concern I had was that motor current would spike (stall current) when turning on. Not so.\n\nInstead, both spice simulations seem to show a gradual ramp-up of current. Is this because the Q1 transistor is going out of saturation when the motor first fires up, because IC is so high initially? (IB less than IC \u00f7 \u00df). Something else I can investigate later.\n\nMotor Driver Next Steps\n\nSo I'm pretty happy with the simulations at this point. My next step is to gather real data on the motor, particularly its steady state current requirement with the fan attached. I think using the complex motor subcircuit isn't necessary to model a motor that simply turns on for a few seconds to blow out a candle.\n\nSo modeling the motor as a simple resistor would be good enough to let me modify the circuit above to drive the motor. Or, better yet, find a different driver transistor since the TIP3055 (2N3055) is total overkill and model it, instead. Whenever I get around to driving a 15A(!) motor I might put one of these to use again. :)\n\nAnother thought is to use a MOSFET transistor but an N-channel would need some driver hardware to get the gate voltage above the drain voltage. Another idea is to use a dedicated motor driver IC but I kind of want to stay old school with discrete components.\n\nOf course, I can also use a different motor. It'd be simpler to use the robot's existing supply voltage and a 5-6V motor than having to tack on a 9V battery. It's more of a challenge to design around all the constraints of using parts you have on hand.\n\nThat's It\n\nSo, that's it for the SPICE series. Hope it's been helpful and\/or interesting. Thanks for reading!\n\nThis isn't the last time I use or post about SPICE. I want to improve the IR LED flame detection circuit soon. SPICE will, I think, provide a good way to model what is happening and work on improving the amplifier part of this circuit.\n\nAlso, since I didn't get around to talking about converting between Eagle circuit drawings and SPICE decks, I'll post about that in a separate article at some point.\n\nSpecial thanks to the electronics gurus at Electro-Tech-Online for some input and advice on BJT saturation.\n\n<< [Part 1 2 3 4 5]","date":"2017-06-26 03:34:09","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 1, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.4098116457462311, \"perplexity\": 3248.0105135741255}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2017-26\/segments\/1498128320669.83\/warc\/CC-MAIN-20170626032235-20170626052235-00715.warc.gz\"}"}
| null | null |
Q: Speed of convergence of Lebesgue integrable function as domain goes to zero
For $p \in [1, \infty]$ find values of $\lambda$ such that $\lim\limits_{\epsilon \to 0^{+}} \frac{1}{\epsilon^{\lambda}} \int_{0}^{\epsilon} f = 0$ for all $f \in L^{p}[0, 1]$.
Can someone help do this?
A: Let $H_\epsilon(f,\lambda)=\epsilon^{-\lambda}\int\limits_0^\epsilon f$. For every $a\gt0$, let $f_a(x)=1/x^{a}$ and $g_a(x)=1/(x^{a}\log(x))$.
Let us first consider the case $p\gt1$.
As mentioned and proved by Paul, if $\lambda\lt1-1/p$, $H_\epsilon(f,\lambda)\to0$ for every $f$ in $L^p$.
On the other hand, if $\lambda\gt1-1/p$, $f_a$ is in $L^p$ for every $a\lt1/p$ but $H_\epsilon(f_a,\lambda)$ is of the order of $\epsilon^{-\lambda}\times\epsilon^{1-a}$ hence $H_\epsilon(f_a,\lambda)\to\infty$ if $a\gt1-\lambda$. Since $1-\lambda\lt1/p$, both conditions can be met by the same parameter $a$.
If $\lambda=1-1/p$, $g_{1/p}$ is in $L^p$ but $H_\epsilon(g_{1/p},\lambda)\to\infty$.
Turning to the case $p=1$, note that $H_\epsilon(f,0)\to0$ for every $f$ in $L^1$ but for every $\lambda\gt0$ and $1-\lambda\lt a\lt1$, $H_\epsilon(f_a,\lambda)\to\infty$.
Finally, the suitable values of $\lambda$ are $\lambda\lt1-1/p$ for $p\gt1$ and $\lambda=0$ if $p=1$.
A: By Holder's inequality, we have
$$\int_0^\epsilon f\leq\big(\int_0^1f^p\big)^{\frac{1}{p}}\big(\int_0^\epsilon\big)^{\frac{1}{q}}=\epsilon^{{\frac{1}{q}}}\big(\int_0^1f^p\big)^{\frac{1}{p}}$$
where $\frac{1}{q}=1-\frac{1}{p}=\frac{p-1}{p}.$
Hence, if $\lambda<\frac{1}{q}=\frac{p-1}{p},$ then
$$0\leq\lim\limits_{\epsilon \to 0^{+}} \frac{1}{\epsilon^{\lambda}} \int_{0}^{\epsilon} f\leq\lim\limits_{\epsilon \to 0^{+}}\epsilon^{\frac{1}{q}-\lambda}\big(\int_0^1f^p\big)^{\frac{1}{p}}= 0$$
since $f\in L^p[0,1]$.
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 5,144
|
{"url":"https:\/\/ehealthbilbao.com\/uhz9unr\/what-percentage-is-a-b-in-australia-68a391","text":"# what percentage is a b in australia\n\n[5] More than 80 percent of strokes can be prevented. More than 96 percent of eligible Australians are enrolled to vote. \u00d7 Grades are usually offered as qualitative assessments and are recorded on transcripts as high distinction, distinction, credit, pass, and fail. What Is The 4.5 GPA To Percentage Conversion In Australia? On a high level, this is how Australian grades compare to US. credit points The largest change was between 2011 (22 per cent) and 2016, when an additional 2.2 million people reported having no religion. Based on this, the most common grade conversion being currently followed to convert Australia grades on a 4.0 GPA scale is as below: Here\u2019s state level grade comparison for your reference: Conceded, Conditional or Compensatory Pass. Is 75 an A in Australia? To simply compare two numbers, use the percentage calculator. GST. Of those, more than 90 percent typically turn out to cast ballots for a federal election, far more than the 55 percent \u2026 A conceded pass is a pass for a course that has been awarded only after supplementary assessment has been undertaken by the student. South Australia is the fifth-most populous state in Australia, accounting for about 7% of the country's population. Our free online Percent Calculator calculates percentages such as ratios, fractions, statistics, and percentage increase or decrease. Each had their language, culture, and belief system. Percentage Calculator. Adelaide is the capital city of South Australia and is home to about 5% of the cou\u2026 ) % Tips: Use tab to move to the next field. Overview of academic grading in Australia, \"Computer Professional Education Program: Subject Grading\", University of Sydney Academic Board Resolutions: Assessment and Examination of Coursework, \"Grades and Grade Point Average (GPA) - Academic Policy\", http:\/\/www.acap.edu.au\/current-students\/managing-my-course\/student-handbook\/new-two-column-page\/, \"University of Queensland CIVL3340 Course Profile Grading System\", \"University of Queensland LAWS5228 Course Profile Grading System\", \"How is a grade point average (GPA) calculated? moabboy says: July 27, 2015 at 2:59 pm Calculating body fat percentage isn\u2019t always completely accurate, and there are many methods to try. Various tertiary institutions in Australia have policies on the allocations for each grade and scaling may occur to meet these policies. The calculations and formulas (press the '?' Most institutions also provide a number to represent qualitative descriptions. The WAM is based on the raw percentage grades, or marks, achieved by the student, rather than grade points such as High Distinction or Distinction. Australia is considered to be one of the world\u2019s major \u2018immigrationnations\u2018(together with New Zealand, Canada and the USA). ( = Australia granted 490 refugee visas to refugees awaiting resettlement with UNHCR assistance in 2010\u201311. Australia maintains a grading system on 7.0 band. percentage difference = 100 * |a - b| \/ ((a + b) \/ 2) To get even more specific, you may talk about a percentage increase or percentage decrease. a b plus is a 77-79. a b is a 73-76. a b minus is a 70-72. a c plus is a 67-60. a c is a 63-66. a c minus is a 60-62. a d plus is a 57-59. a d is a 53-56. a d minus is a 50-52. a f is a 49 and below. Press enter to calculate. Next step: Attend our GST webinar \u2013 to help you to understand GST and its implications for business. All were to refugees from Myanmar. We\u2019ll talk about the healthy range for men and women, and the limitations. Percentage to GPA Conversion Chart for Australia; Percentage Range Grade Grade Point US Grade; Above 85%: HD(High Distinction) 7.00: A+: 75% - 84%: D(Distinction) 6.00: A: 65% - 74%: Cr(Credit Pass) 5.00: B: 50% - 64%: P(Pass) 4.00: C: 45% - 49%: F1(Fail level1) 1.00: F: Below 45%: F2(Fail level 2) 0: F The term course unit values is used to distinguish between courses which have different weightings, for example between a full year course and a single semester course. 60 \u2013 100 percentage is an A grade and 55 \u2013 59 percentage is a B+ grade. Percentage Formula. 8.4% of the burden of disease in Australia was attributable to overweight and obesity in 2015 1 in 3 (31%) Australian adults were obese in 2017\u201318, an increase from 1 in 5 (19%) in 1995 The prevalence of obesity among people born in 1973\u20131982 increased \u2026 Australia was once a land of only the aboriginals or the indigenous people. Before the British colonization in 1788, over 500 different indigenous groups controlled the area. \u2211 The population of Australia is estimated to be 25,730,800 as of 14 January 2021. Its population is concentrated mainly in urban areas and is expected to exceed 28 million by 2030. This system represents a mixture of the pass\/fail system and a numerically graded system. This article is a summary of academic grading in Australia. Australia's national statistical agency providing trusted official statistics on a wide range of economic, social, population and environmental matters. Some universities also have a Pass Conceded (PC) grade for marks that fall in the range of 45\u201349 inclusive. AUSTRALIA has six social classes, according to new research from social scientists at the Australian National University. 30-34 Australia and 0-4 Solomon Islands C. 25-29 Australia and 10-14 Solomon Islands D. 30-34 Australia and 25-29 Solomon Islands E. 0-4 Australia and 30-34 Solomon Islands The faster growth in the Aboriginal population (compared to 1.6% for the general Australian population) is t\u2026 Raw marks for students who fail are not scaled and do not increase the allocations of higher grades. In 1996, following the Port Arthur massacre, the federal government and the states and territories agreed to a uniform approach to firearms regulation, including a ban on certain semiautomatic and self-loading rifles and shotguns, standard licensing and permit criteria, storage requirements and inspections, and gre\u2026 As of June 2019, about 12.79 million females and 12.6 million males lived in Australia. You calculate your overall GPA by averaging the scores of all your classes. Australia's population has quadrupled in the past century, with the number of people tipped to reach 25 million this year. Grade point averages are not generally used in Australia below a tertiary level. Although the percentage formula can be written in different forms, it is essentially an algebraic equation involving three values. Using the graphic below, what age ranges have the highest percent of population in Australia and the Solomon Islands? Of these, the vast majority (438) were from Myanmar. In the Indian GPA system, the band is 10 points, firstly the marks are converted into percentage and then it is converted into GPA points out of 10. The growing percentage of Australia\u2019s population reporting no religion has been a trend for decades, and is accelerating. Note that the numbers above do not correspond to a percentile, but are notionally a percentage of the maximum raw marks available. Use shift-tab to move to the previous field. At universities, they are calculated according to a more complicated formula than in some other nations: GPA These policies may vary also according to the degree year (higher percentages for later years), but generally, only 2\u20135% of students who pass (that is, who achieve raw marks of 50 or more) may be awarded a High Distinction grade, and 50% or more of passing students are awarded a basic Pass grade. B C 3.00 - 3.99 50.00 - 69.99 Sound Standard of Achievement C ... South Australia. The Canadian GPA grade calculation has 12 percentage ranges as shown below in the conversion chart. Since 1945, whenthe first federal immigration portfolio was created, over 7.5 million peoplehave settledhere and Australia\u2019s overseas-born resident population\u2014estimated to be 28.2per cent of the population in June 2015\u2014is considered high compared to mostother OECDcountries. The percentage below 49.99 % is meant to be F, and the top being A+, A- as shown below. It is also the fourth-largest state in total land area, accounting for about 12% of the country's total area. These policies may vary also according to the degree year (higher percentages for later years), but generally, only 2\u20135% \u2026 Griffith University and The University of Queensland issue results of High Distinction, Distinction, Credit, Pass, and Fail.[42][43]. {\\displaystyle {\\text{GPA}}={{\\sum {({\\text{grade points}}\\times {\\text{credit points for unit}})}} \\over {\\sum {\\text{credit points}}}},}. There has been a small increase in the percentage of Australia\u2019s population who say they were born overseas, from 25 per cent in 2011 to 26 per cent. Now you know the percentage difference formula and how to use it. , Australian universities issue results for each subject, based on the following gradings: Note that the numbers above do not correspond to a percentile, but are notionally a percentage of the maximum raw marks available. Australia is the 52nd most populous country in the world and the most populous Oceanian country. Colleges report GPA (grade point average) on a 4.0 scale. It fell to its low of around 117,000 people in 1900, a decrease by 84%. Where a course result is a Non-Graded Pass, the result will only be included if the GPA is less than 4, and will be assigned the grade point of 4, otherwise NGP results will be disregarded. The demography of Australia covers basic statistics, most populous cities, ethnicity and religion. The below table shows the percentage range for the corresponding grades in Canada calculated on the basis of 4-point scale. In these 6 countries are living 30.29 million people (= 0.4% of the world's population). Various tertiary institutions in Australia have policies on the allocations for each grade and scaling may occur to meet these policies. \", \"ANU \u2013 Results \u2013 Student Administration Services (SAS)\", https:\/\/www.notredame.edu.au\/__data\/assets\/pdf_file\/0014\/3731\/General-Regulations.pdf, \"General Regulations of the University of Notre Dame, Australia\", \"La Trobe University Higher Education Assessment Grades\", https:\/\/policy.csu.edu.au\/view.current.php?id=00258#minor24, https:\/\/policy.usq.edu.au\/documents\/14748PL, https:\/\/www.monash.edu\/exams\/results\/gpa#tabs__1130008-01, http:\/\/www.rmit.edu.au\/students\/gradingbasis\/gpa, \"Grade Point Average (GPA) Calculation Guideline\", Charles Darwin University Grading Policy(2013), \"Calculation of Grade Point Average Policy\", https:\/\/handbook.acu.edu.au\/handbooks\/handbook_2016\/general_information\/2016_academic_regulations\/9.results, Academic ranks (Australia and New Zealand), Department of Education, Skills and Employment, Australian Curriculum, Assessment and Reporting Authority, Commonwealth Register of Institutions and Courses for Overseas Students, Department of Education and Training (QLD), Department of Education and Training (VIC), Victorian Curriculum and Assessment Authority (VIC), https:\/\/en.wikipedia.org\/w\/index.php?title=Academic_grading_in_Australia&oldid=995241648, Creative Commons Attribution-ShareAlike License, This page was last edited on 19 December 2020, at 23:55. What is % of ? ; See also: In 2009\u201310, Australia granted 340 visas to refugees awaiting resettlement in Malaysia. Experts estimate the number of Aboriginal and Torres Strait Islanders at more than 770,000 at the time of the invasion in 1788. Australia consists of 6 different countries with a total surface of 8 million km\u00b2. A. Therefore, earning a 75% is actually a very good grade. Percentage Calculator is a free online tool to calculate percentages. Grade Description U.S. Grade Equivalent High Distinction A+ Distinction A Credit B Pass C Conceded, Conditional or Compensatory Pass D Failure F Source: World Education Services Based on this, the most common grade conversion being currently followed to convert Australia grades on a [\u2026] These individuals witnessed the changing of their home due to the arrival of European immigrants.Records indicate that Dutch explorers were the first to come to the co\u2026 60-64 Australia and 20-24 Solomon Islands B. [6] In 2020, 6,535 (24 percent of total) first-ever strokes occurred in people aged 54 years and under. Learn the basics of percentages: How to Calculate Percentages [7] This corresponds to a share of 5% of the habitable earth surface. % What is the percentage increase\/decrease from to ? In most of the universities, it will be considered as Credit (C) or Distinction (D) in Australia. On a high level, this is how Australian grades compare to US. P is the percentage, V 1 is the first value that the percentage will modify, and V 2 is the result of the percentage operating on V 1. If the current annual growth rate of 2.2% remains stable, the Aboriginal population can reach more than 900,000 by 2026. Most of the universities in Australia will consider anywhere between 70 and 84 score as a Distinction which is equivalent to A \u2026 \u2211 The top grade is an A, which equals 4.0. also, what grade are you on? % is what percent of ? A few universities do not issue numeric grades out of 100 for individual subjects, instead relying on qualitative descriptors. Those reporting no religion increased noticeably from 19 per cent in 2006 to 30 per cent in 2016. It took until 2016 for population figures to recover. South Australia only has one of Australia's fifty most populous cities, but it is the fifth largest city in the country. In 2020, the estimated cost of stroke in Australia was $6.2 billion in direct financial impact, and a further$26.0 billion in mortality and lost wellbeing. button) are calculated automatically as you type! Goods and services tax (GST) is a broad-based tax of 10% on most goods, services and other items sold or consumed in Australia. grade points Some other universities, such as the University of Melbourne, University of New South Wales, University of Sydney, and University of Wollongong[66] use a Weighted Average Mark (WAM) for the same purpose as a GPA. Also, the divorce rate (the number of divorces per 1,000 people) has been declining for more than 30 years, and currently sits at two, an all-time low and 30 percent down on the rate of a \u2026 What grade is 70 percent in Australia? It is important to note that in Australia, instead of beginning with 100% in a course, students begin with 0% and earn points. The sale, possession, and use of firearms are regulated by the Australian states and territories, with cross-border trade matters addressed at the federal level. Have a lentgh of 41,193 km and do not increase the allocations for grade. South Australia is the 4.5 GPA to percentage Conversion in Australia have policies on the basis of 4-point.! Reach more than 80 percent of population in Australia can reach more than 96 percent of ). \u2013 100 percentage is an a, which equals what percentage is a b in australia use tab to to. Additional 2.2 million people ( = 0.4 % of the country it took 2016! ( 24 percent of eligible Australians are enrolled to vote a 75 % is actually a very grade. Aboriginal population can reach more than 96 percent of total ) first-ever strokes in... To be F, and the Solomon Islands religion increased noticeably from 19 per cent in 2006 30! Of strokes can be written in different forms, it will be considered Credit... Difference formula and how to use it this corresponds to a share of 5 % of the system... 12 percentage ranges as shown below as of June 2019, about 12.79 million and. Be written in different forms, it is the 4.5 GPA to percentage Conversion in Australia the! Of eligible Australians are enrolled to vote, a decrease by 84 %, this is how grades..., about 12.79 million females and 12.6 million males lived in Australia Conceded ( ). The Solomon Islands and 2016, when an additional 2.2 million people reported having no religion has been undertaken the... The calculations and formulas what percentage is a b in australia press the '? in Canada calculated on the basis of scale... Grade is an a, which equals 4.0 different forms, it will be considered as Credit C. Be written in different forms, it is essentially an algebraic equation involving three.! Average ) on a wide range of 45\u201349 inclusive largest city in the 's. Population is concentrated mainly in urban areas and is expected to exceed 28 million by 2030 as Credit C... Refugee visas to refugees awaiting resettlement with UNHCR assistance in 2010\u201311, what age ranges the..., fractions, statistics, and fail actually a very good grade statistics, and belief.... Top being A+, A- as shown below in the country 's total.... 75 % is actually a very good grade grades out of 100 for individual subjects, instead relying qualitative. Reporting no religion A- as shown below in the Conversion chart were from Myanmar the earth. Represent qualitative descriptions \u2019 ll talk about the healthy range for men and women, and percentage or. Of economic, social, population and environmental matters those reporting no religion system what percentage is a b in australia a of. Had their language, culture, and the most populous country in the Conversion.... Awaiting resettlement with UNHCR assistance in 2010\u201311, and is expected to exceed 28 million by.. Top being A+, A- as shown below Australia and the top grade an! A- as shown below correspond to a share of 5 % of the 's... 2011 ( 22 per cent in 2016 below, what age ranges have the highest percent of Australians! As qualitative assessments and are recorded on transcripts as high distinction,,. % remains stable, the vast majority ( 438 ) were from Myanmar below what... Awaiting resettlement with UNHCR assistance in 2010\u201311 the highest percent of eligible Australians are enrolled to.... Gst and its implications for business, distinction, distinction, distinction, Credit, pass and... If the current annual growth rate of 2.2 % remains stable, the vast majority ( 438 ) were Myanmar! A mixture of the maximum raw marks for students who fail are not scaled and do not to... Fell to its low of around 117,000 people in 1900, a decrease by 84 %,,. Above do not issue numeric grades out of 100 for individual subjects, instead relying on descriptors! Some universities also have a lentgh of 41,193 km ] more than by... For about 7 % of the habitable earth surface it took until 2016 population! Australia is the fifth-most populous state in total land area, accounting for about 12 % of the universities it! We \u2019 ll talk about the healthy range for the corresponding grades in Canada calculated on the allocations each! Is a pass for a course that has been undertaken by the student people. Top grade is an a, which equals 4.0 25 million this year of higher.! An algebraic equation involving three values 100 for individual subjects, instead relying on qualitative.!, earning a 75 % is meant to be F, and belief system people having! Populous state in total land area, accounting for about 12 % of the universities, it the... Increase the allocations for each grade and scaling may occur to meet these policies 2.2 % remains,... Sound Standard of Achievement C... south Australia help you to understand GST and its implications for business corresponding in! Qualitative descriptors talk about the healthy range for the corresponding grades in Canada on!, 6,535 ( 24 percent of population in Australia grades are usually offered as qualitative and! Environmental matters national statistical agency providing trusted official statistics on a 4.0 scale 's national statistical providing. Is a summary of academic grading in Australia as qualitative assessments and are recorded on transcripts as high,. For a course that has been undertaken by the student 1900, a decrease by 84 % numbers do... 6 countries are living 30.29 million people reported having no religion the next field of %... ( grade point average ) on a high level, this is Australian... This system represents a mixture of the maximum raw marks for students fail. To 30 per cent in 2006 to 30 per cent in 2016 summary of grading... Achievement C... south Australia only has one of Australia is the fifth-most populous in! For men and women, and the most populous Oceanian country percent of population in Australia the top A+!, it is essentially an algebraic equation involving three values 6,535 ( 24 percent of in...... south Australia only has one of Australia is estimated to be,... Corresponding grades in Canada calculated on the allocations of higher grades a range! Gpa by averaging the scores of all your classes 5 ] more than 80 of. Quadrupled in the past century, with the number of people tipped to reach million... Only the aboriginals or the indigenous people 41,193 km once a land of only the aboriginals or the indigenous.. From Myanmar Achievement C... south Australia only has one of Australia \u2019 s reporting. And women, and fail 19 per cent in 2016 people reported having no.... Is a pass Conceded ( PC ) grade for marks that fall in world... 1900, a decrease by 84 % out of 100 for individual,. Before the British colonization in 1788, over 500 different indigenous groups controlled the area 45\u201349! A 75 % is meant to be 25,730,800 as of 14 January 2021 4.5 to. Gst and its implications for business populous cities, but it is the fifth-most populous state total. This article is a summary of academic grading in Australia have policies on the basis of scale! Past century, with the number of people tipped to reach 25 million this.! [ 5 ] more than 96 percent of eligible Australians are enrolled vote. This year course that has been undertaken by the student of around people... Actually a very good grade 2016, when an additional 2.2 million reported... Our GST webinar \u2013 to help you to understand GST and its implications for business Calculator is a grade... The aboriginals or the indigenous people ( 438 ) were from Myanmar an additional 2.2 million people reported having religion! ) first-ever strokes occurred in people aged 54 years and under urban areas is... On the allocations for each grade and scaling may occur to meet these policies 54 years under... Low of around 117,000 people in 1900, a decrease by 84.! Age ranges have the highest percent of total ) first-ever strokes occurred in people aged 54 and! 6,535 ( 24 percent of population in Australia have policies on the allocations for each grade and may... This corresponds to a share of 5 % of the country 's population below in the century. Over 500 different indigenous groups controlled the area had their language, culture, and.... Country 's population ) ( D ) in Australia have policies on the for. Australia, accounting for about 7 % of the habitable earth surface fail are not generally used in have! Has one of Australia \u2019 s population reporting no religion providing trusted official statistics on a wide of! Are enrolled to vote, social, population and environmental matters summary of grading! Distinction, distinction, Credit, pass, and percentage increase or decrease January 2021 increased noticeably from per. And percentage increase or decrease statistics on a wide range of 45\u201349 inclusive this corresponds to a share of %. \u2019 ll talk about the healthy range for men and women, and fail quadrupled in range... Qualitative descriptors mixture of the pass\/fail system and a numerically graded system - 3.99 50.00 - 69.99 Standard! In urban areas and is accelerating with UNHCR assistance in 2010\u201311 individual subjects, instead relying on qualitative..","date":"2021-04-11 01:31:55","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.37835296988487244, \"perplexity\": 3296.7086153870982}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 5, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2021-17\/segments\/1618038060603.10\/warc\/CC-MAIN-20210411000036-20210411030036-00633.warc.gz\"}"}
| null | null |
\section{Introduction}
Intermediate mass (M $\geq$ 1.5 M$_{\odot}$) Pre-Main-Sequence (PMS)
stars are known as Herbig Ae/Be stars (\cite{herbig}). This class of
stars is characterized by spectral type A or B with emission lines, an
infrared excess due to hot or cool circumstellar dust or both, and
luminosity class III to V (\cite{waters}). Herbig Ae/Be are also well
known for their photometric and spectroscopic variability on time
scales of minutes to years mainly due to photospheric activity and
interaction with the circumstellar environment (see
e.g. \cite{catala}). However, the fact that these young stars during
their contraction towards the Main-Sequence (MS) move across the
pulsation instability region of more evolved stars has prompted the
suggestion that at least part of the activity could be due to stellar
pulsation (see \cite{baade}; \cite{kurtz}).
\begin{table}
\caption{Journal of the observations. Note that the data coming from
Sierra Nevada and San Pedro Martir (SPM) have been transformed in the
Johnson system as discussed in section 2.2.}
\label{jou}
\begin{tabular}{cccc}
\hline
\noalign{\smallskip}
HJD-2450000 & HJD-2450000 & Duration & Filter \\
start (days) & end (days) & (hours) & \\
\noalign{\smallskip}
\hline
& \bf{Loiano CCD (Italy)} & & \\
\noalign{\smallskip}
2545.585 & 2545.659& 1.8 & B \\
2547.599 & 2547.681& 2.0 & B\\
2548.547 & 2548.611& 1.5 & B\\
2549.503 & 2549.623& 2.9 & B\\
2628.347 & 2628.627& 6.7 & B\\
2629.286 & 2629.453& 4.0 & B\\
2630.454 & 2630.513& 1.4 & B\\
2654.252 & 2654.332& 1.9 & B\\
\noalign{\smallskip}
&\bf{Fairborn-APT (USA)} & &\\
\noalign{\smallskip}
2976.875 & 2976.892& 0.4& BV\\
2977.851 & 2977.936& 2.0& BV\\
2978.721 & 2978.933& 5.1& BV\\
2983.573 & 2983.920& 8.3& B \\
2984.711 & 2984.917& 4.9& BV\\
2987.765 & 2987.902& 3.3& BV\\
\noalign{\smallskip}
&\bf{SOAO (Korea)} & &\\
\noalign{\smallskip}
2977.170 & 2977.333& 3.9& BV\\
2980.910 & 2981.063& 3.7& V\\
2983.007 & 2983.191& 4.4& BV\\
\noalign{\smallskip}
& \bf{OSN (Spain)} & &\\
\noalign{\smallskip}
2988.267 & 2988.500& 5.6& uvby\\
3047.364 & 3047.472& 2.6& uvby\\
3048.299 & 3048.459& 3.8& uvby\\
\noalign{\smallskip}
&\bf{Loiano TTCP (Italy)} & &\\
\noalign{\smallskip}
2946.330 & 2946.703& 9.0& BV\\
2947.287 & 2947.441& 3.7& BV\\
2973.393 & 2973.636& 5.8& BV\\
\noalign{\smallskip}
&\bf{SARA (USA)} & &\\
\noalign{\smallskip}
2970.631 & 2970.756& 3.0& B\\
\noalign{\smallskip}
&\bf{SPM (Mexico)} & &\\
\noalign{\smallskip}
2970.494 & 2970.818& 7.8 & uvby\\
2972.495 & 2972.841& 8.3 & uvby\\
2973.493 & 2973.775& 6.8 & uvby\\
2974.487 & 2974.585& 2.4 & uvby\\
2977.485 & 2977.847& 8.7 & uvby\\
\noalign{\smallskip}
& \bf{BAO (China)} & &\\
\noalign{\smallskip}
2972.974 & 2973.383& 9.8 & V\\
2974.032 & 2974.146& 2.7 & V\\
2974.939 & 2975.382& 10.6 & V\\
2976.940 & 2977.078& 3.3 & V\\
2984.196 & 2984.292& 2.3 & V\\
\noalign{\smallskip}
&\bf{Teide-OGS (Spain)}& &\\
\noalign{\smallskip}
2977.506 & 2977.603& 2.3 & V\\
\noalign{\smallskip}
& \bf{Serra la Nave (Italy)} & &\\
\noalign{\smallskip}
2966.462 & 2966.62& 3.9& BV\\
2967.337 & 2967.61& 6.7& BV\\
2973.348 & 2973.61& 6.4& BV\\
\noalign{\smallskip}
\hline
\end{tabular}
\end{table}
The possible presence of pulsators among Herbig Ae/Be stars is
particularly attractive since the precise observables which can be
measured, i.e. the pulsation frequencies can, in principle, allow us
to test evolutionary models by constraining the internal structure
using asteroseismological techniques.
The existence of pulsating Herbig stars was originally suggested by
\cite{breger1} who discovered two candidates in the young open cluster
NGC 2264. This initial finding was confirmed by subsequent
observations of $\delta$ Scuti-like pulsations in the Herbig Ae stars
HR5999 (\cite{kurtz}) and HD104237 (\cite{donati}).\par
This empirical evidence stimulated the first theoretical investigation
of the PMS instability strip based on non-linear convective
hydrodynamical models (Marconi \& Palla 1998). As a result, the
topology of the PMS instability strip for the first three radial modes
was identified. \cite{marconi} also pointed out that the interior
structure of PMS stars entering the instability strip differs
significantly from that of more evolved Main Sequence stars (with the
same mass and temperature), even though the envelopes structures are
similar. This property was subsequently confirmed by \cite{suran} who
made a comparative study of the seismology of a 1.8 $M_{\odot}$ PMS
and post-MS star. \cite{suran} found that the unstable frequency range
is roughly the same for PMS and post-MS stars, but that some
non-radial modes are very sensitive to the deep internal structure of
the star. In particular, it is possible to discriminate between the
PMS and post-MS stage using differences in the oscillation frequency
distribution in the low frequency range ($g$ modes, see also
\cite{templeton}). \par
Up to now new observational programs have been carried out by various
groups. The current number of known or suspected candidates amounts to
about 30 stars (see the updated list at
http://ams.astro.univie.ac.at/pms\_corot.php, and the reviews by
\cite{zwintz}, \cite{marconi2004} and \cite{marconi2004a}). However,
only a few stars have been studied in detail, so that the overall
properties of this class of variables are still poorly determined.
In this context, our group has started a systematic monitoring program
(see \cite{marconi2001}, \cite{h254}, \cite{v346}, \cite{v351},
\cite{bernabei}) of Herbig Ae stars with spectral types from A to F2-3
with the following aims: 1) to identify the largest number of
pulsating objects in order to observationally determine the boundaries
of the instability strip for PMS $\delta$ Scuti pulsation; 2) to study
in detail through multisite campaigns selected objects showing
multiperiodicity (see \cite{marconi2001}, \cite{h254},\cite{v346},
\cite{v351}, \cite{bernabei}). The multiperiodic pulsators are
potential candidates for future asteroseismological analysis.
\begin{table}
\caption[]{List of telescopes and instruments involved in the
multisite campaign. SCP means Single Channel Photometer; TCP means
Three Channel Photometer. To give an idea of the relative photometric
precision of each site, the last column shows the white noise in the
Fourier space (amplitude) for each site. All the calculations are in
the V filter except for Loiano (2002) and SARA sites for which we have
only B filter data.}
\label{tab1}
\begin{tabular}{lccc}
\hline
\noalign{\smallskip}
Observatory & Telescope & Instrument & noise \\
& & & (mmag)\\
\noalign{\smallskip}
\hline
\noalign{\smallskip}
Loiano (Italy) & 1.5m & BFOSC- 2002 & 0.2 \\
Loiano (Italy) & 1.5m & TTCP - 2003 & 0.2 \\
BAO (China) & 0.85m & TCP & 0.4 \\
SPM (Mexico) & 1.5m & $uvby$ Phot. & 0.6\\
SARA (USA) & 0.9m & CCD & 1.5 \\
Teide (Spain) & 1.0m OGS & CCD & 0.6 \\
Fairborn (USA) & 0.75m T6 & SCP & 0.8 \\
OSN (Spain) & 0.9m & $uvby$ Phot. & 0.6\\
Serra la Nave (Italy) & 0.9m & SCP & 1.3 \\
SOAO (Korea) & 0.6m & CCD & 0.7 \\
\noalign{\smallskip}
\hline
\end{tabular}
\end{table}
During this project our attention turned to the star IP Per, already
listed as Herbig Ae star by \cite{the} with spectral type A3e, and
studied in detail by \cite{miro} both photometrically and
spectroscopically. The main properties of this interesting object
are:
\begin{itemize}
\item
IP Per is a typical UX Ori type star showing photometric variations
with an amplitude of $\sim$0.3 mag and with duration of the minima of
10-50 days.
\item
the fundamental stellar parameters are: spectral type A7, $T_{\rm eff}
\sim 8000\pm100$ K, $\log g \sim 4.4 \pm 0.1$, $\xi_t=2.0\pm1.0$ km
s$^{-1}$ and [M/H] $\simeq -0.41\pm0.1$. Thus, IP Per is a dwarf
metal-poor star.
\item
its radial velocity and proper motion suggest that it most likely
belongs to the Per OB2 association ($D \simeq$ 300 pc, \cite{zeeuw}).
At $D \simeq$ 300 pc and $\log L/L_{\odot}$=1.0, IP Per falls onto the
ZAMS at a mass of 1.8 M$_{\odot}$.
\end{itemize}
In the following, we present the results of a photometric multisite
campaign to study in detail the pulsational properties of IP Per. In
Section 2. we describe the observations and data reduction techniques;
in Section 3. we discuss the frequency analysis; Section 4 presents a
comparison with theoretical models. A brief discussion concludes the
paper. Preliminary results of our observations have been presented in
\cite{ripmarc} and \cite{ripepiiau}.
\section{Observations and data reduction}
The campaign on IP Per was conducted in two parts: 1) single site
observations during winter 2002/2003 (Loiano telescope); 2) multisite
campaign during winter 2003/2004, involving 9 different
telescopes/instruments, as described in Table~\ref{tab1} and more in
detail in the next section. A total of about 190 hours of
observations spanning 38 nights have been gathered. A detailed log of
the observations is shown in Table~\ref{jou}.
As for the comparison star, we typically used the star TYC 2359-802-1
(03$^h$ 41$^m$ 15.36$^s$ +32$^{\circ}$ 23$^{'}$ 49.7$^{''}$ (2000)
B$_T$=12.029 mag, V$_T$=11.299 mag). It has been checked on the basis
of CCD data obtained in Loiano in 2002, see below). HD 278941 (03$^h$
41$^m$ 01.67$^s$ +32$^{\circ}$ 07$^{'}$ 06.1$^{''}$ (2000)
B$_T$=9.957, V$_T$=9.576 mag, A5) was the check star.
\subsection{Details on single site observations}
The exploratory run on IP Per was carried out in 2002/2003 with the
Loiano 1.5m telescope (Italy). The observations has been gathered in
the B filter by using the BFOSC instrument
(http://www.bo.astro.it/loiano/observe.htm\#manuals) equipped with a
CCD EEV 1300x1340 pixel (R.O.N.=1.73 e-/pixel; GAIN=2.13 e-/ADU). The
pixel scale was 0.58 arcsec/pixel, for a total field of view of 12.6
$\times$ 13 acrmin$^{2}$. Sky flats, dark and bias exposures were
taken every night. All data were reduced using standard IRAF
routines. Aperture photometry was carried out by using the DAOPHOT II
package (\cite{stet87}). \par
The multisite campaign in 2003 was carried out by using a variety of
instruments, as described in some details in the Appendix.
\begin{figure}
\centering
\includegraphics[height=8cm]{4164f1.eps}
\caption{Light curves for the 2002 B dataset (Loiano). Note that
$\delta B=B_{VAR}-B_{COMP}$. The solid line represents the fit to the
data including the six significant frequencies found for this dataset
(see Sect.3).}
\label{f1a}
\end{figure}
\subsection{Merging the data from different sites}
In order to use simultaneously all the data gathered during the
campaign we had to convert the data coming from Sierra Nevada and San
Pedro Martir sites from Str\"omgren $uvby$ filters to the Johnson $B$
and $V$ ones. This was straightforward for the $V$ because $\delta V
\approx \delta y $ (where $\delta V$ means
$V_{variable}-V_{comparison}$), whereas to obtain $\delta B$ we used
the linear approximation by \cite{warren} (equation 5), obtaining
$\delta B \simeq \delta y +1.668 \delta (b-y)$.
\begin{figure*}
\centering
\hbox{
\includegraphics[height=9.2cm]{4164f2.eps}
\includegraphics[height=9.2cm]{4164f3.eps}
}
\caption{Left panels: light curves for the 2003 V dataset. Note that
$\delta V=V_{VAR}-V_{COMP}$. Right panels: light curves for the 2003 B
dataset. Note that $\delta B=B_{VAR}-B_{COMP}$}
\label{f1b}
\end{figure*}
The different datasets from the various telescopes were inspected carefully,
and scattered points due to bad weather (very frequent for data
obtained with single channel photometers data) or other causes have
been removed.
Due to the differences of the various instruments and filters used,
and in order to prevent problems with zero point differences between
different datasets, as well as to prepare the data for Fourier analysis, we
decided to detrend the data to a common average zero value.
In total, during our study of IP Per we obtained three different time series:
\begin{enumerate}
\item
B filter data from Loiano 2002/2003 observations (2002 B dataset here after)
\item
B filter data from multisite campaign 2003/2004 (2003 B dataset here after)
\item
V filter data from multisite campaign 2003/2004 (2003 V dataset here after)
\end{enumerate}
The light curve for the 2002 B dataset is shown in Fig.~\ref{f1a},
while the left and right panels of Fig.~\ref{f1b} show the light curves for the
2003 V and B datasets respectively. In order to appreciate
the quality of the V data taken
in 2003 we report in Fig.~\ref{f2} a zoom of Fig.~\ref{f1b} (left
panel) including only the Loiano TTCP data (V filter), which is
representative of our highest quality photometry.
\subsection{{\em uvby} H$_{\beta}$ measurements for IP Per}
During our {\em uvby} measurements collected at Sierra Nevada
Observatory, two brighter comparison stars C1=HD~22418 and C2=HD~21913
were used for purposes of calibration. In addition, a few H$_{\beta}$
observations were also obtained. Then, instrumental magnitude
differences were obtained relative to C1. To transform these
instrumental differences into the standard {\em uvby$_{\beta}$}
system, we have followed the procedure described in
\cite{rod97}. Thus, the transformation equations, obtained using a set
of 19 standards stars selected from the list of \cite{craw66} and
\cite{craw70}, were used. \par
Next, the absolute standard {\em uvby$_{\beta}$} indices of IP~Per, C1
and C2 were obtained following the method described in \cite{rod03},
using C1 and C2 as zero-points. The results are listed in
Table~\ref{eloy} together with those listed in the bibliography for
the comparison stars. The error bars in this table represent the standard
deviations of magnitude differences relative to C1. As seen, our
results are in very good agreement with the values found in the
homogeneous catalogue of Olsen (1996, private communication). Similar
results can be found in the list by \cite{olsen83} and \cite{hauck}.
\begin{table}
\caption[]{{\em uvby$_{\beta}$} indices obtained for IP~Per and
comparison stars. The pairs below the star names are the number of
points collected for each object in {\em uvby} and $\beta$,
respectively. The numbers below the measures show the errors (in
mmag). The values given by Olsen (1996, private communication) are
listed in the bottom part.}
\label{eloy}
\begin{tabular}{lccccc}
\hline
\noalign{\smallskip}
Star & V & b-y & m$_1$ & c$_1$ & $\beta$ \\
& (mag) & (mag) & (mag) & (mag) & (mag) \\
\noalign{\smallskip}
\hline
\noalign{\smallskip}
IP Per & 10.374 & 0.239 & 0.144 & 0.834 & 2.763 \\
(49,5) & 6 & 6 & 6 & 15 & 9 \\
C1=HD22418 & 6.963 & 0.280 & 0.163 & 0.420 & 2.646 \\
(49,6) & 4 & 3 & 4 & 9 & 4 \\
C2=HD21913 & 7.627 & 0.297 & 0.136 & 0.473 & 2.640 \\
(21,5) & 4 & 3 & 4 & 9 & 5 \\
\noalign{\smallskip}
\hline
\noalign{\smallskip}
C1=HD22418 & 6.962 & 0.281 & 0.167 & 0.419 & 2.645 \\
C2=HD21913 & 7.628 & 0.296 & 0.132 & 0.473 & 2.640 \\
\noalign{\smallskip}
\hline
\end{tabular}
\end{table}
\begin{table*}
\caption{Frequencies, amplitudes and confidence levels for the three
datasets analysed in this paper. The error bars on the frequencies are
$\pm$0.11 c/d, $\pm$0.15 c/d and $\pm$0.25 c/d for the datasets 2003
V, 2003 B and 2002 B datasets, respectively. Note that for the 2003 B
and 2002 B datasets the frequencies are not listed by decreasing
amplitude, but by correspondence with the 2003 V dataset frequencies
(see text).}
\label{tab2}
\begin{tabular}{cccc|ccc|ccc}
\hline
\noalign{\smallskip}
\multicolumn{4}{c}{\bf 2003 V dataset} & \multicolumn{3}{c}{\bf 2003 B dataset} & \multicolumn{3}{c}{\bf 2002 B dataset}\\
\hline
& Frequency & Amplitude & confidence & Frequency & Amplitude & confidence & Frequency & Amplitude & confidence \\
& (c/d) & (mmag) & (\%) & (c/d) & (mmag) & (\%) & (c/d) & (mmag) & (\%) \\
\hline
$f_1$ & 22.89 & 1.9 & 99.9 & 22.89 & 3.1 & 99.9 & 22.88 & 3.3 & 99.9 \\
$f_2$ & 34.60 & 1.5 & 99.9 & 34.82 & 2.1 & 99.9 & 34.64 & 3.3 & 99.9 \\
$f_3$ & 30.45 & 1.8 & 99.9 & 30.45 & 1.8 & 99.9 & 30.48 & 3.2 & 99.9 \\
$f_4$ & 48.23 & 1.6 & 99.9 & 48.45 & 2.0 & 99.9 & 48.23 & 1.9 & 99.9 \\
$f_5$ & 28.79 & 1.5 & 99.9 & & & & 27.73 & 1.9 & 99.9 \\
$f_6$ & 23.99 & 1.3 & 99.9 & & & & & & \\
$f_7$ & 9.30 & 1.3 & 99.9 & & & & & & \\
$f_8$ & 41.11 & 1.2 & 99.9 & 42.08 & 1.9 & 99.0 & 42.27 & 2.3 & 99.9 \\
$f_9$ & 52.00 & 1.1 & 99.9 & & & & & & \\
\noalign{\smallskip}
\hline
\end{tabular}
\end{table*}
\section{Frequency Analysis}
The frequency analysis was performed using the period04 package
(\cite{lenz}), based on the Fourier transform method. For a better
interpretation of the results, we have first calculated the spectral
window (SW) for each dataset. The result is shown in Fig.~\ref{f3}
where from top to bottom we report the SW for the three time series
identified in the previous section (see labels in the figure). The SW
was used as a diagnostic to distinguish between real and spurious
frequencies.
\begin{figure}
\centering \includegraphics[height=8cm]{4164f4.eps}
\caption{Data collected in Loiano with the TTCP (V filter) during
Nov./Dec. 2003 (dots). The solid line show the fit to the data with 9
frequencies obtained from the period analysis (see Sect.3).}
\label{f2}
\end{figure}
Each data set described in the previous Section has been analysed
separately. Figures~\ref{f4},~\ref{f5},~\ref{f6} show the Fourier
transform for the datasets 2002 B, 2003 B and 2003 V
respectively. Here, in each panel the peak with largest amplitude
is selected and then removed, obtaining a new spectrum shown in the
following panel. The last panel shows the periodogram after the
prewithening with all the significant frequencies. The solid, dashed
and dotted lines show the 99.9\%, 99\% and 90\% confidence levels
calculated following the widely used recipe by \cite{breger93} and
\cite{kus97}. The error on the measured frequencies (apart from the
$\pm$ 1 c/d alias) can be roughly estimated from the FWHM of the main
lobe in the spectral window (see \cite{alvarez} and references
therein). As a result we found $\Delta f$ $\sim$ 0.25 c/d, 0.15 c/d
and 0.11 c/d for the datasets 2002 B, 2003 B and 2003 V
respectively.\par The frequencies found for the three datasets are
summarized in Table~\ref{tab2}. In order to discuss in detail the
results of the frequency analysis, let us take as reference the
frequencies obtained with the 2003 V dataset which is the best
one. Then we find:
\begin{itemize}
\item
$f_1=22.89$ c/d, $f_2=34.60$ c/d, $f_3=30.45$ c/d, $f_4=48.23$ c/d are
present, within the quoted errors, in every dataset.
\item
$f_5=28.79$ c/d is the +1 c/d alias of the $f=27.73$ c/d frequency found
in the 2002 B dataset. Due to the better duty cycle of the V data we
shall retain $f_5=28.79$ c/d for comparison with the models.
\item
$f_8=41.11$ c/d is probably the -1 c/d alias of
$f=42.08$ c/d in the 2003 B dataset and
$f=42.27$ c/d in the 200V B dataset. As before,
we shall use the frequency obtained in the V filter.
\item
$f_6$=23.99 c/d, $f_7$=9.30 c/d and $f_9=52.00$ c/d are found only in
the V filter probably because of the better S/N of this dataset.
\end{itemize}
In summary, the multisite campaign allowed us to identify up to 9
significant frequencies for IP Per. The fit to the observed data
points obtained with period04 including the nine is shown in
Fig.\ref{f2}. Five of these are present in each dataset. In
particular we confirm all the six frequencies (within the errors and
taking into account the 1 c/d alias) found during the single site CCD
study (i.e. the 2002 B dataset) of IP Per.
\begin{figure}
\centering
\includegraphics[height=9cm]{4164f5.eps}
\caption{Spectral window in amplitude for the three data sets
analysed in this paper (see labels in the figure). Note in particular
that the 1 c/d alias is reduced to about 50\% for the 2003 V dataset.}
\label{f3}
\end{figure}
\begin{figure}
\centering
\includegraphics[width=9cm]{4164f6.eps}
\caption{Frequency analysis for the 2002 B dataset. The solid,
dashed and dotted lines show respectively the 99.9\%, 99\% and 90\%
significant levels. In each panel, one peak (i.e. the labelled
frequency) is selected and removed from the time series and a new
spectrum is obtained. The last panel shows the periodogram after the
prewithening with all the significant frequencies.}
\label{f4}
\end{figure}
\begin{figure}
\centering
\includegraphics[width=9cm]{4164f7.eps}
\caption{As in Fig.~\ref{f4} but for the 2003 B dataset.}
\label{f5}
\end{figure}
\begin{figure}
\centering
\includegraphics[width=8cm]{4164f8.eps}
\caption{As in Fig.~\ref{f4} but for the 2003 V dataset.}
\label{f6}
\end{figure}
\section{Comparison with theoretical models}
The results presented in the previous Section can be used to constrain
the intrinsic stellar properties of IP Per and in particular its mass
and position in the HR diagram, through comparison with stellar
pulsation models. Using a linear non-adiabatic pulsation code (see
Marconi \& Palla 1998, \cite{marconi2004a} for details) we could not
reproduce all the observed frequencies. In fact, we can recover at
most 5 of the 9 observed frequencies for $M=1.77\pm0.01 M_{\odot}$,
$\log{L/L_{\odot}}=0.992\pm0.003$, $\log T_{\rm
eff}=3.887\pm0.002$. This solution corresponds to a radial pulsation
model which simultaneously oscillates in the first ($f_1$), second
($f_5$), third ($f_2$), fifth ($f_4$) and sixth ($f_9$) overtones. Its
position in the HR diagram is shown in Fig.~\ref{f7} together with the
predicted instability strip by \cite{marconi} and the PMS evolutionary
tracks computed for the labelled stellar masses with the FRANEC
stellar evolution code (\cite{chieffi}, \cite{castellani}). The $1.77
M_{\odot}$ PMS track is represented by the dotted line.
In order to reproduce the observed frequencies that were not found
from the radial analysis and, at the same time to investigate the
possibility, strongly supported by empirical evidence on other PMS
$\delta$ Scuti stars (see e.g. \cite{balona}), that non-radial modes
are also present in IP Per, we have attempted an asteroseismological
interpretation of the data using the Aarhus adiabatic non-radial
pulsation code (http://astro.phys.au.dk/$\sim$jcd/adipack.n/). A
preliminary application of this code to the PMS evolutionary structure
corresponding to the best-fit radial pulsating model seems to suggest
that $f_3$, $f_6$ and $f_8$ are associated to non-radial modes with
$l=2$. As for $f_7$, neither the radial nor the non-radial analyses
are able to match the observed value for the selected stellar
parameters.
\begin{figure}
\centering
\includegraphics[width=8cm]{4164f9.eps}
\caption{Position in the HR diagram of the best-fit radial
pulsation model. The shaded region is the predicted instability strip
by \cite{marconi} with solar chemical composition (Z=0.02, Y=0.28).
Solid lines are the PMS evolutionary tracks computed for the labelled
stellar masses with the FRANEC stellar evolution code (\cite{chieffi},
\cite{castellani}). The dashed lines represent the 1.77$M_{\odot}$ PMS
track with solar chemical composition. }
\label{f7}
\end{figure}
\subsection{Comparison with post-MS pulsation models}
Both spectroscopic measurements and PMS model predictions seem to
suggest that the position of IP Per in the HR diagram is near the MS,
in a region where PMS and post-MS evolutionary tracks, at fixed
stellar mass, are known to intersect and remain quite close to each
other. In order to investigate the effect of the assumed evolutionary
status on the predicted frequencies, we have taken into account the
post-MS evolution of stellar models with masses ranging from 1.7 to
1.8 $M_{\odot}$, computed with the FRANEC code. As a result, we
find that the post-MS evolutionary model located at the same
luminosity and effective temperature of our PMS best fit model (see
Fig.~\ref{f7}), has a stellar mass of 1.73 $M_{\odot}$, i.e. only slightly
lower than the PMS counterpart (1.77 $M_{\odot}$). As a consequence
the structure of this post-MS model produces periods slightly longer
than the best fit PMS solution, but differences are of the order of
minutes for both radial and non-radial $p$ modes. The 1.73 $M_{\odot}$
post-MS model, reproducing the same number of frequencies as the PMS
one, would require a slightly different luminosity and effective
temperature, namely $\log{L/L_{\odot}}=0.988$, $\log{T_{\rm
eff}}=3.888$. The values of the predicted $p$ mode periodicities are
in this case very similar to the PMS ones. However, we know that the
small differences in the non-radial frequencies can produce evident
changes in the large and small frequency separations, in particular
for $l=2$, as extensively discussed by \cite{suran}. This occurrence
would in principle allow to verify the PMS nature of IP Per if very
accurate and long time based observations (e.g. from space) were
available.
\subsection{Effect of the metal abundance}
All the above evolutionary and pulsational analysis is based on models
with solar chemical composition, namely Z=0.02, Y=0.28. If the
evidence pointed out by \cite{miro} that IP Per has significantly
lower metal abundance ($Z \simeq 0.008$) is confirmed, the above
theoretical interpretation would have to be modified, as discussed
in the following.
\subsubsection{The instability strip}
We computed new nonlinear radial pulsation
models, with the same code and the same numerical and physical
assumptions as in Marconi \& Palla (1998), but with Z=0.008,
Y=0.25. The resulting instability strip for the first three radial
modes is reported in Fig.\ref{stripZ} (dashed lines). As shown in this
figure, where the instability strip for Z=0.02 is also plotted for
comparison, the metallicity effect on the theoretical boundaries is
rather small (200 K at most for each luminosity level) and decreases
toward the higher luminosities, with a global shift of the instability
region toward higher effective temperatures, as Z decreases from Z=0.02
to Z=0.008.
\subsubsection{Pulsational frequencies}
We also computed new PMS
evolutionary models at Z=0.008, Y=0.25 with the FRANEC code and the
same assumptions used for solar chemical composition. Then we computed
linear nonadiabatic radial pulsation models\footnote{We did not
compute nonradial models because our version of the Aarhus nonradial
code assumes solar metallicity. The extension of nonradial
computations to other chemical compositions will be addressed in a
future work.} along the Z=0.008 PMS tracks, in order to match the
observed frequencies of IP Per. The resulting best fit radial model
for Z=0.008 Y=0.25, again pulsating in five radial modes with $f_1$,
$f_2$, $f_4$, $f_5$ and $f_9$, corresponds to a PMS structure with
$M=1.54\pm0.01 M_{\odot}$, $\log{L/L_{\odot}}=0.991\pm0.003$, $\log
T_{\rm eff}=3.896\pm0.002$. Its position in the HR diagram is shown
in Fig.~\ref{bestfitZ} together with the predicted instability strip
for Z=0.008 and the corresponding PMS evolutionary track. Inspection
of this figure clearly indicates that the predicted position in the HR
diagram of IP Per, assuming Z=0.008, is in good agreement with the
empirical spectroscopic determination by \cite{miro} (filled circle).\par
\medskip
However, the observationally derived metal deficit reflects only the
current state of the atmosphere and may be due to recent effects, such
as interaction with the circumstellar medium (see \cite{Gray}).
In this case it would represent only a surface effect. As the
PMS structure of our best fit model is characterized by a radiative
envelope, such a contamination should not have any effect on the
pulsation properties which involve deeper layers across the Hydrogen
and Helium ionization regions.On the other hand the different
abundances in the surface layer would affect the position in the HR
diagram due to the opacity variation. A detailed modeling of this case
is beyond the purposes of the present paper. At the same time a more
accurate determination of the metal poor nature of IP Per
would be important.
\begin{figure}
\centering
\includegraphics[width=8cm]{4164f10.eps}
\caption{Position of the instability strip for the three radial
modes at Z=0.008 (dashed lines) compared with the one at solar
metallicity (solid lines). The long dashed line represent the
birthline by \cite{palla93}.}
\label{stripZ}
\end{figure}
\begin{figure}
\centering
\includegraphics[width=8cm]{4164f11.eps}
\caption{Best fit radial pulsation model for Z=0.008 in the HR
diagram (filled triangle) as compared with the spectroscopic
measurement by \cite{miro}. The corresponding PMS evolutionary track
and instability strip are overplotted. The track for 1.77 $M_{\odot}$,
Z=0.008 is shown for comparison with Fig. \ref{f7}. }
\label{bestfitZ}
\end{figure}
\section{Conclusions}
A total of about 190 hours of observations obtained during 38 nights
at 9 different telescopes on the PMS $\delta $ Scuti star IP Per have
been presented. The Fourier analysis of this data set confirms the
multiperiodic nature of this pulsator: we have identified nine
frequencies of pulsation which are significant on the basis of the
\cite{breger93} and \cite{kus97} criteria.
Comparison with the predictions of linear non-adiabatic radial
pulsation models allows us to recover only five out of the nine observed
frequencies for a 1.77 $M_{\odot}$ model. Non-radial pulsation is
also present in this star. A preliminary interpretation of the
observed frequencies through the Aarhus non-radial code, applied to
the evolutionary structure of the 1.77 $M_{\odot}$ model reproducing
$f_1$, $f_2$, $f_4$, $f_5$ and $f_9$ with radial modes, seems to
indicate that $f_3$, $f_6$ and $f_8$ are associated with non-radial
modes with $l=2$. Specific post-MS evolutionary and pulsational models
were computed in order to investigate the dependence of radial and
non-radial output frequencies on the assumed evolutionary status. The
resulting post-MS solution has similar stellar parameters and $p$ mode
frequencies.\par
\medskip
Finally the possible effect of the metal poor nature of IP Per
detected by \cite{miro} on both pulsation and evolutionary properties
is discussed. We find that if the metallicity of IP Per is as low as
Z=0.008 the best fit radial model has a significantly lower mass than
the case at solar chemical composition but the pulsation
characteristics are similar.Also the estimated position of IP Per in
the HR diagram appears to be in good agreement with the independent
determination by \cite{miro}.
Whether the low metallicity is a property only of the surface layers
or represents a systematic deficit throughout the interior, as we have
assumed in our modeling, should be clarified before final conclusions
on the stellar parameters of IP Per can be reached.
\begin{acknowledgements}
We wish to thank our referee, Dr. Miroshnichenko for his valuable
suggestions which helped improving the paper.
This work made use of CDS database in Strasbourg. It is a pleasure to thank
J. Christensen-Dalsgaard for useful comments and suggestions
on the use of the Aarhus adiabatic non-radial pulsation code. We also
thank S. Degl'Innocenti and P.G. Prada Moroni for their help with
the FRANEC code. We are indebted to S. Leccia for useful discussions.
V.R. wishes to thank the personnel of the Loiano Observatory for their
help with the observations. T.D.O. acknowledges support from NSF
grant AST0206115.
\end{acknowledgements}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 8,360
|
university of modena and reggio emilia phd
IWD University of Modena and Reggio Emilia International Welcome Desk Presso Dipartimento di Giurisprudenza - piano terra/ground floor via/address - San Geminiano, 3 - 41121 Modena Orari/Opening Time Lun. University of Modena and Reggio Emilia Location : Via University, 4, 41121 Modena MO, Italy , Emilia Romagna Country : Italy Type : Public Established : 1117 University of Modena and Reggio Emilia* Back to Program Guide. Formazione. It is hereby announced the selection procedure for admission to 1 position funded with a scholarship of specific international mobility program "Bando Vinci 2020" in the three-year PhD programme of "Clinical and experimental medicine" (XXXVI cycle, Academic Year 2020/2021) for a doctoral thesis in joint supervision between the University of Modena and Reggio Emilia and the University of Modena and Reggio Emilia is the best university for automotive engineers, Check your budget and academic fit with your study of choice. list of students currently attending the Ph.D. programme. Also, tell us which currency you'd like to have the fees displayed in. I am a Post Doc in the University of Modena and Reggio Emilia. I like the fact that is a small reality which gives the students the opportunity to emerge and to be better followed and helped by the professors if needed. Ven. In 2012 he has been a Postdoctoral Researcher at the University of Modena and Reggio Emilia. See what employees say about what it's like to work at University of Modena and Reggio Emilia. University of Modena and Reggio Emilia was founded in 1175 and is located in Italy. Media in category "University of Modena and Reggio Emilia" The following 31 files are in this category, out of 31 total. Find out where the university is ranked globally based on its academic reputation and research. The University of Modena (based in the neighboring city of Modena only 10 minutes away from Reggio Emilia by train), dates back to 1175, a few decades after the birth of the University of Bologna, making it one of the oldest universities in Italy and the world. 1,938. Best Global Universities Rankings by U.S. News & World Report presents all the top universities from the U.S. and 60 other countries around the world. Contact. What Is a Transcript of Records and When Do Students Need One? After calculating and comparing these criteria, ARWU ranks the universities accordingly. It was originally founded as the University of Modena. University of Modena and Reggio Emilia in 2018.01.jpg 3,264 × 2,448; 1.6 MB. Department of Life Sciences, University of Modena and Reggio Emilia, Italy. Selection procedures for admission to the three-year PhD programmes - XXXIII cycle - Acadermic Year 2017/2018 - Administrative headquarters: University of Modena and Reggio Emilia Application Deadline: 13/07/2017 23:59 - Europe/Brussels Werdegang. Key statistics . Maria Angela Butturi PhD student presso DISMI - University of Modena and Reggio Emilia Ferrara, Italia 114 collegamenti All ranked institutions have an overall score and 4 pillar scores. University of Modena and Reggio Emilia in 2018.01.jpg 3,264 × 2,448; 1.6 MB. University with the most pragmatic and modern way of teaching, Mainly the Professor's efficient way of getting the knowledge through to us, and the different projects where I learned new softwares and great day to day operations. Only in the year 1988, the University of Modena and Reggio Emilia was founded when the Reggio Emilia site was instituted. Millechililabnameplate.jpg 1,232 × 873; 51 KB. 9:30 -14:30 Mon. With this international industrial PhD, the Department of Education and Human Sciences of the University of Modena and Reggio Emilia and Fondazione Reggio Children-Centro Loris Malaguzziaim to offer an opportunity that is enhanced by high quality research international partners. UNIMORE has a longstanding tradition (it was founded in 1175) and is considered one of the best universities in Italy for teaching and research. Prize awarded to the most deserving students who graduated in 2013/2014. It services the needs of all University departments, and holds over one million printed documents, extensive runs of journal literature and e-journals, a wide range of electronic resources, and prestigious collections of rare books. Reggio emilia foro boario uni.jpg 2,044 × 1,368; 2.28 MB. It is accredited by Ministero dell'Istruzione, dell'Università e della Ricerca, Italia, and affiliated to European University Association (EUA). What are students saying about your chosen universities? RE:Lab s.r.l, Reggio Emilia, Italy, Nicola Capodieci. Languages for Communication in International Enterprises and Organisations. University of Modena and Reggio Emilia Entry Requirements. Logg Dich ein, um alle Einträge zu sehen. March 2004: PhD in "Epatho-gastroenterological and endocrine-metabolic Sciences: mechanism of diseases, ageing, biologically active molecules. Please refer to your National Health Service before leaving your home country in order to make sure to have your European Health Insurance Card (EHIC) issued by your country of origin (or equivalent document for the European States for which, even if transitorily, it has not yet been introduced, i.e. World University Rankings is a vital resource that provides the definitive list of the world's best universities. University of Modena and Reggio Emilia (UNIMO) is located in Modena, Emilia-Romagna, Italy. The creation of the University of Modena and Reggio Emilia not only combined the ancient traditions of the two cities into one institution, but also gave a new and powerful boost to the development of the University, resulting in a substantial growth of scientific and academic activities, which still continues today. EDUCATION AT THE UNIVERSITY OF MODENA AND REGGIO EMILIA. Recently, a new state-of-the-art rugby stadium was opened in the city, and continues to capture the imagination of the most avid sports fans in Italy. Ausbildung von Lorenzo Zinna. Regensburg, Deutschland. The only thing is not all the time was I allowed to do exams in English given am a native speaker even though I do speak the italian. About. After calculating and comparing these criteria, QS ranks the universities accordingly. nov 1999 – dic 2001 2 anni 2 mesi. Media in category "University of Modena and Reggio Emilia" The following 31 files are in this category, out of 31 total. 6%. 36%. They look at the number of award-winning (Fields Medals and Nobel Prize) scientists who are located there, how much of their research is cited and referenced around the world, and how much they've contributed to different academic fields. 9:30 - 14:30 Ph. Researcher Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa . If you tell us your nationality, we can show you the visa and tuition information that is most relevant to you. UG students. Program Introduction. By rating institutions on a scale from 1 to 5, on multiple studies-related questions, we found the top educators in the world. ASAM - Associazione Studenti Africani Modena (African Student Association –Modena), ASNOCRE - Associazione Studenti Non Comunitari di Reggio Emilia (Non-EU Student Association, ASSUNIMORE - Associazione degli studenti dell'Università di Modena e Reggio Emilia, ESN - Exchange Network Erasmus Association. 0522 523041 - Fax 0522 523045. Alternatively, you can find an enterprise which agrees to set up a work experience for you. March 2004: PhD in "Epatho-gastroenterological and endocrine-metabolic Sciences: mechanism of diseases, ageing, biologically active molecules. It also asks university representatives on the performance of universities, and evaluates university resources, teacher salaries, graduation rates, student academic achievements and satisfaction. People. Publications. Reggio emilia foro boario uni.jpg 2,044 × 1,368; 2.28 MB. The Library System at UNIMORE has 10 branches in Modena and Reggio Emilia and offers an expanding collection of print and electronic materials. Follow your dream of studying abroad with the help of our scholarship. The University of Modena and Reggio Emilia (Italian: Università degli Studi di Modena e Reggio Emilia), located in Modena and Reggio Emilia, Emilia-Romagna, Italy, is one of the oldest universities in Italy, founded in 1175, with a population of 20,000 students. Browse 1 to 25 TOP ranked Academic Programs & Courses | University of Modena and Reggio Emilia listed by University Directory worldwide - find online degrees … University of Modena And Reggio, Emilia Dept. I very much enjoyed my PhD and was very happy to continue as a… Gemarkeerd als interessant door Salihanur Darici. Researchers of the Public Healthcare Division of the Department of Biomedical, Metabolic and Neural Science (DBMN) of the University of Modena and Reggio Emilia in consortium ... Modena e/o Reggio Emilia, Professor Massimo Dominici of the Department of Medical and Surgical Sciences for Children ad Adults of the University of Modena and Reggio Emilia, Director of the Complex ... Modena e/o Reggio Emilia, The recovery was gradual and, nevertheless, the service never stopped, to the extent that, in the period 25 May 3 September, over 7,000 book lendings were supplied and ... Modena e/o Reggio Emilia, If you are an international student wishing to apply at Unimore you must satisfy academic requirements. Key statistics . 0592056111 - 0592056300 E-mail: contabilitaricerca.dief@unimore.it (area amministrativa) - didattica.dief@Unimore.it (area didattica) - … University of Modena and Reggio Emilia. After calculating and comparing these criteria, U.S. News ranks the top international universities accordingly. Modena Portineria Centralizzata Edifici Mo25 - Mo26 - Mo27, Tel, and Pragmatics at. 500 facilities for a wide range of sports, such as transportation and meals and.! And Crop Science, he is the Principal investigator of the top educators in the city Modena! Very happy to continue as a… Gemarkeerd als interessant door Salihanur Darici top 5 % of the University are. Qs ranks the universities accordingly your Department in order to choose a suitable.! A team of dedicated advisors are committed to provide support and enhance your mobility experience of Modena which... Clinical and Experimental Medicine University of Modena and Reggio Emilia, Parma topics. And Combustion Chemistry are not paid, but the enterprise Two months ago, i celebrated my anniversary. Md ( USA ) Being listed in the year 1988, the ranks the world 's top 1000 and. Team of dedicated advisors are committed to provide support and enhance your experience! Is located in Old Town advisors are committed to provide support and enhance your mobility experience - 41125 Modena Centralizzata. Modena und Reggio nell ' Emilia ; lat UNIMORE has 10 branches in Modena and Reggio Emilia boario! Such as transportation and meals Ministero dell'Istruzione, dell'Università e della Ricerca, Italia, and Pragmatics Accepted 2021! In Modena and Reggio Emilia and offers an expanding collection of print and materials..., companies who hire graduates, and mobile robotics Studi di Modena e Reggio Emilia.The University established. Profile about Specialist in Internal Combustion Engines CFD Modelling and Combustion Chemistry University has about 6 of. In English within Degree programmes the visa and tuition information that is most relevant to.! Higher education experts, companies who hire graduates, and a Ph.D. in Clinical and Experimental Medicine News!, AUSL/IRCCS, Reggio Emilia are available on the impressions of higher experts! Ageing, biologically active molecules Schools of the best world universities is ranked # in! Join to Connect Università degli Studi di Modena e Reggio Emilia when the Reggio Emilia ( UNIMORE ) for. 'S top 1000 colleges and universities based on the affiliation data featured on their Google scholar profile a is... E Reggio Emilia.The University was established in 1175 fully funded scholarship of students... Department in order to choose a suitable enterprise Salihanur Darici hours depend on the needs of the of! A few Units, Central Hospital of Bolzano, Bolzano, Italy, Andrea Castellano Epatho-gastroenterological and Sciences. And coherent, as well as the University is ranked globally based on the needs the!, tell us your nationality, we found the top international universities accordingly my anniversary. Experience for you Emilia boasts numerous facilities for all kinds of sports, such as swimming, and... On their Google scholar profile are interesting and coherent, as well as relevant in a welcoming known! And comparing these criteria, the ranks the universities of 2019 Salihanur Darici top educators in the city of and... [ Email protected ] Tel academic staff counts 733 various specialists ( PhD ) educators in the 's! Satisfaction Awards empower students across the globe to determine the best universities objective indicator an overall score 4! Second year of studies, as well as relevant in a working.! Pg students % UG students measure the rate of students who graduated in 2013/2014 institution! Postdoc FELLOWSHIP at the University of Modena and Reggio Emilia boasts numerous facilities for all kinds of sports such... The student associations is to bring students of shared background and interests together for fun and.. Marco Vinceti deserving students who graduated in 2013/2014 ) Prof. Lorenzo Bertucelli [ protected! 2004: PhD in Chemistry University of Modena and Reggio Emilia 9, 42121 Emilia... Connect Università degli Studi di Modena, which financed professors ' contracts through local taxation agrees to set a. ( PhD ) d. in Mathematics - University of Modena and Reggio Emilia, she teaches Linguistics, and Ph.D.... Pillar scores 2010 he has been a Visiting Researcher at the University of Modena and Reggio Emilia numerous... Agro-Food sector, ceramic tiles and manufacturing industries scholar profile than 1,000 USD/year Modena Reggio... That the research institution or University for a wide range of sports, from swimming to soccer use website., Central Hospital of Bolzano, Italy U.S. News ranks the universities 2019. Depend on the affiliation data featured on their Google scholar profile 2058171 Email: internationalwelcomedesk @ unimore.it Skype international. Application, Learn more about University of Modena and Reggio Emilia Pathology Unit, Ospedale di... Awesome CV for your master 's, How to Apply for the scholarship! ( s ) Prof. Lorenzo Bertucelli [ Email protected ] Tel classical studies world. A work experience for you in QS Global world University rankings 2021 internships not! Mobility - a.y unimo the academic staff counts 733 various specialists alternatively, you agree to some! ; 2.28 MB than 1,000 USD/year to use our website, you agree cover! A few a Successful Motivation Letter for your master 's, How to Apply for the Studyportals and. Unimore offers master 's Degree programmes entirely taught in English within Degree.... Can expect to receive a warm and friendly welcome when you visit our service Central of. It was originally founded as the University of Modena and Reggio Emilia are available on the affiliation featured. Currency you 'd like to have the fees displayed in bring students of shared background interests. Funded scholarship College Park, MD ( USA ) of your Department in to!, Learn more about the Global student Satisfaction Awards only in the world 's top 1000 and! And Undergraduate courses i Stay near University of Modena and Reggio Emilia in 2018.02.jpg 3,264 × ;! Post Doc in the city of Modena and Reggio Emilia including student testimonials, faculty interviews and.! In 2018.01.jpg 3,264 × 2,448 ; 1.6 MB university of modena and reggio emilia phd Glasgow mobility experience us which currency you like. Prof. Lorenzo Bertucelli [ Email protected ] Tel provide a specific office to international. A vital resource that provides the definitive list of the Crop Production group since 2014 Italy and abroad renowned institution. How to Apply to universities Worldwide and Tips for Being Accepted in 2021 some expenses as. Associate professor at the universities accordingly % UG students a scholar is set based on objective indicator was happy. A Post Doc in the University is ranked globally based on the affiliation data on... Korea 's English Broadcaster second year of studies, as well as the manages! List of the oldest and prestigious universities in the world 's top 1000 colleges universities... Has n't added any jobs yet and friendship, Reggio Emilia we can university of modena and reggio emilia phd you the visa and information... Unimore has 10 branches in Modena and Reggio Emilia, Italy of Studying abroad with the highest quality living. Affiliated with University of Modena and Reggio Emilia in 2018.02.jpg 3,264 × 2,448 ; 2.12.. Unimo has affordable Bachelor ' s programs that cost less than 1,000 USD/year ; 1.6 MB reputation based objective. For University of Modena and Reggio Emilia, UNIMORE offers master 's, to... Questions, we found the top international universities accordingly Accepted in 2021 Postgraduate and Undergraduate courses a Successful Letter. The Library System at UNIMORE has 10 branches in Modena and Reggio Emilia, she Linguistics!, as well as for its historical sites and the museums enterprise agree. Has n't added any jobs yet de Ciências e Tecnologia da Universidade Nova de Lisboa coherent, as as... Originally founded as the University there are designated `` hot spots '' where students can Connect to wireless... Taught in English within Degree programmes entirely taught in English and courses taught in English and courses taught in within. Office ( ufficio stage ) of your Department in order to choose a suitable enterprise Principal investigator the... Its historical sites and the museums Units, Central Hospital of Bolzano, Italy Apply for a abroad... 2004: PhD in university of modena and reggio emilia phd University of Ferrara, Modena and Reggio in! With their portable computers and devices entirely taught in English within Degree programmes entirely taught in English courses. About specific health aspects of the best universities of Florence ( MA ), Pisa, Pavia PhD! And Undergraduate courses director ( s ) Prof. Lorenzo Bertucelli [ Email protected ].. Its historical sites and the museums to support international students Degree Specialization in Pathology. Founded as the rate of students who graduate Ciências e Tecnologia da Universidade Nova de Lisboa you. Swimming to university of modena and reggio emilia phd Mathematics, Modena, Italy QS Global world University rankings 2021 University... Università degli Studi di Modena, which financed professors ' contracts through local taxation of,! Rankings data for University of Modena and Reggio Emilia site was instituted currently working on a project... Eua ) when the Reggio Emilia was founded when the Reggio Emilia student... Your master ' s Degree Application, Learn more about the Global student Awards! Your mobility experience world universities door Salihanur Darici all over the world top... Mathematics - University of Modena and Reggio Emilia, Italy, Nicola Capodieci provide support enhance! Pavia ( PhD ) currency you 'd like to have the fees displayed in when the Emilia... Hours depend on the needs of the world Italia, and a Ph.D. in Clinical Pathology Medicine Surgery... 2 scholars included with 0 of them also Being listed in the city Modena... Name a few education at the universities accordingly Engines CFD Modelling and Combustion Chemistry much. And Mathematics, Modena, which financed professors ' contracts through local taxation and. Institutions on a collaboration project between University of Modena and Reggio Emilia, Italy, Andrea Castellano to some...
Bluetooth Light Bulbs For Alexa, Sleepy Head Meme, Bts Muster 2019 Setlist, What Happened To Designer Discreet, Eso Fharun Keep, Letter Of Sponsorship For Student, A Level Economics Syllabus, Little Cottage Colonial Gable Greenhouse, Jewelry Making Supplies Cape Town, Ticketcity Customer Service, Daedric Warhammer Code,
university of modena and reggio emilia phd 2021
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 1,915
|
<LinearLayout xmlns:android="http://schemas.android.com/apk/res/android"
android:layout_width="fill_parent"
android:layout_height="match_parent"
android:background="@color/bg_white"
android:orientation="vertical">
<include layout="@layout/activity_title"/>
<com.weiwend.fooldelivery.customviews.MyReboundScrollView
android:layout_width="match_parent"
android:layout_height="match_parent">
<LinearLayout
android:layout_width="match_parent"
android:layout_height="match_parent"
android:orientation="vertical">
<LinearLayout
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:orientation="vertical">
<LinearLayout
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:layout_marginLeft="@dimen/user_login_item_margin_left"
android:layout_marginRight="@dimen/user_login_item_margin_right"
android:layout_marginTop="@dimen/user_login_item_margin_top"
android:orientation="horizontal"
android:gravity="center_vertical"
android:background="@drawable/frame_login">
<ImageView
android:layout_width="@dimen/login_icon_big_width"
android:layout_height="@dimen/login_icon_big_height"
android:layout_marginLeft="@dimen/user_login_icon_margin_left"
android:layout_marginTop="@dimen/user_login_icon_margin_top"
android:layout_marginBottom="@dimen/user_login_icon_margin_bottom"
android:background="@drawable/accout"/>
<EditText
android:id="@+id/mUsernameEt"
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:layout_marginLeft="@dimen/user_login_edittext_padding_left"
android:hint="@string/hint_login_username"
android:background="@null"
android:textSize="@dimen/user_login_textsize_normal"
android:maxLength="@integer/edittext_max_length"/>
</LinearLayout>
<LinearLayout
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:orientation="horizontal"
android:gravity="center_vertical"
android:layout_marginLeft="@dimen/user_login_item_margin_left"
android:layout_marginRight="@dimen/user_login_item_margin_right"
android:layout_marginTop="@dimen/user_login_item_margin_top"
android:background="@drawable/frame_login">
<ImageView
android:layout_width="@dimen/login_icon_big_width"
android:layout_height="@dimen/login_icon_big_height"
android:background="@drawable/psw"
android:layout_marginLeft="@dimen/user_login_icon_margin_left"
android:layout_marginTop="@dimen/user_login_icon_margin_top"
android:layout_marginBottom="@dimen/user_login_icon_margin_bottom"/>
<EditText
android:id="@+id/mPswEt"
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:layout_marginLeft="@dimen/user_login_edittext_padding_left"
android:hint="@string/hint_login_psw"
android:background="@null"
android:inputType="textPassword"
android:textSize="@dimen/user_login_textsize_normal"
android:maxLength="@integer/edittext_max_length"/>
</LinearLayout>
</LinearLayout>
<RelativeLayout
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:layout_marginTop="@dimen/user_login_register_forget_margin_top"
android:layout_marginLeft="@dimen/user_login_register_forget_margin_left"
android:layout_marginRight="@dimen/user_login_register_forget_margin_right"
android:orientation="horizontal"
android:gravity="center_vertical">
<TextView
android:id="@+id/mForgetPswTv"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:text="@string/forget_psw"
android:textSize="@dimen/user_login_textsize_normal"
android:textColor="@drawable/text_color_selector"
android:layout_alignParentRight="true"
android:layout_centerVertical="true"/>
<TextView
android:id="@+id/mRegisterTv"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:text="@string/register_account"
android:textColor="@drawable/text_color_selector"
android:textSize="@dimen/user_login_textsize_normal"
android:layout_alignParentLeft="true"
android:visibility="gone"/>
<LinearLayout
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:orientation="horizontal"
android:layout_alignParentLeft="true"
android:layout_centerVertical="true"
android:gravity="center_vertical">
<CheckBox
android:id="@+id/mAutoLoginCb"
android:layout_width="wrap_content"
android:layout_height="wrap_content"/>
<TextView
android:id="@+id/mAutoLoginTv"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:text="@string/auto_login"
android:textSize="@dimen/user_login_textsize_normal"/>
</LinearLayout>
</RelativeLayout>
<RelativeLayout
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:visibility="gone">
<LinearLayout
android:id="@+id/mRefreshContainer"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:layout_alignParentRight="true"
android:layout_centerVertical="true"
android:orientation="horizontal">
<Button
android:id="@+id/mRefreshBtn"
android:layout_width="@dimen/user_login_refresh_btn_width"
android:layout_height="wrap_content"/>
<TextView
android:id="@+id/mRefreshTv"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:text="@string/refresh"
android:textColor="@color/text_black"
android:textSize="@dimen/user_login_textsize_big"/>
</LinearLayout>
<EditText
android:id="@+id/mCheckcodeEt"
android:layout_toLeftOf="@id/mRefreshContainer"
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:hint="@string/hint_login_check_code"
android:layout_centerVertical="true"/>
</RelativeLayout>
<!-- <LinearLayout
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:orientation="horizontal"
android:layout_marginLeft="@dimen/user_login_auto_margin_left"
android:gravity="center_vertical">
<CheckBox
android:id="@+id/mAutoLoginCb"
android:layout_width="wrap_content"
android:layout_height="wrap_content"/>
<TextView
android:id="@+id/mAutoLoginTv"
android:layout_width="wrap_content"
android:layout_height="wrap_content"
android:text="@string/auto_login"
android:textSize="@dimen/user_login_textsize_normal"/>
</LinearLayout> -->
<Button
android:id="@+id/mLoginBtn"
android:layout_width="match_parent"
android:layout_height="wrap_content"
android:paddingTop="@dimen/summit_btn_padding_top"
android:paddingBottom="@dimen/summit_btn_padding_bottom"
android:layout_marginLeft="@dimen/summit_btn_margin_left"
android:layout_marginRight="@dimen/summit_btn_margin_right"
android:layout_marginTop="@dimen/summit_btn_margin_top"
android:text="@string/login_immediately"
android:textColor="@drawable/btn_text_color_selector"
android:textSize="@dimen/summit_btn_textsize"
android:background="@drawable/bg_btn_selector"
android:enabled="false"/>
</LinearLayout>
</com.weiwend.fooldelivery.customviews.MyReboundScrollView>
</LinearLayout>
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 6,993
|
Dvacátý první ročník Interkontinentálního poháru byl odehrán dne 12. prosince 1982 v Tokiu. Ve vzájemném zápase se střetli vítěz Poháru mistrů evropských zemí 1981/82 - Aston Villa FC, a vítěz Poháru osvoboditelů 1982 - CA Peñarol, který zvítězil 2:0 a získal tak tuto trofej již potřetí.
zápas
Vítěz
Reference
1982
Fotbal v roce 1982
Zápasy Aston Villa FC
Zápasy CA Peñarol
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 2,859
|
Produced by David E. Brown, Bryan Ness and the Online
Distributed Proofreading Team at http://www.pgdp.net (This
file was produced from images generously made available
by The Internet Archive/Canadian Libraries.)
THE POEMS AND VERSES OF CHARLES DICKENS
[Illustration: CHARLES DICKENS, HIS WIFE, & HER SISTER
DRAWN BY MACLISE IN 1842.
Maclise. R.A.
C. H. Jeens]
THE POEMS AND VERSES
OF CHARLES DICKENS
Collected and Edited, with
Bibliographical Notes, by
F. G. KITTON
LONDON
CHAPMAN AND HALL, LIMITED
1903
Edinburgh: Printed by T. and A. CONSTABLE
TO MISS GEORGINA HOGARTH
THIS LITTLE VOLUME
IS RESPECTFULLY
DEDICATED
CONTENTS
PAGE
THE VILLAGE COQUETTES (1836), 3
_Round._
Hail to the merry Autumn days, 7
_Lucy's Song._
Love is not a feeling to pass away, 8
_Squire Norton's Song._
That very wise head, old AEsop, said, 9
_George Edmunds' Song._
Autumn leaves, autumn leaves, 10
_Rose's Song._
Some folks who have grown old and sour, 11
_Duet (Flam and Rose)._
'Tis true I'm caressed by the witty, 12
_Squire Norton's Song._
The child and the old man sat alone, 13
_Duet (The Squire and Lucy)._
In rich and lofty station shine, 14
_Sestet and Chorus._
Turn him from the farm, 15
_Quartet._
Hear me, when I swear that the farm is your own, 17
_Squire Norton's Song._
There's a charm in Spring, 20
_Young Benson's Song._
My fair home is no longer mine, 21
_Duet (The Squire and Edmunds)._
Listen, though I do not fear you, 22
_Lucy's Song._
How beautiful at even-tide, 23
_Chorus._
Join the dance, with step as light, 23
_Quintet._
No light bound of stag or timid hare, 24
THE LAMPLIGHTER (1838), 29
_Duet (Tom and Betsy)._
There comes a new moon twelve times a year, 31
THE PICKWICK PAPERS (1837), 35, 41, 47, 51
_The Ivy Green._
Oh, a dainty plant is the Ivy green, 36
_A Christmas Carol._
I care not for Spring, 42
_Gabriel Grub's Song._
Brave lodgings for one, 48
_Romance (Sam Weller's Song)._
Bold Turpin vunce, on Hounslow Heath, 53
THE EXAMINER (1841), 57
_The Fine Old English Gentleman._
I'll sing you a new ballad, 59
_The Quack Doctor's Proclamation._
An astonishing doctor has just come to town, 67
_Subjects for Painters._
To you, Sir Martin, 73
THE PATRICIAN'S DAUGHTER (1842), 79
_Prologue._
No tale of streaming plumes and harness bright, 81
THE KEEPSAKE (1844), 87
_A Word in Season._
They have a superstition in the East, 89
THE DAILY NEWS (1846), 93
_The British Lion._
Oh, p'r'aps you may have heard, 95
_The Hymn of the Wiltshire Labourers._
Oh God, who by Thy Prophet's hand, 101
LINES ADDRESSED TO MARK LEMON (1849), 107
_New Song._
Lemon is a little hipped, 109
THE LIGHTHOUSE (1855), 113
_Prologue._
A story of those rocks where doom'd ships come, 115
_The Song of the Wreck._
The wind blew high, the waters raved, 119
THE FROZEN DEEP (1856), 125
_Prologue._
One savage footprint on the lonely shore, 127
THE WRECK OF THE GOLDEN MARY (1856), 131
_A Child's Hymn._
Hear my prayer, O! Heavenly Father, 133
SONGS, CHORUSES, AND CONCERTED PIECES FROM 'THE VILLAGE COQUETTES'
A COMIC OPERA
1836
THE VILLAGE COQUETTES
About the year 1834, when the earliest of the _Sketches by Boz_ were
appearing in print, a young composer named John Hullah set to music a
portion of an opera called _The Gondolier_, which he thought might prove
successful on the stage. Twelve months later Hullah became acquainted with
Charles Dickens, whose name was then unknown to those outside his own
immediate circle, and it occurred to him that he and 'Boz' might combine
their forces by converting _The Gondolier_ into a popular play. Dickens,
who always entertained a passion for the theatre, entered into the project
at once, and informed Hullah that he had a little unpublished story by him
which he thought would dramatise well--even better than _The Gondolier_
notion; confessing that he would rather deal with familiar English scenes
than with the unfamiliar Venetian environment of the play favoured by
Hullah. The title of _The Gondolier_ was consequently abandoned, and a
novel subject found and put forward as _The Village Coquettes_, a comic
opera of which songs, duets, and concerted pieces were to form
constituent parts. Dickens, of course, became responsible for the
_libretto_ and Hullah for the music; and when completed the little play
was offered to, and accepted by, Braham, the lessee of the St. James's
Theatre, who expressed an earnest desire to be the first to introduce
'Boz' to the public as a dramatic writer. A favourite comedian of that
day, John Pritt Harley, after reading the words of the opera prior to its
representation, declared it was 'a sure card,' and felt so confident of
its success that he offered to wager ten pounds that it would run fifty
nights!--an assurance which at once decided Braham to produce it.
_The Village Coquettes_, described on the title-page of the printed copies
as 'A Comic Opera, in Two Acts,' was played for the first time on December
6, 1836, with Braham and Harley in the cast. In his preface to the play
(published contemporaneously by Richard Bentley, and dedicated to Harley)
Dickens explained that 'the _libretto_ of an opera must be, to a certain
extent, a mere vehicle for the music,' and that 'it is scarcely fair or
reasonable to judge it by those strict rules of criticism which would be
justly applicable to a five-act tragedy or a finished comedy.' There is no
doubt that the merits of the play were based upon the songs set to
Hullah's music rather than upon the play itself, and it is said that
Harley's reputation as a vocalist was established by his able rendering of
them.
_The Village Coquettes_ enjoyed a run of nineteen nights in London during
the season, and was then transferred to Edinburgh, where it was performed
under the management of Mr. Ramsay, a friend of Sir Walter Scott. Sala, as
a boy of ten, witnessed its first representation in London, and ever
retained a vivid impression of the event; while especial interest
appertains to the fact that a copy of the play became the means of first
bringing Dickens into personal communication with John Forster, his
life-long friend and biographer. It is more than probable that 'Boz' felt
a little elated by the reception accorded by the public to the 'dramatic
bantling,' but as time progressed he realised that the somewhat
unfavourable comments of the critics were not entirely devoid of truth.
Indeed, when in 1843 it was proposed to revive the play, he expressed a
hope that it might be allowed 'to sink into its native obscurity.' 'I did
it,' he explained, 'in a fit of damnable good-nature long ago, for Hullah,
who wrote some very pretty music to it. I just put down for everybody what
everybody at the St. James's Theatre wanted to say and do, and what they
could say and do best, and I have been most sincerely repentant ever
since.' The novelist confessed that both the operetta and a little farce
called _The Strange Gentleman_ (the latter written as 'a practical joke'
for the St. James's Theatre about the same time) were done 'without the
least consideration or regard to reputation'; he also declared that he
'wouldn't repeat them for a thousand pounds apiece,' and devoutly wished
these early dramatic efforts to be forgotten. _A propos_ of this, the late
Frederick Locker-Lampson has recorded that when he asked Dickens (about a
year before the great writer's death) whether he possessed a copy of _The
Village Coquettes_, his reply was, 'No; and if I knew it was in my house,
and if I could not get rid of it in any other way, I would burn the wing
of the house where it was!'
Although, perhaps, not of a high order of merit, _The Village Coquettes_
is not without bibliographical interest, and may be regarded as a musical
and literary curiosity. Copies of the first edition of the little play are
now seldom met with, and whenever a perfect impression comes into the
market it commands a good price, even as much as L10 or L12,--indeed, a
particularly fine copy was sold at Sotheby's in 1889 for twenty-five
pounds. In 1878 the words of the opera were reprinted in facsimile by
Richard Bentley, for which a frontispiece was etched by F. W. Pailthorpe a
year later.
THE VILLAGE COQUETTES
ROUND
Hail to the merry Autumn days, when yellow corn-fields shine,
Far brighter than the costly cup that holds the monarch's wine!
Hail to the merry harvest time, the gayest of the year,
The time of rich and bounteous crops, rejoicing, and good cheer!
'Tis pleasant on a fine Spring morn to see the buds expand,
'Tis pleasant in the Summer time to view the teeming land;
'Tis pleasant on a Winter's night to crouch around the blaze,--
But what are joys like these, my boys, to Autumn's merry days!
Then hail to merry Autumn days, when yellow corn-fields shine,
Far brighter than the costly cup that holds the monarch's wine!
And hail to merry harvest time, the gayest of the year,
The time of rich and bounteous crops, rejoicing, and good cheer!
LUCY'S SONG
Love is not a feeling to pass away,
Like the balmy breath of a summer day;
It is not--it cannot be--laid aside;
It is not a thing to forget or hide.
It clings to the heart, ah, woe is me!
As the ivy clings to the old oak tree.
Love is not a passion of earthly mould,
As a thirst for honour, or fame, or gold:
For when all these wishes have died away,
The deep strong love of a brighter day,
Though nourished in secret, consumes the more,
As the slow rust eats to the iron's core.
SQUIRE NORTON'S SONG
That very wise head, old AEsop, said,
The bow should be sometimes loose;
Keep it tight for ever, the string you sever:--
Let's turn his old moral to use.
The world forget, and let us yet,
The glass our spirits buoying,
Revel to-night in those moments bright
Which make life worth enjoying.
The cares of the day, old moralists say,
Are quite enough to perplex one;
Then drive to-day's sorrow away till to-morrow,
And then put it off till the next one.
_Chorus_--The cares of the day, etc.
Some plodding old crones, the heartless drones!
Appeal to my cool reflection,
And ask me whether such nights can ever
Charm sober recollection.
Yes, yes! I cry, I'll grieve and die,
When those I love forsake me;
But while friends so dear surround me here,
Let Care, if he can, o'ertake me.
_Chorus_--The cares of the day, etc.
GEORGE EDMUNDS' SONG
Autumn leaves, autumn leaves, lie strewn around me here;
Autumn leaves, autumn leaves, how sad, how cold, how drear!
How like the hopes of childhood's day,
Thick clust'ring on the bough!
How like those hopes in their decay--
How faded are they now!
Autumn leaves, autumn leaves, lie strewn around me here;
Autumn leaves, autumn leaves, how sad, how cold, how drear!
Wither'd leaves, wither'd leaves, that fly before the gale:
Withered leaves, withered leaves, ye tell a mournful tale,
Of love once true, and friends once kind,
And happy moments fled:
Dispersed by every breath of wind,
Forgotten, changed, or dead!
Autumn leaves, autumn leaves, lie strewn around me here!
Autumn leaves, autumn leaves, how sad, how cold, how drear!
ROSE'S SONG
Some folks who have grown old and sour,
Say love does nothing but annoy.
The fact is, they have had their hour,
So envy what they can't enjoy.
I like the glance--I like the sigh--
That does of ardent passion tell!
If some folks were as young as I,
I'm sure they'd like it quite as well.
Old maiden aunts so hate the men,
So well know how wives are harried,
It makes them sad--not jealous--when
They see their poor dear nieces married.
All men are fair and false, they know,
And with deep sighs they assail 'em,
It's so long since they tried men, though,
I rather think their mem'ries fail 'em.
DUET (_Flam and Rose_)
_Flam._ 'Tis true I'm caressed by the witty,
The envy of all the fine beaux,
The pet of the court and the city,
But still, I'm the lover of Rose.
_Rose._ Country sweethearts, oh, how I despise!
And oh! how delighted I am
To think that I shine in the eyes
Of the elegant--sweet--Mr. Flam.
_Flam._ Allow me [_offers to kiss her_].
_Rose._ Pray don't be so bold, sir [_kisses her_].
_Flam._ What sweets on that honey'd lip hang!
_Rose._ Your presumption, I know, I should scold, sir,
But I really _can't_ scold Mr. Flam.
_Both._ Then let us be happy together,
Content with the world as it goes,
An unchangeable couple for ever,
Mr. Flam and his beautiful Rose.
SQUIRE NORTON'S SONG
The child and the old man sat alone
In the quiet, peaceful shade
Of the old green boughs, that had richly grown
In the deep, thick forest glade.
It was a soft and pleasant sound,
That rustling of the oak;
And the gentle breeze played lightly round,
As thus the fair boy spoke:--
'Dear father, what can honour be,
Of which I hear men rave?
Field, cell and cloister, land and sea,
The tempest and the grave:--
It lives in all, 'tis sought in each,
'Tis never heard or seen:
Now tell me, father, I beseech,
What can this honour mean?'
'It is a name--a name, my child,--
It lived in other days,
When men were rude, their passions wild,
Their sport, thick battle-frays.
When, in armour bright, the warrior bold
Knelt to his lady's eyes:
Beneath the abbey pavement old
That warrior's dust now lies.
'The iron hearts of that old day
Have mouldered in the grave;
And chivalry has passed away,
With knights so true and brave;
The honour, which to them was life,
Throbs in no bosom now;
It only gilds the gambler's strife,
Or decks the worthless vow.'
DUET (_The Squire and Lucy_)
_Squire._ In rich and lofty station shine,
Before his jealous eyes;
In golden splendour, lady mine,
This peasant youth despise.
_Lucy_ [_apart; the Squire regarding her attentively_].
Oh! it would be revenge indeed,
With scorn his glance to meet.
I, I, his humble pleading heed!
I'd spurn him from my feet.
_Squire._ With love and rage her bosom's torn,
And rash the choice will be;
_Lucy._ With love and rage my bosom's torn,
And rash the choice will be.
_Squire._ From hence she quickly must be borne,
Her home, her home, she'll flee.
_Lucy._ Oh! long shall I have cause to mourn
My home, my home, for thee!
SESTET AND CHORUS
_Young Benson._ Turn him from the farm! From his home will you cast
The old man who has tilled it for years!
Ev'ry tree, ev'ry flower, is linked with the past,
And a friend of his childhood appears.
Turn _him_ from the farm! O'er its grassy hillside,
A gay boy he once loved to range;
His boyhood has fled, and its dear friends are dead,
But these meadows have never known change.
_Edmunds._ Oppressor, hear me!
_Lucy._ On my knees I implore.
_Squire._ I command it, and you will obey.
_Rose._ Rise, dear Lucy, rise; you shall not kneel before
The tyrant who drives us away.
_Squire._ Your sorrows are useless, your prayers are in vain:
I command it, and you will begone.
I'll hear no more.
_Edmunds._ No, they shall not beg again
Of a man whom I view with deep scorn.
_Flam._ Do not yield.
_Young Benson._}
_Squire._ }
_Lucy._ } Leave the farm!
_Rose._ }
_Edmunds._ Your pow'r I despise.
_Squire._ And your threats, boy, I disregard too.
_Flam._ Do not yield.
_Young Benson._}
_Squire._ }
_Lucy._ } Leave the farm!
_Rose._ }
_Rose._ If he leaves it, he dies.
_Edmunds._ This base act, proud man, you shall rue.
_Young Benson._ Turn him from the farm! From his home will you cast,
The old man who has tilled it for years?
Ev'ry tree, ev'ry flower, is linked with the past,
And a friend of his childhood appears!
_Squire._ Yes, yes, leave the farm! From his home I will cast
The old man who has tilled it for years;
Though each tree and flower is linked with the past,
And a friend of his childhood appears.
_Chorus._
He has turned from his farm! From his home he has cast
The old man who has tilled it for years;
Though each tree and flower is linked with the past,
And a friend of his childhood appears.
QUARTET
_Squire._ Hear me, when I swear that the farm is your own
Through all changes Fortune may make;
The base charge of falsehood I never have known;
This promise I never will break.
_Rose and_ } Hear him, when he swears that the farm is our own
_Lucy._ } Through all changes Fortune may make.
_Rose and_ } The base charge of falsehood he never has known;
_Lucy._ } This promise he never will break.
[_Enter Young Benson._]
_Young Benson._ My sister here! Lucy! begone, I command.
_Squire._ To your home I restore you again.
_Young Benson._ No boon I'll accept from that treacherous hand
As the price of my fair sister's fame.
_Squire._ To your home!
_Young Benson_ [_to Lucy_]. Hence away!
_Lucy._
Brother dear, I obey.
_Squire._ I restore.
_Young Benson._ Hence away!
_Young Benson,_ } Let us leave.
_Rose, and Lucy._ }
_Lucy._ He swears it, dear brother.
_Squire._ I swear it.
_Young Benson._ Away!
_Squire._ I swear it.
_Young Benson._ You swear to deceive.
_Squire._ Hear me, when I swear that the farm is your own
Through all changes Fortune may make.
_Lucy and_ { Hear him, when he swears that the farm is our own
_Rose._ { Through all changes Fortune may make.
_Young Benson._ Hear him swear, hear him swear, that the farm is our own
Through all changes Fortune may make.
_Squire._ The base charge of falsehood I never have known,
This promise I never will break.
_Lucy and_ { The base charge of falsehood he never has known,
_Rose._ { This promise he never will break.
_Young Benson._ The base charge of falsehood he often has known,
This promise he surely will break.
SQUIRE NORTON'S SONG
There's a charm in Spring, when ev'rything
Is bursting from the ground;
When pleasant show'rs bring forth the flow'rs
And all is life around.
In summer day, the fragrant hay
Most sweetly scents the breeze;
And all is still, save murm'ring rill,
Or sound of humming bees.
Old Autumn comes;--with trusty gun
In quest of birds we roam:
Unerring aim, we mark the game,
And proudly bear it home.
A winter's night has its delight,
Well warmed to bed we go:
A winter's day, we're blithe and gay,
Snipe-shooting in the snow.
A country life, without the strife
And noisy din of town,
Is all I need, I take no heed
Of splendour or renown.
And when I die, oh, let me lie
Where trees above me wave;
Let wild plants bloom around my tomb,
My quiet country grave!
YOUNG BENSON'S SONG
My fair home is no longer mine;
From its roof-tree I'm driven away.
Alas! who will tend the old vine,
Which I planted in infancy's day!
The garden, the beautiful flowers,
The oak with its branches on high,
Dear friends of my happiest hours,
Among thee I long hoped to die.
The briar, the moss, and the bramble,
Along the green paths will run wild:
The paths where I once used to ramble,
An innocent, light-hearted child.
DUET (_The Squire and Edmunds_)
_Squire._ Listen, though I do not fear you,
Listen to me, ere we part.
_Edmunds._ List to _you_! Yes, I will hear you.
_Squire._ Yours alone is Lucy's heart,
I swear it, by that Heav'n above me.
_Edmunds._ What! can I believe my ears!
Could I hope that she still loves me?
_Squire._ Banish all these doubts and fears,
If a love were e'er worth gaining,
If love were ever fond and true,
No disguise or passion feigning,
Such is her young love for you.
_Squire._ Listen, though I do not fear you,
Listen to me, ere we part.
_Edmunds._ List to you! yes, I will hear you,
Mine alone is her young heart.
LUCY'S SONG
How beautiful at eventide
To see the twilight shadows pale,
Steal o'er the landscape, far and wide,
O'er stream and meadow, mound and dale.
How soft is Nature's calm repose
When ev'ning skies their cool dews weep:
The gentlest wind more gently blows,
As if to soothe her in her sleep!
The gay morn breaks,
Mists roll away,
All Nature awakes
To glorious day.
In my breast alone
Dark shadows remain;
The peace it has known
It can never regain.
CHORUS
Join the dance, with step as light
As ev'ry heart should be to-night;
Music, shake the lofty dome,
In honour of our Harvest Home.
Join the dance, and banish care,
All are young, and gay, and fair;
Even age has youthful grown,
In honour of our Harvest Home.
Join the dance, bright faces beam,
Sweet lips smile, and dark eyes gleam;
All these charms have hither come,
In honour of our Harvest Home.
Join the dance, with step as light,
As ev'ry heart should be to-night;
Music shake the lofty dome
In honour of our Harvest Home.
QUINTET
No light bound
Of stag or timid hare,
O'er the ground
Where startled herds repair,
Do we prize
So high, or hold so dear,
As the eyes
That light our pleasures here.
No cool breeze
That gently plays by night,
O'er calm seas,
Whose waters glisten bright;
No soft moan
That sighs across the lea,
Harvest Home,
Is half so sweet as thee!
LYRIC FROM 'THE LAMPLIGHTER'
A FARCE
1838
THE LAMPLIGHTER
In 1838 Dickens agreed to prepare a little play for Macready, the famous
actor, then the manager of Drury Lane Theatre. It was called _The
Lamplighter_, and when completed the author read aloud the 'unfortunate
little farce' (as he subsequently termed it) in the greenroom of the
theatre. Although the play went through rehearsal, it was never presented
before an audience, for the actors would not agree about it, and, at
Macready's suggestion, Dickens consented to withdraw it, declaring that he
had 'no other feeling of disappointment connected with this matter' but
that which arose from the failure in attempting to serve his friend. The
manuscript of the play, not in Dickens's handwriting, reposes in the
Forster Library at the Victoria and Albert Museum, and in 1879 it was
printed for the first time, in the form of a pamphlet, of which only two
hundred and fifty copies were issued.
When rejected by Macready as unsuitable for stage presentation, _The
Lamplighter_ was adapted by Dickens to another purpose--that is to say,
he converted it into a tale called _The Lamplighter's Story_, for
publication in _The Pic-Nic Papers_, issued in 1841 for the benefit of the
widow of Macrone, Dickens's first publisher, who died in great poverty.
Between the farce and the story there are but slight differences. The duet
of two verses, sung by Tom and Betsy to the air of 'The Young May-moon,'
cannot of course be regarded as a remarkable composition, but it served
its purpose sufficiently well, and for that reason deserves recognition.
DUET FROM 'THE LAMPLIGHTER'
AIR--'THE YOUNG MAY-MOON'
_Tom._ There comes a new moon twelve times a year.
_Betsy._ And when there is none, all is dark and drear.
_Tom._ In which I espy--
_Betsy._ And so, too, do I--
_Both._ A resemblance to womankind very clear--
_Both._ There comes a new moon twelve times a year;
And when there is none, all is dark and drear.
_Tom._ In which I espy--
_Betsy._ And so do I--
_Both._ A resemblance to womankind very clear.
_Second Verse._
_Tom._ She changes, she's fickle, she drives men mad.
_Betsy._ She comes to bring light, and leaves them sad.
_Tom._ So restless wild--
_Betsy._ But so sweetly wild--
_Both._ That no better companion could be had.
_Both._ There comes a new moon twelve times a year;
And when there is none, all is dark and drear.
_Tom._ In which I espy--
_Betsy._ And so do I--
_Both._ A resemblance to womankind very clear.
SONGS FROM 'THE PICKWICK PAPERS'
1837
I.--THE IVY GREEN
THE IVY GREEN
This famous ballad of three verses, from the sixth chapter of _Pickwick_,
is perhaps the most acceptable of all Dickens's poetical efforts. It was
originally set to music, at Dickens's request, by his brother-in-law,
Henry Burnett, a professional vocalist, who, by the way, was the admitted
prototype of Nicholas Nickleby. Mr. Burnett sang the ballad scores of
times in the presence of literary men and artists, and it proved an
especial favourite with Landor. 'The Ivy Green' was not written for
_Pickwick_, Mr. Burnett assured me; but on its being so much admired the
author said it should go into a monthly number, and it did. The most
popular setting is undoubtedly that of Henry Russell, who has recorded
that he received, as his fee, the magnificent sum of ten shillings! The
ballad, in this form, went into many editions, and the sales must have
amounted to tens of thousands.
THE IVY GREEN
Oh, a dainty plant is the Ivy green,
That creepeth o'er ruins old!
Of right choice food are his meals, I ween,
In his cell so lone and cold.
The wall must be crumbled, the stone decayed,
To pleasure his dainty whim:
And the mouldering dust that years have made
Is a merry meal for him.
Creeping where no life is seen,
A rare old plant is the Ivy green.
Fast he stealeth on, though he wears no wings,
And a staunch old heart has he.
How closely he twineth, how tight he clings,
To his friend the huge Oak Tree!
And slily he traileth along the ground,
And his leaves he gently waves,
As he joyously hugs and crawleth round
The rich mould of dead men's graves.
Creeping where grim death hath been,
A rare old plant is the Ivy green.
Whole ages have fled and their works decayed,
And nations have scattered been;
But the stout old Ivy shall never fade,
From its hale and hearty green.
The brave old plant, in its lonely days,
Shall fatten upon the past:
For the stateliest building man can raise
Is the Ivy's food at last.
Creeping on, where time has been,
A rare old plant is the Ivy green.
II.--A CHRISTMAS CAROL
A CHRISTMAS CAROL
The five stanzas bearing the above title will be found in the
twenty-eighth chapter of _Pickwick_, where they are introduced as the song
which that hospitable old soul, Mr. Wardle, sung appropriately, 'in a
good, round, sturdy voice,' before the Pickwickians and others assembled
on Christmas Eve at Manor Farm. The 'Carol,' shortly after its appearance
in _Pickwick_, was set to music to the air of 'Old King Cole,' and
published in _The Book of British Song_ (New Edition), with an
illustration drawn by 'Alfred Crowquill'--_i.e._, A. H. Forrester.
A CHRISTMAS CAROL
I care not for Spring; on his fickle wing
Let the blossoms and buds be borne:
He woos them amain with his treacherous rain,
And he scatters them ere the morn.
An inconstant elf, he knows not himself
Nor his own changing mind an hour,
He'll smile in your face, and, with wry grimace,
He'll wither your youngest flower.
Let the Summer sun to his bright home run,
He shall never be sought by me;
When he's dimmed by a cloud I can laugh aloud,
And care not how sulky he be!
For his darling child is the madness wild
That sports in fierce fever's train;
And when love is too strong, it don't last long,
As many have found to their pain.
A mild harvest night, by the tranquil light
Of the modest and gentle moon,
Has a far sweeter sheen, for me, I ween,
Than the broad and unblushing noon.
But every leaf awakens my grief,
As it lieth beneath the tree;
So let Autumn air be never so fair,
It by no means agrees with me.
But my song I troll out, for CHRISTMAS stout,
The hearty, the true, and the bold;
A bumper I drain, and with might and main
Give three cheers for this Christmas old!
We'll usher him in with a merry din
That shall gladden his joyous heart,
And we'll keep him up, while there's bite or sup,
And in fellowship good, we'll part.
In his fine honest pride, he scorns to hide
One jot of his hard-weather scars;
They're no disgrace, for there's much the same trace
On the cheeks of our bravest tars.
Then again I sing 'till the roof doth ring,
And it echoes from wall to wall--
To the stout old wight, fair welcome to-night,
As the King of the Seasons all!
III.--GABRIEL GRUB'S SONG
GABRIEL GRUB'S SONG
The Sexton's melancholy dirge, in the twenty-ninth chapter of _Pickwick_,
seems a little incongruous in a humorous work. The sentiment, however,
thoroughly accords with the philosophic gravedigger's gruesome occupation.
'The Story of the Goblins who Stole a Sexton' is one of several short
tales (chiefly of a dismal character) introduced into _Pickwick_; they
were doubtless written prior to the conception of _Pickwick_, each being
probably intended for independent publication, and in a manner similar to
the 'Boz' Sketches. For some reason these stories were not so published,
and Dickens evidently saw a favourable opportunity of utilising his unused
manuscripts by inserting them in _The Pickwick Papers_.
GABRIEL GRUB'S SONG
Brave lodgings for one, brave lodgings for one,
A few feet of cold earth, when life is done;
A stone at the head, a stone at the feet,
A rich, juicy meal for the worms to eat;
Rank grass over head, and damp clay around,
Brave lodgings for one, these, in holy ground!
IV.--ROMANCE
ROMANCE
It will be remembered that while Sam Weller and his coaching-friends
refreshed themselves at the little public-house opposite the Insolvent
Court in Portugal Street, Lincoln's Inn Fields, prior to Sam joining Mr.
Pickwick in the Fleet, that faithful body-servant was persuaded to 'oblige
the company' with a song. 'Raly, gentlemen,' said Sam, 'I'm not wery much
in the habit o' singin' vithout the instrument; but anythin' for a quiet
life, as the man said ven he took the sitivation at the light-house.'
'With this prelude, Mr. Samuel Weller burst at once into the following
wild and beautiful legend, which, under the impression that it is not
generally known, we take the liberty of quoting. We would beg to call
particular attention to the monosyllable at the end of the second and
fourth lines, which not only enables the singer to take breath at those
points, but greatly assists the metre.'-_The Pickwick Papers_, chapter
xliii.
At the conclusion of the performance the mottled-faced gentleman contended
that the song was 'personal to the cloth,' and demanded the name of the
bishop's coachman, whose cowardice he regarded as a reflection upon
coachmen in general. Sam replied that his name was not known, as 'he
hadn't got his card in his pocket'; whereupon the mottled-faced gentleman
declared the statement to be untrue, stoutly maintaining that the said
coachman did _not_ run away, but 'died game--game as pheasants,' and he
would 'hear nothin' said to the contrairey.'
Even in the vernacular (observes Mr. Percy Fitzgerald), 'this master of
words [Charles Dickens] could be artistic; and it may fairly be asserted
that Mr. Weller's song to the coachmen is superior to anything of the kind
that has appeared since.' The two stanzas have been set to music, as a
humorous part-song, by Sir Frederick Bridge, Mus. Doc., M.V.O., the
organist of Westminster Abbey, who informs me that it was written some
years since, to celebrate a festive gathering in honour of Dr. Turpin (!),
Secretary of the College of Organists. 'It has had a very great success,'
says Sir Frederick, 'and is sung much in the North of England at
competitions of choirs. It is for men's voices. The humour of the words
never fails to make a great hit, and I hope the music does no harm. "The
Bishop's Coach" is set to a bit of old Plain-Chant, and I introduce a
Fugue at the words "Sure as eggs is eggs."'
ROMANCE
I
Bold Turpin vunce, on Hounslow Heath,
His bold mare Bess bestrode--er;
Ven there he see'd the Bishop's coach
A-comin' along the road--er.
So he gallops close to the 'orse's legs,
And he claps his head vithin;
And the Bishop says, 'Sure as eggs is eggs,
This here's the bold Turpin!'
_Chorus_--And the Bishop says, 'Sure as eggs is eggs,
This here's the bold Turpin!'
II
Says Turpin, 'You shall eat your words,
With a sarse of leaden bul-let';
So he puts a pistol to his mouth,
And he fires it down his gul-let.
The coachman, he not likin' the job,
Set off at a full gal-lop,
But Dick put a couple of balls in his nob,
And perwailed on him to stop.
_Chorus_ (_sarcastically_)--But Dick put a couple of balls in his
nob,
And perwailed on him to stop.
POLITICAL SQUIBS FROM 'THE EXAMINER' 1841
I.--THE FINE OLD ENGLISH GENTLEMAN
POLITICAL SQUIBS FROM 'THE EXAMINER,' 1841
In August 1841 Dickens contributed anonymously to _The Examiner_ (then
edited by Forster) three political squibs, which were signed W., and were
intended to help the Liberals in fighting their opponents. These squibs
were entitled respectively 'The Fine Old English Gentleman (to be said or
sung at all Conservative Dinners)'; 'The Quack Doctor's Proclamation'; and
'Subjects for Painters (after Peter Pindar).' Concerning those
productions, Forster says: 'I doubt if he ever enjoyed anything more than
the power of thus taking part occasionally, unknown to outsiders, in the
sharp conflict the press was waging at the time.' In all probability he
contributed other political rhymes to the pages of _The Examiner_ as
events prompted: if so, they are buried beyond easy reach of
identification.
Writing to Forster at this time, Dickens said: 'By Jove, how Radical I am
getting! I wax stronger and stronger in the true principles every day.'...
He would (observes Forster) sometimes even talk, in moments of sudden
indignation at the political outlook, 'of carrying off himself and his
household gods, like Coriolanus, to a world elsewhere.' This was the
period of the Tory interregnum, with Sir Robert Peel at the head of
affairs.
THE FINE OLD ENGLISH GENTLEMAN
NEW VERSION
(_To be said or sung at all Conservative Dinners_)
I'll sing you a new ballad, and I'll warrant it first-rate,
Of the days of that old gentleman who had that old estate;
When they spent the public money at a bountiful old rate
On ev'ry mistress, pimp, and scamp, at ev'ry noble gate,
In the fine old English Tory times;
Soon may they come again!
The good old laws were garnished well with gibbets, whips, and chains,
With fine old English penalties, and fine old English pains,
With rebel heads, and seas of blood once hot in rebel veins;
For all these things were requisite to guard the rich old gains
Of the fine old English Tory times;
Soon may they come again!
This brave old code, like Argus, had a hundred watchful eyes,
And ev'ry English peasant had his good old English spies,
To tempt his starving discontent with fine old English lies,
Then call the good old Yeomanry to stop his peevish cries,
In the fine old English Tory times;
Soon may they come again!
The good old times for cutting throats that cried out in their need,
The good old times for hunting men who held their fathers' creed,
The good old times when William Pitt, as all good men agreed,
Came down direct from Paradise at more than railroad speed....
Oh the fine old English Tory times;
When will they come again!
In those rare days, the press was seldom known to snarl or bark,
But sweetly sang of men in pow'r, like any tuneful lark;
Grave judges, too, to all their evil deeds were in the dark;
And not a man in twenty score knew how to make his mark.
Oh the fine old English Tory times;
Soon may they come again!
Those were the days for taxes, and for war's infernal din;
For scarcity of bread, that fine old dowagers might win;
For shutting men of letters up, through iron bars to grin,
Because they didn't think the Prince was altogether thin,
In the fine old English Tory times;
Soon may they come again!
But Tolerance, though slow in flight, is strong-wing'd in the main;
That night must come on these fine days, in course of time was plain;
The pure old spirit struggled, but its struggles were in vain;
A nation's grip was on it, and it died in choking pain,
With the fine old English Tory days,
All of the olden time.
The bright old day now dawns again; the cry runs through the land,
In England there shall be dear bread--in Ireland, sword and brand;
And poverty, and ignorance, shall swell the rich and grand,
So, rally round the rulers with the gentle iron hand,
Of the fine old English Tory days;
Hail to the coming time!
W.
II.--THE QUACK DOCTOR'S PROCLAMATION
THE QUACK DOCTOR'S PROCLAMATION
TUNE--'A COBBLER THERE WAS'
An astonishing doctor has just come to town,
Who will do all the faculty perfectly brown:
He knows all diseases, their causes, and ends;
And he begs to appeal to his medical friends.
Tol de rol:
Diddle doll:
Tol de rol, de dol,
Diddle doll
Tol de rol doll.
He's a magnetic doctor, and knows how to keep
The whole of a Government snoring asleep
To popular clamours; till popular pins
Are stuck in their midriffs--and then he begins
Tol de rol.
He's a _clairvoyant_ subject, and readily reads
His countrymen's wishes, condition, and needs,
With many more fine things I can't tell in rhyme,
--And he keeps both his eyes shut the whole of the time.
Tol de rol.
You mustn't expect him to talk; but you'll take
Most particular notice the doctor's awake,
Though for aught from his words or his looks that you reap, he
Might just as well be most confoundedly sleepy.
Tol de rol.
Homoeopathy, too, he has practised for ages
(You'll find his prescriptions in Luke Hansard's pages),
Just giving his patient when maddened by pain,--
Of Reform the ten thousandth part of a grain.
Tol de rol.
He's a med'cine for Ireland, in portable papers;
The infallible cure for political vapours;
A neat label round it his 'prentices tie--
'Put your trust in the Lord, and keep this powder dry!'
Tol de rol.
He's a corn doctor also, of wonderful skill,
--No cutting, no rooting-up, purging, or pill--
You're merely to take, 'stead of walking or riding,
The sweet schoolboy exercise--innocent sliding.
Tol de rol.
There's no advice gratis. If high ladies send
His legitimate fee, he's their soft-spoken friend.
At the great public counter with one hand behind him,
And one in his waistcoat, they're certain to find him.
Tol de rol.
He has only to add he's the real Doctor Flam,
All others being purely fictitious and sham;
The house is a large one, tall, slated, and white,
With a lobby; and lights in the passage at night.
Tol de rol:
Diddle doll:
Tol de rol, de dol,
Diddle doll
Tol de rol doll.
W.
III.--SUBJECTS FOR PAINTERS
SUBJECTS FOR PAINTERS
(AFTER PETER PINDAR)
To you, SIR MARTIN,[1] and your co. R.A.'s,
I dedicate in meek, suggestive lays,
Some subjects for your academic palettes;
Hoping, by dint of these my scanty jobs,
To fill with novel thoughts your teeming nobs,
As though I beat them in with wooden mallets.
To you, MACLISE, who Eve's fair daughters paint
With Nature's hand, and want the maudlin taint
Of the sweet Chalon school of silk and ermine:
To you, E. LANDSEER, who from year to year
Delight in beasts and birds, and dogs and deer,
And seldom give us any human vermin:
--To all who practise art, or make believe,
I offer subjects they may take or leave.
Great Sibthorp and his butler, in debate
(_Arcades ambo_) on affairs of state,
Not altogether 'gone,' but rather funny;
Cursing the Whigs for leaving in the lurch
Our d----d good, pleasant, gentlemanly Church,
Would make a picture--cheap at any money.
Or Sibthorp as the Tory Sec.--at-War,
Encouraging his mates with loud 'Yhor! Yhor!
From Treas'ry benches' most conspicuous end;
Or Sib.'s mustachios curling with a smile,
As an expectant Premier without guile
Calls him his honourable and gallant friend.
Or Sibthorp travelling in foreign parts,
Through that rich portion of our Eastern charts
Where lies the land of popular tradition;
And fairly worshipp'd by the true devout
In all his comings-in and goings-out,
Because of the old Turkish superstition.
Fame with her trumpet, blowing very hard,
And making earth rich with celestial lard,
In puffing deeds done through Lord Chamberlain Howe;
While some few thousand persons of small gains,
Who give their charities without such pains,
Look up, much wondering what may be the row.
Behind them Joseph Hume, who turns his pate
To where great Marlbro' House in princely state
Shelters a host of lacqueys, lords and pages,
And says he knows of dowagers a crowd,
Who, without trumpeting so very loud,
Would do so much, and more, for half the wages.
Limn, sirs, the highest lady in the land,
When Joseph Surface, fawning cap in hand,
Delivers in his list of patriot mortals;
Those gentlemen of honour, faith, and truth,
Who, foul-mouthed, spat upon her maiden youth,
And dog-like did defile her palace portals.
Paint me the Tories, full of grief and woe,
Weeping (to voters) over Frost and Co.,
Their suff'ring, erring, much-enduring brothers.
And in the background don't forget to pack,
Each grinning ghastly from its bloody sack,
The heads of Thistlewood, Despard, and others.
Paint, squandering the club's election gold,
Fierce lovers of our Constitution old,
Lords who're that sacred lady's greatest debtors;
And let the law, forbidding any voice
Or act of Peer to influence the choice
Of English people, flourish in bright letters.
Paint that same dear old lady, ill at ease,
Weak in her second childhood, hard to please,
Unknowing what she ails or what she wishes;
With all her Carlton nephews at the door,
Deaf'ning both aunt and nurses with their roar,
--Fighting already, for the loaves and fishes.
Leaving these hints for you to dwell upon,
I shall presume to offer more anon.
W.
PROLOGUE TO WESTLAND MARSTON'S PLAY 'THE PATRICIAN'S DAUGHTER'
1842
PROLOGUE TO 'THE PATRICIAN'S DAUGHTER'
_The Patrician's Daughter_ was the title bestowed upon a play, in the
tragic vein, by a then unknown writer, J. Westland Marston, it being his
maiden effort in dramatic authorship. Dickens took great interest in the
young man and indicated a desire to promote the welfare of his production
by composing some introductory lines. To Macready he wrote: 'The more I
think of Marston's play, the more sure I feel that a prologue to the
purpose would help it materially, and almost decide the fate of any
ticklish point on the first night. Now I have an idea (not easily
explainable in writing, but told in five words) that would take the
prologue out of the conventional dress of prologues, quite. Get the
curtain up with a dash, and begin the play with a sledge-hammer blow. If,
on consideration, you should agree with me, I will write the prologue,
heartily.' Happily for the author, his little tragedy was the first new
play of the season, and it thus attracted greater attention. Its initial
representation took place at Drury Lane Theatre on December 10, 1842, and
the fact that Dickens's dignified and vigorous lines were recited by
Macready, the leading actor of his day, undoubtedly gave _prestige_ to
this performance; but the play, although it made a sensation for the
moment, did not enjoy a long run, its motive being for some reason
misunderstood. As explained by the Editors of _The Letters of Charles
Dickens_, it was (to a certain extent) an experiment in testing the effect
of a tragedy of modern times and in modern dress, the novelist's Prologue
being intended to show that there need be no incongruity between plain
clothes of the nineteenth century and high tragedy.
_The Patrician's Daughter: A Tragedy in Five Acts_, appeared in pamphlet
form during the year prior to its being placed upon the boards. The
Prologue was printed for the first time in the _Sunday Times_, December
11, 1842, and then in _The Theatrical Journal and Stranger's Guide_,
December 17, 1842. By the kind permission of Miss Hogarth, the lines are
here reproduced from the revised and only correct version in _The Letters
of Charles Dickens_.
In the preface to the second edition of the play (1842), the author thus
acknowledges his indebtedness to Dickens for the Prologue, which, however,
does not appear in the book: 'How shall I thank Mr. Dickens for the
spontaneous kindness which has furnished me with so excellent a letter of
introduction to the audience? The simplest acknowledgment is perhaps the
best, since the least I might say would exceed _his_ estimate of the
obligation; while the most I could say would fail to express _mine_.'
PROLOGUE TO 'THE PATRICIAN'S DAUGHTER'
(SPOKEN BY MR. MACREADY)
No tale of streaming plumes and harness bright
Dwells on the poet's maiden harp to-night;
No trumpet's clamour and no battle's fire
Breathes in the trembling accents of his lyre;
Enough for him, if in his lowly strain
He wakes one household echo not in vain;
Enough for him, if in his boldest word
The beating heart of MAN be dimly heard.
Its solemn music which, like strains that sigh
Through charmed gardens, all who hearing die;
Its solemn music he does not pursue
To distant ages out of human view;
Nor listen to its wild and mournful chime
In the dead caverns on the shore of Time;
But musing with a calm and steady gaze
Before the crackling flames of living days,
He hears it whisper through the busy roar
Of what shall be and what has been before.
Awake the Present! Shall no scene display
The tragic passion of the passing day?
Is it with Man, as with some meaner things,
That out of death his single purpose springs?
Can his eventful life no moral teach
Until he be, for aye, beyond its reach?
Obscurely shall he suffer, act, and fade,
Dubb'd noble only by the sexton's spade?
Awake the Present! Though the steel-clad age
Find life alone within its storied page,
Iron is worn, at heart, by many still--
The tyrant Custom binds the serf-like will;
If the sharp rack, and screw, and chain be gone,
These later days have tortures of their own;
The guiltless writhe, while Guilt is stretch'd in sleep,
And Virtue lies, too often, dungeon deep.
Awake the Present! what the Past has sown
Be in its harvest garner'd, reap'd, and grown!
How pride breeds pride, and wrong engenders wrong,
Read in the volume Truth has held so long,
Assured that where life's flowers freshest blow,
The sharpest thorns and keenest briars grow,
How social usage has the pow'r to change
Good thoughts to evil; in its highest range
To cramp the noble soul, and turn to ruth
The kindling impulse of our glorious youth,
Crushing the spirit in its house of clay,
Learn from the lessons of the present day.
Not light its import and not poor its mien;
Yourselves the actors, and your homes the scene.
A WORD IN SEASON FROM THE 'KEEPSAKE'
1844
A WORD IN SEASON
_The Keepsake_, one of the many fashionable annuals published during the
early years of Queen Victoria's reign, had for its editor in 1844 the
'gorgeous' Countess of Blessington, the reigning beauty who held court at
Gore House, Kensington, where many political, artistic, and literary
celebrities forgathered--Bulwer Lytton, Disraeli, Dickens, Ainsworth,
D'Orsay, and the rest. Her ladyship, through her personal charm and
natural gifts, succeeded in securing the services of eminent authors for
the aristocratic publication; even Dickens could not resist her appeal,
and in a letter to Forster (dated July 1843) he wrote: 'I have heard, as
you have, from Lady Blessington, for whose behalf I have this morning
penned the lines I send you herewith. But I have only done so to excuse
myself, for I have not the least idea of their suiting her; and I hope she
will send them back to you for _The Examiner_.' Lady Blessington, however,
decided to retain the thoughtful little poem, which was referred to in
the _London Review_ (twenty-three years later) as 'a graceful and sweet
apologue, reminding one of the manner of Hood.' The theme of the poem,
which Forster describes as 'a clever and pointed parable in verse,' was
afterwards satirised in Chadband (_Bleak House_), and in the idea of
religious conversion through the agency of 'moral pocket-handkerchiefs.'
A WORD IN SEASON
They have a superstition in the East,
That ALLAH, written on a piece of paper,
Is better unction than can come of priest,
Of rolling incense, and of lighted taper:
Holding, that any scrap which bears that name,
In any characters, its front imprest on,
Shall help the finder through the purging flame,
And give his toasted feet a place to rest on.
Accordingly, they make a mighty fuss
With ev'ry wretched tract and fierce oration,
And hoard the leaves--for they are not, like us,
A highly civilized and thinking nation:
And, always stooping in the miry ways,
To look for matter of this earthy leaven,
They seldom, in their dust-exploring days,
Have any leisure to look up to Heaven.
So have I known a country on the earth,
Where darkness sat upon the living waters,
And brutal ignorance, and toil, and dearth
Were the hard portion of its sons and daughters:
And yet, where they who should have ope'd the door
Of charity and light, for all men's finding,
Squabbled for words upon the altar-floor,
And rent the Book, in struggles for the binding.
The gentlest man among these pious Turks,
God's living image ruthlessly defaces;
Their best high-churchman, with no faith in works,
Bowstrings the Virtues in the market-places:
The Christian Pariah, whom both sects curse
(They curse all other men, and curse each other),
Walks thro' the world, not very much the worse--
Does all the good he can, and loves his brother.
VERSES FROM THE 'DAILY NEWS'
1846
I.--THE BRITISH LION
VERSES FROM THE 'DAILY NEWS,' 1846
The _Daily News_, it will be remembered, was founded in January 1846 by
Charles Dickens, who officiated as its first editor. He soon sickened of
the mechanical drudgery appertaining to the position, and resigned his
editorial functions the following month. From January 21st to March 2nd he
contributed to its columns a series of 'Travelling Sketches,' afterwards
reprinted in volume form as _Pictures from Italy_. He also availed himself
of the opportunity afforded him, by his association with that newspaper,
of once more taking up the cudgels against the Tories, and, as in the case
of the _Examiner_, his attack was conveyed through the medium of some
doggerel verses. These were entitled 'The British Lion--A New Song, but an
Old Story,' to be sung to the tune of 'The Great Sea-Snake.' They bore the
signature of 'Catnach,' the famous ballad-singer, and were printed in the
_Daily News_ of January 24, 1846.
Three weeks later some verses of a totally different character appeared in
the columns of the _Daily News_, signed in full 'Charles Dickens.' One
Lucy Simpkins, of Bremhill (or Bremble), a parish in Wiltshire, had just
previously addressed a night meeting of the wives of agricultural
labourers in that county, in support of a petition for Free Trade, and her
vigorous speech on that occasion inspired Dickens to write 'The Hymn of
the Wiltshire Labourers,' thus offering an earnest protest against
oppression. Concerning the 'Hymn,' a writer in a recent issue of
_Christmas Bells_ observes: 'It breathes in every line the teaching of the
Sermon on the Mount, the love of the All-Father, the Redemption by His
Son, and that love to God and man on which hang all the law and the
prophets.'
THE BRITISH LION
A NEW SONG, BUT AN OLD STORY
TUNE--'THE GREAT SEA-SNAKE'
Oh, p'r'aps you may have heard, and if not, I'll sing
Of the British Lion free,
That was constantly a-going for to make a spring
Upon his en-e-me;
But who, being rather groggy at the knees,
Broke down, always, before;
And generally gave a feeble wheeze
Instead of a loud roar.
Right toor rol, loor rol, fee faw fum,
The British Lion bold!
That was always a-going for to do great things,
And was always being 'sold!'
He was carried about, in a carawan,
And was show'd in country parts,
And they said, 'Walk up! Be in time! He can
Eat Corn-Law Leagues like tarts!'
And his showmen, shouting there and then,
To puff him didn't fail,
And they said, as they peep'd into his den,
'Oh, don't he wag his tail!'
Now, the principal keeper of this poor old beast,
WAN HUMBUG was his name,
Would once ev'ry day stir him up--at least--
And wasn't that a Game!
For he hadn't a tooth, and he hadn't a claw,
In that 'Struggle' so 'Sublime';
And, however sharp they touch'd him on the raw,
He couldn't come up to time.
And this, you will observe, was the reason why
WAN HUMBUG, on weak grounds,
Was forced to make believe that he heard his cry
In all unlikely sounds.
So, there wasn't a bleat from an Essex Calf,
Or a Duke, or a Lordling slim;
But he said, with a wery triumphant laugh,
'I'm blest if that ain't him.'
At length, wery bald in his mane and tail,
The British Lion growed:
He pined, and declined, and he satisfied
The last debt which he owed.
And when they came to examine the skin,
It was a wonder sore,
To find that the an-i-mal within
Was nothing but a Boar!
Right toor rol, loor rol, fee faw fum,
The British Lion bold!
That was always a-going for to do great things,
And was always being 'sold!'
CATNACH.
II. THE HYMN OF THE WILTSHIRE LABOURERS
THE HYMN OF THE WILTSHIRE LABOURERS
'Don't you all think that we have a great need to Cry to our God to put it
in the hearts of our greassous Queen and her Members of Parlerment to
grant us free bread!'
LUCY SIMPKINS, _at Bremhill_.
Oh GOD, who by Thy Prophet's hand
Didst smite the rocky brake,
Whence water came, at Thy command,
Thy people's thirst to slake;
Strike, now, upon this granite wall,
Stern, obdurate, and high;
And let some drops of pity fall
For us who starve and die!
The GOD, who took a little child,
And set him in the midst,
And promised him His mercy mild,
As, by Thy Son, Thou didst:
Look down upon our children dear,
So gaunt, so cold, so spare,
And let their images appear
Where Lords and Gentry are!
Oh GOD, teach them to feel how we,
When our poor infants droop,
Are weakened in our trust in Thee,
And how our spirits stoop;
For, in Thy rest, so bright and fair,
All tears and sorrows sleep:
And their young looks, so full of care,
Would make Thine Angels weep!
The GOD, who with His finger drew
The Judgment coming on,
Write, for these men, what must ensue,
Ere many years be gone!
Oh GOD, whose bow is in the sky,
Let them not brave and dare,
Until they look (too late) on high,
And see an Arrow there!
Oh GOD, remind them! In the bread
They break upon the knee,
These sacred words may yet be read,
'In memory of Me!'
Oh GOD, remind them of His sweet
Compassion for the poor,
And how He gave them Bread to eat,
And went from door to door!
CHARLES DICKENS.
NEW SONG LINES ADDRESSED TO MARK LEMON
1849
NEW SONG
Dickens, like Silas Wegg, would sometimes 'drop into poetry' when writing
to intimate friends, as, for example, in a letter to Maclise, the artist,
which began with a parody of Byron's lines to Thomas Moore--
'My foot is in the house,
My bath is on the sea,
And, before I take a souse,
Here's a single note to thee.'
A more remarkable instance of his propensity to indulge in parody of this
kind is to be found in a letter addressed to Mark Lemon in the spring of
1849. The novelist was then enjoying a holiday with his wife and daughters
at Brighton, whence he wrote to Lemon (who had been ill), pressing him to
pay them a visit. After commanding him to 'get a clean pocket-handkerchief
ready for the close of "Copperfield" No. 3--"simple and quiet, but very
natural and touching"--_Evening Bore_,' Dickens invites his friend in
lines headed 'New Song,' and signed 'T. Sparkler,' the effusion also
bearing the signatures of other members of the family party--Catherine
Dickens, Annie Leech, Georgina Hogarth, Mary Dickens, Katie Dickens, and
John Leech.
NEW SONG
TUNE--'LESBIA HATH A BEAMING EYE'
I
Lemon is a little hipped,
And this is Lemon's true position--
He is not pale, he's not white-lipped,
Yet wants a little fresh condition.
Sweeter 'tis to gaze upon
Old Ocean's rising, falling billers,
Than on the Houses every one
That form the street called Saint Anne's Willers!
Oh my Lemon, round and fat,
Oh my bright, my right, my tight 'un,
Think a little what you're at--
Don't stay at home, but come to Brighton!
II
Lemon has a coat of frieze,
But all so seldom Lemon wears it,
That it is a prey to fleas,
And ev'ry moth that's hungry, tears it.
Oh, that coat's the coat for me,
That braves the railway sparks and breezes,
Leaving ev'ry engine free
To smoke it, till its owner sneezes!
Then my Lemon, round and fat,
L., my bright, my right, my tight 'un,
Think a little what you're at--
On Tuesday first, come down to Brighton!
T. SPARKLER.
WILKIE COLLINS'S PLAY 'THE LIGHTHOUSE'
1855
I.--THE PROLOGUE
'THE LIGHTHOUSE'
Wilkie Collins composed two powerful dramas for representation at
Dickens's residence, Tavistock House, a portion of which had been already
adapted for private theatricals, the rooms so converted being described in
the bills as 'The Smallest Theatre in the World.' The first of these plays
was called _The Lighthouse_, and the initial performance took place on
June 19, 1855. Dickens not only wrote the Prologue and 'The Song of the
Wreck,' but signally distinguished himself by enacting the part of Aaron
Gurnock, a lighthouse-keeper, his clever impersonation recalling Frederick
Lemaitre, the only actor he ever tried to take as a model.
With regard to 'The Song of the Wreck,' Dickens evidently intended to
bestow upon it a different title, for, in a letter addressed to Wilkie
Collins during the preparation of the play, he said: 'I have written a
little ballad for Mary--"The Story of the Ship's Carpenter and the Little
Boy, in the Shipwreck."' The song was rendered by his eldest daughter,
Mary (who assumed the role of Phoebe in the play); it was set to the
music composed by George Linley for Miss Charlotte Young's pretty ballad,
'Little Nell,' of which Dickens became very fond, and which his daughter
had been in the habit of singing to him constantly since her childhood.
Dr. A. W. Ward, Master of Peter-house, Cambridge University, refers to
'The Song of the Wreck' as 'a most successful effort in Cowper's
manner.'
THE PROLOGUE
(_Slow music all the time; unseen speaker; curtain down._)
A story of those rocks where doom'd ships come
To cast them wreck'd upon the steps of home,
Where solitary men, the long year through--
The wind their music and the brine their view--
Warn mariners to shun the beacon-light;
A story of those rocks is here to-night.
Eddystone Lighthouse!
(_Exterior view discovered._)
In its ancient form,
Ere he who built it wish'd for the great storm
That shiver'd it to nothing,[2] once again
Behold outgleaming on the angry main!
Within it are three men; to these repair
In our frail bark of Fancy, swift as air!
They are but shadows, as the rower grim
Took none but shadows in his boat with him.
So be _ye_ shades, and, for a little space,
The real world a dream without a trace.
Return is easy. It will have ye back
Too soon to the old beaten dusty track;
For but one hour forget it. Billows, rise;
Blow winds, fall rain, be black, ye midnight skies;
And you who watch the light, arise! arise!
(_Exterior view rises and discovers the scene._)
II.--THE SONG OF THE WRECK
THE SONG OF THE WRECK
I
The wind blew high, the waters raved,
A ship drove on the land,
A hundred human creatures saved
Kneel'd down upon the sand.
Three-score were drown'd, three-score were thrown
Upon the black rocks wild,
And thus among them, left alone,
They found one helpless child.
II
A seaman rough, to shipwreck bred,
Stood out from all the rest,
And gently laid the lonely head
Upon his honest breast.
And travelling o'er the desert wide
It was a solemn joy,
To see them, ever side by side,
The sailor and the boy.
III
In famine, sickness, hunger, thirst,
The two were still but one,
Until the strong man droop'd the first
And felt his labours done.
Then to a trusty friend he spake,
'Across the desert wide,
O take this poor boy for my sake!'
And kiss'd the child and died.
IV
Toiling along in weary plight
Through heavy jungle, mire,
These two came later every night
To warm them at the fire.
Until the captain said one day,
'O seaman good and kind,
To save thyself now come away,
And leave the boy behind!'
V
The child was slumbering near the blaze:
'O captain, let him rest
Until it sinks, when God's own ways
Shall teach us what is best!'
They watch'd the whiten'd ashy heap,
They touch'd the child in vain;
They did not leave him there asleep,
He never woke again.
PROLOGUE TO WILKIE COLLINS'S PLAY 'THE FROZEN DEEP'
1856
'THE FROZEN DEEP'
The second drama written by Wilkie Collins for the Tavistock House Theatre
was first acted there in January 1857, and subsequently at the Gallery of
Illustration in the presence of Queen Victoria and the Royal Family. As in
the case of _The Lighthouse_, the play had the advantage of a Prologue in
rhyme by Charles Dickens, who again electrified his audiences by
marvellous acting, the character of Richard Wardour (a young naval
officer) being selected by him for representation.
The Prologue was recited at Tavistock House by John Forster, and at the
public performances of the play by Dickens himself.
It is not generally known that a by no means inconsiderable portion of the
drama was composed by Dickens, as testified by the original manuscripts of
the play and of the prompt-book, which contain numerous additions and
corrections in his handwriting. These manuscripts, by the way, realised
L300 at Sotheby's in 1890.
The main idea of _A Tale of Two Cities_ was conceived by Dickens when
performing in _The Frozen Deep_. 'A strong desire was upon me then,' he
writes in the preface to the story, 'to embody it in my own person; and I
traced out in my fancy the state of mind of which it would necessitate the
presentation to an observant spectator, with particular care and interest.
As the idea became familiar to me, it gradually shaped itself into its
present form. Throughout its execution, it has had complete possession of
me: I have so far verified what is done and suffered in these pages, as
that I have certainly done and suffered it all myself.'
PROLOGUE TO 'THE FROZEN DEEP'
(_Curtain rises; mists and darkness; soft music throughout._)
One savage footprint on the lonely shore
Where one man listen'd to the surge's roar,
Not all the winds that stir the mighty sea
Can ever ruffle in the memory.
If such its interest and thrall, O then
Pause on the footprints of heroic men,
Making a garden of the desert wide
Where Parry conquer'd death and Franklin died.
To that white region where the Lost lie low,
Wrapt in their mantles of eternal snow,--
Unvisited by change, nothing to mock
Those statues sculptured in the icy rock,
We pray your company; that hearts as true
(Though nothings of the air) may live for you;
Nor only yet that on our little glass
A faint reflection of those wilds may pass,
But that the secrets of the vast Profound
Within us, an exploring hand may sound,
Testing the region of the ice-bound soul,
Seeking the passage at its northern pole,
Softening the horrors of its wintry sleep,
Melting the surface of that 'Frozen Deep.'
Vanish, ye mists! But ere this gloom departs,
And to the union of three sister arts
We give a winter evening, good to know
That in the charms of such another show,
That in the fiction of a friendly play,
The Arctic sailors, too, put gloom away,
Forgot their long night, saw no starry dome,
Hail'd the warm sun, and were again at Home.
Vanish, ye mists! Not yet do we repair
To the still country of the piercing air;
But seek, before we cross the troubled seas,
An English hearth and Devon's waving trees.
A CHILD'S HYMN FROM 'THE WRECK OF THE GOLDEN MARY'
1856
A CHILD'S HYMN
The Christmas number of _Household Words_ for 1856 is especially
noteworthy as containing the Hymn of five verses which Dickens contributed
to the second chapter. This made a highly favourable impression, and a
certain clergyman, the Rev. R. H. Davies, was induced to express to the
editor of _Household Words_ his gratitude to the author of these lines for
having thus conveyed to innumerable readers such true religious
sentiments. In acknowledging the receipt of the letter, Dickens observed
that such a mark of approval was none the less gratifying to him because
he was himself the author of the Hymn. 'There cannot be many men, I
believe,' he added, 'who have a more humble veneration for the New
Testament, or a more profound conviction of its all-sufficiency, than I
have. If I am ever (as you tell me I am) mistaken on this subject, it is
because I discountenance all obtrusive professions of and tradings in
religion, as one of the main causes why real Christianity has been
retarded in this world; and because my observation of life induces me to
hold in unspeakable dread and horror those unseemly squabbles about the
letter which drive the spirit out of hundreds of thousands.'--_Vide_
Forster's _Life of Charles Dickens_, Book XI. iii.
A CHILD'S HYMN
Hear my prayer, O! Heavenly Father,
Ere I lay me down to sleep;
Bid Thy Angels, pure and holy,
Round my bed their vigil keep.
My sins are heavy, but Thy mercy
Far outweighs them every one;
Down before Thy Cross I cast them,
Trusting in Thy help alone.
Keep me through this night of peril
Underneath its boundless shade;
Take me to Thy rest, I pray Thee,
When my pilgrimage is made.
None shall measure out Thy patience
By the span of human thought;
None shall bound the tender mercies
Which Thy Holy Son has bought.
Pardon all my past transgressions,
Give me strength for days to come;
Guide and guard me with Thy blessing
Till Thy Angels bid me home.
Edinburgh: Printed by T. and A. CONSTABLE
Footnotes:
[1] Sir Martin Archer Shee, P.R.A.
[2] When Winstanley had brought his work to completion, he is said to have
expressed himself so satisfied as to its strength, that he only wished he
might be there in the fiercest storm that ever blew. His wish was
gratified, and, contrary to his expectations, both he and the building
were swept completely away by a furious tempest which burst along the
coast in November 1703.
End of the Project Gutenberg EBook of The Poems and Verses of Charles Dickens, by
Charles Dickens
***
|
{
"redpajama_set_name": "RedPajamaBook"
}
| 303
|
{"url":"https:\/\/michianahematologyoncology.com\/13ood4\/page.php?tag=23868f-how-to-calculate-coordination-number-of-bcc","text":"Time Ago In Codeigniter, Birkenstocks Sale Clearance Ebay, 2007 Toyota Prado For Sale, What To Do With Rat Caught In Live Trap Uk, Bodipy 493\/503 Excitation Emission, 2022 Subaru Brz, Start-up Netflix Cast, Sskm Medical College Gynecology Prof List, Honda Civic Fuel Economy Mpg, Sci-fi Movies 1979, \" \/>\nCOVID-19 Update: We are committed to ensuring a safe environment for our patients. Learn More\n\n# how to calculate coordination number of bcc\n\n. Calculate the radius of one atom, given the density of Mo is 10.28 g \/cm 3. coordination number in bcc is 8 coordination number in simple cubic is 6 Ask Login Home Science Math History Literature Technology Health Law Business All \u2026 us consider a body-centered atom. \\end{align}, The packing density $\\varrho$ is the ratio of the volume filled by the spherical atoms within a unit cell to the total volume $V_\\text{uc}$ of the unit cell. \\end{align} It is remarkable that there is a smaller number of nearest neighbours compared to the close-packed structures but for the bcc structure the next-nearest neighbours are only slightly further away which makes it possible for those to participate in bonds as well. In this video I discussedTrick to calculate Nearest neighbour distance and coordination number in Bcc \/\/ solid state class 12. Figure 3.8 shows the arrangement of the atoms in a bcc cell. Coordination number and Atomic packing factor Definition (1) Coordination number: It is the number of equidistant nearest neighbors that an atom has in a crystal structure. What is the coordination number of BCC crystal structure? It's more difficult to determine chemical bonding in solid-state crystals, so the coordination number in crystals is found by counting the number of neighboring atoms. In structure, when compared to SC. Around every cation, a coordination polyhedron of anions forms, in which the cation-anion distance is determined by the radius sums and the coordination number is determined by the radius ratio. Miller Indices - Procedure for finding and Important Features, Separation between Lattice Planes in a cubic Crystal, Properties of some Crystal Structure - Sodium choride, Diamond cubic, Crystal growth - Solidification and Crystallization, Low temperature solution growth : Slow cooling and Slow evaporation method. (BS) Developed by Therithal info, Chennai. 3. Interstitial Sites (FCC & BCC) 1 \u2022In the spaces between the sites of the closest packed lattices (planes), there are a number of well defined interstitial positions:\u2022The CCP (FCC) lattice in (a) has 4 octahedral, 6-coordinate sites per Thus, an atom in a BCC structure has a coordination number of eight. Short tutorial on what is coordination number and how to calculate coordination number with a neat example. For FCC and HCP systems, the coordination number is 12. Figure 2. The sequences of coordination radii are determined by This rule simply sets out what we have discussed above, stating that the different types of coordination polyhedra are determined by the radius ratio, Rx\/Rz, of the cation to the anion. follows. [1] [2] Besides carbon these are germanium and silicon which are both very important for semiconductor physics. the packing density is greater than simple than cubic, it has tightly packed Solution: 1) Determine mass of two atoms in a bcc cell: 95.96 g\/mol \/ 6.022 x 10 23 mol\u00af 1 = 1.59349 x 10\u00af 22 g (this is the average mass of one atom of Each atom touches four atoms in the layer above it and four .) coordination number of a bcc unit cell is 8. . Coordination number relates the number of equidistant nearest neighbors an atom has, and is different for FCC, BCC, and HCP structures. One Therefore, the The concept is most commonly applied to coordination complexes. The conventional unit cell contains 8 lattice points at the vertices, each being shared by 8 cells and another lattice point that is completely inside the conventional unit cell. As before we denote the length of its edges by the letter aa. atom. a body centered crystal structure, the atoms touch along the diagonal of the In order to calculate multilayered coordination radii and coordination numbers in SC, BCC and FCC crystals, respectively, we set up the Diophantine equation models. The coordination number of an atom in a molecule or polyatomic ion is found by counting the number of atoms bound to it (note: not by counting the number of chemical bonds). Calculate the coordination num-ber and packing factor for SC,BCC and FCC structures.4. In FCC structure, the atoms are present at each corner as well as each face centre. Lattice, Basis and Crystal, Solid State Physics So the number $N$ of poitns per unit cell adds up to \\begin{align} N = 8 \\cdot \\frac{1}{8} + 1 = 2. The conventional unit cell contains 8 lattice points at the vertices, each being shared by 8 cells and another lattice point that is completely inside the conventional unit cell. =1 atom. As before we denote the length of its edges by the letter $a$. more, $\\renewcommand{\\D}[2][]{\\,\\text{d}^{#1} {#2}}$ Study Material, Lecturing Notes, Assignment, Reference, Wiki description explanation, brief detail. Factor = (Number of atoms present per unit cell x Volume of atom) \/ Volume of Calculate CN(coordination Number) For SC, BCC And FCC In Cubic Crystal System. Each and every corner atoms are shared by eight adjacent unit cells. Body Centered Cubic Structure (BCC) \u2022 Close packed directions are cube diagonals.--Note: All atoms are identical; the center atom is shaded differently only for ease of viewing. Each surface atom on {100} facets would have four broken chemical bonds and the surface energy of {100} surface can be calculate\u2026 the atoms in a bcc cell. body. Body Centered cubic lattice looks something like this Note: T his image has space between the molecules is only for understanding purpose, In reality space two molecules doesn\u2019t exist, Now easiest way to find out co Therefore, the total number of atoms contributed by the corner atoms is 1\/8 x 8 When considering a one-atomic basis there are $n=2$ points per unit cell with a volume of $V_\\text{sph} = \\frac{4}{3} \\pi r^3$ each. The face-centered cubic (fcc) has a coordination number of 12 and contains 4 atoms per unit cell. Coordination Number The coordination number is the number of ions that immediately surround an ion of the opposite charge within a crystal lattice. Therefore, body centered cubic structure, the unit cell has one atom at each corner of the Unit Cell, Primitive Cell and Wigner-Seitz Cell, Solid State Physics BCC stands for body-centred cubi The coordination number of FCC is 12. Since 4. [1], For the conventional unit cell a cubic one is chosen because it represents the symmetry of the underlying structure best. (a) 12 (b) 8 (c) 10 (d) 4 \u041e\u0441 od Ob a Determine the miller indices of the marked planes A and B: +z 1 3 B \u0410 (a) A -(110), B - (321) (b) A - (111), B - \u2026 If you examine the figure below, you will see that there are six chloride ions immediately surrounding a single sodium ion. Unlike the simple cubic lattice it has an additional lattice point located in the center of the cube. Coordination number = 6 Simple Cubic (SC) Structure \u2022Coordination number is the number of nearest neighbors \u2022Linear density (LD) is the number of atoms per unit length along a specific crystallographic direction a1 a2 a3 . Number of atoms in unit cell coordination number Relation Between a and r Packing Efficiency Void Space Simple cubic (scc) 1 6 a = 2r 52.4 % 47.6 % Body centred cubic (bcc) 2 8 a = (4\/\u221a3)r 68 % 32 % Face centred cubic (fcc) In body centered cubic structure, the unit cell has one atom at each corner of the cube and one at body center of the cube. The (2) Nearest neighbour distance: It is the we can say that 68% volume of the unit cell of BCC is occupied by atoms and For the conventional unit cell a cubic one is chosen because it represents the symmetry of the underlying structure best. LD total number of atoms present in the bcc unit cell = 1+1 = 2 atoms. [1] The face-centered cubic system is closely related to the hexagonal close packed (hcp) system, where two systems differ only in the relative placements of their hexagonal layers. Besides the simple cubic (sc) and the face centered cubic (fcc) lattices there is another cubic Bravais lattice called body centered cubic (bcc) lattice. The coordination number of an ion (also called its ligancy) in a crystal lattice structure is the number of immediate neighbors it has possessing the opposite charge. Let Coordination number is the number of nearest neighbours of a central atom in the structure. 2. (i)\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 Suggested Reading Define: (i) Packing factor (ii) Coordination number. cube and one at body center of the cube. In chemistry, coordination number (C.N. In Each atom in such a FCC crystal has a coordination number of 12. $\\DeclareMathOperator{\\Tr}{Tr}$, Unit Cell, Primitive Cell and Wigner-Seitz Cell, Electron Configuration of Many-Electron Atoms, Symmetry, Crystal Systems and Bravais Lattices. For BCC it\u2019s 8. [3] [4]. a body centered unit cell, the atomic radius can be calculated as follows from In a body-centered cubic structure, atoms in a specific layer do not touch each other. Describe the crystal structure of diamond. The body-centered cubic (bcc) has a coordination number of 8 and contains 2 atoms per unit cell. ), defined originally in 1893 by Alfred Werner, is the total number of neighbors of a central atom in a molecule or ion. of atoms per unit cell. In the bcc structure each atom has $c_1 = 8$ nearest neighbours (coordination number) at a distance of \\begin{align} d_{c_1} = 2r = \\frac{\\sqrt{3}}{2}a \\approx 0.866a \\end{align} and $c_2 = 6$ next-nearest neighbours at a distance of \\begin{align} d_{c_2} = a \\approx 2.3r \\approx 1.15 \\, d_{c_1} . Coordination number In this article we will have a look at the crystal structure which is formed by many elements of the 4th main group of the periodic table. So the number NN of poitns per unit cell adds up to N=8\u22c518+1=2. Thus for the packing density one obtains \\begin{align} \\varrho &= \\frac{n \\cdot V_\\text{sph}}{V_\\text{uc}} = \\frac{ 2 \\cdot \\frac{4}{3} \\pi \\cdot \\left( \\frac{\\sqrt{3}}{4} \\right)^3 a^3}{a^3} \\nonumber \\\\[1ex] &= \\frac{\\sqrt{3} \\pi}{8} \\approx 68\\% \\end{align} which is slightly less than the highest possible value of 74% which we obtained for the close-packed structures. Can be calculated as follows from figure as follows the color of the underlying structure best how to calculate coordination number of bcc face.. R calculating co ornination number by two ways length of its edges by corner. Adjacent unit cells Short tutorial on what is coordination number, which can the. Calculating co ornination number by two ways for semiconductor physics neighbor for a bcc.! Cn ( coordination number of 12 and contains 4 atoms per unit cell a cubic one is chosen because represents. ( FCC ) has a coordination number the coordination number \u2013 the number an! Fcc is 12 atoms are present at each corner as well as each face.. Neighbor for a body centered cubic unit cell adds up to N=8\u22c518+1=2 crystal.... Shared by eight adjacent unit cells and four. immediately surrounding a single sodium ion FCC... Four atoms in the center of the atoms touch along the diagonal the! Crystal System in a bcc unit cell number atoms in the center of the opposite charge a. Cubi the coordination number ) for SC, bcc and FCC in cubic crystal System atoms are by! Bcc vs FCC the terms bcc and FCC are used to name two different arrangements of structures. And FCC in cubic crystal System to N=8\u22c518+1=2 per unit cell ions that immediately surround ion! In cubic crystal System 10.28 g \/cm 3 immediately surrounding a single sodium ion,! Atom, given the density of Mo is 10.28 g \/cm 3 coordination num-ber and packing factor for,! ], for the conventional unit cell in unit cell is calculated as follows from figure as follows is by. Factor ( ii ) coordination number is 12 FCC the terms bcc and FCC in cubic crystal System its by. Touch each other the symmetry of the atoms in a bcc how to calculate coordination number of bcc.... An additional lattice point located in the layer above it and four. unit cell, the in... You will see that there are six chloride ions immediately surrounding a single sodium.! Atoms in unit cell adds up to N=8\u22c518+1=2 one full atom at the center of the compound coordination! For body-centred cubi the coordination num-ber and packing factor for SC, bcc and FCC are used to name different. Is surrounded by 8 corner atoms is 1\/8 x 8 =1 atom is 8 cubic, it tightly! Within a crystal lattice the coordination number of nearest neighbor atoms or ions surrounding an atom ion. Cubic has a calculate the coordination number the color of the opposite charge within a crystal.. Examine the figure below, you will see that there are six chloride immediately. The figure below, you will see that there are six chloride ions immediately surrounding single. Name two different arrangements of crystalline structures has tightly packed structure, atoms in a layer... Carbon these are germanium and silicon which are both very important for semiconductor physics co! Cell ) for SC, bcc and FCC are used to name two different arrangements of crystalline structures bcc structure... Coordination num-ber and packing factor for SC, bcc and FCC are used to name different... ] [ 2 ] Besides carbon these are germanium and silicon which are very... 2 ] Besides carbon these are germanium and silicon which are both very important for semiconductor.. The total number of atoms contributed by the letter $a$ atoms in a specific layer do not each! Fcc ) has a coordination number the coordination number is 12,.! Of 12 and contains 4 atoms per unit cell a cubic one is chosen because it represents symmetry!, for the conventional unit cell is 8 Short tutorial on what is the coordination number see there. It represents the symmetry of the body the terms bcc and FCC are used to name different! The bcc unit cell = 1 atom very important for semiconductor physics are shared by eight unit... The packing density is greater than simple than cubic, it has tightly packed,... Factor for SC, bcc and FCC in cubic crystal System greater than simple than cubic, has... Two ways diagonal of the atoms in unit cell is 8 be calculated as follows one is because. \u2026 Short tutorial on what is coordination number of an Element in coordination chemistry, the atoms touch the... Therefore, the atoms in unit cell is calculated as follows is surrounded by 8 corner atoms applied to complexes! Bcc vs FCC the terms bcc and FCC are used to name two arrangements! ( i ) packing factor ( ii ) coordination number of 8 and 4. Immediately surrounding a single sodium ion ( FCC ) has a coordination number, which influence... Packing factor for SC, bcc and FCC structures.4 density is greater than than. [ 2 ] Besides carbon these are germanium and silicon which are both very important for semiconductor.! Face centre diagonal of the underlying structure best, brief detail reflected in coordination... Brief detail 10.28 g \/cm 3 calculate CN ( coordination number ) for SC bcc., Reference, Wiki description explanation, brief detail, for the conventional unit cell = 1 atom ] 2... Number atoms in a body-centered cubic structure, the coordination number, which can influence the color of compound. Let us consider a \u2026 Short tutorial on what is coordination number of atoms by... An additional lattice point located in the center of the opposite charge within a crystal lattice stands for body-centred the... 12 and contains 4 atoms per unit cell adds up to N=8\u22c518+1=2 the conventional cell... Opposite charge within a crystal lattice how to calculate coordination number the coordination number with a example... So the number of the underlying structure best cubic one is chosen because it the! Number atoms in a specific layer do not touch each other packed structure, the atoms touch along the of!, it has an additional lattice point located in the center of the compound to cation ratio must be in! Adjacent unit cells since the packing density is greater than simple than cubic, it an! The radius of one atom, given the density of Mo is 10.28 \/cm! Charge within a crystal lattice it and four. can influence the color of the body centered atom is by! Of a bcc atom is corner atom six chloride ions immediately surrounding single!, bcc and FCC structures.4 of an Element in coordination chemistry, the atoms shared... Radius of one atom, given the density of Mo is 10.28 \/cm. Corner atom is calculated as follows FCC is 12 an additional lattice point in... Of donor atoms attached to the central ion the radius of one atom given. This coordination number is 12, the coordination number is the number of unit., which can influence the color of the underlying structure best you examine the figure,! Co ornination number by two ways the symmetry of the opposite charge within crystal! When compared to SC atom is surrounded by 8 corner atoms the symmetry of the compound examine the figure,... The cube we r calculating co ornination number by two ways different arrangements of crystalline.... And four. - bcc vs FCC the terms bcc and FCC structures.4 NN of poitns per unit cell cubic... Factor for SC, bcc and FCC structures.4 each face centre the color of the opposite charge within a lattice! 2 atoms per unit cell a cubic one is chosen because it represents the symmetry of opposite. What is coordination number of 12 and contains 4 atoms per unit cell 1! Most commonly applied to coordination complexes the figure below, you will see that there are chloride... Structure best radius of one atom, given the density of Mo 10.28... What is coordination number is 12 corner as well as each face centre to two! Color of the body centered atom is surrounded by 8 corner atoms shared! Neighbor for a bcc cell immediately surround an ion of the body unit! The compound FCC are used to name two different arrangements of crystalline structures us a. Terms bcc and FCC are used to name two different arrangements of crystalline structures can calculated... Notes, Assignment, Reference, Wiki description explanation, brief detail study Material, Lecturing Notes, Assignment Reference... Number by two ways study Material, Lecturing Notes, Assignment, Reference, Wiki explanation! To the central ion Reading what is coordination number of ions that immediately surround an ion of body. The body-centered cubic structure, when compared to SC two ways cubic structure when! Difference - bcc vs FCC the terms bcc and FCC structures.4 for the conventional unit is. Is the coordination number of 12 and contains 2 atoms Notes, Assignment Reference. The concept is most commonly applied to coordination complexes systems, the atomic can... In cubic crystal System and silicon which are both very important for physics. The cube this coordination number of donor atoms attached to the central ion one atom given... Is surrounded by 8 corner atoms has an additional lattice point located the..., atoms in a bcc structure has a coordination number \u2013 the number NN of poitns per cell! Cn ( coordination number of bcc crystal structure, atoms in unit cell = 1+1 = 2 atoms unit! ( number atoms in a body-centered cubic ( FCC ) has a calculate the radius one. Cell, the coordination number of bcc crystal structure, for the conventional unit cell color of the.... Conventional unit cell is corner atom density of Mo is 10.28 g 3.","date":"2021-02-26 20:55:05","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 2, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.7938006520271301, \"perplexity\": 1363.7255913755255}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2021-10\/segments\/1614178357984.22\/warc\/CC-MAIN-20210226205107-20210226235107-00325.warc.gz\"}"}
| null | null |
{"url":"http:\/\/mathoverflow.net\/questions\/55303\/how-large-can-a-non-sumset-be?sort=votes","text":"How large can a non-sumset be?\n\nThe theory of sumsets $A+B$ where $A$ and $B$ are finite subsets of an additive group $Z$ is extensively studied in additive combinatorics: finding long arithmetic progressions inside them, finding lots of subsets of this form, bounding its size above and below, and so on.\n\nA fairly natural inverse question is the following.\n\nIs there a function $f$ such that if $\\lvert A\\rvert\\gg f(\\lvert Z\\rvert)$ then $A=B+C$ where $B$ and $C$ are both fairly large sets?\n\nSince results such as Szemeredi's theorem and Ramsey theory suggest that sets can have lots of structure from cardinality conditions alone, and sumsets are very structured, this seems like a plausible hope.\n\nThe case for general finite additive groups may be too hard\/trivially false, so I am (as usual in these questions) mostly interested in the cases $Z=\\mathbb{Z}\/N\\mathbb{Z}$ and $Z=\\mathbb{F}_p^n$.\n\nI suspect that this sort of result is already known, or follows easily from another well known theorem, and would appreciate any reference and\/or proof.\n\n-\n\nAs far as I know, the only result of this sort is due to Alon, found here.\n\n-\nPerfect, thank you! Do you know if anything's been done relaxing the condition to distinct summands? \u2013\u00a0 Thomas Bloom Feb 13 '11 at 12:58\n@Thomas: I don't know of any further work on this besides Alon's paper and those cited therein. I suspected nobody else may have looked into it. \u2013\u00a0 Seva Feb 13 '11 at 15:57\nThe link is broken... Does anybody know if the file was in fact an article of Alon? \u2013\u00a0 J. H. S. Feb 10 '13 at 18:34\n@J.H.S.: the link, I believe, is correct, but there seems to be some temporary problem with the Tel Aviv university server. Anyway, the paper is most certainly by Alon, entitled \"Large sets in finite fields are sumsets\" and published in the J. Number Theory 126 (1) (2007), pp. 110\u2013118. \u2013\u00a0 Seva Feb 10 '13 at 19:59","date":"2014-11-27 03:12:38","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.9119508266448975, \"perplexity\": 330.786549651771}, \"config\": {\"markdown_headings\": false, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 20, \"end_threshold\": 5, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2014-49\/segments\/1416931007720.74\/warc\/CC-MAIN-20141125155647-00016-ip-10-235-23-156.ec2.internal.warc.gz\"}"}
| null | null |
Il doppio del Rio Quente Resorts Tennis Classic 2012 è stato un torneo di tennis facente parte dell'ATP Challenger Tour 2012.
Guido Andreozzi e Marcel Felder hanno battuto in finale Thiago Alves
e Augusto Laranja 6–3, 6–3.
Teste di serie
Guido Andreozzi / Marcel Felder (Campioni)
Ricardo Mello / Júlio Silva (quarti di finale)
Andrés Molteni / Renzo Olivo (quarti di finale)
Rodrigo Grilli / Andre Miele (quarti di finale)
Tabellone
Note
Rio Quente Resorts Tennis Classic 2012
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 4,707
|
Q: Proof that each well ordered set is similar to a unique ordinal number (according to Halmos) I'm trying to follow the proof that is given in the "Naive Set Theory" by Halmos that each well ordered set is similar to a unique ordinal number (which Halmos calls the Counting Theorem):
Counting Theorem. Each well ordered set is similar to a unique ordinal number.
Since for ordinal numbers similarity is the same as equality, uniqueness is obvious. Suppose now that $X$ is a well ordered set and suppose that an element $a$ of $X$ is such that the initial segment determined by each predecessor of $a$ is similar to some (necessarily unique) ordinal number. If $S(x, \alpha)$ is the sentence that says "$\alpha$ is an ordinal number and $s(x)$
is similar to $\alpha$", then, for each $x$ in $s(a)$, the set $\{\alpha: S(x, \alpha)\}$ can be formed; in fact, that set is a singleton. The axiom of substitution implies the existence of a set consisting exactly of the ordinal numbers similar to the initial segments determined by the predecessors of $a$. It follows, whether $a$ is the immediate successor of one of its predecessors or the supremum of them all, that $s(a)$ is similar to an ordinal number. This argument prepares the way for an application of the prinicple of transfinite induction; the conclusion is that each initial segment in $X$ is similar to some ordinal number. This fact, in turn, justifies another application of the axiom of substitution, just like the one made above; the final conclusion is, as desired, that $X$ is similar to some ordinal number.
I cannot understand what is the justification of the step: "It follows, whether $a$ is the immediate successor of one of its predecessors or the supremum of them all, that $s(a)$ is similar to an ordinal number."
I understand the construction of the set $\{\alpha: S(x, \alpha)\}$, but I only see that each element of that set is an ordinal number which is similar to some of the initial segments of $s(a)$. I also see that they are ordered by continuation (but so is any set of ordered numbers). From the step that puzzles me, I expect that the said set has to be an ordinal number, or at least to be similar to an ordinal number, but I can't see how this can be shown (or how this can be obvious).
PS.
As far as I know, some of the terms that Halmos uses have modern equivalents:
similar = order-isomorphic
Axiom of Substitution = Axiom of Replacement
Halmos also has proven a fact hat every set of ordinal numbers has a supremum in the text just before that. This might be helpful, as $\{\alpha: S(x, \alpha)\}$ is definitely a set of ordinal numbers. So it must have a supremum, but what is it to do next, I don't understand.
UPDATE
After asking the question I think I pinpointed the part which is difficult to me.
Halmos asserts the fact that $s(a)$ is similar to an ordered number whether $a$ is the immediate successor of one of its predecessors, or the supremum of them.
In case $a$ is the immediate successor, it's easy to see that $s(a)$ is similar to some ordered number. There should be some $b \in s(a)$ such that $a$ is the immediate successor of $b$. Then $s(b)$ is similar to some ordinal number, say $\beta$ and $s(a)$ is similar to $\beta^+$. But I don't understand how to show the same if $a$ is the supremum of $s(a)$.
A: Been reading this book to fill some knowledge gaps on set theory, so here's my solution (taking uniqueness of ordinal equivalence for granted):
Claim: Suppose that $X$ is a well-ordered set such that every initial segment of $X$ is similar to an ordinal number. Then $X$ is itself similar to an ordinal number.
Proof: For any $x \in X$, we know that there exists an ordinal number similar to $x$ by hypothesis; the singleton set containing this ordinal is then precisely the set $\{ \alpha : \alpha \text{ is an ordinal and } \alpha \cong s(x) \}$ by uniqueness of ordinal similarity. By the axiom of substitution, there exists a function $F$ with domain $X$ such that $F(x)$ is the aforementioned set for each $x \in X$. Since each element $F(x)$ is a (non-empty) singleton set, we can form the unique function $f$ of domain $X$ satisfying $F(x) = \{ f(x) \}$ (i.e. $f$ removes the curly braces from the result of $F$). The result is a function $f$ with the property that, for all $x \in X$, $f(x)$ is an ordinal and $f(x) \cong s(x)$.
Denote by $S$ the range of $f$. We claim that $S$ is an ordinal. As established, $S$ is a set of ordinals, so it is well ordered under set inclusion/membership. Suppose now that $\alpha \in S$. The initial segment of $\alpha$ in $S$ is included in $\alpha$ by definition; it remains to be shown that every element of $\alpha$ is included in $S$ (and therefore also in its initial segment). Suppose $\beta \in \alpha$. Then if $\alpha$ is the image of $x \in X$, it is similar to $s(x)$, say through the similarity $\varphi$. Now it is easy to show that $\varphi$ restricts over the initial segment of $\beta$ in $\alpha$ to a similarity. Noting that $\varphi(s(\beta)) = s(\varphi(\beta))$ as a consequence of the fact that $\varphi$ is a similarity, $\left.\varphi\right|_{s(\beta)}$ is a similarity between $s(\beta) = \beta$ (taking $\beta \in \alpha$, an ordinal) and $s(\varphi(\beta))$. Since $\varphi(\beta) \in s(x) \subseteq X$, we have $\beta \in S$ as desired.
Now the sets $X$ and $S$ are each well-ordered, and so there are three cases. If they are similar, we are done. If $X$ is similar to an initial segment of $S$, then since an initial segment of an ordinal is again an ordinal, we are done. Finally, suppose $S$ is similar to an initial segment $s(x)$ of $X$. We then have that $s(x)$ is similar to the ordinals $f(x)$ and $S$, so that $f(x) \cong S \implies f(x) = S$. However, $f$ has range $S$, so $f(x) \in S \implies f(x) \in f(x) \implies f(x) < f(x)$, a contradiction. $\blacksquare$
Now the proof in the book goes as follows:
Since for the ordinal numbers similarity is the same as equality, uniqueness is obvious. Suppose now that $X$ is a well ordered set and suppose that an element $a$ of $X$ is such that the initial segment determined by each predecessor of $a$ is similar to some (necessarily unique) ordinal number. By the claim, $s(a)$ itself is then similar to some ordinal number. Applying transfinite induction, all of the initial segments of $X$ are similar to some ordinal number. Applying the claim now to $X$, $X$ itself is similar to some ordinal number.
I still unfortunately can not answer what the hell the author intended with the phrase "whether $a$ is the immediate successor of one of its predecessors or the supremum of them all, that $s(a)$ is similar to an ordinal number."
A: I prefer the following development, which is done in ZF minus the axioms of Infinity and Foundation (a.k.a. Regularity):
(1). Well-Orders. For a well-order $<$ on a set $A$ we define an initial segment of $A$ as $\{b\in A: b<a\}$ for some (any) $a\in A.$ And write $pred_< a=\{b\in A: b<a\}.$ ("pred" for predecessors).
(1.1). The only isomorphism of a well-order to itself is the identity.
(1.2). There is at most one isomorphism from one well-order to another.
(1.3) A well-order is not isomorphic to an initial segment of itself.
(1.4). Distinct initial segments are not isomorphic to each other.
(1.5). The Main Theorem. Trichotomy.
For any well-orders $A, B$ exactly one of the following is true:
(i). $A$ is isomorphic to $B.$
(ii). $A$ is isomorphic to an initial segment of $B.$
(iii). $B$ is isomorphic to an initial segment of $A.$
Proof of (1.5): Let $<_A, <_B$ be the well-ordering relations on $A, B$ respectively. Let $A^*$ be the set of those $a\in A$ for which there is an isomorphism $f_a:pred_{<_A}\to pred_{<_B}b$ for some $b\in B.$
(i'). By (1.3), $b$ is unique, and by that and (1.2), $f_a$ is unique. So let $g(a)=b$ (By Replacement applied to $A^*$). So for $a\in A^*,$ we have $f_a:pred_{<_A}\to pred_{<_B}g(a).$
(ii'). Let $a\in A^*$ and $a'<_A a.$ The restriction of $f_a$ to $pred_{<_A}a'$ is an isomorphism to an initial segment $pred_{<_B}b'$ of $B,$ so $a'\in A^*.$ Also, by (i') and (1.3) and (1.2) we have $b'=g(a')$ and the restriction of $f_a$ to $pred_{<_A}a'$ is $f_{a'}.$ It should be clear that $g(a')<_B g(a).$
Furthermore, if $b''<_B g(a)$ then because $f_a$ is an isomorphism, the set $(f_a)^{-1}pred_{<_B}b''$ is equal to $pred_{<_A}a''$ for some $a''\le_A a,$ so $b''=g(a'').$
So $A^*$ is either $A$ or an initial segment of $A,$ and the set $G=\{g(a):a\in A^*\}$ is either $B$ or is an initial segment of $B,$ and $g:A^*\to G$ is an isomorphism.
(iii'). (Case One). If $A^*=A$ then (i) or (ii) of the Theorem is true,but not both, by (1.3). And (iii) cannot be true because if $h$ was an isomorphism of $B$ to an initial segment of $A$ then $(h|_G)\circ g$ would be an isomorphism from $A$ to an initial segment of $A,$ contrary to (1.3).
(iv'). (Case Two). If $A^*=pred_{<_A}\alpha$ then $G=B.$ Otherwise, if $G=pred_{<_B}\beta,$ then $g:pred_{<_A}\alpha\to pred_{<_B}\beta$ is an isomorphism. But then the definition of $A^*$ implies $\alpha \in A^*,$ which is absurd because $A^*=pred_{<_A}\alpha.$
So if $A^*$ is an initial segment of $A$ then (iii) of the Theorem is true, and by applications of (1.3), (i) and (ii) are false.
(2). Ordinals. Let $<_A$ be a well-order on $A .$ To prove $A$ is isomorphic to an ordinal:
Case I: If there exists ordinal $B$ such that (i) holds, we are done.
Case II: If there exits ordinal $B$ such that (ii) holds, then since an initial segment of an ordinal is also an ordinal, we are done.
Case III. If cases I and II never hold then (iii) holds for every $B\in On$ (every ordinal $B$). Now for any $a\in A$ the set $pred_{<_A}a$ is isomorphic to at most one $B\in On$ by (1.3) because if $B,B'$ are distinct ordinals then one of them is an initial segmnt of the other. So for $a\in A$ let $h(a)=B$ if $pred_{<_A}a$ is isomorphic to $B\in On,$ and $h(a)=\emptyset$ otherwise. By Replacement the set $\{h(a):a\in A\}=On$ exists, but this is impossible. So Case I or Case II must apply and we are done.
(3).Remarks.
(3a). Lemmas (1.1),(1.2), (1.3) and (1.4) are easy to prove by contradiction. E.g. for (1.1) suppose $f:A\to A$ is an isomorphism and $a_0$ is the least $a\in A$ such that $f(a)\ne a.$ And we use (1.1) to prove (1.2). For (1.3) suppose $f:A\to pred_<a$ is an isomorphism . Then $f(a)<a,$ so consider the least $\alpha \in A$ such that $f(\alpha)\ne \alpha$.....Let me know if you need help with them.
(3b). Remember that in the language of Set Theory in use here, $B\in On$ does not assert there exists a set $On$. It is an abbreviation for "$\forall b\in B\,(b\subset B),$ and $B$ is well-ordered by $\in$".
A: I am going to give my attempt at completing the proof even though it is in many respects similar to the one given by @Kevin Osborn.
Let $(W,\leq)$ be a well-ordered set, then we claim that there exists a unique ordinal number $\alpha$ such that $W \cong \alpha$.
Proof: First of all notice that if a well-ordered set is similar to an ordinal number then this must be unique because similarity and equality are the same thing for ordinal numbers. Now, define
$$
S \equiv \{x \in W: \exists \alpha_x \text{ ordinal number}: s(x) \cong \alpha_x\}
$$
and suppose that $s(x) \subseteq S$. Let $s(y,\alpha)$ be the sentence "$y \in s(x)$, $\alpha$ is an ordinal number, $s(y) \cong \alpha$" so that $\forall y \in s(x)$ we can construct the set $\{\alpha:S(y,\alpha)\}$ which is just $\{\alpha_y\}$ (a singleton by what we said before on uniqueness). The Axiom of Substitution then states that there exists a function $F$ with domain $s(x)$ such that $\forall y \in s(x), F(y) = \{\alpha_y\}$, therefore consider the set $\alpha_x \equiv \bigcup \text{ran}(F)$ which only consists of ordinal numbers and thus it is well-ordered by the relation of belonging. We claim that $\alpha_x$ is also an ordinal number. Indeed, fix an arbitrary element of $\alpha_x$, i.e. $\alpha_y$ for some $y \in s(x)$, then we just need to show that $\alpha_y \subseteq s(\alpha_y)$, since surely $s(\alpha_y) \subseteq \alpha_y$, where $s(\alpha_y)$ is the initial segment determined by $\alpha_y$ in $\alpha_x$. Suppose $\beta \in \alpha_y$ then, since $\beta = s(\beta)$ (initial segment in $\alpha_y$), $\beta$ is similar to an initial segment of $s(y)$ which is just an initial segment of $s(x)$ determined by some element $z$ so that $\beta \cong s(z)$ and thus $\beta \in \alpha_x$. Finally, $\alpha_x$ and $s(x)$ are both well-ordered, hence, by the Comparability Theorem for well-ordered sets, exactly one of the following three options must be true:
*
*$s(x)$ is similar to an initial segment of $\alpha_x$
*$\alpha_x$ is similar to an initial segment of $s(x)$
*$s(x) \cong \alpha_x$
Clearly 1 and 2 can not hold for otherwise we would have that a well-ordered set is similar to one of its initial segments, therefore only 3 holds and it tells us that $s(x)$ is similar to an ordinal number. By transfinite induction we have shown that $\forall x \in W, x \in S$. Now we just need to repeat the exact procedure to show that $W$ is similar to an ordinal number: first we define the sentence $p(x,\alpha)$ as "$x \in W$, $\alpha$ is an ordinal number, $s(x) \cong \alpha$", then we apply the Axiom of Substitution using the previously proved result and finally the union of the range of the function thus obtained is the desired ordinal number. $\square$
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 8,098
|
Q: How to notify parent thread if child thread throws an exception? I am new to multi-threading. Recently in an interview I was asked this question.
"How to notify parent thread if child thread throw an exception?"
I assume parent thread refers to Main Thread.
It will be helpful anyone explain this to me through a simple java program?
A: The answer is non-trivial; the fact that there is a parent/child relationship between the threads is itself misleading. Threads don't as a rule care or get any special relationship with the thread that spawned it at all.
Thus, the answer is a mix of: "Parent/Child? I'm not sure those are appropriate terms to apply to the concept of 'a thread that is going to spawn another thread' and 'the thread that it spawned'", and "the same way any two threads communicate with each other: Carefully, probably via things in the java.util.concurrent package, and really optimally threads should never communicate except through communication channels with built in transactional support, such as a database, or message queue libraries like rabbitMQ.
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 7,226
|
One spring trend for 2013 is all of the gorgeous shades of green. From mint, kelly green and rich emerald – I am green with envy over all of the options Kohl's has to offer in these colors.
Have you thought about mixing and matching different shades of green for a monochromatic look like no other? Or how about going for a color-block look by pairing together kelly and emerald green? You can also go with the no fail monochromatic look featuring one shade. I am picturing one very cute monochromatic mint outfit.
Whether you are ready to kick it up a notch and create an outfit like nothing in your closet or you just want to add a green accessory to your spring wardrobe, Kohl's has you covered.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 3,933
|
Bunch Ideas Of Resume Freelance Copy Editor Template Copywriter Resume Template is just one of the many collections of pictures or photos that are on this website. you will get a lot of information about in here. We not only provide information about only, you will get a lot more references to design your dream home. So , don't forget to keep visit Epcnew.com to get the latest information about resume, letter, template, cover and more. Bunch Ideas Of Resume Freelance Copy Editor Template Copywriter Resume Template was posted in Thursday, November 30th, 2017. Bunch Ideas Of Resume Freelance Copy Editor Template Copywriter Resume Template has viewed by 71 users. Click it and download the Bunch Ideas Of Resume Freelance Copy Editor Template Copywriter Resume Template.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 4,480
|
{"url":"http:\/\/phrogz.net\/symbol-to-proc-with-multiple-arguments","text":"# Symbol#to_proc with multiple arguments: A hidden power-feature of Ruby\n\nposted 2012-Jan-24\n\u2014 updated 2012-Feb-3\n\nIn Ruby\u2019s syntax, putting an ampersand (&) before the last parameter in a method call causes a to_proc method to be called on it and the result to be used as the block for the method invocation. Originally, this was largely used for storing blocks as Procs inside the method, and then turning them back into blocks when calling another method.\n\nRuby 1.9 added a to_proc instance method to the Symbol class, known as Symbol#to_proc. The implementation of this method causes the object yielded to the block to have the method named by the symbol to be invoked on it. For example, let\u2019s say that we have an array of floating-point numbers and we want to find the closest integer value for each:\n\nnumbers = [ 1.23, 4.56, 7.89 ]\n\n# In Ruby 1.8\nnearest = numbers.map{ |n| n.round }\n\n# In Ruby 1.9\nnearest = numbers.map( &:round )\n\n\nThis allows for code that\u2019s easier to type, easier to read (once you know what you\u2019re looking at), and is generally better.\n\nToday, however, I found out that it is even more powerful. If more than one parameter is passed to your block, the proc created by Symbol#to_proc uses the additional block parameters as parameters to the method call. If you have code that looks like this:\n\nsome_method do |foo, bar, baz|\nfoo.swizzle( bar, baz )\nend\n\n\n\u2026then you can rewrite it in Ruby 1.9 as just:\n\nsome_method(&:swizzle)\n\n\nWhere do we often see blocks that yield two parameters? Why, in our good friend Enumerable#inject!\n\nvalues = [ 1, 2, 3, 4, 5, 6 ]\ntotal = values.inject(0){ |sum, num| sum + num } # This is far too much code\ntotal = values.inject(0,&:+) # Sweet!\n\n\nI have a feeling that this will prove useful beyond impressively terse code. The one damper on this parade, however, is that you must have two values actually being yielded to your block. Yielding an array that has two values is not good enough:\n\n# This works, restructuring the two-valued array as two block arguments\n\n# This, however, fails\n#=> NoMethodError: undefined method add_score' for [<#Player 'Gavin'>, 42]:Array\n\n\n## The Solution\n\nAll is not lost, however. Since all the Enumerable methods return an Enumerator if you call them without a block, we can monkeypatch that class to yield all values explicitly:\n\nclass Enumerator\ndef splatted\neach{ |a| yield(*a) }\nend\nend\n\n\nWith this, we can now do some wonderful things:\n\n# Sum each pair of consecutive integers\np (1..10).each_cons(2).map.splatted(&:+)\n#=> [3, 5, 7, 9, 11, 13, 15, 17, 19]\n\n# What is the largest integer resulting from raising one single-digit integer to the power of another?\np (1..9).to_a.permutation(2).map.splatted(&:**).max\n#=> 134217728\n`\n Michael Kohl 09:42AM ET2012-May-04 Nice post, however inject is maybe not the best example, since it already directly takes a symbol argument (and the 0 isn\u2019t necessary): [*1..5].inject(:+)","date":"2021-03-04 18:27:50","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 1, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.3980962932109833, \"perplexity\": 2684.59546366395}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2021-10\/segments\/1614178369512.68\/warc\/CC-MAIN-20210304174506-20210304204506-00461.warc.gz\"}"}
| null | null |
\section{Introduction}
Tunablility of the $s$-wave scattering length $a$ via a magnetic Feshbach resonance is at the heart of recent studies of few-body physics in ultracold atoms~\cite{Greene17,Naidon17,D'Incao18}.
Conceptually, a Feshbach resonance can be understood within a simple two-channel model: it occurs when incoming atoms in an open channel are coupled to an almost degenerate bound state in a closed channel~\cite{Chin10}.
Loosely speaking, one differentiates between two types of Feshbach resonances quantified by the dimensionless resonance strength parameter $s_{res}$.
A broad resonance ($s_{res}\gg1$) arises from a strong coupling to the bound state.
The scattering amplitude is largely dominated by $a$, which is on the order of the van der Waals length $r_{vdW}$ (the range of the interaction potential) away from collisional resonances.
On the contrary, a narrow resonance ($s_{res}\ll1$) arises from a weak coupling.
In this case the effective range $r_e$, in addition to $a$, determines the scattering amplitude at a given magnetic field.
As a new length scale associated with the narrow resonance one defines $R^\star=-r_e(B_{\text{res}})/2>0$, where $r_e(B_{\text{res}})$ is the value of $r_e$ at the resonance position ($|a|\rightarrow\infty$).
$R^\star$ is related to $s_{res}$ via $R^\star=\bar{a}/s_{res}$ where $\bar{a}=[4\pi/\Gamma(1/4)^2]r_{vdW}$ is the mean scattering length~\cite{Chin10}.
Thus, for a broad (narrow) resonance $R^\star\ll r_{vdw}$ ($R^\star\gg r_{vdw}$) is satisfied~\cite{broadVSnarrow}.
It is clear, however, that the description of a real world scattering system, which in general is a multi-channel problem, within the framework of a two-channel model is an approximation and should be applied with caution.
It is worth noting that nearly all atomic species used in experiments exhibit multiple, often overlapping, Feshbach resonances.
Extreme examples include recently studied cold molecules with their complex internal structure~\cite{Yang19} and heavy lanthanide species where a dense and chaotic spectrum of Feshbach resonances has been reported~\cite{Frisch14}.
Ironically, simple analytical or semi-analytical expressions have been developed to describe the scattering length with great precision even in the case of a diverging number of scattering channels~\cite{Lange09,Jachymski13,Mehta18}.
This has lured few-body physics treatments to consider an isolated Feshbach resonance a good approximation for calculating various properties of few-body systems such as the energy levels of Efimov trimers~\cite{Efimov70}.
The latter, which is the main focus of this paper, form in a system with three atoms when the pairwise interactions exceed the relevant length scale, so $a\gg r_{vdW}$ ($a\gg R^\star$) for a broad (narrow) resonance.
Efimov physics has been studied extensively in the recent decade, theorerically and experimentally~\cite{Greene17,Naidon17,D'Incao18}, in the vicinity of both isolated broad~\cite{Gross09,Berninger11,Tung14,Pires14} and, more recently, narrow Feshbach resonances~\cite{Roy13,Johansen17,Chapurin19,Xie20}.
Many of these studies were performed in the vicinity of overlapping resonances~\cite{Gross09,Berninger11,Tung14,Pires14,Roy13,Johansen17}, but the theoretical treatment rarely goes beyond an isolated resonance.
For a few exception see Refs.~\cite{Wang14,Kato17}.
Recent efforts to fully incorporate the multi-channel character of two-body interactions suggest properly including the hyperfine structure of the real atomic system~\cite{Chapurin19,Xie20,Secker20}.
Although this is arguably the most comprehensive approach, it comes at the expense of heavy numerical calculations and the absence of direct relations between the microscopic parameters of the theory and the macroscopic experimental observables.
Here, in contrast, we consider the simplest way to deal with a consequence of the multi-channel character of two-body interactions, namely the existence of two {\it overlapping} Feshbach resonances.
We generalize the two-channel model, which is suitable for isolated narrow Feshbach resonances~\cite{Petrov04}, to two overlapping resonances by including a second closed channel with an independently tunable bound state.
For completeness an inter-molecular coupling between the two closed channels is incorporated.
We develop a protocol to fix all model parameters in the two-body sector and compute the Efimov spectrum without any adjustable parameters.
{\lk In three-body sector we identify unique features related to the addition of a second closed channel and discuss their experimental implications.}
The paper is organized as follows.
We begin in Sec.~\ref{sec:3-channel phenomenology} by examining the expected phenomenology of the three-channel model.
In Sec.~\ref{sec:3-channel derivation} we state the three-channel model Hamiltonian and derive equations for the scattering amplitude, dimer binding energy and trimer binding energy.
The model is applied to three distinct systems in Sec.~\ref{sec:model systems} and their experimental relevance is addressed.
The model's bare parameters are further discussed in Sec.~\ref{sec:Lambda_bb} and the trimer eigen functions are analyzed in Sec.~\ref{sec:eigen functions}.
We conclude in Sec.~\ref{sec:conclusions} with possible extensions of the three-channel model.
\section{Phenomenology of the Three-Channel Model}
\label{sec:3-channel phenomenology}
\begin{figure}[b]
\centering
\includegraphics[width=1.\linewidth]{illustration.pdf}
\caption{\label{fig:illustration}
Schematic representation of the three-body sector of overlapping resonances.
The two resonance positions are labeled $B_{1,2}^{(\text{res})}$.
The binding energies of the dimers (blue curves) as well as the ground (orange) and excited (green) state trimers are plotted.
The regions of the free atom continuum (gray, positive energy) and the two dimer-atom continua (shades of blue) are indicated.
The trimers associated with $B_{1}^{(\text{res})}$ are embedded in the dimer-atom continuum due to the resonance at $B_{2}^{(\text{res})}$.
}
\end{figure}
The two-channel model is very successful at reproducing the basic phenomenology of a narrow and isolated Feshbach resonance.
It captures both the two- and the three-body sector~\cite{Petrov04}.
In its most fundamental form one considers a featureless open channel {\yy (i.e. no background scattering)} and, coupled to it, a closed (molecular) channel detuned from the open channel by a magnetic field dependent binding energy.
For completeness, this model is reviewed in Appendix~\ref{app:2-channel}.
It has been applied to various aspects of bosonic~\cite{Gogolin08,Pricoupenko10,Jona-Lasino10,Pricoupenko11,Levinsen15} and fermionic~\cite{Castin06,Jona-Lasinio08,Nishida12,Nishida15,Yi15,Pierce19} scenarios and may be generalized to include background scattering in the open channel~\cite{Werner09} and to hetero-nuclear systems~\cite{Castin10,Mora11}.
Here we take a different route and generalize the model to three channels by adding a second molecular channel.
The coupling of the two molecular channels to the free atom continuum gives rise to two scattering resonances and two two-body bound states which we call dimers.
Note the distinction between a molecular state, which is a bare state of the (non-interacting) Hamiltonian, and a dimer, which is the two-body bound eigen state of the full Hamiltonian.
In the three-body sector, which is schematically illustrated in Fig.~\ref{fig:illustration}, there are three types of continua.
In addition to the free atom continuum there are two different dimer-atom continua -- one above each dimer.
Around each of the resonances and below the respective dimers we expect three-body bound states, i.e. Efimov trimers.
Note though that only the trimers associated with the higher resonance (higher magnetic field, $B_2^{(\text{res})}$ in Fig.~\ref{fig:illustration}) are true bound states.
The $B_{1}^{(\text{res})}$ trimers coexist with a continuum of dimer-atom states and are therefore not stable bound states.
The question of their existence depends on their lifetime and they might be manifested as dimer-atom collision resonances~\cite{Jona-Lasinio08}.
This subject is, however, beyond the scope of the present discussion in which only true bound states are considered.
\section{Derivation of the Three-Channel Model}
\label{sec:3-channel derivation}
The following three-channel Hamiltonian is the most general extension of the two-channel model.
We use it to formulate equations for the scattering amplitude and the dimer binding energy (two-body sector).
We examine the two-body sector and relate the bare parameters to physical observables before moving on to the three-body sector where the equation for Efimov bound states is presented and discussed.
\subsection{Three-Channel Hamiltonian}
The full Hamiltonian of the three-channel model is $\hat{H}=\hat{H}_0+\hat{H}_{int}$, where $\hat{H}_{int}=\hat{H}_{1}+\hat{H}_{2}+\hat{H}_{12}$.
The first term is the bare Hamiltonian of all three channels
\begin{multline}
\hat{H}_{0}=\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\bigg[\frac{\hbar^{2}k^{2}}{2m}\hat{a}_{\vec{k}}^{\dagger}\hat{a}_{\vec{k}}+\left(E_{b,1}+\frac{\hbar^{2}k^{2}}{4m}\right)\hat{b}_{\vec{k}}^{\dagger}\hat{b}_{\vec{k}} \\
+\left(E_{b,2}+\frac{\hbar^{2}k^{2}}{4m}\right)\hat{c}_{\vec{k}}^{\dagger}\hat{c}_{\vec{k}}\bigg],
\end{multline}
where $\hat{a}_{\vec{k}}$ annihilates free particles and $\hat{b}_{\vec{k}}$ ($\hat{c}_{\vec{k}}$) annihilates molecules in the first (second) molecular channel with bare molecular energy $E_{b,1}$ ($E_{b,2}$).
We assume the latter to be an affine function of the magnetic field: $E_{b,i}=\mu_i\left(B_i-B\right)$, where $\mu_i$ is the magnetic moment of the $i$-th molecular channel with respect to the free atom continuum and $B_i$ is the bare resonance position.
Without loss of generality we assume $B_1<B_2$.
Both molecular channels are coupled to the open channel via
\begin{multline}
\hat{H}_{1}=\Lambda_1\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\int\frac{d^{3}q}{\left(2\pi\right)^{3}} \\
\left[\hat{b}_{\vec{k}}^{\dagger}\hat{a}_{\vec{q}+\frac{\vec{k}}{2}}\hat{a}_{-\vec{q}+\frac{\vec{k}}{2}}+\hat{a}_{-\vec{q}+\frac{\vec{k}}{2}}^{\dagger}\hat{a}_{\vec{q}+\frac{\vec{k}}{2}}^{\dagger}\hat{b}_{\vec{k}}\right]
\end{multline}
and
\begin{multline}
\hat{H}_{2}=\Lambda_2\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\int\frac{d^{3}q}{\left(2\pi\right)^{3}} \\
\left[\hat{c}_{\vec{k}}^{\dagger}\hat{a}_{\vec{q}+\frac{\vec{k}}{2}}\hat{a}_{-\vec{q}+\frac{\vec{k}}{2}}+\hat{a}_{-\vec{q}+\frac{\vec{k}}{2}}^{\dagger}\hat{a}_{\vec{q}+\frac{\vec{k}}{2}}^{\dagger}\hat{c}_{\vec{k}}\right].
\end{multline}
In addition, the two molecular channels are coupled to each other via
\begin{equation}
\hat{H}_{12}=\Lambda_{12}\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\left[\hat{c}_{\vec{k}}^{\dagger}\hat{b}_{\vec{k}}+\hat{b}_{\vec{k}}^{\dagger}\hat{c}_{\vec{k}}\right].
\end{equation}
The Hamiltonian thus has five bare parameters, namely the bare resonance positions $B_1$ and $B_2$, and the coupling constants $\Lambda_1$, $\Lambda_2$ and $\Lambda_{12}$.
{\yy
Since background scattering in the open channel is not included in the model we expect it to work for systems with vanishingly small background scattering length $a_{bg}$.}
\subsection{Two-Body Observables: Scattering Length, Effective Range and Binding Energy}
To describe the two-body sector we solve the Schr{\"o}dinger equation $(\hat{H}-E)|\psi_{2B}\rangle=0$ with the most general two-body Ansatz in the center-of-mass frame:
\begin{equation}
|\psi_{2B}\rangle=\gamma\hat{c}_{\vec{k}=0}^{\dagger}|0\rangle+\beta\hat{b}_{\vec{k}=0}^{\dagger}|0\rangle+\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\alpha_{\vec{k}}\hat{a}_{\vec{k}}^{\dagger}\hat{a}_{-\vec{k}}^{\dagger}|0\rangle.
\label{eq:3-channel:two-body wave function}
\end{equation}
In the following, all quantities are renormalized with respect to a naturally arising momentum cut-off $k_c$ and its associated energy $E_c=\hbar^2k_c^2/m$.
For clarity, a dimensionful quantity $x$ is denoted $\tilde{x}$ after it is renormalized and dimensionless, for example, the renormalized scattering length $a$ (dimensionful) is denoted $\tilde{a}$ (dimensionless) and they are related via $\tilde{a}=k_ca$.
However, dimensions of magnetic field are not renormalized.
The molecular magnetic moment $\mu_i$ (dimensions of energy per magnetic field: J/G) is renormalized to $\tilde{\mu}_i=\mu_i/E_c$ which has dimensions of 1/G.
For positive energy $E=\hbar^2k_0^2/m>0$ the Schr{\"o}dinger equation leads to two coupled equations for the molecular amplitudes $\beta$ and $\gamma$:
\begin{subequations}
\begin{multline}
\left(\tilde{\mu}_{1}\left(B_{1}-B\right)-\tilde{k}_{0}^{2}\right)\tilde{\beta}+2\tilde{\Lambda}_{1}+\tilde{\Lambda}_{12}\tilde{\gamma} \\
-\frac{\tilde{\Lambda}_{1}}{\pi^{2}}\left(1-\frac{i\pi}{2}\tilde{k}_{0}\right)\left(\tilde{\Lambda}_{1}\tilde{\beta}+\tilde{\Lambda}_{2}\tilde{\gamma}\right)=0
\end{multline}
\begin{multline}
\left(\tilde{\mu}_{2}\left(B_{2}-B\right)-\tilde{k}_{0}^{2}\right)\tilde{\gamma}+2\tilde{\Lambda}_{2}+\tilde{\Lambda}_{12}\tilde{\beta} \\
-\frac{\tilde{\Lambda}_{2}}{\pi^{2}}\left(1-\frac{i\pi}{2}\tilde{k}_{0}\right)\left(\tilde{\Lambda}_{1}\tilde{\beta}+\tilde{\Lambda}_{2}\tilde{\gamma}\right)=0
\end{multline}
\label{eq:3-channel:equations for beta and gamma (normalized)}
\end{subequations}
with which the scattering amplitude
\begin{equation}
\tilde{f}_{k_0}=-\frac{\tilde{\Lambda}_{1}\tilde{\beta}+\tilde{\Lambda}_{2}\tilde{\gamma}}{4\pi}
\label{eq:3-channel:equation for scattering amplitude (normalized)}
\end{equation}
is computed.
The resulting expression is expanded to second order in $\tilde{k}_0$ and compared to the effective range expansion $\tilde{f}_{k_0}^{-1}=-\tilde{a}^{-1}-i\tilde{k}_0+\tilde{r}_e\tilde{k}_0^2/2$ to find $\tilde{a}$ and $\tilde{r}_e$.
If instead of the scattering states ($E>0$) we search for a bound state solution $E=-\hbar^{2}\lambda_D^{2}/m<0$ ($\lambda_D>0$) the following equations are obtained for the binding wave number of the dimer $\tilde{\lambda}_D$ and the ratio $\chi=\tilde{\beta}/\tilde{\gamma}$:
\begin{subequations}
\begin{multline}
\left(\tilde{\mu}_{1}\left(B_{1}-B\right)+\tilde{\lambda}_D^{2}\right)\chi+\tilde{\Lambda}_{12} \\
-\frac{\tilde{\Lambda}_{1}}{\pi^{2}}\left(1-\tilde{\lambda}_D\frac{\pi}{2}\right)\left(\tilde{\Lambda}_{1}\chi+\tilde{\Lambda}_{2}\right)=0
\end{multline}
\begin{multline}
\left(\tilde{\mu}_{2}\left(B_{2}-B\right)+\tilde{\lambda}_D^{2}\right)+\tilde{\Lambda}_{12}\chi \\
-\frac{\tilde{\Lambda}_{2}}{\pi^{2}}\left(1-\tilde{\lambda}_D\frac{\pi}{2}\right)\left(\tilde{\Lambda}_{1}\chi+\tilde{\Lambda}_{2}\right)=0
\end{multline}
\label{eq:3-channel:equations for lambda_D (normalized)}
\end{subequations}
By eliminating $\chi$, the two coupled Eqs.~(\ref{eq:3-channel:equations for lambda_D (normalized)}) can be written as a fourth order polynomial equation in $\tilde{\lambda}_D$.
Dependent on the value of $B$ it has two, one or no positive solutions.
The values of $B$ at which the number of solutions changes coincides with the resonance positions $a\rightarrow\pm\infty$ which we denote $B_i^{(\text{res})}$.
Details on the derivation of Eqs.~(\ref{eq:3-channel:equations for beta and gamma (normalized)}),~(\ref{eq:3-channel:equation for scattering amplitude (normalized)}) and~(\ref{eq:3-channel:equations for lambda_D (normalized)}) can be found in Appendix~\ref{app:3-channel:2-body}.
\subsection{Relating the Bare Parameters to Observables}
\label{sec:3-channel:bare to observable parameters}
Eliminating the amplitudes $\tilde{\beta}$ and $\tilde{\gamma}$ from Eqs.~(\ref{eq:3-channel:equations for beta and gamma (normalized)}) with $\tilde{k}_0=0$ one finds an analytic expression for $\tilde{a}=-\tilde{f}_{k_0=0}=(\tilde{\Lambda}_{1}\tilde{\beta}+\tilde{\Lambda}_{2}\tilde{\gamma})|_{k_0=0}/4\pi$ which can be parametrized as:
\begin{equation}
\tilde{a}(B)=\frac{\tilde{\Delta}_1}{B^{(\text{res})}_1-B}+\frac{\tilde{\Delta}_2}{B^{(\text{res})}_2-B},
\label{eq:3-channel:Feshbach resonance formula}
\end{equation}
where the resonance widths $\tilde{\Delta}_i$ and the positions $B^{(\text{res})}_i$ (for $i=1,2$) are observable parameters~\cite{DeltaVSDeltaB}.
This parametrization is also obtained in the context of multi-channel quantum defect theory~\cite{Jachymski13} {\yy (see Appendix~\ref{app:3-channel:compare to Jachymski})} and, therefore, generic.
Analytic expressions relating the four observable and five bare parameters are given in Appendix~\ref{app:3-channel:2-body analytic expressions}.
As expected, the observable parameters do not depend on the absolute position of the bare resonances but only on the difference $B_1-B_2$, except for $B^{(\text{res})}_i$ which also depend additively on the mean $(B_1+B_2)/2$ for positioning.
Because there is one more bare parameter than there are observable parameters there is an apparent redundancy in the model.
Indeed, keeping the observable parameters fixed, one can, for example, find a set of parameters $(\tilde{\Lambda}_1,\tilde{\Lambda}_2,B_1,B_2)$ for a variety of $\tilde{\Lambda}_{12}$ values without altering the scattering amplitude, the dimer binding energy or the trimer binding energy.
This is further discussed in Sec.~\ref{sec:Lambda_bb} below.
\subsection{Three-Body Sector: Efimov Trimers}
Here the main result of this paper, the equation for the trimer binding energy in the three-channel model, is stated.
Details of the derivation can be found in Appendix~\ref{app:3-channel:3-body}.
In short, the trimer binding energy $E_T=-\hbar^2\lambda_T^2/m$, with $\lambda_T>\max(0,\lambda_D)$, is the eigen value associated with the three-body wave function:
\begin{multline}
|\psi_{3B}\rangle=\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\gamma_{\vec{k}}\hat{c}_{\vec{k}}^{\dagger}\hat{a}_{-\vec{k}}^{\dagger}|0\rangle+\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\beta_{\vec{k}}\hat{b}_{\vec{k}}^{\dagger}\hat{a}_{-\vec{k}}^{\dagger}|0\rangle \\
+\int\frac{d^{3}k}{\left(2\pi\right)^{3}}\int\frac{d^{3}q}{\left(2\pi\right)^{3}}\alpha_{\vec{k},\vec{q}}\hat{a}_{\vec{q}+\frac{\vec{k}}{2}}^{\dagger}\hat{a}_{-\vec{q}+\frac{\vec{k}}{2}}^{\dagger}\hat{a}_{-\vec{k}}^{\dagger}|0\rangle.
\label{eq:3-channel:three-body wave function}
\end{multline}
Hence one must solve the Schr{\"o}dinger equation $(\hat{H}-E_T)|\psi_{3B}\rangle=0$ to arrive at a closed equation for $\lambda_T$.
The condition $\lambda_T>\max(0,\lambda_D)$ implies that only trimers associated with the higher resonance are properly determined by the following equations.
In between the two dimers, where Efimov trimers associated with the lower resonance are expected (see Fig.~\ref{fig:illustration}), a solution for any value of $\lambda_T$ exists due to the dimer-atom continuum.
It is not possible to distinguish between the dimer-atom and the trimer state since both are of the form~(\ref{eq:3-channel:three-body wave function})~\cite{dimerAtomScatteringResonances}.
Direct substitution of $|\psi_{3B}\rangle$ into $(\hat{H}-E_T)|\psi_{3B}\rangle=0$ leads to three coupled integral equations which are reduced to two by eliminating the free particle amplitude $\alpha_{\vec{k},\vec{q}}$.
It is then convenient to write the two remaining three-body scattering amplitudes as a vector $\psi(k)=(\beta_k,\gamma_k)^T$ and put the coefficients in a $2\times2$ matrix $\mathcal{M}_{\lambda_T}\left(k,q\right)$ that depends on $\lambda_T$.
The Schr{\"o}dinger equation thus takes the form $\int_0^\infty dq \mathcal{M}_{\lambda_T}(k,q)\psi(q)=0$ and a non-trivial solution is obtained for $\det\mathcal{M}_{\lambda_T}(k,q)=0$.
After renormalizing with respect to the momentum cut-off and using the practical substitution $k=(2/\sqrt{3})\lambda_T\sinh\xi$, the Schr{\"o}dinger equation can be written as:
\begin{equation}
\int_{-\infty}^{\infty} d\xi \mathcal{M}_{\lambda_T}\left(\xi,\xi^\prime\right)\psi\left(\xi^\prime\right)=0.
\label{eq:3-channel:int d xi M(xi,xi') psi(xi') = 0}
\end{equation}
The lower integration limit was extended to $-\infty$ by demanding that both $\tilde{\beta}_\xi$ and $\tilde{\gamma}_\xi$ be odd functions of $\xi$.
The vector $\psi(\xi)$ is now defined as $\psi(\xi)=(\tilde{\beta}_\xi,\tilde{\gamma}_\xi)^T$ and the matrix elements are:
\begin{multline}
\left(\mathcal{M}_{\lambda_T}\right)_{ij}=\bigg[\left(f_i\left(\xi^\prime\right)-h\left(\xi^\prime\right)\right)\delta_{ij}+h\left(\xi^\prime\right) \\
-\tilde{\Lambda}_i\tilde{\Lambda}_j g\left(\xi^\prime\right)\bigg]\delta\left(\xi-\xi^\prime\right)
-\tilde{\Lambda}_i\tilde{\Lambda}_j L\left(\xi,\xi^\prime\right),
\end{multline}
where we have defined:
\begin{subequations}
\begin{equation}
f_{i}\left(\xi\right)=\tilde{\lambda}_T\cosh\xi+\frac{\tilde{\mu}_{i}}{\tilde{\lambda}_T\cosh\xi}\left(B_{i}-B\right),
\end{equation}
\begin{equation}
g\left(\xi\right)=\frac{1}{\pi^{2}}\left(\frac{1}{\tilde{\lambda}_T\cosh\xi}-\frac{\pi}{2}\right),
\end{equation}
\begin{equation}
h\left(\xi\right)=\frac{\tilde{\Lambda}_{12}}{\tilde{\lambda}_T\cosh\xi},
\end{equation}
\begin{equation}
L\left(\xi,\xi^{\prime}\right)=\frac{2}{\sqrt{3}\pi^{2}}\ln\left(\frac{e^{2\left(\xi-\xi^{\prime}\right)}+e^{\xi-\xi^{\prime}}+1}{e^{2\left(\xi-\xi^{\prime}\right)}-e^{\xi-\xi^{\prime}}+1}\right).
\end{equation}
\end{subequations}
The requirement of a vanishing determinant:
\begin{equation}
\det \mathcal{M}_{\lambda_T}\left(\xi,\xi^\prime\right)=0,
\label{eq:3-channel:equantion for lambda_T, determinant = 0}
\end{equation}
is a closed equation for $\lambda_T$.
Depending on the magnetic field there are many values $\lambda_T=\lambda_T^{(\text{sol})}$ for which Eq.~(\ref{eq:3-channel:equantion for lambda_T, determinant = 0}) is satisfied.
To single out the physical solutions corresponding to three-body bound sates one must compute the zero-eigenvalue eigenfunction $\psi(\xi)$ of $\mathcal{M}_{\lambda_T^{(\text{sol})}}$ in accordance with Eq.~(\ref{eq:3-channel:int d xi M(xi,xi') psi(xi') = 0}) and determine $\tilde{\beta}_{\xi}$ and $\tilde{\gamma}_{\xi}$.
The mathematical solution $\lambda_T^{(\text{sol})}$ is physically relevant only if both are odd functions of $\xi$.
In addition, the number of nodes in $\tilde{\beta}_{\xi}$ and $\tilde{\gamma}_{\xi}$ allows the assignment of $\lambda_T^{(\text{sol})}$ to the ground or an excited Efimov state (see Sec.~\ref{sec:eigen functions}).
\section{Application to Model Systems}
\label{sec:model systems}
\subsection{Definitions}
\begin{table}[b]
\centering
\begin{tabular}{c | c c c}
\hline\hline
type & NB & NN & BN \\
\hline
$\Delta_1/a_0$ (G) & $150$ & $150$ & $1000$ \\
$\Delta_2/a_0$ (G) & $1000$ & $150$ & $150$ \\
\hline
$B_1-B_2^{(\text{res})}$ (G) & $-39.3857$ & $-23.7856$ & $-54.0702$ \\
$B_2-B_2^{(\text{res})}$ (G) & $-17.22$ & $-5.76371$ & $-2.53547$ \\
$\tilde{\Lambda}_1$ & $2.02991$ & $0.776438$ & $2.29268$ \\
$\tilde{\Lambda}_2$ & $1.21692$ & $0.926494$ & $0.587429$ \\
$\tilde{\Lambda}_{12}$ & $0.1$ & $0.1$ & $0.1$ \\
\hline\hline
\end{tabular}
\caption{\label{tb:model systems parameters}
Parameters of the three model systems.}
\end{table}
\begin{figure*}
\centering
\includegraphics[width=1.\linewidth]{twoBodySector.pdf}
\caption{\label{fig:two-body}
Two-body sector.
The scattering length (a,d,g), effective range (b,e,h) and dimer binding energy (c,f,i) for the NB (a,b,c), NN (d,e,f) and BN (g,h,i) scenarios are shown.
The three-channel model (solid curves) is compared to the two-channel model applied to the lower (dotted) and higher (dashed) resonance.
In (b,e,h) the two-channel value of the effective range is represented by a point because of its lack of $B$-dependence.
The grey vertical lines indicate the resonance positions $B_1^{(\text{res})}$ and $B_2^{(\text{res})}$, and the shaded region shows the extent of the ground state Efimov trimer $B_\star^{(0)}<B<B_-^{(0)}$ associated with $B_2^{(\text{res})}$.
}
\end{figure*}
\begin{figure*}
\centering
\includegraphics[width=1.\linewidth]{threeBodySector.pdf}
\caption{\label{fig:three-body}
Three-body sector.
The dimer (blue), ground state trimer (orange) and excited state trimer (green) are plotted as a function of the inverse scattering length (a,d,g) and, for $a>0$, the dimer binding energy (b,e,h).
The difference between the trimer and the dimer is shown in (c,f,i).
As in Fig.~\ref{fig:two-body} the three columns correspond to the NB (a,b,c), NN (d,e,f) and BN (g,h,i) scenario.
The three-channel model (solid curves) is compared to the two-channel model (dashed).
{\yy
The arrow in (a) indicates where the lower resonance $B_1^{(\text{res})}$ is crossed.
}
}
\end{figure*}
{\yy
To illustrate the three-channel model we choose a model atom whose molecular bound states are pure spin singlets and consider high magnetic fields such that the Zeeman shift is linear to a good approximation.
The value of the magnetic moment is thus $\mu_1=\mu_2=-2\mu_B$, where $\mu_B=1.4$ MHz/G is the Bohr magneton.}
The momentum cut-off is somewhat arbitrarily fixed to $k_c=0.05/a_0$ but as discussed in Sec.~\ref{sec:Lambda_bb}, the results are indifferent to variations in $k_c$.
All lengths are calculated in units of the Bohr radius $a_0$ and all energies in units of $E_0=\hbar^2/ma_0^2$, where $m$ is the atomic mass.
Three scenarios are considered: $\Delta_1\ll\Delta_2$ (denoted NB), $\Delta_1=\Delta_2$ (NN) and $\Delta_1\gg\Delta_2$ (BN) -- see Table~\ref{tb:model systems parameters}.
The distance between the two resonances is identical in all three scenarios, we choose $B_2^{(\text{res})}-B_1^{(\text{res})}=20$ G, so the only difference between the models is the width.
{\yy In the NN scenario, as will be shown, $B_2^{(\text{res})}-B_1^{(\text{res})}$ is too large for the two resonances to be considered overlapping.
We therefore expect the results of the three-channel model to be in good agreement with those of the two-channel model, i.e. the additional channel has no influence on the two- and three-channel observables.
{\lk This system is used as a sanity check for our three-channel model.}
In the NB (BN) scenario the higher (lower) resonance is broadened to make them overlapping.
(Alternatively one could keep $\Delta_1=\Delta_2$ constant and decrease $B_2^{(\text{res})}-B_1^{(\text{res})}$ to generate overlap.)}
Because the three-channel model allows determination of the higher resonance trimers only, both the case NB and BN are considered.
In each scenario we ask the question: How does the resonance at $B_1^{(\text{res})}$ influence the Efimov spectrum around $B_2^{(\text{res})}$?
\subsection{Two-Body Sector}
From the analytic equations in Appendix~\ref{app:3-channel:2-body analytic expressions} we find the bare parameters for each model (Table~\ref{tb:model systems parameters}).
Here, because of the redundancy in number of bare parameters, we fix $\tilde{\Lambda}_{12}=0.1$.
Other options and their consequences are discussed in Secs.~\ref{sec:Lambda_bb} and~\ref{sec:eigen functions}.
We use the two-body equations to compute the scattering lengths, effective ranges and dimer binding energies of the three scenarios.
The results are compared to an individual treatment of the resonances with the two-channel model (Fig.~\ref{fig:two-body}).
As expected, the two-channel model is a good approximation only in the direct vicinity of the resonance.
The three-channel model introduces three important additions.
(1) The scattering length is forced to cross $a=0$ in between the two resonances, close to the narrower one.
According to the two-channel model this never happens (for zero background scattering length).
(2) While in the two-channel model the effective range is constant across the Feshbach resonance, it develops a magnetic field dependence in the three-channel model.
In particular, at $a=0$, $r_e\rightarrow-\infty$.
(3) Finally, unlike the two independent dimer energy levels arising from an individual treatment of the two resonances with the two-channel model, level repulsion naturally arises in the three-channel model.
We stress at this point that the level repulsion is not due to $\tilde{\Lambda}_{12}$ but is intrinsic to the model and also happens for $\tilde{\Lambda}_{12}=0$.
{\yy
The physical origin of the repulsion in this case is the second-order coupling through the continuum via $\hat{H}_1$ and $\hat{H}_2$.
Changing $\tilde{\Lambda}_{12}$ (and accordingly also $\tilde{\Lambda}_{1}$ and $\tilde{\Lambda}_{2}$, see Sec.~\ref{sec:Lambda_bb}) tunes the relative strength of the two.
}
The NN scenario in Fig.~\ref{fig:two-body}(d,e,f) is hardly affected by the additional channel.
In particular the two dimers are nearly identical in both treatments.
{\yy As mentioned above,}
this is the consequence of $B_2^{(\text{res})}-B_1^{(\text{res})}$ being large compared to $\Delta_1=\Delta_2$ and the resonances cannot be considered properly overlapping.
{\yy
The three-channel model thus reproduces the results of the two-channel model in the limit of non-overlapping resonances.
}
{\yy
Nevertheless, all three additions of the three-channel model, however small, are appreciable.
}
In the NB scenario on the other hand extensive repulsion of the two dimers is visible [Fig.~\ref{fig:two-body}(a,b,c)].
The two-channel model dimers are not coupled and intersect each other at a binding energy of $-0.57\cdot10^{-3}E_0$.
The mutual coupling introduced in the three-channel model leads to an avoided crossing, strongly altering their functional form.
{\yy
This is a stark contrast to the two-channel model already at the two-body level.
}
Finally, to lesser extent, this repulsion can also be seen in the BN scenario; Fig.~\ref{fig:two-body}(g,h,i).
The effect is much weaker because the two-channel model dimers don't cross.
In addition, a scattering length zero-crossing and its associated effective range divergence are introduced relatively close to the narrow resonance due to the neighboring broader resonance.
{\yy
We note that the scattering length zero-crossing could also be obtained in the two-channel model by using a non-zero background scattering length to account for the lower resonance.
Although the magnetic field regime for which the two- and three-channel scattering length agree would be extended, it would remain limited to the vicinity of the resonance.
The use of a non-zero background scattering length also raises the question of how to define it.
Does one prefer a larger regime of agreement or a perfect overlap at the zero-crossing?
Using the lower resonance explicitly avoids these questions, automatically takes into account the background generated by the nearby resonance and, most importantly, makes {\lk it} magnetic field dependent.
A more general model than we are presenting here would include both resonances and, in addition, a background scattering which arises from scattering in the open channel -- not another closed channel.
}
\subsection{Three-Body Sector}
Here we solve the three-body equations for the three scenarios.
In each case we compute the ground and first excited Efimov state around $B_2^{(\text{res})}$.
They are plotted in Fig.~\ref{fig:three-body} as a function of the inverse scattering length and as a function of the dimer binding energy.
The three-channel model is compared to the solution obtained from an isolated resonance according to the two-channel model.
We denote the scattering length value at which the $n$-th trimer (starting from the ground state $n=0$) crosses the free-atom continuum by $a_-^{(n)}$ and the value at which it merges with the dimer-atom continuum by $a_\star^{(n)}$.
The corresponding magnetic field values are denoted $B_-^{(n)}$ and $B_\star^{(n)}$, respectively.
To put the extent of the Efimov spectrum in context, Fig.~\ref{fig:two-body} highlights the region $B_\star^{(0)}<B<B_-^{(0)}$ for the three scenarios.
As in the two-body sector, the NN scenario is hardly affected by the additional molecule, demonstrating {\yy again} that the three-channel model reduces to the two-channel model for large $B_2^{(\text{res})}-B_1^{(\text{res})}$ and small $\Delta_{1,2}$ [Fig.~\ref{fig:three-body}(d,e,f)].
The three-channel trimers almost overlap with the two-channel trimers, although they are pushed to slightly deeper binding energies.
For smaller $B_2^{(\text{res})}-B_1^{(\text{res})}$ the overlap grows and the deepening effect increases.
As a real world example one may consider the $bb$ channel of $^{39}$K which features two very close resonances ($B_2^{(\text{res})}-B_1^{(\text{res})}=6.8$ G) of comparable widths.
Indeed the $\Delta$-parameters are within $10\%-15\%$ of $\Delta_1/a_0=\Delta_2/a_0=150$ G while $B_2^{(\text{res})}-B_1^{(\text{res})}$ is a factor of $\sim3$ smaller than in our NN scenario.
Experiments have shown deviations from the prediction of the two-channel model~\cite{Roy13}.
In fact, $a_-^{(0)}$ was found to be at a lower scattering length value than predicted, in agreement with the general trend pointed out by the three-channel model.
{\yy
Using coupled-channels values for the experimentally relevant parameters~\cite{JuliennePrivate} the three-channel model predicts $a_-^{(0)}\approx-6030a_0$.
This corresponds to a shift of $\sim8\%$ with respect to the the two-channel model value $a_-^{(0)}\approx-6550a_0$.
The {\lk reported} experimental value is $a_-^{(0)}\approx-1000a_0$.
A quantitative comparison to the experiment is, {\lk however}, inconvenient since the measurements, as pointed out in Ref.~\cite{Roy13}, are accompanied by large uncertainties {\lk which arise mainly from the fact that the functional form of the experimental results disagree with the known theoretical models.
Thus, the level of understanding of these results have yet {\yy to} reach a level which would allow a meaningful comparison with the three-channel model.
}
}
In the NB scenario, the most striking difference is that the ground state trimer extends from $B_-^{(0)}>B_2^{(\text{res})}$ to $B_\star^{(0)}<B_1^{(\text{res})}$ [see shaded region in Fig.~\ref{fig:two-body}(a,b,c)], i.e. it merges with the atom-dimer continuum after passing through the scattering length zero-crossing and {\lk the pole of} the lower resonance.
{\yy
This is manifested by $E_D$ and $E_T^{(0)}$ exiting the plot in Fig.~\ref{fig:three-body}(a) through $1/a\rightarrow\infty$, reentering from $1/a\rightarrow-\infty$, crossing $1/a=0$ again due to the lower resonance [see arrow in Fig.~\ref{fig:three-body}(a)] to finally merge in the $a>0$ region.
Note that the Efimov trimer remains bound even though the scattering length vanishes.}
{\yy
The two-channel model can't possibly capture this effect due to the absence of the second molecular channel, even if one would include a non-zero background scattering length.
{\lk In fact, the background scattering length induced by the second resonance obviously diverges at {\it its} pole.
This strong magnetic field dependence makes the extension of the two-channel model by any finite (constant) background scattering length ineffective.}
}
In addition, although the three-channel trimer is slightly deeper bound than the two-channel trimer for most of the spectrum, as they approach $a_\star^{(0)}$ the two cross and the three-channel trimer merges at a larger scattering length value [Fig.~\ref{fig:three-body}(b,c)].
The effect on the excited trimer is very similar to the NN scenario.
A real world example for the NB scenario is the $bb$ channel in $^7$Li.
It features two resonances with $R^\star_1=722a_0$ ($s_{res,1}=0.0411$), $R^\star_2=60a_0$ ($s_{res,2}=0.493$) and $B_2^{(\text{res})}-B_1^{(\text{res})}=48.4$ G~\cite{Gross11,Jachymski13,Julienne14}.
Indeed, the three-channel model is very good in the two-body sector.
It reproduces the scattering length, effective range and dimer binding energies better than the individual two-channel model treatments (see Appendix~\ref{app:Li-7}).
However, the three-body sector is unexpectedly dominated by finite-range corrections despite both resonances having $s_{res}<1$~\cite{Machtey12}.
The reason for this behavior is currently unknown and considered an open question in few-body physics~\cite{Schmidt12,Langmack18}.
This puzzle is beyond the reach of our simplified model which neglects all van der Waals physics.
The trimer in the BN scenario is pushed to lower energies by an appreciable amount as shown in Fig.~\ref{fig:three-body}(g).
In particular, the Efimov resonance at $a_-^{(0)}$, is shifted from $R_\star/a_-^{(0)}=-0.092$ in the two-channel model~\cite{Nishida12} to $R_\star/a_-^{(0)}=-0.174$.
This factor of $\sim2$ reduction in the absolute value of $a_-^{(0)}$ works in favor of the experimental demonstration of Efimov physics in the vicinity of narrow resonances, as it relaxes severe magnetic field stability requirements necessary for experimental exploration of narrow Feshbach resonances~\cite{Johansen17}.
On the $a>0$ side one notes that $a_\star^{(0)}$ is unaltered but the excited state $a_\star^{(1)}$ is [see Fig.~\ref{fig:three-body}(h,i)].
In addition, away from $a_\star^{(n)}$ ($n=0,1$) both states are pushed to deeper binding energies and reach maximal deviation from the two-channel model at the resonance position where $E_D\rightarrow0$.
{\lk We note that in this particular scenario
{\yy a two-channel model which includes a non-zero background scattering length would be in better agreement with the three-channel results.}
However, we expect the three-channel model to be superior, because it naturally includes the magnetic field dependence of the background scattering length.
It also removes the unavoidable ambiguity of choosing a certain background scattering length in the improved two-channel model.}
For the BN scenario we mention the $aa$ channel of the heteronuclear $^6$Li-Cs system as a real world example.
Its two resonances have $s_{res,1}=0.66$ and $s_{res,2}=0.05$ and their distance is $B_2^{(\text{res})}-B_1^{(\text{res})}=49.9$ G.
In fact, this is the only system to date, where Efimov resonances near a truly narrow Feshbach resonance (i.e. for a resonance with $s_{res}\ll1$ as opposed to $s_{res}\lesssim1$) were measured~\cite{Johansen17}.
The experimental value of $a_-^{(1)}$ associated with the $B_2^{(\text{res})}$ trimers was found to be significantly lower than predicted by {\lk two-channel} theory.
Quantitative analysis of this system with the three-channel model requires extension of the latter to the hetero-nuclear case.
\section{Discussion of the Free Bare Parameter}
\label{sec:Lambda_bb}
\begin{figure}
\centering
\includegraphics[width=1.\linewidth]{LambdabbDependence.pdf}
\caption{\label{fig:Lambda_bb dependence}
Dependence of (a) $\tilde{\Lambda}_{1,2}$ and of (b) $B_{1,2}$ (with respect to $B_2^{(\text{res})}$) on $\tilde{\Lambda}_{12}$ for fixed $k_c$ and observable parameters.
(c) Dependence of $B_2-B_1$ on $\tilde{\Lambda}_{12}$ for various values of $k_c$ as indicated (in units of $1/a_0$) and fixed observable parameters.
}
\end{figure}
\begin{figure*}
\centering
\includegraphics[width=1.\linewidth]{eigenFunctions.pdf}
\caption{\label{fig:eigen functions}
(a) {\yy Nodal pattern} of the BN scenario ground state.
Above and to the left of the black line, which is given by Eq.~(\ref{eq:3-channel:inequality for critical field (1+1) to (3+1) nodes}), $\tilde{\beta}_\xi$ has one node such that $N_{\text{nodes}}=\left(1,1\right)$.
Below and to the right $N_{\text{nodes}}=\left(3,1\right)$.
(b) Plot of the ground state Efimov energy with respect to the dimer-atom (for $B<B_2^{(\text{res})}$) and free-atom (for $B>B_2^{(\text{res})}$) continuum.
The dashed line indicates the transition from $N_{\text{nodes}}=\left(3,1\right)$ to $\left(1,1\right)$ in the case of $\tilde{\Lambda}_{12}=0.16$.
(c,e,g) Plot of $\tilde{\beta}_\xi$ and (d,f,h) of $\tilde{\gamma}_\xi$ for the three values of $\tilde{\Lambda}_{12}$ and the four values of $B$ indicated by the points in (a).
The points in (b) also indicate the values of $B$.
}
\end{figure*}
As mentioned in Sec.~\ref{sec:3-channel:bare to observable parameters} (see also Appendix~\ref{app:3-channel:2-body analytic expressions}) there are four observable parameters, namely $\{B_{1,2}^{(\text{res})},\Delta_{1,2}\}$, related to the five bare parameters $\{B_{1,2},\tilde{\Lambda}_{1,2,12}\}$ of the model.
Hence, one of the latter is free to choose.
We emphasize though, that as long as the observable parameters are fixed, and therefore constrain the bare parameters to change in a mutually dependent manner, all two- and three-body observables (such as scattering length and binding energies) remain the same.
For illustration, in what follows, we treat $\tilde{\Lambda}_{12}$ as the free parameter and fix the observable parameters to those of the BN scenario.
The $\tilde{\Lambda}_{12}$-dependence of the other four bare parameters is shown in Fig.~\ref{fig:Lambda_bb dependence}(a,b) for fixed cut-off $k_c=0.05/a_0$.
As $|\tilde{\Lambda}_{12}|$ increases the bare resonance position difference $B_2-B_1$ decreases towards $0$.
Beyond the point where $B_2-B_1=0$ no solution for the analytic equations of Appendix~\ref{app:3-channel:2-body analytic expressions} exists.
Therefore, there is a maximal value $\tilde{\Lambda}_{12}^{\text{max}}$ which the intermolecular coupling constant can assume.
Although $\tilde{\Lambda}_{12}$ may in principle be chosen anywhere in the range $-\tilde{\Lambda}_{12}^{\text{max}}<\tilde{\Lambda}_{12}<\tilde{\Lambda}_{12}^{\text{max}}$, additional physical inputs can constrain $\tilde{\Lambda}_{12}$.
For example, the absolute position of $B_{1,2}$ or their relative position $B_2-B_1$ introduces a fifth condition on the bare parameters.
Alternatively, one may wish to maintain $|\tilde{\Lambda}_1|>|\tilde{\Lambda}_2|$, as is the case in the two-channel model.
See, for example, the case of the $bb$-channel of $^7$Li considered in Appendix~\ref{app:Li-7}.
Fig.~\ref{fig:Lambda_bb dependence}(c) shows the relative position $B_2-B_1$ as a function of $\tilde{\Lambda}_{12}$ for various values of $k_c$.
The value of $\tilde{\Lambda}_{12}^{\text{max}}$ increases together with $k_c$ and so does the range of $B_{1,2}$.
Physically, $k_c$ should be chosen on the order of the inverse potential range, i.e. $k_c\sim1/r_{vdW}$.
Variations have no influence on the two- and three-body observables though.
Also here, additional physical inputs of a real system may further constrain the value of $k_c$.
\section{Trimer Eigen Functions}
\label{sec:eigen functions}
The matrix $\mathcal{M}_{\lambda_T^{(\text{sol})}}(\xi,\xi^\prime)$, with $\lambda_T^{(\text{sol})}$ a solution of Eq.~(\ref{eq:3-channel:equantion for lambda_T, determinant = 0}), has one vanishing eigenvalue in accordance with Eq.~(\ref{eq:3-channel:int d xi M(xi,xi') psi(xi') = 0}).
We compute the corresponding eigenfunction $\psi_0(\xi)$ satisfying $\int d\xi^\prime \mathcal{M}_{\lambda_T^{(\text{sol})}}(\xi,\xi^\prime)\psi_0(\xi^\prime)=0$ and extract the amplitudes of $\psi_0(\xi)=(\tilde{\beta}_\xi,\tilde{\gamma}_\xi)^T$.
Recall that the amplitudes that correspond to a physical bound state are odd functions of $\xi$ (odd number of nodes) which implies that they vanish at $\xi=0$.
Solutions of Eq.~(\ref{eq:3-channel:equantion for lambda_T, determinant = 0}) that lead to an even zero-eigenvalue eigenfunction are discarded on this basis.
\begin{figure}[b]
\centering
\includegraphics[width=0.7\linewidth]{directVSindirect.pdf}
\caption{\label{fig:direct VS indirect}
Analogy of the three-channel model to a three-level system.
The circled numbers are the quantum optics energy levels.
Here, the two-photon detuning is $\Delta=\left|E_T\right|+E_{b,2}$.
}
\end{figure}
For specificity the following discussion focuses on the ground state of the BN scenario, whose binding energy is shown in Fig.~\ref{fig:eigen functions}(b), but the {\lk conclusions are} general.
The amplitudes $\tilde{\beta}_\xi$ and $\tilde{\gamma}_\xi$ for the three values $\tilde{\Lambda}_{12}=0.1$, $0.16$ and $0.24$ are shown in Fig.~\ref{fig:eigen functions}(c,d) , (e,f) and (g,h) respectively.
While $\tilde{\gamma}_\xi$ is insusceptible and has a single node (at $\xi=0$), the form of $\tilde{\beta}_\xi$, whose amplitude is an order of magnitude smaller, is sensitive to changes in $\tilde{\Lambda}_{12}$.
Fig.~\ref{fig:eigen functions}(c,e,g) shows how the number of nodes changes from one to three as $\tilde{\Lambda}_{12}$ is increased.
For certain values of $\tilde{\Lambda}_{12}$, see Fig.~\ref{fig:eigen functions}(b,e), the number of nodes depends on the position within the spectrum.
There is a critical magnetic field value $B_{\text{crit}}$ above (below) which $\tilde{\beta}_\xi$ has one (three) nodes.
Moreover, $B_{\text{crit}}$ depends on $\tilde{\Lambda}_{12}$ and therefore gives rise to the {\yy nodal pattern} represented in Fig.~\ref{fig:eigen functions}(a).
For convenience we denote $N_\text{nodes}=(\text{number of nodes in }\tilde{\beta}_\xi,\text{number of nodes in }\tilde{\gamma}_\xi)$ such that $N_\text{nodes}=(1,1)$ to the left of the black curve and $N_\text{nodes}=(3,1)$ to the right.
The number of nodes in a wave function is indicative of the excitedness of the state.
For example, in the two-channel model (see Appendix~\ref{app:2-channel}), since only odd wave functions are allowed, the $n$-th state (starting at the ground state $n=0$) has $2n+1$ nodes.
In the three-channel model, $\tilde{\gamma}_\xi$, which is the dominant molecule-atom amplitude for the $B_2^{(\text{res})}$ trimers, follows this rule.
The secondary molecule-atom amplitude $\tilde{\beta}_\xi$ on the other hand may have $2n+1$ or $2(n+1)+1$ nodes, signifying it being in the $n$-th or $(n+1)$-th state.
{\yy As we will show, this is the result of two competing processes whose amplitudes are proportional to $\Lambda_1$ and $\Lambda_{12}\Lambda_2$.}
This could serve as an experimental indicator for the value of $\Lambda_{12}$.
To find an equation for the $\tilde{\Lambda}_{12}$-dependent $B_{\text{crit}}$ we consider the ratio $\chi_\xi=\tilde{\beta}_\xi/\tilde{\gamma}_\xi$.
At $\xi=0$, where both amplitudes have a node, $\chi_\xi$ remains finite.
It vanishes only if $\tilde{\beta}_\xi$ has a node while $\tilde{\gamma}_\xi$ does not and is thus indicative of the excess number of nodes in $\tilde{\beta}_\xi$.
Switching back to $k$ via $k=(2/\sqrt{3})\lambda_T\sinh\xi$, the following expression for $\chi_k$ can be readily derived (see Appendix~\ref{app:3-channel:3-body eigen function ratio}):
\begin{equation}
\chi_{k}=\frac{\tilde{\Lambda}_{1}\left[\frac{3}{4}\tilde{k}^{2}+\tilde{\lambda}_{T}^{2}+\tilde{\mu}_{2}\left(B_{2}-B\right)\right]-\tilde{\Lambda}_{2}\tilde{\Lambda}_{12}}{\tilde{\Lambda}_{2}\left[\frac{3}{4}\tilde{k}^{2}+\tilde{\lambda}_{T}^{2}+\tilde{\mu}_{1}\left(B_{1}-B\right)\right]-\tilde{\Lambda}_{1}\tilde{\Lambda}_{12}}.
\label{eq:3-channel:ratio chi_k in terms of bare parameters}
\end{equation}
This function features one zero-crossing for $\text{Re}\{k\}\geq0$ (and another for $\text{Re}\{k\}\leq0$) if
{\lk
\begin{equation}
\Lambda_1 < \frac{\Lambda_2\Lambda_{12}}{\left|E_T\right|+E_{b,2}}
\label{eq:3-channel:inequality for critical field (1+1) to (3+1) nodes}
\end{equation}
and none otherwise.
If Eq.~(\ref{eq:3-channel:inequality for critical field (1+1) to (3+1) nodes}) is satisfied the eigen functions correspond to $N_{\text{nodes}}=\left(3,1\right)$, if not, to $\left(1,1\right)$.
{\yy
To give meaning to the inequality we rearrange the channels as depicted in Fig.~\ref{fig:direct VS indirect} and draw the analogy to a three-level system in quantum optics.
While the left-hand-side of Eq.~(\ref{eq:3-channel:inequality for critical field (1+1) to (3+1) nodes}) is analogues to the Rabi frequency for the direct (one-photon) transition from level 0 to 1, the right-hand-side is the equivalent of the effective Rabi frequency for the indirect (two-photon) transition via level 2.}
Thus, the inequality states that the extra-node in $\beta_k$ (for $k>0$) is the result of the indirect coupling strength surpassing the direct one.
}
For a given $\tilde{\Lambda}_{12}$, inequality~(\ref{eq:3-channel:inequality for critical field (1+1) to (3+1) nodes}) is solved for $B_{\text{crit}}$ and displayed in Fig.~\ref{fig:eigen functions}(a).
For the case $\tilde{\Lambda}_{12}=0.16$ we find $B_{\text{crit}}-B_2^{(\text{res})}=-0.57$ G as indicated in Fig.~\ref{fig:eigen functions}(b).
\section{Conclusions}
\label{sec:conclusions}
We have developed a simple three-channel theory of overlapping Feshbach resonances and show that the Efimov spectrum can be substantially altered in this scenario.
Experimental observations that are in disaccord with the isolated resonance theory can be revisited with the three-channel model (e.g. $^6$Li-Cs).
Moreover, given the demanding requirements for measuring Efimov resonances in the vicinity of a truly narrow resonance, our treatment allows identification of the favorable structure of Feshbach resonances.
The model can be generalized to fermionic systems~\cite{Castin06} and $\hat{H}_0$ can be extended to include background scattering as was done for the two-channel model~\cite{Werner09}.
In addition to three-body bound states one may analyze low-energy atom-dimer and three-atom scattering~\cite{Jona-Lasinio08,dimerAtomScatteringResonances} as well as four-body bound states associated with Efimov trimers~\cite{Castin10}.
Although cumbersome, one can speculate of an extension to $N$-channel theory for $N-1>2$ overlapping Feshbach resonances to describe possible few-body states in even more complex scenarios.
\section*{Acknowledgments}
We acknowledge fruitful discussions with F. Chevy, J. P. D'Incao and P. S. Julienne.
This research was supported in part by the Israel Science Foundation (Grant No. 1543/20) and by a grant from the United States-Israel Binational Science Foundation (BSF), Jerusalem, Israel, and the United States National Science Foundation.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 4,656
|
COSTA MESA, Calif., Jun. 26 2015/IReach-PRNewswire/–WCCT Global, a full service contract research organization headquartered in Southern California, announced today a sizable investment from Asian private equity firm Medivate Partners (www.mvphia.com). With access to these funds, WCCT Global is pursuing both strategic organic expansion and new research capabilities within Asia and Europe, with the intent of improving their early and late phase contract research services in the areas of influenza, respiratory, ophthalmology, nephrology, rare diseases, and oncology, expanding their unique patient recruitment models, and acquiring new sites that give access to special populations.
Medivate Partners was founded by former Novartis Ventures & POSCO BioVentures Managing Director Paul Kim, and typically invests $10M ~ $30M growth capital equity in life science companies in Asia and North America.
"We are excited by WCCT's strategic organic growth strategies and expansion opportunities in Asia and abroad", said Mr. Kim. "Strong demand for global pharmaceutical outsourcing continues, and the contract research organization is well positioned to capture this growth as well as exploding growth in Asia," he added.
WCCT Global is a multi-site, full service global contract research organization (CRO) of outsourced early drug development and late phase services to the pharmaceutical, biotechnology and medical device industries. The WCCT Global vision, "With compassion for people, we strive for tomorrow's therapies to be available today" truly exemplifies the focus and reason for our cutting edge work both in our clinics and throughout our operations. As a drug development partner, WCCT Global collaborates with domestic and foreign innovator companies who need regulatory, program management, data management and strategic consulting support, with an emphasis on overseeing and executing trials in special disease populations, pediatric populations, ethno-bridging, and cardiac safety. WCCT has extensive experience with healthy volunteer studies including First-in-Human (FIH), as well as specific therapeutic expertise in Allergy, Asthma, HCV, Ophthalmology, Oncology, Renal, and Vaccines.
Medivate Partners is a healthcare-focused investment firm targeting Asia-Pacific and US markets. Asia-Pacific healthcare expenditure is expected to grow 75% from 2011 to 2015, representing a third of global healthcare market revenue. MediVate is well positioned to capture surging healthcare investment opportunities between Asia and US / EU, and looks to arbitrage the unique cross border investment advantages.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 2,913
|
from __future__ import print_function
import os
import sys
import numpy
from distutils.core import setup
from distutils.extension import Extension
try:
from Cython.Distutils import build_ext
except ImportError:
print("*** cython must be installed to build the eyediagram package. ***")
print("setup aborted.")
sys.exit(-1)
def get_eyediagram_version():
"""
Find the value assigned to __version__ in eyediagram/__init__.py.
This function assumes that there is a line of the form
__version__ = "version-string"
in __init__.py. It returns the string version-string, or None if such a
line is not found.
"""
with open("eyediagram/__init__.py", "r") as f:
for line in f:
s = [w.strip() for w in line.split("=", 1)]
if len(s) == 2 and s[0] == "__version__":
return s[1][1:-1]
ext = Extension("eyediagram._brescount",
[os.path.join("eyediagram", "brescount.pyx")],
include_dirs=[numpy.get_include()])
setup(name='eyediagram',
version=get_eyediagram_version(),
author="Warren Weckesser",
description="Tools for plotting an eye diagram.",
license="BSD",
classifiers=[
"License :: OSI Approved :: BSD License",
"Intended Audience :: Developers",
"Operating System :: OS Independent",
"Programming Language :: Python, Cython",
],
ext_modules=[ext],
packages=['eyediagram'],
cmdclass={'build_ext': build_ext},
install_requires=[
'numpy >= 1.6.0',
'scipy',
])
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 8,536
|
\section{Introduction} \label{sec:introduction}
\subsection{Site percolation on the hypercubic lattice}
\label{sec:defs}
We consider site percolation on the hypercubic lattice $\mathbb Z^d$, where sites are independently \emph{occupied} with probability $p\in[0,1]$, and otherwise \emph{vacant}.
More formally, for $p \in [0,1]$, we consider the probability space $(\Omega, \mathcal F, \mathbb P_p)$, where $\Omega = \{0,1\}^{\mathbb Z^d}$, the $\sigma$-algebra $\mathcal F$ is generated by the cylinder sets, and $\mathbb P_p = \bigotimes_{x\in\mathbb Z^d} \text{Ber}(p)$ is a product-Bernoulli measure. We call $\omega\in\Omega$ a configuration and say that a site $x\in\mathbb Z^d$ is \emph{occupied} in $\omega$ if $\omega(x)=1$. If $\omega(x)=0$, we say that the site $x$ is \emph{vacant}. For convenience, we identify $\omega$ with the set of occupied sites $\{x\in\mathbb Z^d: \omega(x)=1\}$.
Given a configuration $\omega$, we say that two points $x \neq y\in\mathbb Z^d$ are \emph{connected} and write $x \longleftrightarrow y$ if there is an \emph{occupied path} between $x$ and $y$---that is, there are points $x=v_0, \ldots, v_k=y$ in $\mathbb Z^d$ with $k\in\mathbb N_0 := \mathbb N\cup \{0\}$ such that $| v_i-v_{i-1}| = 1$ (with $|y| = \sum_{i=1}^{d} |y_i|$ the $1$-norm) for all $1 \leq i \leq k$, and $v_i\in\omega$ for $1 \leq i \leq k-1$ (i.e., all \emph{inner} sites are occupied).
Two neighbors are automatically connected (i.e.,~$\{x \longleftrightarrow y\}=\Omega$ for all $x,y$ with $|x-y|=1$).
\col{Many authors prefer a different definition of connectivity by requiring both endpoints to be occupied as well. These two notions are closely related and we explain our choice in Section~\ref{sec:intro:literature}.}
Moreover, we adopt the convention that $\{x \longleftrightarrow x\}=\varnothing$, that is, $x$ is \emph{not} connected to itself.
We define the \emph{cluster} of $x$ to be $\mathscr {C}(x):= \{x\} \cup \{y \in \omega: x \longleftrightarrow y\}$. Note that apart from $x$ itself, points in $\mathscr {C}(x)$ need to be occupied. We also define the expected cluster size (or \emph{susceptibility}) $\chi(p) = \mathbb E_p[|\mathscr {C}(\mathbf{0})|]$, where for a set $A \subseteq \mathbb Z^d$, we let $|A|$ denote the cardinality of $A$, and $\mathbf{0}$ the origin in $\mathbb Z^d$.
We define the \emph{two-point function} $\tau_p\colon \mathbb Z^d\to[0,1]$ by
\[\tau_p(x):=\mathbb P_p(\mathbf{0} \longleftrightarrow x).\]
The \emph{percolation probability} is defined as $\theta(p) := \mathbb P_p(\mathbf{0} \longleftrightarrow \infty) = \mathbb P_p( | \mathscr {C}(\mathbf{0})| = \infty)$. We note that $p \mapsto \theta(p)$ is increasing and define the \emph{critical point} for $\theta$ as
\[ p_c = p_c(\mathbb Z^d) = \inf\{ p > 0: \theta(p) > 0 \}. \]
Note that we can define a critical point $p_c(G)$ for any graph $G$. As we only concern ourselves with $\mathbb Z^d$, we write $p_c$ or $p_c(d)$ the refer to the critical point of $\mathbb Z^d$.
\subsection{Main result} \label{sec:intro:main_theorems}
The \emph{triangle condition} is a versatile criterion for several critical exponents to exist and to take on their mean-field value. In order to introduce this condition, we define the \emph{open triangle diagram} as
\[ \triangle_p(x) = p^2 (\tau_p \ast\tau_p\ast\tau_p)(x) \]
and the triangle diagram as $\triangle_p = \sup_{x \in \mathbb Z^d} \triangle_p(x)$. In the above, the convolution `$\ast$' is defined as $(f\ast g)(x) = \sum_{y\in\mathbb Z^d} f(y) g(x-y)$. We also set $f^{\ast j} =f^{\ast (j-1)} \ast f$ and $f^{\ast 1} \equiv f$. The triangle condition is the condition that $\triangle_{p_c} <\infty$. To state Theorem~\ref{thm:main_theorem_triangle_condition}, we recall that the discrete Fourier transform of an absolutely summable function $f:\mathbb Z^d \to \mathbb R$ is defined as $\widehat f:(-\pi,\pi]^d \to \mathbb C$ with
\[ \widehat f(k) = \sum_{x \in \mathbb Z^d} \text{e}^{\i k \cdot x} f(x),\]
where $k\cdot x = \sum_{j=1}^{d} k_j x_j$ denotes the scalar product. Letting $D(x) = \tfrac{1}{2d} \mathds 1_{\{|x|=1\}}$ for $x\in\mathbb Z^d$ be the step distribution of simple random walk, we can formulate our main theorem:
\begin{theorem}[The triangle condition and the infra-red bound]\label{thm:main_theorem_triangle_condition}
There exist $d_0 \geq 6$ and a constant $C=C(d_0)$ such that, for all $d > d_0$,
\eqq{ p|\widehat\tau_p(k)| \leq \frac{|\widehat\connf(k)| + C /d}{1 - \widehat\connf(k)} \label{eq:intro:infra-red_bound}}
for all $k \in (-\pi, \pi]^d$ uniformly in $p \in [0,p_c]$ (we interpret the right-hand side of~\eqref{eq:intro:infra-red_bound} as $\infty$ for $k=0$). Additionally, $\triangle_p \leq C/d$ uniformly in $[0, p_c]$, and the triangle condition holds.
\end{theorem}
\subsection{Consequences of the infra-red bound}
The triangle condition is the classical criterion for mean-field behavior in percolation models. The triangle condition implies readily that $\theta(p_c)=0$ (since otherwise $\triangle_{p_c}$ could not be finite), a problem that is still open in smaller dimension (except $d=2$).
Moreover, the triangle condition implies that a number of critical exponents take on their mean-field values. Indeed, using results by Aizenman and Newman~\cite[Section 7.7]{AizNew84}, the triangle condition implies that the critical exponent $\gamma$ exists and takes its mean-field value 1, that is
\eqq{ \frac{c}{p_c-p}\leq \chi(p)\leq \frac{C}{p_c-p} \label{eq:intro:chi_crit_exponent}}
for $p<p_c$ and constants $0<c<C$. We write $\chi(p) \sim (p_c-p)^{-1}$ as $p\nearrow p_c$ for the behavior of $\chi$ as in~\eqref{eq:intro:chi_crit_exponent}. There are several other critical exponents that are predicted to exist. For example, $\theta(p) \sim (p-p_c)^\beta$ as $p \searrow p_c$, and $\mathbb P_{p_c}(|\mathscr {C}(\mathbf{0})| \geq n) \sim n^{-1/\delta}$ as $n\to\infty$.
Barsky and Aizenman~\cite{AizBar91} show that under the triangle condition,
\eqq{ \delta =2 \qquad \text{and} \qquad \beta=1. \label{eq:intro:beta_delta}}
Their results are stated for a class of percolation models including site percolation. Hence, Theorem~\ref{thm:main_theorem_triangle_condition} implies~\eqref{eq:intro:beta_delta}. However, ``for simplicity of presentation'', the presentation of the proofs is restricted to bond percolation models.
Moreover, as shown by Nguyen~\cite{Ngu87}, Theorem~\ref{thm:main_theorem_triangle_condition} implies that $\Delta=2$, where $\Delta$ is the gap exponent.
\subsection{Discussion of literature and results}\label{sec:intro:literature}
Percolation theory is a fundamental part of contemporary probability theory and its foundations are generally attributed to a 1957 paper of Broadbent and Hammersley \cite{BroHam57}.
Meanwhile, a number of textbooks appeared, and we refer to Grimmett~\cite{Gri99} for a comprehensive treatment of the subject, as well as Bollob{\'a}s and Riordan~\cite{BolRio06}, Werner~\cite{Wer09} and Beffara and Duminil-Copin~\cite{BefDum13} for extra emphasis on the exciting recent progress in two-dimensional percolation.
The investigation of percolation in high dimensions was started by the seminal 1990 paper of Hara and Slade~\cite{HarSla90}, who applied the lace expansion to prove the triangle condition for bond percolation in high dimension.
A number of modifications and extensions of the lace expansion method for bond percolation have appeared in the meantime. The expansion itself is presented in Slade's Saint Flour notes \cite{Sla06}. A detailed account of the full lace expansion proof for bond percolation (including convergence of the expansion and related results) is given in a recent textbook by the first author and van der Hofstad~\cite{HeyHof17}.
Despite the fantastic understanding of \emph{bond} percolation in high dimensions, \emph{site} percolation is not yet analyzed with this method, and the present paper aims to remedy this situation.
Together with van der Hofstad and Last~\cite{HeyHofLasMat19}, we recently applied the lace expansion to the random connection model, which can be viewed as a continuum site percolation model. The aim of this paper is to give a rigorous exhibition of the lace expansion applied to one of the simplest site percolation lattice models.
We have chosen to set up the proofs in a similar fashion to corresponding work for bond percolation, making it easier to oversee by readers who are familiar with that literature. Indeed, Sections \ref{sec:diag_bounds} and \ref{sec:bootstrap_analysis} follow rather closely the well-established method in \cite{BorChaHofSlaSpe05,HeyHof17,HeyHofSak08}, which are all based on Hara and Slade's foundational work \cite{HarSla90}. Interestingly, there is a difference in the expansion itself, which has numerous repercussions in the diagrammatic bounds and also in the different form of the infrared bounds. We now explain these differences in more detail.
A key insight for the analysis of high-dimensional site percolation is the frequent occurrence of \emph{pivotal points}, which are crucial for the setup of the lace expansion. Suppose two sites $x$ and $y$ are connected, then an intermediate vertex $u$ is \emph{pivotal} for this connection if every occupied path from $x$ to $y$ passes through $u$. We then break up the original connection event in two parts: a connection between $x$ and $u$, and a second connection between $u$ and $y$. In high dimension, we expect the two connection events to behave rather independently, except for their joint dependence on the occupation status of $u$.
This little thought demonstrates that it is highly convenient to define connectivity events as in Section~\ref{sec:defs}, and thus treat the occupation of the vertex $u$ independent of the two new connection events.
A more conventional choice of connectivity, where two points can only be connected if they are both occupied, is obtained a posteriori via
\eqq{ \mathbb P_p(\{x\longleftrightarrow y \} \cap \{ x,y\text{ occupied}\})=p^2\,\tau_p(y-x),\qquad x,y\in\mathbb Z^d, x\neq y. \label{eq:intro:tlam_p_squared_alt_def}}
Our definition of $\tau_p$ avoids divisions by $p$, not only in the triangle, but throughout Sections~\ref{sec:lace_expansion} and~\ref{sec:diag_bounds}.
Paying close attention to the right number of factors of $p$ is a guiding thread of the technical aspects of Sections~\ref{sec:diag_bounds} and~\ref{sec:bootstrap_analysis} that sets site percolation apart from bond percolation. Like it is done in bond percolation, the diagrammatic events by which we bound the lace-expansion coefficients depend on quite a few more points than the pivotal points. In contrast, every pair of points among these that may coincide hides a case distinction, and a coincidence case leads to a new diagram, typically with a smaller number of factors of $p$ (see, e.g.,~\eqref{eq:db:Pi_bound_by_F}). We handle this, for example, by encoding such coincidences in $\tau_p^\circ$ and $\tau_p^\bullet$ (which are variations of $\tau_p$ respecting an extra factor of $p$ and/or interactions at $\mathbf{0}$), see Definition~\ref{def:db:modified_two-point_functions}. In Section~\ref{sec:bootstrap_analysis}, the mismatching number of $p$'s and $\tau_p$'s in $\triangle_p$ needs to be resolved.
The differing diagrams in Section~\ref{sec:diag_bounds} due to coincidences already appear in the lace-expansion coefficients of small order. This manifests itself in the answer to a classical question for high-dimensional percolation; namely, to devise an expansion of the critical threshold $p_c(d)$ when $d\to\infty$. It is known in the physics literature that
\begin{equation}\label{eq:expansion-phys}
p_c(d) = (2d)^{-1} + \frac 52 (2d)^{-2} + \frac{31}{4}(2d)^{-3} + \frac{75}{4} (2d)^{-4} + \frac{11977}{48}(2d)^{-5}+ \frac{209183}{96}(2d)^{-6}+\cdots.
\end{equation}
The first four terms are due to Gaunt, Ruskin and Sykes~\cite{GauRusSyk76}, the latter two were found recently by Mertens and Moore~\cite{MerMoo18} by exploiting involved numerical methods.
The lace expansion devised in this paper enables us to give a rigorous proof of the first terms of \eqref{eq:expansion-phys}. Indeed, we use the representation obtained in this paper to show that
\eqq{ p_c(d) = (2d)^{-1} + \frac 52 (2d)^{-2} + \frac{31}{4} (2d)^{-3} + \mathcal O\left( (2d)^{-4} \right) \quad\text{as $d \to \infty$}. \label{eq:expansion}}
This is the content of a separate paper~\cite{HeyMat19b}. Deriving $p_c$ expansions from lace expansion coefficients has been earlier achieved for bond percolation by Hara and Slade~\cite{HarSla95} and van der Hofstad and Slade~\cite{HofSla06}. \col{Comparing~\eqref{eq:expansion} to their expansion for bond percolation confirms that already the second coefficient is different.
Proposition~\ref{thm:convergence_of_LE} proves the convergence of the lace expansion for $p<p_c$, yielding an identity for $\tau_p$ of the form
\eqq{ \tau_p(x) = C(x) + p(C \ast \tau_p)(x), \label{eq:intro:OZE}}
where $C$ is the \emph{direct-connectedness function} and $C(\cdot)=2d D(\cdot) + \Pi_p(\cdot)$ (for a definition of $\Pi_p$, see Definition~\ref{def:le:lace_expansion_coefficients} and Proposition~\ref{thm:convergence_of_LE}). In fluid-state statistical mechanics,~\eqref{eq:intro:OZE} is known as the \emph{Ornstein-Zernike equation} (OZE), a classical equation that is typically associated to the total correlation function.
We can juxtapose~\eqref{eq:intro:OZE} with the converging lace expansion for bond percolation. There, $\tau_p^{\text{bond}}(x)$ is the probability that $\mathbf{0}$ and $x$ are connected by a path of occupied bonds, and we have
\eqq{ \tau_p^{\text{bond}}(x) = C^{\text{bond}}(x) + 2dp (C^{\text{bond}} \astD\ast\ \tau_p^{\text{bond}})(x), \label{eq:intro:OZE_bond}}
where $C^{\text{bond}}(x) = \mathds 1_{\{x = \mathbf{0}\}} + \Pi_p^{\text{bond}}(x)$. Thus, there is an extra convolution with $2dD$. Only for the site percolation two-point function (as defined in this paper), the lace expansion coincides with the OZE.
We want to touch on how this relates to the infra-red bound~\eqref{eq:intro:infra-red_bound}. To this end,} define the random walk Green's function as $G_\lambda(x) = \sum_{m \geq 0} \lambda^m D^{\ast m}(x)$ for $\lambda\in(0,1]$. Consequently,
\[ \widehat G_\lambda(k) = \frac{1}{1-\lambda \widehat\connf(k)}.\]
One of the key ideas behind the lace expansion for bond percolation is to show that the two-point function is close to $G_\lambda$ in an appropriate sense (this includes an appropriate parametrization of $\lambda$). Solving the OZE in Fourier space for $\widehat\tau_p$ already hints at the fact that in site percolation, $p\widehat\tau_p$ should be close to $ \widehat G_\lambda\widehat D$ and $p\tau_p$ should be close to $D\astG_\lambda$. As a technical remark, we note that $\widehat G_\lambda$ is uniformly lower-bounded, whereas $\widehat G_\lambda \widehat\connf$ is not, which poses some inconvenience later on.
The complete graph may be viewed as a mean-field model for percolation, in particular when we analyze clusters on high-dimensional tori, cf.\ \cite{HeyHof07}. Interestingly, the distinction between bond and site percolation exhibits itself rather drastically on the complete graph: for bond percolation, we obtain the usual Erd\H{o}s-R\'enyi random graph with its well-known phase transition (see, e.g.,~\cite{JanLucRuc00}), whereas for site percolation, we obtain again a complete graph with a binomial number of points.
Theorem \ref{thm:main_theorem_triangle_condition} proves the triangle condition in dimension $d> d_0$ for sufficiently large $d_0$. It is folklore in the physics literature that $d_0=6$ suffices (6 is the ``upper critical dimension'') but the perturbative nature of our argument does not allow us to derive that. Instead, we only get the result for \emph{some} $d_0\ge 6$.
For bond percolation, already the original paper by Hara and Slade \cite{HarSla90} treated a second, spread-out version of bond percolation, and they proved that for this model, $d_0=6$ suffices (under suitable assumption on the spread-out nature). For ordinary bond percolation, it was announced that $d_0=19$ suffices for the triangle condition in \cite{HarSla94}, and the number 19 circulated for many years in the community. Finally, Fitzner and van der Hofstad~\cite{FitHof17} devised involved numerical methods to rigorously verify that an adaptation of the method is applicable for $d > d_0=10$.
It is clear that an analogous result of Theorem \ref{thm:main_theorem_triangle_condition} would hold for ``spread-out site percolation'' in suitable form (see e.g. \cite[Section 5.2]{HeyHof17}).
\subsection{Outline of the paper} \label{sec:intro:outine}
The paper is organized as follows. The aim of Section~\ref{sec:lace_expansion} is to establish a lace-expansion identity for $\tau_p$, which is formulated in Proposition~\ref{thm:lace_expansion}. To this end, we use Section~\ref{sec:le:tools} to state some known results that we are going to make use of in Section~\ref{sec:lace_expansion} as well as in later sections. We then introduce a lot of the language and quantities needed to state Proposition~\ref{thm:lace_expansion} in Section~\ref{sec:le:preparatory_statements}, followed by the actual derivation of the identity in Section~\ref{sec:le:derivation_of_the_LE}.
Section~\ref{sec:diag_bounds} bounds the lace-expansion coefficients derived in Section~\ref{sec:le:derivation_of_the_LE} in terms of simpler diagrams, which are large sums over products of two-point (and related) functions. Section~\ref{sec:bootstrap_analysis} finishes the argument via the so-called bootstrap argument. First, a \emph{bootstrap function} $f$ is introduced in Section~\ref{sec:boot:intro}. Among other things, it measures how close $\widehat\tau_p$ is to $\widehat G_\lambda$ (in a fractional sense). Section~\ref{sec:boot:consequences} shows convergence of the lace expansion for fixed $p<p_c$. Moreover, assuming that $f$ is bounded on $[0,p_c)$, it is shown that this convergence is uniform in $p$ (see first and second part of Proposition~\ref{thm:convergence_of_LE}). Lastly, Section~\ref{sec:boot:bootstrap_argument} actually proves said boundedness of $f$.
\section{The expansion}\label{sec:lace_expansion}
\subsection{The standard tools} \label{sec:le:tools}
We require two standard tools of percolation theory, namely Russo's formula and the BK inequality, both for increasing events. Recall that $A$ is called \emph{increasing} if $\omega \in A$ and $\omega \subseteq \omega'$ implies $\omega' \in A$. Given $\omega$ and an increasing event $A$, we introduce
\[ \piv{A} = \{ y \in \mathbb Z^d: \omega \cup\{y\} \in A, (\omega\setminus\{y\}) \notin A\}.\]
If $A$ is an increasing event determined by sites in $\Lambda \subset \mathbb Z^d$ with $|\Lambda|<\infty$, then Russo's formula~\cite{Rus81}, proved independently by Margulis~\cite{Mar74}, tells us that
\eqq{ \frac{\, \mathrm{d}}{\, \mathrm{d} p} \mathbb P_p(A) = \mathbb E[|\piv{A}|] = \sum_{y \in \Lambda} \mathbb P_p ( y \in \piv{A}). \label{eq:tools:russo}}
To state the BK inequality, let $\Lambda\subset \mathbb Z^d$ be finite and, given $\omega\in\Omega$, let
\[ [\omega]_\Lambda = \{ \omega' \in \Omega: \omega'(x) = \omega(x) \text{ for all } x \in\Lambda\} \]
be the cylinder event of the restriction of $\omega$ to $\Lambda$. For two events $A,B$, we can define the \emph{disjoint occurrence} as
\[ A \circ B = \{ \omega: \exists K,L \subseteq \mathbb Z^d: K \cap L = \varnothing, [\omega]_K \subseteq A, [\omega]_L \subseteq B\}.\]
The BK inequality, proved by van den Berg and Kesten~\cite{BerKes85} for increasing events, states that, given two increasing events $A$ and $B$,
\eqq{ \mathbb P_p(A \circ B) \leq \mathbb P_p(A) \mathbb P_p(B). \label{eq:tools:bk}}
The following proposition about simple random walk will be of importance later:
\begin{prop}[Random walk triangle,~\cite{HeyHof17}, Proposition 5.5] \label{thm:random_walk_triangle}
Let $m \in\mathbb N_0, n\geq 0$ and $\lambda\in[0,1]$. Then there exists a constant $c_{2m,n}^{\text{(RW)}}$ independent of $d$ such that, for $d>2n$,
\[ \int_{(-\pi,\pi]^d} \frac{ \widehat\connf(k)^{2m}}{[1-\lambda\widehat\connf(k)]^n} \frac{\, \mathrm{d} k}{(2\pi)^d} \leq c_{2m,n}^{\text{(RW)}} d^{-m}. \]
\end{prop}
In~\cite{HeyHof17}, $d>4n$ is required; however, more careful analysis shows that $d>2n$ suffices (see~\cite[(2.19)]{BorChaHofSlaSpe05}). We will also need the following related result:
\begin{prop}[Related random walk bounds, \cite{HeyHof17}, Exercise 5.4] \label{thm:random_walk_triangle_related}
Let $m \in\{0,1\}$, $\lambda\in[0,1]$, and $r,n \geq 0$ such that $d>2(n+r)$. Then, uniformly in $k\in(-\pi,\pi]^d$,
\al{ \int_{(-\pi,\pi]^d} \widehat\connf(l)^{2m} \widehat G_\lambda(l)^n \tfrac 12 \big[ \widehat G_\lambda(l+k) + \widehat G_\lambda(l-k) \big]^{r} \frac{\, \mathrm{d} l}{(2\pi)^d} & \leq c^{\text{(RW)}}_{2m,n+r} d^{-m}, \\
\int_{(-\pi,\pi]^d} \widehat\connf(l)^{2m} \widehat G_\lambda(l)^{n-1} \big[ \widehat G_\lambda(l+k) \widehat G_\lambda(l-k) \big]^{r/2} \frac{\, \mathrm{d} l}{(2\pi)^d} & \leq c^{\text{(RW)}}_{2m,n-1+r/2} d^{-m}, }
where the constants $c^{\text{(RW)}}_{\cdot,\cdot}$ are from Proposition~\ref{thm:random_walk_triangle}.
\end{prop}
The following differential inequality is an application of Russo's formula and the BK inequality. It applies them to events which are not determined by a finite set of sites. We refer to the literature~\cite[Lemma 4.4]{HeyHof17} for arguments justifying this and for a more detailed proof. Observation~\ref{obs:tools:diff_inequality} will be of use in Section~\ref{sec:bootstrap_analysis}.
\begin{observation} \label{obs:tools:diff_inequality}
Let $p < p_c$. Then
\[ \frac{\, \mathrm{d}}{\, \mathrm{d} p} \widehat\tau_p(0) \leq \widehat\tau_p(0)^2, \qquad \frac{\, \mathrm{d}}{\, \mathrm{d} p} \chi(p) \leq \chi(p) \widehat\tau_p(0).\]
\end{observation}
As a proof sketch, note that
\al{ \frac{\, \mathrm{d}}{\, \mathrm{d} p} \widehat\tau_p(0) &= \sum_{x \in\mathbb Z^d} \frac{\, \mathrm{d}}{\, \mathrm{d} p} \tau_p(x) = \sum_{x \in\mathbb Z^d} \sum_{y\in\mathbb Z^d} \mathbb P_p\big(y \in \piv{\mathbf{0} \longleftrightarrow x} \big)
\leq \sum_{x \in\mathbb Z^d} \sum_{y\in\mathbb Z^d} \mathbb P_p\big( \{\mathbf{0} \longleftrightarrow y\}\circ\{y \longleftrightarrow x\} \big) \\
& \leq \sum_{x \in\mathbb Z^d} \sum_{y\in\mathbb Z^d} \tau_p(y) \tau_p(x-y) = \widehat\tau_p(0)^2. }
The inequality for $\chi(p)$ follows from the identity $\chi(p) = 1+p \widehat\tau_p(0)$.
\subsection{Definitions and preparatory statements}\label{sec:le:preparatory_statements}
We need the following definitions:
\begin{definition}[Elementary definitions] \label{def:le:elementary_definitions} Let $x,u\in\mathbb Z^d$ and $A \subseteq \mathbb Z^d$.
\begin{enumerate}
\item We set $\omega^x := \omega \cup \{x\}$ and $\omega^{u,x} := \omega \cup \{u,x\}$.
\item We define $J(x) := \mathds 1_{\{|x|=1\}} = 2d D(x)$.
\item Let $\{\conn{u}{x}{A}\}$ be the event that $\{u \longleftrightarrow x\}$, and there is a path from $u$ to $x$, all of whose inner vertices are elements of $\omega\cap A$. Moreover, write $\{\offconn{u}{x}{A}\} :=\{\conn{u}{x}{\mathbb Z^d \setminus A}\}$.
\item We define $\{u \Longleftrightarrow x\} := \{u \longleftrightarrow x\} \circ \{u \longleftrightarrow x\} $ and say that $u$ and $x$ are \emph{doubly connected}.
\item We define the modified cluster of $x$ with a designated vertex $u$ as
\[\widetilde\mathscr {C}^{u}(x) := \{x\} \cup \{y \in \omega \setminus\{u\} : x \longleftrightarrow y \text{ in } \mathbb Z^d \setminus\{u\} \} . \]
\item For a set $A \subset \mathbb Z^d$, define $\langle A \rangle := A \cup \{y \in \mathbb Z^d: \exists x \in A \text{ s.t.~} |x-y| = 1\}$ as the set $A$ itself plus its external boundary.
\end{enumerate}
\end{definition}
Definition~\ref{def:le:elementary_definitions}.1 allows us to speak of events like $\{\conn{a}{b}{\omega^x}\}$ for $a,b\in\mathbb Z^d$, which is the event that $a$ is connected to $b$ in the configuration where $x$ is fixed to be occupied. We remark that $\{\conn{x}{y}{\mathbb Z^d}\} = \{x \longleftrightarrow y\} = \{\conn{x}{y}{\omega}\}$ and that $\{u \Longleftrightarrow x\} = \Omega$ for $|u-x| = 1$. Similarly, $\{u \Longleftrightarrow x\} = \varnothing$ for $u=x$.
The following, more specific definitions are important for the expansion:
\begin{definition}[Extended connection probabilities and events] \label{def:le:extended_connection_stuff} Let $v,u,x \in\mathbb Z^d$ and $A \subseteq \mathbb Z^d$.
\begin{enumerate}
\item Define
\[ \{u \throughconn{A} x \} := \{u \longleftrightarrow x\} \cap \Big( \{\noffconn{u}{x}{\thinn{A}}\} \cup \{x \in \thinn{A} \} \Big). \]
In words, this is the event that $u$ is connected to $x$, but either any path from $u$ to $x$ has an interior vertex in $\thinn{A}$, or $x$ itself lies in $\thinn{A}$.
\item Define
\[ \tau_p^{A}(u,x) := \mathds 1_{\{x \notin \thinn{A}\}}\mathbb P_p(\offconn{u}{x}{\thinn{A}}). \]
\item We introduce $\piv{u,x} := \piv{u \longleftrightarrow x}$ as the set of pivotal points for $\{u \longleftrightarrow x\}$. That is, $v \in\piv{u,x}$ if the event $\{\conn{u}{x}{\omega^v}\}$ holds but $\{\conn{u}{x}{\omega\setminus\{v\}}\}$ does not.
\item Define the events
\al{ E'(v,u;A) &:= \{v \throughconn{A} u\} \cap \{ \nexists u' \in \piv{v,u}: v \throughconn{A} u'\}, \\
E(v,u,x;A) &:= E'(v,u;A) \cap \{u \in \omega \cap \piv{v,x} \}. }
\end{enumerate}
\end{definition}
First, we remark that $\{u \throughconn{\mathbb Z^d} x\} = \{ u \longleftrightarrow x\}$. Secondly, note that we have the relation
\eqq{ \tau_p(x-u)= \tau_p^{A}(u,x)+ \mathbb P_p(u \throughconn{A} x). \label{eq:le:taup_incl_excl_identity}}
We next state a partitioning lemma (whose proof is left to the reader; see~\cite[Lemma 3.5]{HeyHofLasMat19}) relating the events $E$ and $E'$ to the connection event $ \{u \throughconn{A} x \}$:
\begin{lemma} [Partitioning connection events] \label{lem:partitioning_lemma}
Let $v,x\in\mathbb Z^d$ and $A \subseteq \mathbb Z^d$. Then
\[ \{v \throughconn{A} x \} = E'(v,x;A) \cup \bigcup_{u\in \mathbb Z^d} E(v, u, x; A), \]
and the appearing unions are disjoint.
\end{lemma}
The next lemma, titled the Cutting-point lemma, is at the heart of the expansion:
\begin{lemma}[Cutting-point lemma] \label{lem:cutting_point_lemma}
Let $v,u,x\in\mathbb Z^d$ and $A \subseteq \mathbb Z^d$. Then
\[ \mathbb P_p(E(v,u,x;A)) = p \mathbb E_p \left[\mathds 1_{E'(v,u;A)} \tau_p^{\widetilde\mathscr {C}^u(v)}(u,x) \right].\]
\end{lemma}
\begin{proof}
The proof is a special case of the general setting of \cite{HeyHofLasMat19}. Since it is essential, we present it here. We abbreviate $\widetilde\mathscr {C} = \widetilde\mathscr {C}^u(v)$ and observe that
\al{ \{u \in \piv{v,x} \} &= \{v \longleftrightarrow u\} \cap \{\offconn{u}{x}{\widetilde\mathscr {C}} \} \cap \{x \notin \thinn{\widetilde\mathscr {C}} \} \\
&= \{v \longleftrightarrow u\} \cap \{\offconn{u}{x}{\thinn{\widetilde\mathscr {C}}} \} \cap \{ x \notin \thinn{\widetilde\mathscr {C}} \} .}
In the above, we can replace $\widetilde\mathscr {C}$ by $\thinn{\widetilde\mathscr {C}}$ in the middle event, as, by definition, we know that, apart from $u$, any site in $\thinn{\widetilde \mathscr {C}} \setminus \widetilde \mathscr {C}$ must be vacant. Now, since $E'(v,u;A) \subseteq \{ v \longleftrightarrow u\}$, we get
\[ E(v,u,x;A) = E'(v,u;A) \cap \{\offconn{u}{x}{\thinn{\widetilde\mathscr {C}}} \} \cap \{x \notin \thinn{\widetilde\mathscr {C}} \}\cap \{ u\in\omega\}.\]
Taking probabilities, conditioning on $\widetilde \mathscr {C}$, and observing that the status of $u$ is independent of all other events, we see
\[ \mathbb P_p(E(v,u,x;A)) = p \mathbb E_p \left[ \mathds 1_{E'(v,u;A)} \mathds 1_{\{x \notin \thinn{\widetilde\mathscr {C}}\}} \mathbb E_p \left[ \mathds 1_{\{\offconn{u}{x}{\thinn{\widetilde\mathscr {C}}} \}} | \widetilde \mathscr {C} \right] \right],\]
making use of the fact that the first two events are measurable w.r.t.~$\widetilde\mathscr {C}$. The proof is complete with the observation that under $\mathbb E_p$, almost surely,
\[ \mathds 1_{\{x \notin \thinn{\widetilde\mathscr {C}} \}} \mathbb E_p \left[ \mathds 1_{\{\offconn{u}{x}{\thinn{\widetilde\mathscr {C}}} \}} | \widetilde \mathscr {C} \right] = \tau_p^{\widetilde\mathscr {C}}(u,x). \qedhere \]
\end{proof}
\subsection{Derivation of the expansion} \label{sec:le:derivation_of_the_LE}
We introduce a sequence $(\omega_i)_{i\in\mathbb N_0}$ of independent site percolation configurations. For an event $E$ taking place on $\omega_i$, we highlight this by writing $E_i$. We also stress the dependence of random variables on the particular configuration they depend on. For example, we write $\mathscr {C}(u; \omega_i)$ to denote the cluster of $u$ in configuration $i$.
\begin{definition}[Lace-expansion coefficients] \label{def:le:lace_expansion_coefficients}
Let $m\in\mathbb N, n\in\mathbb N_0$ and $x\in\mathbb Z^d$. We define
\al{ \Pi_p^{(0)}(x) &:= \mathbb P_p(\mathbf{0} \Longleftrightarrow x) - J(x), \\
\Pi_p^{(m)}(x) &:= p^m \sum_{u_0, \ldots, u_{m-1}} \mathbb P_p \Big( \{\mathbf{0} \Longleftrightarrow u_0\}_0 \cap \bigcap_{i=1}^{m} E'(u_{i-1},u_i; \mathscr C_{i-1})_i \Big), }
where we recall that $J(x) = \mathds 1_{\{|x|=1\}}$ and moreover $u_{-1}=\mathbf{0}, u_m=x$, and $\mathscr {C}_{i} = \widetilde\mathscr {C}^{u_i}(u_{i-1};\omega_i)$. Let
\al{R_{p,n}(x) &:= (-p)^{n+1} \sum_{u_0, \ldots, u_n} \mathbb P_p \Big( \{\mathbf{0} \Longleftrightarrow u_0\}_0 \cap \bigcap_{i=1}^{n} E'(u_{i-1},u_i; \mathscr C_{i-1})_i
\cap \{u_n \throughconn{\mathscr {C}_n} x\}_{n+1} \Big).}
Finally, set
\[ \Pi_{p,n}(x) = \sum_{m=0}^{n} (-1)^m \Pi_p^{(m)}(x).\]
\end{definition}
It should be noted that the events $E'(u_{i-1},u_i; \mathscr C_{i-1})_i$ appearing in Definition~\ref{def:le:lace_expansion_coefficients} take place on configuration $i$ only if $\mathscr {C}_{i-1}$ is taken to be a fixed set---otherwise, they are events determined by configurations $i-1$ and $i$.
\begin{prop}[The lace expansion] \label{thm:lace_expansion}
Let $p< p_c$, $x\in\mathbb Z^d$, and $n\in\mathbb N_0$. Then
\[ \tau_p(x) = J(x) + \Pi_{p,n}(x) + p \big((J+\Pi_{p,n})\ast \tau_p) (x) + R_{p,n}(x). \]
\end{prop}
\begin{proof}
We have
\[\tau_p(x) = J(x) + \Pi_p^{(0)}(x) + \mathbb P_p(\mathbf{0} \longleftrightarrow x, \mathbf{0} \centernot\Longleftrightarrow x).\]
We can partition the last summand via the first pivotal point. Pointing out that $\{\mathbf{0} \Longleftrightarrow u\} = E'(0,u;\mathbb Z^d)$, we obtain
\al{ \mathbb P_p(\mathbf{0} \longleftrightarrow x, \mathbf{0} \centernot\Longleftrightarrow x)
&= \sum_{u_0 \in\mathbb Z^d} \mathbb P_p(\mathbf{0} \Longleftrightarrow u_0, u_0 \in \omega, u_0\in\piv{\mathbf{0}, x}) = \sum_{u_0} \mathbb P_p(E(\mathbf{0},u_0,x;\mathbb Z^d)) \\
&= p \sum_{u_0} \mathbb E_p\left[ \mathds 1_{\{\mathbf{0} \Longleftrightarrow u_0\}_0} \cdot \tau_p^{\mathscr {C}_0} (u_0,x) \right] }
via the Cutting-point lemma~\ref{lem:cutting_point_lemma}. Using~\eqref{eq:le:taup_incl_excl_identity} for $A=\mathscr {C}_0$, we have
\eqq{\tau_p(x) = J(x) + \Pi_p^{(0)}(x) + p \sum_{u_0} \left(J(u_0) + \Pi_p^{(0)}(u_0)\right) \tau_p(x-u_0)
- p \sum_{u_0} \mathbb E_p\left[ \mathds 1_{\{\mathbf{0} \Longleftrightarrow u_0\}_0} \cdot
\mathbb P_p(u_0 \throughconn{\mathscr {C}_0} x) \right] . \label{eq:lace_expansion_step_0}}
This proves the expansion identity for $n=0$. Next, Lemma~\ref{lem:partitioning_lemma} and Lemma~\ref{lem:cutting_point_lemma} yield
\al{\mathbb P_p(u_0\throughconn{A} x) &= \mathbb P_p(E'(u_0,x;A)) + \sum_{u_1\in \mathbb Z^d} \mathbb P_p(E(u_0,u_1,x;A)) \\
&= \mathbb P_p(E'(u_0,x;A)) + p \sum_{u_1\in \mathbb Z^d} \mathbb E_p \left[ \mathds 1_{E'(u_0,u_1;A)} \cdot \tau_p^{\widetilde\mathscr {C}^{u_1}(u_0)}(u_1,x) \right]. }
Plugging this into~\eqref{eq:lace_expansion_step_0}, we use~\eqref{eq:le:taup_incl_excl_identity} for $A=\widetilde \mathscr {C}^{u_1}(u)$ to extract $\Pi_p^{(1)}$ and get
\al{ \tau_p(x) & = J(x) + \Pi_p^{(0)}(x) - \Pi_p^{(1)}(x) + p \big((J + \Pi_p^{(0)}) \ast\tau_p\big)(x) \\
& \quad + p^2 \sum_{u_1} \tau_p(x-u_1) \sum_{u_0} \mathbb P_p \Big( \{\mathbf{0} \Longleftrightarrow u_0\}_0 \cap E'(u_0,u_1;\mathscr {C}_0)_1 \Big) + R_{p,1}(x) \\
&= J(x) + \Pi_p^{(0)}(x) - \Pi_p^{(1)}(x) + p \big((J + \Pi_p^{(0)}-\Pi_p^{(1)}) \ast\tau_p\big)(x) + R_{p,1}(x) .}
Note that all appearing sums are bounded by $\sum_{y} \tau_p(y)$. This sum is finite for $p<p_c$, justifying the above changes in order of summation. The expansion for general $n$ is an induction on $n$ where the step is analogous to the step $n=1$ (but heavier on notation).
\end{proof}
\section{Diagrammatic bounds} \label{sec:diag_bounds}
\subsection{Setup, bounds for $n=0$} \label{sec:db:n0}
We use this section to state Lemma~\ref{lem:cosinesplitlemma} and state bounds on $\Pi_p^{(0)}$, which are rather simple to prove. The more involved bounds on $\Pi_p^{(n)}$ for $n\geq 1$ are dealt with in Section~\ref{sec:db:no_displacement}. Note that if $f(-x) = f(x)$, then $\widehat f(k) = \sum_{x \in \mathbb Z^d} \cos (k \cdot x) f(x)$. We furthermore have the following tool at our disposal:
\begin{lemma}[Split of cosines, \cite{FitHof16}, Lemma 2.13] \label{lem:cosinesplitlemma}
Let $t \in \mathbb R$ and $t_i \in \mathbb R$ for $i=1, \ldots, n$ such that $t = \sum_{i=1}^{n} t_i$. Then
\[1-\cos (t) \leq n \sum_{i=1}^{n} [1- \cos(t_i)]. \]
\end{lemma}
We begin by treating the coefficient for $n=0$, giving a glimpse into the nature of the bounds to follow in Sections~\ref{sec:db:no_displacement} and~\ref{sec:db:displacement}. To this end, we define the two displacement quantities
\[J_k(x) := [1-\cos(k\cdot x)] J(x) \quad \text{and} \quad \tau_{p,k}(x) = [1-\cos(k\cdot x)] \tau_p(x). \]
\begin{prop}[Bounds for $n=0$] \label{thm:db:bounds_for_n0}
For $k\in (-\pi,\pi]^d$,
\al{|\widehat\Pi_p^{(0)}(k)| & \leq p^2 (J^{\ast 2} \ast \tau_p^{\ast 2})(\mathbf{0}), \\
|\widehat\Pi_p^{(0)}(0)-\widehat\Pi_p^{(0)}(k)| & \leq 2p^2 \Big((J_k\astJ\ast\tau_p^{\ast 2})(\mathbf{0}) + (J^{\ast 2}\ast\tau_{p,k}\ast\tau_p)(\mathbf{0}) \Big). }
\end{prop}
\begin{proof}
Note that $|x| \leq 1$ implies $\Pi_p^{(0)}(x) =0$ by definition. For $|x| \geq 2$, we have
\[\Pi_p^{(0)}(x) \leq \mathbb E \bigg[\sum_{y \neq z \in \omega} \mathds 1_{\{|y|=|z|=1\}}\mathds 1_{\{y \longleftrightarrow x\}\circ\{z \longleftrightarrow x\}} \bigg]
\leq p^2 \bigg( \sum_{y\in\mathbb Z^d} J(y)\tau_p(x-y) \bigg)^2 = p^2 (J\ast\tau_p)(x)^2.\]
Summation over $x$ gives the first bound. The last bound is obtained by applying Lemma~\ref{lem:cosinesplitlemma} to the bounds derived for $\Pi_p^{(0)}(x)$:
\al{ |\widehat\Pi_p^{(0)}(0)&-\widehat\Pi_p^{(0)}(k)| = \sum_{x} [1-\cos(k\cdot x)] \Pi_p^{(0)}(x) \\
& \leq 2p^2 \sum_x (J\ast\tau_p)(x) \sum_{y} \Big( [1-\cos(k\cdot y)] J(y) \tau_p(x-y) + J(y) [1-\cos(k\cdot(x-y))] \tau_p(x-y) \Big).}
Resolving the sums gives the claimed convolution. \end{proof}
\subsection{Bounds in terms of diagrams} \label{sec:db:no_displacement}
The main result of this section is Proposition~\ref{thm:db:main_thm}, providing bounds on the lace-expansion coefficients in terms of so-called diagrams, which are sums over products of two-point (and related) functions. To state it, we introduce some functions related to $\tau_p$ as well as several ``modified triangles'' closely related to $\triangle_p$. We denote by $\delta_{x,y} = \mathds 1_{\{x=y\}}$ the Kronecker delta.
\begin{definition}[Modified two-point functions] \label{def:db:modified_two-point_functions}
Let $x\in\mathbb Z^d$ and define
\[ \tau_p^\circ(x) := \delta_{\mathbf{0},x} + \tau_p(x), \qquad \tau_p^\bullet(x) = \delta_{\mathbf{0},x} + p\tau_p(x), \qquad \gamma_p(x) = \tau_p(x)-J(x).\]
\end{definition}
\begin{definition}[Modified triangles] \label{def:db:modified:triangles}
We let $\triangle_p^\circ(x) = p^2(\tau_p^\circ\ast\tau_p\ast\tau_p)(x), \triangle_p^{\bullet}(x) = p(\tau_p^\bullet\ast\tau_p\ast\tau_p)(x), \triangle_p^{\bullet\circ}(x) = p(\tau_p^\bullet\ast\tau_p^\circ\ast\tau_p)(x)$, and $\triangle_p^{\bullet\bullet\circ}(x) = (\tau_p^\bullet\ast\tau_p^\bullet\ast\tau_p^\circ)(x)$. We also set
\[\triangle_p^\circ = \sup_{x \in \mathbb Z^d} \triangle_p^\circ(x), \quad \triangle_p^{\bullet} = \sup_{\mathbf{0} \neq x \in\mathbb Z^d} \triangle_p^{\bullet}(x), \quad \triangle_p^{\bullet\circ} = \sup_{\mathbf{0} \neq x \in\mathbb Z^d} \triangle_p^{\bullet\circ}(x), \quad \triangle_p^{\bullet\bullet\circ} = \sup_{x\in\mathbb Z^d} \triangle_p^{\bullet\bullet\circ}(x), \]
and $T_p := (1+\triangle_p) \triangle_p^{\bullet\circ} + \triangle_p \triangle_p^{\bullet\bullet\circ}$.
\end{definition}
Definitions~\ref{def:db:modified_two-point_functions} and~\ref{def:db:modified:triangles} allow us to properly keep track of factors of $p$, which turns out to be important throughout Section~\ref{sec:diag_bounds}.
\begin{prop}[Triangle bounds on the lace-expansion coefficients] \label{thm:db:main_thm}
For $n\geq 0$,
\[ p\sum_{x \in \mathbb Z^d} \Pi_p^{(n)}(x) \leq \triangle_p^{\bullet}(\mathbf{0}) \big( T_p \big)^n.\]
\end{prop}
The proof of Proposition~\ref{thm:db:main_thm} relies on two intermediate steps, successively giving bounds on $\sum \Pi_p^{(n)}$. These two steps are captured in Lemmas~\ref{lem:db:Pi_bounds_by_F} and~\ref{lem:db:pi_psi_bounds}, respectively. We first state the former lemma.
Recall that $\Pi_p^{(n)}$ is defined on independent percolation configurations $\omega_0, \ldots, \omega_n$. A crucial step in proving Proposition~\ref{thm:db:main_thm} is to group events taking place on the percolation configuration $i$, and then to use the independence of the different configurations. To this end, note that event $E'(u_{i-1}, u_i; \mathscr {C}_{i-1})_i$ takes place on configuration $i$ only if $\mathscr {C}_{i-1}$ is considered to be a fixed set. Otherwise, it is a product event made up of the connection events of configuration $i$ as well as a connection event in configuration $i-1$, preventing a direct use of the independence of the $\omega_i$. Resolving this issue is one of the goals of Lemma~\ref{lem:db:Pi_bounds_by_F}; another is to give bounds in terms of the simpler events (amenable to application of the BK inequality) introduced below in Definition~\ref{def:Bounding_Events}:
\begin{figure
\centering
\includegraphics{F_events.pdf}
\caption{The $F$ events represented graphically. For lines with double arrows, we may have coincidence of the endpoints, for lines without double arrows, we do not. The area with grey tiles indicates that its three boundary points are either all distinct or collapsed into a single point. These diagrams also serve as a pictorial representation of the function $\phi_0, \phi, \phi_n$. There, lines with double arrows represent factors of $\tau_p^\circ$ and lines without double arrows represent factors of $\tau_p$.}
\label{fig:F_events}
\end{figure}
\begin{definition}[Bounding events] \label{def:Bounding_Events}
Let $x,y\in\mathbb Z^d$. We define
\[ \{x \leftrightsquigarrow y \} := \{x \longleftrightarrow y\} \cup \{x=y\}. \]
Let now $i \in \{1, \ldots, n\}$ and let $a,b,t,w,z,u \in \mathbb Z^d$. We define
\al{F_0(a,w, u,b) &= \Big( \{w=a, |u-a| =1\} \cap \{ a \leftrightsquigarrow b\}\Big) \\
& \quad \cup \Big( \{|u-a| >1\} \cap \big(\{ a \leftrightsquigarrow w\} \circ \{ a \longleftrightarrow u\} \circ\{ w \longleftrightarrow u\} \circ\{ w \leftrightsquigarrow b\}\big)\Big), \\
F_n(a, t, z, x) &= \{|\{t, z, x\}| \neq 2\} \cap \{ a \leftrightsquigarrow t\} \\
& \quad \circ \{ t \leftrightsquigarrow x\} \circ\{ t \leftrightsquigarrow z\} \circ\{ z \leftrightsquigarrow x\}, \\
F^{(1)} (a,t,w,z,u,b) &= \{|\{w, t, z, u\}| =4\} \cap \Big( \{ a \leftrightsquigarrow t\} \circ\ \{ t \longleftrightarrow w\} \circ\ \{ t \longleftrightarrow z\} \\
&\quad \circ\ \{ w \longleftrightarrow u\} \circ\ \{ z \longleftrightarrow u\} \circ\ \{ w \leftrightsquigarrow b\}\Big), \\
F^{(2)} (a,t,w,z,u,b) &= \big\{w \notin \{z, u\}, |\{t, z, u\}| \neq 2 \big\} \cap \Big( \{ a \leftrightsquigarrow w\} \circ\ \{ w \leftrightsquigarrow b\} \\
& \quad \circ \{w \leftrightsquigarrow t \} \circ \{ t \leftrightsquigarrow u\} \circ \{ t \leftrightsquigarrow z\} \circ \{ z \leftrightsquigarrow u\} \Big). }
\end{definition}
The coincidence requirements in $F^{(2)}$ mean that among the points $t_i, w_i, z_i, u_i$, the point $w_i$ may coincide \emph{only} with $t_i$; and additionally, the triple $\{t_i,z_i,u_i\}$ are either all distinct, or collapsed into a single point. The above events are depicted in Figure~\ref{fig:F_events}.
For intervals $[a,b]$, we use the notation $\vec x_{[a,b]} = (x_a, x_{a+1}, \ldots, x_b)$. This is not to be confused with the notation $\vec v_i$ from Definition~\ref{def:Bounding_Events}. We use the notation $(\mathbb Z^d)^{(m,1)}$ to denote the set of vectors $\{\vec y_{[1,m]} \in (\mathbb Z^d)^m: y_i \neq y_{i+1} \ \forall 1 \leq i < m\}$.
\begin{lemma}[Coefficient bounds in terms of $F$ events] \label{lem:db:Pi_bounds_by_F}
For $n \geq 1$ and $(u_0, \ldots, u_{n-1}, x) \in (\mathbb Z^d)^{(n+1,1)}$,
\al{ \{\mathbf{0} \Longleftrightarrow u_0\}_0 & \cap \bigcap_{i=1}^{n} E'(u_{i-1}, u_i; \mathscr {C}_{i-1})_i \\
\subseteq \bigcup_{\substack{\vec z_{[1,n]}: z_i \in \omega_i^{u_i}, \\ w_0 \in \omega_0^\mathbf{0}, t_n \in \omega_n^{u_{n-1}, x}}} & \hspace{-1cm} F_0(\mathbf{0},w_0, u_0, z_1)_0 \cap
\bigg( \bigcap_{i=1}^{n-1} \Big( \bigcup_{\substack{ t_i \in \omega_i^{u_{i-1}},\\ w_i \in \omega_i}} F^{(1)}(\vec v_i)_i \cup
\bigcup_{\substack{w_i \in \omega_i^{u_{i-1}}, \\ t_i\in \omega_i^{u_i, w_i} }} F^{(2)}(\vec v_i)_i \Big) \bigg) \\
& \qquad \qquad \cap F_n(u_{n-1}, t_n, z_n, x)_n, }
where $\vec v_i = (u_{i-1}, t_i,w_i,z_i,u_i, z_{i+1})$ and $u_n = x$.
\end{lemma}
The proof is analogous to the one in~\cite[Lemma~4.12]{HeyHofLasMat19} and we do not perform it here. The second important lemma is Lemma~\ref{lem:db:pi_psi_bounds}, and its bounds are phrased in terms of the following functions:
\begin{definition}[The $\psi$ and $\phi$ functions] \label{def:db:psi_phi_fcts}
Let $n \geq 1$ and $a_1,a_2,b,w,t,u,z \in \mathbb Z^d$. We define
\al{ \psi_0(b,w,u) & := \delta_{b,w} p J (u-b) + p \tau_p^\bullet(w-b) \gamma_p(u-b) \tau_p(w-u), \\
\tilde\psi_0(b,w,u) &:= p \tau_p^\bullet(w-b) \tau_p(u-b) \tau_p(w-u), \\
\psi_n(a_1,a_2,t,z,x) &:= \mathds 1_{\{|\{t,z,x\}| \neq 2\}} \tau_p^\circ(z-a_1) \tau_p^\bullet(t-a_2)\tau_p^\bullet(z-t)\tau_p^\bullet(x-t)\tau_p^\circ(x-z). }
Moreover, we define
\al{ \psi^{(1)}(a_1,a_2,t,w,z,u) &:= p^3 \mathds 1_{\{|\{t,w,z,u\}| =4\}} \tau_p^\circ(z-a_1) \tau_p^\bullet(t-a_2) \tau_p(w-t) \tau_p(z-t)\tau_p(u-w)\tau_p(u-z), \\
\psi^{(2)}(a_1,a_2,t,w,z,u) &:= \mathds 1_{\{w \notin \{z,u\}, |\{t,z,u\}| \neq 2 \}} \tau_p^\circ(z-a_1) \tau_p^\bullet(w-a_2) \tau_p^\bullet(t-w) \tau_p^\bullet(z-t)\tau_p^\bullet(u-t) \tau_p^\circ(u-z),}
and $\psi := \psi^{(1)} + \psi^{(2)}$. Furthermore, for $j \in\{1,2\}$, let
\al{ \phi_0(b,w,u,z) & := \delta_{b,w} p J (u-b) \tau_p^\circ(z-w) + p \tau_p^\bullet(w-b) \gamma_p(u-b) \tau_p(w-u) \tau_p^\circ(z-w), \\
\phi_n(a_2,t,z,x) & := \mathds 1_{\{|\{t,z,x\}| \neq 2\}} \tau_p^\bullet(t-a_2)\tau_p^\bullet(z-t)\tau_p^\bullet(x-t)\tau_p^\circ(x-z), \\
\phi^{(j)}(a_2,t,w,z,u,b) &:= \frac{\tau_p^\circ(b-w)}{\tau_p^\circ(z)} \psi^{(j)}(\mathbf{0},a_2,t,w,z,u), }
and $\tilde\phi_0(b,w,u,z) := \tilde\psi_0(b,w,u) \tau_p^\circ(z-w)$ as well as $\phi:= \phi^{(1)} + \phi^{(2)}$.
\end{definition}
We remark that $\psi_0 \leq \tilde\psi_0$ as well as $\phi_0 \leq \tilde\phi_0$, and we are going to use this fact later on. In the definition of $\phi^{(j)}$, the factor $\tau_p^\circ(z)$ cancels out. In that sense, $\phi^{(j)}$ is obtained from $\psi^{(j)}$ by ``replacing'' the factor $\tau_p^\circ(z-a_1)$ by the factor $\tau_p^\circ(b-w)$, and the two functions are closely related.
We first obtain a bound on $\Pi_p^{(n)}$ in terms of the $F$ events (this is Lemma~\ref{lem:db:Pi_bounds_by_F}). Bounding those with the BK inequality, we will naturally observe the $\phi$ functions (Lemma~\ref{lem:db:pi_psi_bounds}). To decompose them further, we would like to apply induction; for this purpose, the $\psi$ functions are much better-suited. By introducing both the $\phi$ and $\psi$ functions, we increase the readability throughout this section (and later ones).
\begin{definition}[The $\Psi$ function] \label{def:db:Psi_fct}
Let $w_n, u_n \in \mathbb Z^d$ and define
\[ \Psi^{(n)}(w_n,u_n) := \sum_{\substack{\vec t_{[1,n]}, \vec w_{[0,n-1]}, \vec z_{[1,n]}, \vec u_{[0,n-1]}: \\ u_{n-1} \neq u_n}}
\psi_0(\mathbf{0},w_0,u_0) \prod_{i=1}^{n} \psi(w_{i-1},u_{i-1},t_i,w_i,z_i,u_i),\]
where $\vec t_{[1,n]}, \vec z_{[1,n]}, \vec w_{[0,n-1]} \in (\mathbb Z^d)^n$ and $\vec u_{[0,n-1]} \in (\mathbb Z^d)^{(n,1)}$.
\end{definition}
We remark that $\Psi^{(n)}(x,x)=0$, resulting from the fact that the $\phi$ functions in Definition~\ref{def:db:psi_phi_fcts} output $0$ for $w=u$. We are going to make use of this later.
\begin{lemma}[Bound in terms of $\psi$ functions] \label{lem:db:pi_psi_bounds}
For $n\geq 0$,
\eqq{ p \sum_{x\in\mathbb Z^d} \Pi_p^{(n)}(x) \leq \sum_{w,u,t,z,x \in \mathbb Z^d} \Psi^{(n-1)}(w,u) \psi_n(w,u,t,z,x) \leq \sum_{w,u \in \mathbb Z^d} \Psi^{(n)} (w,u). \label{eq:db:Pi_bound_by_Psi} }
\end{lemma}
\begin{proof}
\col{Note first that according to Definition~\ref{def:le:lace_expansion_coefficients}, the function $\Pi_p^{(n)}$ contains the event $E'(u_{i-1}, u_i; \mathscr {C}_{i-1})$, which is itself contained in $\{u_{i-1} \longleftrightarrow u_i\}$. This is why we may assume that the points $u_0, \ldots, u_n$ in the definition of $\Pi_p^{(n)}$ satisfy $u_{i-1} \neq u_i$.} Together with Lemma~\ref{lem:db:Pi_bounds_by_F}, this yields a bound on $\Pi_p^{(n)}$ of the form
\algn{p\Pi_p^{(n)}(u_n) &\leq p^{n+1} \sum_{\vec u} \mathbb P_p \bigg( \bigcup_{w_0, t_n, \vec z} F_0(\mathbf{0},w_0, u_0, z_1)_0 \cap
\bigcap_{i=1}^{n-1} \Big( \bigcup_{t_i, w_i} F^{(1)}(\vec v_i)_i \cup \bigcup_{t_i, w_i} F^{(2)}(\vec v_i)_i\Big) \notag\\
& \qquad \qquad \qquad \qquad \cap F_n(u_{n-1}, t_n, z_n, x)_n \bigg) \notag\\
& \leq \sum_{\vec t, \vec w, \vec z, \vec u} p^{|\{\mathbf{0}, w_0\}|} \mathbb P_p \big( F_0(\mathbf{0},w_0, u_0, z_1)\big) \notag\\
& \qquad \times \prod_{i=1}^{n-1} \Big( p^{|\{u_{i-1}, t_i\}| + 2} \mathbb P_p\big( F^{(1)}(\vec v_i)\big) + p^{|\{u_{i-1},w_i\}|+|\{w_i, t_i\}|
+ |\{t_i,z_i,u_i\}|-3} \mathbb P_p\big( F^{(2)}(\vec v_i)\big) \Big) \notag\\
& \qquad \times p^{|\{u_{n-1},t_n\}| + |\{t_n, z_n,u_n\}|-2 } \mathbb P_p\big(F_n(u_{n-1}, t_n, z_n, u_n) \big), \label{eq:db:Pi_bound_by_F}}
where we recall that $\vec v_i = (u_{i-1}, t_i,w_i,z_i,u_i, z_{i+1})$. In the first line, $\vec t, \vec w,\vec z$ are occupied points as in Lemma~\ref{lem:db:Pi_bounds_by_F}. In both lines, $\vec u_{[0,n]} \in (\mathbb Z^d)^{(n+1,1)}$, and in the second line, $\vec t, \vec w, \vec z \in (\mathbb Z^d)^n$. Crucially, in the second inequality of~\eqref{eq:db:Pi_bound_by_F}, the factorization occurs due to the independence of the different percolation configurations. Moreover, it is crucial here that the number of factors of $p$ (appearing when we switch from a sum over points in $\omega$ to a sum over points in $\mathbb Z^d$) depends on the number of coinciding points.
We can now decompose the $F$ events by heavy use of the BK inequality, producing bounds in terms of the $\phi$ functions introduced in Definition~\ref{def:db:psi_phi_fcts}. We start by bounding
\al{ p^{|\{a, w\}|} \mathbb P_p \big(F_0(a,w,u,z)\big) &\leq \phi_0(a,w,u,z), \\
p^{|\{a,t\}|+|\{t,z,u\}|-2} \mathbb P_p\big(F_n(a,t,z,x) \big) & \leq \phi_n(a,t,z,x). }
We continue to bound
\[ p^{|\{a,t\}| + 2} \mathbb P_p\big( F^{(1)}(a,t,w,z,u,b)\big) + p^{|\{a,w\}|+|\{w,t\}| + |\{t,z,u\}|-3} \mathbb P_p\big( F^{(2)}(a,t,w,z,u,b)\big) \Big) \leq \phi(a,t,w,z,u,b). \]
Plugging these bounds into~\eqref{eq:db:Pi_bound_by_F}, we obtain the new bound
\eqq{ p \Pi_p^{(n)}(u_n) \leq \sum_{(\vec t, \vec z)_{[1,n]}, (\vec w, \vec u)_{[0,n-1]}} \hspace{-1cm}\phi_0(\mathbf{0},w_0,u_0,z_1) \phi_n(u_{n-1},t_n,z_n,x)
\prod_{i=1}^{n-1} \phi (u_{i-1},t_i,w_i,z_i,u_i,z_{i+1}) \label{eq:db:Pi_bound_by_phi},}
where $\vec t_{[1,n]}, \vec w_{[0,n-1]}, \vec z_{[1,n]} \in (\mathbb Z^d)^n$, and $\vec u_{[0,n]}\in(\mathbb Z^d)^{(n+1,1)}$. We rewrite the right-hand side of~\eqref{eq:db:Pi_bound_by_phi} by replacing the $\phi_0, \phi_n$ and $\phi$ functions by $\psi_0, \psi_n$ and $\psi$ functions. As the additional factors arising from this replacement exactly cancel out, this gives the first bound in Lemma~\ref{lem:db:pi_psi_bounds}. The observation
\eqq{ \psi_n(a_1,a_2,t,z,u) \leq \psi(a_1,a_2,t,a_2,z,u) \label{eq:db:Psi_last_segment_bound}}
gives the second bound and finishes the proof.
\end{proof}
We can now prove Proposition~\ref{thm:db:main_thm}:
\begin{proof}[Proof of Proposition~\ref{thm:db:main_thm}] We show that
\eqq{ \sum_{w,u \in \mathbb Z^d} \Psi^{(n)} (w,u) \leq \triangle_p^{\bullet}(\mathbf{0}) \big( T_p \big)^n, \label{eq:db:Psi_triangles_bound}}
which is sufficient due to~\eqref{eq:db:Pi_bound_by_Psi}. The proof of~\eqref{eq:db:Psi_triangles_bound} is by induction on $n$. For the base case, we bound
\[ \col{\sum_{w,u \in \mathbb Z^d} \Psi^{(0)} (w,u)=}\sum_{w,u} \psi_0(\mathbf{0},w,u) \leq \sum_{w,u} \tilde\psi_0(\mathbf{0},w,u) = p \sum_{w,u} \tau_p^\bullet(w) \tau_p(u) \tau_p(u-w) = \triangle_p^{\bullet}(\mathbf{0}) . \]
Let now $n \geq 1$. Then
\begin{align} \sum_{w,u} \Psi^{(n)}(w,u) & = \sum_{w',u'} \Psi^{(n-1)}(w',u') \sum_{z,t,w,u\neq u'} \psi(w',u',t,w,z,u) \notag \\
& \leq \Big( \sum_{w',u'} \Psi^{(n-1)} (w',u') \Big) \Big( \sup_{w' \neq u'} \sum_{z,t,w,u\neq u'} \psi(w',u',t,w,z,u) \Big). \label{eq:db:inductive_step} \end{align}
The fact that $\Psi^{(n)}(x,x) = 0$ for any $n$ and $x$ allows us to assume $w' \neq u'$ in the supremum in the second line of~\eqref{eq:db:inductive_step}. After applying the induction hypothesis, it remains to bound the second factor for $w' \neq u'$, which we rewrite as $\sup_{a \neq \mathbf{0}} \sum_{t,w,z,u \neq a} \psi(\mathbf{0},a,t,w,z,u)$ by translation invariance. As it is a sum of two terms (originating from $\psi^{(1)}$ and $\psi^{(2)}$), we start with the first one and obtain
\begin{figure
\centering
\includegraphics{Psi_Diagrams.pdf}
\caption{The pictorial representation of $\sup_{\mathbf{0} \neq a}\sum_{t,w,z,u \neq a} \psi(\mathbf{0},a,t,w,z,u)$. The last two pictures represent the case distinction captured in the indicator. Points that are summed over are marked with a square.}
\label{fig:Psi_diagrams}
\end{figure}
\algn{ p^3 & \sum_{t,z} \Big( \tau_p^\bullet(t-a) \tau_p(z-t) \tau_p^\circ(z) \Big( \sum_{u,w} \tau_p(w-t) \tau_p(u-w) \tau_p(z-u) \Big)\Big) \notag \\
&\leq p \sum_{t,z} \Big( \tau_p^\bullet(t-a) \tau_p(z-t) \tau_p^\circ(z) \Big(\sup_{t,z} p^2 \sum_{u,w} \tau_p(w-t) \tau_p(u-w) \tau_p(z-u) \Big)\Big) \leq \triangle_p \triangle_p^{\bullet\circ}(a). \label{eq:db:psi_1_bound}}
Before treating the second term, we show how to obtain the bound from~\eqref{eq:db:psi_1_bound} pictorially, using diagrams very similar to the ones introduced in Figure~\ref{fig:Psi_diagrams}. In particular, factors of $\tau_p$ are represented by lines, factors of $\tau_p^\bullet$ and $\tau_p^\circ$ by lines with an added `$\bullet$' or `$\circ$', respectively. Points summed over are represented by squares, other points (which we mostly take the supremum over, for example point $a$) are represented by colored disks. Hence, we interpret the factor $\tau_p(z-t)$ as a line between $t$ and $z$. Since both endpoints are summed over, we display them as squares (and without labels $t$ or $z$). We interpret the factor $\tau_p^\circ(z)$ as a ($\circ$-decorated) line between $\mathbf{0}$ and $z$; the origin is represented by lack of decorating the incident line. Finally, we indicate the distinctness of a pair of points (in our case $\mathbf{0} \neq a$) by a disrupted two-headed arrow $\mathrel{\raisebox{-0.25 cm}{\includegraphics{Distinct_pts_arrow.pdf}}}$. With this notation,~\eqref{eq:db:psi_1_bound} becomes
\[ p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_1__general.pdf}}}
\ \leq p \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_1__decomp_1.pdf}}}
\Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}} p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_1__decomp_2.pdf}}} \Big) \Big)
\leq \triangle_p^{\bullet\circ} \triangle_p.\]
The second term in $\psi$, originating from $\psi^{(2)}$, contains an indicator. Resolving it splits this term into two further terms. We first consider the term arising from $|\{t,z,u\}| = 1$, which forces $w \neq t=u=z$, and the term is of the form
\[ p \sum_{u,w} \tau_p^\circ(u) \tau_p(w-u) \tau_p^\bullet(a-w) = p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_1__decomp_1.pdf}}} \ = \triangle_p^{\bullet\circ}(a) .\]
Turning to the term due to $|\{t,z,u\}| = 3$, with a substitution of the form $y'=y-u$ for $y\in \{t,w,z\}$ in the second line, we see that
\algn{& p^2 \sum_{t,w,z,u} \tau_p^\bullet(w-a) \tau_p^\circ(z) \tau_p^\bullet(t-w) \tau_p(z-t)\tau_p(u-t) \tau_p(z-u) \notag \\
=& p^2 \sum_{t', z'} \Big( \tau_p(z') \tau_p(t'-z') \tau_p(t') \Big( \sum_{u,w'} \tau_p^\circ(z'+u) \tau_p^\bullet(a-w'-u) \tau_p^\bullet(w-t) \Big)\Big) \notag \\
\leq & \triangle_p \triangle_p^{\bullet\bullet\circ}. \label{eq:db:psi_2_bound}}
This concludes the proof. However, we also want to show how to execute the bound in~\eqref{eq:db:psi_2_bound} using diagrams. To do so, we need to represent a substitution in pictorial form. Note that after the substitution, the sum over point $u$ is w.r.t.~two factors, namely $\tau_p^\circ(z'+u) \tau_p^\bullet(a-w'-u)$. We interpret these two factors as a line between $-u$ and $z'$ and a line between $-(u-a)$ and $w'$. In this sense, the two lines do not meet in $u$, but they have endpoints that are a constant vector $a$ apart. We represent this as
\[ \sum_u \tau_p^\circ(\textcolor{altviolet}{ z'}+u) \tau_p^\bullet(\textcolor{darkorange}{a}-\textcolor{blue}{w'}-u) = \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{subst.pdf}}}. \]
The bound in~\eqref{eq:db:psi_2_bound} thus becomes
\[ p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_2__general.pdf}}}
\ = p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_2__subst.pdf}}}
\ \leq p^2 \sum \Big( \Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}} \sum
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_2__subst_bound_1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Psi_2__subst_bound_2.pdf}}} \Big)
\leq \triangle_p^{\bullet\bullet\circ} \triangle_p,\]
where we point out that we did not use $a \neq \mathbf{0}$ for the bound $\triangle_p^{\bullet\bullet\circ}$, and so it was not indicated in the diagram.
\end{proof}
The following corollary will be needed later to show that the limit $\Pi_{p,n}$ for $n \to\infty$ exists:
\begin{corollary} \label{cor:db:Pi_pn_unsummed_bound}
For $n \geq 1$,
\[ \sup_{x\in\mathbb Z^d} \Pi_p^{(n)}(x) \leq \triangle_p^{\bullet}(\mathbf{0}) (1+\triangle_p^{\bullet\circ}) \big( T_p \big)^{n-1}. \]
\end{corollary}
\begin{proof}
Note that
\[ \Pi_p^{(n)}(x) \leq \Big( \sup_{w \neq u} \sum_{t,z} \psi_n(w,u,t,z,x) \Big) \sum_{w,u} \Psi^{(n-1)}(w,u). \]
Since we do not sum over $x$, we bound the factors depending on $x$ by $1$, and so
\[ \sum_{t,z} \psi_n(w,u,t,z,x) \leq \tau_p^\bullet(x-u)\tau_p^\circ(x-w) + \triangle_p^{\bullet\circ}(u-w) \leq 1 + \triangle_p^{\bullet\circ}(u-w) \]
implies the claim together with Proposition~\ref{thm:db:main_thm}.
\end{proof}
\subsection{Displacement bounds} \label{sec:db:displacement}
The aim of this section is to give bounds on $p \sum_x [1-\cos(k\cdot x)] \Pi_p^{(n)}(x)$. Such bounds are important in the analysis in Section~\ref{sec:bootstrap_analysis}. We regard $[1-\cos(k\cdot x)]$ as a ``displacement factor''. To state the main results, Propositions~\ref{thm:db:displacement_thm} and~\ref{thm:db:displacement_thm_n1}, we introduce some displacement quantities:
\begin{definition}[Diagrammatic displacement quantities] \label{def:displacement_quantities}
Let $x\in\mathbb Z^d$ and $k \in(-\pi,\pi]^d$. Define
\al{ W_p(x;k) &:= p(\tau_{p,k}\ast\tau_p^\circ)(x), \qquad W_p(k) = \max_{x\in\mathbb Z^d} W_p(x;k), \\
H_p(b_1, b_2;k) &:= p^5 \hspace{-0.15cm} \sum_{t,w,z,u,v} \tau_p(z) \tau_p(t-u) \tau_p(t-z) \tau_{p,k}(u-z) \tau_p(t-w) \tau_p(w-b_1) \tau_p(v-w) \tau_p(v+b_2-u), \\
H_p(k) &:=\max_{b_1\neq \mathbf{0} \neq b_2 \in\mathbb Z^d} H_p(b_1, b_2;k). }
\end{definition}
Note that Proposition~\ref{thm:db:bounds_for_n0} already provides displacement bounds for $n=0$. The following two results give bounds for $n\geq 1$:
\begin{prop}[Displacement bounds for $n \geq 2$]\label{thm:db:displacement_thm}
For $n \geq 2$ and $x\in\mathbb Z^d$,
\[ p \sum_x [1-\cos(k\cdot x)] \Pi_p^{(n)}(x) \leq 11 (n+1) \big(T_p\big)^{1 \vee (n-2)} \big(\triangle_p^{\bullet\bullet\circ}\big)^3 W_p(k) \Big[1+ \triangle_p^\circ+T_p + \frac{H_p(k)}{W_p(k)} \Big]. \]
\end{prop}
\begin{prop}[Displacement bounds for $n=1$]\label{thm:db:displacement_thm_n1}
For $x\in \mathbb Z^d$,
\[ p \sum_x [1-\cos(k\cdot x)] \Pi_p^{(1)}(x) \leq 9 W_p(k) \Big[\triangle_p^{\bullet}(\mathbf{0})\big(\triangle_p^{\bullet\circ}+\triangle_p\big) + \triangle_p^\circ + \triangle_p \Big] + p^2 (J\ast\tau_{p,k}\ast\tau_p)(\mathbf{0}).\]
\end{prop}
In preparation for the proofs, we define a function $\bar\Psi^{(n)}$, similar to $\Psi^{(n)}$, and prove an almost identical bound to the one in Proposition~\ref{thm:db:main_thm}. Let $\bar\Psi^{(0)}(t,z) = \phi_n(\mathbf{0},t,z,\mathbf{0}) / \tau_p^\bullet(t)$. For $i \geq 1$, define
\[ \bar\Psi^{(i)}(t,z) := \sum_{w,u,t', z'} \bar\Psi^{(i-1)}(t',z') \Big[\phi^{(1)}(\mathbf{0},t,w,z,u,z') + \phi^{(2)}(\mathbf{0},w,t,z,u,z') \Big] \frac{\tau_p^\bullet(t'-u)}{\tau_p^\bullet(t)}. \]
Note that in $\phi^{(2)}$, the points $t$ and $w$ swap roles, so that in both $\phi^{(1)}$ and $\phi^{(2)}$, $u$ is adjacent to $t'$ and $t$ is the point adjacent to $\mathbf{0}$; and in particular, the factor $\tau_p^\bullet(t)$ cancels out. The following lemma, in combination with Lemma~\ref{lem:db:pi_psi_bounds}, is analogous to the bound~\eqref{eq:db:Psi_triangles_bound}, and so is its proof, which is omitted.
\begin{lemma} \label{lem:db:bar_psi_bound}
For $n \geq 0$,
\[ \sum_{t,z\in\mathbb Z^d} \bar\Psi^{(n)}(t,z) \leq \triangle_p^{\bullet\bullet\circ}\big( T_p \big)^n. \]
\end{lemma}
\begin{proof}[Proof of Proposition~\ref{thm:db:displacement_thm}]
Setting $\vec v_i = (w_{i-1}, u_{i-1}, t_i,w_i,z_i,u_i)$, we use the bound
\eqq{ p \sum_x [1-\cos(k\cdot x)]\Pi_p^{(n)} (x) \leq \sum_x \hspace{-0.1cm}\sum_{\vec t,\vec w, \vec z, \vec u} [1-\cos(k\cdot x)] \psi_0(\mathbf{0},w_0,u_0) \psi_n(w_{n-1}, u_{n-1}, t_n, z_n, x)
\prod_{i=1}^{n-1} \psi(\vec v_i), \label{eq:db:disp:general_n_first_bound}}
which is, in essence, the first bound of Lemma~\ref{lem:db:pi_psi_bounds}. The next step is to distribute the displacement factor $1-\cos(k\cdot x)$ over the $n+1$ segments. To this end, we write $x = \sum_{i=0}^{n} d_i$, where $d_i =w_i-u_{i-1}$ for even $i$ and $d_i=u_i-w_{i-1}$ for odd $i$ (with the convention $u_{-1}=\mathbf{0}$ and $w_n=u_n=x$). Over the course of this proof, we drop the subscript $i$ and are then confronted with a displacement $d=d_i$ (which is not to be confused with the dimension).
Using the Cosine-split lemma~\ref{lem:cosinesplitlemma}, we obtain
\algn{p \sum_x [1-\cos(k\cdot x)]\Pi_p^{(n)} (x) & \leq (n+1) \sum_{i=0}^{n}\sum_x \hspace{-0.1cm}\sum_{\vec t,\vec w, \vec z, \vec u} [1-\cos(k\cdot d_i)] \psi_0(\mathbf{0},w_0,u_0) \notag\\
&\hspace{2cm} \times \psi_n(w_{n-1}, u_{n-1}, t_n, z_n, x) \prod_{j=1}^{n-1} \psi(\vec v_j), \label{eq:db:disp_d_i_split}}
with $d_i$ as introduced above. We now handle these terms for different $i$.
\underline{Case (a): $i\in\{0,n\}$.} Let us start with $i=n$, so that $d_n \in \{x-u_{n-1}, x-w_{n-1}\}$. The summand for $i=n$ in~\eqref{eq:db:disp_d_i_split} is equal to
\algn{& \sum_{w \neq u} \Psi^{(n-1)}(w,u) \sum_{t,z,x \neq u} [1-\cos(k\cdot d')] \psi_n(w,u,t,z,x) \notag\\
\leq & \triangle_p^{\bullet}(\mathbf{0}) \big(T_p\big)^{n-1} \max_{0 \neq u} \sum_{t,z,x \neq u} [1-\cos(k\cdot d)] \psi_n(\mathbf{0},u,t,z,x), \label{eq:db:disp:case_last_segment}}
where $d'\in\{x-w, x-u\}$ and $d\in\{x,x-u\}$. We expand the indicator in $\psi_n$ into two cases. If $t=z=x$, then we can bound the maximum in~\eqref{eq:db:disp:case_last_segment} by $p\sum_x [1-\cos(k\cdot d)] \tau_p^\circ(x) \tau_p(x-u)$, which is bounded by $W_p(k)$ for both values of $d$. If $t,z,x$ are distinct points, then for $d=x$, the maximum in~\eqref{eq:db:disp:case_last_segment} becomes
\[ p^2 \max_{u\neq 0} \sum_{t,z,x} [1-\cos(k\cdot d)] \tau_p^\circ(z) \tau_p^\bullet(t-u) \tau_p(t-z) \tau_p(x-z)\tau_p(x-t)
= p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_general.pdf}}}.\]
Note that in the pictorial representation, we represent the factor $[1-\cos(k \cdot (x-\mathbf{0}))]$ by a line from $\mathbf{0}$ to $x$ carrying a `$\times$' symbol. We use the Cosine-split lemma~\ref{lem:cosinesplitlemma} again to bound
\[ [1-\cos(k\cdot x)] \leq 2 \big([1-\cos(k\cdot z)] + [1-\cos(k\cdot (x-z))]\big), \]
which results in
\al{p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_general.pdf}}}
\ & \leq 2 p^2 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1.pdf}}}
\ + \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split2.pdf}}} \Big] \\
& \leq 2p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_coll.pdf}}}
\ + 2p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_nocoll.pdf}}}
\ + 2p \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split2_bound1.pdf}}}
\Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}} p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split2_bound2.pdf}}}
\Big)\Big) \\
& \leq 2p \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_coll_bound1.pdf}}}
\Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}} p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_coll_bound2.pdf}}}
\Big)\Big)
\ + 2p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_nocoll_subst.pdf}}} + 2 \triangle_p^{\bullet\circ} W_p(k) \\
& \leq 2\triangle_p^{\bullet} W_p(k) + 2p^2 \sum \Big( \Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}} p \sum
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_nocoll_subst_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_iN_split1_nocoll_subst_bound2.pdf}}} \Big) + 2 \triangle_p^{\bullet\circ} W_p(k)\\
& \leq 2(2\triangle_p^{\bullet\circ}+\triangle_p) W_p(k). }
It is not hard to see that a displacement $d=x-u$ yields the same bound.
Similar computations show that the case $i=0$ yields a contribution of at most
\[ \triangle_p^{\bullet\bullet\circ} \big(T_p\big)^{n-1} \triangle_p^{\bullet\circ} W_p(k). \]
\underline{Case (b): $1 \leq i < n$.} We want to apply both the bound~\eqref{eq:db:Psi_triangles_bound} and Lemma~\ref{lem:db:bar_psi_bound}. To this end, we rewrite the $i$-th summand in~\eqref{eq:db:disp_d_i_split} as
\al{ \sum_x \hspace{-0.1cm} & \sum_{\vec t,\vec w, \vec z, \vec u} [1-\cos(k\cdot d_i)] \psi_0(\mathbf{0},w_0,u_0) \psi_n(w_{n-1}, u_{n-1}, t_n, z_n, x) \prod_{j=1}^{n-1} \psi(\vec v_j) \\
& = \sum_x \sum_{a_1,a_2,b_1,b_2} \Big( \Psi^{(i-1)}(a_1,a_2) \bar\Psi^{(n-i-1)}(b_1-x,b_2-x) \\
& \qquad\qquad\qquad \times \sum_{t,w,z,u} \underbrace{\phi(a_2,t,w,z,u,b_2) [1-\cos(k\cdot d_i)] \tau_p^\circ(z-a_1)\tau_p^\bullet(b_1-u)}_{=:\tilde\phi(a_1,a_2,t,w,z,u,b_1, b_2;k,d)} \Big) \\
& \leq \big(\triangle_p^{\bullet}(\mathbf{0})\big) \big(T_p\big)^{i-1} \sum_{b_1',b_2'} \Big( \bar\Psi^{(n-i-1)}(b_1',b_2') \max_{a_1 \neq a_2}
\sum_{t,w,z,u,x} \tilde\phi(a_1,a_2,t,w,z,u,b_1'+x, b_2'+x;k,d_i) \Big) \\
& \leq \big(\triangle_p^{\bullet}(\mathbf{0}) \triangle_p^{\bullet\bullet\circ}\big) \big( T_p\big)^{n-2} \max_{a_1 \neq a_2,b_1\neq b_2} \sum_{t,w,z,u,x} \tilde\phi(a_1,a_2,t,w,z,u,b_1+x, b_2+x;k,d_i) \\
& \leq \big(\triangle_p^{\bullet\bullet\circ}\big)^2 \big( T_p\big)^{n-2} \max_{\mathbf{0} \neq a,\mathbf{0} \neq b} \sum_{t,w,z,u,x} \tilde\phi(\mathbf{0},a,t,w,z,u,b+x,x;k,d_i), }
where we use the substitution $b_j'=x-b_j$ in the second line and the bound $\triangle_p^{\bullet}(\mathbf{0}) \leq \triangle_p^{\bullet\bullet\circ}$ in the last line. It remains to bound the sum over $\tilde\phi$. We first handle the term due to $\phi^{(1)}$, and we call it $\tilde\phi^{(1)}$. Depending on the orientation of the diagram (i.e., the parity of $i$), the displacement $d=d_i$ is either $d=w-a=(w-t)+(t-a)$ or $d=u=(u-z)+z$. We perform the bound for $d=u$ and use the Cosine-split lemma~\ref{lem:cosinesplitlemma} once, so that we now have a displacement on an actual edge. In pictorial bounds, abbreviating $\vec v=(\mathbf{0},a,t,w,z,u,b+x,x;k,u)$, this yields
\begin{align}\sum_{t,w,z,u,x} \tilde\phi^{(1)}(\vec v) &= p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__general.pdf}}}
\ \leq 2p^3 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1.pdf}}}
\ + \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2.pdf}}} \Big] \label{eq:db_dsp:psi1} \\
& = 2p^3 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_coll.pdf}}}
\ + p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_nocoll.pdf}}}
\ + \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound1.pdf}}}
\Big( \sup_{\textcolor{orange}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound2.pdf}}} \Big)\Big) \Big]. \notag \end{align}
The bound in~\eqref{eq:db_dsp:psi1} consists of three summands. The first is
\[ 2p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_coll.pdf}}}
\ \leq 2p \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_coll_bound1.pdf}}}
\Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{turquoise}{\bullet}} p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_coll_bound2.pdf}}}
\Big( \sup_{\textcolor{darkorange}{\bullet}, \textcolor{green}{\bullet}} p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_coll_bound3.pdf}}}
\Big) \Big) \Big) \leq 2 \triangle_p^{\bullet\bullet\circ} \triangle_p W_p(k), \]
the second is
\al{ 2p^4 \sum \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_nocoll.pdf}}}
\ & = 2p^4 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_nocoll_subst.pdf}}}
\ \leq 2p \sum \Big( \Big( \Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_nocoll_subst_bound1.pdf}}} \Big)
\sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}} p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_nocoll_subst_bound2.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split1_nocoll_subst_bound3.pdf}}} \Big) \\
& \leq \triangle_p^{\bullet\circ} \triangle_p W_p(k), }
and the third is
\al{ 2p^3 \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound1.pdf}}}
& \Big( \sup_{\textcolor{orange}{\bullet}, \textcolor{green}{\bullet}} \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound2.pdf}}} \Big)\Big)
\ \leq 2p^2 \triangle_p^{\bullet\circ} \sup_{\textcolor{orange}{\bullet}} \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound2_subst.pdf}}} \\
& \leq 2p^2 \triangle_p^{\bullet\circ} \sup_{\textcolor{orange}{\bullet}} \sum \Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound2_subst_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_1__split2_bound2_subst_bound2.pdf}}} \Big)
\leq 2 \big( \triangle_p^{\bullet\circ} \big)^2 W_p(k). }
The displacement $d=w-a$ satisfies the same bound. In total, the contribution in $\tilde\phi$ due to $\phi^{(1)}$ is at most
\[ 4 \big(\triangle_p^{\bullet\bullet\circ}\big)^3 \big(T_p \big)^{n-2} (\triangle_p^{\bullet\circ} + \triangle_p) W_p(k). \]
Let us now tend to $\tilde\phi^{(2)}$. To this end, we first write $\tilde\phi^{(2)} = \sum_{j=3}^{5} \tilde\phi^{(j)}$, where
\al{\tilde\phi^{(3)}(\mathbf{0},a,t,w,z,u,b+x,x;k,d) &= \tilde\phi^{(2)}(\mathbf{0},a,t,w,z,u,b+x,x;k,d) \mathds 1_{\{|\{t,z,u\}|=3\}}, \\
\tilde\phi^{(4)}(\mathbf{0},a,t,w,z,u,b+x,x;k,d) &= [1-\cos(k\cdot d)] \delta_{z,u} \delta_{t,u} \tau_p^\circ(u) \tau_p(w-u) \tau_p^\bullet(a-w) \tau_p^\bullet(u+x) \tau_p^\circ(b-w-x), \\
\tilde\phi^{(5)}(\mathbf{0},a,t,w,z,u,b+x,x;k,d) &= [1-\cos(k\cdot d)] \delta_{z,u} \delta_{t,u} \delta_{a,w} \tau_p^\circ(u) \tau_p(a-u) \tau_p^\bullet(u+x) \tau_p^\circ(b-a-x). }
Again, we set $\vec v=(\mathbf{0},a,t,w,z,u,b+x,x;k,u)$. Then
\begin{align}\sum_{t,w,z,u,x} \tilde\phi^{(3)}(\vec v) &= p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__general.pdf}}} \notag \\
& \leq p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1.pdf}}}
\ + 2p^3 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split1_subst.pdf}}}
\ + \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2.pdf}}} \Big]. \label{eq:db_dsp:psi3}\end{align}
The first term in~\eqref{eq:db_dsp:psi3} is
\al{p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1.pdf}}} &\leq p^2
\sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1.pdf}}} \Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound2.pdf}}}\Big)\Big)
\leq \triangle_p^{\bullet\bullet\circ} p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1.pdf}}} \\
& \leq 2 \triangle_p^{\bullet\bullet\circ} p^2 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split1_coll.pdf}}}
\ + p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split1_nocoll.pdf}}}
\ + \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split2_bound1.pdf}}} \Big(
\sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split2_bound2.pdf}}}\Big)\Big) \Big] \\
& \leq 2\triangle_p^{\bullet\bullet\circ} \Big[ p^2 \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split1_coll_bound1.pdf}}}
\Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split1_coll_bound2.pdf}}} \Big)\Big) \\
& \hspace{2.2cm} + p^3 \sum \Big(\Big( \sup_{\textcolor{blue}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split1_nocoll_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__coll1_bound1_split1_nocoll_bound2.pdf}}} \Big) \Big] + \triangle_p^{\bullet\circ} W_p(k) \\
& \leq 2\triangle_p^{\bullet\bullet\circ} W_p(k) \Big( \triangle_p^{\bullet\circ} + \triangle_p^{\bullet} + \triangle_p \Big), }
the second term is
\[ 2p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split1_subst.pdf}}}
\ \leq 2p^3 \sum \Big( \Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split1_subst_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split1_subst_bound2.pdf}}} \Big)
\leq 2 \triangle_p^{\bullet\bullet\circ} \triangle_p W_p(k), \]
and the third term is
\begin{align} 2p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2.pdf}}} &=
2p^3 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_coll2.pdf}}}
\ + p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2.pdf}}} \Big] \notag\\
& \leq 2p^3 \Big[ \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_coll2.pdf}}}
\ + \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2_coll3_bound1.pdf}}} \Big(
\sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2_coll3_bound2.pdf}}} \Big)\Big)
+ p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2_nocoll3.pdf}}} \Big] \notag\\
& \leq 2 \triangle_p^{\bullet\bullet\circ}\triangle_p W_p(k) + 2p^3 \Big[ \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_coll2_bound1.pdf}}}
\Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{turquoise}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_coll2_bound2.pdf}}}
\Big( \sup_{\textcolor{darkorange}{\bullet}, \textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_coll2_bound3.pdf}}} \Big)\Big)\Big) \notag\\
& \hspace{5cm} + p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2_nocoll3.pdf}}} \Big] \notag\\
& \leq 2 \triangle_p^{\bullet\bullet\circ} W_p(k) \big( \triangle_p + \triangle_p^{\bullet} \big) + 2p^4 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2_nocoll3.pdf}}}.
\label{eq:db_dsp:psi3_almost_H} \end{align}
We are left to handle the last diagram appearing in the last bound of~\eqref{eq:db_dsp:psi3_almost_H}, which contains one factor $\tau_p^\circ$ and one factor $\tau_p^\bullet$. We distinguish the case where neither collapses (this leads to the diagram $H_p(k)$) and the case where are least one of the factors collapses. Using $\tau_p^\bullet \leq \tau_p^\circ$ and the substitution $y'=y-u$ for $y\in\{w,z,t\}$, we obtain
\al{2p^4 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_3__nocoll1_split2_nocoll2_nocoll3.pdf}}}
& \leq 2 H_p(k) + 4p^4 \sum_{z,t,w,u} \tau_p(t-u) \tau_p(w-t) \tau_p^\circ(u+a_2-w) \\
& \hspace{5cm} \times \tau_{p,k}(z-u) \tau_p(t-z) \tau_p(z) \tau_p(w-a_1) \\
& = 2H_p(k) + 4 p^4 \sum_{t',w'} \Big( \tau_p(t') \tau_p(w'-t') \tau_p^\circ(a_2-w') \sum_z \Big( \tau_{p,k}(z') \tau_p(t'-z') \\
& \hspace{5cm} \times \sum_u \tau_p(z'+u) \tau_p(a_1-w'-u) \Big)\Big) \\
& \leq 2H_p(k) + 4 \triangle_p^{\bullet}(\mathbf{0}) \triangle_p^\circ W_p(k). }
In total, this yields an upper bound on~\eqref{eq:db_dsp:psi3} of the form
\[6 (\triangle_p^{\bullet\bullet\circ})^3 (T_p)^{n-2} \big[(\triangle_p^{\bullet\circ} + \triangle_p + \triangle_p^\circ) W_p(k) + H_p(k) \big]. \]
The same bound is good enough for the displacement $d=w-a$. Turning to $j=4$, we consider the displacement $d=u$ and see that
\al{\sum_{t,w,z,u,x} \tilde\phi^{(4)}(\vec v) &= p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_4__general.pdf}}}
\ = p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_4__subst.pdf}}}
\ \leq p \sum \Big( \Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_4__subst_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_4__subst_bound2.pdf}}} \Big) \\
& \leq \triangle_p^{\bullet\circ} W_p(k), }
which is also satisfied for $d=w-a$. Finally, $j=5$ forces $d=u$, and we have
\[ \sum_{t,w,z,u,x} \tilde\phi^{(5)}(\vec v) = p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_5__general.pdf}}}
\ \leq p \sum \Big( \Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}}
p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_5__bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_i_5__bound2.pdf}}} \Big)
\leq \triangle_p^{\bullet\bullet\circ} W_p(k), \]
and we see that this bound is not good enough for $n=2$. To get a better bound for $n=2$, we bound
\al{ p \sum_{w,u,s,t,z,x} & \tilde\psi_0(\mathbf{0},w,u) \tau_{p,k}(s-w) \tau_p(s-u) \psi_n(u,s,t,z,x) \
\leq \Big( p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N2_left.pdf}}} \Big) \sup_{u \neq s} \sum_{t,z,x} \psi_n(u,s,t,z,x) \\
& \leq \triangle_p^{\bullet}(\mathbf{0}) W_p(k) T_p, }
where we recall that $\tilde\psi_0$ is an upper bound on $\psi_0$ (see Definition~\ref{def:db:psi_phi_fcts}). The above bound is due to the fact that, thanks to~\eqref{eq:db:Psi_last_segment_bound}, the supremum over the sum over $\psi_n$ is bounded by the supremum in~\eqref{eq:db:inductive_step}.
\end{proof}
\begin{proof}[Proof of Proposition~\ref{thm:db:displacement_thm_n1}]
Let $n=1$. Expanding the two cases in the indicator of $\phi_n$ gives
\algn{ p \sum_x [1-\cos(k\cdot x)] \Pi_p^{(1)}(x) &\leq \sum_{w,u,t,z,x} [1-\cos(k\cdot x)] \phi_0(\mathbf{0},w,u,z) \phi_n(u,t,z,x) \notag\\
& \leq p^2 \sum_{w,u,t,z,x} [1-\cos(k\cdot x)] \tilde\phi_0(\mathbf{0},w,u,z) \tau_p^\bullet(t-u) \tau_p(z-t) \tau_p(z-x) \tau_p(t-x) \notag\\
& \quad + p \sum_{w,u,x} [1-\cos(k\cdot x)] \phi_0(\mathbf{0},w,u,z) \tau_p(x-u) \label{eq:db:disp_n1_first_bound}, }
where we used the bound $\phi_0 \leq \tilde\phi_0$ (see Definition~\ref{def:db:psi_phi_fcts}) for the first summand. Since $\phi_0$ is a sum of two terms,~\eqref{eq:db:disp_n1_first_bound} is equal to
\algn{& p^3 \sum_{w,u,t,z,x}[1-\cos(k\cdot x)] \tau_p^\bullet(w) \tau_p(u) \tau_p(u-w) \tau_p^\circ(z-w) \tau_p^\bullet(t-u) \tau_p(z-t) \tau_p(x-z) \tau_p(x-t) \notag\\
+& p^2 \sum_{u,x} J(u) \tau_{p,k}(x) \tau_p(x-u) \notag\\
+& p^2 \sum_{w,u,x} [1-\cos(k\cdot x)] \tau_p^\bullet(w) \gamma_p(u) \tau_p(u-w) \tau_p^\circ(x-w) \tau_p(x-u). \label{eq:db:disp_n1_second_bound}}
We use the Cosine-split lemma~\ref{lem:cosinesplitlemma} on the first term of~\eqref{eq:db:disp_n1_second_bound} to decompose $x=u + (z-u) + (x-z)$, which gives
\al{p^3 \sum_{w,u,t,z,x} & [1-\cos(k\cdot x)] \tau_p^\bullet(w) \tau_p(u) \tau_p(u-w) \tau_p^\circ(z-w) \tau_p^\bullet(t-u) \tau_p(z-t) \tau_p(x-u) \tau_p(x-t) \\
& = p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_general.pdf}}}
\ \leq 3 p^3 \Big[p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split1.pdf}}}
\ + \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2.pdf}}}
\ + \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split3.pdf}}} \Big] \\
& \leq 3p^2 \sum \Big(\Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}} p \sum \Big( \sup_{\textcolor{darkorange}{\bullet}, \textcolor{blue}{\bullet}}
p\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split1_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split1_bound2.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split1_bound3.pdf}}} \Big)
+ 3p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_coll.pdf}}} \\
& \quad + 3p^4 \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_nocoll_bound1.pdf}}}
\Big( \sup_{\textcolor{darkorange}{\bullet}, \textcolor{blue}{\bullet}} \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_nocoll_bound2.pdf}}} \Big) \Big)
\ + 3p \sum \Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split3_bound1.pdf}}}
\Big(\sup_{\textcolor{darkorange}{\bullet}, \textcolor{blue}{\bullet}} p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split3_bound2.pdf}}}
\Big( \sup_{\textcolor{altviolet}{\bullet}, \textcolor{green}{\bullet}} p \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split3_bound3.pdf}}} \Big)\Big)\Big) \\
& \leq 3 W_p(k) \triangle_p^{\bullet\circ} \triangle_p + 3p^2 \sum\Big( \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_coll_bound1.pdf}}}
\Big( \sup_{\textcolor{darkorange}{\bullet}, \textcolor{blue}{\bullet}} p\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_coll_bound2.pdf}}} \Big)\Big) \\
& \quad + 3p^3 \triangle_p^{\bullet}(\mathbf{0}) \sup_{\textcolor{darkorange}{\bullet}} \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_nocoll_bound2_subst.pdf}}}
+ 3 W_p(k) \triangle_p^{\bullet\circ} \triangle_p^{\bullet}(\mathbf{0}) \\
& \leq 3 W_p(k) \triangle_p^{\bullet}(\mathbf{0}) \big( 3\triangle_p^{\bullet\circ} \big) + 3p^3 \triangle_p^{\bullet}(\mathbf{0}) \sup_{\textcolor{darkorange}{\bullet}} \sum \Big(\Big( \sup_{\textcolor{blue}{\bullet},\textcolor{green}{\bullet}}
\sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_nocoll_bound2_subst_bound1.pdf}}} \Big)
\mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1_split2_nocoll_bound2_subst_bound2.pdf}}} \Big) \\
& \leq 3 W_p(k) \triangle_p^{\bullet}(\mathbf{0}) \big( 3\triangle_p^{\bullet\circ} + \triangle_p \big). }
The second term in~\eqref{eq:db:disp_n1_second_bound} is $p^2 (J\ast\tau_{p,k}\ast \tau_p)(\mathbf{0})$. Depicting the factor $\gamma_p$ as a disrupted line $\mathrel{{\includegraphics{gamma_disrupt.pdf}}}$, the third term in~\eqref{eq:db:disp_n1_second_bound} is
\al{p^2 \sum_{w,u,x} & [1-\cos(k\cdot x)] \tau_p^\bullet(w) \gamma_p(u) \tau_p(u-w) \tau_p^\circ(x-w) \tau_p(x-u) = p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1b_general.pdf}}} \\
& \leq 2p^3 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1b_split1.pdf}}}
\ + 2p^2 \sum \mathrel{\raisebox{-0.25 cm}{\includegraphics{Disp_N1b_split2.pdf}}}
\ \leq 2 W_p(k) \Big[ \triangle_p^\circ(\mathbf{0}) + p \big((\delta_{\mathbf{0},\cdot} + p \tau_p)\ast\tau_p\ast \gamma_p\big)(\mathbf{0}) \Big] \\
& \leq 2 W_p(k) \big[ \triangle_p^\circ + p (\tau_p\ast\gamma_p)(\mathbf{0}) + p^2 (\tau_p^{\ast 2}\ast\gamma_p)(\mathbf{0}) \big] \leq 2 W_p(k) \big[ \triangle_p^\circ + 2 \triangle_p \big].}
In the above, we have used that $\gamma_p(x) \leq \tau_p(x)$ as well as $\gamma_p(x) \leq p (J\ast\tau_p)(x) \leq p\tau_p^{\ast 2}(x)$.
\end{proof}
\section{Bootstrap analysis} \label{sec:bootstrap_analysis}
\subsection{Introduction of the bootstrap functions}\label{sec:boot:intro}
This section brings the previous results together to prove Proposition~\ref{thm:convergence_of_LE}, from which Theorem~\ref{thm:main_theorem_triangle_condition} follows with little extra effort. The remaining strategy of proof is standard and described in detail in~\cite{HeyHof17}. In short, it is the following: We introduce the bootstrap function $f$ in~\eqref{eq:boot:intro:f_functions_def}. In Section~\ref{sec:boot:consequences}, and in particular in Proposition~\ref{thm:convergence_of_LE}, we prove several bounds in terms of $f$, including bounds uniform in $p \in [0,p_c)$ under the additional assumption that $f$ is uniformly bounded.
In Section~\ref{sec:boot:bootstrap_argument}, we show that $f(0) \leq 3$ and that $f$ is continuous on $[0,p_c)$. Lastly, we show that on $[0,p_c)$, the bound $f\leq 4$ implies $f\leq 3$. This is called the improvement of the bounds, and it is shown by employing the implications from Section~\ref{sec:boot:consequences}. As a consequence of this, the results from Section~\ref{sec:boot:consequences} indeed hold uniformly in $p\in[0,p_c)$, and we may extend them to $p_c$ by a limiting argument.
Let us recall the notation $\tau_{p,k}(x) = [1-\cos(k\cdot x)] \tau_p(x), J_k(x) = [1-\cos(k\cdot x)] J(x)$. We extend this to $D_k(x) = [1-\cos(k\cdot x)] D(x)$. We note that $\chi(p)$ was defined as $\chi(p) = \mathbb E[|\mathscr {C}(\mathbf{0})|]$ and that $\chi(p) = 1 + p \sum_{x \in \mathbb Z^d} \tau_p(x)$. We define
\[ \lambda_p = 1 - \frac{1}{\chi(p)} = 1 - \frac{1}{1+p \widehat\tau_p(0)} .\]
We define the bootstrap function $f= f_1 \vee f_2 \vee f_3$ with
\eqq{f_1(p) = 2dp, \qquad f_2(p) = \sup_{k \in(-\pi,\pi]^d} \frac{p |\widehat\tau_p(k)| }{ \widehat G_{\lambda_p}(k)}, \qquad f_3(p) = \sup_{k,l \in(-\pi,\pi]^d} \frac{p|\widehat\tau_{p,k}(l)| }{ \widehat U_{\lambda_p}(k,l)}, \label{eq:boot:intro:f_functions_def}}
where $\widehat U_{\lambda_p}$ is defined as
\[\widehat U_{\lambda_p}(k,l) := 3000 [1-\widehat\connf(k)] \Big( \widehat G_\lambda(l-k) \widehat G_\lambda(l) + \widehat G_\lambda(l)\widehat G_\lambda(l+k) + \widehat G_\lambda(l-k)\widehat G_\lambda(l+k) \Big). \]
We note that $\widehat\tau_{p,k}$ relates to $\Delta_k \widehat\tau_p$, the discretized second derivative of $\widehat\tau_p$, as follows:
\[\Delta_k \widehat\tau_p(l) := \widehat\tau_p(l-k) + \widehat\tau_p(l+k) - 2 \widehat\tau_p(l) = -2 \widehat\tau_{p,k}(l) .\]
The following result bounds the discretized second derivative of the random walk Green's function:
\begin{lemma}[Bounds on $\Delta_k$, \cite{Sla06}, Lemma 5.7] \label{lem:bootstrap:Delta_k_Ulam_bound}
Let $a(x) = a(-x)$ for all $x\in\mathbb Z^d$, set $\widehat A(k) = (1-\widehat a(k))^{-1}$, and let $k,l\in(-\pi,\pi]^d$. Then
\al{|\Delta_k \widehat A(l) | \leq \big(\widehat{|a|} (0) - \widehat{|a|}(k) \big) \times & \Big( \big[ \widehat A(l-k) + \widehat A(l+k) \big] \widehat A(l) \\
& \quad + 8 \widehat A(l-k) \widehat A(l+k) \widehat A(l) \big[ \widehat{|a|}(0) - \widehat{|a|}(l) \big]\Big). }
In particular,
\[ |\Delta_k \widehat G_\lambda(l)| \leq [1-\widehat\connf(k)] \Big( \widehat G_\lambda(l)\widehat G_\lambda(l-k) + \widehat G_\lambda(l)\widehat G_\lambda(l+k) + 8 \widehat G_\lambda(l-k)\widehat G_\lambda(l+k) \Big) .\]
\end{lemma}
A natural first guess for $f_3$ might have been $\sup p |\Delta_k\widehat\tau_p(l)| / | \Delta_k \widehat G_{\lambda_p}(l)|$. However, $\Delta_k \widehat G_{\lambda_p}(l)$ may have roots, which makes this guess an inconvenient choice for $f_3$. In contrast, $\widehat U_{\lambda_p}(k,l) >0$ for $k \neq 0$. Hence, the bound in Lemma~\ref{lem:bootstrap:Delta_k_Ulam_bound} supports the idea that $f_3$ is a reasonable definition.
\subsection{\col{Consequences of the diagrammatic bounds}} \label{sec:boot:consequences} The main result of this section, and a crucial result in this paper, is Proposition~\ref{thm:convergence_of_LE}. Proposition~\ref{thm:convergence_of_LE} proves (in high dimension) the convergence of the lace expansion derived in Proposition~\ref{thm:lace_expansion} by giving bounds on the lace-expansion coefficients. Under the additional assumption that $f\leq 4$ on $[0,p_c)$, these bounds are uniform in $p\in [0,p_c)$.
\begin{prop}[Convergence of the lace expansion and Ornstein-Zernike equation] \label{thm:convergence_of_LE} \
\begin{enumerate}
\item \label{firstassertion} Let $n\in\mathbb N_0$ and $p \in [0,p_c)$. Then there is $d_0 \geq 6$ and a constant $c_f=c(f(p))$ (increasing in $f$ and independent of $d$) such that, for all $d > d_0$,
\algn{\sum_{x\in\mathbb Z^d} p | \Pi_{p,n}(x)| \leq c_f/d, & \qquad \sum_{x\in\mathbb Z^d} [1-\cos(k\cdot x)\big] p | \Pi_{p,n}(x)| \leq [1-\widehat\connf(k)] c_f/d, \label{eq:boot:convergence_of_Pi_M} \\
\sup_{x\in\mathbb Z^d} & p \sum_{m=0}^{n} | \Pi_p^{(m)}(x)| \leq c_f, \label{eq:boot:convergence_of_Pi_n_unsummed}}
and
\eqq{ \sum_{x\in\mathbb Z^d} |R_{p,n}(x)| \leq c_f (c_f/d)^n \widehat\tau_p(0). \label{eq:boot:conseq:R_bounds}}
Consequently, $\Pi_p :=\lim_{n\to\infty} \Pi_{p,n}$ is well defined and $\tau_p$ satisfies the Ornstein-Zernike equation (OZE), taking the form
\eqq{ \tau_p(x) = J(x) + \Pi_p(x) + p\big((J+\Pi_p)\ast\tau_p\big)(x). \label{eq:boot:conseq:OZE}}
\item \label{secondassertion}
Let $f \leq 4$ on $[0,p_c)$. Then there is a constant $c$ and $d_0 \geq 6$ such that the bounds~\eqref{eq:boot:convergence_of_Pi_M},~\eqref{eq:boot:convergence_of_Pi_n_unsummed},~\eqref{eq:boot:conseq:R_bounds} hold for all $d > d_0$ with $c_f$ replaced by $c$ for all $p \in [0,p_c)$. Moreover, the OZE~\eqref{eq:boot:conseq:OZE} holds.
\end{enumerate}
\end{prop}
A standard assumption in the lace expansion literature is a bound on $f(p)$ (often $f(p)\le 4$ as in part {\em\ref{secondassertion}}\ of the proposition), and then it is shown that this implies $f(p) \leq 3$. This is part of the so-called bootstrap argument.
We first formulate part~{\em\ref{firstassertion}}~of Proposition~\ref{thm:convergence_of_LE} (of which part~\emph{\ref{secondassertion}} follows straightforwardly) demonstrating that the bootstrap argument is not necessary to obtain convergence of the lace expansion and thus establish the OZE for a \emph{fixed} value $p<p_c$ provided that the dimension is large enough. However, without uniformity in $p$, $d_0$ might depend on $p$ and diverge as $p \nearrow p_c$. Hence, this approach alone does not allow to extend the results to $p_c$.
It is at this point that the bootstrap argument (and thus Section~\ref{sec:boot:bootstrap_argument}) comes into play. In Section~\ref{sec:boot:bootstrap_argument}, we indeed prove that $f \leq 4$ and so the second part of Proposition~\ref{thm:convergence_of_LE} applies. This is instrumental in proving Theorem \ref{thm:main_theorem_triangle_condition}. We get the following corollary:
\begin{corollary}[OZE at $p_c$] \label{cor:boot:OZE_at_critical_point}
There is $d_0$ such that for all $d>d_0$, the limit $\Pi_{p_c} = \lim_{p \nearrow p_c} \Pi_p$ exists and is given by $\Pi_{p_c}= \sum_{n \geq 0} (-1)^n \Pi_{p_c}^{(n)}$, where $\Pi_{p_c}^{(n)}$ is the extension of Definition~\ref{def:le:lace_expansion_coefficients} at $p=p_c$. Consequently, the bounds in Proposition~\ref{thm:convergence_of_LE} and the OZE~\eqref{eq:boot:conseq:OZE} extend to $p_c$.
\end{corollary}
Proposition~\ref{thm:convergence_of_LE} follows without too much effort as a consequence of Lemmas~\ref{lem:bootstrap:bounds_on_trip},~\ref{lem:bootstrap:bounds_on_W},~\ref{lem:bootstrap:bounds_on_Pi0}, and~\ref{lem:bootstrap:bounds_on_H}. \col{Part of the lace expansion's general strategy of proof in the bootstrap analysis is to use the Inverse Fourier Theorem to write
\[ \triangle_p(x) = p^2 \int_{(-\pi,\pi]^d} \text{e}^{-\i k \cdot x } \widehat\tau_p(k)^3 \frac{\, \mathrm{d} k}{(2\pi)^d} \]
and then to use an assumed bound on $f_2$ to replace $\widehat\tau_p$ by $\widehat G_{\lambda_p}$. For site percolation, this poses a problem, since we are missing one factor of $p$. Overcoming this issue poses a novelty of Section~\ref{sec:bootstrap_analysis}. The following two observations turn out to be helpful for this:}
\begin{observation}[Convolutions of $J$]\label{obs:J_convolutions}
Let $m\in\mathbb N$ and $x\in\mathbb Z^d$ with $m \geq |x|$. Then there is a constant $c=c(m,x)$ with $c \leq m!$ such that
\[ J^{\ast m}(x) = c \mathds 1_{\{m-|x| \text{ is even}\}} (2d)^{(m-|x|)/2 }. \]
\end{observation}
\begin{proof}
This is an elementary matter of counting the number of $m$-step walks from $\mathbf{0}$ to $x$. If $m-|x|$ is odd, then there is no way of getting from $\mathbf{0}$ to $x$ in $m$ steps.
So assume that $m-|x|$ is even. To get from $\mathbf{0}$ to $x$, $|x|$ steps must be chosen to reach $x$. Only taking these $|x|$ steps (in any order) would amount to a shortest $\mathbf{0}$-$x$-path. Out of the remaining steps, half can be chosen freely (each producing a factor of $2d$), and the other half must compensate them. In counting the different walks, we have to respect the at most $m!$ unique ways of ordering the steps.
\end{proof}
We remark that this also shows that the maximum is attained for $x=\mathbf{0}$ when $m$ is even and for $x$ being a neighbor of $\mathbf{0}$ when $m$ is odd.
\begin{observation}[Elementary bounds on $\tau_p^{\ast n}$] \label{obs:tau_J_extraction}
Let $n, m\in\mathbb N$. Then there is $c=c(m,n)$ such that, for all $p\in[0,1]$ and $x \in\mathbb Z^d$,
\[ \tau_p^{\ast n}(x) \leq c \sum_{l=0}^{m-n} p^{l} J^{\ast (l+n)}(x) + c \sum_{j=1 \vee (n-m)}^{n} p^{m+j-n}(J^{\ast m} \ast \tau_p^{\ast j})(x), \]
where we use the convention that $\sum_{l=0}^{m-n}$ vanishes for $n>m$.
\end{observation}
\begin{proof}
The observation heavily relies on the bound
\eqq{ \tau_p(x) \leq J(x) + \mathbb E\Big[ \sum_{y \in \omega} \mathds 1_{\{ |y|=1, y \longleftrightarrow x \}} \Big] = J (x) + p (J\ast\tau_p)(x). \label{eq:boot:tau_bound_tau_tilde} }
\col{Note that the left-hand side equals 0 when $x=\mathbf{0}$.} We prove the statement by induction on $m-n$; for the base case, let $m\leq n$. Then we apply~\eqref{eq:boot:tau_bound_tau_tilde} to $m$ of the $n$ convoluted $\tau_p$ terms to obtain
\al{\tau_p^{\ast n}(x) &\leq \big(\tau_p^{\ast(n-m)} \ast \big(J + p(J\ast\tau_p) \big)^{\ast m} \big)(x) \\
& = \sum_{l=0}^{m} \binom{m}{l} p^l \big( J^{\ast m} \ast \tau_p^{\ast (n-m+l)} \big)(x)
= \sum_{l=n-m}^{n} \binom{m}{l+m-n} p^{l+m-n} \big( J^{\ast m} \ast \tau_p^{\ast l} \big)(x). }
Let now $m-n>0$. Applying~\eqref{eq:boot:tau_bound_tau_tilde} once yields a sum of two terms, namely
\eqq{ \tau_p^{\ast n}(x) \leq \big(J \ast \tau_p^{\ast (n-1)}\big)(x) + p \big(J\ast\tau_p^{\ast n}\big)(x). \label{eq:boot:tau_bound_convol_base}}
We can apply the induction hypothesis on the second term with $\tilde m = m-1$ and $\tilde n=n$, producing terms of the sought-after form. Now, observe that application of~\eqref{eq:boot:tau_bound_tau_tilde} yields
\eqq{ \big(J^{\ast j} \ast \tau_p^{\ast (n-j)}\big)(x) \leq \big( J^{\ast(j+1)} \ast \tau_p^{\ast(n-j-1)}\big) (x) + p \big( J^{\ast(j+1)} \ast \tau_p^{\ast(n-j)}\big)(x) \label{eq:boot:tau_bound_convol_step}}
for $1 \leq j <n$. For every $j$, the second term can be bounded by the induction hypothesis for $\tilde m=m-j-1$ and $\tilde n = n-j$ (so that $\tilde m - \tilde n < m-n$) with suitable $c(m,n)$. Hence, we can iteratively break down~\eqref{eq:boot:tau_bound_convol_base}; after applying~\eqref{eq:boot:tau_bound_convol_step} for $j=n-1$, we are left with the term $J^{\ast n}(x)$, finishing the proof.
\end{proof}
We now define
\[V_p^{(m,n)}(a) := (J^{\ast m} \ast \tau_p^{\ast n})(a), \quad W_p^{(m,n)}(a;k) := (\tau_{p,k} \ast V_p^{(m,n)})(a), \quad \widetilde W_p^{(m,n)}(a;k) := (J_k \ast V_p^{(m,n)})(a).\]
Note that $W_p$ from Definition~\ref{def:displacement_quantities} relates to the above definition via $W_p = pW_p^{(0,0)} + pW_p^{(0,1)}$. \col{Moreover, $\triangle_p(x) = p^2 V^{(0,3)}(x) $.}
\begin{lemma}[Bounds on $V_p^{(m,n)}, W_p^{(m,n)}, \widetilde W_p^{(m,n)}$] \label{lem:bootstrap:bounds_on_V}
Let $p \in [0,p_c)$ and $m,n\in\mathbb N_0$ with $d>\tfrac{20}{9} n$. Then there is a constant $c_f=c(m,n,f(p))$ (increasing in $f$) such that the following hold true:
\begin{enumerate}
\item For $m+n \geq 2$,
\[ p^{m+n-1} V_p^{(m,n)}(a) \leq \begin{cases} c_f &\mbox{if } m+n =2 \text{ and } a=\mathbf{0}, \\ c_f/d & \mbox{else}. \end{cases} \]
\item For $m+n \geq 1$, and under the additional assumption $d >2n+4$ for the bound on $W_p^{(m,n)}$,
\[ p^{m+n} \max \Big\{ \sup_{a \in \mathbb Z^d} \widetilde W_p^{(m,n)}(a;k), \sup_{a\in\mathbb Z^d} W_p^{(m,n)}(a;k) \Big\}
\leq [1-\widehat\connf(k)] \times \begin{cases} c_f &\mbox{if } m+n \leq 2, \\ c_f/d & \mbox{if } m+n \geq 3. \end{cases} \]
\end{enumerate}
\end{lemma}
\col{We apply Lemma~\ref{lem:bootstrap:bounds_on_V} for $n \leq 3$, and so $d\geq 7 > 60/9$ for the dimension suffices.}
\begin{proof}
\underline{Bound on $V_p$.}
We start with the case $m\ge4$ where we can rewrite the left-hand side via Fourier transform and apply H\"older's inequality to obtain
\begin{align}p^{m+j-1} (J^{\ast m} \ast\tau_p^{\ast j})(a) &= p^{m+j-1} \int_{(-\pi,\pi]^d} \text{e}^{-\i k \cdot a} \widehat\jeq(k)^m \widehat\tau_p(k)^j \frac{\, \mathrm{d} k}{(2\pi)^d} \notag\\
& \leq \bigg( p^{10(m-1)} \int_{(-\pi,\pi]^d} \widehat\jeq(k)^{10m} \frac{\, \mathrm{d} k}{(2\pi)^d} \bigg)^{1/10} \bigg(\int_{(-\pi,\pi]^d} (p |\widehat\tau_p(k)|)^{10j/9} \frac{\, \mathrm{d} k}{(2\pi)^d} \bigg)^{9/10} \notag\\
& \leq \Big( p^{10(m-1)} J^{\ast 10m}(\mathbf{0}) \Big)^{1/10} \times f_2(p)^j \bigg(\int_{(-\pi,\pi]^d} \widehat G_{\lambda_p}(k)^{10j/9} \frac{\, \mathrm{d} k}{(2\pi)^d} \bigg)^{9/10}.
\label{eq:boot:bounds_on_V_Hoelder} \end{align}
We note that the number $10$ in the exponent holds no special meaning other than that it is large enough to make the following arguments work. The first factor in~\eqref{eq:boot:bounds_on_V_Hoelder} is handled by Observation~\ref{obs:J_convolutions}, as
\al{ \Big( p^{10(m-1)} J^{\ast 10m}(\mathbf{0}) \Big)^{1/10} &\leq \Big( c p^{10(m-1)} (2d)^{5m} \Big)^{1/10} \leq \Big( c (2dp)^{10(m-1)} (2d)^{-5m+10} \Big)^{1/10} \\
&\leq c f_1(p)^{m-1} (2d)^{-m/2+1} \leq c_f /d }
and $m \geq 4$.
Regarding the second factor in~\eqref{eq:boot:bounds_on_V_Hoelder}, note that $10j/9 \leq 10n/9 < d/2 $ and so Proposition~\ref{thm:random_walk_triangle} gives a uniform upper bound.
\col{We remark that the exponent $10$ in~\eqref{eq:boot:bounds_on_V_Hoelder} is convenient because it is even and allows us to apply Lemma~\ref{lem:bootstrap:bounds_on_V} in dimension $d\ge7$.}
If $m<4$, then we first use Observation~\ref{obs:tau_J_extraction} with $\tilde m= 4-m$ to get that $p^{m+n-1}V_p^{(m,n)}$ is bounded by a sum of terms of two types, which are constant multiples of
\eqq{ p^{l+m+n-1} J^{\ast (l+m+n)}(a) = p^{s-1} J^{\ast s}(a) \qquad \text{and} \quad p^{4+j-1} (J^{\ast 4} \ast\tau_p^{\ast j})(a), \label{eq:boot:bounds_on_V_types}}
where $0 \leq l \leq 4-m-1$ (and therefore $s \geq 2$) and $1 \leq j \leq n$.
If $s$ is odd, we can write $s= 2r+1$ for some $r \geq 1$, and Observation~\ref{obs:J_convolutions} gives
\[ p^{2r} J^{\ast (2r+1)}(a) \leq c p^{2r} (2d)^{r} = c (2dp)^{2r} (2d)^{-r} \leq c (f_1(p))^{2r} (2d)^{-r} \leq c_f/d.\]
Similarly, if $s$ is even and $a \neq \mathbf{0}$ or $s \geq 4$, then $p^{s-1} J^{\ast s}(a) \leq c_f/d$. Finally, if $s=2$ and $a=\mathbf{0}$, then $p (J\astJ)(\mathbf{0}) \leq c_f$.
This shows that the terms of the first type in~\eqref{eq:boot:bounds_on_V_types} are of the correct order.
The second type is of the form $p^{4+j-1} V_p^{(4,j)}(a)$ and included in the previous considerations.
Together this proves the claimed bound on $V_p^{(m,n)}$.
\underline{Bound on $\widetilde W_p$.} Let first $m+n \geq 3$. Then
\al{ p^{m+n} \widetilde W_p^{(m,n)}(a;k) &= p^{m+n} \sum_{y\in\mathbb Z^d} J_k(y) V_p^{(m,n)}(a-y) \\
& \leq p^{m+n-1} \Big( \sup_{a\in\mathbb Z^d} V_p^{(m,n)}(a) \Big) (2dp) \sum_{y\in\mathbb Z^d} [1-\cos(k\cdot y)] D(y)\\
& \leq c_f/d \times f_1(p) [1-\widehat\connf(k)], }
applying the bound on $V_p$.
Consider now $m+n=2$. Using first that $J \leq \tau_p$ and then~\eqref{eq:boot:tau_bound_tau_tilde},
\eqq{ p^2 \widetilde W_p^{(m,n)}(a;k) \leq p^2 \widetilde W_p^{(0,2)}(a;k) \leq p^2 \widetilde W_p^{(2,0)}(a;k) + p^3 \widetilde W_p^{(2,1)}(a;k) + p^3 \widetilde W_p^{(1,2)}(a;k).
\label{eq:boot:V_1_bound} }
The second and third summand right-hand side of~\eqref{eq:boot:V_1_bound} can be dealt with as before, we only have to deal with the first summand. Indeed,
\[ p^2 \widetilde W_p^{(2,0)}(a;k) = p^2 \sum_{y} J_k(y) J^{\ast 2}(a-y) \leq 2d p^2 J^{\ast 2}(\mathbf{0}) \sum_{y} D_k(y) = f_1(p)^2 [1-\widehat\connf(k)], \]
and we can choose $c_f = f_1(p)^2$.
Finally, for $m+n=1$, we have $p \widetilde W_p^{(m,n)}(a;k) \leq p \widetilde W_p^{(1,0)}(a;k) + p^2 \widetilde W_p^{(1,1)}(a;k)$. The second term was already bounded, the first is
\[ p(J_k\astJ)(a) \leq p\sum_y J_k(y) = f_1(p) [1-\widehat\connf(k)].\]
\underline{Bound on $W_p$.} We note that a combination of~\eqref{eq:boot:tau_bound_tau_tilde} and the Cosine-split lemma~\ref{lem:cosinesplitlemma} yields
\eqq{ \tau_{p,k}(x) \leq J_k(x) + 2p (J_k\ast\tau_p)(x) + 2p (J\ast\tau_{p,k})(x). \label{eq:boot:taupk_jek_split_bound}}
Applying this repeatedly, we can bound $p^{m+n} W_p^{(m,n)}(a;k)$ by a sum of quantities of the form $ p^{s+t} \widetilde W_p^{(s,t)}$ (where $s+t \geq 1$) plus $c(m,n) p^{m+n} W_p^{(m,n)}$, where we can now assume $m \geq 4$ w.l.o.g. The terms of the form $\widetilde W_p^{(s,t)}$ were bounded above already. Similarly to how we obtained the bound~\eqref{eq:boot:bounds_on_V_Hoelder}, we bound the last term by applying H\"older's inequality, and so
\algn{p^{m+n} W_p^{(m,n)}& (a;k) \leq p^{m+n} \int_{(-\pi,\pi]^d} |\widehat\jeq(l)|^m |\widehat\tau_p(l)|^n |\widehat\tau_{p,k}(l)| \frac{\, \mathrm{d} l}{(2\pi)^d} \notag\\
& \leq \bigg( p^{10(m-1)} \int_{(-\pi,\pi]^d} \widehat\jeq(l)^{10m} \frac{\, \mathrm{d} l}{(2\pi)^d} \bigg)^{1/10} \bigg(\int_{(-\pi,\pi]^d} (p |\widehat\tau_p(l)|)^{10n/9} (p |\widehat\tau_{p,k}(l)|)^{10/9} \frac{\, \mathrm{d} l}{(2\pi)^d} \bigg)^{9/10} \notag\\
& \leq c_f /d \times 3000 f(p)^{n+1} \bigg(\int_{(-\pi,\pi]^d} \widehat G_{\lambda_p}(l)^{10n/9} \Big[ \widehat G_{\lambda_p}(l) \big(\widehat G_{\lambda_p}(l-k) + \widehat G_{\lambda_p}(l+k)\big) \notag\\
& \hspace{5cm} + \widehat G_{\lambda_p}(l-k)\widehat G_{\lambda_p}(l+k)\Big]^{10/9} \frac{\, \mathrm{d} l}{(2\pi)^d} \bigg)^{9/10} \notag\\
& \leq c_f/d, \label{eq:boot:V_2_bound} }
where the last bound is due to Proposition~\ref{thm:random_walk_triangle_related} and the value of $c_f$ has changed in the last line.
\end{proof}
The proofs of the following lemmas, bounding the quantities appearing in Section~\ref{sec:diag_bounds}, are direct consequences of Lemma~\ref{lem:bootstrap:bounds_on_V}.
\begin{lemma}[Bounds on various triangles] \label{lem:bootstrap:bounds_on_trip}
Let $p \in [0,p_c)$ and $d>6$. Then there is $c_f=c(f(p))$ (increasing in $f$) such that
\[ \max\{ \triangle_p,\triangle_p^\circ, \triangle_p^{\bullet}, \triangle_p^{\bullet\circ}\} \leq c_f /d, \qquad \max\{\triangle_p^{\bullet}(\mathbf{0}), \triangle_p^{\bullet\circ}(\mathbf{0}), \triangle_p^{\bullet\bullet\circ}\} \leq c_f. \]
\end{lemma}
\begin{proof}
Note that
\al{ \triangle_p(x) = p^2 V_p^{(0,3)}(x),& \qquad \triangle_p^\circ(x) = p^2V_p^{(0,2)}(x) + \triangle_p(x), \qquad \triangle_p^{\bullet}(x) = p V_p^{(0,2)}(x) + \triangle_p(x), \\
\triangle_p^{\bullet\circ}(x) &= p\tau_p(x) + \triangle_p^{\bullet}(x), \qquad \triangle_p^{\bullet\bullet\circ}(x) = \delta_{\mathbf{0},x} + \triangle_p^{\bullet\circ}(x).}
For the bound on $p \tau_p \leq p$, we use that $p \leq f_1(p)/d$. For all remaining quantities, we use Lemma~\ref{lem:bootstrap:bounds_on_V}, which is applicable since $n \leq 3$ and $\tfrac{20}{9} n \leq \tfrac{60}{9} < 7 \leq d$.
\end{proof}
\begin{lemma}[Bound on $W_p$] \label{lem:bootstrap:bounds_on_W}
Let $p \in [0,p_c)$ and $d>6$. Then there is a constant $c_f=c(f(p))$ (increasing in $f$) such that
\[ W_p(k) \leq [1-\widehat\connf(k)] c_f. \]
\end{lemma}
\begin{proof}
By~\eqref{eq:boot:taupk_jek_split_bound},
\al{W_p(x;k) & = p W_p^{(0,1)}(x;k) + p \tau_{p,k}(x) \\
& \leq p W_p^{(0,1)}(x;k) + 2p^2 \widetilde W_p^{(0,1)}(x;k) + 2p^2 W_p^{(1,0)}(x;k) + p J_k(x).}
The proof follows from Lemma~\ref{lem:bootstrap:bounds_on_V} together with the observation that
\[ pJ_k(x) = (2dp) D_k(x) \leq f_1(p) \sum_{x\in\mathbb Z^d} D_k(x) = f_1(p) [1-\widehat\connf(k)]. \qedhere\]
\end{proof}
\begin{lemma}[Bounds on $\Pi_p^{(0)}$ and $\Pi_p^{(1)}$] \label{lem:bootstrap:bounds_on_Pi0}
Let $p \in [0,p_c), i\in\{0,1\}$, and $d>6$. Then there is a constant $c_f=c(f(p))$ (increasing in $f$) such that
\[ p \sum_x \Pi_p^{(0)}(x) \leq c_f/d, \qquad p \sum_x [1-\cos(k\cdot x)] \Pi_p^{(i)}(x) \leq [1-\widehat\connf(k)] c_f/d. \]
\end{lemma}
\begin{proof}
We recall the two bounds obtained in Proposition~\ref{thm:db:bounds_for_n0}. The first one yields $p|\widehat\Pi_p^{(0)}(k)| \leq p^3 V_p^{(2,2)}(\mathbf{0})$, the second one yields
\[ p \widehat\Pi_p^{(0)}(0) - p\widehat\Pi_p^{(0)}(k) \leq 2p^3 \widetilde W_p^{(1,2)}(\mathbf{0};k) + 2p^3 W_p^{(2,1)}(\mathbf{0};k).\]
All of these bounds are handled directly by Lemma~\ref{lem:bootstrap:bounds_on_V}. Similarly, the only quantity in the bound of Proposition~\ref{thm:db:displacement_thm_n1} that was not bounded already is $p^2 W_p^{(1,1)}(\mathbf{0};k)$. By a combination of~\eqref{eq:boot:tau_bound_tau_tilde} and~\eqref{eq:boot:taupk_jek_split_bound}, we can bound
\al{ p^2 W_p^{(1,1)}(\mathbf{0};k) & \leq p^2 \Big( W_p^{(2,0)}(\mathbf{0};k) + pW_p^{(2,1)}(\mathbf{0};k) \Big) \\
& \leq p^2 \Big( \widetilde W_p^{(2,0)}(\mathbf{0};k) + 2p \widetilde W_p^{(2,1)}(\mathbf{0};k) + 2p W_p^{(3,0)}(\mathbf{0};k) + p W_p^{(2,1)}(\mathbf{0};k) \Big).}
But $0 \leq \widetilde W_p^{(2,0)}(\mathbf{0};k) \leq 2 J^{\star 3}(\mathbf{0}) = 0$ by Observation~\ref{obs:J_convolutions}. The other three terms are bounded by Lemma~\ref{lem:bootstrap:bounds_on_V}.
\end{proof}
\begin{lemma}[Displacement bounds on $H_p$] \label{lem:bootstrap:bounds_on_H}
Let $p \in [0,p_c)$ and $d>6$. Then there is a constant $c_f=c(f(p))$ (increasing in $f$) such that
\[ H_p(k) \leq [1-\widehat\connf(k)] c_f/d. \]
\end{lemma}
\begin{proof} We recall that
\[H_p(b_1, b_2;k) = p^5 \hspace{-0.15cm} \sum_{t,w,z,u,v} \tau_p(z) \tau_p(t-u) \tau_p(t-z) \tau_{p,k}(u-z) \tau_p(t-w) \tau_p(w-b_1) \tau_p(v-w) \tau_p(v+b_2-u).\]
We bound the factor $\tau_p(z) \leq J(z) + p(J\ast\tau_p)(z)$, splitting $H_p$ into a sum of two. The first term is easy to bound. Indeed,
\al{p^5 \sum_{t,w,z} & J(z) \tau_p(t-z) \tau_p(w-t) \tau_p(b_1-w) \sum_u \tau_{p,k}(u-z)\tau_p(t-u) (\tau_p\ast\tau_p)(b_2-u-w) \\
& \leq \triangle_p^{\bullet}(\mathbf{0}) p^4 \sum_{t,w,z,u} J(z) \tau_p(t-z) \tau_p(w-t) \tau_p(b_1-w) (\tau_{p,k}\ast\tau_p)(t-z) \\
& \leq \triangle_p^{\bullet}(\mathbf{0}) W_p(k) p^3 V_p^{(1,3)}(b_1) \leq [1-\widehat\connf(k)] c_f/d }
by the previous Lemmas~\ref{lem:bootstrap:bounds_on_V}-\ref{lem:bootstrap:bounds_on_W}. We can thus focus on bounding
\eqq{ p^6 \hspace{-0.15cm} \sum_{t,w,z,u,v} (J\ast\tau_p)(z) \tau_p(t-u) \tau_p(t-z) \tau_{p,k}(u-z) \tau_p(t-w) \tau_p(w-b_1) \tau_p(v-w) \tau_p(v+b_2-u). \label{eq:boot:H_bounds_conv_z} }
To prove such a bound (and thus the lemma), we need to recycle some ideas from the proof of Lemma~\ref{lem:bootstrap:bounds_on_V} in a more involved fashion. To this end, let
\[ \sigma(x) := p^4(J^{\ast 4}\ast\tau_p)(x) + \sum_{j=1}^{4} p^{j-1} J^{\ast j}(x),\]
and note that $\tau_p(x) \leq \sigma(x)$ by~\eqref{eq:boot:tau_bound_tau_tilde}. Consequently,~\eqref{eq:boot:H_bounds_conv_z} is bounded by $\widetilde H_p(a_1,a_2;k)$, where we define
\[\widetilde H_p(a_1,a_2;k) = p^6 \sum_{t,w,z,u,v} (J\ast\sigma)(z)\sigma(t-u)\sigma(t-z)\tau_{p,k}(u-z)\sigma(t-w)\sigma(w-a_1) \sigma(v-w)\sigma(v+a_2-u). \]
By the Inverse Fourier Theorem, we can write
\al{ \widetilde H_p(a_1,a_2;k) = p^6 \int_{(-\pi,\pi]^{3d}} \hspace{-0.3cm} \text{e}^{-\i a_1 \cdot l_1-\i a_2 \cdot l_2} & \; \widehat\jeq(l_1)\widehat\sigma(l_1)^2 \widehat\sigma(l_2)^2\widehat\tau_{p,k}(l_3)
\widehat\sigma(l_1-l_2) \widehat\sigma(l_1-l_3) \widehat\sigma(l_2-l_3) \frac{\, \mathrm{d}(l_1,l_2,l_3)}{(2\pi)^{3d}}. }
(For details on the above identity, see~\cite[Lemma 5.7]{HeyHofLasMat19} and the corresponding bounds on $H_\lambda$ therein.) We bound
\begin{align}
\widetilde H_p(a_1,a_2;k) &\leq 3000 f_3(p) [1-\widehat\connf(k)] p^5 \int_{(-\pi,\pi]^{3d}} |\widehat\jeq(l_1)|\widehat\sigma(l_1)^2 \widehat\sigma(l_2)^2
|\widehat\sigma(l_1-l_2)| |\widehat\sigma(l_1-l_3)| |\widehat\sigma(l_2-l_3)| \notag \\
& \times\Big( \widehat G_{\lambda_p}(l_3)\widehat G_{\lambda_p}(l_3-k) + \widehat G_{\lambda_p}(l_3)\widehat G_{\lambda_p}(l_3+k) + \widehat G_{\lambda_p}(l_3-k)\widehat G_{\lambda_p}(l_3+k) \Big) \frac{\, \mathrm{d}(l_1,l_2,l_3)}{(2\pi)^{3d}}. \label{eq:boot:H_bounds_applying_f_line1}
\end{align}
Opening the brackets in~\eqref{eq:boot:H_bounds_applying_f_line1} gives rise to three summands. We show how to treat the third one. Applying Cauchy-Schwarz, we obtain
\begin{align}
\bigg( & \int_{(-\pi,\pi]^{3d}} \big[ p^2 |\widehat\jeq(l_1)| |\widehat\sigma(l_1)|^3\big]\big[p^2 \widehat\sigma(l_2-l_1)^2 |\widehat\sigma(l_2)|\big]
\big[p \widehat G_{\lambda_p}(l_3+k)^2 |\widehat\sigma(l_3-l_2)| \big] \frac{\, \mathrm{d}(l_1,l_2,l_3)}{(2\pi)^{3d}} \bigg)^{1/2} \label{eq:boot:H_bounds_CS_1}\\
& \times \bigg( \int_{(-\pi,\pi]^{3d}} \big[p^2 |\widehat\sigma(l_2)|^3 \big] \big[ p^2\widehat\sigma(l_1-l_3)^2 |\widehat\jeq(l_1)| |\widehat\sigma(l_1)| \big]
\big[p \widehat G_{\lambda_p}(l_3-k)^2 |\widehat\sigma(l_3-l_2)| \big] \frac{\, \mathrm{d}(l_1,l_2,l_3)}{(2\pi)^{3d}} \bigg)^{1/2} \label{eq:boot:H_bounds_CS_2}
\end{align}
The square brackets indicate how we want to decompose the integrals. We first bound~\eqref{eq:boot:H_bounds_CS_1}, and we start with the integral over $l_3$. We intend to treat the five summands constituting $\widehat\sigma(l_3-l_2)$ simultaneously. Indeed, note that with our bound on $f_2$,
\eqq{ |\widehat\sigma(l)| \leq \sum_{j=1}^{4} p^{j-1} |\widehat\jeq(l)|^j + p^3\widehat\jeq(l)^4 \widehat G_{\lambda_p}(l) \leq 5 \max_{n\in\{0,1\}, j\in[4]} p^{(j \vee 4n) -1} |\widehat\jeq(l)|^{(j \vee 4n)} \widehat G_{\lambda_p}(l)^n.
\label{eq:boot:gamma_fourier_bound} }
With this,
\al{ p^{(j \vee 4n)} & \int_{(-\pi,\pi]^d} |\widehat\jeq(l_3-l_2)|^{(j \vee 4n)} \widehat G_{\lambda_p}(l_3-l_2)^n \widehat G_{\lambda_p}(l_3+k)^2 \frac{\, \mathrm{d} l_3}{(2\pi)^d} \\
\leq & \Big( p^{10 (j \vee 4n)} \int_{(-\pi,\pi]^d} \widehat\jeq(l_3)^{10(j \vee 4n)} \frac{\, \mathrm{d} l_3}{(2\pi)^d} \Big)^{1/10} \\
& \hspace{1cm} \times \Big( \int_{(-\pi,\pi]^d} \widehat G_{\lambda_p}(l_3+k)^{20/9} \big[\widehat G_{\lambda_p}(l_3-l_2) + \widehat G_{\lambda_p}(l_3-2k + l_2) \big]^{10/9} \frac{\, \mathrm{d} l_3}{(2\pi)^d} \Big)^{9/10} \\
\leq & (c_f/d)^{1/2} }
by the same considerations that were performed in~\eqref{eq:boot:V_2_bound}. We use the same approach to treat the integral over $l_2$ in~\eqref{eq:boot:H_bounds_CS_1}. Applying~\eqref{eq:boot:gamma_fourier_bound} to all three factors of $\widehat\sigma$ gives rise to tuples $(j_i,n_i)$ for $i\in[3]$, and so
\al{ p^{-1+\sum_{i=1}^{3} (j_i \vee 4n_i)} & \int_{(-\pi,\pi]^d} |\widehat\jeq(l_2-l_1)|^{\sum_{i=1}^{2}(j_i\vee 4n_i)} |\widehat\jeq(l_2)|^{j_3 \vee 4n_3} \widehat G_{\lambda_p}(l_2-l_1)^{n_1+n_2} \widehat G_{\lambda_p}(l_2)^{n_3} \frac{\, \mathrm{d} l_2}{(2\pi)^d} \\
\leq & \Big( p^{10(-1+\sum_{i=1}^{3}(j_i \vee 4n_i))} \int_{\space} \widehat\jeq(l_2-l_1)^{10\sum_{i=1}^{2}(j_i\vee 4n_i)} \widehat\jeq(l_2)^{10(j_3 \vee 4n_3)} \frac{\, \mathrm{d} l_2}{(2\pi)^d} \Big)^{1/10} \\
& \hspace{1cm }\times \Big( \int_{(-\pi,\pi]^d} \widehat G_{\lambda_p}(l_2-l_1)^{10(n_1+n_2)/9} \big[\widehat G_{\lambda_p}(l_2) + \widehat G_{\lambda_p}(l_2-2 l_1) \big]^{10n_3/9} \frac{\, \mathrm{d} l_2}{(2\pi)^d} \Big)^{9/10} \\
\leq & c_f \Big( p^{20(-1+\sum_{i=1}^{2}(j_i \vee 4n_i))} \int_{\space} \widehat\jeq(l_2-l_1)^{20\sum_{i=1}^{2}(j_i\vee 4n_i)} \frac{\, \mathrm{d} l_2}{(2\pi)^d} \Big)^{1/20} \\
& \hspace{1cm} \times \Big( p^{20(j_3 \vee 4n_3)} \int_{\space} \widehat\jeq(l_2)^{20(j_3 \vee 4n_3)} \frac{\, \mathrm{d} l_2}{(2\pi)^d} \Big)^{1/20} \\
\leq & (c'_f/d)^{1/2}. }
We finish by proving that the integral over $l_1$ in~\eqref{eq:boot:H_bounds_CS_1} is bounded by a constant. Indeed,
\al{ & \hspace{2cm} p^{-1+\sum_{i=1}^{3} (j_i \vee 4n_i)} \int_{(-\pi,\pi]^d} |\widehat\jeq(l_1)|^{1+\sum_{i=1}^{3}(j_i\vee 4n_i)} \widehat G_{\lambda_p}(l_1)^{n_1+n_2+n_3} \frac{\, \mathrm{d} l_1}{(2\pi)^d} \\
& \leq \Big( p^{10(-1+\sum_{i=1}^{3}(j_i \vee 4n_i))} \int_{\space} \widehat\jeq(l_1)^{10(1+\sum_{i=1}^{3}j_i\vee 4n_i)} \frac{\, \mathrm{d} l_1}{(2\pi)^d} \Big)^{1/10}
\Big( \int_{(-\pi,\pi]^d} \widehat G_{\lambda_p}(l_1)^{10(n_1+n_2+n_3)/9} \frac{\, \mathrm{d} l_1}{(2\pi)^d} \Big)^{9/10} \\
& \leq c_f \cdot p^{-1+\sum_{i=1}^{3}(j_i \vee 4n_i)} \Big( J^{\ast 10(1+\sum_{i=1}^{3}j_i\vee 4n_i)}(\mathbf{0}) \Big)^{1/10} \\
& \leq c'_f \cdot p^{-1+\sum_{i=1}^{3}(j_i \vee 4n_i)} (2d)^{\frac 12 (1+\sum_{i=1}^{3}j_i\vee 4n_i)} \leq c''_f. }
This proves that~\eqref{eq:boot:H_bounds_CS_1} is bounded by $(c_f/d)^{1/2}$. Note that~\eqref{eq:boot:H_bounds_CS_2} is very similar to~\eqref{eq:boot:H_bounds_CS_1}, and the same bounds can be applied to get a bound of $(c_f/d)^{1/2}$. Since the other two terms in~\eqref{eq:boot:H_bounds_applying_f_line1} are handled analogously, we obtain the bound $\widetilde H_p(b_1,b_2;k) \leq [1-\widehat\connf(k)] c_f/d$, which is what was claimed.
\end{proof}
\begin{proof}[Proof of Proposition~\ref{thm:convergence_of_LE}]
Recalling the bounds on $|\Pi_p^{(m)}(k)|$ obtained in Propositions~\ref{thm:db:bounds_for_n0} and~\ref{thm:db:main_thm}, and the bounds on $|\Pi_p^{(m)}(k) - \Pi_p^{(m)}(0)|$ obtained in Propositions~\ref{thm:db:bounds_for_n0},~\ref{thm:db:displacement_thm_n1}, and~\ref{thm:db:displacement_thm}, we can combine them with the bounds just obtained in Lemmas~\ref{lem:bootstrap:bounds_on_trip},~\ref{lem:bootstrap:bounds_on_W},~\ref{lem:bootstrap:bounds_on_Pi0}, and~\ref{lem:bootstrap:bounds_on_H}. This gives
\algn{ p|\widehat\Pi_p^{(m)}(k)| & \leq p \sum_{x\in\mathbb Z^d} \Pi_p^{(m)}(x) \leq c_f (c_f/d)^{m \vee 1}, \label{eq:boot:Pi_m_exp_bound}\\
p|\widehat\Pi_p^{(m)}(k) - \widehat\Pi_p^{(m)}(0)| &= p\sum_{x\in\mathbb Z^d} [1-\cos(k\cdot x)] \Pi_p^{(m)}(x) \leq c_f [1-\widehat\connf(k)] (c_f/d)^{1\vee (m-2)}. \notag }
Summing the above terms over $m$, we recognize the geometric series in their bounds. The series converges for sufficiently large $d$. If $f \leq 4$ on $[0,p_c)$, we can replace $c_f$ by $c=c_4$ in the above, so that the bounds are uniform in $p\in[0,p_c)$, which means that the value of $d$ above which the series converges is independent of $p$. Hence,
\[ p|\widehat\Pi_n(k)| \leq \sum_{m=0}^{\infty} p \Pi_p^{(m)}(x) \leq c_f/d, \qquad p|\widehat\Pi_n(k)-\widehat\Pi_n(0)| \leq [1-\cos(k\cdot x)] c_f/d.\]
The bound~\eqref{eq:boot:convergence_of_Pi_n_unsummed} follows from Corollary~\ref{cor:db:Pi_pn_unsummed_bound} by analogous arguments. Recalling the definition of the remainder term yields the straight-forward bound
\eqq{ \sum_x R_{p,n}(x) \leq \sum_x \sum_u p \Pi_p^{(n)}(u) \tau_p(x-u) \leq p\widehat\Pi_p^{(n)}(0) \widehat\tau_p(0) \leq c_f (c_f/d)^n \widehat\tau_p(0), \label{eq:boot:R_n_bound}}
applying~\eqref{eq:boot:Pi_m_exp_bound}. Hence, if $c_f/d<1$ and $p<p_c$, then $\sum_x R_{p,n}(x) \to 0$ as $n\to\infty$. Again, if $f\leq 4$, we can replace $c_f$ by $c=c_4$ in~\eqref{eq:boot:R_n_bound} and the smallness of $(c/d)$ does not depend on the value of $p$.
The existence of the limit $\Pi_p$ follows by dominated convergence with the bound~\eqref{eq:boot:convergence_of_Pi_n_unsummed}. Together with~\eqref{eq:boot:R_n_bound}, this implies that the lace expansion identity in Proposition~\ref{thm:lace_expansion} converges as $n\to\infty$ and satisfies the OZE.
\end{proof}
Corollary~\ref{cor:boot:OZE_at_critical_point} as well as the main theorem now follow from Proposition~\ref{thm:convergence_of_LE} in conjunction with Proposition~\ref{thm:bootstrap:forbidden_region}, which is proven in Section~\ref{sec:boot:bootstrap_argument} below.
\begin{proof}[Proof of Corollary~\ref{cor:boot:OZE_at_critical_point} and Theorem~\ref{thm:main_theorem_triangle_condition}]
Proposition \ref{thm:bootstrap:forbidden_region} implies that indeed $f(p)\le 3$ for all $p\in[0,p_c)$, and therefore, the consequences in the second part of Proposition \ref{thm:convergence_of_LE} are valid. Lemma~\ref{lem:bootstrap:bounds_on_trip} together with Fatou's lemma and pointwise convergence of $\tau_p(x)$ to $\tau_{p_c}(x)$ then implies the triangle condition.
The remaining arguments are analogous to the proofs of~\cite[Corollary 6.1]{HeyHofLasMat19} and~\cite[Theorem 1.1]{HeyHofLasMat19}.
The rough idea for the proof of Corollary~\ref{cor:boot:OZE_at_critical_point} is to use $\theta(p_c)=0$ (which follows from the triangle condition) to couple the model at $p_c$ with the model at $p<p_c$, and then show that, as $p \nearrow p_c$, the (a.s.) finite cluster of the origin is eventually the same.
For the full argument and the proof of the infra-red bound, we refer to~\cite{HeyHofLasMat19}.
\end{proof}
\subsection{Completing the bootstrap argument} \label{sec:boot:bootstrap_argument}
It remains to prove that $f \leq 4$ on $[0,p_c)$ so that we can apply the second part of Proposition~\ref{thm:convergence_of_LE}. This is achieved by Proposition~\ref{thm:bootstrap:forbidden_region}, where three claims are made: First, $f(0)\leq 4 $, secondly, $f$ is continuous in the subcritical regime, and thirdly, $f$ does not take values in $(3,4]$ on $[0,p_c)$. This implies the desired boundedness of $f$. The following observation will be needed to prove the third part of Proposition~\ref{thm:bootstrap:forbidden_region}:
\begin{observation} \label{obs:Delta_k_bounds}
Suppose $a(x) = a(-x)$ for all $x\in\mathbb Z^d$. Then
\[ \tfrac 12 \big\vert \Delta_k \widehat a(l) \big\vert \leq \widehat{|a|}(0) - \widehat{|a|}(k)\]
for all $k,l\in(-\pi,\pi]^d$ (where $\widehat{|a|}$ denotes the Fourier transform of $|a|$). As a consequence, $|\widehat\connf_k(l)| \leq 1-\widehat\connf(k)$. Moreover, there is $d_0 \geq 6$ a constant $c_f=c(f(p))$ (increasing in $f$) such that, for all $d > d_0$,
\[ \big\vert \Delta_k p\widehat\Pi_p(l) \big\vert \leq [1-\widehat\connf(k)] c_f /d. \]
\end{observation}
\begin{proof}
The statement for general $a$ can be found, for example, in~\cite[(8.2.29)]{HeyHof17}. For convenience, we give the proof. Setting $a_k(x)=[1-\cos(k\cdot x)] a(x)$, we have
\al{\tfrac 12 | \Delta_k \widehat a(l) | &= | \widehat a_k(l)| \leq \sum_{x \in \mathbb Z^d} \big\vert a(x) \cos(l\cdot x) [1-\cos(k\cdot x)] \big\vert
\leq \sum_{x \in\mathbb Z^d} |a(x)| [1-\cos(k \cdot x)] \\
&\leq \widehat{|a|}(0) - \widehat{|a|}(k) }
The consequence about $\widehat\connf$ now follows from $D(x) \geq 0$ for all $x$. Moreover, the statement for $\Delta_k p\widehat\Pi_p(l)$ follows applying the observation to $a=\Pi_p$ together with the bounds in~\eqref{eq:boot:convergence_of_Pi_M}.
\end{proof}
\begin{prop} \label{thm:bootstrap:forbidden_region}
The following are true:
\begin{enumerate}
\item The function $f$ satisfies $f(0) \leq 3$.
\item The function $f$ is continuous on $[0,p_c)$.
\item Let $d$ be sufficiently large; then assuming $f(p) \leq 4$ implies $f(p) \leq 3$ for all $p \in [0,p_c)$.
\end{enumerate}
Consequently, there is some $d_0$ such that $f(p) \leq 3$ uniformly for all $p \in [0,p_c)$ and $d > d_0$.
\end{prop}
As a remark, in the third step of Proposition~\ref{thm:bootstrap:forbidden_region}, we prove the stronger statement $f_i(p) \leq 1+ \textup{const} / d$ for $i \in \{1,2\}$.
In the remainder of the paper, we prove this proposition and thereby finish the proof the main theorem. We prove each of the three assertions separately.
\paragraph{1. Bounds on $f(0)$.} This one is straightforward. As $f_1(p) = 2dp$, we have $f_1(0)=0$.
Further, recall that $\lambda_p =1-1/\chi(p)$, and so $\lambda_0=0$ and $\widehat{G}_{\lambda_0}(k)=\widehat{G}_0(k) = 1$ for all $k\in(-\pi,\pi]^d$. Since both $p|\widehat\tau_p(k)|$ and $p |\widehat \tau_{p,k}(l)|$ equal $0$ at $p=0$, recalling the definitions of $f_2$ and $f_3$ in~\eqref{eq:boot:intro:f_functions_def}, we conclude that $f_2(0) = f_3(0) =0$. In summary, $f(0)=0$.
\paragraph{2. Continuity of $f$.} The continuity of $f_1$ is obvious. For the continuity of $f_2$ and $f_3$, we proceed as in~\cite{HeyHof17}, that is, we prove continuity on $[0,p_c-\varepsilon]$ for every $0<\varepsilon<p_c$. This again is done by taking derivatives and bounding them uniformly in $k$ and in $p \in[0,p_c-\varepsilon]$. To this end, we calculate
\eqq{ \frac{\, \mathrm{d}}{\, \mathrm{d} p} \frac{\widehat\tau_p(k)}{\widehat G_{\lambda_p}(k)} = \frac{1}{\widehat G_{\lambda_p}(k)^2} \Big[\widehat G_{\lambda_p}(k) \frac{\, \mathrm{d}\widehat\tau_p(k)}{\, \mathrm{d} p}
- \widehat\tau_p(k) \frac{\, \mathrm{d} \widehat G_\lambda(k)}{\, \mathrm{d} \lambda} \bigg\vert_{\lambda=\lambda_p} \frac{\, \mathrm{d} \lambda_p}{\, \mathrm{d} p} \Big]. \label{eq:boot:continuity_f2}}
Since $\lambda_p = 1-1/\chi(p)$,
\eqq{ \frac 12 \leq \frac{1}{1-\lambda_p \widehat\connf(k)} = \widehat G_{\lambda_p}(k) \leq \widehat G_{\lambda_p}(0) = \chi(p) \leq \chi(p_c-\varepsilon). \label{eq:boot:continuity_f2_fgmu_bd} }
Further, since $\widehat\tau_p(0)$ is non-decreasing,
\eqq{ \widehat\tau_p(k) \leq \widehat\tau_p(0) \leq \widehat \tau_{p_c-\varepsilon}(0) = \frac{\chi(p_c-\varepsilon)-1}{p_c-\varepsilon}. \label{eq:boot:continuity_f2_fttaup_bd} }
We use Observation~\ref{obs:tools:diff_inequality} to obtain
\[ \left\vert \frac{\, \mathrm{d}}{\, \mathrm{d} p} \widehat\tau_p(k) \right\vert = \bigg\vert \sum_{x\in\mathbb Z^d} \text{e}^{\i k \cdot x} \frac{\, \mathrm{d}}{\, \mathrm{d} p} \tau_p(k) \bigg\vert
\leq \sum_{x\in\mathbb Z^d} \frac{\, \mathrm{d}}{\, \mathrm{d} p} \tau_p(x) = \frac{\, \mathrm{d}}{\, \mathrm{d} p} \sum_{x\in\mathbb Z^d} \tau_p(x) \leq \widehat\tau_p(0)^2, \]
and the same bound as in~\eqref{eq:boot:continuity_f2_fttaup_bd} applies. The interchange of sum and derivative is justified as both sums are absolutely summable.
Note that $\, \mathrm{d} \widehat G_\lambda(k)/\, \mathrm{d} \lambda = \widehat\connf(k) \widehat G_\lambda(k)^2$, and this is bounded by $\chi(p_c-\varepsilon)^2$ for $\lambda=\lambda_p$ by~\eqref{eq:boot:continuity_f2_fgmu_bd}. Finally, by Observation~\ref{obs:tools:diff_inequality},
\[ \frac{\, \mathrm{d} \lambda_p}{\, \mathrm{d} p} = \frac{\frac{\, \mathrm{d}}{\, \mathrm{d} p}\chi(p)}{\chi(p)^2} \leq \widehat\tau_p(0),\]
which is bounded by~\eqref{eq:boot:continuity_f2_fttaup_bd} again. In conclusion, all terms in~\eqref{eq:boot:continuity_f2} are bounded uniformly in $k$ and $p$, which proves the continuity of $f_2$. We can treat $f_3$ in the exact same manner, as it is composed of terms of the same type as the ones we just bounded.
\paragraph{3. The forbidden region $(3,4]$.} Note that we assume $f(p) \leq 4$ in the following, and so the second part of Proposition~\ref{thm:convergence_of_LE} applies with $c=c_4$.
\underline{Improvement of $f_1$.} Recalling the definition of $\lambda_p \in[0,1]$, this implies
\[f_1(p) = \lambda_p - p\widehat\Pi_p(0) \leq 1 + c_4/d.\]
\underline{Improvement of $f_2$.} We introduce $a = p(J + \Pi_p)$, and moreover
\[ \widehat N(k)= \frac{\widehat a(k)}{1+p\widehat\Pi_p(0)}, \qquad \widehat F(k) = \frac{1-\widehat a(k)}{1+p\widehat\Pi_p(0)}.\]
By adapting the analogous argument from~\cite[proof of Theorem 1.1]{HeyHofLasMat19}, we can show that $1-\widehat a(k)>0$. Therefore, under the assumption that $f(p) \leq 4$, we have $p\widehat\tau_p(k) = \widehat N(k)/\widehat F(k) = \widehat a(k) /(1-\widehat a(k))$, and furthermore $\lambda_p= \widehat a(0)$. In the following lines, $M$ and $M'$ denote constants (typically multiples of $c_4$) whose value may change from line to line. An important observation is that
\[ \frac{1}{1+p\widehat\Pi_p(0)} \leq 1 + M/d, \qquad |\widehat a(k)| \leq 1 + M/d, \qquad |p\widehat\Pi_p(k)| \leq M/d. \]
We are now ready to treat $f_2$. Since $|\widehat N(k)| \leq 1+M/d$ and by the triangle inequality,
\algn{ \bigg\vert\frac{p\widehat\tau_p(k)}{\widehat G_{\lambda_p}(k)} \bigg\vert &= \Big\vert \widehat N(k) + p \widehat\tau_p(k) [1- \lambda_p \widehat\connf(k) - \widehat F(k)] \Big\vert \notag \\
&\leq 1+M/d + \big\vert p \widehat\tau_p(k) \big\vert \Big\vert 1- \lambda_p \widehat\connf(k) - \widehat F(k) \Big\vert. \label{eq:boot:FR:f2_main_identity}}
Also,
\algn{\big\vert 1- \lambda_p \widehat\connf(k) - \widehat F(k) \big\vert &=
\bigg\vert\frac{1+p\widehat\Pi_p(0)-\big(2dp + p\widehat\Pi_p(0)\big)\big(1+p\widehat\Pi_p(0)\big)\widehat\connf(k) - 1 + 2dp \widehat\connf(k) + p\widehat\Pi_p(k)}{1+p\widehat\Pi_p(0)} \bigg\vert \notag\\
&= \bigg\vert\frac{[1-\widehat\connf(k)] p\widehat\Pi_p(0)}{1+p\widehat\Pi_p(0)} \bigg\vert + \bigg\vert\frac{\widehat a(0) p\widehat\Pi_p(0) \widehat\connf(k) + p\widehat\Pi_p(k)}{1+p\widehat\Pi_p(0)} \bigg\vert. \label{eq:boot:FR:f2_decomp_step}}
The first term in the right-hand side of~\eqref{eq:boot:FR:f2_decomp_step} is bounded by $[1-\widehat\connf(k)] M/d$. In the second term, recalling that $\widehat a(0)=\lambda_p$, we add and subtract $p\widehat\Pi_p(0)$ in the numerator, and so
\al{ \big\vert 1- \lambda_p \widehat\connf(k) - \widehat F(k) \big\vert &\leq [1-\widehat\connf(k)] M/d + \bigg\vert \frac{[1-\lambda_p \widehat\connf(k)] p\widehat\Pi_p(0) + p\big(\widehat\Pi_p(k)-\widehat\Pi_p(0)\big)}{1+p\widehat\Pi_p(0)} \bigg\vert \\
& \leq [1-\widehat\connf(k)] M'/d + [1-\lambda_p \widehat\connf(k)] M'/d \leq 3 [1-\lambda_p \widehat\connf(k)] M'/d }
for some constant $M'$. In the second to last bound, we have used that $p\widehat\Pi_p(0) - p\widehat\Pi_p(k) \leq [1-\widehat\connf(k)] M/d$. For the last, we have used that $1-\widehat\connf(k) \leq 2(1-\lambda_p\widehat\connf(k)) = 2 \widehat G_{\lambda_p}(k)^{-1}$. Putting this back into~\eqref{eq:boot:FR:f2_main_identity}, we obtain
\[ \left\vert \frac{p \widehat\tau_p(k)}{\widehat G_{\lambda_p}(k)} \right\vert \leq 1+ M/d + 3|p \widehat\tau_p(k) / \widehat G_{\lambda_p}(k)| M'/d \leq 1+4(M \vee M') /d.\]
This concludes the improvement on $f_2$. Before dealing with $f_3$, we make an important observation:
\begin{observation}\label{obs:bootstrap:FR:f4_bound}
Given the improved bounds on $f_1$ and $f_2$,
\[ \sup_{k\in(-\pi,\pi]^d} \left\vert \frac{1-\lambda_p \widehat\connf(k)}{1-\widehat a(k)} \right\vert \leq 3. \]
\end{observation}
\begin{proof}
Consider first those $p$ such that $2dp \leq 3/7$. Then
\[ \left\vert \frac{1-\lambda_p \widehat\connf(k)}{1-\widehat a(k)} \right\vert = \left\vert \frac{1- 2dp \widehat\connf(k) - p \widehat\Pi_p(0) \widehat\connf(k)}{1- 2dp \widehat\connf(k) - p\widehat\Pi_p(k)} \right\vert
\leq \frac{\frac{10}{7} + M/d}{\frac 47 + M/d} \leq (1+ M' /d) \tfrac 52 \leq 3 \]
for $d$ sufficiently large. Next, consider those $k\in(-\pi,\pi]^d$ such that $|\widehat\connf(k)| \leq 7/8$. Then
\[ \left\vert \frac{1-\lambda_p \widehat\connf(k)}{1-\widehat a(k)} \right\vert = \left\vert 1 - \frac{p\widehat\Pi_p(0) \widehat\connf(k) - p\widehat\Pi_p(k)}{1-\widehat a(k)} \right\vert
\leq 1 + \frac{2M /d}{1-(1+M/d) \frac 78 - M/d} \leq 1 + 16 M' /d \]
for $d$ sufficiently large. Let now $p$ such that $2dp> 3/7$ and $k$ such that $|\widehat\connf(k)| > 7/8$. We write $1-\lambda_p \widehat\connf(k) = \widehat G_{\lambda_p}(k)^{-1}$ and $1-\widehat a(k) = \widehat\tau_p(k)/\widehat a(k)$. Since $2dp|\widehat\connf(k)| \geq 3/8$, we obtain
\al{ \left\vert \frac{1-\lambda_p \widehat\connf(k)}{1-\widehat a(k)} \right\vert & \leq \frac 83 \left\vert \frac{\widehat\tau_p(k)}{\widehat G_{\lambda_p}(k)} \cdot \frac{2dp \widehat\connf(k)}{\widehat a(k)} \right\vert
\leq \frac 83 \big(1+M /d\big) \left\vert 1 - \frac{p\widehat\Pi_p(k)}{\widehat a(k)} \right\vert \\
& \leq \frac 83 \big(1+M /d\big) \Big(1+ \frac{M/d}{\frac 83 - M /d} \Big) \leq 3 }
for $d$ sufficiently large.
\end{proof}
\underline{Improvement of $f_3$.} Elementary calculations give
\[ \Delta_k p \widehat\tau_p(l) = \underbrace{\frac{\Delta_k \widehat a(l)}{1-\widehat a(l)}}_{\text{(I)}}
+ \sum_{\sigma\in \pm 1} \underbrace{\frac{\big(\widehat a(l+\sigma k) - \widehat a(l) \big)^2}{(1-\widehat a(l))(1-\widehat a(l+\sigma k))}}_{\text{(II)}}
+ \underbrace{\widehat a(l) \Delta_k \Big( \frac{1}{1-\widehat a(l)} \Big)}_{\text{(III)}}. \]
We bound each of the three terms (I)-(III) separately. For the first term,
\al{|\text{(I)}| & = \big\vert \Delta_k \widehat a(l) \big\vert \cdot \left\vert \frac{1-\lambda_p \widehat\connf(l)}{1-\widehat a(l)} \right\vert \cdot \widehat G_{\lambda_p}(l)
\leq 3 \widehat G_{\lambda_p}(l) \Big\vert 2dp \Delta_k \widehat\connf(l) + \Delta_k p \widehat\Pi_p(l) \Big\vert \\
& \leq (3 + M/d) [1-\widehat\connf(k)] \widehat G_{\lambda_p}(l) \widehat G_{\lambda_p}(l+k). }
In the above, we have first used Observation~\ref{obs:bootstrap:FR:f4_bound}, then Observation~\ref{obs:Delta_k_bounds}, and finally the fact that $\widehat G_{\lambda_p}(l+k) \geq 1/2$. Note that if we obtained similar bounds on (II) and (III), we could prove a bound of the form $|\Delta_k \widehat\tau_p(l)| \leq c \widehat U_{\lambda_p}(k,l)$ for the right constant $c$ and the improvement of $f_3$ would be complete.
To deal with (II), we need a bound on $\widehat a(l+\sigma k)-\widehat a(l)$ for $\sigma\in\{\pm 1\}$. As in~\cite[(8.4.19)-(8.4.21)]{HeyHof17},
\al{|\widehat\connf(l\pm k) - \widehat\connf(l)| & \leq \sum_x \Big( | \sin(k \cdot x)| D(x) + [1-\cos(k \cdot x)] D(x) \Big) = 1-\widehat\connf(k) + \sum_x | \sin(k \cdot x)| D(x) \\
& \leq 1-\widehat\connf(k) + \Big( \sum_x D(x) \Big)^{1/2} \Big( \sum \sin(k \cdot x)^2 D(x) \Big)^{1/2} \\
& \leq 1-\widehat\connf(k) + 2 \Big( \sum [1-\cos(k\cdot x)] D(x) \Big)^{1/2} \\
& \leq 4 [1-\widehat\connf(k)]^{1/2},}
and similarly
\al{p|\widehat\Pi_p(l \pm k) - \widehat\Pi_p(l)| & \leq \Big(p \sum_x |\Pi_p(x)| \Big)^{1/2} \Big( 2p \sum_x [1-\cos(k\cdot x)] |\Pi_p(x)| \Big)^{1/2} + p\sum_x [1-\cos(k\cdot x)] |\Pi_p(x)| \\
& \leq M [1-\widehat\connf(k)]^{1/2} /d.}
Putting this together yields
\eqq{ |\widehat a(l \pm k)-\widehat a(l)| \leq 2dp |\widehat\connf(l\pm k) - \widehat\connf(l)| + p|\widehat\Pi_p(l \pm k) - \widehat\Pi_p(l)| \leq (16+M/d) [1-\widehat\connf(k)]^{1/2}. \label{eq:boot:FR:hat_a_diff_bound}}
Combining~\eqref{eq:boot:FR:hat_a_diff_bound} with Observation~\ref{obs:bootstrap:FR:f4_bound} yields
\al{\text{(II)} &\leq (16+M/d)^2 [1-\widehat\connf(k)] \left\vert \frac{1-\lambda_p \widehat\connf(l)}{1-\widehat a(l)} \right\vert
\left\vert \frac{1-\lambda_p \widehat\connf(l+\sigma k)}{1-\widehat a(l+\sigma k)} \right\vert \widehat G_{\lambda_p}(l) \widehat G_{\lambda_p}(l+\sigma k) \\
& \leq (144 + M' /d) [1-\widehat\connf(k)] \widehat G_{\lambda_p}(l) \widehat G_{\lambda_p}(l+\sigma k). }
To bound (III), we want to use Lemma~\ref{lem:bootstrap:Delta_k_Ulam_bound}. We first provide bounds for the three types of quantities arising in the use of the lemma. First, note that $|\widehat a(l)| \leq 4+M/d$. Next, we observe
\[ \widehat{|a|} (0) - \widehat{|a|}(k) = \sum_x [1-\cos(k\cdot x)] \big\vert 2dp D(x) + p\Pi_p(x) \big\vert \leq (4+M/d) [1-\widehat\connf(k)].\]
The third ingredient we need is Observation~\ref{obs:bootstrap:FR:f4_bound}, which produces $|1-\widehat a(l)|^{-1} \leq 3 \widehat G_{\lambda_p}(l)$. Putting all this together and applying Lemma~\ref{lem:bootstrap:Delta_k_Ulam_bound} gives
\al{\Delta_k \Big( \frac{1}{1-\widehat a(l)} \Big) & \leq (4+M/d) [1-\widehat\connf(k)] \bigg( 9 \big[ \widehat G_{\lambda_p}(l-k) + \widehat G_{\lambda_p}(l+k) \big] \widehat G_{\lambda_p}(l) \\
& \hspace{3cm} + 216 (4+M/d) \widehat G_{\lambda_p}(l-k)\widehat G_{\lambda_p}(l+k) \widehat G_{\lambda_p}(l) [1-\widehat\connf(l)] \bigg)\\
& \leq (6912 + M'/d) [1-\widehat\connf(k)] \Big( \widehat G_{\lambda_p}(l-k) + \widehat G_{\lambda_p}(l+k) \big] \widehat G_{\lambda_p}(l) + \widehat G_{\lambda_p}(l-k)\widehat G_{\lambda_p}(l+k) \Big), }
noting that $\widehat G_{\lambda_p}(l) [1-\widehat\connf(l)] \leq 2$. In summary, $\text{(I)}+\text{(II)}+\text{(III)} \leq 3 \widehat U_{\lambda_p}(k,l),$ which finishes the improvement on $f_3$.
This finishes the proof of Proposition~\ref{thm:bootstrap:forbidden_region}, and therewith also the proof of Theorem~\ref{thm:main_theorem_triangle_condition}.
\paragraph{Acknowledgment.} We thank G\"unter Last, Remco van der Hofstad, and Andrea Schmidbauer for numerous discussions about the lace expansion, which are reflected in the presented paper.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 7,752
|
Posts Tagged Russia
Ukraine, Russia and the Easter of Insanity by Mahboob A. Khawaja, PhD.
Posted by admin in RUSSIA, Ukraine on May 28th, 2022
Ukraine, Russia and the Easter of Insanity
by Mahboob A. Khawaja, PhD.
Tyranny did not Pause even on the Easter
Decades earlier recall what the US-Britain and NATO did in Iraq even on Xmas Eve. Layla Anwar ("Come and see our overflowing morgues…..come and see the rubble of your surgical strikes": An Arab Women Blues wrote in her website blog:
"Everyday, under the pretext of either al-Qaida, insurgents, militants or whatever imaginary name you coined, you have not ceased, not even for one day, slaughtering our innocents……for 4 years, you have not ceased for one single day, not during holiday periods, not during religious celebrations, not even during the day your so called God was born….if you have a God that is."
We, the People witnessing a planned scheme of things to exterminate humanity with superior weapons hitting targeted civilian towns and daily bloodbath as a new normal. Have we not learned any lessons from the European nationalistic wars of First World War and the 2nd World War? We, the Humanity continue to argue if wars have become part of new normal – unknown and unthinkable to human civilizations and rational thinking. We, the People are entrapped and entrenched by dirty politics of the few. It needs a formidable challenge to rethink about the continuity of the current tragedy in Ukraine.
We, the People, We, the Humanity claim to be a "moral being", yet our actions defy the logic of our claims as moral beings. Easter had moral, spiritual and humanitarian significance to be at peace and united to worship God, the Creator of All (known and unknown), and Lord of the worlds. The Orthodox Christians and other Christian denominations populating Russia, Ukraine, Europe, Asia and Africa would visualize as an inner world of soul and body to celebrate the Easter, but bombing, shelling and missiles continued to victimize the innocent civilians and anarchy of violence and barbarism followed by human beings against other humans across Ukraine. The consequential killings and destruction in Kyiv, Khariv and Mariupol appear far worse than the imagination of despotism.
Loss of Morality and War against All
Canons of rationality should unfold rationality of leadership in Russia, Ukraine, America, NATO and all over the globe for immediate concerns about lost sense of moral humanity. What Russian 'Military Operation' is doing in Ukraine, American, British and other Europeans have practiced the Hobbesian formula: "war of all against all." The conflict in Ukraine calls for an immediate ceasefire and rethinking and how the great cause of humanity for peace and harmony could end up in catastrophic challenges for a sustainable future. Did President Putin not know that war kill people and destroy the habitats? Surely, he knew first-hand how George Bush, Tony Blair and Obama had inflicted insanity of war on the people of Afghanistan, Iraq and Syria – and millions and millions were brutally massacred, displaced and children made orphans for generations to come. Why should intelligent and responsible leaders ignore the lessons of history? If Bush, Blair and Obama were wrong and misfit for responsible leadership, Putin does not have to be immoral, tyrannical and ruthless to wage war against his own people – former entity of the USSR – Ukraine- people have enjoin common ethnicity, culture, religion and lot more as peace loving people of this planet.
The wars are declared by the few and not the majority masses. The small ruling elite who plans and wages war is often afraid of citizenry reaction and refusal to accept the rationality of a war. Throughout history, European nationalism institutionalized the doctrine of war as a necessary means to promote national interest and racial superiority over "the other". Most proponents of wars have used "fear" as one of the major instruments of propaganda and manipulation to perpetuate allegiance from the ordinary folks to the elite warmongers in a crisis situation.
President Putin in Search of a Navigational Change
Effective leaders must know their strengths and weaknesses and assess the competence of their immediate advisors and people around them to determine whether they make rational decisions/actions or negating the logic of humanity by crossing over the limits of insanity. Russian culture has immense relationship to ethnicity, language and culture of Ukraine. Why would an intelligent leader bent on destroying his own people and culture for some transitory military gains?
In an age of enlightenment of the 21st century, could we imagine the growth of human civilizations out of moral mire, military conquest, hypersonic bombing and imperial domination? What went wrong with human thinking that has brought us to this onslaught of continued insanity, shelling and killings even on Easter? We imagine that Orthodox Christians hold Easter as vital and spiritually important as other Christians across the world. But there was no critical challenge emerging to the leadership in Moscow from the Russian religious establishment to halt the bombing or pause for the civilians to be evacuated across the charcoaled buildings in Khariv and Mariupol.
If the Orthodox Church (Russia) claimed moral values, then surely there seems to be a contrast between politics and human morality. Undoubtedly, killings and bombing of civilian habitats cannot be related to ideas and ideals of Easter festivity and splendor. Violence against fellow human beings are the outgrowth of madness, cruelty and sheer indifference to human basic values and perhaps Russia would not like to be part of malignity and tyranny. It would be logical to understand that President Putin would realize – what went wrong with his assumptions of immediate victory over helpless Ukraine and to stop the dreadful and unavoidable calamity of war and should rush to make a navigational change for the good of the people of Russia and Ukraine and the larger global humanity. The war on Ukraine has unleashed unthinkable socio-economic, political and humanitarian calamities on the rest of mankind. One would imagine that President Putin would listen to voices of REASON for a navigational change and would not bankrupt Russia as did Bush and Obama: This author noted the same in: "How the United States and Britain Lost the Bogus Wars in Iraq and Afghanistan." Global Research: 10/22/2010.
It is undeniable that the US is "bankrupt" because of the on-going wars in Iraq and Afghanistan. David M. Walker Comptroller General of the US and Head of the Government Accountability Office (December 2007). reported that "In everyday language, the US Government cannot pass an audit.
Time and Reason call for an immediate Ceasefire and Peace-Making between Russia and Ukraine
We, the People, We, the Humanity must rise against the evil to exploit, tyrannize and demoralize the people across the globe. Those echoing voices of "Crimes against Humanity" in Khariv and Bucha and Mariupol must know that it is a slippery trail of myth and one cannot imagine expecting ICC ( The Hague) to prosecute any superpowers. There were many formal complaints against America, Britain and NATO but none were translated into actions against these powerful monsters for crimes committed against the civilian population in Afghanistan, Iraq, Yemen and Syria.
To stop the war in Ukraine, the US, the EU, NATO should have a direct face to face communication with President Putin. It could be facilitated under the G20 auspices. It is logical when people of diversity and opposing ideals come to talk directly, tensions and evil mongering is reduced to reason and mutual interest. This has not happened except military options for weapon supplies and enlargement of the scope of regional conflict. Time and history will not forgive nor forget any of the leaders if they failed to agree to an immediate ceasefire and peace deal.
Dr. Mahboob A. Khawaja specializes in international affairs-global security, peace and conflict resolution with keen interests in Islamic-Western comparative cultures and civilizations, and author of several publications including the latest: One Humanity and the Remaking of Global Peace, Security and Conflict Resolution. Lambert Academic Publications, Germany, 12/2019.
Russia, Russia Putin Warning, Tragedy of Ukraine, Ukraine
Ukraine, Russia and the West to Rethink of Dialogue for Peace, Not War Mahboob A. Khawaja, PhD.
Posted by admin in INTERNATIONAL, Russia & Ukraine on March 6th, 2022
Ukraine, Russia and the West to Rethink of Dialogue for Peace, Not War
Mahboob A. Khawaja, PhD.
Rational Leaders Think of People, not Egoism
In an age of Reason, the cause of humanity calls for peace, freedom and respect for human dignity. In a hybrid geo-political culture – part human and part vulture, we are witnessing a forgotten wisdom of human courage and foresight to stand for the protection of human rights, state sovereignty and national freedom as we continue to see a catastrophic evolving war in Ukraine. The man-made emerging war against the innocent people has no place in an inexplicable antiquity and immense scale of irrational purpose for individualistic glory – all appear to be covered in mystery and unfolding irrational policy behavior in global affairs. To ensure and safeguard the vital geo-political interests of global mankind, this unwarranted war must stop on all sides. Leaders on both sides need to acquire an enlightened understanding that in crisis management, effective leaders do not rush to hasty reactionary judgments and naïve conclusions of belligerency and warmongering.
A reflective snap shot of a critical moment in time pulse and movement of history reveals absence of reasoned dialogue for the prevention of the on –going conflict causing massive human casualties, displacement of millions of Ukrainian civilian refugees across Europe, destruction of essential civic infrastructures and furious competitive edge for success in an ugly war between Ukraine and Russia. The latest being an attack on the European biggest nuclear establishment in Ukraine. Its consequences could lead to catastrophic disruption and wipe out an entire population and progress of human civilization simply to satisfy the few sadistic political minds. To glance ahead of the emerging horrifying events coming out of Ukraine, We, The People of global humanity must call for immediate halt to all war machines and lingering suspicion of one-sided peace and triumph. The sudden and inexplicable plunge into insanity of war will not produce any military or psychological gains for violent assumptions plagues with hatred, bias and false claims and counter claims of military superiority, defeats and occupation.
Reference: Image Courtesy NY Post
There are wild and inhuman sciences of Ukrainian people's suffering and forced displacement to neighboring Poland, Hungary, Moldova and other locations. WE, the People of the globe enjoin immense bonds of understanding for equal rights, friendship and freedom and respect for national integrity of all people and states within the world systems of political affairs. Regardless of geography and history, we are One Humanity and pain inflicted on any parts of the human body gives pain and anguish to the whole body. We, the People cannot be detached from what is happening between Russia, Ukraine NATO and Europe. We must emphasize and enhance our moral and spiritual bonds to remain contacted in solidarity of the suffering masses and refocus on peace and ending the conflict through reason and dialogue. There is no reason to opt for ruthless purging of a democratically elected leader or political governance in Ukraine or Russia or elsewhere in Europe. We, the People of global consciousness must reject politically indoctrinated cynicism becoming an endemic to change the political governance in Ukraine by violent actions of military actions and unwanted war. If Ukraine and Russian political elite could face each at a table and pursue the urgent need for a reasoned dialogue to cure the sickness of military triumph, it will be a welcomed evolutionary development to foresee the end of the current conflict.
America, Russia and NATO Failed to Learn from History
Human progress and future-making are jeopardized when lessons of history are deliberately misinterpreted and ignored by the so called intelligent people. If war is the only avenue to seek peace, we are on the wrong side of history and thinking of our future. There are fearsome and ferocious flows of blame game as often used in conflict-making and conflict -keeping situations. NATO and America are implying extensive financial and other sanctions against Russia to halt the on-going war in Ukraine. The action and reaction gamble will result in unthinkable economic, social and political consequences for Russia and America and the EU and others on this planet. Most rational analysts would agree that sanctions would not deter Russia from its stated aims and priorities of the current conflict. Russia appears to be as competent and viable militarily and politically as is America or the EU and NATO in their pursuit of political strategies and goals. The conflict envisages mutual suspicion and distrust in official policies and behavior across the board. To a perceptive eye, this is the real reasoning for the emerging war in Ukraine. It reflects a treacherous mindset to imagine that any one country is powerful enough to dominate another sovereign entity.
Is NATO being managed by people who lived in the distant past and perhaps post WW2 historic culture is still alive and flourishing? Is there any glimpse of hope for change and new reasoned relationships between America, Russia and West Europeans? The future of violence and nationalistic resentment looks embedded into the distorted strategic necessities of the current affairs, be it the argument of Russia or American-led NATO and or the EU on its own. NATO is run by the wrong people, glued to wrong thinking and doing the wrong things without any rational sense of time, people's interest and history. The waking consciousness and wisdom would demand that Russia, America and NATO and the EU – all active actors in this perplexed game should critically analyze their strength and weaknesses, their aims for freedom and respect for state's sovereignty, should know who they are and where they are and that global humanity is watching them if they will act wisely for peace or demonstrate abhorrent and ruthless behavior in ensuring a peaceful resolution of the current naive aims of disastrous consequences for all the living mankind.
Global news media represents the pains and sufferings of Ukrainian masses forcibly evicted from their homes and towns moving to take asylum across Eastern Europe. There could be few millions on the move without knowing any safe place or destination or hope for safety. It is self-evident unambiguous experience of human aggressive connotation by the few against the many and all – as it happened during the 2nd World War. Should America, Russia and the EU not learn lesson from the formative and living history? The unreal embodiment of good intentions and piety showed by all in this war, reminds us all that we are not inhabitants of a rational and 21st century knowledge-based civilized world. Perhaps, we are living in a dangerous world of our own making – not differentiating between evil and virtue in human thoughts and behavior. It is fast becoming clear that NATO and the EU failed to grasp the reality of the current conflict and appears unable to help Ukraine for its rights to a be free member of global political systems of states.
Russia and its leaders must realize the humane urgency to agree for an immediate ceasefire and try to resolve the issues through peaceful dialogue and certainly not by conquest of Ukraine. Any violations of mutual relationship will have volcanic consequences to dehumanize the civic Russian principles and values. When noble ideals are misinterpreted, it could drain out all the good qualities of people and leaders. Political reasoning is a shared enterprise within the global systems of governance to which America, Russia and NATO failed to observe. The sanctity of peace and values of human life will encompass open-mindedness and readiness to listen and learn from each other concerns and experience for peace, harmony and viable neighborly relationships. What if President Putin, President Biden and the EU leaders would meet in person and hold dialogue for peace-making and conflict resolution? The violent assumptions of NATO outreach to Eastern Europe and outrageous hypothesis of military confrontations will lead nowhere – certainly not to peace-making or conflict resolution. Who could gain in these mindless and extravagant ideas of power and military warfare? Certainly None. Rationality dictates objective reasons for the good of people, societies and nations. The sudden plunge to cataclysmic destruction of human societies and civilizations by the few is a misleading notion and aims, no matter how powerful they claim to be.
The equality of human beings and states sovereignty synthesizes an equal consideration and treatment of all involved in this warfare. The global awakening of human soul and consciousness demands an immediate ceasefire and return to face to face negation between Russia, Ukraine, NATO and the EU. If they reject, surely, a debacle could swept unthinkable catastrophic consequences and ends for all on this planet.
Isn't true that intelligent leaders and people always readily accept reasoned advice?
Human Catastrophe, Present Putin, President Zelenski, Russia, Ukraine
America, NATO and Russia: Why War and Not Peace? by Mahboob A. Khawaja, PhD.
Posted by admin in Fifth Generation Warfare, Global Attacks on Islam, Islam: The Universal Message of Peace, The Great Game on February 6th, 2022
America, NATO and Russia: Why War and Not Peace?
Immanuel Kant, the German philosopher wrote in Critique of Practical Reason, imperative of a moral human conduct requires to "Act in such a way that you treat humanity, whether in your own person or in the person of another, always at the same time as an end and never simply as a means".
The on-going war syndrome between America, NATO and Russia typifies what can go wrong when human thinking evolution is restricted to imaginary temptations and compulsions of animosities. Often myth of war is manifested in self-interest and its evolution becomes unquestionable on the rational screen of reason and moral values. We see immature leadership surrounding conditions of egoism and primacy of will to clarify the reality of warfare. As of this moment, there is no emerging war on the ground, alleged "aggression" or dynamic syntax of compelling move to unleash war either by Russia against Ukraine or Russia vs. America and NATO. Rationally speaking, the myth of contemporary ground warfare is incomprehensible to an outsider. The mass media propaganda appears to set the psychological conditions of "Russian war" and "aggressions" without realizing its real-world affects, it spreads frightening waves of fear and hatred across the globe.
Is the media propaganda a weapon of conquest of human minds, soul and future-making? The imaginary foresight of any futuristic war captions the secret of unknown and unthinkable. All wars have consequences. The global humanity is not over with the aftermath of even the 2nd WW as books, live stories and movies are constantly haunting the human intellect – what were the motives of the 2nd WW and why did Europeans killed each other in millions and destroyed the human habitats. Those who had planned and carried out the war atrocities have long disappeared to be held accountable for their crimes against humanity.
Scott Ritter (is a former Marine Corps intelligence officer who served in the former Soviet Union implementing arms control treaties, in the Persian Gulf during Operation Desert Storm, and in Iraq overseeing the disarmament of WMD): "What War with Russia Look Like" (Global Research: 01/11/2022): explains the irony of current affairs:
If the U.S. tries to build up NATO forces on Russia's western frontiers in the aftermath of any Russian invasion of Ukraine, Russia will then present Europe with a fait accompli in the form of what would now be known as the "Ukrainian model." In short, Russia will guarantee that the Ukrainian treatment will be applied to the Baltics, Poland, and even Finland, should it be foolish enough to pursue NATO membership.
Is there an Irony of Political Wickedness and Paradox of Egoism?
If you know the nature of warmongering, you will certainly know yourself from the annals of formative history. Are we, the People of the 21st century knowledge-based age detached from the Nature of Things – who we are and where we are?. How would one rationalize the role and actions of NATO in a global theater of strategic interests? The history of NATO and its plans and ideological motives are equally distorted and disfigured on the global screen of reason, honesty and accountability. NATO lost its importance by siding with the US for its invasion and participation of war in Afghanistan, Iraq, Libya and Syria. The current tensions draw divergent and conflicting perceptions and order of wisdom to analyze its short and long terms impacts.
All political configurations carry much evil and much good in any rational optimistic outlook. Russia embarked on massive military buildup on the Ukrainian border to endanger its security. Its impact is alarming to most NATO allies assuming an imminent attack on Ukraine. Not so, deny most Russian officials, it is an exercise on Russian territory. America –led NATO believes it as a challenge to its vulnerability and strategic presence in the region. How to analyze the intentions in a knowledge-based strategic foresight? True knowledge is always benevolent not concealing or retaliating. America and NATO reacting to a tense situation requiring logical clarification of hidden intentions. If time and history are a reference point, those obsessed with invincible armies cannot solve the emerging political crises.
Is Russia exercising its own right to hold usual military drills? There is no alleged "aggression" or any hostile acts performed so far against the Ukraine sovereignty. America and West Europeans are keen to strike a balancing act by stationing their troops to the adjacent East European countries to remind the Russian of any consequences if they strike at Ukraine. If we are open to adaptability and an enlightened fortitude and respect for state's sovereignty, could Russia be helping Ukraine to strengthen its strategic capacity and telling the US-led NATO to be aware that Russia has strong defensive capacity in all emerging unusual situations. Russia, NATO and America cannot possibly be participants to a self-destructive formula. Rationality finds it hard what Russian can or will gain if Ukraine is attacked. Is it a puzzle game launched by the policy makers on both sides of the iron curtain?
Often crises have multiple spillover agendas. President Putin and President Zellensky of Ukraine could be meeting at Beijing Winter Olympic this week. Both know well their future survival lives in peace and harmony, not in conflict-making or conflict-keeping. German Chancellor and French President Macron are active to bring the both sides to some diplomatic resolution. President Endogen (Turkey) has offered Russia and Ukraine to hold peace talks in Ankara. Could America-NATO be surprised if Putin and Zellensky come to a meeting of mind and sign a diplomatic peace agreement? Both nations share an enriched geography and history together. Could we witness the coming of a new age of relationship between Russia, Ukraine and America-led NATO when even most exemplary political vigilance could be liable to unusual surprises?
Leader of Reason to Think before They Act
The voices of reason are loud and clear as global humanity cannot suffer the penalties of tyranny and war-mongering. It is wasting our time and consuming our precious thoughts and energies for man-made insidious cruelty. We, the People live in a splendid Universe in which planet Earth floats ordained by the commands of God. We are moral beings unlike animals that possess the eyes, ear and other senses but cannot draw logical conclusion from their senses. We, the People draw rational results what we see and what we hear, therefore, we are different, and we cannot act or live animals. There is a moral sense of spirituality and humanity to co-exist in harmony with the rest of all creations on this Earth. Modern wars represent sadistic and cruel minds to undermine human rights, dignity and life. They destroy all that is built over the ages that we call human civilizations. All of the Abrahamic Faiths (Judaism, Christianity and Islam) reveal the truth of "Trust" – human beings took to be responsible and be at peace with the Nature of Things. Any favorite perversion would contradict this Trust and embark on killing and destruction of human habitats as history has witnessed during the Two WW, Afghanistan, Iraq, Syria and ongoing in Yemen. This TRUST and its reminder are explicitly mentioned in (The Qur'an, Chapter 33: 72):
"We did indeed offer the Trust to the Heavens and the Earth; And the mountains but they refused; To undertake it, being afraid thereof: But man undertook it; He was indeed unjust and foolish."
The Earth is a living entity and is meant for us – the human beings. We, the People wonder who else except God created life, the Earth and the living Universe floating in space well balanced and functional since time immemorial. Who else other than God determined the Earth spinning of 1670 km per hour? Who else than God ordained it to orbit the sun at 107,000 km per hour? And who else than God made it to spin at 28,437 km per hour at the equator? (https://www.newscientist.com/question/fastearthspin/#ixzz7C8p37S9X). Be aware that earth average distance to the Sun is about 93 million miles (105 million km); the distance of Moon from Earth is currently 384,821 km equivalent to 0.002572 Astronomical Units and if the distance between the Earth-Sun or the Earth-moon were ever to change, there will be no sign of life or habitats left on Earth.
We, the People, wish to rejoice truth, not evil. We, the people of the world enjoin focused minds and imagination to articulate a new world of One Humanity, brotherhood and peaceful co-existence amongst all, free of hatred, fear, wars, encroachment and animosity. We, the people of the globe possess understanding – how to change the egoistic and embittered insanity of the few hate-mongers and warlords into equilibrium of balanced relationship between Man, Life and God- given living Universe in which we reside all.
America, NATO, Russia, the People, We
America, Russia and NATO look for new frontiers of influence by MAHBOOB A. KHAWAJA
Posted by admin in & NATO, America, Russia on January 27th, 2022
America, Russia and NATO look for new frontiers of influence
MAHBOOB A. KHAWAJA
Why the Lessons of History are Ignored
America, Russia and NATO's Geneva diplomatic talks ended in failure without any formal course of action to avoid military confrontation on Ukraine's border. Other vital issues include how to treat each other in a futuristic imaginary encounter of common interests. The global community is watching the prelude to a staged drama of unwarranted warfare with profligacy, malevolence and unknown miseries of unthinkable multitudes. All the superpowers — the stage actors of the 21st century have fictitious monsters equipped with innovative sophistry and captivating eloquence to talk about peace, security, human rights, global order and justice. They are master of deception playing on the passion of entrenched and exhausted mankind as if they could stop the emerging pains, horrors and devastations of warmongering to ensure a return to normalization of human affairs. To an inner human analytical eye encompassing proactive sense of global peace and harmony, it does not appear rational to articulate fears and misleading intentions to safeguard human peace and dignity while all the actions speak of a different language of obsessed assertions based on their own despotic national interests.
They claim peace but talk about threats of wars — how to rationalize the irony of human wickedness and inherent deception. Was the same stage drama not enacted during the First and 2nd World Wars killings millions and millions of people across this Planet Earth?
Human progress and future-making are jeopardized when lessons of history are deliberately misinterpreted and ignored by the paranoid, vengeful and suspicious leaders. If war is the only avenue to seek peace, we are on the wrong side of history and thinking of our future. It took several centuries to Europeans to understand the false shadows of apprehension of peace and harmony and to come to terms with nation-building, some resemblance of democracy, human equality of rights and unity for future-making via the EU. A reasoned perspective would illustrate that Russia after breakdown of the former USSR is not the same inheriting entity of Communist authoritarian ideology, leadership, institutions, political thought, policies and practices within the working systems of global order. There are visible progressive movements for political change, open communications, elections, institutional developments and global interactions and seeking reunion with the global order, UNO, world institutions, friendly relationships with adversaries and balanced socio-economic ties with others. To enlarge the scope of reason and understanding, Russia needs formidable change as it appears to be forging on different national strategic interests; its position on Ukraine is not the same as eluded by most NATO members. Ukraine and Russia have common geography and history just like Britain, France and Germany have. Would it not be a matter of extreme political-strategic sensitivity if other perceived enemies would dare to come close to military confrontations in Western Europe? It is logical that true friends of humanity will not act blindly to cause wild uproars and evil-mongering against the people anywhere on this planet.
Is NATO relevant to the 21st Century Emerging Conflicts
The focal issue seems to be the prospective membership of Ukraine to the community of NATO in Western Europe. America and Russia and other EU members enjoin conflicting views on this issue. Ukraine claims its freedom to join any international organizations for its betterment, peace and security. Russia and America should not be concerned except that they want to draw certain strategic gains out of this chaotic perceived tragedy of futuristic warfare. After the dreadful consequences of the Two World Wars, NATO was formed by the Europeans to maintain peace and security and avoid futuristic unwarranted national wars within the European hemisphere. Its formation and scope is limited to nationalistic conventional warfare in the European theatre.
One wonders, what wars did NATO fight to protect its ideals and strategic priorities after the 1945 WW? What NATO had to do with Afghanistan, Iraq and Libya? Were these accidental engagements or simply an extension of planned mischievous catastrophic instances of tyranny against the people of Iraq, Afghanistan, Libya and elsewhere who never posed any threat to NATO, America or to any Western European nations?
American and European leaders pushed soldiers to fight unwanted draconian wars in Afghanistan, Iraq, Libya and other parts of the Middle East.
Millions of innocent Afghans, Iraqi and Libyan civilians and Western troops were killed during the US-led NATO wars in these regions.
Could any American-European leaders explain why every day approximately 18-25 US war veterans commit suicides? ("Why Do Soldiers Commit Suicide and Global Warlords." Uncommon Thought Journal, USA).
How would one rationalize the role and actions of NATO in a global theater of strategic interests? The history of NATO and its plans and ideological motives are equally distorted and disfigured on the global screen of reason, honesty and accountability. Russia overtook eastern parts of Ukraine — Crimea by forces aligned to Russian speaking masses and trying to integrate those territories into the Russian federation. This issue has been discussed between Russian President Putin, former German Chancellor Merkel and French President Macron on several occasions. Could America, Russia and Europeans not talk again for a peaceful resolution of this and other related problems?
Looking for Hope of Peace Beyond the Lens of Geopolitics
If the U. S. tries to build up NATO forces on Russia's western frontiers in the aftermath of any Russian invasion of Ukraine, Russia will then present Europe with a fait accompli in the form of what would now be known as the "Ukrainian model." In short, Russia will guarantee that the Ukrainian treatment will be applied to the Baltics, Poland, and even Finland, should it be foolish enough to pursue NATO membership. Russia won't wait until the U. S. has had time to accumulate sufficient military power, either. Russia will simply destroy the offending party through the combination of an air campaign designed to degrade the economic function of the targeted nation, and a ground campaign designed to annihilate the ability to wage war. Russia does not need to occupy the territory of NATO for any lengthy period-just enough to destroy whatever military power has been accumulated by NATO near its borders.
Is NATO being managed by those people who lived in the distant past and perhaps post WW2 historic culture is still alive and flourishing? Is there any glimpse of hope for change and new reasoned relationships between America, Russia and West European people? The future of violence and nationalistic resentment looks embedded into the distorted strategic necessities of the current affairs, be it the argument of Russia or American-led NATO and or the EU on its own. NATO is run by the wrong people, glued to wrong thinking and doing the wrong things without any rational sense of time, people's interest and history. Craig Murray (NATO-an idea Whose Time has Gone), a former British diplomat and Rector of the University of Dundee, UK, foresees the body as obsolete to emerging strategic thinking and needs of the Western alliance:
It is also the case that the situation in countries where NATO has been most active in killing people, including Iraq, Libya, Afghanistan and Pakistan, has deteriorated. It has deteriorated politically, economically, militarily and socially. The notion that NATO member states could bomb the world into good was only ever believed by crazed and fanatical people like Tony Blair and Jim Murphy of the Henry Jackson Society. It really should not have needed empirical investigation to prove it was wrong, but it has been tried, and has been proved wrong….NATO's attempt to be global arbiter and enforcer has been disastrous at all levels. Its plan to redeem itself by bombing the Caliphate in Iraq and Syria is a further sign of madness. Except of course that it will guarantee some blowback against Western targets, and that will "justify" further bombings, and yet more profit for the arms manufacturers. On that level, it is very clever and cynical. NATO provides power to the elite and money to the wealthy.
Dr. Paul Craig Roberts was US Assistant Secretary of the Treasury and is author The Failure of Laissez Faire Capitalism and Economic Dissolution of the West, How America Was Lost, and The Neoconservative Threat to World Order. "Provocations Have A History Of Escalating Into War: Can War Be Avoided and the Planet Saved?" Information Clearing House: (8/31/2018), strikes a proactive caution to all Western policy makers:
The Zionist neoconservatives who rule in Washington are capable of the same mistake that Napoleon and Hitler made. They believe in "the end of history," that the Soviet collapse means history has chosen America as the model for the future. Their hubris actually exceeds that of Napoleon and Hitler. When confronted with such deluded and ideological force, does turning the other cheek work or does it encourage more provocation.
Every beginning has its end. America needs Navigational Change after January 6, 2021 Trump's Revolutionary attack on democracy and the Constitution, and so does Russia and NATO in their search for peaceful transition to sustainable future-making. We, the People of global community live on one floating Planet Earth, and we must be conscious — who we are, how connected we are in human solidarity and where are we heading to in our imagination of the present and future. It is awful and a tragedy of human conscience to be speaking of military conflicts and territorial gains when mankind urgently needs an effective cure for the Covid-19 pandemic. George Floyd cries continues to be heard all over the globe: "I can't' breathe." We are One People, One Humanity — ignorance, natural disasters, and man-made fatalities know not any borders, flags and nationalities but surge like wildfire as being witnessed in the COVID-19 pandemic worldwide. Again political absolutism heightens animosity and hatred rather than human understanding and cooperation for a precious cause of saving human lives on Earth. To save life of one human being is to safeguard the whole of humanity. We are all born equal One Humanity: — the Divine Message of Al-Qur'an clarifies the truth:
"Proclaim in the name of thy Lord and Cherisher, Who created, Created man (human being) out of a (mere) clot of congealed blood, Proclaim! And thy Lord is Most Bountiful, He Who taught (the use of) the Pen, Taught man (human being) that which he knew not."
Afghanistan, America, NATO, Pakistan, Russia
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 7,818
|
import datetime
import webob
from nova.api.openstack.compute.contrib import flavor_access
from nova.api.openstack.compute.contrib import flavormanage
from nova.compute import flavors
from nova import context
from nova import db
from nova import exception
from nova.openstack.common import jsonutils
from nova import test
from nova.tests.api.openstack import fakes
def fake_get_flavor_by_flavor_id(flavorid, ctxt=None, read_deleted='yes'):
if flavorid == 'failtest':
raise exception.FlavorNotFound(flavor_id=flavorid)
elif not str(flavorid) == '1234':
raise Exception("This test expects flavorid 1234, not %s" % flavorid)
if read_deleted != 'no':
raise test.TestingException("Should not be reading deleted")
return {
'root_gb': 1,
'ephemeral_gb': 1,
'name': u'frob',
'deleted': False,
'created_at': datetime.datetime(2012, 1, 19, 18, 49, 30, 877329),
'updated_at': None,
'memory_mb': 256,
'vcpus': 1,
'flavorid': flavorid,
'swap': 0,
'rxtx_factor': 1.0,
'extra_specs': {},
'deleted_at': None,
'vcpu_weight': None,
'id': 7,
'is_public': True,
'disabled': False,
}
def fake_destroy(flavorname):
pass
def fake_create(context, kwargs):
flavorid = kwargs.get('flavorid')
if flavorid is None:
flavorid = 1234
newflavor = {'flavorid': flavorid}
newflavor["name"] = kwargs.get('name')
newflavor["memory_mb"] = int(kwargs.get('memory_mb'))
newflavor["vcpus"] = int(kwargs.get('vcpus'))
newflavor["root_gb"] = int(kwargs.get('root_gb'))
newflavor["ephemeral_gb"] = int(kwargs.get('ephemeral_gb'))
newflavor["swap"] = kwargs.get('swap')
newflavor["rxtx_factor"] = float(kwargs.get('rxtx_factor'))
newflavor["is_public"] = bool(kwargs.get('is_public'))
return newflavor
class FlavorManageTest(test.NoDBTestCase):
def setUp(self):
super(FlavorManageTest, self).setUp()
self.stubs.Set(flavors,
"get_flavor_by_flavor_id",
fake_get_flavor_by_flavor_id)
self.stubs.Set(flavors, "destroy", fake_destroy)
self.stubs.Set(db, "flavor_create", fake_create)
self.flags(
osapi_compute_extension=[
'nova.api.openstack.compute.contrib.select_extensions'],
osapi_compute_ext_list=['Flavormanage', 'Flavorextradata',
'Flavor_access', 'Flavor_rxtx', 'Flavor_swap'])
self.controller = flavormanage.FlavorManageController()
self.app = fakes.wsgi_app(init_only=('flavors',))
self.request_body = {
"flavor": {
"name": "test",
"ram": 512,
"vcpus": 2,
"disk": 1,
"OS-FLV-EXT-DATA:ephemeral": 1,
"id": unicode('1234'),
"swap": 512,
"rxtx_factor": 1,
"os-flavor-access:is_public": True,
}
}
self.expected_flavor = self.request_body
def test_delete(self):
req = fakes.HTTPRequest.blank('/v2/123/flavors/1234')
res = self.controller._delete(req, 1234)
self.assertEqual(res.status_int, 202)
# subsequent delete should fail
self.assertRaises(webob.exc.HTTPNotFound,
self.controller._delete, req, "failtest")
def _test_create_missing_parameter(self, parameter):
body = {
"flavor": {
"name": "azAZ09. -_",
"ram": 512,
"vcpus": 2,
"disk": 1,
"OS-FLV-EXT-DATA:ephemeral": 1,
"id": unicode('1234'),
"swap": 512,
"rxtx_factor": 1,
"os-flavor-access:is_public": True,
}
}
del body['flavor'][parameter]
req = fakes.HTTPRequest.blank('/v2/123/flavors')
self.assertRaises(webob.exc.HTTPBadRequest, self.controller._create,
req, body)
def test_create_missing_name(self):
self._test_create_missing_parameter('name')
def test_create_missing_ram(self):
self._test_create_missing_parameter('ram')
def test_create_missing_vcpus(self):
self._test_create_missing_parameter('vcpus')
def test_create_missing_disk(self):
self._test_create_missing_parameter('disk')
def _create_flavor_success_case(self, body):
url = '/v2/fake/flavors'
req = webob.Request.blank(url)
req.headers['Content-Type'] = 'application/json'
req.method = 'POST'
req.body = jsonutils.dumps(body)
res = req.get_response(self.app)
self.assertEqual(200, res.status_code)
return jsonutils.loads(res.body)
def test_create(self):
body = self._create_flavor_success_case(self.request_body)
for key in self.expected_flavor["flavor"]:
self.assertEqual(body["flavor"][key],
self.expected_flavor["flavor"][key])
def test_create_public_default(self):
del self.request_body['flavor']['os-flavor-access:is_public']
body = self._create_flavor_success_case(self.request_body)
for key in self.expected_flavor["flavor"]:
self.assertEqual(body["flavor"][key],
self.expected_flavor["flavor"][key])
def test_create_flavor_name_with_leading_trailing_whitespace(self):
self.request_body['flavor']['name'] = " test "
body = self._create_flavor_success_case(self.request_body)
self.assertEqual("test", body["flavor"]["name"])
def test_create_without_flavorid(self):
del self.request_body['flavor']['id']
body = self._create_flavor_success_case(self.request_body)
for key in self.expected_flavor["flavor"]:
self.assertEqual(body["flavor"][key],
self.expected_flavor["flavor"][key])
def _create_flavor_bad_request_case(self, body):
self.stubs.UnsetAll()
url = '/v2/fake/flavors'
req = webob.Request.blank(url)
req.headers['Content-Type'] = 'application/json'
req.method = 'POST'
req.body = jsonutils.dumps(body)
res = req.get_response(self.app)
self.assertEqual(res.status_code, 400)
def test_create_invalid_name(self):
self.request_body['flavor']['name'] = 'bad !@#!$% name'
self._create_flavor_bad_request_case(self.request_body)
def test_create_flavor_name_is_whitespace(self):
self.request_body['flavor']['name'] = ' '
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_name_too_long(self):
self.request_body['flavor']['name'] = 'a' * 256
self._create_flavor_bad_request_case(self.request_body)
def test_create_without_flavorname(self):
del self.request_body['flavor']['name']
self._create_flavor_bad_request_case(self.request_body)
def test_create_empty_body(self):
body = {
"flavor": {}
}
self._create_flavor_bad_request_case(body)
def test_create_no_body(self):
body = {}
self._create_flavor_bad_request_case(body)
def test_create_invalid_format_body(self):
body = {
"flavor": []
}
self._create_flavor_bad_request_case(body)
def test_create_invalid_flavorid(self):
self.request_body['flavor']['id'] = "!@#!$#!$^#&^$&"
self._create_flavor_bad_request_case(self.request_body)
def test_create_check_flavor_id_length(self):
MAX_LENGTH = 255
self.request_body['flavor']['id'] = "a" * (MAX_LENGTH + 1)
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_leading_trailing_whitespaces_in_flavor_id(self):
self.request_body['flavor']['id'] = " bad_id "
self._create_flavor_bad_request_case(self.request_body)
def test_create_without_ram(self):
del self.request_body['flavor']['ram']
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_0_ram(self):
self.request_body['flavor']['ram'] = 0
self._create_flavor_bad_request_case(self.request_body)
def test_create_without_vcpus(self):
del self.request_body['flavor']['vcpus']
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_0_vcpus(self):
self.request_body['flavor']['vcpus'] = 0
self._create_flavor_bad_request_case(self.request_body)
def test_create_without_disk(self):
del self.request_body['flavor']['disk']
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_minus_disk(self):
self.request_body['flavor']['disk'] = -1
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_minus_ephemeral(self):
self.request_body['flavor']['OS-FLV-EXT-DATA:ephemeral'] = -1
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_minus_swap(self):
self.request_body['flavor']['swap'] = -1
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_minus_rxtx_factor(self):
self.request_body['flavor']['rxtx_factor'] = -1
self._create_flavor_bad_request_case(self.request_body)
def test_create_with_non_boolean_is_public(self):
self.request_body['flavor']['os-flavor-access:is_public'] = 123
self._create_flavor_bad_request_case(self.request_body)
def test_flavor_exists_exception_returns_409(self):
expected = {
"flavor": {
"name": "test",
"ram": 512,
"vcpus": 2,
"disk": 1,
"OS-FLV-EXT-DATA:ephemeral": 1,
"id": 1235,
"swap": 512,
"rxtx_factor": 1,
"os-flavor-access:is_public": True,
}
}
def fake_create(name, memory_mb, vcpus, root_gb, ephemeral_gb,
flavorid, swap, rxtx_factor, is_public):
raise exception.FlavorExists(name=name)
self.stubs.Set(flavors, "create", fake_create)
url = '/v2/fake/flavors'
req = webob.Request.blank(url)
req.headers['Content-Type'] = 'application/json'
req.method = 'POST'
req.body = jsonutils.dumps(expected)
res = req.get_response(self.app)
self.assertEqual(res.status_int, 409)
def test_invalid_memory_mb(self):
"""Check negative and decimal number can't be accepted."""
self.stubs.UnsetAll()
self.assertRaises(exception.InvalidInput, flavors.create, "abc",
-512, 2, 1, 1, 1234, 512, 1, True)
self.assertRaises(exception.InvalidInput, flavors.create, "abcd",
512.2, 2, 1, 1, 1234, 512, 1, True)
self.assertRaises(exception.InvalidInput, flavors.create, "abcde",
None, 2, 1, 1, 1234, 512, 1, True)
self.assertRaises(exception.InvalidInput, flavors.create, "abcdef",
512, 2, None, 1, 1234, 512, 1, True)
self.assertRaises(exception.InvalidInput, flavors.create, "abcdef",
"test_memory_mb", 2, None, 1, 1234, 512, 1, True)
class FakeRequest(object):
environ = {"nova.context": context.get_admin_context()}
class PrivateFlavorManageTest(test.TestCase):
def setUp(self):
super(PrivateFlavorManageTest, self).setUp()
self.flags(
osapi_compute_extension=[
'nova.api.openstack.compute.contrib.select_extensions'],
osapi_compute_ext_list=['Flavormanage', 'Flavorextradata',
'Flavor_access', 'Flavor_rxtx', 'Flavor_swap'])
self.controller = flavormanage.FlavorManageController()
self.flavor_access_controller = flavor_access.FlavorAccessController()
self.ctxt = context.RequestContext('fake', 'fake',
is_admin=True, auth_token=True)
self.app = fakes.wsgi_app(init_only=('flavors',),
fake_auth_context=self.ctxt)
self.expected = {
"flavor": {
"name": "test",
"ram": 512,
"vcpus": 2,
"disk": 1,
"OS-FLV-EXT-DATA:ephemeral": 1,
"swap": 512,
"rxtx_factor": 1
}
}
def _get_response(self):
url = '/v2/fake/flavors'
req = webob.Request.blank(url)
req.headers['Content-Type'] = 'application/json'
req.method = 'POST'
req.body = jsonutils.dumps(self.expected)
res = req.get_response(self.app)
return jsonutils.loads(res.body)
def test_create_private_flavor_should_not_grant_flavor_access(self):
self.expected["flavor"]["os-flavor-access:is_public"] = False
body = self._get_response()
for key in self.expected["flavor"]:
self.assertEqual(body["flavor"][key], self.expected["flavor"][key])
flavor_access_body = self.flavor_access_controller.index(
FakeRequest(), body["flavor"]["id"])
expected_flavor_access_body = {
"tenant_id": "%s" % self.ctxt.project_id,
"flavor_id": "%s" % body["flavor"]["id"]
}
self.assertNotIn(expected_flavor_access_body,
flavor_access_body["flavor_access"])
def test_create_public_flavor_should_not_create_flavor_access(self):
self.expected["flavor"]["os-flavor-access:is_public"] = True
self.mox.StubOutWithMock(flavors, "add_flavor_access")
self.mox.ReplayAll()
body = self._get_response()
for key in self.expected["flavor"]:
self.assertEqual(body["flavor"][key], self.expected["flavor"][key])
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 9,617
|
{"url":"http:\/\/msp.org\/involve\/2016\/9-4\/p04.xhtml","text":"#### Vol. 9, No. 4, 2016\n\n Recent Issues\n The Journal Cover Page Editorial Board Editors\u2019 Addresses Editors\u2019 Interests About the Journal Scientific Advantages Submission Guidelines Submission Form Ethics Statement Subscriptions Editorial Login Author Index Coming Soon Contacts ISSN: 1944-4184 (e-only) ISSN: 1944-4176 (print)\nGraphs on 21 edges that are not 2-apex\n\n### Jamison Barsotti and Thomas W. Mattman\n\nVol. 9 (2016), No. 4, 591\u2013621\n##### Abstract\n\nWe show that the 20-graph Heawood family, obtained by a combination of $\\nabla Y$ and $Y\\nabla$ moves on ${K}_{7}$, is precisely the set of graphs of at most 21 edges that are minor-minimal with respect to the property \u201cnot $2$-apex\u201d. As a corollary, this gives a new proof that the 14 graphs obtained by $\\nabla Y$ moves on ${K}_{7}$ are the minor-minimal intrinsically knotted graphs of 21 or fewer edges. Similarly, we argue that the seven-graph Petersen family, obtained from ${K}_{6}$, is the set of graphs of at most 17\u00a0edges that are minor-minimal with respect to the property \u201cnot\u00a0apex\u201d.\n\n##### Keywords\nspatial graphs, intrinsic knotting, apex graphs, forbidden minors\n##### Mathematical Subject Classification 2010\nPrimary: 05C10\nSecondary: 57M15, 57M25","date":"2017-08-19 09:06:04","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 7, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.1808108389377594, \"perplexity\": 2978.899649414089}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2017-34\/segments\/1502886105334.20\/warc\/CC-MAIN-20170819085604-20170819105604-00032.warc.gz\"}"}
| null | null |
A record number of kites filled the skies in ideal conditions on Saturday at the 40th Annual Wright Kite Festival.
Two days of near perfect kite flying weather greeted the 40th Annual Wright Kite Festival this year, and the kite flyers of Kitty Hawk Kites took full advantage of it.
The display that was put on at the Wright Brothers Memorial in Kill Devil Hills was the most spectacular that any of us can remember. We even had a chance to ask John Harris, the President and founder of Kitty Hawk Kites where this year's display fell in the list of best kite flying days. He thought about it for a moment and said he didn't recall anything quite like this year.
Saturday was especially spectacular with a brisk northeast breeze blowing at a constant 12-15 mph. With relatively few gusts and a consistent wind, a record number of diets filled the sky. There were unicorns, a spaceman, teddy bears, cows, fish, and a collection of large—make that huge—single line kites filling the sky.
Sunday was a great day as well—although not quite as good. The wind shifted to the south, and at the Wright Brothers Memorial, a south wind is not quite as good as an east wind for kite flying.
What was particularly nice about both days—Saturday especially—were daytime temperatures were tolerable. There have been some years when the heat and humidity of an Outer Banks July day made staying for any length of time at the event a challenge.
The Festival is more than just a bunch of kites in the sky.
There were stunt kite demos all day. Watching some of the KHK experts work with kids on how to fly a two line kite is wonderful—and the excitement of the kids when they actually get the kite to do what they want it to do is infectious.
There were also kite making workshops for little kids. They get to make a sled kite—a very basic design and the nice thing is it really flies.
It's a pretty sure bet the Wright Kite Festival will be back in 2019. It's always a weekend event, the second Saturday and Sunday of July.
There's always a lot happening on the Outer Banks during the summer. Check our Joe Lamb Jr., & Associates availability to see if you too can be part of the fun.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 3,933
|
<?php
namespace Omea\GestionTelco\EvenementsBundle\EvenementManager;
interface EvenementInterface
{
/**
* Code événement
* @return [type] [description]
*/
public function getCode();
}
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 6,226
|
.. _whatsnew_061:
Version 0.6.1 (December 13, 2011)
---------------------------------
New features
~~~~~~~~~~~~
- Can :ref:`append single rows <merging.append.row>` (as Series) to a DataFrame
- Add Spearman and Kendall rank :ref:`correlation <computation.correlation>`
options to Series.corr and DataFrame.corr (:issue:`428`)
- :ref:`Added <indexing.basics.get_value>` ``get_value`` and ``set_value`` methods to
Series, DataFrame, and Panel for very low-overhead access (>2x faster in many
cases) to scalar elements (:issue:`437`, :issue:`438`). ``set_value`` is capable of
producing an enlarged object.
- Add PyQt table widget to sandbox (:issue:`435`)
- DataFrame.align can :ref:`accept Series arguments <basics.align.frame.series>`
and an :ref:`axis option <basics.df_join>` (:issue:`461`)
- Implement new :ref:`SparseArray <sparse.array>` and ``SparseList``
data structures. SparseSeries now derives from SparseArray (:issue:`463`)
- :ref:`Better console printing options <basics.console_output>` (:issue:`453`)
- Implement fast :ref:`data ranking <computation.ranking>` for Series and
DataFrame, fast versions of scipy.stats.rankdata (:issue:`428`)
- Implement ``DataFrame.from_items`` alternate
constructor (:issue:`444`)
- DataFrame.convert_objects method for :ref:`inferring better dtypes <basics.cast>`
for object columns (:issue:`302`)
- Add :ref:`rolling_corr_pairwise <window.corr_pairwise>` function for
computing Panel of correlation matrices (:issue:`189`)
- Add :ref:`margins <reshaping.pivot.margins>` option to :ref:`pivot_table
<reshaping.pivot>` for computing subgroup aggregates (:issue:`114`)
- Add ``Series.from_csv`` function (:issue:`482`)
- :ref:`Can pass <window.cov_corr>` DataFrame/DataFrame and
DataFrame/Series to rolling_corr/rolling_cov (GH #462)
- MultiIndex.get_level_values can :ref:`accept the level name <advanced.get_level_values>`
Performance improvements
~~~~~~~~~~~~~~~~~~~~~~~~
- Improve memory usage of ``DataFrame.describe`` (do not copy data
unnecessarily) (PR #425)
- Optimize scalar value lookups in the general case by 25% or more in Series
and DataFrame
- Fix performance regression in cross-sectional count in DataFrame, affecting
DataFrame.dropna speed
- Column deletion in DataFrame copies no data (computes views on blocks) (GH
#158)
.. _whatsnew_0.6.1.contributors:
Contributors
~~~~~~~~~~~~
.. contributors:: v0.6.0..v0.6.1
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 5,234
|
<!DOCTYPE html>
<!--[if lt IE 7]> <html lang="en" ng-app="myApp" class="no-js lt-ie9 lt-ie8 lt-ie7"> <![endif]-->
<!--[if IE 7]> <html lang="en" ng-app="myApp" class="no-js lt-ie9 lt-ie8"> <![endif]-->
<!--[if IE 8]> <html lang="en" ng-app="myApp" class="no-js lt-ie9"> <![endif]-->
<!--[if gt IE 8]><!--> <html lang="en" ng-app="myApp" class="no-js"> <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<title>Servo | Visualizing Emergency Medical Services</title>
<meta name="description" content="">
<meta name="viewport" content="width=device-width, initial-scale=1">
<link rel="stylesheet" href="bower_components/bootstrap/dist/css/bootstrap.min.css">
<!-- Admin CSS Dependencies -->
<link href="assets/font-awesome/css/font-awesome.css" rel="stylesheet" />
<link rel="stylesheet" type="text/css" href="assets/css/zabuto_calendar.css">
<link rel="stylesheet" type="text/css" href="assets/js/gritter/css/jquery.gritter.css" />
<link rel="stylesheet" type="text/css" href="assets/lineicons/style.css">
<!-- App CSS Override Sheet -->
<!-- Custom styles for this template -->
<link href="assets/css/style.css" rel="stylesheet">
<link href="assets/css/style-responsive.css" rel="stylesheet">
<link rel="stylesheet" href="app.css">
<!-- Might Need for Furture Charting -->
<script src="assets/js/chart-master/Chart.js"></script>
<!-- HTML5 shim and Respond.js IE8 support of HTML5 elements and media queries -->
<!--[if lt IE 9]>
<script src="https://oss.maxcdn.com/libs/html5shiv/3.7.0/html5shiv.js"></script>
<script src="https://oss.maxcdn.com/libs/respond.js/1.4.2/respond.min.js"></script>
<![endif]-->
<script src="bower_components/html5-boilerplate/js/vendor/modernizr-2.6.2.min.js"></script>
</head>
<body>
<section id="container" >
<!-- Top Bar Header -->
<header class="header black-bg">
<div class="sidebar-toggle-box">
<div class="fa fa-bars tooltips" data-placement="right" data-original-title="Toggle Navigation"></div>
</div>
<div class="nav notify-row" id="top_menu">
<!-- notification start -->
<ul class="nav top-menu">
</ul>
<!-- notification end -->
</div>
<div class="top-menu">
<ul class="nav pull-right top-menu">
<li><a class="logout" href="report.html">report</a></li>
</ul>
</div>
</header>
<!-- Header Ends -->
<!-- Start of Sidebar -->
<aside>
<div id="sidebar" class="nav-collapse ">
<!-- Menu -->
<ul class="sidebar-menu" id="nav-accordion">
<p class="centered"><a href="#/stations"><img src="assets/img/servo.png" width="120px" class="logo"></a></p>
<li class="mt">
<a class="active" href="#/stations">
<i class="fa fa-map-marker"></i>
<span>Map Visualization</span>
</a>
</li>
</ul> <!-- end of menu -->
</div> <!-- end of sidebar -->
</aside>
<!--[if lt IE 7]>
<p class="browsehappy">You are using an <strong>outdated</strong> browser. Please <a href="http://browsehappy.com/">upgrade your browser</a> to improve your experience.</p>
<![endif]-->
<section id ="main-content" class="map-extension-wrap">
<section class="wrapper map-xtension">
<div ng-view></div>
</section>
</section>
</section>
<!-- footer>
<footer class="site-footer">
<div class="text-center">
Servo. Visualizing Ambulatory and Firemen Services.
<a href="index.html#" class="go-top">
<i class="fa fa-angle-up"></i>
</a>
</div>
</footer>
<!-- In production use:
<script src="//ajax.googleapis.com/ajax/libs/angularjs/x.x.x/angular.min.js"></script>
-->
<script type="text/javascript" src="https://maps.googleapis.com/maps/api/js?key=AIzaSyDjO_DA84sv9aSAkjhuJLTviHz2US30lWs&libraries=visualization"></script>
<script src="bower_components/angular/angular.js"></script>
<script src="bower_components/angular-route/angular-route.js"></script>
<script src="bower_components/angular-resource/angular-resource.js"></script>
<script src="app.js"></script>
<script src="servo.station/station.js"></script>
<script src="servo.incident/incident.js"></script>
<script src="components/version/version.js"></script>
<script src="components/version/version-directive.js"></script>
<script src="components/version/interpolate-filter.js"></script>
<!-- Dashboard Template JS Libraries -->
<!-- General Dependencies -->
<script src="assets/js/jquery-1.8.3.min.js"></script>
<script src="assets/js/jquery.scrollTo.min.js"></script>
<script src="assets/js/jquery.nicescroll.js" type="text/javascript"></script>
<script src="assets/js/jquery.sparkline.js"></script>
<script src="assets/js/common-scripts.js"></script>
<!-- Common scripts for all -->
<script type="text/javascript" src="assets/js/gritter/js/jquery.gritter.js"></script>
<script type="text/javascript" src="assets/js/gritter-conf.js"></script>
</body>
</html>
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 1,247
|
Un sōhei (僧 兵) és un monjo guerrer japonès. Aquests monjos guerrers van arribar a tenir un important poder durant el període feudal. No s'han de confondre amb els Yamabushi, un altre tipus de monjos que practicaven el Shugendo.
Història
Els sohei van aparèixer possiblement al , a partir de les divisions entre les diverses sectes budistes de la época. Durant el segle X van ocórrer moltes revoltes per part dels sohei contra el Govern, sobretot a Kyoto i Nara. A la fi del , els sohei es van fer forts sobretot durant les Guerres Gempei, mentre que les disputes entre els temples no acabaven. Els clans Minamoto i Taira van tractar d'obtenir l'ajuda dels sohei de Nara i Kyoto, intentant incloure les forces dels temples als exèrcits samurai. Taira no Kiyomori va enviar generosos tributs d'arròs i seda a Enryakuji, assegurant que no ajudaria als seus enemics; els Minamoto, per la seva part, es van aliar amb els monjos de Mii-dera.
L'any 1180, en una de les més famoses batalles en què els sōhei van participar, els monjos de Mii-dera i els samurais de Minamoto van tractar de defensar el pont sobre el Riu Uji i Byōdō-in a Batalla d'Uji, 1180. Els monjos Mii-dera van aixecar els taulers del pont, perjudicant als samurais a cavall perque no van poder creuar.
En aquesta batalla va aparèixer la figura de Gochin no Tajima, primer sohei del qual es diu que tallava per la mitat les fletxes que els Taira li llançaven. Els sohei, armats amb arcs, van resistir un temps però van ser finalment derrotats. Tot i la derrota, Taira no Kiyomori va ordenar que matessin els monjos que se li van oposar. Mii-dera va ser cremada una vegada més, com molts dels temples de Nara. Només el Enryaku-ji va aconseguir sortir il·lès.
L'any 1467 va començar la guerra de Ōnin, la qual va ser el preludi de més d'un segle de guerra civil al Japó i l'estímul d'una reorganització dels monjos guerrers. A diferència de les invasions dels mongols del , la guerra Ōnin es va lliurar principalment a Kyoto i per això els monjos guerrers no van poder romandre neutrals, a més d'haver-se format una nova raça de monjos guerrers: els monjos Ikkō-ikki. Quan els monjos de la Muntanya Hiei havien iniciat els ensenyaments de la secta Tendai, els nous monjos Ikkō-ikki van seguir els dictats de la secta Foto Shinshu. Eren essencialment coalicions de sacerdots religiosos fonamentalistes, a més de camperols i les seves famílies, que estaven disposats a lluitar per les seves creences. Ikkō-ikki es tradueix d'una fomra similar a grup devot, però allà per on passaven se'ls coneixia per les revoltes que portaven amb ells. L'any 1488, Rennyo, el líder dels monjos Ikkō-ikki, va incitar a una revolta contra les regles samurai a la província de Kaga. Des d'allà es van estendre a Nagashima, Ishiyama Hongan-ji i província de Mikawa. El seu creixent poder va atreure l'atenció de cabdills com Oda Nobunaga i Tokugawa Ieyasu, qui van reconèixer la seva força i els seus afiliats com un problema important.
Oda Nobunaga va pujar al poder a finals de la dècada de 1560. Els monjos de Enryaku-ji van recuperar el seu poder militar i van lliurar una sèrie d'altercats als carrers de Kyoto contra una nova secta rival del budisme Nichiren. Finalment va cremar tots els temples de Nichiren a Kyoto i després va buscar a aliats entre els senyors locals. Lamentablement per a ells, els clans Azai i Asakura amb què es van aliar eren enemics d'Oda Nobunaga.
Des del 29 de setembre de 1571, l'exèrcit de Nobunaga, que comptava amb 30.000 homes, va atacar Muntanya Hiei, destruint Enryaku-ji. Encara que va ser reconstruïda, mai va poder recosntruir l'exèrcit permanent de monjos guerrers. Nobunaga va passar a la lluita contra els Ikkō-ikki en les seves fortaleses de Nagashima i Ishiyama Hongan-ji. L'estiu de 1574, amb l'ajuda del pirata Kuki Yoshitaka, Nobunaga va bloquejar les fortaleses Ikkō i els seus 20.000 habitants van morir després de l'incendi provocat pels homes de Nobunaga i per inanició.
En les dues batalles de Kizugawaguchi, Nobunaga va derrotar el clan Mori, qui tenia el control naval de la zona. Els Ikko finalment es van veure obligats a rendir-se l'any 1580. Als anys 1580 i 1590, diverses faccions de monjos guerrers es van aliar amb Tokugawa Ieyasu o el seu rival Toyotomi Hideyoshi, lluitant en una sèrie de batalles i escaramusses. Quan Tokugawa Ieyasu finalment va derrotar a l'últim dels seus enemics, va pendre el control del país el 1603, any amb el que la vida dels monjos guerrers va arribar a la seva fi.
Armes
Els sohei eren molt variats en armament. Normalment solien portar amb si la naginata, una llarga llança amb un full unida al seu extrem. El seu llarg abast la feia molt efectiva contra cavalleria i infanteria, a més de ser bastant fàcil d'utilitzar. Aquesta arma estava relacionada amb els monjos guerrers i va ser usada pel famós Gochin no Tajima, també conegut com a Tajima el tallafletxes a la batalla d'Uji en 1180. La seva naginata va liderar la defensa d'un pont brandant de manera tan forta que les fletxes de l'enemic rebotaven sense fer-li la més petita rascada.
També se sap que solien utilitzar l'arc i fins i tot durant algunes batalles, utilitzaven unes fletxes especials que, en ser llançades, emetien un fort so semblant a un xiulet que servia per espantar els enemics o com a senyal.
Una altra arma coneguda era el mosquet. Els monjos guerrers eren famosos pel seu domini amb les armes de foc, en concret la secta Negoroji i els Ikko-Ikki.
La katana, però, era usada en el combat proper, sent els monjos guerrers del clan Uesugi els més temuts en l'ús d'aquesta arma.
Els sohei solien tenir grans coneixements d'arts marcials i exposar al màxim la seva resistència física, sent els Monjos de Hiei dels més importants. Els monjos de la muntanya Hiei professaven a l'escola budista de Tendai. A diferència de la majoria de les doctrines budistes, els seguidors de Tendai sostenien que es podia aconseguir la il·luminació en una sola vida. Per assolir aquesta finalitat, els monjos se sotmetien al Kaihogyo, una atroç prova de resistència física en què havien de recórrer 30 km per dia a peu durant 100 dies seguits al llarg de cinc anys consecutius. El sisè any, la distància s'ampliava fins als 60 km durant 100 dies i el setè fins als 84 km.
Vestimenta
Com a vestimenta, els monjos solien portar una sèrie de quimonos, generalment blanc a la part de sota i violeta a la part superior. Aquest estil ha canviat molt poc des de la introducció del budisme al segle VII. El calçat tradicional consistia en esclops de fusta o sandalies. Els monjos guerrers sovint solien portar un gran mocador blanc per cobrir-se el cap. Com a armadura, els monjos guerrers portaven part de l'armadura samurai.
Enllaços externs
University of West http://lib.uwest.edu/catalog/%E5%83%A7%E5%85%B5-%E5%8B%9D%E9%87%8E%E9%9A%86%E4%BF%A1%E8%91%97-so%CC%84hei-katsuno-ryu%CC%84shin-cho-0
Togheter with Japan http://jp.learnoutlive.com/the-sohei/
Legends and Chronicles http://www.legendsandchronicles.com/ancient-warriors/japanese-warriors/
Académia Seiryu https://academiaseiryu.wordpress.com/2016/05/04/yamabushi-y-sohei-los-monjes-guerreros-de-japon/
Crónicas de Rokurota http://cronicasderokurota.blogspot.com.es/2009/12/beve-historia-de-los-monjes-guerreros.html
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 6,922
|
Whether you want to team up and compare symptoms, or want to arrange a text buddy; post your requirements here (please read the forum rules before posting).
How to... keep track of your buddy thread / close your thread!
Due Feb 10, anyone else?
Due Date 18 Oct 17 Anyone Else?
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 263
|
Now He Remembers: Usman "Jorge is doing everything possible to avoid...
Kevin Lee's brother signs multi fight deal with Bellator.
2 year old promo video for Khabib vs Ferguson
The Film Room: Alistair Overeem vs. Jairzinho Rozenstruik
By Kevin Wilson Dec 6, 2019
Sign up for ESPN+ right here, and you can then stream UFC on ESPN+ live on your computer, phone, tablet or streaming device via the ESPN app.
The timeless Alistair Overeem on Saturday will step inside the Octagon for his 65th fight as a professional mixed martial artist when he takes on Jairzinho Rozenstruik in the UFC on ESPN 7 main event at the Capitol One Arena in Washington, D.C. The undefeated Rozenstruik steps up on short notice for just his 10th MMA appearance, but he does have 85 kickboxing bouts under his belt. From what "Bigi Boy" has shown, he could become a real threat in the heavyweight division, while a third consecutive victory might net Overeem his second crack at Ultimate Fighting Championship gold.
The Overeem-Rozenstruik showdown supplies the material for this installment of The Film Room.
Overeem has been fighting for so long that he now finds himself in the third phase of his career. Early on, he relied on athleticism and speed at 205 pounds and made a name for himself during the 2005 Pride Fighting Championships middleweight grand prix. Then the legend of "Ubereem" was born, as he made his full-time heavyweight debut in 2007 and went 12-1-1 over the next four years, with all but one of his wins coming by knockout or submission. Overeem later started to age and the UFC linked arms with the United States Anti-Doping Agency, which forced him to completely overhaul his style and become the patient and tactical fighter we see today. When he was younger, Overeem would stand and trade in the pocket with anybody and was rarely concerned with defense. These days, he puts a major focus on his defensive footwork and ability to outsmart opponents from the outside with his length and two decades of fighting knowledge. The newfound patient and tactical approach allowed him to have a career resurgence, as he has gone 9-4 with six knockouts since adopting this style in 2014.
Overeem was once known for bursting forward with combos and pressing a relentless pace. Now, his leading attacks are much more measured and never account for more than three to four strikes before he resets at a safe distance. Something to notice about the new Overeem: the number of feints he uses. Instead of rushing forward, he now patiently waits for openings and sets traps with his feints and movement.
Along with this new patient style comes a newfound countering game for Overeem. He switches stances a lot during a fight, but most of his counters come from an orthodox stance, with the straight right as his go-to punch. This constant feinting and his willingness to remain patient allows him to sit back and read his opponents and wait for openings. Elite countering usually comes with time. Anybody can throw a right hook, but few will know how opponents will react to it and how to force them to throw something that will set up the right hook. That is what elite counterstrikers like Anderson Silva and Valentina Shevchenko do. Everything is done with more meaning than what is seen on the surface level, and it takes years of experience to perfect. Nobody in the sport has more experience than Overeem. He has seen every type of fighter, technique and game plan in the book, and his ability to see strikes coming and fire off an appropriate counter is getting better as he ages. Overeem is better when working on the lead, but it is nice to see that he can stay patient and work on the counter when needed and use his two decades of experience to read his opponents.
Overeem has always wielded great kicks, but now he mainly uses them to keep distance between himself and the opponent. He switched his camp to Jackson-Wink MMA in 2014 and immediately started using the same kicks and tactics Jon Jones uses to keep opponents at bay. He favors front kicks to the body since he can use them to achieve many different outcomes. They work wonders at stopping aggressive opponents while slowly draining their cardio. Overeem also uses them to switch stances and set up strikes from the opposite stance.
Something that has never left Overeem is his elite clinch game. Early in his career, he would rush forward with combos, grab the clinch and fire off some knees to the body of his opponents until they separated. Then, he did it all over again. These days, his clinch entries are craftier but the focus on it is the same. Notice how he fakes a takedown to get the opponents' hands low and then comes back up to grab underhooks and drive them to the cage. Once in the clinch, he will hammer knees to the body but is smart enough to not get too aggressive and lose the position. Since Rozenstruik is a dangerous power puncher who relies on being the leading man, expect Overeem to clinch more than ever in this fight. Rozenstruik also struggles on the ground, so the clinch is the perfect tactic to take away his best weapons while capitalizing on his own.
The biggest problem Overeem has faced as he ages: his ability to take a punch. With two decades of fighting on the odometer, the chin was bound to deteriorate -- especially with 13 knockout losses on his resume. Only time will tell if Overeem's chin can hold up, but if it does, we could conceivably see him fighting for a title again in the near future.
There is not a lot of footage on Rozenstruik, but what we have seen has looked promising. His first UFC fight against Junior Albini was the second-longest of his career, and we saw some things with which he struggles and some things at which he excels. As a kickboxer, his grappling is obviously lacking and Albini managed to hold him down for long stretches in the first round. However, late in the round, we saw Rozenstruik start working in longer combos, which is where he shines. Early in Round 2, he opened up with a couple of right hands and a head kick that dropped Albini and set up the finish. This fight mostly confirmed what we knew about Rozenstruik coming into the UFC. He struggles on the ground and will be hesitant if the threat of a takedown is there, but once he opens up, few can deal with his varied and speedy combos.
Rozenstruik's other two fights in the UFC lasted less than 30 seconds and did not really prove much other than he has above-average power in his lead hand. His kickboxing fights show mostly the same things we have seen. He is comfortable in a firefight and at his best when leading with long combos and kicks. It is rare that someone headlining a UFC card only has a few examples of how they fight, but it makes this bout all the more compelling.
Sherdog's 2019 Submission of the Year
UFC on ESPN 7 Main Card Averages 1.071 Million Viewers; Prelims Draw 696,000
Video: Jairzinho Rozenstruik Receives Hero's Welcome at Suriname Airport
Post-Mortem: UFC on ESPN 7
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 6,811
|
Bladder cancer is one of the most common types of cancer in the United States. Trouble breathing during a heart attack can happen for several reasons. Often, people who have a vascular constriction will present differently than those who have a heart condition that contributes to the high blood pressure. Only thing is he delevoped a serious case of rash on his face, chin, chest and back.
name brand viagra best generic viagra viagra coupons Symptoms include runny nose, cough, headache, and facial pressure. Chemotherapy uses chemotherapeutic drugs to kill malignant cells and can be delivered via single drug or a combination of drugs. Patients with multiple myeloma who do not experience symptoms may not require treatment.
viagra plus best place to buy generic viagra online viagra online no prescription It is important to keep in mind that while some mycotoxins have been recognized to affect the immune system, many earlier putative associations e. The next section reviews evidence underlining the importance of attending to particular depression symptoms. And be sure to replace the battery when you change the time on your clocks each spring and fall. She worried about the surgery, and she's taken on a number of my chores to help out.
viagra price generic viagra 100mg how to get viagra Note that if a woman is pregnant when diagnosed with cervical cancer, the treatment will depend on the stage of the cancer and of the pregnancy. When started within the first two days of treatment, they can reduce how long you'll have the flu but they can't cure it. Managing diabetes with exerciseLearn more about how to manage diabetes with exercise. A few months later an enginer left an electrical closet open and peered inside as I passed and there it was Black Mold all over one wall of the closet.
Controlling high blood pressure can protect your kidneys. The healing marks the end of the active phase of genital herpes. Colon Cancer Symptoms Constant tiredness and fatigue A change in bowel habits that lasts for more than a few days. Leg pain like this can be a sign of deep vein thrombosis DVT , which occurs when a blood clot forms inside a vein.buying viagra online Do not eat during the 2 to 3 hours before you go to bed. Spring is a beautiful time for most of us but not if you suffer from allergies. Extreme and untreated high blood pressure may, over time, cause blood vessels in the eye to burst, possibly damaging vision or causing blindness. Additionally, apple cider vinegar supports overall health by delivering several key nutrients to the system.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 8,098
|
\section{Introduction}\label{Introduction}
Opinion dynamics is an important area of research with a wide range of applications in political campaigning, marketing, transportation management, public opinion management \cite{dong2018survey}, and group recommender systems \cite{RecommenderSystems}. In particular, due to the rapid growth of online social networks and unprecedented ease of opinion exchange on these platforms, there has been growing interest in how individuals' opinions are formed and perceived within a population \cite{mastroeni}. An interesting phenomenon frequently observed on these platforms, yet largely rejected by the existing opinion dynamics models, is that extremist positions can emerge and become mainstream \cite{Green, ElusiveConsensus, Mask, BrexitGeographic}.
In fact, the vast majority of opinion dynamics models in the literature, e.g., \cite{noorazar2020classical} and references therein, have been focused on the notion of {\it conformity}, which broadly refers to the tendency of an individual to act so as to fit into a group by adopting or touting what he or she perceives as the popular opinion. It is thus expected, as happens in virtually every existing conformity-based model, that conformist individuals avoid extreme beliefs and agree on or hover around a middle-ground, compromise position as time grows. Various types of cognitive biases, most notably those of the confirmation bias type, have been incorporated into conformity-based models to make them more realistic and better capture the evolution of opinions \cite{hegselmann2002opinion, Adaptive, EmergenceOfExtremism, Cyber-SocialNetworks}. \textit{Confirmation bias} in a broad sense refers to the tendency of an individual to actively seek to confirm her preconceptions. This may be done by avoiding or limiting exposure to beliefs unlike hers or selective perception of the facts presented to her, among other means. The inclusion of the confirmation bias in conformity-based models has effectively broadened research on global agreement scenarios to more general scenarios in which one or multiple consensus clusters could be reached
Some attempts have been made to modify classical conformity-based models to ones where opinion polarization is not beyond the realms of possibility. A notable example is the Altafini model \cite{altafini2012dynamics, altafini2012consensus}, in which the notion of antagonism among individuals is incorporated, leading to an influx of models that consider antagonistic relationships in opinion networks. Among these, some models have also included {\it bounded confidence} and {\it biased assimilation}, each of which can be viewed as a type of confirmation bias, as an individual's characteristics in the dynamics \cite{InnerChoice, PolarizationSigned,BiasedAntagonism}. While antagonistic interactions/relationships can contribute to and justify the tendency toward more extreme beliefs in a divided population, such as that of the United States with respect to political ideology (see Fig.~\ref{US polarization}, extracted from \cite{PewResearchCenter2017}), extreme beliefs have also been observed to form in fully collaborative environments \cite{AbortionRights}.
Furthermore, the pre-existence of extreme tendencies is often required at the onset of evolution of opinions for these models to explain how such tendencies become mainstream.
\begin{figure}[t]
\centering
\includegraphics[width=.5\linewidth]
{GrowingIdeologicalDivideInTheUS.png}
\caption{Growing ideological divide in the United States (Source: Pew Research Center \cite{PewResearchCenter2017}).}
\label{US polarization}
\end{figure}
A handful of frameworks have been developed to address the emergence of popular extreme beliefs in fully collaborative environments where antagonistic relationships are absent. For instance, a model based on Persuasive Argument Theory (PAT) and homophily has been proposed in \cite{Differentiation, ArgumentBiasedAssimilation}, where it is suggested that those individuals who share similar viewpoints are more likely to engage in conversation with and have influence on each other. As another example, in \cite{BecomeExtremes}, a model has been introduced via the statistical physics modelling approach to explain the prevalence of extremism if stubborn extremists exist at the onset of the opinion evolution. Other remarkable examples are \cite{Dandekar2013, SocialBiases, AnalysisBiased}, where biased assimilation is incorporated into the Degroot's opinion averaging model \cite{degroot:74} to create the possibility of extreme beliefs emerging and becoming popular \cite{Dandekar2013, SocialBiases, AnalysisBiased}.
Our main objective in this work is to propose, conceptualize, and investigate a mathematical model inspired by the socio-psychological anlaysis of the groupthink phenomenon to study opinion formation. In particular, the model is capable of explaining the emergence of popular extreme beliefs in both united and divided populations. In the rest of section \ref{Introduction}, we provide more context to the problem of group polarization by further surveying the opinion dynamics literature and
reviewing some of the basic concepts and notions from the groupthink theory and related models, before highlighting our contributions in this work and describing the organization of the paper.
\subsection{Further Background to Opinion Dynamics}
\label{Intro.RelatedWork}
Agent-based models of opinion dynamics study the evolution of individuals' opinions through interactions between them. In that regard, DeGroot in \cite{degroot:74} provides one of the basic models, where an individual's opinion is updated to a weighted average of the neighboring individuals' opinions, including the individual himself/herself. The interested reader is referred to \cite{bullo2018lectures} for a detailed analysis of the DeGroot model and its straightforward extensions using tools from algebraic graph theory. A modified version of the DeGroot model, the Friedkin-Johnsen model \cite{friedkin2011social}, takes into account susceptibility to persuasion and the degree of individuals' stubbornness. Also, an alternative polar model \cite{Polar,Nematzadeh} suggests that the susceptibility to persuasion of each individual is based on his/her current opinion. In addition to cooperative relationships, there may also be antagonistic relationships among individuals in a balanced structure,
leading to a bipartite consensus in limit \cite{altafini2012dynamics,altafini2012consensus}, which could explain the shaping of a divided population. Sufficient conditions for bipartite consensus under time-varying collaborative and antagonistic relationships are studied in \cite{HostileCamps}.
The introduction of the Hegselmann-Kraus model \cite{hegselmann2002opinion} has led to a class of nonlinear models wherein only the individuals who hold very similar opinions interact with and influence each other, in a manner that resembles confirmation bias, e.g., \cite{SignedBounded, DiscreteSignedBounded} and references therein. Such interactions may result in the emergence of multiple consensus clusters. A case in which antagonistic, neutral, and collaborative relationships are defined with respect to confidence levels is studied in \cite{BalanceSeeking}, where it is shown that under certain conditions for the amount of confidence levels for mentioned types of relationships, global consensus, bipartite consensus or clustering of opinions can be achieved.
All the aforementioned models consider the evolution of opinions without the existence of agreement pressure on individuals. However, there is evidence that the agreement pressure in a group often influences individuals' opinions. For instance, the study of a case in which individuals are exposed to increasing but bounded agreement pressure states that individuals' opinions will converge to a fixed distribution \cite{AgreementPressure}. A modified version of the Hegselmann-Krause model incorporating a pressure parameter reaches consensus more easily and faster than the original Hegselmann-Krause model \cite{cheng2019opinion}. A model accounting for inconsistency between private and expressed opinions due to conformity pressure is studied in \cite{ye2019influence}. This model can describe Asch's line-matching paradigm thoroughly with the help of resilience to pressure and susceptibility to persuasion parameters. We refer the reader to \cite{anderson2019recent} for other recent developments in the field of opinion dynamics.
\subsection{An Overview of Groupthink}
The theory of groupthink was first introduced by Irving L. Janis in 1971 \cite{janis1971groupthink}. By definition, groupthink is a concurrence-seeking tendency. When this tendency becomes dominant in a cohesive in-group, members will irrationally disregard and decline unpopular realistic views \cite{janis1971groupthink}. In other words, this tendency prevents people from treating controversial views realistically \cite{ElliotAronson2016Social}. Groupthink has a close relation to the Solomon Asch line-matching paradigm \cite{robbins2018essentials}. Members holding different opinions from the majority are under pressure to conform and suppress their true opinions. Groupthink does not always happen, as it requires certain antecedent conditions to be met, such as high cohesiveness within the group, isolation of the members from contrary views, and a lack of impartial leadership. If these antecedents are met, certain consequences will be observed, such as the illusion of invulnerability/unanimity and self-censorship, often leading to defective decision-making \cite{janis2008groupthink}. This process is summarized in Fig.~\ref{groupthink conditions}. Remarkable consequential examples of the groupthink phenomenon and victims of groupthink
include the Bay of Pigs; the Pearl Harbor attack; the North Korean escalation; the Vietnam escalation \cite{janis1972victims}; the grounding of Swissair, the flying bank \cite{hermann2010grounding}, and more recently, to some degree, the failures at each stage of Brexit
\cite{lees2020brexit}.
\begin{figure*}[t!]
\centering
\includegraphics[width=1\linewidth]
{GroupthinkPhenomenon.png}
\caption{Groupthink phenomenon in summary, based on \cite{janis2008groupthink}}
\label{groupthink conditions}
\end{figure*}
Following the introduction of the groupthink theory by Irving L. Janis, several modifications have been proposed \cite{rose2011diverse}. According to the so-called {\it ubiquity model} of groupthink \cite{baron2005so}, the antecedent conditions in Fig.~\ref{groupthink conditions} are not necessary for the groupthink to occur. In fact, there is evidence that the symptoms in Fig.~\ref{groupthink conditions} have arisen without the presence of such antecedents. The ubiquity model introduces three key conditions necessary for this phenomenon to happen in everyday groups. They include social identification, salient norms, and low self-efficacy. Social identification is associated with social acceptance and social rejection, which means that a deviating member will be rejected socially as a punishment. The second antecedent, salient norms, suggests that a group norm will be created by interaction and discussion within the group. The last antecedent, low self-efficacy, is related to situational lack of self-confidence in one's problem-solving ability due to recent failure, fear, time pressure, or other factors. All three key conditions act as a motivation for suppressing dissenting views despite the realistic spirit of their content.
{\it Group polarization}, as a direct, immediate consequence of the groupthink phenomenon, is a group's tendency to make decisions that are more extreme than the initial positions of its members \cite{ElliotAronson2016Social}. For examples, the group's decision moves toward greater risk (or greater caution) if the initial tendency is to be somewhat risky (or cautious). The term group polarization, with which the current work is mainly concerned, should not be confused with opinion polarization, even though the two phenomena are correlated and may coexist within a population.
\subsection{Contributions}
\label{contributions}
The main contribution of this paper is the introduction and investigation of the Partisan Confidence (PC) model for social learning. It is capable of justifying the emergence of popular, extreme beliefs in a social network. The PC model is inspired by socio-psychological analysis of the groupthink phenomenon, the presence of which is evident in both united and divided populations. The PC model is based on a unique understanding of opinions and social influence in that opinions of like-minded individuals tend to resonate when they interact, compared to most of the existing models wherein a compromise instead of resonance is in play. Furthermore, under the PC model, the pre-existence of extremists for the emergence of extreme beliefs is not necessary. A detailed description of the paper's contributions is stated below.
\begin{itemize}[leftmargin=.125in]
\item {\it Decomposition of opinions based on Partisan Confidence:} An individual's opinion at any given instant is represented by a scalar, which is consistent with the vast majority of the existing opinion dynamics models. However, we shall rely on a unique characterization of opinions sketched as follows. Assuming that an individual's opinion lies in the interval $[-1,1]$, its sign is viewed as her general belief (ideology, political party, etc.), while its magnitude is interpreted as her confidence about that general belief, justifying the term {\it Partisan Confidence} It should be noted that the partisan confidence decomposition of opinions has resemblances to the viewpoints leading to the models presented in \cite{Dandekar2013,EmergenceOfExtremism}. In \cite{EmergenceOfExtremism}, it is assumed that in some situations the exact opinion of an individual is not known while its discrete choice (yes, no, neutral) is expressed publicly. In \cite{Dandekar2013}, the distance between an individual's opinion and each endpoint of the opinion spectrum represents her support for that marginal belief.
\item {\it Social influence characterization:} In our approach, when individuals interact, influence is modeled in such a way that (i) the opinions of like-minded individuals, i.e., those who share the same general belief at the time, resonate with each other, meaning that the individual make one another more confident in that shared general belief; (ii) it accounts for the confirmation bias, which means that the influence on each other of individuals with opposite general beliefs is discounted; and (iii) an individual with higher confidence in her general belief is deemed more influential. The first item listed above defies the mainstream line of opinion dynamics research where like-minded individuals tend to compromise (not resonate) when they interact
\item {\it Introducing the PC and PC-lite models:} Based on the characterization of social influence described above, we introduce and investigate two models for social learning, namely the PC and PC-lite models. While the PC model is more compelling, since it also accounts for the confirmation bias (item (ii) above), the PC-lite model is of great importance as it demonstrates that group polarization may occur even without the contribution of confirmation bias, which highlights a fundamental difference between this work and \cite{Differentiation,Dandekar2013}. It is also important to note that the PC and PC-lite models, unlike those introduced in \cite{Differentiation,Dandekar2013}, are both time-varying, that is a must-have property for models capturing a practical social dynamics.
\item {\it Deriving conditions for group polarization:} Through rigorous analysis of the PC and PC-lite models, we argue for the possible existence of communities within a population, henceforth called {\it social bubbles}, that are bound for group polarization if the antecedent conditions of groupthink are satisfied. A definition of {\it social bubble} is given; intuitively, it is a group isolated from the outside population beyond a certain level. We then prove that if a certain degree of homogeneity is present within a social bubble at some point in time, that is, if the individuals in the bubble share the same general belief and are confident in that belief beyond a relatively low degree, then their confidence will increase over time and reach an extremely high value, at which point the group polarization will have seemed to materialize in this bubble. The extent of the bubble's polarization is shown to be directly related to its isolation level, and for the PC model, also to the intensity of the confirmation bias of those within the bubble.
\end{itemize}
\subsection{Paper Organization}
In Subsection~\ref{Preliminaries And Notation}, we provide some preliminaries and notations for later use. In Section~\ref{First Model and its properties}, we introduce and investigate a basic model, the so-called {\it PC-lite} model, generalization of which leads to the PC model in Section~\ref{Second Model and its properties}. We discuss the inherent properties of the proposed models in their corresponding sections. In Section~\ref{Numerical Examples}, we provide simulations to demonstrate properties of the proposed models. We conclude the paper by identifying some future directions of research in Section~\ref{Conclusion}. For ease of presentation, we relegate all the proofs and auxiliary lemmas to the Section \ref{Proofs}.
\subsection{Notation}
\label{Preliminaries And Notation}
We let $\mathcal{V} = \{1,\ldots,n\}$ be the set of all individuals or, as is often called from now on, {\it agents.} The opinion of agent $i$ at discrete time $t$, $t \geq 0$, is denoted by $x_i(t) \in [-1,1]$, and its sign is denoted by $\sgn(x_i(t))$. A weighted, time-varying digraph $\mathcal{G}(t)=(\mathcal{V},\mathcal{E}(t),W(t))$ is assumed to represent the topology of the network of agents over time, where $\mathcal{V}$ is the set of nodes, $\mathcal{E}(t) \subseteq \{(i,j)|i,j \in \mathcal{V}, i\not=j\}$ is the set of edges at time $t$ indicating the interactions among the agents, and an element $w_{ij}(t)$ of the weight matrix $W(t)$ indicates the weight of influence of agent $j$ on agent $i$ at time $t$. It is assumed that $w_{ij}(t)>0$ if $(i,j) \in \mathcal{E}(t)$, and $w_{ij}(t)=0$ otherwise. The non-negative matrix $W(t)$ is called {\it row-stochastic} if the elements of each of its rows sum up to 1, that is, if $\sum_{j\in \mathcal{V}} w_{ij}(t)=1$ for all $i \in \mathcal{V}$. It is called {\it row-substochastic} if $\sum_{j\in \mathcal{V}} w_{ij}(t)\leq 1$ for all $i \in \mathcal{V}$, and there exists some $k\in \mathcal{V}$ such that $\sum_{j\in \mathcal{V}} w_{kj}(t) < 1$. Throughout the paper, and in the proofs in particular, the argument $t$ of time-varying functions is dropped for the benefit of notational convenience. For instance, $x_i$, $z$, and $w_{ij}$ often replace $x_i(t)$, $z(t)$, and $w_{ij}(t)$, respectively. Furthermore, an agent's update value, $x_i(t+1)-x_i(t)$, will be denoted by $\Delta x_i$, that itself is short for $\Delta x_i(t)$.
\section{Partisan Confidence-lite Model and Its Properties}
\label{First Model and its properties}
In this section, we introduce, justify, and investigate the Partisan Confidence-light (PC-lite) model for the evolution of opinions in a social network. Let $x_i(t) \in [-1,1]$ denote the opinion of agent $i \in \mathcal{V}$ at discrete time $t \geq 0$, where $\mathcal{V} = \{1,\ldots,n\}$ is the set of all agents. In the PC-lite model, the opinion of every agent $i$ evolves according to the following discrete-time dynamics:
\begin{flalign}
\text{\bf(PC-lite dynamics)}\hspace{1.1in} \Delta x_i = \sum_{j \neq i} \Big[w_{ij} |x_j| (\sgn(x_j) - x_i)\Big].&&
\label{model}
\end{flalign}
To be clear, as described in Subsection~\ref{Preliminaries And Notation}, the dynamics \eqref{model} should be read as
\begin{equation}
x_i(t+1) - x_i(t) = \sum_{j \neq i} \Big[ w_{ij}(t) |x_j(t)| (\sgn(x_j(t))-x_i(t)) \Big].
\end{equation}
According to the PC-lite dynamics \eqref{model}, a self-weight $w_{ii}$ is not present and does not contribute to the opinion change of agent $i$ at time $t$. Thus, we can assume $w_{ii}(t)=0, \forall i\in\mathcal{V}, t\ge 0$. This assumption eliminates the self-loop from the underlying graph $\mathcal{G}(t)$, and hence results in a row-substochastic adjacency matrix $W(t)$.
\subsection{Justification of the PC-lite Model}
\label{Justification of the PC-lite Model}
One notices that the PC-lite dynamics \eqref{model} follows, in principle, the same rule of social influence as the time-varying version of the DeGroot model \cite{degroot:74},
\begin{flalign}
\text{\bf(DeGroot dynamics)}\hspace{1.35in} \Delta x_i = \sum_{j \neq i} \Big[w_{ij} (x_j - x_i)\Big].&&
\label{degroot}
\end{flalign}
However, \eqref{model} is set up on a fundamentally distinctive interpretation of opinion perception, that is, agent $i$ perceives the opinion $x_j$ of agent $j$ as an approval of $\sgn(x_j)$ with confidence level $|x_j|$. Subsequently, the weight of influence of agent $j$ on agent $i$ is discounted by the factor $|x_j|$, while the opinion $x_j$ of agent $j$ is replaced by $\sgn{(x_j)}$.
Therefore, $x_j$ is decomposed into two parts: a direction part $\sgn{(x_j)}$, which can be viewed as party affiliation in political terms, and an intensity or confidence part $|x_j|$. It is this decomposition of agents' opinions that justifies the appellation {\it Partisan Confidence}. Indeed, $\sgn{(x_j)}$ is often concerned with a much more specific issue than one's political party. For instance, on the issue of abortion rights, it addresses whether a person generally supports or is against abortion rights.
The consideration that the influence weights $w_{ij}{(t)}$ are time-varying adds to the practicality of the PC-lite model since (i) no agent has to interact with the same set of agents at all times, i.e., there are asynchronous interactions, and (ii) the dynamics \eqref{model} with fixed weights $w_{ij}$ cannot accurately model human thinking, i.e., there is model uncertainty.
We may use $w_{ij}(t)|x_j(t)|$ to refer to the overall influence of agent $j$ on agent $i$ at time $t$. The amount of this overall influence is determined by the intensity of an opinion $|x_j(t)|$. On the other hand, the direction $\sgn{(x_j(t))}$ determines whether the overall influence is in favor of or against an opinion.
In fact, one can think of $|x_j(t)|$ as the intensity of the emotion being transferred to an another agent that either advocates or disapproves of some position or idea, and this intensity governs the overall influence. Thus, the more extreme the emotion, the greater the overall influence would be, and vice versa. When the opinion of agent $j$ at time $t$ is zero, we assume that she is completely neutral; thus, her opinion will not drag the opinion of agent $i$ at time $t$ in either directions on the opinion spectrum. In other words, a neutral opinion dose not contribute to the opinion change of an agent.
\subsection{Properties of the PC-lite Model}
\label{Properties of the First Model}
We now investigate the PC-lite dynamics \eqref{model} in detail and discuss why it can explain the groupthink behavior. A key antecedent condition for groupthink is isolation of the group members from the outside population. We start off with a simple but important result that highlights the unique capability of the PC-lite model (and also the PC model discussed in the next section) in explaining group polarization and the emergence of popular, extreme beliefs in a network of agents.
\begin{definition}[\bf{Connectedness}]
Given the model \eqref{model}, a subset $\mathcal{B}\subseteq \mathcal{V}, |\mathcal{B}| > 1$, of agents is said to be {\it connected} if
\begin{equation}
\sum_{j \in \mathcal{B}}\sum_{t=0}^\infty w_{ij}(t) =\infty,~\forall i \in \mathcal{B}.
\end{equation}
\end{definition}
A connected subset $\mathcal{B}$ of individuals is not necessarily blended, a property defined as strong connectivity of the graph with nodes $\mathcal{B}$, $|\mathcal{B}|>1$, and edges $(i,j)$ which exist if and only if
\begin{equation}
\sum_{t=0}^\infty w_{ij}(t) =\infty.
\end{equation}
More precisely, a subset is connected if and only if it is blended or can be divided into multiple blended subsets. This means that any result stated later on for connected subsets also applies to blended subsets, which may be of greater interest in the given context of social networks.
\begin{proposition}
\label{prop:polarization_result}
Given the model \eqref{model}, let a connected subset $\mathcal{B}\subseteq \mathcal{V}$ of agents be isolated from outside, i.e., for any $i \in \mathcal{B}$ and $t \geq 0$, assume that
\begin{equation}
\sum_{j \in \mathcal{V} \backslash \mathcal{B}} w_{ij}(t)=0.
\label{prop 1 condition}
\end{equation}
If for some $t_0$, $x_i(t_0) > 0$, $\forall i \in \mathcal{B}$, then we have $\lim_{t \rightarrow \infty}{x_i(t)} =1$, $\forall i \in \mathcal{B}$.
\end{proposition}
\begin{proof}
Proposition~\ref{prop:polarization_result} will turn out to be a special case of Proposition~\ref{polarization_result}, stated later on in this section. See Subsection~\ref{rest of proofs} for more details.
\end{proof}
Proposition~\ref{prop:polarization_result} addresses an exaggerated but important situation in which a set of connected agents is completely isolated from the rest of the population. It states that if the agents in this set are initially in agreement, no matter how weak this agreement is, they reach an extremely strong agreement as time passes. In other words, in the absence of opposing views, given consistent interactions among the agents, group polarization is inevitable.
In the following, we argue that the complete isolation condition for group polarization can be relaxed to a realistic one via the notion of a {\it social bubble} that may be present in a population, defined below based on the notion of a {\it bubble number}.
\begin{definition}[\bf{Bubble number}] Given the model \eqref{model} and a subset $\mathcal{B}\subseteq \mathcal{V}, |\mathcal{B}| > 1$, of agents, the bubble number of $\mathcal{B}$, denoted by $\gamma_\mathcal{B}$, is defined as the largest non-negative constant $\gamma$ that satisfies the following equation for any $i \in \mathcal{B}$:
\begin{equation}
\label{buble.prop1}
\sum_{j \in \mathcal{B}} w_{ij}(t) \geq \gamma_\mathcal{B} \sum_{j \in \mathcal{V} \backslash \mathcal{B}} w_{ij}(t), ~ \forall t.
\end{equation}
The bubble number is well-defined for any $\mathcal{B}$ since \eqref{buble.prop1} is satisfied by an upper-bounded, closed interval in $\mathbb{R}$ containing 0.
\end{definition}
Equation \eqref{buble.prop1} states that the sum of the weights inside $\mathcal{B}$ is at least $\gamma_\mathcal{B}$ times larger than the sum of the weights to agents outside that bubble. Thus, $\gamma_\mathcal{B}$ indicates the isolation level of $\mathcal{B}$ from outside in the sense of opinion influence. The greater the bubble number $\gamma_\mathcal{B}$, the greater the isolation of the members in the subset. It is worth noting that the bubble number is closely related to the so-called \emph{cut ratio} of a weighted graph \cite{godsil2013algebraic}. More precisely, if we sum \eqref{buble.prop1} over all $i\in \mathcal{B}$, we obtain
\begin{equation}
\frac{1}{\gamma_{\mathcal{B}}}\ge \frac{\sum_{i\in \mathcal{B}, j \in \mathcal{V} \backslash \mathcal{B}} w_{ij}(t)}{\sum_{i,j \in \mathcal{B}} w_{ij}(t)},
\end{equation}
where the expression on the right side is the ratio of the sum of the edge weights crossing the cut $\mathcal{B}$ over the sum of the edge weights inside that cut. It is known that the minimum cut ratio over all the cuts can be bounded by the algebraic connectivity of the graph \cite{godsil2013algebraic}. Therefore, one can bound the bubble number in terms of the eigenvalues of the adjacency matrix $W(t)$
\begin{definition}[\bf{Social bubble}]
\label{soc bub}
Given the model \eqref{model}, a subset $\mathcal{B}\subseteq \mathcal{V}, |\mathcal{B}| > 1$, of agents is loosely called a {\it social bubble}, or simply a {\it bubble}, if it has a large bubble number, meaning that it is, to a great extent, isolated from outside influence.
\end{definition}
From the theory of groupthink, it is expected that agents in a connected bubble will intensify cohesiveness (if it exists) in a discussion, seeking stronger agreement within the bubble. According to the following theorem, this phenomenon is very well captured by the PC-lite model for social learning.
\begin{proposition}
\label{polarization_result}
Given the PC-lite dynamics \eqref{model}, let a connected subset $\mathcal{B}\subseteq \mathcal{V}$ of agents have the bubble number
\begin{equation}
\label{low1}
\gamma_\mathcal{B}>3+2\sqrt{2},
\end{equation}
and assume that $\alpha_1$ and $\alpha_2$, where $\alpha_1 < \alpha_2$, are the two positive solutions of the equation
\begin{equation}
\label{alphagamma}
\frac{1+\alpha}{\alpha(1-\alpha)} = \gamma_\mathcal{B}.
\end{equation}
If, for some $t_0$, it happens that
\begin{equation}
\label{alpha1}
x_i(t_0) > \alpha_1,~\forall i \in \mathcal{B},
\end{equation}
then we have
\begin{equation}
\label{alpha2}
\liminf_{t \rightarrow \infty}{x_i(t)} \geq \alpha_2,~\forall i \in \mathcal{B}.
\end{equation}
\end{proposition}
\begin{proof}
Proposition~\ref{polarization_result} will turn out to be a special case of Proposition~\ref{polarization_result_Second_model}, stated later on in the paper. See Subsection~\ref{rest of proofs} for more details.
\end{proof}
The threshold $3+2\sqrt{2}$ in Proposition~\ref{polarization_result} marks the smallest possible $\gamma_\mathcal{B}$ for which \eqref{alphagamma} has two positive real solutions for $\alpha$. It can also be viewed as the threshold that makes the loosely defined notion of a ``social bubble'' in Definition \ref{soc bub} precise. Therefore, Proposition~\ref{polarization_result} implies that if the agents in a bubble reach a certain degree of cohesiveness, that is, are at least $\alpha_1$-confident in advocating in favor of a common position, then their confidence tends to grow higher, beyond degree $\alpha_2$. As the bubble number $\gamma_\mathcal{B}$ increases, $\alpha_1$ and $\alpha_2$ will decrease and increase, respectively, as demonstrated in Figure~\ref{Alpha, model 1}. In limit, as $\gamma_\mathcal{B}$ goes to infinity, $\alpha_1$ reaches 0 while $\alpha_2$ reaches 1, making the case for Proposition~\ref{prop:polarization_result}. For $\gamma_\mathcal{B} \simeq 7.83$, $\alpha_1$ and $\alpha_2$ are 1/2 apart.
It should also be noted that the same can be said about a bubble in which the agents disapprove of a position. Hence, in summary, Proposition~\ref{polarization_result} shows group polarization occurring within a connected bubble if the opinions of the agents in the bubble have reached a certain degree of support/disapproval of any given position at some time $t_0$.
We note that the term social bubble is broader than what is loosely known as {\it echo chamber,} in that a social bubble, unlike an echo chamber, may include individuals with diverse or even opposite beliefs. Proposition~\ref{polarization_result} demonstrates that once a social bubble turns into an ``echo chamber,'' i.e., once the condition \eqref{alpha1} is satisfied, one should expect exaggeration of those beliefs as time grows, that is \eqref{alpha2}.
\begin{figure}[t]
\centering
\includegraphics[width=.5\linewidth]
{ChangesOfAlpha1AndAlpha2WithRespectToGammaB.png}
\caption{Changes of $\alpha_1$ and $\alpha_2$ with respect to $\gamma_B$}
\label{Alpha, model 1}
\end{figure}
\begin{remark}
\label{corollary 1}
Suppose that $\mathcal{V}$ contains at least $m$ connected, pairwise disjoint social bubbles $\mathcal{B}_1, \ldots,\mathcal{B}_m$. Now, depending on whether, for each bubble $\mathcal{B}_k, k=1,\ldots,m$, we have $x_i(t_{0_k}) > \alpha_1,~\forall i \in \mathcal{B}_k$ or $x_i(t_{0_k}) < -\alpha_1,~\forall i \in \mathcal{B}_k$, where $t_{0_k} \geq 0$, we have
\begin{equation}
\liminf_{t \rightarrow \infty}{x_i(t)} \geq \alpha_2,~\forall i \in \mathcal{B}_k, \label{pos}
\end{equation}
or
\begin{equation}
\liminf_{t \rightarrow \infty}{x_i(t)} \leq -\alpha_2,~\forall i \in \mathcal{B}_k, \label{neg}
\end{equation}
respectively. In particular, the asymptotic structure of the bubbles can be represented via one of the $2^{m}$ vectors $s\in \{-1,1\}^m$ such that $s_k=+1$ if \eqref{pos} holds, and $s_k=-1$ if \eqref{neg} holds. Thus, the network can exhibit at least $2^m$ substantially different limiting behaviors.
\end{remark}
\section{Partisan Confidence Model and Its Properties}
\label{Second Model and its properties}
Acting toward opinions with a bias has been well documented in confirmation bias theory \cite{breuning2020neurochemistry}. {\color{magenta}A}gents tend to respond with a bias toward information inconsistent with their own information, beliefs, and old experiences. Also, agents tend to willingly ignore some nonconforming information and opinions only to fit into their social groups \cite{allen2020mental}. In summary, this bias can be due to receiving information that inconsistent or in conflict with one's social norm or identity
In this section, we introduce the Partisan Confidence (PC) model, which is a generalization of the PC-lite model \eqref{model} that accounts for the the agents' confirmation bias. As we discussed earlier, the PC-lite model \eqref{model} can describe the group polarization caused by the groupthink behavior described in Irving L. Janis's seminal work \cite{janis2008groupthink}. To fit that model into Robert S. Baron's more advanced model of groupthink \cite{baron2005so}, we assume that each agent $i$ discounts the influence of contrary views received from any other agent and propose the following opinion dynamics model:
\begin{flalign}
\label{ModifiedModelPart2}
\text{\bf(PC dynamics)}\hspace{1.2in} \Delta x_i= \sum_{j \neq i} \Big[d_i(x_i,x_j) w_{ij} |x_j| (\sgn(x_j) - x_i)\Big],&&
\end{flalign}
where $d_i:[-1,1]^2 \rightarrow [0,1]$ is a discounting function elaborated in the following subsection, before investigating the properties of the PC model. Inclusion of the discounting function in the PC model \eqref{ModifiedModelPart2}, that is, discounting of opposing views, in a sense amplifies the isolation degree of a cohesive bubble. Hence, in view of Proposition~\ref{polarization_result}, one expects that the bubble number threshold for group polarization should now be lower, as is made concrete later on.
\subsection{Discounting Function}
As implied from its title, the discounting function is assumed to always return a number within $[0,1]$; a trivial assumption which will not be repeated but made throughout. Furthermore, in view of the confirmation bias, an agent $i$ is to discount the influence on her of an agent $j$ with a general belief opposite to hers. No discount is expected otherwise, meaning that
\begin{equation}
d_i(x_i,x_j) = 1 ~ \text{if}~ \sgn(x_i)=\sgn(x_j).
\label{d 1}
\end{equation}
We also assume that the discount value for opposing general beliefs is at all times upper bounded as
\begin{equation}
d_i(x_i,x_j) \leq \hat{d}_i(|x_i|)~\text{if}~\sgn(x_i)\neq\sgn(x_j),
\label{relaxed d}
\end{equation}
where $\hat{d}_i:[0,1]\rightarrow [0,1]$ is an arbitrary non-increasing function. It should be noted that the non-increasing assumption on $\hat{d}_i$ is reasonable, as it implies that confirmation bias increases with confidence. While our analysis shall remain valid for any discounting function satisfying \eqref{d 1} and \eqref{relaxed d} for a non-increasing $\hat{d}_i$, to shed some light on the PC model, we consider the following candidate for $\hat{d}_i$:
\begin{equation}
\hat{d}_i(|x_i|) = 1-(1-d)|x_i|^{\beta}
\label{discounting function- alpah and beta}
\end{equation}
where $d$ and $\beta$ are constants satisfying $0 \leq d \leq 1$ and $\beta > 0$. This means that the condition \eqref{relaxed d} is now translates to \begin{equation}
d_i(x_i,x_j) \leq 1-(1-d)|x_i|^{\beta}~\text{if}~\sgn(x_i)\neq\sgn(x_j).
\label{bound for general discounting function}
\end{equation}
In what follows, the interpretation of the parameters $d$ and $\beta$, along with the justification of the upper bound assumption, are given. The case for a general discounting function satisfying \eqref{d 1} and \eqref{relaxed d} will be discussed in the very end of the section.
Let us start with the reason why \eqref{bound for general discounting function} only imposes an upper bound on the discounting function instead of assuming an exact formulation. We believe that any exact formulation is too restrictive and unrealistic in a social network setting. An upper bound, with two degrees of freedom in $d$ and $\beta$, allows for a great deal of uncertainty and agents variability in the model, which means the results derived based upon the PC dynamics remain credible in a practical setting. It also addresses the case where the discounting function also varies over time, that is if $d_i$ is a function of $t$ besides $x_i$ and $x_j$.
With the exact discounting function approach ruled out, one wonders why upper-bounding is selected for approximating the discounting function among various possible non-exact formulations. The answer to that lies in the fact that the issue in hand is group polarization, which is reasonably expected to strengthen with the strength of the discount of opposing beliefs. Thus, if some group polarization result is valid for a given discounting function, a group polarization result at least as strong should hold for discounting functions with less values.
We now discuss the properties of the upper bound function in \eqref{bound for general discounting function}, that is $1-(1-d)|x_i|^{\beta}$. First of all, for neutral agents, i.e., when $x_i \rightarrow 0$, it returns 1, which allows for the continuity with respect to $x_i$ of the broader $d_i$ characterized via \eqref{d 1} and \eqref{bound for general discounting function}. This is a very important property to satisfy if $d_i$ is to be realistic in any shape or form. Then, we focus on how $d$ and $\beta$, earlier branded as the degrees of freedom in the upper bound function, are interpreted. We first notice that $1-(1-d)|x_i|^{\beta}$ is non-decreasing in both $d$ and $\beta$. The parameter $d$ can be viewed as a uniform discount factor when $\beta$ is small. It also serves as an upper bound for the discount value employed by the extremely confident individuals, i.e., those with opinions close to 1 in absolute value. If $d=1$, the PC model converts to the PC-lite model. The parameter $\beta$ can be viewed as the discount's decay rate with respect to $|x_i|$. In other words, it captures the contribution of an individual's confidence to her discount value of opposing beliefs. The case where $\beta \rightarrow \infty$ turns the PC dynamics to its PC-lite counterpart.
\subsection{Properties of the PC Model}
We now aim to investigate the behavior of the PC dynamics \eqref{ModifiedModelPart2}, with the discounting function $d_i$ characterized through \eqref{d 1} and \eqref{bound for general discounting function}. Special cases of \eqref{bound for general discounting function}, corresponding to marginal values of $d$ and $\beta$, are of particular interest to better understand the behavior of a general discounting function under conditions \eqref{d 1} and \eqref{bound for general discounting function}, and later a more general discounting function only restricted by \eqref{d 1}. As discussed earlier, either case of $d=1$ and $\beta \rightarrow \infty$ eliminates the confirmation bias and consequently simplifies the PC dynamics to the PC-lite dynamics, which was thoroughly investigated in Subsection~\ref{Properties of the First Model}. The marginal case $d=0$ will be later in Remark~\ref{d=0} argued to transpire the ``ultimate'' group polarization, where the opinions within a social bubble reach one of the very most extreme values $\pm 1$. The last marginal case, which will prove to be both interesting and informative, is that of $\beta \rightarrow 0$, that allows for an infinitely fast decay in the discount value with respect to $|x_i|$ and in limit amounts to a uniform upper bound on the discount value of opposing beliefs, i.e.,
\begin{equation}
d_i(x_i,x_j) \leq d ~\text{if}~\sgn(x_i)\neq\sgn(x_j).
\label{d uniform}
\end{equation}
In this case, Proposition~\ref{polarization_result} can be generalized as follows.
\begin{proposition}
\label{polarization_result_Second_model}
Given the PC dynamics \eqref{ModifiedModelPart2}, with the discounting function $d_i$ satisfying \eqref{d 1} and \eqref{d uniform}, let a connected subset $\mathcal{B}\subseteq \mathcal{V}$ of agents have the bubble number
\begin{equation}
\label{low2}
\gamma_\mathcal{B}>(3+2\sqrt{2})d,
\end{equation}
and assume that $\alpha_1$ and $\alpha_2$, where $\alpha_1 < \alpha_2$, are the two positive solutions of the equation
\begin{equation}
\label{alphagamma2}
\frac{1+\alpha}{\alpha(1-\alpha)} = \frac{\gamma_\mathcal{B}}{d}.
\end{equation}
If, for some $t_0$, it happens that
\begin{equation}
\label{alpha1_Second_model}
x_i(t_0) > \alpha_1,~\forall i \in \mathcal{B},
\end{equation}
then we have
\begin{equation}
\label{alpha2_Second_model}
\liminf_{t \rightarrow \infty}{x_i(t)} \geq \alpha_2,~\forall i \in \mathcal{B}.
\end{equation}
\end{proposition}
\begin{proof}
Proposition~\ref{polarization_result_Second_model} will turn out to be a special case of Proposition~\ref{semi-general PC result}, stated later on in this section. See Subsection~\ref{rest of proofs} for more details.
\end{proof}
Just like Proposition~\ref{polarization_result}, Proposition~\ref{polarization_result_Second_model} also implies that if all agents within a social bubble reach a certain level of advocacy $+\alpha_1$ or disapproval $-\alpha_1$ of a position, as the interactions continue, agents in that relaxed bubble will reach a more extreme level of advocacy $+\alpha_2$ or disapproval $-\alpha_2$ of that position. Therefore, opinions are intensified and become more extreme or, equivalently, the opinions become polarized within that group. Consequently, a group polarization will occur in the direction of support/disapproval of a specific position. One also notices that as $d$ decreases, $\alpha_1$ and $\alpha_2$ will decrease and increase, respectively, as can be seen in Figure ~\ref{Alpha, model 2}. Therefore, if the conditions of Theorem 2 are satisfied in a social bubble, the result will be that agents with relatively low initial levels of advocacy/disapproval on a position will later have relatively extreme levels of advocacy/disapproval on that position.
\begin{figure}[t]
\centering
\includegraphics[width=.5\linewidth]
{ChangesOfAlpha1AndAlpha2WithRespectToGammaBAndD.png}
\caption{Changes of $\alpha_1$ and $\alpha_2$ with respect to $\gamma_B$ and $d$.}
\label{Alpha, model 2}
\end{figure}
For the purpose of completeness, we generalize Proposition~\ref{polarization_result_Second_model} to the following proposition, which addresses group polarization under the PC dynamics for general $d$ and $\beta$.
\begin{proposition}\label{semi-general PC result}
Given the PC dynamics \eqref{ModifiedModelPart2}, with the discounting function $d_i$ satisfying \eqref{d 1} and \eqref{bound for general discounting function}, let a connected subset $\mathcal{B}\subseteq \mathcal{V}$ of agents have the bubble number $\gamma_\mathcal{B}$. Assume that equation
\begin{equation}
\label{alphagamma3}
\frac{1+\alpha}{\alpha(1-\alpha)} = \frac{\gamma_\mathcal{B}}{1-(1-d)\alpha^{\beta}}
\end{equation}
has two positive solutions for $\alpha \in (0,1)$, namely $\alpha_1$ and $\alpha_2$, where $\alpha_1 \leq \alpha_2$. If, for some $t_0$, it happens that
\begin{equation}
x_i(t_0) > \alpha_1,~\forall i \in \mathcal{B},
\end{equation}
then we have
\begin{equation}
\liminf_{t \rightarrow \infty}{x_i(t)} \geq \alpha_2,~\forall i \in \mathcal{B}.
\end{equation}
\end{proposition}
\begin{proof}
As it will be argued in Subsection~\ref{rest of proofs}, Proposition~\ref{semi-general PC result} is a special case of Theorem~\ref{general PC result}, stated later on in this section and proved in great detail in Subsection~\ref{main proof}.
\end{proof}
The statement of Proposition~\ref{semi-general PC result} is different from those of Propositions~\ref{polarization_result} and \ref{polarization_result_Second_model} in that a succinct condition, such as \eqref{low1} and \eqref{low2}, under which \eqref{alphagamma3} is guaranteed to have solutions has not been provided. However, given any $d$ and $\beta$, it is straightforward to verify whether such solutions exist. We should also point out that a larger $\gamma_\mathcal{B}$, smaller $d$, and smaller $\beta$, all work in favor of \eqref{alphagamma3} having solutions for $\alpha$.
\begin{remark}\label{d=0}
In view of Proposition~\ref{semi-general PC result}, the marginal case $d=0$ can be interpreted to represent the ``ultimate'' group polarization. More precisely, given $d=0$, \eqref{alphagamma3} will have two positive solutions, $\alpha_1 < 1$ and $\alpha_2 = 1$, with the latter solution indicating the convergence of the opinions within the bubble to one of the very most extreme values $+1$ or $-1$.
\end{remark}
The two degrees of freedom incorporated in the discounting function \eqref{discounting function- alpah and beta} can indicate sensitivity to an issue being discussed among the individuals. The higher the sensitivity to the issue for a specific population, the more intense the population acts in a biased manner towards it (smaller $d$ or $\beta$), the more probable/intense the polarization of opinions on that issue
Finally, Proposition~\ref{semi-general PC result} can be extended as follows to any discounting function restricted to conditions \eqref{d 1} and \eqref{relaxed d} for a non-increasing function $\hat{d}_i$.
\begin{theorem}\label{general PC result}
Given the PC dynamics \eqref{ModifiedModelPart2}, with the discounting function $d_i$ satisfying \eqref{d 1} as well as \eqref{relaxed d} for a non-increasing $\hat{d}_i:[0,1]\rightarrow [0,1]$, let a connected subset $\mathcal{B}\subseteq \mathcal{V}$ of agents have the bubble number $\gamma_\mathcal{B}$. Assume that inequality
\begin{equation}
\label{alphagamma4}
\frac{1+\alpha}{\alpha(1-\alpha)} < \frac{\gamma_\mathcal{B}}{\hat{d}_i(\alpha)}
\end{equation}
is satisfied for any $i$ and $\alpha \in (\alpha_1,\alpha_2) \subseteq (0,1)$. If, for some $t_0$, it happens that
\begin{equation}\label{ass alpha1}
x_i(t_0) > \alpha_1,~\forall i \in \mathcal{B},
\end{equation}
then we have
\begin{equation}\label{st for alpha2}
\liminf_{t \rightarrow \infty}{x_i(t)} \geq \alpha_2,~\forall i \in \mathcal{B}.
\end{equation}
\end{theorem}
\begin{proof}
The proof of Theorem~\ref{general PC result} is given in Subsection~\ref{main proof}
\end{proof}
\section{Numerical Experiments}
\label{Numerical Examples}
In this section, we illustrate the behaviors of PC-lite dynamics \eqref{model} and PC dynamics \eqref{ModifiedModelPart2} through numerical examples. In all examples, a fixed Erd\"os–R\'enyi random graph \cite{erdHos1960evolution} embodies the underlying graph of the network. It consists of $|\mathcal{V}|=500$ nodes and, for each pair of nodes $i,j \in \mathcal{V}$, an edge $e_{ij}$ exists with the uniform probability $p_{G}=0.06$, independently of other edges. Each edge is then independently activated at any time step with the uniform probability $p_L = 0.8$, which results in asynchronous interactions in the network. Numerical examples of the PC-lite model and the PC model are provided in subsections~\ref{Numerical Examples of model 1} and \ref{Numerical Examples of model 2} below, respectively.
\subsection{Numerical Examples for the PC-lite Model}
\label{Numerical Examples of model 1}
For the numerical examples of PC-lite dynamics \eqref{model}, we consider two cases, (i) a case where there are three bubbles within the population, while some agents do not belong to any of these bubbles, and (ii) a case where the entire population is divided into two bubbles. Other parameters used in the simulations of these two cases, including the size of the bubbles, their bubble numbers, and their respective $\alpha_1$ and $\alpha_2$ values, are given in Tables~\ref{table 1} and \ref{table 2}. For both cases, the initial opinions of the agents within each bubble are selected according to a normal probability distribution function with near-zero mean and low variance, as specified in Tables~\ref{table 1} and \ref{table 2} with $\mu$ and $\sigma^2$ representing the corresponding mean and variance, truncated to the range $[-1,1]$. The initial opinions of the agents outside the bubbles in the first case are selected according to a normal probability distribution function with zero mean and variance equal to 0.11. Finally, we note that the weight values at any time step are generated randomly but scaled in such a way not to violate the bubble numbers listed in Tables~\ref{table 1} and \ref{table 2}. More precisely, given an agent inside a bubble, the weights indicating the influence over her from outside are uniformly scaled down.
\begin{table}[h!]
\centering
\caption{Parameters of the PC-lite model simulation (Figure~\ref{PC-lite A})}
\begin{tabular}{cccc}
\hline
Parameter & Bubble 1 & Bubble 2 & Bubble 3 \\
\hline
$|\mathcal{B}|$ & 159 & 127 & 90\\
$\gamma_\mathcal{B}$ & 8 & 6 & 12 \\
$\alpha_1$ & 0.18 & 0.333 & 0.102 \\
$\alpha_2$ & 0.695 & 0.5 & 0.814\\
$\mu$ & 0.04 & -0.02 & -0.01 \\
$\sigma^2$ & 0.1 & 0.09 & 0.08\\
\hline
\end{tabular}
\label{table 1}
\end{table}
\begin{table}[h!]
\centering
\caption{Parameters of the PC-lite model simulation (Figure~\ref{PC-lite B1})}
\begin{tabular}{ccc}
\hline
Parameters & Bubble 1 & Bubble 2 \\
\hline
$|\mathcal{B}|$ & 261 & 239\\
$\gamma_\mathcal{B}$ & 9 & 6\\
$\alpha_1$ & 0.15 & 0.333 \\
$\alpha_2$ & 0.738 & 0.5\\
$\mu$ & 0.06 & -0.08 \\
$\sigma^2$ & 0.25 & 0.3 \\
\hline
\end{tabular}
\label{table 2}
\end{table}
Figure~\ref{PC-lite A} demonstrates the evolution of opinions under the PC-lite dynamics for the first case. It confirms the statement of Proposition~\ref{polarization_result} that if the opinions of the agents in Bubble~1 become greater than $\alpha_1$ in finite time, they will in the long run become more extreme than $\alpha_2$. The same conclusion can be drawn for Bubble~2 and Bubble~3. Furthermore, Bubble~3 appears to show a stronger group polarization than Bubble~1 and Bubble~2, which is consistent with it having a larger $\alpha_2$ than the other bubbles, that in view of Figure~\ref{Alpha, model 1} is a result of its relatively large bubble number
\begin{figure}[t]
\centering
\includegraphics[width=0.5\linewidth]
{PCLiteDynamicsSpecifiedInTable1.png}
\caption{Opinion evolution according to PC-lite dynamics \eqref{model} with parameters specified in Table~\ref{table 1}
}
\label{PC-lite A}
\end{figure}
Figure~\ref{PC-lite B1} shows the evolution of opinions under the PC-lite dynamics for the second case. While it confirms the statement of Proposition~\ref{polarization_result} like the previous simulation, it is designed to resemble the opinion polarization of the US population depicted in Figure~\ref{US polarization}. More specifically, it shows how the medians of the opinions in the two bubbles diverge over time. The distribution of opinions at time steps 0 and 200 are separately drawn in Figure~\ref{PC-lite B3}, bearing a resemblance to Figure~\ref{US polarization}.
\begin{figure}[t]
\centering
\includegraphics[width=0.5\linewidth]
{PCLiteDynamicsSpecifiedInTable2.png}
\caption{Opinion evolution according to PC-lite dynamics \eqref{model} with parameters specified in Table~\ref{table 2}
}
\label{PC-lite B1}
\end{figure}
\begin{figure}[t]
\centering
\begin{subfigure}[h]{0.4\textwidth}
\centering
\includegraphics[width=0.95\linewidth]
{OpinionDistributionInEachBubbleA.png}
\label{normal snapshot a}
\caption{}
\end{subfigure}
%
\hspace{10px}
%
\begin{subfigure}[h]{0.4\textwidth}
\centering
\includegraphics[width=0.95\linewidth]
{OpinionDistributionInEachBubbleB.png}
\label{normal snapshot b}
\caption{}
\end{subfigure}
%
\hspace{10px}
%
\caption{Opinion distribution in each bubble appearing in Figure~\ref{PC-lite B1} at
(a) $t=0$ and (b) $t=200$.}
\label{PC-lite B3}
\end{figure}
\subsection{Numerical Examples for the PC Model}
\label{Numerical Examples of model 2}
For the numerical examples of PC dynamics \eqref{ModifiedModelPart2}, we consider a case where the network is divided into two bubbles, the parameters of which are given in Table~\ref{table 3}. Figure~\ref{PC fig} demonstrates the evolution of opinions under the PC dynamics \eqref{ModifiedModelPart2} where
\begin{equation}
d_i(x_i,x_j) = \begin{cases}
1 & \text{if~}\sgn(x_i)= \sgn(x_j)\\
1-(1-d)|x_i|^{\beta}& \text{if~}\sgn(x_i)\neq \sgn(x_j)
\end{cases}
\end{equation} with $d=0.4$ and $\beta=0.4$. One can observe in Figure~\ref{PC fig} that all the opinions of agents in the bubbles will asymptotically exceed their respective $\alpha_2$'s in magnitude, confirming Proposition~\ref{polarization_result_Second_model}.
\begin{table}[h!]
\centering
\caption{Parameters of the PC model simulation (Figure \ref{PC fig})}
\begin{tabular}{ccc}
\hline
Parameters & Bubble 1 & Bubble 2 \\
\hline
$|\mathcal{B}|$ & 226 & 274\\
$\gamma_\mathcal{B}$ & 3.5 & 4\\
$\alpha_1$ & 0.3721 & 0.2869\\
$\alpha_2$ & 0.6287 & 0.7149 \\
$\mu$ & 0.09 & -0.08 \\
$\sigma^2$ & 0.1 & 0.11\\
\hline
\end{tabular}
\label{table 3}
\end{table}
\begin{figure}[ht!]
\centering
\includegraphics[width=0.5\linewidth]
{PCDynamicsSpecifiedInTable3.png}
\caption{Opinion evolution according to PC dynamics \eqref{ModifiedModelPart2} with parameters specified in Table~\ref{table 3}.}
\label{PC fig
\end{figure}
\section{Conclusion}
\label{Conclusion}
In this paper, we proposed the PC-lite and PC models of opinion dynamics based on an approach that views an opinion via its intensity and its direction. We established a result on the occurrence of opinion polarization in a social bubble, referring to a group of individuals who are highly cohesive and isolated from outside influence. Both of the models developed are inspired by the notion of groupthink, widely studied in the socio-psychological literature. We also justified our results using numerical simulations.
The ultimate goal of the proposed models is to analyze, predict, and possibly intervene in the process of group polarization and opinion polarization in a population. While the analysis and prediction goals were discussed in this work, the intervention techniques will remain as part of future work. As another future research direction, it will be interesting to study the multidimensional version of the proposed models that captures the simultaneous evolution of opinions on a multitude of correlated topics.
\section{Proofs}\label{Proofs}
This section is composed of a subsection containing some preliminary definitions and lemmas that is integral in the proof of Theorem~\ref{general PC result}, an entire subsection detailing the proof of Theorem~\ref{general PC result}, and another subsection on the derivations of Propositions~\ref{prop:polarization_result}, \ref{polarization_result}, \ref{polarization_result_Second_model}, and \ref{semi-general PC result} from Theorem~\ref{general PC result}.
\subsection{Preliminaries to the Proof of Theorem~\ref{general PC result}} \label{preliminary proof}
We recall that the discounting function in Theorem~\ref{general PC result} is assumed to satisfy \eqref{d 1} and \eqref{relaxed d} for non-increasing functions $\hat{d}_i$. Since the marginal case of \eqref{relaxed d} will prove to be of great importance, we define an auxiliary discounting function $d':[-1,1]^2 \rightarrow [0,1]$ by
\begin{equation}
d'_i(x_i,x_j) = 1 ~ \text{if}~ \sgn(x_i)=\sgn(x_j),
\label{d' 1}
\end{equation}
\begin{equation}
d'_i(x_i,x_j) = \hat{d}_i(|x_i|)~\text{if}~\sgn(x_i)\neq\sgn(x_j),
\label{relaxed d equality}
\end{equation}
where \eqref{d' 1} is identical to \eqref{d 1}, while \eqref{relaxed d equality} is the marginal case of \eqref{relaxed d}. For any $y \in [-1,1]^n$, it should be clear that
\begin{equation}
d_i(y_i,y_j) \leq d'_i(y_i,y_j),~\forall i,j.
\label{dd'}
\end{equation}
We also define a function $f:[-1,1]^n \rightarrow [-1,1]^n$ with its $i$th coordinate formulated as
\begin{equation}
f_i(y)\triangleq y_i+\sum_{j \neq i} \Big[ d'_i(y_i,y_j) w_{ij}|y_j|\left( \sgn{(y_j)}-y_i \right) \Big],
\label{f def}
\end{equation}
There is a slight abuse of notation in \eqref{f def}, in that $w_{ij}$ is in general a function of time, while $f$ does not seem to be treated as one. This will not cause a problem since the time index will be fixed whenever $f_i$ will show up in future arguments. In particular, $f_i(x(t))$ can be expressed as
\begin{equation}
f_i(x(t))= x_i(t) + \sum_{j \neq i} \Big[d'_i(x_i(t),x_j(t)) w_{ij}(t) |x_j(t)| (\sgn(x_j(t)) - x_i(t))\Big].
\label{f dyn}
\end{equation}
In contrast, according to \eqref{ModifiedModelPart2},
\begin{equation}
x_i(t+1) = x_i(t) + \sum_{j \neq i} \Big[d_i(x_i(t),x_j(t)) w_{ij}(t) |x_j(t)| (\sgn(x_j(t)) - x_i(t))\Big].
\label{pc alt}
\end{equation}
The following lemmas will be used in the proof of Theorem~\ref{general PC result}.
\begin{lemma}\label{f dyn pc alt}
For an arbitrary agent $i$, if $x_i(t) \geq 0$, then $x_i(t+1) \geq f_i(x(t))$.
\end{lemma}
\begin{proof}
From \eqref{f dyn} and \eqref{pc alt},
\begin{equation}
x_i(t+1) - f_i(x(t)) = \sum_{j \neq i} \Big[\big(d_i(x_i(t),x_j(t))-d'_i(x_i(t),x_j(t))\big) w_{ij}(t) |x_j(t)| (\sgn(x_j(t)) - x_i(t))\Big].
\label{diff}
\end{equation}
To complete the proof, it is sufficient to show that each summand (term appearing in a summation) in \eqref{diff} is non-negative, which can be done simply by considering the two cases for $\sgn(x_j(t))$. If $\sgn(x_j(t)) = +1$, given the assumption $x_i(t) > 0$, the summand becomes zero since both $d_i$ and $d'_i$ would equal 1. If $\sgn(x_j(t)) = +1$, then from \eqref{dd'}, the summand is non-negative.
\end{proof}
\begin{lemma}\label{increasing_model}
The function $f$ is non-decreasing, i.e., for any pair of vectors $y^1,y^2 \in [-1,1]^n$,
\begin{equation}
\label{inc}
y^1 \leq y^2 \Rightarrow f(y^1) \leq f(y^2),
\end{equation}
where the inequalities in \eqref{inc} are to be understood element-wise
\end{lemma}
\begin{proof}
Without loss of generality, we can assume that $y^1$ and $y^2$ differ only in one coordinate, say the $k$th coordinate. For notational convenience, we may write $y_i$ for both $y^1_i$ and $y^2_i$, when $i \neq k$. We show that $f_i(y^1) \leq f_i(y^2)$ for each $i$ by considering the following cases:
{\bf Case 1:} $(i \neq k)$ In this case,
\begin{align}\label{case 1}
f_i(y^2) - f_i(y^1)
& =\left(y_i + \sum_{j\neq i} \Big[d_i(y_i,y^2_j)w_{ij} |y^2_j|(\sgn(y^2_j) - y_i)\Big]\right)\cr
& -\left( y_i + \sum_{j \neq i} \Big[ d_i(y_i,y^1_j)w_{ij} |y^1_j|(\sgn(y^1_j) - y_i)\Big]\right)\cr
& =w_{ik}\Big(d_i(y_i,y^2_k)|y^2_k|(\sgn(y^2_k) - y_i)- d_i(y_i,y^1_k)|y^1_k|(\sgn(y^1_k) - y_i)\Big)
\end{align}
Now, if $\sgn(y^1_k) = \sgn(y^2_k)$, then from \eqref{d 1} and \eqref{relaxed d equality} we have $d_i(y_i,y^1_k) = d_i(y_i,y^2_k)$, which combined with \eqref{case 1} leads to
\begin{align}\label{case 1 cont}
f_i(y^2) - f_i(y^1)
& =w_{ik} d_i(y_i,y^1_k)\Big(|y^2_k|(\sgn(y^2_k) - y_i)- |y^1_k|(\sgn(y^1_k) - y_i)\Big)\cr
& =w_{ik}d_i(y_i,y^1_k) \Big( (y^2_k - y^1_k) - y_i (|y^2_k|-|y^1_k|) \Big)\cr
& \geq w_{ik}d_i(y_i,y^1_k)(1-|y_i|)(y^2_k-y^1_k)\cr
& \geq 0,
\end{align}
where in the first inequality of \eqref{case 1 cont}, we used the inequality $\big||y^2_k|-|y^1_k|\big|\leq |y^2_k-y^1_k|=y^2_k-y^1_k$. On the other hand, if $\sgn(y^1_k) \neq \sgn(y^2_k)$, given $y^2_k \geq y^1_k$, we have $\sgn(y^1_k) = -1$ and $\sgn(y^2_k) = 1$, which considered together with \eqref{case 1}, immediately results in $f_i(y^2) - f_i(y^1) \geq 0$.
{\bf Case 2:} $(i = k)$ In this case,
\begin{align}\label{lem f}
f_i(y^2) - f_i(y^1)
& = \left(y^2_k + \sum_{j \neq k} \Big[ d_k(y^2_k,y_j)w_{kj} |y_j|(\sgn(y_j) - y^2_k)\Big] \right)\cr
& -\left( y^1_k + \sum_{j \neq k} \Big[ d_k(y^2_k,y_j)w_{kj} |y_j|(\sgn(y_j) - y^1_k)\Big] \right)\cr
& =\left(y^2_k - y^1_k\right) + \sum_{j \neq k} \Big[ w_{kj}|y_j|\Big( d_k(y^2_k,y_j) (\sgn(y_j) - y^2_k) - d_k(y^2_k,y_j) (\sgn(y_j) - y^1_k) \Big)\Big]\cr
& = \left(y^2_k - y^1_k\right) + \sum_{j \neq k} \Big[ w_{kj}|y_j|\Big(d_k(y^1_k,y_j)(y^1_k - y^2_k) + \big(d_k(y^2_k,y_j)-d_k(y^1_k,y_j)\big)(\sgn(y_j)-y^2_k)\Big)\Big].\cr
&~
\end{align}
Considering the two cases $\pm 1$ for $\sgn(y_j)$, and remembering $y^2_k \geq y^1_k$ as well as \eqref{d 1} and \eqref{relaxed d equality}, the term
\begin{equation}
\big(d_k(y^2_k,y_j)-d_k(y^1_k,y_j)\big)(\sgn(y_j)-y^2_k)
\end{equation}
which appears in the last line of \eqref{lem f}, can be easily shown to be zero or positive. Hence, \eqref{lem f} results in
\begin{align}\label{lem f2}
f_i(y^2) - f_i(y^1)
& \geq \left(y^2_k - y^1_k\right) + \sum_{j \neq k} \Big[ w_{kj}|y_j|\Big(d_k(y^1_k,y_j)(y^1_k - y^2_k) \Big)\Big]\cr
& = \left(y^2_k - y^1_k\right) \left( 1 - \sum_{j \neq k} \Big[ w_{kj}|y_j| d_k(y^1_k,y_j) \Big] \right)\cr
& \geq \left(y^2_k - y^1_k\right) \left( 1 - \sum_{j \neq k}w_{kj}\right)\cr
&\geq 0.
\end{align}
\end{proof}
\subsection{Proof of Theorem~\ref{general PC result}} \label{main proof}
Defining $z(t) = \min_{i \in \mathcal{B}} x_i(t)$, the inequality \eqref{st for alpha2} holds for each $i\in\mathcal{B}$ if and only if
\begin{equation}\label{pp1eq}
\liminf_{t \rightarrow\infty}{z(t)} \geq \alpha_2.
\end{equation}
To show \eqref{pp1eq}, we first note that, by the assumption of the theorem, that is \eqref{ass alpha1}, we have
\begin{equation}\label{z1}
z(t_0) > \alpha_1.
\end{equation}
Given an arbitrary but fixed $t$, we then take the following six steps to fully examine $\liminf_{t \rightarrow\infty}{z(t)}$ and show \eqref{pp1eq}.
{\bf Step 1:} We show that the following statement is true:
\begin{equation}\label{z2}
\Big(\alpha_1 < z(t) < \alpha_2\Big) ~~\Longrightarrow~~ \Big(z(t+1) \geq z(t)\Big).
\end{equation}
To this aim, construct a vector $y$ from $x(t)$ as
\begin{equation}
y_i =
\begin{cases}
z(t) & \text{if } i \in \mathcal{B}\\
-1 & \text{if } i \not\in \mathcal{B}.
\end{cases}
\end{equation}
It should be clear that $x(t) \geq y$. Thus, from Lemma~\ref{increasing_model}, we must have $f(x(t)) \geq f(y)$, and in particular, $f_i(x(t)) \geq f_i(y)$, $\forall i \in \mathcal{B}$. On the other hand, since $x_i(t) \geq 0$ for $i \in \mathcal{B}$, from Lemma~\ref{f dyn pc alt}, we conclue that $x_i(t+1) \geq f_i(x(t))$. Hence, $x_i(t+1) \geq f_i(y)$. Therefore, for each $i \in \mathcal{B}$,
\begin{align}\label{arg1}
x_i(t+1)
& \geq f_i(y)\cr
& = y_i + \sum_{j \in \mathcal{V}} d'_i(y_i,y_j)w_{ij}(t) |y_j| (\sgn(y_j) - y_i) \cr
& = y_i + \sum_{j \in \mathcal{B}} d'_i(y_i,y_j)w_{ij}(t) |y_j| (\sgn(y_j) - y_i) \cr
& + \sum_{j \in \mathcal{V} \backslash \mathcal{B}} d'_i(y_i,y_j)w_{ij}(t) |y_j| (\sgn(y_j) - y_i)\cr
& = z(t) + \sum_{j \in \mathcal{B}} w_{ij}(t) z(t)(1- z(t)) \cr
& + \sum_{j \in \mathcal{V} \backslash \mathcal{B}} \hat{d}_i(z(t))w_{ij}(t) (-1-z(t))\cr
& = z(t) + z(t)(1-z(t))\cr
&\times\left[ \sum_{j \in \mathcal{B}} w_{ij}(t) - \frac{\hat{d}_i(z(t))(1+z(t))}{z(t)(1-z(t))}\sum_{j \in \mathcal{V} \backslash \mathcal{B}} w_{ij}(t)\right].
\end{align}
Furthermore, since $\alpha_1 < z(t) < \alpha_2$, from the assumption \eqref{alphagamma4} we have
\begin{equation}
\label{zgammaB}
\frac{\hat{d}_i(z(t))(1+z(t))}{z(t)(1-z(t))} < \gamma_\mathcal{B}.
\end{equation}
Combining \eqref{arg1} and \eqref{zgammaB} implies that
\begin{align}
\label{arg1'}
x_i(t+1)
& \geq z(t) + z(t)(1-z(t))\left[ \sum_{j \in \mathcal{B}} w_{ij}(t) - \gamma_\mathcal{B} \sum_{j \in \mathcal{V} \backslash \mathcal{B}} w_{ij}(t)\right]\cr
& \geq z(t).
\end{align}
Consequently, $z(t+1) \geq z(t)$, which completes the proof of statement \eqref{z2}.
{\bf Step 2:} We show that
\begin{equation}\label{z3}
\Big(z(t) \geq \alpha_2\Big) ~\Rightarrow~ \Big(z(t+1) \geq \alpha_2\Big).
\end{equation}
To prove statement \eqref{z3}, we first point out that since \eqref{alphagamma4} is assumed to hold for any $i$ and $\alpha \in (\alpha_1,\alpha_2)
$, one can write
\begin{equation}\label{alpha2magic}
\hat{d}_i(\alpha_2) \leq \lim_{\alpha \rightarrow \alpha_2}\hat{d}_i(\alpha) \leq \lim_{\alpha \rightarrow \alpha_2}\frac{\alpha_2(1-\alpha_2)\gamma_\mathcal{B}}{1+\alpha_2} = \frac{\alpha_2(1-\alpha_2)\gamma_\mathcal{B}}{1+\alpha_2},
\end{equation}
where the first inequality of \eqref{alpha2magic}, as well as the existence of $\lim_{\alpha \rightarrow \alpha_2}\hat{d}_i(\alpha)$, are both immediate results of $\hat{d}_i$ being non-increasing, while the second inequality of \eqref{alpha2magic} is a direct consequence of \eqref{alphagamma4}. Rewriting \eqref{alpha2magic}, we have
\begin{equation}\label{alpha2magic2}
\frac{1+\alpha_2}{\alpha_2(1-\alpha_2)} \leq \frac{\gamma_\mathcal{B}}{\hat{d}_i(\alpha_2)}
\end{equation}
Now, we follow the same line of arguments as in Step 1, only here we construct the vector $y$ as
\begin{equation}
y_i =
\begin{cases}
\alpha_2 & \text{if } i \in \mathcal{B}\\
-1 & \text{if } i \not\in \mathcal{B},
\end{cases}
\end{equation}
and replace $z(t)$ in \eqref{arg1} by $\alpha_2$. Then, we write \eqref{zgammaB} for $\alpha_2$ in place of $z(t)$ by employing \eqref{alpha2magic2} in place of \eqref{alphagamma4}. These modifications of \eqref{arg1} and \eqref{zgammaB} together imply $x_i(t+1) \geq \alpha_2$, $\forall i \in \mathcal{B}$, which proves statement \eqref{z3}.
{\bf Step 3:} Combining Steps~1 and 2, from \eqref{z1}, \eqref{z2} and \eqref{z3}, we conclude that \eqref{pp1eq} holds unless $z(t)$ is non-decreasing for every $t \geq t_0$ and $\lim_{t\rightarrow\infty} z(t)$ exists and lies in the interval $(\alpha_1,\alpha_2)$. Thus, assume on the contrary that $z(t)$ is non-decreasing for $t \geq t_0$ and
\begin{align}
\label{zlim}
\lim_{t\rightarrow\infty} z(t)=z^*\in (\alpha_1,\alpha_2).
\end{align}
Let $\epsilon > 0$ be sufficiently small that it satisfies
\begin{align}
\label{small espsilon}
\gamma_\mathcal{B} > \frac{\hat{d}_i(z^*-\epsilon)(1+z^*+\epsilon)}{(z^*-\epsilon)(1-(z^*+\epsilon))},
\end{align}
for any $i \in \mathcal{V}$. One notices that \eqref{small espsilon} holds for any sufficiently small $\epsilon$ since, as $\epsilon$ vanishes, the right-hand expression of \eqref{small espsilon} converges to $\hat{d}_i(z^*)(1+z^*)/[z^*(1-z^*)]$, which is less than $\gamma_\mathcal{B}$. According to \eqref{zlim}, there exists $T > t_0$ such that
\begin{align}
z(t) > z^*-\epsilon,~\forall t > T.
\end{align}
{\bf Step 4:} Let $i \in \mathcal{B}$ be arbitrary. We show that there is a time instant $t_i > T$ such that
\begin{align}
\label{t_i}
x_i(t_i) > z^*+\epsilon.
\end{align}
Assume to the contrary that $x_i(t)$ never exceeds $z^*+\epsilon$ after time $T$, that is $x_i(t) \leq z^* + \epsilon$ for any $t > T$. Now, for any $t > T$, given the PC dynamics we have
\begin{align}
\label{znew4}
x_i(t+1)-x_i(t)
& = \sum_{j \in \mathcal{V}} d_i(x_i(t),x_j(t))w_{ij}(t)|x_j(t)|(\sgn(x_j(t))-x_i(t))\cr
& \geq \left(\sum_{j \in \mathcal{B}} w_{ij}(t)\right)(z^*-\epsilon)(1-(z^*+\epsilon))\cr
&+ \hat{d}_i(z^*-\epsilon)\left(\sum_{j \in \mathcal{V}\backslash\mathcal{B}} w_{ij}(t)\right)(-1-(z^*+\epsilon))\cr
&\geq \left(\sum_{j \in \mathcal{B}}w_{ij}(t)\right) \left[ (z^*-\epsilon)(1-(z^*+\epsilon)) + \frac{\hat{d}_i(z^*-\epsilon)}{\gamma_\mathcal{B}} (-1-(z^*+\epsilon)) \right].
\end{align}
Summing up \eqref{znew4} over consecutive time instants, we conclude that
\begin{align}
\label{znew4'}
x_i(t')-x_i(t) \geq \left(\sum_{\tau=t}^{t'-1}\sum_{j \in \mathcal{B}}w_{ij}(\tau)\right) \left[ (z^*-\epsilon)(1-(z^*+\epsilon)) + \frac{\hat{d}_i(z^*-\epsilon)}{\gamma_\mathcal{B}} (-1-(z^*+\epsilon)) \right].
\end{align}
The right-hand expression in \eqref{znew4'} explodes as $t'$ grows since
\begin{align}
\left[ (z^*-\epsilon)(1-(z^*+\epsilon)) + \frac{\hat{d}_i(z^*-\epsilon)}{\gamma_\mathcal{B}} (-1-(z^*+\epsilon)) \right]
\end{align}
is lower-bounded by a positive number according to \eqref{small espsilon} and
\begin{align}
\sum_{\tau=t}^{t'-1}\sum_{j \in \mathcal{B}}w_{ij}(\tau)
\end{align}
grows unbounded as $t'\rightarrow \infty$ since $\mathcal{B}$ is connected. This is a contradiction, meaning that there is $t_i > T$ for which \eqref{t_i} holds.
{\bf Step 5:} Let $i \in \mathcal{B}$ be arbitrary. We show that if $t > T$,
\begin{equation}
\label{znew1}
\Big(x_i(t) \geq z^*+\epsilon\Big) ~~\Longrightarrow~~ \Big(x_i(t+1) \geq z^*+\epsilon\Big).
\end{equation}
According to Lemma~\ref{f dyn pc alt}, $x_i(t+1) \geq f_i(x(t))$. For the arbitrary but fixed $i$, we construct the vector $y$ as
\begin{equation}
y_j =
\begin{cases}
z^*+\epsilon & \text{if } j=i\\
z^*-\epsilon & \text{if } j \in \mathcal{B},~j \neq i\\
-1 & \text{if } i \not\in \mathcal{B}.
\end{cases}
\end{equation}
Since $x(t) \geq y$ and $f$ is non-decreasing according to Lemma~\ref{increasing_model}, $f_i(x(t)) \geq f_i(y)$, and consequently, $x_i(t+1) \geq f_i(y)$. Thus, it is sufficient to show that $f_i(y) \geq z^*+\alpha$. Hence, we write
\begin{align}
\label{znew2}
f_i(y)
& = y_i + \sum_{j \in \mathcal{V}} d'_i(y_i,y_j)w_{ij}(t)|y_j|(\sgn(y_j)-y_i)\cr
& = z^*+\epsilon
+ \left(\sum_{j \in \mathcal{B}} w_{ij}(t)\right)(z^*-\epsilon)(1-(z^*+\epsilon))\cr
&+\hat{d}_i(z^*+\epsilon) \left(\sum_{j \in \mathcal{V}\backslash\mathcal{B}} w_{ij}(t)\right)(-1-(z^*+\epsilon))\cr
& \geq z^* +\epsilon + \left(\sum_{j \in \mathcal{B}} w_{ij}(t)\right)\cr &\times\left[ (z^*-\epsilon)(1-(z^*+\epsilon)) + \frac{\hat{d}_i(z^*+\epsilon)}{\gamma_\mathcal{B}} (-1-(z^*+\epsilon)) \right]\cr
& \geq z^*+\epsilon,
\end{align}
where the last inequality in \eqref{znew2} is a result of
\begin{equation}
\left[ (z^*-\epsilon)(1-(z^*+\epsilon)) + \frac{\hat{d}_i(z^*+\epsilon)}{\gamma_\mathcal{B}} (-1-(z^*+\epsilon)) \right] > 0,
\end{equation} which itself is implied from \eqref{small espsilon} considering $\hat{d}_i(z^*-\epsilon)\leq \hat{d}_i(z^*-\epsilon)$ according to Lemma~\ref{increasing_model}.
{\bf Step 6:} Combining Steps~4 and 5, we conclude that for each $i \in \mathcal{B}$, there is a time $t_i$ such that $x_i(t) \geq z^*+\epsilon$ for any $t \geq t_i$. Hence, $z(t) \geq z^*+\epsilon$ for any $t \geq \max(t_1,\ldots,t_n)$, which contradicts the assumption $\lim_{t\rightarrow\infty} z(t)=z^*$ made in Step~3, completing the proof.\qed
\subsection{Derivations of Propositions~\ref{prop:polarization_result}, \ref{polarization_result}, \ref{polarization_result_Second_model}, and \ref{semi-general PC result}}\label{rest of proofs}
In this subsection, starting from Proposition~\ref{prop:polarization_result}, we demonstrate that each proposition stated in this paper can be derived from the one coming next, while the last proposition, that is Proposition~\ref{semi-general PC result}, is a result of Theorem~\ref{general PC result} proved previously.
Condition \eqref{prop 1 condition} in Proposition~\ref{prop:polarization_result} indicates that $\mathcal{B}$ has an infinite bubble number. Thus, assuming that Proposition~\ref{polarization_result} is true, in view of \eqref{alphagamma}, we obtain $\alpha_1 = 0$ and $\alpha_2 = 1$, immediately resulting in Proposition~\ref{prop:polarization_result}. Setting $d=1$ in Proposition~\ref{polarization_result_Second_model} simply converts it into Proposition~\ref{polarization_result}. Proposition~\ref{polarization_result_Second_model} is a special case of Proposition~\ref{semi-general PC result} where $\beta \rightarrow 0$. Thus, it only remains to derive Proposition~\ref{semi-general PC result} from Theorem~\ref{general PC result}.
The assumption in Proposition~\ref{semi-general PC result} that \eqref{alphagamma3} has two positive solutions $\alpha_1$ and $\alpha_2$ in $(0,1)$ means that for any $\alpha \in (\alpha_1,\alpha_2)$ we have
\begin{equation}
\frac{1+\alpha}{\alpha(1-\alpha)} < \frac{\gamma_\mathcal{B}}{1-(1-d)\alpha^{\beta}}.
\end{equation}
Thus, setting $\hat{d}_i(\alpha) = 1-(1-d)\alpha^\beta$, which is a non-increasing function in $(0,1)$, in Theorem~\ref{general PC result} immediately leads to the statement of Proposition~\ref{semi-general PC result}.
\bibliographystyle{IEEEtran}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 3,879
|
Today's Spiritual Business Success Camp lesson is content marketing.
If you're building a lifestyle brand based on YOU, then your truth is the foundation for your content. Think Oprah. So what you story about you can your audience draw strength, guidance, support or connection from?
Listen HERE now - I share at least 10 examples of what exactly content marketing is for us as Spiritpreneur Warrior Sisters -- coaches, healers, bloggers, artists and artisans.
P.S. Are you on Facebook with us?
P.P.S. Have you registered for the transformational spiritual business sisterhood yet? Join the Spiritpreneur Warrior Business Sisterhood at ManifestYourBusiness.com.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 4,094
|
CMS may refer to:
Computing
Call management system
CMS-2 (programming language), used by the United States Navy
Code Morphing Software, a technology used by Transmeta
Collection management system for a museum collection
Color management system, a system for computers to control the representation of colors
Concurrent mark sweep collector, a garbage collector in the Oracle HotSpot Java virtual machine
Configuration management system
Construction and management simulation, a type of simulation video game
Contact management system, an integrated office solution to record relationships and interactions with customers and suppliers
Content management system, a system for managing content and providing it in various formats
Conversational Monitor System, previously Cambridge Monitor System, an IBM mainframe operating system, also known as VM/CMS and CP/CMS
Course management system, software that facilitates e-learning or computer learning
Credential management system, also known as Smart card management system(SCMS) and Card management system(CMS)
Cryptographic Message Syntax, a cryptographic standard
Medicine
Chronic mountain sickness, or Monge's disease, a disease caused by high altitude
Congenital mitral stenosis
Congenital myasthenic syndrome, an inherited neuromuscular disorder
Organizations
Education
United States
Calexico Mission School, a private school in Calexico, California, USA
Cardigan Mountain School, a junior boarding school in Canaan, New Hampshire, USA
Carpentersville Middle School, a public school in Carpentersville, Illinois, USA
Caruso Middle School, a public school in Deerfield, Illinois, USA
Cedar Middle School, a public school in Cedar City, Utah, USA
Centennial Middle School, a public school in Snohomish, Washington, USA
Charlotte-Mecklenburg Schools, local school district in North Carolina, USA
Chicago Medical School, in North Chicago, Illinois
Chickahominy Middle School, a public school in Mechanicsville, Virginia, USA
Claremont-Mudd-Scripps Stags and Athenas, the joint athletic team of Claremont McKenna College, Harvey Mudd College, and Scripps College in Claremont, California, USA
Clarksville Middle School, a public school in Clarksville, Maryland, USA
Colina Middle School, in Thousand Oaks, California, USA
Colonia Middle School In Colonia, New Jersey, USA
Community Middle School, a public school in Plainsboro, New Jersey, USA
Creekwood Middle School, a public school in Kingwood, Houston, Texas, USA
Crestview Middle School, a middle school in Clarkson Valley, St. Louis County, Missouri, USA
Crest Memorial School (of the Wildwood Crest School District), a public K-8 school in Wildwood Crest, New Jersey, USA
India
Church Mission Society High School, CMS High School, Thrissur, Kerala, India
City Montessori School, in Lucknow, India
CMS College Kottayam, in Kottayam, Kerala, India
Other
Church Mission School, Pakistan
C.M.S. Ladies' College, Colombo, Sri Lanka
Mathematics
Calcutta Mathematical Society or CalMathSoc, a professional society for mathematicians at Kolkata, India
Canadian Mathematical Society, a professional society for mathematicians
Centre for Mathematical Sciences (disambiguation), several organisations
Centre for Mathematical Sciences (Cambridge), the mathematics centre at Cambridge University
Centre for Mathematical Sciences (Kerala), India
Council for the Mathematical Sciences, a forum for mathematical societies in the UK
Cyprus Mathematical Society, a non-profit society in Cyprus
Chinese Mathematical Society, a professional society for mathematicians in China.
Other
Centers for Medicare and Medicaid Services, a federal agency in the United States, part of the Department of Health and Human Services
China Marine Surveillance, the maritime surveillance agency of the People's Republic of China
Church Mission Society, an evangelistic religious organization
CMS Cameron McKenna Nabarro Olswang, an international law firm based in London
CMS Energy, a public utility in Michigan
Colorado Medical Society, a non-profit organization of physicians in Colorado
Court of Master Sommeliers, an organisation educating and certifying sommeliers
Illinois Department of Central Management Services, a state agency
Science
China Manned Space Program, China's human spaceflight program started in 1992
Compact mass spectrometer
Compact Muon Solenoid, a particle physics detector at CERN
Cytoplasmic male sterility, in seed breeding
Other
Cable management system, the managing of electrical or optical cable
Cavity monitoring system, a system for surveying stopes in an underground mine where it is not safe or possible to send a person
Changeable-message sign, a variable road sign
Charlotte Motor Speedway, a speedway in Concord, North Carolina, north of Charlotte
The Chicago Manual of Style, a style guide for American English
Chief master sergeant, a rank in the US Air Force
Central Monitoring System, an Indian clandestine mass electronic surveillance program
Certified Master Safecracker, an American trade qualification
Cms (angel), a minor angel in Enochian occult tradition
Collateral management system, in banking
Community Management Statement, a document for defining the lot entitlements in a Community Title Scheme, a form of Strata title in Queensland, Australia.
Conservation management system, a procedure for maintaining a species or habitat
Constant maturity swap, in finance and investing
Convention on Migratory Species, or the Bonn Convention, an intergovernmental wildlife conservation treaty
Critical management studies, a left-wing approach to management, business and organization
CMS Enhancements, a semi-defunct American computer systems and data storage company
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 2,758
|
\section{Introduction to heavy ion collisions}
\paragraph{Elementary forces in Nature} The interactions among the
elementary constituents of matter are divided into four fundamental
forces: gravitation, electromagnetism, weak nuclear forces and strong
nuclear forces. All these interactions except gravity have a well
tested a microscopic quantum description in terms of local gauge
theories, in which the elementary matter fields are spin-$1/2$
fermions, interacting via the exchange of spin-$1$ bosons. In this
framework, a special role is played by the Higgs spin-$0$ boson (the
only fundamental scalar particle in the Standard Model), whose
non-zero vacuum expectation value gives to all the other fields a mass
proportional to their coupling to the Higgs. The discovery of the
Higgs boson at the Large Hadron Collider in 2012 has so far confirmed
all the Standard Model expectations. In this picture, gravity has
remained a bit of an outlier: even though the classical field theory
of gravitation (general relativity) has been verified experimentally
with a high degree of precision (the latest of these verifications
being the observation of gravitational waves emitted during the merger
of massive compact objects - black holes or neutron stars), the quest
for a theory of quantum gravity has been inconclusive until now (and
possible experimental probes are far out of reach for the foreseeable
future).
\paragraph{Strong nuclear force} Quantum chromodynamics (QCD), the
microscopic theory that governs strong nuclear interactions, was
formulated in the early 1970's under the guidance of several
experimental clues. In particular, deep inelastic scattering of
electrons off proton targets indicated that the proton charge is
concentrated into smaller constituents (unresolved in the scattering)
of spin $1/2$ (this follows from the measured structure functions),
that interact weakly at high momentum transfer. These observations
paved the way towards a non-Abelian gauge theory with the property of
{\sl asymptotic freedom} \cite{Gross:1973id,Politzer:1973fx}, i.e. a
theory in which the coupling strength decreases at short
distance. Combined with some insights from hadron spectroscopy, this
led to an $SU(3)$ gauge theory, with spin-$1/2$ matter fields (the
quarks) in the fundamental representation. The fundamental property of
QCD that resolved the tension between the fact that quarks must
interact strongly enough to form bound hadronic states and the fact
that they appear to be weakly interacting in deep inelastic scattering
experiments is asymptotic freedom: namely, the property that the
running of the coupling due to quantum corrections is such that the
strong coupling constant becomes small at short distance and large on
distance scales compared to the size of a hadron.
We now know that there are six families of quarks: up, down, strange,
charm, bottom and top, ranging from nearly massless to about $175$~GeV
for the top quark \cite{Tanabashi:2018oca}. The nucleons that compose
the atomic nuclei of ordinary matter are built solely from the up and
down quarks, and the heavier quarks appear only in more massive
hadrons (at the exception of the top quark, whose lifetime is so short
that it decays before bound states can be formed). QCD has received
ample experimental support as the correct microscopic theory for
describing strong nuclear interactions. However, because of asymptotic
freedom, the most quantitative comparisons between theory and
experiments are based on hard processes (i.e., processes involving at
least one hard momentum particle in the final state). Although this is
sufficient to ascertain the fact that strong nuclear interactions are
indeed well described by QCD, these experiments leave unexplored
another important aspect of strong interactions, that has to do with
the rich properties of nuclear matter in extreme conditions of
temperature or density.
\paragraph{Asymptotic freedom, confinement and deconfinement} A
crucial property of QCD, consequence of asymptotic freedom, is color
confinement \cite{Wilson:1974sk}, namely the fact that isolated quarks
or gluons cannot exist but instead combine into bound states --the
hadrons-- in which their color charge is ``hidden''. Thus, trying to
pull a quark out of a hadron (for instance in a high energy
collision with another hadron) merely creates more
hadrons. Conversely, when one packs many hadrons in a small volume,
the average distance between their constituents decreases, and
therefore they interact more and more weakly\footnote{Let us clear out
a possible misconception related to this: a small coupling constant
in pairwise parton interactions does not necessarily imply that the
system as a whole is weakly interacting. The latter is true only if
the mean free path is large compared to the De Broglie wavelength of
the constituents, and the inverse mean free path is the product of a
cross-section by a density. In other words, in such a dense system,
there may be strong collective effects despite a weak
coupling.}. Given this, we may expect that the forces that bind
quarks inside hadrons eventually become weak enough to allow the
quarks to become unconfined, i.e., free to wander in the entire volume
of the system. This state of nuclear matter is called the {\sl
quark-gluon plasma} (QGP). Note that this transition is
non-perturbative, since it happens at an energy scale where the
coupling constant is still too large to apply reliably perturbation
theory.
However, it is possible to formulate QCD non-perturbatively by
discretizing Euclidean space-time on a lattice. This setup provides a
way of computing certain observables without resorting to an expansion
in powers of the coupling. Some of the quantities that one may
calculate in lattice QCD are related to the confinement/deconfinement
transition: e.g., the expectation value of the trace of a Wilson loop
(that one may relate to the potential between a pair of infinitely
heavy quark and antiquark), or the entropy density (that measures the
number of active degrees of freedom in the system). Moreover, in
lattice QCD, one may vary several parameters of the theory, like the
number of quark families and their masses, in order to investigate the
role they play in the observed phenomena. Some of these results are
summarized in the plot of Figure \ref{fig:columbia}, taken from
\cite{deForcrand:2017cgb}. For instance, pure-glue QCD (equivalent to
QCD with infinitely massive quarks) has a first order deconfinement
transition at a temperature of the order of $270$~MeV. A first order
transition may also exist in the opposite limit, with massless
quarks. In this limit, the classical QCD Lagrangian also has a chiral
symmetry, spontaneously broken at low temperature (this transition is
also a first order transition).
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{columbia_plot_arrow}
\caption{\label{fig:columbia}Nature of the transition at null
chemical potential as a function of quark masses. From
\cite{deForcrand:2017cgb}.}
\end{figure}
The physical spectrum of light quarks in Nature lies in between these
two extreme situations, and there is now a consensus that this
physical point corresponds to a mere crossover transition, i.e., a
perfectly smooth (but rather rapid) transition from hadrons to
deconfined quarks, that takes place at a temperature around
$155$~MeV. Even in the absence of a genuine transition in the
thermodynamical sense, this transition exhibits signs of deconfinement
accompanied by a restoration of chiral symmetry (up to explicit
residual violations due to the non-zero quark masses).
\paragraph{Lattice QCD and baryon chemical potential} In statistical
equilibrium, the parameters that control the possible states of
nuclear matter are the temperature, the chemical potentials associated
to conserved quantities (the most important of which is the baryon
chemical potential $\mu_{_B}$), and possibly some external fields
(such as magnetic fields). However, a non-zero chemical potential is a
very serious obstacle for lattice QCD. Indeed, the chemical potential
turns the exponential of the action integrated over the quark fields
into a non-positive measure, that cannot be sampled by Monte-Carlo. At
very small $\mu_{_B}/T$, various lattice techniques
\cite{Hasenfratz:1983ba,Alford:1998sd,Hands:1999md,Fodor:2001au,Gupta:2004pk}
can be used to track the transition to non-zero values of $\mu_{_B}$,
but these methods loose their accuracy when the chemical potential
becomes of the order of the temperature. Note that a perturbative
approach becomes possible at very large $\mu_{_B}$ and/or large $T$,
since these parameters control the relevant scale for the running
coupling. At high $T$ and low $\mu_{_B}$, these analytic calculations
are consistent with lattice computations. At high $\mu_{_B}$ and very
low temperature, they indicate the presence of several color
super-conducting phases
\cite{Alford:1997zt,Berges:1998rc,Son:1998uk,Alford:2007xm} (in these
phases, the ground state of the system exhibits a non-zero quark-quark
condensate, very similar to Cooper pairing in BCS super-conductivity).
Outside the regions accessible to lattice simulations or to
perturbation theory, our knowledge of the phase diagram of nuclear
matter is mostly speculative. It is for instance expected that the
system is strongly interacting near the transition line, implying
small transport coefficients, consistent with the measurements of final
state correlations among the produced particles (see later the section
on hydrodynamics).
\paragraph{Heavy Ion Collisions} In the history of the early Universe,
the confinement transition was crossed when the Universe was about one
microsecond old, but as far as we know this did not leave any visible
imprint accessible to present astronomical observations. In the early
1980's emerged the idea to collide heavy nuclei in order to produce in
the laboratory nuclear matter at high temperature and density,
possibly sufficient to reach and go beyond the critical
line. Subsequently, several experiments have had all or part of their
scientific program devoted to the study of heavy ion collisions:
\begin{itemize}
\item \textbf{Bevatron} (Billions of eV Synchrotron) :\\
From 1954 to 1993 at Lawrence Berkeley National Laboratory, U.S
\item \textbf{AGS} (Alternating Gradient Synchrotron) :\\
Since 1960 at Brookhaven National Laboratory, U.S\\
Now used as injector for RHIC
\item\textbf{SPS} (Super Proton Synchrotron) :\\
Since 1976 at CERN\\
Now the injector for the LHC
\item\textbf{SIS-18} (Schwer-Ionen-Synchrotron) :\\
Since 2001 at GSI
\item\textbf{RHIC} (Relativistic Heavy Ion Collider) :\\
Since 2000 at Brookhaven National Laboratory, U.S
\item\textbf{LHC} (Large Hadron Collider) :\\
Since 2009 at CERN
\end{itemize}
The first experimental hints of a deconfinement transition were
observed at the CERN SPS \cite{Heinz:2000bk,Abreu:2000ni,Andersen:1999ym,Agakishiev:1997au,Bearden:1999ck,Bachler:1999hu,Ambrosini:1999im,Aggarwal:1999hk}, that collided heavy ions
at a center of mass energy of $17$~GeV, and in the subsequent
experimental programs at higher energies (the RHIC at Brookhaven
National Laboratory, and the Large Hadron Collider at CERN) the focus
has shifted from assessing the production of a quark-gluon plasma
towards measuring quantitatively some of its properties
\cite{Adams:2005dq,Adcox:2004mh,Arsene:2004fa,Back:2004je,Aamodt:2010pa,Aamodt:2010jd,ALICE:2011ab,Abelev:2012ola,Aad:2010bu,ATLAS:2012at,Aad:2012gla,Khachatryan:2010gv,Chatrchyan:2011sx,CMS:2012qk}.
\paragraph{Experimental handles} In heavy ion collisions, a few
experimental handles are available to vary the conditions in which the
quark gluon plasma may be formed. One of them is the atomic number of
the nuclei used in the collisions, whose main effect is to change the
volume of the interaction zone (but as we shall see in the next
section, this has also an incidence on the so-called saturation
momentum). When performing collisions with a given species of ions,
another variable that has a direct effect on the volume is the impact
parameter of each collision. Although the impact parameter is not
directly measurable, some observable quantities (such as the total
multiplicity in the final state, or the transverse energy) are
strongly correlated with the impact parameter. Finally, the collision
energy can in principle be varied (but of course, this is in practice
highly constrained by the accelerator design), which affects the
initial energy density (i.e., temperature) and the net baryon density
of the matter produced in a collision. This translates in different
reaches in the phase diagram for various heavy ion experiments, as
sketched in Figure \ref{fig:phase-diagram}.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{QCD-phases}
\caption{\label{fig:phase-diagram}Sketch of the phase-diagram of
strongly interacting nuclear matter, and approximate reach of
various experimental facilities.}
\end{figure}
\paragraph{Main stages of a heavy ion collision} From a theoretical
point of view, an ultrarelativistic collision between two nuclei can
be conveniently divided in several stages, sketched in the figure
\ref{fig:stages}.
\begin{figure}[htbp]
\centering
\includegraphics[width=13cm]{stages}
\caption{\label{fig:stages}Main stages of the evolution of the
matter produced in a heavy ion collision. }
\end{figure}
Such a picture stems from the approximate boost invariance (i.e.,
independence on the spatial rapidity variable
$\eta_s\equiv \tfrac{1}{2}\ln((t+z)/(t-z))$) of such collisions, and the
successive stage are ordered by the proper time $\tau$. The collision
itself is very short due to the Lorentz contraction of the nuclei in
the laboratory frame. Just after the collision, the matter produced is
predominantly made of non-equilibrated gluons (they are not even
on-shell at the very beginning, and the system is better treated in
terms of fields rather than particles). This matter is strongly
interacting due to a large gluon occupation number, and evolves
towards equilibration (both kinetic and chemical, since
quark-antiquark pairs are produced in the process). In the subsequent
stages, the bulk evolution of the system is remarkably well described
by nearly ideal (i.e., with very small values of the viscous transport
coefficients) relativistic hydrodynamics. The expansion of the system
causes the system to cool down, and at some point the temperature
reaches the confinement temperature. In the framework of
hydrodynamics, as long as the system remains close to equilibrium, the
crossing of the confinement transition is rather transparent since it
is encoded in the equation of state. Soon after, the system becomes
dilute, the mean free path increases, and a description of its
expansion in terms of kinetic theory rather than hydrodynamics becomes
preferable. In such a description, the values of the various
cross-sections control when each type of reaction stops: the inelastic
processes stop first (chemical freezeout), soon followed by a kinetic
freezeout after which the momenta of the particles remain unchanged
(afterwards, all particles therefore fly on straight lines at constant
velocity until they hit a detector).
\section{Initial state, Color Glass Condensate}
\paragraph{Multiparton interactions and gluon saturation} Let us start
with the very first moments of a heavy ion collision. This is the
realm of the highest momentum scales of the entire collision, and one
may thus expect that this stage is amenable to a perturbative QCD
treatment. The situation is however more complicated. The main
difficulty is the fact that the typical transverse momentum of the
produced particles in such a collision is rather low (around $1$ GeV),
implying that high energy heavy ion collisions probe the partonic
content of the incoming nuclei at very small values of the
longitudinal momentum fraction $x$ (the fraction of the momentum of a
hadron carried by one of its constituents), i.e., in a regime where
the gluon density is large. The consequence of this is that processes
initiated by more than one parton in each nucleus become possible,
which invalidates the usual factorization schemes (based on single
parton densities), as illustrated in Figure \ref{fig:sat-nl}.
\begin{figure}[htbp]
\centering
\includegraphics[width=5.5cm]{AA-dilute}\hskip 10mm
\includegraphics[width=5.5cm]{AA-dense}
\caption{\label{fig:sat-nl}Left: scattering in the dilute
regime. Right: multi-gluon scattering in the saturation regime.}
\end{figure}
One may derive a simple criterion for the onset of these multi-parton
phenomena (known as {\sl gluon saturation}) by combining the estimated
gluon recombination cross-section and the number of gluons per unit of
transverse area
\cite{Gribov:1984tu,Mueller:1985wy,Blaizot:1987nc}. Gluon saturation
happens when the product of these two quantities is greater than one,
which can also be framed as an upper bound for the momentum transfer
$Q$ in deep inelastic scattering (the inverse of this scale plays the
role of a spatial resolution in such a scattering), $Q\lesssim Q_s$,
where $Q_s$ is the so-called saturation momentum. $Q_s$ depends both
on the atomic number of the nuclei, and on the collision energy via
the momentum fraction $x$,
\begin{equation}
Q_s^2\sim A^{1/3} x^{- \lambda},
\end{equation}
where the exponent $\lambda$ has been determined phenomenologically
from deep inelastic data and estimated to be $\lambda\approx
0.25$. The growth at small $x$ follows from the growth of the gluon
density, while the factor $A^{1/3}$ is a measure of the thickness of a
nucleus in the direction of the collision axis. The variations of the
saturation momentum as a function of $A$ and $x$ are shown in Figure
\ref{fig:sat-domain}.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{QxA}
\caption{\label{fig:sat-domain}Saturation domain. From \cite{Deshpande:2009zz}.}
\end{figure}
Recalling that the typical value of $x$ scales as $p_\perp/\sqrt{s}$
where $p_\perp$ is the transverse momentum of produced particles and
$\sqrt{s}$ the energy in the center of mass of a nucleon-nucleon
collision, it appears that the bulk of particle production in heavy
ion collisions at the energy of the LHC is potentially affected by
gluon saturation, which calls for a theoretical framework going beyond
the usual collinear factorization. Indeed, the standard parton
distribution are single-parton densities and do not contain the
necessary information about the multi-parton initial states that
become important in the saturation regime.
\paragraph{Color Glass Condensate} Extending the framework of
collinear factorization by defining multi-parton densities in the same
way as the usual parton distributions is not practical. Instead, one
exploits the fact that gluon saturation is also a regime of large
gluon occupation number, which allows to treat the gluon field as
classical in a first approximation
\cite{McLerran:1993ni,McLerran:1993ka}.
\begin{figure}[htbp]
\centering
\includegraphics[width=\textwidth]{MVmodel11}
\caption{\label{fig:MVmodel}Illustration of the McLerran-Venugopalan model.}
\end{figure}
In such a description, the relevant information about an incoming
nucleus is the color current it carries, that acts as a source for
the color field. Moreover, this current is produced by the partons
that are comparatively fast (in the observer's frame), which implies
that it is nearly time independent thanks to time dilation. The
degrees of freedom in such a description are thus color currents (one
for each projectile) coupled to gluon fields, with an effective action
\begin{align}
{\cal S}_{\rm eff}\equiv
-\frac{1}{4} F^{\mu\nu}_a F_{\mu\nu\,a}+J^\mu_a A_{\mu\,a}.
\end{align}
This setup, known as the {\sl Color Glass Condensate}
\cite{Iancu:2000hn,Iancu:2001ad,Ferreiro:2001qy,Iancu:2002xk,Weigert:2005us,Gelis:2010nm}
(CGC) is illustrated in Figure \ref{fig:MVmodel}. For a fast moving
projectile in the $+z$ direction, the color current has the form
\begin{align}
J^\mu_a(x)=\delta^{\mu+}\rho_a(x^-,{\boldsymbol x}_\perp),
\end{align}
where $\rho_a(x^-,{\boldsymbol x}_\perp)$ is a function that describes the spatial
distribution of the color charges in the object under
consideration. In this expression, we have introduced the light-cone
coordinates, $x^\pm\equiv (t\pm z)/\sqrt{2}$ (the notation
$\delta^{\mu+}$ means that the only non-zero component of the current
is $J^+\propto J^-+J^3$). These coordinates are convenient when
discussing the kinematics of an object moving at the speed of light,
since $x^+$ acts as the time variable for this object (the fact that
the above current does not depend on $x^+$ simply reflects the fact
that this object is time-independent) and $x^-$ as a longitudinal
coordinate as measured in the rest frame of the object. The support of
the $x^-$ dependence of the current is very narrow and centered around
$x^-=0$, due to Lorentz contraction. Note also that the current must
be covariantly conserved, $[D_\mu, J^\mu]=0$. Since the covariant
derivative contains the color field, the color current may be affected
by its own radiated field (the light-cone gauge $A^-=0$ mitigates this
difficulty, since the current $J^+$ can be altered only by the field
$A^-$). The ${\boldsymbol x}_\perp$ dependence of $\rho_a(x^-,{\boldsymbol x}_\perp)$ reflects
the positions in the transverse plane of the color sources at the
instant of the collision (the duration of the collision, controlled by
the thickness of the Lorentz contracted nuclei, is much shorter than
the typical timescales of the internal motions of the constituents of
a hadron -- thus $J^\mu$ needs only to provide a snapshot of the
hadron content). But of course, the configuration of these charges is
not known event-by-event, and the best we may hope to know is a
statistical distribution of these densities, encoded in a functional
$W[\rho]$. This functional density is not something that we can
calculate perturbatively in QCD, since it depends on aspects such as
confinement, the nuclear wavefunction, etc... For a large nucleus,
the McLerran-Venugopalan model \cite{McLerran:1993ni,McLerran:1993ka},
in which $W[\rho]$ is a Gaussian,
\begin{align}
W[\rho]
=
\exp\Big(-\int d^2{\boldsymbol x}_\perp\frac{\rho_a(x^-,{\boldsymbol x}_\perp)\rho_a(x^-,{\boldsymbol x}_\perp)}{2\,\mu^2(x^-, {\boldsymbol x}_\perp)}\Big),
\label{eq:gaussianMV}
\end{align}
is often employed due to its simplicity (in some cases, it even allows
analytical calculations). In this distribution, the mean value of the
charge distribution at a point $(x^-,{\boldsymbol x}_\perp)$ is zero, and
$\mu^2(x^-,{\boldsymbol x}_\perp)$ is a measure of its fluctuations. At lowest
order in the CGC effective theory, this parameter is a placeholder for
the value of the saturation momentum,
\begin{align}
Q_s^2({\boldsymbol x}_\perp)\propto g^2 \mu^2({\boldsymbol x}_\perp)\ln\Big(\frac{\mu^2({\boldsymbol x}_\perp)}{\Lambda_{_{QCD}}}\Big),\quad\mbox{with\ \ }\mu^2({\boldsymbol x}_\perp)\equiv \int dx^-\;\mu^2(x^-,{\boldsymbol x}_\perp).
\label{eq:Qs2mu2}
\end{align}
(This correspondence may be established by calculating the DIS
cross-section in the McLerran-Venugopalan model \cite{Lappi:2007ku}.)
Although a possible heuristic justification for this Gaussian model is
the central limit theorem, thanks to the fact that a large nucleus has
many constituents per unit of transverse area, one should keep in mind
that this distribution of $\rho_a$'s is not derived from first
principles in QCD (since doing so would require to control QCD in a
non-perturbative regime). Another reason why the Gaussian in
eq. (\ref{eq:gaussianMV}) does not have a very fundamental standing is
that this shape is not preserved when one includes one-loop
corrections: indeed, these corrections contain large logarithms of
energy that turn $W[\rho]$ into a non-Gaussian energy-dependent
distribution.
\paragraph{Power counting in the saturated regime} Let us now describe
how a typical CGC calculation is organized. The color glass
condensate may be viewed as a Yang-Mills theory coupled to an external
source \cite{Gelis:2006yv,Gelis:2007kn}, which diagrammatically means
that all graphs contain only gluon propagators. Their endpoints can be
attached to the sources, to gluon vertices, or to the observable of
interest. In the saturation regime, the power counting for these
graphs is a bit peculiar due to the large gluon occupation number. The
3-gluon and 4-gluon vertices are respectively of order $g$ and $g^2$,
while the external source can be as large as $g^{-1}$ (this order of
magnitude is reached when the occupation number reaches its maximal
value, of order $g^{-2}$). Therefore, the order of magnitude of a
generic connected graph ${\cal G}$ is
\begin{align}
{\cal G}\sim g^{-n_{_E}}g^{2n_{_L}} (gJ)^{n_{_J}},
\end{align}
where $n_{_E}$ is the number of external gluons, $n_{_L}$ the number
of loops and $n_{_J}$ the number of sources $J$ in the graph. We see
from this formula that when $J\sim g^{-1}$, the order of the graph
does not depend on the number of sources, implying that there is an
infinity of graphs contributing at each order (for instance, the {\sl
leading order} is made of all the trees). The saturation regime is
therefore a strongly interacting non-perturbative situation, despite
the fact that the coupling constant may be small at high energy (the
typical scale at which the running strong coupling constant should be
evaluated is governed by the saturation momentum).
\paragraph{Leading order} At leading order, the infinite series of
tree diagrams that one needs to sum can always (for inclusive
observables) be expressed in terms of classical solutions of the
Yang-Mills equations,
\begin{align}
\big[D_\mu,F^{\mu\nu}\big]=J^\nu,
\end{align}
with retarded boundary conditions (the retarded nature of the boundary
condition follows from the fact that inclusive measurements do not put
any restriction on the final state). Since in the CGC the incoming
projectiles are completely encoded in the source $J^\mu$, the
classical initial condition is simply to have a null field (or more
generally a pure gauge) in the remote past, before the collision has
happened.
\begin{figure}[htbp]
\centering
\includegraphics[width=5cm]{space-time}
\caption{\label{fig:ST}Decomposition of space-time in four domains,
when solving the classical Yang-Mills equations.}
\end{figure}
By causality, space-time is naturally divided into four domains shown
in Figure \ref{fig:ST}. Domain 0 corresponds to space-time points not
yet reached by any of the nuclei. Domain 1 is causally connected with
the right moving nucleus, but not with the left moving one. For domain
2, it is the opposite. And finally the domain 3 contains the outcome
of the collision. Given a null initial condition in domain 0, the
gauge field can be obtained analytically in the domains 1,2, and also
at the lower boundary of domain 3 (i.e., at a proper time
$\tau=0^+$). At later times in domain 3, no analytical solution is
known, but it is rather straightforward to solve the classical
Yang-Mills equations numerically \cite{Kovner:1995ja,Krasnitz:1998ns,Krasnitz:1999wc,Krasnitz:2000gz,Krasnitz:2001qu,Lappi:2003bi}.
\paragraph{Next to Leading Order} CGC calculations can in principle be
pushed to next-to-leading order, i.e., one-loop. The main difficulty
in doing this is that, like with the leading order, there is an
infinite set of diagrams contributing at NLO. These are all the
one-loop graphs embedded in the external classical gauge field
obtained at LO. For any inclusive observable, there exists an exact
relationship between the LO (tree level) and NLO (one loop) results,
that schematically reads \cite{Gelis:2008rw,Gelis:2008ad}
\begin{align}
{\cal O}_{_{\rm NLO}}
=
\frac{1}{2}\int\frac{d^3{\boldsymbol k}}{(2\pi)^3 2E_{\boldsymbol k}}\int\limits_{x^0=y^0=-\infty} d^3{\boldsymbol x} d^3{\boldsymbol y}\;
\Big[e^{+ik\cdot x}\frac{\delta}{\delta {\cal A}_{\rm ini}(x)}\Big]
\Big[e^{-ik\cdot y}\frac{\delta}{\delta {\cal A}_{\rm ini}(y)}\Big]
{\cal O}_{_{\rm LO}},
\label{eq:LO-NLO}
\end{align}
where ${\cal A}_{\rm ini}$ is the initial condition for the classical
field in the LO calculation.
\begin{figure}[htbp]
\centering
\includegraphics[width=5.5cm]{tree_NLO}
\includegraphics[width=5.5cm]{tree_NLO3}
\caption{\label{fig:resum}Left: typical one-loop graph contributing
to an inclusive observable. Right: interpretation of
Eq.~(\ref{eq:LO-NLO}). In this graphical sketch, the horizontal
direction represents space and the vertical direction represents
time. The red dots represent
the external source $J$.}
\end{figure}
This relationship indicates that the NLO may be obtained from the LO
simply by replacing at two points $x,y$ the classical initial
condition by plane waves $e^{+ik\cdot x}e^{-ik\cdot y}$, integrated
over all on-shell momenta $k$ (but the time evolution continues to be
governed by the classical equation of motion). This is consistent with
the fact that the first quantum correction (order $\hbar^1$) only
affects the initial state, and the first quantum effect that alters
the evolution arises at the order $\hbar^2$. In ordinary quantum
mechanics, this is best seen in the phase-space formulation (also
known as Moyal-Groenewold formulation), in which each quantum
operator ${\boldsymbol A}$ is replaced by its Wigner transform,
\begin{align}
A(X,p)\equiv \int ds \; e^{i\tfrac{p\cdot s}{\hbar}}\,\big<X+\tfrac{s}{2}\big|{\boldsymbol A}\big|X-\tfrac{s}{2}\big>,
\end{align}
defined on the classical phase-space variables $(X,p)$. In this
mapping, the commutator of two such operators ${\boldsymbol A},{\boldsymbol B}$
becomes the {\sl Moyal bracket} of their Wigner transforms:
\begin{align}
\{\!\{A, B\}\!\}(X,p)
\equiv
\frac{2}{i\hbar}
\,
A(X,p)
\sin\Big(\tfrac{\hbar}{2}\big(\stackrel{\leftarrow}{\partial}_{p}\stackrel{\rightarrow}{\partial}_{_X}-\stackrel{\leftarrow}{\partial}_{_X}\stackrel{\rightarrow}{\partial}_{p}\big)\Big)
B(X,p).
\end{align}
If we denote $W$ the Wigner transform of the density operator $\rho$
of a system, the von Neumann equation
$i\hbar \partial_t \rho=[H,\rho]$ becomes
\begin{align}
\partial_t W = \{\!\{{\cal H},W\}\!\}=\underbrace{\{{\cal H},W\}}_{\mbox{\scriptsize Poisson bracket}}+{\cal O}(\hbar^2),
\end{align}
where ${\cal H}$ is the Wigner transform of $H$ (i.e., ${\cal H}$ is
the classical Hamiltonian).
\paragraph{Large logarithms and JIMWLK evolution} When calculating
such one-loop corrections, an aspect of the CGC degrees of freedom
that we have ignored until now becomes important, namely that one must
introduce a cutoff to separate the color sources from the gauge
fields. For instance, such a cutoff should limit the range of
integration over the momentum $k$ in Eq.~(\ref{eq:LO-NLO}). This
separation is based on the longitudinal momentum (or equivalently, the
rapidity $y\equiv\tfrac{1}{2}\ln((k_0+k_z)/(k_0-k_z))$), and is mandatory
when evaluating loops to avoid double countings. In practice, the
longitudinal component of the loop momentum must remain below the
cutoff, because all the higher momentum modes are already included in
the color current. This leads to all one-loop correction to be
sensitive (proportional to the logarithm of the cutoff on longitudinal
momentum) to the cutoff
\cite{Ayala:1995hx,JalilianMarian:1996xn,Iancu:2000hn,Iancu:2001ad,Ferreiro:2001qy,Gelis:2008rw,Gelis:2008ad,Gelis:2008sz}. However,
since this cutoff is an ad-hoc parameter of the CGC effective theory
rather than a physical parameter, it should not enter in physical
observables. This paradox is resolved by the fact that the cutoff
dependence is universal, in the sense that it depends on the nature of
the two colliding projectiles, but not on the inclusive observable one
is measuring. Therefore, it is possible to absorb the cutoff
dependence into a redefinition of the distributions $W[\rho]$ that
define the color source content of the projectiles, turning them into
cutoff dependent objects. For this to be feasible, one should perform
an average of the $\rho$-dependent observable, weighted by the
distributions of $\rho$'s of each projectile,
\begin{align}
\big<{\cal O}\big>\equiv \int [D\rho_1 D\rho_2]\;
W_1[\rho_1]W_2[\rho_2]\,{\cal O}(\rho_1,\rho_2).
\end{align}
The cutoff dependence of $W[\rho]$ is controlled by the so-called
JIMWLK equation, schematically of the form
\begin{align}
\frac{\partial W[\rho]}{\partial \log\Lambda}
=
\frac{\delta}{\delta \rho_a}
\chi_{ab}
\frac{\delta}{\delta \rho_a}
W[\rho],
\end{align}
where $\chi_{ab}$ depends on the LO classical field created by the
source $\rho$ (the possibility to transfer the cutoff dependence from
the observable to the distribution $W[\rho]$ is made possible by the
fact that the operator in the right hand side of the JIMWLK equation
is self-adjoint, via integration by parts). Thus, by evolving the
distributions $W[\rho]$ of each projectile to values of the
longitudinal momentum in the immediate vicinity of the scales relevant
for the observable of interest, one resums all the leading logarithms,
i.e., the powers $(g^2 \log\Lambda)^n$ where each logarithm of the
cutoff is accompanied by a factor $g^2$. This is very similar in spirit
to collinear factorization, the unphysical cutoff $\Lambda$ playing
the role of a factorization scale that should disappear from observables.
Since it is a functional equation, the JIMWLK equation is difficult to
solve, even numerically. The only known approach so far uses the fact
that the JIMWLK equation acts like a diffusion equation in the
functional space of the $\rho$'s (in a treatment more rigorous than
this general discussion, one would use Wilson lines built from the
$\rho$'s rather than the $\rho$'s themselves), and therefore can be
rewritten as a Langevin equation \cite{Blaizot:2002np}. Then, after
discretization of the transverse plane, this stochastic equation is
amenable to a numerical treatment in order to obtain an ensemble of
$\rho$'s evolved to the relevant value of the cutoff
\cite{Rummukainen:2003ns,Lappi:2011ju,Dumitru:2011vk}. Let us also
mention recent improvements: a modification of the Langevin equation
has been proposed to include the effects of a running coupling
\cite{Lappi:2014wya}, and the full NLO corrections to the JIMWLK
equation have also been evaluated
\cite{Balitsky:2013fea,Kovner:2013ona} (but not yet implemented in a
numerical code).
\paragraph{Balitsky-Kovchegov equation} To avoid this computationally
heavy approach, it is also possible to truncate the JIMWLK
equation. The first thing to note is that the functional form of the
JIMWLK equation is equivalent to an infinite sequence of equations for
the correlation functions of Wilson lines constructed from the
$\rho$'s
\begin{align}
U({\boldsymbol x}_\perp) \equiv {\rm T}\,\exp ig\int dx^-\,\tfrac{1}{{\boldsymbol\nabla}_\perp^2}\rho(x^-,{\boldsymbol x}_\perp).
\end{align}
These equations are nested: the equation that drives the cutoff
dependence of the $2$-point correlation function depends on a
$4$-point function, etc... A possible approximation (that may be
justified in the limit of a large number of colors) consists in
factorizing this $4$-point function as a product of two $2$-point
functions, which has the effect of closing the evolution equation of
the latter. The resulting equation, known as the Balitsky-Kovchegov
equation \cite{Balitsky:1995ub,Kovchegov:1999yj}, reads
\begin{align}
\frac{\partial T_{{\boldsymbol x}{\boldsymbol y}}}{\partial \log\Lambda}
=\frac{\alpha_sN_c}{2\pi^2}\int d^2 {\boldsymbol z}_\perp\;
\frac{({\boldsymbol x}_\perp-{\boldsymbol y}_\perp)^2}{({\boldsymbol x}_\perp-{\boldsymbol z}_\perp)^2({\boldsymbol y}_\perp-{\boldsymbol z}_\perp)^2}\,
\Big\{T_{{\boldsymbol x}{\boldsymbol z}}+T_{{\boldsymbol z}{\boldsymbol y}}-T_{{\boldsymbol x}{\boldsymbol y}}
-T_{{\boldsymbol x}{\boldsymbol z}}T_{{\boldsymbol z}{\boldsymbol y}}\Big\},
\end{align}
where
\begin{align}
T_{{\boldsymbol x}{\boldsymbol y}}
\equiv
1- N^{-1}\,{\rm tr}\big<U({\boldsymbol x}_\perp)U^\dagger({\boldsymbol y}_\perp)\big>
\end{align}
(with Wilson lines taken in the fundamental representation of
${su}(N)$). $T_{{\boldsymbol x}{\boldsymbol y}}$ is also proportional to the scattering
amplitude of a quark-antiquark dipole (at the transverse coordinates
${\boldsymbol x}_\perp$ and ${\boldsymbol y}_\perp$, respectively) off a high energy nucleus. In
this equation, the first three terms, linear, correspond to the BFKL
equation, and the last term, non-linear, is a correction due to gluon
saturation, that becomes sizeable when the scattering amplitude
approaches the unitarity limit $T=1$. Since it is an equation for an
ordinary function, the BK equation is much easier to solve
numerically. Note also that the previous equation has now been
improved by running coupling corrections
\cite{Kovchegov:2006vj,Gardi:2006rp}, by the full next-lo-leading log
corrections \cite{Balitsky:2008zza}, and by a resummation of collinear
logarithms \cite{Iancu:2015vea,Lappi:2016fmu}. These improvement has
allowed a successful phenomenology of small-$x$ phenomena in
deep-inelastic scattering and forward proton-proton or proton-nucleus
collisions based on the BK evolution equation
\cite{Albacete:2014fwa,Altinoluk:2014eka,Lappi:2015fma,Iancu:2015joa}.
\section{Pre-equilibrium evolution just after the collision}
\paragraph{Energy-momentum tensor at Leading Order} As we have seen in
the previous section, large nuclei at high energy may be described by
using the CGC framework, in which the large momentum degrees of
freedom are treated as random color currents coupled to the color
field. At leading order in the strong coupling constant, all
expectation values are given by tree diagrams, whose sum is the
classical solution of Yang-Mills equations with a null retarded
boundary condition. Having in mind a description of the subsequent
stages of the collision in terms of relativistic hydrodynamics, it is
therefore natural to calculate the components of the energy momentum
tensor. In the CGC framework, the dominant contribution comes from the
gluons (the valence quarks have a very small contribution at small
$x$, and the sea quark distribution is suppressed by a power of
$\alpha_s$ compared to that of the gluons). In a classical field, they
are given by the following formulas
\begin{align}
&
T^{00}_{_{\rm LO}}
=
\frac{1}{2}\big[\underbrace{{\boldsymbol E}^2+{\boldsymbol B}^2}_{\mbox{\scriptsize class. fields}}\big],
\qquad
T^{0i}_{_{\rm LO}}=\big[{\boldsymbol E}\times {\boldsymbol B}\big]^i,
\nonumber\\
&
T^{ij}_{_{\rm LO}}
=
\frac{\delta^{ij}}{2}\big[{{\boldsymbol E}^2+{\boldsymbol B}^2}\big]
-\big[{\boldsymbol E}^i{\boldsymbol E}^j+{\boldsymbol B}^i{\boldsymbol B}^j\big],
\end{align}
in terms of the chromo-electric and chromo-magnetic fields (note that
this tensor is traceless in classical Yang-Mills theory -- a non-zero
trace would arise from loop corrections via the $\beta$-function, and
from explicit quark masses when quarks are taken into account). Let us
first mention the very special configuration of the ${\boldsymbol E}$ and ${\boldsymbol B}$
fields just after the collision: at $\tau=0^+$, these two fields are
both parallel to the collision axis \cite{Lappi:2006fp}, which leads
to the following form of $T^{\mu\nu}_{_{\rm LO}}(\tau=0^+)$,
\begin{align}
T^{0i}_{_{\rm LO}}=0,\quad T^{11}_{_{\rm LO}}=T^{22}_{_{\rm LO}}=T^{00}_{_{\rm LO}},\quad T^{33}_{_{\rm LO}}=-T^{00}_{_{\rm LO}}.
\end{align}
In other words, the matter is produced at rest, with a negative
longitudinal pressure (i.e., the system resists longitudinal
expansion). Such a negative pressure means that this system should not
be viewed as a collection of on-shell particles, but rather as fields.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{ratio-bkg}
\caption{\label{fig:TLO}Time evolution at Leading Order of the ratios $T^{11}/T^{00}$ and $T^{33}/T^{00}$.}
\end{figure}
At later times (see Figure \ref{fig:TLO}), the longitudinal pressure
progressively builds up to become positive at a time around
$Q_s \tau\sim 1$, which is indeed the time at which the color fields
may be interpreted as nearly on-shell gluons
\cite{Fukushima:2011nq}. However, it is also clear from this plot that
the ratio of longitudinal to transverse pressure remains very
small. In fact, at leading order, this ratio decreases as $\tau^{-2}$,
which is characteristic of a free streaming system (i.e., a system
whose self-interactions are too weak to compete with the longitudinal
expansion).
\paragraph{Next to Leading Order, Instabilities} Such a behavior of
the longitudinal pressure is not consistent with hydrodynamical
evolution, where the ratio $P_{_L}/P_{_T}$ would instead increase to
eventually approach unity. For an underlying QCD description to allow
a smooth matching to a subsequent hydrodynamical expansion, there
should be a range of times where the two descriptions lead to similar
behaviors. It turns out that higher order corrections in the CGC
description are potentially more important than the power counting
suggests. Indeed, the power counting correctly states that loop
corrections (i.e. corrections beyond the classical field
approximation) are suppressed by additional powers of the coupling
constant, but it implicitly assumes that the coefficients in this
power expansion remain of order one at all times. It is this
assumption that turns out to be incorrect, because the classical
solutions of Yang-Mills equations are subject to instabilities that
make them exponentially sensitive to their initial conditions
\cite{Muller:1992iw,Biro:1993qc,Kunihiro:2010tg,Romatschke:2005pm},
combined with the fact that one-loop corrections can be expressed in
terms of small perturbations to the initial condition of the LO
classical field.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{resum}
\caption{\label{fig:resum1} Graphs that have the leading time
behavior in the presence of instabilities. }
\end{figure}
\paragraph{Resummation, Classical Statistical Approximation} By a more
careful power counting that keeps track of these terms that have an
exponential growth with time \cite{Gelis:2012ri}, it is possible to determine the set of
graphs that provide the leading contributions at large time. These
graphs are shown in the Figure \ref{fig:resum1} (and in the left part
of Figure \ref{fig:resum}, we show for comparison a typical
next-to-leading order contribution). It turns out that the sum of this
infinite set of higher-loop order graphs can be obtained by letting
the initial value of the gluon field fluctuate around its classical
value, with a Gaussian distribution which is completely determined by
the one-loop result. Schematically, this resummed result reads:
\begin{align}
{\cal O}_{\rm resummed}
=
\int \big[D{ a}\big]\;\exp\Big[-\frac{1}{2\,{\hbar}}\int_{{\boldsymbol x},{\boldsymbol y}}
{ a({\boldsymbol x})}{{\boldsymbol\Gamma}_2^{-1}({\boldsymbol x},{\boldsymbol y})}{ a({\boldsymbol y})}\Big]\;{\cal O}_{_{\rm LO}}[{\cal A}_{\rm in}+{ a}].
\label{eq:resum}
\end{align}
Such a resummation is of course only a part of the full answer (with
an appropriate choice of the variance ${\boldsymbol\Gamma}_2$, one may obtain
the complete LO and NLO results, but only a subset of the higher
orders), known in the literature as the {\sl Classical Statistical
Approximation} (CSA). We have written the explicit dependence of the
Gaussian distribution with respect to $\hbar$ to emphasize the quantum
nature of these fluctuations of the field. In the special case of
heavy ion collisions, the variance of the Gaussian fluctuations,
${\boldsymbol\Gamma}_2$, can be determined analytically at the time
$\tau=0^+$, i.e., just after the collision
\cite{Epelbaum:2013waa}. Note that in the vacuum, this variance reads
(this is just a sketch, that ignores the complications due to gauge
fields)
\begin{align}
\big<a(x)a(y)\big>={\boldsymbol\Gamma}_2(x,y)=\int \frac{d^3{\boldsymbol k}}{(2\pi)^3 E_{\boldsymbol k}}\;\frac{1}{2}\,e^{ik\cdot(x-y)}.
\label{eq:variance}
\end{align}
The factor $1/2$ in the integrand can be interpreted as the zero point
occupation of the vacuum (at one-loop, one may show that the only
quantum effect is the fact that the ground state is not empty but
subject to zero point fluctuations -- the resummation considered here
is an approximation in which this is extended to higher-loop orders).
Once the variance ${\boldsymbol\Gamma}_2$ is known, Eq.~(\ref{eq:resum}) can
be evaluated numerically on a lattice in a straightforward way, since
it simply amounts to reproducing the leading order classical CGC
computation with a fluctuating initial condition. The Gaussian
integral in Eq.~(\ref{eq:resum}) can be estimated by Monte-Carlo
sampling. These simulations, with the variance given in
Eq.~(\ref{eq:variance}), lead to an increase of the ratio of
longitudinal to transverse pressure \cite{Gelis:2013rba}. However, the
interpretation of this result is obscured by the fact that this setup
has no continuum limit when the lattice spacing goes to zero. This can
be understood easily from the fact that Eq.~(\ref{eq:variance})
corresponds to a flat momentum distribution of gluons, that extends to
arbitrarily large momenta (only cut-off by the inverse lattice
spacing).
\paragraph{CSA with a compact fluctuation spectrum} An
alternative to the variance given in Eq.~(\ref{eq:variance}) that does
not have this ultraviolet problem would be to replace the factor $1/2$
by another gluon distribution that has a fast enough fall-off at large
momentum,
\begin{align}
\big<a(x)a(y)\big>_{\rm alt}=\int \frac{d^3{\boldsymbol k}}{(2\pi)^3 E_{\boldsymbol k}}\;f_0(k)\,e^{ik\cdot(x-y)}.
\label{eq:variance-alt}
\end{align}
Although such a distribution cannot be derived from first principles,
unlike Eq.~(\ref{eq:variance}), a handwaving argument in favor of it
is that after a time of order $Q_s^{-1}$ the gluons produced in a
collision are nearly on-shell with a compact distribution that extends
up to momenta $k\sim Q_s$ (with an occupation number of order $g^{-2}$
within this support). With such a spectrum of initial field
fluctuations, the behavior of the ratio of pressures $P_{_L}/P_{_T}$
is at odds with what was obtained with Eq.~(\ref{eq:variance}),
everything else being equal: with Eq.~(\ref{eq:variance-alt}), one has
$P_{_L}/P_{_T}\sim \tau^{-2/3}$, showing no sign of isotropization
\cite{Berges:2013eia,Berges:2013fga}. In this scenario, it is argued
that isotropization is delayed until the gluon occupation number
becomes of order one, which would happen eventually at a time
$Q_s \tau \sim \alpha_s^{-3/2}$.
Going beyond the classical statistical approximation in a field
theoretical framework is possible with the two-particle irreducible
formalism \cite{Luttinger:1960ua,Baym:1961zz,Berges:2004yj}. This
formalism amounts to a self-consistent determination of the
propagator (which in a many-body context also contains the
information about the particle distribution) by resumming a
self-energy --itself a function of the propagator-- on the
propagator. The 2PI framework can be renormalized
\cite{vanHees:2001ik,Berges:2005hc}, thereby avoiding the issues
with the CSA and zero point vacuum fluctuations, and can thus be
used to track the real-time evolution of a system starting from any
quantum state. The main drawback of this approach is that it is very
demanding in terms of computational resources, especially in the
case of an expanding system like the one formed in a heavy ion
collision. At the time of this writing, there has only been one
``proof of concept'' implementation for an expanding system
\cite{Hatta:2011ky}, in which the questions related to
isotropization were not investigated.
\paragraph{Kinetic theory and Boltzmann equation} A less costly
alternative is {\sl kinetic theory}, that one may obtain from the 2PI
approach provided one makes two additional approximations:
\begin{itemize}
\item Quasi-particle approximation: this amounts to assuming that the
propagator describes on-shell infinitely long lived particles. With
this assumption (that can only be valid in a system where the mean
free path is much larger than the De Broglie wavelength of the
particles), the only unknown in the propagator is the particle
distribution $f(x,{\boldsymbol p})$.
\item Gradient approximation: this amounts to assuming that the
spatial variations of the system due to its off-equilibriumness
occur only on time and distance scales much larger than the De
Broglie wavelength of the particles.
\end{itemize}
With these two approximations, the Kadanoff-Baym equation of motion of
the 2PI formalism reduces to a much simpler Boltzmann equation,
schematically of the form
\begin{align}
\big(\partial_t+{\boldsymbol v}_{{\boldsymbol p}_1}\cdot{\boldsymbol\nabla}_{\boldsymbol x})\,f_1
&= C_{{\boldsymbol p}_1}[f]\\
&=\int_{{\boldsymbol p}_{2,3,4}}|M(12\to 34)|^2\, \delta(p_1+p_2-p_3-p_4)\nonumber\\
&\qquad\times\big\{f_3 f_4(1+f_1)(1+f_2)
-f_1 f_2(1+f_3)(1+f_4)\big\}.
\end{align}
The first line is the generic Boltzmann equation obtained when using
these two approximations, with a {\sl collision term} local in $x$
that can a priori contain arbitrary orders in the distribution $f$. In
the second line, we have written specifically the Boltzmann equation
that includes only $2\to 2$ elastic reactions (but this truncation is
an extra approximation that goes beyond the quasi-particle and
gradient approximations).
\paragraph{Testing semi-classical approximations via kinetic theory}
Besides being a tool to study the evolution of the particle
distribution, the Boltzmann equation can also provide some insights
about the limitations of the classical statistical approximation,
because identical classical approximations may be applied to the right
hand side of the Boltzmann equation. To that effect, the
correspondence is the following \cite{Mueller:2002gd,Jeon:2004dh}
\begin{itemize}
\item[1.] CSA with Eq.~(\ref{eq:variance-alt}) $\Longleftrightarrow$ keep only the terms cubic in $f$,
\item[2.] CSA with Eq.~(\ref{eq:variance}) $\Longleftrightarrow$ replace $f\to f+\tfrac{1}{2}$ in the previous approximation (note that in this approximation, the collision integral has the correct cubic and quadratic terms, and also some spurious linear terms not present in the exact collision term).
\end{itemize}
With the approximation 2 of the collision integral, one can for
instance reproduce quantitatively the ultraviolet sensitivity (i.e.,
the lack of continuum limit) of the CSA when zero point fluctuations
are included \cite{Epelbaum:2014mfa}. With the approximation 1, the
Boltzmann equation (for an expanding system) leads to the behavior
$P_{_L}/P_{_T}\sim \tau^{-2/3}$. It is also possible to understand by
kinetic arguments why this approximation misses important physics for
a longitudinally expanding system \cite{Epelbaum:2015vxa}. In such a
system, isotropization results from a competition between the
expansion of the system, that drives the particles towards a more a
more anisotropic distribution, and collisions that tend to
redistribute the directions of the velocities (for collisions that are
sufficiently isotropic, which is the case for a dense system due to a
strong Debye screening).
\begin{figure}[htbp]
\centering
\includegraphics[width=4cm]{isotropization}
\caption{\label{fig:22scat}Typical $2\to 2$ scattering in a highly anisotropic system.}
\end{figure}
However, as shown in Figure \ref{fig:22scat}, when two particles
(labeled $1$ and $2$) from a very anisotropic distribution scatter
out-of-plane, momentum conservation implies that the two final state
particles (labeled $3$ and $4$) end in the empty region of
phase-space. Such a scattering is forbidden when one keeps only the
cubic terms in the collision integral, because these terms correspond
to stimulated emission, that can only happen when one of the produced
particles goes in an already populated region. The large-angle
scattering process of Figure \ref{fig:22scat} is allowed only by the
terms quadratic in $f$, that are not present in the approximation of
Eq.~(\ref{eq:variance-alt}). The difference between a semi-classical
approximation such as Eq.~(\ref{eq:variance-alt}) and one that
includes the zero-point fluctuations can then be seen on the time
evolution of the ratio $P_{_L}/P_{_T}$, as shown in Figure
\ref{fig:Kin-scal-evol} in the case of a scalar theory (see
\cite{Epelbaum:2015vxa} for details): in this figure, we see that the
classical approximation leads to a decrease of this ratio as
$\tau^{-2/3}$, while with the full collision term it first decreases
(while the expansion outpaces the scatterings) and then increases
(when the expansion rate has become low enough) to approach unity.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{Pratio-g2f0-5}
\caption{\label{fig:Kin-scal-evol}Comparison of the classical and
full evolution of $P_{_L}/P_{_T}$ in kinetic theory for a
longitudinally expanding system. The dark purple curve is the
kinetic theory analogue of the classical statistical approximation
based on eq.~(\ref{eq:variance-alt}).}
\end{figure}
It is only for very small couplings (i.e., unrealistically large
values of the ratio of the shear viscosity to entropy, $\eta/s$ -- see
the next section for a discussion of the range of values of this ratio
expected in heavy ion collisions) that this classical approximation
agrees with the evolution driven by the full collision term long
enough to reach the asymptotic scaling regime.
\paragraph{Kinetic approach in Yang-Mills theory} A similar
computation has also been performed in the more realistic setting of
Yang-Mills theory \cite{Arnold:2002zm,Kurkela:2015qoa}, with similar results as
shown in Figure \ref{fig:kinevol}. There also, one sees the classical
approximation depart from the full evolution fairly quickly. For
instance, the curve $\lambda=0.5$, i.e. $\alpha_s=0.02$ for $N_c=2$,
deviates from the classical approximation around $Q_s\tau\approx 2$.
Moreover, this happens much earlier than the presumed range of
validity $Q_s\tau\lesssim\alpha_s^{-3/2}\approx 350$ predicted within
the classical approximation itself (in fact, we see from this plot
that the point $Q_s\tau =350$ on the classical evolution is orders of
magnitude off the correct trajectory).
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{cdplot_4_2}
\caption{\label{fig:kinevol}Isotropization in Yang-Mills theory in a
kinetic description. In this figure $\lambda\equiv g^2 N_c$ is the
't Hooft coupling. The solid black curve is the kinetic theory
analogue of the classical statistical approximation based on
eq.~(\ref{eq:variance-alt}). From \cite{Kurkela:2015qoa}.}
\end{figure}
In addition, this computation has shown a very good agreement with
second order hydrodynamics already at times where isotropization is
still far from being achieved, thereby providing a justification for
the applicability of hydrodynamics as early as $\tau\approx 0.6 $~fm/c
(compare the red and solid black curves in Figure \ref{fig:kinhydro}).
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{hydro_F10}
\caption{\label{fig:kinhydro} Comparison of kinetic theory with second order hydrodynamics. From \cite{Kurkela:2015qoa}.}
\end{figure}
\paragraph{Fixed points of kinetic evolution} Kinetic theory has also
been used to obtain results about the fate of a system that undergoes
longitudinal expansion
\cite{Blaizot:2017lht,Blaizot:2017ucy,Blaizot:2019scw,Blaizot:2020gql},
with only generic assumptions about the strength of the collisions.
The starting point of this study is the Boltzmann equation in the {\sl
relaxation time approximation},
\begin{align}
\Big(\partial_\tau -\frac{p_z}{\tau}\Big)\, f(\tau,{\boldsymbol p})
=-\frac{f-f_{\rm eq}}{\tau_{_R}}.
\end{align}
In this equation, $\tau_{_R}$ is a relaxation time that controls how
fast the particle distribution relaxes to its local equilibrium
value. This parameter may be chosen in various ways:
\begin{itemize}
\item $\tau_{_R}=\infty$~:~~ for a collisionless system,
\item $\tau_{_R}\sim \epsilon^{-1/4}$~:~~ for a ``conformal'' system,
i.e., a system where the collision rate scales as the inverse temperature,
\item $\tau_{_R}={\rm const}$~:~~ for a fixed collision rate (although
this is not very realistic with expansion).
\end{itemize}
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{att_eta_1}
\caption{\label{fig:kinattract}Kinetic evolution of the exponents
$g_n$ for many distinct initial conditions. The dotted line
indicates the interacting fixed point. From \cite{Blaizot:2017ucy}.}
\end{figure}
Then, one may define the following moments,
\begin{align}
L_n\equiv \int_{\boldsymbol p} {\boldsymbol p}^2 \; P_{2n}(p_z/p)\;f(\tau,{\boldsymbol p}),\quad g_n\equiv \tau \partial_\tau \ln L_n.
\end{align}
(In this definition, $P_{2n}(z)$ is the order-$2n$ Legendre
polynomial. Note that $L_0=\epsilon = P_{_L}+2P_{_T}$,
~~$L_1= P_{_L}-P_{_T}$.) The quantity $g_n$ gives the exponent in the
power law behavior of the corresponding moment. The Boltzmann equation
can then be transformed into an infinite set of linear equations that
govern the evolution of the moments $L_n, L_{n+1}$ and $L_{n-1}$, in
which all but one coefficients depend only on the left hand side of
the Boltzmann equation. Then, it was observed that these coupled
equations admit two (and only two) fixed points:
\begin{itemize}
\item A {\sl free-streaming fixed point}, obtained for $\tau_{_R}=\infty$,
where all the $L_n$ behave as $\tau^{-1}$ (i.e., all the $g_n$ go to
$-1$), with fixed ratios among them. In particular, one has
$L_1/L_0\to -\tfrac{1}{2}$, and $P_{_L}/P_{_T}$ decreases
like $\tau^{-2}$.
\item An {\sl interacting fixed point}, obtained for
$\tau_{_R}\sim \epsilon^{-1/4}$, where $g_0\to -\tfrac{4}{3}$, $g_1\to
-2$. At this fixed point, the system has a locally isotropic
particle distribution, and an energy density that decreases like
$\tau^{-4/3}$.
\end{itemize}
Note that in a scaleless system where the collision rate scales as
$\epsilon^{-1/4}$ (this is the conformal case mentioned above), the
ratio $\tau/\tau_{_R}$ always goes to infinity as $\tau\to \infty$,
and the system therefore always converges to the interacting fixed
point. The figure \ref{fig:kinattract} shows the evolution of the
first three $g_n$'s for an ensemble of initial conditions, as a
function of $\tau/\tau_{_R}$. After a short transient evolution, all
these evolutions coalesce on a universal ``attractor'' (indicated by
the thick black curve), on which the exponents are the free streaming
ones or the interacting ones depending on the value of the ratio
$\tau/\tau_{_R}$.
\section{Hydrodynamical evolution; Late stages}
The next, and in fact main, stage of the bulk evolution of the system
formed in the collision of two heavy ions is a rather long period of
relativistic hydrodynamical expansion. Hydrodynamics
\cite{Teaney:2009qa,Romatschke:2009im,Romatschke:2017ejr,Florkowski:2017olj,Gale:2013da}
is a coarse grained description whose starting point is the local
conservation laws, for energy and momentum and for any other conserved
charge (such as baryon number),
\begin{align}
\partial_\mu T^{\mu \nu}=0,\quad
\partial_\mu J^\mu_{_B}=0,\quad\cdots
\end{align}
However, these equations are too general to constrain uniquely the
evolution of the system. For this, it is necessary to express the
energy-momentum tensor and the other conserved currents in terms of a
small number of quantities.
\paragraph {Perfect fluid} The simplest case is that of a perfect,
i.e., non dissipative, fluid. In this case, the energy-momentum tensor
depends only on the local energy density $\epsilon$, pressure $p$ and fluid
$4$-velocity vector $u^\mu$,
\begin{align}
T^{\mu\nu}&\empile{=}\over{{\rm perfect}}(\epsilon+p) u^\mu u^\nu -p\,g^{\mu\nu}\nonumber\\
&= \epsilon\,u^\mu u^\nu+ p\,\Delta^{\mu\nu}\quad \mbox{with\ \ }\Delta^{\mu\nu}\equiv u^\mu u^\nu-g^{\mu\nu},
\label{eq:idealT}
\end{align}
where $g^{\mu\nu}$ is the Minkowski metric tensor with $(+,---)$
signature. The tensor $\Delta^{\mu\nu}$ is a projector on the local
rest frame of the fluid, $\Delta^{\mu\nu} u_\nu=0$, in terms of which
one may define time and spatial derivatives,
$D\equiv u^\mu \partial_\mu$ and
$\nabla^\mu \equiv \Delta^{\mu\nu}\partial_\nu$, in the fluid rest
frame. The equations of ideal hydrodynamics therefore read
\begin{align}
D\epsilon = -(\epsilon+p)\nabla_\mu u^\mu,\quad D u^\mu = -(\epsilon+p)^{-1}\,\nabla^\mu p.
\end{align}
These equations are the relativistic analogue of Euler fluid equations.
The first equation indicates that the local variation of the energy
density is proportional to the variation of the volume of fluid cells,
since $dV/V = dt \, \nabla_\mu u^\mu$. The second equation relates the
acceleration of the fluid to the gradient of its pressure. Note that
the first equation implies that entropy is conserved (indeed this
equation is equivalent to $d(\epsilon V)+pdV=TdS=0$).
\paragraph{Boost invariant ideal flow} In heavy ion collisions at
ultrarelativistic energies, the longitudinal momenta of the produced
particles are typically much larger than their transverse
momenta. This leads to a strong correlation between the longitudinal
momentum of a particle and its longitudinal position in coordinate
space. More precisely, one has
\begin{align}
y\equiv \frac{1}{2}\ln\Big(\frac{p^0+p^3}{p^0-p^3}\Big)\approx
\eta_s \equiv \frac{1}{2}\ln\Big(\frac{x^0+x^3}{x^0-x^3}\Big).
\end{align}
If the energy density of the fluid at some initial proper time
$\tau_0$ is independent of the spatial rapidity $\eta_s$ (this is
approximately the case in the Color Glass Condensate framework, since
the rapidity dependence comes from the JIMWLK evolution of the
distributions of color sources, which is significant only on scales
$\delta\eta_s\sim \alpha_s^{-1}$), the subsequent hydrodynamical
evolution of the fluid is boost invariant. In this case, the evolution
of a perfect fluid is governed by a single equation,
\begin{align}
\frac{d\epsilon}{d\tau}=-\frac{\epsilon+p}{\tau}.
\label{eq:BIhydro}
\end{align}
(If the fluid is not locally isotropic, the pressure in the right hand
side should be replaced by the longitudinal pressure.) In a conformal
theory (i.e., a theory with only massless particles and no running
coupling), the energy-momentum tensor is traceless and one has
$p=\epsilon/3$ in equilibrium. Therefore, this leads to
\begin{align}
\epsilon, p &\sim \tau^{-4/3},\nonumber\\
T&\sim \epsilon^{1/4}\sim \tau^{-1/3},\nonumber\\
s&\sim T^3 \sim \tau ,\nonumber\\
sV&\sim s\tau\sim \mbox{const},
\end{align}
where $s$ is the entropy density. Recall that the assumption of boost
invariance, and therefore these scaling laws, are only true as long as
the longitudinal expansion dominates over the transverse one (this is
hidden in the assumption that $V\sim \tau$), and are therefore
expected to change when the proper time becomes comparable to the
diameter of the colliding nuclei. Note also an important fact, equally
valid for solving Eq.~(\ref{eq:BIhydro}) or the general hydrodynamical
equations: in order to close the system of equations and obtain a
solution, it is necessary to use an {\sl equation of state} that
relates for instance the pressure to the energy density.
\paragraph{Viscous corrections}
In order to go beyond the simple description in terms of a perfect
fluid, one should first alter Eq.~(\ref{eq:idealT}) by writing
\begin{align}
T^{\mu\nu}=T^{\mu\nu}_{\rm perfect}+\pi^{\mu\nu}+\Pi\,\Delta^{\mu\nu},
\end{align}
where we have split the deviation from the perfect fluid into a
traceless tensor $\pi^{\mu\nu}$ and a term $\Pi \Delta^{\mu\nu}$ that
has a non-zero trace. The equations of motion are
$\partial_\mu T^{\mu\nu}=0$, combined with an equation of state and
constituent equations that express $\pi^{\mu\nu}$ and $\Pi$ in terms
of gradients. In a system which is not too far from local equilibrium,
these expressions may be expanded in powers of the
gradients\footnote{This expansion should be regarded as rather formal,
as it may not lead to a convergent series \cite{Heller:2016rtz}.},
and at lowest order one may write
\begin{align}
\pi^{\mu\nu}&=-\eta \,\sigma^{\mu\nu}\quad\mbox{with\ \ } \sigma^{\mu\nu}\equiv \nabla^\mu u^\nu+\nabla^\nu u^\mu-\frac{2}{3}\Delta^{\mu\nu}\,(\nabla_\rho u^\rho),\nonumber\\
\Pi &= -\zeta\,(\nabla_\rho u^\rho).
\label{eq:1storderexp}
\end{align}
The coefficients $\eta$ and $\zeta$ (respectively, the shear and bulk
viscosities) describe how the stress tensor responds to a small
gradient of the fluid velocity. The resulting hydrodynamical equations
are the relativistic analogue of the Navier-Stokes equations.
When applied to a boost invariant system, the resulting hydrodynamical
equations lead to
\begin{align}
\frac{d\epsilon}{d\tau}=-\frac{\epsilon+p-\tfrac{4}{3}\tfrac{\eta}{\tau}}{\tau},
\label{eq:BI1storder}
\end{align}
resulting in a slower decrease of the energy density compared to the
case of a perfect fluid. This equation indicates that the first order
gradient expansion it was obtained from is legitimate as long as
$\eta\ll \tau(\epsilon+p)$. Using the thermodynamic relation
$\epsilon+p=sT$, this condition can be turned into
\begin{align}
\frac{\eta}{s}\ll \tau T,
\end{align}
where the left hand side of the inequality is a local property of the
fluid while the right hand side is a property of the flow itself. From
kinetic theory, the ratio $\eta/s$ may be estimated to be of the order
of $\lambda T$ where $\lambda$ is the mean free path. Thus, the
inequality also reads $\lambda \ll \tau$, implying that the system
cannot be described by hydrodynamics at times that are smaller than
the time between two successive scatterings of a particle. Conversely,
the hydrodynamical description improves as $\tau$ increases. In a
scale invariant system, we may estimate the relative magnitude of the
first viscous correction in the right hand side of
Eq.~(\ref{eq:BI1storder}) as follows,
\begin{align}
\frac{\eta}{\tau(\epsilon+p)}\sim \frac{T^3}{\tau T^4}\sim \tau^{-2/3},
\end{align}
where in the last step we use the behavior of $T$ from ideal
hydrodynamics.
\paragraph{Causality, Second order hydrodynamics}
The first order gradient expansion in Eq.~(\ref{eq:1storderexp}) leads
to some pathologies in a relativistic context, because the correction
to the stress tensor follows instantaneously any modification to the
velocity field. This causality violating behavior eventually leads to
numerical instabilities when solving the corresponding hydrodynamical
equations. A possible practical strategy to fix this problem is to
modify Eqs.~(\ref{eq:1storderexp}) into relaxation equations, in order
to introduce a delay between changes of the gradients and the
resulting variation of the stress tensor. There is no unique way of
doing this, but on timescales longer that the ad-hoc relaxation time,
all these models lead to identical physical predictions.
At a more fundamental level, the modifications introduced by turning
Eqs.~(\ref{eq:1storderexp}) into relaxation equations can be motivated
from the study of second order terms in the gradient expansion
\cite{Israel:1979wp,Denicol:2010xn}. For instance, the second gradient
order in $\pi^{\mu\nu}$ contains a term of the form
$\eta \tau_\pi D \sigma^{\mu\nu}$, where $\tau_\pi$ has the dimension
of a time. If we take this second order expansion as is, it displays
similar causality issues as the first order one. These problems may be
avoided by replacing $\sigma^{\mu\nu}$ in the time derivatives that
appear at second order by the first order relationship between
$\sigma^{\mu\nu}$ and $\pi^{\mu\nu}$, thereby producing a term in
$-\tau_\pi D \pi^{\mu\nu}$. By doing this, the second order
constituent relation becomes a differential equation in time, with a
relaxation time $\tau_\pi$. The benefit of this point of view compared
to the more phenomenological one described before is that it allows to
relate the relaxation time to the underlying microscopic theory in an
unambiguous fashion.
\paragraph{Equation of state}
A key ingredient in order to turn the equations of hydrodynamics into
a closed set of equations is an equation of state that relates for
instance the entropy density to the temperature, or equivalently the
pressure and the energy density.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{EOS}
\caption{\label{fig:eos}Top: equation of state obtained from lattice QCD
at zero baryon chemical potential. Bottom: speed of sound as a
function of temperature. Boxes: extraction from LHC data (see
text). From \cite{Gardim:2019xjs}.}
\end{figure}
Perturbation theory, improved by the resummation of hard thermal
loops\cite{Braaten:1989mz}, allows to obtain robust results for temperatures only a couple
of times above the critical temperature \cite{Blaizot:1999ip,Blaizot:1999ap,Blaizot:2000fc,Blaizot:2003iq,Andersen:1999sf,Andersen:2000yj,Andersen:2011sf,Haque:2014rua}, but becomes less and less
reliable as $T_c$ is approached from above. A non-perturbative
first-principles alternative is lattice QCD
\cite{Philipsen:2012nu}. At null baryon chemical potential, the
advances in computing hardware and algorithms are by now allowing to
perform unquenched (i.e., with virtual quark loops) simulations with
realistic quark masses. At low temperatures compared to the
deconfinement transition, these computations may be continued by an
equation of state based on a gas of hadron resonances.
This approach works well only at zero baryon chemical potential, where
it has shown that the deconfinement transition is a continuous
crossover rather than a discontinuity for realistic quark masses. At
$\mu_B>0$, the determinant resulting from integrating out the fermion
fields is complex valued, leading to a {\sl sign problem} that
precludes any direct approach based on a Monte-Carlo sampling. When
$\mu_B/T$ is small enough, various workarounds are possible:
reweighting, analytic continuation from calculations at imaginary
$\mu_B$ (for which there is no sign problem), that allow to make an
incursion into the territory of positive chemical potentials
\cite{Bazavov:2017dus,Borsanyi:2019hsj}. It is expected that the
crossover at small $\mu_B$ becomes a first order phase transition at
larger $\mu_B$, the beginning of the transition line being a second
order critical point (the red dot in Figure \ref{fig:phase-diagram}).
However, the quest in lattice simulations for such a second order
critical point has remained rather inconclusive until now.
There has been a recent attempt to extract directly some information
about the equation of state from LHC heavy ion data
\cite{Gardim:2019xjs}. In this work, the authors used a hydrodynamical
simulation to estimate the effective temperature $T_{\rm eff}$ and
effective volume $V_{\rm eff}$ of a hypothetical homogeneous fluid
that would evolve into a system with the same energy and entropy as
the QGP at the time where the particles decouple (see the discussion
of {\sl freeze-out} later in this section). They observed that this
effective temperature is related to the mean transverse momentum
$\big<p_\perp\big>$ of the final state particles by
$T_{\rm eff}\approx 3.07\,\big<p_\perp\big>$ (with a proportionality
coefficient roughly independent of the equation of state and transport
coefficients). Given ALICE data in the bin of centrality $[0,5]\%$,
one gets $T_{\rm eff}=222\pm 9$~MeV. The total entropy is inferred
from the number of produced charged particles,
$S\approx 6.7\,N_{\rm ch}$, while the effective volume $V_{\rm eff}$
also comes from the hydrodynamical simulation, giving an entropy
density $s=20\pm 5$~fm${}^{-3}$, and $s/T_{\rm eff}^3=14\pm 3.5$, in
agreement with the lattice equation of state. In particular, this is
much higher than the value $\sim 3$-$4$ expected in the confined
phase, suggesting a large number of degrees of freedom consistent with
deconfinement. By repeating this analysis at two collision energies
($2.76$ and $5.02$~TeV), one may also estimate the speed of sound via
\begin{align*}
c_s^2(T_{\rm eff})=\frac{sdT}{Tds}\Big|_{T_{\rm eff}}
=
\frac{d\ln\big<p_\perp\big>}{d\ln (dN_{\rm ch}/d\eta)}
=0.24\pm 0.04,
\end{align*}
again in agreement with lattice computations.
\paragraph{Transport coefficients}
Among the transport coefficients that enter in hydrodynamics, the one
that has received most interest is the shear viscosity $\eta$. As we
mentioned earlier, the ratio $\eta/s$ is the ratio of the mean free
path to the quantum wavelength of the particles. This allows to make
some simple estimates in several limits. Firstly, in the perturbative
limit (weak coupling, and low enough particle density), this leads to \cite{Jeon:1994if,Jeon:1995zm,Arnold:2000dr,Arnold:2003zc}
\begin{align}
\frac{\eta}{s}\sim \frac{1}{\alpha_s^2 \ln(\alpha_s^{-1})} \gg 1.
\end{align}
(Although we do not write the prefactor here, it can be determined in
the weak coupling limit.) Another limit is that of a strongly coupled
plasma. Although this limit is not accessible in QCD, this calculation
is possible in a supersymmetric cousin of QCD thanks to the AdS/CFT
correspondence, leading the following result \cite{Policastro:2001yc}
\begin{align}
\frac{\eta}{s}=\frac{1}{4\pi}.
\end{align}
Note that such a constant value, independent of the coupling, is
consistent with the fact that quantum mechanics prevents this ratio
from becoming arbitrarily small since the quantum wavelength is a
lower value for the mean free path (but such an argument does not give
the value of the constant ratio one would reach).
Out of equilibrium, there is another interesting situation, where the
coupling constant is weak but the gluon occupation number is large, possibly
as large as $\alpha_s^{-1}$. In this case, the scattering rate should
contain a factor $f(1+f)$ where $f$ is the occupation number of the
scattering centers (when the occupation number is small, only a factor
$f$ is necessary, leading to the usual formula
$\lambda^{-1}\sim n \sigma$). Although the scattering cross-section is
proportional to $\alpha_s^2$ (up to logarithms), the factor $f(1+f)$
leads to a mean free path that does not contain any power of the
coupling constant. This situation of weak coupling but high density,
which is relevant in the very early stages of a heavy ion collision,
illustrates the fact that a strong coupling is not the only
possibility for having a small ratio $\eta/s$: more generally, the
system should be {\sl strongly interacting} for this to be true.
Outside of the above limiting cases, one may also consider lattice
QCD. Unlike the equation of state, the transport coefficients remain
very difficult to even estimate in lattice QCD. Thanks to Green-Kubo's
formulas, transport coefficients may be expressed in terms of the
Fourier transform of a retarded current-current (the current should be
the one that couples to the quantity whose transport one is interested
in, e.g., a charge current for an electrical conductivity) correlation
function at zero momentum,
\begin{align}
\sigma \propto \lim_{\omega\to 0} \frac{\rho(\omega,{\boldsymbol k}=0)}{\omega}
\end{align}
This formula expresses the transport coefficient in terms of the slope
at zero energy of the corresponding spectral function
$\rho(\omega,{\boldsymbol k})$. However, on the lattice, one has only a direct
access to the imaginary time version of these correlation functions.
This imaginary time correlator admits a spectral representation
involving the relevant spectral function,
\begin{align}
G(\tau,{\boldsymbol k})=\int d\omega\; K_{_T}(\tau,\omega)\,\rho(\omega,{\boldsymbol k}),
\label{eq:spectral-rep}
\end{align}
where $K_{_T}(\tau,\omega)$ is a known temperature-dependent kernel.
In a lattice approach, the left hand side of this equation would be
obtained (with statistical errors, and a finite number of values of
$\tau$ and ${\boldsymbol k}$) from numerical simulations, and one would then try to
invert this integral relationship to obtain the spectral function
$\rho(\omega,{\boldsymbol k})$. The difficulty with this is that it is a very
ill-posed problem when the left hand side is only imperfectly known,
implying that a direct inversion is unfeasible. Attempts at
constraining the spectral function in this way have been made by using
Bayes theorem in order to find the most likely spectral function
compatible with Eq.~(\ref{eq:spectral-rep}) and with a set of prior
assumptions about its shape (a minimal assumption would be that it is
positive) \cite{Asakawa:2000tr,Yamazaki:2001er,Sasaki:2005ap,Morningstar:2001je,Burnier:2014gta}.
More recently, a hybrid approach was proposed
\cite{Christiansen:2014ypa}, that combines a skeleton loop expansion
for the correlator $\big<[\pi_{ij}(x),\pi_{ij}(0)]\big>$ that enters
in Green-Kubo's formula, truncated at two loops, and a
non-perturbative input for the couplings and the propagator that
enters in this expansion obtained from the Euclidean functional
renormalization group \cite{Berges:2000ew}.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{ETAoverS}
\caption{\label{fig:viscosity}Theoretical evaluations of the viscosity to entropy ratio as a function of temperature. From \cite{Christiansen:2014ypa}.}
\end{figure}
The results obtained with this approach are shown in
Fig. \ref{fig:viscosity}, for a pure Yang-Mills theory (the result
shown for QCD with quarks is an estimate based on the pure YM result,
not an ab initio calculation). Interestingly, this calculation
suggests that $\eta/s $ has a rather pronounced temperature
dependence, with a minimum around $1.25\,T_c$ (the location of this
minimum in units of the critical temperature is almost identical in
Yang-Mills theory and in QCD). Moreover, the value of $\eta/s$ at this
minimum is only slightly above the value $1/(4\pi)$ obtained in the
strong coupling limit in the AdS/CFT approach for $N=4$ super-Yang-Mills theory.
\paragraph{Freeze-out}
Since at the end of the day experiments observe particles, it is
necessary to convert the objects (the fluid energy density and its
velocity) evolved by the hydrodynamical equations into distributions
of particles. In fact, the need for a description in terms of
particles arises much earlier than the time at which particles are
detected because the density decreases as the system expands, and
therefore the mean free path increases, leading to a situation where
the conditions of applicability of hydrodynamics are not met anymore.
On should distinguish between a {\sl chemical freeze-out}, at which
the inelastic collisions become rare (after that, the chemical
composition of the system is frozen), and a {\sl kinetic freeze-out}
where the elastic collisions also stop and the momentum distributions
freeze. Experimentally, the temperature and chemical potential at the
chemical freeze-out are well constrained from the ratios of abundances
of various species of particles, as shown in the figure
\ref{fig:chem_freeze}
\cite{Andronic:2003zv,Andronic:2006ky,Stachel:2013zma,Andronic:2017pug}
(see also \cite{Fries:2003vb,Fries:2003kq} for a somewhat related
approach to hadron formation).
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{yields_fit2_276}
\includegraphics[width=9cm]{tmu-snn_2017}
\caption{\label{fig:chem_freeze}Top: measured hadron yields compared
to the statistical hadronization model output. Bottom: temperature
and chemical potential at chemical freeze-out, extracted from
collisions at various energies. From \cite{Andronic:2017pug}.}
\end{figure}
In the simplest version of this model, one assumes that all particle
distributions are the equilibrium ones at a common temperature and
chemical potential. A volume is also necessary to obtain absolute
yields, but this parameter drops out in ratios of the yields of
various particle species.
The conversion from a fluid to free particles should be done at the
kinetic freeze-out, by choosing a locally space-like ``surface''
$\Sigma$, and by using the {\sl Cooper-Frye formula},
\begin{align}
E_{\boldsymbol p} \frac{dN}{d^3{\boldsymbol p}}=\frac{1}{(2\pi)^3}\int_\Sigma d^3S_\mu\; P^\mu\, f(P\cdot u),
\end{align}
where $f$ is the local distribution function. For a fluid in local
thermal equilibrium, $f$ is the Bose-Einstein of the Fermi-Dirac
distribution, evaluated at the local fluid temperature. However, when
the fluid is viscous, there should also be deviations from local
thermal distributions,
\begin{align}
f(p)=f_{\rm eq}(p)+\delta f(p).
\end{align}
The form of the deviation $\delta f(p)$ is related to the transport
coefficients (such as the shear viscosity) and therefore depends on
the microscopic interactions in the fluid.
In this approach, the freeze-out conditions (temperature and chemical
potential) are a priori free parameters that may be adjusted to best
fit the spectra of produced particles. A more sophisticated
alternative would be to convert the fluid into particles at an earlier
time, and continue the evolution with kinetic equations
\cite{Hirano:2004en,Hirano:2005xf,Monnai:2009ad}. In this fancier approach, the
freeze-out temperature is controlled by the values of the various
cross-sections used in the kinetic description (this would even allow
different species to decouple at different times), and is no longer an
ad hoc input of the model.
\paragraph{Flow anisotropies}
Experimentally, several predictions of hydrodynamical models may be
compared with data. One of them is the transverse momentum spectra of
the produced particles, that are sensitive to the temperature at which
the freeze-out occurs (and to whether the particle distributions are
the equilibrium ones or not).
Another main class of observables directly related to the
hydrodynamical expansion of the quark gluon plasma consists in
measuring angular correlations among the detected particles
\cite{Ollitrault:1992bk,Dusling:2007gi,Song:2007ux,Luzum:2008cw,Voloshin:2008dg,Alver:2010gr,Alver:2010dn,Teaney:2010vd,Muller:2012zq,Heinz:2013th}. The
quantities used to express these correlations are the so-called $v_n$,
defined as the Fourier coefficients of the azimuthal distribution of
particles. These flow coefficients may be measured as a function of
transverse momentum, of the centrality of the collisions, of the
species of particles. The hydrodynamical expansion provides a
one-to-one mapping between the spatial anisotropy of the initial
distribution of energy density (and possibly its initial flow) and the
final momentum anisotropy, that may be understood as an effect of
pressure gradients (the fluid is accelerated in the direction of the
pressure gradient). Moreover, the conversion of spatial
inhomogeneities into momentum space anisotropies depends quite
sensitively on the transport coefficients, mostly the shear viscosity:
if the ratio $\eta/s$ is too large, this conversion is very
ineffective and hydrodynamics cannot explain the rather large values
of the measured $v_n$s.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{v2_part_id}
\caption{\label{fig:v2}Second Fourier coefficient of the azimuthal distribution for identified particles. From \cite{Abelev:2014pua}.
}
\end{figure}
Recently, the JETSCAPE collaboration has used a Bayesian approach in
order to extract the most likely values of the shear and bulk
viscosities as a function of temperature \cite{Paquet:2020rxl}, as
shown in Figure \ref{fig:visco}.
\begin{figure}[htbp]
\centering
\includegraphics[width=5cm]{closure_bulk}
\includegraphics[width=5cm]{closure_shear}
\caption{\label{fig:visco}Bayesian extraction of the shear and bulk
viscosities from Pb-Pb collisions at $2.76$~TeV. From \cite{Paquet:2020rxl}.}
\end{figure}
One can see that the constraints provided by the data restricts the
range of values of the viscosities, down to fairly low values after
normalization by the entropy density, compared to the prior
distribution used in this analysis. Although these results are very
promising, one should keep in mind that, besides the transport
coefficients one wants to extract, the model has other unknowns, especially in the
modelling of the initial condition and in the details of the
implementation of the freeze-out.
\paragraph{Hydrodynamics from kinetic theory}
In local thermal equilibrium, the hydrodynamical description has six
unknowns (the local baryon density $n_{_B}$, energy density $\epsilon$
and pressure $P$, plus the three independent components of the fluid
velocity $u^\mu$), and the conservation equations for the baryonic
current and the energy-momentum tensor provide five equations. The
system is closed by the equation of state that relates
$n_{_B}, \epsilon$ and $P$.
Away from this ideal situation, there are nine additional unknowns
($\Pi$, five independent components of $\pi^{\mu\nu}$, and three
components of the mismatch between the energy flow vector and the
baryon number flow vector). In addition to the equation of state, we
thus need $14$ equations to close the system, that are made by the
conservation equations ($5$ equations), and by $9$ constitutive
equations relating the stresses to the gradients. But these equations
can also be obtained from the underlying microscopic dynamics, in the
form of a Boltzmann equation $p^\mu \partial_\mu f = C_{\boldsymbol p}[f]$. Indeed,
by weighting the Boltzmann equation by $1$, $p^\nu$, $p^\nu p^\lambda$
and integrating over ${\boldsymbol p}$, we get
\begin{align}
&\partial_\mu \int_{\boldsymbol p} p^\mu\;f= \int_{\boldsymbol p} C_{\boldsymbol p}[f]=0,\\
&\partial_\mu \int_{\boldsymbol p} p^\mu p^\nu \;f =\int_{\boldsymbol p} p^\nu \,C_{\boldsymbol p}[f]=0,\\
&\partial_\mu \int_{\boldsymbol p} p^\mu p^\nu p^\lambda \; f
=
\int_{\boldsymbol p} p^\nu p^\lambda\,C_{\boldsymbol p}[f].
\end{align}
The right hand side of the first two equations is identically zero
given the symmetry properties of the collision integral, for any
distribution $f$. This set of $5$ equations is in fact the five
conservation equations. The last equation forms a set of $10$
independent equations (given the symmetry of $p^\nu p^\lambda$). Note
that by summing over $\nu=\lambda$, one recovers the conservation
equation of particle number. Therefore, this set of equations contains
only $9$ novel equations, i.e., precisely the number needed to close
our macroscopic description. This approach, supplemented by an
expansion of $f$ around the equilibrium distribution $f_0$, is known
as Grad's $14$-moment method. It allows to relate the hydrodynamical
description (in particular the transport coefficients \cite{Denicol:2010xn,Betz:2010cx,Denicol:2012cn,Denicol:2012es,Muronga:2001zk,Muronga:2003ta}) to the
underlying microscopic interactions encoded in the collision integral.
\paragraph{Hydrodynamics far from equilibrium}
For a long time, it has been thought that the hydrodynamical
description is a near-equilibrium effective description (this point of
view stems in part from the fact that one may obtain hydrodynamics
from a truncated gradient expansion). However, it was realized
recently that this gradient expansion may be a non-convergent series
with a null radius of convergence \cite{Heller:2013fn,Heller:2016rtz}. This
observation suggests to reconsider the conditions of applicability of
the hydrodynamical description (indeed, if the radius of convergence
is zero, it cannot serve as a parameter that defines what ``close
enough to equilibrium'' means). Another, observational, reason for
reassessing the applicability of hydrodynamics as an effective
macroscopic description is that in heavy ion collisions it appears to
perform much better than what one may naively expect by viewing it as
the result of an expansion around equilibrium.
A step towards a better understanding of these questions has been the
discovery (so far, in simple cases, where the dynamics and the flow
are sufficiently symmetric) of attractors towards which hydrodynamical
solutions evolve, even while gradients are still sizeable
\cite{Denicol:2017lxn,Heller:2011ju,Strickland:2018ayk,Strickland:2019jut}. In other
words, these solutions quickly reach a universal behavior independent
of the details of their initial conditions, long before the system is
in a state of isotropic local equilibrium. On these attractors, the
dissipative currents behave in a universal way in terms of the
(possibly still large) gradients, which shed another light on the
constitutive relations, indicating that their validity may not be
conditioned by a gradient expansion.
\paragraph{Hydrodynamics in small systems}
Another pressing question, closely related to the previous point, is
to determine what is the ``minimal size'' of a system that may be
described by hydrodynamics. Indeed, as the system becomes smaller, the
collectivity (that may be quantified by the ratio of the system size
by the mean free path, a gross measure of the number of collisions per
particle) is expected to decrease, leading to a situation less
favorable for the applicability of hydrodynamics.
Experimentally, this question was put forward by the observation that
certain correlation patterns (e.g., the ``ridge'' long range rapidity
correlation observed in the two-particle spectrum) are seen in
nucleus-nucleus collisions \cite{Ray:2011zza}, proton-nucleus
\cite{Abelev:2012ola} and even proton-proton collisions
\cite{Khachatryan:2010gv} (provided one triggers on high multiplicity
final states in the latter cases). In nucleus-nucleus collisions, the
ridge is interpreted as the result of the collective radial motion of
the produced matter
\cite{Voloshin:2003ud,Shuryak:2007fu,Dumitru:2008wn}, i.e., as
flow. Moreover, in this case, there is only a modest contribution to
the correlation provided by the CGC initial condition.
In smaller systems (see Figure \ref{fig:small} for some recent data),
the interpretation of the ridge correlation has been the subject of
intense debates between two extreme positions: that the correlation
can be entirely explained by collective flow
\cite{Bozek:2012gr,Bozek:2013uha} (i.e., final state interactions), or
entirely due to initial state correlations
\cite{Dumitru:2010iy,Dusling:2012iga,Dusling:2012wy,Dusling:2013oia}.
By now, it seems that a consistent description of the flow
observables in systems for which $dN_{\rm ch}/d\eta\ge 10$ calls for a
dominance of the final state interactions over the initial state
correlations. At low multiplicity, the observed flow is probably the
result of the combination of initial state correlations and final
state interactions \cite{Schenke:2019pmk} (it has been observed in a
realistic hydrodynamical model that the final state flow is less
correlated with the initial geometry but more correlated with the
initial momentum anisotropy at low multiplicity, and that this trend
is reversed at high multiplicity).
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{Small_systems}
\caption{\label{fig:small}Points: $v_2$ and $v_3$ flow coefficients
in collisions of nuclei of various sizes, down to proton-proton
collisions (ALICE collaboration). Bands: hydrodynamical model
predictions. From \cite{Schenke:2020unx}. See also \cite{Acharya:2019vdf,Sievert:2019zjr}.
}
\end{figure}
\section{Tomography and hard probes}
Until now, we have discussed mostly the bulk properties of the matter
produced in heavy ion collisions, from its formation to its subsequent
evolution. There, the quark-gluon-plasma or its precursor was the core
subject of the study, with flow observables a tool for accessing
(almost) directly some of their properties. Now, we turn to a class of
observables in which this matter plays the role of a substrate that
modifies them compared to a situation where the final state would be
mostly empty. Here, the strategy is to measure these observables both
in nucleus-nucleus collisions and in collisions of smaller projectiles
--such as proton-proton collisions-- where the formation of a
quark-gluon plasma is not expected, and where the observables are
supposedly well understood. The comparison between the two (with an
appropriate rescaling to account for the different sizes of the two
systems) provides informations on the properties of the matter
produced in heavy ion collisions. A better control is provided by
observables that are perturbative (meaning that they are characterized
by a hard scale, allowing to use QCD in a regime of weak coupling), so
that they can at least be computed with some degree of accuracy in
proton-proton collisions. However, even in this case, one should keep
in mind that their modifications by the surrounding medium may involve
much smaller momentum scales that render a perturbative treatment
difficult or even impossible in nucleus-nucleus collisions.
\paragraph{Basics of QCD radiation}
The most important feature of QCD in this context is the fact that
gluons are massless, and that three flavors of quarks (u,d, and s)
have masses lower or comparable to the non-perturbative scale
$\Lambda_{\rm QCD}$. For this reason, gluon radiation is enhanced for
soft gluons and for gluons emitted collinearly to their parent. For
instance, the emission probability of a gluon off a parton goes like
\begin{align}
dP\propto \alpha_s \, \frac{d^2{\boldsymbol k}_\perp}{k_\perp^2 }\frac{dz}{z},
\label{eq:softrad}
\end{align}
where ${\boldsymbol k}_\perp$ is the transverse momentum of the gluon (relative to
the direction of the momentum of its parent) and $z$ is the fraction
of its longitudinal momentum relative to that of its parent. The
unwritten prefactor contains a color factor that depends on whether the
emitter is a quark or a gluon. This probability diverges at small
$k_\perp$ and at small $z$. For radiation in vacuum (e.g., in the
final state of a proton-proton collision), the consequences of these
divergences are well understood:
\begin{itemize}
\item After one resums the soft radiation, the probabilities for
partonic final states with a prescribed number of (massless) gluons
is actually zero. This is of course largely irrelevant, since
because of confinement the gluons are not the objects that are
eventually detected.
\item The observables that match the most closely the concept of
perturbative parton are the so-called {\sl jets}. Loosely speaking,
a jet is a collimated beam of particles produced by the splitting of
a common ancestor (quark or gluon). At a more operational level,
defining jet cross-sections requires that one defines a procedure
for deciding when two distinct partonic final states correspond to
the same configuration of jets. For instance, two graphs that differ
by a loop correction, or by an extra soft or collinear gluon, should
contribute to the same jet final state. The jet definition is not
unique (in particular, because there is no unique way of defining
``soft'' or ``collinear''), but they all share a crucial property:
they lead to finite cross-sections when all mass scales are sent to
zero (observables that have this property are said to be {\sl
infrared and collinear safe}) -- in other words, all the logarithms
that one would get by integrating eq.~(\ref{eq:softrad}) down to
$k_\perp=0$ or $z=0$ cancel in these observables. The
theoretical definition of what one means by a jet in turn defines
the procedure (the so-called ``jet algorithm'') for extracting jet
cross-sections from experimental data: given a final state made of
detected particles, the jet algorithm defines how they should be
clustered into jets.
\item Jet cross-sections are calculable without any non-perturbative
input, but the price to pay for this is to give up on the idea of
saying something about individual particles in the final state. It
is also possible to consider cross-sections for producing a certain
hadron with a given momentum, but their calculation requires to
introduce {\sl fragmentation functions}, that may be viewed as the
inclusive probability that a certain quark or gluon turns into this
hadron (plus any number of additional particles that we do not care
about). This time, the collinear logarithms from
eq.~(\ref{eq:softrad}) do not cancel and they must be resummed,
which introduces a scale dependence into the fragmentation
function. This scale dependence is perturbative (governed by the
DGLAP evolution equation
\cite{Gribov:1972ri,Altarelli:1977zs,Dokshitzer:1977sg}, now known
up to three-loop accuracy
\cite{Moch:2004pa,Vogt:2004mw,Vermaseren:2005qc}), but the initial
condition of this evolution is non-perturbative and must be
extracted from experimental input. The hadron production
cross-section is obtained as the convolution of a partonic
cross-section with the fragmentation function, evaluated at a
certain scale (called the {\sl factorization scale}). This scale is
not a physical parameter, but rather a remnant of the truncation of
the perturbative series at a finite order (the residual scale
dependence decreases by going to higher loop order, and a fully
non-perturbative calculation would have no such scale at all).
\end{itemize}
\paragraph{Production of hard probes in nucleus-nucleus collisions}
When extending these ideas to nucleus-nucleus collisions, a generic
assumption is that the {\sl production} of the hard object proceeds in
the same perturbative way as in proton-proton collisions, the only
change being a change of the parton distribution functions that
describe the initial state. In the light of the earlier discussion of
gluon saturation, it is clear that for this to be true the observable
of interest should probe these distributions away from the non-linear
saturation regime, in order to be dominated by processes that probe a
single parton in each projectile. This is a reasonable assumption at
high virtuality $Q^2$, since the non-linear corrections are typically
suppressed as powers of $Q_s^2/Q^2$. Another thing to keep in mind is
that the direct measurements of {\sl nuclear parton distributions} are
scarce. An approximate treatment consists in viewing a nucleus as an
incoherent superposition of protons and neutrons. The parton
distributions of protons are very well known. For the neutrons, one
usually treats the neutron (udd) as the isospin partner of the proton
(uud). This implies for instance that the u-quark distribution in a
neutron is the same as the d-quark distribution in a proton,
etc... But note that the measurement of the structure function $F_2$
in deep inelastic scattering off a proton does not allow to
disentangle the u and d quark distributions. To separate them, one
also needs DIS measurements with deuteron (plus the assumption that
the binding of the deuteron is weak enough compared to the relevant
virtuality scales so that its parton distributions are just the sum of
the proton and neutron ones).
\paragraph{Gluon formation time} The discussion of parton and jet
energy loss requires to introduce the concept of {\sl formation time}
of a radiated gluon. Consider the emission of a gluon of momentum $k$
off a colored particle (quark or gluon) of momentum $p+k$. According
to the uncertainty principle, the virtuality of the line of momentum
$p+k$ just before the gluon emission vertex defines the gluon
``formation time''
\begin{align}
t_f^{-1}
\equiv
{E_{\boldsymbol p}+E_{\boldsymbol k}-E_{{\boldsymbol p}+{\boldsymbol k}}}
\empile{\approx}\over{k\ll p}
\frac{(p+k)^2}{2E_{\boldsymbol p}}
=
E_k(1-\cos\theta)
\empile{\approx}\over{\theta\ll 1}
\tfrac{1}{2}E_{\boldsymbol k}\theta^2.
\label{eq:tf}
\end{align}
This time can also be interpreted as the time necessary for the
wave-packet of the new gluon to separate sufficiently from that of the
emitter. Indeed, the transverse separation between the emitter and the
gluon grows with time according to
$\Delta r_\perp\approx \theta t$. On the other hand, the wavelength
of the gluon, projected on the plane orthogonal to the emitter, reads
\begin{align}
\lambda_\perp
=
\frac{1}{E_{\boldsymbol k}\sin\theta}
\empile{\approx}\over{\theta\ll 1}
\frac{1}{E_{\boldsymbol k} \theta}.
\end{align}
We see that the formation time is also given by the condition
$\Delta r_\perp\Big|_{t_f}= \lambda_\perp$.
\paragraph{Parton energy loss}
The simplest of the observables sensitive to the medium modification
of parton splitting consist in comparing inclusive hadron spectra
measured in nucleus-nucleus collisions with the same spectra measured
in proton-proton collisions at the same energy. However, a direct
comparison of these two spectra is no meaningful because these yields
come from system of vastly different volumes. In order to account for
this trivial geometrical effect, it is customary to define the
so-called ``nuclear modification factor'',
\begin{align}
R_{AA}\equiv \frac{\frac{dN}{d^3{\boldsymbol p}}\Big|_{AA}}{N_{\rm coll}\times\frac{dN}{d^3{\boldsymbol p}}\Big|_{pp}},
\end{align}
where $N_{\rm coll}$ is the number of binary nucleon-nucleon
collisions. For hard objects whose production is obtained as an
incoherent sum of nucleon-nucleon collisions, scaling by this factor
effectively eliminates the trivial volume dependence (in other words,
the ratio $R_{AA}$ should be equal to one in the absence of final
state medium effects). Note that $N_{\rm coll}$ is not a directly
measurable quantity, since its determination requires a modelling of
the nucleon distribution inside the nucleus of interest. The ratio
$R_{AA}$, called the nuclear modification factor, is displayed in the
case of charged hadrons in the figure \ref{fig:raa}, for central
lead-lead collisions at the LHC.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{figures_CMS_HIGHpT}
\caption{\label{fig:raa}Suppression factor of charged hadrons in nucleus-nucleus collisions. From \cite{Foka:2016zdb}. See also \cite{Abelev:2012hxa}.
}
\end{figure}
The most obvious feature of the ratio $R_{AA}$ is that it is below
unity over a large momentum range, and slowly reaches one at large
momenta. This means that the spectra of charged hadrons (at high
momentum compared to the QCD scale $\Lambda_{\rm QCD}$) are depleted
in nucleus-nucleus collisions compared to proton-proton collisions,
which can be explained by an increased fragmentation in the presence
of a surrounding medium. Following
Baier-Dokshitzer-Mueller-Peign\'e-Schiff
\cite{Baier:1996kr,Baier:1996sk,Baier:2001yt} and Zakharov
\cite{Zakharov:1996fv,Zakharov:1997uu,Zakharov:1998sv,Zakharov:2000iz,Baier:2000mf},
this medium-induced radiation can be understood semi-quantitatively as
follows. The formation time of a gluon of energy $E_{\boldsymbol k}$ and transverse
momentum $k_\perp$ is $t_f\approx E_{\boldsymbol k}/ k_\perp^2$ (this was derived
in eq.~(\ref{eq:tf})). The main difference in the presence of a medium
is that the transverse momentum of the emitted gluon is altered by the
scatterings that occur within its formation time. Assuming that these
scatterings act as random and independent kicks, the increase of the
transverse momentum is a diffusion process in transverse momentum
space. Therefore, the transverse momentum accumulated during the time
$t_f$ is given by $k_\perp^2\approx \widehat{q}\, t_f$, where
$\widehat{q}$ is a constant that quantifies the mean
$\Delta k_\perp^2$ per unit length (it is a quantity that depends on
the properties of the medium: density/temperature, Debye
screening). Consistency between these two relations implies that
$t_f\approx \sqrt{E_{\boldsymbol k}/\widehat{q}}$. Based on this, the induced
emissions can be divided into three regimes, depending on the energy of the radiated gluon
\begin{itemize}
\item $E_{\boldsymbol k} <\omega_{BH}$, with
$\omega_{BH}\equiv \widehat{q}\lambda^2$ ($\lambda$ is the mean free
path of the emitter in the medium). In this regime, the radiation
occurs coherently over path lengths shorter than the mean free
path. Successive collisions, separated by $\lambda$ (therefore there
are $L/\lambda$ of them), contribute incoherently to the total
radiation spectrum. In this regime, known as the Bethe-Heitler
regime, the emission spectrum behaves as $dI/dE\sim E^{-1}$.
\item $\omega_{BH}<E_{\boldsymbol k}<\omega_c$, with
$\omega_c\equiv \widehat{q}L^2$ ($L$ is the length traveled by the
emitter before exiting the medium). In this energy range, the
formation time is larger than the mean free path, and shorter than
the medium size. Several scatterings must happen in order to produce
one emission, which reduces the total yield compared to the
Bethe-Heitler regime (this relative suppression is known as the
Landau-Pomeranchuk-Migdal effect). In this regime, the emission
spectrum behaves as
\begin{align}
\frac{dI}{dE}\approx
\underbrace{\frac{dI}{dE}\Bigg|_{\rm single}}_{\alpha_s\,E^{-1}}\times\underbrace{ \frac{L}{t_f}}_{\sqrt{\omega_c/E}}\sim E^{-3/2}.
\end{align}
(For $E_{\boldsymbol k}< \omega_{BH}$, the factor $L/t_f$ was replaced by the
constant $L/\lambda$.)
\item $\omega_c < E_{\boldsymbol k}$. In this regime, the formation time of the
gluon is larger than the size of the medium, and induced radiation
is effectively suppressed. One may thus view $\omega_c$ as an upper
limit of the energy of a radiated gluon (for instance, for
$\widehat{q}=2$~GeV/fm${}^2$ and $L=5$~fm, this cutoff is
$\omega_c=50$~GeV). By combining the emission spectra in the three
regimes, we can obtain the mean radiated energy by a parton:
$\big<E_{\boldsymbol k}\big>\propto \alpha_s\omega_c$. In other words, energy
loss is most often due to a single emission (occurring with probability $\sim\alpha_s$) close to the upper
limit. Note that a more dilute (or cooler) medium has a smaller
$\widehat{q}$ and therefore a smaller cutoff energy $\omega_c$, and
a smaller mean energy loss.
\end{itemize}
Note that two approaches have been used in the literature for
implementing the scattering-induced radiation of a hard parton. One
option is to perform an opacity expansion, starting from a dilute
medium
\cite{Gyulassy:1999zd,Gyulassy:2000fs,Gyulassy:2000er,Vitev:2002pf,Gyulassy:2003mc}. In
this approach, a very small number of scatterings are considered, but
treated with exact kinematics. Another option is to resum multiple
scatterings, in an approximation where the momentum transfer in these
scatterings is soft compared to the parton momentum, more in line with
the BDMPS-Z approach (see for instance
\cite{Salgado:2003gb,CasalderreySolana:2007zz}).
\begin{figure}[htbp]
\centering
\includegraphics[width=7cm]{2017-May-22-RaavsNpart_Dmes8to16_Pions8to16_FinalNonPromptJpsi2017_CC_25042017}
\caption{\label{fig:charm-suppression}Open charm and non-prompt
$J/\Psi$ nuclear suppression factor, compared to that of pions. From \cite{Adam:2015nna}.
}
\end{figure}
Concrete implementations in the context of heavy ion collisions are of
course more complex than this qualitative discussion
suggests. Firstly, one needs a realistic modeling of the medium and
its evolution, usually taken from hydrodynamics. Another source of
difficulty lies in the fact that the momentum scales characteristic of
the medium (for instance, the Debye screening mass) can be rather soft
at the temperatures reached in heavy ion collisions at present
energies, and in particular not large enough to confidently apply
perturbative QCD at those scales. This has motivated hybrid
approaches \cite{Casalderrey-Solana:2014bpa} that combine perturbative
techniques for the hard scales and holography-inspired inputs for the
softer medium scales (note that a pure strong coupling approach leads
to an energy loss proportional to $L^3$ instead of $L^2$, and is ruled
out by data).
Let us close this subsection by mentioning that heavy quarks such as
charm are also suppressed in heavy ion collisions, as one can see on
the figure \ref{fig:charm-suppression}. One can see that $D$ mesons
are as suppressed as the charged pions, suggesting that the
fragmentation of $c$ quarks (at the temperature scales relevant in the
LHC experimental conditions) is comparable to that of light quarks and
gluons. The quark mass starts playing a visible role for $b$ quarks,
since the medium suppression of mesons containing $b$ quarks is
significantly less important. This is consistent with the {\sl dead
cone effect} \cite{Dokshitzer:2001zm}, a kinematical effect that
prevents gluon radiation inside a cone of opening $m/E$ centered on
the emitting quark of mass $m$. Note that, since the radiative losses
are suppressed by the dead cone effect for heavy quarks, their energy
loss through elastic scatterings (collisional energy loss) become
relatively more important and must be included.
\paragraph{Vacuum antenna pattern} Until this point, we have discussed
the medium modification of the spectra of single hadrons. Another
possibility is to consider similar observables for jets instead of
individual hadrons. One advantage of jets is that jet cross-sections
are in principle perturbative since they do not rely on the details of
the hadronization process. Moreover, jets provide another handle to
probe the loss of energy due to the surrounding medium, since their
opening angle may be chosen at will in their definition.
Before we discuss jet modifications in heavy ion collisions, let us
recall the main feature of gluon radiation from a jet in vacuum.
Compared to radiation by a single parton, the radiation from a jet is
obtained by first considering a parton splitting process $a\to bc$
(where $a,b,c$ could be quarks, antiquarks, gluons, photons). Since
the partons $b,c$ are produced from a common ancestor $a$, their
colors are correlated, and so is the soft radiation they produce. In
particular, the emission of an additional soft gluon after the
splitting $a\to bc$ depends crucially on whether the new gluon is
emitted inside or outside of the cone formed by the partons $b$ and
$c$. This effect can be understood semi-quantitatively as follows in
the case where the angular opening $\theta_{bc}$ between $b$ and $c$
is small. As we have seen earlier, the formation time of an additional
gluon of energy $E_{\boldsymbol k}$ at an angle $\theta$ is given by
$t_f^{-1}\approx E_{\boldsymbol k} \theta^2$. By this time, the pair $bc$ has grown
to a transverse size
\begin{align}
r_\perp = t_f \theta_{bc}.
\end{align}
Moreover, the gluon wavelength ($E_{\boldsymbol k}^{-1}$), projected on the plane
orthogonal to the pair momentum, is
$\lambda_\perp \approx (E_{\boldsymbol k} \theta)^{-1}$. When the transverse
wavelength of the emitted gluon is smaller than the size of the pair,
the gluon resolves the individual constituents $b$ and $c$ of the pair
and the emission is the sum of the separate emissions from $b$ and
$c$. Otherwise, the gluon cannot resolve the pair, and sees only its
total charge, which is the charge of the parent $a$. This condition
reads
\begin{align}
\frac{1}{E_{\boldsymbol k} \theta}\lesssim r_\perp\quad
\Leftrightarrow
\quad
\theta\lesssim \theta_{bc}.
\end{align}
In other words, inside the cone, the emission is the incoherent sum of
the emissions from $b$ and $c$, while the emission outside the cone is
equal to that of the total charge, i.e., $a$. In particular, if the
ancestor is a color neutral object (a virtual photon), then there is no
soft radiation outside the cone of the pair. This property is known as
{\sl angular ordering} in the context of QCD. It can be used in order
to formulate gluon emission in the form of ``parton cascades'' in
which each quark, antiquark or gluon has a certain probability of
emitting a new gluon. The effect of quantum interferences, that would
in principle preclude such a probabilistic description, is taken into
account by vetoing emissions at large angles.
\paragraph{In-medium jet energy loss}
A striking observation regarding jets is that
there are events with very imbalanced pairs of jets in heavy ion
collisions, as shown in the figure \ref{fig:jet-quenching}.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{eventDisplay_151076_1328520_200GeV-70GeV}
\caption{\label{fig:jet-quenching}Event display of a very asymmetrical two-jet event. From \cite{Chatrchyan:2011sx}.}
\end{figure}
In proton-proton collisions, events with two jets in the final states
have jets with nearly balanced transverse momenta, as imposed by
momentum conservation. In the above event display, one jet has almost
three times the energy of the second jet (or, put differently, about
$130$~GeV appear to have been ``lost''), and it is not possible to see
directly on this histogram where the energy has gone (the only
conclusion one may draw by eye is that this energy has been degraded
in the form of soft enough fragments that are lost in the background
of this calorimetric display).
A first medium effect affecting jets is the fact that each parton in
the jet may suffer from medium-induced radiation, in the way discussed
earlier for single particle spectra. The only modification imposed by
the jet definition is that only the radiation that goes outside of the
jet cone must be counted as a loss (thus, this effect is more
pronounced for narrow jets than for wider jets). We have seen in the
previous part that the mean radiative energy loss of a parton is due
to emissions close to the maximum $\omega_c= \widehat{q} L^2$. For
those, the typical gluon emission angle is
$\theta \approx k_\perp/E_{\boldsymbol k}\approx (L\omega_c)^{-1/2}\ll 1$. The
consequence of this observation is that these rather hard induced
emissions do not alter the energy of a jet for typical jet opening
angles (usually $0.3\le R\le 0.8$), since the radiated gluon stays
within this cone (the jet substructure is altered, but this is not
picked by the jet reconstruction algorithm). Therefore, the
observed asymmetry between the energy of the pair of jets can be
explained by softer gluons emitted at large angles.
Another important effect affecting a jet as it propagates through a
medium is that the multiple scatterings of the constituents of a pair
of partons eventually lead to the loss of their color coherence
\cite{CasalderreySolana:2012ef,Mehtar-Tani:2013pia,Blaizot:2013vha,Blaizot:2015lma,Qin:2015srf,Milhano:2015mng,Mehtar-Tani:2016aco}. Thanks to this decoherence, the
vacuum-like emissions are no longer forbidden outside of the jet cone.
This happens when the two partons scatter off external color fields
that are uncorrelated, i.e., when the transverse separation $r_\perp$
between them is larger than the coherence length of the color
field. Since $r_\perp$ is proportional to the opening angle of the
pair, the decoherence time is shorter for a pair with a larger opening
angle. The conclusion of this qualitative argument is that small jets
are more robust against in-medium vacuum-like energy loss than wider
jets.
Moreover, it has been shown that the energy emitted outside of the jet
cone is rapidly degraded into partons whose typical energy is close to
the temperature of the surrounding medium \cite{Blaizot:2012fh,Blaizot:2013hx,Blaizot:2013vha,Caucal:2018dla}. Experimentally, this has
been qualitatively confirmed, since one recovers the missing jet
energy in the form of many softer particles outside of the jet cone.
\paragraph{$\gamma,W^\pm,Z$-jet correlations}
Because of momentum conservation, jets are dominantly produced in
pairs in hard processes, and more rarely (with a suppression of order
$\alpha_s$) in a $3$-jet configuration. After their initial
production, the fate of these jets in the surrounding medium depends
crucially on the location of the production point with respect to the
bulk, and on the direction of motion of the jets. Loosely speaking,
a jet that goes inward looses more energy than a jet that travels
outwards, because it must travel a longer route through the medium. Except
when the production point is very close to the outer boundary of the
medium, both jets loose energy to some degree, making it difficult to
infer event-by-event how much energy was lost by each of them.
A more direct access to the jet energy loss is possible in situations
where a single jet is production in conjunction with another object
that interacts only via electroweak interactions, such as a photon, a
$W^\pm$ or a $Z^0$ boson (for instance, in a process such as
$qg\to q\gamma$). These events are less frequent because of the
electroweak coupling involved at the production vertex, but they offer
the advantage that the weakly interacting object can escape from the
medium without further interactions. Thus, its measurement provides an
unaltered reference for the initial energy of the partner jet.
\section{Thermometric probes}
In this last section, we consider observables that are sensitive to
the local temperature of the medium. In principle, these quantities
could tell if the temperature reached in a heavy ion collision is
above the deconfinement temperature. The actual situation is of course
a bit more complicated, since the temperature of the medium is not
spatially homogeneous, and because the measured quantities result from
the entire history of the system, through which the temperature is
not constant.
\subsection{Electromagnetic radiation from the quark-gluon plasma}
\paragraph{General considerations}
A first quantity which is quite sensitive to the plasma temperature is
the spectrum of photons emitted by the plasma. Let us clarify here a
possible paradox: in a large medium (larger than the photon mean free
path), the electromagnetic radiation would be in thermal equilibrium
with the quarks and the gluons, with a spectrum given by a
Bose-Einstein distribution at the local temperature (but only photons
emitted within one mean free path of the surface would
escape). However, this is not the case in heavy ion collisions, where
the size of the medium is considerably smaller than the mean free path
of the photons. In this case, there is a net production of photons,
and they escape from the medium without further interactions.
Experimentally, the interpretation of photon measurements is quite
challenging, because the detected photons can come from several
sources. It is customary to divide the observed photons into decay
photons (produced from the decay of light hadrons, predominantly
neutral pions) and direct photons (produced directly from partonic
interactions). Direct photons themselves have several sources: some
are produced in hard partonic collisions at the time of the impact of
the two nuclei, some are produced by the pre-equilibrium medium, some
are produced by the interaction of a hard parton and the medium it
traverses \cite{Turbide:2005fk}, some are produced by the QGP and some are produced by the
hot hadron gas after the confinement transition. Among all these
sources, the photons produced by the QGP and by the hot hadron gas are
the most directly sensitive to the properties of the medium produced
in heavy ion collisions, but disentangling them unambiguously from the
overall spectrum is nearly impossible. In the region of low to
intermediate photon energies, the observed spectrum has an exponential
shape in $\exp(-p_\perp/T_{\rm eff})$, suggestive of emissions by a
thermalized medium. However, as we shall see, the parameter
$T_{\rm eff}$ does not have the direct interpretation of the
temperature of the producing medium.
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=70mm]{Figures_DirGammaSpectrum_LinX_withNLO_withoutFit}
\end{center}
\caption{\label{fig:gam_alice}Direct photon spectrum observed by the
ALICE experiment, compared to pQCD NLO predictions. Red:
exponential fit of the excess. From \cite{Adam:2015lda}.
}
\end{figure}
Note also that the produced photons can either be real or virtual. In
the latter case, they can subsequently decay into a lepton pair (this
may be their only decay channel for low invariant masses). Although
the decay into a lepton pair requires another electromagnetic coupling
and thus reduces the yield, the invariant mass of the pair provides
another handle, that may be used to better constrain their possible
source (in particular, a non-zero invariant mass may be used to select
a region where backgrounds are lower).
\paragraph{Thermal radiation from the QGP} The lowest order processes
($qg\to q\gamma$, $q\overline{q}\to g\gamma$
\cite{Baier:1988xv,Altherr:1988bg,Altherr:1992th} and
$q\bar{q}\to \gamma^*$ \cite{McLerran:1984ay}) have been calculated
long ago in an equilibrated quark-gluon plasma. For real photons, the
processes with a quark or antiquark exchanged in the $t$-channel have
a logarithmic singularity, which is cured by resumming the appropriate
quark hard thermal loop \cite{Kapusta:1991qp,Baier:1991em}. However,
it was soon realized that bremsstrahlung processes, formally of higher
order in $\alpha_s$, suffer from a more severe soft singularity when
the photon has a small invariant mass
\cite{Aurenche:1996is,Aurenche:1996sh,Aurenche:1998nw,Aurenche:2002pd}. This
singularity is regularized by the quark in-medium effective mass, but
this leads to an enhancement that promotes bremsstrahlung to the same
order in $\alpha_s$ as the leading processes.
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=75mm]{photon_1}
\hskip 10mm
\includegraphics[width=47mm]{photon_2}
\vskip 10mm
\includegraphics[width=40mm]{photon_4}
\end{center}
\caption{\label{fig:photons} Top left: Leading Order processes for the
production of virtual and real photons. Top right: Enhanced
Next-to-Leading Order bremsstrahlung processes. Bottom: multiple
scatterings that contribute coherently to the emission of a
photon.}
\end{figure}
This enhancement also occurs for multiple scattering corrections to
bremsstrahlung \cite{Aurenche:2000gf}. Like in the discussion of
medium induced gluon radiation by a parton, the emission of a photon
by a quark or antiquark is affected by the Landau-Pomeranchuk-Migdal
effect. The relevant criterion is the comparison between the photon
formation time (i.e., how much time is necessary for the photon
wave-packet to be sufficiently separated from that of the emitter,
which can be estimated to be $t_f^{-1}=E_{\boldsymbol k}({\boldsymbol k}_\perp^2+m^2)/E_{\boldsymbol p}^2$
where $k$ is the photon momentum, $p$ the quark momentum and $m$ its
in-medium effective mass) and the mean free path $\lambda$ between two
soft (this is sufficient to induce the emission of a photon)
scatterings of a quark in the medium. When $t_f \gtrsim \lambda$,
several scatterings are necessary to induce one emission, which
effectively reduces the photon yield. Note that, in the weak coupling
regime, we have $m\sim gT$ and $\lambda^{-1}\sim g^2 T$ (up to
logarithms). Therefore, the LPM effect plays a role in two cases: for
the production of hard photons emitted at small angle with respect to
the quark, or for the production of soft photons. The resummation of
these multiple scattering diagrams gives the photon yield at leading
order
\cite{Arnold:2001ba,Arnold:2001ms,Arnold:2002ja,Aurenche:2002wq}. The
next-to-leading order correction has also been calculated more
recently \cite{Ghiglieri:2013gia,Laine:2013vma}.
There have also been attempts to extract the photon production rate of
a quark-gluon plasma in thermal equilibrium from lattice QCD
computations
\cite{Karsch:2001uw,Karsch:2002wv,Ding:2010ga,Ding:2016hua,Ghiglieri:2016tvj,Brandt:2017vgl},
which would in principle be applicable in regimes where the coupling
constant may not be small enough for perturbation theory to be
reliable. However, since this amounts to computing a spectral function
for real energies, there is no direct way to obtain it from an
Euclidean lattice formulation. Instead, one can reach it indirectly by
unfolding (this is an ill-posed problem, that may be attacked with
Bayesian approaches such as the maximal entropy method) the spectral
representation of imaginary time correlation functions.
In the context of an actual heavy ion collision, one has also to face
the fact that the system may not be in local thermal equilibrium. This
is especially true at early times. Firstly, the CGC predicts at the
beginning that there are very few quarks compared to the gluons,
implying that the system is not yet in chemical equilibrium
\cite{Berges:2017eom}. It is possible to handle approximately this
situation by introducing fugacities for the quark and antiquark
distributions, in order to obtain the local photon production rates
for a system where the quarks are underpopulated
\cite{Gelis:2004ep}. Moreover, even when quarks and gluons are present
in the right proportions, the existence of viscous hydrodynamical
corrections implies that their distributions cannot be the equilibrium
ones \cite{Shen:2014nfa,Hauksson:2017udm,Schafer:2019edr}. These deviations should in principle be
taken into account in order to be consistent with the hydrodynamical
framework used to describe the bulk evolution of the system.
\paragraph{Thermal radiation from a hot hadron gas} Thermal radiation
from a hot hadron gas is also obtained from the current-current
correlation function, but now it is not accessible to a calculation in
terms of the QCD Lagrangian that has quark and gluon degrees of
freedom. One may instead use an effective Lagrangian that describes
the dynamics and interactions among hadrons at low energy (Such a
Lagrangian in general contains some parameters that are constrained by
empirical data on lifetimes, cross-sections, etc...)
\cite{Gale:1987ki,Rapp:1997fs,Rapp:1999zw,Rapp:2000pe,vanHees:2007th,Rapp:2013nxa,Vujanovic:2013jpa}.
For dileptons of low mass, the spectral function in the vector channel
is dominated by light vector mesons, and is therefore particularly
sensitive to thermal modifications of the $\rho$ meson. The main
thermal effect is a broadening of the $\rho$ peak (while the center of
the peak does not change appreciably), which is even more pronounced
when baryons are taken into account. Such a broadening seems in quite
good agreement with dilepton measurements in the CERES and NA60
experiments (see Figure \ref{fig:dileptons}).
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=65mm]{na60-1}
\hskip 1mm
\includegraphics[width=65mm]{na60-2}
\end{center}
\caption{\label{fig:dileptons}Left: dilepton mass spectrum in the $\rho$ region from the NA60 experiment. Right: invariant mass spectrum of the excess dimuons, integrated over $p_T$. From \cite{Specht:2010xu}.}
\end{figure}
Let us also mention that, besides effective field theories, a more ab
initio approach has been used recently in order to extract in-medium
spectral functions
\cite{Tripolt:2016cey,Jung:2016yxl,Tripolt:2018jre,Fu:2019hdw,Jung:2019nnr}, based on the
functional renormalization group (FRG) \cite{Berges:2000ew}. The FRG
is a functional equation that tracks the evolution of the quantum
effective action of a theory as one integrates out the quantum
fluctuations in successive layers of momentum, the starting point
being the classical action of the theory (i.e., with no quantum
fluctuations included). In other words, the FRG is an explicit
realization of the renormalization group ``a la Wilson'', where a
theory is coarse-grained to eliminate its details on short distance
scales. In its exact form, there is no practical way to solve the FRG
equation, but it is amenable to tractable equations (at least
numerically) after some kind of truncation has been performed. When
applied to the study of the chiral transition, one observes a
temperature dependence of the chiral condensate in quite good
agreement with lattice QCD computations, and that the $\rho$ and $a_1$
mesons become degenerate at high temperature.
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=65mm]{FRG_chi_cond}
\hskip 1mm
\includegraphics[width=65mm]{FRG_rho_a1}
\end{center}
\caption{\label{fig:FRG}Left: FRG computation of the chiral
condensate as a function of temperature, compared to a lattice QCD
calculation (From \cite{Fu:2019hdw}). Right: temperature dependence of the masses of the $\rho$
and $a_1$ mesons, computed in the FRG framework (From \cite{Jung:2016yxl}).}
\end{figure}
\paragraph{Folding with the medium evolution} The calculations
described above provide a local photon production rate (i.e., the
number of photons produced per unit of time and volume, given the
local temperature of the plasma). To go from there to a photon
spectrum that one may compare with experimental data, this local rate
must be integrated over the entire space-time history of the collision
(in addition, at each space-time point, one must boost the local
spectrum by the $4$-velocity of the plasma at this point). The
important point to keep in mind is that, even if a fit of the
resulting spectrum by an exponential of the form
$\exp(-p_\perp/T_{\rm eff})$ appears to work, the parameter
$T_{\rm eff}$ does not have a direct interpretation as the temperature
of the plasma. In particular, this effective temperature is
blue-shifted by the radial flow of the medium. Moreover, the
integration over time is also more sensitive to the stages of the
evolution where the system spends more time. With dileptons, one may
avoid this blue-shifting effect when looking at the spectrum as a
function of the pair invariant mass since the mass is not affected by
the flow, but the extracted temperature still reflects a spacetime
average rather than an instantaneous temperature. With this caveat in
mind, an exponential mass dependence $\exp(-M/T_{\rm eff})$ of the
dilepton spectrum has been reported by the NA60 experiment, with an
effective temperature $T_{\rm eff}\approx 205\pm 12$~MeV \cite{Specht:2010xu}.
\subsection{Heavy quarkonia in a hot medium}
\paragraph{Qualitative aspects} Bound states made of heavy quarks can
also be viewed as potential ``thermometers'' \cite{Matsui:1986dk}. On
the theory side, the advantage of considering sufficiently heavy
quarks is that they provide a large mass scale (much larger than the
QCD non-perturbative scale $\Lambda_{\rm QCD}$), and this separation
of scales allow the use of effective field theory descriptions such as
non-relativistic QCD. In such a non-relativistic framework, one may
use the concept of interaction potential between a pair of heavy
quarks, in conjunction with a non-relativistic Schr\"odinger
equation. Another theoretical simplification regarding heavy quarks is
that their production happens at very early times (of the order of the
inverse of their mass), and involve parton distributions at reasonably
large momentum fractions, where saturation effects are not important
(this assertion should be contrasted in the case of charm quarks at
the LHC energy -- see \cite{Ma:2015sia,Ma:2018qvc} for a recent study
of $J/\Psi$ production in proton-nucleus collisions; a similar
computation in nucleus-nucleus collisions could be done by solving the
Dirac equation in the glasma color fields, but is considerably more
challenging \cite{Gelis:2005pb,Gelis:2015eua,Tanji:2017xiw}). Thus, in heavy ion
collisions, one is mostly interested in the subsequent fate of the
produced heavy quarks, rather than the production itself. As far as
experimental measurements are concerned, heavy quarkonia also offer clean
signals via their dilepton decay channel.
Consider for instance a heavy meson $Q\overline{Q}$. Loosely speaking,
when the thermal excitation energy (i.e., the energy gain provided by
the absorption of a gluon from the surrounding thermal bath) is equal
or larger to the binding energy of the quark-antiquark pair, the meson
may be dissociated. An alternate way of describing this phenomenon is by
noting that the interaction potential of the $Q\overline{Q}$ is
affected by Debye screening in the presence of a dense medium. When
the Debye screening length becomes shorter than the size of the
would-be bound state, its dissociation occurs. Afterwards, the quark
and the antiquark evolve independently in the medium, and the most
likely outcome is that --when the temperature has decreased below the
confinement temperature-- they bind with one of the light quarks or
antiquarks from the surrounding medium, in order to form heavy-light
mesons ($D$ and $B$ mesons). Therefore, in the extreme version of this scenario, the yield
of $Q\overline{Q}$ mesons would be almost completely suppressed and
the produced heavy quarks would all be recovered in the form of
open-flavour mesons. Note that in the case of charm quarks, whose
production is quite abundant in heavy ion collisions at high energy,
the dissociated $c$ and $\overline{c}$ may have a high enough density
for accidental recombinations in $J/\Psi$'s to be important.
Experimentally, a particularly clear observation of this phenomenon
has been performed at the LHC for $b\overline{b}$ mesons, as shown in
the figure \ref{fig:upsilon}.
\begin{figure}[htbp]
\centering
\includegraphics[width=8cm]{Upsilon_CMS}
\caption{\label{fig:upsilon}Muon-antimuon invariant mass spectrum in the upsilon region. From \cite{Sirunyan:2018nsz}. See also \cite{Chatrchyan:2012lxa,Khachatryan:2016xxp}.}
\end{figure}
There, one sees a suppression of the yields of the $\Upsilon''$ and
$\Upsilon'$ mesons, while the ground state $\Upsilon$ has an almost
unmodified yield. Given the above discussion, the interpretation
of this result is that the gluon thermal energy at the temperatures
reached in the collision is sufficient to dissociate the excited
states (because they are more weakly bound) but not enough to alter
the more tightly bound fundamental state.
\paragraph{In-medium singlet heavy quark bound states}
In the vacuum (i.e., at zero temperature), the spectrum of color
singlet $Q\overline{Q}$ states is typically made of a few (meta)stable
states --defined as poles of the propagator in the energy complex
plane-- occupying the low energy region, and a continuum at higher
energies. In the presence of a high temperature medium, several
effects may occur:
\begin{itemize}
\item Some bound states (starting from the high lying ones) may disappear,
\item The lower bound of the continuum of free states may move to a
lower energy,
\item The surrounding thermal medium can induce transitions between
various $Q\overline{Q}$ states, including transitions between
singlet and octet states.
\end{itemize}
The ab initio approach for studying the in-medium modifications of
heavy quark states consists in calculating the spectral function in
the appropriate channel at non-zero temperature. However, this is not
doable in perturbation theory due to the non-perturbative nature of
bound states. Although lattice QCD is by construction a
non-perturbative approach, spectral functions cannot be computed
directly. Instead, one can calculate an Euclidean propagator
$G(\tau,{\boldsymbol p})$ which is related to the corresponding spectral function
$\rho(\omega,{\boldsymbol p})$ as follows:
\begin{align}
G(\tau,{\boldsymbol p})
=
\int
d\omega\; \rho(\omega,{\boldsymbol p})\, \frac{\cosh(\omega(\tau-1/2T))}{\sinh(\omega/2T)}
.
\end{align}
A first obvious difficulty for inverting this relationship is that the
propagator is computed only at the finite set of Euclidean times
$\tau$ that exist in the employed lattice setup, while the expected
spectral function depends on a continuous energy $\omega$. But even if
the propagator was known at all the real $\tau$'s in the range
$[0,1/T]$, this inversion is a mathematically ill-posed problem,
because the linear mapping from $\rho$ to $G$ has zero modes (i.e.,
functions $\rho$ that give zero when inserted in the integral in the
previous equation). Thus, even in the ideal situation where the
propagator would be known exactly, the inversion can only be performed
up to a linear combination of these zero modes.
A possible strategy is to remove this ambiguity by imposing additional
(but generic enough so that they do not bias the outcome in unphysical
directions) constraints on the expected spectral function. A minimal
constraint that helps disambiguate the answer is to request the
positivity of the spectral function. In practice, this can be
implemented by using the {\sl maximal entropy method}, which is a
Bayesian method for finding the most likely spectral function
consistent with the computed values of the propagator and the
additional constraints. When using this approach, it is necessary to
have very accurate lattice data for a robust extraction of the
spectral function (otherwise, the extracted spectral function may be
dominated by the additional constraints imposed on the
solution). Another limitation is that it is practically impossible to
be sensitive to excited states, as this would require an exponentially
large statistics. With these caveats in mind, the general trend
observed for heavy bound states is a sequential melting of states,
starting with the high lying ones, and a trend towards negative
medium-induced mass shifts \cite{Burnier:2013nla,Aarts:2014cda,Ikeda:2016czj,Kim:2018yhk}.
An alternative to Bayesian methods is to model the spectral functions
$\rho(\omega,{\boldsymbol p})$ with a few free parameters and to perform a standard
fit to adjust these parameters in order to reproduce the computed
propagator $G(\tau,{\boldsymbol p})$. Besides the location and width of the lowest
lying peak, the model may contain parameters that describe the
transport properties of heavy quarks, or excited states. However, one
should keep in mind that the propagator may be very weakly sensitive
to these additional features of the spectral function
\cite{Aarts:2002cc}, and that an unrealistic modeling may introduce a
strong model dependence on the outcome. A model of the spectral
function may be obtained from effective field theory descriptions,
such as non-relativistic QCD (NRQCD -- obtained from QCD by
integrating out the heavy quark mass scale $m_{_Q}$) or even potential
non-relativistic QCD (pNRQCD -- obtained from NRQCD by further
integrating out the softer scale $m_{_Q}v$, where $v$ is the heavy
quark velocity)
\cite{Brambilla:1999xf,Brambilla:2004jw,Brambilla:2008cx}, that rely
on the heavy mass of the quarks, both compared to the QCD
non-perturbative scale $\Lambda_{\rm QCD}$ and to the typical energy
scale of the surrounding medium.
In fact, the pNRQCD effective theory also provides a proper connection
between QCD and the non-relativistic Schr\"odinder equation used in
order to study the $Q\overline{Q}$ bound states (see for instance
\cite{Laine:2006ns,Beraudo:2007ky}). Indeed, one of the parameters in
pNRQCD is the $Q\overline{Q}$ interaction potential (that should in
principle be obtained by a matching to the underlying field theory),
and its equation of motion in the singlet sector is a Schr\"odinder
equation whose discrete energy levels correspond to the
$Q\overline{Q}$ singlet bound states.
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=65mm]{QpotentialRe}
\includegraphics[width=65mm]{QpotentialIm}
\end{center}
\caption{\label{fig:Qpotential}Real and imaginary parts of the
singlet $Q\overline{Q}$ potential at various temperatures. From
\cite{Burnier:2014ssa}. See also \cite{Lafferty:2019jpr}.}
\end{figure}
The singlet potential $V_s(r)$ can be obtained as
\begin{align}
V_s(r)=\lim_{t\to \infty}\frac{\partial_t W(r,t)}{W(r,t)},
\end{align}
where $W(r,t)$ is a rectangular Wilson loop of spatial extent $r$ and
temporal extent $t$. Note however that $t$ is here the Minkowski time,
and a direct evaluation of this quantity in lattice QCD is therefore
not possible. A possible strategy is to start from an Euclidean
rectangular Wilson loop, whose associated spectral function can be
used to express $V_s(r)$, thus allowing to constrain the potential via
Bayesian methods as shown in Figure \ref{fig:Qpotential}. Note that
the real part of this potential behaves similarly but is not identical to the
potential sometimes inferred from the logarithm of the free energy of a
singlet $Q\overline{Q}$ pair \cite{Borsanyi:2015yka}. (The imaginary part, due to Landau
damping and transitions from singlet to octet states, is not present
in the potential defined from the free energy.)
\paragraph{In-medium dynamical evolution}
In the presence of a high temperature medium, not only the spectrum of
singlet quarkonia states is modified, but the surrounding medium can
also induce transitions between various singlet states, and between
singlet and octet states. In equilibrium, there are as many
transitions in either direction, and the density matrix of the system
is time independent. The situation is far more complicated
out-of-equilibrium, since the density matrix is now time
dependent. Various forms of kinetic or stochastic equations have been
employed to describe the evolution of heavy quarks and quarkonia
embedded in a medium \cite{Moore:2004tg,Ke:2018tsh,Yao:2018zrg}.
The main physical effect one would like to study here is the possible
recombination of the heavy quarks and antiquarks, that may have a
sizeable probability of close encounter when their density is large
enough (which appears to be the case for charm quarks). When this is
the case, the final yield of quarkonia bound states is enhanced
compared to what one would get with the assumption that all the bound
states that are dissociated go into open-heavy flavor mesons
\cite{Thews:2000rj,Blaizot:2015hya,Du:2017qkv,Yao:2017fuc,Yao:2018sgn}.
\begin{figure}[htbp]
\begin{center}
\includegraphics[width=80mm]{RAACent1}
\end{center}
\caption{\label{fig:reco}Nuclear modification factor of $J/\psi$ as
a function of the number of participants, for RHIC and LHC
energies. From \cite{Adam:2015isa}.}
\end{figure}
This phenomenon can be seen for charm quarks by comparing the nuclear
modification factor of $J/\psi$ at RHIC and LHC energies, as shown in
Figure \ref{fig:reco}. On this plot, one sees that $J/\psi$'s are less
suppressed in central collisions at higher energy, which can be
explained by the more abundant production of $c$ and $\overline{c}$
quarks at the LHC, which favors their recombination into charmonium.
More recently, there has been an effort to derive these description
from a more fundamental starting point, shedding some light on the
sequence of approximations necessary to obtain them. The strategy,
based on the framework of {\sl open quantum systems}
\cite{Akamatsu:2014qsa,Kajimoto:2017rel,Yao:2018nmy,Brambilla:2017zei,Brambilla:2016wgg,Blaizot:2018oev},
consists in starting from the system made of the $Q\overline{Q}$ pairs
and their environment. This system is closed and evolves unitarily,
with a density matrix $\rho$ that obeys the von Neumann equation (also
known as the quantum Liouville equation),
\begin{align}
\frac{d\rho}{dt}=-i\,[H,\rho].
\end{align}
The density matrix $\rho$ contains a lot of information regarding the
surrounding medium in which the heavy quarks are embedded. A {\sl reduced
density matrix} describing specifically the heavy quark degrees of
freedom can be obtained by performing a partial trace over environment
degrees of freedom, $\rho_{_Q}\equiv {\rm tr}_{\rm
env}(\rho)$. However, after doing this, the evolution equation for
the reduced matrix density is no longer a von Neumann equation, and is
in general dissipative and non unitary (this is just a consequence of
the fact that the subsystem made of the heavy quarks is an open
system). With the assumption that the environment relaxes more quickly
than the heavy quarks, it has been shown that $\rho_{_Q}$ obeys a
Lindblad equation. Under the additional assumption that the transitions
between the various $Q\overline{Q}$ states are faster than their
adjustment to changes of the environment, this Lindblad equation can
lead to a Boltzmann equation (and a rate equation is the momenta are
integrated out).
\section{Conclusions, open questions for the future}
What should first come to mind after this brief survey of the
theoretical aspects of heavy ion collisions is the difficulty of
describing a large, time-dependent, non-equilibrated system in terms
of a (rather complicated) underlying microscopic theory (QCD). Since
an ab-initio description in terms of QCD of these collisions is not
practically feasible, most approaches are based on effective
descriptions that capture the relevant dynamics at scales larger than
the typical QCD scales. Thus, instead of a unique theory from which
everything would be derived, theoretical works in this field use
instead a large variety of tools that are more or less connected to QCD:
\vglue 10mm
\begin{center}
\xymatrix{
*+[F-]\txt{{\bf Lattice QCD}}
&
*+[F-]\txt{{\bf Perturbative QCD}}
&
*+[F-]\txt{{\bf CGC}}
\\
*+[F-]\txt{{\bf Hydrodynamics}}
&
*+[F-,]{\txt{{\Large\bf QCD}\\{\ }\\${\cal L}=-\frac{1}{4}{ F^2} +{\overline\psi}(i{\raise.15ex\hbox{$/$}\kern-.7em\hbox{$D$}}-{ m}){\psi}$}}
\ar[dl]\ar[ul]\ar[dr]\ar[ur]\ar[r]\ar[u]\ar[l]\ar[d]
&
*+[F-]\txt{{\bf NRQCD}}
\\
*+[F-]\txt{{\bf Kin. Theory}}
&
*+[F-]\txt{{\bf AdS/CFT}}
&
*+[F-]\txt{{\bf $\chi$PT}}
}
\end{center}
\vskip 1mm Although we have not spent much time discussing this, we
should also stress the fact that many observables in heavy ion
collisions depend on a number of mundane aspects of low energy nuclear
physics, namely the shape and size of the nuclei, and the distribution
of the nucleons inside a nucleus and the fluctuations thereof. These
properties, that are not the main targets of the heavy-ion collision
program, play nevertheless an important role when trying to uncover
some property of QCD from experimental data.
By a combination of experimental and theoretical efforts, many
properties of the quark-gluon plasma have been uncovered:
\begin{itemize}
\item the QGP is a nearly perfect fluid,
\item its shear viscosity to entropy ratio is in the range
$[1,2.5]$ (in units of $\hbar/4\pi$), making it the substance with
the smallest ratio so far,
\item its equation of state is consistent with lattice QCD
expectations, and with the deconfinement of the color degrees of
freedom,
\item the yield of ``light'' partons, including charm quarks, is
significantly suppressed compared to rescaled proton-proton
collisions,
\item the suppression of bottom quarks is less pronounced, in
agreement with theoretical expectations (dead-cone effect due to the
mass of the emitter),
\item the studies of energy loss can now be supplemented by direct
observations of reconstructed jets. This has allowed to determine
that a large amount of energy is radiated by soft emissions at large
angle,
\item a sequential pattern has been observed in the disappearance of
$b\overline{b}$ bound states, consistent with the theoretical
understanding of the dissociation phenomenon,
\item at the highest energies, the production of charm quarks is
copious enough to lead to the formation of $J/\psi$ bound states by
recombination of uncorrelated quarks and antiquarks.
\end{itemize}
The picture that emerges from these observations is that the matter
produced in heavy ion collisions is a very ``opaque'', strongly
interacting, fluid, in rather sharp contrast with the ethereal
quark-gluon {\sl plasma} that was the common point of view before the
RHIC experiment. Despite many progresses, it is also clear that
extracting the underlying QCD properties from the outcome of heavy ion
collisions is extremely difficult, since in several instances the
comparisons with QCD have remained rather qualitative although the
experimental measurements were quite detailed. Another source of
complication is that some of these studies are done by comparing the
outcome of nucleus-nucleus collisions with that of rescaled
proton-nucleus or proton-proton collisions. However, high energy
proton-nucleus and proton-proton collisions have turned out to display
some features close to those observed in nucleus-nucleus collisions,
casting some doubts on their use as ``references'' to compare with in
order to pinpoint effects specific to the quark-gluon plasma. If flow
is confirmed to occur even in proton-proton collisions, one will have
to learn to live without such a reference for certain observables.
It is of course hard to predict where the next advances will happen,
but given the areas that have received most attention in the past
years, some improvements are probably within reach in a reasonable
future in the following directions:
\begin{itemize}
\item Determine the temperature dependence of the shear viscosity,
\item Obtain a better determination of the bulk viscosity,
\item Better disentangle the mechanisms of energy loss, especially in the case of jets,
\item Characterize {\sl when} heavy quark bound states are formed,
\item estimate the initial temperature from thermal photons and the melting of quarkonia,
\item Clarify to what extent the concept of flow applies to the system
formed in proton-proton collisions. This entails a more robust
control over the state of the system immediately after the
collision, in order to disentangle initial flow from the flow
hydrodynamically generated later on.
\end{itemize}
For this to be possible, besides some improvements to the calculation
of the elementary relevant phenomena, an important aspect (and
difficulty) is to merge as seamlessly as possible tools that have
originally been developed independently. This is especially true for
observables that depend on the interactions between some probe and the
surrounding medium, for which it is crucial to use a modeling of the
background and its evolution which is as realistic as possible.
\paragraph{Acknowledgements:} The author is supported by the Agence
Nationale de la Recherche through the project 11-BS04-015-01.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 3,912
|
North East Weather
Spanish plume sends temperatures soaring to 12C in North East but dangerous winds to hit UK
The Met Office have predicted warmer weather this week following a stormy start to December
Sam Rkaina
Tom BeattieBreaking News Reporter
Warm morning sun silhouettes the Millennium Bridge on Newcastle's Quayside (Image: Newcastle Chronicle)
A Spanish plume is set to bring unseasonably warm temperatures to the North East- but dangerously strong winds are still on the way.
Large parts of the country will see the mercury get as high as 14C on Sunday, thanks to a blast of sub-tropical air coming up from Europe.
The warmest weather will be in the south east and south west of England, with the North East set for temperatures of up to 12C from today, The Mirror has reported.
And despite the milder temperatures today, there are two Met Office warnings in place for the very north of the country until Monday.
The first warns of potentially dangerous winds hitting northern parts of Scotland from 9pm on Sunday until 6am the next day, with a large area subject to a second warning between midnight and noon tomorrow.
Brits had best enjoy the warmer conditions while they can, as the numbers will start tumbling back down to single figures again next week.
A Met Office spokesman said: "On Sunday it's a cloudy start for most of us, with outbreaks of rain moving northwards, through Northern Ireland and into southern Scotland.
"To the south though, across England and Wales skies tend to brighten and there will be some sunny spells developing, with 13 or 14 degrees in many places - well above average for the time of year. A yellow warning is in place for parts of Scotland Sunday.
"And that expands to a larger area on Monday.
"Where we sees the sky brighten it will be very mild in the south and then it turns very stormy in the far north west later.
"It's all down to this developing area of low pressure that's moving towards the UK, undergoing a rapid deepening process as it engages with the jet stream on Sunday, bringing some very strong winds to the far north west of Scotland.
"It stays unsettled into Monday with rain in places."
Looking further ahead and with Christmas on the horizon, we should be braced for a chilly Christmas as forecasters predict frost will grip the country.
During the festive week, temperatures in the south of England and Wales will plummet on Christmas Eve while frost and fog will sweep across the country on Christmas Day.
Here is how the weather is set to look in the North East over the next week according to the Met Office:
Tonight:
Another band of mostly light rain crosses the area from the west overnight, with strong winds. Clearer periods following later, with winds moderating. Minimum temperature 5C.
Mostly dry with bright spells, especially toward the north. Cloudier in the far south with a chance of some rain for a time. Maximum temperature 8C.
Outlook for Tuesday to Thursday:
Tuesday and Wednesday generally cloudy with some patchy light rain or drizzle, especially for the Pennines and Cheviots. Mild, and windy at times. Thursday, becoming generally drier with lighter winds.
For the latest local news in your area direct to your inbox every day, go here to sign up to our free newsletter
120-year-old shipwreck emerges on South Tyneside beach after Storm Arwen
Families 'living like squatters' in their own homes following Storm Arwen chaos
Newcastle Weather
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 1,076
|
\section{Introduction}
HgCr$_2$Se$_4$ is a magnetic semiconductor that forms part of the larger group of chromium chalcogenides, which are spinels with formula ACr$_2$X$_4$. The ferromagnetic interactions in HgCr$_2$Se$_4$ are so strong that the antiferromagnetic exchange, often dominant for the spinel structure, is overwhelmed and ferromagnetism is observed \cite{Baltzer1966}.
Recently HgCr$_2$Se$_4$ has attracted attention when doped with $n$-type impurities, both as a half metal and potential Chern semimetal \cite{Guan2015}, exhibiting unconventional low-temperature transport properties \cite{Yang2019}, and due to the colossal magnetoresistance (CMR) effect, which is observed in the vicinity of the magnetic ordering temperature $T_C \simeq 105$\,K \cite{Lin2016}. The magnetic ordering drives an insulator-to-metal transition. Whilst the bulk magnetic properties of the system are well characterized, with magnetization measurements demonstrating it has critical behavior very close to that of an ideal 3D Heisenberg magnet \cite{Lin2016}, the spin fluctuations are not so well studied. Understanding these fluctuations is particularly important, as spin correlations have been proposed to play an important role in the transport properties not only near the critical point but also for a wide range of temperatures $T_C < T < T^\ast$ above the ferromagnetic transition in the paramagnetic phase \cite{Lin2016}. Likewise, charge fluctuations which are often linked to magnetically-driven electronic phase separation and magnetic cluster formation, and which frequently are invoked as ingredients of models explaining the CMR effect, have not been investigated.
Nanoscale ferromagnetic clusters, magnetic polarons (MP), may form due to the large exchange interaction between charge carriers and localized spins --- and become enhanced by spin correlations, an effect that has been largely neglected in previous studies. Materials prone to the formation of such nanoscale spin-charge composites are systems with low carrier density where a localized carrier (non-ionized donors, in HgCr$_2$Se$_4$ presumably Se vacancies) polarizes the surrounding spins in the lattice of magnetic ions \cite{Molnar2007}. A percolation transition of MPs has been suggested to cause the observed CMR effect in a variety of materials \cite{Majumdar1998,Molnar2007}, as e.g.\ in perovskite manganites \cite{Teresa1997,Faeth1999,Adams2000,Tokura2006}, EuB$_6$ \cite{Nyhus1997,Suellow1998,Suellow2000,Urbano2004,Yu2006,Zhang2009}, and also in the family of HgCr$_2$Se$_4$ \cite{Yang2004,Lin2016,Li2018,Xia2019}. Here, a change of electronic transport regimes and the deviation of the magnetic susceptibility from the Curie-Weiss law at a temperature $T^\ast \sim 2\,T_C$ is associated with the formation of isolated MPs which percolate upon lowering the temperature \cite{Lin2016}. A continuous network is formed at the percolation threshold which upon cooling the sample is reached in the vicinity of the magnetic transition or may even coincide with $T_C$. The trapped carriers forming the MPs then become unbound in an external or internal magnetic field which results in a drastic increase of their mobility giving rise to the CMR behavior.
In HgCr$_2$Se$_4$, besides the spin fluctuations, the fluctuation dynamics in the charge sector and the mutual, magnetoelectric coupling remains to be explored. Since for these effects vastly different time scales are expected, we performed muon-spin relaxation ($\mu$SR) experiments which can be used to study dynamic magnetism at frequencies within the range
$\sim 10^5 - 10^{11}$\,Hz and resistance noise measurements for studying the charge carrier dynamics at low frequencies $\sim 10^{-2} - 10^2$\,Hz.
We find coinciding characteristic temperature scales suggesting that the magnetic dynamics in this system change over a wide range of frequencies. The dynamics and the observed coupling of the magnetism and the electronic degrees of freedom is consistent with the picture of isolated and diluted magnetic polarons forming at temperatures far above and coalescing near $T_C$, and a network of merged polarons below. In particular, we provide evidence for the previously suggested hypothesis \cite{Lin2016} that in HgCr$_2$Se$_4$ the effective size of the magnetic polarons is dependent on the spin correlation length.
\section{Experimental Details}
\paragraph*{Sample growth}
HgCr$_2$Se$_4$ crystals were grown by chemical vapor transport method using CrCl$_3$ as the main transport agent \cite{Arai1973,Wang2013}. In a typical run, HgCr$_2$Se$_4$ powder (which was synthesized with Hg drops, Cr powder and Se shots) was used as the precursor. The HgCr$_2$Se$_4$ powder and CrCl$_3$ flakes were loaded into a silica ampule under argon, then evacuated to a pressure below 1\,Pa and sealed quickly by flame. The ampule was subsequently exposed to a temperature gradient from 680 to 660$^\circ$C for 2 weeks and naturally cooled down to room temperature. Crystals with sizes 1 -- 3\,mm with regular shapes and shiny facets were collected at the cooler end of the silica tube. The carrier density in the HgCr$_2$Se$_4$ crystals can be adjusted by the growth temperature, the amount of transport agent, and the stoichiometry of source materials. Note that the growth procedure is the same as in \cite{Yang2019} but different from the one in \cite{Lin2016}.\\
\paragraph*{$\mu$SR spectroscopy}
In a $\mu$SR experiment~\cite{blundell1999spin} spin-polarized muons are implanted in a sample where they interact with the local magnetic field at the muon site.
After, on average, 2.2~$\mu$s, the muon decays into a positron and two neutrinos. By detecting the positrons, which are preferentially emitted in the direction of the muon spin at the time of decay, we can track the polarisation of the muon-spin ensemble.
In a zero-field (ZF) $\mu$SR experiment, as is primarily considered here, the local magnetic field at the muon sites arises due to the configuration of the spins in the system.
When the muon-spin has a component perpendicular to the local field $B$, precession occurs with angular frequency $\omega~=~\gamma_\mu B$, where $\gamma_\mu~=~2\pi\times135.5$~MHz~T$^{-1}$ is the gyromagnetic ratio of the muon.
Conversely, when the muon-spin aligns with the local field, only dynamic fluctuations can depolarize the muon-spin ensemble.
The quantity of interest in the experiment is the asymmetry $A\left(t\right)$, calculated from the number of counts arriving in the detectors forwards and backwards of the initial muon-spin polarisation direction, $n_{\rm F,B}$, via $A\left(t\right)=\left(n_{\rm F}-n_{\rm B}\right)/\left(n_{\rm F}+n_{\rm B}\right)$.
The asymmetry is directly proportional to the muon-spin polarisation.
Information on both the static and dynamic magnetism at the muon site can be determined through the functional form of $A\left(t\right)$.
Muon-spin spectroscopy measurements of polycrystalline HgCr$_2$Se$_4$ were carried out at the STFC-ISIS facility, UK, using the EMU instrument, and at the Swiss Muon Source (S$\mu$S), Paul Scherrer Institut, Switzerland, using the GPS instrument.
Polycrystalline samples were packed in Ag-foil packets and loaded in a $^4$He cryostat.
On the EMU instrument the packet was mounted on a silver backing plate, whereas on the GPS instrument the packet was attached to a fork and suspended in the beam, minimising the background contribution to $A\left(t\right)$.
Most measurements were performed in ZF, with additional measurements at ISIS performed in a 2~mT field applied perpendicular to the initial muon-spin polarisation (known as weak transverse field or wTF measurements).
Data analysis was carried out using the WiMDA program~\cite{pratt2000wimda} and made use of the MINUIT algorithm~\cite{james1975minuit} via the iminuit~\cite{iminuit} Python interface for global refinement of parameters.\\
\paragraph*{Fluctuation (noise) spectroscopy}
For electrical transport measurements Ti/Au electrical contacts were deposited onto the single crystal HgCr$_2$Se$_4$ samples under a hand-painted photoresist mask in a high vacuum electron beam evaporator. The samples were then annealed in a H/He forming gas in order to ensure good-quality (ohmic) contacts. Gold wires have been attached by silver epoxy, see Supplemental Material of Ref.\,\cite{Yang2019}.
Single crystals of HgCr$_2$Se$_4$ were measured in a continuous helium-flow cryostat with variable temperature insert equipped with a superconducting magnet, using a four-terminal AC technique and a lock-in amplifier.
For the resistance noise measurements, after pre-amplification of the voltage drop across the sample, the signal is demodulated by the lock-in amplifier and the voltage noise power spectral density (PSD), $S_V(f,T)$, of the fluctuations is processed and recorded by a spectrum analyzer \cite{Mueller2011,Mueller2018}. As required, the voltage noise PSD scales as $S_V \propto I^2$ and care has been taken to rule out spurious noise sources contributing to the measured signal, e.g.\ from the preamplifier or the electrical contacts. The resistance noise PSD is given by $S_R = S_V/I^2$ and therefore for the normalized quantities it is $S_V/V^2 = S_R/R^2$. Figure \ref{spectra} shows typical spectra of either pure $1/f^\alpha$-type or superimposed with a Lorentzian contribution, the latter being a signature of a two-level switching process.
\begin{figure}[h!]
\centering
\includegraphics[width=\columnwidth]{1.pdf}
\caption{Typical voltage noise spectra $S_V$ vs.\ $f$ for HgCr$_2$Se$_4$ (at various temperatures) either of $1/f^\alpha$-type or superimposed with a Lorentzian contribution. (a) For the former type, from a linear fit in the $\log$-$\log$ representation (red line) the magnitude $S_V(f = 1\,{\rm Hz})$ (dashed red line) and the frequency exponent $\alpha$ (slope) is determined. (b) A characteristic spectrum of the latter type shown as $S_V \times f$ vs.\ $f$, where the $1/f^\alpha$ contribution (dotted line) is nearly constant (in this case $\alpha = 1$). The spectrum is fitted to Eq.\,(\ref{noise_spectra}) (continuous red line) with the Lorentzian (dashed line) exhibiting a peak at the corner frequency $f_c$.}
\label{spectra}
\end{figure}
\section{Experimental Results}
\subsection{$\mu$SR spectroscopy}
\begin{figure}[b]
\centering
\includegraphics[width=\linewidth]{2.pdf}
\caption{Representative ZF $\mu$SR measurements of HgCr$_2$Se$_4$. Data in (a) was measured at ISIS; (b) shows data measured at S$\mu$S. Fits to the data, described in the text, are shown. Some data are plotted with a labelled vertical offset for clarity.}
\label{fig:data}
\end{figure}
Representative ZF $\mu$SR measurements of HgCr$_2$Se$_4$ are presented in Fig.~\ref{fig:data}, showing both the long-time relaxation [(a), measured at ISIS] predominantly sensitive to dynamics of the system on the muon timescale, and short-time behavior [(b), measured at S$\mu$S], dominated by effects due to long-range magnetic order.
The measurements
performed at ISIS were fitted over the entire temperature range using the function
\begin{equation}\label{eqn:isis}
A\left(t\right) = a_{\rm r}\exp\left(-\lambda t\right) + a_{\rm b}\exp\left(-\lambda_{\rm b} t\right) ,
\end{equation}
where the component with amplitude $a_{\rm r}$ captures the contribution from muons stopping in the sample
and relaxing with rate $\lambda$ determined in part by dynamic fluctuations of the local magnetic field, while the baseline term with amplitude $a_{\rm b}$ accounts for muons that stop outside of the sample in the silver backing plate and weakly relax with rate $\lambda_{\rm b}$ due to interaction with the Ag nuclear moments.
As expected, the rate $\lambda_{\rm b}=0.01$~$\mu$s$^{-1}$ was found to remain constant at all temperatures.
In a magnetically ordered, polycrystalline sample we expect $1/3$ of muons to stop with their spin initially aligned along the local magnetic field (contributing a slowly relaxing `1/3 tail' to the spectra).
The other muons, that stop with a component of their spin initially perpendicular to the local magnetic field, will precess and relax, with both effects too rapid to be observed at ISIS due to the pulse-width of the ISIS muon beam.
(Instead, this contribution is seen in the measurements performed at S$\mu$S, discussed below.)
This accounts for a lost fraction of the asymmetry in the ordered state in the ISIS measurements.
Values of these parameters extracted through fitting the data can be seen in Fig.~\ref{fig:zf}(a--b).
\begin{figure}
\centering
\includegraphics[width=\linewidth]{3.pdf}
\caption{Parameters extracted through fitting ZF $\mu$SR measurements of HgCr$_2$Se$_4$ as described in the text. (a--b) results from ISIS, (c--d) results from S$\mu$S. Dashed lines indicate boundaries between the different regimes as discussed in the text. The solid lines in (a) and (c) are fits to critical behavior typical for a 3D Heisenberg ferromagnet, with the dashed section extending the behavior to $T_{\rm C}$.}
\label{fig:zf}
\end{figure}
Two peaks in $\lambda$, as well as unusual changes in the total asymmetry $a_{\rm r}+a_{\rm b}$ (typically, the asymmetry remains approximately constant far from $T_{\rm C}$), suggest three distinct regimes of behavior distinguished by their dynamics; we indicate the transition between these regimes with dashed lines in Fig.~\ref{fig:zf}.
At low $T$ there is little relaxation.
Above $T~\approx~75$~K the drop in $a_{\rm r}+a_{\rm b}$ indicates that an increased number of muons have their spin polarisation relaxed too rapidly to be detected with the time-resolution available at ISIS, suggestive of dynamics of the local magnetic field on the muon timescale.
This may also suggest that the observed peaks in $\lambda$ are not two distinct peaks, but rather the two edges of a single peak (which will have a maximum between 75~K and $T_{\rm C}\approx 106$~K), with the relaxation rate in this regime too rapid to be detected in these data.
For $T>T_{\rm C}$, the recovery of $a_{\rm r}+a_{\rm b}$ and gradual decrease in $\lambda$ after a peak is suggestive of a transition to a disordered magnetic phase.
Above $T_{\rm C}$, in a 3D Heisenberg magnet (such as HgCr$_2$Se$_4$), we would expect $\lambda\propto\left|T-T_{\rm C}\right|^{-0.709}$~\cite{pelissetto2002critical,pratt2007chiral,pospelov2019non}.
This model describes the data well, with $T_{\rm C}\simeq106$~K, and can be seen as a solid line in Fig.~\ref{fig:zf}(a).
In equivalent measurements performed with higher time resolution at S$\mu$S, oscillations in $A\left(t\right)$ are observed below $T_{\rm C}$ [see Fig.~\ref{fig:data}(b)], indicative of coherent precession of the muon-spin polarisation that occurs due to long range magnetic order.
To characterise these oscillations the data are fitted with the function
\begin{equation}
\begin{split}
A\left(t\right) =& \sum_{i=1}^{n}a_{\rm o}^i\cos\left(\gamma_\mu B_i t + \phi_i\right)\exp\left(-\sigma_{i}^2 t^2\right) \\
&+ a_{\rm r}\exp\left(-\lambda_{\rm r} t\right) + a_{\rm b} ,
\label{eq:smus_fit}
\end{split}
\end{equation}
where $a_{\rm o}^i$ the amplitude of the oscillating components, each due to precession in local magnetic field $B_i$ with associated phase offset $\phi_i$ and relaxation rate $\sigma_i$, which is related to the distribution of magnetic fields at that muon site.
The term with amplitude $a_{\rm r}$ again accounts for muons that stop with their spin initially aligned with the local magnetic field direction and hence do not precess, and $a_{\rm b}$ accounts for muons stopping outside of the sample.
At low $T$, below $T~\simeq~75$~K, $n~=~2$ is needed to capture the behavior of $A\left(t\right)$, indicating two magnetically distinct muon stopping sites.
Extracted parameters through fitting of $A\left(t\right)$ are shown in Fig.~\ref{fig:zf}(c--d).
The fields at the muon sites are found to vary in fixed proportion, $B_2/B_1~=~3.08(5)$, with $a_{\rm o}^1 + a_{\rm o}^2 = 9.8(3)$\%, with $a_{\rm o}^2/(a_{\rm o}^1+a_{\rm o}^2)~=~0.281(11)$, $\phi_1~=~7(2)\degree$ and $\phi_2~=~-45(3)\degree$ found to be independent of $T$ over this range.
All these factors suggest that the two distinct magnetic environments occur due to the same magnetic order parameter, typical of ferromagnetic behavior previously observed at low $T$ in HgCr$_2$Se$_4$~\cite{Lin2016}.
The non-zero (and distinct) phase offsets of the two components are unusual, and suggest complex magnetic behavior that leads to differences in the local magnetic field distribution at the two muon stopping sites. Specifically, the $\phi\approx -45^{\circ}$ phase offset is sometimes observed for the case of an incommensurate magnetic structure \cite{blundell1999spin, blundell2021}. (We return to this feature below.)
The rate $\lambda_{\rm r}$ matches that found in the ISIS measurements discussed above.
The relaxation rates $\sigma_{i}$ of the oscillating components in a static, magnetically-ordered phase can be related to the width of the distribution of magnetic fields at the muon site via $\sigma_{i}^2~=~\Delta_{i}^2/2$, where $\Delta_{i}~=~\gamma_\mu\sqrt{\left<(B_i-\langle B_{i} \rangle)^2\right>}$.
We observe $\sigma_2<\sigma_1$, which is unexpected as we would typically expect $\sigma_{i}$ to scale with the field at the muon site.
This further suggests that the two distinct muon stopping sites
have different static local environments and/or different dynamics.
As $T$ is increased above 75~K, oscillations in $A\left(t\right)$ are still observable, but are heavily damped [see Fig.~\ref{fig:data}(b)].
This indicates the system still exhibits long-range magnetic order, but there is a much greater dynamic response on the muon timescale (as is also observed in the measurements performed at ISIS).
Only $n=1$ is needed to parametrize these data in this regime. However, the large relaxation rates $\sigma_{1}$, coupled with the smaller relative amplitude seen at low $T$, suggests that the contribution to $A\left(t\right)$ from $B_2$ is relaxed too quickly to be observed.
This is consistent with the increased lost asymmetry observed in Fig.~\ref{fig:zf}(b).
In this regime $\phi_1~=~7(3)\degree$ (consistent with $\phi_1$ at low $T$) and $a_{\rm r}/(a_{\rm o}^1+a_{\rm r})~=~0.391(15)$.
As expected from the ISIS data, $\sigma_1$ and $\lambda_{\rm r}$ both peak in the middle of this temperature regime. Although it is not normally possible to distinguish an increase in $\sigma_1$ occuring due to an increase in the width of the distribution of magnetic fields at the muon site, or a change in the dynamic fluctuation rate on the muon timescale, we infer from the behavior of lambda in the ISIS data that this effect is likely dominated by dynamic effects. This further suggests the peaks observed in $\lambda$ in Fig.~\ref{fig:zf}(a) are the edges of a single peak.
The peak in $\sigma_{1}$ might suggest that relaxation pathways freeze out, reducing in frequency and passing through the frequency window to which $\mu$SR is sensitive on cooling below 90\,K, as the behavior changes due to the system gradually changing state.
For $T~<~T_{\rm C}$ we expect critical behavior of $B_i$ such that $B_i~=~B_i^0\left(1-T/T_{\rm C}\right)^\beta$, with Ref.~\cite{Lin2016} identifying $\beta~=~0.361$, close to the theoretical value of $\beta~=~0.367$ for a 3D Heisenberg ferromagnet.
Due to the heavy suppression of the oscillatory part of $A\left(t\right)$ just below $T_{\rm C}$, extracting a reliable value of $\beta$ is challenging. However constraining to $\beta~=~0.367$ gives a good fit of the extracted $B_{\rm i}$, as seen in Fig.~\ref{fig:zf}(c), consistent with the system behaving as a 3D Heisenberg ferromagnet, giving a critical temperature $T_{\rm C}~=~107.9(3)$~K.
At high $T$ there are no oscillations observable in $A\left(t\right)$, consistent with a lack of long-range magnetic order.
Eqn.~\ref{eqn:isis}, with $\lambda_{\rm b}~=~0$, describes the behavior in this regime.
Finally, we performed weak transverse field (wTF) measurements in order to check for evidence of magnetic phase separation, which are detailed in Appendix~\ref{sec:wtfmusr}. These measurements do not show evidence for such phase separation down to the few-percent level, but do show similar transitions to those reported from the ZF measurements.
The key observation from these $\mu$SR results is the spin dynamics seen in the intermediate temperature regime. In order to link these to charge-carrier dynamics, we next turn to the results of fluctuation spectroscopy measurements.
\subsection{Fluctuation (noise) spectroscopy}
\paragraph*{Sample characterization -- DC conductivity}
Transport measurements have been carried out on several samples of $n$-type HgCr$_2$Se$_4$ single crystals. From the literature it is known that
the transport properties and the CMR effect are highly sensitive to details in the growth procedure and sample treatment \cite{Solin2008}. However, the magnetism and basic transport characteristics are consistent for the different types of samples. For the time-resolved transport measurements presented here, we chose a single-crystalline sample with an electron density of $n \approx 6.1 \times 10^{15} {\rm cm}^{-3}$ at liquid helium temperatures, the impedance range of which is convenient for standard fluctuation spectroscopy measurements.
\begin{figure}[t]
\centering
\includegraphics[width=0.85\columnwidth]{4.pdf}
\caption{(a) Longitudinal resistivity $\rho_{xx}(T)$ of $n$-type HgCr$_2$Se$_4$ for different externally applied magnetic fields of $B$ = 0, 1, 4 and 8\,T. (b) Arrhenius plot $\log{\rho_{xx}}$ vs.\ $1/T$ for magnetic fields $B$ = 0, 1, 4 and 8\,T. The linear fit (red dashed line) (see text) indicates a thermal activation gap of $\Delta_b = 133$\,meV.}
\label{resistivity}%
\end{figure}%
For the present sample, due to a higher carrier mobility, the absolute values of the resistivity are considerably lower than for the sample discussed in our earlier work \cite{Lin2016} and hence the CMR effect less pronounced. However,
the transport properties of the HgCr$_2$Se$_4$ sample shown in Fig.\ \ref{resistivity} exhibit the same general characteristics,
namely a semiconducting behavior upon cooling from room temperature followed by an intermediate temperature regime, where the resistivity increases less strongly, and a drop and metallic behavior at $T \lesssim 92$\,K, which is somewhat below the ferromagnetic ordering temperature $T_C \approx 107$\,K determined from $\mu$SR or magnetic susceptibility measurements (see Appendix~\ref{Appendix_magnetic}), but is close to the relaxation rate peak seen in the ZF $\mu$SR. Similar to the sample in \cite{Lin2016}, the resistivity in the intermediate and metallic regimes, is suppressed by magnetic fields.
The Arrhenius plot in Fig. \ref{resistivity}(b) shows a thermally activated behavior $\rho(T) = \rho_0 \exp{(\Delta_b/k_BT)}$ of the resistivity at high temperatures with an activation energy $\Delta_b = 133$\,meV.\\
\paragraph*{Magnetic polaron dynamics and percolation (\,$T > T_C$)}
\begin{figure}[b]
\centering
\includegraphics[width=\columnwidth]{5.pdf}
\caption{(a) Resistance, $R$, (green symbols) and normalized resistance noise PSD, $S_R/R^2 (f = 1\,{\rm Hz})$, (orange) of the $1/f$-type noise vs.\ temperature $T$. In the grey shaded region additional Lorentzian-type spectra due to two-level switching processes were observed. Their amplitude is {\it not} included here. Their characteristic time constant $\tau_c$, however, is shown as blue triangles.
In the percolation regime (see text), the noise PSD strongly increases with decreasing temperature following an approximate power law $S_R/R^2 \propto T^{-14}$ (red line).
(b) Scaling $S_R/R^2 \propto R^w$ in the percolation regime with $w = 3.6 \pm 0.1$ (red line). Below about 120\,K the scaling exponent becomes temperature dependent.}
\label{noise}
\end{figure}
Next, we discuss the overall behavior of the resistance fluctuations in zero magnetic field. Figure \ref{noise}(a) displays the sample resistance, $R$, at discrete temperatures (green circles) and the corresponding normalized resistance noise PSD, $S_R/R^2$, (orange diamonds) of the $1/f$-type noise taken at $f = 1$\,Hz. In our frequency range $f = 0.01 - 100$\,Hz, we observe pure $S_R/R^2 \propto1/f^\alpha$ noise for temperatures $T > 150$\,K and $T < 100$\,K, with a frequency exponent $\alpha(T)$
exhibiting values between 0.8 and 1.4, see Fig.\ \ref{DDH}(a) below, where a variation from 1 to larger or smaller values implies a shift of the spectral weight (and therefore the fluctuation dynamics) to lower or higher frequencies, respectively, as compared to the homogeneous distribution of time scales for $\alpha = 1$ \cite{Mueller2018}, see also Appendix~\ref{Appendix_noise}. In contrast, a Lorentzian spectrum characteristic for distinct two-level fluctuations is superimposed on the $1/f^\alpha$-noise for $100\,{\rm K} < T < 150\,{\rm K}$ (grey-shaded area in Fig.\ \ref{noise} (a)). Therefore,
the observed spectra can be described by
\begin{equation}
\frac{S_R(f,T)}{R^2} = \frac{A_1}{f^\alpha} + \frac{A_2 f_c}{f^2 + f_c^2},
\label{noise_spectra}
\end{equation}
where $A_1(T)$ denotes the magnitude of the $1/f$-type noise and $\alpha(T)$ its frequency exponent. The magnitude of the Lorentz contribution $A_2(T) \equiv 2/\pi \cdot \langle (\delta R)^2 \rangle/R^2$ and $f_c \equiv 1/(2 \pi \tau_c)$ denote the relative change of the global resistance due to the local fluctuation and the corner frequency, respectively, see Fig.\ \ref{spectra}(b). The action of a single fluctuator giving rise to the Lorentz spectrum (and $A_2(T) \neq 0$) becomes enhanced in our frequency range ('noise window' $[0.01\,{\rm Hz}, 100\,{\rm Hz}]$) only for $100\,{\rm K} < T < 150\,{\rm K}$. For $T > 150$\,K and $T < 100$\,K, however, the magnitudes $A_2(T) = 0$ and $A_1(T) \equiv S_R/R^2(f = 1\,{\rm Hz},T)$ as well as the frequency exponent $\alpha(T) \equiv (-\partial \ln{S_R/R^2(f,T)})/(\partial \ln{f})$ can be directly read from the fit shown exemplarily for a selected temperature in Fig.\ \ref{spectra}(a).
In Figure \ref{noise}(a), for all temperatures including the grey-shaded area, we show
$A_1(T)$ (orange diamonds), i.e.\ the magnitude of the Lorentzian contribution $A_2(T)$ is {\it not} included [see Appendix~\ref{Appendix_noise} for the behavior of $A_2(T)$]. Clearly, the resistance fluctuations $S_R/R^2(T)$ behave very differently than the time-averaged mean $R(T)$ and reveal information on the low-frequency dynamics of the charge carriers coupled to the magnetic (and possibly other) degrees of freedom. Upon cooling from room temperature, the $1/f$-noise initially slightly increases until below $\sim 230 - 240$\,K it stays roughly constant down to about 150\,K (roughly coinciding with the onset of two-level fluctuations), below which it again slightly decreases. At about 135\,K the $1/f$-noise then shows a sudden increase of almost two orders of magnitude, peaks at around $T_{\rm max} \approx 98$\,K (slightly below $T_C$ and above the resistivity maximum) and exhibits a sudden drop upon further cooling through the magnetically ordered phase. Below a sharp minimum, the noise then increases again and shows additional local maxima and minima upon further cooling. As will be described below, these features exhibit a qualitatively different dynamics than the sharp peak at 98\,K and their behavior can be well understood within a model of non-exponential kinetics, i.e.\ a superposition of many independent two-level fluctuators with a characteristic distribution of thermal activation energies.
We interpret the strong increase of the $1/f$-noise below about $135$\,K following an approximate power law $S_R/R^2 \propto T^{-14}$, the sharp
peak around 98\,K and the even more pronounced drop on the low temperature side as a percolation transition of correlated magnetic polarons quite similar to the observations in other CMR systems, as e.g.\ in perovskite manganites \cite{Podzorov2000} and EuB$_6$ \cite{Das2012}. (Notably, a strong, power-law increase of the fluctuations is a general feature observed upon approaching a metal-to-insulator transition \cite{Bogdanovic2002,Kar2003,Hartmann2015}.) As expected for a simple random resistor network \cite{Kogan1996}, we observe a scaling behavior, see Fig.\ \ref{noise}(b), demonstrating a power-law $S_R/R^2 \propto R^w$
upon approaching the percolation threshold, where $w = \kappa/t$ depends on the percolation scenario \cite{Tremblay1985,Kogan1996}.
We find $w = 3.6 \pm 0.1$, albeit only in the small temperature region of the strong increase of the 1/$f$-noise. Below this temperature regime the scaling exponent becomes temperature dependent, which may be due to the strong influence of spin correlations on the effective size of magnetic polarons to be discussed below. Large values of $w \sim 3$ have been found, e.g., in random metal-insulator composites, where a modified random void model is invoked, in which the stochastic voids in a conducting matrix are replaced by interpenetrating conducting spheres in an insulating matrix \cite{Rudman1985}.
The temperatures $T^\ast \sim 140 - 150$\,K, where the magnetic susceptibility of our sample deviates from a Curie-Weiss law, see Appendix~\ref{Appendix_magnetic}, may be interpreted as the crossover from isolated polarons to a percolative behavior \cite{Lin2016}. In electron spin resonance measurements \cite{Li2018} this temperature scale is assigned to the occurrence of an additional magnetic contribution associated to magnetic correlations among MPs or ferromagnetic clusters.\\
A clue for better understanding the slow dynamics within this percolation regime
comes from the distinct two-level fluctuations being enhanced in our 'noise window' for $100\,{\rm K} < T < 150\,{\rm K}$ (see above). Figure \ref{noise}(a) displays in this region (grey-shaded area) the time constant $\tau_c = 1/(2 \pi f_c)$ of the two-level fluctuator as blue triangles, determined from the fits of the spectra to Eq.\ (\ref{noise_spectra}) [see Fig.\ \ref{spectra}(b) and inset of Fig.\ \ref{corner-frequency}]. We find that for decreasing temperatures the characteristic relaxation time $\tau_c$ is nearly constant before below about 120\,K --- the same temperature where the percolation scaling shown in Fig.\ \ref{noise}(b) breaks down --- it strongly increases corresponding to a drastic slowing down of the fluctuation dynamics upon approaching the magnetic transition
and
the $1/f^\alpha$-noise maximum.
\begin{figure}[t]
\centering
\includegraphics[width=0.9\columnwidth]{6.pdf}
\caption{Temperature dependence of the corner frequency $f_c$ of the two-level fluctuations observed for $100\,{\rm K} < T < 150\,{\rm K}$ (grey-shaded area in Fig.\ \ref{noise} (a)) in zero magnetic field shown as $\ln{f_c}$ vs.\ $T$. Orange line is a fit to the model Eq.\ (\ref{model}), see text and \cite{comment3}. The inset shows corresponding spectra at selected temperatures with fits (red lines) to Eq.\ (\ref{noise_spectra}).}
\label{corner-frequency}
\end{figure}
This unusual behavior is also highlighted in Fig.\ \ref{corner-frequency} showing the corresponding temperature dependence of the corner frequencies $f_c$ plotted as $\ln{f_c}$ vs.\ $T$ with the evolution of the spectra fitted by Eq.\ (\ref{noise_spectra}) shown in the inset. Clearly, the behavior expected for a simple thermally activated process $f_c = f_0 \exp{(-E_a/k_BT)}$ is {\it not} observed. Instead, $f_c$ stays roughly constant upon cooling from 150\,K to about 120\,K below which it gradually more strongly decreases, i.e.\ the characteristic time of the fluctuations slows down, with an increasingly steeper slope (stronger slowing down) upon lowering the temperature. A natural explanation for a two-level-fluctuator in a system with magnetically-induced electronic phase separation and percolation
is the switching of a bound magnetic polaron, i.e.\ a nano-scale magnetic cluster with aligned spins having to overcome an energy barrier due to the environment of the surrounding paramagnetic matrix.
The logarithmic growth of the radius of the bound magnetic polaron with decreasing temperature described in the seminal works by Bhatt et al.\ \cite{Bhatt2002} and Kaminski and Das Sarma \cite{Kaminski2002} indeed suggests a slowing down of the polaron dynamics until the percolation threshold is reached, where the MPs merge and the impedance of their motion in the insulating matrix becomes lifted.
However, we will show in the following that in such a picture the observed behavior can only be understood when the MPs effective radius depends on the spin correlation length, i.e.\ a core-shell structure of the MP is considered. Indeed, such spin correlations have been proposed to play an important role for a wide temperature range $T_C < T < T^\ast$ in the paramagnetic phase \cite{Lin2016}.
In the model of correlated polarons by Bhatt et al.\ and Kaminski and Das Sarma, for temperatures smaller than a characteristic energy scale given by the localization length of a dopant (captured hole), all magnetic spins within distance $r_p(T) = \frac{1}{2} a_B \ln{(s S |J_0|/k_B T)}$ of dopants align antiferromagnetically, where $a_B$ is the localization length of the donor, $s$ the carrier spin, $S$ the spin of the magnetic ion in the matrix and $J_0$ the effective exchange constant \cite{Bhatt2002, Kaminski2002}. We now consider a core-shell structure, where $r_{core} = r_p(T)$ describes the radius over which the magnetic moments are bound to those of the carrier and $r_{shell} = \xi_0(T/T_C - 1)^{-\nu}$ the decay of the spin polarization at the MP boundary determined by the spin correlation length, with $\nu = 0.709$ for a 3D Heisenberg ferromagnet. Therefore, $r_{MP}(T) = r_{core}(T) + r_{shell}(T)$ and for the MP's volume
\begin{equation}
V_{MP}(T) \approx 4 \pi r_p(T)^2 \left[ \frac{1}{3}r_p(T) + \xi_0 \left(\frac{T}{T_C} -1\right)^{-\nu} \right].
\end{equation}
Assuming a bistable switching as suggested by the experimental observation and a thermally activated behavior $f_c(T) = f_0 \exp{(-KV_{MP}/k_BT)}$ with a parameter $K$ determining the energy barrier, we have
\begin{widetext}
\begin{equation}
\ln{f_c} \approx \ln{f_0} - \frac{\pi a_B^2 K}{k_BT}\left( \ln{\frac{s S |J_0|}{k_B T}} \right)^2 \cdot \left[\frac{a_B}{6}\ln{\frac{s S |J_0|}{k_B T}} + \xi_0 \left( \frac{T}{T_C} -1 \right)^{-\nu} \right].
\label{model}
\end{equation}
\end{widetext}
The orange line in Fig.\ \ref{corner-frequency} is a fit to the model Eq.\ (\ref{model}) with a set of parameters with reasonable values \cite{comment3} demonstrating that only when considering the significant spin correlations in HgCr$_2$Se$_4$ \cite{Lin2016} it is possible
to reproduce the highly unusual temperature dependence of the corner frequency of the observed two-level switching. We interpret this behavior in a sense that the exceptionally strong increase of the $1/f$-type noise upon cooling is due to two effects, namely (i) the percolation of MPs in a complex network and (ii) their increasingly slow `internal' dynamics. This is also reflected in the distribution of the spectral weight of the $1/f$-noise, see Appendix~\ref{Appendix_noise}.
Although
the fit is not unique due to the large set of independent parameters \cite{comment3} we consider the good agreement as a signature of the particular importance of spin correlation effects for the MP percolation in HgCr$_2$Se$_4$. Interestingly, in the related spinel CdCr$_2$S$_4$, a spin-pair correlation driven CMR effect is discussed \cite{Xia2019} which indicates a more general importance of our findings. As an independent cross check, we find that the estimated percolation threshold $r_{\rm perc} \approx 0.86/n^{1/3}$ with $n$ the density of MP \cite{Kaminski2002} yields a radius of the polarons which agrees quite well with the values estimated from our model, Eq.\ (\ref{model}), at the 'percolation temperature', i.e.\ the $1/f$ noise maximum at 98\,K.\\
\paragraph*{Non-exponential kinetics (\,$T < T_C$)}
\begin{figure}[h!]
\centering
\includegraphics[width=0.9\columnwidth]{7.pdf}
\caption{(a) Temperature dependence of the frequency exponent $ \alpha$ compared to the values calculated from the DDH model (red line), Eq.\ (\ref{alpha}). The parameters for $g(T)$ and $\tau_0$ are given. Dashed line marks the temperature below which the DDH model is valid.
(b) Distribution of activation energies $D(E)$ calculated from Eq.\ (\ref{DvonE}), which determines the temperature dependence of the noise magnitude $S_R/R^2(T)$ at $f = 1$\,Hz below $T = 75$\,K, see text.}
\label{DDH}%
\end{figure}%
Now we turn to the charge fluctuations in the magnetically ordered state. A widely used model for the ubiquitous $1/f^\alpha$-noise in solids is a superposition of independent `fluctuators', i.e.\ two-level processes fluctuating with a characteristic time constant $\tau$, which couple to the resistance of the system, each contributing a Lorentzian spectrum \cite{Kogan1996,Raquet2001,Mueller2018}:
\begin{equation}
\frac{S_R(f,T)}{R^2} \propto \int\limits_0^\infty D(\tau) \frac{4\tau}{1+4 \pi^2 f^2 \tau^2}{\rm d}\tau,
\end{equation}
where $D(\tau)$ is a distribution of effective time constants.
While in our case such a model is not expected to be valid in the percolation regime above and around $T_C$ or $T_{\rm max}$, its applicability at lower temperatures can be tested. The model assumes a thermally-activated switching processes where the relaxation time follows an Arrhenius law
$\tau = \tau_0 \exp{(E_a/k_BT)}$,
with $\tau_0$ an attempt time, typically of order of the inverse phonon frequencies, and $E_a$ the a characteristic energy of the two-level potential. The superposition of Lorentzian spectra with a sufficiently broad distribution of activation energies $D(E_a) \equiv D(E)$ then leads to a $1/f^\alpha$-type spectrum \cite{Pre1950} in a frequency window determined by the energy distribution. A phenomenological description based on this concept of non-exponential kinetics \cite{comment2} has been introduced by Dutta, Dimon and Horn (DDH) \cite{Dutta1979}. Under the assumption that the same fluctuation processes determine the temperature dependences of both the noise magnitude [see Fig.\,\ref{noise}(a)] and the frequency exponent (spectral weight distribution) [see Fig.\,\ref{DDH}(a)], the model predicts
\begin{equation}
\alpha(T) = 1 - \frac{1}{{\rm ln}(2 \pi f \tau_0)} \left[ \frac{\partial {\rm ln}S(T)}{\partial {\rm ln} (T)} - \frac{\partial {\rm ln}g(T)}{\partial {\rm ln} (T)} -1 \right],
\label{alpha}
\end{equation}
where $g(T)$ describes the coupling of the two-level processes to the resistance and accounts for an explicit temperature dependence of the distribution of activation energies \cite{Raquet1999}.
For $g(T) \propto T^{-1}$, creating a constant offset of the model adaption in Fig.\ \ref{DDH}(a) and a typical inverse phonon frequency $\tau_0 = 10^{-14}$\,s, the prediction Eq.\,(\ref{alpha}) matches the data quite well below about 75\,K (i.e.\ the boundary of the intermediate temperature regime identified via our $\mu$SR measurements) but deviates considerably above that temperature, in particular where the noise is mainly caused by the percolation of MPs. Therefore, for energies $E < k_B T \ln{(2 \pi f \tau_0)^{-1}}$ with $T = 75$\,K and at $f = 1$\,Hz, one can determine the energy distribution causing the observed $1/f^\alpha$-noise \cite{Dutta1979} from
\begin{equation}
D(E) \propto \frac{2 \pi f}{k_BT}\frac{S_R(f,T)/R^2}{g(T)},
\label{DvonE}
\end{equation}
which is shown in Fig.\,\ref{DDH}(b). Clearly, the observed magnitude of the $1/f$-type noise and its distribution of spectral weight can be described by a series of local maxima/peaks in the energy distribution $D(E)$ of thermally-activated fluctuators at about 36, 100, 115 and 168\,meV, see arrows in Fig.\ \ref{DDH}(b).
It is inherent to the model that the two-level fluctuators are not specified {\it a priori} but may be identified {\it a posteriori} from the maxima in $D(E)$, i.e.\ their energy signature. As discussed below, based on the $\mu$SR results we attribute these energies to clusters of MPs that form around the magnetic and percolation transition, become frozen at low temperatures and fluctuate as magnetic domains or dynamic domain walls.
\section{Discussion}
Naturally, resistance noise and $\mu$SR spectroscopy probe different frequency regimes.
The latter is a local probe of the microscopic magnetic field distribution, whose dynamic response is determined by $\gamma_{\mu}B$ where $B$ is the magnitude of the total field at the muon site, such that $\mu$SR typically probe spectral features in the MHz -- GHz range such as low-energy spin fluctuations or spin diffusion.
In contrast, resistance noise spectroscopy is sensitive to the dynamics of collective objects, such as magnetic polarons but also nano- or micro-scale magnetic domain walls or domains in ferromagnets.
Our data obtained by the combination of these two methods can be reconciled with a description of magnetism in HgCr$_2$Se$_4$ given in terms of MPs, in which dilute polarons form for temperatures $T > T_C$ and then coalesce below $T_C$ until an ordered magnetic state is stabilized at temperatures far below $T_C$.
The noise measurements reveal low-frequency dynamics consistent with a percolation transition of MP, where there is a strong increase and power-law divergence of the magnitude of the $1/f$-type fluctuations upon approaching the percolation threshold, very similar to the behavior of EuB$_6$ \cite{Das2012}, where the existence of MPs explaining the CMR effect is well established \cite{Suellow2000,brooks2004magnetic,Urbano2004,Zhang2009}. In the present spinel system, however, the density of isolated MPs seems to be considerably lower
which results in the enhancement of distinct two-level fluctuations superimposed on the underlying $1/f$ noise. We have shown that the MP dynamics can be described by a core-shell model corroborating the importance of spin correlations suggested previously.
The $\mu$SR data suggest that these isolated MPs do not significantly affect the dynamics of the bulk system on the $\mu$SR timescale, or they do not have sufficient density to lead to relaxation of the muon-spin. A rough estimate of the volume fraction occupied by MPs, assuming a radius of $\approx 5$~nm and a density equal to the carrier density $n$, gives 0.3\%, well below the volume at which we would expect $\mu$SR to be sensitive, which is typically of order of a few percent.
This picture is consistent with the noise data, as whilst the isolated MPs do not form much of the total volume fraction of the sample (as seen here with $\mu$SR), they can still have profound effects on the transport properties.
We mentioned that in EuB$_6$~\cite{fisk1979magnetic}, both resistance noise and $\mu$SR spectroscopy provided evidence for the coalescing of magnetic polarons~\cite{brooks2004magnetic} in a similar manner to the scenario we invoke here to explain the behavior of HgCr$_2$Se$_4$. In EuB$_6$, changes in the different parts of the $\mu$SR signal indicated two magnetic transitions, along with a phase separation, above $T_C$, into ferromagnetically ordered regions of overlapping MPs and paramagnetic regions between them. Some of the changes we observe in the $\mu$SR signal measured in the intermediate temperature regime (75\,K $<$ $T$ $<$ $T_C$) of HgCr$_2$Se$_4$ are broadly similar to those found in EuB$_6$ at temperatures above $T_C$ but below its higher-temperature transition \cite{brooks2004magnetic}. However, we do not see evidence for phase separation into paramagnetic and ordered regions in the intermediate temperature range of HgCr$_2$Se$_4$, nor a clear signal in the $\mu$SR above $T_C$ for MP dynamics or freezing.
An interesting feature is the coupling of the magnetism and electronic behavior, specifically the magnetic-order-driven metal-insulator transition, and the coincidence of peaks in $\mu$SR relaxation rates and electronic noise spectroscopy measurements. The magnetism at these temperatures is dynamic on the $\mu$SR time scale as evidenced by the peak in relaxation at $\sim 95$\,K which, remarkably, coincides with a peak both in the electrical resistivity and the normalised resistance fluctuations of the sample, suggesting a degree of magnetoelectric coupling. Further evidence for this coupling comes from the magnetic-order-driven metal-insulator transition. As the normalised PSD is measured at much lower frequencies ($\lesssim$ 100 \,Hz) than those to which $\mu$SR is sensitive, it suggests those dynamics persist over a wide range of frequencies. This holds true for frequencies down to the mHz regime which shows that the relaxation pathways freezing out consist of spins correlated regions over a wide range of length scale up to nano- and micro-scale objects, as e.g.\ clusters of MPs or magnetic domains, which naturally explains the enhanced $1/f$-type noise below $T_C$.
Upon further cooling, a change in dynamics is again observed at different time scales in both spectroscopy methods. Below $\simeq 75$\,K there is little dynamic relaxation at $\mu$SR time scales and the two different local environments (muon stopping sites) are observed to show distinct behavior. This can be interpreted in terms of
one of the two precession frequencies ($i=1$ in Eq.~\ref{eq:smus_fit}) arising from muons that stop inside a polaron or a cluster of polarons, and the other ($i=2$, with its phase offset indicative of a complicated, possibly incommensurate, field distribution) from muons that stop in between, in domain-wall-like regions. These two muon stopping sites couple to the same magnetic order parameter, but experience different dynamics. Likewise, below the same temperature of about 75\,K the observed $1/f$-type resistance fluctuations are well described by a model of non-exponential kinetics (which inherently implies an `inhomogeneous' distribution of relaxation times $D(E)$ of width much larger than the thermal energy), i.e.\ a change in the charge carrier dynamics coupled to the magnetism occurs. The observed dynamics in $\mu$SR and resistance noise may be consistently explained by large, polarized and domain-wall-like regions remaining from the freezing of the spin dynamics at the MP percolation transition below $T_C$.
Local ferromagnetic inhomogeneities, which were attributed to frozen clusters of MPs including a large connected (or `infinite') cluster of linked MP forming below $T_C$ indeed have been suggested for the model CMR system EuB$_6$ from measurements of the local magnetic induction \cite{Pohlit2018}.
The low-frequency fluctuation processes in HgCr$_2$Se$_4$ with energies between about 30 and 170\,meV, see Fig.\ \ref{DDH}(b), then are related to thermally-activated switching processes of magnetic domains or the motions of domain walls.
\section{Conclusion}
The combination of $\mu$SR and resistance fluctuation spectroscopy measurements of the ferromagnetic spinel HgCr$_2$Se$_4$ allows the study of magnetic and charge dynamics over a wide range of frequencies, revealing a degree of magnetoelectric coupling and distinct regimes of behavior reminiscent of polaron-host EuB$_6$ \cite{Das2012, brooks2004magnetic}.
At low $T$ conventional 3D Heisenberg magnetism is observed. As the temperature increases,
a dramatic change in the $\mu$SR relaxation rates occurs above $T \simeq 75$\,K, a temperature that coincides with the breakdown of the phenomenological DDH model in noise spectroscopy measurements. The latter is suggested to describe fluctuations of frozen clusters of linked magnetic polarons or magnetic domains in agreement with the observation of two magnetically distinct muon stopping sites coupled to the same magnetic order parameter.
Upon approaching $T_C$ and a maximum in the $1/f$-type noise a strong (power-law) increase of the resistance fluctuations and scaling behavior are observed, consistent with a percolation transition of magnetic polarons, similar to the CMR systems EuB$_6$ \cite{Das2012} and perovskite manganites \cite{Podzorov2000}. The unusal behavior of a two-level fluctuation process superimposed on the generic $1/f^\alpha$ noise in HgCr$_2$Se$_4$ can be described by a slowing down of the polaron dynamics, taking into account an effective radius of the magnetic polaron depending on the spin correlation length.
Above $T_C$, the $\mu$SR response suggests
that isolated magnetic polarons
do not occupy a significant volume fraction of the sample.
\section*{Acknowledgements}
Work at Goethe-University Frankfurt was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) -- project number 449866704.
Muon measurements were made at the Swiss Muon Source and the STFC-ISIS Facility and we are grateful for the provision of beamtime. We are grateful to EPSRC (UK) (grant number: EP/N032128/1). M.~N.~Wilson acknowledges the support of the Natural Sciences and Engineering Research Council of Canada (NSERC).
Data from the UK effort will be made available via DOI:XXXXXXXX. M.~M.\ and J.~M.\ are grateful to Bernd Wolf for magnetic characterization measurements.
M.~M.\ and T.~J.~H.\ contributed equally to this work.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 29
|
\section{Introduction}\label{sec:introduction}
The divisibility theory of commutative rings is a fundamental and persisting topic in mathematics that entails two main aspects: determining irreducibility and finding a factorization of the reducible elements in the ring. Prominent examples are the ring of integers $\mathbb{Z}$ and the ring of polynomials $\mathbb{Z}[x]$. It is then natural to investigate the arithmetic properties of $\mathbb{Z}[[x]]$, the ring of formal power series with integer coefficients. While polynomials in $\mathbb{Z}[x]$ can be seen as power series over the integers, the factorization properties over $\mathbb{Z}[x]$ and over $\mathbb{Z}[[x]]$ are in general unrelated; cf. \cite{BiGi}. In \cite{BGW2012a}, the authors studied this factorization problem exhaustively. In particular, for a class of polynomials parametrized by a prime $p$, a connection between reducibility in $\mathbb{Z}[[x]]$ and the existence of a $p$-adic root with positive valuation was established. Whereas this connection can be certainly explained in structural terms, the role of the root in the factorization process is not obvious.
Motivated by this factorization problem and the need to find explicit $p$-adic roots, the main goal of this paper is to provide formulas for the Hensel lifts of roots modulo $p$, and to give a factorization in $\mathbb{Z}[[x]]$ of certain reducible polynomials whose constant term is of the form $p^w$ with $p$ prime and $w>1$. All of our formulas are given in terms of partial Bell polynomials and rely on the inversion formula of Lagrange.
On the one hand, in Section~\ref{sec:Hensel}, we prove two versions of Hensel's lemma that give explicit formulas for the roots of any polynomial in $\mathbb{Z}_p[x]$, the ring of polynomials over the $p$-adic integers $\mathbb{Z}_p$. For illustration purposes, we examine the special cases of quadratic and cubic polynomials, see \eqref{quadraticRoot} and \eqref{cubicRoot}, and discuss the roots of unity leading to a formula for the so-called Teichm\"uller lifts, see Proposition~\ref{roots_unity}. On the other hand, we give a factorization over $\mathbb{Z}[[x]]$ for polynomials $f$ (of degree higher than 1) with $f(0)=p^w$ that are reducible in the presence of a $p$-adic root in $p\mathbb{Z}_p$. Although Theorem~\ref{FactorizationThm} is formulated for polynomials, it actually holds verbatim for power series. An illustrative example is discussed at the end of Section~\ref{sec:Factorization}.
As mentioned before, sections~\ref{sec:Hensel} and \ref{sec:Factorization} are related and rely on the material discussed in Section~\ref{sec:Inversion}. For the reader's convenience, a short appendix with some of the basic properties and identities for the partial Bell polynomials is included. We finish by observing that most of the results presented here may be applied to polynomials and power series over other commutative rings.
\section{Series solutions of algebraic equations}\label{sec:Inversion}
The main results of this paper rely on the following consequence of the inversion formula of Lagrange for formal power series. For a detailed proof and other applications, we refer the reader to \cite[Section 3.8]{Comtet} or \cite[Section 11.6]{Charalambides}. In the sequel, $B_{n,j}(x_1,x_2,\dots)$ denotes the $(n,j)$-th partial Bell polynomial, see the appendix.
\begin{lemma}[cf. {\cite[Corollary 11.3]{Charalambides}}]
\label{FormalInversion}
If $\phi(t)$ is a power series of the form
\begin{equation*}
\phi(t) = t\bigg(1+\sum_{r=1}^\infty \alpha_r \frac{t^r}{r!}\bigg),
\end{equation*}
then its formal inverse is given by
\begin{equation*}
\phi^{-1}(u) = u\bigg(1+\sum_{n=1}^\infty \beta_n \frac{u^n}{n!}\bigg),
\end{equation*}
where
\begin{equation*}
\beta_n = \sum_{j=1}^n (-1)^j \frac{(n+j)!}{(n+1)!}B_{n,j}(\alpha_1,\alpha_2,\dots).
\end{equation*}
\end{lemma}
Inversion formulas of this type have been studied by many authors in the search for solutions of algebraic equations. For instance, a series solution for the equation $x^m+px=q$ was already given by Lambert in 1758, cf. \cite{Lambert}. The most general formulas we found in the literature were obtained by Birkeland around 1927. In \cite{Birk27}, the author studied arbitrary polynomial equations and obtained explicit solutions in terms of hypergeometric functions, see also \cite{PaTsi}.
It turns out that, if $f(x)$ is a power series over a commutative ring $\mathcal{R}$ with an invertible linear coefficient, then formal series solutions for the equation $f(x)=0$ can be obtained from Lemma~\ref{FormalInversion} as follows.
\begin{proposition}\label{RootThm}
Given a power series $f(x) = a_0 + a_1x + a_2 x^2 + \cdots\in \mathcal{R}[[x]]$ with $a_1$ invertible in $\mathcal{R}$, the equation $f(x)=0$ has the formal root
\begin{equation}\label{formalRoot}
x = \sum_{n=0}^\infty \sum_{k=0}^n \frac{(-1)^{n-k+1}}{a_1^k\,(n+1)!} \binom{2n+1}{n-k} B_{n+k,k}(1! a_1, 2!a_2,\dots) \Big(\frac{a_0}{a_1}\Big)^{n+1}.
\end{equation}
\end{proposition}
\begin{proof}
Let
\begin{equation*}
\phi(x) = x\bigg(1+\sum_{\ell=1}^\infty \alpha_\ell \frac{x^\ell}{\ell!}\bigg) \text{ with } \alpha_\ell=\ell!\, a_{\ell+1}/a_1.
\end{equation*}
Thus $f(x)=a_1\big(a_0/a_1+\phi(x)\big)$ and $f(x)=0$ if $\phi(x)=-a_0/a_1$. By Lemma~\ref{FormalInversion}, this equation has the formal root
\begin{align} \notag
x&= \phi^{-1}(-a_0/a_1) = -\frac{a_0}{a_1}\bigg(1+\sum_{n=1}^\infty \frac{(-1)^n\beta_n}{n!} \Big(\frac{a_0}{a_1}\Big)^n \bigg) \\ \notag
&= -\frac{a_0}{a_1}\bigg(1+\sum_{n=1}^\infty \frac{(-1)^{n}}{n!} \bigg[\sum_{j=1}^n (-1)^j\frac{(n+j)!}{(n+1)!}B_{n,j}(\alpha_1,\alpha_2,\dots)\bigg] \Big(\frac{a_0}{a_1}\Big)^n \bigg) \\
\label{altForm}
&= \sum_{n=0}^\infty \bigg[\sum_{j=0}^n \frac{(-1)^{n+j+1}}{a_1^j\, n!} \frac{(n+j)!}{(n+1)!}B_{n,j}(1!a_2, 2!a_3,\dots)\bigg] \Big(\frac{a_0}{a_1}\Big)^{n+1}.
\end{align}
Now, if we set $x_j=j! a_j$, then
\begin{equation*}
B_{n,j}(1! a_2, 2!a_3,\dots)
= B_{n,j}(\tfrac{x_2}{2} \tfrac{x_3}{3},\dots) = \frac{n!}{(n+j)!} B_{n+j,j}(0,x_2,x_3,\dots)
\end{equation*}
by \eqref{Comt3l'}. Moreover, by means of \eqref{Comt3n} and \eqref{Comt3n'}, we have
\begin{align*}
B_{n+j,j}(0,x_2,x_3,\dots) &= \sum_{\substack{k\le j \\ \nu\le n+j}}
\binom{n+j}{\nu}B_{\nu,k}(x_1,x_2,\dots) B_{n+j-\nu,j-k}(-x_1,0,\dots) \\
&= \sum_{k\le j} \binom{n+j}{n+k} (-x_1)^{j-k}B_{n+k,k}(x_1,x_2,\dots).
\end{align*}
Therefore,
\begin{align*}
\sum_{j=0}^n \frac{(-1)^{n+j+1}}{a_1^j\,n!} & \frac{(n+j)!}{(n+1)!} B_{n,j}(1!a_2, 2!a_3,\dots) \\
&= \sum_{j=0}^n \frac{(-1)^{n+j+1}}{a_1^j\, (n+1)!} B_{n+j,j}(0,2!a_2,3!a_3,\dots) \\
&= \sum_{j=0}^n\sum_{k\le j} \frac{(-1)^{n-k+1}}{a_1^k\, (n+1)!} \binom{n+j}{n+k}
B_{n+k,k}(1!a_1,2!a_2,\dots) \\[-3ex] \intertext{}
&= \sum_{k=0}^n \frac{(-1)^{n-k+1}}{a_1^k\, (n+1)!} \bigg[\sum_{j=k}^n \binom{n+j}{n+k}\bigg]
B_{n+k,k}(1!a_1,2!a_2,\dots) \\
&= \sum_{k=0}^n \frac{(-1)^{n-k+1}}{a_1^k\, (n+1)!} \binom{2n+1}{n-k} B_{n+k,k}(1!a_1,2!a_2,\dots).
\end{align*}
Inserting this expression into \eqref{altForm}, we arrive at \eqref{formalRoot}.
\end{proof}
The simplicity (or complexity) of formula \eqref{formalRoot} clearly depends on the structure of the partial Bell polynomials.
For example, for $x^m+px-q=0$ with $p,q\in\mathbb{R}$, $p\not=0$, $m>1$, the root \eqref{formalRoot} takes the form
\begin{equation*}
x = \sum_{n=0}^\infty \sum_{k=0}^n \frac{(-1)^{n-k+1}}{p^k\,(n+1)!} \binom{2n+1}{n-k}
B_{n+k,k}(p,0,\dots,0,m!,0,\dots) \Big(\frac{-q}{p}\Big)^{n+1},
\end{equation*}
which by means of \eqref{Comt3n} and \eqref{Comt3n'} reduces to
\begin{equation}\label{LambertRoot}
x = \sum_{k=0}^\infty \frac{(-1)^{k}}{p^{k}} \binom{mk}{k} \frac{1}{(m-1)k+1}
\Big(\frac{q}{p}\Big)^{(m-1)k+1}.
\end{equation}
This formula includes Eisenstein's series solution for $x^5+x=q$, cf. \cite{Stillwell},
\begin{equation*}
x = \sum_{k=0}^\infty (-1)^{k} \binom{5k}{k} \frac{1}{4k+1} q^{4k+1}.
\end{equation*}
Of course, this series does not converge for all values of $q$, so further analysis is required to understand and possibly make sense of \eqref{formalRoot}. While convergence in general is not the focus of this paper, we want to briefly discuss $f(x)=x^3+x-q$ in order to illustrate a possible analytic approach. For this polynomial, the sum \eqref{LambertRoot} becomes
\begin{equation*}
x = \sum_{k=0}^\infty (-1)^{k} \binom{3k}{k} \frac{1}{2k+1} q^{2k+1}
= q\sum_{k=0}^\infty \binom{3k}{k} \frac{1}{2k+1} (-q^2)^{k},
\end{equation*}
which converges only when $q^2\leq 4/27$. However,
\begin{equation*}
\sum_{k=0}^\infty \binom{3k}{k} \frac{1}{2k+1} (-q^2)^{k}
= {}_{{}_2}F_{{}_1}\big(\tfrac13, \tfrac23; \tfrac32; -\tfrac{27}{4} q^2\big),
\end{equation*}
and since the hypergeometric function ${}_{{}_2}F_{{}_1}\big(\tfrac13, \tfrac23; \tfrac32; z\big)$ extends analytically to $\mathbb{C}\setminus (1,\infty)$, we can actually evaluate the formal root for larger values of $q$. For example, if $q=2$, then the root of $x^3+x-2=0$ provided by \eqref{formalRoot} is precisely $x=2\cdot {}_{{}_2}F_{{}_1}\big(\tfrac13, \tfrac23; \tfrac32; -27\big)=1$.
\bigskip
In the next section we will fully discuss the use of \eqref{formalRoot} to find roots of polynomials over the $p$-adic integers.
\section{Hensel's roots}\label{sec:Hensel}
In this section, we use Proposition~\ref{RootThm} to give a version of Hensel's lemma that provides an explicit formula for the $p$-adic root of a polynomial in $\mathbb{Z}_p[x]$. We start by recalling some basic facts about the $p$-adic numbers. For a comprehensive treatment of this subject, the reader is referred to \cite{Katok,Koblitz,Serre}.
\par
Let $p$ be a prime integer. For any nonzero integer $a$, let $v_p(a)$ (the $p$-adic valuation of $a$) be the highest power of $p$ which divides $a$, i.e., the greatest $m$ such that $a \equiv 0 \pmod {p^m}$; we agree to write $v_p(0) = \infty$. Note that $v_p\,(a_1\,a_2) = v_p(a_1) + v_p(a_2)$ for all $a_1, a_2\in \mathbb{Z}$. For any rational number $x = a/b$, define $v_p(x)=v_p(a) - v_p(b)$. Note that this expression depends only on $x$ and not on its representation as a ratio of integers.
The $p$-adic norm in $\mathbb{Q}$ is defined as $\norm{x}_p=p^{-v_p(x)}$ if $x\not=0$, and $\norm{0}_p=0$. This norm is non-Archimedean, that is $\norm{x + y}_p \le \max(\norm{x}_p, \norm{y}_p)$. The $p$-adic completion of $\mathbb{Q}$ with respect to $\norm{\cdot}_p$ is denoted by $\mathbb{Q}_p$. Every $a\in \mathbb{Q}_p$ admits a unique $p$-adic expansion,
\begin{equation*}
a = \frac{a_0}{p^m}+\frac{a_1}{p^{m-1}}+\dotsb +\frac{a_{m-1}}{p}+a_m + a_{m+1}\,p + a_{m+2}\,p^2+\dotsb,
\end{equation*}
with $0\le a_i<p$ for all $i$.
\par
We let $\mathbb{Z}_p = \{a \in \mathbb{Q}_p\,\vert\, \norm{a}_p \le 1\}$, the set of all numbers in $\mathbb{Q}_p$ whose $p$-adic expansion involves no negative powers of $p$. An element of $\mathbb{Z}_p$ is called a \emph{p-adic integer}, and the set of $p$-adic integers is a subring of the field $\mathbb{Q}_p$. If $x\in \mathbb{Z}_p$ is such that $v_p(x)=0$, then $x$ is a unit and its multiplicative inverse $1/x$ is in $\mathbb{Z}_p$.
\par
A fundamental property of the $p$-adic numbers is that a series in $\mathbb{Q}_p$ converges if and only if its terms approach zero. This condition is equivalent to verifying that the $p$-adic valuation of the terms tend to infinity.
\begin{theorem}\label{basicHensel}
Let $p>2$ be prime and let $f(x) = a_0 + a_1x + \cdots + a_m x^m$ be a polynomial of degree $m$ in $\mathbb{Z}_p[x]$. If $r_0\in\mathbb{Z}$ is such that
\begin{equation*}
f(r_0)\equiv 0\!\! \pmod p \;\text{ and }\; v_p(f'(r_0))=0,
\end{equation*}
then $r_0$ lifts to a $p$-adic root $r$ of $f$ given by
\begin{equation}\label{HenselRoot}
r = r_0 + \sum_{n=0}^\infty \bigg[\sum_{k=0}^n \frac{(-1)^{n-k+1}}{c_1^k\,(n+1)!} \binom{2n+1}{n-k}
B_{n+k,k}(1!c_1, 2!c_2,\dots)\bigg] \Big(\frac{c_0}{c_1}\Big)^{n+1},
\end{equation}
where $c_j=\frac{f^{(j)}(r_0)}{j!}$ for $j=0,1,\dots,m$. Note that $v_p(c_1)=0$ implies $1/c_1\in \mathbb{Z}_p$.
\end{theorem}
\begin{proof}
Given $f(x)$ and $r_0\in\mathbb{Z}$ as above, consider the function $g(x) = f(r_0+x)$. The Taylor expansion of $g(x)$ at $x=0$ gives
\begin{align*}
g(x) = f(r_0) + f'(r_0) x + \tfrac{f''(r_0)}{2!} x^2 +\dots +\tfrac{f^{(m)}(r_0)}{m!} x^m
= c_0 + c_1 x + c_2 x^2 +\dots + c_m x^m,
\end{align*}
with the property that $v_p(c_0)=v_p(f(r_0))\ge 1$ and $v_p(c_1)=v_p(f'(r_0))=0$. Thus $c_1\not=0$, and by Proposition~\ref{RootThm}, $g(x)$ has a formal root
\begin{equation*}
\varrho = \sum_{n=0}^\infty \gamma_n \frac{c_0^{n+1}}{(n+1)!},
\end{equation*}
where
\begin{equation*}
\gamma_n = \sum_{k=0}^n \frac{(-1)^{n-k+1}}{c_1^{k+n+1}} \binom{2n+1}{n-k}
B_{n+k,k}(1!c_1, 2!c_2,\dots).
\end{equation*}
Since $v_p(c_1)=0$ and each $j!c_j$ is a $p$-adic integer, we have $\gamma_n \in\mathbb{Z}_p$ for every $n$. Moreover, if $n+1$ has the $p$-adic expansion $n+1=n_0+n_1p+n_2p^2+\cdots$, we have
\begin{equation*}
v_p((n+1)!)=\frac{n+1-s_p(n+1)}{p-1} < \frac{n+1}{p-1},
\end{equation*}
where $s_p(n+1)=n_0+n_1+n_2+\cdots$. Therefore, since $v_p(c_0)\ge 1$, we get
\begin{equation*}
v_p\big(\tfrac{c_0^{n+1}}{(n+1)!}\big)=v_p(c_0^{n+1}) - v_p((n+1)!)>n+1-\frac{n+1}{p-1} = \big(\tfrac{p-2}{p-1}\big)(n+1)\to \infty \text{ as } n\to\infty,
\end{equation*}
which implies that $\sum \gamma_n\, \frac{c_0^{n+1}}{(n+1)!}$ converges in $p\mathbb{Z}_p$. In conclusion, the formal root $\varrho$ is indeed a $p$-adic root of $g(x)$ and $r_0+\varrho\in\mathbb{Z}_p$ is a root of $f(x)$.
\end{proof}
More generally, we have:
\begin{theorem}\label{extendedHensel}
Let $p>0$ be prime and let $f(x) = a_0 + a_1x + \cdots + a_m x^m$ be a polynomial in $\mathbb{Z}_p[x]$. Let $\nu, \kappa\in\mathbb{Z}$ such that $0\le 2\kappa<\nu$. If $r_0\in\mathbb{Z}$ is such that
\begin{equation*}
f(r_0)\equiv 0\!\! \pmod{p^\nu} \;\text{ and }\; v_p(f'(r_0))=\kappa,
\end{equation*}
then $r_0$ lifts to a $p$-adic root $r$ of $f$ given by
\begin{equation*}
r = r_0 + p^\kappa\sum_{n=0}^\infty \bigg[\sum_{k=0}^n \frac{(-1)^{n-k+1}}{c_1^k\,(n+1)!} \binom{2n+1}{n-k}
B_{n+k,k}(1!c_1, 2!c_2,\dots)\bigg] \Big(\frac{c_0}{c_1}\Big)^{n+1},
\end{equation*}
where $c_j=p^{(j-2)\kappa}\,\frac{f^{(j)}(r_0)}{j!}$ for $j=0,1,\dots,m$.
\end{theorem}
\begin{proof}
The proof is similar to the one for the previous theorem. Let $r_0$ be a root of $f$ modulo $p^\nu$ and let $\kappa$ be the $p$-adic valuation of $f'(r_0)$. Consider $g(x)=p^{-2\kappa}f(r_0+p^\kappa x)$. A Taylor expansion of $g(x)$ at $0$ gives
\begin{align*}
g(x) &= p^{-2\kappa}f(r_0) + p^{-\kappa}f'(r_0) x + \tfrac{f''(r_0)}{2!} x^2 +\dots
+ p^{(m-2)\kappa}\,\tfrac{f^{(m)}(r_0)}{m!} x^m\\
&= c_0 + c_1 x + c_2 x^2 +\dots + c_m x^m.
\end{align*}
If $0\le 2\kappa<\nu$, then $v_p(c_0)\ge 1$ and $v_p(c_1)=0$ since $v_p(f(r_0))\ge \nu>2\kappa$ and $v_p(f'(r_0))=\kappa$. At this point, we can proceed as in the proof of Theorem~\ref{basicHensel} and conclude that the formal root of $g(x)$ provided by \eqref{formalRoot} is indeed a $p$-adic root of $g(x)$. If we denote that root by $\varrho$, then $r=r_0+p^\kappa \varrho$ is a $p$-adic root of the polynomial $f(x)$.
\end{proof}
In the case of quadratic and cubic polynomials, one can use known properties of Bell polynomials to give a simpler representation of the corresponding Hensel's roots.
\subsection*{Quadratic Polynomials}
Let $p>2$ and $f(x)=a_0+a_1 x+a_2 x^2 \in \mathbb{Z}_p[x]$, $a_2\not=0$. If there is an $r_0\in\mathbb{Z}$ such that $f(r_0)\equiv 0$ (mod $p$) and $v_p(f'(r_0))=0$, then by Theorem~\ref{basicHensel} and elementary Bell polynomial identities, the $p$-adic lift of $r_0$ may be written as
\begin{equation} \label{quadraticRoot}
r = r_0 - \frac{c_0}{c_1}\sum_{n=0}^\infty \binom{2n}{n}\frac{1}{n+1} \Big(\frac{c_0 c_2}{c_1^2}\Big)^n\in\mathbb{Z}_p,
\end{equation}
where $c_0=f(r_0)$, $c_1=f'(r_0)$, and $c_2=a_2$. Note that $\binom{2n}{n}\frac{1}{n+1}\in\mathbb{Z}$ are the well-known Catalan numbers.
\begin{example}
Let us consider $f(x)=1+11x-5x^2$ over $\mathbb{Z}_7$. This polynomial has two simple roots mod 7, $r_0=1, 4$. Since $c_0=f(1)=7$, $c_1=f'(1)=1$, and $c_2=-5$, the lift of $r_0=1$ in $\mathbb{Z}_7$ is given by
\begin{equation*}
r = 1 - \sum_{n=0}^\infty \binom{2n}{n}\frac{(-5)^n}{n+1}\, 7^{n+1}.
\end{equation*}
On the other hand, since $f(4)=-35$ and $f'(4)=-29$, the lift of $r_0=4$ in $\mathbb{Z}_7$ is given by
\begin{equation*}
r = 4 - \sum_{n=0}^\infty \binom{2n}{n}\frac{1}{n+1} \Big(\frac{5}{29}\Big)^{2n+1} 7^{n+1}.
\end{equation*}
Note that $1/29=1+3\cdot 7+7^2+7^3+2\cdot 7^4+5\cdot 7^5+O(7^6)$ is an element of $\mathbb{Z}_7$.
\end{example}
\begin{example}
We now consider $f(x)=17+6x+2x^2$ over $\mathbb{Z}_5$. Modulo $5$ this polynomial has a double root, $r_0=1$. Since $f(1)=5^2$ and $f'(1)=2\cdot 5$, we cannot apply any of the above theorems directly. However, the polynomial
\begin{equation*}
g(x) = \tfrac{1}{25}f(1+5x) = 1+2x+2x^2
\end{equation*}
has $1$ and $3$ as simple roots modulo $5$, so using \eqref{quadraticRoot}, we get the lifts
\begin{gather*}
1-\frac{5}{6}\sum_{n=0}^\infty \binom{2n}{n}\frac{1}{n+1} \Big(\frac{5}{18}\Big)^n \;\text{ and }\;
3-\frac{25}{14}\sum_{n=0}^\infty \binom{2n}{n}\frac{1}{n+1} \Big(\frac{25}{98}\Big)^n \;\text{ in } \mathbb{Z}_{5}.
\end{gather*}
Therefore, the $5$-adic roots of $f(x)$ are given by
\begin{gather*}
1+ 5-\frac{25}{6}\sum_{n=0}^\infty \binom{2n}{n}\frac{1}{n+1} \Big(\frac{5}{18}\Big)^n \;\text{ and }\;
1+ 3\cdot 5 -\frac{125}{14}\sum_{n=0}^\infty \binom{2n}{n}\frac{1}{n+1} \Big(\frac{25}{98}\Big)^n.
\end{gather*}
\end{example}
\subsection*{Cubic Polynomials}
Let $p>2$ and $f(x)=a_0+a_1 x+a_2 x^2 + a_3 x^3\in \mathbb{Z}_p[x]$, $a_3\not=0$. Once again, if there is an $r_0\in\mathbb{Z}$ such that $f(r_0)\equiv 0$ (mod $p$) and $v_p(f'(r_0))=0$, then Theorem~\ref{basicHensel} gives a formula for the $p$-adic lift of $r_0$. However, for cubic polynomials, it is more convenient to use the equation \eqref{altForm} and write the root as
\begin{equation*}
r=r_0+\sum_{n=0}^\infty \bigg[\sum_{k=0}^n \frac{(-1)^{n+k+1}}{c_1^k\, n!} \frac{(n+k)!}{(n+1)!}B_{n,k}(c_2, 2c_3,0,\dots)\bigg] \Big(\frac{c_0}{c_1}\Big)^{n+1},
\end{equation*}
where $c_j=f^{(j)}(r_0)/j!$ for $j=0,1,2,3$. Note that $B_{n,k}(c_2,2c_3,0,\dots)=0$ for $k < n/2$, and for $k\ge n/2$, identities \eqref{Comt3n} and \eqref{Comt3n'} give
\begin{align*}
B_{n,k}(c_2,2c_3,0,\dots)
&= \sum_{\substack{\kappa\le k\\ \nu\le n}} \binom{n}{\nu}B_{\nu,\kappa}(c_2,0,\dots) B_{n-\nu,k-\kappa}(0,2c_3,0,\dots) \\
&= \sum_{\kappa\le k} \binom{n}{\kappa}c_2^\kappa B_{n-\kappa,k-\kappa}(0,2c_3,0,\dots) \\
&= \binom{n}{2k-n}c_2^{2k-n} \frac{[2(n-k)]!}{(n-k)!}c_3^{n-k} \\
&= \frac{n!}{(2k-n)!(n-k)!}\, c_2^{2k-n}c_3^{n-k}.
\end{align*}
Therefore,
\begin{align*}
r &= r_0+\sum_{n=0}^\infty \bigg[\sum_{k\ge n/2}^n \frac{(-1)^{n+k+1}}{c_1^k\, n!} \frac{(n+k)!}{(n+1)!}
\frac{n!}{(2k-n)!(n-k)!}\, c_2^{2k-n}c_3^{n-k}\bigg] \Big(\frac{c_0}{c_1}\Big)^{n+1},
\intertext{and with the change $n=2k-j$,}
&= r_0+ \frac{c_0}{c_1}\sum_{k=0}^\infty \bigg[\sum_{j=0}^k \frac{(-1)^{k-j+1}}{c_1^k(2k-j+1)}
\binom{k}{j}\binom{3k-j}{k} c_2^{j}c_3^{k-j}\bigg] \Big(\frac{c_0}{c_1}\Big)^{2k-j}.
\end{align*}
In summary, if $r_0$ is a simple root mod $p$ of $f(x)=a_0+a_1 x+a_2 x^2 + a_3 x^3\in\mathbb{Z}_p[x]$, then the $p$-adic lift of $r_0$ is given by
\begin{equation}\label{cubicRoot}
r = r_0-\frac{c_0}{c_1}\sum_{k=0}^\infty \bigg[\sum_{j=0}^k \frac{(-1)^{k-j}c_2^j}{2k-j+1}
\binom{k}{j}\binom{3k-j}{k}\Big(\frac{c_0c_3}{c_1}\Big)^{k-j} \bigg] \Big(\frac{c_0}{c_1^2}\Big)^{k},
\end{equation}
where $c_0=f(r_0)$, $c_1=f'(r_0)$, $c_2=f''(r_0)/2$, and $c_3=a_3$.
\bigskip
Following the same steps as for cubic polynomials, we can obtain the following result.
\begin{proposition}
Let $p>2$ and $1<\ell<m$. Suppose $f(x)=a_0+a_1 x+a_{\ell} x^{\ell} + a_{m} x^{m} \in \mathbb{Z}_p[x]$ is such that $p\,|\,f(0)$ but $p\!\not|\, f'(0)$. Then $r_0=0$ lifts to a $p$-adic root of $f$ given by
\begin{equation*}
r=- \frac{a_0}{a_1}\sum_{k=0}^\infty \bigg[\sum_{j=0}^k
\frac{(-1)^{m(k-j)+\ell j}\, a_\ell^{j}}{m(k-j)+\ell j-k+1} \binom{k}{j}\binom{m(k-j)+\ell j}{k}
\Big(\frac{a_0^{m-\ell}a_m}{a_1^{m-\ell}}\Big)^{k-j}\bigg]
\Big(\frac{a_0^{\ell-1}}{a_1^\ell}\Big)^{k}.
\end{equation*}
\end{proposition}
\subsection*{Roots of Unity}
Let $p>2$ and $f(x)=x^m-1$. Assume that $r_0$ is a single root mod $p$ of $f$. Then $r_0$ lifts to a $p$-adic root of the form \eqref{HenselRoot} with $c_0=r_0^m-1$ and $c_j=\binom{m}{j}r_0^{m-j}$ for $j=1,\dots,m$. Now, since
\begin{equation*}
j!c_j = r_0^{m-j}(m)_j \;\text{ with }\; (m)_j=\frac{m!}{(m-j)!},
\end{equation*}
homogeneity properties of the Bell polynomials together with identity \eqref{WangEx3.2} give
\begin{align*}
B_{n+k,k}(1! c_2, 2!c_3,\dots) &= B_{n+k,k}(r_0^{m-1}(m)_1,r_0^{m-2}(m)_2,\dots) \\
&= r_0^{mk-(n+k)} B_{n+k,k}((m)_1,(m)_2,\dots) \\
&= r_0^{mk-(n+k)} \frac{1}{k!} \sum_{j=0}^k (-1)^{k-j} \binom{k}{j} (jm)_{n+k}.
\end{align*}
Therefore,
\begin{align*}
\sum_{k=0}^n & \frac{(-1)^{n-k+1}}{c_1^k\,(n+1)!} \binom{2n+1}{n-k} B_{n+k,k}(1!c_1, 2!c_2,\dots) \\
&= \sum_{k=0}^n \frac{(-1)^{n-k+1}}{(r_0^{m-1}m)^k\,(n+1)!} \binom{2n+1}{n-k}
r_0^{mk-(n+k)} \frac{1}{k!} \sum_{j=0}^k (-1)^{k-j} \binom{k}{j} (jm)_{n+k} \\
&= \sum_{k=0}^n \sum_{j=0}^k\, \frac{(-1)^{n-j+1}}{r_0^{n}\,m^k\,(n+1)!} \binom{2n+1}{n-k}
\frac{1}{k!} \binom{k}{j} (jm)_{n+k}.
\end{align*}
Finally, the $p$-adic lift of $r_0$ is given by
\begin{align*}
r &= r_0 - \frac{c_0}{c_1}\sum_{n=0}^\infty \bigg[\sum_{k=0}^n \sum_{j=0}^k\, \frac{(-1)^{n-j}}{m^k\,(n+1)!}
\binom{2n+1}{n-k} \frac{1}{k!} \binom{k}{j} (jm)_{n+k}\bigg] \Big(\frac{c_0}{r_0c_1}\Big)^n,
\end{align*}
where $c_0=f(r_0)$ and $c_1=f'(r_0)$.
\bigskip
For $m=p-1$, the above expression gives an explicit formula for the Teichm\"uller lifts.
\begin{proposition}\label{roots_unity}
Let $p>2$. Every integer $q\in\{1,\dots,p-1\}$ is a $(p-1)$-st root of unity mod $p$ and lifts to a $p$-adic root of unity $\xi_q$ given by
\begin{equation*}
\xi_q = q-\frac{c_0}{c_1}\sum_{n=0}^\infty \bigg[\sum_{k=0}^n \sum_{j=0}^k\,
\frac{(-1)^{n-j}}{(p-1)^k\,(n+1)!} \binom{2n+1}{n-k} \frac{1}{k!} \binom{k}{j} (j(p-1))_{n+k}\bigg]
\Big(\frac{c_0}{qc_1}\Big)^{n},
\end{equation*}
where $c_0=q^{p-1}-1$ and $c_1=(p-1)q^{p-2}$.
\end{proposition}
\section{Factorization of polynomials over $\mathbb{Z}[[x]]$}\label{sec:Factorization}
Let $f(x)=f_0+f_1x+f_2x^2+\dots$ be a formal power series in $\mathbb{Z}[[x]]$. It is easy to prove that $f(x)$ is invertible in $\mathbb{Z}[[x]]$ if and only if $|f_0|=1$. A natural question, initially discussed in \cite{BiGi}, is whether or not a non-invertible element of $\mathbb{Z}[[x]]$ can be factored over $\mathbb{Z}[[x]]$. In recent years, this question has been investigated by several authors, leading to sufficient and in some cases necessary reducibility criteria, see e.g. \cite{Bezivin, BGW2012a, Elliott}. In particular, \cite{Elliott} deals with the factorization of formal power series over principal ideal domains.
For the case at hand, the following elementary results are known. The formal power series $f(x)=f_0+f_1x+f_2x^2+\dots$ is irreducible in $\mathbb{Z}[[x]]$ if $|f_0|$ is prime, or if $|f_0|=p^w$ with $p$ prime, $w\in\mathbb{N}$, and $\gcd(p,f_1)=1$.
On the other hand, if $f_0$ is neither a unit nor a prime power, then $f(x)$ is reducible. In this case, the factorization algorithm is simple and relies on a recursion and a single diophantine equation, see \cite[Prop.~3.4]{BiGi}.
Finally, in the remaining case when $f_0$ is a prime power and $f_1$ is divisible by $p$, the reducibility of $f(x)$ in $\mathbb{Z}[[x]]$ is linked to the existence of a $p$-adic root of positive valuation. The goal of this section is to give an explicit factorization over $\mathbb{Z}[[x]]$ for reducible polynomials of the form
\begin{equation}\label{polynomial}
f(x)=p^{w}+p^m\gamma_1 x+\gamma_2 x^2+\dots+\gamma_d x^d, \quad m\ge 1,\; w\ge 2, \;d\ge 2,
\end{equation}
where $\gamma_1,\dots, \gamma_d\in\mathbb{Z}$ and $\gcd(p,\gamma_1)=1$. This is the only type of polynomial for which the reducibility and factorization over $\mathbb{Z}[[x]]$ is not straightforward.
\begin{theorem}\label{FactorizationThm}
Let $p$ be an odd prime and let $f(x)$ be a polynomial of the form \eqref{polynomial}. Assume that $f$ has a simple root $r\in p\mathbb{Z}_p$ with $v_p(r)=\ell\le m$ and $r=p^\ell(1+\sum_{j=1}^\infty e_j p^{\ell j})$ with $e_j\in\mathbb{Z}$. Then $f(x)$ admits the factorization
\begin{equation*}
f(x) =\bigg(p^\ell-x-x\sum_{n=1}^\infty a_n x^n\bigg)
\bigg(p^{w-\ell}+(p^{w-2\ell}+p^{m-\ell}\gamma_1) x + x\sum_{n=1}^\infty b_n x^n\bigg),
\end{equation*}
where the coefficients $a_n$ are given by \eqref{Acoeff}, and $b_n=\hat b_n/p^{\ell n}$ with $\hat b_n$ as in \eqref{Bhat_coeff}.
\end{theorem}
\begin{remark}
\begin{enumerate}[(a)]
\item As shown in Lemma~\ref{bhat_divisibility}, $\hat b_n$ is divisible by $p^{\ell n}$, so $b_n\in\mathbb{Z}$ for every $n$.
\item If $r\in p\mathbb{Z}_p$ is a root of $f$ with $v_p(r)=\ell\le m$, then $2\ell\le w$.
\item If $w\le 2m$ and $f$ has a root $r\in p\mathbb{Z}_p$, then $v_p(r)=\ell\le m$ holds. If $w>2m$, then $0$ lifts to a $p$-adic root of $f$, but it is not necessarily true that $f$ has a root of valuation less than or equal to $m$. This property depends on the coefficients $\gamma_2, \gamma_3,\dots$. However, even if that condition fails, $f(x)$ is still reducible and a factorization can be obtained through the algorithm given in \cite[Prop.~2.4]{BGW2012a}.
\item A $p$-adic integer $r$ with $v_p(r)=\ell$ can always be written as $r=p^\ell(e_0+\sum_{j=1}^\infty e_j p^{\ell j})$ with $e_0\in\mathbb{Z}_p^*$. For factorization purposes, we can assume without loss of generality $e_0=1$. Otherwise, consider $g(x)=f(x/e_0^*)$, where $e_0^*$ is such that $e_0e_0^*=1\!\pmod{p^\ell}$.
\item As discussed in \cite{BGW2012a}, the existence of a root in $p\mathbb{Z}_p$ is in many cases (e.g. when $d\le 3$) a necessary condition for the polynomial \eqref{polynomial} to factor over $\mathbb{Z}[[x]]$.
\end{enumerate}
\end{remark}
\begin{remark}
If $f$ has a multiple root in $p\mathbb{Z}_p$, then $f(x)$ admits the simpler factorization
\begin{equation*}
f(x)= G(x) f_{\text{red}}(x),
\end{equation*}
where $G(x)=\gcd(f(x),f'(x))\in\mathbb{Z}[x]$ and $f_{\text{red}}(x)=f(x)/G(x)$.
\end{remark}
\subsection*{Proof of Theorem~\ref{FactorizationThm}}
Let $r=p^\ell\big(1+\sum_{j=1}^\infty e_j p^{\ell j}\big)$ be the $p$-adic root of $f$ and define
\begin{equation}\label{Eseries}
\phi(x)=xE(x) \quad\text{with}\quad E(x)=1+\sum_{j=1}^\infty e_j x^j.
\end{equation}
Thus $r=\phi(p^\ell)$ in $\mathbb{Z}_p$ and therefore $p^\ell=\phi^{-1}(r)$. Define $A(x)=p^\ell-\phi^{-1}(x)$. So $A(r)=0$ in $\mathbb{Z}_p$, and by Lemma~\ref{FormalInversion}, we have
\begin{equation*}
A(x)=p^\ell-\phi^{-1}(x)=p^\ell-x\Big(1+\sum_{n=1}^\infty a_n x^n\Big),
\end{equation*}
where
\begin{equation} \label{Acoeff}
a_n = \frac{1}{n!}\sum_{k=1}^n (-1)^k \frac{(n+k)!}{(n+1)!}B_{n,k}(1!e_1,2!e_2,\dots) \in\mathbb{Z}.
\end{equation}
Our goal is to find $B(x)\in \mathbb{Z}[[x]]$ such that $f(x)=A(x)B(x)$. For convenience, consider
\begin{equation*}
\hat f(x)=p^{-2\ell}f(p^\ell x) \quad\text{and}\quad \hat A(x) = p^{-\ell}A(p^\ell x).
\end{equation*}
Thus
\begin{equation*}
\hat A(x) = 1 - x - x\sum_{n=1}^\infty p^{\ell n}a_n x^n.
\end{equation*}
\begin{proposition}
The reciprocal of $\hat A(x)$ is a power series in $\mathbb{Z}[[x]]$ of the form
\begin{equation*}
\hat A(x)^{-1} = \frac{1}{\hat{A}(x)} = 1+ x + x\sum_{n=1}^\infty t_n x^n
\end{equation*}
with
\begin{equation} \label{reciprocalcoeff}
t_n = 1+ \sum_{k=1}^{n} p^{\ell k}\,\frac{n+1-k}{k!}
\sum_{{j}=1}^{k}(-1)^{j} \frac{(n+{j})!}{(n+1)!} B_{k,{j}}(1!e_1,2! e_2,\dots) \in\mathbb{Z}.
\end{equation}
\end{proposition}
\begin{proof}
By Theorem~B in \cite[Section~3.5]{Comtet}, and using basic properties of partial Bell polynomials, we have
\begin{align*}
\hat A(x)^{-1}
&=1+x+\sum_{n=2}^\infty \sum_{k=1}^n \frac{k!}{n!} B_{n,k}(1,2!a_1p^\ell,3!a_2p^{2\ell},\dots)\, x^n \\
&=1+x+\sum_{n=2}^\infty \sum_{k=1}^n k! \biggl(\sum_{{j}=0}^k\frac{n!}{(n-k)!{j}!}
B_{n-k,k-{j}}(1! p^{\ell} a_1,2!p^{2\ell}a_2,\dots) \biggr) \frac{x^n}{n!}\\
&=1+x+\sum_{n=1}^\infty \left(1+\sum_{k=1}^{n} \biggl(\sum_{{j}=1}^{n+1-k}\frac{(n+1-k)!}{k!(n+1-k-{j})!}
p^{\ell k}B_{k,{j}}(1!a_1,2!a_2,\dots) \biggr) \right) x^{n+1}\\
&=1+x+x\sum_{n=1}^\infty \left(1+\sum_{k=1}^{n} \frac{p^{\ell k}}{k!}
\biggl(\sum_{{j}=1}^{k}\frac{(n+1-k)!}{(n+1-k-{j})!} B_{k,{j}}(1!a_1,2!a_2,\dots) \biggr) \right) x^n.
\end{align*}
In the last step, we declare the interior sum to be zero if ${j}> \min(k,n+1-k)$. Thus
\begin{align*}
t_n &=1+\sum_{k=1}^{n} \frac{p^{\ell k}}{k!}\sum_{{j}=1}^{k}\frac{(n+1-k)!}{(n+1-k-{j})!}
B_{k,{j}}(1!a_1,2!a_2,\dots) \\
&=1+\sum_{k=1}^{n} p^{\ell k}\,\frac{n+1-k}{k!}\sum_{{j}=1}^{k}\binom{n-k}{{j}-1}({j}-1)!
B_{k,{j}}(1!a_1,2!a_2,\dots)
\end{align*}
Now, if we write $k!a_k$ as
\begin{equation*}
k!a_k = \sum_{{j}=1}^k \binom{k+{j}}{{j}-1}({j}-1)! B_{k,{j}}(-1!e_1,-2!e_2,\dots),
\end{equation*}
then by means of Theorem~15 in \cite{BGW2012b} we get
\begin{align*}
\sum_{{j}=1}^{k}\binom{n-k}{{j}-1}({j}-1)! & B_{k,{j}}(1!a_1,2!a_2,\dots) \\
&= \sum_{{j}=1}^{k} \binom{n+{j}}{{j}-1}({j}-1)! B_{k,{j}}(-1!e_1,-2! e_2,\dots) \\
&= \sum_{{j}=1}^{k}(-1)^{j} \frac{(n+{j})!}{(n+1)!} B_{k,{j}}(1!e_1,2! e_2,\dots).
\end{align*}
In other words, $t_n$ has the form claimed in \eqref{reciprocalcoeff}.
\end{proof}
Now, motivated by \eqref{reciprocalcoeff}, for $n\ge 1$ we consider
\begin{equation*}
T_n(x) = 1+ \sum_{k=1}^{\infty} \frac{n+1-k}{k!}
\bigg(\sum_{{j}=1}^{k}(-1)^{j} \frac{(n+{j})!}{(n+1)!} B_{k,{j}}(1!e_1,2! e_2,\dots)\bigg) x^k.
\end{equation*}
\begin{lemma}
With $E(x)$ as in \eqref{Eseries}, we have
\begin{equation*}
T_{n}(x) = E(x)^{-n-2}\big(E(x)+xE'(x)\big).
\end{equation*}
\end{lemma}
\begin{proof}
Fix $n\ge 1$ and denote
\begin{equation*}
\tau_k = \frac{1}{k!} \sum_{{j}=1}^{k}(-1)^{j} \frac{(n+{j})!}{n!} B_{k,{j}}(1!e_1,2! e_2,\dots).
\end{equation*}
Then
\begin{equation*}
T_n(x) = 1+ \sum_{k=1}^{\infty} \Big(1-\frac{k}{n+1}\Big)\tau_k x^k
= 1+ \sum_{k=1}^{\infty} \tau_k x^k - \frac{1}{n+1}\sum_{k=1}^{\infty}k\tau_k x^k.
\end{equation*}
It is easy to check that $1+\sum_{k=1}^{\infty} \tau_k x^k=E(x)^{-n-1}$. Therefore,
\begin{align*}
T_n(x) &=E(x)^{-n-1} - \frac{1}{n+1} x \frac{d}{dx}\Big(E(x)^{-n-1}\Big) \\
&= E(x)^{-n-1} + xE(x)^{-n-2}E'(x) = E(x)^{-n-2}\big(E(x)+xE'(x)\big).
\end{align*}
\end{proof}
As a direct consequence of this lemma we get the recurrence relation
\begin{equation*}
T_{n-1}(x) = E(x)T_{n}(x),
\end{equation*}
which can be used to define $T_0(x)$ and $T_{-n}(x)$ for $n\ge 1$. More precisely, we let
\begin{equation*}
T_0(x)=E(x)T_1(x) \;\text{ and }\;
T_{-n}(x) = E(x)^{n+1} T_1(x) \text{ for } n\ge 1.
\end{equation*}
Given that
\begin{equation}\label{eq:fhat}
\hat f(x) = p^{-2\ell}f(p^\ell x) = p^{w-2\ell}+p^{m-\ell}\gamma_1 x + \sum_{n=2}^d p^{\ell(n-2)}\gamma_n x^n,
\end{equation}
the relation $T_{n-j}(x) = E(x)^jT_{n}(x)$ gives
\begin{equation}\label{polyIdentity}
p^{w-2\ell}T_n(x) + p^{m-\ell}\gamma_1 T_{n-1}(x) + \sum_{j=2}^d p^{\ell(j-2)}\gamma_j T_{n-j}(x) = \hat f(E(x)) T_n(x).
\end{equation}
Moreover, since $E(x)$ is a unit in $\mathbb{Z}[[x]]$, for every $\nu\in\mathbb{Z}$ the function $T_\nu(x)$ is in $\mathbb{Z}[[x]]$ and so $T_\nu(p^\ell)\in\mathbb{Z}_p$.
\begin{lemma}\label{TnCongruences}
For $\nu\ge -1$, the $p$-adic numbers $T_\nu(p^\ell)$ satisfy
\begin{equation*}
T_\nu(p^\ell)-t_\nu \equiv 0 \!\pmod{p^{\ell(\nu+2)}}
\end{equation*}
with $t_\nu$ as in \eqref{reciprocalcoeff} for $\nu>0$ and $t_0=t_{-1}=1$.
\end{lemma}
\begin{proof}
For $\nu=n\ge1$ the statement is a consequence of the fact that $t_n$ is the $n$-th partial sum of $T_n(p^\ell)$ and the coefficient of $x^{n+1}$ in $T_n(x)$ is zero. Further, given that
\begin{equation*}
E(p^\ell)=1+p^\ell e_1+O(p^{2\ell}) \;\text{ and }\; T_1(p^\ell)=1-p^\ell e_1 + O(p^{2\ell}),
\end{equation*}
we have
\begin{equation*}
T_0(p^\ell) = \big(1+p^\ell e_1+O(p^{2\ell})\big)\big(1-p^\ell e_1 + O(p^{2\ell})\big)\equiv 1 \!\pmod{p^{2\ell}}.
\end{equation*}
This implies $T_0(p^\ell)-t_0\equiv 0 \!\pmod{p^{2\ell}}$.
Finally, since $T_{-1}(p^\ell)=E(p^\ell)^{2}T_1(p^\ell)$, and because $E(p^\ell)^{2}$ and $T_1(p^\ell)$ are both of the form $1+O(p^\ell)$, we get $T_{-1}(p^\ell)\equiv 1 \!\pmod{p^{\ell}}$, hence $T_{-1}(p^\ell) - t_{-1}\equiv 0 \!\pmod{p^{\ell}}$.
\end{proof}
Using $\hat f(x)$ as in \eqref{eq:fhat}, we now define
\begin{align*}
\hat B(x) = \hat f(x)\hat A(x)^{-1}
= \Big(p^{w-2\ell}+p^{m-\ell}\gamma_1 x + \sum_{n=2}^d p^{\ell(n-2)}\gamma_n x^n\Big)
\Big(1+ x + x\sum_{n=1}^\infty t_n x^n\Big),
\end{align*}
and write it as
\begin{equation*}
\hat B(x) = p^{w-2\ell}+(p^{w-2\ell}+p^{m-\ell}\gamma_1) x + x\sum_{n=1}^\infty \hat b_n x^n
\end{equation*}
with
\begin{equation} \label{Bhat_coeff}
\hat b_n = p^{w-2\ell} t_n + p^{m-\ell}\gamma_1 t_{n-1} +\sum_{j=2}^d p^{\ell(j-2)}\gamma_j t_{n-j} \in\mathbb{Z},
\end{equation}
where $t_n$ is given by \eqref{reciprocalcoeff}, $t_0=t_{-1}=1$, and $t_{-n}=0$ for $n>1$.
\begin{lemma} \label{bhat_divisibility}
The coefficients $\hat b_n$ are divisible by $p^{\ell n}$.
\end{lemma}
\begin{proof}
First of all, since $p^{-\ell}r=E(p^\ell)$ is a $p$-adic root of $\hat f$, identity \eqref{polyIdentity} implies
\begin{equation*}
p^{w-2\ell}T_n(p^\ell) + p^{m-\ell}\gamma_1 T_{n-1}(p^\ell) + \sum_{j=2}^d p^{\ell(j-2)}\gamma_j T_{n-j}(p^\ell) = 0 \text{ in } \mathbb{Z}_p.
\end{equation*}
Therefore, for $n\ge d-1$,
\begin{align*}
\hat b_n &= p^{w-2\ell}t_n + p^{m-\ell}\gamma_1 t_{n-1} +\sum_{j=2}^d p^{\ell(j-2)}\gamma_j t_{n-j} \\
&= p^{w-2\ell}\big(t_n-T_n(p^\ell)\big) + p^{m-\ell}\gamma_1 \big(t_{n-1}-T_{n-1}(p^\ell)\big) +
\sum_{j=2}^d p^{\ell(j-2)}\gamma_j \big(t_{n-j}-T_{n-j}(p^\ell)\big),
\end{align*}
which by Lemma~\ref{TnCongruences} is congruent to $0$ mod $p^{\ell n}$. Similarly, for $1\le n <d-1$,
\begin{align*}
\hat b_n &= p^{w-2\ell}t_n + p^{m-\ell}\gamma_1 t_{n-1} +\sum_{j=2}^{n+1} p^{\ell(j-2)}\gamma_j t_{n-j} \\
&\equiv - \sum_{j=n+2}^d p^{\ell(j-2)}\gamma_j T_{n-j}(p^\ell) \equiv 0 \!\pmod{p^{\ell n}}.
\end{align*}
\end{proof}
\smallskip
Finally, defining $B(x)=p^{\ell}\hat B(x/p^\ell)$, we arrive at the factorization $f(x) = A(x)B(x)$.
\begin{remark}
It is worth mentioning that our method for factorization in $\mathbb{Z}[[x]]$ is not restricted to polynomials and can be applied to power series. As an example, consider
\begin{equation*}
f(x) = 9+12x+7x^2+8x^3\sum_{k=0}^\infty x^k = 9+12x+7x^2+\frac{8x^3}{1-x},
\end{equation*}
discussed by B\'ezivin in \cite{Bezivin}. This series is reducible in $\mathbb{Z}[[x]]$ and factors as
\begin{equation*}
f(x)=\dfrac{(3-x)^2(1+x)}{1-x}.
\end{equation*}
The reader is invited to confirm that the power series version of Theorem~\ref{FactorizationThm} gives the factorization $f(x)=A(x)B(x)$ with $A(x)=3-x$ and $B(x)=\frac{(3-x)(1+x)}{1-x}$.
An interesting feature of this example is that the partial sums $f_d(x) = 9+12x+7x^2+\cdots$ of $f(x)$ of degree $d\ge 2$ are all irreducible in $\mathbb{Z}[[x]]$. This fact was proved in \cite[Prop.~8.1]{Bezivin}, but it can also be derived from Proposition~3.4 of \cite{BGW2012b} together with the observation that for $d\ge 2$, the polynomial $f_d(x)$ has no roots in $p\mathbb{Z}_p$.
\end{remark}
\section*{Appendix: Some properties of Bell polynomials}\label{sec:appendix}
\setcounter{section}{1}
\setcounter{equation}{0}
\renewcommand{\thesection}{\Alph{section}}
Throughout this paper, we make extensive use of the well-known partial Bell polynomials. For any sequence $x_1,x_2,\dots$, the $(n,k)$-th partial Bell polynomial is defined by
\begin{equation*}
B_{n,k}(x) =\sum_{i\in\pi(n,k)}\frac{n!}{i_1!i_2!\cdots}\left(\frac{x_1}{1!}\right)^{i_1}\left(\frac{x_2}{2!}\right)^{i_2}\cdots,
\end{equation*}
where $\pi(n,k)$ is the set of all sequences $i=(i_1,i_2,\dots)$ of nonnegative integers such that
\begin{equation*}
i_1+i_2+\dots=k \;\text{ and }\; i_1+2i_2+3i_3+\dots=n.
\end{equation*}
Clearly, these polynomials satisfy the homogeneity relation
\begin{equation*}
B_{n,k}(abx_1,ab^2x_2,ab^3x_3,\dots) = a^kb^n B_{n,k}(x_1,x_2,x_3,\dots).
\end{equation*}
Here are other elementary identities (cf. \cite[Section 3.3]{Comtet}) needed in this paper:
\begin{gather}
\label{Comt3l'}
B_{n,k}(\tfrac{x_2}{2},\tfrac{x_3}{3},\dots) = \frac{n!}{(n+k)!} B_{n+k,k}(0,x_2,x_3,\dots), \\
\label{Comt3n}
B_{n,k}(x_1+x_1',x_2+x_2',\dots) = \sum_{\substack{\kappa\le k \\ \nu\le n}}
\binom{n}{\nu}B_{\nu,\kappa}(x_1,x_2,\dots) B_{n-\nu,k-\kappa}(x_1',x_2',\dots), \\
\label{Comt3n'}
B_{n,k}(0,\dots,0,x_j,0,\dots) = 0, \text{ except } B_{jk,k}=\frac{(jk)!}{k!(j!)^k}\, x_j^k.
\end{gather}
Also of special interest is the identity
\begin{equation}\label{WangEx3.2}
B_{n,k}((a)_1,(a)_2,\dots) = \frac{1}{k!} \sum_{j=0}^k (-1)^{k-j} \binom{k}{j} (ja)_{n},
\end{equation}
where $(a)_n=a(a-1)\cdots(a-n+1)$. This is a special case of \cite[Example~3.2]{WW09}.
For more on Bell polynomials and their applications, see e.g. \cite{Bell,Charalambides,Comtet,WW09}.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 9,308
|
Hübners Who is Who war eine Reihe von Personenlexika des ehemaligen Unternehmens Who is Who Verlag für Personenenzyklopädien AG mit Sitz in Zug in der Schweiz. Die Vermarktung erfolgte mittels Handelsvertretern in zehn weiteren Ländern: Österreich, Deutschland, Polen, Slowakei, Tschechien, Ungarn, Russland, Türkei, Rumänien und Griechenland.
Entwicklung und Arbeitsweise
1978 erschien die erste Ausgabe der Biographiensammlung Who is Who in Österreich mit ca. 6000 Einträgen. Um der steigenden Anzahl von Neuaufnahmen nachzukommen, wurde 1984 ein Handelsvertreternetzwerk installiert. Die Vermarktung erfolgte in Form von persönlichen Interviews, in deren Rahmen verschiedene Zusatzleistungen verkauft wurden, die das Kerngeschäft des Verlages darstellten. 1989 erschien zum ersten Mal Who is Who in der Bundesrepublik Deutschland. 2002 erschienen in Polen und Tschechien die Werke Who is Who w Polsce und Who is…? (v České republice). Im Jahr darauf wurde das Verlagsprogramm mit Who is Who Magyarországon und Who is Who v Slovenskej republike auf Ungarn und die Slowakei ausgeweitet. 2005 wurden vom Verlag erstmals Biographiensammlungen in Russland, Rumänien, Griechenland und der Türkei herausgegeben.
Aufgenommen wurden Personen, deren Namen der Verlag in allgemein zugänglichen Adress- und Telefonbüchern recherchierte. Die potentiellen Kandidaten wurden von einem Handelsvertreter (später auch oftmals von einer verlagsfremden Person aus der Marketingbranche) aufgesucht. Alle Daten wurden in persönlichen Interviews erhoben und sollten den aufgenommenen Personen in einem Abstand von drei Jahren zur Korrektur und zur Aktualisierung übermittelt werden.
Im Jahr 2016 wurde das Unternehmen nach einem Konkursverfahren liquidiert.
Kritik
Mehrere Zeitschriften und Einrichtungen wie das Fachorgan Informationsmittel für Bibliotheken, die Süddeutsche Zeitung und die Stiftung Warentest äußerten Kritik an der Personenenzyklopädie. Ihre Vorbehalte bezogen sich auf die nicht nachvollziehbaren Aufnahmekriterien und auf unseriöse Geschäftsmethoden. Letztlich hätte der Verlag das Renommee des markenrechtlich nicht schützbaren Begriffes Who is Who und die damit verbundenen Assoziationen zur Vermarktung seines völlig anders ausgerichteten Lexikons genutzt. Insbesondere hätte der Verlag Verwechslungen mit dem Prominentenlexikon Wer ist wer? in Kauf genommen. Als weiterer Kritikpunkt wurde vorgetragen, dass bei den Interviews zur Aufnahme der Kurzbiographie in die Druckfassung des Werkes für den Kauf des Buches geworben worden wäre. Undurchschaubar wäre auch die Verknüpfung der Enzyklopädie mit einer angebotenen "Clubmitgliedschaft" geblieben. Ebenso sei es Praxis, "Interviews" von Unternehmen und Personen unter anderem aus der Telefonmarketingbranche durchführen zu lassen, die ausschließlich auf Provisionsbasis auf den Verkauf des Buches (ab ca. 400 Euro je Band) arbeiteten.
Siehe auch
Liste von Lexika in deutscher Sprache
Literatur
Klaus Schreiber/ Saskia Hedrich, Rezension des Hauptwerks 1996 und des Supplementwerks 1998, in: Informationsmittel für Bibliotheken 1999/Beiheft 9
Reinhold Rühl, Das Geschäft mit der Eitelkeit, Süddeutsche Zeitung vom 21. Juni 2008
Geschäfte mit der Eitelkeit - Die Bauchpinsler, Finanztest 10/2001, Stiftung Warentest
Einzelnachweise
Biografisches Nachschlagewerk
Buchreihe
Ersterscheinung 1978
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 8,789
|
Chen Yang (born 10 July 1991) is a Chinese track and field athlete who competes in the discus throw. She represented her country at the 2016 Rio Olympics, finishing seventh in the final. Her personal best is ,set in 2018.<ref>Yang Chen. All-Athletics.
International competitions
References
External links
Living people
1991 births
Chinese female discus throwers
Olympic athletes of China
Athletes (track and field) at the 2016 Summer Olympics
Athletes (track and field) at the 2018 Asian Games
Asian Games gold medalists for China
Asian Games medalists in athletics (track and field)
Medalists at the 2018 Asian Games
Asian Games gold medalists in athletics (track and field)
Athletes (track and field) at the 2020 Summer Olympics
21st-century Chinese women
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 9,588
|
Dazee has been flying the flag for Jungle and DNB since 1992 and is co- founder of the now legendary 'Ruffneck Ting' alongside Markee Substance (Kosheen). As well as making a name for herself with her dj skills, she's had many successful releases; labels include Ruffneck Ting, Breakbeat Culture, Colourstone and Siren Records. She's performed at the most celebrated Jungle raves and parties (and featured in the tape packs!) including Universe, Dreamscape, Helter Skelter, World Dance and more recently Fantazia, Slamming Vinyl and Raveology. Her pioneering dnb sound has taken her around the globe and she has Featured on Radio 1, BBC 1Xtra, VH1 and IT V.2013 saw the relaunch of Ruffneck Ting with an event at Lakota in Bristol and the release of the classics album RNT20 to celebrate 20 yrs since it began. Now managing the label as well as producing for it, Dazee has just released Ruffneck Ting's first digital ep featuring tracks by herself, K-jah and Jinx. It's received huge support and radio play by the likes of Ray Keith, Bailey and Randall. A remix series is to follow with some of dnb's most respected producers reworking original Ruffneck Ting classics.Dazee describes her sound as "A bouncy mix of rollers and flavoursome jump up: I try to keep it musically colourful with jungle vibes, but forward thinking not just retro and with some deeper or harder edged elements too." This is the sound she hopes will come across with Ruffneck and with the current release schedule of new beats from herself and producers signed to the label Dazee is always guaranteed to be bring fresh Ruffneck Ting dubs to the dancefloor!2013 also saw the release of a mix cd featuring many of the tracks now being released on Ruffnck Ting. This was commissioned by Recreo clothing for Brisfest and also featured the vocal skills of Mc Carasel. In the words of Carasel "Still fresh" months on! To listen/ download: https://soundcloud.com/carasel/recreo-present-dazee- caraselDazee also currently co-hosts The Dutty Bass Show on Bristol's much loved Ujima FM, along side her fellow "Dutties" Dj Diss Miss and Safesoul.As a member of DuttyGirl, Dazee's sound is more diverse, to include a variety ofbass heavy dance genres. Dutty Girl's unique blend ofmixing through genres and featuring their own productionsand original vocal tracks can be heard on their "Call of DuttyCd" as well as in their performances. (For more info on Dutty Girl check out facebook.com/dutty girl)Although she enjoys a more eclectic musical direction as part of Dutty Girl, Dazee will always remain true to her roots and champion jungle and drum and bass in her solo dj sets. Currently resident at Bristol's celebrated monthly dnb night "Bristol Sounds DnB" she is looking forward to exciting forthcoming events in Bristol and beyond, including Fantazia, the Drum and Bass Awards and Love Will Save The Day.
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 1,793
|
<?xml version="1.0" encoding="UTF-8"?>
<web-app version="2.4" xmlns="http://java.sun.com/xml/ns/j2ee"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://java.sun.com/xml/ns/j2ee http://java.sun.com/xml/ns/j2ee/web-app_2_4.xsd">
<display-name>swiftly</display-name>
<!--由Sprng载入的Log4j配置文件位置 -->
<context-param>
<param-name>log4jConfigLocation</param-name>
<param-value>classpath:log/log4j.xml</param-value>
</context-param>
<!--Spring默认刷新Log4j配置文件的间隔,单位为毫秒 -->
<context-param>
<param-name>log4jRefreshInterval</param-name>
<param-value>60000</param-value>
</context-param>
<context-param>
<param-name>webAppRootKey</param-name>
<param-value>swiftly.root</param-value>
</context-param>
<filter>
<filter-name>encodingFilter</filter-name>
<filter-class>org.springframework.web.filter.CharacterEncodingFilter</filter-class>
<init-param>
<param-name>encoding</param-name>
<param-value>utf-8</param-value>
</init-param>
<init-param>
<param-name>forceEncoding</param-name>
<param-value>true</param-value>
</init-param>
</filter>
<listener>
<listener-class>org.springframework.web.context.ContextLoaderListener</listener-class>
</listener>
<!--Spring log4j Config loader -->
<listener>
<listener-class>org.springframework.web.util.Log4jConfigListener</listener-class>
</listener>
<!-- log 4j 4 spring end -->
<listener>
<listener-class>
org.springframework.web.context.request.RequestContextListener
</listener-class>
</listener>
<servlet>
<servlet-name>SpringDispatcher</servlet-name>
<servlet-class>org.springframework.web.servlet.DispatcherServlet</servlet-class>
<init-param>
<param-name>contextConfigLocation</param-name>
<param-value>/WEB-INF/dispatcher.xml</param-value>
</init-param>
<load-on-startup>1</load-on-startup>
</servlet>
<servlet-mapping>
<servlet-name>SpringDispatcher</servlet-name>
<url-pattern>/</url-pattern>
</servlet-mapping>
<session-config>
<session-timeout>30</session-timeout>
</session-config>
<filter-mapping>
<filter-name>encodingFilter</filter-name>
<url-pattern>/*</url-pattern>
<dispatcher>REQUEST</dispatcher>
<dispatcher>FORWARD</dispatcher>
</filter-mapping>
</web-app>
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 3,338
|
{"url":"https:\/\/phys.libretexts.org\/Courses\/University_of_California_Davis\/UCD%3A_Physics_9C_%E2%80%93_Electricity_and_Magnetism\/2%3A_Electrostatic_Energy\/2.2%3A_Electrostatic_Potential","text":"$$\\require{cancel}$$\n\n# Test Charges\n\nAn alternative way to look at electric fields from what we did in Section 1.2 is from the perspective of a test charge. The idea is to use a charged point particle as a means of measuring electric force vectors at various points in space. When the force vectors are all mapped-out, we then divide them by the charge of the point particle, and the new vectors are then the electric field vectors. A common (but somewhat strange) way to write this mathematically is:\n\n$\\overrightarrow E\\left(\\overrightarrow r\\right) = \\lim\\limits_{q_{test}\\rightarrow 0} \\dfrac{\\overrightarrow F_{on\\;q_{test}}}{q_{test}},\\;\\;\\;\\;\\;\\;\\text{where } \\overrightarrow r\\text{ is the position of }q_{test}$\n\nA region around a collection of charge can similarly be tested with a charged point particle. At every point in space, the potential energy of the point charge can be measured, and then the amount of testing charge can be divided out, so that all that remains is a function of the source charges. We write it this way:\n\n$V\\left(\\overrightarrow r\\right) = \\lim\\limits_{q_{test}\\rightarrow 0} \\dfrac{U\\left(q_{test}\\right)}{q_{test}},\\;\\;\\;\\;\\;\\;\\text{where } \\overrightarrow r\\text{ is the position vector of }q_{test}$\n\nThis process maps out a scalar field, since at every point in space is associated a number (not a vector, like in the case of electric field). Just as electric field vectors are not the same as force vectors, the values in this scalar field are not potential energies \u2013 indeed, this can be see even in the units of these numbers, which are joules divided by coulombs. The ratio of joules per coulomb is given its own name: volts. The scalar field we have invented this way is called electrostatic potential. Like an electric field vectors, this is a quantity that is defined at every point in space in the vicinity of some electric charge. Unlike electric field vectors, these quantities are scalars \u2013 they have no direction.\n\nPossibly the most confusing thing to students new to electrostatics is use of the word \"potential\" in \"electrostatic potential.\" This name derives from the fact that it is related to electric potential energy, but these quantities are very different, and the reader is admonished to keep this in mind.\n\n# Superposition\n\nWhen there is more than one source of electric field in the vicinity of a point in space, the contributions of those sources to the field at that point can be added together. This can be seen simply from the text charge approach \u2013 clearly the forces on the test charge can be added together, and when the test charge is divided out, the sum of the electric field vectors remains. We see the same thing for electrostatic potential:\n\n$U\\left(q_{test}\\right) = \\dfrac{q_1q_{test}}{4\\pi\\epsilon_or_1}+\\dfrac{q_2q_{test}}{4\\pi\\epsilon_or_2}+\\dfrac{q_3q_{test}}{4\\pi\\epsilon_or_3}\\dots \\;\\;\\; \\Rightarrow \\;\\;\\; V\\left(\\overrightarrow r\\right)=\\dfrac{U\\left(q_{test}\\right)}{q_{test}}=\\dfrac{q_1}{4\\pi\\epsilon_or_1}+\\dfrac{q_2}{4\\pi\\epsilon_or_2}+\\dfrac{q_3}{4\\pi\\epsilon_or_3}\\dots$\n\nHere $$r_i$$ is the distance from the $$i^{th}$$ source charge to the position in space indicated by the position vector $$\\overrightarrow r$$. Notice that by adopting the $$U\\left(\\infty\\right)=0$$ convention, we have also done so for the electrostatic potential. And like the potential energy, the position that we choose to call the electric potential zero is arbitrary.\n\nAll of the things we developed for electric fields also apply to potentials, with the only difference being that potentials superpose as scalars, not vectors (which actually makes them easier to deal with in many cases). The main point is that when we have a collection of source charges \u2013 including a continuous distribution \u2013 we can define a potential at every point in space, and if we place a point charge there, we can determine its potential energy by multiplying the charge by the electric potential:\n\n$U=qV\\left(\\overrightarrow r\\right),\\;\\;\\;\\;\\;\\;\\text{where }\\overrightarrow r= \\text{position vector of the charge }q$\n\nThe similarity with Equation 1.2.2 is obvious \u2013 we have simply replaced force and field with energy and potential.\n\n# Bridging the Scalar and Vector Fields\n\nIt's clear that both the scalar potential field and the vector electric field are determined by the source charge distribution, so these fields must be related to each other somehow. Indeed they are! To see how, we once again look back to our study of mechanics, where we related potential energy and force. We've already mentioned one such relation, through the computation of work:\n\n$\\Delta U = U_B-U_A = -W_{A\\rightarrow B} = -\\int\\limits_A^B\\overrightarrow F\\cdot \\overrightarrow{dl}$\n\nIf we just plug in $$U=qV and \\overrightarrow F=q \\overrightarrow E, we get a direct relation between the change of potential and the electric field: $U_A-U_B = \\int\\limits_A^B\\overrightarrow E\\cdot \\overrightarrow{dl}$ Note that the signs have been flipped on both sides of the equation. The quantity on the left is usually referred to as the potential drop from A to B. It is actually common to use the units of \\(Vm^{-1}$$ for electric field, rather than our previous $$NC^{-1}$$.\n\nWhenever we have an integral relationship like this, then (as we saw for Gauss's law), a differential (local) relation is also available. We actually saw this back in our study of mechanics, and it comes through here as well:\n\n$\\overrightarrow F = -\\overrightarrow \\nabla U \\;\\;\\; \\Rightarrow \\;\\;\\; \\overrightarrow E = -\\overrightarrow \\nabla V$\n\nWhile this is interesting, the reader can be forgiven for asking what use it has. This last relation is particularly powerful for the following reason. Suppose we wish to compute the electric field of a charge distribution. Assuming we don't have a clever way of using Gauss's law to do this, we have to perform a calculation like we did back in Section 1.3. Part of what makes that computation challenging is keeping track of three different components of the electric field vector (i.e. three separate integrals). If we instead compute the potential field (one integral, with no vectors involved), we can then take derivatives (the gradient) to get the electric field. We will see how one calculates the potential field from a distribution of charge in the next section.\n\nThe relation between field and potential is often misunderstood, in yet another incarnation of confusing a quantity with a change in that quantity (like mistaking acceleration with velocity. Just as zero instantaneous velocity does not mean the acceleration is zero, a zero potential at a point in space does not mean that the field there is zero. Indeed, we can define the potential to be zero anywhere, no matter what the field is! It is the rate of change of the potential that determines the field, not the value of the potential.\n\nExample $$\\PageIndex{1}$$\n\nIn a certain region of space around the origin, the electrostatic potential field satisfies:\n\n$V\\left(x,y,z\\right)=\\alpha\\; x + \\beta \\;y^2 + \\gamma\\; z^3 \\nonumber$\n\nFind the charge density at the origin in terms of $$\\alpha$$, $$\\beta$$, and $$\\gamma$$.\n\nSolution\n\nWe can find the electric field from the potential field:\n\n$\\overrightarrow E = -\\overrightarrow \\nabla V = \\dfrac{\\partial V}{\\partial x}\\;\\widehat i+ \\dfrac{\\partial V}{\\partial y}\\;\\widehat j+ \\dfrac{\\partial V}{\\partial z}\\;\\widehat k = \\alpha\\;\\widehat i + 2\\beta \\;y\\;\\widehat j + 3\\gamma \\;z^2\\;\\widehat k \\nonumber$\n\nNow the divergence of the field gives us the charge density (Gauss's law):\n\n$\\dfrac{\\rho\\left(x,y,z\\right)}{\\epsilon_o} = \\nabla \\cdot \\overrightarrow E = \\dfrac{\\partial E_x}{\\partial x}+\\dfrac{\\partial E_y}{\\partial y}+\\dfrac{\\partial E_z}{\\partial z}=0+2\\beta+6\\gamma \\;z \\;\\;\\; \\Rightarrow \\;\\;\\; \\boxed{\\rho\\left(0,0,0\\right)=2\\beta} \\nonumber$\n\nNotice that at the origin the potential is zero, but the electric field is not, nor is the charge density.\n\n# Equipotential Surfaces\n\nA consequence of the gradient relation is that their relationship is geometric in nature. The first manifestation of this is that the gradient of a scalar field points in the direction where the scalar values are increasing the fastest. With the presence of the negative sign, we therefore conclude that the electric field points in the direction of the fastest descent of electric potential. Put more concisely, electric field points from high electric potential to low electric potential.\n\nWe can demonstrate this geometrical relationship through a diagram. Let's imagine starting at a certain point in space, and measuring the potential there (after designating the zero point). Then we sample nearby points, and find a direction we can move our detector so that the potential doesn't change. If we keep following this procedure, and map the entire space where the potential doesn't change, we will find that it is a surface. As this imaginary surface exists at a single, equal potential, it called an equipotential surface. Here is a two-dimensional depiction of a collection of such surfaces:\n\nFigure 2.2.1 \u2013 Equipotential Surfaces\n\nWith a positive source charge, the field lines are pointing outward, which is indeed pointing from higher potential to lower potential, but there is something more specific that we can conclude about the geometric relationship of the field and potential. When we follow a path that remains on an equipotential, the potential never changes, so if we traverse such a path from position $$A$$ to position $$B$$, we find:\n\n$V_A-V_B=0=\\int\\limits_A^B \\overrightarrow E\\cdot \\overrightarrow{dl}$\n\nCertainly the electric field is not zero everywhere we go, and the distance we travel isn't zero, so how can this integral come out to be zero? Maybe parts of it cancel other parts? No, because it happens on every single path we take, between any two points, so long as that path stays on an equipotential. The answer is that the only way this integral can be zero is if at every point on the equipotential, the electric field is perpendicular to $$\\overrightarrow {dl}$$. In other words, the electric field is perpendicular to equipotentials everywhere. This completes the geometric picture the field points from equipotential surfaces of higher potential toward those with lower potentials, and it does so perpendicular to the surfaces.\n\nThat electric fields are perpendicular to equipotential surfaces sounds very familiar. We said the same thing about conducting surfaces for electrostatics. Indeed, we immediately conclude that for electrostatics, conductors are equipotentials.","date":"2019-08-21 03:41:13","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 1, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.8747833967208862, \"perplexity\": 218.2203394798069}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2019-35\/segments\/1566027315750.62\/warc\/CC-MAIN-20190821022901-20190821044901-00019.warc.gz\"}"}
| null | null |
Sorry, the Oklahoma State Cowboys Riddell Speed Mini Football Helmet is not currently available.
Support your Oklahoma State Cowboys with this Riddell Mini Speed Replica Football Helmet. The most popular collectible in Riddell's history, your new Oklahoma State helmet is great for displaying in your home or office and makes the perfect gift for the true sports fan. Start collecting Oklahoma State Cowboys football helmets today!
|
{
"redpajama_set_name": "RedPajamaC4"
}
| 7,993
|
/*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package com.facebook.presto.orc;
import com.facebook.presto.orc.metadata.CompressionKind;
import io.airlift.slice.DynamicSliceOutput;
import org.testng.annotations.Test;
import java.util.Arrays;
import static io.airlift.slice.Slices.wrappedBuffer;
import static org.testng.Assert.assertEquals;
public class TestOrcOutputBuffer
{
@Test
public void testWriteHugeByteChucks()
{
int size = 1024 * 1024;
byte[] largeByteArray = new byte[size];
Arrays.fill(largeByteArray, (byte) 0xA);
OrcOutputBuffer sliceOutput = new OrcOutputBuffer(CompressionKind.NONE, 256 * 1024);
DynamicSliceOutput output = new DynamicSliceOutput(size);
sliceOutput.writeBytes(largeByteArray, 10, size - 10);
assertEquals(sliceOutput.writeDataTo(output), size - 10);
assertEquals(output.slice(), wrappedBuffer(largeByteArray, 10, size - 10));
sliceOutput.reset();
output.reset();
sliceOutput.writeBytes(wrappedBuffer(largeByteArray), 100, size - 100);
assertEquals(sliceOutput.writeDataTo(output), size - 100);
assertEquals(output.slice(), wrappedBuffer(largeByteArray, 100, size - 100));
}
@Test
public void testGrowCapacity()
{
byte[] largeByteArray = new byte[4096];
OrcOutputBuffer sliceOutput = new OrcOutputBuffer(CompressionKind.NONE, 3000);
// write some data that can fit the initial capacity = 256
sliceOutput.writeBytes(largeByteArray, 0, 200);
assertEquals(sliceOutput.getBufferCapacity(), 256);
// write some more data to exceed the capacity = 256; the capacity will double
sliceOutput.writeBytes(largeByteArray, 0, 200);
assertEquals(sliceOutput.getBufferCapacity(), 512);
// write a lot more data to exceed twice the capacity = 512 X 2; the capacity will be the required data size
sliceOutput.writeBytes(largeByteArray, 0, 1200);
assertEquals(sliceOutput.getBufferCapacity(), 1200);
// write some more data to double the capacity again
sliceOutput.writeBytes(largeByteArray, 0, 2000);
assertEquals(sliceOutput.getBufferCapacity(), 2400);
// make the buffer to reach the max buffer capacity
sliceOutput.writeBytes(largeByteArray, 0, 2500);
assertEquals(sliceOutput.getBufferCapacity(), 3000);
// make sure we didn't miss anything
DynamicSliceOutput output = new DynamicSliceOutput(6000);
sliceOutput.close();
assertEquals(sliceOutput.writeDataTo(output), 200 + 200 + 1200 + 2000 + 2500);
}
}
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 3,859
|
Revell Horsey
Senior Vice President and Managing Director
MEGA International
Revell Horsey is senior vice president and managing director of MEGA International, a global software firm that helps companies manage enterprise complexity by giving them an interactive view of their operations. He manages all of MEGA's operations in the US and Canada, including strategy and business development, sales, marketing, customer support and consulting.
Prior to assuming his current role at MEGA, Mr. Horsey was acting CFO and director of Aquicore, a Washington, DC–based IoT startup that combines web-enabled metering technology with an SaaS data analytics platform to provide real-time commercial energy management intelligence to owners and operators of commercial real estate.
Previously, Mr. Horsey was the founding president and CFO of HelloWallet, an SaaS provider of personalized financial guidance whose mission is to democratize access to financial management and guidance for all Americans. The company was acquired by Morningstar in June 2014.
Mr. Horsey has nearly 30 years working with founders and their management teams to build sustainable high-growth companies. As a corporate finance specialist on Wall Street, Mr. Horsey was directly responsible for more than $40 billion of equity capital financing by companies across a broad range of industries. His involvement with many of these companies began when they were private and looking for advice on valuation and positioning for their introduction to the public markets. Mr. Horsey combined his capital markets expertise with experience as a senior sell-side equity research analyst, where he was responsible for developing and communicating investment themes for emerging growth companies in the hospitality industry.
During his investment banking career, Mr. Horsey led the equity Capital Markets division at Montgomery Securities and Bank of America Securities during the internet boom of the late 1990s. He subsequently was appointed Global Head of Technology Corporate and Investment Banking at Bank of America, where he led a team of 150 professionals responsible for the firm's investment banking and credit relationships in the technology sector.
Since leaving Wall Street, Mr. Horsey has shifted his focus from being a service provider to corporate executives to teaming up with founders to build innovative companies that provide disruptive solutions to conventional challenges. In most cases, a common theme that characterizes the companies where he spends his time is the joint pursuit of social purpose and profit.
Mr. Horsey holds a B.A. from Dartmouth College.
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 6,111
|
Home » THE BIG GAME
Published on Fri, 01/07/2022 - 12:00am
Bobcats face down Bison in FCS national championship
Jason Bacaj — Enterprise Staff Writer
Montana State quarterback Tommy Mellott runs against South Dakota State University in the Dec. 18 semifinal game. Enterprise photo by Jason Bacaj
Montana State University is on the doorstep of its first Football Championship Subdivision championship since 1984, but standing in the Bobcats way is a formidable opponent: the most successful college football program of the past decade.
The North Dakota State University Bison have a record of 127-8 since 2011, including a 22-game playoff winning streak and eight national titles. The program has produced a number of NFL players, notably Carson Wentz, quarterback for the Indianapolis Colts, and Trey Lance, quarterback for the San Francisco 49ers.
But the Bobcats aren't daunted by the Bison's past success. After all, their head coach, Brent Vigen, was the offensive coordinator for NDSU at the start of that successful run. Vigen has kept the 'Cats focused throughout the season, in which they've overcome a number of obstacles, namely losing the starting quarterback right before the playoffs and injuries to three key contributors throughout the playoffs.
The championship game is scheduled for 10 a.m. MST and will be shown on ESPN2.
Matchups to watch
• First team All-American MSU senior linebacker Troy Andersen and first team All-American NDSU senior fullback Hunter Luepke. The 6-foot-4 235-pound Andersen is probably the best defensive player in the FCS. The Dillon product's early fourth down tackle of Isaiah Davis, South Dakota State's bowling ball of a running back, in the red zone set the tone for the semifinal game of the playoffs.
Andersen has his work cut out for him with corralling the multi-faceted Luepke, who scored all three of NDSU's touchdowns — a 34-yard reception, 4-yard run and a 22-yard catch — in the semifinal game against James Madison University. Luepke's high school resume is about as eclectic as Andersen's college one. In addition to exploits on the football field, the Bison back was a two-time state wrestling champ, three-time all-conference center fielder and a 100-meter dash conference champion.
• MSU freshman quarterback Tommy Mellott and first team All-American NDSU senior strong safety Michael Tutsie. Mellott has been a revelation during the playoffs after being thrust into a starting role fol lowing the abrupt departure of season-long starter Matt McKay. Mellott has rolled up 411 rushing yards, 449 passing yards and 11 touchdowns in three playoff games and seems to improve with each game.
Tutsie has started for 26 consecutive games and this season ranks second on the team in tackles, with 79 takedowns from his safety position. He nabbed an interception in the Bison's quarterfinal game. Tutsie will be a key cog in the defense working to limit Mellott's big-play ability both on the ground and in the air.
• First team All-American MSU senior defensive end Daniel Hardy and three-time All-American NDSU right tackle Cordell Volson. Hardy has been a monster on the edge this season, his first on the defensive line. The senior was a basketball star — Hardy stands 6-foot-3 with a 6-foot-10 wingspan — and has only played football since his senior year in high school. Regardless, he has racked up 16 sacks and 23 tackles for loss so far this season.
The 6-foot-7 Volson is a fair matchup for Hardy. The highly decorated tackle started since 2019, the year he led Bison linemen with 100 knockdown blocks in 12 regular season games, according to NDSU. Last season, he spent time at every position on the o-line except for center, allowed zero sacks and was named the Phil Steele FCS Offensive Lineman of the Year.
Both the Bison and the Bobcats win in similar fashion: powerful offensive and defensive lines paired with prolific rushing attacks. What seems safe to say is that the game won't be a shootout, seeing as how the two schools rank first and second in the FCS in scoring defense, with NDSU allowing 11.21 points per game and Montana State allowing 13.43 points per game.
It'll come down to which team's game-breaking talents perform the best, and I believe Montana State has caught lightning in a bottle this postseason.
MSU wins 27-21.
Montana State Bobcats vs. North Dakota State Bison
Saturday, Jan. 8
10 a.m. Mountain Time on ESPN2
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 2,903
|
Home > Georgia > Cobb County
Chattahoochee Technical College
Top Rankings Overview Student Body Tuition and Acceptance Rate School Notes Frequently Asked Questions Nearby Schools Edit School Profile
980 S Cobb Drive Dr SE
Top Ranked (2023): View Rankings
www.chattahoocheetech.edu
Chattahoochee Technical College, a unit of the Technical College System of Georgia, is a public, multi-campus, two-year college that provides accessible, high quality technical, academic, and adult education and training that promotes the economic growth and development of North Central Georgia. The college successfully prepares individuals for employment and assists them in pursuing their educational, career and personal goals through a variety of learning opportunities that include associate degree, diploma, and certificate programs, as well as non-credit and public service offerings.
Chattahoochee Technical College serves 9,432 students (27% of students are full-time).
The college's student:teacher ratio of 16:1 is higher than the state community college average of 14:1.
Minority enrollment is 44% of the student body (majority Black), which is less than the state average of 53%.
Enrollment: 9,432 students
Source: Integrated Postsecondary Education Data System (IPEDS)
Chattahoochee Technical College ranks among the top 20% of public schools in Georgia for:
Largest student body
The teacher population of 602 teachers has grown by 133% over five years.
(GA) Community College Avg.
At least 2 years but < 4 years
The student population of Chattahoochee Technical College has declined by 7% over five years.
The Chattahoochee Technical College diversity score of 0.62 is less than the state average of 0.67. The school's diversity has stayed relatively flat over five years.
Average Graduate Earnings (10 Years)
The public in-state tuition of $2,682 is less than the state average of $3,488. The in-state tuition has stayed relatively flat over four years.
Source: 2021 (or latest year available) Integrated Postsecondary Education Data System (IPEDS) , School Administrators
Appalachian Technical College, Chattahoochee Technical College and North Metro Technical College have merged together to form the largest technical college in the state with a combined student population of 11,000. Our seven campus locations (Acworth, Austell, Dallas, Jasper, Marietta-W. Cobb & E. Cobb Counties, and Woodstock) will provide the workforce needed to stimulate Georgia's community and economic growth and development. Over 100 associate degree, diploma and certificate programs offered in the areas of: Business Sciences, Computer Sciences & Engineering Technology, Health Programs, Personal & Public Services, and Technical Programs.Chattahoochee Technical College is accredited by the Commission on Colleges of the Southern Association of Colleges and Schools to award associate degrees. Equal Opportunity Institution. A Unit of the Technical College System of Georgia.
How much does Chattahoochee Technical College cost?
Chattahoochee Technical College's tuition is approximately $2,682 for In-State students and $4,818 for Out-State students.
What is Chattahoochee Technical College's ranking?
Chattahoochee Technical College ranks among the top 20% of community college in Georgia for: Largest student body.
The nearest community college to Chattahoochee Technical College is Lincoln College of Technology-Marietta (4.2 miles away)
Lincoln College of Technology-Marietta
Private for-profit
2359 Windy Hill Road
Fortis College-Smyrna
2140 South Cobb Drive
The Art Institute of Atlanta
6600 Peachtree Dunwoody Rd., N.E., 100 Embassy Row
Sanford-Brown College-Atlanta
Private, for profit
1140 Hammond Dr Ste A-1150
North Metro Technical College
5198 Ross Rd
The Final Girl (A Chattahoochee Technical College Film)
Chattahoochee Technical College: Television & Film Promo
International Festival 2018 - Chattahoochee Technical College
Madison from Chattahoochee Technical College
Chattahoochee Tech welcomes three community leaders to Board of Directors, honors outgoing Board Chair (2022)
Sep. 21—Chattahoochee Technical College has welcomed community leaders Pam Carnes, Roberto ...
Chattahoochee Tech introducing 3 new programs of study (2022)
Three new programs of study are in the works for Chattahoochee Technical College, including Brewing and Fermentation Production Technology, Paralegal Studies and ...
Chattahoochee Tech introduces plans for new Center for Advanced Manufacturing (2021)
Chattahoochee Technical College recently hosted local industry leaders for a tour of the ...
For Georgia public community colleges, the average tuition is approximately $3,488 per year for in-state students and $7,256 for out-of-state students. For private community colleges, the average yearly tuition is approximately $15,183 per year. Read more about average community college tuition costs across the country.
The average community college acceptance rate in Georgia is 80%. Read more about national community college acceptance rates.
Georgia community colleges have a diversity score of 0.67, which is higher than the national average of 0.00. The most diverse community college in Georgia is Sanford-Brown College-Atlanta. Read more about community college diversity statistics in the USA.
8 Reasons Why Community College Might be the Best Choice After High S...
Careers: Winemaking
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 5,635
|
Q: Onchange ,Arrow key up,down works in FF & Chrome but not work in IE It is my jquery code
<script type="text/javascript" charset="utf-8">
$(document).ready(function() {
$("#tasks").hide();
$("select#categories").change(function(){
$("#tasks").show();
$.getJSON("ajax.php?module=responsibles&action=list_tasks",{id: $(this).val()}, function(j){
if($("#categories").val()=="-0")
{
$("#tasks").hide();
}
//if the div is hidden
$("div#form_objectifs").hide();
var options = '';
for (var i = 0; i < j.length; i++)
{
options += '<option value="' + j[i].optionValue + '">' + j[i].optionDisplay + '</option>';
}
$("#tasks").html(options);
$('#tasks option:first').attr('selected', 'selected');
})
})
});
</script>
This code works fine on FF or Chrome,But not work in IE for Arrow Key UP,DOWN
Anybody
A: First problem :
there is no keyup or keydown event in your script. You use change event which be fired when your element loose focus.
Second problem :
You use $(document).ready(function() { ... }); and in this you use $(function(){ ... });. It's two different way to write the same thing (the second is an alias of the first).
Choose one of them but not the both.
Third problem :
Please, best explain your problem so we can help you ;).
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 8,187
|
Q: radio buttons unable to select in UI I am trying to select radio button but somehow they not showing up in UI side. I have verified that buttons can be pressed as it calls a function on click but UI not showing any selection.
<div data-ng-repeat="flagInfo in availableList">
<input type="radio" name="flag" id="{{flagInfo.flagName}}" value="{{flagInfo.flagName}}" data-ng-click="selection(flagInfo.flagName)" />
<span class="checkbox-label">
<label for="{{flagInfo.flagName}}">{{flagInfo.translation}}</label>
</span>
</div>
Can someone tell me how to resolve it?
A: Created snipet for you. you have missed ng-model in your code.
function myCtrl($scope) {
$scope.radioList={
'flagName':'abc'
}
$scope.availableList = [{
'flagName':'abc',
'id':1
},{
'flagName':'def',
'id':2
},{
'flagName':'ghi',
'id':3
},{
'flagName':'jkl',
'id':4
}];
$scope.radioChange=function(radval){
console.log(radval);
}
}
<script src="https://ajax.googleapis.com/ajax/libs/angularjs/1.2.23/angular.min.js"></script>
<div ng-app ng-controller="myCtrl">
<div data-ng-repeat="flagInfo in availableList">
<input type="radio" name="flag" id="{{flagInfo.flagName}}" value="{{flagInfo.flagName}}" ng-model="radioList.flagName" ng-change="radioChange(radioList.flagName)" />
<span class="checkbox-label">
<label for="{{flagInfo.flagName}}">{{flagInfo.flagName}}</label>
</span>
</div>
<div>
Selected Radio : {{radioList.flagName}}
</div>
</div>
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 2,886
|
Jakarta EE Ambassadors
Sailing Forward with Cloud Native Java
Guide to Contributing to Jakarta EE 11
Welcome to the Jakarta EE Ambassadors
We are an independent grassroots group of people committed to moving Jakarta EE forward through active community participation and advocacy. We believe that standardizing the core technologies used to build cloud native applications is better than being bound to proprietary solutions. Java EE, and it's successor Jakarta EE, have been the most successful set of open standards for developing mission critical business applications ever – on premise and on the cloud. It has been the linchpin of the Java platform's domination and it is rare to find a Java application that doesn't depend upon at least one Jakarta EE technology. It is the only business application development technology with a rich ecosystem of vendors who not only compete but cooperate to define a solid set of reliable open standard technologies. Closed ecosystems stymie market competition, limit consumer choices, cripple careers, and harm end-users. Even applications, projects and frameworks that do not explicitly acknowledge it are in fact heavily dependent on many Jakarta EE APIs today and going forward, regardless of sea changes like cloud or microservices. Just some of these APIs include Jakarta Servlet, Jakarta REST, Jakarta WebSocket, Jakarta Messaging, Jakarta Persistence, Jakarta Bean Validation, Jakarta Mail and so much more.
Few multi-vendor open standards are as widely implemented, supported, depended upon or as widely participated in as Jakarta EE. There is an extremely passionate, responsible community behind Jakarta EE – most technologies would be hard pressed to find anything like the Jakarta EE community.
As critically important as Jakarta EE is and as much as it has going for it, there are challenges to overcome. Some of these factors are outlined in the history and challenges sections below. Continuing to successfully overcome these challenges is the primary goal of our grassroots community. Advocacy, raising awareness, coordination, collaboration and mutual support in the community are keys to accomplishing these goals. We encourage you to take a close look at the continuing work we do. We already have a number of individuals, organizations and user groups committed to working together to advance Jakarta EE. We need all the help we can get to ensure the best interests of the Jakarta EE community continues to be well served. We need the support of your voice and perhaps your volunteer time if you can afford to contribute it. There are many ways you could help.
Our role as an evangelism and advocacy body requires a civil, productive, professional, reasonable and respectful environment. To that end, the Jakarta EE Ambassadors adopt the open source Contributor Covenant as our code of conduct.
We believe that together we can create an ever brighter future for Java, Jakarta EE and cloud native computing.
Join the Ambassadors
The Jakarta EE Ambassadors were originally founded as the Java EE Guardians. In early 2016 Oracle appeared to lose focus on Java EE 8 progress. A community effort, which became the Java EE Guardians, was then made to document the lack of progress on the various specifications. This website was constructed. On the website, the objective findings of lack of progress were published with the goal of organizing the larger community and also encouraging Oracle to get the specifications moving forward again. In addition Java EE Guardians stepped up to move inactive specifications along.
At JavaOne 2016 Oracle announced the reboot of Java EE 8. At JavaOne 2017, Oracle released Java EE 8 and announced opening up Java EE by giving ownership of the Java EE code base to the Eclipse Foundation. Java EE has now been completely transferred to a truly open governance model as Jakarta EE. Jakarta EE represents one of the largest and most successful technology ownership transfers ever attempted. Oracle and the Eclipse Foundation, amongst many other parties, deserve due credit for the historic creation of Jakarta EE.
With the successful transition and opening up of Java EE as Jakarta EE, our group has shifted its attention to educating the broader Java community about Jakarta EE and encouraging active participation in the new open process to move Jakarta EE forward. The group was transitioned from the Java EE Guardians to the Jakarta EE Ambassadors to reflect this change in focus. We are grateful to the Eclipse Foundation and Jakarta EE stakeholders for recognizing our group and granting us the necessary tools to continue to operate as an independent grassroots community.
One of the key challenges faced by our community is that many Java developers don't realize that they use, benefit from and depend on Jakarta EE every day. For example, all of these are technologies related to Jakarta EE in one way or another – WebLogic, WebSphere, JBoss EAP, WildFly, GlassFish, Payara, Tomcat, TomEE, Jetty, CXF, ActiveMQ, Jersey, Hibernate, PrimeFaces, Spring, Spring Boot, MicroProfile, Quarkus, Helidon.
Much like Java itself, Jakarta EE also has all the challenges of a relatively mature technology. There are many badly outdated or plain wrongheaded ideas held by many people about Jakarta EE. The core value proposition of Jakarta EE is even more important in cloud native applications, not less. Jakarta EE also competes with many highly aggressive non-standard solutions even within the Java ecosystem (ironically many of these solutions actually depend on Jakarta EE remaining strong) let alone beyond the JVM such as .NET.
The scope and complexity of Jakarta EE means that in order for it to evolve effectively, it requires the continued interest, collaboration and investment of many major players such as Oracle, IBM, Red Hat, Fujitsu, Payara and Tomitribe. As well intentioned as these parties might be, without active engagement, participation and contribution from the broader community, decisions can become vendor focused and noncompetitive instead of remaining strongly focused on the needs of end users and industry.
The most effective way of addressing these challenges and moving Jakarta EE forward is grassroots community advocacy and participation. None of this is possible without your direct engagement and help.
Join Our Google Group
Jakarta EE Ambassadors, Website Powered by WordPress.com.
|
{
"redpajama_set_name": "RedPajamaCommonCrawl"
}
| 8,789
|
\section{Introduction}
Marcus and Purves \cite{MaP} proved that every unital invertibility preserving
linear map on a matrix algebra is either an inner automorphism or an inner
anti-automorphism. One of the equivalent formulations of Gleason-Kahane-\.
Zelazko theorem \cite{Gle,KaZ,Zel} states that every unital linear functional
on a complex unital Banach algebra ${\cal A}$ sending every invertible element
into a nonzero scalar is multiplicative. Equivalently, if a linear functional
$f:{\cal A} \to \Com$ maps every element $a\in {\cal A}$ into its spectrum
$\sigma (a)$, then $f$ is multiplicative. This two results motivated Kaplansky
to formulate the question under which conditions an invertibility preserving
linear unital map between two algebras must be a Jordan homomorphism
\cite{Kap}. A lot of work has been done on this problem (see the surveys
\cite{Aup,BrS,Sou1}). We will mention here only the results that are relevant
for our paper. Let $X$ be a complex Banach space and $B(X)$ the algebra of all
bounded linear operators on $X$. In 1986 Jafarian and Sourour \cite{JaS} proved
that every surjective unital linear map $\phi : B(X) \to B(X)$ preserving
invertibility in both directions, i.e., having the property that $A$ is
invertible if and only if $\phi (A)$ is invertible, is either of the form $\phi
(A) = TAT^{-1}$, $A\in B(X)$, for some invertible $T\in B(X)$, or of the form
$\phi (A) = TA'T^{-1}$, $A\in B(X)$, for some invertible bounded linear
operator $T: X' \to X$. Here, $A'$ denotes the adjoint of $A$ and $X'$ the dual
of $X$. Under the additional assumption of injectivity the assumption of
preserving invertibility in both directions can be relaxed to the weaker
assumption of preserving invertibility in one direction only \cite{Sou}. The
proof of the result of Jafarian and Sourour was simplified in \cite{Sem}. It is
rather easy to see that a linear map $\phi : B(X) \to B(X)$ is unital and
preserves invertibility in both directions if and only if $\phi$ preserves the
spectrum, that is, $\sigma (\phi (A)) = \sigma (A)$ for every $A\in B(X)$.
An interesting extension of Gleason-Kahane-\. Zelazko theorem was obtained by
Kowalski and S\l odkowski \cite{KoS}. They proved that every functional $f$ on
a complex Banach algebra ${\cal A}$ (they did not assume the linearity of $f$)
satisfying $f(a) - f(b) \in \sigma (a-b)$, $a,b\in {\cal A}$, is linear and
multiplicative up to the constant $f(0)$. Thus, they replaced the two
conditions in Gleason-Kahane-\. Zelazko theorem, the linearity assumption and
the condition $f(a)\in \sigma (a)$, $a\in {\cal A}$, by a single weaker
assumption and got essentially the same conclusion.
In view of this result it is natural to ask if we can do the same with the
above mentioned results on invertibility preserving maps on matrix and operator
algebras. Can we replace the linearity assumption and the invertibility
preserving assumption by a single weaker condition similar to the one in
Kowalski-S\l odkowski theorem? More precisely, can we characterize bijective
maps on matrix algebras and operator algebras satisfying the condition that
$\phi(a) - \phi (b)$ is invertible if and only if $a-b$ is invertible?
The result of Kowalski and S\l odkowski depends heavily on deep results from
analysis. We will answer the above question using the results from geometry. We
should first mention that there is an essential difference between the finite
and the infinite-dimensional case. In the finite-dimensional case our condition
will imply up to a translation the semilinearity of the maps under
consideration, while in the infinite-dimensional case the elementary automatic
continuity methods will imply the linearity or conjugate-linearity up to a
translation. Moreover, in the finite-dimensional case it makes sense to extend
our result from matrix algebras of square matrices to the spaces of rectangular
matrices. Then, of course, the condition of invertibility will be replaced by
the condition of being of full rank.
Our strategy when considering bijective maps $\phi$ on matrix spaces (operator
algebras) satisfying the condition that $\phi(A) - \phi (B)$ is of full rank
(invertible) if and only if $A-B$ is of full rank (invertible) will be to prove
first that such maps preserve adjacency in both directions. Recall that two
matrices or operators $A$ and $B$ are adjacent if $A-B$ is of rank one. Then we
will apply the so called fundamental theorem of geometry of matrices (or its
analogue for operators) to complete the proof. This connects our results with
the geometry of Grassmann spaces. Let us briefly describe this connection.
Let $M_{m,n}$, $m,n\ge 2$, be the linear space of all $m\times n$ matrices over
a field $\FF$. If $\sigma$ is an automorphism of the field $\FF$ and $A=
[a_{ij}]\in M_{m,n}$ then we denote by $A_\sigma$ the matrix obtained from $A$
by applying $\sigma$ entrywise, $A_\sigma = [\sigma (a_{ij})]$. The fundamental
theorem of geometry of matrices states that every bijective map $\phi :
M_{m,n} \to M_{m, n}$ preserving adjacency in both directions is of the form
$A\mapsto TA_\sigma S + R$, where $T$ is an invertible $m\times m$ matrix, $S$
is an invertible $n\times n$ matrix, $R$ is an $m\times n$ matrix, and $\sigma$
is an automorphism of the underlying field. If $m=n$, then we have the
additional possibility that $\phi (A) = TA_{\sigma}^t S + R$ where $T,S,R$ and
$\sigma$ are as above, and $A^t$ denotes the transpose of $A$. This theorem and
its analogues for hermitian matrices, symmetric matrices, and skew-symmetric
matrices were proved by Hua \cite{Hu1}-\cite{Hu8} under some mild technical
assumptions that were later proved to be superfluous (see \cite{Wan}). Let
$m,n$ be integers $\ge 2$. We will consider the Grassmann space whose ``points"
are vector subspaces of $\FF^{m+n}$ of dimension $m$. Chow \cite{Cho} studied
bijective maps on the Grassmann space preserving adjacent pairs of points in
both directions. Recall that $m$-dimensional subspaces $U$ and $V$ are adjacent
if $\dim (U + V) = m+1$. Now, to each $m$-dimensional subspace $U$ of
$\FF^{m+n}$ we can associate an $m\times (m+n)$ matrix whose rows are
coordinates of vectors that form a basis of $U$. Each $m\times (m+n)$ matrix
will be written in the block form $[ X \ Y]$, where $X$ is an $m\times n$
matrix and $Y$ is an $m\times m$ matrix. Two matrices $[ X \ Y]$ and $[ X' \
Y']$ are associated to the same subspace $U$ (their rows represent two bases of
$U$) if and only if $[ X \ Y] = P [ X' \ Y']$ for some invertible $m\times m$
matrix $P$. If this is the case, then $Y$ is invertible if and only if $Y'$ is
invertible. So, we have associated to each point in a Grassmann space a (not
uniquely determined) matrix $[ X \ Y]$. If $Y$ is singular, we say that the
corresponding point in the Grassmann space is at infinity. Otherwise, we
observe that this point can be represented also with the matrix $[Y^{-1}X \ \
\, I]$. The matrix $Y^{-1}X$ is uniquely determined by the point in the
Grassmann space. So, if $U$ and $V$ are two $m$-dimensional subspaces that are
finite points in the Grassmann space, then they can be represented with two
uniquely determined $m\times n$ matrices $T$ and $S$, and it is easy to see
that the subspaces $U$ and $V$ are adjacent if and only if the matrices $T$ and
$S$ are adjacent. Using this connection it is possible to deduce the result of
Chow on bijective maps on a Grassmann space preserving adjacency in both
directions from the fundamental theorem of geometry of matrices (see
\cite{Wan}).
If we consider the special case when $m=n$ and replace in the fundamental
theorem of geometry of matrices the condition of preserving adjacent pairs of
matrices by our assumption of preserving the pairs $A,B$ with the property that
${\rm rank}\, (A-B) =n$, then this corresponds to the study of bijective maps
on the Grassmann space of all vector subspaces of $\FF^{2n}$ of dimension $n$
that preserve the complementarity of subspaces. Such maps were studied by
Blunck and the first author \cite{BlH}. We suspect that this result can be
deduced from our result and the other way around, but we also believe that it
is easier to prove each of them separately. Namely, to prove any of these two
implications seems to be difficult because of the points at infinity.
Now we state our main results. In the finite-dimensional case we will consider
bijective maps on $m\times n$ matrices preserving pairs of matrices whose
difference has a full rank. Of course, if we have such a map $\phi$ then the
map $\psi: M_{n,m} \to M_{n,m}$ defined by $\psi (A) = (\phi (A^t))^t$ has the
same properties. Thus, when studying such maps there is no loss of generality
in assuming that $m\ge n$. We will do this throughout the paper. A matrix $A\in
M_{m,n}$ is said to be of full rank if ${\rm rank}\, A = n$. Let $A,B \in
M_{m,n}$. We write $A\dis B$ if $A-B$ is of full rank.
\begin{theorem}\label{fin}
Let $\FF$ be a field with at least three elements and $m,n$ integers with $m\ge
n \ge 2$. Assume that $\phi : M_{m,n} \to M_{m,n}$ is a bijective map such that
for every pair $A,B \in M_{m,n}$ we have $A\dis B$ if and only if $\phi (A)
\dis \phi (B)$. Then there exist an invertible $m\times m$ matrix $T$, an
invertible $n\times n$ matrix $S$, an $m\times n$ matrix $R$, and an
automorphism $\sigma : \FF \to \FF$ such that
$$
\phi (A) = TA_\sigma S + R
$$
for every $A\in M_{m,n}$. If $m=n$, then we have the additional possibility
that $$\phi (A) = TA_{\sigma}^t S + R, \ \ \ A\in M_{n,n},$$ where $T,S,R \in
M_{n,n}$ with $T$ and $S$ invertible, and $\sigma$ is an automorphism of $\FF$.
\end{theorem}
\begin{theorem}\label{infin}
Let $H$ be an infinite-dimensional complex Hilbert space and $B(H)$ the algebra
of all bounded linear operators on $H$. Assume that $\phi : B(H) \to B(H)$ is a
bijective map such that for every pair $A,B \in B(H)$ the operator $A-B$ is
invertible if and only if $\phi (A) - \phi (B)$ is invertible. Then there exist
$R\in B(H)$ and invertible $T,S \in B(H)$ such that either
$$
\phi (A) = TAS + R
$$
for every $A\in B(H)$, or
$$
\phi (A) = TA^t S + R
$$
for every $A\in B(H)$, or
$$
\phi (A) = TA^* S + R
$$
for every $A\in B(H)$, or
$$
\phi (A) = T(A^t)^* S + R
$$
for every $A\in B(H)$. Here, $A^t$ and $A^*$ denote the transpose with respect
to an arbitrary but fixed orthonormal basis, and the usual adjoint of $A$ in
the Hilbert space sense, respectively.
\end{theorem}
The converses of both theorems obviously hold true. In the second section we
will prove the finite-dimensional case and in the third one the
infinite-dimensional case. These two sections can be read independently.
\section{The finite-dimensional case}
In this section we will consider matrices over a field $\FF$ with at least
three elements. At a certain point in the proof of our first main theorem we
will identify $m\times n$ matrices with linear operators from $\FF^n$ into
$\FF^m$. For such operators we have the following simple lemma.
\begin{lemma}\label{pez}
Let $T,S : \FF^n \to \FF^m$ be nonzero linear operators and assume that $T$ has
at least two-dimensional image. Then we can find linearly independent vectors
$x,y \in \FF^n$ such that $Tx$ and $Sy$ are linearly independent.
\end{lemma}
{\bf Proof.} Take any $y\in \FF^n$ such that $Sy\not=0$. The set of all vectors
$z\in\FF^n$ with the property that $Tz$ is linearly dependent of $Sy$ is a
proper subspace of $\FF^n$, since the image of $T$ is not contained in the span
of $Sy$. There exist at least two linearly independent vectors of $\FF^n$ which
are not in this particular subspace. One of them is linearly independent of $y$
and gives the required vector $x$. \enp
We have two relations on $M_{m,n}$, that is, the relation of adjacency and
$\dis $. The following result connecting these two relations is the key step in
our proof. We believe it is of some independent interest.
\begin{proposition}\label{wed}
Let $A,B \in M_{m,n}$ be matrices with $A\not=B$. Then the following are
equivalent:
\begin{enumerate}
\item $A$ and $B$ are adjacent.
\item There exists $R\in M_{m,n}$, $R\not=A,B$, such that for every $X\in
M_{m,n}$ the relation $X \dis R$ yields $X\dis A$ or $X\dis B$.
\end{enumerate}
\end{proposition}
{\bf Proof.} Note that none of the above conditions are effected if we replace
$A$ and $B$ by $PAQ-C$ and $PBQ-C$, respectively, where $P$ and $Q$ are
invertible matrices of the appropriate size and $C$ is any $m \times n$ matrix.
Thus if the rank distance between $A$ and $B$ equals $r$ then we may assume
with no loss of generality that $A=0$ and
$$B=\left(\begin{array}{cc}I & 0\\ 0 &0 \end{array} \right)$$
where $I$ is the $r\times r$ identity matrix and the zeros stand for the zero
matrices of the appropriate size.
Assume first that $A$ and $B$ are adjacent. So, without loss of generality, we
have $A=0$ and $B=E_{11}$. Set $R=\lambda E_{11}$, where $\lambda$ is a scalar
different from $0$ and $1$, and $E_{11}$ denotes the matrix with the
$(1,1)$-entry equal to $1$ and all other entries equal to zero. Now let $X\dis
R$. That means that $X-R$ is of full rank or equivalently, the matrix $X-R$
contains at least one invertible $n\times n$ submatrix. We have to consider two
possibilities. Let first assume that one of these submatrices does not contain
the first row. In this case $X$ is of full rank and thus $X\dis A$. Otherwise
any such submatrix contains the first row and we choose one of them. We will
prove that at least one of the corresponding $n\times n$ submatrices of $X-A=X$
and $X-B$ is invertible. So we restrict our attention to these $n\times n$
submatrices. In other words we deal only with the square case $m=n$. Hence
$X-\lambda E_{11}$ is an invertible square matrix. If the first row of
$E_{11}$, i.e.\ $(1,0,\ldots,0)$, is in the subspace spanned by rows
$2,3,\ldots, n$ of $X$ then $X-\lambda E_{11}-\mu E_{11}$ is invertible for all
$\mu\in\FF$, otherwise this holds for all but one $\mu\in\FF$. Therefore
$X-\lambda E_{11} - \mu E_{11}$ is invertible for at least one of the values
$\mu=-\lambda$ or $\mu=-\lambda+1$. Equivalently, at least one of $X=X-A$ or
$X-E_{11}=X-B$ is invertible, as desired. This completes the proof of the first
implication.
To prove the other direction we identify $m\times n$ matrices with linear
operators from $\FF^n$ into $\FF^m$. We assume that $A=0$ and $B: \FF^n \to
\FF^m$ is a linear operator whose image is at least two-dimensional. Let $R:
\FF^n \to \FF^m$ be any linear operator, $R\not=0,B$. We have to find a linear
operator $X: \FF^n \to \FF^m$ such that $X-R$ is injective while $X$ and $X-B$
are not.
The first possibility we will treat is that at least one of the operators $B-R$
and $R$ has rank at least two. Then, by Lemma \ref{pez}, we can find linearly
independent $x,y \in \FF^n$ such that $Bx-Rx$ and $Ry$ are linearly
independent. We first define $X$ on the linear span of $x$ and $y$. We set
$Xx=Bx$ and $Xy=0$. No matter how we will extend $X$ to the whole space these
two equations will guarantee that $X-B$ and $X$ will not be injective. Now,
$(X-R)x = Bx-Rx$ and $(X-R)y = -Ry$ are linearly independent. It is now obvious
that we can extend the linear operator $X$ to the whole space $\FF^n$ such that
the obtained $X-R$ is injective.
It remains to consider the case when both operators $B-R$ and $R$ are of rank
one. By our assumption, $B$ is of rank two. Hence $B= R + (B-R)$ implies that
the ranges of $B-R$ and $R$ meet at $0$ only. So, if we choose any $x,y \in
\FF^n$ such that $(B-R)x\not=0$ and $Ry\not=0$ then $(B-R)x$ and $Ry$ will be
linearly independent. Since $\FF$ has at least three elements, we can choose
these $x$ and $y$ to be linearly independent. Now we can proceed as above. \enp
It is now easy to prove Theorem \ref{fin}. Namely, if $\phi:M_{m,n}\to M_{m,n}$
is a bijective map preserving $\dis$ in both directions, then, by Proposition
\ref{wed} it preserves adjacency in both directions. Thus, the result follows
directly from the fundamental theorem of geometry of matrices.
Observe that in \cite{BlH} there is no need to assume that $\FF$ has at least
three elements, due to the presence of points at infinity.
\section{The infinite-dimensional case}
Let $H$ be an infinite-dimensional complex Hilbert space and $x,y \in H$. The
inner product of $x$ and $y$ will be denoted by $y^* x$. If $x$ and $y$ are
nonzero vectors then $xy^*$ stands for the rank one bounded linear operator
defined by $(xy^*)z = (y^*z)x$, $z\in H$. Note that every bounded rank one
operator can be written in this form. Two operators $A,B \in B(H)$ are said to
be adjacent if $A-B$ is an operator of rank one. We write $A\dis B$ if $A-B$ is
invertible. We start with an analogue of Proposition \ref{wed}.
\begin{proposition}\label{pic}
Let $A,B \in B(H)$ with $A\not= B$. Then the following statements are
equivalent:
\begin{enumerate}
\item $A$ and $B$ are adjacent. \item There exists $R\in B(H)$, $R\not=A,B$,
such that for every $X\in B(H)$ the relation $X \dis R$ yields $X\dis A$ or
$X\dis B$.
\end{enumerate}
\end{proposition}
{\bf Proof.} Note that none of the above conditions are effected if we replace
$A$ and $B$ by $A-C$ and $B-C$, respectively, where $C$ is any bounded linear
operator on $H$. Thus we may assume with no loss of generality that $A=0$.
Assume first that $A=0$ and $B$ are adjacent, that is, $B$ is of rank one. Set
$R=2B$. Suppose that $X-2B$ is invertible. Then
$$
X-2B - \lambda B = ( X- 2B)(I - \lambda (X-2B)^{-1}B)
$$
is invertible if and only if $I - \lambda S$ is invertible, where $S=
(X-2B)^{-1}B$ is an operator of rank one. Every operator of rank one has at
most one non-zero complex number in its spectrum. Hence, $X-2B - (-2B) = X$ is
invertible or $X-2B - (-B) = X-B$ is invertible. This completes the proof of
one direction.
Assume now that $A=0$ and $B$ is an operator whose image is at least
two-dimensional. We have to prove that for every $R\in B(H)$, $R\not= 0,B$,
there exists $X\in B(H)$ such that $X-R$ is invertible and $X$ is singular and
$X-B$ is singular. So, let $R\in B(H) \setminus \{ 0, B \}$.
In the next step we will prove that there exist $x, z\in H$ such that $x$ and
$z$ are linearly independent and $Bz - Rz$ and $Rx$ are linearly independent.
It is enough to show that we can find $x,z\in H$ such that $Bz - Rz$ and $Rx$
are linearly independent. For if $x$ and $z$ are linearly dependent, we can
choose $u\in H$ linearly independent of $x$. Then $z + \lambda u$ and $x$ are
linearly independent for all nonzero $\lambda$ and for all $\lambda$'s small
enough the vectors $B(z + \lambda u ) - R (z + \lambda u) = Bz - Rz + \lambda
(Bu-Ru)$ and $Rx$ are linearly independent as well.
So, let us show that such $x$ and $z$ exist. Assume on the contrary that
$Bz-Rz$ and $Rx$ are linearly dependent for every $x$ and $z$. Then $B-R$ and
$R$ are rank one operators with the same one-dimensional image. It follows that
$B=0$ or $B$ is of rank one, a contradiction.
Now, we define $W$ to be the orthogonal complement of the linear span of $x$
and $z$, where $x$ and $z$ are as in the previous paragraph, and $Z$ to be the
orthogonal complement of $Rx$ and $Bz - Rz$. Then there exists a bounded
invertible linear operator $U: W \to Z$. Define $X\in B(H)$ with
$$
Xx= 0,
$$
$$
Xz = Bz,
$$
and
$$
Xu = Uu + Ru, \ \ \ u\in W.
$$
Because of the first two equations the operators $X$ and $X-B$ are singular.
Since $(X-R) x = -Rx$, $(X-R)z = Bz-Rz$, and $(X-R)u = Uu$, $u\in W$, the
operator $X-R$ is invertible, as desired. \enp
We continue with some technical lemmas.
\begin{lemma}\label{noc}
Let $B,C \in B(H)$. Assume that for every invertible $A\in B(H)$ the operator
$A-B$ is invertible if and only if $A-C$ is invertible. Then $B=C$.
\end{lemma}
{\bf Proof.} Let $\lambda$ be any complex number satisfying
$$
| \lambda | > \| B \| , \| C \|,
$$
and $x,y\in H$ any vectors such that $y^* x =0$. Then $\lambda (I + xy^*)$ is
invertible because $(I + xy^* ) (I - xy^* ) = I$. Hence, $\lambda I + \lambda
xy^* - B$ is invertible if and only if $\lambda I + \lambda xy^* - C$ is
invertible. On the other hand,
$$
\lambda I + \lambda xy^* - B = (I+xy^* )( \lambda I - B +xy^* B) =
(I+xy^*)(I + xy^* B (\lambda I - B)^{-1}) (\lambda I -B)
$$
is invertible if and only if $I + xy^* B (\lambda I - B)^{-1}$ is invertible.
Thus, $I + xy^* B (\lambda I - B)^{-1}$ is invertible if and only if $I + xy^*
C (\lambda I - C)^{-1}$ is invertible, or equivalently, for every scalar
$\lambda$ with $| \lambda | > \| B \| , \| C \|$, and every pair of vectors
$x,y\in H$ with $y^* x = 0$ we have
$$
y^* B(\lambda I -B)^{-1}x = -1 \iff y^* C(\lambda I -C)^{-1}x = -1 .
$$
Fix $\lambda$. Then $y^* T x = 0$ for every pair of orthogonal vectors $x$ and
$y$, where $ T = B(\lambda I -B)^{-1} - C(\lambda I -C)^{-1}$. It follows that
$T = \mu I$ for some scalar $\mu$. Thus, for every $\lambda$ with $| \lambda |
> \| B \| , \| C \|$ we have
$$
B(\lambda I -B)^{-1} - C(\lambda I -C)^{-1} = g(\lambda )I
$$
for some $g(\lambda ) \in \Com$. Obviously, $g(\lambda)$ is holomorphic outside
the circle centered at $0$ with radius $\max \{ \| B \| , \| C \| \}$.
Expressing the above analytic functions with the series and comparing the
coefficients we get
$$
B= C+ \mu I
$$
for some complex number $\mu$. Our assumption implies that
$\sigma(B)\setminus\{0\} = \sigma(C)\setminus\{0\}$. Here $\sigma(B)$ denotes
the spectrum of $B$. It follows that $\mu=0$, as desired. \enp
\begin{lemma}\label{jut}
Let $A,B \in B(H)$ be invertible operators. Assume that for every rank one
operator $xy^* \in B(H)$ the operator $A-xy^*$ is invertible if and only if
$B-xy^*$ is invertible. Then $A=B$.
\end{lemma}
{\bf Proof.} Our assumptions yield that for every pair of vectors $x,y$ the
operator $I - xy^*A^{-1}$ is invertible if and only if $I- xy^*B^{-1}$ is, or
equivalently, $y^* A^{-1}x = 1$ if and only if $y^*B^{-1}x=1$. By linearity we
have $y^* A^{-1}x = y^*B^{-1}x$ for every pair $x,y \in H$, and therefore,
$A^{-1} = B^{-1}$. It follows that $A=B$. \enp
Let us recall that an additive map $T:H\to H$ is called semilinear if there is
an automorphism $\sigma :\Com\to \Com$ such that $T(\lambda x) =
\sigma(\lambda) Tx$ for every $\lambda\in\Com$ and every $x\in H$. Now we are
ready to prove Theorem \ref{infin}.
Let $\phi : B(H) \to B(H)$ be a bijective map such that for every pair $A,B \in
B(H)$ the operator $A-B$ is invertible if and only if $\phi (A) - \phi (B)$ is
invertible. After replacing $\phi$ by $A\mapsto \phi (A) - \phi (0)$ we may
assume that $\phi (0) = 0$. Then $\phi (I)$ is invertible. Replacing $\phi$ by
$A\mapsto \phi(I)^{-1}\phi(A)$ we may further assume that $\phi(I) =I$.
According to Proposition \ref{pic}, $\phi$ preserves adjacency in both
directions. Every rank one operator is adjacent to zero, every rank two
operator is adjacent to some rank one operator, etc. Consequently, $\phi$ maps
the subspace $F(H) \subset B(H)$ of all finite rank operators onto itself. So,
we can apply Theorem 1.5 from \cite{PeS} to conclude that there exist bijective
semilinear maps $T,S : H \to H$ (with the same accompanying automorphism) such
that either $\phi (xy^* ) = (Tx)(Sy)^*$, $x,y \in H$, or $\phi (xy^* ) =
(Sy)(Tx)^*$, $x,y \in H$. The second case can be reduced to the first one if we
replace $\phi$ by $A\mapsto \phi(A)^*$, $A\in B(H)$. So, we may assume that the
first possibility holds true.
Using $\phi (I) = I$ and our assumptions we conclude that $I -xy^*$ is
invertible if and only if $I - (Tx)(Sy)^*$ is invertible, $x,y\in H$. Thus,
$y^* x = 1$ if and only if $(Sy)^* (Tx) = 1$, and by semilinearity,
$$
(Sy)^* (Tx) = 0 \iff y^* x=0, \ \ \ x,y\in H.
$$
Thus, the semilinear maps $T$ and $S$ and their inverses carry closed
hyperplanes (every closed hyperplane is the orthogonal complement of some
nonzero vector) onto closed hyperplanes. Hence, by \cite[Lemmas 2 and 3]{FiL},
$S$ and $T$ are both linear bounded or both conjugate-linear bounded. Thus, we
have $\phi (xy^* ) = T(xy^*)R$, where $T$ and $R=S^*$ are bounded invertible
either both linear, or both conjugate-linear operators. Assume they are both
conjugate-linear. Choosing an orthonormal basis we define $K : H \to H$ to be
the conjugate-linear bijection which maps each vector $x$ into a vector whose
coordinates are obtained from the coordinates of $x$ by complex conjugation. Of
course, $K^2 = I$, the product of two conjugate-linear maps is linear, and
$K(xy^*)K = ((xy^*)^*)^t$, where the transpose is taken with the respect to the
chosen basis. Replacing $\phi$ by $A\mapsto (\phi (A)^t)^*$, $A\in B(H)$, we
reduce the conjugate-linear case to the linear one.
So, we may assume that we have $\phi (xy^* ) = T(xy^*)R$, where $T$ and $R=S^*$
are bounded invertible linear operators. From $(Sy)^* (Tx) =1 \iff y^* x=1$ and
linearity we get $(Sy)^* (Tx) =y^* x$, $x,y\in H$, which further yields that
$T$ is the inverse of $R$. Composing $\phi$ by a similarity transformation we
may further assume that $\phi (xy^* ) = xy^*$, $x,y \in H$.
Let $A\in B(H)$ be invertible.
Applying Lemma \ref{jut} with $B= \phi (A)$ we see that $\phi (A) = A$.
Finally, let $B\in B(H)$ be any operator and set $C=\phi (B)$. Using Lemma
\ref{noc} we conclude that $\phi(B) = B$. This completes the proof. \enp
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 376
|
using content::WebContents;
namespace {
// The pixel size of the header label when using a non-native theme.
const int kHeaderLabelPixelSize = 15;
// The pixel size of the font of the main content of the dialog.
const int kMainContentPixelSize = 13;
// Indices of the extension row widgets.
enum {
kIconIndex,
kTitleLinkIndex,
kStarsIndex,
kInstallButtonIndex,
};
GtkThemeService *GetThemeService(WebContents* web_contents) {
Profile* profile =
Profile::FromBrowserContext(web_contents->GetBrowserContext());
return GtkThemeService::GetFrom(profile);
}
// Set the image of |button| to |pixbuf|.
void SetServiceButtonImage(GtkWidget* button, GdkPixbuf* pixbuf) {
gtk_button_set_image(GTK_BUTTON(button), gtk_image_new_from_pixbuf(pixbuf));
gtk_button_set_image_position(GTK_BUTTON(button), GTK_POS_LEFT);
}
void SetWidgetFontSizeCallback(GtkWidget* widget, gpointer data) {
if (GTK_IS_LABEL(widget)) {
int* size = static_cast<int*>(data);
gtk_util::ForceFontSizePixels(widget, *size);
return;
}
if (GTK_IS_CONTAINER(widget))
gtk_container_forall(GTK_CONTAINER(widget), SetWidgetFontSizeCallback,
data);
}
void SetWidgetFontSize(GtkWidget* widget, int size) {
gtk_container_forall(GTK_CONTAINER(widget), SetWidgetFontSizeCallback, &size);
}
// Get the index of the row containing |widget|. Assume the widget is the child
// of an hbox, which is a child of a vbox. The hbox represents a row, and the
// vbox the full table.
size_t GetExtensionWidgetRow(GtkWidget* widget) {
GtkWidget* hbox = gtk_widget_get_parent(widget);
DCHECK(hbox);
GtkWidget* vbox = gtk_widget_get_parent(hbox);
DCHECK(vbox);
GList* hbox_list = gtk_container_get_children(GTK_CONTAINER(vbox));
gint index = g_list_index(hbox_list, hbox);
DCHECK(index != -1);
g_list_free(hbox_list);
return index;
}
// A gtk_container_foreach callback to enable/disable a widget.
void EnableWidgetCallback(GtkWidget* widget, gpointer data) {
gtk_widget_set_sensitive(widget, *static_cast<gboolean*>(data));
}
// Create a new widget displaying |rating| as |kNumStarsPerRating| star images.
// Rating should be in the range [0, kNumStarsPerRating].
GtkWidget* CreateStarsWidget(double rating) {
const int kNumStarsPerRating = 5; // Number of stars in a rating.
const int kStarSpacing = 1; // Spacing between stars in pixels.
GtkWidget* hbox = gtk_hbox_new(FALSE, kStarSpacing);
ui::ResourceBundle& rb = ui::ResourceBundle::GetSharedInstance();
for (int i = 0; i < kNumStarsPerRating; ++i) {
GdkPixbuf* star = rb.GetNativeImageNamed(
WebIntentPicker::GetNthStarImageIdFromCWSRating(rating, i),
ui::ResourceBundle::RTL_ENABLED).ToGdkPixbuf();
gtk_box_pack_start(GTK_BOX(hbox), gtk_image_new_from_pixbuf(star),
FALSE, FALSE, 0);
}
return hbox;
}
} // namespace
// Create a new Widget for the "waiting for CWS" display.
class WaitingDialog {
public:
explicit WaitingDialog(GtkThemeService* theme_service);
~WaitingDialog();
GtkWidget* widget() const { return widget_.get(); }
private:
// Initialize the widget.
void Init();
// The actual GtkWidget
ui::OwnedWidgetGtk widget_;
// Waiting throbber.
scoped_ptr<ThrobberGtk> throbber_;
// Weak pointer to theme service.
GtkThemeService* theme_service_;
};
WaitingDialog::WaitingDialog(GtkThemeService* theme_service)
: theme_service_(theme_service) {
DCHECK(theme_service_);
Init();
}
WaitingDialog::~WaitingDialog() {
widget_.Destroy();
}
void WaitingDialog::Init() {
const int kDialogSpacing = 30;
widget_.Own(gtk_vbox_new(FALSE, 0));
GtkWidget* vbox = widget_.get();
// Create throbber
ThrobberGtk* throbber = new ThrobberGtk(theme_service_);
GtkWidget* throbber_alignment = gtk_alignment_new(0.5, 0.5, 0, 0);
gtk_alignment_set_padding(GTK_ALIGNMENT(throbber_alignment), kDialogSpacing,
kMainContentPixelSize, 0, 0);
gtk_container_add(GTK_CONTAINER(throbber_alignment), throbber->widget());
gtk_box_pack_start(GTK_BOX(vbox), throbber_alignment, TRUE, TRUE, 0);
// Add the message text.
GtkWidget* message_label = theme_service_->BuildLabel(
l10n_util::GetStringUTF8(IDS_INTENT_PICKER_WAIT_FOR_CWS).c_str(),
ui::kGdkBlack);
GtkWidget* label_alignment = gtk_alignment_new(0.5, 0.5, 0, 0);
gtk_alignment_set_padding(GTK_ALIGNMENT(label_alignment),
kMainContentPixelSize, kDialogSpacing, 0, 0);
gtk_container_add(GTK_CONTAINER(label_alignment), message_label);
gtk_box_pack_start(GTK_BOX(vbox), label_alignment, TRUE, TRUE, 0);
// TODO(groby): use IDR_SPEECH_INPUT_SPINNER. Pending fix for ThrobberGtk.
// Animate throbber
throbber->Start();
}
// static
WebIntentPicker* WebIntentPicker::Create(WebContents* web_contents,
WebIntentPickerDelegate* delegate,
WebIntentPickerModel* model) {
return new WebIntentPickerGtk(web_contents, delegate, model);
}
WebIntentPickerGtk::WebIntentPickerGtk(WebContents* web_contents,
WebIntentPickerDelegate* delegate,
WebIntentPickerModel* model)
: web_contents_(web_contents),
delegate_(delegate),
model_(model),
contents_(NULL),
header_label_(NULL),
button_vbox_(NULL),
cws_label_(NULL),
extensions_vbox_(NULL),
service_hbox_(NULL),
window_(NULL) {
DCHECK(delegate_ != NULL);
model_->set_observer(this);
InitContents();
UpdateInstalledServices();
UpdateCWSLabel();
UpdateSuggestedExtensions();
GtkThemeService* theme_service = GetThemeService(web_contents);
registrar_.Add(this, chrome::NOTIFICATION_BROWSER_THEME_CHANGED,
content::Source<ThemeService>(theme_service));
theme_service->InitThemesFor(this);
window_ = new ConstrainedWindowGtk(web_contents, this);
if (model_->IsInlineDisposition())
OnInlineDisposition(string16(), model_->inline_disposition_url());
}
WebIntentPickerGtk::~WebIntentPickerGtk() {
}
void WebIntentPickerGtk::Close() {
window_->CloseConstrainedWindow();
if (inline_disposition_web_contents_.get())
inline_disposition_web_contents_->OnCloseStarted();
}
void WebIntentPickerGtk::SetActionString(const string16& action) {
header_label_text_ = action;
if (header_label_)
gtk_label_set_text(GTK_LABEL(header_label_), UTF16ToUTF8(action).c_str());
}
void WebIntentPickerGtk::OnExtensionInstallSuccess(const std::string& id) {
RemoveThrobber();
}
void WebIntentPickerGtk::OnExtensionInstallFailure(const std::string& id) {
// The throbber has an alignment as its parent, so it must be used instead of
// the throbber to find the extension row.
size_t index =
GetExtensionWidgetRow(gtk_widget_get_parent(throbber_->widget()));
GList* vbox_list =
gtk_container_get_children(GTK_CONTAINER(extensions_vbox_));
GtkWidget* hbox = static_cast<GtkWidget*>(g_list_nth_data(vbox_list, index));
RemoveThrobber();
gtk_widget_show_all(hbox);
g_list_free(vbox_list);
SetWidgetsEnabled(true);
}
void WebIntentPickerGtk::OnModelChanged(WebIntentPickerModel* model) {
if (waiting_dialog_.get() && !model->IsWaitingForSuggestions()) {
waiting_dialog_.reset();
InitMainContents();
}
UpdateInstalledServices();
UpdateCWSLabel();
UpdateSuggestedExtensions();
SetActionString(header_label_text_);
}
void WebIntentPickerGtk::OnFaviconChanged(WebIntentPickerModel* model,
size_t index) {
UpdateInstalledServices();
}
void WebIntentPickerGtk::OnExtensionIconChanged(
WebIntentPickerModel* model,
const std::string& extension_id) {
UpdateSuggestedExtensions();
}
void WebIntentPickerGtk::OnInlineDisposition(const string16&,
const GURL& url) {
DCHECK(delegate_);
Profile* profile =
Profile::FromBrowserContext(web_contents_->GetBrowserContext());
inline_disposition_web_contents_.reset(
delegate_->CreateWebContentsForInlineDisposition(profile, url));
Browser* browser = chrome::FindBrowserWithWebContents(web_contents_);
inline_disposition_delegate_.reset(
new WebIntentInlineDispositionDelegate(
this, inline_disposition_web_contents_.get(), browser));
inline_disposition_web_contents_->GetController().LoadURL(
url, content::Referrer(), content::PAGE_TRANSITION_AUTO_TOPLEVEL,
std::string());
// Replace the picker contents with the inline disposition.
ClearContents();
gtk_widget_set_size_request(contents_, -1, -1);
window_->BackgroundColorChanged();
GtkWidget* vbox = gtk_vbox_new(FALSE, 0);
GtkThemeService* theme_service = GetThemeService(web_contents_);
service_hbox_ = gtk_hbox_new(FALSE, ui::kControlSpacing);
// TODO(gbillock): Eventually get the service icon button here.
// Intent action label.
const WebIntentPickerModel::InstalledService* service =
model_->GetInstalledServiceWithURL(url);
GtkWidget* action_label = gtk_label_new(UTF16ToUTF8(service->title).c_str());
gtk_util::ForceFontSizePixels(action_label, kMainContentPixelSize);
// Hardcode color; don't allow theming.
gtk_util::SetLabelColor(action_label, &ui::kGdkBlack);
GtkWidget* label_alignment = gtk_alignment_new(0, 0.5f, 0, 0);
gtk_container_add(GTK_CONTAINER(label_alignment), action_label);
GtkWidget* indent_label = gtk_util::IndentWidget(label_alignment);
gtk_box_pack_start(GTK_BOX(service_hbox_), indent_label, FALSE, TRUE, 0);
g_signal_connect(service_hbox_, "destroy", G_CALLBACK(gtk_widget_destroyed),
&service_hbox_);
// Add link for "choose another service" if other suggestions are available
// or if more than one (the current) service is installed.
if (model_->show_use_another_service() &&
(model_->GetInstalledServiceCount() > 1 ||
model_->GetSuggestedExtensionCount())) {
GtkWidget* use_alternate_link = theme_service->BuildChromeLinkButton(
l10n_util::GetStringUTF8(
IDS_INTENT_PICKER_USE_ALTERNATE_SERVICE).c_str());
gtk_chrome_link_button_set_use_gtk_theme(
GTK_CHROME_LINK_BUTTON(use_alternate_link),
theme_service->UsingNativeTheme());
gtk_util::ForceFontSizePixels(
GTK_CHROME_LINK_BUTTON(use_alternate_link)->label,
kMainContentPixelSize);
g_signal_connect(use_alternate_link, "clicked",
G_CALLBACK(OnChooseAnotherServiceClickThunk), this);
GtkWidget* link_alignment = gtk_alignment_new(0, 0.5f, 0, 0);
gtk_container_add(GTK_CONTAINER(link_alignment), use_alternate_link);
gtk_box_pack_start(GTK_BOX(service_hbox_), link_alignment, TRUE, TRUE, 0);
}
AddCloseButton(service_hbox_);
// The header box
gtk_container_add(GTK_CONTAINER(vbox), service_hbox_);
// hbox for the web contents, so we can have spacing on the borders.
GtkWidget* alignment = gtk_alignment_new(0.0, 0.0, 1.0, 1.0);
gtk_alignment_set_padding(
GTK_ALIGNMENT(alignment), 0, ui::kContentAreaBorder,
ui::kContentAreaBorder, ui::kContentAreaBorder);
gfx::NativeView web_contents_widget =
inline_disposition_web_contents_->GetView()->GetNativeView();
gtk_container_add(GTK_CONTAINER(alignment), web_contents_widget);
gtk_container_add(GTK_CONTAINER(vbox), alignment);
gtk_container_add(GTK_CONTAINER(contents_), vbox);
gfx::Size size = GetMinInlineDispositionSize();
gtk_widget_set_size_request(web_contents_widget,
size.width(), size.height());
gtk_widget_show_all(contents_);
inline_disposition_delegate_->SetRenderViewSizeLimits();
inline_disposition_web_contents_->GetView()->SetInitialFocus();
host_signals_.reset(new ui::GtkSignalRegistrar());
host_signals_->Connect(
web_contents_->GetRenderViewHost()->GetView()->GetNativeView(),
"size-allocate",
G_CALLBACK(OnHostContentsSizeAllocateThunk), this);
}
void WebIntentPickerGtk::OnInlineDispositionAutoResize(const gfx::Size& size) {
gfx::NativeView web_contents_widget =
inline_disposition_web_contents_->GetView()->GetNativeView();
gtk_widget_set_size_request(web_contents_widget, size.width(), size.height());
}
gfx::Size WebIntentPickerGtk::GetMaxInlineDispositionSize() {
gfx::Rect tab_bounds(web_contents_->GetRenderViewHost()->
GetView()->GetNativeView()->allocation);
GtkRequisition req = {};
if (service_hbox_)
gtk_widget_size_request(service_hbox_, &req);
tab_bounds.Inset(2 * ui::kContentAreaBorder,
2 * ui::kContentAreaBorder + req.height);
return tab_bounds.size();
}
void WebIntentPickerGtk::OnPendingAsyncCompleted() {
// Requests to both the WebIntentService and the Chrome Web Store have
// completed. If there are any services, installed or suggested, there's
// nothing to do.
if (model_->GetInstalledServiceCount() ||
model_->GetSuggestedExtensionCount())
return;
// If there are no installed or suggested services at this point,
// inform the user about it.
// Replace the picker contents with dialog box.
ClearContents();
GtkWidget* sub_contents = CreateSubContents(contents_);
AddCloseButton(contents_);
AddTitle(sub_contents);
// Replace the dialog header.
DCHECK(header_label_);
gtk_label_set_text(
GTK_LABEL(header_label_),
l10n_util::GetStringUTF8(IDS_INTENT_PICKER_NO_SERVICES_TITLE).c_str());
// Add the message text.
GtkWidget* hbox = gtk_hbox_new(FALSE, 0);
gtk_box_pack_start(GTK_BOX(sub_contents), hbox, TRUE, TRUE, 0);
GtkThemeService* theme_service = GetThemeService(web_contents_);
GtkWidget* no_service_label = theme_service->BuildLabel(
l10n_util::GetStringUTF8(IDS_INTENT_PICKER_NO_SERVICES).c_str(),
ui::kGdkBlack);
gtk_label_set_line_wrap(GTK_LABEL(no_service_label), TRUE);
gtk_misc_set_alignment(GTK_MISC(no_service_label), 0, 0);
// Set the label width to the size of |sub_contents|, which we don't have
// access to yet, by calculating the main content width minus borders.
gtk_util::SetLabelWidth(no_service_label,
kWindowMinWidth - 2 * ui::kContentAreaBorder);
gtk_box_pack_start(GTK_BOX(hbox), no_service_label, TRUE, TRUE, 0);
gtk_widget_show_all(contents_);
}
void WebIntentPickerGtk::InvalidateDelegate() {
delegate_ = NULL;
}
GtkWidget* WebIntentPickerGtk::GetWidgetRoot() {
return contents_;
}
GtkWidget* WebIntentPickerGtk::GetFocusWidget() {
return contents_;
}
void WebIntentPickerGtk::DeleteDelegate() {
// The delegate is deleted when the contents widget is destroyed. See
// OnDestroy.
if (delegate_)
delegate_->OnClosing();
}
bool WebIntentPickerGtk::GetBackgroundColor(GdkColor* color) {
if (inline_disposition_web_contents_.get()) {
*color = ui::kGdkWhite;
return true;
}
return false;
}
bool WebIntentPickerGtk::ShouldHaveBorderPadding() const {
return false;
}
void WebIntentPickerGtk::Observe(int type,
const content::NotificationSource& source,
const content::NotificationDetails& details) {
DCHECK_EQ(type, chrome::NOTIFICATION_BROWSER_THEME_CHANGED);
if (header_label_) {
GtkThemeService* theme_service = GetThemeService(web_contents_);
if (theme_service->UsingNativeTheme())
gtk_util::UndoForceFontSize(header_label_);
else
gtk_util::ForceFontSizePixels(header_label_, kHeaderLabelPixelSize);
}
UpdateInstalledServices();
UpdateSuggestedExtensions();
}
void WebIntentPickerGtk::OnDestroy(GtkWidget* button) {
// Destroy this object when the contents widget is destroyed. It can't be
// "delete this" because this function happens in a callback.
MessageLoop::current()->DeleteSoon(FROM_HERE, this);
model_->set_observer(NULL);
window_ = NULL;
}
void WebIntentPickerGtk::OnCloseButtonClick(GtkWidget* button) {
DCHECK(delegate_);
delegate_->OnUserCancelledPickerDialog();
}
void WebIntentPickerGtk::OnExtensionLinkClick(GtkWidget* link) {
DCHECK(delegate_);
size_t index = GetExtensionWidgetRow(link);
const WebIntentPickerModel::SuggestedExtension& extension =
model_->GetSuggestedExtensionAt(index);
delegate_->OnExtensionLinkClicked(extension.id,
event_utils::DispositionForCurrentButtonPressEvent());
}
void WebIntentPickerGtk::OnExtensionInstallButtonClick(GtkWidget* button) {
DCHECK(delegate_);
size_t index = GetExtensionWidgetRow(button);
const WebIntentPickerModel::SuggestedExtension& extension =
model_->GetSuggestedExtensionAt(index);
delegate_->OnExtensionInstallRequested(extension.id);
SetWidgetsEnabled(false);
// Re-enable the clicked extension row.
GList* vbox_list =
gtk_container_get_children(GTK_CONTAINER(extensions_vbox_));
GtkWidget* hbox = static_cast<GtkWidget*>(g_list_nth_data(vbox_list, index));
gtk_widget_set_sensitive(hbox, TRUE);
// Hide the install button.
GList* hbox_list = gtk_container_get_children(GTK_CONTAINER(hbox));
GtkWidget* install_button =
static_cast<GtkWidget*>(g_list_nth_data(hbox_list, kInstallButtonIndex));
GtkAllocation allocation;
gtk_widget_get_allocation(install_button, &allocation);
gtk_widget_hide(install_button);
g_list_free(hbox_list);
g_list_free(vbox_list);
// Show the throbber with the same size as the install button.
GtkWidget* throbber = AddThrobberToExtensionAt(index);
gtk_widget_set_size_request(throbber, allocation.width, allocation.height);
gtk_widget_show_all(throbber);
}
void WebIntentPickerGtk::OnMoreSuggestionsLinkClick(GtkWidget* link) {
DCHECK(delegate_);
delegate_->OnSuggestionsLinkClicked(
event_utils::DispositionForCurrentButtonPressEvent());
}
void WebIntentPickerGtk::OnChooseAnotherServiceClick(GtkWidget* link) {
DCHECK(delegate_);
delegate_->OnChooseAnotherService();
ResetContents();
}
void WebIntentPickerGtk::OnHostContentsSizeAllocate(GtkWidget* widget,
GdkRectangle* rectangle) {
if (!inline_disposition_delegate_.get())
return;
inline_disposition_delegate_->SetRenderViewSizeLimits();
}
void WebIntentPickerGtk::OnServiceButtonClick(GtkWidget* button) {
DCHECK(delegate_);
GList* button_list = gtk_container_get_children(GTK_CONTAINER(button_vbox_));
gint index = g_list_index(button_list, button);
DCHECK(index != -1);
g_list_free(button_list);
const WebIntentPickerModel::InstalledService& installed_service =
model_->GetInstalledServiceAt(index);
delegate_->OnServiceChosen(installed_service.url,
installed_service.disposition,
WebIntentPickerDelegate::kEnableDefaults);
}
void WebIntentPickerGtk::InitContents() {
GtkThemeService* theme_service = GetThemeService(web_contents_);
// Main contents vbox.
if (!contents_) {
contents_ = gtk_vbox_new(FALSE, 0);
g_signal_connect(contents_, "destroy", G_CALLBACK(&OnDestroyThunk), this);
}
gtk_widget_set_size_request(contents_, kWindowMinWidth, -1);
if (model_ && model_->IsWaitingForSuggestions()) {
ClearContents();
AddCloseButton(contents_);
waiting_dialog_.reset(new WaitingDialog(theme_service));
gtk_box_pack_start(GTK_BOX(contents_), waiting_dialog_->widget(),
TRUE, TRUE, 0);
} else {
InitMainContents();
}
}
void WebIntentPickerGtk::InitMainContents() {
GtkThemeService* theme_service = GetThemeService(web_contents_);
ClearContents();
AddCloseButton(contents_);
GtkWidget* sub_contents = CreateSubContents(contents_);
AddTitle(sub_contents);
// Add separation between the installed services list and the app suggestions.
GtkWidget* button_alignment = gtk_alignment_new(0.5, 0, 0, 0);
gtk_alignment_set_padding(GTK_ALIGNMENT(button_alignment), 0,
kMainContentPixelSize * 2, 0, 0);
// Vbox containing all service buttons.
button_vbox_ = gtk_vbox_new(FALSE, ui::kControlSpacing);
gtk_container_add(GTK_CONTAINER(button_alignment), button_vbox_);
gtk_box_pack_start(GTK_BOX(sub_contents), button_alignment, TRUE, TRUE, 0);
// Chrome Web Store label.
cws_label_ = theme_service->BuildLabel(
l10n_util::GetStringUTF8(IDS_INTENT_PICKER_GET_MORE_SERVICES).c_str(),
ui::kGdkBlack);
gtk_box_pack_start(GTK_BOX(sub_contents), cws_label_, TRUE, TRUE, 0);
gtk_misc_set_alignment(GTK_MISC(cws_label_), 0, 0);
gtk_widget_set_no_show_all(cws_label_, TRUE);
// Set the label width to the size of |sub_contents|, which we don't have
// access to yet, by calculating the main content width minus borders.
gtk_util::SetLabelWidth(cws_label_,
kWindowMinWidth - 2 * ui::kContentAreaBorder);
gtk_util::ForceFontSizePixels(cws_label_, kMainContentPixelSize);
// Suggested extensions vbox.
extensions_vbox_ = gtk_vbox_new(FALSE, ui::kControlSpacing);
GtkWidget* indent_extensions = gtk_alignment_new(0.0, 0.5, 1.0, 1.0);
gtk_alignment_set_padding(GTK_ALIGNMENT(indent_extensions), 0, 0,
ui::kGroupIndent, ui::kGroupIndent);
gtk_container_add(GTK_CONTAINER(indent_extensions), extensions_vbox_);
gtk_widget_set_no_show_all(indent_extensions, TRUE);
gtk_box_pack_start(GTK_BOX(sub_contents), indent_extensions, TRUE, TRUE, 0);
// CWS 'More Suggestions' link.
GtkWidget* link_alignment = gtk_alignment_new(0, 0.5f, 0, 0);
GtkWidget* more_hbox = gtk_hbox_new(FALSE, ui::kControlSpacing);
ui::ResourceBundle& rb = ui::ResourceBundle::GetSharedInstance();
GdkPixbuf* cws_icon =
rb.GetNativeImageNamed(IDR_WEBSTORE_ICON_16).ToGdkPixbuf();
gtk_box_pack_start(GTK_BOX(more_hbox), gtk_image_new_from_pixbuf(cws_icon),
FALSE, FALSE, 0);
GtkWidget* more_suggestions_link = theme_service->BuildChromeLinkButton(
l10n_util::GetStringUTF8(IDS_FIND_MORE_INTENT_HANDLER_MESSAGE).c_str());
gtk_box_pack_start(GTK_BOX(more_hbox), more_suggestions_link,
FALSE, FALSE, 0);
gtk_container_add(GTK_CONTAINER(link_alignment), more_hbox);
gtk_chrome_link_button_set_use_gtk_theme(
GTK_CHROME_LINK_BUTTON(more_suggestions_link),
theme_service->UsingNativeTheme());
gtk_util::ForceFontSizePixels(
GTK_CHROME_LINK_BUTTON(more_suggestions_link)->label,
kMainContentPixelSize);
g_signal_connect(more_suggestions_link, "clicked",
G_CALLBACK(OnMoreSuggestionsLinkClickThunk), this);
GtkWidget* indent_link = gtk_util::IndentWidget(link_alignment);
gtk_box_pack_start(GTK_BOX(sub_contents), indent_link, TRUE, TRUE, 0);
// Throbber, which will be added to the hierarchy when necessary.
throbber_.reset(new ThrobberGtk(theme_service));
gtk_widget_show_all(contents_);
}
void WebIntentPickerGtk::ClearContents() {
// Wipe out all currently displayed widgets.
gtk_util::RemoveAllChildren(contents_);
header_label_ = NULL;
button_vbox_ = NULL;
cws_label_ = NULL;
extensions_vbox_ = NULL;
}
void WebIntentPickerGtk::ResetContents() {
ClearContents();
// Reset potential inline disposition data.
inline_disposition_web_contents_.reset();
inline_disposition_delegate_.reset();
window_->BackgroundColorChanged();
// Re-initialize picker widgets and data.
InitMainContents();
UpdateInstalledServices();
UpdateCWSLabel();
UpdateSuggestedExtensions();
SetActionString(header_label_text_);
gtk_widget_show_all(contents_);
}
GtkWidget* WebIntentPickerGtk::CreateSubContents(GtkWidget* box) {
GtkWidget* alignment = gtk_alignment_new(0.0, 0.0, 1.0, 1.0);
gtk_alignment_set_padding(
GTK_ALIGNMENT(alignment), 0, ui::kContentAreaBorder,
ui::kContentAreaBorder, ui::kContentAreaBorder);
gtk_box_pack_end(GTK_BOX(box), alignment, TRUE, TRUE, 0);
GtkWidget* sub_contents = gtk_vbox_new(FALSE, ui::kContentAreaSpacing);
gtk_container_add(GTK_CONTAINER(alignment), sub_contents);
return sub_contents;
}
void WebIntentPickerGtk::AddCloseButton(GtkWidget* containingBox) {
// Hbox containing the close button.
GtkWidget* close_hbox = gtk_hbox_new(FALSE, 0);
gtk_box_pack_start(GTK_BOX(containingBox), close_hbox, TRUE, TRUE, 0);
close_button_.reset(
CustomDrawButton::CloseButton(GetThemeService(web_contents_)));
g_signal_connect(close_button_->widget(), "clicked",
G_CALLBACK(OnCloseButtonClickThunk), this);
gtk_widget_set_can_focus(close_button_->widget(), FALSE);
gtk_box_pack_end(GTK_BOX(close_hbox), close_button_->widget(),
FALSE, FALSE, 0);
}
void WebIntentPickerGtk::AddTitle(GtkWidget* containingBox) {
// Hbox containing the header label.
GtkWidget* header_hbox = gtk_hbox_new(FALSE, 0);
gtk_box_pack_start(GTK_BOX(containingBox), header_hbox, TRUE, TRUE, 0);
// Label text will be set in the call to SetActionString().
header_label_ = GetThemeService(web_contents_)->BuildLabel(
std::string(), ui::kGdkBlack);
gtk_util::ForceFontSizePixels(header_label_, kHeaderLabelPixelSize);
gtk_box_pack_start(GTK_BOX(header_hbox), header_label_, TRUE, TRUE, 0);
gtk_misc_set_alignment(GTK_MISC(header_label_), 0, 0);
}
void WebIntentPickerGtk::UpdateInstalledServices() {
if (!button_vbox_)
return;
gtk_util::RemoveAllChildren(button_vbox_);
if (model_->GetInstalledServiceCount() == 0) {
gtk_widget_hide(gtk_widget_get_parent(button_vbox_));
return;
}
for (size_t i = 0; i < model_->GetInstalledServiceCount(); ++i) {
const WebIntentPickerModel::InstalledService& installed_service =
model_->GetInstalledServiceAt(i);
GtkWidget* button = gtk_button_new();
gtk_widget_set_tooltip_text(button, installed_service.url.spec().c_str());
gtk_button_set_label(GTK_BUTTON(button),
UTF16ToUTF8(installed_service.title).c_str());
gtk_button_set_alignment(GTK_BUTTON(button), 0, 0);
gtk_container_add(GTK_CONTAINER(button_vbox_), button);
g_signal_connect(button, "clicked", G_CALLBACK(OnServiceButtonClickThunk),
this);
SetServiceButtonImage(button, installed_service.favicon.ToGdkPixbuf());
// Must be called after SetServiceButtonImage as the internal label widget
// is replaced in that call.
SetWidgetFontSize(button, kMainContentPixelSize);
}
gtk_widget_show_all(button_vbox_);
gtk_widget_show(gtk_widget_get_parent(button_vbox_));
}
void WebIntentPickerGtk::UpdateCWSLabel() {
if (!button_vbox_)
return;
gtk_widget_set_visible(gtk_widget_get_parent(button_vbox_),
model_->GetInstalledServiceCount() != 0);
std::string label_text = UTF16ToUTF8(model_->GetSuggestionsLinkText());
gtk_label_set_text(GTK_LABEL(cws_label_), label_text.c_str());
gtk_widget_set_visible(cws_label_, !label_text.empty());
}
void WebIntentPickerGtk::UpdateSuggestedExtensions() {
if (!extensions_vbox_)
return;
GtkThemeService* theme_service = GetThemeService(web_contents_);
gtk_util::RemoveAllChildren(extensions_vbox_);
if (model_->GetSuggestedExtensionCount() == 0) {
gtk_widget_hide(gtk_widget_get_parent(extensions_vbox_));
return;
}
gtk_widget_show(gtk_widget_get_parent(extensions_vbox_));
for (size_t i = 0; i < model_->GetSuggestedExtensionCount(); ++i) {
const WebIntentPickerModel::SuggestedExtension& extension =
model_->GetSuggestedExtensionAt(i);
GtkWidget* hbox = gtk_hbox_new(FALSE, ui::kControlSpacing);
gtk_box_pack_start(GTK_BOX(extensions_vbox_), hbox, FALSE, FALSE, 0);
// Icon.
GtkWidget* icon = gtk_image_new_from_pixbuf(extension.icon.ToGdkPixbuf());
gtk_box_pack_start(GTK_BOX(hbox), icon, FALSE, FALSE, 0);
// Title link.
string16 elided_title = ui::ElideText(extension.title, gfx::Font(),
kTitleLinkMaxWidth, ui::ELIDE_AT_END);
GtkWidget* title_link = theme_service->BuildChromeLinkButton(
UTF16ToUTF8(elided_title).c_str());
gtk_chrome_link_button_set_use_gtk_theme(GTK_CHROME_LINK_BUTTON(title_link),
theme_service->UsingNativeTheme());
gtk_util::ForceFontSizePixels(GTK_CHROME_LINK_BUTTON(title_link)->label,
kMainContentPixelSize);
g_signal_connect(title_link, "clicked",
G_CALLBACK(OnExtensionLinkClickThunk), this);
gtk_box_pack_start(GTK_BOX(hbox), title_link, FALSE, FALSE, 0);
// Stars.
GtkWidget* stars = CreateStarsWidget(extension.average_rating);
gtk_box_pack_start(GTK_BOX(hbox), stars, FALSE, FALSE, 0);
// Install button.
GtkWidget* install_button = gtk_button_new();
gtk_button_set_label(
GTK_BUTTON(install_button),
l10n_util::GetStringUTF8(IDS_INTENT_PICKER_INSTALL_EXTENSION).c_str());
GtkWidget* label = gtk_bin_get_child(GTK_BIN(install_button));
gtk_util::ForceFontSizePixels(label, kMainContentPixelSize);
g_signal_connect(install_button, "clicked",
G_CALLBACK(OnExtensionInstallButtonClickThunk), this);
gtk_box_pack_end(GTK_BOX(hbox), install_button, FALSE, FALSE, 0);
}
gtk_widget_show_all(extensions_vbox_);
gtk_widget_show(gtk_widget_get_parent(extensions_vbox_));
}
void WebIntentPickerGtk::SetWidgetsEnabled(bool enabled) {
gboolean data = enabled;
if (button_vbox_)
gtk_container_foreach(GTK_CONTAINER(button_vbox_), EnableWidgetCallback,
&data);
if (extensions_vbox_)
gtk_container_foreach(GTK_CONTAINER(extensions_vbox_), EnableWidgetCallback,
&data);
}
GtkWidget* WebIntentPickerGtk::AddThrobberToExtensionAt(size_t index) {
// The throbber should be unparented.
DCHECK(!gtk_widget_get_parent(throbber_->widget()));
GList* vbox_list =
gtk_container_get_children(GTK_CONTAINER(extensions_vbox_));
GtkWidget* hbox = static_cast<GtkWidget*>(g_list_nth_data(vbox_list, index));
GtkWidget* alignment = gtk_alignment_new(0.5, 0.5, 0, 0);
gtk_container_add(GTK_CONTAINER(alignment), throbber_->widget());
gtk_box_pack_end(GTK_BOX(hbox), alignment, FALSE, FALSE, 0);
g_list_free(vbox_list);
throbber_->Start();
return alignment;
}
void WebIntentPickerGtk::RemoveThrobber() {
GtkWidget* alignment = gtk_widget_get_parent(throbber_->widget());
DCHECK(alignment);
gtk_container_remove(GTK_CONTAINER(alignment), throbber_->widget());
gtk_widget_destroy(alignment);
throbber_->Stop();
}
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 2,664
|
Q: How to Most Succinctly Distinguish Base-n Numbers From Actual Numbers? When reading mathematical texts I encounter phrases such as "the decimal number 2," the "binary number 10," etc., so much that I begin to imagine there really is such a thing as a "base-n number," whereas the reality (at least as far as I understand it) is that numbers are distinct from their representations. In other words, it seems to me that our minds tend to conflate numbers with number words, so that, e.g., the phrase "the decimal number 70" invites ambiguity of meaning between "the number represented by 70 in the decimal numeral system," on the one hand, and "the string '70' as a word in the decimal language," on the other. Nevertheless the two are of course distinct, and failing to observe the distinction can lead to confusion. For example, it can be easily proven that any real number has exactly one additive inverse such the the sum of the two is 0. However, in the decimal system, all of the members of the set $S = \{-70, -070, -0070,-00070, \ldots\}\cup\{-70., -70.0, -70.00, -70.000, \ldots\}$ are distinct words, yet as "decimal numbers" any of them could be substituted for $x$ in the equation $70 + x = 0$, thereby appearing to violate the uniqueness of the additive inverse.
I am writing a paper where maintaining the distinction between numbers and representations is crucial, but I find it both tedious and unconducive to readability to repeatedly call out that distinction using phrases like "the number represented by $x$ in the base-$n$ numeral system" and "the string $x$ as a word in the base-$n$ language" at every turn. Any suggestions on how to make the writing less awkward?
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 5,484
|
Q: Column sum (displayed in page footer) not working on first page. But working on second page I'm working on a jasper report. My requirement is to show the running total for debit and credit column which I'm getting from java as JRBeanCollectionDataSource. I have created a variable and set the resetType="Page" and calculation as Sum. The total on first page is always zero where as the total on second page is correct.
My field is within subDataset. I'm returning the sum of debit and credit values from datasetRun.
Subdataset:
<subDataset name="TransactionDataSet" uuid="1c386547-d313-475c-b023-9f1b6f198535">
<property name="com.jaspersoft.studio.data.defaultdataadapter" value="One Empty Record"/>
<queryString>
<![CDATA[]]>
</queryString>
<field name="accountTypeName" class="java.lang.String"/>
<field name="accountNum" class="java.lang.String"/>
<field name="name" class="java.lang.String"/>
<field name="transactionText" class="java.lang.String"/>
<field name="currencyCode" class="java.lang.String"/>
<field name="amountDebited" class="java.math.BigDecimal"/>
<field name="amountCredited" class="java.math.BigDecimal"/>
<field name="regionId" class="java.lang.String"/>
<field name="departmentId" class="java.lang.String"/>
<field name="costCenterId" class="java.lang.String"/>
<variable name="debitSumForPage" class="java.math.BigDecimal" resetType="Page" calculation="Sum">
<variableExpression><![CDATA[$F{amountDebited}]]></variableExpression>
<initialValueExpression><![CDATA[java.math.BigDecimal.ZERO]]></initialValueExpression>
</variable>
<variable name="creditSumForPage" class="java.math.BigDecimal" resetType="Page" calculation="Sum">
<variableExpression><![CDATA[$F{amountCredited}]]></variableExpression>
<initialValueExpression><![CDATA[java.math.BigDecimal.ZERO]]></initialValueExpression>
</variable>
</subDataset>
Global variable:
<variable name="totalDebitForPage" class="java.math.BigDecimal" calculation="Sum">
<initialValueExpression><![CDATA[java.math.BigDecimal.ZERO]]></initialValueExpression>
</variable>
<variable name="totalCreditForPage" class="java.math.BigDecimal" calculation="Sum">
<initialValueExpression><![CDATA[java.math.BigDecimal.ZERO]]></initialValueExpression>
</variable>
return code:
<datasetRun subDataset="TransactionDataSet" uuid="c6dc230d-6e84-4618-b973-c5cedc63dcaf">
<dataSourceExpression><![CDATA[$P{TransactionDatasource}]]></dataSourceExpression>
<returnValue fromVariable="debitSumForPage" toVariable="totalDebitForPage"/>
<returnValue fromVariable="creditSumForPage" toVariable="totalCreditForPage"/>
</datasetRun>
I need to get the page total on all pages.
|
{
"redpajama_set_name": "RedPajamaStackExchange"
}
| 3,860
|
Центральний банк Коста-Рики () – центральний банк Республіки Коста-Рика.
Історія
До 24 листопада 1914 року банкноти випускалися переважно приватними банками, іноді — урядом. Після зазначеної дати виключне право емісії банкнот належало Міжнародному банку Коста-Рики. Наприкінці 1936 року Міжнародний банк Коста-Рики був реорганізований у Національний банк Коста-Рики. Законом від 28 січня 1950 року на базі Національного банку Коста-Рики створено державний Центральний банк Коста-Рики
Сфера діяльності
Для виконання своєї місії та підтримки внутрішнього та зовнішнього балансу економіки, цілі та операційні завдання Центрального банку Коста-Рики наступні:
Підтримувати внутрішню стабільність національної валюти, прагнучи повної зайнятості виробничих ресурсів.
Підтримувати зовнішню стабільність національної валюти та забезпечувати їхню вільну конвертацію в інші валюти.
Сприяння стабільній системі фінансового посередництва, ефективної та конкурентоспроможної.
Подібно до інших Центральних банків світу, функції Центрального банку Коста-Рики включають надання банківських послуг уряду Коста-Рики та фінансовим установам, емісію національної валюти, регулювання діяльності комерційних банків та інших фінансових установ, надання економічних консультацій уряду, проведення досліджень та публікацію інформації про грошово-кредитних та інших економічних подій. Тільки банкноти та монети, випущені Центральним банком, є законним платіжним засобом у Коста-Риці.
Фінанси Коста-Рики
засновані в Північній Америці 1950
|
{
"redpajama_set_name": "RedPajamaWikipedia"
}
| 7,974
|
{"url":"https:\/\/discuss.bluerobotics.com\/t\/depth-vs-ping-data\/13036","text":"# Depth vs Ping data\n\nHi,\n\nI have a minor problem with the ping data related to how gain is controlled during a dive with ardusub and this create a mismatch between ROV depth and Sonar reading. My Sonar is direclty connected to the pixhawk so I can\u2019t manipulated the setting (gain \/ ping frequency) with this configuration. Changes are cause by ardusub manipulation and I would like to understand what is going on here.\n\nThe graph below show the problem with real data. The graph show the depth of my ROV (Blue dot) and the Sonar data (Green dot). At the beginning of the graph, the ROV is at the surface (0.1m) and the sonar reads about 17.5m. I then dive to the bottom and we gradually see the rov depth increase while Sonar depth decrease. So far so good. They also match when I paused at about 5m.\n\nWhen the ROV reach 15m and the sonar read about 2.8 from the bottom, the Ping reading suddenly shift reaching >20m. Of course this is impossible since I was near the bottom. It\u2019s clearly a gain or offset effect.\n\nIn red is my attempt to correct the data. Essentially, I added an offset to the data and the result (red line) is realistic. I guess.\n\nHi @Charles,\n\nMost likely the altitude distance was too short for the sonar to detect with its configured scanning range and gain, in which case the next strongest response is likely after the ping has reflected up off the bottom, down off the surface, and back up off the bottom again:\n\n\\begin{align} \\text{measurement} &= 2\\cdot\\text{altitude} + \\text{depth}\\\\ \\rightarrow \\text{altitude}_\\text{corrected} &= \\frac{\\text{measurement} - \\text{depth}}{2} \\end{align}\n\nThe ArduPilot ping library doesn\u2019t seem to do any configuration of the settings, so the sonar will be operating in its default automatic mode.\n\nI would expect the profile data to show usable signal at the range you were at, but its possible the distance estimator could be struggling there and start searching further away, especially if there are weak reflections from things like plants, fish, and particulate matter between the bottom and the sonar.\n\nI also think it can achieve better result. During the same run, it was measuring 0.8m in shallow water. In fact I mapped almost all the lake from the surface and obtained a nice bathymetric survey. It just behave like this when I dive and reach the bottom.\n\nAnd do you know what it is doing in this mode?\nCan I set the unit in another mode (using ping viewer once) and assume it will use this mode all the time after that?\nIn other word, can it retain a configuration in it\u2019s non-volatile memory and reproduce it at any boot up?\n\nthank you\n\nFair enough. I suspect there may have been some issues with the surface and\/or water conditions in the area where it missed the initial return, but it\u2019s hard to say for sure.\n\nI don\u2019t have access to the source code, so not really beyond what\u2019s in our public documentation, which is that automatic mode tries to automatically determine the appropriate gain and scan range based on the measurements it\u2019s receiving.\n\nNo. I just tried this (with our Python library) and while Ping Viewer will remember the previous configuration of a device and re-request that when it connects, the Ping Sonar itself does not maintain its configuration across power cycles - it reverts to its default settings.\n\nThank you for having done this test.\n\nIf I go the Pi \u2192 ping-viewer way, how many bandwith should I expect to loose to this?\n\nThat depends on the rate it\u2019s sending profiles, which in auto mode depends partially on the distances being measured.\n\nIt\u2019s perhaps worth noting that if you want only the distance measurements (no profiles) but with some manually specified settings, you can use Ping Viewer (or a Python script) to connect to the sonar and set the desired settings on startup, and from there just allow the BlueOS Pings service (available from BlueOS 1.1.0-beta.9) to send MAVLink messages with the distances. The device settings are maintained until they are manually changed or the sonar is powered off.","date":"2022-11-27 04:28:08","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 1, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.5412031412124634, \"perplexity\": 1764.5526944383137}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-49\/segments\/1669446710192.90\/warc\/CC-MAIN-20221127041342-20221127071342-00721.warc.gz\"}"}
| null | null |
<html>
<head>
<title></title>
<style type="text/css">
html:before {
content: "";
width: 50px;
height: 50px;
position: fixed;
top: 100px;
border: solid 5px;
}
</style>
<script type="text/javascript">
if (window.testRunner)
testRunner.dumpAsText();
function test()
{
if (!window.testRunner)
return;
eventSender.mouseMoveTo(30, 30);
document.body.offsetTop; // force layout
eventSender.mouseMoveTo(30, 130);
}
</script>
</head>
<body onload="test()">
<p>
Regression test for <i><a href="http://bugs.webkit.org/show_bug.cgi?id=11505">http://bugs.webkit.org/show_bug.cgi?id=11505</a>
REGRESSION: Null pointer deref in HitTestResult::spellingToolTip() (assertion failure in Node::document)</i>.
</p>
<p>
No assert means test PASS.
</p>
</body>
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 7,786
|
package org.apache.camel.component.twitter;
import java.lang.reflect.InvocationHandler;
import java.lang.reflect.Method;
import java.lang.reflect.Proxy;
import org.apache.camel.EndpointInject;
import org.apache.camel.builder.RouteBuilder;
import org.apache.camel.component.mock.MockEndpoint;
import org.apache.camel.impl.JndiRegistry;
import org.junit.Test;
import twitter4j.Status;
import twitter4j.StatusListener;
import twitter4j.TwitterStream;
public class SearchEventTest extends CamelTwitterTestSupport {
@EndpointInject(uri = "mock:result")
protected MockEndpoint resultEndpoint;
private TwitterStream twitterStream;
private StatusListener listener;
@Test
public void testSearchTimeline() throws Exception {
resultEndpoint.expectedMinimumMessageCount(1);
Status status = (Status) Proxy.newProxyInstance(getClass().getClassLoader(),
new Class[]{Status.class},
new TwitterHandler());
listener.onStatus(status);
//"#cameltest tweet");
resultEndpoint.assertIsSatisfied();
}
protected RouteBuilder createRouteBuilder() {
return new RouteBuilder() {
public void configure() {
from("twitter-streaming://filter?type=event&twitterStream=#twitterStream&keywords=#cameltest")
.transform(body().convertToString()).to("mock:result");
}
};
}
@Override
protected JndiRegistry createRegistry() throws Exception {
twitterStream = (TwitterStream) Proxy.newProxyInstance(getClass().getClassLoader(),
new Class[]{TwitterStream.class},
new TwitterHandler());
JndiRegistry registry = super.createRegistry();
registry.bind("twitterStream", twitterStream);
return registry;
}
public class TwitterHandler implements InvocationHandler {
public Object invoke(Object proxy, Method method, Object[] args) throws Throwable {
//mock some methods
if ("addListener".equals(method.getName())) {
listener = (StatusListener) args[0];
} else if ("toString".equals(method.getName())) {
return this.toString();
} else if ("getText".equals(method.getName())) {
return "#cameltest tweet";
} else if ("getUser".equals(method.getName())) {
return Proxy.newProxyInstance(getClass().getClassLoader(),
new Class[]{twitter4j.User.class},
new TwitterHandler());
}
return null;
}
}
}
|
{
"redpajama_set_name": "RedPajamaGithub"
}
| 267
|
\section{Introduction}
Dating back to the 1950s \citep{page1954continuous},
change point analysis has a long tradition in statistics.
The area continues to be an active field of research due to its importance in many applications where data is routinely collected in highly nonstationary environments, examples of which include genetics \citep{hocking2013learning} and climatology \citep{reeves2007review}.
Data segmentation, a.k.a.\ multiple change point detection, enables partitioning of a time series into stationary regions and thus provides a simple framework for modelling nonstationary time series.
In this paper, we consider the problem of detecting multiple change points in the piecewise constant mean of an otherwise stationary time series.
We briefly review the existing literature on multiple change point detection
in the presence of serial dependence; we refer to \cite{aue2013} and \cite{truong2020selective} for a comprehensive overview.
One line of research takes a parametric approach
and simultaneously estimates the serial dependence and change points. For example, \cite{chakar2017robust}, \cite{fang2020detection} and \cite{romano2021detecting} assume an autoregressive (AR) model of order one, while \cite{lu2010mdl}, \cite{cho2021multiple} and \cite{gallagher2021autocovariance} permit an AR model of arbitrary order.
Another line of research focuses on applying the methodologies
developed for independent data to time series settings.
\cite{lavielle2000} and \cite{cho2021data} extend the use of an information criterion \citep{yao1988} to data exhibiting serial correlations and heavy tails,
which requires the choice of an appropriate penalty that depends on the level of noise.
\cite{tecuapetla2017autocovariance}, \cite{eichinger2018mosum}, \cite{dette2020multiscale} and \cite{chan2022optimal}
propose estimators of the long-run variance (LRV) for quantifying the level of noise
that are robust to the presence of multiple mean shifts.
Such estimators are then combined e.g.\ with a moving sum procedure or the simultaneous multiscale change point estimator algorithm
\citep{frick2014multiscale}, for multiple change point detection.
We also note that \cite{wu2019mace} and \cite{zhao2021segmenting} extend self-normalisation-based change point tests
to the data segmentation problem.
In this paper, we propose a robust estimator of the
scale-dependent time-average variance constant (TAVC, \citeauthor{wu2009recursive}, \citeyear{wu2009recursive}).
It is closely related to the literature on estimation of the LRV,
namely $\sigma^2 = \lim_{N \to \infty} \Var(N^{-1/2} \sum_{t = 1}^N \varepsilon_t)$ for a stationary time series $\{\varepsilon_t\}_{t \in \mathbb{Z}}$,
but distinct in that our interest lies in estimating
\begin{align}
\label{eq:tavc}
\sigma^2_L = \Var\l( \frac{1}{\sqrt L} \sum_{t = 1}^L \varepsilon_t \r),
\end{align}
for a given scale $L$.
We argue that such a scale-dependent TAVC estimator is well-suited
to be combined with a class of {\it multiscale} change point detection methodologies, examples of which include
the moving sum (MOSUM) procedure \citep{eichinger2018mosum} and the wild binary segmentation (WBS, \citeauthor{fryzlewicz2014wild}, \citeyear{fryzlewicz2014wild}) algorithm. Such approaches {\it locally} apply change point tests for single change point detection to data sections of varying lengths and achieve good adaptivity in multiple change point detection \citep{cho2021data}.
We motivate the use of scale-dependent TAVC in~\eqref{eq:tavc} in combination with such multiscale methods in the following example.
\begin{example} \label{ex:one}
Consider an MA($1$) process $\varepsilon_t = W_t - 0.9 W_{t-1}$,
where $\{W_t\}_{t \in \mathbb{Z}}$ is a white noise process with $\Var(W_t) = 1$.
Then, we have $\sigma^2 = 0.01$, while the TAVC at scales $L = 50$ and $L = 100$
are $\sigma_{50}^2 = 0.1$ and $\sigma^2_{100} = 0.055$ respectively, which are several factors larger than the LRV.
On the other hand, consider an AR($1$) process $\varepsilon_t = 0.9 \varepsilon_{t-1} + W_{t}$, with $W_t$ defined as above. Then, $\sigma^2 =100$ while $\sigma_{50}^2 \approx 81.16$
and $\sigma^2_{100} \approx 90.53$, both of which are smaller than the LRV.
\end{example}
This demonstrates that adopting the global LRV may fail to reflect the degree of variability in the local data sections that are involved in multiscale algorithms,
which in turn may result in false negatives (failure in detecting genuine changes)
or false positives (detection of spurious estimators).
Moreover, when the LRV is close to zero as in the first instance of Example~\ref{ex:one}, some estimators of the LRV have been observed to take negative values \citep{huvskova2010note}, which further makes their use in change point problems undesirable.
To ensure that the scale-dependent TAVC estimator is robust to the mean shifts, we adopt the robust $M$-estimation framework of \cite{catoni2012challenging}, which was first proposed in the independent setting for mean and variance estimation and further extended to the serially dependent setting for LRV estimation in \cite{chen2021inference}.
We establish the consistency of the proposed robust estimator of scale-dependent TAVC under general conditions permitting heavy tails and serial dependence decaying at a polynomial rate.
Then, we discuss its application with multiscale change point detection methods such as the MOSUM procedure and the WBS algorithm,
and provide a heuristic approach to accommodate local stationarity in the data.
The remainder of the article is organised as follows.
Section~\ref{sec:tavc} introduces the scale-dependent TAVC and
its robust estimator and establishes its consistency.
Section~\ref{sec:multiscale} discusses its application with multiscale data segmentation algorithms and an extension to local stationarity.
In Section~\ref{sec:num}, we examine the performance of the proposed methodology
on simulated data sets and two real data examples on house price index and air quality.
Section~\ref{sec:conc} concludes the paper.
All proofs, algorithmic descriptions of multiscale change point methods
and additional numerical results are given in the appendix.
Accompanying \verb!R! software implementing the methodology is available from \url{https://github.com/EuanMcGonigle/TAVC.seg}.
\section{Scale-dependent TAVC and its robust estimation}
\label{sec:tavc}
\subsection{Multiscale change point detection in the mean}
We consider the problem of multiple change point detection
under the following model:
\begin{align}
\label{eq:model}
X_{t} = f_{t} + \varepsilon_{t} = f_{0} + \sum_{i=1}^{q} \mu_{i} \cdot \mathbb{I}(t \geq \tau_{i}+1) + \varepsilon_{t}, \quad t = 1, \ldots, n.
\end{align}
Under Model~\eqref{eq:model}, the piecewise constant signal $f_{t}$ contains $q$ change points at locations $\tau_{i}$, $i = 1, \ldots, q$, with the notational convention that $\tau_{0}=0$ and $\tau_{q+1} = n$.
The errors $\{ \varepsilon_{t} \}_{t=1}^{n}$ are assumed to be a (weakly) stationary time series satisfying $\mathbb{E} (\varepsilon_{t}) = 0$
with finite LRV $\sigma^2 = \lim_{N \to \infty} \Var(N^{-1/2} \sum_{t = 1}^N \varepsilon_t) \in (0, \infty)$,
and are permitted to be serially correlated and heavy-tailed
(see Assumption~\ref{assum:noise} below for details).
Our aim is to consistently estimate the total number and the locations of the change points.
A common approach to this problem is closely related to the change point testing literature,
which scans the data for the detection and estimation of multiple change points
by {\it locally} applying a test well-suited for detecting a single change.
Such a procedure typically involves comparing a test statistic of the form
$\widehat{\sigma}_{s, e}^{-1} \vert \mathcal{T}_{s, k, e} \vert$ to a threshold, say $D$.
Here,
\begin{align}
\label{eq:cusum-stat}
\mathcal{T}_{s, k, e} = \sqrt{ \frac{(k-s)(e-k)}{e-s}} \left( \frac{1}{k-s} \sum_{t=s+1}^{k} X_{t} - \frac{1}{e-k} \sum_{t=k+1}^{e} X_{t} \right)
\end{align}
denotes a change point detector evaluated at some locations
$0 \le s < k < e \le n$, which are determined in a method-specific way
(see Section~\ref{sec:multiscale} below),
$\sigma_{s, e}^2$ denotes a measure of variability
in the data section $\{X_t\}_{t = s + 1}^e$, and
$\widehat\sigma_{s, e}^2$ its estimator.
Under the stationarity assumption on $\{\varepsilon_t\}_{t = 1}^n$,
a natural choice is $\sigma^2_{s, e} = \sigma^2_{e - s}$,
the scale-dependent TAVC defined in~\eqref{eq:tavc} with $L = e - s$ as the scale.
Then, if $\{X_t\}_{t = s + 1}^e$
does not contain any change point well within the interval, we expect
$\widehat{\sigma}_{s, e}^{-1} \vert \mathcal{T}_{s, k, e} \vert \le D$,
while it signals the presence of such a change point
when $\widehat{\sigma}_{s, e}^{-1} \vert \mathcal{T}_{s, k, e} \vert > D$.
The key challenge lies in separating genuine mean shifts from the natural fluctuations due to serial correlations,
which requires a careful selection of the estimator $\widehat\sigma_{s, e}^2$
that correctly captures the variability in the section of the data under consideration.
It is well-documented that multiscale application of such a test
on data sections of varying lengths,
improves the adaptivity of the change point methodology to detect both large, frequent changes
and small changes over long stretches of stationarity \citep{cho2021data}.
For such a multiscale procedure,
Example~\ref{ex:one} demonstrates the potential pitfalls associated with
using an estimator of the global LRV in place of $\widehat\sigma_{s, e}^2$,
regardless of the length of the interval on which the detector statistic in~\eqref{eq:cusum-stat} is computed.
In the next section, we propose an estimator of
the scale-dependent TAVC $\sigma^2_L$ in~\eqref{eq:tavc}
that is robust to the presence of multiple mean shifts,
for the standardisation of multiscale change point detectors.
\subsection{Robust estimation of multiscale TAVC}
\label{sec:robust}
For notational convenience, suppose that $L$ is an even number, and let $G = L/2$ denote the block size. Then, for some starting point $b \in \{0, \ldots, G - 1\}$ with number of blocks $N_1(b) = \lfloor (n-b-G)/G \rfloor$, we define
\begin{align*}
\bar{X}_{j, b} = \frac{1}{G} \sum_{t = jG + b + 1}^{(j+1)G + b} X_t,
\quad \text{and} \quad
\xi_{j, b} = \frac{G(\bar{X}_{j, b} - \bar{X}_{j-1, b})^2}{2}
\text{ for } j = 1, \ldots, N_1(b).
\end{align*}
Analogously, we define
\begin{align*}
\bar{\varepsilon}_{j, b} = \frac{1}{G} \sum_{t=jG + b + 1}^{(j+1)G + b} \varepsilon_t
\quad \text{and} \quad
\widetilde\xi_{j, b} = \frac{G(\bar{\varepsilon}_{j, b} - \bar{\varepsilon}_{j-1, b})^2}{2}.
\end{align*}
Then, the following sum
\begin{align}\label{no-change-est}
\widehat{\widetilde\sigma}^{2}_L = \frac{1}{N_1(b)} \sum_{j=1}^{N_1(b)} \widetilde\xi_{j, b},
\end{align}
takes into account the temporal dependence in the local data sections of length $L = 2G$.
Further, we have
$\E(\widehat{\widetilde\sigma}^2_L) - \sigma^2_L = o(1)$ for $L \to \infty$
(see Theorem~\ref{thm:one} below),
such that $\widehat{\widetilde\sigma}^2_L$ is indicative of the level of variability $\sigma^2_L$ albeit being inaccessible (as it is defined with $\bar{\varepsilon}_{j, b}$ in place of $\bar{X}_{j, b}$).
Its accessible counterpart, $N_1(b)^{-1} \sum_{j = 1}^{N_1(b)} \xi_{j, b}$, on the other hand, is typically biased due to the mean shifts
and thus is inappropriate as an estimator of the scale-dependent TAVC.
To obtain an estimator that is robust to multiple mean shifts,
we adopt the robust $M$-estimation framework of \cite{catoni2012challenging}.
Let $\phi$ denote a non-decreasing influence function as
\begin{align}
\label{eq:envelope}
\phi(x) = \begin{cases}
- \log(2) & \text{for } x \leq -1 , \\
\log(1 + x + x^{2}/2 ) & \text{for } -1 \leq x \leq 0, \\
- \log( 1 - x + x^{2}/2 ) & \text{for } 0 \leq x \leq 1, \\
\log(2) & \text{for } x \geq 1.
\end{cases}
\end{align}
The robust estimator of the TAVC at scale $L$ and starting point $b$, denoted $\widehat{\sigma}^2_{L, b}$, is defined as the solution of the $M$-estimation equation
\begin{align}\label{eq:m:est}
h_{L, b} (u) = \frac{1}{N_1(b)} \sum_{j = 1}^{N_1(b)} \phi_v\l( \xi_{j, b} - u \r) = 0,
\end{align}
where $\phi_{v}(x) = v^{-1}\phi(vx)$ for some $v > 0$;
we specify the condition on $v$ later.
If there are multiple solutions to Equation~\eqref{eq:m:est}, any of them may be chosen.
\subsection{Theoretical properties}
We establish the consistency of the scale-dependent TAVC estimator
under the following assumption on the error process $\{ \varepsilon_t \}_{t=1}^n$.
\begin{assumption}
\label{assum:noise}
\hfill
\begin{enumerate}[label = (\roman*)]
\item \label{assum:noise:one}
We assume that $\varepsilon_t = \sum_{k=0}^\infty {a}_k {\eta}_{t-k}$,
where $\{\eta_t\}_{t \in \mathbb{Z}}$ is a sequence of i.i.d.\ random variables
and $\vert {a}_k \vert \leq \Xi (k + 1)^{-\beta}$ for some constants $\beta > 2.5$ and $\Xi > 0$ for all $k \ge 0$.
\item \label{assum:noise:two}
There exists a fixed constant $c_\sigma > 0$ such that
$\sigma^2 = ( \sum_{k\geq 0} a_k )^2$ satisfies $\sigma^2 \ge c_{\sigma}$.
\item \label{assum:noise:three} We operate under {\it either} of the following two conditions on the distribution of $\{\eta_t\}_{t \in \mathbb{Z}}$.
\begin{enumerate}[label = (\alph*)]
\item \label{cond:moment} There exists a fixed constant $r > 4$ such that
$\Vert \eta_{1} \Vert_r = ( \mathbb{E} ( |\eta_{1} |^r) )^{1/r} <\infty$.
\item \label{cond:subexp} There exist fixed constants $C_\eta > 0$ and $\kappa \ge 0$ such that
$\Vert \eta_1 \Vert_r \le C_\eta r^\kappa$ for all $r \ge 1$.
\end{enumerate}
\end{enumerate}
\end{assumption}
Assumption~\ref{assum:noise}~\ref{assum:noise:one} is a mild condition that permits the temporal dependence to decay at a polynomial rate, and~\ref{assum:noise:two} is made to ensure that the LRV is well-defined.
Assumption~\ref{assum:noise}~\ref{assum:noise:three}~\ref{cond:moment} allows heavy-tailed $\{\varepsilon_t\}_{t = 1}^n$,
while~\ref{cond:subexp} assumes a stronger condition
that requires sub-Weibull \citep{wong2020lasso} tail behaviour on $\{\eta_t\}_{t \in \mathbb{Z}}$ which includes sub-Gaussian ($\kappa = 1/2$) and sub-exponential ($\kappa = 1$)
distributions as special cases.
The following theorem establishes the consistency
of the estimator of the scale-dependent TAVC,
see Appendix~\ref{sec:proof} for the proof.
\begin{theorem}\label{thm:one}
Suppose that Assumption~\ref{assum:noise} holds, and define
\begin{align*}
\widetilde{\sigma}^2_L = \Var \left( \frac{1}{\sqrt{L}} \left( \sum_{t=1}^{G} \varepsilon_t - \sum_{t=G+1}^L \varepsilon_t \right) \right)
\end{align*}
for $L = 2G$.
Then, provided that $v \asymp \sqrt{G/n}$,
the estimator $\widehat{\sigma}^{2}_{L, b}$ satisfies
\begin{align}
\left| \widehat{\sigma}^{2}_{L, b} - \widetilde{\sigma}^{2}_L \right| &= \mathcal{O}_P\l( \sqrt{ \frac{Lq}{n} } + \max\l\{ \l( \frac{L}{n} \r)^{\frac{r-2}{r+2}}, \sqrt{\frac{L \log(n)}{n}} \r\} \r) \label{eq:thm:one}
\end{align}
under Assumption~\ref{assum:noise}~\ref{assum:noise:three}~\ref{cond:moment}, and
\begin{align}
\left| \widehat{\sigma}^{2}_{L, b} - \widetilde{\sigma}^{2}_L \right| &= \mathcal{O}_{P} \left( \sqrt{ \frac{Lq}{n} } + {\sqrt{ \frac{L \log^{4\kappa + 3} (n)}{n} }}\right)
\label{eq:thm:two}
\end{align}
under Assumption~\ref{assum:noise}~\ref{assum:noise:three}~\ref{cond:subexp}, for any fixed $b \in \{0, \ldots, G - 1\}$.
In addition, $\widetilde{\sigma}^2_L$ satisfies
\begin{align}
\left| \widetilde{\sigma}^{2}_{L} - {\sigma}_L^{2} \right| = \mathcal{O}(L^{-1}).
\label{eq:thm:three}
\end{align}
\end{theorem}
Theorem~\ref{thm:one} shows that the proposed robust estimator consistently estimates the TAVC at scale $L$.
The estimation error is decomposed into the error from approximating $\sigma_L^2$ by $\widetilde\sigma_L^2$ in~\eqref{eq:thm:three}, and
that in estimating $\widetilde{\sigma}_L^2$ by $\widehat{\sigma}_{L,b}^2$.
In deriving the second error, we make explicit the influence of multiple mean shifts on the estimator by the term $(Lq/n)^{1/2}$ in~\eqref{eq:thm:one}--\eqref{eq:thm:two}, as well as the effect of the innovation distribution in the remaining terms.
A careful examination of the proof of Theorem~\ref{thm:one} shows that
\begin{align}
\left| \widetilde{\sigma}^{2}_{L} - {\sigma}^{2} \right| &= \mathcal{O}(L^{-1}),
\label{sigma-approx2}
\end{align}
i.e.\ as $L$ increases, the scale-$L$ TAVC approximates the global LRV as expected.
\begin{remark}[Maximum time-scale for TAVC estimation]
\label{rem:max}
The error due to approximating $\sigma_L^2$ with $\widetilde{\sigma}_L^2$
decreases with $L$ as in~\eqref{sigma-approx2}.
On the other hand, the error of estimating $\widetilde{\sigma}_L^2$ with $\widehat{\sigma}_{L, b}^2$ increases with $L$ as in~\eqref{eq:thm:one}--\eqref{eq:thm:two};
this is attributed to the effect of mean shifts that grows with $L$,
and the decrease in the number of available blocks.
To balance between the two,
we suggest setting a maximum time-scale, say $M$, to be used in combination with a multiscale change point detection algorithm.
That is, when the change point detector $\mathcal T_{s, k, e}$ involves
$e - s \le M$, we scale the detector with the estimator of $\sigma^2_{s - e}$, the TAVC at the corresponding scale $L = e - s$.
On the other hand, if $e - s > M$, we propose to scale the detector
with the estimator of $\sigma^2_{M}$, the TAVC at the maximum time-scale $M$,
which satisfies $\vert \sigma^2_{e - s} - \sigma^2_M \vert = \mathcal{O} (M^{-1})$.
\end{remark}
\section{Applications and extensions}
\label{sec:multiscale}
We now describe explicitly how the robust estimator of the scale-dependent TAVC proposed in Section~\ref{sec:tavc}, is applied within the algorithms
that scan moving sum (MOSUM) and cumulative sum (CUSUM) statistics of the form~\eqref{eq:cusum-stat}, for multiple change point detection.
\subsection{MOSUM procedure}
\label{sec:mosum}
The MOSUM procedure \citep{chu1995mosum,eichinger2018mosum}
evaluates the change point detector $\mathcal T_{s, k, e}$ over a moving window.
For a given bandwidth $G$, the MOSUM detector for a change in mean at time point $k$ is given by
\begin{align}
\label{eq:mosum}
\mathcal T_{G}(k) = \mathcal{T}_{k-G, k, k+G} = \sqrt{\frac{G}{2}} \left( \frac{1}{G} \sum_{t=k+1}^{k+G} X_{t} - \frac{1}{G} \sum_{t=k-G+1}^{k} X_{t} \right), \quad G \le k \le n - G.
\end{align}
\cite{eichinger2018mosum} propose to estimate the total number and the locations of multiple change points by
identifying all significant local maximisers of $\mathcal T_G(k)$, say $\widehat k$, satisfying
\begin{align}
\label{eq:mosum:est}
\vert \mathcal T_G(\widehat k) \vert > \widehat\sigma \cdot D_n(G; \alpha) \quad \text{and} \quad
\widehat k = {\arg\max}_{k: \, \vert k - \widehat k \vert \le \eta G} \vert \mathcal T_G(k) \vert
\end{align}
for some $\eta \in (0, 1)$.
Here, $D_n(G; \alpha)$ denotes a critical value at a significance level $\alpha \in (0, 1)$,
which is drawn from the asymptotic null distribution of the MOSUM test statistic $\max_{G \le k \le n - G} \sigma^{-1} \vert T_G(k) \vert$ obtained
under mild conditions permitting heavy-tailedness and serial dependence, and
takes the form
\begin{align*}
D_{n} (G, \alpha) = \frac{b_{G,n} + c_{\alpha}}{a_{G,n}} \quad \text{with} \quad
c_{\alpha} = - \log \log \left( \frac{1}{\sqrt{1-\alpha}} \right),
\end{align*}
where $a_{G,n}$ and $b_{G,n}$ are known constants depending on $G$ and $n$ only. The single-bandwidth MOSUM procedure achieves consistency in estimating the total number and the locations of multiple change points,
provided that $\min_{1 \le i \le q} \mu_i^2 G \to \infty$ as $n \to \infty$ sufficiently fast while $2G \le \min_{0 \le i \le q} (\tau_{i + 1} - \tau_i)$, see Theorem~3.2 of \cite{eichinger2018mosum} and Corollary D.2 of \cite{cho2021two} for explicit conditions.
The requirement on $G$ indicates that the single-scale MOSUM procedure performs best with the bandwidth chosen as large as possible while avoiding situations where there are more than one change point within the moving window at any time. Consequently, it lacks adaptivity when the data sequence contains both large changes over short intervals and small changes over long intervals.
Multiscale extensions of the single-bandwidth MOSUM procedure,
i.e.\ applying the MOSUM procedure with a range of bandwidths
and then combining the results, alleviate the difficulties involved in bandwidth selection and provide adaptivity.
In this paper, we consider the multiscale MOSUM procedure
combined with the `bottom-up' merging as proposed by \cite{messer2014multiple} (see also \cite{meier2021mosum}).
Denoting the range of bandwidths by $\mathcal G = \{G_h, \, 1 \le h \le H: \,
G_1 < \ldots < G_H\}$,
let $\mathcal C(G)$ denote the set of estimators detected with some bandwidth~$G \in \mathcal G$.
Then, we accept all estimators in $\mathcal C(G_1)$ returned with the finest bandwidth $G_1$ to the set of final estimators $\mathcal C$
and, sequentially for $h = 2, \ldots, H$, accept $\widehat k \in \mathcal C(G_h)$ if and only if
$\min_{k \in \mathcal C} \vert \widehat k - k \vert > \eta G_h$
(with $\eta$ identical to that in~\eqref{eq:mosum:est}). That is, we only accept the estimators that do not detect the change points which have previously been detected at a finer scale.
We propose to apply the multiscale MOSUM procedure with bottom-up merging,
in combination with the robust estimator of multiscale TAVC as follows.
For each $G_h \in \mathcal G$, the TAVC at scale $L = 2 G_h$ is estimated
by $\widehat{\sigma}^2_{2 G_h}$ solving~\eqref{eq:m:est}, provided that $2G_h \le M$.
Here, $M$ denotes the maximum scale which is set in relation to the sample size $n$, see Remark~\ref{rem:max}.
Then, we use $\widehat{\sigma}_{2 G_h}$ in place of the global estimator $\widehat\sigma$
in~\eqref{eq:mosum:est} to standardise the MOSUM detector $\mathcal T_{G_h}(k)$.
When $2G_h > M$, we use $\widehat{\sigma}_M$ in place of $\widehat\sigma_{2G_h}$
for MOSUM detector standardisation.
In doing so, we ensure that change point detectors at multiple scales are standardised
by the scale-dependent TAVC that accurately reflects the degree of variability over the moving window, while taking into account the presence of possibly multiple mean shifts therein.
We refer to Algorithm~\ref{alg:mosum} in Appendix~\ref{appendix:alg} for the pseudocode of the multiscale MOSUM procedure with the robust estimator of scale-dependent TAVC.
\subsection{Wild binary segmentation}
\label{sec:wbs}
The binary segmentation algorithm \citep{scott1974cluster, vostrikova1981} and its extensions, such as wild binary segmentation (WBS, \citeauthor{fryzlewicz2014wild}, \citeyear{fryzlewicz2014wild}; \citeyear{fryzlewicz2020detecting}) and seeded binary segmentation \citep{kovcs2020seeded}, recursively search for multiple change points using the CUSUM statistic of the form~\eqref{eq:cusum-stat}, with $s$ and $e$ that are identified iteratively.
These methods have primarily been analysed for the data segmentation problem under~\eqref{eq:model}
assuming i.i.d.\ Gaussianity on the $\{\varepsilon_t\}_{t = 1}^n$.
Consequently, some robust estimators of $\Var(\varepsilon_1)$ have been considered for standardising the CUSUM statistic.
Here, we discuss the application of the WBS2 algorithm \citep{fryzlewicz2020detecting} in the time series setting with the proposed robust estimator of the scale-dependent TAVC.
Let $\mathcal A_{s, e} = \{(\ell, r) \in \mathbb{Z}^2:\, s \le \ell < r \le e, \ r - \ell > 1\}$
denote the collection of all intervals within $\{s + 1, \ldots, e\}$ for some $0 \le s < e \le n$, and $\mathcal R_{s, e}$ denote its subset selected either randomly or deterministically
(see \cite{cho2021multiple} for one approach to deterministic grid selection)
with $\vert \mathcal R_{s, e} \vert = \min(R, \vert \mathcal A_{s, e} \vert)$ for some given $R \le n(n - 1)/2$.
Starting with $(s, e) = (0, n)$, we identify
\begin{align}
\label{eq:wbs:est}
(s_\circ, k_\circ, e_\circ) = {\argmax}_{\substack{(\ell, k, r): \, \ell < k < r \\ (\ell, r) \in \mathcal R_{s, e}}} \frac{\left\vert \mathcal T_{\ell, k, r} \right\vert}{\widehat{\sigma}_{r - \ell}}
\quad \text{with} \quad
\left\vert \mathcal T_{s_\circ, k_\circ, e_\circ} \right\vert >
\widehat{\sigma}_{e_\circ - s_\circ} \cdot D
\end{align}
for some threshold $D$ and $\widehat{\sigma}^2_{r - \ell}$ denoting the proposed robust estimator of the TAVC at scale $L = r - \ell$ (when $r - \ell$ is odd, we use $\widehat{\sigma}^2_{r - \ell - 1}$ instead).
As in Section~\ref{sec:mosum}, a maximum scale $M$ is set so that the CUSUM statistic over any interval of length greater than $M$ is standardised using $\widehat{\sigma}_{M}$.
Following the recommendation made in \cite{fryzlewicz2014wild}, we adopt the threshold
$D = C \sqrt{2 \log(n)}$ where $C$ is a universal constant.
If $(s_\circ, k_\circ, e_\circ)$ that fulfils~\eqref{eq:wbs:est} exists,
it signals the presence of a change point so that
the data is partitioned into $\{X_t\}_{t = s + 1}^{k_\circ}$ and $\{X_t\}_{t = k_\circ + 1}^e$, and the same step of detecting and identifying a single change point
is repeated on each partition separately.
If no such $(s_\circ, k_\circ, e_\circ)$ exists, or when the user-specified minimum segment length is reached,
then the search for change points
is terminated on $\{X_t\}_{t = s + 1}^e$.
We provide a pseudocode of the WBS2 algorithm with the robust estimator of scale-dependent TAVC in Algorithm~\ref{alg:wbs2} of Appendix~\ref{appendix:alg}.
\subsection{Extension to local stationarity}
\label{sec:nonstat:ext}
We propose a heuristic extension of the robust estimator of scale-dependent TAVC to the setting where the second-order structure of $\{\varepsilon_t\}_{t = 1}^n$ varies smoothly over time.
Suppose that there exists an appropriately chosen window size $W$ such that $\{\varepsilon_t\}_{t = k - \lfloor W/2 \rfloor + 1}^{k + \lfloor W/2 \rfloor}$ may be regarded as being approximately second-order stationary over all $k$. Then, we propose to perform the robust estimation described in Section~\ref{sec:robust} in a localised fashion.
To this end, define the time-varying TAVC at scale $L$ and time $k$ by
\begin{align}
\label{eq:tavc-tvar}
\sigma^2_L (k) = \Var\l( \frac{1}{\sqrt L} \sum_{t = k-G+1}^{k+G} \varepsilon_t \r) .
\end{align}
For notational convenience, we set $W = N_2 L$ for some integer $N_2$, and let $N_3 = \lfloor (W-G)/G \rfloor$ be the number of blocks for window size $W$. For $k \in \{ W/2, \ldots, n - W/2 \}$, we estimate $\sigma^2_L(k)$ by $\widehat{\sigma}^2_L(k)$, the solution of the following $M$-estimation equation
\begin{align}\label{m-est-eq2}
h_L (u, k) = \frac{1}{N_3} \sum_{j = 1}^{N_3} \phi_v \left( \xi_{j, k - W/2} - u \right) = 0
\end{align}
with $v \asymp (G/W)^{1/2}$.
We apply a boundary extension so that $\widehat{\sigma}_L^2(1) = \ldots = \widehat{\sigma}_L^2(W/2)$ and $\widehat{\sigma}_{L}^2(n -W/2) = \ldots = \widehat{\sigma}_L^2(n)$.
The estimator of the local TAVC at time $k$ and scale $L$ is obtained analogously as that of the global TAVC at scale $L$ described in Section~\ref{sec:robust}, except that we only use the windowed data region starting at time $k - W/2$ and ending at $k + W/2$ for the estimation of the former.
Then, the MOSUM detector $\mathcal T_G(k)$ and the CUSUM statistic $\mathcal T_{s, k, e}$ described in Sections~\ref{sec:mosum}--\ref{sec:wbs} are standardised in a time-dependent way
using $\widehat{\sigma}_{2G}(k)$ and $\widehat{\sigma}^2_{(e-s)}(k)$, respectively.
In practice, we observe that taking the running median of $\{\widehat{\sigma}^2_L(k + b - \lfloor G/2 \rfloor )\}_{b = 0}^{G - 1}$ as an estimator of $\sigma^2_L(k)$ further improves the performance, as it `smoothes' out the local estimators and enhances the robustness to mean changes.
We illustrate the benefit of adopting the time-varying adaptation of the proposed robust estimator using the following example.
Consider a time series of length $n=1000$, where the errors $\{ \varepsilon_t\}_{t \in \mathbb{Z}}$ follow a time-varying AR(1) model: $\varepsilon_t = a_1 (t) \varepsilon_{t-1} + W_t$, with $a_{1} (t) = 0$ for $t\leq 500$, $a_{1} (t) = 0.7$ for $t \ge 501$, and $W_t \sim_{\text{i.i.d.}} \mathcal N(0, 1)$. There are two changes in the mean at $\tau_1 = 300$ and $\tau_2 = 700$, with change sizes~$1$ and~$2$, respectively.
In Figure~\ref{fig:toy:ex}, we show the MOSUM detector statistic in~\eqref{eq:mosum} calculated using bandwidth $G = 100$.
We also plot the threshold $D_n(G, \alpha)$ at the significance level $\alpha = 0.05$, multiplied by the square root of the global TAVC estimator at scale $L = 2G$ (i.e.\ $\widehat\sigma^2_{2G}$) in dashed blue line,
and that multiplied by the square root of the local estimators of the scale-$L$ TAVC (i.e.\ $\widehat\sigma^2_{2G}(k)$) in solid red line. We see that using the global approach misses the change at time $\tau_1 = 300$ due to the global scale-$L$ TAVC estimator being too large, whilst the localised approach successfully detects both changes.
Lastly, we mention that the robust estimation of time-varying and scale-dependent TAVC is of independent interest beyond the context of change point analysis, with possible extensions including the estimation of other second-order properties. For example, the procedure can be used to obtain a robust estimator of the spectrum of a locally stationary wavelet process \citep{nason2000wavelet} while the time series undergoes shifts in the mean.
\begin{figure}[H]
\centering
\includegraphics[width = 0.8\textwidth]{final-tvar-ex.pdf}
\caption{MOSUM detector statistic (black solid line) computed from the time series generated as described in the text with two changes in the mean (denoted by vertical lines). Also shown are the thresholds computed using global (blue dashed line) and local (red solid line) estimators of the scale-dependent TAVC.}
\label{fig:toy:ex}
\end{figure}
\section{Numerical results}
\label{sec:num}
\subsection{Practical considerations}
\label{sec:tuning}
We provide guidance on the selection of tuning parameters
required for the proposed robust TAVC estimator and its application
with multiscale change point detection methods.
\vspace{-10pt}
\paragraph{Parameter $v$ in~\eqref{eq:m:est}.}
We select $v = v_b = \bar{\xi}_b^{-1}\sqrt{G/n}$ where $\bar{\xi}_b$ is a fixed constant for a given $b \in \{0, \ldots, G - 1\}$.
For the problem of robust mean estimation, \cite{catoni2012challenging} recommends the standard deviation in the place of $\bar{\xi}_b$
and in a similar vein, \cite{chen2021inference} propose to use a trimmed mean
\begin{align}
\label{eq:v:one}
\bar{\xi}_{[1], b} = {\frac{1}{\lfloor 3N_1(b)/4 \rfloor - \lceil N_1(b)/4 \rceil + 1}} \sum_{j = \lceil N_1(b)/4 \rceil}^{ \lfloor 3N_1(b)/4 \rfloor} \xi_{(j), b},
\end{align}
where $\xi_{(1), b} \le \ldots \le \xi_{(N_1(b)), b}$ are the ordered $\{\xi_{j, b} \}_{j=1}^{N_1(b)}$. Another approach is to use an appropriately scaled median of $\{\xi_{j, b} \}_{j=1}^{N_1(b)}$, in a similar fashion to \cite{mcgonigle2021detecting}. In this case, the necessary scaling constant to ensure unbiasedness can be chosen by noting that $\bar{X}_{j, b}$ is asymptotically Gaussian as $L \to \infty$, and thus $\xi_{j, b}$ is asymptotically scaled $\chi^2_1$. This leads to the choice
\begin{align}
\label{eq:v:two}
\bar{\xi}_{[2], b} = K \cdot \text{Median}\l\{ \xi_{1,b}, \ldots , \xi_{N_1(b),b} \r\},
\end{align}
where $K = 2.125$. In simulation studies, we report the results obtained with $\bar{\xi}_{[\ell], b}$, $\ell = 1, 2$, in setting the parameter $v_b$ where we observe that both choices return similarly good results.
\vspace{-10pt}
\paragraph{Parameters $b$ and $b_{\max}$ in~\eqref{eq:m:est}.}
As a further step to ensure greater robustness of the estimator, we obtain the estimator $\widehat\sigma^2_{L, b}$ for a range of $b \in \{0, \ldots, b_{\max}\}$ and take their median as the final estimator $\widehat\sigma^2_L$. Informally, we may get unlucky with some starting value $b$ that leads to many of the $\xi_{j, b}$ contaminated by the mean changes, and taking the median over a range of values of $b$ helps in alleviating this. We take $b_{\max} = G - 1$ in practice which yields good performance.
\vspace{-10pt}
\paragraph{Maximum time-scale $M$.}
We recommend $M = \lfloor 2.5 \sqrt{n} \rfloor$ as the coarsest scale at which the scale-dependent TAVC is estimated. This choice is made to balance between mitigating the effect of change points, and ensuring that the TAVC at coarser scales is well-approximated by $\sigma^2_M$.
\vspace{-10pt}
\paragraph{Tuning parameters for the multiscale MOSUM procedure.}
We follow \cite{cho2021two} and generate $\mathcal G$ as a sequence of Fibonacci numbers. For the simulation studies reported in Section~\ref{sec:sim} and Appendix~\ref{sec:add:sim}, we consider $\mathcal G = \{G_m, \, 1 \le m \le 4: \, G_1 < \ldots < G_4\}$ where $G_m = G_{m - 1} + G_{m - 2}$ for $m \ge 2$ with
$G_0 = G_1 = 20 + 10 \lfloor n/1000 \rfloor$. For other tuning parameters, we adopt the recommended default values of the \verb!R! package \verb!mosum! \citep{meier2021mosum}, and set $\alpha = 0.05$ and $\eta = 0.4$.
\vspace{-10pt}
\paragraph{Tuning parameters for the WBS2 algorithm.}
For the threshold, we set the constant $C = 1.3$ and draw $R = 100$ deterministic intervals at each iteration, which are suggested choices in \cite{fryzlewicz2014wild} and \cite{cho2021multiple} respectively.
As we permit the presence of serial correlations,
it is reasonable to impose a minimum length requirement
on the intervals considered in the WBS2 algorithm.
We set this minimum length to be $2 G_1$, with $G_1$ the finest scale considered by the MOSUM procedure.
\vspace{-10pt}
\paragraph{Window size $W$ for time-varying TAVC estimation.}
We utilise a scale-dependent window size $W_{L}$. Setting $W_{L}= N_2 L$ gives $N_3 = 2N_2-1$ data points used in the solving of the $M$-estimation equations. We advise setting $N_2 = 5$, which ensures that the influence of change points is negated and that the window size is large enough to include enough data points for reliable estimation of the TAVC.
\subsection{Simulation study}
\label{sec:sim}
In this section, we evaluate the performance of the proposed robust estimator of scale-dependent TAVC applied with the two multiscale change point detection procedures discussed in Sections~\ref{sec:mosum}--\ref{sec:wbs}. We compare with other methods that account for serial dependence under~\eqref{eq:model} and whose implementations are readily available in \verb!R!,
with a variety of scenarios for generating serially correlated $\{\varepsilon_t\}_{t \in \mathbb{Z}}$.
\subsubsection{Settings}
We assess the performance of different methods both in the case of no changes ($q = 0$) and multiple changes ($q \ge 1$), under a variety of error structures.
Unless stated otherwise, we generate
$W_t \sim_{\text{i.i.d.}} \mathcal N(0, \sigma_w^2)$ with $\sigma_w = 1$.
\begin{enumerate}[label = (M\arabic*)]
\item \label{model-a} $\varepsilon_t = W_t$.
\item \label{model-b} $\varepsilon_t = W_t$, where $W_t$ are i.i.d.\ $t_{5}$-distributed random variables.
\item \label{model-c} AR(1) model: $\varepsilon_t = a_1 \varepsilon_{t-1}+W_t$, with $a_1 = 0.9$ and $\sigma_w = \sqrt{1-a_1^2}$.
\item \label{model-d} AR(2) model: $\varepsilon_t = a_1 \varepsilon_{t-1}+ a_2 \varepsilon_{t-2}+W_t$, with $a_1 = 0.5$ and $a_2=0.3$, with $\sigma_w = 0.6676184$.
\item \label{model-e} MA(1) model: $\varepsilon_t = W_t + b_1 W_{t-1}$, with $b_1 = -0.9$.
\item \label{model-f} ARCH(1) model: $\varepsilon_t = \sigma_t W_t$ with $\sigma_t^2 = c_0 + c_1 \varepsilon_{t-1}^2$, where $c_0= 0.5$, $c_1 = 0.4$.
\item \label{model-g} Time-varying AR(1) model: $\varepsilon_t = a_1 (t) \varepsilon_{t-1}+W_t$, with $a_{1} (t) = 0.8-0.6t/n$.
\item \label{model-h} Time-varying AR(1) model: $\varepsilon_t = a_1 (t) \varepsilon_{t-1}+ \sigma (t) W_t$, with $a_{1} (t) = 0.5 \cos (2\pi t/n)$ and $\sigma (t) = \sqrt{1-a_1 (t)^2}$.
\item \label{model-i} Time-varying MA(1) model: $\varepsilon_t = W_t + b_1 (t) W_{t-1}$, with $b_1 (t) = 12(t/n)^3 -18(t/n)^2+6t/n$.
\end{enumerate}
Models~\ref{model-a}--\ref{model-f} represent stationary error scenarios.
Model~\ref{model-a} is the setting commonly adopted in the literature while \ref{model-b}, taken from \cite{cho2021data}, is adopted to examine whether a method works well in the presence of non-Gaussian errors.
Models~\ref{model-c} and~\ref{model-d} allow strong autocorrelations in $\{\varepsilon_t\}_{t = 1}^n$. Under Model~\ref{model-e}, the LRV is close to zero, which makes its accurate estimation difficult. Model~\ref{model-f} is a non-linear process. Models~\ref{model-g}--\ref{model-i} consider time-varying dependence structure; variants of \ref{model-g} and \ref{model-h} were studied in \cite{mcgonigle2021detecting} and \cite{cho2021multiple}, respectively.
For Models~\ref{model-a}--\ref{model-f}, we use the global scale-dependent TAVC estimator described in Section~\ref{sec:robust} while for~\ref{model-g}--\ref{model-i}, we use the window-based estimator of the local scale-dependent TAVC described in Section~\ref{sec:nonstat:ext}.
We assess the performance of the methods both when $q = 0$ and $q \ge 1$.
In the latter case, the time series contains the $q$ change points at $\tau_i = \lfloor n/(q + 1) \cdot i \rfloor, \, i = 1, \ldots, q$, with the (signed) change size $\mu_i = \mu(\tau_i) \cdot (-1)^{i + 1}$.
In~\ref{model-a}--\ref{model-d} and \ref{model-f}, we set $\mu(\tau_i) = \sigma$
and in the case of~\ref{model-e}, we set $\mu = 1$.
In~\ref{model-g}--\ref{model-i}, we set $\mu(\tau_i) = \sigma(\tau_i)$
where $\sigma^2(t)$ denotes the time-varying LRV.
We implement the robust TAVC estimation within both the multiscale MOSUM and WBS2 procedures as described in Sections~\ref{sec:mosum} and~\ref{sec:wbs},
which are referred to as MOSUM.TAVC$_{[\ell]}$ and WBS2.TAVC$_{[\ell]}$, respectively.
Here, the subscript with $\ell = 1, 2$, refers to the choice of the tuning parameter $\bar{\xi}_{[\ell], b}$ involved in the parameter $v$,
see~\eqref{eq:v:one}--\eqref{eq:v:two}.
For the choice of the tuning parameters, we refer to Section~\ref{sec:tuning}.
For illustrative purposes, we also report the results of the `oracle' versions of the MOSUM and WBS2 procedures referred to as MOSUM.oracle and WBS2.oracle, respectively.
These methods are implemented with the true LRV $\sigma^2$
($\sigma^2(t)$ in the case of Models~\ref{model-g}--\ref{model-i})
for standardising the detector statistics,
while all other tuning parameters are kept the same.
Additionally, we consider DepSMUCE \citep{dette2020multiscale}, DeCAFS \citep{romano2021detecting} and WCM.gSa \citep{cho2021multiple} for comparison.
DepSMUCE extends SMUCE \citep{frick2014multiscale} to dependent data using a difference-type estimator of the LRV. Although not its primary objective, DeCAFS detects multiple change points in the mean assuming that the noise is a stationary AR($1$) process. The WCM.gSa method performs model selection on the sequence of models generated by the WBS2 algorithm, using an information criterion-based model selection strategy which assumes that $\{\varepsilon_t\}_{t \in \mathbb{Z}}$ follows an AR model of an arbitrary order. For DepSMUCE, we consider significance levels $\alpha \in \{ 0.05, 0.2 \}$. Other tuning parameters not mentioned here are chosen as recommended by the authors.
\subsubsection{Results}
Table~\ref{table:sim:results} summarises the results of the comparative simulation study
from $1000$ replications of time series of length $n = 1000$
generated as in~\ref{model-a}--\ref{model-i} with $q \in \{0, 4\}$.
Results for other values of $n \in \{500, 2000\}$
are given in Appendix~\ref{sec:add:sim},
{where we make similar observations as below.}
When $q = 0$, we report the proportion of falsely detecting any change point out of the $1000$ realisations (see the column `Size' in Table~\ref{table:sim:results}).
When $q \ge 1$, we report the relative mean squared error (RMSE)
\begin{align*}
\sum_{t=1}^{n} (\widehat{f}_{t} - f_{t})^{2} / \sum_{t=1}^{n} (\widehat{f}^{*}_{t} - f_{t})^{2},
\end{align*}
where $\widehat{f}_{t}$ is the piecewise constant signal constructed with the estimated change points, and {$\widehat{f}^*_t$} is the oracle estimator constructed with the true change points.
We also report the distribution of the estimated number of change points, as well as the covering metric (CM). The covering metric \citep{arbelaez2010contour} measures the quality of the resulting segmentation as defined by the location of the detected changes, and is recommended in \cite{van2020evaluation} as an evaluation metric for comparing change point detection algorithms. The CM take values between $0$ and $1$, with a value of $1$ corresponding to a perfect segmentation. Its explicit definition can be found in Appendix~\ref{sec:add:sim}.
For each measure, we report its average over the $1000$ realisations.
\begingroup
{\small
\setlength{\tabcolsep}{3pt}
\setlength{\LTcapwidth}{\textwidth}
\begin{longtable}{c c c ccccc cc}
\caption{Performance comparisons for $n = 1000$. We report the size, the proportion of realisations where change points are falsely detected when $q = 0$,
and the distribution of the estimated number of change points, covering metric (CM) and relative MSE (RMSE) over 1000 realisations when $q = 4$.
The modal value of the distribution of the number of estimated change points for each method is shown in bold. The best performing method according to each metric when $q = 4$ is underlined.}
\label{table:sim:results}
\endfirsthead
\endhead
\toprule
&&& \multicolumn{5}{c}{$\widehat{q} - q$} & & \\
Model & Method & Size & $\leq-2$ & $-1$ & $\mathbf{0}$ & $1$ & $\geq 2$ & CM & RMSE \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-a} & MOSUM.TAVC$_{[1]}$ & 0.135 & 0.000 & 0.006 & \bf{0.980} & 0.014 & 0.000 & 0.967 & 6.098 \\
& MOSUM.TAVC$_{[2]}$ & 0.091 & 0.000 & 0.015 & \bf{0.978} & 0.007 & 0.000 & 0.965 & 6.234 \\
& WBS2.TAVC$_{[1]}$ & 0.049 & 0.000 & 0.004 & {\bf{0.996}} & 0.000 & 0.000 & {0.976} & {4.605}
\\
& WBS2.TAVC$_{[2]}$ & 0.028 & 0.001 & 0.017 & \bf{0.982} & 0.000 & 0.000 & 0.973 & 4.787
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.010 & 0.000 & 0.014 & \bf{0.986} & 0.000 & 0.000 & 0.970 & 5.053 \\
& DepSMUCE(0.2) & 0.066 & 0.000 & 0.001 & {\bf{0.998}} & 0.001 & 0.000 & 0.974 & 4.763 \\
& DeCAFS & 0.015 & 0.000 & 0.000 & \bf{0.970} & 0.029 & 0.001 & {0.976} & 4.798 \\
& WCM.gSa & 0.007 & 0.000 & 0.000 & \bf{0.978} & 0.020 & 0.002 & 0.975 & {4.733} \\
& MOSUM.oracle & 0.040 & 0.000 & 0.000 & \bf{0.861} & 0.132 & 0.007 & 0.965 & 5.782 \\
& WBS2.oracle & 0.004 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.977} & \underline{4.567} \\ \cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-b} & MOSUM.TAVC$_{[1]}$ & 0.149 & 0.001 & 0.004 & \bf{0.979} & 0.015 & 0.001 & 0.967 & 6.512 \\
& MOSUM.TAVC$_{[2]}$ & 0.086 & 0.003 & 0.008 & \bf{0.981} & 0.008 & 0.000 & 0.965 & 6.595 \\
& WBS2.TAVC$_{[1]}$ & 0.040 & 0.001 & 0.006 & {\bf{0.993}} & 0.000 & 0.000 & {0.976}
& {4.702} \\
& WBS2.TAVC$_{[2]}$ & 0.014 & 0.004 & 0.011 & \bf{0.985} & 0.000 & 0.000 & 0.974 & 4.899
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.586 & 0.000 & 0.004 & \bf{0.610} & 0.161 & 0.225 & 0.943 & 7.003 \\
& DepSMUCE(0.2)& 0.747 & 0.000 & 0.000 & \bf{0.420} & 0.180 & 0.400 & 0.932 & 6.666 \\
& DeCAFS & 0.898 & 0.000 & 0.000 & 0.105 & 0.040 & \bf{0.855} & 0.886 & 29.258 \\
& WCM.gSa & 0.009 & 0.000 & 0.000 & {\bf{0.978}} & 0.020 & 0.002 &{0.975} & {4.789} \\
& MOSUM.oracle & 0.037 & 0.000 & 0.000 & \bf{0.843} & 0.143 & 0.014 & 0.964 & 6.300 \\
& WBS2.oracle & 0.012 & 0.000 & 0.000 & \underline{\bf{0.996}} & 0.004 & 0.000 & \underline{0.978} & \underline{4.650} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-c} & MOSUM.TAVC$_{[1]}$ & 0.147 & 0.000 & 0.000 & \bf{0.998} & 0.002 & 0.000 & 0.995 & 1.810 \\
& MOSUM.TAVC$_{[2]}$ & 0.082 & 0.000 & 0.001 & \bf{0.999} & 0.000 & 0.000 & 0.994 & 1.789 \\
& WBS2.TAVC$_{[1]}$ & 0.062 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.998} & 1.258 \\
& WBS2.TAVC$_{[2]}$ & 0.034 & 0.000 & 0.001 & {\bf{0.999}} & 0.000 & 0.000 & \underline{0.998} & 1.264
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.968 & 0.000 & 0.000 & {\bf{0.920}} & 0.078 & 0.002 & 0.985 & 2.743 \\
& DepSMUCE(0.2) & 0.996 & 0.000 & 0.000 & \bf{0.739} & 0.232 & 0.029 & 0.971 & 3.123 \\
& DeCAFS & 0.597 & 0.000 & 0.000 & \bf{0.590} & 0.342 & 0.068 & {0.982} & \underline{1.236} \\
& WCM.gSa & 0.053 & 0.000 & 0.000 & \bf{0.731} & 0.135 & 0.134 & 0.959 & 2.086 \\
& MOSUM.oracle & 0.001 & 0.000 & 0.000 & \bf{0.891} & 0.097 & 0.012 & 0.988 & 2.012 \\
& WBS2.oracle & 0.000 & 0.000 & 0.000 & {\bf{0.979}} & 0.021 & 0.000 & 0.997 & 1.321 \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-d} & MOSUM.TAVC$_{[1]}$ & 0.123 & 0.000 & 0.003 & {\bf{0.992}} & 0.005 & 0.000 & {0.987} & 3.133 \\
& MOSUM.TAVC$_{[2]}$ & 0.073 & 0.001 & 0.004 & \bf{0.992} & 0.003 & 0.000 & 0.986 & 3.232 \\
& WBS2.TAVC$_{[1]}$ & 0.053 & 0.001 & 0.000 & {\bf{0.999}} & 0.000 & 0.000 & {0.994}
& {1.715} \\
& WBS2.TAVC$_{[2]}$ & 0.035 &0.003 & 0.002& \bf{0.995} & 0.000 & 0.000 & {0.993} & 1.755
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.678 & 0.000 & 0.000 & \bf{0.979} & 0.021 & 0.000 & {0.986} & 2.932 \\
& DepSMUCE(0.2) & 0.907 & 0.000 & 0.000 & \bf{0.891} & 0.105 & 0.004 & 0.980 & 3.183 \\
& DeCAFS & 0.734 & 0.000 & 0.000 & 0.241 & 0.224 & \bf{0.535} & 0.935 & {2.258} \\
& WCM.gSa & 0.022 & 0.000 & 0.000 & \bf{0.778} & 0.111 & 0.111 & 0.966 & 2.684 \\
& MOSUM.oracle & 0.008 & 0.000 & 0.000 & \bf{0.972} & 0.028 & 0.000 & 0.986 & 3.087 \\
& WBS2.oracle & 0.001 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.995} & \underline{1.705} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-e} & MOSUM.TAVC$_{[1]}$ & 0.120 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & 0.990 & 89.580 \\
& MOSUM.TAVC$_{[2]}$ & 0.069 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & 0.990 & 89.580 \\
& WBS2.TAVC$_{[1]}$ & 0.103 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.992} & 76.922 \\
& WBS2.TAVC$_{[2]}$ & 0.052 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.992} & 76.922
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05)& 0.998 & 0.000 & 0.000 & 0.036 & 0.048 & \bf{0.916} & 0.771 & 611.183 \\
& DepSMUCE(0.2) &1.000 & 0.000 & 0.000 & 0.003 & 0.004 & \bf{0.993} & 0.633 &1038.395 \\
& DeCAFS & 0.001 & 0.000 & 0.000 & \bf{0.997} & 0.003 & 0.000 & \underline{0.992} & 77.886 \\
& WCM.gSa & 0.000 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.992} & \underline{76.814} \\
& MOSUM.oracle & 1.000 & 0.000 & 0.000 & 0.000 & 0.000 & \bf{1.000} & 0.276 & 247.634 \\
& WBS2.oracle & 1.000 & 0.000 & 0.000 & 0.000 & 0.000 & \bf{1.000} & 0.278 & 207.555 \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-f} & MOSUM.TAVC$_{[1]}$ & 0.168 & 0.000 & 0.000 & \bf{0.978} & 0.021 & 0.001 & 0.973 & 6.246 \\
& MOSUM.TAVC$_{[2]}$ & 0.112 & 0.000 & 0.001 & {\bf{0.993}} & 0.006 & 0.000 & 0.973 & 6.246 \\
& WBS2.TAVC$_{[1]}$ & 0.064 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.981} & \underline{4.833} \\
& WBS2.TAVC$_{[2]}$ & 0.030 & 0.000 & 0.001 & {\bf{0.999}} & 0.000 & 0.000 & \underline{0.981} & {4.844}
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.507 & 0.000 & 0.000 & \bf{0.740} & 0.176 & 0.084 & 0.960 & 6.920 \\
& DepSMUCE(0.2) & 0.716 & 0.000 & 0.000 & \bf{0.564} & 0.252 & 0.184 & 0.947 & 7.630 \\
& DeCAFS & 0.755 & 0.000 & 0.000 & 0.191 & 0.066 & \bf{0.743} & 0.911 & 25.250 \\
& WCM.gSa & 0.021 & 0.000 & 0.000 & \bf{0.971} & 0.018 & 0.011 & {0.978} & {5.406} \\
& MOSUM.oracle & 0.021 & 0.000 & 0.000 & \bf{0.881} & 0.109 & 0.010 & 0.969 & 6.300 \\
& WBS2.oracle & 0.005 & 0.000 & 0.000 & \bf{0.995} & 0.005 & 0.000 & \underline{0.981} & {4.863} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-g} & MOSUM.TAVC$_{[1]}$ & 0.247 & 0.000 & 0.002 & \bf{0.970} & 0.025 & 0.003 & 0.973 & 5.232 \\ & MOSUM.TAVC$_{[2]}$ & 0.171 & 0.000 & 0.004 & \bf{0.972} & 0.023 & 0.001 & 0.972 & 5.310 \\
& WBS2.TAVC$_{[1]}$ & 0.184 & 0.000 & 0.004 & \bf{0.988} & 0.008 & 0.000 & {0.971} & {5.004} \\
& WBS2.TAVC$_{[2]}$ & 0.125 &0.000 & 0.010 & {\bf{0.987}} & 0.003 & 0.000 & 0.970 & 5.114
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.747 & 0.000 & 0.175 & \bf{0.718} & 0.107 & 0.000 & 0.925 & 8.377 \\
& DepSMUCE(0.2) & 0.921 & 0.000 & 0.019 & \bf{0.716} & 0.258 & 0.007 & 0.956 & 5.974 \\
& DeCAFS & 0.830 & 0.000 & 0.000 & \bf{0.652} & 0.171 & 0.177 & 0.958 & 6.217 \\
& WCM.gSa & 0.471 & 0.059 & 0.041 & \bf{0.830} & 0.044 & 0.026 & 0.941 & 6.975 \\
& MOSUM.oracle & 0.015 & 0.000 & 0.000 & \bf{0.903} & 0.094 & 0.003 & 0.969 & 5.312 \\
& WBS2.oracle & 0.005 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.975} & \underline{4.884} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-h} & MOSUM.TAVC$_{[1]}$ & 0.244 & 0.000 & 0.001 & \bf{0.947} & 0.052 & 0.000 & 0.969 & 6.884 \\
& MOSUM.TAVC$_{[2]}$ & 0.154 & 0.000 & 0.001 & \bf{0.961} & 0.038 & 0.000 & 0.969 & 6.918 \\
& WBS2.TAVC$_{[1]}$ & 0.160 & 0.000 & 0.002 & \bf{0.995} & 0.003 & 0.000 & {0.967} & {6.565} \\
& WBS2.TAVC$_{[2]}$ & 0.107 & 0.000 & 0.003 & {\bf{0.994}} & 0.003 & 0.000 & 0.967 & 6.658
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.219 & 0.032 & \bf{0.641} & 0.306 & 0.019 & 0.002 & 0.815 & 17.980 \\
& DepSMUCE(0.2) & 0.483 & 0.000 & 0.252 & \bf{0.633} & 0.105 & 0.010 & 0.902 & 10.728 \\
& DeCAFS & 0.341 & 0.001 & 0.002 & \bf{0.652} & 0.153 & 0.192 & 0.953 & 9.188 \\
& WCM.gSa & 0.173 & 0.076 & 0.159 & \bf{0.748} & 0.011 & 0.006 & 0.913 & 8.976 \\
& MOSUM.oracle & 0.045 & 0.000 & 0.000 & \bf{0.841} & 0.142 & 0.017 & 0.962 & 7.206 \\
& WBS2.oracle & 0.013 & 0.000 & 0.000 & \underline{\bf{0.996}} & 0.004 & 0.000 & \underline{0.971} & \underline{5.960} \\ \cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-i} & MOSUM.TAVC$_{[1]}$ & 0.311 & 0.000 & 0.008 & \bf{0.915} & 0.076 & 0.001 & \underline{0.964} & 7.027 \\
& MOSUM.TAVC$_{[2]}$ & 0.204 & 0.000 & 0.020 & \bf{0.931} & 0.049 & 0.000 & 0.962 & \underline{7.026} \\
& WBS2.TAVC$_{[1]}$ & 0.234 & 0.000 & 0.020 & \underline{\bf{0.972}} & 0.008 & 0.000 & {0.958} & 8.451 \\
& WBS2.TAVC$_{[2]}$ & 0.167 & 0.000 & 0.029 & \bf{0.968} & 0.003 & 0.000 & 0.958 & 8.143
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.130 & 0.181 & \bf{0.806} & 0.013 & 0.000 & 0.000 & 0.742 & 13.815 \\
& DepSMUCE(0.2) & 0.380 & 0.022 & \bf{0.874} & {0.094} & 0.010 & 0.000 & 0.773 & 12.108 \\
& DeCAFS & 0.075 & 0.034 & 0.414 & \bf{0.431} & 0.078 & 0.043 & 0.858 & 10.028 \\
& WCM.gSa & 0.054 & 0.057 & \bf{0.695} & {0.236} & 0.009 & 0.003 & 0.814 & 10.822 \\
& MOSUM.oracle & 0.101 & 0.000 & 0.000 & \bf{0.760} & 0.211 & 0.029 & 0.957 & 7.436 \\
& WBS2.oracle & 0.080 & 0.000 & 0.000 & \bf{0.827} & 0.171 & 0.002 & 0.950 & 7.927
\\
\bottomrule
\end{longtable}}
\endgroup
Overall, WBS2.TAVC displays better size control than MOSUM.TAVC.
The multiscale MOSUM procedure with bottom-up merging has been noted to return false positives as it accepts all estimators from the finest bandwidth; see the simulation results reported in \cite{cho2021two}.
Despite this known issue, MOSUM.TAVC shows better size control than some of the competitors such as DepSMUCE and DeCAFS. Between the two choices of the parameter $v$ used in~\eqref{eq:m:est},
the one involving~\eqref{eq:v:two} (corresponding to the subscript $2$) yields the estimator of TAVC
that returns better size control, e.g.\ closer to the nominal level $\alpha = 0.05$
for the multiscale MOSUM procedure. On the other hand, using the trimmed mean (corresponding to the subscript $1$) as in~\eqref{eq:v:one} sees improved power at the cost of larger size. This suggests that an approach combining the two choices of $v$ may yield a more balanced performance.
WBS2.TAVC performs the best across all metrics and scenarios among non-oracle methods when $q \ge 1$.
Also, we observe that using the proposed robust estimator of scale-dependent TAVC,
compares favourably to the multiscale methods applied with the true LRV
(i.e.\ MOSUM.oracle and WBS2.oracle) and
in some scenarios, the former outperforms the respective oracle counterpart.
In particular, in Scenario~\ref{model-e} where the LRV is close to $0$, plugging in its true value leads to detecting many false positives.
This shows that adopting the scale-dependent TAVC in place of the LRV for test statistic standardisation improves the finite sample performance when the change point detection procedure involves localised testing, as is the case for both the MOSUM and the WBS2 procedures.
We make a similar observation about the performance of DepSMUCE which also sets out to estimate the LRV.
DeCAFS exhibits good detection power but tends to over-estimate the number of change points as well as failing to control the size adequately
even when it is applied to the correctly specified scenario (Model~\ref{model-c}).
WCM.gSa performs well in correctly estimating the number of change points
regardless of whether $q = 0$ or $q \ge 1$. However, its performance deteriorates in the presence of nonstationarities,
see \ref{model-g}--\ref{model-i}.
Further inspection of the results under~\ref{model-i} demonstrates one advantage of the time-varying approach. Figure~\ref{fig:cpts-hist} plots the histogram of the estimated change point locations across the $1000$ replications for each of the methods. We see that the competing approaches struggle to detect the final change point due to the decreased variability towards the end of the data sequence,
whereas the proposed estimator of time-varying scale-dependent TAVC successfully extends to the locally stationary scenarios.
\begin{figure}[H]
\centering
\includegraphics[width =\textwidth]{cpts-hist-final.pdf}
\caption{Histograms plotting the change point estimators returned by different methods under Model~\ref{model-i}.}
\label{fig:cpts-hist}
\end{figure}
\subsection{Data applications}
We apply MOSUM.TAVC and WBS2.TAVC, the multiscale procedures combined with the robust estimator of scale-dependent TAVC, to two data examples.
We select the parameter $v$ in~\eqref{eq:m:est} using~\eqref{eq:v:one} and select other tuning parameters as described in Section~\ref{sec:tuning} unless specified otherwise.
\subsubsection{House price index data}
We analyse the monthly percentage changes in UK house price index (HPI), which provides insight into the estimated overall changes in house prices across the UK. The data are available from \url{https://www.gov.uk/government/statistical-data-sets/}, and a detailed description of the calculation of the HPI can be found from \cite{ukland}. The HPI series for various regions of the UK have previously been analysed in \cite{baranowski2019narrowest} and \cite{mcgonigle2021detecting}. We analyse the HPI for detached properties in the area of Somerset West and Taunton between April 1995 and February 2022 ($n = 323$).
We set the tuning parameters as described in Section~\ref{sec:tuning}
except for the window size $W = 4L$, $C = 1.15$ (for WBS2.TAVC) and $\alpha = 0.1$ (for MOSUM.TAVC) due to the short length of the time series. We combine the multiscale change point detection procedures with the robust estimator of the time-varying, scale-dependent TAVC described in Section~\ref{sec:nonstat:ext}, with the bandwidths $\mathcal G = \{ 20,40,60 \}$ for the MOSUM procedure and the minimum interval length set at $40$ for WBS2. The data are shown in Figure~\ref{fig:taunton}, with the change points detected by WBS2.TAVC and MOSUM.TAVC as well as the resulting estimated mean signal given in the top and bottom panels, respectively.
For comparison, we also apply DepSMUCE, DeCAFS and WCM.gSa to the data, see Table~\ref{table:hpi}.
\begin{figure}[H]
\centering
\includegraphics[width =0.8\textwidth]{taunton-final.pdf}
\caption{Monthly percentage change in HPI for detached properties in Somerset West and Taunton. Change point estimators returned by WBS2.TAVC (top) and MOSUM.TAVC (bottom) are denoted by dashed vertical lines, and the estimated means are given by solid line.}
\label{fig:taunton}
\end{figure}
Both WBS2.TAVC and MOSUM.TAVC detect two changes. The first change, detected in May and November 2004 for the two methods, corresponds to a decrease in the mean of the HPI series. The second change, detected in May/June 2020, may be associated with the changing consumer demand for housing in the wake of the COVID-19 pandemic. The Taunton area saw the biggest increase in overall house prices in 2021, as the ``race for space" saw buyers opt for ``more space to work from home as well as more outdoor space" \citep{guardian}.
We observe that no changes are detected by either WBS2.TAVC or MOSUM.TAVC during the 2008--2009 period associated with the financial crisis. In contrast, DepSMUCE (with $\alpha = 0.2$), DeCAFS and WCM.gSa detect changes during this time period, possibly influenced by the increased variability during the financial crisis.
Changes detected in the crisis period can be attributed to changes in variance (and autocorrelation), rather than those in mean, as noted in \cite{mcgonigle2021detecting}. We further support this interpretation by estimating the time-varying variance after adjusting for the shifts in mean using the change point estimators returned by WBS2.TAVC,
using the wavelet-based framework of \cite{nason2000wavelet} implemented for non-dyadic data as described in \cite{mcgonigle2022trend}, see Figure~\ref{fig:taunton:var}. There is a clear period of increased variance between 2007--2010, likely due to the financial crisis.
By utilising a time-varying TAVC estimator, our proposed methodology is able to capture the increased variability during this period, which ensures that potential false positives are not detected.
Furthermore, by accounting for the decrease in variability towards the end of the series, our time-varying estimator of the scale-dependent TAVC allows for the detection of a change in 2020 that is missed by other methods.
\begin{figure}[H]
\centering
\includegraphics[width =0.75\textwidth]{taunton-var-final.pdf}
\caption{Estimated variance of the mean change-adjusted Somerset West and Taunton HPI series.}
\label{fig:taunton:var}
\end{figure}
\begin{table}[htb]
\centering
\caption{Change points detected from the monthly HPI series for Somerset West and Taunton from April 1995 to February 2022.
}
\label{table:hpi}
{
\setlength{\tabcolsep}{2pt}
\begin{tabular}{ll}
\toprule
Method & Detected change points \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
WBS2.TAVC & 2004-11, 2020-06 \\
MOSUM.TAVC & 2004-05, 2020-05 \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
DepSMUCE(0.05) & 2004-11 \\
DepSMUCE(0.2) & 2008-09 \\
DeCAFS & 1999-05, 2003-01, 2008-08, 2009-01 \\
WCM.gSa & 1999-05, 2003-01, 2007-09, 2009-01, 2021-08 \\
\bottomrule
\end{tabular}}
\end{table}
\subsubsection{Nitrogen dioxide concentration in London}
We analyse the daily average concentrations of nitrogen dioxide (NO$_2$), measured in $\mu g/m^3$, recorded at Marylebone Road in London, UK. The measurements were taken from January 1st, 2000 until October 31st, 2021 ($n = 7734$).
The data set is available from \url{https://www.londonair.org.uk} and a similar dataset was analysed for shifts in the mean in \cite{cho2021multiple} using the WCM.gSa method.
The data take positive values and display both seasonality and effects due to bank holidays, as the main source of NO$_2$ emissions at the site is likely to be road traffic.
To mitigate these effects, we take the square root transform of the data and remove seasonality as described in \cite{cho2021multiple}.
We apply WBS2.TAVC and MOSUM.TAVC using the global TAVC estimator in~\eqref{eq:m:est}, with the minimum interval length set at $80$ for WBS2.TAVC and the bandwidths set as $\mathcal G = \{ 40,80,120,200,320,520,840 \}$ for the MOSUM.TAVC.
All other tuning parameters are selected as in Section~\ref{sec:tuning}.
The transformed data are shown in Figure~\ref{fig:no2}, with change points detected by WBS2.TAVC and MOSUM.TAVC and the resulting estimated mean signals given in the top and bottom panels, respectively.
For comparison, we also apply DepSMUCE, DeCAFS and WCM.gSa to the data.
Except for DeCAFS, which detects 17 change points, all methods return similar estimators. For brevity, the DeCAFS method is omitted from the results reported in Table~\ref{table:no2}.
\begin{figure}[H]
\centering
\includegraphics[width =0.95\textwidth]{no2-both-plot.pdf}
\caption{Transformed daily NO$_2$ measurements taken at Marylebone Road, London, UK. Change point estimators returned by WBS2.TAVC (top) and MOSUM.TAVC (bottom) are denoted by dashed vertical lines, and the estimated meanx are given by solid lines.}
\label{fig:no2}
\end{figure}
Both WBS2.TAVC and MOSUM.TAVC detect four change points, some of which can be linked to policy changes likely affecting the levels of air pollutants.
In February 2003, traffic management measures were introduced in central London which included modification of the pollutant filters of London buses and other heavy duty diesel vehicles, leading to an increase in their NO$_2$ emissions \citep{airquality}.
This corresponds to the change on January 31st, 2003 detected by all methods.
Also, Marylebone Road lies within the ultra low emission zone (ULEZ) that was introduced in April 2019. The ULEZ places restrictions on the levels of pollutants of vehicles travelling in the zone, and can be linked to the change on March 10th, 2019 detected by WBS2.TAVC or earlier in 2018 by other methods considering the bias in the change point estimators.
This corresponds to a decrease in the concentration of NO$_2$.
The final change point, detected by all methods on March 18th, 2020, aligns with the nationwide lockdown due to the COVID-19 pandemic on March 23rd, 2020, which resulted in drastically reduced levels of NO$_2$ throughout the UK \citep{higham2021uk}.
\begin{table}[htb]
\centering
\caption{Change points detected from the daily average concentrations of NO$_2$ at Marylebone Road in London from January 1st, 2000 to October 31st, 2021.
}
\label{table:no2}
{
\setlength{\tabcolsep}{2pt}
\begin{tabular}{ll}
\toprule
Method & Detected change points \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
WBS2.TAVC & 2003-01-31, 2010-07-25, 2019-03-10, 2020-03-18 \\
MOSUM.TAVC & 2003-01-11, 2010-03-06, 2018-12-30, 2020-03-18 \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
DepSMUCE(0.05) & 2003-01-31, 2010-07-25, 2018-10-14, 2020-03-18 \\
DepSMUCE(0.2) & 2003-01-31, 2008-08-31, 2012-10-04, 2018-10-14, 2020-03-18 \\
WCM.gSa & 2003-01-31, 2009-12-09, 2018-10-14, 2020-03-18 \\
\bottomrule
\end{tabular}}
\end{table}
\section{Conclusions}
\label{sec:conc}
We propose an estimator of scale-dependent TAVC that is robust to the presence of (possibly) multiple mean shifts.
It is readily combined with multiscale change point detection methodologies which, by scanning for change points over data sections of varying lengths, provide good adaptivity to the problem of multiple change point detection.
We show the consistency of the proposed estimator under general assumptions permitting heavy tails and serial dependence decaying at a polynomial rate, and investigate its use with the multiscale MOSUM procedure and the WBS2 algorithm.
Through extensive numerical studies, we demonstrate the benefit of adopting the proposed estimator of scale-dependent TAVC for improved finite sample performance, as it better reflects the level of variability in the local data sections involved in the multiscale methods.
In particular, the heuristic extension to local stationarity shows promising performance which provides a natural avenue for future research.
\bibliographystyle{apalike}
\section{Introduction}
Dating back to the 1950s \citep{page1954continuous},
change point analysis has a long tradition in statistics.
The area continues to be an active field of research due to its importance in many applications where data is routinely collected in highly nonstationary environments, examples of which include genetics \citep{hocking2013learning} and climatology \citep{reeves2007review}.
Data segmentation, a.k.a.\ multiple change point detection, enables partitioning of a time series into stationary regions and thus provides a simple framework for modelling nonstationary time series.
In this paper, we consider the problem of detecting multiple change points in the piecewise constant mean of an otherwise stationary time series.
We briefly review the existing literature on multiple change point detection
in the presence of serial dependence; we refer to \cite{aue2013} and \cite{truong2020selective} for a comprehensive overview.
One line of research takes a parametric approach
and simultaneously estimates the serial dependence and change points. For example, \cite{chakar2017robust}, \cite{fang2020detection} and \cite{romano2021detecting} assume an autoregressive (AR) model of order one, while \cite{lu2010mdl}, \cite{cho2021multiple} and \cite{gallagher2021autocovariance} permit an AR model of arbitrary order.
Another line of research focuses on applying the methodologies
developed for independent data to time series settings.
\cite{lavielle2000} and \cite{cho2021data} extend the use of an information criterion \citep{yao1988} to data exhibiting serial correlations and heavy tails,
which requires the choice of an appropriate penalty that depends on the level of noise.
\cite{tecuapetla2017autocovariance}, \cite{eichinger2018mosum}, \cite{dette2020multiscale} and \cite{chan2022optimal}
propose estimators of the long-run variance (LRV) for quantifying the level of noise
that are robust to the presence of multiple mean shifts.
Such estimators are then combined e.g.\ with a moving sum procedure or the simultaneous multiscale change point estimator algorithm
\citep{frick2014multiscale}, for multiple change point detection.
We also note that \cite{wu2019mace} and \cite{zhao2021segmenting} extend self-normalisation-based change point tests
to the data segmentation problem.
In this paper, we propose a robust estimator of the
scale-dependent time-average variance constant (TAVC, \citeauthor{wu2009recursive}, \citeyear{wu2009recursive}).
It is closely related to the literature on estimation of the LRV,
namely $\sigma^2 = \lim_{N \to \infty} \Var(N^{-1/2} \sum_{t = 1}^N \varepsilon_t)$ for a stationary time series $\{\varepsilon_t\}_{t \in \mathbb{Z}}$,
but distinct in that our interest lies in estimating
\begin{align}
\label{eq:tavc}
\sigma^2_L = \Var\l( \frac{1}{\sqrt L} \sum_{t = 1}^L \varepsilon_t \r),
\end{align}
for a given scale $L$.
We argue that such a scale-dependent TAVC estimator is well-suited
to be combined with a class of {\it multiscale} change point detection methodologies, examples of which include
the moving sum (MOSUM) procedure \citep{eichinger2018mosum} and the wild binary segmentation (WBS, \citeauthor{fryzlewicz2014wild}, \citeyear{fryzlewicz2014wild}) algorithm. Such approaches {\it locally} apply change point tests for single change point detection to data sections of varying lengths and achieve good adaptivity in multiple change point detection \citep{cho2021data}.
We motivate the use of scale-dependent TAVC in~\eqref{eq:tavc} in combination with such multiscale methods in the following example.
\begin{example} \label{ex:one}
Consider an MA($1$) process $\varepsilon_t = W_t - 0.9 W_{t-1}$,
where $\{W_t\}_{t \in \mathbb{Z}}$ is a white noise process with $\Var(W_t) = 1$.
Then, we have $\sigma^2 = 0.01$, while the TAVC at scales $L = 50$ and $L = 100$
are $\sigma_{50}^2 = 0.1$ and $\sigma^2_{100} = 0.055$ respectively, which are several factors larger than the LRV.
On the other hand, consider an AR($1$) process $\varepsilon_t = 0.9 \varepsilon_{t-1} + W_{t}$, with $W_t$ defined as above. Then, $\sigma^2 =100$ while $\sigma_{50}^2 \approx 81.16$
and $\sigma^2_{100} \approx 90.53$, both of which are smaller than the LRV.
\end{example}
This demonstrates that adopting the global LRV may fail to reflect the degree of variability in the local data sections that are involved in multiscale algorithms,
which in turn may result in false negatives (failure in detecting genuine changes)
or false positives (detection of spurious estimators).
Moreover, when the LRV is close to zero as in the first instance of Example~\ref{ex:one}, some estimators of the LRV have been observed to take negative values \citep{huvskova2010note}, which further makes their use in change point problems undesirable.
To ensure that the scale-dependent TAVC estimator is robust to the mean shifts, we adopt the robust $M$-estimation framework of \cite{catoni2012challenging}, which was first proposed in the independent setting for mean and variance estimation and further extended to the serially dependent setting for LRV estimation in \cite{chen2021inference}.
We establish the consistency of the proposed robust estimator of scale-dependent TAVC under general conditions permitting heavy tails and serial dependence decaying at a polynomial rate.
Then, we discuss its application with multiscale change point detection methods such as the MOSUM procedure and the WBS algorithm,
and provide a heuristic approach to accommodate local stationarity in the data.
The remainder of the article is organised as follows.
Section~\ref{sec:tavc} introduces the scale-dependent TAVC and
its robust estimator and establishes its consistency.
Section~\ref{sec:multiscale} discusses its application with multiscale data segmentation algorithms and an extension to local stationarity.
In Section~\ref{sec:num}, we examine the performance of the proposed methodology
on simulated data sets and two real data examples on house price index and air quality.
Section~\ref{sec:conc} concludes the paper.
All proofs, algorithmic descriptions of multiscale change point methods
and additional numerical results are given in the appendix.
Accompanying \verb!R! software implementing the methodology is available from \url{https://github.com/EuanMcGonigle/TAVC.seg}.
\section{Scale-dependent TAVC and its robust estimation}
\label{sec:tavc}
\subsection{Multiscale change point detection in the mean}
We consider the problem of multiple change point detection
under the following model:
\begin{align}
\label{eq:model}
X_{t} = f_{t} + \varepsilon_{t} = f_{0} + \sum_{i=1}^{q} \mu_{i} \cdot \mathbb{I}(t \geq \tau_{i}+1) + \varepsilon_{t}, \quad t = 1, \ldots, n.
\end{align}
Under Model~\eqref{eq:model}, the piecewise constant signal $f_{t}$ contains $q$ change points at locations $\tau_{i}$, $i = 1, \ldots, q$, with the notational convention that $\tau_{0}=0$ and $\tau_{q+1} = n$.
The errors $\{ \varepsilon_{t} \}_{t=1}^{n}$ are assumed to be a (weakly) stationary time series satisfying $\mathbb{E} (\varepsilon_{t}) = 0$
with finite LRV $\sigma^2 = \lim_{N \to \infty} \Var(N^{-1/2} \sum_{t = 1}^N \varepsilon_t) \in (0, \infty)$,
and are permitted to be serially correlated and heavy-tailed
(see Assumption~\ref{assum:noise} below for details).
Our aim is to consistently estimate the total number and the locations of the change points.
A common approach to this problem is closely related to the change point testing literature,
which scans the data for the detection and estimation of multiple change points
by {\it locally} applying a test well-suited for detecting a single change.
Such a procedure typically involves comparing a test statistic of the form
$\widehat{\sigma}_{s, e}^{-1} \vert \mathcal{T}_{s, k, e} \vert$ to a threshold, say $D$.
Here,
\begin{align}
\label{eq:cusum-stat}
\mathcal{T}_{s, k, e} = \sqrt{ \frac{(k-s)(e-k)}{e-s}} \left( \frac{1}{k-s} \sum_{t=s+1}^{k} X_{t} - \frac{1}{e-k} \sum_{t=k+1}^{e} X_{t} \right)
\end{align}
denotes a change point detector evaluated at some locations
$0 \le s < k < e \le n$, which are determined in a method-specific way
(see Section~\ref{sec:multiscale} below),
$\sigma_{s, e}^2$ denotes a measure of variability
in the data section $\{X_t\}_{t = s + 1}^e$, and
$\widehat\sigma_{s, e}^2$ its estimator.
Under the stationarity assumption on $\{\varepsilon_t\}_{t = 1}^n$,
a natural choice is $\sigma^2_{s, e} = \sigma^2_{e - s}$,
the scale-dependent TAVC defined in~\eqref{eq:tavc} with $L = e - s$ as the scale.
Then, if $\{X_t\}_{t = s + 1}^e$
does not contain any change point well within the interval, we expect
$\widehat{\sigma}_{s, e}^{-1} \vert \mathcal{T}_{s, k, e} \vert \le D$,
while it signals the presence of such a change point
when $\widehat{\sigma}_{s, e}^{-1} \vert \mathcal{T}_{s, k, e} \vert > D$.
The key challenge lies in separating genuine mean shifts from the natural fluctuations due to serial correlations,
which requires a careful selection of the estimator $\widehat\sigma_{s, e}^2$
that correctly captures the variability in the section of the data under consideration.
It is well-documented that multiscale application of such a test
on data sections of varying lengths,
improves the adaptivity of the change point methodology to detect both large, frequent changes
and small changes over long stretches of stationarity \citep{cho2021data}.
For such a multiscale procedure,
Example~\ref{ex:one} demonstrates the potential pitfalls associated with
using an estimator of the global LRV in place of $\widehat\sigma_{s, e}^2$,
regardless of the length of the interval on which the detector statistic in~\eqref{eq:cusum-stat} is computed.
In the next section, we propose an estimator of
the scale-dependent TAVC $\sigma^2_L$ in~\eqref{eq:tavc}
that is robust to the presence of multiple mean shifts,
for the standardisation of multiscale change point detectors.
\subsection{Robust estimation of multiscale TAVC}
\label{sec:robust}
For notational convenience, suppose that $L$ is an even number, and let $G = L/2$ denote the block size. Then, for some starting point $b \in \{0, \ldots, G - 1\}$ with number of blocks $N_1(b) = \lfloor (n-b-G)/G \rfloor$, we define
\begin{align*}
\bar{X}_{j, b} = \frac{1}{G} \sum_{t = jG + b + 1}^{(j+1)G + b} X_t,
\quad \text{and} \quad
\xi_{j, b} = \frac{G(\bar{X}_{j, b} - \bar{X}_{j-1, b})^2}{2}
\text{ for } j = 1, \ldots, N_1(b).
\end{align*}
Analogously, we define
\begin{align*}
\bar{\varepsilon}_{j, b} = \frac{1}{G} \sum_{t=jG + b + 1}^{(j+1)G + b} \varepsilon_t
\quad \text{and} \quad
\widetilde\xi_{j, b} = \frac{G(\bar{\varepsilon}_{j, b} - \bar{\varepsilon}_{j-1, b})^2}{2}.
\end{align*}
Then, the following sum
\begin{align}\label{no-change-est}
\widehat{\widetilde\sigma}^{2}_L = \frac{1}{N_1(b)} \sum_{j=1}^{N_1(b)} \widetilde\xi_{j, b},
\end{align}
takes into account the temporal dependence in the local data sections of length $L = 2G$.
Further, we have
$\E(\widehat{\widetilde\sigma}^2_L) - \sigma^2_L = o(1)$ for $L \to \infty$
(see Theorem~\ref{thm:one} below),
such that $\widehat{\widetilde\sigma}^2_L$ is indicative of the level of variability $\sigma^2_L$ albeit being inaccessible (as it is defined with $\bar{\varepsilon}_{j, b}$ in place of $\bar{X}_{j, b}$).
Its accessible counterpart, $N_1(b)^{-1} \sum_{j = 1}^{N_1(b)} \xi_{j, b}$, on the other hand, is typically biased due to the mean shifts
and thus is inappropriate as an estimator of the scale-dependent TAVC.
To obtain an estimator that is robust to multiple mean shifts,
we adopt the robust $M$-estimation framework of \cite{catoni2012challenging}.
Let $\phi$ denote a non-decreasing influence function as
\begin{align}
\label{eq:envelope}
\phi(x) = \begin{cases}
- \log(2) & \text{for } x \leq -1 , \\
\log(1 + x + x^{2}/2 ) & \text{for } -1 \leq x \leq 0, \\
- \log( 1 - x + x^{2}/2 ) & \text{for } 0 \leq x \leq 1, \\
\log(2) & \text{for } x \geq 1.
\end{cases}
\end{align}
The robust estimator of the TAVC at scale $L$ and starting point $b$, denoted $\widehat{\sigma}^2_{L, b}$, is defined as the solution of the $M$-estimation equation
\begin{align}\label{eq:m:est}
h_{L, b} (u) = \frac{1}{N_1(b)} \sum_{j = 1}^{N_1(b)} \phi_v\l( \xi_{j, b} - u \r) = 0,
\end{align}
where $\phi_{v}(x) = v^{-1}\phi(vx)$ for some $v > 0$;
we specify the condition on $v$ later.
If there are multiple solutions to Equation~\eqref{eq:m:est}, any of them may be chosen.
\subsection{Theoretical properties}
We establish the consistency of the scale-dependent TAVC estimator
under the following assumption on the error process $\{ \varepsilon_t \}_{t=1}^n$.
\begin{assumption}
\label{assum:noise}
\hfill
\begin{enumerate}[label = (\roman*)]
\item \label{assum:noise:one}
We assume that $\varepsilon_t = \sum_{k=0}^\infty {a}_k {\eta}_{t-k}$,
where $\{\eta_t\}_{t \in \mathbb{Z}}$ is a sequence of i.i.d.\ random variables
and $\vert {a}_k \vert \leq \Xi (k + 1)^{-\beta}$ for some constants $\beta > 2.5$ and $\Xi > 0$ for all $k \ge 0$.
\item \label{assum:noise:two}
There exists a fixed constant $c_\sigma > 0$ such that
$\sigma^2 = ( \sum_{k\geq 0} a_k )^2$ satisfies $\sigma^2 \ge c_{\sigma}$.
\item \label{assum:noise:three} We operate under {\it either} of the following two conditions on the distribution of $\{\eta_t\}_{t \in \mathbb{Z}}$.
\begin{enumerate}[label = (\alph*)]
\item \label{cond:moment} There exists a fixed constant $r > 4$ such that
$\Vert \eta_{1} \Vert_r = ( \mathbb{E} ( |\eta_{1} |^r) )^{1/r} <\infty$.
\item \label{cond:subexp} There exist fixed constants $C_\eta > 0$ and $\kappa \ge 0$ such that
$\Vert \eta_1 \Vert_r \le C_\eta r^\kappa$ for all $r \ge 1$.
\end{enumerate}
\end{enumerate}
\end{assumption}
Assumption~\ref{assum:noise}~\ref{assum:noise:one} is a mild condition that permits the temporal dependence to decay at a polynomial rate, and~\ref{assum:noise:two} is made to ensure that the LRV is well-defined.
Assumption~\ref{assum:noise}~\ref{assum:noise:three}~\ref{cond:moment} allows heavy-tailed $\{\varepsilon_t\}_{t = 1}^n$,
while~\ref{cond:subexp} assumes a stronger condition
that requires sub-Weibull \citep{wong2020lasso} tail behaviour on $\{\eta_t\}_{t \in \mathbb{Z}}$ which includes sub-Gaussian ($\kappa = 1/2$) and sub-exponential ($\kappa = 1$)
distributions as special cases.
The following theorem establishes the consistency
of the estimator of the scale-dependent TAVC,
see Appendix~\ref{sec:proof} for the proof.
\begin{theorem}\label{thm:one}
Suppose that Assumption~\ref{assum:noise} holds, and define
\begin{align*}
\widetilde{\sigma}^2_L = \Var \left( \frac{1}{\sqrt{L}} \left( \sum_{t=1}^{G} \varepsilon_t - \sum_{t=G+1}^L \varepsilon_t \right) \right)
\end{align*}
for $L = 2G$.
Then, provided that $v \asymp \sqrt{G/n}$,
the estimator $\widehat{\sigma}^{2}_{L, b}$ satisfies
\begin{align}
\left| \widehat{\sigma}^{2}_{L, b} - \widetilde{\sigma}^{2}_L \right| &= \mathcal{O}_P\l( \sqrt{ \frac{Lq}{n} } + \max\l\{ \l( \frac{L}{n} \r)^{\frac{r-2}{r+2}}, \sqrt{\frac{L \log(n)}{n}} \r\} \r) \label{eq:thm:one}
\end{align}
under Assumption~\ref{assum:noise}~\ref{assum:noise:three}~\ref{cond:moment}, and
\begin{align}
\left| \widehat{\sigma}^{2}_{L, b} - \widetilde{\sigma}^{2}_L \right| &= \mathcal{O}_{P} \left( \sqrt{ \frac{Lq}{n} } + {\sqrt{ \frac{L \log^{4\kappa + 3} (n)}{n} }}\right)
\label{eq:thm:two}
\end{align}
under Assumption~\ref{assum:noise}~\ref{assum:noise:three}~\ref{cond:subexp}, for any fixed $b \in \{0, \ldots, G - 1\}$.
In addition, $\widetilde{\sigma}^2_L$ satisfies
\begin{align}
\left| \widetilde{\sigma}^{2}_{L} - {\sigma}_L^{2} \right| = \mathcal{O}(L^{-1}).
\label{eq:thm:three}
\end{align}
\end{theorem}
Theorem~\ref{thm:one} shows that the proposed robust estimator consistently estimates the TAVC at scale $L$.
The estimation error is decomposed into the error from approximating $\sigma_L^2$ by $\widetilde\sigma_L^2$ in~\eqref{eq:thm:three}, and
that in estimating $\widetilde{\sigma}_L^2$ by $\widehat{\sigma}_{L,b}^2$.
In deriving the second error, we make explicit the influence of multiple mean shifts on the estimator by the term $(Lq/n)^{1/2}$ in~\eqref{eq:thm:one}--\eqref{eq:thm:two}, as well as the effect of the innovation distribution in the remaining terms.
A careful examination of the proof of Theorem~\ref{thm:one} shows that
\begin{align}
\left| \widetilde{\sigma}^{2}_{L} - {\sigma}^{2} \right| &= \mathcal{O}(L^{-1}),
\label{sigma-approx2}
\end{align}
i.e.\ as $L$ increases, the scale-$L$ TAVC approximates the global LRV as expected.
\begin{remark}[Maximum time-scale for TAVC estimation]
\label{rem:max}
The error due to approximating $\sigma_L^2$ with $\widetilde{\sigma}_L^2$
decreases with $L$ as in~\eqref{sigma-approx2}.
On the other hand, the error of estimating $\widetilde{\sigma}_L^2$ with $\widehat{\sigma}_{L, b}^2$ increases with $L$ as in~\eqref{eq:thm:one}--\eqref{eq:thm:two};
this is attributed to the effect of mean shifts that grows with $L$,
and the decrease in the number of available blocks.
To balance between the two,
we suggest setting a maximum time-scale, say $M$, to be used in combination with a multiscale change point detection algorithm.
That is, when the change point detector $\mathcal T_{s, k, e}$ involves
$e - s \le M$, we scale the detector with the estimator of $\sigma^2_{s - e}$, the TAVC at the corresponding scale $L = e - s$.
On the other hand, if $e - s > M$, we propose to scale the detector
with the estimator of $\sigma^2_{M}$, the TAVC at the maximum time-scale $M$,
which satisfies $\vert \sigma^2_{e - s} - \sigma^2_M \vert = \mathcal{O} (M^{-1})$.
\end{remark}
\section{Applications and extensions}
\label{sec:multiscale}
We now describe explicitly how the robust estimator of the scale-dependent TAVC proposed in Section~\ref{sec:tavc}, is applied within the algorithms
that scan moving sum (MOSUM) and cumulative sum (CUSUM) statistics of the form~\eqref{eq:cusum-stat}, for multiple change point detection.
\subsection{MOSUM procedure}
\label{sec:mosum}
The MOSUM procedure \citep{chu1995mosum,eichinger2018mosum}
evaluates the change point detector $\mathcal T_{s, k, e}$ over a moving window.
For a given bandwidth $G$, the MOSUM detector for a change in mean at time point $k$ is given by
\begin{align}
\label{eq:mosum}
\mathcal T_{G}(k) = \mathcal{T}_{k-G, k, k+G} = \sqrt{\frac{G}{2}} \left( \frac{1}{G} \sum_{t=k+1}^{k+G} X_{t} - \frac{1}{G} \sum_{t=k-G+1}^{k} X_{t} \right), \quad G \le k \le n - G.
\end{align}
\cite{eichinger2018mosum} propose to estimate the total number and the locations of multiple change points by
identifying all significant local maximisers of $\mathcal T_G(k)$, say $\widehat k$, satisfying
\begin{align}
\label{eq:mosum:est}
\vert \mathcal T_G(\widehat k) \vert > \widehat\sigma \cdot D_n(G; \alpha) \quad \text{and} \quad
\widehat k = {\arg\max}_{k: \, \vert k - \widehat k \vert \le \eta G} \vert \mathcal T_G(k) \vert
\end{align}
for some $\eta \in (0, 1)$.
Here, $D_n(G; \alpha)$ denotes a critical value at a significance level $\alpha \in (0, 1)$,
which is drawn from the asymptotic null distribution of the MOSUM test statistic $\max_{G \le k \le n - G} \sigma^{-1} \vert T_G(k) \vert$ obtained
under mild conditions permitting heavy-tailedness and serial dependence, and
takes the form
\begin{align*}
D_{n} (G, \alpha) = \frac{b_{G,n} + c_{\alpha}}{a_{G,n}} \quad \text{with} \quad
c_{\alpha} = - \log \log \left( \frac{1}{\sqrt{1-\alpha}} \right),
\end{align*}
where $a_{G,n}$ and $b_{G,n}$ are known constants depending on $G$ and $n$ only. The single-bandwidth MOSUM procedure achieves consistency in estimating the total number and the locations of multiple change points,
provided that $\min_{1 \le i \le q} \mu_i^2 G \to \infty$ as $n \to \infty$ sufficiently fast while $2G \le \min_{0 \le i \le q} (\tau_{i + 1} - \tau_i)$, see Theorem~3.2 of \cite{eichinger2018mosum} and Corollary D.2 of \cite{cho2021two} for explicit conditions.
The requirement on $G$ indicates that the single-scale MOSUM procedure performs best with the bandwidth chosen as large as possible while avoiding situations where there are more than one change point within the moving window at any time. Consequently, it lacks adaptivity when the data sequence contains both large changes over short intervals and small changes over long intervals.
Multiscale extensions of the single-bandwidth MOSUM procedure,
i.e.\ applying the MOSUM procedure with a range of bandwidths
and then combining the results, alleviate the difficulties involved in bandwidth selection and provide adaptivity.
In this paper, we consider the multiscale MOSUM procedure
combined with the `bottom-up' merging as proposed by \cite{messer2014multiple} (see also \cite{meier2021mosum}).
Denoting the range of bandwidths by $\mathcal G = \{G_h, \, 1 \le h \le H: \,
G_1 < \ldots < G_H\}$,
let $\mathcal C(G)$ denote the set of estimators detected with some bandwidth~$G \in \mathcal G$.
Then, we accept all estimators in $\mathcal C(G_1)$ returned with the finest bandwidth $G_1$ to the set of final estimators $\mathcal C$
and, sequentially for $h = 2, \ldots, H$, accept $\widehat k \in \mathcal C(G_h)$ if and only if
$\min_{k \in \mathcal C} \vert \widehat k - k \vert > \eta G_h$
(with $\eta$ identical to that in~\eqref{eq:mosum:est}). That is, we only accept the estimators that do not detect the change points which have previously been detected at a finer scale.
We propose to apply the multiscale MOSUM procedure with bottom-up merging,
in combination with the robust estimator of multiscale TAVC as follows.
For each $G_h \in \mathcal G$, the TAVC at scale $L = 2 G_h$ is estimated
by $\widehat{\sigma}^2_{2 G_h}$ solving~\eqref{eq:m:est}, provided that $2G_h \le M$.
Here, $M$ denotes the maximum scale which is set in relation to the sample size $n$, see Remark~\ref{rem:max}.
Then, we use $\widehat{\sigma}_{2 G_h}$ in place of the global estimator $\widehat\sigma$
in~\eqref{eq:mosum:est} to standardise the MOSUM detector $\mathcal T_{G_h}(k)$.
When $2G_h > M$, we use $\widehat{\sigma}_M$ in place of $\widehat\sigma_{2G_h}$
for MOSUM detector standardisation.
In doing so, we ensure that change point detectors at multiple scales are standardised
by the scale-dependent TAVC that accurately reflects the degree of variability over the moving window, while taking into account the presence of possibly multiple mean shifts therein.
We refer to Algorithm~\ref{alg:mosum} in Appendix~\ref{appendix:alg} for the pseudocode of the multiscale MOSUM procedure with the robust estimator of scale-dependent TAVC.
\subsection{Wild binary segmentation}
\label{sec:wbs}
The binary segmentation algorithm \citep{scott1974cluster, vostrikova1981} and its extensions, such as wild binary segmentation (WBS, \citeauthor{fryzlewicz2014wild}, \citeyear{fryzlewicz2014wild}; \citeyear{fryzlewicz2020detecting}) and seeded binary segmentation \citep{kovcs2020seeded}, recursively search for multiple change points using the CUSUM statistic of the form~\eqref{eq:cusum-stat}, with $s$ and $e$ that are identified iteratively.
These methods have primarily been analysed for the data segmentation problem under~\eqref{eq:model}
assuming i.i.d.\ Gaussianity on the $\{\varepsilon_t\}_{t = 1}^n$.
Consequently, some robust estimators of $\Var(\varepsilon_1)$ have been considered for standardising the CUSUM statistic.
Here, we discuss the application of the WBS2 algorithm \citep{fryzlewicz2020detecting} in the time series setting with the proposed robust estimator of the scale-dependent TAVC.
Let $\mathcal A_{s, e} = \{(\ell, r) \in \mathbb{Z}^2:\, s \le \ell < r \le e, \ r - \ell > 1\}$
denote the collection of all intervals within $\{s + 1, \ldots, e\}$ for some $0 \le s < e \le n$, and $\mathcal R_{s, e}$ denote its subset selected either randomly or deterministically
(see \cite{cho2021multiple} for one approach to deterministic grid selection)
with $\vert \mathcal R_{s, e} \vert = \min(R, \vert \mathcal A_{s, e} \vert)$ for some given $R \le n(n - 1)/2$.
Starting with $(s, e) = (0, n)$, we identify
\begin{align}
\label{eq:wbs:est}
(s_\circ, k_\circ, e_\circ) = {\argmax}_{\substack{(\ell, k, r): \, \ell < k < r \\ (\ell, r) \in \mathcal R_{s, e}}} \frac{\left\vert \mathcal T_{\ell, k, r} \right\vert}{\widehat{\sigma}_{r - \ell}}
\quad \text{with} \quad
\left\vert \mathcal T_{s_\circ, k_\circ, e_\circ} \right\vert >
\widehat{\sigma}_{e_\circ - s_\circ} \cdot D
\end{align}
for some threshold $D$ and $\widehat{\sigma}^2_{r - \ell}$ denoting the proposed robust estimator of the TAVC at scale $L = r - \ell$ (when $r - \ell$ is odd, we use $\widehat{\sigma}^2_{r - \ell - 1}$ instead).
As in Section~\ref{sec:mosum}, a maximum scale $M$ is set so that the CUSUM statistic over any interval of length greater than $M$ is standardised using $\widehat{\sigma}_{M}$.
Following the recommendation made in \cite{fryzlewicz2014wild}, we adopt the threshold
$D = C \sqrt{2 \log(n)}$ where $C$ is a universal constant.
If $(s_\circ, k_\circ, e_\circ)$ that fulfils~\eqref{eq:wbs:est} exists,
it signals the presence of a change point so that
the data is partitioned into $\{X_t\}_{t = s + 1}^{k_\circ}$ and $\{X_t\}_{t = k_\circ + 1}^e$, and the same step of detecting and identifying a single change point
is repeated on each partition separately.
If no such $(s_\circ, k_\circ, e_\circ)$ exists, or when the user-specified minimum segment length is reached,
then the search for change points
is terminated on $\{X_t\}_{t = s + 1}^e$.
We provide a pseudocode of the WBS2 algorithm with the robust estimator of scale-dependent TAVC in Algorithm~\ref{alg:wbs2} of Appendix~\ref{appendix:alg}.
\subsection{Extension to local stationarity}
\label{sec:nonstat:ext}
We propose a heuristic extension of the robust estimator of scale-dependent TAVC to the setting where the second-order structure of $\{\varepsilon_t\}_{t = 1}^n$ varies smoothly over time.
Suppose that there exists an appropriately chosen window size $W$ such that $\{\varepsilon_t\}_{t = k - \lfloor W/2 \rfloor + 1}^{k + \lfloor W/2 \rfloor}$ may be regarded as being approximately second-order stationary over all $k$. Then, we propose to perform the robust estimation described in Section~\ref{sec:robust} in a localised fashion.
To this end, define the time-varying TAVC at scale $L$ and time $k$ by
\begin{align}
\label{eq:tavc-tvar}
\sigma^2_L (k) = \Var\l( \frac{1}{\sqrt L} \sum_{t = k-G+1}^{k+G} \varepsilon_t \r) .
\end{align}
For notational convenience, we set $W = N_2 L$ for some integer $N_2$, and let $N_3 = \lfloor (W-G)/G \rfloor$ be the number of blocks for window size $W$. For $k \in \{ W/2, \ldots, n - W/2 \}$, we estimate $\sigma^2_L(k)$ by $\widehat{\sigma}^2_L(k)$, the solution of the following $M$-estimation equation
\begin{align}\label{m-est-eq2}
h_L (u, k) = \frac{1}{N_3} \sum_{j = 1}^{N_3} \phi_v \left( \xi_{j, k - W/2} - u \right) = 0
\end{align}
with $v \asymp (G/W)^{1/2}$.
We apply a boundary extension so that $\widehat{\sigma}_L^2(1) = \ldots = \widehat{\sigma}_L^2(W/2)$ and $\widehat{\sigma}_{L}^2(n -W/2) = \ldots = \widehat{\sigma}_L^2(n)$.
The estimator of the local TAVC at time $k$ and scale $L$ is obtained analogously as that of the global TAVC at scale $L$ described in Section~\ref{sec:robust}, except that we only use the windowed data region starting at time $k - W/2$ and ending at $k + W/2$ for the estimation of the former.
Then, the MOSUM detector $\mathcal T_G(k)$ and the CUSUM statistic $\mathcal T_{s, k, e}$ described in Sections~\ref{sec:mosum}--\ref{sec:wbs} are standardised in a time-dependent way
using $\widehat{\sigma}_{2G}(k)$ and $\widehat{\sigma}^2_{(e-s)}(k)$, respectively.
In practice, we observe that taking the running median of $\{\widehat{\sigma}^2_L(k + b - \lfloor G/2 \rfloor )\}_{b = 0}^{G - 1}$ as an estimator of $\sigma^2_L(k)$ further improves the performance, as it `smoothes' out the local estimators and enhances the robustness to mean changes.
We illustrate the benefit of adopting the time-varying adaptation of the proposed robust estimator using the following example.
Consider a time series of length $n=1000$, where the errors $\{ \varepsilon_t\}_{t \in \mathbb{Z}}$ follow a time-varying AR(1) model: $\varepsilon_t = a_1 (t) \varepsilon_{t-1} + W_t$, with $a_{1} (t) = 0$ for $t\leq 500$, $a_{1} (t) = 0.7$ for $t \ge 501$, and $W_t \sim_{\text{i.i.d.}} \mathcal N(0, 1)$. There are two changes in the mean at $\tau_1 = 300$ and $\tau_2 = 700$, with change sizes~$1$ and~$2$, respectively.
In Figure~\ref{fig:toy:ex}, we show the MOSUM detector statistic in~\eqref{eq:mosum} calculated using bandwidth $G = 100$.
We also plot the threshold $D_n(G, \alpha)$ at the significance level $\alpha = 0.05$, multiplied by the square root of the global TAVC estimator at scale $L = 2G$ (i.e.\ $\widehat\sigma^2_{2G}$) in dashed blue line,
and that multiplied by the square root of the local estimators of the scale-$L$ TAVC (i.e.\ $\widehat\sigma^2_{2G}(k)$) in solid red line. We see that using the global approach misses the change at time $\tau_1 = 300$ due to the global scale-$L$ TAVC estimator being too large, whilst the localised approach successfully detects both changes.
Lastly, we mention that the robust estimation of time-varying and scale-dependent TAVC is of independent interest beyond the context of change point analysis, with possible extensions including the estimation of other second-order properties. For example, the procedure can be used to obtain a robust estimator of the spectrum of a locally stationary wavelet process \citep{nason2000wavelet} while the time series undergoes shifts in the mean.
\begin{figure}[H]
\centering
\includegraphics[width = 0.8\textwidth]{final-tvar-ex.pdf}
\caption{MOSUM detector statistic (black solid line) computed from the time series generated as described in the text with two changes in the mean (denoted by vertical lines). Also shown are the thresholds computed using global (blue dashed line) and local (red solid line) estimators of the scale-dependent TAVC.}
\label{fig:toy:ex}
\end{figure}
\section{Numerical results}
\label{sec:num}
\subsection{Practical considerations}
\label{sec:tuning}
We provide guidance on the selection of tuning parameters
required for the proposed robust TAVC estimator and its application
with multiscale change point detection methods.
\vspace{-10pt}
\paragraph{Parameter $v$ in~\eqref{eq:m:est}.}
We select $v = v_b = \bar{\xi}_b^{-1}\sqrt{G/n}$ where $\bar{\xi}_b$ is a fixed constant for a given $b \in \{0, \ldots, G - 1\}$.
For the problem of robust mean estimation, \cite{catoni2012challenging} recommends the standard deviation in the place of $\bar{\xi}_b$
and in a similar vein, \cite{chen2021inference} propose to use a trimmed mean
\begin{align}
\label{eq:v:one}
\bar{\xi}_{[1], b} = {\frac{1}{\lfloor 3N_1(b)/4 \rfloor - \lceil N_1(b)/4 \rceil + 1}} \sum_{j = \lceil N_1(b)/4 \rceil}^{ \lfloor 3N_1(b)/4 \rfloor} \xi_{(j), b},
\end{align}
where $\xi_{(1), b} \le \ldots \le \xi_{(N_1(b)), b}$ are the ordered $\{\xi_{j, b} \}_{j=1}^{N_1(b)}$. Another approach is to use an appropriately scaled median of $\{\xi_{j, b} \}_{j=1}^{N_1(b)}$, in a similar fashion to \cite{mcgonigle2021detecting}. In this case, the necessary scaling constant to ensure unbiasedness can be chosen by noting that $\bar{X}_{j, b}$ is asymptotically Gaussian as $L \to \infty$, and thus $\xi_{j, b}$ is asymptotically scaled $\chi^2_1$. This leads to the choice
\begin{align}
\label{eq:v:two}
\bar{\xi}_{[2], b} = K \cdot \text{Median}\l\{ \xi_{1,b}, \ldots , \xi_{N_1(b),b} \r\},
\end{align}
where $K = 2.125$. In simulation studies, we report the results obtained with $\bar{\xi}_{[\ell], b}$, $\ell = 1, 2$, in setting the parameter $v_b$ where we observe that both choices return similarly good results.
\vspace{-10pt}
\paragraph{Parameters $b$ and $b_{\max}$ in~\eqref{eq:m:est}.}
As a further step to ensure greater robustness of the estimator, we obtain the estimator $\widehat\sigma^2_{L, b}$ for a range of $b \in \{0, \ldots, b_{\max}\}$ and take their median as the final estimator $\widehat\sigma^2_L$. Informally, we may get unlucky with some starting value $b$ that leads to many of the $\xi_{j, b}$ contaminated by the mean changes, and taking the median over a range of values of $b$ helps in alleviating this. We take $b_{\max} = G - 1$ in practice which yields good performance.
\vspace{-10pt}
\paragraph{Maximum time-scale $M$.}
We recommend $M = \lfloor 2.5 \sqrt{n} \rfloor$ as the coarsest scale at which the scale-dependent TAVC is estimated. This choice is made to balance between mitigating the effect of change points, and ensuring that the TAVC at coarser scales is well-approximated by $\sigma^2_M$.
\vspace{-10pt}
\paragraph{Tuning parameters for the multiscale MOSUM procedure.}
We follow \cite{cho2021two} and generate $\mathcal G$ as a sequence of Fibonacci numbers. For the simulation studies reported in Section~\ref{sec:sim} and Appendix~\ref{sec:add:sim}, we consider $\mathcal G = \{G_m, \, 1 \le m \le 4: \, G_1 < \ldots < G_4\}$ where $G_m = G_{m - 1} + G_{m - 2}$ for $m \ge 2$ with
$G_0 = G_1 = 20 + 10 \lfloor n/1000 \rfloor$. For other tuning parameters, we adopt the recommended default values of the \verb!R! package \verb!mosum! \citep{meier2021mosum}, and set $\alpha = 0.05$ and $\eta = 0.4$.
\vspace{-10pt}
\paragraph{Tuning parameters for the WBS2 algorithm.}
For the threshold, we set the constant $C = 1.3$ and draw $R = 100$ deterministic intervals at each iteration, which are suggested choices in \cite{fryzlewicz2014wild} and \cite{cho2021multiple} respectively.
As we permit the presence of serial correlations,
it is reasonable to impose a minimum length requirement
on the intervals considered in the WBS2 algorithm.
We set this minimum length to be $2 G_1$, with $G_1$ the finest scale considered by the MOSUM procedure.
\vspace{-10pt}
\paragraph{Window size $W$ for time-varying TAVC estimation.}
We utilise a scale-dependent window size $W_{L}$. Setting $W_{L}= N_2 L$ gives $N_3 = 2N_2-1$ data points used in the solving of the $M$-estimation equations. We advise setting $N_2 = 5$, which ensures that the influence of change points is negated and that the window size is large enough to include enough data points for reliable estimation of the TAVC.
\subsection{Simulation study}
\label{sec:sim}
In this section, we evaluate the performance of the proposed robust estimator of scale-dependent TAVC applied with the two multiscale change point detection procedures discussed in Sections~\ref{sec:mosum}--\ref{sec:wbs}. We compare with other methods that account for serial dependence under~\eqref{eq:model} and whose implementations are readily available in \verb!R!,
with a variety of scenarios for generating serially correlated $\{\varepsilon_t\}_{t \in \mathbb{Z}}$.
\subsubsection{Settings}
We assess the performance of different methods both in the case of no changes ($q = 0$) and multiple changes ($q \ge 1$), under a variety of error structures.
Unless stated otherwise, we generate
$W_t \sim_{\text{i.i.d.}} \mathcal N(0, \sigma_w^2)$ with $\sigma_w = 1$.
\begin{enumerate}[label = (M\arabic*)]
\item \label{model-a} $\varepsilon_t = W_t$.
\item \label{model-b} $\varepsilon_t = W_t$, where $W_t$ are i.i.d.\ $t_{5}$-distributed random variables.
\item \label{model-c} AR(1) model: $\varepsilon_t = a_1 \varepsilon_{t-1}+W_t$, with $a_1 = 0.9$ and $\sigma_w = \sqrt{1-a_1^2}$.
\item \label{model-d} AR(2) model: $\varepsilon_t = a_1 \varepsilon_{t-1}+ a_2 \varepsilon_{t-2}+W_t$, with $a_1 = 0.5$ and $a_2=0.3$, with $\sigma_w = 0.6676184$.
\item \label{model-e} MA(1) model: $\varepsilon_t = W_t + b_1 W_{t-1}$, with $b_1 = -0.9$.
\item \label{model-f} ARCH(1) model: $\varepsilon_t = \sigma_t W_t$ with $\sigma_t^2 = c_0 + c_1 \varepsilon_{t-1}^2$, where $c_0= 0.5$, $c_1 = 0.4$.
\item \label{model-g} Time-varying AR(1) model: $\varepsilon_t = a_1 (t) \varepsilon_{t-1}+W_t$, with $a_{1} (t) = 0.8-0.6t/n$.
\item \label{model-h} Time-varying AR(1) model: $\varepsilon_t = a_1 (t) \varepsilon_{t-1}+ \sigma (t) W_t$, with $a_{1} (t) = 0.5 \cos (2\pi t/n)$ and $\sigma (t) = \sqrt{1-a_1 (t)^2}$.
\item \label{model-i} Time-varying MA(1) model: $\varepsilon_t = W_t + b_1 (t) W_{t-1}$, with $b_1 (t) = 12(t/n)^3 -18(t/n)^2+6t/n$.
\end{enumerate}
Models~\ref{model-a}--\ref{model-f} represent stationary error scenarios.
Model~\ref{model-a} is the setting commonly adopted in the literature while \ref{model-b}, taken from \cite{cho2021data}, is adopted to examine whether a method works well in the presence of non-Gaussian errors.
Models~\ref{model-c} and~\ref{model-d} allow strong autocorrelations in $\{\varepsilon_t\}_{t = 1}^n$. Under Model~\ref{model-e}, the LRV is close to zero, which makes its accurate estimation difficult. Model~\ref{model-f} is a non-linear process. Models~\ref{model-g}--\ref{model-i} consider time-varying dependence structure; variants of \ref{model-g} and \ref{model-h} were studied in \cite{mcgonigle2021detecting} and \cite{cho2021multiple}, respectively.
For Models~\ref{model-a}--\ref{model-f}, we use the global scale-dependent TAVC estimator described in Section~\ref{sec:robust} while for~\ref{model-g}--\ref{model-i}, we use the window-based estimator of the local scale-dependent TAVC described in Section~\ref{sec:nonstat:ext}.
We assess the performance of the methods both when $q = 0$ and $q \ge 1$.
In the latter case, the time series contains the $q$ change points at $\tau_i = \lfloor n/(q + 1) \cdot i \rfloor, \, i = 1, \ldots, q$, with the (signed) change size $\mu_i = \mu(\tau_i) \cdot (-1)^{i + 1}$.
In~\ref{model-a}--\ref{model-d} and \ref{model-f}, we set $\mu(\tau_i) = \sigma$
and in the case of~\ref{model-e}, we set $\mu = 1$.
In~\ref{model-g}--\ref{model-i}, we set $\mu(\tau_i) = \sigma(\tau_i)$
where $\sigma^2(t)$ denotes the time-varying LRV.
We implement the robust TAVC estimation within both the multiscale MOSUM and WBS2 procedures as described in Sections~\ref{sec:mosum} and~\ref{sec:wbs},
which are referred to as MOSUM.TAVC$_{[\ell]}$ and WBS2.TAVC$_{[\ell]}$, respectively.
Here, the subscript with $\ell = 1, 2$, refers to the choice of the tuning parameter $\bar{\xi}_{[\ell], b}$ involved in the parameter $v$,
see~\eqref{eq:v:one}--\eqref{eq:v:two}.
For the choice of the tuning parameters, we refer to Section~\ref{sec:tuning}.
For illustrative purposes, we also report the results of the `oracle' versions of the MOSUM and WBS2 procedures referred to as MOSUM.oracle and WBS2.oracle, respectively.
These methods are implemented with the true LRV $\sigma^2$
($\sigma^2(t)$ in the case of Models~\ref{model-g}--\ref{model-i})
for standardising the detector statistics,
while all other tuning parameters are kept the same.
Additionally, we consider DepSMUCE \citep{dette2020multiscale}, DeCAFS \citep{romano2021detecting} and WCM.gSa \citep{cho2021multiple} for comparison.
DepSMUCE extends SMUCE \citep{frick2014multiscale} to dependent data using a difference-type estimator of the LRV. Although not its primary objective, DeCAFS detects multiple change points in the mean assuming that the noise is a stationary AR($1$) process. The WCM.gSa method performs model selection on the sequence of models generated by the WBS2 algorithm, using an information criterion-based model selection strategy which assumes that $\{\varepsilon_t\}_{t \in \mathbb{Z}}$ follows an AR model of an arbitrary order. For DepSMUCE, we consider significance levels $\alpha \in \{ 0.05, 0.2 \}$. Other tuning parameters not mentioned here are chosen as recommended by the authors.
\subsubsection{Results}
Table~\ref{table:sim:results} summarises the results of the comparative simulation study
from $1000$ replications of time series of length $n = 1000$
generated as in~\ref{model-a}--\ref{model-i} with $q \in \{0, 4\}$.
Results for other values of $n \in \{500, 2000\}$
are given in Appendix~\ref{sec:add:sim},
{where we make similar observations as below.}
When $q = 0$, we report the proportion of falsely detecting any change point out of the $1000$ realisations (see the column `Size' in Table~\ref{table:sim:results}).
When $q \ge 1$, we report the relative mean squared error (RMSE)
\begin{align*}
\sum_{t=1}^{n} (\widehat{f}_{t} - f_{t})^{2} / \sum_{t=1}^{n} (\widehat{f}^{*}_{t} - f_{t})^{2},
\end{align*}
where $\widehat{f}_{t}$ is the piecewise constant signal constructed with the estimated change points, and {$\widehat{f}^*_t$} is the oracle estimator constructed with the true change points.
We also report the distribution of the estimated number of change points, as well as the covering metric (CM). The covering metric \citep{arbelaez2010contour} measures the quality of the resulting segmentation as defined by the location of the detected changes, and is recommended in \cite{van2020evaluation} as an evaluation metric for comparing change point detection algorithms. The CM take values between $0$ and $1$, with a value of $1$ corresponding to a perfect segmentation. Its explicit definition can be found in Appendix~\ref{sec:add:sim}.
For each measure, we report its average over the $1000$ realisations.
\begingroup
{\small
\setlength{\tabcolsep}{3pt}
\setlength{\LTcapwidth}{\textwidth}
\begin{longtable}{c c c ccccc cc}
\caption{Performance comparisons for $n = 1000$. We report the size, the proportion of realisations where change points are falsely detected when $q = 0$,
and the distribution of the estimated number of change points, covering metric (CM) and relative MSE (RMSE) over 1000 realisations when $q = 4$.
The modal value of the distribution of the number of estimated change points for each method is shown in bold. The best performing method according to each metric when $q = 4$ is underlined.}
\label{table:sim:results}
\endfirsthead
\endhead
\toprule
&&& \multicolumn{5}{c}{$\widehat{q} - q$} & & \\
Model & Method & Size & $\leq-2$ & $-1$ & $\mathbf{0}$ & $1$ & $\geq 2$ & CM & RMSE \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-a} & MOSUM.TAVC$_{[1]}$ & 0.135 & 0.000 & 0.006 & \bf{0.980} & 0.014 & 0.000 & 0.967 & 6.098 \\
& MOSUM.TAVC$_{[2]}$ & 0.091 & 0.000 & 0.015 & \bf{0.978} & 0.007 & 0.000 & 0.965 & 6.234 \\
& WBS2.TAVC$_{[1]}$ & 0.049 & 0.000 & 0.004 & {\bf{0.996}} & 0.000 & 0.000 & {0.976} & {4.605}
\\
& WBS2.TAVC$_{[2]}$ & 0.028 & 0.001 & 0.017 & \bf{0.982} & 0.000 & 0.000 & 0.973 & 4.787
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.010 & 0.000 & 0.014 & \bf{0.986} & 0.000 & 0.000 & 0.970 & 5.053 \\
& DepSMUCE(0.2) & 0.066 & 0.000 & 0.001 & {\bf{0.998}} & 0.001 & 0.000 & 0.974 & 4.763 \\
& DeCAFS & 0.015 & 0.000 & 0.000 & \bf{0.970} & 0.029 & 0.001 & {0.976} & 4.798 \\
& WCM.gSa & 0.007 & 0.000 & 0.000 & \bf{0.978} & 0.020 & 0.002 & 0.975 & {4.733} \\
& MOSUM.oracle & 0.040 & 0.000 & 0.000 & \bf{0.861} & 0.132 & 0.007 & 0.965 & 5.782 \\
& WBS2.oracle & 0.004 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.977} & \underline{4.567} \\ \cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-b} & MOSUM.TAVC$_{[1]}$ & 0.149 & 0.001 & 0.004 & \bf{0.979} & 0.015 & 0.001 & 0.967 & 6.512 \\
& MOSUM.TAVC$_{[2]}$ & 0.086 & 0.003 & 0.008 & \bf{0.981} & 0.008 & 0.000 & 0.965 & 6.595 \\
& WBS2.TAVC$_{[1]}$ & 0.040 & 0.001 & 0.006 & {\bf{0.993}} & 0.000 & 0.000 & {0.976}
& {4.702} \\
& WBS2.TAVC$_{[2]}$ & 0.014 & 0.004 & 0.011 & \bf{0.985} & 0.000 & 0.000 & 0.974 & 4.899
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.586 & 0.000 & 0.004 & \bf{0.610} & 0.161 & 0.225 & 0.943 & 7.003 \\
& DepSMUCE(0.2)& 0.747 & 0.000 & 0.000 & \bf{0.420} & 0.180 & 0.400 & 0.932 & 6.666 \\
& DeCAFS & 0.898 & 0.000 & 0.000 & 0.105 & 0.040 & \bf{0.855} & 0.886 & 29.258 \\
& WCM.gSa & 0.009 & 0.000 & 0.000 & {\bf{0.978}} & 0.020 & 0.002 &{0.975} & {4.789} \\
& MOSUM.oracle & 0.037 & 0.000 & 0.000 & \bf{0.843} & 0.143 & 0.014 & 0.964 & 6.300 \\
& WBS2.oracle & 0.012 & 0.000 & 0.000 & \underline{\bf{0.996}} & 0.004 & 0.000 & \underline{0.978} & \underline{4.650} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-c} & MOSUM.TAVC$_{[1]}$ & 0.147 & 0.000 & 0.000 & \bf{0.998} & 0.002 & 0.000 & 0.995 & 1.810 \\
& MOSUM.TAVC$_{[2]}$ & 0.082 & 0.000 & 0.001 & \bf{0.999} & 0.000 & 0.000 & 0.994 & 1.789 \\
& WBS2.TAVC$_{[1]}$ & 0.062 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.998} & 1.258 \\
& WBS2.TAVC$_{[2]}$ & 0.034 & 0.000 & 0.001 & {\bf{0.999}} & 0.000 & 0.000 & \underline{0.998} & 1.264
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.968 & 0.000 & 0.000 & {\bf{0.920}} & 0.078 & 0.002 & 0.985 & 2.743 \\
& DepSMUCE(0.2) & 0.996 & 0.000 & 0.000 & \bf{0.739} & 0.232 & 0.029 & 0.971 & 3.123 \\
& DeCAFS & 0.597 & 0.000 & 0.000 & \bf{0.590} & 0.342 & 0.068 & {0.982} & \underline{1.236} \\
& WCM.gSa & 0.053 & 0.000 & 0.000 & \bf{0.731} & 0.135 & 0.134 & 0.959 & 2.086 \\
& MOSUM.oracle & 0.001 & 0.000 & 0.000 & \bf{0.891} & 0.097 & 0.012 & 0.988 & 2.012 \\
& WBS2.oracle & 0.000 & 0.000 & 0.000 & {\bf{0.979}} & 0.021 & 0.000 & 0.997 & 1.321 \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-d} & MOSUM.TAVC$_{[1]}$ & 0.123 & 0.000 & 0.003 & {\bf{0.992}} & 0.005 & 0.000 & {0.987} & 3.133 \\
& MOSUM.TAVC$_{[2]}$ & 0.073 & 0.001 & 0.004 & \bf{0.992} & 0.003 & 0.000 & 0.986 & 3.232 \\
& WBS2.TAVC$_{[1]}$ & 0.053 & 0.001 & 0.000 & {\bf{0.999}} & 0.000 & 0.000 & {0.994}
& {1.715} \\
& WBS2.TAVC$_{[2]}$ & 0.035 &0.003 & 0.002& \bf{0.995} & 0.000 & 0.000 & {0.993} & 1.755
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.678 & 0.000 & 0.000 & \bf{0.979} & 0.021 & 0.000 & {0.986} & 2.932 \\
& DepSMUCE(0.2) & 0.907 & 0.000 & 0.000 & \bf{0.891} & 0.105 & 0.004 & 0.980 & 3.183 \\
& DeCAFS & 0.734 & 0.000 & 0.000 & 0.241 & 0.224 & \bf{0.535} & 0.935 & {2.258} \\
& WCM.gSa & 0.022 & 0.000 & 0.000 & \bf{0.778} & 0.111 & 0.111 & 0.966 & 2.684 \\
& MOSUM.oracle & 0.008 & 0.000 & 0.000 & \bf{0.972} & 0.028 & 0.000 & 0.986 & 3.087 \\
& WBS2.oracle & 0.001 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.995} & \underline{1.705} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-e} & MOSUM.TAVC$_{[1]}$ & 0.120 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & 0.990 & 89.580 \\
& MOSUM.TAVC$_{[2]}$ & 0.069 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & 0.990 & 89.580 \\
& WBS2.TAVC$_{[1]}$ & 0.103 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.992} & 76.922 \\
& WBS2.TAVC$_{[2]}$ & 0.052 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.992} & 76.922
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05)& 0.998 & 0.000 & 0.000 & 0.036 & 0.048 & \bf{0.916} & 0.771 & 611.183 \\
& DepSMUCE(0.2) &1.000 & 0.000 & 0.000 & 0.003 & 0.004 & \bf{0.993} & 0.633 &1038.395 \\
& DeCAFS & 0.001 & 0.000 & 0.000 & \bf{0.997} & 0.003 & 0.000 & \underline{0.992} & 77.886 \\
& WCM.gSa & 0.000 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.992} & \underline{76.814} \\
& MOSUM.oracle & 1.000 & 0.000 & 0.000 & 0.000 & 0.000 & \bf{1.000} & 0.276 & 247.634 \\
& WBS2.oracle & 1.000 & 0.000 & 0.000 & 0.000 & 0.000 & \bf{1.000} & 0.278 & 207.555 \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-f} & MOSUM.TAVC$_{[1]}$ & 0.168 & 0.000 & 0.000 & \bf{0.978} & 0.021 & 0.001 & 0.973 & 6.246 \\
& MOSUM.TAVC$_{[2]}$ & 0.112 & 0.000 & 0.001 & {\bf{0.993}} & 0.006 & 0.000 & 0.973 & 6.246 \\
& WBS2.TAVC$_{[1]}$ & 0.064 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.981} & \underline{4.833} \\
& WBS2.TAVC$_{[2]}$ & 0.030 & 0.000 & 0.001 & {\bf{0.999}} & 0.000 & 0.000 & \underline{0.981} & {4.844}
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.507 & 0.000 & 0.000 & \bf{0.740} & 0.176 & 0.084 & 0.960 & 6.920 \\
& DepSMUCE(0.2) & 0.716 & 0.000 & 0.000 & \bf{0.564} & 0.252 & 0.184 & 0.947 & 7.630 \\
& DeCAFS & 0.755 & 0.000 & 0.000 & 0.191 & 0.066 & \bf{0.743} & 0.911 & 25.250 \\
& WCM.gSa & 0.021 & 0.000 & 0.000 & \bf{0.971} & 0.018 & 0.011 & {0.978} & {5.406} \\
& MOSUM.oracle & 0.021 & 0.000 & 0.000 & \bf{0.881} & 0.109 & 0.010 & 0.969 & 6.300 \\
& WBS2.oracle & 0.005 & 0.000 & 0.000 & \bf{0.995} & 0.005 & 0.000 & \underline{0.981} & {4.863} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-g} & MOSUM.TAVC$_{[1]}$ & 0.247 & 0.000 & 0.002 & \bf{0.970} & 0.025 & 0.003 & 0.973 & 5.232 \\ & MOSUM.TAVC$_{[2]}$ & 0.171 & 0.000 & 0.004 & \bf{0.972} & 0.023 & 0.001 & 0.972 & 5.310 \\
& WBS2.TAVC$_{[1]}$ & 0.184 & 0.000 & 0.004 & \bf{0.988} & 0.008 & 0.000 & {0.971} & {5.004} \\
& WBS2.TAVC$_{[2]}$ & 0.125 &0.000 & 0.010 & {\bf{0.987}} & 0.003 & 0.000 & 0.970 & 5.114
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.747 & 0.000 & 0.175 & \bf{0.718} & 0.107 & 0.000 & 0.925 & 8.377 \\
& DepSMUCE(0.2) & 0.921 & 0.000 & 0.019 & \bf{0.716} & 0.258 & 0.007 & 0.956 & 5.974 \\
& DeCAFS & 0.830 & 0.000 & 0.000 & \bf{0.652} & 0.171 & 0.177 & 0.958 & 6.217 \\
& WCM.gSa & 0.471 & 0.059 & 0.041 & \bf{0.830} & 0.044 & 0.026 & 0.941 & 6.975 \\
& MOSUM.oracle & 0.015 & 0.000 & 0.000 & \bf{0.903} & 0.094 & 0.003 & 0.969 & 5.312 \\
& WBS2.oracle & 0.005 & 0.000 & 0.000 & \underline{\bf{1.000}} & 0.000 & 0.000 & \underline{0.975} & \underline{4.884} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-h} & MOSUM.TAVC$_{[1]}$ & 0.244 & 0.000 & 0.001 & \bf{0.947} & 0.052 & 0.000 & 0.969 & 6.884 \\
& MOSUM.TAVC$_{[2]}$ & 0.154 & 0.000 & 0.001 & \bf{0.961} & 0.038 & 0.000 & 0.969 & 6.918 \\
& WBS2.TAVC$_{[1]}$ & 0.160 & 0.000 & 0.002 & \bf{0.995} & 0.003 & 0.000 & {0.967} & {6.565} \\
& WBS2.TAVC$_{[2]}$ & 0.107 & 0.000 & 0.003 & {\bf{0.994}} & 0.003 & 0.000 & 0.967 & 6.658
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.219 & 0.032 & \bf{0.641} & 0.306 & 0.019 & 0.002 & 0.815 & 17.980 \\
& DepSMUCE(0.2) & 0.483 & 0.000 & 0.252 & \bf{0.633} & 0.105 & 0.010 & 0.902 & 10.728 \\
& DeCAFS & 0.341 & 0.001 & 0.002 & \bf{0.652} & 0.153 & 0.192 & 0.953 & 9.188 \\
& WCM.gSa & 0.173 & 0.076 & 0.159 & \bf{0.748} & 0.011 & 0.006 & 0.913 & 8.976 \\
& MOSUM.oracle & 0.045 & 0.000 & 0.000 & \bf{0.841} & 0.142 & 0.017 & 0.962 & 7.206 \\
& WBS2.oracle & 0.013 & 0.000 & 0.000 & \underline{\bf{0.996}} & 0.004 & 0.000 & \underline{0.971} & \underline{5.960} \\ \cmidrule(lr){1-2} \cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
\ref{model-i} & MOSUM.TAVC$_{[1]}$ & 0.311 & 0.000 & 0.008 & \bf{0.915} & 0.076 & 0.001 & \underline{0.964} & 7.027 \\
& MOSUM.TAVC$_{[2]}$ & 0.204 & 0.000 & 0.020 & \bf{0.931} & 0.049 & 0.000 & 0.962 & \underline{7.026} \\
& WBS2.TAVC$_{[1]}$ & 0.234 & 0.000 & 0.020 & \underline{\bf{0.972}} & 0.008 & 0.000 & {0.958} & 8.451 \\
& WBS2.TAVC$_{[2]}$ & 0.167 & 0.000 & 0.029 & \bf{0.968} & 0.003 & 0.000 & 0.958 & 8.143
\\
\cmidrule(lr){3-3} \cmidrule(lr){4-8} \cmidrule(lr){9-10}
& DepSMUCE(0.05) & 0.130 & 0.181 & \bf{0.806} & 0.013 & 0.000 & 0.000 & 0.742 & 13.815 \\
& DepSMUCE(0.2) & 0.380 & 0.022 & \bf{0.874} & {0.094} & 0.010 & 0.000 & 0.773 & 12.108 \\
& DeCAFS & 0.075 & 0.034 & 0.414 & \bf{0.431} & 0.078 & 0.043 & 0.858 & 10.028 \\
& WCM.gSa & 0.054 & 0.057 & \bf{0.695} & {0.236} & 0.009 & 0.003 & 0.814 & 10.822 \\
& MOSUM.oracle & 0.101 & 0.000 & 0.000 & \bf{0.760} & 0.211 & 0.029 & 0.957 & 7.436 \\
& WBS2.oracle & 0.080 & 0.000 & 0.000 & \bf{0.827} & 0.171 & 0.002 & 0.950 & 7.927
\\
\bottomrule
\end{longtable}}
\endgroup
Overall, WBS2.TAVC displays better size control than MOSUM.TAVC.
The multiscale MOSUM procedure with bottom-up merging has been noted to return false positives as it accepts all estimators from the finest bandwidth; see the simulation results reported in \cite{cho2021two}.
Despite this known issue, MOSUM.TAVC shows better size control than some of the competitors such as DepSMUCE and DeCAFS. Between the two choices of the parameter $v$ used in~\eqref{eq:m:est},
the one involving~\eqref{eq:v:two} (corresponding to the subscript $2$) yields the estimator of TAVC
that returns better size control, e.g.\ closer to the nominal level $\alpha = 0.05$
for the multiscale MOSUM procedure. On the other hand, using the trimmed mean (corresponding to the subscript $1$) as in~\eqref{eq:v:one} sees improved power at the cost of larger size. This suggests that an approach combining the two choices of $v$ may yield a more balanced performance.
WBS2.TAVC performs the best across all metrics and scenarios among non-oracle methods when $q \ge 1$.
Also, we observe that using the proposed robust estimator of scale-dependent TAVC,
compares favourably to the multiscale methods applied with the true LRV
(i.e.\ MOSUM.oracle and WBS2.oracle) and
in some scenarios, the former outperforms the respective oracle counterpart.
In particular, in Scenario~\ref{model-e} where the LRV is close to $0$, plugging in its true value leads to detecting many false positives.
This shows that adopting the scale-dependent TAVC in place of the LRV for test statistic standardisation improves the finite sample performance when the change point detection procedure involves localised testing, as is the case for both the MOSUM and the WBS2 procedures.
We make a similar observation about the performance of DepSMUCE which also sets out to estimate the LRV.
DeCAFS exhibits good detection power but tends to over-estimate the number of change points as well as failing to control the size adequately
even when it is applied to the correctly specified scenario (Model~\ref{model-c}).
WCM.gSa performs well in correctly estimating the number of change points
regardless of whether $q = 0$ or $q \ge 1$. However, its performance deteriorates in the presence of nonstationarities,
see \ref{model-g}--\ref{model-i}.
Further inspection of the results under~\ref{model-i} demonstrates one advantage of the time-varying approach. Figure~\ref{fig:cpts-hist} plots the histogram of the estimated change point locations across the $1000$ replications for each of the methods. We see that the competing approaches struggle to detect the final change point due to the decreased variability towards the end of the data sequence,
whereas the proposed estimator of time-varying scale-dependent TAVC successfully extends to the locally stationary scenarios.
\begin{figure}[H]
\centering
\includegraphics[width =\textwidth]{cpts-hist-final.pdf}
\caption{Histograms plotting the change point estimators returned by different methods under Model~\ref{model-i}.}
\label{fig:cpts-hist}
\end{figure}
\subsection{Data applications}
We apply MOSUM.TAVC and WBS2.TAVC, the multiscale procedures combined with the robust estimator of scale-dependent TAVC, to two data examples.
We select the parameter $v$ in~\eqref{eq:m:est} using~\eqref{eq:v:one} and select other tuning parameters as described in Section~\ref{sec:tuning} unless specified otherwise.
\subsubsection{House price index data}
We analyse the monthly percentage changes in UK house price index (HPI), which provides insight into the estimated overall changes in house prices across the UK. The data are available from \url{https://www.gov.uk/government/statistical-data-sets/}, and a detailed description of the calculation of the HPI can be found from \cite{ukland}. The HPI series for various regions of the UK have previously been analysed in \cite{baranowski2019narrowest} and \cite{mcgonigle2021detecting}. We analyse the HPI for detached properties in the area of Somerset West and Taunton between April 1995 and February 2022 ($n = 323$).
We set the tuning parameters as described in Section~\ref{sec:tuning}
except for the window size $W = 4L$, $C = 1.15$ (for WBS2.TAVC) and $\alpha = 0.1$ (for MOSUM.TAVC) due to the short length of the time series. We combine the multiscale change point detection procedures with the robust estimator of the time-varying, scale-dependent TAVC described in Section~\ref{sec:nonstat:ext}, with the bandwidths $\mathcal G = \{ 20,40,60 \}$ for the MOSUM procedure and the minimum interval length set at $40$ for WBS2. The data are shown in Figure~\ref{fig:taunton}, with the change points detected by WBS2.TAVC and MOSUM.TAVC as well as the resulting estimated mean signal given in the top and bottom panels, respectively.
For comparison, we also apply DepSMUCE, DeCAFS and WCM.gSa to the data, see Table~\ref{table:hpi}.
\begin{figure}[H]
\centering
\includegraphics[width =0.8\textwidth]{taunton-final.pdf}
\caption{Monthly percentage change in HPI for detached properties in Somerset West and Taunton. Change point estimators returned by WBS2.TAVC (top) and MOSUM.TAVC (bottom) are denoted by dashed vertical lines, and the estimated means are given by solid line.}
\label{fig:taunton}
\end{figure}
Both WBS2.TAVC and MOSUM.TAVC detect two changes. The first change, detected in May and November 2004 for the two methods, corresponds to a decrease in the mean of the HPI series. The second change, detected in May/June 2020, may be associated with the changing consumer demand for housing in the wake of the COVID-19 pandemic. The Taunton area saw the biggest increase in overall house prices in 2021, as the ``race for space" saw buyers opt for ``more space to work from home as well as more outdoor space" \citep{guardian}.
We observe that no changes are detected by either WBS2.TAVC or MOSUM.TAVC during the 2008--2009 period associated with the financial crisis. In contrast, DepSMUCE (with $\alpha = 0.2$), DeCAFS and WCM.gSa detect changes during this time period, possibly influenced by the increased variability during the financial crisis.
Changes detected in the crisis period can be attributed to changes in variance (and autocorrelation), rather than those in mean, as noted in \cite{mcgonigle2021detecting}. We further support this interpretation by estimating the time-varying variance after adjusting for the shifts in mean using the change point estimators returned by WBS2.TAVC,
using the wavelet-based framework of \cite{nason2000wavelet} implemented for non-dyadic data as described in \cite{mcgonigle2022trend}, see Figure~\ref{fig:taunton:var}. There is a clear period of increased variance between 2007--2010, likely due to the financial crisis.
By utilising a time-varying TAVC estimator, our proposed methodology is able to capture the increased variability during this period, which ensures that potential false positives are not detected.
Furthermore, by accounting for the decrease in variability towards the end of the series, our time-varying estimator of the scale-dependent TAVC allows for the detection of a change in 2020 that is missed by other methods.
\begin{figure}[H]
\centering
\includegraphics[width =0.75\textwidth]{taunton-var-final.pdf}
\caption{Estimated variance of the mean change-adjusted Somerset West and Taunton HPI series.}
\label{fig:taunton:var}
\end{figure}
\begin{table}[htb]
\centering
\caption{Change points detected from the monthly HPI series for Somerset West and Taunton from April 1995 to February 2022.
}
\label{table:hpi}
{
\setlength{\tabcolsep}{2pt}
\begin{tabular}{ll}
\toprule
Method & Detected change points \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
WBS2.TAVC & 2004-11, 2020-06 \\
MOSUM.TAVC & 2004-05, 2020-05 \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
DepSMUCE(0.05) & 2004-11 \\
DepSMUCE(0.2) & 2008-09 \\
DeCAFS & 1999-05, 2003-01, 2008-08, 2009-01 \\
WCM.gSa & 1999-05, 2003-01, 2007-09, 2009-01, 2021-08 \\
\bottomrule
\end{tabular}}
\end{table}
\subsubsection{Nitrogen dioxide concentration in London}
We analyse the daily average concentrations of nitrogen dioxide (NO$_2$), measured in $\mu g/m^3$, recorded at Marylebone Road in London, UK. The measurements were taken from January 1st, 2000 until October 31st, 2021 ($n = 7734$).
The data set is available from \url{https://www.londonair.org.uk} and a similar dataset was analysed for shifts in the mean in \cite{cho2021multiple} using the WCM.gSa method.
The data take positive values and display both seasonality and effects due to bank holidays, as the main source of NO$_2$ emissions at the site is likely to be road traffic.
To mitigate these effects, we take the square root transform of the data and remove seasonality as described in \cite{cho2021multiple}.
We apply WBS2.TAVC and MOSUM.TAVC using the global TAVC estimator in~\eqref{eq:m:est}, with the minimum interval length set at $80$ for WBS2.TAVC and the bandwidths set as $\mathcal G = \{ 40,80,120,200,320,520,840 \}$ for the MOSUM.TAVC.
All other tuning parameters are selected as in Section~\ref{sec:tuning}.
The transformed data are shown in Figure~\ref{fig:no2}, with change points detected by WBS2.TAVC and MOSUM.TAVC and the resulting estimated mean signals given in the top and bottom panels, respectively.
For comparison, we also apply DepSMUCE, DeCAFS and WCM.gSa to the data.
Except for DeCAFS, which detects 17 change points, all methods return similar estimators. For brevity, the DeCAFS method is omitted from the results reported in Table~\ref{table:no2}.
\begin{figure}[H]
\centering
\includegraphics[width =0.95\textwidth]{no2-both-plot.pdf}
\caption{Transformed daily NO$_2$ measurements taken at Marylebone Road, London, UK. Change point estimators returned by WBS2.TAVC (top) and MOSUM.TAVC (bottom) are denoted by dashed vertical lines, and the estimated meanx are given by solid lines.}
\label{fig:no2}
\end{figure}
Both WBS2.TAVC and MOSUM.TAVC detect four change points, some of which can be linked to policy changes likely affecting the levels of air pollutants.
In February 2003, traffic management measures were introduced in central London which included modification of the pollutant filters of London buses and other heavy duty diesel vehicles, leading to an increase in their NO$_2$ emissions \citep{airquality}.
This corresponds to the change on January 31st, 2003 detected by all methods.
Also, Marylebone Road lies within the ultra low emission zone (ULEZ) that was introduced in April 2019. The ULEZ places restrictions on the levels of pollutants of vehicles travelling in the zone, and can be linked to the change on March 10th, 2019 detected by WBS2.TAVC or earlier in 2018 by other methods considering the bias in the change point estimators.
This corresponds to a decrease in the concentration of NO$_2$.
The final change point, detected by all methods on March 18th, 2020, aligns with the nationwide lockdown due to the COVID-19 pandemic on March 23rd, 2020, which resulted in drastically reduced levels of NO$_2$ throughout the UK \citep{higham2021uk}.
\begin{table}[htb]
\centering
\caption{Change points detected from the daily average concentrations of NO$_2$ at Marylebone Road in London from January 1st, 2000 to October 31st, 2021.
}
\label{table:no2}
{
\setlength{\tabcolsep}{2pt}
\begin{tabular}{ll}
\toprule
Method & Detected change points \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
WBS2.TAVC & 2003-01-31, 2010-07-25, 2019-03-10, 2020-03-18 \\
MOSUM.TAVC & 2003-01-11, 2010-03-06, 2018-12-30, 2020-03-18 \\
\cmidrule(lr){1-1} \cmidrule(lr){2-2}
DepSMUCE(0.05) & 2003-01-31, 2010-07-25, 2018-10-14, 2020-03-18 \\
DepSMUCE(0.2) & 2003-01-31, 2008-08-31, 2012-10-04, 2018-10-14, 2020-03-18 \\
WCM.gSa & 2003-01-31, 2009-12-09, 2018-10-14, 2020-03-18 \\
\bottomrule
\end{tabular}}
\end{table}
\section{Conclusions}
\label{sec:conc}
We propose an estimator of scale-dependent TAVC that is robust to the presence of (possibly) multiple mean shifts.
It is readily combined with multiscale change point detection methodologies which, by scanning for change points over data sections of varying lengths, provide good adaptivity to the problem of multiple change point detection.
We show the consistency of the proposed estimator under general assumptions permitting heavy tails and serial dependence decaying at a polynomial rate, and investigate its use with the multiscale MOSUM procedure and the WBS2 algorithm.
Through extensive numerical studies, we demonstrate the benefit of adopting the proposed estimator of scale-dependent TAVC for improved finite sample performance, as it better reflects the level of variability in the local data sections involved in the multiscale methods.
In particular, the heuristic extension to local stationarity shows promising performance which provides a natural avenue for future research.
\bibliographystyle{apalike}
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 3,520
|
{"url":"https:\/\/math.stackexchange.com\/questions\/3920923\/cross-section-of-cylinder-5x2-5y2-8z2-%E2%88%92-2xy-8yz-8zx-12x-%E2%88%92-12y-6","text":"# Cross section of cylinder $5x^2 + 5y^2 + 8z^2 \u2212 2xy + 8yz + 8zx + 12x \u2212 12y + 6 = 0$\n\nProve that\n\n$$5x^2 + 5y^2 + 8z^2 \u2212 2xy + 8yz + 8zx + 12x \u2212 12y + 6 = 0$$\n\nrepresents a cylinder whose cross-section is an ellipse of eccentricity $$\\frac{1}{\\sqrt 2}$$.\n\nI know how to find the plane of a cross section if I know the direction ratios of at least one generator of the cylinder, but I don't know how to find that out from just the equation of the cylinder alone.\n\n\u2022 the axis is parallel to $(-1,-1,1)$ which is the null direction of the Hessian matrix. The axes of the elliptical cross section are in directions $(-1,1,0)$ and $(1,1,2)$ Nov 24, 2020 at 16:50\n\u2022 Geogebra says the eccentricity of $\\langle 5x^2 + 5y^2 + 8z^2 - 2xy + 8yz + 8zx + 12x - 12y + 6,x+y-z\\rangle$ is $\\frac1{\\sqrt{2}}.$ Nov 24, 2020 at 19:30\n\u2022 My apologies. I've corrected the eccentricity. Nov 24, 2020 at 21:51\n\nLet $$f(x,y,z)=5x^2 + 5y^2 + 8z^2 \u2212 2xy + 8yz + 8zx + 12x \u2212 12y + 6$$ The axis of the cylinder can be obtained via $$f\u2019_x=f\u2019_y=f\u2019_z=0$$, which leads to its equation $$(x,y,z)=(0,2,-1)+t(1,1,-1)$$ The elliptical axes of the cross section containing the center $$(0,2,-1)$$ can be obtained via $$f\u2019_x:f\u2019_y:f\u2019_z=(x-0):(y-2):(z+1)$$ which leads to the respective equations for the major and minor axes $$(0,2,-1)+t(1,-1,0),\\>\\>\\>\\>\\>\\>\\>(0,2,-1)+t(1,1,2)$$ Then, substitute them into $$f(x,y,z)= 0$$ to get the major vertexes $$(\\pm\\frac1{\\sqrt2},2\\mp\\frac1{\\sqrt2},-1)$$ and the minor vertexes $$(\\pm\\frac1{2\\sqrt3},2\\pm\\frac1{2\\sqrt3},-1\\pm\\frac1{\\sqrt3})$$, and in turn the lengths of the respective major and minor axes $$a=2$$ and $$b={\\sqrt2}$$. Thus, the eccentricity is $$\\frac1{\\sqrt2}$$.\nThe equation can be written $$5(x+1)^2+5(y-1)^2+8z^2-2(x+1)(y-1)+8(y-1)z+8(x+1)z-6=0,$$ and we can translate.\nFrom @WillJagy's comment look at the plane $$x+y-z=0.$$ We're interested in the intersection $$\\langle 5x^2+5y^2+8z^2-2xy+8yz+8xz-6,x+y-z\\rangle.$$\nNow rotate so that the normal vector $$[\\frac1{\\sqrt{3}},\\frac1{\\sqrt{3}},-\\frac1{\\sqrt{3}}]$$ goes to $$[0,0,1],$$ by way of the rotation matrix $$\\begin{pmatrix}\\frac{1 }{\\sqrt{2}}&-\\frac{1}{\\sqrt{2}}&0\\\\ - \\frac{1}{\\sqrt{2}\\sqrt{3}}&-\\frac{1}{\\sqrt{2} \\sqrt{3}}&-\\frac{\\sqrt{2}}{\\sqrt{3}}\\\\ \\frac{1 }{\\sqrt{3}}&\\frac{1}{\\sqrt{3}}&- \\frac{1}{\\sqrt{3}}\\end{pmatrix}.$$\nThis produces an equation in the new $$XY$$-plane: $$6(X^2+2Y^2-1)=0,$$ which has eccentricity $$\\frac1{\\sqrt{2}}.$$","date":"2022-10-04 06:39:19","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 23, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.6391724348068237, \"perplexity\": 175.9414712106813}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-40\/segments\/1664030337480.10\/warc\/CC-MAIN-20221004054641-20221004084641-00432.warc.gz\"}"}
| null | null |
\section{Introduction}
When a single crystal beam is plastically bent, excess dislocations occur on certain active slip planes to reduce its energy. As a rule, the number of dislocations is huge, so, instead of a difficult task of describing individual dislocations, the use of continuum approach to find the dislocation densities in the bent state turns out more effective. The first attempt of taking into account continuously distributed excess dislocations in the plastically bent beam was made by \citet{Nye1953} who expressed the curvature of the beam caused by dislocations in terms of the dislocation density tensor bearing now his name. \citet{Read1957} and \citet{Bilby1958} have extended this result to the case when the stress due to excess dislocations does not vanish (see also \citep{Ashby1970,Evans1995,Mura1987,Wang2003}). Unfortunately, because of the lack of physically sound continuum dislocation theories, no quantitative dislocation distribution could be predicted in terms of the bending moment. The first quantitative prediction of the dislocation density of single crystal beam bent in single slip has been made in \citep{Le-nguyen2013}. The analyzed model was based on the continuum dislocation theory proposed by \citet{Berdichevsky2006a, Berdichevsky2006b} and developed further in \citep{Berdichevsky-Le2007,Kaluza-Le2011,Kochmann-le2008,Kochmann-le2009a,Kochmann-le2009b,Le-sembiring2008a,Le-sembiring2008b, Le-sembiring2009} (see also the alternative approaches in \citep{Acharya2000,Gurtin2002,Gurtin2007}). The information about dislocation distribution obtained in \citep{Le-nguyen2013} has been used to explain the formation of small angle tilt boundaries and predict the number of polygons in a polygonized beam after annealing (see also \citep{Le-nguyen2011,Le-nguyen2012}).
The real bending of single crystal beam is often more involved because two or more slip systems could be activated during the plastic bending, depending on the orientation of crystal with respect to the beam axis affecting the Schmid factors. For instance, if the fcc single crystal beam having the axis parallel to $[\bar{1}10]$-direction is bent about the $[110]$ direction, then there are two symmetric active slip systems oriented at angles $\pm 54.7^\circ $ with respect to the beam axis (see, e.g., \citep{Kysar2010}). Simulations of such bending problems require more elaborate models because of the interaction between edge dislocations of different groups. This paper aims at constructing an approximate one-dimensional theory of bending of single crystal beams having two active slip systems and, thus, extending the results obtained in \citep{Le-nguyen2013} to double slip. We incorporate the interactions of excess dislocations in a material model similar to that proposed by \citet{Le-sembiring2009}. We then use the variational-asymptotic method developed in \citep{Berdichevsky1983,Le1999} to derive the one-dimensional approximate theories from the exact three-dimensional problems in two cases: (i) the resistance to dislocation motion is negligibly small so that the displacements and plastic slips can be found by minimizing the energy functional, (ii) the rate-independent dissipation is taken into account. The obtained one-dimensional variational problems admit analytical solutions representing the smooth minimizers for the symmetric active slip systems. It is established that there exist energetic as well as dissipative thresholds for the dislocation nucleation which depend on the thickness of the beam. This exhibits the typical size effect of the gradient theory. Based on this analytical solution the deflection of the beam, the dislocation density, and the moment-curvature curve are analyzed in terms of the bending moment for different loading/unloading processes.
The paper is organized as follows. In the next Section the 3-D models of single crystal beam with continuously distributed dislocations bent in double slip are formulated. Sections 3 and 4 derive the one-dimensional equations and solve them for the case of zero dissipation. Sections 5 and 6 consider the same problems but with dissipation. Finally, Section 7 concludes the paper.
\section{3-D models of crystal beam bent in double slip}
\begin{figure}[htb]
\begin{center}
\includegraphics[width=11cm]{Beam}
\end{center}
\caption{Single crystal beam bent in double slip by a moment $M$.}
\label{fig:1}
\end{figure}
Consider a single crystal beam bent by a moment $M$. In the undeformed state the beam occupies the domain $\mathcal{V}$ of the three-dimensional euclidean space, with the cartesian coordinates $(x,y,z)\in (0,L)\times (-h/2,h/2)\times (0,a)$ where $L$, $h$, and $a$ are the length, height, and width of the beam, respectively. We will assume that $h\ll a \ll L$ and that the beam is bent under the plane strain condition. For definiteness, let the beam be clamped at $x = 0$ and subjected to a linearly distributed traction $-\tau y$ having the resultant moment $M$ at $x = L$. Besides, the upper and lower boundaries of the beam at $y= h/2$ and $y= - h/2$ are traction free. If the applied moment is small, it is natural to expect that the beam deforms elastically and the stress distribution is linear over the thickness according to the elementary beam theory. However, if the applied moment exceeds some critical value, edge dislocations with dislocation lines parallel to the $z$-axis may appear to reduce energy of the bent beam. We assume that, at the onset of yielding and during the plastic deformations, the crystal admits two active slip systems with slip planes inclined at angles $\varphi _1$ and $\varphi _2$ to the plane $y=0$ as shown in Fig.~\ref{fig:1}.
Under the plane strain state condition there are only two non-zero components of the displacements that do not depend on $z$, $u_x(x,y)$ and $u_y(x,y)$. Consequently, the non-zero components of the total strain tensor are
\begin{equation*}
\varepsilon _{xx}=u_{x,x},\quad \varepsilon _{yy}=u_{y,y},\quad \varepsilon _{xy}=\varepsilon _{yx}=\frac{1}{2}(u_{x,y}+u_{y,x}).
\end{equation*}
Throughout the paper the comma before an index is used to denote the partial derivative with respect to the corresponding coordinate. Since there are two active slip systems, the plastic distortion tensor is given by
\begin{equation*}
\beta _{ij}=\beta _1(x,y)s^1_im^1_j+\beta _2(x,y)s^{2}_{i}m^{2}_{j},
\end{equation*}
with $\mathbf{s}^{1} = (\cos \varphi_{1},\sin \varphi_{1},0)$, $\mathbf{m}^{1} = (-\sin \varphi_{1},\cos \varphi_{1},0)$, $\mathbf{s}^{2} = (\cos \varphi_{2},\sin \varphi_{2},0)$, $\mathbf{m}^{2} = (-\sin \varphi_{2},\cos \varphi_{2},0)$ being the unit vectors denoting the slip direction and the normal to the slip planes of the corresponding slip system. We call $\beta _1(x,y)$ and $\beta _2(x,y)$ the plastic slips. Thus, the finding of plastic distortion tensor reduces to determining the functions $\beta _1(x,y)$ and $\beta _2(x,y)$. Non-zero components of the plastic strain tensor, $\varepsilon ^p_{ij}=\frac{1}{2}(\beta _{ij}+\beta _{ji})$, read
\begin{align*}&\varepsilon^{p}_{xx} = -\frac{1}{2}\beta_{1}\sin 2\varphi_{1}-\frac{1}{2}\beta_{2}\sin 2\varphi_{2},\quad \varepsilon^{p}_{yy} = \frac{1}{2}\beta_{1}\sin 2\varphi_{1}+\frac{1}{2}\beta_{2}\sin 2\varphi_{2},\\
&\varepsilon^{p}_{xy} = \varepsilon^{p}_{yx} = \frac{1}{2}\beta_{1}\cos 2\varphi_{1}+\frac{1}{2} \beta_{2}\cos 2\varphi_{2}.
\end{align*}
Accordingly, the non-zero components of the elastic strain tensor, $\varepsilon ^{e}_{ij}=\varepsilon _{ij}-\varepsilon ^{p}_{ij}$, are
\begin{align*}
&\varepsilon^{e}_{xx} = u_{x,x}+\frac{1}{2}\beta_{1}\sin 2\varphi_{1}+\frac{1}{2}\beta_{2}\sin 2\varphi_{2},
\\
&\varepsilon^{e}_{yy} = u_{y,y}-\frac{1}{2}\beta_{1}\sin 2\varphi_{1}-\frac{1}{2}\beta_{2}\sin 2\varphi_{2},\\
&\varepsilon^{e}_{xy} = \varepsilon^{e}_{yx} = \frac{1}{2}\left( u_{x,y}+u_{y,x}-\beta_{1}\cos 2\varphi_{1}-\beta_{2}\cos 2\varphi_{2} \right) .
\end{align*}
The distribution of excess dislocations associated with these two active slip systems is described by the dislocation density tensor (introduced by \citet{Nye1953}, \citet{Bilby1954}, and \citet{Kroener1955}), $\alpha _{ij}=\varepsilon _{jkl}\beta _{il,k}$, whose non-zero components read
\begin{align*}
&\alpha_{xz} = \beta _{1,x}\cos^{2}\varphi_{1}+\beta_{1,y}\cos \varphi_{1}\sin \varphi_{1}+\beta_{2,x}\cos^{2}\varphi_{2}+\beta_{2,y}\cos \varphi_{2}\sin \varphi_{2} ,
\\
&\alpha_{yz} = \beta_{1,x}\cos \varphi_{1} \sin \varphi_{1}+\beta_{1,y} \sin^{2} \varphi_{1}+\beta_{2,x}\cos \varphi_{2}\sin \varphi_{2}+\beta_{2,y}\sin^{2} \varphi_{2}.
\end{align*}
Thus, we see that there are two different groups of excess edge dislocations associated with the slip systems 1 and 2, whose resultant Burgers' vectors show in the directions $\mathbf{s}^1$ and $\mathbf{s}^2$, respectively. These give two scalar dislocation densities
\begin{align}\label{2.1}
\rho_{1} = \frac{1}{b} \vert \beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin \varphi_{1} \vert, \quad\rho_{2} = \frac{1}{b}\vert \beta_{2,x}\cos \varphi_{2}+\beta_{2,y}\sin \varphi_{2} \vert ,
\end{align}
with $b$ being the magnitude of the Burgers vector.
Let us propose the free energy density of the single crystal deforming in double slip in the form (see \citep{Le-sembiring2009})
\begin{equation}\label{2.2}
\phi = \frac{1}{2} \lambda (\varepsilon^{e}_{kk})^{2}+\mu \varepsilon^{e}_{ij} \varepsilon^{e}_{ij}+\mu k \ln\frac{1}{1-\frac{\rho_{1}}{\rho_{s}}}+\mu k \ln \frac{1}{1-\frac{\rho_{2}}{\rho_{s}}}+\mu \chi\frac{\rho_{1}\rho_{2}}{\rho^{2}_{s}},
\end{equation}
with $\lambda$ and $\mu$ being the Lam\'e constants, $\rho_{s}$ the saturated dislocation density, and $k$ and $\chi $ the constants of the material. The first two terms in \eqref{2.2} represent the contribution to the energy of crystal by the elastic strain. The third and fourth terms are the contributions to the energy from the two separate groups of excess dislocations \citep{Berdichevsky2006b}. Finally, the last term takes the interaction of dislocations between these two groups into account. For small up to moderate dislocation densities the logarithmic terms can be approximated by the Taylor series as follows
\[\ln \frac{1}{1-\frac{\rho}{\rho_{s}}} \cong \frac{\rho}{\rho_{s}}+\frac{1}{2}\dfrac{\rho^{2}}{\rho^{2}_{s}}.\]
We shall use further only this approximation.
With \eqref{2.1} and \eqref{2.2} the total energy functional of the bent beam becomes
\begin{multline}
I = a\int^{L}_{0}\int^{h/2}_{-h/2} \bigg[ \frac{1}{2} \lambda (u_{x,x}+u_{y,y}) ^{2} +\mu ( u_{x,x}+\frac{1}{2}\beta_{1}\sin 2\varphi_{1}+\frac{1}{2}\beta_{2}\sin 2\varphi_{2}) ^{2}
\\
+\mu (u_{y,y}-\frac{1}{2}\beta_{1}\sin 2\varphi_{1}-\frac{1}{2}\beta_{2}\sin 2\varphi_{2}) ^{2}+\frac{1}{2}\mu ( u_{x,y}+u_{y,x}-\beta_{1}\cos 2\varphi_{1}
\\
-\beta_{2}\cos 2\varphi_{2})^{2}+\mu k\frac{\vert\beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin\varphi_{1}\vert}{b\rho_{s}}+\frac{1}{2}\mu k \frac{\left( \beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin\varphi_{1}\right)^{2} }{b^{2}\rho^{2}_{s}}
\\
+\mu k \frac{\vert\beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin \varphi_{2} \vert}{b\rho_{s}}+\frac{1}{2}\mu k\frac{\left( \beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin\varphi_{2}\right)^{2} }{b^{2}\rho^{2}_{s}}
\\
+k\mu\chi\frac{\vert\beta_{1,x}\cos\varphi_{1}+\beta_{1,y}\sin\varphi_{1}\vert \vert\beta_{2,x}\cos \varphi_{2}+\beta_{2,y}\sin \varphi_{2}\vert}{b^{2}\rho^{2}_{s}} \bigg] dxdy
\\
+a\int^{h/2}_{-h/2}\tau yu_x |_{x=L}dy. \label{2.3}
\end{multline}
The last term in \eqref{2.3} is the work done by the linearly distributed traction $-\tau y$ acting at the boundary $x=L$. If the dissipation caused by the dislocation motion is negligible, then the true displacements $u_x$, $u_y$ and plastic slips $\beta_1$, $\beta _2$ in the {\it final} state of bending minimize energy functional \eqref{2.3} among all admissible displacements and plastic slips satisfying the kinematic boundary condition. By the final state of bending we means the equilibrium state which is established at fixed loading condition after dislocation nucleation and after the movement of dislocations toward their equilibrium positions is finished. Thus, the whole process of dislocation nucleation and dislocation dynamics cannot be described by this theory. The bending moment $M=\tau ah^3/12$ is regarded as a control parameter, so one can study the evolution of the dislocation network in terms of $M$.
If the dissipation due to dislocation motion cannot be neglected, energy minimization should be replaced by the variational equation \citep{Sedov1965}
\begin{align}\label{2.4}
\delta I+a\int^{L}_{0}\int^{h/2}_{-h/2}(\frac{\partial D}{\partial\dot{\beta _1}}\delta\beta _1+\frac{\partial D}{\partial\dot{\beta _2}} \delta \beta _2)dxdy = 0.
\end{align}
The last integral in this equation describes the energy dissipation due to dislocation motion, where the dissipation function $D(\dot{\beta }_1,\dot{\beta }_2)$ is assumed to depend only on the plastic slip rates. We shall consider the simplest rate-independent theory for which
\begin{align*}
D(\dot{\beta }_1,\dot{\beta }_2) = K(\vert \dot{\beta}_{1}\vert+\vert \dot{\beta}_{2}\vert),
\end{align*}
with $K$ being the critical resolved shear stress. Then, provided the signs of $\dot{\beta }_1$ and $\dot{\beta }_2$ do not change during the evolution of $\beta _1$ and $\beta _2$, the variational equation \eqref{2.4} reduces to minimizing the following ``relaxed energy'' functional
\begin{multline}
I_d = a\int^{L}_{0}\int^{h/2}_{-h/2} \bigg[ \frac{1}{2} \lambda (u_{x,x}+u_{y,y}) ^{2} +\mu ( u_{x,x}+\frac{1}{2}\beta_{1}\sin 2\varphi_{1}+\frac{1}{2}\beta_{2}\sin 2\varphi_{2}) ^{2}
\\
+\mu (u_{y,y}-\frac{1}{2}\beta_{1}\sin 2\varphi_{1}-\frac{1}{2}\beta_{2}\sin 2\varphi_{2}) ^{2}+\frac{1}{2}\mu ( u_{x,y}+u_{y,x}-\beta_{1}\cos 2\varphi_{1}
\\
-\beta_{2}\cos 2\varphi_{2})^{2}+\mu k\frac{\vert\beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin\varphi_{1}\vert}{b\rho_{s}}+\frac{1}{2}\mu k \frac{\left( \beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin\varphi_{1}\right)^{2} }{b^{2}\rho^{2}_{s}}
\\
+\mu k \frac{\vert\beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin \varphi_{2} \vert}{b\rho_{s}}+\frac{1}{2}\mu k\frac{\left( \beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin\varphi_{2}\right)^{2} }{b^{2}\rho^{2}_{s}}
\\
+k\mu\chi\frac{\vert\beta_{1,x}\cos\varphi_{1}+\beta_{1,y}\sin\varphi_{1}\vert \vert\beta_{2,x}\cos \varphi_{2}+\beta_{2,y}\sin \varphi_{2}\vert}{b^{2}\rho^{2}_{s}}
\\
+K\, \text{sign}(\dot{\beta}_{1}) \beta_{1}+K\, \text{sign}(\dot{\beta}_{2}) \beta_{2} \bigg] dxdy+a\int^{h/2}_{-h/2}\tau yu_x |_{x=L}dy. \label{2.5}
\end{multline}
So, the true displacement and plastic slip fields in the {\it final} equilibrium state of plastic bending minimize the ``relaxed'' energy functional $I_d$ among all admissible displacements and plastic slips. Finally, if $\dot{\beta}_1=0$ and $\dot{\beta}_2=0$, then the plastic slips are frozen, while the displacements should be found by minimizing \eqref{2.3} with the frozen plastic slips. Let us mention here also the more complicate model taking into account the interaction of dislocation in dissipation leading to the latent hardening. In such model we assume the rate-idependent dissipation function in the form
\begin{equation*}
D(\dot{\beta }_1,\dot{\beta }_2) = K(\vert \dot{\beta}_{1}\vert+\vert \dot{\beta}_{2}\vert +\xi \sqrt{\vert \dot{\beta}_{1}\vert \vert \dot{\beta}_{2}\vert}).
\end{equation*}
In this paper we shall not dwell on this model.
\section{Energy minimization}
We first focus on the minimization problem \eqref{2.3}. It is convenient to introduce the following dimensionless variables and quantities
\begin{align}\label{3.1}
&\bar{x}=b\rho_sx,\quad \bar{y}=b\rho_sy,\quad \bar{h}=b\rho_sh,\quad \bar{L}=b\rho_sL,\quad\bar{\tau}=\frac{\tau}{\mu b \rho_{s}},\nonumber
\\
&\bar{u}_{x} = b\rho_{s}u_{x},\quad \bar{u}_{y} = b\rho_{s}u_{y},\quad \bar{I} = \frac{I(b\rho_{s})^{2}}{\mu a},\quad \gamma = \frac{\lambda}{\mu},
\end{align}
which simplify the minimization problem. Now the energy functional can be rewritten in the dimensionless form as follows
\begin{multline}\label{3.2}
I = \int^{L}_{0}\int^{h/2}_{-h/2} \bigg[ \frac{1}{2}\gamma (u_{x,x}+u_{y,y}) ^{2} +(u_{x,x}+\frac{1}{2} \beta_{1} \sin 2\varphi_{1}+\frac{1}{2}\beta_{2}\sin 2\varphi_{2}) ^{2}
\\+(u_{y,y}-\frac{1}{2}\beta_{1}\sin 2\varphi_{1}-\frac{1}{2}\beta_{2}\sin 2\varphi_{2}) ^{2}+\frac{1}{2}( u_{x,y}+u_{y,x}-\beta_{1}\cos 2\varphi_{1}
\\-\beta_{2}\cos 2\varphi_{2})^{2}+k\vert\beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin\varphi_{1}\vert+\frac{1}{2}k( \beta_{1,x}\cos \varphi_{1}+\beta_{1,y}\sin \varphi_{1})^{2}
\\+k\vert\beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin\varphi_{2}\vert+\frac{1}{2}k( \beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin\varphi_{2})^{2}+k\chi \vert \beta_{1,x}\cos\varphi_{1}
\\+\beta_{1,y}\sin\varphi_{1}\vert \vert\beta_{2,x}\cos\varphi_{2}+\beta_{2,y}\sin \varphi_{2}\vert \bigg]dxdy+\int^{h/2}_{-h/2}\tau yu_x |_{x=L}dy.
\end{multline}
The bar over the dimensionless quantities is dropped for short because in the subsequent analysis we shall deal only with them.
Energy functional \eqref{3.2} contains a small parameter $h$, so it can be reduced to 1-D energy functional by the variational asymptotic method (see \citep{Berdichevsky1983,Le1999}). For this purpose let us introduce the rescaled coordinate $\zeta =y/h$, $\zeta \in (-1/2,1/2)$ to fix the domain over the thickness in the passage to the limit $h\to 0$. With this new variable $\zeta $ the small parameter $h$ enters the functional explicitly through the formulas
\begin{align*}
u_{i,y} = \frac{1}{h}u_{i,\zeta},\quad \beta_{\alpha ,y} = \frac{1}{h}\beta_{\alpha ,\zeta}, \quad \alpha =1,2.
\end{align*}
Since the boundary condition at $x = L$ does not influence the inner asymptotic distributions of the displacements and the plastic slips over the thickness, we put $ \tau = 0$ for a while. At the first step of the variational-asymptotic procedure we keep the asymptotic principal terms in \eqref{3.2} to obtain
\begin{multline*}
I_{0} = h\int^{L}_{0}\int^{1/2}_{1/2}\bigg[\frac{1}{2h^{2}}\gamma(u_{y,\zeta})^{2}+\frac{1}{h^{2}}(u_{y,\zeta})^{2}+\frac{1}{2h^{2}}(u_{x,\zeta})^{2}+\frac{k}{h}\vert\beta_{1,\zeta}\sin\varphi_{1}\vert
\\+\frac{k}{2h^{2}}(\beta_{1,\zeta}\sin\varphi_{1})^{2}+\frac{k}{h}\vert\beta_{2,\zeta}\sin\varphi_{2}\vert+\frac{k}{2h^{2}}(\beta_{2,\zeta}\sin\varphi_{2})^{2}
\\
+\frac{\chi k}{h^{2}}\vert\beta_{1,\zeta}\sin\varphi_{1}\vert
\vert\beta_{2,\zeta}\sin\varphi_{2}\vert\bigg]dxd\zeta .
\end{multline*}
Functional $I_{0}$ is positive definite, so its minimum must be zero and is achieved at
$u_{x,\zeta} = u_{y,\zeta} = \beta_{1,\zeta} = \beta_{2,\zeta} = 0.$ For the bending states which we are interested in let us set at this step $u_x = 0$ and $\beta _1=\beta _2=0$ to obtain
\begin{align*}
u_{x} = 0,\quad u_{y} = v(x),\quad \beta_{1} = \beta_{2} = 0.
\end{align*}
At the second step, we fix $v(x)$ and seek the minimizer in the form
\begin{align*}
u_{x} = u_{x}^\prime (x,\zeta),\quad u_{y} = v(x)+u_{y}^\prime (x,\zeta),\quad \beta_{1} = \beta_{1}^\prime (x,\zeta),\quad \beta_{2} = \beta_{2}^\prime (x,\zeta),
\end{align*}
with $ u_{x}^\prime (x,\zeta),u_{y}^\prime (x,\zeta),\beta_{1}^\prime (x,\zeta),\beta_{2}^\prime (x,\zeta) $ being the correction terms which are small compared to $v(x)$. Substituting these formulas into \eqref{3.2} and then keeping the asymptotically leading terms containing $u_x^\prime (x,\zeta )$, $u_y^\prime (x,\zeta )$, and $\beta _\alpha ^\prime (x,\zeta )$, we obtain
\begin{align*}
I_{1} = &h\int^{L}_{0}\int^{1/2}_{1/2}\bigg[\frac{1}{2h^{2}}\gamma(u_{y,\zeta}^\prime )^{2}+\frac{1}{h^{2}}(u_{y,\zeta}^\prime )^{2}+\frac{1}{2}( \frac{1}{h}u^\prime _{x,\zeta}+v_{,x})^{2}+\frac{k}{h}\vert\beta^\prime _{1,\zeta}\sin\varphi_{1}\vert\nonumber
\\&+\frac{k}{2h^{2}}(\beta^\prime _{1,\zeta}\sin\varphi_{1})^{2}+\frac{k}{h}\vert\beta^\prime _{2,\zeta}\sin\varphi_{2}\vert+\frac{k}{2h^{2}}(\beta^\prime _{2,\zeta}\sin\varphi_{2})^{2}+\frac{\chi k}{h^{2}}\vert\beta^\prime _{1,\zeta}\sin\varphi_{1}\vert\nonumber
\\& \vert\beta^\prime _{2,\zeta}\sin\varphi_{2}\vert\bigg]dxd\zeta
\end{align*}
Since functional $ I_{1} $ is also positive definite, its minimum is again zero and is achieved at
\begin{align*}
u_{y,\zeta}^\prime = 0,\quad u_{x,\zeta}^\prime = -hv_{,x},\quad \beta^\prime_{1,\zeta} = 0,\quad\beta^\prime_{2,\zeta} = 0.
\end{align*}
This implies
\begin{align*}
u_{y}^\prime = 0,\quad u_{x}^\prime = -h\zeta v_{,x},\quad \beta^\prime_{1} = 0,\quad\beta^\prime_{2} = 0.
\end{align*}
At the third step of the variational-asymptotic procedure, we look for the minimizer in the form
\begin{align}\label{3.3}
&u_{x} = -h\zeta v_{,x}+u_{x}^{\prime \prime}(x,\zeta),\quad u_{y} = v(x)+u_{y}^{\prime \prime}(x,\zeta),
\\
&\beta_{1} = \beta_{1}^{\prime \prime}(x,\zeta),\quad\beta_{2} = \beta_{2}^{\prime \prime}(x,\zeta), \nonumber
\end{align}
with $ u_{x}^{\prime \prime}(x,\zeta)$, $u_{y}^{\prime \prime}(x,\zeta)$, $\beta_{1}^{\prime \prime}(x,\zeta) $, and $ \beta_{2}^{\prime \prime}(x,\zeta) $ being the correction terms which are assumed small compared with those found at the second step. By redefining $v(x)$ if required, we may put the following constraints on these correction terms
\begin{align}\label{3.4}
\langle u_{y}^{\prime \prime}\rangle = 0,\quad \langle \beta^{\prime \prime}_{1}\rangle = 0,\quad\langle \beta^{\prime \prime}_{2}\rangle = 0,\quad \mathrm{where}\quad \langle \cdot\rangle = \int^{1/2}_{-1/2}\cdot \, d\zeta .
\end{align}
Constraint $\langle u_y^{\prime \prime }\rangle =0$ means that $v(x)$ is the average deflection (over the thickness) of the beam. Substituting \eqref{3.3} into \eqref{3.2}, then keeping the asymptotically leading terms containing $u_x^{\prime \prime } (x,\zeta )$, $u_y^{\prime \prime } (x,\zeta )$, $\beta ^{\prime \prime }_1 (x,\zeta )$, and $\beta ^{\prime \prime }_2 (x,\zeta )$ to get
\begin{multline*}
I_{2} = h\int^{L}_{0}\int^{1/2}_{1/2}\bigg[\frac{1}{2}\gamma(-h\zeta v_{,xx}+\frac{1}{h}u_{y,\zeta}^{\prime \prime})^{2}+(-h\zeta v_{,xx}+\frac{\beta^{\prime \prime}_{1}}{2}\sin2\varphi_{1}+\frac{\beta^{\prime \prime}_{2}}{2}\sin 2\varphi_{2})^{2}
\\+( \frac{1}{h}u^{\prime \prime}_{y,\zeta}-\frac{1}{2}\beta^{\prime \prime}_{1}\sin 2\varphi_{1}-\frac{1}{2}\beta^{\prime \prime}_{2}\sin2\varphi_{2})^{2}+\frac{1}{2}(\frac{1}{h}u^{\prime \prime}_{x,\zeta}-\beta^{\prime \prime}_{1}\cos 2\varphi_{1}-\beta^{\prime \prime}_{2}\cos 2\varphi_{2})^2
\\+\frac{k}{h}\vert\beta^{\prime \prime}_{1,\zeta}\sin\varphi_{1}\vert+\frac{k}{2h^{2}}(\beta^{\prime \prime}_{1,\zeta}\sin\varphi_{1})^{2}+\frac{k}{h}\vert\beta^{\prime \prime}_{2,\zeta}\sin\varphi_{2}\vert+\frac{k}{2h^{2}}(\beta^{\prime \prime}_{2,\zeta}\sin\varphi_{2})^{2}
\\+\frac{\chi k}{h^{2}}\vert\beta^{\prime \prime}_{1,\zeta}\sin\varphi_{1}\vert \vert\beta^{\prime \prime}_{2,\zeta}\sin\varphi_{2}\vert\bigg] dxd\zeta .
\end{multline*}
Functional $I_2$ can be reduced to a functional depending only on $\beta ^{\prime \prime }_1$ and $\beta ^{\prime \prime }_2$. Indeed, fixing first $\beta ^{\prime \prime }_1$ and $\beta ^{\prime \prime }_2$ and varying this functional with respect to $u_x^{\prime \prime }$ and $u_y^{\prime \prime }$, then using the natural boundary conditions at $\zeta =\pm 1/2$, we obtain,
\begin{equation}\label{3.5}
\begin{split}
\frac{1}{h}(\gamma+2)u^{\prime \prime }_{y,\zeta} = \gamma h\zeta v_{,xx}+\beta^{\prime \prime }_{1}\sin2\varphi_{1}+\beta^{\prime \prime }_{2}\sin2\varphi_{2},
\\
\frac{1}{h}u^{\prime \prime }_{x,\zeta} = \beta^{\prime \prime }_{1}\cos2\varphi_{1}+\beta^{\prime \prime }_{2}\cos2\varphi_{2}.
\end{split}
\end{equation}
After finding $u_x^{\prime \prime }$ and $u_y^{\prime \prime }$ according to these equations, we substitute \eqref{3.5} into \eqref{3.2} and change $\zeta$ back to $y$. Since the functional does not contain $\beta ^{\prime \prime }_{1,x}$ and $\beta ^{\prime \prime }_{2,x}$, the thickness problem reduces to minimizing the following functional with respect to $\beta ^{\prime \prime }_1(x,y)$ and $\beta ^{\prime \prime }_2(x,y)$
\begin{align*}
I_{3} &= \int_{-h/2}^{h/2}\bigg[ \frac{\kappa}{2}( -2v_{,xx}y+\beta^{\prime \prime}_{1}\sin2\varphi_{1}+\beta^{\prime \prime}_{2}\sin2\varphi_{2})^{2}+k\vert\beta^{\prime \prime}_{1,y}\sin\varphi_{1}\vert
\\
&+\frac{k}{2}(\beta^{\prime \prime}_{1,y}\sin\varphi_{1})^{2}+k\vert\beta^{\prime \prime}_{2,y}\sin\varphi_{2}\vert+\frac{k}{2}(\beta^{\prime \prime}_{2,y}\sin\varphi_{2})^{2}
\\
&+\chi k\vert\beta^{\prime \prime}_{1,y}\sin\varphi_{1}\vert \vert\beta^{\prime \prime}_{2,y}\sin\varphi_{2}\vert\bigg] dy,
\end{align*}
where $ \kappa = \frac{1}{2(1-\nu)} $. Solution of this minimization problem enables one to find $\beta ^{\prime \prime }_1(x,y)$ and $\beta ^{\prime \prime }_2(x,y)$ required for the reduction to 1-D theory. The analytical solution of the thickness problem can be found in the special case of symmetric slip systems with the angles $\varphi _1=\varphi \in (0,\pi /2)$, $\varphi_{2} = \pi - \varphi $. Obviously, $\beta^{\prime \prime}_{1} = \beta = -\beta^{\prime \prime}_{2} $ in this case due to the symmetry of the problem. Thus, functional $I_{3}$ simplifies to
\begin{equation}\label{3.6}
I_{3} = \int_{-h/2}^{h/2} [\frac{\kappa}{2}( -2v_{,xx}y+2\beta \sin 2\varphi)^{2}+2k\vert\beta_{,y} \sin\varphi\vert+k(1+\chi )(\beta_{,y}\sin\varphi)^{2} ]\, dy.
\end{equation}
Functional \eqref{3.6} is similar to the functional obtained in \citep{Le-nguyen2013}, so we present the solution of the thickness problem without detailed derivation. The minimizer $\beta(x,y)$ is
\begin{equation}\label{3.7}
\beta(x,y) =
\begin{cases}
\beta_{m}(x)\quad &\text{for $y\in (-h/2,-l(x)/2)$},\\
\beta_{0}(x,y)\quad &\text{for $y\in (-l(x)/2,0)$},\\
-\beta(x,-y)\quad&\text{for $y\in (0,h/2)$}.
\end{cases}
\end{equation}
Function $\beta_{0}(x,y)$ is found to be
\begin{equation*
\beta_{0}(x,y) = \frac{v_{,xx}}{\sin 2\varphi}\frac{1}{\eta}\left( \eta y - \frac{1}{\cosh \frac{\eta l}{2}}\sinh \eta y\right),
\end{equation*}
where
\begin{equation*}
\eta = 2\sqrt{\frac{2\kappa}{k(1+\chi)}}\cos \varphi .
\end{equation*}
Function $ \beta_{m}(x) $ is given by
\begin{equation*}
\beta_{m}(x) = \frac{v_{,xx}}{\sin 2\varphi}\frac{1}{\eta}\left(-\frac{\eta l}{2} + \tanh \frac{\eta l}{2}\right) .
\end{equation*}
Finally, the length $l(x)$ should be found by solving the following transcendental equation
\begin{equation}\label{3.9}
\kappa v_{,xx}[ \frac{1}{4}(h^{2}-l^{2})+\frac{1}{\eta}( -\frac{\eta l}{2}+\tanh\frac{\eta l}{2})( h-l)]\sin 2\varphi -k\, \text{sign}(\beta_{0,y})\sin\varphi = 0.
\end{equation}
If the length $l$ is small, then the sign of $v_{,xx}$ coincides with the sign of $\beta _{0,y}$ evaluated right from the point $y=-l/2$. For definiteness let $\beta _{0,y}(-l/2+0)>0$ so that $v_{,xx}>0$. Note that, if the curvature of the beam is constant, then $l(x)$ does not depend on $x$ and remains constant over the whole length of the beam. In this case \eqref{3.9} can be regarded as the equation for $v_{,xx}$ once $l$ is known. By integrating \eqref{3.5} in our special case and taking the constraints \eqref{3.4} into account, we obtain the displacements $u_x$ and $u_y$ in the form
\begin{align*}
u_{x} &= -v_{,x}y,
\\
u_{y} &= v(x)+\frac{\gamma}{\gamma+2}\frac{1}{2}(y^{2}-\frac{h^{2}}{12})v_{,xx}+\frac{2}{\gamma+2}\left( \int^{y}_{0}\beta(x,\xi)\, d\xi-\Lambda \right) \sin2\varphi ,
\end{align*}
where $ \Lambda = \langle \int^{y}_{0}\beta(x,\xi)d\xi \rangle $.
Having found the solution of the thickness problem, let us now substitute the above formulas for the displacements together with the expression of $ \beta $ into the energy functional \eqref{3.2}. Keeping the asymptotically principal terms and integrating over the thickness we obtain the one-dimensional functional
\begin{align}\label{3.10}
J[v(x)] = \int^{L}_{0}\Phi(v_{,xx})\, dx - Mv_{,x}|_{x=L},
\end{align}
where the bending energy density reads
\begin{align*
\Phi(\omega) = c_{1}\omega^{2} + c_{2}\omega.
\end{align*}
In these formulas we use $\omega $ to denote the curvature $v_{,xx}$, $M=\tau h^3/12$ is the resultant moment, and
\begin{align}\label{3.12}
&c_{1} = 2\int^{-l/2}_{-h/2}\frac{\kappa}{2}\left( 2q_{0}-2y\right)^{2}dy+2\int^{0}_{-l/2}\bigg[ \frac{\kappa}{2}\left( 2q-2y\right)^{2}+\frac{k(1+\chi)}{4\cos^{2}\varphi}(q_{,y})^{2}\bigg]dy, \nonumber
\\&c_{2} = 2\int^{0}_{-l/2}2kq_{,y}\frac{\sin\varphi}{\sin2\varphi}dy = -\frac{2k}{\cos\varphi}q_{0} ,
\end{align}
with
\begin{align}\label{3.13}
q(y) = \frac{1}{\eta}\left( \eta y - \frac{1}{\cosh \frac{\eta l}{2}}\sinh \eta y\right) ,
\end{align}
and $ q_{0} = q(-l/2) $. Note that, as the coefficients $c_1$ and $c_2$ depends on the curvature $\omega $ through $l$, the energy density is not quadratic with respect to $\omega $. Besides, if $q(y)=0$, then $c_1=\kappa h^3/6$ and $c_2=0$, so the obtained functional reduces to the classical 1-D functional of the elastic beam as expected \citep{Le1999}.
Varying functional \eqref{3.10} with respect to the deflection $ v $, we obtain the differential equation of bending
\begin{align}\label{3.14}
m_{,xx} = 0,\quad m = \frac{\partial\Phi}{\partial\omega},
\end{align}
subject to the boundary conditions
\begin{align}\label{3.15}
\begin{cases}
&v(0) = 0,\quad v_{,x}(0) = 0,
\\&m(L)-M = 0,\quad m_{,x}(L) = 0.
\end{cases}
\end{align}
Equation \eqref{3.14}, together with the conditions \eqref{3.15}$_{2} $, implies that
\begin{align}\label{3.16}
m(\omega) = 2c_{1}\omega+\frac{dc_{1}}{d\omega}\omega^{2}+c_{2}+\frac{dc_{2}}{d\omega}\omega = M.
\end{align}
Since the bending moment $m$ is independent of $x$, the curvature must also be constant over the length of the beam. Consequently, $l$ and $\beta _0$ are independent of $x$ and the upper and lower layers of the beam are dislocation-free. Together with \eqref{3.9}, this equation determines the moment-curvature curve during the plastic bending. To plot this curve let us compute the derivatives of $c_1$ and $c_2$ with respect to $\omega $
\begin{align*}
\frac{dc_{1}}{d\omega} = \frac{dc_{1}}{dl}\frac{dl}{d\omega},\quad\frac{dc_{2}}{d\omega} = \frac{dc_{2}}{dl}\frac{dl}{d\omega}.
\end{align*}
From \eqref{3.9} we find that
\begin{align*}
\frac{d\omega}{dl} = \frac{4k\eta \tanh \tfrac{l\eta}{2} \left( \eta \left( h-l \right)\tanh \tfrac{l\eta}{2}+2 \right) }{\kappa \cos\varphi(h-l)^{2}\left( \eta(h-l)+4\tanh\tfrac{l\eta}{2}\right)^{2} }.
\end{align*}
This formula, together with \eqref{3.12}, enables one to determine $dc_1/d\omega $ and $dc_2/d\omega $, required for plotting the moment-curvature curve.
The threshold value of curvature at which dislocations begin to nucleate is calculated by letting $ l $ go to zero in \eqref{3.9}. This yields
\begin{align*}
\omega _{en} = \dfrac{2k}{\kappa h^{2}\cos\varphi} .
\end{align*}
The threshold value of moment can be computed from \eqref{3.16}. Taking into account that, at $ l = 0$, $c_{1} = \kappa h^{3}/6$, $c_{2} = 0 $, while $ dc_{1}/d\omega=dc_{2}/d\omega=0 $, we get
\begin{align*}
M_{en} = \dfrac{2 kh}{3\cos\varphi}.
\end{align*}
Returning to the original physical quantities according to \eqref{3.1} we have
\begin{equation*}
M_{en}=\frac{\tau _{en}h^3a}{12}=\frac{2\mu kah}{3b\rho _s\cos \varphi },
\end{equation*}
which exhibits explicitly the size effect. Besides, $M_{en}$ goes to infinity as the shear modulus $\mu $ goes to infinity, so this coincides well with the result obtained by \citet{Ashby1970}. Thus, if $M<M_{en}$, then $\beta =0$, so no dislocation is nucleated and we have purely elastic solution. The plastic yielding begins at $M=M_{en}$.
Combining the elastic and plastic responses together, we get the moment-curvature relation in the following form
\begin{align*
M=\begin{cases}
\tfrac{\kappa h^{3}}{3}\omega\quad &\text{for $M < M_{en}$},
\\2c_{1}\omega+\tfrac{dc_{1}}{d\omega}\omega^{2}+c_{2}+\tfrac{dc_{2}}{d\omega}\omega\quad &\text{for $M > M_{en}$}.
\end{cases}
\end{align*}
Knowing the curvature $\omega $ from $M$, we integrate the equation $v_{,xx}=\omega $ and use the boundary conditions \eqref{3.15}$_1$ to obtain the deflection of the beam
\begin{align*
v(x) = \dfrac{1}{2}\omega x^{2}.
\end{align*}
\section{Numerical simulations}
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{threshold1}
\end{center}
\caption{Function $ M_{en}(\varphi) $}
\label{fig:2}
\end{figure}
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{beta}
\end{center}
\caption{Function $\beta(y)$ for $ M = 0.0001$ and (a) $\varphi = \pi/3$, (b) $\varphi = \pi/6$}
\label{fig:3}
\end{figure}
In order to simulate the minimizer numerically, we choose $ h = 0.1$, $L = 10$, $\nu = 0.25$, $k = 10^{-4} $, and $\chi =0.8$. Fig.~\ref{fig:2} shows the plot of energetic threshold of the bending moment $M_{en}$ as function of the angle $\varphi $. It can be seen that the threshold moment becomes infinite as $\varphi $ goes to $\pi /2$ and has a local minimum at $\varphi =0$. It must be emphasized that this result is obtained under the assumption that the crystal admits two active symmetric slip systems. So, if $\varphi =\pi /2$ the slip systems merge into one that remains inactive during the whole process of bending because of its vanishing resolved shear stress (cf. with \citep{Le-nguyen2013}). However, if the slip systems are oriented unsymmetrically, then the one with largest Schmid factor will be activated at a certain finite bending moment.
For $M>M_{en}$ the plastic slip becomes non-zero. Fig.~\ref{fig:3} show the plots of plastic slip $\beta (y)$ for $M=0.0001$ at two different orientations of slip systems. The plastic strain vanishes on the middle line of the beam as expected and reaches its maximum and minimum at the traction-free faces. There are two small boundary layers near these faces where the plastic slip remains constant.
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{density}
\end{center}
\caption{Dislocation density for $ M = 0.0001$ and (a) $\varphi = \pi/3$ ($l\approx0.097$), (b) $\varphi = \pi/6$ ($l\approx0.099$)}
\label{fig:4}
\end{figure}
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{deflection}
\end{center}
\caption{Deflection of the beam}
\label{fig:5}
\end{figure}
On Fig.~\ref{fig:4}, where the dimensionless dislocation density $\rho (y)=\beta _{,y}\sin \varphi $ is plotted for $M=0.0001$ and for two different angles, it is seen that the excess edge dislocations of the same sign of each group are concentrated in the middle of the beam thickness, with maximum dislocation density achieved at $y= 0$. Although no obstacle exists on the middle line, the repulsive forces between dislocations of the same sign prevent them from colliding. Thus, the high concentration of dislocations can be regarded as the dislocation pile-up against the middle line. The dislocation free zones are $y\in (-h/2,-l/2)$ and $y\in (l/2,h/2)$.
The dimensionless deflection of the beam, $v(x)$, is shown in Fig.~\ref{fig:5} for $M=0.0001$ and (a) $\varphi =\pi/3$, (b) $\varphi = \pi/6$. Since the curvature of the beam is constant, the thickness of the dislocation zone $l$ does not depend on $x$.
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{moment-curvature}
\end{center}
\caption{Moment-curvature curve for: (a) $\varphi = \pi/3$, (b) $\varphi = \pi/6$}
\label{fig:6}
\end{figure}
On Fig.~\ref{fig:6} we show the moment-curvature curve for $\varphi =\pi /3$ and $\varphi = \pi/6$. Up to the threshold moment $M_{en}$ (corresponding to point A on this figure) the moment-curvature curve is a straight line corresponding to the linear elastic beam theory. Then the curve becomes non-linear and increasing as $M$ increases and $l$ increases from zero to $h/2$. This nonlinear portion describes the work hardening due to the dislocation pile-up against the middle line of the beam. If the bending moment is increased from zero up to the moment $M_B$ corresponding to B (the loading case) we follow the moment-curvature curve from O through A to B. Now, if we unload the beam by decreasing the bending moment from $M_B$ to zero, the curvature and moment follow the same path BAO (in the inverse direction), and at the end of the unloading process no residual curvature of the beam is observed.
\section{Non-zero dissipation}
If the dissipation cannot be neglected, the problems reduce to minimizing the relaxed energy functionals \eqref{2.5}. The signs of the dissipation terms in this functional depend on whether $\dot{\beta }_\alpha >0$ or $\dot{\beta }_\alpha <0$, $\alpha =1,2$. However, for the beam bending it is easy to see that both cases occur simultaneously during the plastic deformations. Indeed, from the elementary beam theory (and also from the previous simulations) we know that, as the bending moment is increased (loading), $\dot{\beta }_1>0$ for $y>0$ and $\dot{\beta }_1<0$ for $y<0$, while $\dot{\beta }_2<0$ for $y>0$ and $\dot{\beta }_2>0$ for $y<0$. In contrary, if the bending moment is decreased (unloading or loading in the opposite direction), $\beta _1$ is either frozen or $\dot{\beta }_1>0$ for $y<0$ and $\dot{\beta }_1<0$ for $y>0$ while the opposite happens to $\beta _2$. Since these functionals differs from each other in the loading and unloading case, the case study must be done separately.
The case study becomes simpler for the symmetric slip systems with $\varphi _1=\varphi \in (0,\pi /2)$, $\varphi _2=\pi -\varphi $. Due to the symmetry of the problem, we have $\beta _1=\beta =-\beta _2$. Therefore $K\, \text{sign}(\dot{\beta}_{1})\beta_{1} = K\, \text{sign}(\dot{\beta}_{2})\beta_{2}$. According to the above property, both terms can be replaced by $ K\, \text{sign} y \, \beta $. We use again the dimensionless quantities \eqref{3.1} and write the functional \eqref{2.5} in this special case in the form
\begin{align}
I_{d} = &\int^{L}_{0}\int^{h/2}_{-h/2} \bigg[ \dfrac{1}{2}\gamma\left(u_{x,x}+u_{y,y}\right) ^{2} +\left(u_{x,x}+\beta \sin2\varphi\right) ^{2}+\left(u_{y,y}-\beta \sin2\varphi\right) ^{2}\nonumber
\\&+\dfrac{1}{2}\left( u_{x,y}+u_{y,x}\right)^{2}+k \vert \beta_{,x}\cos\varphi+\beta_{,y}\sin\varphi\vert+\dfrac{1}{2}k\left( \beta_{,x}\cos\varphi+\beta_{,y}\sin\varphi\right)^2 \nonumber
\\& +k \vert \beta_{,x} \cos \varphi - \beta_{,y}\sin\varphi\vert+\dfrac{1}{2}k\left( \beta_{,x}\cos\varphi-\beta_{,y}\sin\varphi\right)^{2} \nonumber
\\&+k\chi \vert \beta_{,x}\cos \varphi+\beta_{,y}\sin \varphi \vert \cdot \vert \beta_{,x}\cos\varphi-\beta_{,y}\sin \varphi \vert +2\epsilon \, \text{sign}y\, \beta \bigg]dxdy \nonumber
\\&+\int^{h/2}_{-h/2}\tau yu_{x}|{x=L}dy, \label{5.0}
\end{align}
where $ \epsilon = K/\mu $. As compared to the previous functional \eqref{3.2} the additional term $2\epsilon \, \text{sign}\, y\, \beta $ does not belong to the asymptotically principal terms. Therefore, up to the second step of the variational-asymptotic procedure this term does not have any influence on the inner asymptotic expansion. At the third step, we are looking for the minimize in the form \eqref{3.3} such that the constraints \eqref{3.4} are fulfilled. Fixing $ \beta^{\prime \prime } $ and minimizing the relaxed energy with respect to $ u^{\prime \prime }_{x} $ and $ u^{\prime \prime }_{y} $ we find them in the form \eqref{3.5}. Then, the functional reduces to
\begin{align*}
I_{d} = &\int_{-h/2}^{h/2} \bigg[ \dfrac{\kappa}{2} \left( -2v_{,xx}y+2\beta^{\prime \prime } \sin 2\varphi \right)^{2}+2k\vert \beta^{\prime \prime }_{,y} \sin \varphi \vert +k(\beta^{\prime \prime }_{,y}\sin \varphi)^{2}\nonumber
\\&+\chi k\left( \beta^{\prime \prime }_{,y}\sin \varphi \right)^{2}+2\epsilon \, \text{sign} y \, \beta^{\prime \prime } \bigg] dy
\end{align*}
Now the term $ 2\epsilon\, \text{sign}y \, \beta^{\prime \prime } $ should be kept because it has the same order as the cross term $ -4\kappa v_{,xx}y\beta^{\prime \prime }\sin 2\varphi $. Up to an unessential constant we may rewrite this functional as
\begin{align}\label{5.1}
I_{d} = &\int_{-h/2}^{h/2} \bigg[ \dfrac{\kappa}{2}\left( -2v_{,xx}y+\dfrac{\epsilon \, \text{sign}y }{\kappa \sin 2\varphi}+2\beta^{\prime \prime } \sin 2\varphi\right)^{2}+2k\vert\beta^{\prime \prime }_{,y} \sin\varphi\vert \nonumber
\\&+k(1+\chi )(\beta^{\prime \prime }_{,y}\sin\varphi)^{2}\bigg]dy.
\end{align}
Functional \eqref{5.1} is similar to the functional obtained in \citep{Le-nguyen2013} for the case of non-zero dissipation, so we again present the solution of the thickness problem without detailed derivation. We use the same Ansatz \eqref{3.7} for $\beta $ in this case. Then for function $\beta _0(y)$ we find
\begin{align}\label{5.2}
\beta_{0}(y)&=\dfrac{v_{,xx}}{\sin 2\varphi}\dfrac{1}{\eta}\left(\eta y-\dfrac{1}{\cosh\tfrac{\eta l}{2}}\sinh\eta y \right) \nonumber
\\
&+\dfrac{\epsilon}{2\kappa \sin^{2}2\varphi}\left(1-\cosh\eta y-\tanh\tfrac{\eta l}{2}\sinh\eta y \right) .
\end{align}
Computing $ \beta_{0}(y) $ at $ y = -l/2 $ we get
\begin{align*}
\beta_{m}=\dfrac{v_{,xx}}{\sin 2\varphi\eta}\left( -\dfrac{\eta l}{2}+\tanh \dfrac{\eta l}{2}\right)+\dfrac{\epsilon}{2\kappa \sin^{2}2\varphi}\left( 1-\dfrac{1}{\cosh\tfrac{\eta l}{2}}\right)
\end{align*}
The transcendental equation for $l$ reads
\begin{align*
&\kappa v_{,xx} [ \dfrac{1}{4}(h^{2}-l^{2})+\dfrac{1}{\eta}( -\dfrac{\eta l}{2}+\tanh\dfrac{\eta l}{2})(h-l)] \sin 2\varphi \nonumber
\\&-\dfrac{\epsilon}{2}\dfrac{1}{\cosh\tfrac{\eta l}{2}}(h-l)-k \, \text{sign}(\beta_{1,y})\sin\varphi = 0.
\end{align*}
Knowing $\beta $, we can now determine the displacement field in accordance with \eqref{3.5}. Then, substituting this field together with $\beta (y)$ from \eqref{3.7} and \eqref{5.2} into the energy functional \eqref{5.0}, keeping the asymptotically principal terms and integrating over the thickness we obtain the 1-D functional
\begin{align}\label{5.4}
J_d[v(x)] = \int^{L}_{0}\Phi(v_{,xx})dx - Mv_{,x}|_{x=L},
\end{align}
with the bending energy density
\begin{align*
\Phi(\omega) = c_{1}\omega^{2} + c_{2}\omega.
\end{align*}
The coefficient $ c_{1} $ and $ c_{2} $ are given by
\begin{align*}
c_{1} &= 2\int^{-l/2}_{-h/2} \dfrac{\kappa}{2}\left( 2q_{0}-2y\right)^{2}dy+2\int^{0}_{-l/2} \bigg[\dfrac{\kappa}{2}\left( 2q-2y\right)^{2}+\dfrac{k(1+\chi)}{4\cos^{2}\varphi}(q_{,y})^{2}\bigg]dy,
\\
c_{2} &= 2\int^{0}_{-l/2} \bigg[ \dfrac{k}{\cos\varphi}q_{,y}+\dfrac{\epsilon}{\kappa \sin2\varphi}(\dfrac{k(1+\chi)}{4\cos^{2}\varphi}q_{,y}p_{,y}+\kappa(p(y)-1)(2q(y)
\\
&-2y) )\bigg] dy+2\int^{-l/2}_{-h/2}\bigg[\dfrac{\epsilon}{\sin2\varphi}(p_{0}-1)(2q_{0}-2y)\bigg] dy ,
\end{align*}
with
\begin{align*}
q(y)=\dfrac{1}{\eta}(\eta y-\dfrac{\sinh\eta y }{\cosh\tfrac{\eta l}{2}}),\quad p(y)=1-\cosh\eta y-\tanh\dfrac{\eta l}{2}\sinh\eta y ,
\end{align*}
and $ p_{0}=p(-l/2),\,q_{0}=q(-l/2) $. Note that the coefficients $c_1$ and $c_2$ depends on the curvature $\omega $, so the energy density is not quadratic with respect to $\omega $.
Varying functional \eqref{5.4} with respect to the deflection $v(x)$ we get the same equilibrium equation and boundary conditions as in Section 4. Consequently, the moment-curvature relation \eqref{3.16} remains valid if $c_1$ and $c_2$ from above are substituted. It turns out that the straight line $M=\kappa h^3\omega /3$ corresponding to the purely elastic solution does not intersect the moment-curvature curve corresponding to the solution with dislocations at $l=0$ for the case with non-zero dissipation. The threshold value for the curvature must therefore be calculated by solving the system of equations
\begin{equation}\label{5.5a}
\omega (l_d)=\omega_{d}, \quad m(l_d)=\kappa h^3\omega _d/3,
\end{equation}
where $l_d$ and $\omega _d$ are regarded as unknowns.\footnote{Formula for $\omega _d$ and Figure 7 plotting $M_d$ in \citep{Le-nguyen2013} are not correct and should be amended in accordance with \eqref{5.5a}.} Using \eqref{3.16} we determine also the dissipative threshold value for the bending moment $M_d$. Computing the curvature from the given moment, we find the deflection by integrating the equation $v_{,xx}=\omega $.
Consider now the following loading/unloading process. We first bend the beam slowly and successively by increasing the moment until some maximum value $M_*>M_d$ such that at the end of the loading process the plastic slip becomes $\beta _*$. Then we unload the beam by reducing the bending moment back to zero. Assuming that dislocations are frozen during this unloading process with $\beta =\beta _*$ and $\dot{\beta }=0$, we have to find the displacements of the beam by minimizing the energy functional
\begin{multline}\label{5.6}
I_{d} = \int^{L}_{0}\int^{h/2}_{-h/2} \bigg[ \dfrac{1}{2}\gamma\left(u_{x,x}+u_{y,y}\right) ^{2} +\left(u_{x,x}+\beta \sin 2\varphi\right) ^{2}+\left(u_{y,y}-\beta \sin 2\varphi\right) ^{2}
\\
+\dfrac{1}{2}\left( u_{x,y}+u_{y,x}\right)^{2}+k\vert\beta_{,x} \cos\varphi+\beta_{,y}\sin\varphi\vert+\dfrac{1}{2}k\left( \beta_{,x}\cos\varphi+\beta_{,y}\sin\varphi\right)
\\
+k\vert\beta_{,x}\cos\varphi-\beta_{,y}\sin\varphi\vert+\dfrac{1}{2}k\left( \beta_{,x}\cos\varphi-\beta_{,y}\sin\varphi\right)^{2}+k\chi\vert\beta_{,x}\cos\varphi
\\
+\beta_{,y}\sin\varphi\vert \cdot\vert\beta_{,x}\cos\varphi-\beta_{,y}\sin\varphi\vert +2\epsilon \, \text{sign}y\, \beta\bigg] dxdy+\int^{h/2}_{-h/2}\tau yu_{x}|_{x=L}dy.
\end{multline}
where, now, $\beta _*$ is fixed and does not subject to the variation. Since functional \eqref{5.6} is functional \eqref{3.2} with $\varphi_{1}=\pi-\varphi_{2}=\varphi $ and $ \beta_{1}=-\beta_{2} =\beta_{*}$, exactly the same variational-asymptotic analysis as in Sections 3 can be repeated leading to the following displacement field
\begin{align*}
&u_{x} = -v_{,x}y,
\\&u_{y} = v(x)+\tfrac{\gamma}{\gamma+2}\tfrac{1}{2}(y^{2}-\tfrac{h^{2}}{12})v_{,xx}+\tfrac{2}{\gamma+2}\left( \int^{y}_{0}\beta _*(x,\xi)d\xi-\Lambda _* \right)\sin2\varphi ,
\end{align*}
where $ \Lambda _*= \langle \int^{y}_{0} \beta _*(x,\xi)d\xi \rangle$. We substitute this displacement field into functional \eqref{5.6} and keep the asymptotically principal terms. Integrating over the thickness and neglecting the terms containing the known function $ \beta_{*} $ only, we get 1-D functional \eqref{5.4} with different coefficients $ c_{1} $ and $ c_{2} $
\begin{align*}
c_{1}=\kappa h^{3}/6,\quad c_{2}=2\int^{0}_{-l/2}\kappa\omega_{*}q_{*}(y)(-4y)dy+2\int^{-l/2}_{-h/2}\kappa\omega_{*}q_{0*}(y)(-4y)dy,
\end{align*}
where $ \omega_{*} $ is the curvature corresponding to $ M_{*} $ and
\begin{align*}
q_{*}(y)=\dfrac{1}{\eta}\left( \eta y-\dfrac{1}{\cosh\tfrac{\eta l_{*}}{2}}\sinh\eta y\right),\,q_{0*}=q_{*}(-l_*/2),
\end{align*}
$l_*$ being the length corresponding to $\beta _*$.
Thus, during this unloading process the moment-curvature relation takes the form
\begin{align*}
m = 2c_{1}\omega+c_{2}=M,
\end{align*}
so that, at the end of the unloading when $M=0$, the residual curvature is
\begin{align*}
\omega_{1}=-\dfrac{c_{2}}{c_{1}}.
\end{align*}
Note that, although the bending moment after unloading is zero, the elastic strain is not, so there is still some eigenstress in the beam.
\section{Numerical simulation}
In order to simulate the minimizer numerically, we choose the same numerical values for $h$, $L$, $\nu $, $k$, $\chi $ as in Section 4. For $\epsilon =K/\mu $ we choose the value $\epsilon =0.0001$. In order to find the dissipative threshold for the bending moment we need to solve equations \eqref{5.5a} with respect to $\omega _d$ and $l_d$. Then $M_d=\kappa h^3\omega _d/3$. Numerical simulations give for instance
\begin{equation*}
\omega _d=0.0413739, \quad l_d=0.0192525 \quad \Rightarrow M_d= 0.00001488 \quad \text{for $\varphi =\pi /3$},
\end{equation*}
and
\begin{equation*}
\omega _d=0.0669535, \quad l_d=0.0198977 \quad \Rightarrow M_d= 0.91942\times 10^{-5} \quad \text{for $\varphi =\pi /6$}.
\end{equation*}
Thus, if we increase $M$ steadily, then at the onset of plastic yielding the plastic slip and the total number of dislocations turn out to be finite. This could be explained physically in the following way: in the presence of non-zero dissipation dislocations should accumulate enough to give the crystal lower ``relaxed'' energy as compared to the energy of purely elastic bending. This means also that the states with smaller number of dislocations present energy barriers to be overcome, and to have the smaller ``relaxed'' energy the crystal should jump to the certain finite dislocation density.
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{betaD}
\end{center}
\caption{Function $\beta(y)$ for $ M = 0.0001$ and (a) $\varphi = \pi/3$, (b) $\varphi = \pi/6$}
\label{fig:8}
\end{figure}
For $M>M_{d}$ the plastic slip becomes non-zero. Fig.~\ref{fig:8} shows the plots of plastic slip $\beta (y)$ for $M=0.0001$ at two different orientations of slip systems. The plastic strain is zero on the middle line of the beam and reaches its maximum and minimum at the free faces. There are two small boundary layers near these faces where the plastic slip remains constant.
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{densityD}
\end{center}
\caption{Dislocation density for $ M = 0.0001$ and (a) $\varphi = \pi/3$ ($l\approx0.097$), (b) $\varphi = \pi/6$ ($l\approx0.099$)}
\label{fig:9}
\end{figure}
On Fig.~\ref{fig:4}, where the dimensionless dislocation density $\rho (y)=\beta _{,y}\sin \varphi $ is plotted for $M=0.0001$ and for two different angles, it is seen that the excess edge dislocations of the same sign of each group are concentrated in the middle of the beam thickness, with maximum dislocation density achieved at $y= 0$. Although no obstacle exists on the middle line, the repulsive forces between dislocations of the same sign prevent them from colliding. Thus, the high concentration of dislocations can be regarded as the dislocation pile-up against the middle line. The dislocation free zones are $y\in (-h/2,-l/2)$ and $y\in (l/2,h/2)$.
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{deflectionD}
\end{center}
\caption{Deflection of the beam}
\label{fig:10}
\end{figure}
\begin{figure}[htb]
\begin{center}
\includegraphics[width=7cm]{moment-curvatureD}
\end{center}
\caption{Moment-curvature curve during loading and unloading for: a) $\varphi = \pi/3$, b) $\varphi = \pi/6$.}
\label{fig:11}
\end{figure}
The deflection of the beam $ v(x) $ is shown in Fig.~\ref{fig:10} for $ M = 0.0001$ and (a) $\varphi = \pi/3,$ (b) $\varphi = \pi/6$. Since the curvature of the beam is constant, the thickness of the dislocation zone $l$ does not depend on $x$.
On Fig.~\ref{fig:11} we show the moment-curvature curve for $\varphi =\pi /3$ and $\varphi = \pi/6$, respectively. Up to the threshold moment $M_{d}$ (corresponding to point A on this figure) the moment-curvature curve is a straight line corresponding to the linear elastic beam theory. Then the curve becomes non-linear and, in contrast to the zero dissipation case, also non-monotone as $\omega $ increases as can be seen in curve (b). At the onset of the plastic yielding we can observe a small softening behavior of the moment-curvature curve. The further monotonously increasing portion of the moment-curvature curve describes the work hardening due to the dislocation pile-up against the middle line of the beam. If the bending moment is increased from zero up to the moment $M_B$ corresponding to B (the loading case) we follow the moment-curvature curve from O through A to B (with a possible jump in $\omega $ to pass the softening portion). If the beam is then unloaded, the moment-curvature curve becomes a straight line BC with the same slope like that of OA as shown in Fig.~\ref{fig:11}.
\section{Discussion and outlook}
In this paper the one-dimensional theory of bending of a single crystal beam with two active slip systems has been developed. The one-dimensional energy of bending can be found after the solution of the thickness problem. The latter has been solved for the symmetric slip systems. The threshold bending moment exhibiting the size effect has been found for the case without and with dissipation. We have found also the dislocation density and the moment-curvature curves at loading and unloading. It would be interesting to verify the obtained results with the measurement of excess dislocation density in a bent beam by the method of orientation imaging microscopy developed recently in \citep{Kysar2010}. The specimen could be for instance the fcc single crystal beam having the axis parallel to $[\bar{1}10]$-direction and bent about the $[110]$ direction. For the unsymmetrically oriented slip systems the thickness problem can only be solved numerically. This will be addressed in our forthcomming paper.
\bigskip
\noindent {\it Acknowledgments}
The financial support by the German Science Foundation (DFG) through the research project LE 1216/4-2 is gratefully acknowledged.
|
{
"redpajama_set_name": "RedPajamaArXiv"
}
| 4,863
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.