| { |
| "paper_id": "Y13-1022", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T13:32:12.586341Z" |
| }, |
| "title": "Some Formal Properties of Higher Order Anaphors", |
| "authors": [ |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Zuber", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "Laboratoire de Linguistique Formelle", |
| "institution": "CNRS and University Paris-Diderot", |
| "location": {} |
| }, |
| "email": "richard.zuber@linguist.univ-paris-diderot.fr" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "Formal properties of functions denoted by higher order anaphors like each other and syntactically complex expressions containing each other are studied. A partial comparison between these functions and functions denoted by (simple and complex) reflexives is draw. In particular it is shown that both types of function are predicate invariant (in a generalised sense). These results allows us to understand the anaphoric character of both reflexive and reciprocal expressions.", |
| "pdf_parse": { |
| "paper_id": "Y13-1022", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "Formal properties of functions denoted by higher order anaphors like each other and syntactically complex expressions containing each other are studied. A partial comparison between these functions and functions denoted by (simple and complex) reflexives is draw. In particular it is shown that both types of function are predicate invariant (in a generalised sense). These results allows us to understand the anaphoric character of both reflexive and reciprocal expressions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "By higher order anaphor, I mean expressions like each other, sometimes called basic higher order anaphors, and various complex expressions syntactically containing each other. These complex anaphors include Boolean compounds like each other and most students, each other and themselves and various modifications of each other like only each other or at least each other. Higher order anaphors are also expressions formed by the application of a higher order anaphoric determiner like each other's or every ...except each other to a common noun (CN). All such expressions will be called reciprocals and sentences containing them (in object position) will be called reciprocal sentences.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Higher order anaphors can be opposed to (logically) simple anaphors whose basic example is the reflexive pronoun himself/herself/themselves. This simple basic anaphora can also occur in Booleanly complex anaphors like himself and most students or in modified expressions like only himself, even themselves. So the distinction between simple and higher order anaphors is of logical nature: as we will see below functions denoted by higher order anaphors take binary relations (or binary relations and sets) as arguments and give sets of type 1 quantifiers as output whereas simple anaphors have arguments of the same type as higher order ones (that is their arguments are binary relations or sets and binary relations) but their output are sets (of individuals).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "The semantics of reciprocal sentences is a complex matter (as shown for instance in Dalrymple et al., 1998; Cable, forthcoming; Dotla\u010dil, forthcoming; Mari, forthcoming) . In fact there does not seem to be any general agreement concerning the data and the interpretation of reciprocal constructions (cf. Beck, 2000) . In this paper I am not, strictly speaking, interested in the semantics of higher order anaphors but in the formal properties of functions denoted by higher order anaphors. Two types of such properties will be discussed: those which are similar to properties of functions denoted by simple anaphors and those which make them different from functions denoted by simple anaphors. Formal properties of functions denoted by simple anaphors have been studied in Keenan (2007) , Zuber (2010b) and Zuber (2011) and some formal properties of higher order anaphors are given in Sabato and Winter (2012) and Peters and Westerst\u00e5hl (2006) . As far as I can tell, no comparison between the two types of function have been made. Moreover, only basic anaphors (that is syntactically simple anaphors) have been taken into consideration.", |
| "cite_spans": [ |
| { |
| "start": 84, |
| "end": 107, |
| "text": "Dalrymple et al., 1998;", |
| "ref_id": "BIBREF2" |
| }, |
| { |
| "start": 108, |
| "end": 127, |
| "text": "Cable, forthcoming;", |
| "ref_id": null |
| }, |
| { |
| "start": 128, |
| "end": 150, |
| "text": "Dotla\u010dil, forthcoming;", |
| "ref_id": null |
| }, |
| { |
| "start": 151, |
| "end": 169, |
| "text": "Mari, forthcoming)", |
| "ref_id": "BIBREF6" |
| }, |
| { |
| "start": 304, |
| "end": 315, |
| "text": "Beck, 2000)", |
| "ref_id": "BIBREF0" |
| }, |
| { |
| "start": 774, |
| "end": 787, |
| "text": "Keenan (2007)", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 790, |
| "end": 803, |
| "text": "Zuber (2010b)", |
| "ref_id": "BIBREF10" |
| }, |
| { |
| "start": 808, |
| "end": 820, |
| "text": "Zuber (2011)", |
| "ref_id": "BIBREF11" |
| }, |
| { |
| "start": 886, |
| "end": 910, |
| "text": "Sabato and Winter (2012)", |
| "ref_id": "BIBREF8" |
| }, |
| { |
| "start": 915, |
| "end": 944, |
| "text": "Peters and Westerst\u00e5hl (2006)", |
| "ref_id": "BIBREF7" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": "1" |
| }, |
| { |
| "text": "We will consider binary relations and functions over universe E which is supposed to be finite. If a function takes only a binary relation as argument, its type is noted 2 : \u03c4 , where \u03c4 is the type of the output; if a function takes a set and a binary relation as arguments, its type is noted 1, 2 : \u03c4 . If \u03c4 = 1 then the output of the function is a set of individuals and thus the type of the function is 2 : 1 . For instance the function SELF , defined as SELF (R) = {x : x, x \u2208 R}, is of this type. The case we will basically consider here is when \u03c4 corresponds to a set of type 1 quantifiers and thus \u03c4 equals, in Montagovian notation, e, t t t . In short, the type of such functions will be noted either 2 : 1 (functions from binary relations to sets of type 1 quantifiers) or 1, 2 : 1 (functions from sets and binary relations to sets of type 1 quantifiers).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "Let R be a binary relation. Then : dom(R) = {x : \u2203 y x, y \u2208 R} and rg(R) = {x : \u2203 y y, x \u2208 R}. Furthermore, for any a \u2208 E, aR = {x : a, x \u2208 R} and Ra = {x : x, a \u2208 R}. The relation R \u22121 is the converse of R (that is R \u22121 = { x, y : y, x \u2208 R}) and the relation R S is the maximal symmetric relation included in R, that is", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "R S = R \u2229 R \u22121 . A type 2 : 1 or type 2 : 1 function F is convertible iff F (R) = F (R \u22121 ). Relation I is defined as I = { x, x : x \u2208 E}.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "The relation R t is the transitive closure of the relation R, that is the smallest transitive relation in which R is included.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "Basic type 1 quantifiers are functions from sets (sub-sets of E) to truth-values. In this case they are denotations of subject NPs. However, NPs can also occur in oblique positions and in this case their denotations do not take sets (denotations of verb phrases) as arguments but rather denotations of intransitive verb phrases, that is relations, as arguments. To account for this eventuality it has been proposed to extend the domain of application of basic type 1 quantifiers so that they apply to n-ary relations and act as arity reducers, that is have as output an (n-1)-ary relation. Since we are basically interested in binary relations, the domain of application of basic type 1 quantifiers will be extended by adding to their domain the set of binary relations. In this case the quantifier Q can act as a\"subject\" quantifier or a \"direct object\" quantifier giving rise to the nominative case extension Q nom and accusative case extension Q acc respectively. They are defined as follows (Keenan, 1987; Keenan and Westerst\u00e5hl, 1997) :", |
| "cite_spans": [ |
| { |
| "start": 995, |
| "end": 1009, |
| "text": "(Keenan, 1987;", |
| "ref_id": null |
| }, |
| { |
| "start": 1010, |
| "end": 1039, |
| "text": "Keenan and Westerst\u00e5hl, 1997)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "D1:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "For each type 1 quantifier Q, Q nom (R) = {a : Q(Ra) = 1}.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Formal preliminaries", |
| "sec_num": "2" |
| }, |
| { |
| "text": "For each type 1 quantifier Q, Q acc (R) = {a : Q(aR) = 1}.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "From now on Q nom (R) will be noted Q(R). Nominative and accusative extensions can thus be considered as functions from binary relations to sets. By type 1 quantifiers I will mean basic type 1 quantifiers as well as their nominative and accusative extensions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "Given that type 1 quantifiers and their arguments form Boolean algebras, every quantifier Q has its Boolean complement, denoted by \u00acQ, and its post-complement Q\u00ac, defined as follows: Q\u00ac = {P : P \u2286 E \u2227 P \u2208 Q} (where P is the Boolean complement of P ). The dual Q d of the quantifier Q is, by definition,", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "Q d = \u00ac(Q\u00ac) = (\u00acQ)\u00ac. A quantifier Q is self-dual iff Q = Q d", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": ". These definitions work also for extended type 1 quantifiers. It easy to see for instance that \u00ac(Q acc ) = (\u00acQ) acc and (", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "Q d ) acc = (Q acc ) d . A type 1 quantifier Q is positive iff Q(\u2205) = 0.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "A special class of type 1 quantifiers is formed by individuals, that is ultrafilters generated by an element of E. Thus I a is an individual (generated by a \u2208 E) iff I a = {X : a \u2208 X}. Ultrafilters are special (principal) filters. A (principal) filter generated by the set A \u2286 E is the following quantifier: F t(A) = {X : X \u2286 E \u2227 A \u2286 X}. Thus ultrafilters are principal filters generated by singletons.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "One property that we will use is the property of living on. The basic type 1 quantifier lives on the set A (where A \u2286 E) iff for all X \u2286 E, Q(X) = Q(X \u2229 A). If E is finite then there is always the smallest set on which a quantifier Q lives: it is the meet of all sets on which Q lives. The fact that A is the smallest set on which the quantifier Q lives will be noted Li(Q, A). If A \u2208 Q then A is called the witness set of Q:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "A = wt(Q). The quantifier Q is called plural, noted Q \u2208 P L, iff \u2203 a,b\u2208E such that Q \u2286 I a \u2229 I b .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "Functions from pairs of sets to truth-values or binary relations between sets are type 1, 1 quantifiers. In NLs they are denoted by (unary) nominal determiners, that is expressions which take one CN as argument and give a NP as output. Denotations of nominal determiners obey various constraints. Recall first the constraint of conservativity for type 1, 1 quantifiers. A well-known definition of conservativity is given in D5:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "D3: F \u2208 CON S iff for any property X, Y one has F (X, Y ) = F (X, X \u2229 Y )", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "Definition D3 can be generalised so that it applies to type 1, 2 : \u03c4 functions (cf. Zuber 2010a):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "D4: A function F of type 1, 2 : \u03c4 is conservative iff F (X, R) = F (X, (E \u00d7 X) \u2229 R)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "Observe that the above definition does not depend on the type \u03c4 of the result of the application of the function. So obviously it can be used with higher order functions. Type 2 : 1 functions can also be (predicate or argument) invariant and invariance is a property depending on the type of the output of the function. Thus (see Keenan and Westerst\u00e5hl, 1997) :", |
| "cite_spans": [ |
| { |
| "start": 330, |
| "end": 359, |
| "text": "Keenan and Westerst\u00e5hl, 1997)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "D5: A type 2 : 1 function F is predicate invariant iff a \u2208 F (R) \u2261 a \u2208 F (S) whenever aR = aS.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "For instance the function SELF is predicate invariant. The following definitions are generalisations of predicate invariance applying to type 2 : 1 and type 1, 2 : 1 functions: D6: A type 2 : 1 function F satisfies HPI (higher order predicate invariance) iff for any positive type 1 quantifier Q, any A \u2286 E, any binary relations R, S, if A = W t(Q) and F t(A)R = F t(A)S then Q \u2208 F (R) iff Q \u2208 F (S). D7: A type 1, 2 : 1 function F satisfies D1HPI (higher order predicate invariance for unary determiners) iff for any positive type 1 quantifier Q, any A \u2286 E, any binary relations R and S, if", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "A = W t(Q 1 ) and F t(A)R \u2229 X = F t(A)S \u2229 X then Q \u2208 F (X, R) iff Q \u2208 F (X, S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D2:", |
| "sec_num": null |
| }, |
| { |
| "text": "In this section I briefly present simple syntactic, or categorial, similarities and, possibly, differences, between reflexives and reciprocals, both simple and syntactically complex.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "We will consider sentences of the form given in (1):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "(1) NP TVP GNP In this schema, GNP is a generalised noun phrase.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "GNPs are linguistic objects that can play the role of syntactic arguments of transitive verb phrases (TVPs). So \"ordinary\" NPs or DPs (determiner phrases) are GNPs. However there are genuine GNPs which differ from \"ordinary\" NPs in that they cannot play the role of all verbal arguments; in particular they cannot occur in subject position. This is the case of anaphoric expressions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "The GNPs related to reflexives and reciprocals are anaphoric noun phrases (ANPs). Roughly, their (\"referential\") meaning depends on the meaning of another expression in the sentence, the so-called antecedent of the anaphor, by which it is bound. In the simplest case the antecedent is the subject NP. Thus a more specific form of sentences that we will consider of the form given in (2) instantiated in (3) and (4):(2) NP TVP ANP.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "(3) Most students washed themselves. (4) Leo and Lea hate each other.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Thus the GNPs we consider are ANPs. In the above examples we have syntactically simple ANPs. Such ANPs can occur as syntactic parts of complex GNPs; in particular they can be parts of Boolean compounds and can be modified by categorially polyvalent modifiers such as only, also, even, at least, let alone, etc. : (5a) Leo and Lea admire themselves and most teachers. (5b) Leo and Lea admire each other, themselves and two teachers. (6) Two monks hug each other only.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "A special class of complex ANPs is formed by the application of anaphoric determiners (ADets), to CNs. Again, this can be done both with reflexive determiners and with reciprocal ones. Many languages have possessive anaphoric determiners. This is the case with Slavic languages which have the possessive \"determiner-pronoun\" SV OJ (meaning, roughly 'ones own') which can be considered as ADet with reflexive meaning (cf. Zuber, 2011) . Similarly, marking the simple reciprocal each other in English by the possessive marker results in a ADet with reciprocal meaning. This possibility is indicated in the following examples:", |
| "cite_spans": [ |
| { |
| "start": 421, |
| "end": 433, |
| "text": "Zuber, 2011)", |
| "ref_id": "BIBREF11" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "(7) Leo and Lea admire their own books. (8) Leo and Lea admire each other's books.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Thus their own in (7) is a ADet with a reflexive meaning and each other's in (8) is an ADet with reciprocal meaning.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "More interestingly it is possible to use an ordinary determiner (or its \"part\") and the simple ANPs himself/herself/themselves to form a ADet with reflexive meaning and to use an ordinary determiner and the simple ANPs each other to form an ADet with reciprocal meaning. Thus, roughly speaking (Zuber, 2010a) , if D is an ordinary one place determiner, denoting monotone increasing function, then D..., including himself or D...in addition to themselves are ADets with the reflexive meaning. If D is a determiner denoting monotone decreasing functions then D, not even himself is an ADet as well. The following sentences contain various complex ADets with reflexive meaning:", |
| "cite_spans": [ |
| { |
| "start": 294, |
| "end": 308, |
| "text": "(Zuber, 2010a)", |
| "ref_id": "BIBREF9" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "(9) Two students admire most teachers in addition to themselves and Picasso (10) Leo and Lea washed some vegetarians including at least themselves. (11) Leo and Lea admires no philosophers, not even themselves or Socrates.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "Quite similar procedure can be applied, though probably somewhat less productively, to (syntactically) simple and complex reciprocals in order to obtain ADets with reciprocal meaning. The following examples illustrate this possibility:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "(12) Two students shaved most students including each other. (13) Leo and Lea admire most logicians in addition to each other. (14) Leo and Lea admire no philosopher, let alone each other.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "As the following examples show simple and complex reflexives and reciprocals can occur also in other than direct object positions. The following example show that reflexives and reciprocals can be arguments of a verb taking three arguments: The difference pointed out by the above examples is related to the difference in the categorial status of reflexives on the one hand and reciprocals on the other. Thus it is usually assumed that ANPs with reflexive meaning are \"argument\" reducers: when applied to a di-transitive verb phrase they give a transitive verb phrase, and when applied to a transitive verb phrase they give just a VP.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "The situation with reciprocals is different. Recall that ANPs are GNPs. GNPs apply to TVPs and give VPs as result. So what is the category of such VPs. Ignoring directionality, the subject NPs in the constructions we are interested in are of the category S/(S/N P ). This means that, in order to avoid type mismatch, verb phrases must be raised and have the category S/(S/(S/N P )). Then their denotational type is e, t t t . Consequently, sentences of the form (1) are true iff the quantifier denoted by the N P is an element of the set denoted by T V P GN P . Thus ANPs with reciprocal meaning denote type 2 : 1 functions. This categorial difference is related to the following semantic difference. Consider the following examples: Clearly (25a) in conjunction with (25b) entails (26) whereas (27a) in conjunction with (27b) does not entail (28). This means that the quantifiers denoted by the subject NPs in (27a) and (27b) do not apply to the predicate denoted by the complex VPs in these sentences and that the GNPs like each other denote type 2 : 1 functions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "There are of course genuine type e, t t t (or type 2 : 1 in our notation) functions, that is such that they are not lifts of simple type 2 : 1 functions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Reciprocals and reflexives", |
| "sec_num": "3" |
| }, |
| { |
| "text": "We have seen that higher order anaphors denote type 2 : 1 functions. Any type 2 : 1 function whose output is denoted by a VP can be lifted to the type 2 : 1 function. This is in particular the case with the acusative and nominative extensions of a type 1 quantifier. For instance the accusative extension of a type 1 quantifier can be lifted to type 2 : 1 function in the way indicated in (29). Such functions will be called accusative lifts. More generally iff F is a type 2 : 1 function, its lift F L , a type 2 : 1 function, is defined in (30):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(29) Q L acc (R) = {Z : Z(Q acc (R)) = 1}. (30) F L (R) = {Z : Z(F (R)) = 1}", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The variable Z above runs over the set of type 1 quantifiers.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "As we have seen, simple reflexives are interpreted by the function SELF . This function is of type 2 : 1 , that is a function which takes binary relations as argument and gives a set as result. Complex reflexives are interpreted by corresponding Boolean combination of SELF with (lifted) denotations of NPs being a part Boolean compounds or, in the case of modification by categorially polyvalent particles, by modifications of SELF . Obviously, they are also of type 2 : 1 . These functions satisfy predicate invariance defined in D5. The function SELF , but not the functions denoted by complex reflexives, also satisfies the left predicate invariance: D 8: A type 2 : 1 function F is left predicate invariant iff for any a \u2208 E and any binary relations R, S, if Ra = Sa then a \u2208 F (R) iff a \u2208 F (S) where Ra = {x : x, a \u2208 R}.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Accusative extensions of type 1 quantifiers, which can also be considered as type 2 : 1 functions, satisfy a stronger condition than predicate invariance. They satisfy so-called accusative extension condition AE (Keenan and Westerst\u00e5hl, 1997) : D 9: A type 2 : 1 function F satisfies AC iff for any a, b \u2208 E and any binary relations R, S, if aR = bS then a \u2208 F (R) iff b \u2208 F (S).", |
| "cite_spans": [ |
| { |
| "start": 212, |
| "end": 242, |
| "text": "(Keenan and Westerst\u00e5hl, 1997)", |
| "ref_id": "BIBREF5" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "It is important (Keenan, 2007) that functions denoted by reflexive expressions, simple and complex, do not satisfy AC and thus they are different from accusative extensions of type 1 quantifiers denoted by \"ordinary\" NPs in the object position. . In that sense, reflexive expressions are also genuine GNPs.", |
| "cite_spans": [ |
| { |
| "start": 16, |
| "end": 30, |
| "text": "(Keenan, 2007)", |
| "ref_id": "BIBREF4" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The corresponding higher order extension condition is defined in D10: D10: A type 2 : 1 function F satisfies HEC (higher order extension condition) iff for any positive type 1 quantifiers Q 1 and Q 2 , any A, B \u2286 E, any binary relations R, S, if A = W t(Q 1 ) and B = W t(Q 2 ), and", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "F t(A)R = F t(B)S then Q 1 \u2208 F (R) iff Q 2 \u2208 F (S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Functions which are accusative lifts satisfy HEC. We will see that functions denoted by higher order anaphors do not satisfy HEC because functions satisfying HEC have the following obvious property:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 1: Let F be a type 2 : 1 function which satisfies HEC and let R = E \u00d7 C, for C \u2286 E arbitrary. Then for any X \u2286 E either F t(X) \u2208 F (R) or for any X, F t(X) / \u2208 F (R).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "In order to present various properties of functions denoted by higher order anaphors I will discuss only some such functions and not define all functions which constructions discussed in the PACLIC-27 previous section denote. Some other functions are discussed in Zuber (2012) .", |
| "cite_spans": [ |
| { |
| "start": 264, |
| "end": 276, |
| "text": "Zuber (2012)", |
| "ref_id": "BIBREF12" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Consider first the function given in (31):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(31) RF L-RECIP (R) = {Q :", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2203 A\u2286E A = W t(Q) \u2227 Q(Dom(A \u00d7 A) \u2229 (R \u2229 R \u22121 )) = 1}", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Informally, this function can be considered as the denotation of an anaphor like each other or oneself or themselves. In other words it does not make a priori a distinction between \"purely\" reflexive and \"purely\" reciprocal interpretation, as apparently it happens in many languages. Observe in particular that individuals can be in the output of this function. Furthermore, the meet of two individuals can be in the output of this function even if they are in the relation R with themselves only.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The following function excludes the \"reflexive part\" and interprets purely reciprocal anaphors (in their strong logical reading):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(32) SEA(R) = {Q : A = W t(Q) \u2227 |A| \u2265 2 \u2227 Q(Dom((A \u00d7 A) \u2229 (R \u2229 R \u22121 ) \u2229 I )) = 1},", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "where I is the complement of the identity relation I.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Consider now example (33), where, clearly, a Boolean composition of two higher order functions is involved, one of which is an accusative lift:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(33) Leo and Lea admire each other and most teachers.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We want to give a function interpreting the complex anaphor each other and most teachers. Obviously this function has to entail the function SEA above and be completed by the part corresponding to most teachers. It is given in (34):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(34) SEA Q (R) = {Z : Z \u2208 SEA \u2227 Z \u2208 Q L acc }", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The above functions are based on the relation R S . Sentences in (35) have somewhat illogical interpretation. Functions corresponding to these interpretations are given in (36):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(35a) Five students followed each other. (35b) All pupils followed each other and two teachers. (36a) ILEA(R) = {Z :", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2203 A\u2286E (Li(Z, A) \u2227 A \u00d7 A \u2229 I \u2286 R t } (36b) ILEA Qconj (R) = ILEA(R) \u2229 Q L acc (R)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Let us see now some constraints on the above functions. First we have:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 2: Functions RF L-RECIP , SEA and SEA Q satisfy HPI.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proof We prove only that RF L-RECIP satisfies HPI. Suppose that A = W t(Q) and that Q \u2208 REF -RECIP (R).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "We have to show that if for some binary relation S (i) holds (i):", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "\u2200 x\u2208A (xR = xS) then Q \u2208 RF L-RECIP (S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Given the definition of RF L-RECIP this happens when Q(Dom((A\u00d7A)\u2229(S \u2229S \u22121 ) = 1. But if (i) holds then (A\u00d7A)\u2229(R \u2229R \u22121 ) = (A\u00d7A)\u2229(S \u2229S \u22121 ). Hence Q \u2208 RF L-RECIP (S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "It is easy to prove, using proposition 1, that:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 3: Functions RF L-RECIP , SEA and SEA Q do not satisfy HAI.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proof : We prove only that the function RF L-RECIP does not satisfy HAI. Given its definition in (31) we can see that for C \u2286 E arbitrary , for any C 1 such that C \u2286 C 1 we have F t(C 1 ) / \u2208 RF L-RECIP (E \u00d7 C) and for any C 2 \u2286 C we have F t(C 2 ) \u2208 RF L-RECIP (E \u00d7 C). Hence, given proposition 1, RF L-RECIP does not satisfy HPI.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Here are some other properties:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 4: Let F \u2208 {RF L- RECIP, SEA, ILEA} and R = S \u22121 . Then F (R) = F (S)", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 4 has an interesting consequence: since R = (R \u22121 ) \u22121 , it follows from Proposition 2 that functions RF L-RECIP, SEA and ILEA are convertible.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The above properties do not hold for complex higher order functions that is functions denoted by syntactically complex reciprocals. For higher order functions based on the relation R S the following proposition holds:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 5:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Let F \u2208 {RF L-RECIP, SEA, SEA Q }, R = S \u22121 and Dom(R) = Dom(S). Then F (R) = F (S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "To illustrate Proposition 5 consider the following examples:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(37a) Five students followed each other. (37b) Five students preceded each other.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "If we consider that the relation expressed by follow is the converse of the relation expressed by precede that (37a) and (37b) are equivalent.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Observe that the property of functions expressed in Proposition 6 does not depend on the type of the output of the function. It is easy to see, for instance that many reflexives functions denoted by reflexives have a similar property. More precisely we have:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proposition 6: Let F \u2208 {SELF, SELF \u2297 Q acc },", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "where \u2297 is a Boolean connective, R = S \u22121 and Dom(R) = Dom(S). Then F (R) = F (S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Thus Propositions 5 and 6 express, informally, properties of functions sensitive to some aspects of their arguments only.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Conservativity, as defined in D4 is such a property. Definition of conservativity given in D4 naturally applies to functions denoted by anaphoric determiners. The conservativity of anaphoric determiners giving rise to reflexives is discussed in Zuber (2010b). We are not directly interested here in the semantics of anaphoric determiners but it would be easy to show that the anaphoric determiner Every...except each other as it occurs in (38) denotes a conservative function:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "(38) Two students washed every student except each other.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "To conclude let us see some other properties of functions denoted by anaphors. These functions are \"sensitive\" to some aspects of their arguments, that is to some properties of the binary relations to which they apply. Consider the following definition:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "D11: A type 2 : \u03c4 function F is symmetry sensitive, F \u2208 SY M S, iff F (R) = F (S) whenever R \u2229 R \u22121 = S \u2229 S \u22121 .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Functions SELF , RF L-RECIP and SEA are symmetry sensitive. Functions denoted by complex anaphors (reflexive or reciprocal) do not have this property. They have the following property: D12: A type 2 : \u03c4 function F is symmetry and range sensitive, F \u2208 SY M RS iff F (R) = F (S) whenever R \u2229 R \u22121 = S \u2229 S \u22121 and Rg(R) = Rg(S).", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Note that SY M S \u2286 SY M RS. Thus not only functions denoted by complex anaphors but also those denoted by simple anaphors are symmetry and range sensitive. This is what all anaphors have in common. In order to distinguish anaphors with purely reflexive meaning from those with purely reciprocal meaning the following definitions can be used: For instance only each other denotes a symmetry only sensitive function and himself or himself and most students denote reflexivity and range sensitive functions.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "D13: A type 2 : \u03c4 function F is symmetry only sensitive, F \u2208 SY M OS, iff F (R) = F (S) whenever R \u2229 R \u22121 \u2229 I = S \u2229 S \u22121 \u2229 I and", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Observe that SY M OS \u2286 SY M RS and REF LS \u2286 SY M RS. Similarly SY M S \u2286 SY N RS. Thus purely reflexive anaphors denote functions which are not symmetry only sensitive and purely reciprocal anaphors denote functions which are not reflexivity sensitive but both classes are symmetry and range sensitive. quantifiers) but also with expressions which denote functions which are not quantifiers (or their extensions). In that respect the reciprocals are similar to reflexives since functions interpreting reflexives (like function SELF , its modifications and its Boolean compounds) are neither quantifiers not extensions of a type 1 quantifier.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "It is well-known (Keenan 2007; Zuber 2010b ) that the existence of anaphors in NLs shows that the expressive power of natural languages would be less than it is if the only noun phrases we needed were those interpretable as subjects of main clause intransitive verbs. The reason is that anaphors like himself, herself must be interpreted by functions from relations to sets which lie outside the class of generalised quantifiers as classically defined. In this paper some preliminary results are presented to show that the existence of higher order anaphors even further extends the expressive power of NLs.", |
| "cite_spans": [ |
| { |
| "start": 17, |
| "end": 30, |
| "text": "(Keenan 2007;", |
| "ref_id": "BIBREF4" |
| }, |
| { |
| "start": 31, |
| "end": 42, |
| "text": "Zuber 2010b", |
| "ref_id": "BIBREF10" |
| } |
| ], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The move to consider that higher order anaphors denote genuine type 2 : 1 functions allows us to understand the \"non-Boolean\" behaviour of the conjunction and in their context. Observe, for instance, that (39a) in conjunction with (39b) does not entail 40 Functions denoted by higher order anaphors satisfy higher order invariance: they are predicate invariant in a generalised sense. They are different from quantifiers denoted by NPs on the direct object position because they do not satisfy the higher order accusative extension which accusative lifts satisfy. In that respect they similar to functions denoted by simple anaphors (reflexives) which are predicate invariant and do not satisfy the accusative extension condition.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Various conservativity-like properties of functions denoted by reciprocals have been also exhibited. Thus it has been indicated that both types of anaphoric determiners, those giving rise to reflexives and those giving rise to reciprocals, denote conservative functions. Moreover, it has been formally expressed how both types of functions are \"sensitive'\" only to some aspects of binary relations which are their arguments.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Higher order anaphors", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Conclusive remarksIt has been shown that it is preferable to treat simple and complex reciprocal expressions, belonging to the class of higher order anaphors, as denoting type 2 : 1 functions (that is functions having relations as arguments and sets of type 1 quantifiers as result) and not as denoting type 1, 2 quantifiers, as usually proposed. The main reason for this treatment is the fact that the basic reciprocal expression each other can combine not only with NPs (which denote (extensions of) type 1 PACLIC-27", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Exceptions in Relational Plurals", |
| "authors": [ |
| { |
| "first": "Sigrid", |
| "middle": [], |
| "last": "Beck", |
| "suffix": "" |
| } |
| ], |
| "year": 2000, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "1--16", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sigrid Beck. 2000. Exceptions in Relational Plurals, in Brendan Jackson, and Tanya Matthews (eds), SALT X, Cornell University, 1-16", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Seth Cable. forthcoming. Reflexives, Reciprocals and Contrast, forthcoming in Journal of Semantics", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Seth Cable. forthcoming. Reflexives, Reciprocals and Contrast, forthcoming in Journal of Semantics", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Reciprocal expressions and the concept of reciprocity", |
| "authors": [ |
| { |
| "first": "Mary", |
| "middle": [], |
| "last": "Dalrymple", |
| "suffix": "" |
| } |
| ], |
| "year": 1998, |
| "venue": "Linguistics and Philosophy", |
| "volume": "21", |
| "issue": "2", |
| "pages": "159--210", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Mary Dalrymple, et al. 1998. Reciprocal expressions and the concept of reciprocity, Linguistics and Phi- losophy 21(2):159-210", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "Jakub Dotla\u010dil. forthcoming. Reciprocals Distribute over Information States, forthcoming in Journal of Semantics Edward L. Keenan. 1987. Semantic Case Theory", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Jakub Dotla\u010dil. forthcoming. Reciprocals Distribute over Information States, forthcoming in Journal of Semantics Edward L. Keenan. 1987. Semantic Case Theory, in Groenendijk, J. and Stokhof, M. (eds.) Proc. of the Sixth Amsterdam Colloquium", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "On the denotations of anaphors", |
| "authors": [ |
| { |
| "first": "Edward", |
| "middle": [ |
| "L" |
| ], |
| "last": "Keenan", |
| "suffix": "" |
| } |
| ], |
| "year": 2007, |
| "venue": "Research on Language and Computation", |
| "volume": "5", |
| "issue": "1", |
| "pages": "5--17", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Edward L. Keenan. 2007. On the denotations of anaphors. Research on Language and Computation 5(1):5-17", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Generalized Quantifiers in Linguistics and Logic", |
| "authors": [ |
| { |
| "first": "L", |
| "middle": [], |
| "last": "Edward", |
| "suffix": "" |
| }, |
| { |
| "first": "Dag", |
| "middle": [], |
| "last": "Keenan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Westerst\u00e5hl", |
| "suffix": "" |
| } |
| ], |
| "year": 1997, |
| "venue": "Handbook of logic and language", |
| "volume": "", |
| "issue": "", |
| "pages": "837--893", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Edward L. Keenan and Dag Westerst\u00e5hl. 1997. Gener- alized Quantifiers in Linguistics and Logic, in Johan van Benthem and Alice ter Meulen (eds). Handbook of logic and language, Elsevier, Amsterdam, 837- 893", |
| "links": null |
| }, |
| "BIBREF6": { |
| "ref_id": "b6", |
| "title": "Each other, Asymmetry and Reasonable Futures", |
| "authors": [ |
| { |
| "first": "Alda", |
| "middle": [ |
| "Mari" |
| ], |
| "last": "Forthcoming", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "forthcoming in Journal of Semantics", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Alda Mari. forthcoming. Each other, Asymmetry and Reasonable Futures, forthcoming in Journal of Se- mantics", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Quantifiers in Language and Logic", |
| "authors": [ |
| { |
| "first": "S", |
| "middle": [], |
| "last": "Peters", |
| "suffix": "" |
| }, |
| { |
| "first": "Dag", |
| "middle": [], |
| "last": "Westerst\u00e5hl", |
| "suffix": "" |
| } |
| ], |
| "year": 2006, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Peters, S. and Dag Westerst\u00e5hl. 2006. Quantifiers in Language and Logic, Oxford U.P.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "Relational domains and the interpretation of reciprocals", |
| "authors": [ |
| { |
| "first": "Sivan", |
| "middle": [], |
| "last": "Sabato", |
| "suffix": "" |
| }, |
| { |
| "first": "Yoad", |
| "middle": [], |
| "last": "Winter", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "Linguistics and Philosophy", |
| "volume": "35", |
| "issue": "3", |
| "pages": "191--241", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sivan Sabato, and Yoad Winter. 2012. Relational do- mains and the interpretation of reciprocals. Linguis- tics and Philosophy 35(3):191-241", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Generalising Conservativity", |
| "authors": [ |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "Zuber", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "WoLLIC 2010", |
| "volume": "6188", |
| "issue": "", |
| "pages": "247--258", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Richard Zuber, 2010a. Generalising Conservativity, in Anuj Dawar, and Ruiz de Queiroz. (eds). WoLLIC 2010, LNAI 6188, Springer Verlag, 247-258", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "Semantic Constraints on Anaphoric Determiners", |
| "authors": [ |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "Zuber", |
| "suffix": "" |
| } |
| ], |
| "year": 2010, |
| "venue": "Research on Language and Computation", |
| "volume": "8", |
| "issue": "4", |
| "pages": "255--271", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Richard Zuber. 2010b. Semantic Constraints on Anaphoric Determiners, Research on Language and Computation 8(4): 255-271", |
| "links": null |
| }, |
| "BIBREF11": { |
| "ref_id": "b11", |
| "title": "Semantics of Slavic anaphoric possessive determiners", |
| "authors": [ |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "Zuber", |
| "suffix": "" |
| } |
| ], |
| "year": 2011, |
| "venue": "Proceedings of SALT 19, e-Language", |
| "volume": "", |
| "issue": "", |
| "pages": "464--477", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Richard Zuber. 2011. Semantics of Slavic anaphoric possessive determiners, in Ed Cormany, et ali. (eds). Proceedings of SALT 19, e-Language, 464-477", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "Reciprocals as higher order functions", |
| "authors": [ |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "Zuber", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "Proceedings of the Ninth International Workshop of Logic and Engineering of Natural Language Semantics 9", |
| "volume": "", |
| "issue": "", |
| "pages": "118--129", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Richard Zuber. 2012. Reciprocals as higher order func- tions, in Proceedings of the Ninth International Workshop of Logic and Engineering of Natural Lan- guage Semantics 9 (LENLS 9), 118-129", |
| "links": null |
| }, |
| "BIBREF13": { |
| "ref_id": "b13", |
| "title": "Some Higher Order Functions on Binary Relations", |
| "authors": [ |
| { |
| "first": "Richard", |
| "middle": [], |
| "last": "Zuber", |
| "suffix": "" |
| } |
| ], |
| "year": 2012, |
| "venue": "LNCS", |
| "volume": "8036", |
| "issue": "", |
| "pages": "277--291", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Richard Zuber. 2013. Some Higher Order Functions on Binary Relations, in Glyn Morrill, and Mark- Jan Nederhof (eds). Formal Grammar 2012/2013, LNCS 8036, Springer-Verlag, 277-291", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "text": "Leo protected himself/himself and Lea from Al. (16) Leo and Lea protected every students from themselves. (17) Most philosophers protect themselves from themselves. (18) Most philosophers protect themselves and the president from themselves. (19) Two monks protect themselves from the guru and themselves. (20) Five philosophers protected each other from themselves. (21) Leo and Lea/every student protected each other from Al. (22) Leo and Lea protected every philosopher from each other. This shows that reflexives can occur twice in a sentence in two different argumental positions of the verb. This is not the case with reciprocals: ?(23) Leo and Lea prevented each other from each other ?(24) Leo and Lea gave each other each other's book. The above sentences are not acceptable, or at least not interpretable.", |
| "num": null, |
| "uris": null, |
| "type_str": "figure" |
| }, |
| "FIGREF1": { |
| "text": "25a) Leo and Lea washed themselves (25b) Bill and Sue washed themselves. PACLIC-27 (26) Four persons, Leo, Lea, Bill and Sue washed themselves. (27a) Leo and Lea hug each other. (27b) Bill and Sue hug each other. (28) Four persons, Leo, Lea, Bill and Sue hug each other.", |
| "num": null, |
| "uris": null, |
| "type_str": "figure" |
| }, |
| "FIGREF2": { |
| "text": "Rg(R) = Rg(S). D14: A type 2 : \u03c4 function F is reflexivity and range sensitive, F \u2208 REF LRS, iff F (R) = F (S) whenever R \u2229 I = S \u2229 S \u2229 I and Rg(R) = Rg(S).", |
| "num": null, |
| "uris": null, |
| "type_str": "figure" |
| }, |
| "FIGREF3": { |
| "text": ": (39a) Leo and Lea hug each other. (39b) Bill and Sue hug each other. (40) Four people hug each other.", |
| "num": null, |
| "uris": null, |
| "type_str": "figure" |
| } |
| } |
| } |
| } |
