| { |
| "paper_id": "W89-0121", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T03:45:17.922192Z" |
| }, |
| "title": "Erferenheter av Swetra-ett svenskt M T-experiment", |
| "authors": [ |
| { |
| "first": "Bengt", |
| "middle": [], |
| "last": "Sigurd", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "Swetra is primarily a research project, but its computer programs and reports may result in commercial products. The research group at Lund consists of a few persons only, which has advantages eind disadvantages. The main languages studied are Russian, English and Swedish. The gram matical model is Referent Grammar (R G) and the project is a spin-off of the development of that grammar. Referent Grammar is inspired by generalized phrase structure grammar (G PSG) and written directly in Prolog (D C G), which is why the grammars can run directly on computers. Arity Prolog and P C 's are used. The work within Swetra consists mainly of writing grammar modules and lexicons for different languages. These modules are connected in automatic translation. Referent grammars are bidirectional and can be used both for analysis and synthesis (generation). A certain number of transfer rules are also needed in translation, however, in order to make necessary changes o f functional representations, add features such as definiteness when translating from Russian into English etc. The paper gives a survey of the experience from Swetra reseeirch so far. 1 Inledning S w etra (S w edish C o m p u t e r T ra n sla tion R e s e a rch) h ar p \u00e5 g \u00e5 tt cirk a t v \u00e5 \u00e5 r, o c h d e t finns en d e l e rfa ren h eter a tt f\u00f6 r m e d la-e rfa ren h e ter s o m a n d r a s o m g e r sig ut p \u00e5 m a sk in \u00f6v e rs\u00e4 ttn in g e n s g u n g fly kan h a g l\u00e4 d je a v a tt k \u00e4 n n a till. S w etra , so m s t\u00f6 d s a v S venska F orsk n in gsr\u00e5 d et f\u00f6 r Scim h \u00e4llveten skap o c h H u m a n io ra , \u00e4r ett m in im a lt p r o je k t ifr\u00e5 g a o m ek on om isk a , p e rs o n e lla o c h m a sk in e lla resurser. I p r o je k te t a rb e ta r-f\u00f6 r u to m u n d e rte ck n a d-M a ts E e gO lo fs s o n (h a lv t id), L ars G u sta fsso n (k v a r ts tid), B arb rira G a w ron sk a-W ern g ren (h a lv t id). P ro g ra m m e rin gen h ar fr\u00e4 m st sk ett p \u00e5 P C m ed a n v \u00e4 n d a n d e a v A r ity P r o lo g. J a g skall m e d d e la m in a erfa ren h ete r a v p r o je k te t u n d er f\u00f6 lja n d e h u v u d ru b ri ker: O rg a n isa tio n , V \u00e4 g v a l (t e x t ty p , sp r\u00e5 k , in st\u00e4 lln in g till sy n ta x , g ra m m a tik m o d e ll), V a d S w etra kan. P u b lik try ck e t.", |
| "pdf_parse": { |
| "paper_id": "W89-0121", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "Swetra is primarily a research project, but its computer programs and reports may result in commercial products. The research group at Lund consists of a few persons only, which has advantages eind disadvantages. The main languages studied are Russian, English and Swedish. The gram matical model is Referent Grammar (R G) and the project is a spin-off of the development of that grammar. Referent Grammar is inspired by generalized phrase structure grammar (G PSG) and written directly in Prolog (D C G), which is why the grammars can run directly on computers. Arity Prolog and P C 's are used. The work within Swetra consists mainly of writing grammar modules and lexicons for different languages. These modules are connected in automatic translation. Referent grammars are bidirectional and can be used both for analysis and synthesis (generation). A certain number of transfer rules are also needed in translation, however, in order to make necessary changes o f functional representations, add features such as definiteness when translating from Russian into English etc. The paper gives a survey of the experience from Swetra reseeirch so far. 1 Inledning S w etra (S w edish C o m p u t e r T ra n sla tion R e s e a rch) h ar p \u00e5 g \u00e5 tt cirk a t v \u00e5 \u00e5 r, o c h d e t finns en d e l e rfa ren h eter a tt f\u00f6 r m e d la-e rfa ren h e ter s o m a n d r a s o m g e r sig ut p \u00e5 m a sk in \u00f6v e rs\u00e4 ttn in g e n s g u n g fly kan h a g l\u00e4 d je a v a tt k \u00e4 n n a till. S w etra , so m s t\u00f6 d s a v S venska F orsk n in gsr\u00e5 d et f\u00f6 r Scim h \u00e4llveten skap o c h H u m a n io ra , \u00e4r ett m in im a lt p r o je k t ifr\u00e5 g a o m ek on om isk a , p e rs o n e lla o c h m a sk in e lla resurser. I p r o je k te t a rb e ta r-f\u00f6 r u to m u n d e rte ck n a d-M a ts E e gO lo fs s o n (h a lv t id), L ars G u sta fsso n (k v a r ts tid), B arb rira G a w ron sk a-W ern g ren (h a lv t id). P ro g ra m m e rin gen h ar fr\u00e4 m st sk ett p \u00e5 P C m ed a n v \u00e4 n d a n d e a v A r ity P r o lo g. J a g skall m e d d e la m in a erfa ren h ete r a v p r o je k te t u n d er f\u00f6 lja n d e h u v u d ru b ri ker: O rg a n isa tio n , V \u00e4 g v a l (t e x t ty p , sp r\u00e5 k , in st\u00e4 lln in g till sy n ta x , g ra m m a tik m o d e ll), V a d S w etra kan. P u b lik try ck e t.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "J a g tr o r a tt en liten forsk a rg ru p p lok aliserad p \u00e5 en p la ts h ar s to r a f\u00f6rd ela r. s\u00e4 ttn in g . J a g h a d e l\u00e4 n g e ta la t o c h sk riv it o m hur sv \u00e5 rt m a sk in \u00f6v e rs\u00e4 ttn in g \u00e4r u ta n a tt s j\u00e4 lv h a n \u00e5 g o n r ik tig d a to re rfa re n h e t a v d e t-b a r a p a p p e rssp e k u la tio ner. S w etra \u00e4r en b ip r o d u k t a v g ra m m a tik fo rsk n in g o c h u p p s to d sed a n ja g in sett a tt d e t g ick a tt a n v \u00e4 n d a refere n tg ra m m a tik f\u00f6 r \u00f6 v e rs\u00e4 ttn in g . S w etra \u00e4r snarast e tt f\u00f6 r s \u00f6 k a tt se h u r l\u00e5 n g t m a n kan k o m m a m ed sm \u00e5 m ed el. V i \u00e4 g n a r f\u00f6 g a tid \u00e5t a tt v is a v a d m a n in te kan k la ra eller b e v is a a tt m a sk in \u00f6 v e rs\u00e4 ttn in g \u00e4 r o m \u00f6 jlig t, m est tid \u00e5 t a tt v is a v a d m a n kan klara. ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Organization", |
| "sec_num": "2" |
| }, |
| { |
| "text": "D e t \u00e4r n \u00f6 d v \u00e4 n d ig t a tt g \u00f6 r a e tt a n ta l v \u00e4 g v a l n \u00e4r m a n g \u00e5 r in i m a sk in \u00f6v ers\u00e4 ttn in g sb ra n sch e n . F le r a a v d essa va l h \u00f6 r ih o p . M a n m \u00e5 ste v \u00e4 lja m ellan a tt f\u00f6 rs\u00f6k a sig p \u00e5 a tt \u00f6 v e r s \u00e4 tta g o d ty c k lig te x t (fr it e x t ) eller te x t in o m n \u00e5 g o t s p e cia lo m r\u00e5 d e s o m te k n ik (t .e x . b ila r, d a t o r e r ), m e d icin , a il\u00e4 rsk on tra k t, v \u00e4 d e rle k sra p p orter, e tc . In le d n in g sv is tr\u00e4 fia d e v i in te n \u00e5 g o t s\u00e5 d a n t va l-m a n \\'ar g la d a a tt d et gick a tt \u00f6 v e r s \u00e4 tt a n \u00e5 g ra m e n in g a r \u00f6 v e r h u v u d ta g e t. D e m en in g a r v i f\u00f6rst \u00f6 v e r s a tte var ty p is k a lin g v is tm e n in g a r, ty p \" E n flick a k o m \" , \" P o jk e n s lo g h u n d en s o m sp ra n g \" , f\u00f6 r a tt \u00f6v ers\u00e4 ttn in g en skall fu n g e ra . J a g skall sed a n v is a ty p is k a S w e tra te x te r so m v i bru k ar k \u00f6 ra i v \u00e5 r D e m o (se o c k s \u00e5 S ig u rd G a w ro d sk a -W e rn g re n 1 9 8 8 ) .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ".1 T exttyp", |
| "sec_num": "3" |
| }, |
| { |
| "text": "S j\u00e4lvk lart m \u00e5 ste m a n v \u00e4 lja vilka sp r\u00e5 k m a n v ill \u00f6 v e r s \u00e4 tt a m ella n , m en m \u00e4 rk lig t till o c h fr\u00e5 n fran ska, p olsk a , g eorgisk a, sa m oa n sk a , sven ska, en gelsk a, rysk a. P \u00e5 sen are tid h ar v i a lltm e r k on ce n trera t oss p \u00e5 rysk a, en gelska, sven sk a o c h fr\u00e4 m st p \u00e5 \u00f6 v e rs\u00e4 ttn in g fr\u00e5n ry sk a till sven ska-sed a n B a r b a r a G a w ro \u00e5 sk a -W e rn g re n st\u00e4 lld e sin a sp r\u00e5 k k u n sk a p er till f\u00f6 r fo g a ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Spr\u00e5k", |
| "sec_num": "3.2" |
| }, |
| { |
| "text": "E tt a n n a t v \u00e4 g v a l r \u00f6 r in st\u00e4 lln in g en till sy n ta x . M \u00e5 n g a -i sy n n erh e t tid igare- av o v a n n \u00e4 m n d a ty p ( \" E n flick a sp r a n g \" , \" H u n d en , s o m , p o jk e n , s o m flickan k \u00e4 n de s lo g sp ra n g \" ). M a n ser t.e x . a ld rig en G B -g r a m m a tik e r v isa hur a lla m en in g a rn a p \u00e5 en viss s id a te x t skall a n a ly sera s, vilk a tr ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ".Inst\u00e4llning til syntenc", |
| "sec_num": "3" |
| }, |
| { |
| "text": "S w e tra kan a n a ly se ra , sy n te tis e ra (g e n e re ra ) o c h \u00f6 v e r s \u00e4 tta ty p isk a lin gv istm e n in g a r m ed in tra n sitiv a , tra n sitiv ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Vad Swetra kan", |
| "sec_num": "4" |
| }, |
| { |
| "text": "Proceedings of NODALIDA 1989, pages 229-236", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "Proceedings of NODALIDA 1989", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "A referent grcunmatical analysis of Polish relative clau ses", |
| "authors": [ |
| { |
| "first": "B", |
| "middle": [], |
| "last": "Gawronska^werngren", |
| "suffix": "" |
| } |
| ], |
| "year": 1988, |
| "venue": "Studia Linguistica", |
| "volume": "42", |
| "issue": "1", |
| "pages": "18--48", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Gawronska^Werngren, B. 1988. A referent grcunmatical analysis of Polish relative clau ses. Studia Linguistica 42(1):18-48", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "Using Referent Grammar (R G ) in computer analysis, generation and translar tion o f sentences", |
| "authors": [ |
| { |
| "first": "B", |
| "middle": [], |
| "last": "Sigurd", |
| "suffix": "" |
| } |
| ], |
| "year": 1987, |
| "venue": "Nordic Journal o f Linguistics", |
| "volume": "41", |
| "issue": "2", |
| "pages": "95--115", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sigurd, B. 1987. Referent grammar (R G ). A generalized phrase structure grammar with built-in referents. Studia Linguistica 41(2):115-135 -1988. Using Referent Grammar (R G ) in computer analysis, generation and translar tion o f sentences. Nordic Journal o f Linguistics 11(1-2):129-150. -1989. A referent grammatical analysis of relative clauses. Acta Linguistica Hafniensia 21(2):95-115", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "The potentials of Swetra, a multilanguage MT-system", |
| "authors": [ |
| { |
| "first": "B", |
| "middle": [], |
| "last": "Sigurd", |
| "suffix": "" |
| }, |
| { |
| "first": "B", |
| "middle": [], |
| "last": "Gawrohska^werngren", |
| "suffix": "" |
| } |
| ], |
| "year": 1988, |
| "venue": "Computers and Translation", |
| "volume": "3", |
| "issue": "", |
| "pages": "238--250", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Sigurd, B. & c Gawrohska^Werngren, B. 1988. The potentials of Swetra, a multilanguage MT-system. Computers and Translation 3:238-250", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "M a n v e t v a d d e a n d r a g \u00f6 r o c h b e h \u00f6 v e r in te resa ru nt eller sam las till sto ra sem in a rier. L e d n in g e n f\u00f6ren k la s. D e t r\u00e4ck er o ft a m e d n \u00e5 g ra o r d till d e a n d ra f\u00f6 r a tt d e skall v a ra m e d p \u00e5 n o te rn a . U p p e n b a rlig e n s\u00e4 g er ja g d e tta m er eller m in d r e o m e d v e te t m e d ta n k e p \u00e5 E u r o tr a , d \u00e4 r m e d a r b e ta r n a syn es h a tv in g a ts a n v \u00e4 n d a m y ck e t tid p \u00e5 a tt resa f\u00f6 r a tt in fo rm era v a ra n d ra o c h k o m m a \u00f6veren s o m v ilk a lin g v istisk a te o r ie r , p ro g ra m m e rin g s sp r\u00e5 k , fo r m a t o c h d a to r e r m an skall a n v \u00e4 n d a . D e n lilla t\u00e4 t t s a m m a n sv e tsa d e g ru p p e n h a r s\u00e4k ert s to r a f\u00f6 rd e la r n\u00e4r d e t g \u00e4 lle r a tt u tv e c k la d e n g ru n d l\u00e4 g g a n d e teorien o c h p r o to ty p e n , m en n\u00e4r det se d a n g \u00e4 lle r a tt im p le m e n te ra teo rien o c h g \u00e5 fr\u00e5n p r o t o t y p till p r o d u k t, a tt skriva s t o r a le x ik o n , a tt f\u00f6 r b \u00e4 t t r a p ro g ra m m e n o c h a tt te s tk \u00f6 r a sy ste m e t d \u00e5 h ar st\u00f6rre a r b e ts g r u p p e r s\u00e4 k ert f\u00f6 rd e la r. V i h a d e i sj\u00e4 lv a v erk et in g et val i L u n d , n \u00e4r v i b \u00f6 r ja d e p r \u00f6 v a p \u00e5 m a sk in \u00f6v er" |
| }, |
| "FIGREF1": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "A n p a s s n in g e n till P C b e s t\u00e4 m d e v i o ss tid ig t f\u00f6 r. D e n u n d e rl\u00e4 tta r fly ttn in g o ch d e m o n s tr a tio n a v p ro g ra m m e n . V i t\u00e4 n k er nu o c k s \u00e5 a n v \u00e4 n d a s n a b b a M a cin to shd a to r e r s o m o c k s \u00e5 kan l\u00e4 sa D O S -d is k e tte r o c h m \u00f6 jlig g \u00f6 r fly ttn in g m ellan olika m ilj\u00f6 e r." |
| }, |
| "FIGREF2": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "H u n d e n , s o m p o jk e n , s o m flickan k \u00e4 n d e s lo g sp ra n g \" . E fte r a tt h a sa tt u p p v a d v i ty c k te v a r sy n ta k tis k a reg ler s o m b o r d e k la ra g ru n d l\u00e4 g g a n d e stru k tu rer f\u00f6 r s \u00f6 k te v i p r \u00f6 v a d e m p \u00e5 em p irisk a te x te r , b l.a . en liten G r e e n p e a c e -te x t o m k \u00e4 rn k ra ftv e rk i S ellafleld s o m fo r tfa r a n d e s l\u00e4 p p e r u t p lu to n iu m o c h sedan n \u00e5 gra n o tis e r fr\u00e5 n P r a v d a . M \u00f6 t e t m e d em p irisk a te x te r var o m tu m la n d e ; m a n inser att te o re tisk a lin g v is te r lev er i e tt reservat (m e d e tt sven sk t u ttry ck : en sk y d d a d v e r k s ta d ). V a r e n d a m e n in g le d d e till a tt v i fick re v id e r a g ra m m a tik e n , \u00e4 n d ra eller l\u00e4 g g a till reg ler o c h in te b a r a u t\u00f6 k a le x ik o n e t s o m m a n ju h op p a s . M a n kan t\u00e4 n k a sig a tt rita d ia g ra m s o m v isa r \u00f6 k n in g e n a v d e g ra m m a tisk a reg lern a o c h le x ik on f\u00f6 r v a rje n y m en in g s o m skall \u00f6 v e rs\u00e4 tta s. I b \u00f6 r ja n m \u00e5 ste m \u00f6 te t m ed en n y m en in g n \u00f6 d v \u00e4 n d ig tv is le d a till s to r \u00f6 k n in g b \u00e5 d e a v g r a m m atiken o c h le x ik o n e t, m en ku rvan b \u00f6 r s \u00e5 sm \u00e5 n in g o m p la n a u t, n \u00e4 r n \u00e4 sta n a lla m en in ga r klaras a v u ta n a tt n \u00e5 g on f\u00f6 r b \u00e4 t tr in g b e h \u00f6 v e r g \u00f6 ra s. E tt f\u00e4 r d ig t sy ste m b \u00f6 r k la ra a llt o c h in te k r\u00e4 v a m e ra u tb y g g n a d . O m d e t \u00e4r e tt in te ra k tiv t s y ste m b \u00f6 r d et in te fr \u00e5 g a o p e r a t\u00f6 r e n a lltf\u00f6 r o ft a . J a g kan in te s\u00e4 g a a tt v i k o m m it s\u00e5 l\u00e5 n gt i S w etra a tt v i k \u00e4 n n er a tt ku rvan h \u00e5 ller p \u00e5 a tt p la n a u t; v a re n d a liten te x t v i f\u00f6 rs\u00f6k er o ss p \u00e5 k r\u00e4 v er flera k o m p le ttrin g a r-o c h d e t \u00e4 r in te a lltid b a r a fr \u00e5 g a o m a tt s \u00e4 tta in fler o r d i le x ik on . D e t val i fr \u00e5 g a o m te x te r v i avser a tt tr\u00e4 ffa in n e b \u00e4 r a tt v i s\u00e4 g er a tt S w etra in te kan \u00f6 v e r s \u00e4 tt a g o d ty c k lig te x t, b a r a sp e cie ll te x t , m en v i vill s p e cifice ra vilk en sp e cie ll te x t S w etra kan \u00f6 v e r s \u00e4 tta p \u00e5 e tt lin g v istisk t s\u00e4 tt. V i vill in te s p e cifice ra d e ssa te x te r g e n o m h \u00e4 n v isn in g till en g e n re s o m t.e x . tek n isk a m a n u aler, v \u00e4 d e rle k sra p p o rte r, u ta n g e sp ecifik a tion e n m e r a lin g v istisk t o c h s\u00e4 g a a tt m en in g a rn a in te f\u00e5r in n e h \u00e5 lla s a m o r d n a d e re la tiv sa tse r, in te m er \u00e4n t v \u00e5 sa tsa d v erb ia l o c h tr e a n d r a a d v e rb ia l, in te in b \u00e4 d d a d e g e n itiv e r s o m \" p o jk e n s h u n ds svan s\" e tc . L exik a lt kan sp e cifik a tio n e n g \u00f6 r a s g e n o m a tt m a n t.e x . s\u00e4ger a tt o rd e n m \u00e5 ste tillh \u00f6 r a d e n m a rin a sf\u00e4 ren o c h r \u00f6 r a s ig o m o lik a t y p e r a v b \u00e5 ta r som r\u00f6 r sig i \u00d6 s te rsj\u00f6 n . I sj\u00e4 lv a v erk et b o r d e n o g m \u00e5 n g a M T -s y s te m h a s p e c ific e r a t sin a b e g r\u00e4 n sn in g a r p \u00e5 d e tta s\u00e4 tt. S \u00e5 v it t ja g f\u00f6 r s t\u00e5 tt a v d e d e m o n s tr a tio n e r E u r o t r a f\u00f6 r e ta g it i B riissel h a r E u r o tr a eg en tlig e n o c k s \u00e5 g \u00e5 tt p \u00e5 d e n n a lin je: D e lta g a r n a fick b a ra f\u00f6re sl\u00e5 m en in g a r s o m h a d e h \u00f6 g st 7 o r d , b a r a e tt a d je k tiv i n om in a lfra sen , en d a st s u b je k tiv a rela tiv sa tser, in g a s a m o r d n in g a r e tc . D e t su b sp r\u00e5 k s o m S w etra vill k u n n a \u00f6 v e r s \u00e4 tta d efin ieras a llts \u00e5 i f\u00f6 r s t a h a n d g e n o m v is s a sy n ta k tisk a b eg r\u00e4 n sn in g a r. S edan \u00e4 r d e t en sj\u00e4 lv k la rh e t a tt o r d e n m \u00e5 s te fin n as i le x ik o n e t" |
| }, |
| "FIGREF3": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "n o g h ar v i tv ek a t l\u00e4 n g e p \u00e5 d e n n a p u n k t i S w etra . D e t b e r o r n a tu rlig tv is p \u00e5 a tt v i \u00e5 ta g it oss a tt fo rsk a i m a sk in \u00f6v ers\u00e4 ttn in g , in te a tt le v e re ra e tt f\u00e4 rd ig t p r o g r a m .D e t \u00e4 r i sj\u00e4 lv a v erk et in tressan t a tt p r \u00f6 v a hur b r a d e n g ra m m a tisk a m o d e lle n referen tg ra m m a tik fu n g e ra r p \u00e5 olik a sp r\u00e5 k . V i h a r erfa ren h et a v \u00f6 v e r s \u00e4 ttn in g" |
| }, |
| "FIGREF4": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "n d e . D e olik a referen tg ra m m a tisk a m o d u lern a so m p r o g ra m m e ra s d irek t i P r o lo g (D C G = D e f in it e C la u se G ra m m a x ) \u00e4 r i p rin cip b id ire k tio n e lla o c h kan k \u00f6ra s b \u00e5 d e i a n a ly s o c h sy n tes (g e n e r e r in g ). M e n vill m an g \u00f6 r a m era a v a n cera d \u00f6 v e r s \u00e4 ttn in g , a n v \u00e4 n d a sp ecifik a tra n sferreg ler, a n k n y ta d isk u rssem an tisk a p r o c e d u r e r m .m . o c h g \u00f6 r a e tt e ffe k tiv t p r o g r a m \u00e4r d e t b \u00e4 st a tt b e s t\u00e4 m m a sig f\u00f6 r en rik tn in g . V i hax b e d \u00f6 m t \u00f6 v e r s \u00e4 ttn in g m ella n rysk a o c h en gelska so m intressaxit d \u00e4 rf\u00f6r a tt d e t t a p a r l\u00e4 n g e in tresserat M T -fo rs k a rn a . E tt fra m g\u00e5 n g rik t sy stem so m kan \u00f6 v e r s \u00e4 tt a m e lla n ry sk a o c h en gelsk a h a r u p p e n b a ra p rak tisk a till\u00e4 m p n in ga r. S p r\u00e5 k en s ty p o lo g is k a sk illn a d er g \u00f6 r \u00f6 v e rs\u00e4 ttn in g o c k s \u00e5 lin gv istiskt intressant. E fte r s o m m \u00e5 n g a p e rs o n e r in te kan rysk a b lir d e m o n s tra tio n e r d essu tom m era im p o n e r a n d e . \u00c5 s k \u00e5 d a r n a b lir m e ra f\u00f6 r b r y lla d e av a tt se hur en o b e g r ip lig tex t \u00f6 v e r s \u00e4 tte s till en b e g r ip lig \u00e4n a tt se t.e x . en gelska, s o m d e flesta kan, \u00f6vers\u00e4 tta s till sven sk a. ( \u00c5 a n d r a sid a n b lir \u00e5 sk \u00e5 d a rn a s m \u00f6 jlig h e te r a tt b e d \u00f6 m a \u00f6v ers\u00e4 ttn in g e n s k v a litet m e ra b e g r\u00e4 n s a d e )." |
| }, |
| "FIGREF5": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "s y s te m b e tr a k ta r o r d \u00f6 v e r s \u00e4 ttn in g so m d e n p rim \u00e4 ra o p e r a tio n e n v id \u00f6v ers\u00e4 ttn in g . M a n f\u00f6 rs \u00f6 k e r sed a n se k u n d \u00e4 rt se till a tt o rd e n s k o n g ru e n sb \u00f6 jn in g st\u00e4 m m e r o ch a tt o r d f\u00f6 ljd e n \u00e4r r \u00e4 tt g e n o m a tt f\u00f6 rs \u00f6 k a id en tifiera vilka o r d so m h \u00f6 r ih o p i fraser o c h vilk a o r d (fra s e r) s o m \u00e4 r s u b je k t, p red ik a t, o b je k t o c h a d v e rb ia l-n \u00e5 g o t som o ft a \u00e4 r e tt m in im ik ra v f\u00f6 r a tt k la ra b \u00f6 jn in g o c h o r d f\u00f6 ljd . S w etra h ar in te o rd s o m p r im \u00e4 r a en h e te r, u ta n sa tser o c h fra ser i olik a fu n k tio n e r (s\u00e5 so m su b je k t, o b je k t , p red ik a t, s a tsa d v e rb ia l, a n d ra a d v e rb ia l, a tt r ib u t). S w etra \u00e4r p \u00e5 d e tta s \u00e4 tt en ty p is k p r o d u k t a v sin tid -den tid i lin gv istik en d \u00e5 s y n ta x skriven s\u00e5 som g e n e r a tiv g r a m m a tik st\u00e5 r i fok u s. J a g skall f\u00f6rk la ra d e tta n \u00e4rm a re sen are n \u00e4r ja g b e s k riv e r d en g ra m m a tisk a m o d e ll (R e fe r e n tg r a m m a tik ) s o m S w e tra h ela tiden a r b e ta t m ed . E tt sy ste m s o m p r io rite ra r o r d \u00f6 v e r s \u00e4 ttn in g kan o ft a \u00f6 v e r s \u00e4 tta m \u00e5 n g a sorters te x te r s n a b b t g iv e t e tt s to r t le x ik o n , m en k v a liteten b lir lid a n d e o m m an inte h \u00e5 ller r e d a p \u00e5 sa tsstru k tu re n i d e ta lj. F \u00f6 r v \u00e5 n a n d e m \u00e5 n g a tv e ty d ig a o r d blir e n ty d ig a o m m a n ta r u t sa tsd ela rn a -sta n d a rd e x e m p le t \u00e4r: V a r var d et var? B \u00f6 jn in g a v o r d e n i m \u00e5 lsp r\u00e5 k e t g er sig lik so m o r d f\u00f6 ljd e n o m m an h ar d eta ljera d e u p p ly s n in g a r o m o rd e n s fu n k tio n e r i fra ser o c h sa tsd ela r. A a n d ra sidan \u00e4r sa ts l\u00f6 s n in g (p a r s n in g ) tid s \u00f6 d a n d e , o c h a lla p r o b le m kan in te l\u00f6sa s m ed sy n ta x\u00f6 k \u00e4 n d a \u00e4 r p re p o s itio n s fra s e rn a vars a n k n y tn in g till v e r b eller n om in a lfra s o fta e n d a s t kan a v g \u00f6 ra s p \u00e5 sem a n tisk v \u00e4 g . N \u00e4 r S w e tra g \u00e5 r g ra m m a tik v \u00e4 g e n , in n e b \u00e4 r d e t a tt m a n e n d a s t a c c e p te r a r o c h \u00f6 v e r s \u00e4 tte r m en in g a r vars gra m m a tisk a str u k tu r s y ste m e t h a r reg ler f\u00f6 r a tt id en tifiera . E tt lex ik on orien tera t sy stem kan a lltid f\u00f6 r e s l\u00e5 en \u00f6 v e r s \u00e4 ttn in g \u00e4 v en o m sy ste m e t in te h a r g jo r t n \u00e5 g o n d eta ljera d sa tsl\u00f6 sn in g ." |
| }, |
| "FIGREF6": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "n m o d e r n lin g v is t h ar till sy n es flera v \u00e4 l g e n o m d is k u te ra d e fo rm e lla gram m a^ tiska m o d e lle r a tt v \u00e4 lja p \u00e5 , t.e x . T ra n s fo rm a tio n e n g e n e ra tiv g ra m m a tik m ed fra sstru k tu r o c h tr a n s fo r m a tio n e r (E S T , G o v e rn m e n t a n d B in d in g ), L e x ica l-F u n c tio n a l G r a m m a r (L F G ), G en era lisera d fra stru k tu rg ra m m a tik (G P S G ), D ep e n d e n s g r a m m a tik . F u n k tion ell G ra m m a tik a v H a llid a y s ty p . D e t \u00e4r p\u00e5 fa lla n d e a tt d e ssa g ra m m a tik e r n \u00e4 sta n b a r a h a r p r \u00f6 v a ts p \u00e5 ty p isk a lin gv istm en in g a r 232 Proceedings of NODALIDA 1989" |
| }, |
| "FIGREF7": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "\u00e4 d d ia g r a m m a n b \u00f6 r rita o c h vilk a p r o b le m m a n m \u00f6 te r . G \u00e4 n g se te o re tisk a g r a m m a tik m o d e lle r v isa r fr\u00e4 m st hur v issa fo rm e lla id \u00e9 e r s o m t.e x . fra sstru k tu r, tr a n s fo r m a tio n e r , d e p e n d e n s , s u b ja ce n cy , c -c o m m a n d fu n gerar. D e h ar in te d et p rim \u00e4 ra s y ite t a tt v isa hur v an liga m en in g a r skall a n alyseras s\u00e5 a tt b e sk riv n in g e n f\u00f6 rk la ra r v a rf\u00f6 r o rd e n h ar den fo r m d e h a r o c h st\u00e5 r i d en o r d n in g d e g \u00f6 r . A llr a m in st f\u00f6 rk la ra r d e va rf\u00f6r d en f\u00f6 re lig g a n d e m en in gen b e ty d e r v a d d en g \u00f6 r. S \u00e5 som fr a m g \u00e5 tt v a r d et in te s\u00e5 a tt m e d a r b e ta r n a i S w etra b e s t\u00e4 m d e a tt d e skulle \u00f6 v e r s \u00e4 tta m e d d a to r m ellan t v \u00e5 b e s t\u00e4 m d a sp r\u00e5 k o c h sed a n b \u00f6 r ja d e leta efter en l\u00e4 m p lig g ra m m a tik . I st\u00e4 llet v a r d et s\u00e5 a tt ja g h \u00f6 ll p \u00e5 m e d d e p ilreg ler som D C G e r b ju d e ro c h in s\u00e5 g a tt m a n s o m resu lta t a v a n a ly sen (sa tsl\u00f6 sn in g e n , p a rsn in g en ) k u n d e f\u00e5 en sorts u n iversell sem a n tisk re p resen ta tion : en p re d ik a tslog isk fo rm e l eller en fu n k tio n e ll rep rese n ta tion , en re p resen ta tion so m ta la r o mvad s o m \u00e4 r s u b je k t, p red ik a t, o b je k t , a d v e rb ia l, e tc . o c h v a d o r d e n b e ty d e r .(C h o m s k y s klassiska g e n e ra tiv a g ra m m a tik re g le r g e r e n d a st e tt S s o m resu lta t av a n alysen , vilket b a r a s\u00e4 g er a tt sa tsen var g r a m m a tis k ). S ed an ja g sk rivit en s\u00e5dan g ra m m a tik f\u00f6 r sven sk a o c h en f\u00f6 r en g elsk a in s\u00e5 g ja g a tt m a n m \u00e5 stek u nn a \u00f6 v e r s \u00e4 tta v ia d en g e m e n s a m m a fu n k tio n e lla re p resen ta tion en o m m a n sta n d a rd isera d e d en . D e t \u00e4r i p r in c ip d e t v i h \u00e5llit p \u00e5 m e d sed a n i S w etra . D en fu n k tion ella rep re se n ta tion en fu n g era r s o m e tt m ella n sp r\u00e5 k , in terlin g u a . D en g ra m m a tisk a m o d e lle n f\u00f6 r S w etra va r a llts\u00e5 g iv e n , m en s a m tid ig t s o m vi a rb e ta t m ed \u00f6 v e rs \u00e4 ttn in g s p ro g ra m m e n h a r v i u tv e ck la t r e fe re n tg ra m m a tik o c h ta g it b eslu t so m sta n d a rd isera t d en o c h fra m f\u00f6 ra llt d ess fu n k tio n e lla rep re sen ta tio n o c h le x ik o n fo rm a t. R e fe re n tg ra m m a tik ( R G , se referen ser i litte ra tu rlista n ) \u00e4r en so rts g en eraliserad fra sstru k tu rg ra m m a tik ( G P S G ), m en i m o ts a ts till G P S G a r b e ta r R G i d e g e n e ra tiv a reg lern a m ed t v \u00e5 rep re sen ta tion er: d en y tlig a k a tegorirep resen ta tion en b e te c k n a d o -r e p r e s e n ta tio n e n , o c h d en fu n k tio n e lla rep rese n ta tion en b e te ck n a d f-re p re s e n ta tio n e n . D e s sa b e te c k n in g a r p \u00e5 m in n e r o m d e m s o m a n v \u00e4 n d es i le x ic a l-fu n c tio n a l g ra m m a r (L F G ), en g r a m m a tik s o m n o g o c k s \u00e5 kan a n v \u00e4 n d a s v id a u to m a tis k \u00f6 v e rs\u00e4 ttn in g . R e fe re n tg ra m m a tik \u00e4r b id ire k tio n e ll, d v s kan a n v \u00e4 n d a s b \u00e5 d e i a n a ly s o c h syn tes (g e n e re rin g ). D e t b e ty d e r a tt d en st\u00e4 ller sa m m a h \u00e5 r d a k ra v p \u00e5 sitt in p u t o c h sitt o u tp u t. E n k orrek t sk riven R G -m o d u l g e n e re ra r b a r a k o rre k ta m en in ga r o c h a c c e p te r a r b a r a k orrek ta m en in g a r. F \u00f6 r a tt en g r a m m a tik skall v a ra a n v \u00e4 n d b a r f\u00f6 r M T k r\u00e4 v s a tt d en kan sp e c ific e r a k o rrek th et p \u00e5 a lla n iv \u00e5 e r, g en ererar k orrek ta fo r m e r av artik la r, r\u00e4 k n e ord , p r o n o m e n , a d je k tiv , s u b s ta n tiv , v e rb , s \u00e4 tta o r d e n i r \u00e4 tt o rd n in g , e tc . R G g \u00f6 r d e t-d e n g en erera r t.e x . o c k s \u00e5 k orrek ta fo r m e r a v re la tiv p r o n o m e n i ry sk a o c h p o lsk a en u p p g ift s o m k r\u00e4 v er a tt flera fa k to re r ta s h \u00e4 n sy n till. R G \u00e4r d e s s u to m l\u00e4 tt a tt sk riv a f\u00f6 r en lin g v ist o c h v i h ar in te fu n n it n \u00e5 g o t s o m d e n in te kan h a n tera . D e n k larar o c k s \u00e5 s.k. u nbounded d ep en d en cies, d v s fly ttn in g a r s o m kan v a ra h u r l\u00e5 n g a s o m h elst. R G a n v \u00e4 n d e r lik som G P S G d e fe k ta sy n ta k tisk a k a teg orier, m e n d e fo rm a liseras in te s o m \" sla sh -ka tegprier\" i R G u ta n b e n \u00e4 m n s s d a e n t = s u b je k ts d e fe k t sats, o d p p = o b je k td e fe k t p re p o s itio n s fra s e tc . I p r o g r a m m e t fly tta s d e n sa k n a d e koDStituenten till r\u00e4tt position, vilket g\u00f6r att man kan s\u00e4ga att R G har dolda transformationer, precis som G P S G . F \u00f6 r a tt en g r a m m a tik m o d e ll skall k u n n a a n v \u00e4 n d a s i m a sk in \u00f6v ers\u00e4 ttn in g m \u00e5 ste d e n o c k s \u00e5 h a ta g it st\u00e4 lln in g till hur o r d b e ty d e ls e r skall besk riv a s o c h v ilket fo r m a t d e skall h a i le x ik o n e t. M \u00e5 n g a teoretisk a m o d e lle r l\u00e4 m n a r d e n fr\u00e5gan \u00f6 p p e n , m en d e t d u g e r in te i e tt M T -s y s te m d \u00e4 r a lla k o m p o n e n te r m \u00e5 ste finnas o c h k u n n a sam verk a. R G h a r e tt b e s t\u00e4 m t le x ik o n fo rm a t (se artik la r i litter\u00e4r tu r lis ta n ) o c h le x ik o n e t \u00e4r en s\u00e4 rsk ild fil p \u00e5 vilk en m a n kan l\u00e5 ta d iv erse operar tio n e r a r b e ta . O r d b e ty d e ls e n b e sk riv s f.n . i en en g elsk lik n a n d e fo r m (m a c h in e s \u00e9 ), m e n le x ik o n e t h a r u tr y m m e f\u00f6 r m \u00e5 n g a g ra m m a tisk a o c h sem an tisk a u p p g ifter o m o r d e n . R G h a r o c k s \u00e5 u tv e ck la t v issa p ro c e d u r e r f\u00f6 r m o r fo lo g i (o r d b \u00f6 jn in g o c h o r d b ild n in g ) o c h till\u00e4 m p a t s.k . im p lik a tion ell m o r fo lo g i, en m o d e ll d \u00e4 r den m o r fo lo g is k a ku n sk a p en ses s o m ku nsk ap o m v issa o r d fo r m e r k o m p le tte ra d m ed k u n sk a p o m h u r ex isten sen a v en v iss fo r m im p lice ra r ex iste n sen a v en viss a n n a n fo r m . O m m a n kan en fo r m i sp r\u00e5 k et o c h v et v a d d en b e ty d e r , kan m an d r a slu ts a tser o m h u r m \u00e5 n g a a n d r a fo rm e r b \u00f6 r se ut o c h v a d d e b e ty d e r. D e n g ra m m a tisk a d isk u ssio n e n d e sen a ste d e ce n n ie rn a h a r p .g .a . C h om sk y s sta rk a in fly ta n d e m y ck et b e r \u00f6 r t u n iversella sy n ta k tisk a p r in c ip e r f\u00f6 r n atu rliga sp r\u00e5 k , t.e x . b e g r\u00e4 n s n in g a r p \u00e5 fly ttn in g a v k o n stitu en ter. F ly ttn in g a r o c h h o p p a n d e p \u00e5 co m p le m e n tiz e r s (i en g elsk a ) hax t.e x . d isk u tera ts \u00e5 tsk illig t, o c h relar tiv s a ts e r h a r t.e x . b e sk riv its s o m en fly ttn in g a v d en re la tiv isera d e k on stitu en ten till k o m p le m e n tize rn = re la tiv m a rk \u00f6 re n e fte rl\u00e4 m n a n d e e tt o s y n lig t sp \u00e5 r. M e n det fin n s m y ck e t a n n a t s o m \u00e4 r v ik tig t a tt b e sk riv a o m m a n skall k u n n a \u00f6v e rs\u00e4 tta re la tiv sa tser. M a n fin n er em e lle rtid in o m E S T / G B o c h a n d r a teo re tisk a g ra m m a tik m o d e lle r in te t.e x . n \u00e5 g ra k o n k re ta reg ler f\u00f6 r val a v r\u00e4 tt fo r m a v relar tiv p r o n o m e n , n \u00e5 g o t s o m \u00e4 r n \u00f6 d v \u00e4 n d ig t o m m a n v ill im p le m e n te ra M T . D et ta la s o ft a v id p re se n ta tio n e n a v d e ssa m o d e r n a g ra m m a tik m o d e lle r so m o m b \u00f6 jn in g v o r e en triv ia l u p p g ift -m en d e t \u00e4r d en in te en s i en gelskan . En g r a m m a tik s o m skall a n v \u00e4 n d a s f\u00f6 r M T m \u00e5 ste h a regler s o m gen erera r \" w h o \" , re s p e k tiv e \" w h o m \" (f\u00f6 r a tt in te ta la o m \" w h o s e \" ). I e tt sp r\u00e5 k s o m g eorg isk a \u00e4r o r d b \u00f6 jn in g h u v u d u p p g ifte n : o m m a n ly ck a s f\u00e5 v e rb fo rm e n r\u00e4 tt \u00e4r d e t m e sta av sa tsen a v k la ra t. I b \u00e5 d e d e n ry sk a o c h d en sven sk a g ra m m a tik m o d u le n u p p ta s en a n sen lig m \u00e4 n g d a v reg lern a m ed a tt b \u00f6 ja o r d e n r\u00e4 tt." |
| }, |
| "FIGREF8": { |
| "uris": null, |
| "num": null, |
| "type_str": "figure", |
| "text": "a o c h d u b b e lt tra n sitiv a v e rb o c h u p p till tv \u00e5 s a ts a d v e rb ia l o c h tr e a n d r a a d v e rb ia l. S om a d v e rb ia l kan \u00e4ven f\u00f6 r e k o m m a p r e p os itio n sfra se r o c h u n d e r o r d n a d e k o n ju n k tio n sb isa tse r. S w e tra a c c e p te r a r kopular tiv a sa tse r (i ry sk a u ta n k o p u la ) o c h sa tser m e d p a ssiv u m , h j\u00e4 lp v e rb ( \" h a \" , m o d a la h j\u00e4 lp v e r b s o m \" kan \" o c h a sp e k tu e lla v e rb so m \" b \u00f6 r ja \" ). S w etra har o m fa tta n d e k o m p le x a v reg ler f\u00f6 r a tt h a n tera n om in a lfra se r m ed b e st\u00e4 m n in g a r f\u00f6 r e o c h e fte r h u v u d e t o c h k o m p lic e r a d k o n g ru e n s. S \u00e4rsk ild u p p m \u00e4 rk sa m h et har \u00e4 g n a ts \u00e5 t re la tiv sa tse r; d e t re fe re n tb e g re p p s o m g iv it R G dess n a m n v \u00e4 x te i sj\u00e4lva verk et fra m ur a n a ly sen a v rela tiv sa tser (se u p p s a tse r o m re la tiv sa tse r i littera tu rlista n ). S to ra lik v \u00e4 rd ig a g ra m m a tik m o d u le r iin n s f\u00f6 r en gelsk a, ry sk a o c h sven sk a o c h tillh \u00f6ra n d e lex ik on \u00e4 r f\u00f6 r n \u00e4 rv a ra n d e p \u00e5 n \u00e5 g ra tu sen o r d . D e m o n s tr a tio n s p r ogrzim v isa r an^llys a v sa tsen i k\u00e4 llsp r\u00e5 k et, o c h g en e re rin g a v m otsvaJ^lnde sa ts i m \u00e5lsp r\u00e5 k et. \u00d6 v e rs\u00e4 ttn in g e n ta r o ft a b a r a n \u00e5 g ra sek u n d er. D iv e rs e u p p s n a b bn in gsk n ep a n v \u00e4 n d es (s o m v i h \u00e4r in te t\u00e4 n k er a v s l\u00f6 ja ). G r o v \u00f6 v e r s \u00e4 ttn in g g \u00f6 r s v ia g em en sa m fu n k tio n e ll rep re se n ta tion , m en v issa tra n sfe rreg ler h a r d e s s u to m u ta rb e ta ts f\u00f6 r \u00f6 v e r s \u00e4 ttn in g i en viss rik tn in g (s\u00e4 rsk ilt ry sk a till en g elsk a ). T y p isk a em p irisk a p r o v \u00e4 r d esk rip tiv a te x te r , t y p n y h e tsn o tise r. S w e tra h ar b l.a . a n v \u00e4 n t n o tise r u r P r a v d a , o c h d e t \u00e4r tr o lig t a tt S w e tra sp e cia lise ra r sig p \u00e5 n otis-eller b u lle tin te x te r (se S igu rd & G a w ro n sk a -W e rn g re n 1 9 8 8 ). S \u00e5 d a n a te x te r Uknar d e te x te r s o m d e t te x tg e n e re ra n d e d a to r p r o g r a m m e t C o m m e n ta to r k u n d e gen ere ra o c h sam verkan m ellan C o m m e n ta to r o c h S w e tra f\u00f6 r a tt g en e re ra o c h \u00f6 v e r s \u00e4 tta s\u00e5 d a n a te x te r till n \u00e5 g ra sp r\u00e5 k h a r d isk u tera ts s o m e tt fr a m tid a p r o je k t. E tt e x e m p e l p \u00e5 en s\u00e5d an te x t \u00e4r f\u00f6 lja n d e : \" E tt o k \u00e4 n t fly g p la n n \u00e4 rm a d e sig \u00d6 la n d fr\u00e5n \u00f6 s te r ig\u00e5r. D e t g ir a d e s \u00f6 d e ru t in n a n d e t k o m in p \u00e5 sv en sk t o m r \u00e5 d e o c h f\u00f6rsva n n sed a n s \u00f6 d e ru t. In g et sv en sk t fly g p la n fiin n s d \u00e5 in o m o m r \u00e5 d e t. F ly g p la n e t o b se rv e ra d e s a v r a d a r.\" Y tte r lig a r e e x e m p e l p \u00e5 te x te r s o m S w etra rik tar in sig p \u00e5 finns i cirtiklar o m n \u00e4 m n d a i referen slista n . 5 Publiktrykket N \u00e4 r m an s\u00e4 g er a tt m an h \u00e5ller p \u00e5 m e d a u to m a tis k \u00f6 v e r s \u00e4 ttn in g f\u00e5 r m a n o fe lb a r t fr\u00e5 g a n : \" H u r m \u00e5 n g a \u00e5r d r \u00f6 je r d et in n a n d et g \u00e5 r, tr o r D u ? \" S va ret: \" A ld r ig \" a cce p te ra s in te o c h in te h eller svar a v ty p e n : \" V i kan \u00f6 v e r s \u00e4 tt a te x te r m ed 8 0 % k v a litet o m d e lig g e r in o m en v iss d o m \u00e4 n , t.e x . d et m a rin a o m r \u00e5 d e t o c h g ra m m a tik en \u00e4r s\u00e5 b e g r\u00e4 n sa d a tt d en in te till\u00e5 te r s a m o r d n a d e r e la tiv sa tse r o c h in te m er \u00e4n t v \u00e5 p rep osition s& a ser i en n o m in a lfra s e t c .\" D e t finn s e tt p u b lik tr y c k s o m d e s o m sysslax m ed m a sk in \u00f6 v e rs\u00e4 ttn in g a lltid m \u00e5 ste h a k \u00e4 n t. D e t \u00e4 r d e t tr y c k s o m lo ck a r en a tt s\u00e4 g a : \" O m 5 \u00e5 r, eller o m 10 \u00e5 r\" o c h kanske m u m la n \u00e5 g o t o h \u00f6 r b a r t o m te x tu e lla b e g r\u00e4 n sn in g a r. D e t \u00e4r det tr y c k s o m f\u00e5 tt m \u00e5 n g a forsk are a tt lo v a f\u00f6 r m y ck e t o c h d \u00e4 r ig e n o m tid v is f\u00f6 r s t\u00f6 r a m a rk n a d en f\u00f6 r forsk n in g in o m a u to m a tis k \u00f6 v e r s \u00e4 ttn in g . I a n s\u00f6 k n in g a rn a f\u00f6 r S w etra h ar v i a ld rig lo v a t a tt le v erera e tt fu n g e ra n d e \u00f6 v e rs \u00e4 ttn in g s s y s te m , b a ra a tt fo rsk a in o m a u to m a tis k \u00f6 v e rs\u00e4 ttn in g . O fta s\u00e4gs o ck s \u00e5 : \" M e n p o e s i o c h s k \u00f6 n litte ra tu r k o m m e r v \u00e4 l a ld rig a tt k u n n a \u00f6 v e rs \u00e4 tta s ?\" O ch n \u00e5 g ra to k r o lig a v a n d rin g e x e m p e l g es. D e t fin n s e tt tr y c k a tt d \u00e5 s\u00e4ga: \" N ej, d e t k o m m e r a ld rig a tt g \u00e5 .\" M e n h \u00e4 r k \u00e4 n n er ja g m ig o ft a lo ck a d a tt o c k s \u00e5 g \u00e5 p \u00e5 tv \u00e4 re n o c h s\u00e4 g a a tt: \" J o , p o e s i g \u00e5 r lik a b ra , fa st d e t b lir kanske in te rik tig t sa m m a p o e s i u ta n o ft a d j\u00e4 rv a re v \u00e4 n d n in g a r o c h fr\u00e4 sch a re m e ta fo r e r .\" M en \u00e5 h \u00f6 ra rn a v ill in te h eller h \u00f6 r a p \u00e5 d e t \u00f6 r a t. M T \u00e4 r e tt sv \u00e5 rt o m r \u00e5 d e -m en det \u00e4r ja g in te d en f\u00f6 r s te s o m k on sta te ra t." |
| } |
| } |
| } |
| } |
