{ "paper_id": "W89-0117", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T03:45:24.842903Z" }, "title": "", "authors": [ { "first": "Sabine", "middle": [ "K" ], "last": "Irchmeier-A Ndersen", "suffix": "", "affiliation": {}, "email": "" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "The treatment of coordination in a multilingual MT-system as EURO T R A poses two major problems: 1. to provide an algorithm for the monolingual analysis o f coordinate structures in accordance with the general linguistic theory for EU R O T R A which can treat basic coordination and the coordination o f incomplete constituents in the surface analysis and perform the mapping of these constructions onto the deeper levels. 2. to establish a semantic feature system, that captures the meaning of the conjunctions in order to facilitate their translation. Since research in this area within Eurotra on a multilingual basis is still ongoing, what will be presented is the present state of the art, which is a fairly complete theory for basic coordination as well as some ideas for the analysis of the more complex cases o f coordination involving gapping and movement.", "pdf_parse": { "paper_id": "W89-0117", "_pdf_hash": "", "abstract": [ { "text": "The treatment of coordination in a multilingual MT-system as EURO T R A poses two major problems: 1. to provide an algorithm for the monolingual analysis o f coordinate structures in accordance with the general linguistic theory for EU R O T R A which can treat basic coordination and the coordination o f incomplete constituents in the surface analysis and perform the mapping of these constructions onto the deeper levels. 2. to establish a semantic feature system, that captures the meaning of the conjunctions in order to facilitate their translation. Since research in this area within Eurotra on a multilingual basis is still ongoing, what will be presented is the present state of the art, which is a fairly complete theory for basic coordination as well as some ideas for the analysis of the more complex cases o f coordination involving gapping and movement.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "It is w ell k n o w n th a t c o o r d in a tio n is o n e o f th e m o s t p rev a len t a n d c o m p le x co n s tr u c tio n s in E u r o p e a n leinguages w h ich ca u ses g re a t In th e first p a rt o f th is p a p e r, I w o u ld like t o illu stra te h o w c o o r d in a tio n is h a n d led in E u r o tr a in a n a ly sis, in tra n sfer a n d in g e n e ra tio n cind I w ill co m m e n t o n so m e o f th e u n iversal m ech a n ism s o r co n s tra in ts , w h ich w e e x p lo it in o r d e r to a v oid o v e rg e n e ra tio n a n d t o fa cilita te th e tra n sla tio n o f c o o r d in a te d stru ctu re s.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "(F o r in fo rm a tio n o n th e E u r o t r a lin g u istic th e o r y as w ell as th e o v e ra ll o b je c t iv e s o f th e p r o je c t see P ersch k e 1986, J a sp a ert 1986, A r n o ld 1986 a n d R a w et al.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "1 9 8 9 ). In th e s e co n d p a rt I w ill g o in to tw o o f th e d ifficu lt a reas, th a t w e are w o rk in g o n a t th e m o m e n t. Basic Coordination-an Example B a s ic c o o r d in a t io n is d efin ed as c o o r d in a tio n o f co n stitu e n ts w ith th e sa m e ca t eg o ry , i.e. \" J o h n a n d M a r y \" , \" h a p p y o r s a d \" e tc.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "T h e fo llo w in g s e c tio n g iv e s a sh o rt d e s crip tio n o f th e tra n sla tion o f a sen ten ce c o n ta in in g a c o o r d in a te d s tr u c tu r e fr o m D a n ish in to E n glish . T h e fo llo w in g e x a m p le is u sed:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "(1 ) B \u00e5 d e U S A , J a p a n o g K in a v iser in teresse fo r E F b o t h u sa, ja p a n a n d c h in a sh o w in terest fo r eec (lite r a l tr a n s la tio n )", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "1" }, { "text": "T h e fig u res 1 a n d 2 sh o w th e stru ctu ra l rep resen ta tion o f th e sen te n ce (1 ) a t tw o o f th e s y n ta c tic levels in th e E u r o tr a th e o ry : E C S -th e c o n s titu e n t stru ctu re. T h e ir m a in fu n c tio n is t o g r o u p th e c o n ju n c ts a n d th e c o n ju n c tio n s to g e th e r in th e co n s titu e n t a n a ly sis at E C S a n ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "A nalysis", "sec_num": "2.1" }, { "text": "In th e tra n sfer ph a se, th e D a n ish -E n g lish tra n sfer m o d u le tra n sla tes th e le x ica l valu es o f th e c o n ju n c ts in to E n g lish . T h is is p r a c tic a lly a ll th a t h a p p e n s. T h e rest o f th e stru ctu re is tra n sla ted b y a d e fa u lt m e ch a n ism , th a t s im p ly tra n fers it in to th e ta rg e t la n g u a g e IS s tru ctu re w ith o u t a n y ch a n g es (fig u re 3 ). T h e fin al resu lt is th e fo llo w in g sen ten ce:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Transfer", "sec_num": "2.2" }, { "text": "(2 ) T h e U S A , J a p a n a n d C h in a ta k e in terest in th e E E C", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "S \"I____________", "sec_num": null }, { "text": "A t E C S in g e n e ra tio n th e \" c s (3 ) J a p a n a n d th e U S A a n d C h in a ta k e in terest in th e E E C T h is is a m a tte r o f s ty le a n d ta ste . ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "S \"I____________", "sec_num": null }, { "text": "In o rd e r t o p e rfo rm th e tra n sla tio ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Universal Constraints", "sec_num": "3" }, { "text": "1. 2 . 3. X -> csl[X] *cs2[X] csl[X] -> (preconj) X conj cs2[X] -> conj X X is a v", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Beu* Level Constreunts", "sec_num": "3.1" }, { "text": "I f w e w a n t t o v e rg e n e ra tio n at E C S b e fo r e th e v a lid a tio n o f th e c o n s tr u c tio n s c a n ta k e p la ce a t E R S a n d IS. In th e E n g lish a n d D a n ish sen te n ces th e ru les fo r ' A c r o s s -th e -b o a r d e x tr a c t io n ' seem t o b e th e sa m e a n d w e ca n p r o d u c e a sim ilar su rfa ce stru ctu re. In th e tra n s la tio n in to G e r m a n (5 b ) it is o b v io u s th a t th e sa m e m ech a n ism d o e s n o t w o r k a n d th a t th e se n te n ce is u n g ra m m a tica l. In G e rm a n , In ( 8 ) -( 8 b ) an equ iv a len t stru ctu re c a n n o t b e b u ild n eith er in D a n ish n o r in G erm a n . je c t it takes-eith e r t o \" leg en \" / \" lae g g e\" o r t o \" ste lle n \" / \" stille \" in G e rm a n a n d ", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": ".H ier2irchy", "sec_num": "3" }, { "text": "Proceedings of NODALIDA 1989, pages 191-203", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "ComputationalLinguistics -Reykjavik 1989", "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": "General view o f the design methodology", "authors": [ { "first": "D", "middle": [], "last": "Arnold", "suffix": "" } ], "year": 1986, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Arnold, D. 1986. General view o f the design methodology. MxiltUingua 5-3/1986.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "The Mental Representation o f Grammatical Relations", "authors": [ { "first": "J", "middle": [], "last": "Bresnan", "suffix": "" } ], "year": 1982, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Bresnan, J. [ed]. 1982. The Mental Representation o f Grammatical Relations. MIT Press, Cambridge, Mass.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "Aspects o f the theory o f syntax", "authors": [ { "first": "Noam", "middle": [], "last": "Chomsky", "suffix": "" } ], "year": 1965, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Chomsky, Noam. 1965. Aspects o f the theory o f syntax. M IT Press, Cambridge, Mass.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Elementaer Dansk Grammatik", "authors": [ { "first": "Paul", "middle": [], "last": "Diderichsen", "suffix": "" } ], "year": 1946, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Diderichsen, Paul. 1946. Elementaer Dansk Grammatik. Gyldendal, K\u00f8benhavn.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "Coordination. North-HoUand", "authors": [ { "first": "Simon", "middle": [], "last": "Dik", "suffix": "" } ], "year": 1972, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Dik, Simon. 1972. Coordination. North-HoUand, Amsterdam.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Generalized Phrase Structure Grammar", "authors": [ { "first": "G", "middle": [], "last": "Gazdar", "suffix": "" } ], "year": 1985, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Gazdar, G. et. al. 1985. Generalized Phrase Structure Grammar. Harvard University Press.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Moderne dansk. Grafisk Forlag", "authors": [ { "first": "A", "middle": [], "last": "Hansen", "suffix": "" } ], "year": 1967, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Hansen, A. 1967. Moderne dansk. Grafisk Forlag, K\u00f8benhavn.", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "The levels of representation", "authors": [ { "first": "L", "middle": [], "last": "Jaspaert", "suffix": "" } ], "year": 1986, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Jaspaert, L. 1986. The levels of representation. Multilingua 5-3/1986.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Tysk Grammatik I-III", "authors": [ { "first": "P", "middle": [], "last": "J\u00f8rgensen", "suffix": "" } ], "year": 1966, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "J\u00f8rgensen, P. 1966. Tysk Grammatik I-III. Gad, K\u00f8benhavn.", "links": null }, "BIBREF9": { "ref_id": "b9", "title": "The Semantics of Coordination, SLCS 9. John Benjamins", "authors": [ { "first": "E", "middle": [], "last": "Lang", "suffix": "" } ], "year": 1984, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Lang, E. 1984. The Semantics of Coordination, SLCS 9. John Benjamins, Amsterdam.", "links": null }, "BIBREF10": { "ref_id": "b10", "title": "Knowledge betsed M\u00e9ichine Translation. Machine Thinslation", "authors": [ { "first": "S", "middle": [], "last": "Nirenburg", "suffix": "" } ], "year": 1989, "venue": "", "volume": "3", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Nirenburg, S. 1989. Knowledge betsed M\u00e9ichine Translation. Machine Thinslation vol. 3 and 4.", "links": null }, "BIBREF12": { "ref_id": "b12", "title": "An Introduction to the Eurotra Machine Translation System", "authors": [ { "first": "T", "middle": [], "last": "Raw", "suffix": "" } ], "year": 1989, "venue": "Working Papers in Natural Language Processing", "volume": "1", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Raw, T. et al. 1989. An Introduction to the Eurotra Machine Translation System. Working Papers in Natural Language Processing vol. 1, Eurotra-Leuven.", "links": null } }, "ref_entries": { "FIGREF0": { "uris": null, "num": null, "text": "d ifficu lties in all sy n ta c t ic fo rm a lism s. A t p resen t in E u r o tr a , n o t all a s p e c ts o f c o o r d in a tio n are co v ered . T h e cu rren t im p le m e n ta tio n s o f th e g ra m m a rs o f th e m o n o lin g u a l m o du les h a n d le b a s ic c o o r d in a tio n o f alike c o n s titu e n ts , b u t n o t s p e c ia l ca ses like c o o r d in a tio n o f u n lik e o r in c o m p le te c o n s tr u c tio n s . C o o r d in a tio n is still a t o p ic o f o n g o in g research b o t h fr o m a m o n o lin g u a l a n d a c o n tra s tiv e p o in t o f v iew .", "type_str": "figure" }, "FIGREF1": { "uris": null, "num": null, "text": "IS an d IS-th e d e e p s y n ta c tic stru ctu re . B etw ee n th e se levels w e h a v e a th ir d level, E R S -th e su rfa ce s y n ta c tic s tru ctu re , b u t w ith r e sp e ct t o c o o r d in a t io n th e E R S an d IS rep resen ta tion s axe q u ite sim ila r s o E R S is left o u t in th ese ex a m p le s . C o m p a r in g th e tw o stru ctu ra l rep rese n ta tio n s in fig u re 1 a n d fig u re 2, w e ca n n o te th a t th e in te rm e d ia te n o d e s a t E C S , \" c s l \" a n d \" c s 2 \" , h a v e b e e n d e le te d a t IS.", "type_str": "figure" }, "FIGREF2": { "uris": null, "num": null, "text": "d t o p re v en t o th e r c a te g o r ie s fr o m e n te rin g th e c o o r d in a te s tr u c tu r e d u rin g th e p^lrsing p ro ce s s . T h e rep resen ta tio n a t E C S is in s o m e re sp e cts s im ila r t o th e o n e p re sen te d b y G a z d a r et al. (1 9 8 5 ). G P S G uses a co n stitu e n t sim ila r t o th e cs2 n o d e fo r all p a rts o f th e c o o r d in a te d stru ctu re . W e h ave in tr o d u c e d a m o re c o m p le x c o n s titu e n t, c s l , fo r th e first tw o c o n ju n c ts in o rd e r t o sp e e d u p th e p a rsin g p ro ce ss. T h e IS lev el, w h ere th e c o n s titu e n ts axe d efin ed as d e p e n d e n c y stru ctu re s co n s is tin g o f a g o v e r n o r a n d a n u m b e r o f a rg u m en ts a n d m o d ifie rs, is ch a r a c terized b y th e c a n o n ica l o r d e r in g o f th ese co n s titu e n ts . S in ce c o o r d in a tio n is n ot a d e p e n d e n c y rela tio n , th e su rfa ce o r d e r o f th e c o n s titu e n ts in c o o r d in a t e stru ctu re s is n o t ch a n g ed . T h e c o n ju n c tio n , th e c o m m a a n d th e p r e c o n ju n c t io n \" b \u00e5 d e \" a re d e le te d a t IS. T h e ir sem a n tic co n te n t is p reserv ed as a fe a tu re \" [co n ju n c tio n ]\" o n th e to p n o d e o f th e c o o r d in a te d stru ctu re.", "type_str": "figure" }, "FIGREF3": { "uris": null, "num": null, "text": "Figure 4: E C S", "type_str": "figure" }, "FIGREF4": { "uris": null, "num": null, "text": "\" n o d e s h a v e re a p p e a re d . T h e y m ak e it easier t o in sert th e c o n ju n c tio n s in th e righ t p la ce s. In sy n th esis w e d o n o t n eed to d istin g u ish b e tw e e n \" c s l \" a n d \" c s 2 \" , b e c a u s e w e k n o w fr o m th e an alysis, e x a c tly w h ich co n s titu e n ts a re t o b e c o n jo in e d . A n eq u iv a len t t o th e p r e c o n ju n c t io n \" b \u00e5 d e \" h as n o t b e e n cre a te d a n d th is is p a rt o f th e ju s tific a t io n fo r th e fe a tu riz a tio n o f th e c o n ju n c tio n s , sin ce th e o c c u r re n ce o f th ese p r e c o n ju n c t io n s sh o w m o n o lin g u a l v a ria tion s. In E n glish \" b o t h \" is c o n s id e r e d a s tr ic tly b in a r y c o n ju n c tio n , p r o b a b ly b e c a u s e o f th e h o m o n y m y w ith th e q u a n tifie r w ich c o r r e s p o n d s t o \" b e g g e \" in D a n ish . In D a n ish \" b \u00e5 d e \" m a y o c c u r w ith a n y n u m b e r o f c o n ju n c ts . W e c a n a ls o n o te , th a t th e c o m m a h as b e e n tra n sla ted as a c o m m a a n d n o t as a le x ica l c o n ju n c t io n as in (3 ):", "type_str": "figure" }, "FIGREF5": { "uris": null, "num": null, "text": "L ik e a jo k e r in a g a m e o f ca rd s , th e c o m m a c a n ta k e th e p la c e o f a n y c o n ju n c t io n in a c o o r d in a te stru ctu re , p r o v id e d th a t th ere a re a t lea st th re e c o n ju n c ts a n d th a t th e c o m m a d o e s n o t a p p e a r w ith th e first o r th e la st c o n ju n c ts (e x c e p t in e n u m e r a tio n s ). I f w e a llo w e d su ch v a ria tion s w e w o u ld g e t a m u ltip le o u tp u t, a n d sin ce th e o u tp u t o f a m a ch in e tra n sla tion s y s te m p e r fe r a b ly sh o u ld b e o n e s o lu tio n , w e h ave ch osen th e o n e in (2 ).", "type_str": "figure" }, "FIGREF6": { "uris": null, "num": null, "text": "n d e s c r ib e d a b o v e , w e h ave relied o n a n u m b e r o f gen era l p ro p e rtie s , th a t seem t o b e th e sa m e fo r all c o o r d in a te d s tru ctu re s.", "type_str": "figure" }, "FIGREF7": { "uris": null, "num": null, "text": "In o r d e r t o b e a b le t o c o o r d in a te all ca te g o rie s , w e w a n t t o u n d e r s p e c ify th e ca te g o r y valu e. T h e a d v a n ta g e is o f co u rs e th a t w e ca n use o n ly o n e b a s ic set o f rules t o h a n d le all ca ses o f c o o r d in a tio n o f alik e co n s titu e n ts :", "type_str": "figure" }, "FIGREF8": { "uris": null, "num": null, "text": "a ria b le w h ich is in sta n tia te d w ith th e c a te g o r y v alu e o f th e c o n ju n c t an d p e r c o la te d t o th e t o p n o d e o f th e c o o r d in a te d s tru ctu re . U sin g u n ifica tio n w e ca n en su re, th a t X in e v e ry ru le c o n ta in s th e sa m e c a te g o ry . T h e s e v e ry g en era l rules, h ow ev er, ten d t o c o o r d in a te c a te g o rie s a t e v e ry b a r level th u s p e r fo r m in g an an lysis w h ich is o b v io u s ly w ro n g . C o n s id e r ca se s as (4 ): u e n tly , w e a llo w o n ly c o o r d in a tio n o f c o n s titu e n ts a t th e sa m e b a r level a n d n o t a t b a r level z e r o as in (5 ). A sim ila r a p p r o a c h has b e e n d e v e lo p e d b y N ire n b u rg (1 9 8 9 ). or th e in sertion o f th e c o r r e c t c o n ju n c tio n s in sy n th esis it is im p o r ta n t t o ca l cu la te w h e th e r th e c o o r d in a te c o n s tr u c tio n co n sists o f tw o o r m o re th a n tw o elem en ts, as d e m o n s tr a te d fo r \" b o t h \" in (1 ) a n d (2 ). A c o o r d in a t e s tr u c tu r e is b in a ry i f it co n sists o f o n ly tw o c o n ju n c ts as in (6 ): (6 ) B o th X a n d Y Ite ra tiv e c o o r d in a t e stru ctu re s c o n sist o f m o re th a n tw o co n ju n c ts as in (7 ): (7 ) X a n d Y a n d Z a n d ................................. T h e o c c u r r e n c e o f p r e c o n ju n c t io n s in b in a ry a n d ite ra tiv e c o o r d in a tio n is la n g u a g e sp e c ific as sh o w n in th e tra n sla tio n e x a m p le a b o v e .", "type_str": "figure" }, "FIGREF9": { "uris": null, "num": null, "text": "o in tr o d u c e th e c o n c e p t o f h iera rch y fo r c o o r d in a te d stru ctu res, it en ta ils th a t o u r ru les h a v e t o b e recu rsiv e. T h is m ea n s th a t w e m u st b e a b le to c r e a te d e e p re p re se n ta tio n s as in fig u re 5 as w ell as th e flat o n e s presen ted in fig u re 1 a n d fig u re 4.", "type_str": "figure" }, "FIGREF10": { "uris": null, "num": null, "text": "E C S T h is s tr u c tu r e is h iera rch ica l in th e sen se th a t o n e o f th e co n ju n c ts is a c o o r d in a te d s tr u c tu r e itself. A llo w in g th is ty p e o f recu rsio n w ith o u t re striction s w o u ld ca u se h e a v y o v e rg e n e ra tio n a t E C S . T h e n u m b e r o f p o s s ib le stru ctu res g r o w s fa st w ith th e n u m b e r o f c o n ju n c ts . 2 c o n ju n c ts = 1, 3 c o n ju n c ts = 3, 4 c o n ju n c ts = 11 e tc . A g a in w e p refer o n ly o n e s o lu tio n . T h e co n stra in t th a t we p o s e o n th e se stru ctu re s is s im p ly th a t c o o r d in a te stru ctu res m a y n o t im m ed i a te ly d o m in a t e c o o r d in a t e stru ctu re s o f th e sa m e ty p e (se e 3 .4 b e lo w ). In the E u r o t r a fo rm a lis m th is is e x p re sse d in te rm s o f filter o r killer rules w h ich delete th e u n d e sired s tru ctu re .", "type_str": "figure" }, "FIGREF11": { "uris": null, "num": null, "text": "in a lly , w e h a v e t o p o s e c o o c c u r r e n c e r e s trictio n s o n th e c o n ju n c tio n s b o t h w ith re g a rd t o th eir se m a n tic t y p e a n d th eir p a r a d ig m a tic d is trib u tio n . T h e w a y in w h ich c o o c c u r r e n c e re s tr ic tio n s a re ex p re sse d is a m o n o lin g u a l m a tte r. F or E ur o p e a n la n g u a g es, w e ca n d istin g u ish 5 b a s ic u n iv ersa l ty p e s o f co o r d in a tio n . sem a n tic r e s trictio n s im p ly th a t w e c a n n o t h a v e a c o o r d in a t io n like(8 )o n th e sa m e h iera rch ica l level.(8 ) * b \u00e5 d e X eller Y T h e p o s itio n a l re strictio n s fo r th e c o n ju n c tio n s o p e r a te w ith 3 p o s itio n s : in itial, n o n -in itia l a n d final. In b in a ry c o o r d in a tio n o n ly th e in itia l a n d fin al p o s itio n s a re u sed. In ite ra tiv e c o o r d in a tio n a th e o r e tic a lly in fin ite n u m b e r o f c o n ju n c tio n s m a y a d d itio n a lly a p p e a r in n o n -in tia l p o s itio n . c o o r d in a te stru ctu res c o n s is tin g o f alik e c o n s titu e n ts a re th e m o s t fre q u en t on es, th e rules eind co n stra in ts d iscu ssed a b o v e ca n h a n d le m a n y o f th e cases o f c o o r d in a tio n o c c u r r in g in th e te x t ty p e s w e a re w o rk in g w ith . H ow ev er, th ere are still a n u m b e r o f cases w h ere th is b a s ic tre a tm e n t is n o t su fficien t, s o m e o f w h ich w ill b e trea ted in th e s e c o n d p a rt o f th is p a p e r. q u e stion is n o w h o w t o in te g ra te th e c o o r d in a tio n o f u n lik e c o n s titu e n ts in to th e sy ste m sk etch ed a b o v e . B y c o o r d in a tio n o f u nlike co n stitu e n ts w e m ea n c o o r d in a t e d stru ctu re s c o n sistin g o f d ifferen t ca te g o rie s as in ( l ) -e x a m p le s sh o w s o m e o f th e co m b in a tio n s o f c o n s titu e n ts th a t m a y o c c u r in c o o r d in a t e d stru ctu re s in D a n ish te x ts . H ow ever, th ere are certa in restriction s w ith reg a rd t o w h ich u n lik e c o n s titu e n ts ca n b e c o n jo in e d , b u t th e rules fo r th ese re s tr ic tio n s a re n o t e a sy t o d e te rm in e fr o m a su rfa ceorien ted c o n s titu e n t an alysis. u n d e r ly in g p a tte r n e m erg es i f w e lo o k at th e s y n ta c tic fu n c tio n s o f the c o n s titu e n ts in ste a d o f th e ir g r a m m a tic a l ca teg o ry . ( 1 ) : s u b j V m o d ifie r + m o d ifie r ( 2 ) ; s u b j V a tt r .s u b j + a tt r .s u b j ( 3 ) : s u b j V a tt r .s u b j + a tt r .s u b j ( 4 ) : s u b j V m o d ifie r + m o d ifie r ( 5 ) : s u b j V o b j + o b j ( 6 ) : * s u b j V m o d ifie r + in te r n a l.o b j ( 7 ) : * s u b j V a tt r .s u b j + m o d ifie r ( 8 ) : * s u b j V o b j + m o d ifie r D id e rich se n (1 9 4 6 ) n o te s th a t: L e d d e n e i en s id e o r d n e t F o rb in d e lse sta a r h vert fo r sig i sa m m e F o r h o ld til et tr e d je L e d s o m H elh ed en a n d in th is w a y h e in tr o d u c e s th e s y n ta c tic fu n c tio n s o f th e co n stitu e n ts as a c r ite r io n fo r a c o r r e c t c o o r d in a tio n . G a z d a r et al. (1 9 8 5 ) a lso b a se th e ir an alysis o f c o o r d in a te d stru ctu re s o n th is o b s e r v a tio n . In G P S G , c o n s titu e n t s tr u c tu r e in fo rm a tio n as w ell as th e s y n ta c tic in fo r m a tio n ca n b e a cce sse d a t th e sa m e tim e , a u d c o n se q u e n tly in G P S G th e c o o r d in a tio n o f u n lik e c o n s titu e n ts is tre a te dq u ite e leg a n tly : rtu n a tely , w e c a n n o t ju s t tra n sfer th is a n a ly sis d ir e c t ly t o th e E u ro tria nm o d e l, b e ca u se o f th e s tra tifica tio n a l d esig n o f th e s y ste m w h ich im p lie s th a t differen t in fo rm a tio n is c o m p u te d a t differen t levels. T h is m ea n s th a t w e c a n n o t a ccess th e in fo rm a tio n a b o u t th e s y n ta c tic fu n c tio n o f th e c o n s titu e n ts a t E C S b e ca u se it is n o t c o m p u te d b e fo r e th e n e x t lev el. W e h a v e t o w r ite ru les fo r all p o ssib le co m b in a tio n s o f c o n s titu e n ts a t E C S . T h e resu lt is a p ro v is io n a l", "type_str": "figure" }, "FIGREF12": { "uris": null, "num": null, "text": "o st w ell k n o w n a c c o u n t fo r th e c o o r d in a tio n o f in c o m p le te co n s titu e n ts is p r o b a b ly th e g e n e ra tiv e o n e a d v o c a te d b y C h o m s k y in 1965. A c c o r d in g t o C h om sk y , a c o o r d in a t e su rfa ce stru ctu re is d e riv e d b y m e a n s o f c o n ju n c tio n r e d u ctio n fr o m tw o p a ra llel sen ten ces in th e d e e p s tru ctu re . H an elsk ed e h aven S im on D ik (1 9 7 2 ) led th is th e o r y a d a b s u rd u m w ith an e x a m p le like th e follo w in g , w h ich he cla im s lea d s t o 81 differen t sen ten ces in th e d e e p stru ctu re . n o g K a r l o g K u r t sae lger ae b ler o g p ae re r o g b a n a n er i K \u00f8 b e n h a v n , O d e n se o g \u00c5 rh u s m a n d a g , tirsd a g o g o n s d a g J oh n sae lger ae b ler i K \u00f8 b e n h a v n m a n d a g . K a r l sae lg er p ae rer i O d e n s e tirsd a g . K u r t ___ e tc . til 81. T h e q u e s tio n is n o w , w h e th e r c o n ju n c tio n re d u c tio n rea lly is an irrelevant ru le o r w h e th e r th e c o n c e p t o r s o m e in sta n ce o f it c o u ld b e useful. F rom th e p o in t o f v ie w o f m a ch in e tra n s la tio n o n e o f th e first th in g s t o b e in v estig a ted is th e tr a n s la tio n a l rele v a n ce. C o n s id e r th e fo llo w in g e x a m p le s: (5 ) I k n o w th e w o m a n w h o p a in ted _ a n d y o u m et th e m a n w h o s to le the p ic t u r e th a t H a rry w as s o fo n d o f _ (5 a ) je g k en d er d e n k v in d e s o m m a le d e _ o g d u m \u00f8 d te d en m a n d , so m stja l det b ille d e , s o m h a rry va r s\u00e5 g la d fo r _ (5 b ) * ich k en n e d ie F ra u , d ie _ m a lte u n d d u tra fst d en M a n n , d er dafi B ild sta h l, d a s H a rry _ s o g ern m o c h te (5 c ) ich ken n e d ie F rau , d ie d a s B ild , dafi H a rry s o g ern m o c h te , m a lte , u n d du tra & t d e n M a n n , d e r es stah l.", "type_str": "figure" }, "FIGREF13": { "uris": null, "num": null, "text": "d ifferen t op erar tio n s h a v e t o ta k e p la c e i f w e w an t t o cre a te an a d e q u a te tra n sla tion . In ord e r t o d o th is, th e G e r m a n g e n e r a tio n m o d u le m u st h ave a cce ss t o th e c o m p le te in fo r m a tio n . (6 ) J o h n o ffe re d _ a n d H a rry g a v e S a lly a C a d illa c (6 a ) J o h n t ilb \u00f8 d _ o g H a rry g a v S a lly en Cadillac (6 b ) * J o h n b o t _ (a n ) u n d H a rry g a b S a lly ein en C a d illa c In ( 6 ) -( 6 b ) th e sa m e p r o b le m arises, h ere tw o co n stitu e n ts are e x tr a c te d fr o m th e first c o n ju n c t a n d a g a in an eq u iv a len t su rfa ce rea liza tion is im p o ssib le in G e r m a n b e c a u s e o f th e d e ta c h e d p r e p o s itio n \" a n \" . (7 ) sh ow s th a t this ty p e o f e x tr a c tio n is p o s s ib le i f th e v e r b d o e s n o t h ave a d e ta c h e d p refix. (7 ) J o h n v erk a u fte u n d H a rry g a b S ally ein en C a d illa c", "type_str": "figure" }, "FIGREF14": { "uris": null, "num": null, "text": "t th e b o o k aw ay an d _ th e gla ss o n th e ta b le (8 a ) * jo h n le g te d a s b u ch w eg u n d _ d a s g la s a u f d en tisch (8 b ) * jo h n la g d e b o g e n v ae k o g _ gla sset p \u00e5 b o r d e t (8 a a ) J oh n leg te d a s B u c h w eg u n d stellte d a s G la s a u f d en T is c h T h e v e rb \" p u t\" is tra n sla ted d iffere n tly d e p e n d in g o n th e n a tu re o f th e o b", "type_str": "figure" }, "FIGREF15": { "uris": null, "num": null, "text": "an ish . A n ew v e r b m u st b e in tr o d u c e d in th e tw o ta rg e t la n g u a g e s t o en su re th e co r r e c t tra n sla tio n o f th e s e c o n d p a rt o f th e cla u se t o m a tch th e s e m a n tic fea tu res o f th e o b je c t . In th e se ca ses a c o m p le tio n o f th e in c o m p le te c o n s titu e n ts b y fillin g th e g a p s w o u ld m a k e th e tra jisla tio n m u ch easier. rom th e e x a m p le s in s e ctio n 4 it is o b v io u s th a t in s o m e ca ses g a p s h a v e t o b e filled at IS. W e c a n n o t b e su re th a t th e in c o m p le te c o n s titu e n ts h a v e eq u iv a len ts in th e ta rg e t la n g u a g e . T h e p r o c e s s o f r e d u c tio n a n d e x tr a c tio n is m o n o lin g u a lly d e te rm in ed a n d h ea v ily in flu en ced b y p h e n o m e n a as su rfa ce s y n ta x , h o m o n y m y a n d sele ction a l p ro p e rtie s o f le x ica l ite m s. T h is m ea n s th a t w e h a v e t o lea ve it u p t o th e ta rg et level g e n e ra to r t o p r o d u c e th e c o r r e c t d e g ree o f r e d u c tio n o n th e b a sis o f th e m a x im a l stru ctu re a t IS. H ow ev e r, th ere is n o n eed t o tre a t e v e ry ty p e o f c o o r d in a tio n as a r e d u c tio n . T h is w o u ld lea d t o r id icu lo u s m u ltip lic a tio n s o f th e stru ctu re s as a lre a d y n o te d b y D ik . T h e r e fo r e , as a w o rk in g s tra te g y w e a re p u rsu in g th e fo llo w in g a p p r o a c h to th ese p ro b le m s : W e d istin g u ish tw o ty p e s o f c o o r d in a te stru ctu res: 1. C o o r d in a tio n o f a r g u m e n ts /m o d ifie r s 2. C o o r d in a tio n o f g o v e rn o rs In ca se o f c o o r d in a tio n o f a r g u m e n ts /m o d ifie r s th e re d u c e d s tr u c tu r e is n o t re b u ilt a t IS , b u t sim p ly k ep t as th e c o o r d in a tio n o f tw o o r m o r e c o n s titu e n ts as in figu re 6 , w h ich iU ustrates th e fo llo w in g sen ten ce:(9 ) J o h n a n d P e te r left", "type_str": "figure" }, "FIGREF16": { "uris": null, "num": null, "text": "In ca se o f c o o r d in a t io n o f g o v e rn o rs (1 0 ) th e m issin g a rg u m e n ts a re in serted and c o in d e x e d w ith th e c o r r e s p o n d in g co n stitu e n ts in th e first c o n ju n c t in o r d e r to c re a te c o m p le t e stru ctu re s as in fig u re 7, w h ere th e c o o r d in a tio n o f th e tw o v erbs is tr a n s fo r m e d in to th e c o o r d in a tio n o f tw o sen ten ces. O n ly th e v a le n cy b o u n d a rg u m e n ts a re c o p ie d t o en su re th e c o m p le te n e s s a n d co h e re n ce o f th e stru ctu re ciccord in g t o th e d e fin itio n o f th e IS stru ctu re. (1 0 ) W e g a th e re d a n d m a rch e d fo r sev era l h ou rs c o m p le tio n o f th e c o o r d in a te d s tru ctu re a lso takes p la ce i f m o re than o n e g a p is fo u n d in th e c o o r d in a t e d s tr u c tu r e as in (8 ) a b o v e , w h ere th e secon d p a rt o f th e s tr u c tu r e co n s is ts o n ly o f th e o b je c t (th e g la s) a n d th e m o d ifie r (o n th e ta b le ). In th is ca s e b o t h th e s u b je c t a n d th e v e r b a re in serted an d c o in d e x e d . T h u s, in c o m p le te stru ctu res a re o n ly m a d e c o m p le te i f w e a re d e a lin g w ith th e c o o r d in a tio n o f g o v e rn o rs o r sm a ll cla u ses. W e b e lie v e th a t a la rg e a m o u n t o f p ro b le m cases ca n b e s o lv e d in th is way. 6 Conclusion A lth o u g h th e th e o r y fo r b a s ic c o o r d in a tio n is w ell d e v e lo p e d a n d w ell fu n c tio n in g in th e E U R O T R A sy ste m , th ere are still a n u m b er o f p r o b le m s th a t w e h a v e n o t m e n tion ed here i.e. th e ro le o f n e g a tio n in c o o r d in a te stru ctu re s, th e c a lc u la tio n o f featu res, th e d e te r m in a tio n o f th e ca te g o ria l sta tu s o f a c o o r d in a t e s tru ctu re th a t co n sists o f u n lik e co n s titu e n ts e tc . S o m e o f th e se h a v e b e e n s o lv e d w h erea s oth ers still a re th e t o p ic o f o n g o in g research . F or c o m p le x c o o r d in a tio n th e p ic tu r e is m o r e u n clea r sin ce w e h a v e t o d ea l w ith g a p p in g b o th & o m a m o n o lin g u a l a n d a tra n sla tio n a l p o in t o f v ie w . H o w ever, w e b e lie v e th a t th e c o m p a r a tiv e research w ill g iv e fru itfu l in p u t t o th e m o n o lin g u a l an alysis.", "type_str": "figure" } } } }