|
{ |
|
"paper_id": "W98-0124", |
|
"header": { |
|
"generated_with": "S2ORC 1.0.0", |
|
"date_generated": "2023-01-19T06:04:33.939729Z" |
|
}, |
|
"title": "Wh-dependencies in Romanian and TAG", |
|
"authors": [ |
|
{ |
|
"first": "Manuela", |
|
"middle": [], |
|
"last": "Leahu", |
|
"suffix": "", |
|
"affiliation": { |
|
"laboratory": "", |
|
"institution": "Universite", |
|
"location": { |
|
"addrLine": "Paris 2,place Jussieu", |
|
"postCode": "7003, 75251, cedex 05", |
|
"settlement": "Paris" |
|
} |
|
}, |
|
"email": "[email protected]" |
|
} |
|
], |
|
"year": "", |
|
"venue": null, |
|
"identifiers": {}, |
|
"abstract": "", |
|
"pdf_parse": { |
|
"paper_id": "W98-0124", |
|
"_pdf_hash": "", |
|
"abstract": [], |
|
"body_text": [ |
|
{ |
|
"text": "The aim of this paper is to analyse the whmovement for Romanian in TAG formalisrn. Romanian shares free extractibility from tensed clauses with its Rornance sister languages and it has borrowed multiple wh-fronting from the slavic languages. These features of Rornanian are quoted by Kroch (1989) frorn Comorovsky (1986) , where he justifies the analysis of extractions in TAG. This formalism allows a correlation between the absence of wh-islands and the possibility of multiple wh -movement. But the facts of the Romanian language are more complex .We consider here several phenornena like simple questions, unbounded dependencies, wh-islands, multiple wh -movement . Because of order between the free wh-words for the multiple wh-movement , a complete analysis is not possible with TAG. TAG derivation trees do not provide a good ' representation of the dependencies between the words of the sentence, i.e\" of the predicateargument and modification structure. Also, the derivation structures of MCTAG (Joshi,1987 ) cannot be given a linguistically meaningful interpretation (Section 3). We show here that an analysis is possible with DTG formalism (Vijay-Shanker, D. Weir, 0. Rambow,1995) that resolve these problems with the use of a single operation -that we call subsertion -for handling all complementation. 1", |
|
"cite_spans": [ |
|
{ |
|
"start": 284, |
|
"end": 296, |
|
"text": "Kroch (1989)", |
|
"ref_id": null |
|
}, |
|
{ |
|
"start": 303, |
|
"end": 320, |
|
"text": "Comorovsky (1986)", |
|
"ref_id": "BIBREF0" |
|
}, |
|
{ |
|
"start": 1004, |
|
"end": 1015, |
|
"text": "(Joshi,1987", |
|
"ref_id": null |
|
}, |
|
{ |
|
"start": 1151, |
|
"end": 1191, |
|
"text": "(Vijay-Shanker, D. Weir, 0. Rambow,1995)", |
|
"ref_id": null |
|
} |
|
], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Introduction", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "(l) Pe cinei vede Ion ei ? Who; sees Ion e; ? This sentence in the TAG formalism is represented as a transitive tree with object extraction and the initial place of the extraction is marked by a trace . A characteristic feature for questions is the inversion of the subject.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Simple Questions", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "The following sentences illustrate some examples of unbounded dependencies :", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Unbounded dependencies", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "(2) Ce i regreti ca a citit Maria e; ?", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Unbounded dependencies", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "What ido you regret that Mary has read e, ? The wh -pronoun in the initial tree is in the same verb with which it is construed and i ts interpretation as the object of the verb \" to read \u2022\u2022 is thus guaranteed . Following standard conventions we represent the relationship between the fronted constituent and thc position in which phrases with its grammatical role normally appear by coindexing the fronted wh with an empty category. The relationship between an indexed empty category and the categorially identical, c-commanding node with which it is in coindexed , we call \"linking \". The adjunction of the auxiliary tree in the initial tree produces the final tree in which the wh-word is now initial in the matrix sentence. Strikingly, there is no bound on the depth of the embedding: (3)Pe eine i crezi ca Paul a zis ca Ion a placut e;?", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Unbounded dependencies", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "Who ido you think that Paul said that Ion liked e ;? In (3), the wh-word is an argument of the most deeply embedded verb\" like\" , thus causing the non-projectivity. A TAG can capture the longdistance dependency naturally, since the recursive adjunction operation allows an unbounded number of clauses to intervenue between directly dependent lexemes. We first substitute all nominal arguments into their respective verbal trees , and then adjoin thc: intermediate say -clause into the most deeply embedded /ike-clause at the S node immediately dominated by the root. This has the effect of separating the wh-word from its verb. evcn though they originated in the same structure . We than subsequently adjoin the matrix thi11kclause into the intennediate say-clause .", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Unbounded dependencies", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "Islands phenome na can be found in Romanian for relative clauses and adjuncts. (4)* Pe eine; cunosti femeia i care i a inta lnit e; ?", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Wh-islands", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "Who i do you know the woman i which ie; met e;? (5) * Pe eine ; ai plecat inainte sa examineze ei Ionescu ?", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Wh-islands", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "Who i did you leave before that examine e i Ionescu ?", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Wh-islands", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "These violate for locality reasons : there is no way to localize the wh-element and its coindexed base position in the same tree set (MCTAGs) which can then be adjoined into a single elementary tree. But in the case of interrogative clauses are not islands for extraction: (6) Pe eine i crezi ca Paul detesta e;? Who i do you think that Paul detests e ; ? (7) Pentru care clauza vrei sa afli eine ;e; nu a decis inca ce 1: va vota ek ei? (Comorovsky 1986) For which paragraph do you want to leam whoiei has not decided yet whatk he will vote e1: e 1 ?", |
|
"cite_spans": [ |
|
{ |
|
"start": 438, |
|
"end": 455, |
|
"text": "(Comorovsky 1986)", |
|
"ref_id": "BIBREF0" |
|
} |
|
], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Wh-islands", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "For this Kroch suggests an interesting account that reduces the constraint on movement out of an island to a local well-formedness condition on elementary trees.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Wh-islands", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "In In examples 9-12(a) and 14(b) because o f the ordering constraint, the TAG formalism is not able to analyse these cases , given the predicatt'argument coocurrence constraint on t!k men-tary trees. The problem in describing this phenomena with TAG arises from the fact observed by Vijay-Shanker 1992 , that adjoining is an overly restricted way of combining structures.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Multiple Wh -Movement", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "+] DPi CP ~ 1 /'--.... DPj CP cuiC/ IP~ A cemasina ~ ) - D T 1 / \\ PR6 (\\ Paul 1 VP 1 p I~ sa V~ CP* P D j", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Multiple Wh -Movement", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "In Multi-Component TAG (MCT AG) (J oshi, 1987), trees are grouped into sets which must be adjoined together ( multicomponent adjunction). The elementary tree is split up into parts, which are grouped together into sets. All trees from one set must be adjoined at the same time, at different nodes into the single tree representing the embedded clause. However, MCTAG Jack expressive power since, while syntactic relations are invariably subject to c-command or dominance constraints , there is no way to state that two trees from a set must be in a dominance relation in the derived tree. (Figure 2 )", |
|
"cite_spans": [], |
|
"ref_spans": [ |
|
{ |
|
"start": 589, |
|
"end": 598, |
|
"text": "(Figure 2", |
|
"ref_id": null |
|
} |
|
], |
|
"eq_spans": [], |
|
"section": "Multiple Wh -Movement", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "~,.\u2022\u2022\u2022\u2022\u2022\u2022\u2022\u2022\u2022\u2022\u2022\u2022~ DPi S 1 cui DP VP ~ promite S o()\\", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Multiple Wh -Movement", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "ce masina DP VP DTG is designed to overcome this limitation . Subsertion can be viewed as a generalization of adjunction in which components of the clausal complement (the subserted structure) can be interspersed within the structure that is the site cf the subsertion . DTG provide a mechanism involving the use Qf domination links(d-edges) that ensure that parts of the subserted structure that are not substituted dominate those parts that are . Furthermore,there is a need to constrain the way in which the non -substituted components can be interspersed. The derivation proceeds as follows : we first subsert the embedded clause tree into the matrix clause tree. After we subsert the wh-pronoun of In DTG formalism,the ordering constraint on the extractions is marked by the foature \" topic \".The final tree is the desired . semantically motivated , dependency structure : the embedded clause depends on the matrix clause), with respect to the ordering constraint on the wh-pronouns.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Multiple Wh -Movement", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "DTG are designed to share some of the advantages like other fonnalisms in the TAG family, while overcoming some of their limitations . The most distinctive feature of DTG is that there is complete uniformity in the way that the subsertion relate lexical items. Furthermore , DTG can provide a unifonn analysis for wh-movement in Romanian, despite the fact that the wh-elements in Romanian can appear in sentence -initial position and in sentence -second position. Ph.D.Thesis.Univ. Paris 7.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Conclusion", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "A. Abeille, forthcoming : Extraction out of NP and clitic-noun dependencies in French . TAG+3, Paris.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "Conclusion", |
|
"sec_num": null |
|
}, |
|
{ |
|
"text": "We are grateful to Anne Abeille Sylvain Kahane and Owen Rambow.", |
|
"cite_spans": [], |
|
"ref_spans": [], |
|
"eq_spans": [], |
|
"section": "", |
|
"sec_num": null |
|
} |
|
], |
|
"back_matter": [], |
|
"bib_entries": { |
|
"BIBREF0": { |
|
"ref_id": "b0", |
|
"title": "Multiple WH Movement in Romanian, Linguistic Inquiry", |
|
"authors": [ |
|
{ |
|
"first": "I", |
|
"middle": [], |
|
"last": "Comorovsky", |
|
"suffix": "" |
|
} |
|
], |
|
"year": 1986, |
|
"venue": "", |
|
"volume": "", |
|
"issue": "", |
|
"pages": "", |
|
"other_ids": {}, |
|
"num": null, |
|
"urls": [], |
|
"raw_text": "I. Comorovsky ,1986: Multiple WH Movement in Romanian, Linguistic Inquiry.", |
|
"links": null |
|
}, |
|
"BIBREF1": { |
|
"ref_id": "b1", |
|
"title": "The syntax of Romanian", |
|
"authors": [ |
|
{ |
|
"first": "C", |
|
"middle": [], |
|
"last": "Dobrovie", |
|
"suffix": "" |
|
} |
|
], |
|
"year": 1994, |
|
"venue": "", |
|
"volume": "", |
|
"issue": "", |
|
"pages": "", |
|
"other_ids": {}, |
|
"num": null, |
|
"urls": [], |
|
"raw_text": "C. Dobrovie ,1994 :The syntax of Romanian. Mouton de Gruyter .", |
|
"links": null |
|
}, |
|
"BIBREF2": { |
|
"ref_id": "b2", |
|
"title": "Uni v .of Pennsylvania A .Joshi, 0. Rambow,1992: A formal Look at Dependency Grammars and Phrase -Structure Grammars ,with Special Consideration of Word-Order Phenomena", |
|
"authors": [ |
|
{ |
|
"first": "R", |
|
"middle": [], |
|
"last": "Frank", |
|
"suffix": "" |
|
} |
|
], |
|
"year": 1992, |
|
"venue": "Syntactic Locality and TAG : Grammatical , Acquisition and Processing Perspecti ves", |
|
"volume": "", |
|
"issue": "", |
|
"pages": "", |
|
"other_ids": {}, |
|
"num": null, |
|
"urls": [], |
|
"raw_text": "R. Frank , 1992 : Syntactic Locality and TAG : Grammatical , Acquisition and Processing Perspecti ves.Ph.D.thesis.Uni v .of Pennsylvania A .Joshi, 0. Rambow,1992: A formal Look at Dependency Grammars and Phrase -Structure Grammars ,with Special Consideration of Word- Order Phenomena. In Leo Wanner(ed.) Recent trends in Meaning-Text Theory A. Kroch 1989 :Asymmetries in Long Distance Extraction in a TAG. In Alternative conceptions of phrase strncture.", |
|
"links": null |
|
}, |
|
"BIBREF3": { |
|
"ref_id": "b3", |
|
"title": "Univ.of Chicago Press, Chicago A. Kroch and A. Joshi l 986 :Analysing extraposition in a TAG", |
|
"authors": [], |
|
"year": null, |
|
"venue": "Syntax and Semamics : Discontinous Constituents", |
|
"volume": "", |
|
"issue": "", |
|
"pages": "", |
|
"other_ids": {}, |
|
"num": null, |
|
"urls": [], |
|
"raw_text": "M.Baltin and A. Kroch(eds), Univ.of Chicago Press, Chicago A. Kroch and A. Joshi l 986 :Analysing extraposition in a TAG. In G.Huckand A. Ojeda (eds) , Syntax and Semamics : Discontinous Constituents", |
|
"links": null |
|
}, |
|
"BIBREF4": { |
|
"ref_id": "b4", |
|
"title": "D-Tree Grammars, ACL 1995 XTAG for English", |
|
"authors": [ |
|
{ |
|
"first": "K", |
|
"middle": [], |
|
"last": "Rambow", |
|
"suffix": "" |
|
}, |
|
{ |
|
"first": ".", |
|
"middle": [ |
|
"D" |
|
], |
|
"last": "Vijay-Shanker", |
|
"suffix": "" |
|
}, |
|
{ |
|
"first": "", |
|
"middle": [], |
|
"last": "Weir", |
|
"suffix": "" |
|
} |
|
], |
|
"year": 1995, |
|
"venue": "", |
|
"volume": "", |
|
"issue": "", |
|
"pages": "", |
|
"other_ids": {}, |
|
"num": null, |
|
"urls": [], |
|
"raw_text": "Rambow, K.Vijay-Shanker. D. Weir. 1995 :D-Tree Grammars, ACL 1995 XTAG for English 1995, Institute for Research in Cognitive Science, Univ.of Pennsylvania", |
|
"links": null |
|
} |
|
}, |
|
"ref_entries": { |
|
"FIGREF0": { |
|
"uris": null, |
|
"text": "Figure 1 : trees for (14a)", |
|
"num": null, |
|
"type_str": "figure" |
|
}, |
|
"FIGREF1": { |
|
"uris": null, |
|
"text": "MC-trees for (14b)", |
|
"num": null, |
|
"type_str": "figure" |
|
}, |
|
"FIGREF3": { |
|
"uris": null, |
|
"text": "Figure 3 :quasi-trees for (14b)", |
|
"num": null, |
|
"type_str": "figure" |
|
} |
|
} |
|
} |
|
} |