{ "paper_id": "1992", "header": { "generated_with": "S2ORC 1.0.0", "date_generated": "2023-01-19T07:45:13.622413Z" }, "title": "", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "", "affiliation": { "laboratory": "", "institution": "Universit\u00e9 de Paris", "location": { "addrLine": "7 2 Place Jussieu", "postCode": "75251, Cedex 05", "settlement": "Paris", "country": "France" } }, "email": "" }, { "first": "P", "middle": [], "last": "Samvelian", "suffix": "", "affiliation": { "laboratory": "", "institution": "Universit\u00e9 de Paris", "location": { "addrLine": "7 2 Place Jussieu", "postCode": "75251, Cedex 05", "settlement": "Paris", "country": "France" } }, "email": "" } ], "year": "", "venue": null, "identifiers": {}, "abstract": "To translate collocations such as prendre une d\u00e9cision, prendre l'habitude, perdre l'habitude, avoir soif, most MT systems include bilingual rules as prendre (-d\u00e9cision)-> make, prendre (-habitude)-> get into, perdre (-habitude)-> get out of, soif (avoir-)-> thirsty (be-). These numerous rules, which are lexically context-sensitive, are costly, be they used in a rationalist or empiricist approach. This paper presents a solution where they are avoided. This solution is based on linguistics notions such as \"predicative element\", \"aspectual variant\", \"diathetic variant\" which allow us to use only the unavoidable simple bilingual rules such as d\u00e9cision-> decision, habitude-\u2022> habit, thirsty-> soif. This solution, which has been successfully implemented in a rule-based transfer system, i.e. the Eurotra system which handles nine languages, requires a lot of monolingual lexical information, but we will show that this lexical information should be easily extracted from tagged corpora. * We want to thank Jean-Marc Lang\u00e9 for his help on statistical aspects, Ted Briscoe and Annie Meunier for their fruitful comments on earlier drafts. John is fat-> Jean est gros Case 3) Vsup-Npred-> Vsup-Npred Jean a pris une d\u00e9cision-> John made a decision Case 4) be-Prep X-> be-Prep X John is in good spirits-> Jean est de bonne humeur Case 5) Vord-> be-Adj 77\u00bb room adjoins the kitchen ~> La chambre est contigu\u00eb \u00e0 la cuisine Case 6) Vord-> Vsup-Npred John ambushed Mary-> Jean a tendu une embuscade \u00e0 Marie Case 7) Vord-> be-Prep X This theater billed Hamlet-> Hamlet \u00e9tait \u00e0 l'affiche de ce th\u00e9\u00e2tre Case 8) Vsup-Npred-> Vord Jean a fait un croc-en-jambe \u00e0 Marie-> John tripped up Mary Case 9) Vsup-Npred-> be-Adj Jean a de la reconnaissance envers Marie-> John is grateful to Marie Case 10) Vsup-Npred-> be-Prep X Jean a des difficult\u00e9s financi\u00e8res-> John is in financial difficulties Case 11) be-Adj-> Vord It was snowy yesterday-> II a neig\u00e9 hier Case 12) be-Adj-> Vsup-Npred John is thirsty-> Jean a soif Case 13) be-Adj-> be-Prep X John is agog about this letter-> Jean est en \u00e9moi \u00e0 cause de cette lettre Case 14) be-Prep X-> Vord Jean est dans l'attente d'une r\u00e9ponse-> John is waiting for an answer", "pdf_parse": { "paper_id": "1992", "_pdf_hash": "", "abstract": [ { "text": "To translate collocations such as prendre une d\u00e9cision, prendre l'habitude, perdre l'habitude, avoir soif, most MT systems include bilingual rules as prendre (-d\u00e9cision)-> make, prendre (-habitude)-> get into, perdre (-habitude)-> get out of, soif (avoir-)-> thirsty (be-). These numerous rules, which are lexically context-sensitive, are costly, be they used in a rationalist or empiricist approach. This paper presents a solution where they are avoided. This solution is based on linguistics notions such as \"predicative element\", \"aspectual variant\", \"diathetic variant\" which allow us to use only the unavoidable simple bilingual rules such as d\u00e9cision-> decision, habitude-\u2022> habit, thirsty-> soif. This solution, which has been successfully implemented in a rule-based transfer system, i.e. the Eurotra system which handles nine languages, requires a lot of monolingual lexical information, but we will show that this lexical information should be easily extracted from tagged corpora. * We want to thank Jean-Marc Lang\u00e9 for his help on statistical aspects, Ted Briscoe and Annie Meunier for their fruitful comments on earlier drafts. John is fat-> Jean est gros Case 3) Vsup-Npred-> Vsup-Npred Jean a pris une d\u00e9cision-> John made a decision Case 4) be-Prep X-> be-Prep X John is in good spirits-> Jean est de bonne humeur Case 5) Vord-> be-Adj 77\u00bb room adjoins the kitchen ~> La chambre est contigu\u00eb \u00e0 la cuisine Case 6) Vord-> Vsup-Npred John ambushed Mary-> Jean a tendu une embuscade \u00e0 Marie Case 7) Vord-> be-Prep X This theater billed Hamlet-> Hamlet \u00e9tait \u00e0 l'affiche de ce th\u00e9\u00e2tre Case 8) Vsup-Npred-> Vord Jean a fait un croc-en-jambe \u00e0 Marie-> John tripped up Mary Case 9) Vsup-Npred-> be-Adj Jean a de la reconnaissance envers Marie-> John is grateful to Marie Case 10) Vsup-Npred-> be-Prep X Jean a des difficult\u00e9s financi\u00e8res-> John is in financial difficulties Case 11) be-Adj-> Vord It was snowy yesterday-> II a neig\u00e9 hier Case 12) be-Adj-> Vsup-Npred John is thirsty-> Jean a soif Case 13) be-Adj-> be-Prep X John is agog about this letter-> Jean est en \u00e9moi \u00e0 cause de cette lettre Case 14) be-Prep X-> Vord Jean est dans l'attente d'une r\u00e9ponse-> John is waiting for an answer", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Abstract", "sec_num": null } ], "body_text": [ { "text": "In MT systems, whether they are based on a rationalist or empiricist approach, the translation of collocations such as avoir l'habitude, perdre l'habitude, prendre l'habitude, prendre une d\u00e9cision are handled with rules where the translation of the verb is determined by its object 1 : 1aavoir ( -habitude ) --> be in (1b) perdre ( -habitude ) --> get out of (1c) prendre ( -habitude ) --> get into (1d) prendre ( -d\u00e9cision ) -->make", "cite_spans": [ { "start": 399, "end": 403, "text": "(1d)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "These rules, which are numerous and costly, are either hand-written or extracted automatically from corpora (Brown et alii 1988 (Brown et alii , 1991 . Our purpose is to show that they are unnecessary. We will propose a solution which is centered on the noun of these collocations. This noun is the key element on linguistics grounds, and on statistical grounds, it is the most informative element to the extent that e.g. the chances of observing make given decision are more important than the chances of observing decision given make.", "cite_spans": [ { "start": 108, "end": 127, "text": "(Brown et alii 1988", "ref_id": null }, { "start": 128, "end": 149, "text": "(Brown et alii , 1991", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "Our solution, which relies on thorough linguistic studies exposed briefly below, consists in associating to nouns such as habitude or habit a set of \"support verbs\" where each verb is given a \"semantic value\" (written below within brackets):", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "habitude -> {avoir (neuter), perdre (terminative), prendre (inchoative)} habit ->{be in (neuter), get out of (terminative), get into (inchoative)}", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "The translation of a collocation like prendre l'habitude is carried out through the following steps: 1)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "figure out that prendre is one of the support verbs associated with habitude and record its semantic value, i.e. inchoative, 2) use the simple transfer rule habitude --> habit, 3) search in the set of support verbs associated with habit the element with the semantic value recorded in step 1 ), which brings back get into. The solution proposed for translating collocations like prendre l'habitude can be extended to handle translation cases which entail a \"category switching\" like:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "(2) be snowy --> neiger", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "(3) be thirsty <--> avoir soif", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "The implementation of this solution in a rule-based transfer system, i.e. Eurotra, will be presented.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "Finally, we will discuss the feasibility of extracting automatically from corpora monolingual lexical information such as the set of support verbs associated with a noun like habitude or habit. Let us first expose the linguistics basis of our solution.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "1 Linguistic data", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Introduction", "sec_num": "0." }, { "text": "We assume that a sentence is the projection of the argument structure of a lexical head into a syntactic structure. This head will be called the predicative element of the sentence. We note Na the first argument of the predicative element, Nb and Nc a possible second and third argument.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "Let us consider four types of predicative elements as illustrated in the following examples:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "(4) John is reading a book (5) John is proud of Mary (6) John has ascendancy over Mary", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "(7)", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "John is at the mercy of Mary", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "The predicative element of these sentences is respectively:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "-the ordinary verb read in (4) which has the following structure: Na Vord Nb (Vord stands for ordinary verb); -the adjective proud (supported by be ) in (5) which has the following structure: Na be Adj Prepb Nb;", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "-the predicative noun ascendancy (supported by have ) in (6) which has the following structure: Na Vsup Det Npred Prepb Nb where the verb is called a \"support verb\" (noted Vsup). It introduces a \"predicative noun\" (noted Npred ) 2 .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "-the sequence at the mercy (supported by be ) in (7) which has the following structure: Na be Prep X Prepb Nb where the verb be is followed by a preposition (noted Prep ) and a noun phrase (noted X) 3 .", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "In (4) the lexical head of the sentence (i.e. a Vord) corresponds to its syntactic head. In the three other examples, the head (which is not a Vord) is supported by a verb with which it forms a complex syntactic head. We are going to show that using this notion of semantic head allows us to formalize the contrastive data in a way which is fruitful for automatic translation. These four types of predicative elements are noted: Pred i with i \u03a3 {Vord, Vsup-Npred, be-Adj, be-PrepX} A predicative element can be viewed as the lexicalisation of a semantico-logical concept (a notion that we will not attempt to formalize). For example, English commit-suicide and French se suicider can both be viewed as lexicalisation of the language-free concept SUICIDE. As lexicalisation is a purely idiosyncratic monolingual phenomenon, one can expect that a predicative element of type i is translated into another language as a predicative element of type j with a relation between i and j which cannot be predicted. Given the four types of predicative elements we consider here, one can expect sixteen (i.e. 16 = 4 2 ) cases of translation, and in fact, these sixteen cases are encountered when translating from French to English or from English to French, as illustrated bellow:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "Case 1) Vord->Vord", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "John is reading a book --> Jean lit un livre", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "Case 2) be-Adj --> be-Adj", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Predicative element of a sentence", "sec_num": "1.1" }, { "text": "Case 16) be-Prep X --> Vsup-Npred", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Jean est sur la paille --> John is broke", "sec_num": null }, { "text": "The first four cases entail no category switching while the twelve others do. The case (3) Vsup-Npred --> Vsup-Npred entails no category switching but a semi-compositional translation (Danlos 1992 ). The support verb in the target language is determined by the target Npred and not by the source Vsup.This is due to the fact that monolingually the choice of a Vsup depends on the Npred.", "cite_spans": [ { "start": 184, "end": 196, "text": "(Danlos 1992", "ref_id": "BIBREF12" } ], "ref_spans": [], "eq_spans": [], "section": "Jean est d'un caract\u00e8re ferm\u00e9 -> John has a withdrawn nature", "sec_num": null }, { "text": "The following translations of the support verb faire illustrate this phenomenon: However a verb such as begin can be used with a predicative noun without its support verb:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Jean est d'un caract\u00e8re ferm\u00e9 -> John has a withdrawn nature", "sec_num": null }, { "text": "Jean", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Jean est d'un caract\u00e8re ferm\u00e9 -> John has a withdrawn nature", "sec_num": null }, { "text": "(12a) Jean a commenc\u00e9 un croc-en-jambe \u00e0 Marie", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Jean est d'un caract\u00e8re ferm\u00e9 -> John has a withdrawn nature", "sec_num": null }, { "text": "The translation of (12a) is the same as that of (12), so it could be obtained if the equivalence between (12) and (12a) were established by some 'coercion' operation (Pustejovsky 1989 (Pustejovsky , 1991 Boguraev et alii 1990) . Coercion phenomenon is out of the scope of this paper, so it will not be discussed here. The existence of these two cases of translation of Pred i -aspx comes from the fact that the lexicalization of a non-neuter aspect for a given predicative element is a monolingual idiosyncratic phenomenon, as any lexicalization.", "cite_spans": [ { "start": 166, "end": 183, "text": "(Pustejovsky 1989", "ref_id": "BIBREF19" }, { "start": 184, "end": 203, "text": "(Pustejovsky , 1991", "ref_id": "BIBREF20" }, { "start": 204, "end": 226, "text": "Boguraev et alii 1990)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Jean est d'un caract\u00e8re ferm\u00e9 -> John has a withdrawn nature", "sec_num": null }, { "text": "Aspect is one kind of semantic information a sentence can convey. Other types of semantic information are diathesis and modality. As for aspect, they can be expressed either by special diathetic/modal operators or by lexicalized diathetic/modal predicative elements. As an illustration, one diathetic value is causativisation, which modifies the argument structure by adding a new argument with a causative denotation. As for aspect, causativisation may be expressed by two means, either by lexicalisation or by addition of an auxiliary factitive verb, noted Op-caus. The translation of a lexicalized causative operator is similar to that of a lexicalized aspectual variant.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Semantic Values", "sec_num": "1.3." }, { "text": "More precisely, given a predicative element Pred i , which translates into the predicative element Pred j , and which has a causative operator Pred i -caus, the translation of Pred i -caus may induce two cases, represented by the following diagrams: However, a verb such as seem can be used with an adjective or Prep X without its support verb be:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Semantic Values", "sec_num": "1.3." }, { "text": "1) Pred i --> Pred j Pred i -caus --> Pred j -", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Semantic Values", "sec_num": "1.3." }, { "text": "(17a) John seems thirsty", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Semantic Values", "sec_num": "1.3." }, { "text": "The translation of (17a) is the same as that of (17) and it could be obtained if the equivalence between (17) and (17a) were established, in a similar way as the equivalence between (12) and 12a, where a Npred is used without its support verb.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Other Semantic Values", "sec_num": "1.3." }, { "text": "To conclude this section on linguistic data, let us recapitulate the monolingual lexical information which is associated with a predicative element. First, the argument structure of any predicative element, whatever its type, is lexical information which indicates the number of arguments, their semantic features and their syntactic realization. The linguistics data we have examined put forward the additional monolingual lexical information 4 :", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "-for a predicative noun Npred, its neuter support verb and the aspectual, diathetic and modal variants of this support verb (if any). As an illustration, the lexical information associated with the noun ascendancy is the following: it is a predicative noun which has two arguments, none of which is obligatory human, the second one is introduced by the preposition over, the neuter support verb is have, the inchoative variant gain, the terminative variant loose, the durative variant keep and the causativisation inchoative variant give. In the rest of this paper, the neuter support verb of an Npred and its variants will all be called \"support verb\";", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "-for a predicative adjective or a Prep X sequence, the lexical information is similar to that of a Npred, except that the neuter support verb is the same for all predicative elements of these two types, e.g.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "be in English. Moreover, for Prep X, additional information is needed for the preposition and the determiner, in some cases (Samvelian et alii 1992) ;", "cite_spans": [ { "start": 124, "end": 148, "text": "(Samvelian et alii 1992)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "-for an ordinary verb Vord, the linguistic data we have examined involve the issue of derivational morphology, that we are not going to tackle here.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "This lexical information structured the way it is allows us to use it straightforwardly in an MT system, as we are going to illustrate it with the Eurotra system, where the dictionary entry for e.g.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "ascendancy is as described above.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Monolingual lexical information for predicative elements", "sec_num": "1.4" }, { "text": "Let us examine how to compute the translation of predicative elements in a transfer-based translation system and how it is actually handled in Eurotra. Only the design of the intermediary representations and the principles of the analysis, transfer and generation processes will be exposed. There is no room here to present in detail the actual implementation which is described in (Danlos et alii 1988 , Danlos 1992 , Daille and Danlos 1992 , Samvelian 1990 , Samvelian et alii 1992 .", "cite_spans": [ { "start": 382, "end": 402, "text": "(Danlos et alii 1988", "ref_id": null }, { "start": 403, "end": 416, "text": ", Danlos 1992", "ref_id": "BIBREF12" }, { "start": 417, "end": 441, "text": ", Daille and Danlos 1992", "ref_id": null }, { "start": 442, "end": 458, "text": ", Samvelian 1990", "ref_id": "BIBREF21" }, { "start": 459, "end": 483, "text": ", Samvelian et alii 1992", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Automatic translation in a rule-based transfer system", "sec_num": "2" }, { "text": "Eurotra uses a semantic-oriented representation which is a \"lowered governor deep dependency grammar\" (Allegranza et alii 1991 , Durand et alii 1991 . To handle the phenomena we are dealing with, the basic idea is that the governor of a sentence is its predicative element without its support (if any). The semantic information that comes from the support verb is featurized.", "cite_spans": [ { "start": 102, "end": 126, "text": "(Allegranza et alii 1991", "ref_id": null }, { "start": 127, "end": 148, "text": ", Durand et alii 1991", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Automatic translation in a rule-based transfer system", "sec_num": "2" }, { "text": "Therefore, the support verb is not represented nor translated as lexical item. The realization of this basic idea does not go without problems (e.g. how to represent the fact that Npred in a support verb construction is part of a standard NP with determiners and possible modifiers?). These problems are discussed at length in (Danlos 1992 In this representation, the predicative noun is the governor of the sentence which is of type \"svc\" = support verb construction). In analysis, (I) is obtained thanks to the equality between the verb and one of the support verbs which are recorded in the dictionary entry of the predicative noun. The value of the feature \"aspect\", i.e. inchoative, is obtained because gain is recorded as inchoative variant in the entry of ascendancy. Vsup(N) should be made up of the support verbs of N. Each element of this list must be given a semantic value such as neuter, inchoative, etc. For some elements of L-vsup, this value is the same for any Npred, e.g. perdre has always a terminative value. For other elements of L-vsup, this value depends on Npred, e.g. prendre has a neuter value with d\u00e9cision and an inchoative value with habitude . However, it is possible to call upon rules such as the following (deduced from (Viv\u00e8s 1983 ) data): if prendre and perdre both belong to Vsup(N), then prendre has an inchoative value.", "cite_spans": [ { "start": 327, "end": 339, "text": "(Danlos 1992", "ref_id": "BIBREF12" }, { "start": 1253, "end": 1264, "text": "(Viv\u00e8s 1983", "ref_id": "BIBREF24" } ], "ref_spans": [], "eq_spans": [], "section": "Automatic translation in a rule-based transfer system", "sec_num": "2" }, { "text": "To sum up, it should be possible to extract from tagged corpora the monolingual lexical information we need for our solution. This extraction is under study within the ET-10 project 6 (Gaussier et alii 1992).", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Automatic translation in a rule-based transfer system", "sec_num": "2" }, { "text": "There is little hope that a collocation like make-decision be translated correctly with a pure statistical approach when the two elements of the collocation occur in different clauses:", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "Conclusion", "sec_num": null }, { "text": "(18) The decision that the French government is going to make about tobacco is disliked by the majority of the population This limit of the statistical approach can be considered as well known since the statistical models only capture local phenomena (Brown et alii 1991) . Nevertheless, the translation of (18) is likely to be incorrect also in a rule-based system using contextual bilingual rules such as make ( -decision ) --> prendre. In conclusion, linguistics knowledge as that presented here is needed in both rationalist and empiricist approaches.", "cite_spans": [ { "start": 251, "end": 271, "text": "(Brown et alii 1991)", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "Conclusion", "sec_num": null }, { "text": "We assume that the reader knows French and English sufficiently to understand the examples.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "cf Cartel 1984;Danlos 1980Danlos ,1981Giry-Schneider 1978, 1987Gross 1981;G. Gross 1989;Meunier 1981; Viv\u00e8s 1984. 3 cf.Danlos 1980Danlos ,1981Samvelian 1990, Samvelian et alii 1992", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "Such lexical information is roughly recorded in (Melcuk 1984) dictionary where it is indexed under rubrics called \"lexical functions\".", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "The representation of an epithet adjective such as thirsty in a thirsty man is equivalent to the representation of a man who is thirsty. So we obtain the right translation a thirsty man --> un homme qui a soif.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null }, { "text": "ET-10 is a project which is co-funded by the European Economic Community and which is conducted by the Centre Scientifique d'IBM-France with TALANA, University of Essex and University of Lancaster as main participants.", "cite_spans": [], "ref_spans": [], "eq_spans": [], "section": "", "sec_num": null } ], "back_matter": [ { "text": "Consider now the translation of John remains thirsty into French. Below ", "cite_spans": [ { "start": 66, "end": 71, "text": "Below", "ref_id": null } ], "ref_spans": [], "eq_spans": [], "section": "annex", "sec_num": null }, { "text": "The solution we propose is quite economic at the transfer level. However, this solution requires a lot of monolingual lexical information. Let us examine if it would be feasible to extract this lexical information from tagged corpora, taking as illustration the lexical information associated with an Npred.First, one can lay down on statistical grounds that the support verbs of any Npred in a given language L belong to a closed class, noted L-vsup. For exemple, the most common support verbs in French are:Fr-vsup = {avoir, faire, prendre, perdre, donner, recevoir, subir, conserver, garder, entamer, poursuivre}The support verbs which do not belong to L-vsup are stylistic (exotic) variants such as:Jean (a + nourrit + caresse) l'espoir de trouver un boulot John (has + nourishes + entertains + toys with) the hope of getting a job Second, one can assume that there exists an algorithm which produces for a given noun N the list V(N) of the x most frequent verbs V for which N is a (direct) object of V (such an algorithm can be inspired by the one used in (Smadja and McKeown 1990) . For a non-predicative noun such as beer, the list V(N) should contain ordinary verbs which occur frequently with N as a (direct) object,", "cite_spans": [ { "start": 1061, "end": 1086, "text": "(Smadja and McKeown 1990)", "ref_id": "BIBREF23" } ], "ref_spans": [], "eq_spans": [], "section": "Automatic acquisition of monolingual lexical information", "sec_num": "3" } ], "bib_entries": { "BIBREF0": { "ref_id": "b0", "title": "Linguistics for Machine Translation: The Eurotra Linguistic specifications", "authors": [ { "first": "V", "middle": [], "last": "Allegranza", "suffix": "" }, { "first": "P", "middle": [], "last": "Bennett", "suffix": "" }, { "first": "J", "middle": [], "last": "Durand", "suffix": "" }, { "first": "F", "middle": [], "last": "Van Eynde", "suffix": "" }, { "first": "L", "middle": [], "last": "Humphreys", "suffix": "" }, { "first": "P", "middle": [], "last": "Schmidt", "suffix": "" }, { "first": "E", "middle": [], "last": "Steiner", "suffix": "" } ], "year": 1991, "venue": "Studies in Machine Translation and natural language processing", "volume": "1", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Allegranza, V., Bennett, P., Durand, J., Van Eynde, F., Humphreys L., Schmidt, P., Steiner, E., 1991, \"Linguistics for Machine Translation: The Eurotra Linguistic specifications\", in Studies in Machine Translation and natural language processing, Vol 1, Luxembourg: Commission of the European communities.", "links": null }, "BIBREF1": { "ref_id": "b1", "title": "Enjoy the paper: Lexical Semantics via Lexicology", "authors": [ { "first": "B", "middle": [], "last": "Boguraev", "suffix": "" }, { "first": "T", "middle": [], "last": "Briscoe", "suffix": "" }, { "first": "A", "middle": [], "last": "Copestake", "suffix": "" } ], "year": null, "venue": "Proceedings of COLING '90", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Boguraev, B., Briscoe, T., Copestake, A., \"Enjoy the paper: Lexical Semantics via Lexicology\", in Proceedings of COLING '90, Helsinki, Finland.", "links": null }, "BIBREF2": { "ref_id": "b2", "title": "A statistical approach to language translation", "authors": [ { "first": "P", "middle": [], "last": "Brown", "suffix": "" }, { "first": "J", "middle": [], "last": "Cocke", "suffix": "" }, { "first": "S", "middle": [], "last": "Della Pietra", "suffix": "" }, { "first": "V", "middle": [], "last": "Della Pietra", "suffix": "" }, { "first": "F", "middle": [], "last": "Jelinek", "suffix": "" }, { "first": "R", "middle": [], "last": "Mercer", "suffix": "" }, { "first": "P", "middle": [], "last": "Roossin", "suffix": "" } ], "year": 1988, "venue": "Proceedings of COLING '88", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Brown, P., Cocke, J., Della Pietra, S., Della Pietra, V., Jelinek, F., Mercer, R., Roossin, P., 1988, \"A statistical approach to language translation\", in Proceedings of COLING '88, Budapest Hungary.", "links": null }, "BIBREF3": { "ref_id": "b3", "title": "Word-sense disambiguation using statistical methods", "authors": [ { "first": "P", "middle": [], "last": "Brown", "suffix": "" }, { "first": "S", "middle": [], "last": "Della Pietra", "suffix": "" }, { "first": "F", "middle": [], "last": "Della Pietra", "suffix": "" }, { "first": "", "middle": [ "R" ], "last": "Mercer", "suffix": "" } ], "year": 1991, "venue": "Proceedings of ACL '91", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Brown, P., Della Pietra, S., Della Pietra, F., Mercer.R., 1991, \"Word-sense disambiguation using statistical methods\", in Proceedings of ACL '91, Berkeley, USA.", "links": null }, "BIBREF4": { "ref_id": "b4", "title": "Composite Predicates in English", "authors": [ { "first": "R", "middle": [], "last": "Cattel", "suffix": "" } ], "year": 1984, "venue": "Syntax and Semantics", "volume": "17", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Cattel, R., 1984, Composite Predicates in English, Syntax and Semantics, vol. 17, Academic Press, New York.", "links": null }, "BIBREF5": { "ref_id": "b5", "title": "Les constructions \u00e0 verbe support : description linguistique et implementation dans Eurotra", "authors": [ { "first": "B", "middle": [], "last": "Daille", "suffix": "" }, { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": null, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Daille, B., Danlos, L., \"Les constructions \u00e0 verbe support : description linguistique et implementation dans Eurotra\", Rapport technique TALANA, n\u00b0 4, Paris.", "links": null }, "BIBREF6": { "ref_id": "b6", "title": "Repr\u00e9sentation d'informations linguistiques : les constructions N \u00eatre Pr\u00e9p X, Th\u00e8se de 3\u00e8me cycle", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": 1980, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., 1980, Repr\u00e9sentation d'informations linguistiques : les constructions N \u00eatre Pr\u00e9p X, Th\u00e8se de 3\u00e8me cycle, Universit\u00e9 de Paris 7.", "links": null }, "BIBREF7": { "ref_id": "b7", "title": "La morphosyntaxe des expressions fig\u00e9es", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": 1981, "venue": "Langages", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., 1981, \"La morphosyntaxe des expressions fig\u00e9es\", Langages, Larousse, Paris.", "links": null }, "BIBREF8": { "ref_id": "b8", "title": "Une illustration d'\u00e9tude formelle des noms : charg(-e, -er -ement)", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": 1986, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., 1986, \"Une illustration d'\u00e9tude formelle des noms : charg(-e, -er -ement)\", Langue Fran\u00e7aise, 69, Larousse, Paris.", "links": null }, "BIBREF9": { "ref_id": "b9", "title": "Les expressions fig\u00e9es construites avec le verbe support \u00eatre Pr\u00e9p", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": 1988, "venue": "Langages", "volume": "90", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., 1988, \"Les expressions fig\u00e9es construites avec le verbe support \u00eatre Pr\u00e9p \", Langages, 90, Larousse, Paris.", "links": null }, "BIBREF10": { "ref_id": "b10", "title": "Les probl\u00e8mes pos\u00e9s par les verbes support en traduction automatique", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": 1988, "venue": "Actes du Colloque Informatique et Langue Naturelle", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., 1988, \"Les probl\u00e8mes pos\u00e9s par les verbes support en traduction automatique\", in Actes du Colloque Informatique et Langue Naturelle, Universit\u00e9 de Nantes, Nantes.", "links": null }, "BIBREF11": { "ref_id": "b11", "title": "Support verbs and Predicative Nouns in a machine translation system", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" }, { "first": "A", "middle": [], "last": "Bech", "suffix": "" }, { "first": "F", "middle": [], "last": "Caroli", "suffix": "" }, { "first": "B", "middle": [], "last": "Daille", "suffix": "" }, { "first": "N", "middle": [], "last": "Mesli", "suffix": "" }, { "first": "F", "middle": [], "last": "Namer", "suffix": "" }, { "first": "S", "middle": [], "last": "Nohr", "suffix": "" } ], "year": 1989, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., with A. Bech, F. Caroli, B. Daille, N. Mesli, F. Namer, S. Nohr, 1989, \"Support verbs and Predicative Nouns in a machine translation system\", in Eurotra Reference Manual 6.0, Technical report of Eurotra, Luxembourg.", "links": null }, "BIBREF12": { "ref_id": "b12", "title": "Support Verb Constructions: linguistic properties, representation, translation", "authors": [ { "first": "L", "middle": [], "last": "Danlos", "suffix": "" } ], "year": 1992, "venue": "Journal of French Language Studies", "volume": "2", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Danlos, L., 1992, \"Support Verb Constructions: linguistic properties, representation, translation\", Journal of French Language Studies, Vol 2, n\u00b01, Cambridge University Press, Cambridge.", "links": null }, "BIBREF13": { "ref_id": "b13", "title": "The Eurotra Linguistic specifications: An overview", "authors": [ { "first": "J", "middle": [], "last": "Durand", "suffix": "" }, { "first": "V", "middle": [], "last": "Allegranza", "suffix": "" }, { "first": "P", "middle": [], "last": "Bennett", "suffix": "" }, { "first": "F", "middle": [], "last": "Van Eynde", "suffix": "" }, { "first": "L", "middle": [], "last": "Humphreys", "suffix": "" }, { "first": "P", "middle": [], "last": "Schmidt", "suffix": "" }, { "first": "E", "middle": [], "last": "Steiner", "suffix": "" } ], "year": 1991, "venue": "Machine Translation", "volume": "6", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Durand, J., Allegranza, V., Bennett, P., Van Eynde, F., Humphreys L., Schmidt, P., Steiner, E., (1991). The Eurotra Linguistic specifications: An overview. Machine Translation, vol 6, Netherlands: Kluwer Academic Publishers.", "links": null }, "BIBREF14": { "ref_id": "b14", "title": "Towards Bilingual terminology", "authors": [ { "first": "E", "middle": [], "last": "Gaussier", "suffix": "" }, { "first": "J.-M", "middle": [], "last": "Lang\u00e9", "suffix": "" }, { "first": "F", "middle": [], "last": "Meunier", "suffix": "" } ], "year": 1992, "venue": "Proceedings of ALLC/ACH Conference", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Gaussier, E., Lang\u00e9, J.-M., Meunier, F., 1992, Towards Bilingual terminology\", in Proceedings of ALLC/ACH Conference, Oxford, England.", "links": null }, "BIBREF15": { "ref_id": "b15", "title": "Les nominalisations en fran\u00e7ais : l'op\u00e9rateur \"faire", "authors": [ { "first": "J", "middle": [], "last": "Giry-Schneider", "suffix": "" }, { "first": "", "middle": [], "last": "Dans Le Lexique", "suffix": "" }, { "first": "Gen\u00e8ve", "middle": [], "last": "Droz", "suffix": "" }, { "first": "J", "middle": [], "last": "Giry-Schneider", "suffix": "" } ], "year": 1978, "venue": "Les pr\u00e9dicats nominaux en fran\u00e7ais : les phrases simples \u00e0 verbe support", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Giry-Schneider, J., 1978, Les nominalisations en fran\u00e7ais : l'op\u00e9rateur \"faire\" dans le lexique, Droz, Gen\u00e8ve. Giry-Schneider, J., 1987, Les pr\u00e9dicats nominaux en fran\u00e7ais : les phrases simples \u00e0 verbe support, Droz, Gen\u00e8ve.", "links": null }, "BIBREF16": { "ref_id": "b16", "title": "Les bases empiriques de la notion de pr\u00e9dicat s\u00e9mantique", "authors": [ { "first": "M", "middle": [], "last": "Gross", "suffix": "" } ], "year": 1981, "venue": "Langages", "volume": "63", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Gross, M., 1981, \"Les bases empiriques de la notion de pr\u00e9dicat s\u00e9mantique\", Langages 63, Larousse, Paris.", "links": null }, "BIBREF17": { "ref_id": "b17", "title": "Etudes syntaxiques de constructions converses", "authors": [ { "first": "G", "middle": [], "last": "Gross", "suffix": "" }, { "first": "Gen\u00e8ve", "middle": [], "last": "Droz", "suffix": "" }, { "first": "I", "middle": [], "last": "Melcuk", "suffix": "" } ], "year": 1984, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Gross, G., 1989, Etudes syntaxiques de constructions converses, Droz, Gen\u00e8ve. Melcuk, I., 1984, Le Dictionnaire explicatif et combinatoire du fran\u00e7ais contemporain, Presses de l'Universit\u00e9 de Montr\u00e9al, Montr\u00e9al.", "links": null }, "BIBREF18": { "ref_id": "b18", "title": "Nominalisation d'adjectifs par verbes supports, Th\u00e8se de 3\u00e8me cycle", "authors": [ { "first": "A", "middle": [], "last": "Meunier", "suffix": "" } ], "year": 1981, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Meunier, A., 1981, Nominalisation d'adjectifs par verbes supports, Th\u00e8se de 3\u00e8me cycle, Universit\u00e9 de Paris 7, Paris.", "links": null }, "BIBREF19": { "ref_id": "b19", "title": "Issues in computational lexical semantics", "authors": [ { "first": "J", "middle": [], "last": "Pustejovsky", "suffix": "" } ], "year": 1989, "venue": "Proceedings of ACL '89", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Pustejovsky, J., 1989, \"Issues in computational lexical semantics\", in Proceedings of ACL '89, Manchester, England.", "links": null }, "BIBREF20": { "ref_id": "b20", "title": "The generative lexicon", "authors": [ { "first": "J", "middle": [], "last": "Pustejovsky", "suffix": "" } ], "year": 1991, "venue": "Computational Linguistics", "volume": "17", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Pustejovsky, J., 1991, \"The generative lexicon\", in Computational Linguistics, Vol 17, n\u00b04, MIT Press.", "links": null }, "BIBREF21": { "ref_id": "b21", "title": "Les constructions \u00eatre Pr\u00e9p X : formalisation et traitement dans la cadre d'Eurotra , DEA", "authors": [ { "first": "P", "middle": [], "last": "Samvelian", "suffix": "" } ], "year": 1990, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Samvelian, P., 1990, Les constructions \u00eatre Pr\u00e9p X : formalisation et traitement dans la cadre d'Eurotra , DEA, Universit\u00e9 de Paris 7, Paris.", "links": null }, "BIBREF22": { "ref_id": "b22", "title": "Be Pr\u00e9p X constructions in a machine translation system", "authors": [ { "first": "P", "middle": [], "last": "Samvelian", "suffix": "" }, { "first": "M", "middle": [], "last": "Gavrilidou", "suffix": "" }, { "first": "N", "middle": [], "last": "Ruimy", "suffix": "" }, { "first": "U", "middle": [], "last": "Svendsen", "suffix": "" } ], "year": 1992, "venue": "Technical report of Eurotra", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Samvelian, P., Gavrilidou, M., Ruimy, N., Svendsen, U., 1992, \"Be Pr\u00e9p X constructions in a machine translation system\", Technical report of Eurotra, Luxembourg.", "links": null }, "BIBREF23": { "ref_id": "b23", "title": "Automatically extracting and representing collocations for language generation", "authors": [ { "first": "F", "middle": [], "last": "Smadja", "suffix": "" }, { "first": "K", "middle": [], "last": "Mckeown", "suffix": "" } ], "year": 1990, "venue": "Proceedings of ACL'90", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Smadja, F., McKeown, K., 1990, \"Automatically extracting and representing collocations for language generation\", in Proceedings of ACL'90, Pittsburgh, USA.", "links": null }, "BIBREF24": { "ref_id": "b24", "title": "Avoir, prendre, faire. Constructions \u00e0 verbes supports et extensions aspectuelles., Th\u00e8se de 3\u00e8me cycle", "authors": [ { "first": "R", "middle": [], "last": "Viv\u00e8s", "suffix": "" } ], "year": 1983, "venue": "", "volume": "", "issue": "", "pages": "", "other_ids": {}, "num": null, "urls": [], "raw_text": "Viv\u00e8s, R., 1983, Avoir, prendre, faire. Constructions \u00e0 verbes supports et extensions aspectuelles., Th\u00e8se de 3\u00e8me cycle, Universit\u00e9 de Paris 7, Paris.", "links": null } }, "ref_entries": { "FIGREF0": { "uris": null, "type_str": "figure", "num": null, "text": "The translation of the predicative elements just require simple lexical transfer rules such as: {lu = decision, cat =npred} --> {lu = d\u00e9cision, cat = npred} {lu = ascendancy, cat =npred} --> {lu = ascendant, cat = npred} {lu = thirsty, cat =adj} --> {lu = soif, cat = npred} 5 Let us examine the translation of the sentence John gains ascendancy over Mary into French. The transfer of (I) leads to a tree which is identic except for the leaves, i.e. the lexical values. The French generation module accesses the entry of ascendant. This entry indicates that the predicative noun has an inchoative support verb which is prendre and that the second argument must be introduced by the preposition sur. So the French generation module builds up the sentence Jean prend de l'ascendant sur Marie. Let us underline again that no lexical transfer rule such as gain (ascendancy) --> prendre is used. e.g. drink, consume, enjoy for beer. These verbs are eliminated when considering the intersection of V(N) with L-Vsup. This intersection is noted Vsup(N). For a non-predicative noun, Vsup(N) either should be empty (in which case N is known as non-predicative), or it should include verb(s) withwhich N form a collocation such as have a beer for N =: beer, prendre une bi\u00e8re for N=: bi\u00e8re. In the latter case, the wrong assumption that N is a predicative noun is made, but a bilingual rule suchas have a beer --> prendre une bi\u00e8re can be automatically computed. For a predicative noun N," }, "TABREF0": { "html": null, "type_str": "table", "text": "The focus on the duration is called \"durative\" aspect, as in :The \"neuter\" aspect is used when the process or state is designated as a whole without focusing on a specific phase of it. So far we have been dealing only with predicative elements with a neuter aspect. Let us consider now the four non-neuter aspects presented above. A non-neuter aspect is expressed by two means: either by lexicalization or by lexical items addition. Lexicalization is Vord this is mainly, but not exclusively, rendered by means of different prefixes. In French, for instance, the following prefixes have an aspectual role :These prefixes raise the issue of derivational morphology which will not be discussed here, however. A predicative element which is not a Vord (i.e. Vsup-Npred, be-Adj, be-Prep X) denotes a non-neuter aspect when used with a lexical aspectual variant of Vsup or be. For example, the predicative elements gain-ascendancy in (8a), come-out of coma in (9a), and remain-thirsty in (10a) are respectively non-neuter aspectual variants of have-ascendancy, be-in coma and bethirsty. Note that a lexical aspectual variant of Vsup or be does not change the argument structure.The fact that a predicative element of type Pred i denotes a non neuter aspect x is noted Pred iaspx. The lexical items which are added to express a non-neuter aspect are mainly aspectual verbs such as begin in (8b) or continue in (10b), and adverbials such as not anymore in (9b) and again in (11b). Aspectual lexical items are noted Op-aspx and their addition to Pred i is noted Op-aspx + Pred i .The translation of Op-aspx + Pred i seems to be compositional (i.e. it is the translation of Op-aspx added to the translation of Pred i ) :(12) Jean a commenc\u00e9 \u00e0 faire un croc-en-jambe \u00e0 Marie --> John began to trip up Mary", "num": null, "content": "
(10a)John remains thirsty
(10b)John continues to be thirsty
Finally the \"iterative\" aspect expresses the repetition of the same process or state, as in :
(11a)John est retomb\u00e9 amoureux
(11b)John est tomb\u00e9 amoureux de nouveau(John fell in love again)
illustrated in examples (8a), (9a), (10a) and (11 a), while (8b), (9b), (10b) and (11b) are examples of
lexical items addition. By lexicalization we mean that a predicative element of a given type
expresses in itself a non-neuter aspect. With a -en : inchoative s'endormir = commencer \u00e0 dormir (start to sleep)
-d\u00e9 : terminativese d\u00e9sint\u00e9resser = cesser de s'int\u00e9resser (stop to be interested)
-re : iterativeretomber = tomber \u00e0 nouveau (fall again)
inchoative\"
aspect, as in :
(8a)John gains ascendancy over Mary
(8b)John begins to have ascendancy over Mary
The focus on the end is called \"terminative\" aspect, as in :
(9a)John came out of a coma
(9b)John is not anymore in a coma
" }, "TABREF1": { "html": null, "type_str": "table", "text": "Let us, on the other hand, examine the translation of a predicative element with a non-neuter aspect Pred i -aspx, given that Pred i translates as Pred j . The translation of the latter may induce two cases : either there exists a lexicalized non-neuter aspectual variant Prep j -aspx, or there is none. The first case is represented in the following diagram:", "num": null, "content": "
neuterPred i-->Pred j
non-neuterPred i -aspx-->Pred j -aspx
neuter\u00eatre riche--->be rich
inchoativedevenir riche--->grow rich
neuterbe acquainted (with )-->avoir connaissance (de)
inchoativebecome acquainted (with)-->prendre connaissance (de)
The second case is represented in the following diagram:
neuterPred i-->Pred j
non-neuterPred i -aspx-->Op-aspx + Pred j
neuterbe thirsty-->avoir soif
durativeremain thirsty-->continuer \u00e0 avoir soif
neuters'approvisionner-->stock up
iterativese r\u00e9approvisionner-->stock up again
" } } } }