Development of a classifiers/quantifiers dictionary towards French-Japanese MT

18586

Authors

nan

Abstract

Although classifiers/quantifiers (CQs) expressions appear frequently in everyday communi- cations or written documents, they are described neither in classical bilingual paper dictio- naries, nor in machine-readable dictionaries. The paper describes a CQs dictionary, edited from the corpus we have annotated, and its usage in the framework of French-Japanese ma- chine translation (MT). CQs treatment in MT often causes problems of lexical ambiguity, polylexical phrase recog- nition difficulties in analysis and doubtful output in transfer-generation, in particular for distant languages pairs like French and Japanese. Our basic treatment of CQs is to annotate the corpus by UNL-UWs (Universal Network- ing Language - Universal words)1, and then to produce a bilingual or multilingual dictionary of CQs, based on synonymy through identity of UWs. Keywords: classifiers, quantifiers, corpus annotation, UNL, UWs dictionary, phraseology study, Tori Bank, French-Japanese MT

Keywords

nan

Full Content

Development of a classifiers/quantifiers dictionary towards French-Japanese MT Mutsuko Tomokiyo Mutsuko.Tomokiyo@imag.fr Mathieu Mangeot Mathieu.Mangeot@imag.fr Christian Boitet Christian.Boitet@imag.fr UGA, GETALP-LIG, bâtiment IMAG, 700 avenue Centrale, Domaine Universitaire de Saint-Martin-d’Hères, CS 40700 38058 Grenoble CEDEX 9, France Abstract Although classifiers/quantifiers (CQs) expressions appear frequently in everyday communi- cations or written documents, they are described neither in classical bilingual paper dictio- naries, nor in machine-readable dictionaries. The paper describes a CQs dictionary, edited from the corpus we have annotated, and its usage in the framework of French-Japanese ma- chine translation (MT). CQs treatment in MT often causes problems of lexical ambiguity, polylexical phrase recog- nition difficulties in analysis and doubtful output in transfer-generation, in particular for distant languages pairs like French and Japanese. Our basic treatment of CQs is to annotate the corpus by UNL-UWs (Universal Network- ing Language - Universal words)1, and then to produce a bilingual or multilingual dictionary of CQs, based on synonymy through identity of UWs. Keywords: classifiers, quantifiers, corpus annotation, UNL, UWs dictionary, phraseology study, Tori Bank, French-Japanese MT Introduction We call CQs (classifiers/quantifiers) words or phrases which are used in some languages to indicate the class of a noun or a nominal phrase, depending upon the type of its referent or upon speaker’s observation of the referent, when they appear in quantitative expressions. They denote: (a) CQs expressing quantity of the referent by counting. Eg. pièce (piece) (in French), 枚(mai, sheet), 点 (ten, piece) (in Japanese), cm, gram (b) CQs representing quantity concept, based on speaker’s observation or general metonymy. Eg. un brin de (a little), bribes de (scraps of), ひとつまみの (hito-tsumami no, a pinch of), 山盛りの (yama-mori no, a pile of). There are two cases for a CQ: (1) it can belong to only the (a) type or the (b) type, and (2) it can belong at the same time to both the (a) and (b) types. That is because, on the 1The UNL (Universal Networking Language) project was founded at the Institute of Advanced Studies (IAS) of the United Nations University in Tokyo in April 1996 under the aegis of UNU (United Nations University, Tokyo) and with financial support from ASCII corporation (a Japanese publishing company, 1977-2002) and UNL-IAS. http://www.undl.org/unlsys/unl/unl2005/attribute.htm Proceedings of MT Summit XVI, vol.1: Research Track Nagoya, Sep. 18-22, 2017 | p. 216 one hand, there are some CQs that play only the role of classifier or quantifier, and, on the other hand, there are CQs that play both of these roles. Eg. un brin de paille (a wisp of straw), un brin de folie (a touch of madness)2. When we started to deal with CQs expressions in the framework of French-Japanese MT, we met mainly the following difficulties, which were inherent in QCs: 1. Resolution of lexical ambiguity of polysemic nouns Eg. pièce (piece) : (Japanese translation as CQs) 枚(mai, sheet or ϕ3), 点 (ten, ϕ), 頭(tou, ϕ), 樽(taru, cask), etc. 2. Producing adequate CQs in Japanese when they are absent in French Eg. deux livres (two books) : (Japanese translation) 二冊の本 …

Content truncated to 500 words

Labeled Text

Although classifiers/quantifiers (CQs) expressions appear frequently in everyday communications or written documents, they are described neither in classical bilingual paper dictionaries, nor in machine-readable dictionaries. The paper describes a CQs dictionary, edited from the corpus we have annotated, and its usage in the framework of French-Japanese machine translation (MT). CQs treatment in MT often causes problems of lexical ambiguity, polylexical phrase recognition difficulties in analysis and doubtful output in transfer-generation, in particular for distant languages pairs like French and Japanese. Our basic treatment of CQs is to annotate the corpus by UNL-UWs (Universal Networking Language - Universal words)1, and then to produce a bilingual or multilingual dictionary of CQs, based on synonymy through identity of UWs. Keywords: classifiers, quantifiers, corpus annotation, UNL, UWs dictionary, phraseology study, Tori Bank, French-Japanese MT

Publication Details

Paper ID:
18586
Year:
2017
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan