中日同形词意义用法距离的计量研究——以对译比构建的F-measure为尺度  

Measuring the Semantic Distance of Chinese-Japanese Homographs: Approached from the F-measure Constructed on the Basis of Translation Ratio

作　　者：施建军[1] 谯燕[1] 

机构地区：[1]北京外国语大学北京日本学研究中心,北京100089

出　　处：《解放军外国语学院学报》2016年第4期76-84,160,共9页Journal of PLA University of Foreign Languages

基　　金：教育部哲学社会科学研究重大课题攻关项目"东亚国家语言中汉字词汇使用现状研究"(12JZD014);国家社会科学基金项目"中日两国现代语言生活中的同形词汇研究"(12BYY138);国家社会科学基金项目"两岸三地现代汉语对日语借词的吸收及创造性使用研究"(15BYY187)

摘　　要：中日同形词意义用法的差异一直是日语教学的一个重点,也是汉日语对比研究中的一个热点。传统研究试图用同形同义、同形异义、同形近义三分类的方法从意义用法上区分同形词。但是,这种区分方法存在一个难以解决的问题,即同形近义词的用法差异得不到精确描述,因为同形近义词的用法差异有大有小,其意义用法距离的大小无法用"同形近义"来衡量。本文提出了以对译语料库为知识库,以同形词的对译比和在其基础上构建的F-measure为参数描写中日同形词意义用法距离的方法,并用1900余对常用的中日同形词进行验证,取得了很好的效果。The semantic difference of Chinese-Japanese homographs has been the hot topic of Japanese teaching and Chinese-Japanese comparative study. Traditional researches tried to distinguish homographs semantically by means of three classification methods,i. e. the same meaning( S),the different meaning( D) and the overlapping meaning( O). However,this method encounters a big problem,i. e. the semantic differences among homographs with the overlapping meaning cannot be described accurately. This is because the semantic differences range from huge ones to minor ones,which make them hard to measure in terms of'overlapping meaning'. This paper puts forward a method to measure the semantic distance of Chinese-Japanese homographs,which uses translation corpus as knowledge base and takes translation ratio and F-measure constructed on the basis of translation ratio as parameters. The effectiveness of this method was verified by a test based on 1900 pairs of Chinese-Japanese homographs.

关 键 词：中日同形词 意义用法距离 对译比 F-MEASURE 

分 类 号：H365.9[语言文字—日语]