常规对应、数据关联与“谷歌翻译”评价  

Conventional Correspondence,Statistical Correlation and Google Language Tools

作　　者：褚东伟[1] 

机构地区：[1]广东外语外贸大学,广州510420

出　　处：《广东外语外贸大学学报》2010年第4期50-53,共4页Journal of Guangdong University of Foreign Studies

基　　金：广东省哲学社会科学"十一五"规划2009年度学科共建项目"新时期商务英语教学模式创新与发展研究"(批准号:09GK-02)的阶段性成果

摘　　要：从语言的层面上讲,原语和译入语之间习惯性的对应关系叫常规对应,谷歌网络翻译系统用数据统计的方法利用平行语料库通过数据关联形成译文,同时也使初次翻译的结果和译入语单语文本语料库产生数据关联从而使译文更符合译入语规范。经过试用发现,谷歌翻译虽然还不很完善,但它已经不是传统的纯机器翻译。它已经是结合了人工翻译理念的机器翻译,具有多方面的作用。谷歌翻译所取得的进步是谷歌不懈的努力和人类机器翻译经验积累的结果,也是时代的产物,谷歌翻译技术的发展说明我国的翻译工作者和翻译研究者应该具有更广阔的视野和胸怀。The habitual correspondence between source and target language is known as conventional correspondence. Google Language Tools translates via its large parallel and monolingual corpora on the basis of statistical correlations. A test run of Google Language Tools shows that despite its flaws, it marks a breakthrough in machine translation technology for its simulation of human translation. Google＇s progress in its language tools is the result of Google＇s efforts and the accumulative experience of generations of workers in this field. It follows that this situation does entail that Chinese translators and translation scholars have a wider field of vision in their ideas and techniques.

关 键 词：机器翻译 网络翻译 常规对应 数据关联 

分 类 号：H085[语言文字—语言学]