基于语料库的汉语翻译语体特征多维分析

A corpus-based multi-dimensional analysis of the stylistic features of translated Chinese

本文基于汉语翻译和原创类比语料库,对汉语翻译语体进行多维/多特征分析。研究表明:多维分析法可区分汉语翻译和原创语料,找出两者具有区别性的语言特征。文中对多个汉语语言特征频率进行了因子分析,得到两个主要因子。第一个因子可区分文学与非文学语料;第二因子可区分翻译与原创语料。按因子负荷的高低,汉语翻译语体的典型特征有:语法显化程度提高,助、介、连、代词等均存在显化趋势;高频词、习用语、成语及汉语特有结构"被"字句和"把"字句等的使用频率显著提高。

This paper is devoted to a multi-dimensional/multi-feature (MD/MF) analysis of the stylistic features of translated Chinese based on comparable corpora of translated and original Chinese texts.The investigation reveals that the MD/MF analysis is effective in differentiating the translated Chinese texts from the texts originally written in Chinese and in finding out the linguistic features that enable the differentiation.The two factors produced by Factor Analysis can explain most of the linguistic features taken into account.The first factor differentiates the literary from the non-literary texts; the second differentiates the translated from original texts.Based on the factor loads,the following features are found to be the most typical in translated Chinese texts:general tendency towards grammatical explicitation (increased ratio of functional words); higher frequencies of auxiliaries,prepositions and pronouns; higher frequencies of the list-head (most frequent) tokens,idiomatic phrases and some other expressions (like Ba and Bei constructions) that are uniquely Chinese.