從語料庫統計的觀點分析美國之音新聞英文被動句中文譯文

Corpus-Based Analysis for English-Chinese Translation of Passives in VOA News

譯者英翻中時，常常會不自覺受到英文語法的影響，譯出的文字往往反應英文語法、不符合中文的表達習慣，成為西化中文。本文以「被動句型」為例，以科學的方法，來討論美國之音新聞英文「被動句」的中文譯文。本論文使用語料庫研究法，透過中文非譯文語料庫裡的例句，有系統地分析「被動句」中文譯文。中文非譯文語料來自英國「Sketch Engine（SKE）」繁體中文語料庫，中文譯文語料來自「NTU Chinese－English Parallel Corpora」。本文計算譯文語料在SKE語料庫裡的「出現頻率」及「詞彙搭配力（t值及MI值）」，並利用「文本索引」及「詞彙特性素描」功能討論詞語使用型態。本論文提出利用語料庫來檢視譯文是否為道地中文的方法，幫助譯者自我改善英到中譯文的品質。本文的另一個貢獻在於以統計量化方式闡述翻譯學裡語言規範的概念。

When translating English into Chinese, translators are often unconsciously influenced by English syntax, which often results in the westernization of translated Chinese. This study addresses this issue by employing scientific methods to look at the English-Chinese translation of passives in Voice of America News. Based on the “Sketch Engine Traditional Chinese Corpus” and the “NTU Chinese-English Parallel Corpora,” we examine the quantitative characteristics of westernized Chinese in terms of word frequency and two statistical measures for collocations—Mutual Information and t score. Corpus tools such as “concordance” and “word sketch” in Sketch Engine were employed to observe and analyze collocation patterns in authentic Chinese. The goal is to help translators avoid using nonidiomatic and inauthentic Chinese and to advocate the use of both monolingual and bilingual corpora as important resources for improving the quality of translation, as well as the use of statistics to present the norm of the language in translation studies.