漢語翻譯文本中的簡化現象：以語料庫方法探究平均句長

Simplification in Translated Chinese Texts: A Corpus-Based Study on Mean Sentence Length

平均句長是簡化現象的測量方法之一：如果一個文本的句長較短，通常被認為較簡單，讓讀者易於了解。假設翻譯文本比非翻譯文本簡單，翻譯文本的平均句長就比較短。本研究旨在以語料庫方法研究漢語翻譯文本中的簡化現象，而且特別針對平均句長做研究。本研究所使用的語料庫為自建的懸疑小說可比語料庫（The Corpus of Comparable Mystery Fiction），分析平均句長（以詞計算）以及其延伸的測量方法：以字計算的平均句長，以詞計算的平均小句長和以字計算平均小句長。分析平均句長以及其延伸的測量方法的結果皆顯示本語料庫的翻譯文本的句長比非翻譯文本的句長短。

Mean sentence length is proposed as a measure of simplification: if a text has shorter mean sentence length, it is assumed to be simpler for readers to comprehend. Since translated texts are hypothesized to be simpler than non-translated text, they would presumably exhibit shorter mean sentence length. The present research aims to investigate, using corpus-based methods, the phenomenon of simplification in translated, compared to non-translated, Chinese texts. This paper focuses on measuring sentence length of the translated texts in the Corpus of Comparable Mystery Fiction, analyzing the results of mean sentence length and its additional measures: mean sentence length in terms of characters, mean sentence sub-unit length in terms of words and mean sentence sub-unit length in terms of characters. The measure of mean sentence length and its additional measures render consistent results showing that the translated texts of the corpus under study exhibit shorter sentence length than the non-translated texts.