基於語料庫的譯者風格研究：以《老人與海》及其三種中譯本為例

Abstract

譯者風格的研究是一項重要的研究議題，許多學者也對此有所探討。然而，目前對於中文翻譯風格的研究卻不多，因此本研究欲朝此方面研究，探究厄尼斯特·海明威所著小說──《老人與海》的三種中文譯本，分別由張愛玲、趙少偉及余光中所譯。本研究將以實證研究方式分析三譯本，著墨於譯者的單字、句長及用語特徵。本研究將採用多種語料庫分析工具，包含 LF Aligner、CUC_Paraconc、ICTCLAS 2016、CKIP中文斷詞系統、Anymalign、Antconc、Stanford Log-linear Part-Of-Speech Tagger、AntWordProfiler、中文文件分析系統及 Microsoft Excel 2016。透過這些工具的分析，能看出譯者風格差異，如：平均句長、詞性比例、原文譯文關聯係數、用字選擇等資訊。研究結果看出，三譯者用字選擇有所差異，最值得一提的結果是，趙少偉的譯文較為口語和方言，使用如多使用「老漢」、「咱們」及許多感嘆詞和兒化韻詞語。以字元數分析中看出，張愛玲較為貼近原文，因張愛玲與原文的關聯係數最高。而余光中常使用短句，由字數分析可以看出余用字較精簡，而透過成語使用比例也能看出余多使用書面語翻譯。除研究三譯者的風格外，本研究欲為譯者風格研究提出一個較連貫且客觀分析的研究方法，希冀呈現質量並重的內容與證據。

The study of translators’ styles is an important research topic that has attracted increasing attention in translation study (e.g. Baker, 2000; Munday, 2012; Saldanha, 2014, among others). However, there have been very few studies on Chinese translators’ styles so far. To bridge this gap, this study tries to address translators’ styles of the three Chinese translations of The Old Man and the Sea translated by Eileen Chang, Zhao Shao-Wei, and Yu Kwang-Chung, respectively. Drawing on empirical evidence, this paper analyzes the three translations in terms of vocabulary, syntax, and discourse features. Several corpus analysis tools have been employed, including LF Aligner, CUC_Paraconc, ICTCLAS 2016, CKIP Word Segmentation System (中研院中文斷詞系統), Anymalign, Antconc, Stanford Log-linear Part-Of-Speech Tagger, AntWordProfiler, Chinese Text Analysis (中文文件分析系統), and Microsoft Excel 2016. Through these tools, the differences of the three translators’ styles are manifested in terms of features such as average sentence length, part-of-speech ratios, correlation between the original text and target texts, word choices, and many others. Scrutiny of the word alignment based on the original text and the three Chinese translations reveals the differences of word choices by three translators. The most interesting finding is that Zhao Shao-Wei’s translation is more colloquial and vernacular, as evidenced by the abundant use of sentence-final particles and words with regional characteristics such as “老漢,” “咱們” and the suffix “兒.” Eileen Chang’s rendition is considered the most literal one among the three works. This can be seen from the fact that it has the highest correlation coefficient with that of the original text in terms of sentence length. Yu Kwang-Chung tends to use fewer words and more literary words than other two translators, as evidenced by the use of short sentences and Chinese idioms. Aside from analyzing the three translations, this paper also proposes a coherent corpus-based methodology to study translator styles by presenting both quantitative and qualitative evidence.