文学翻译中的语义迁移研究--以基于信息贡献度的主题词提取方法为例

Revisiting Semantic Transfer in Literary Translation Studies:Using Key Keywords to Measure Informativeness

本文尝试提出信息贡献度这一概念,描述翻译文本中的词汇因其对译文特定部分或主题贡献最大的信息量而被赋予相应的权重。信息贡献度有其信息论、语料库语言学、翻译文学、文化传播以及翻译的文化研究等多重理论基础和研究价值。与传统的语料库经典研究方法和大数据研究方法相比,信息贡献度研究方法具有独特作用,可应用于多种与语义相关的研究。借助于Python构建的合理算法,实现了文学翻译作品历时性不同译本之间的文化意象的系统化对比研究。研究结果表明,基于信息贡献度方法的语义迁移研究是以文本语义结构作为研究的出发点,不仅保留了传统语料库研究方法的精髓,还为大数据时代Python工具在翻译文学和语料库翻译学领域里的应用开拓了新前景。

This paper attempts to propose a new concept of informativeness as a key means to measure the contribution of lexis to a certain section or overall theme of translated works. This concept draws on inspirations from multidisciplinary knowledge as wide as information theory,corpus linguistics,translated literature,cultural communication,and cultural research of translation,etc.and bears significant theoretical and empirical values.The key to traditional corpus research methods is the retrieval and identification of target words,while the key to big data research methods is the retrieval of big data and the training of word vector models. The former completes semantic construction in the subsequent process of interpreting the target word,while the latter constructs the semantic relationship through training the word vector model and obtain information for semantic interpretation in the training process.In comparison with the traditional corpus methods and big data methods,informativeness integrates the affordances of these two approaches while avoiding the limitations regarding the huge amount of data. This is because informativeness-informed studies interpret semantic transfers in line with semantic structures of those lexis in texts,and thus have the potential to be applied in various semantic studies.Using established algorithm available in Python packages, this research sets out to compare cultural images constructed in different translations of literary works. Specifically, this research, based on two translations of Lin Yutang ’s English novel Moment in Peking( aka Jinghua Yanyun in Chinese),scrutinizes such terms as "Young Master"( dashaoye),"She Fox"( hulijing), "Gentleman "( zhengren junzi) and"Ordinary People"( laobaixing) from the perspective of informativeness,providing positive evidences to augment applicability of language intelligence in translation studies.Our research showcases a closed loop from language intelligence and literary translation, meaning that comparing informativeness of key keywords can well indicate the directionality and paths of lexical semantic transfer.We argue that informativeness-informed studies on semantic transfer depart from semantic structures of texts,retain the basic tenets of traditional corpus linguistic techniques, but chart new territories by combining Python-based big data methods with literary translation and corpus-based translation studies.