一份社会语言调查对汉英机器翻译中词语切分的启示——汉语分词与汉英机器翻译研究系列之二

Chinese Word Segmentation in Machine Translation in the Light of an Investigation

本论文汇报了一份由两个问卷调查和八个后继采访组成的语言调查,其目的主要是为了探寻人们如何在阅读中进行汉语词语切分。本文作者接着讨论了问卷与采访的结果,将其与机器翻译实验中的相关结果进行了对比分析,并在对比中凸现人与机器在汉语分词过程中所表现出来的方法和策略差异。我们希望,人类独特的分词方式与方法可以给机器翻译中的汉语分词技术带来启示,对将来的机器翻译系统(尤其是词语切分部分)的设计有所帮助。

This paper reports on a language investigation that consists of two surveys and eight interviews,which aims to find out how Chinese speakers segment Chinese sentences into words and how similarly they carry out this word-segmenting process.The author then proceeds to discuss and compare the results from the surveys and interviews with those from some machine translation experiments in a previous study,foregrounding those findings that show methodological and strategic differences of word-segmentation between humans and machines.Based on these findings,some implications are drawn in the hope that the knowledge of humans' unique word-segmentation process may shed light on Chinese word-segmentation techniques in machine translation(MT)and contribute to future MT system designing,especially regarding the part of Chinese word segmentation.