摘要
语用学是研究语言的使用的一门科学,统计机器翻译方法通过词语在短语语境中的使用情况确定歧义词语的具体含义,这也是语用方法在机器翻译中的自然应用。语用的方法还能够通过分析源语言单元的关联关系来确定源语言的逻辑顺序关系,从而确定目标语言的逻辑顺序关系,这就使得源语言结构和目标语言结构具有可计算性。基于这种可计算性的语用自动调序,可以基本解决长距离调序问题,使得像英汉这样差别巨大的语言之间的自动翻译达到基本可读性目标成为可能。本文将论述基于语用方法的语用调序基本原理,并将实验翻译结果与现有统计机器翻译系统的结果进行对比分析,最后概括研究过程中的问题、解决方案。
Pragmatics is a scientific study of the use of languages. In Statistical Machine Translation, according to the use of the ambiguous words and expressions in the context, the exact meaning could be defined by statistical method and this is an unconscious application of the pragmatics. By way of pragmatics, the logic relations of the source language could also be defined by analyzing the relevant relations of the source elements, and hence the logic sequence of the target language could be reached, and the computability between the source and the target languages could be realized. Based on this kind of computability, the pragmatic automatic reordering may address the the issue of long-distance reordering, making it readable for the target language readers of such vastly different languages as English and Chinese. And by comparing the pragmatic-reordered translating results and current popular used MT results, some key issues and the addressing approaches can be revealed.
出处
《西安外国语大学学报》
CSSCI
2017年第4期77-81,共5页
Journal of Xi’an International Studies University
基金
2014年教育部留学回国人员科研启动基金"机器翻译中词汇的语用辨义"(项目编号:2014-1685)
2015年河北省社会科学基金项目"商务英语语用翻译研究"(项目编号:HB15YY026)的部分研究成果
关键词
机器翻译
语用
长距离调序
可读
machine translation
pragmatics
long-distance reordering
readable
