基于语料库及对应词表的英语特异组合发现方法

Corpus and Wordlist Based Identification of Indecomposable English Word Combinations

语料库语言学研究搭配多从单语视角出发,利用统计方法,计算真实文本中词语与其他词语的共现强度。一些语言学家从双语对比和二语学习的角度,提出以目的语中是否存在可预测的对应词作为区分自由结合与搭配的依据。本文采取后者的视角,利用前者的方法,提出"平行(语法模式一致)"和"不对应(构成成分不能直译)"的原则界定英语"特异组合",并提出一项发现特异组合的可操作性强的方法。该方法采用双语对应词表及词语语义知识库、各类语料库(英、汉单语语料、英汉平行语料库)以及浅层语法分析工具,通过一系列步骤,能够有效发现中国学生难以学习和掌握的特异组合。

In corpus linguistics, collocations are often identified with the frequency information of a word and its concurrent words. For some linguists in the contrastive linguistics and second language learning circles, free word combinations and collocations are distinguished on the basis of whether predictable equivalents can be found in the phrase in the target language. In this study, we coin the term ＂specific combination＂, referring to those phrases in the target language which are ＂grammatically parallel＂ but ＂lexically inequivalent＂ to a phrase in the source language. We therefore propose an operationalizable method to identify such combinations. With resources such as a bilingual word list, a thesaurus, as well as monolingual and parallel corpora （English and Chinese monolingual corpora, English and Chinese parallel corpora） and a shallow parser, these ＂specific combinations＂, which are more likely to pose difficulty for Chinese learners of English, are identified through a series of procedures.