말뭉치와 언어연구: 외국의 사례와 경향

Abstract

This paper is a survey of recent trends in the corpus related linguistic studies, adding two specific cases of such research as a way to illustrate how corpora can contribute to the efforts of finding interesting distributional patterns of language. We first start in Section 1 with the question if Corpus Linguistics is popular at all in the first place and some ways to find the answer, suggesting that the ‘Google N-gram viewer’, based on the N-gram data collected from the very large corpus, provides a new and interesting method to such a query, just like corpus linguistics does to the more linguistic questions. In Section 2, the current trends are surveyed from several perspectives such as ‘corpora and theoretical linguistics’, ‘corpus annotation’, ‘corpora and other linguistic resources’, ‘corpus (access) tools’, ‘using corpora for linguistic research’, and ‘linguistic research from a narrow and a wider perspective’. Section 3 discusses the Sketch Engine as a case of showing distributional patterns of the words. Section 4 introduces a case study, following the methods proposed in Resnik (1996), where the selection preference information between a predicate and its object is inductively and automatically derived on the bases of a parsed corpus and a wordnet in Korean.