摘要：This article will argue that intuition-based description by native speakers to some extent may not be accurate in depicting what people really say and think about the language. Since the problem of learning a second language is essentially lexical, the introduction of machine-readable text corpora has been seen as an alternative for describing lexical behaviour of a natural language and media which second language teachers and learners can benefit from (see Fox, 1988, Willis, 1990, Francis, 1993). The corpus programs such as Concordancing, Sentence, and Collocation are capable of providing data of lexical properties, and thus, overcoming the failings of intuition. In addition, it also helps learners and teachers understand the characteristics of lexicon, various meanings of the word and acquire its syntactic properties.
关键词：korpus; studi kebahasaan; pembelajaran bahasa; bahasa kedua
