코퍼스를 활용한 서술어/비서술어 명사의 행동양식 분석

Abstract

This paper analyzes the behaviors of predicative/unpredicative nouns in Korean, based on the corpora. Previous analyses have been focused on the distinction between predicative/unpredicative nouns, mostly focusing on the intuitive data rather than on the corpus. We look into with the behaviors of predicative/unpredicative nouns in the corpora. Our methods reveal that the behaviors on predicative/unpredicative nouns are based on the combination of multi-tiered levels of classification features. We claim that there needs to be the notion of classification of typical vs. non-typical types among predicative/unpredicative nouns, requiring the gradience in grammaticality. Unlike the former linear classification, we use the statistics found in the corpora, leading the multi-dimensional analysis of the multiple combination of features.