한국어 음운이웃 네트워크에 대한 어휘부 규모 효과: 타 언어들과 비교를 위한 한국어 음운이웃 네트워크 분석

초록

This paper analyzes the various sizes of Korean lexicons (39 lexicons ranging from 1,592 words to 101,804 words) in terms of phonological word-similarity in the framework of network theory. 101,804 words are selected to make these lexicons from Kang and Kim(2009). The phonological neighborhood network for each lexicon is constructed. Its giant component (GC), average shortest path length (ASPL), average clustering coefficient (ACC), and assortative mixing by degree (AMD) are measured. The results are compared with those of several languages in previous studies. It is shown that the Korean phonological neighborhood network exhibits some language-universal properties. It satisfies the requirements for being a small-world network. However, language-particular properties are also found in the Korean phonological neighborhood network. The GC size is relatively larger than those in the other languages. In the value-curves of GC size, ASPL, and AMD based on the lexicon size, Korean exhibits the rise and fall shapes, which are not observed in the other languages.

키워드