Using Case Prototypicality as a Semantic Primitive
Dan-Hee Yang, Ik-Hwan Lee, Mansuk Song Song · Institutional Repositories DataBase (IRDB) · 1998
A lexicon is essential to provide a proper semantic analysis of linguistic expressions. Thus, much work has been done to build one automatically. However, the results were not enough to define word meaning, but no more than a kind of thesaurus or taxonomy. In this paper, we present a new method that is not only more practical and efficient for implementing on computers, but also more reasonable in view of psycholinguistics. We first note that componential analysis is not real in our cognitive system. We define word meaning by a set of Case prototypicalities. We show that this knowledge can be constructed by supervised machine learning on the basis of Case particles and the collocational information from a large corpus. 1.