Quantitative methods for identifying systematic polysemy classes

Abstract

In this paper we report the results of four experiments conducted to extract lists of nouns that exhibit inherent polysemy from corpus data following semiautomatic and automatic procedures. We compare the methods used and the results obtained. We argue that quantitative methods can be used to distinguish different classes of polysemous nouns in the language on the basis of the variability of copredication contexts.