Frequency And Multiple Class Membership In Modern Chinese: An Empirical Study

The studies of multiple class membership in Modern Chinese have been trapped in the dilemma of its theory and linguistic reality for a long time. Scholars engaged in Modern Chinese grammar put forward the Principle of Simplicity which requires that multiple class membership be minimized so that Chinese grammar analysis could be under control. Although some studies arrived at the conclusion that higher frequency words are more liable to possess multiple class membership than lower ones by adopting data from dictionaries and vocabulary lists, there remain shortcomings in their sampling and data processing. If the correlation between word frequency and multiple class membership is positive, i.e. the high frequency words bear more multiple parts of speech, the Principle of Simplicity will run counter to this linguistic reality. Is the correlation between frequency and multiple word class positive? Does the Principle of Simplicity workBased on the prudent data collected from The Contemporary Chinese Dictionary (CCD5) (5th edition) (2005) and our DIY database sampled from the Modern Chinese Corpus of the Center for Chinese Linguistics at Peking University (hereinafter called CCL Corpus), this thesis claims that there is a positive correlation between frequency and multiple class membership. According to our usage-pattern survey based on CCL Corpus, the correlation between frequency and multiple word class is positive, which manifests that multiple class membership of high frequency words is more easily conversed and entrenched than that of the lower ones. However, CCD5, under the instruction of the Principle of Simplicity, shows inadequate evidence to support the correlation between frequency and multiple word class membership. In line with Usage-Based Theory, this paper suggests that lexemes of multiple parts of speech, carrying two or more word classes synchronically, come into being through repetition of certain lexemes in different syntactic environments by the mechanism of decategorization and recategorization, which abides by the principle of economy. Obviously, artificially reducing the number of lexemes with more than one part of speech conceals the linguistic realities, and meanwhile disorders the settlement of word class. Hence, POS (part of speech) labeling should take advantage of large and systematic corpora to investigate the usage patterns of every lexeme in terms of the language system comprehensively, rather than abiding by the alleged Principle of Simplicity.This study can not only shed new light on POS labeling for Modern Chinese dictionaries and POS tagging in Modern Chinese corpus construction, but also has theoretical significance for Modern Chinese grammar studies..