An Empirical Study Of Word Class Labeling Of Antonymous Compounds In Chinese-English Dictionaries

The word class categorization of word type of analytic languages like Chinese has always been a hot and difficult issue in linguistics.In Chinese there are many antonymous compounds,which are referred to parallel compound words consisting of two monosyllabic morphemes whose meanings are antonymous or opposite to each other.For example,“呼吸”,“买卖”,and“大小”.Previous studies are mostly focused on the origins,cognitive mechanisms,classifications and other aspects of antonymous compounds which are seldom studied from the perspective of lexicology.This thesis takes two authoritative dictionaries,namely A Chinese-English Dictionary(3rd edition,CED3 for short)and A New Century Chinese-English Dictionary(2004,NCCED for short),as the sources of research data.It is found that there are 257 antonymous compounds shared by those two dictionaries,which will be our research object.It is intended to figure out the authentic word class labeling status of antonymous compounds, and try to analyze the grammatical categories and usage patterns through corpus-based surveys. Based on the theoretical model of two-level word class categorization in analytic languages(TLWCC for short) put forward by Wang Renqiang(2010), this thesis aims to explain the findings, improve the word class representations of antonymous compounds in C-E dictionaries, direct the word class labeling of philological dictionaries(namely, the word class labeling in the lexicon at communal level of langue), and offer theoretical basis and operational criteria for judging the accuracy of the word class labeling of the existing dictionaries. It has conducted a full survey on the word class labeling conditions of the 257 antonymous compounds, and has collected the word class labeling information from Contemporary Chinese Dictionary(6th edition, CCD6 for short) and corpora as a reference. By comparing, this thesis found there are 48 items are labeled differently among the three dictionaries, and there are 60 words are wrongly labeled among the other 209 entries. Among the error types of word class labeling, most are words belonging to noun and verb. The word class labeling accuracies of CED3, NCCED and CCD6 respectively are 65.76%,64.98% and 64.20%. Compared with the rate of multi-category words based on corpus usage pattern surveys stands by 44.36%, it only accounts for 14.01%, 16.34%, and 12.45% respectively in CED3, NCCED and CCD6. Multiple class membership is characteristic of analytic languages, and in the studies of Chinese word classes it is unjustified to follow the “principle of parsimony” which argues for fewest possible multi-category words. According to TLWCC, it is the word classes in the lexicon at communal level of langue which should be included in a dictionary, namely the meaning potentials. And the word class labeling of antonymous compounds in C-E dictionaries should be based on large-scale balanced Chinese corpora by analyzing the degree of conventionalization of propositional speech acts patterns to determine their word class membership, and to provide the English translation of word classes and typical exemplifications.This thesis has not only practical implication for C-E dictionary compilation, but also theoretical values to the Chinese grammar research.