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

Categorization is a core issue in linguistics,and word class categorization is also regarded as a time-honored difficult issue.In the 36th Annual Conference of the German Linguistic Society in 2014,the study of linguistic categories like word classes was seen as the study of“God particles”.And the study of word class categorization of analytical languages like Modern Chinese is taken as the“Goldbach Conjecture”in linguistics.In the study of Modern Chinese word class categorization,multiple class membership is a controversial issue.As a unique word-formation phenomenon in Chinese,antonymous compounds have attracted much attention from academic circles.However,most studies are conducted around the cognitive mechanism,teaching and acquisition of antonymous compounds,and few of them investigate the word class labeling of antonymous compounds in dictionaries based on large data.This thesis takes the 264 shared antonymous compounds as the object of the study selected from two authoritative Chinese dictionaries:The Contemporary Chinese Dictionary(7^th Ed.)?CCD7 for short?and The Modern Chinese Standard Dictionary(3^rd Ed.)?MCSD3 for short?,and adopts the Two-level Word Class Categorization Theory as theoretical foundation to conduct corpus-based surveys by means of the Modern Chinese Corpus developed by the National Language Committee of China?CNC for short?and the Modern Chinese Corpus of the Center for Chinese Linguistics in Peking University?CCL Corpus for short?.It is found that CCD7 and MCSD3 artificially minimize the number of multiple class membership of antonymous compounds.Among the 264 antonymous compounds,the number of antonymous compounds which are improperly labeled in CCD7 and in MCSD3 are 62 and 81 respectively,accounting for 23.48%and 30.68%respectively.The problem of word class omission is striking.The proportion of multiple class membership of antonymous compounds based on corpora is 31.83%while the proportion of multiple class membership of antonymous compounds in CCD7 and in MCSD3 are 11.74%and 12.5%respectively.The study shows that the compilers of CCD7 and MCSD3 do not clearly distinguish between word classes at langue and word classes at parole,following the so-called“Principle of Parsimony”.The corpus-based usage pattern surveys are helpful in making exact word class labeling of lexeme in dictionaries.By conducting corpus-based surveys of the word class labeling of the shared antonymous compounds in CCD7 and MCSD3,this thesis is helpful in improving the word class labeling of antonymous compounds in Chinese dictionaries and deepening Chinese word class studies.