An Empirical Study Of Conjunction Labeling In Modern Chinese Dictionaries

The researches on word classes or parts of speech (POS) have a long history of more than two thousand years. However, there are still controversies on the essence of POS and the criterion or the basis of POS classification etc. The controversies over Chinese word classes have never stopped since the publication of Ma Shi Wen Tong in 1898, as Chinese is believed to be lack of morphological markers and morphological change. Word class, as the basis of grammatical research, is also one of the important elements of dictionary-making. POS labeling is still a tough issue and the weakest links in the compilation of Chinese dictionaries. The Contemporary Chinese Dictionary (5th edition) (CCD5 for short) and Modern Chinese Standard Dictionary (MCSD for short) are the two most influential dictionaries among the present influential ones with POS labeling.The study of Chinese function words has a longer history than Chinese content words, so does the POS labeling of Chinese function words. Does this signify that POS labeling of function words is more accurate and scientific than that of content words? This thesis focuses on Chinese conjunctions and restricts the scope on CCD5 and MCSD, the following problems are intended to be discussed: (1) Do these two dictionaries adopt the same criteria in labeling conjunctions? (2) Is their labeling of conjunctions correct? If not, what problems are there in their labeling of conjunctions? What is the mechanism or motivation behind these problems?We will conduct the empirical research on the 81 differently labeled conjunction lexemes based on the corpus launched by the Center for Chinese Linguistics at PKU. It is found that, for the disyllabic controversial lexemes, CCD5 often neglect the prototypical use of lexemes while MCSD often leave out the extended functions which have already become the norm usage. Grammaticalizaiton is the common character of human language. Therefore, it is clear that both CCD5 and MCSD fail to reflect the actual usage of words. The static and dynamic view of categorization can be combined dialectically to study the word class category. As the development of the human languages, word class category always goes through the process of decategorization. Grammaticalization as the theoretical framework of this thesis, the change of which is gradual, is the mechanism of decategorization and recategorization.Based on statistical analysis and case studies, we can obtain the following findings: (1) The disyllabic controversial conjunction lexemes in the two dictionaries should be treated as having multiple word classes instead of one word class, including the prototypical function and the extended function which are entrenched in modern Chinese. That is to say, they are actually words belonging to both conjunction and other function words, such as preposition, or they are words having conjunction and its content origin such as verb and noun. (2) The norm usage of the word should be treated as one sense in the dictionary, but there are cases that these two dictionaries have left out such usages. (3) Most of the monosyllabic conjunctions in ancient Chinese have lost the usage of conjunction and become morphemes or shifted to other word classes instead of conjunction. (4) The inconsistent treatment of morphemes is evident in these two dictionaries.This study integrates the qualitative and quantitative research methods. On the basis of empirical study, it also finds the deep reason behind conjunction labeling. It has both theoretical significance in lexicography and practical guidance in POS labeling of headwords in Chinese dictionaries.