A Study On The Representation Strategies Of Multi-Category Lexemes In The Contemporary Chinese Dictionary(6~th Edition)

Multiple class membership is so far a difficult and hot issue in the study of linguistic categories.Multiple class membership refers to a multi-functional phenomenon that a polysemous type word or lexeme has two or even more word classes at the level of communal language,which is specified by the headwords with multiple parts-of-speech in dictionaries.As a result,it is reasonable to choose dictionaries as our corpus to study word classes and multiple class membership.Based on the DIY Word Class Labeling Database of The Contemporary Chinese Dictionary(6~th edition),an in-depth survey of the representation strategies of multi-category words is conducted under the guidance of Two-level Word Class Categorization Model(TLWCCM).The findings show that the number and percentage of multi-category words haven't changed much compared with the 5th edition.The newly added multi-category words result from the adding of sense(s)(193),change of POS labeling(9),adding of new headwords(92)and the lumping of headwords(43).Besides,some multi-category words become single-classed,including especially those related to inference method or form of thought,sports and “results expressed in action”,the noun POS label of which are removed.Firstly we discuss the essence and causes of multiple class membership,the standard of identification of words and the standard of part-of-speech or word classes.Based on the discussion we find that the reason to remove self-reference usages is to follow the so-called “Principle of Parsimony”,which sticks to the principle of “fewest possible multi-category words”.However,the date from the Modern Chinese Corpus developed by the National Language Committee of China show that a lot of token words come from a small amount of multi-category words,following the power law.We propose that the “Principle of Parsimony” be modified to Critical Minority Law in multiple class membership,and multiple class membership be treated more tolerantly and wisely.Furthermore,the contradiction in dealing with fuzzy words is caused by the misconception about the relationship of self-reference and transferred-reference.In fact they are inclusive opposites instead of exclusive opposites and there exists the self-designation-like usages,which are caused by the failure to distinguish process and result or content.They are situated in the middle ground between self-reference and transferred-reference.The types of paraphrase patterns of self-reference senses are also explored.In a word,only by following the instruction from TLWCCM,adhering to the right order of judging parts-of-speech or word classes in modern Chinese dictionaries compilation,and distinguishing the word token categorization in syntax at parole and word type categorization in lexicon an langue can we get out of the dilemma of POS labeling.