摘要
否定結構和“Ａ不Ａ”疑問式普遍被看作是句法現像。但是在廣東方言中，這兩種結構卻是在詞匯中形成的，因為否定前綴詞素ｍ“唔”不是在句法中，而是在詞匯中附加子動詞之前的。
ｍ的前綴性質可以由下述事實證明：一、ｍ從來不單獨出現，甚至不能單獨用來回答問題。二、某些動詞（諸如 yau“有”）的肯定形式和否定式是由不同的詞根派生出來的。三、一些詞項（比如 dou“都”）對一般性規則而言是例外現像。四、ｍ的某些用法顯然像英語中的un-一樣，是詞匯前綴。
但是，語言學界一直認為ｍ是由句法規則插入句中，用來否定動詞詞組的。本文試圖證明ｍ所否定的不是詞組，而是全部帶有〔＋Ｖ〕成份的詞，包括動詞，形容詞，前置詞，以及一些副詞。本文還證明，“Ａ不Ａ”問句的形成只能來自詞匯重?規則，根本不可能是句法現像。
This paper argues that two phenomena widely perceived as syntactic, negation and A-not-A question formation, are performed in the lexicon in Cantonese. Specifically, the negative morpheme m 唔 is prefixed in the lexicon, not in the syntax.
Evidence that m 唔 is a prefix includes the following. It never appears in isolation, not even as a response to questions. Certain verbs, like yau, 有 'have', show special suppletive forms in the negative. Some lexical items, like the quantifier dou, 都 all, are exceptions to the general rules. Lastly, in certain uses it is clearly a lexical prefix like the English un-.
However, it has always been thought to be inserted by syntactic rule, and to negate the phrasal category VP. I argue that it negates not phrasal but lexical categories, including all those with the feature [＋V]: verbs, adjectives, prepositions, and some adverbs. I also argue that A-not-A questions must be formed by a lexical reduplication rule, not in the syntax at all.
This proposal will offer a way of dealing with the interaction between negation and aspect in Cantonese.
I have drawn heavily on Yau (1973), a very comprehensive and insightful work on negation in Cantonese.