Classifiers in Chinese pose an interesting testing ground for theories of categorization and cognition. This study takes a psycholinguistic approach to previous theoretical investigations concerning categorization of Chinese classifiers (Tai 1990a, 1990b, 1992), and expands on the neurolinguistic results discussed in Tzeng, Chen and Hung's (1991) study of classifier production in Chinese normals and aphasics. Based on three generalizations found in elicitation experiments on normals, I hypothesize that the neutral classifier is less likely to occur when the relationship between a noun and a certain classifier is unique and elaborative. In addition, I re-evaluate Tzeng et al.'s (l991) data and show that Wernicke's aphasics are not crossing semantic classes when they neutralize to the classifier zhi; instead they are practicing a form of code-switching. The results of my study demonstrate that both normal and aphasic subjects adhere to similar constraints on classifier production.
研究中文量詞對了解認知能力以及範疇理論有許多幫助。這篇文章從心理語言學的角度,擴大前人對範疇理論研究(Dai 1990, 1990b, 1992)以及更深入探討失語症病人使用量詞(Tzeng 1991)的現象。根據正常人elicitation實驗的三個歸納結果,我斷定當量詞跟名詞有一種獨特或elaborative的關係時,那麼量詞的中化(neutralization)就比較不容易出現。而且我再檢查Tzeng(1991)的資料證明Wernicke病人並沒有混淆語意的類別。我的研究結果表示正常人和失語症病人的量詞使用都受同樣的限制。