The semantic difference between Chinese CHUAN and DOU In Chinese, both CHUAN and DOU can express the meaning of "all" as shown in (1a) and (1b).This paper examines the semantic properties of CHUAN and DOU, and argues that they have different semantic functions. I adopt Lin's (1998) idea that DOU is a generalized distributivity operator with quantificational force distributing over the members of a plurality cover (Schwarzchild 1996). I argue that the difference between CHUAN and DOU is that CHUAN is not a distributor and its semantic function is solely to ensure that the value of cover is a good fit in the sense of Brisson (1998). This proposal accounts for the distributional differences between CHUAN and DOU. In my account, since only DOU is a distributor with quantificational force but not CHUAN, the fact that only DOU but not CHUAN can occur with wh-words such as shei "who" and shenme "what" and turn them into universal quantifiers is expected (e.g. (2a) and (2b)).There is another context in which DOU but not CHUAN is used. When DOU used together with LIAN, it acts like a "focus marker" to yield a reading equivalent to English even (e.g. (3a), (3b), (3c)). Following Wu (1999), DOU here is a distributor distributing the property of the predicate over every member of Rooth"s alternative P-set, which is invoked by LIAN. CHUAN is incompatible in this construction because it is not a distributor in my account. The idea that the function of CHUAN is to ensure a good-fitting cover also accounts for the difference between sentences with CHUAN and those without it. Sentences with definite plural NPs sometimes allow the non-maximality reading, which Brisson (1998) attributes to the possibility of allowing ill-formed cover as pragmatic weakening. The claim that CHUAN eliminates all ill-fitting covers from consideration to ensure that the value assigned to cover is a good fit predicts that pragmatic weakening is not allowed in sentences with CHUAN. The prediction is borne out. While sentence (4a) is acceptable in the situation where one of the clothes is still in the laundry basket, sentence (4b) is considered false under the situation. I show that my analysis accounts for the distributional difference between CHUAN and DOU, and distinguishes sentences with CHUAN from those without it. Finally, this proposal has a theoretical implication in that it is an additional empirical argument for Brisson's (1998) notion of good fit. Data (1a) naxie shu, wo dou kan-wan le those book, I all read-finish ASP "I finished reading all of those books." (1b) naxie shu, wo chuan kan-wan le those book, I all read-finish ASP "I finished reading all of those books." (2a) shei dou hui lai who all will come "Everyone will come." (2b) *shei chuan hui lai who all will come "Everyone will come." (3a) lian Zhangsan dou lai le even Zhangsan all lai ASP "Even Zhangsan has come." (3b)*lian Zhangsan chuan lai le even Zhangsan all lai ASP "Even Zhangsan has come." (3c) *lian Zhangsan lai le even Zhangsan lai ASP "Even Zhangsan has come." (4a) ifu, wo xi-huao le clothes, I wash-finish ASP "I have washed (all of) the clothes." (4b) ifu, wo chuan xi-huao le clothes, I all wash-finish ASP "I have washed all of the clothes." References Brisson, C. 1998. Distributivity, maximality and floating quantifiers. Doctoral dissertation, Rutgers University, New Brunswick. Lin, J. 1998. Distributivity in Chinese and its implications. Natural Language Semantics 6: 201-243 Schwarzschild, R. 1996. Pluralities. Dordrecht: Kluwer. Wu, J. 1999. A minimal analysis of Dou-quantification. University of Maryland, College Park. ÿ