On The Asymmetry Between Syntax And Semantics Of Chinese Wh-Phrases

This thesis discusses three issues related to Chinese wh-in-situ:its licensing mechanism, its syntactic movement and the association of adverbs with wh-in-situ.There exist two competing approaches explaining why a wh-in-situ can be interpreted in its scope position. One assumes that a wh-in-situ is an interrogative operator and moves covertly to its scope position at LF while the other assumes that a wh-in-situ is a variable and is bound by an interrogative operator in the scope position. We call the former movement approach and the latter non-movement approach. Movement approach, put forward by Huang (1982), works well in GB, in which the transformational rule QR maps S-Structure and LF. Tsai's series of works (1994,1999,2008) are representative of non-movement approach, which is supported by many scholars. This approach incorporates well in the MP since Merger is more economical than Move. This thesis argues for movement approach based on the following arguments.The first is concerned with intervention effects. Wu (1999) finds that a sentence in which a focus or a quantifier c-commands a wh-in-situ is ungrammatical but provides no explanation. Since then, these data have been left unaccounted for. Cross-linguistic data from many languages such as Japanese, Korean and German (Beck1996a, b,2006; Beck&Kim1997; Ko2005) show that LF-movement of a wh-in-situ is blocked by a scope bearing element. Intervention effects are regarded as strong arguments for movement approach. This thesis argues that Wu's (1999) linguistic data can be attributed to intervention effects. Then more data have been studied and I conclude that Chinese wh-phrases do show intervention effects, interveners being a weak quantifier and a focus.Second, I explore scope ambiguity between a wh-phrase and a strong QP. After pointing out empirical problems with scope principles proposed by May (1985) and Aoun&Li (1993), I put forward the LF-chain based scope principle, which can account for more linguistic data. The LF-chain can not be generated without LF-movement of a wh-in-situ, thus scope ambiguity forms another piece of evidence for movement approach.Third, I offer a novel analysis of island effects from the perspective of presuppositions. Based on different judgments of linguistic data, both movement and non-movement approaches regard island effects as their empirical evidence. Tomioka (2009) holds the opinion that a WHY-question presupposes the truth of a non-WHY-proposition while other wh-questions include variables in their presuppositions. I extend this analysis further and find why has no definite counterpart in its non-interrogative sentence while other wh-phrases have. My analysis coincides with Lin's (1992) observation that only why can not be explained out of an island. My explanation is neither for movement approach nor for non-movement approach.Considering the above evidence, I am for movement approach. In my opinion, LF-movement and syntactic movement differ in driving force and intervention effects. I further analyze problems with non-movement approach which is based on empirical data that wh-movement and question particles are in complementary distribution and Chinese wh-phrases are used as indefinites in some cases. However, these data are not proper according to Bruening (2007) who says that most languages, wh-movement or wh-in-situ, have question particles and that there is no connection between wh-in-situ and wh-indefinites. Thus, non-movement approach can not hold water since its empirical evidence is not proper.The association of adverbs with wh-in-situ is very complex since adverbs can affect the grammaticality or semantics of wh-interrogatives. This thesis investigates dou 'all'-quantification of a wh-noun and its syntactic realization. It has been noted by Chinese grammarians that a wh-noun to the left of dou'all'is interpreted as a universal quantifier while that to the right of dou'all' is interrogative. Although there is an extensive literature dealing with this issue, there has been no consensus as to the semantic function of dou'all'and the quantificational status of a wh-noun. This thesis concludes that dou'all'is a predicate-related distributor and syntactically projected into DistP when associated with elements to its left and dou'all'expresses exhaustiveness and is an adjunct to IP syntactically when associated with a wh-noun to its right. The universal interpretation of a wh-noun to the left of dou'all'has been cited as empirical evidence for non-movement approach, which is against my analysis since a wh-noun to the right of dou'all'is c-commanded by dou'all'but has an interrogative reading.Syntactic movement of a wh-noun is also covered in my study. I provide evidence to show that this movement differs from English wh-movement, Japanese wh-scrambling and topicalization and then argue that a wh-noun is a focus in the sentence and this movement is an instance of focalization.Another contribution in this thesis is that I provide new evidence for the classical dichotomy of QPs. A weak QP is a potential intervener for LF-movement while interaction between a strong QP and a wh-phrase may lead to ambiguity. Furthermore, I provide empirical and theoretical evidence to argue that a plural DP is a strong QP inspired by Daniela's (2006) study of English DPs.