| Summary: | 碩士 === 國立清華大學 === 臺灣研究教師在職進修碩士學位班 === 97 === The thesis deals with wh-questions and left periphery in Saisiyat. We focus on the topics regarding the interpretation and the syntactic distribution of wh's.
The interrogatives in Saisiyat are grouped into two types: nominal and adverbial according to their corresponding syntactic and morphological behaviors. Among them, nominal wh-words like general nouns can either be marked by case markers or function as predicates in equational constructions. On the other hand, the free distribution of adverbial wh-words, which violate subject-sensitive theory, can be reasonably illustrated under the T-feature checking hypothesis proposed by Tsai (2004). Moreover, we will also discuss some wh-words with dual-status, which can serve as a head and as a modifier.
We argue that Saisiyat is a true wh-in-situ language. Since nominal wh-words are licensed by operator-variable binding involving no syntactic movement, it lacks locality effects, and violates ECP and Subjacency. Crucially, the interpretation procedure is through the choice function put forward by Reinhart (1998). That is, wh’s serve as a choice function which selects the set variable denoted by N-restriction. The choice function, in turn, is translated as an f-variable being long-distantly bound by a Q-operator far away. Thus, it solves the problems of incorrect entailment resulting from applying unselective binding mechanism. Furthermore, the indefinite wh construals are due to the intervention between the Q-operator and other quantifiers.
We propose that the adverbial wh-words, i.e. nak ’ino’ ‘how’ and ’am powa’ ‘why’, in Saisiyat have different significance depending on their corresponding syntactic positions. According to the evidence from subjectivity restriction, intervention effects, no-matter constructions, multiple wh-constructions, etc., it follows that causal how and reason why are directly merged into left periphery and function as sentential operators, whereas manner how, instrumental how and purpose why occupy the vP periphery and serve as event predicates. Manner how as an adverb with quantifier feature must undergo LF movement. However, the latter two are licensed by operator-variable binding. Dissimilar licensing procedure results in intervention effects with different degree. Besides, reason how as an argument occurs in the complement position of predicate and is governed by the verb. Adopting the cartographic approach (cf. Rizzi 1997, Ramchand 2003) and inner-outer dichotomy (cf. Tsai 2007), we depict the syntax-semantics map of the wh’s. Thus, the asymmetry of intervention effects on wh’s can get a reasonable explanation under Rizzi’s (1997) Generalized Relativized Minimality (see also Tsai 2008).
|
|---|
