Lexical Courtesy Revisited:Evidence from Tsou and Seediq Wh-Constructions

Abstract

This paper seeks to further substantiateChomsky's (2000) claim that Merge preempts Move. Previous studies have shown that the claim has a great impact on our understanding of A'-dependencies, in that an operator-variable pair must be built before any movement akes place, as encoded by the Lexical Courtesy Hypothesis (Tsai 1999a,b). Crucially, it provides a coherent account of why English, Japanese, and Chinese all employ the (unselective) binding strategy to form whquestions, while differing in the magnitude of binding according to their distinct morpho-syntactic makeups, that is, lexical in English, phrasal in Japanese, and sentential in Chinese. On the empirical front, the proposal crucially relies upon a correlation between indefinite and interrogative wh-construals allowed in a particular language. Based on data from two Formosan languages, Tsou and Seediq, I argue that the correlation is indeed attested in VOS languages, and that Lexical Courtesy can be maintained cross-linguistically if we allow unselective binding as the optimal strategy of assigning wide scope to wh's-in-situ.