A multimodular approach to case assignment in Japanese: A study of complex and stative predicates.

摘要

The primary goal of my dissertation is to challenge "syntactocentrism" in mainstream generative grammar and alternatively to advance a multimodular view of analysis that de-synthesizes much of explanatory role attributed to syntax as empirically more successful and conceptually more reasonable.;With particular reference to various morphological case manifestations in Japanese complex and stative predicates, this dissertation argues for a parallel grammatical architecture, sketched in Sadock (1985, 1991, 2003) and Jackendoff (1997, 2002), in which participant role structure and function-argument structure (among others) stand on equal footing with syntactic structure. I propose that the distribution of morphological case particles is a function of corresponding rules, which hierarchically relate overt nominal expressions to participant roles and arguments across distinct grammatical components. The main components of the dissertation (Chapters 3 and 4) are thus intended to develop a non-derivational, constraint-based case system that generates the canonical case arrays in Japanese and at the same time also regulates theoretical possibilities that are non-existent in reality. I also conclude that Kuroda's (1978) Linear Case Marking hypothesis along with Ostler (1981) is (when coupled with the proposed view of grammar) the correct analysis despite the 30 year neglect in generative studies of Japanese.