@article{oai:nagoya.repo.nii.ac.jp:00027967,
 author = {Osawa, Satoko},
 journal = {IVY},
 month = {Oct},
 note = {The purpose of this paper is to show a set of syntactic behaviors is determined by a certain morphological property. In particular, I will investigate a relation between scrambling and Case systems and show that the possibility of scrambling, the level of the θ-Criterion, and the requirement of adjacency are determined by the property of Case licensing systems. A traditional analysis of scrambling as an optional movement has been problematic within the Minimalist Program, since the Last Resort Principle requires that movement be driven by morphological reasons and hence no optional movement should exist. Recently Bošković and Takahashi (1998) propose a base-generation analysis of scrambling, according to which a scrambled DP is generated in its surface position and undergoes LF lowering for the θ-theoretic reason. This analysis seems to be a promising approach in that it resolves the problem concerning the Last Resort Principle and it also can accounts for many otherwise unexplained phenomenon. Adopting Bošković and Takahashi's base-generation analysis, I will claim that a "scrambled" DP is base-generated freely in an adjoined position and then undergoes LF lowering. However, I will argue LF lowering is triggered by the Case requirement but not by the θ-requirement since the base-generation analysis with θ-driven LF lowering poses some problems. Given the Case-theoretic LF lowering, the difference of scrambling between English and Japanese can be attributed to the difference of Case licensing systems that these two languages use. I will put forth the lexical Case licensing hypothesis for Japanese accusative Case. According to the hypothesis, Japanese accusative Case is licensed within a lexical domain by its predicate head, in contrast to structural Case, which is licensed in a functional domain by checking. This proposal not only accounts for the difference in the possibility of scrambling between English and Japanese but also receives independent supports from the subject-object asymmetry of scrambling in Japanese and the scrambling of nominative DP in Japanese. Moreover, I will show that the difference in Case licensing between English and Japanese determines some other syntactic differences between the two languages. First, it accounts for why θ-positions can be satisfied at LF in Japanese (Saito and Hoshi 1994) but it must be satisfied prior to LF in English (Chomsky 1995). Second, it provides an account of a question as to why English exhibits the adjacency effect but Japanese does not. The investigation leads us to a conclusion that a certain morphological property determines a set of syntactic behaviors. To put it more concretely, the Case licensing systems determine the possibility of scrambling, the level of the θ-positions and the requirement of adjacency.},
 pages = {87--116},
 title = {Case Licensing and Scrambling},
 volume = {31},
 year = {1998}
}