On nominative-genitive conversion

要旨

Nominative-Genitive Conversion (hereafter NGC) in Japanese, which is also often called Ga/No Conversion, is one of the most intriguing syntactic phenomena in the history of generative grammatical study of Japanese (see Harada 1971, 1976, Bedell 1972, Inoue 1976, Shibatani 1977, 1978, Nakai 1980, Saito 1982, Fukui 1986, Fukui and Nishigauchi 1992, Terada 1990, Murasugi 1991, Miyagawa 1989, 1993, Ura 1993, Sakai 1994, Hasegawa 1995, Watanabe 1994, 1996a, 1996b, and Ochi 1999, among many others). Nevertheless, as we will see below, the phenomenon has not yet been provided with a conceptually and empirically adequate theoretical account. The purpose of this paper is to elucidate the syntactic architecture of NGC and to explore an optimal account for the phenomenon within the framework of the Minimalist Program (cf. Chomsky 1995, 1999, 2000). In particular we will reject on much empirical grounds the ECM/Raising analysis of NGC, which assumes the genitive Case feature-checking by an external D head. Our investigation reveals that UG allows another mechanism of genitive Case checking in addition to D. More specifically, it will be argued that the genitive Case can be checked by the φfeature of the C-T-V amalgamate formed via AGREE, independently of D. It is shown that this hypothesis in fact has a number of important consequences for parametric syntax.

主題