Argument Structure and Ditransitive Verbs in Japanese

Abstract

Ditransitive verbs such as send and give appear in two distinct structures in English, the double object and the to-dative constructions. It is well known that the two differ semantically and syntactically. In some recent works, it is suggested that the semantic differences observed by Bresnan (1978), Oehrle (1976) and others, and the structural properties noted by Barss and Lasnik (1986), Larson (1988), and others, can both be captured by postulating an extra head for the Double Object Construction (DOC, e.g., Marantz (1993), Harley (1995), Pylkkanen (2002)). This head, which corresponds to the applicative head in Bantu languages, takes the goal as its specifier and relates it either to the VP that contains the verb and the theme (Marantz (1993)), or directly to the theme (Pylkkanen (2002)). The applicative head contributes the meaning distinct to the DOC, and it gives rise to the hierarchical structure noted by Barss and Lasnik. This applicative head is missing in the to-dative so that this construction has an argument structure distinct from the DOC. In this paper, we will look at the corresponding construction(s) in Japanese. Unlike English, Japanese appears to have only one structure, in which the goal is marked with the dative and the theme with the accusative case marking. The goal-theme order is assumed to be the basic order (Hoji (1985), Takano (1998), Yatsushiro (1998, 2003)). The only variation is that the theme can occur before the goal, but this is viewed simply as an instance of optional scrambling. We will give arguments that the difference between English and Japanese is only apparent. With close scrutiny, we find that the two argument structures corresponding to the DOC and the to-dative in English exist in Japanese.