Why Japanese is different: Nominalization and verbalization in syntax and the distribution of arguments

A new theory of parametric variation is developed to explain certain differences between English and Japanese. Under a syntactic approach to derivational morphology, a fundamental hypothesis is proposed which states that Japanese is a ‘nominal language’, where all verbs are derived from nouns, and thus differs from English, where verbs and their nominal counterparts are derived from common roots in a parallel way. More specifically, it is argued that whereas in English verbs are derived by verbalization of simple roots, in Japanese they are derived by verbalization of nominalized roots. It is proposed that a parameter couched in terms of selectional relations between syntactic heads is responsible for the difference. It is shown that this parametric difference, coupled with syntactic principles, makes it follow that all Japanese nouns are non-θ-markers, which in turn explains unexpected properties related to the distribution of arguments in Japanese nominals. It is also shown that the proposed approach provides a simple account of the properties exhibited by lexical and null arguments in Japanese clauses that would otherwise remain mysterious.