Volume 24, Issue 1

Following the spirit, but not the same mechanism, of Kayne’s (1998, 2000) overt syntactic derivational analysis of focus-related elements such as only, also/too, and even in English, this paper explores the nature of focus feature organization in narrow syntax (NS) with special reference to the additive use of the focus particle mo ‘also’ in Japanese. First, it is argued that, in spite of their attractiveness, at least the versions of LF focus particle movement analysis of association with focus entertained in Aoyagi (1998, 1999, 2006) are faced with theoretical/empirical problems. Then, it is claimed that association with focus in Japanese should be best captured in overt syntax without recourse to any LF movement operations, while keeping to Aoyagi’s (1998, 1999, 2006) original insight of taking association with focus as “focus feature agreement or sharing.” In so doing, a new hypothesis for focus feature organization in NS (Focus Feature-Splitting Hypothesis) is proposed, which dictates that in principle an interpretable focus feature [focus] and an uninterpretable focus feature [uFoc] in the domain of the Foc head occupied by mo can be separated under some “locality condition” based on phases. To the extent that the proposed overt syntactic derivational approach is on the right track, it provides another empirical support for the single-cycle computational system for NS in the faculty of language (FL), entertained in Kayne (1998, 2000), Epstein et al. (1998), Chomsky (2001b, 2004), and Epstein and Seely (2006) inter alia.

It has been noted that the variation in syntactic positioning of quantity expressions in Japanese - the pre-nominal construction and the NP-external construction - reflects different referential domains of the quantified entities (Martin 1975, Inoue 1978, Okutsu 1986, Downing 1996). The distinction between definiteness/indefiniteness as well as the notion of specificity have been used to explain the behavioral patterns of quantity expressions. However, due to the lack of a clearly defined framework that illuminates the relationship between the two separate referential categories of definiteness/indefiniteness and specificity/non-specificity, previous studies could not offer a coherent account of why the two constructions sometimes yield contrasting definite and indefinite readings, sometimes become synonymous, and sometimes allow slight differences in nuance. Using a framework that illuminates the relationship between the two referential categories, this study re-analyzes the observations made in previous studies and ascertains the referential characteristics of the two quantifier constructions. It suggests that the pre-nominal construction and NP-external construction do not express opposing domains of one referential category, but instead express one of the domains of two independent categories of referentiality, i.e. the domains of specificity and indefiniteness respectively. This study also discerns the referential meanings expressed by the constructions themselves from contextually derived meanings.

This paper argues for verb movement in Japanese. In Taro-ga sushi-o mo tabe-ta 'Taro also ate sushi,' for example, we claim that tabe 'eat' actually moves past mo 'also,' as in [Taro-ga sushi-o t tabe mo tabe-ta]. This analysis is supported by the interpretation of the adjunct clitic mo. We further claim that the verb-movement operation, coupled with the so-called Morphological Merger, successfully accounts for various data involving su-support (analogous to do-support in English). As a consequence, the present analysis has an important implication for the language typology on verb movement.