Wa/Ga-Subjects in Japanese and Subdivisions of Tense

Abstract

WA/GA-SUBJECTS IN JAPANESE AND SUBDIVISIONS OF TENSE Shizuka Torii This thesis takes a semantically based tense/aspect approach to the long-standing problem of wa- and ga-markings of 'subjects' in Japanese. It argues for a correlation between wa/ga-markings of 'subjects' and tense/aspect interpretations of clauses, as illustrated in (1) below, to shed light on a new dimension of the problem. (1) a. John-waki-ta. John come-Past 'John came.' <past> b. John-ga ki-ta. John come-Past 'John has just come/arrived.' <'hot news' perfect> The two types of tense/aspect interpretations correlated with wa- and ga-marked subjects are pinned down in terms of (i) two types of 'evaluation time', which are distinguished as 'original' and 'new' (Enç 1987), (ii) two types of R[eference time] (Reichenbach 1947); one that coincides with S[peech time] but not with E[vent time] (R = S/ ≠ E), and the other that coincides with E but not with S (R = E/ ≠ S), and (iii) two types of 'viewpoint aspect' (Smith 1991); one that presents 'part' of a situation manifested at a precise temporal point (View part) and the other that presents 'all' of a situation without decomposing it (View all). In order to provide syntactic mechanisms to account for the correlation between wa/ga-markings of 'subjects' and the two distinct types of tense/aspect interpretations, I propose two subdivisions of Tense in line with Chomsky's (1995: 240) suggestion that Tense might have "further subdivisions and implications about event structure and perhaps other properties". I assume that the two subdivisions of Tense are functional categories making up an articulated tense structure (above VP) and contain distinct semantic features responsible for the distinct tense/aspect interpretations correlated with wa- and ga-markings of subjects in Japanese. Being tense categories, they both have T[ense]-features and D[eterminer]-features to be checked by predicates and subject DPs respectively. Due to the distinct semantic content of the two syntactic categories, depending on which T- and D-features predicates and subjects check, we get two distinct types of tense/aspect interpretations of predicates and two distinct types of subjects (which are morphologically distinguished by wa- and ga-markings in Japanese). In this analysis, the T- and D-features of a tense category ensure that a subject and a predicate are necessarily of the same semantic type. The tense system I propose to account for the wa/ga-phenomena unifies tense and aspect to the extent that the wa/ga-phenomena relate to the interpretation of both tense and aspect. A notable consequence of my analysis is that the syntax and semantics of stage- and individual-level predicates (cf. Carlson 1977, Kratzer 1989 and Diesing 1992) fall under the syntax and semantics of tense. The analysis also exhibits some interesting parallelisms to Davis' (1998), in which person features of subjects are related to a temporal structure. In addition the proposed two subject positions within the articulated tense structure are demonstrated to be tenable across languages. Furthermore I show that the reanalysis is extendable to subordinate clause case markings and interpretations, with special attention to factors such as factivity and the distinctions among propositions, states of affairs, and situation-types.