Topicalization and Truth Conditions: A Categorial Account

Topicalization and Truth Conditions: A Categorial Grammar Account∗ Hiroaki Nakamura a aDepartment of Liberal Arts, 5-1, Wakaba-cho, Kure City, Hiroshima, Japan hiroaki-nakamura@ax2.mopera.ne.jp Abstract. In this paper we will consider how the choice between the topic marker wa and nominative marker for subjects affects truth conditions of sentences. We derive the proper interpretations for sentences with topics and nominative subjects in terms of syntax- semantics interface. The categorial framework adopted here allows us to account for significant difference in meaning with respect to the choice of markers for subjects while maintaining the principle of strong compositionality. Keywords: information structure, focus, topic, exhaustiveness, contrastiveness, syntax- semantics interface. ∗ Variants of this material has been presented in part at several conferences including the past PACLIC. I'd like to thank the audiences at those presentations. I also indebted to two anonymous reviewers for helpful comments and discussion. All errors, of course, are my responsibility. Copyright 2010 by Hiroaki Nakamura. 1 Introduction In Japanese linguistics, contrastive analyses of the topic and nominative markers has long received a great deal of attention, especially in terms of semantic or pragmatic notions like giveness/aboutness/contrastiveness/exhaustiveness, etc, whereas it seems that the truth conditional information they convey have been the subject of much less study. Though this paper examines truth conditional aspects of the use of the topic marker WA and nominative marker GA, I do not agree with the parallel treatment of WA and GA due to the fact that there are significant differences in syntax/semantics between these particles. Before presenting my analysis, let us review some properties of these markers. It has been agreed that sentences expressing categorical judgment (mostly, individual or kind-level sentences) strongly tend to have their subjects marked with WA (see papers in Kuroda 2003, among others), as illustrated in (1): (1) Saikin-no nihonjin-wa/*?nihonjin-ga se-ga takai. Recent-Gen Japanese-Top /Japanese-Nom height-Nom high. 'Recent Japanese are tall.' As often pointed out in the literature, WA marking is usually permitted only in matrix clauses (often called the root phenomena, see Heycock 2008), and even in matrix sentences expressing enduring properties of the subjects, question words answers corresponding to them can never be marked with WA (the complex form of which + N can be followed by WA, with contrastive connotations). (2) a. Dare-ga/*Dare-wa se-ga takai-no? Who-Nom/Who-Top height-Nom tall-Q 'Who is tall?' PACLIC 24 Proceedings 251 b. Tanaka-san-ga/*Tanaka-san-wa takai-desu. Tanaka-Mr.-Nom/Takana-Mr.-Top tall-Pres. (2b) can be acceptable if the topic marked subject is interpreted as a contrastive topic. Since Kuno (1973), it is also assumed that WA marks constituents whose referent is already present in the common ground (in other words, given, familiar, recoverable, etc.), but this generalization does not seem correct. New referents can be introduced into discourse marked with WA, as in (3), and the point is that the subject cannot be replaced with the nominative case. (3) a. Syaku Esyo-wa/*ga Enkou-ji-no syamon nari-ki. Houshi, ikeri-si toki, ... Priest Esyo-Top/Nom Enko-temple-Gen priest be-Past Priest alive-Be when 'Priest Esyo was a monk of Enko temple. When he was alive, he …

Content truncated to 500 words