Head-Internal Relatives in Japanese as Rich Context-Setters

Head-internal Relatives in Japanese as Rich Context-Setters Tohru Seraku St. Catherine’s College Manor Road, Oxford, UK OX1 3UJ tohru.seraku@scatz.ox.ac.uk Abstract Head-Internal Relatives (HIRs) in Japanese are regarded as rich context-setters within Dynamic Syntax (DS): the propositional tree of the HIR clause is mapped onto a ‘partial’ tree, which establishes a rich context for the embedding clause to be parsed. This partial tree contains a situation node decorated with the Relevancy restriction and a node for an internal head. This account handles some new data and makes a novel prediction. Further, it is shown that the past DS analysis of HIRs in fact models change relatives (but not HIRs). 1 Introduction Japanese displays so-called HIRs (Head-Internal Relatives), where the relative clause lacks a gap, the head is found inside the relative clause, and the relative clause ends with the particle no. (1) [Ringo-ga tsukue-no-ue-ni [apple-NOM table-GEN-top-at oite-atta no]-o Kiki-ga tabeta. place-existed NO]-ACC K-NOM ate ‘An apple was on a table and Kiki ate it.’ This paper addresses Japanese HIRs in Dynamic Syntax (DS; Cann et al., 2005; Kempson et al., 2001). Sect. 2 surveys previous studies. Sect. 3 introduces DS. Sect. 4 argues that the past DS account of no (Cann et al., 2005) fails to capture the non-nominality of HIRs. Sect. 5 presents an alternative DS account. Sect. 6 argues that the past DS account of no models change relatives (but not HIRs). Sect. 7 concludes the paper. 2 Previous Studies Several papers collected in Kuroda (1992) as a point of departure, the Japanese HIR has been extensively explored (Kitagawa, 2005; Kuroda, 2005; see references therein). Two approaches stand out. First, some scholars note parallelisms between HIRs and E-type anaphora and make use of the E-type mechanism for HIRs (Hoshi, 1995; Kim, 2007, 2008a/b, 2009; Matsuda, 2002; Shimoyama, 1999, 2001). The most advanced work in this camp is Kim’s analysis. Second, others postulate the null functional head ChR (Choose Role) as a sister to VP, and assume that ChR picks out the internal head by choosing a salient thematic role in the eventuality denoted by VP (Grosu, 2010; Grosu & Landman, 2012). Kim’s E-type analysis and the ChR analysis are the two most influential accounts of HIRs in the literature, but they seem need revisions. First, it is widely held that the head in the HIR denotes a maximal set of individuals that satisfy the HIR clause description (Hoshi, 1995). For instance, for (1) to be felicitous, the situation must be the one where Kiki ate all of the apples on the table. But maximality effects are shown to be derived pragmatically. Thus, for (2) to be felicitous, a situation must be the one where each passenger puts no more than one ticket in the checker, even though he has multiple tickets, provided our world knowledge that the insertion of multiple tickles may cause malfunction of the checker (Kubota & Smith, 2007: 154). (2) Dono-zyookyakui-mo [ei saifu-ni every-passenger-too [ wallet-in kaisuuken-ga haitteita no]-o coupon.ticket-NOM was.present NO]-ACC toridashite kaisatsu-ni ireta. pick.up …

Content truncated to 500 words