Partial Case-Marking in Japanese Stripping/Sluicing: A Dynamic Syntax Account

PACLIC 28 !93 Partial Case-Marking in Japanese Stripping/Sluicing: A Dynamic Syntax Account Tohru Seraku Hankuk University of Foreign Studies 81, Oedae-ro, Cheoin-gu, Yongin-si, Gyeonggi-do 449-791, Korea seraku@hufs.ac.kr Abstract This article presents novel data on partial case- marking in Japanese stripping/sluicing: only the final NP in multiple stripping/sluicing may lack a case particle. These data challenge previous works that assign radically distinct structures to stripping/sluicing depending on whether or not case-marking is involved. These case-marking patterns are reducible to incremental growth of semantic representation, formalised in Dynamic Syntax: each NP is parsed at an ‘unfixed’ node, and this structural uncertainty must be resolved before another unfixed node is introduced. 1 Introduction There is a growing body of research on ellipsis in Japanese (Hiraiwa & Ishihara 2012 and references therein). Stripping is a relatively understudied type of elliptical construction (Fukaya 2007, Fukaya & Hoji 2003, Fukui & Sakai 2003, Sakai 2000; see also Hankamer & Sag 1976). As shown in (1)B, stripping consists of the NP Mary and the copula da, where case-marking of Mary is optional. (1) A: Tom-ga ringo-o tabe-ta-yo. T-NOM apple-ACC eat-PAST-SFP ‘Tom ate apples.’ B: Iya, Mary(-ga) da. no M(-NOM) COP ‘No, Mary.’ (= ‘No, Mary ate apples.’) Japanese also allows “multiple stripping.” That is, the pre-copula part may involve more than one NP: (2) A: Tom-ga ringo-o tabe-ta-yo. T-NOM apple-ACC eat-PAST-SFP ‘Tom ate apples.’ B: Iya, Mary-ga nashi-o da. no M-NOM pear-ACC COP ‘No, Mary, pears.’ (= ‘No, Mary ate pears.’) The most elaborated analysis of stripping is found in Fukaya (2007), the main claim being that case- marked and case-less stripping must be structurally distinguished. According to Fukaya, movement is relevant only to case-marked stripping.1 What has not been noted in previous studies is that when there are multiple NPs in stripping, only the final NP may be case-less (see Section 4 for details). Compare (2)B with (3)B, where the final NP nashi (= ‘pear’) may be case-less, but not the non-final NP Mary.2 (3) A: Tom-ga ringo-o tabe-ta-yo. T-NOM apple-ACC eat-PAST-SFP ‘Tom ate apples.’ B: Iya, Mary*(-ga) nashi da. no M(-NOM) pear COP ‘No, Mary, pears.’ (= ‘No, Mary ate pears.’) 1 This non-uniform analysis is based on the observation that only case-marked stripping is sensitive to “islands” (Fukaya 2007). Seraku (2013) shows that our account captures the island-(in)sensitivity patterns of stripping by means of the ‘LINK’ mechanism (Cann et al. 2005). 2 For some speakers, acceptability slightly drops with the string Mary-ga nashi da, but what is essential is that it is much more acceptable than the string Mary nashi-o da and the string Mary nashi da. The same type of remark also applies to the data in Sections 4 and 5. Copyright 2014 by Tohru Seraku 28th Pacific Asia Conference on Language, Information and Computation pages 93–100 PACLIC 28 !94 This partial case-marking phenomenon raises two problems for previous works. First, (3)B manifests case-marked and case-less stripping at the same time; that is, the single string contains the case- marked NP Mary-ga and …

Content truncated to 500 words