Japanese Pseudo-NPI Dare-mo as an “Unrestricted” Universal Quantifier

Copyright 2012 by Katsuhiko Yabushita 26th Pacific Asia Conference on Language,Information and Computation pages 436–445 Japanese Pseudo-NPI Dare-mo as an “Unrestricted” Universal Quantifier Katsuhiko Yabushita Department of English Naruto University of Education Naruto, Tokushima 772-8502, Japan yabuchan@naruto-u.ac.jp Abstract Japanese dare-mo has been widely acknowledged to be an NPI, furthermore, a “strict” NPI in the sense of Giannakidou (2011) as it seems to be licensed only in an “antiveridical” environment, specifically, with a clausemate negation. However, there is a type of positive sentences in which dare-mo can appear, i.e. non-episodic sentences, which indicates that dare-mo is in fact not an NPI and its NPI-like distribution is an epiphenomenon due to dare-mo’s lexical meaning and the resulting interpretational properties of dare-mo sentences. In the current work, based on novel data we will propose that dare-mo is an “unrestricted” universal quantifier and demonstrate that the proposed meaning of dare-mo and a reasonable assumption about episodic predicates predict that positive episodic dare-mo sentences will be contradictory while negative episodic ones and non- episodic ones, positive or negative will be contingent, nicely characterizing the grammaticality facts of dare-mo sentences. 1 Introduction of Japanese Dare-mo in Question (“NPI” Dare-mo) in Contrast to Dáre-mo (“Non-NPI” Dare-mo) In this section the Japanese expression in question, dare-mo will be introduced in terms of its morphological, phonological, and preliminary semantic features. 1.1 Morphological Features Dare-mo is morphologically composed of indefinite pronoun dare ‘who’ and particle mo, which has been sometimes glossed as ‘also’ and other times as ‘even’.1 1.2 Phonological Features As will be seen in the next subsection, there is another dare-mo distinct from dare-mo in question here syntactically and semantically. In correlation with the syntactic and semantic differences, there is a phonetic and phonological difference between them at least in Tokyo Japanese. In Japanese, a pitch accent language, the placement of accent induces difference in meaning, as is illustrated in (1): 1 In this paper the issue is not addressed whether mo in question is polysemous, there is a unique meaning, or there are two distinct mo’s, in which case, which one is relevant here. In any case, the compositional analysis of the meaning of dare-mo out of that of dare and that of mo will not be dealt with here; thus, throughout this paper, mo will be glossed rather ambiguously simply as ‘MO’. 436 (1) (Adopted from Haraguchi 1995: (6)) Nouns Glosses Placement of Accent a. káki (-ga) ‘oyster’+Nom initial-accented H L L b. kakí (-ga) ‘fence’+Nom final-accented L H L c. kaki (-ga) ‘persimmon’+Nom unaccented L H H In Tokyo Japanese, the accent is placed on the mora before the pitch drop; in other words, the accent is on the H immediately before L. (Haraguchi 1999: 5) As dare-mo in question has the same tone melody as (1c), as is shown in (2), it is considered to be an unaccented word. (2) The tone melody of “NPI” dare-mo dare-mo L H H On the other hand, the near-homonymous, “non- NPI” dare-mo has the tone melody as …

Content truncated to 500 words