Retour vers liste

Détail de la contribution

Auteur: Osamu SAWADA

Co-Auteur(s): Jun SAWADA, Aoyama Gakuin University, Japan

The modal demonstratives in Japanese: a mismatch between at-issue and non-at-issue meanings

Abstract/Résumé: This paper investigates the meaning of the Japanese ‘modal’ demonstratives as exemplified in (1): (1) Ano Federer-ga make-ta. That Federer-NOM lose-PAST At-issue: Federer lost. Implicature: Federer is unlikely to lose. The demonstrative ano in (1) is different from the spatial demonstrative ano in that the former, but not the latter, induces an unlikelihood implicature. In (1) the speaker is not physically pointing to Federer; instead, he/she is construing Federer as being highly unlikely to lose. The important point is that the modal demonstrative in (1) is situated inside the domain of DP, but its meaning has to do with an entire proposition. In this paper we will investigate the meaning/use of modal demonstrative and propose a mechanism that can capture its meaning in a compositional way. The meaning of modal demonstratives is a conventional implicature (CI)(Grice 1975; Potts 2005). This is corroborated by the fact that we can determine the truth value of a given proposition without a modal demonstrative, which is not the case with spatial demonstratives. We consider that the meaning of modal demonstratives is not a presupposition. Unlike presuppositions, the meaning of modal demonstratives can project under a presupposition plug: (2) Taro-wa ano Ziro-ga katu-to sinzi-teiru. Taro-TOP that Ziro-NOM win-that believe-TEIRU ‘Taro believes that that Ziro will win.’ The CI meaning that ‘Ziro is unlikely to win’ is anchored to the speaker. We propose the following denotation for the modal ano: (3) [[ anoMOD]]=λxλP.unlikely(P(x))>!STAND (where x = a speaker’s image of x.) The modal ano in (3) conventionally implicates that the degree of unlikelihood of P(x) much is greater than a contextual norm by a large amount (i.e. P(x) is highly unlikely to be true). The above analysis, however, poses a serious problem for the compositional relationship between an at-issue meaning and a CI. For example, in (1) the predicate make, ‘lose’ takes Federer as an argument in the at-issue dimension. However, in the CI dimension the predicate is an argument for ano (Federer). In order to solve this ‘mismatch’ we propose the revised denotation for the modal ano as in: (4) [[ anoMOD]]=λxλP.P(x)♦λxλP.unlikely(P(x))>!STAND (where x = a speaker’s image of x.) The modal ano in (4) behaves as an ‘overt’ typing shifting (lifting) operator in the at-issue domain (the left side of ♦) in addition to having a CI meaning. The implication of this analysis is that if there is a mismatch between an at-issue dimension and a CI dimension in an argument-predicate structure, the former adjusts to the latter. This can be another fundamental difference between an at-issue meaning and a CI.