A Type-Logical Approach to Potential Consructions in Japanese

31st Pacific Asia Conference on Language, Information and Computation (PACLIC 31), pages 42–49 Cebu City, Philippines, November 16-18, 2017 Copyright c⃝2017 Hiroaki Nakamura A Type-Logical Approach to Japanese Potential Constructions Hiroaki Nakamura Institute for Promotion of Higher Education Hirosaki University 1, Bunkyo-cho, Hirosaki, Aomori Prefecture 036-8560, JAPAN hiroaki@hirosaki-u.ac.jp Abstract Potential constructions have long attracted much attention in Japanese Linguistics, mainly focusing on the case alternation of object NPs. I will point out some important characteristics of the constructions they have missed and propose a completely new analysis from a view point of logical grammar. First, we show significant differences between potential and passive sentences which have been assumed to been projected from one and the same suffix –rare ‘can’. I suggest that these two uses must be distinguished at least in contemporary Japanese. Our type-logical approach to unbounded dependencies has an empirical coverage broader than traditional and generative grammatical approaches and can explain the fact that various arguments including adjuncts can be marked with nominative. We also examine interesting interactions of case alternation with scope alternation. 1 Introduction Potential constructions have long attracted much attention in Japanese Linguistics, mainly focusing on their meanings and case-alternation phenomena. I argue in this paper that the past studies have failed to describe their important characteristics in significant ways and propose a completely new analysis from a formal grammar view point. To show what is wrong with the past analyses, let us observe the points Japanese traditional linguistics have assumed, and show why the potential suffix rare must be distinguished from the passive rare, and then propose an analysis which can properly deal with a broad empirical coverage. Observe the standard active and passive pair in Japanese in (1). (1) a. Hitobito-wa sakuban takusan-no People-Top last-night a lot of banana-o tabeta. bananas-Acc ate. 'People ate a lot of bananas last night.' b. Takusan-no banana-ga sakuban A lot of bananas last night hitobito-niyotte taber-are-ta. people by eat-Pass-Past 'A lot of bananas were eaten by people last night.' Sentences in (1) show a typical active-passive correspondence where the passive suffix -rare is used to form the passive complex verb taber-are-ta ‘were eaten,’ the theme argument banana is subjectivized and the agent argument is demoted to the adjunct marked with oblique case. In Japanese linguistics, it has been assumed that the same suffix -rare is also used to form the potential verbs and that the distinction in interpretation between passives and potentials is dependent on contexts. It is also suggested that complex potential verbs project active or passive potential sentences and the distinctions were made depending on surface case markings of arguments, as exemplified in (2) (see Teramura 1982 for discussion on this dichotomy): (2) a. Kodomo-ga kono banana-o taber-are-u. Children-Nom this banana-Acc eat-can-Pres 42 'The child can eat this banana.' (active) b. Kono banana-ga mou taber-are-ru. This banana-Nom already eat-Can-Pres ‘This banana can be eaten now.’ (passive) Teramura (1982) and his followers call sentences like (2a) ‘active potentials’ and those like (2b) passive potentials. This dichotomy …

Content truncated to 500 words