Representing Events in Japanese Complex Predicates

Representing Events in Japanese Complex Predicates1 Yoshie Yamamori Kobe University yy080707@gmail.com Abstract. This paper presents an event semantics-based investigation of Japanese complex predicates of the form V1+V2, to clarify their syntactic and semantic constitution. The Japanese V1+V2 formation primarily exploits the aspectual modification of a core event; of particular interest here are complex predicates composed of intentional active verbs with te-form + the existential verb a-ru; this complex pattern is known to be ambiguous between what is usually termed a ‘resultative’ and a ‘preparative’ reading. One major problem of the “V-te-a-ru” construction is retrieving the semantic relation linking their components, which is (partially) implicit on the surface. To date, no uniform analysis has been developed to formally account for these facts. The main goal of this study was to develop a uniform analysis of their meaning representations, using Davidson’s (1967) event semantics as a reference model. Basic data revealed some interesting findings, indicating some asymmetry in event structure. This could help identify the semantic relations of components and clarify the semantic representation of a given syntactic paradigm. Keywords: event structure, complex predicate, Japanese, experiencer-agent alternation. 1. Introduction When children acquire the meaning of a verb, they understand the possible state or event it describes. However, many verbs also come with arguments. Do we know what arguments a verb accepts when we understand the verb’s meaning? It is still not clear how arguments are matched with the right kind of verbs; neither are the possible repercussions from theories about argument structured. Clarification of these issues will require thinking about the meanings of verbs in non-traditional ways. For example, verb denotations might not carry all the information about a verb’s argument structure. This study investigated Japanese complex predicates of the kind “V-te-a-ru”, composed of an intentional active verb2 in te-form3 and an existential verb aru, which is taken to be either a 1 This work was supported by a Grant-in-Aid for Scientific Research (C) (21520403). 2 The term “intentional active verbs” refers to verbs that have a control relation with their argument. These may primarily involve transitive verbs, but (unergative) intransitive verbs too. A sentence that includes the intentional verb read in (i) can be paraphrased by (ii). (i) John read a book. (ii) What John did is to read a book. In contrast, “non-intentional active verbs” can not co-occur with “V-te-a-ru”as in (iii): PACLIC 24 Proceedings 311 resultative or preparative predicate (this distinction has been discussed previously by Japanese linguists with regard to agentive argument realization: see, e.g., Masuoka, 1987 and Kinsui, 2009, among others, but these scholars have not provided an appropriate description of the event or argument structures of “V-te-a-ru” constructions). Intuitively, intentional verbs appear to specify as part of their meaning an action carried out by an implicit agent, while the existential verb a-ru specifies the ‘coming about’ of a result state. They instantiate well-formed complex patterns, which are not totally predictable with regard to regular syntactic constructions, or are ambiguous between what is usually termed a ‘resultative’ …

Content truncated to 500 words