Utterance Units and Exchanges in Spontaneous Japanese Dialogue

UTTERANCE UNITS AND EXCHANGES IN SPONTANEOUS JAPANESE DIALOGUE Masahito KAWAMORI Nippon Telegraph and Telephone, NTT Basic Research Laboratories kawamoriOatom.brl.ntt.co.jp Akira SHIMAZU Japan Advanced Institute of Science and Technology shimazu0 j aist . ac . j p ABSTRACT In dialogue, speakers are engaged in various types of speech acts other than just those fre- quently encountered such as question, command, promise. The speaker often tries to make sure that his interlocutor has received the information that he is conveying, that his intentions are understood, and that his attitudes toward his interlocutor and the general surrounding environ- ment are clear. Especially in spontaneously spoken discourse or dialogue, these various speech acts are not necessarily expressed in full-fledged well-formed sentences; they are often couched in rather fragmentary, truncated semi-sentences or phrases. In an analoguous manner to Parsons' subatomic event structures, we call these speech acts corresponding to non-full-fledged sentences SUBATOMIC SPEECH ACTS. We show in this paper how these subatomic speech acts are used to form mutual beliefs and to coordinate discourse. We use examples from actual Japanese spoken dialogue to demonstrate that our theory accounts for some of the characteristic phenomena in discourse and assigns correct interpretations and explanations to these phenomena. 1 INTRODUCTION It is often said that ordinary conversations are full of irregularities. Some say that there is no `grammar' in every day spoken language. In theoretical linguistics, such irregularities in every day conversation has been considered due to various limitations on the performance language use. These phenomena have been regarded as rather peripheral and not a central concern of theoretical linguis- tics, at least that of a grammatical theory. Such an attitude, however, has made it rather difficult to apply the scientific results of theoretical linguistics to the language in actual use. Grammatical theories are particularly vulnerable to. the vulgarity of conversational language because many people speak in incomplete fragmentary sentences. But the fact still remains that people do understand everyday language and that they usually do not find their conversational language anyway more difficult to understand than its written counterpart; on the contrary, it is usually the case that people find the written language spoken to them harder to interpret than spontaneously spoken casual language. So there seems to be a sort of paradox; everyday conversational language seems to lack proper 'grammar' and regularity enjoyed by its written counterpart but is not particularly difficult to understand, while its written counterpart with a proper grammar and regular structure is usually harder, when spoken, to understand. We believe that the paradox is only superficial. The paradox ensues if one assumes that since conversational language does not always follow the rules of its written counterpart, it lacks rules and grammar. If one accepts that conversational language has its own grammar and rules, somewhat different from those of its written counterpart, then there will be no paradox involved. In other words, what we need is a theory of performance, a grammar of language in actual use. We will in this paper explore …

Content truncated to 500 words