Relevance: How Discourse Coherence Is Constructed

This article sets out to investigate discourse coherence in terms of relevance. The hypothesis underlying the present research is that coherence is not an intrinsic property of discourse as a linguistic unit, but is constructed as a consequence of cooperation between the communication participants in the dynamic process of interaction.Discourse coherence has long been a focus of concern in the field of discourse analysis since it has to do with every aspect about how and why a passage can be considered as a discourse. Previous research on coherence, both semantic and pragmatic, has contributed greatly to our understanding of this phenomenon. But that is not the whole story in that it has paid inadequate attention to the cognitive feature of coherence. This article is an attempt to explore the process of constructing coherence from a pragmatic and cognitive perspective. It maintains that discourse is a verbal record of a certain communicative event. In order to see it more clearly, coherence should be analyzed by placing it in the dynamic process of interaction. On the one hand, all the relevant factors, both verbal and non-verbal, and their influence upon the construction of coherence could be specified. On the other hand, we could see the constant efforts made by the participants in their cooperative contribution to coherence.Theoretical foundation comes from Relevance Theory, a cognitive pragmatics proposed by Sperber&Wilson(1986/1995). According to Relevance Theory, human communication is a certain kind of cognition, which is relevance-oriented. That is to say, we tend to pay attention to the information that seems most relevant to us. The fundamental assumption is that cognition tends to maximize effects with minimum amount of efforts. Unsatisfied with the two existing communication models, the code model and inference model, they formulate a new one, termed as Ostensive-lnferential Communication, which is used as the basic theoretical framework for the present research. The major part of the analysis contains two sections. One is concerned with ostension and the other has to do with inference.In the section of ostension, our concern is with the contributions made by the addresser tothe construction of coherence. It is assumed by Sperber& Wilson that each ostensive behavior carries an automatic guarantee of relevance, three types of which are distinguished here: semantic relevance, topical relevance, and contextual relevance. They contribute to coherence in different ways: semantic relevance enables an utterance produced by ostensive act to connect with others linearly; topical relevance unites all utterances sharing the same topic into a whole; and contextual relevance helps discourse fulfill the functions of language in a certain context by integrating it properly with it.The inference in the interpretation of an utterance is proposed to be a search for relevance created by ostensive behavior. Relevance is defined in terms of contextual effects, which are result of the integration of new information and old.information. Conclusions derived would be used as contextual assumptions in the next inference. It is in this way that each utterance is connected with those preceding it and a coherent discourse is finally constructed.