ABSTRACT
Recent studies of conversation from a variety of linguistic, psychological, anthropological, and sociological perspectives, have shed light upon a number of issues important to the study of conversational inference. It is generally agreed that grammatical knowledge is only one of several factors in the interpretation process. Aside from physical setting, participants’ personal background knowledge, and their attitudes toward each other, sociocultural assumptions concerning role and status relationships, as well as social values associated with various message components, also play an important role. So far, however, treatment of such contextual factors has been primarily descriptive. The procedure has been to identify or list what can potentially affect interpretation. With rare exceptions, there have
3 4 5 6222 7 8 9 10111 1 2 3 4 15111 6 7 8 9 20111 1 2 3 4 5 6 7 8 9 30111 1 2 3 4 5 6 7 8 9 40111 1 2 3 4 5
‘
attempts to show how social knowledge is used in situated intercourse of interaction, often without a change in extralinguistic context. Therefore, the social input to conversation is not entirely constant. Assumptions about role and status relationships vary as the conversation progresses, and these changes are signalled through speech itself (Gumperz and Cook-Gumperz 1976). The signals by which this is accomplised can be regarded as a metalanguage or a meta-signalling system. So far, however, we know very little about this metalanguage. In this paper I want to suggest at least the outlines of a theory which deals with the question of how social knowledge is stored in the mind, how it is retrieved from memory, and how it is integrated with grammatical knowledge in the act of conversing.
