Introduction

This chapter deals with the question of how conversationalists use prosody to initiate and sustain verbal encounters. ‘Prosody’ here includes: (a) intonation, i.e. pitch levels on individual syllables and their combination into contours; (b) changes in loudness; (c) stress, a perceptual feature generally comprising variations in pitch, loudness and duration; (d) other variations in vowel length; (e) phrasing, including utterance chunking by pausing, accelerations and decelerations within and across utterance chunks; and (f) overall shifts in speech register. These are conceptual conflations of variations in the three basic phonological dimensions of frequency, amplitude and duration.

Prosodic phenomena have been studied from a variety of perspectives. They have been examined as elements of syntactic and lexical (Bresnan 1971, Berman & Szamosi 1972) as well as pragmatic competence (Bolinger 1972; Brazil & Coulthard 1980). In spite of many basic disagreements, linguists and phoneticians have discovered a great deal about the conventions of English prosodic usage, and about the nature of the semantic information conveyed by prosody. This work, which forms the basis of this analysis of prosody in conversation, will be discussed below. However, my approach to the semantics of prosody will be quite different. The question raised here is one which has hitherto received little direct attention: What sorts of information do speakers in fact rely on prosody to provide in verbal exchanges?

In conversations, we must continually make judgements at simultaneous levels of meaning, through an inferential process which both interprets what has been said and generates expectations about what is to come. The process is always situated or context bound.