Introduction

The research question which this chapter addresses is motivated by the growing awareness that speakers who are engaged in verbal interaction with one another employ adaptive strategies which entail ‘matching’ their speech behaviour in one way or another to that of their interlocutor. On the verbal level this matching may involve repetition of words, expressions or whole utterances, e.g. to contextualize affiliation or support (Tannen 1987, 1989). On the non-verbal level we may find rhythmic matching, as, for example, in English conversation, where adjusting one's rhythm and tempo to that of one's partner at turn transitions counts as a well-timed entry to the floor (Couper-Kuhlen 1993, Auer, Couper-Kuhlen and Müller (to appear)); or we may find melodic matching, as for instance in ‘wheel-spinning’ (Chafe 1988:7), where successive interlocutor turns are occupied with saying the same thing (albeit with different words) and intonationally echoing one another.

The general question with respect to verbal and non-verbal matching of this kind is: when does one speaker's repetition of the words and/or prosody of another become one speaker's mimicry of another? Under what conditions does matching speech behaviour switch from being something that interlocutors do together, to something that one interlocutor does to the other? Related to this is a question concerning the ‘rules of mimicry’ (Goffman 1974:537). Goffman has pointed out that, in quoting a person, we quite naturally ‘quote’ the overlay of accent and gesture as well. However, at some point – for instance, if a male speaker quotes a female speaker and ‘too much’ of the gender expression is taken over – the quoter becomes ‘suspect’, a mimic with presumably disaffiliatory intentions (1974:539).