1. Introduction

The research described in this paper forms part of a larger project which investigates the discourse structure of email discussions, and also investigates the suitability of Conversation Analysis (CA) developed from the work of Sacks, Schegloff and Jefferson (1974), and Discourse Analysis (DA) developed from the work of Sinclair and Coulthard (1975) for the analysis of email discussion data. Since DA in general does not concern itself with repair, the theoretical framework for the investigation of repair is taken from CA.

Although CA was originally used for the investigation of spoken conversation, Sacks et al note that conversation is one of many "speech-exchange systems", others of which include "debates, meetings, press conferences, seminars..." (Sacks et al, 1974, p.729), and that the turn-taking system will vary from one speech exchange system to another. They saw conversation as a standard for comparison with other types of spoken interaction, and indeed CA has been used in this way, in particular for the comparison of various types of institutional talk with the "benchmark" of casual conversation (see for example papers reporting research on a wide range of institutional talk in Drew and Heritage (1992)). Hutchby and Wooffitt (1998, p.148) writing about CA argue that "The significance of this approach is that it succeeds in revealing what is distinctive about interaction in different types of environment". Communicative environments now include virtual environments, so an appropriate extension of the use of CA might therefore be to investigate interaction mediated by computer. The issue is a current one, and a prominent conversation analyst has recently called for the use of CA techniques to analyse Computer-Mediated Communication:

Now that an increasing number of people spend various amounts of their time 'online', chatting with friends or whoever is available, it is time to study Computer-Mediated Conversation (CMC), as we previously studied face-to-face conversation and Telephone (Mediated) Conversation, using the same procedural perspective (ten Have (2000, p.1)).

A central concern of CA is that of repair, and Sacks observes that the conversational system contains within itself the mechanisms for dealing with conversational problems:

The business of remedial exchanges is to handle problems of local order in conversation: failures to understand, failures to hear, interruptions, silences, more than one person starting up at the same time, etc. It turns out that the means for remedying local problems of order in conversation are adjacency pairs. (Sacks (1995), Vol 2, p.525)

This paper investigates the use of repair in email discussions, and, since common sources of trouble in email discussions are turns which are considered to be in some way inappropriate, it investigates how participants deal with inappropriate turns.