This essay will examine the notion of ‘Conversational analysis’ and its contributions to explaining the dynamics of social interaction, then cite possible pathways for discourse in the future. Conversation is the main means by which humans communicate, and is thus vital for full and rich social interaction. An obvious definition of conversation is a process of talking where at least two participants freely alternate in speaking. Certain methods of communication would not fall into this, though – such as a prisoner being questioned by a barrister in a law court, or a tutor lecturing to his eager Cognitive Science students. However, the analysis of conversation is not a simple matter. One of the most fascinating things about it is that it occurs naturally – so to study it in the artificial conditions of a psychologist’s laboratory may repress its spontaneity and devalue the results. Conversational analysis has thus been taken up by pioneering sociologists known as ethnomethodologists. Ethnomethodology was a sociological and pragmatic backlash to the quantitative methods surrounding the area which were seen to yeild somewhat less than objective results. The ethnomethodological stance, then, is to look, in a qualitative manner, at the dynamics of conversation used by agents as they interact with their social environment. A main feature of conversations is that they tend to follow the convention of turn taking. Simply, this is where one person waits for the other to finish his/her utterance before contributing their own. This is as much a utilitarian convention as mere manners – a conversation, given the aforementioned definition, would logically cease to take place if the agents involved insisted on speaking even when it was plain that the other was trying to contribute. It is, additionally, comforting to know that the other person respects your opinions enough not to continually interrupt you. The best example of this occurs in the Houses of Parliament – a supposed ‘debating’ chamber which is often anything but, due to the failure of the members to observe the turn-taking code. Note, however, that a person rarely explictly states that they have finished their utterance and are now awaiting yours. Intriguing exceptions to this are in two-way radios, where many social and psychological cues are lacking, and thus it is more difficult for speakers to follow turn-taking. A more recent example might be computer-mediated communication. I will examine these two later. The potential for one to reply can be missed, deliberately or not, so that the first person may contribute once more. Failure to realise this can result in an awkward ‘pregnant pause’, or, as I have mentioned, a cacophany of competing voices in a large crowd. Sacks et al (1974, 1978) suggested some regulations of such conversation – a ‘local management system’ where, at certain points, the potential for a change of speakers arises. Called transition relevance places, once again, this is not compulsory. Another fundamental feature of conversation is the idea of adjacency pairs. Posited by Goffman (1976), an example would be found in a question-answer session. Both conversing parties are aware that a response is required to a question; moreover, a particular response to a given question. I might invite a friend into my house and ask: “Would you like a biscuit?” To which the adjacency pair response is expected to be either “Yes” or “No”. My friend may be allergic to chocolate, however, and place an insertion sequence into the response: “Do you have any ginger snaps?” the reply to which would cause him to modify his answer accordingly. This assumes, however, a certain equality in terms of power, or authority. For example, if a lecturer was to call a coffee-break mid-lecture, he might ask: “Would you like to return in ten minutes?” Where he would expect a willing response. He would not expect his students to set conditions or make enquiries of what they can do with the allocated ten minutes. How may such observations be utilised in our social interactions? In the above consideration of turn-taking, if the second agent did not take their opportunity to respond to the first, the implication is that they have nothing to say about the topic. But perhaps the transition relevance place was one in which the second agent was in fact selected, but failed to respond, or responded in an inappropriate manner? A contestant on Mastermind is subject to an ‘enforced’ turn-taking scenario, where non-answer of a question-answer adjacency pair (as opposed to ‘Pass’ or ‘Don’t know’) breaches the expectations of the audience and compare – one might choose to sit down in the spotlight and say nothing for two entire minutes. Alternatively, one could respond to the question “What is the capital of Tibet?” with the nonsensical: “A blue fish”; or even with a sardonic: “‘T’ is the capital of Tibet”. This infinity of responses is what makes language so entertaining, and in the above cases the audience might make inferences about the reasons for the ‘incorrect’ response. In the first case, the individual may not have responded because he did not understand the question, or has ’stage fright’, or any number of other reasons. As Atkinson & Drew (1979) note, such a silence often reveals an unwillingness to answer. Sacks et al. accounts for this in examining preference organisation, where a preferred response is the simplest of a number of options. Dispreferred responses tend to be preceded by a pause, and feature a declination component which is the non-acceptance of the first part of the adjacency pair. Not responding at all to the above question is one such – and has been dubbed an attributable silence, ie, a silence which in fact communicates certain information about the non-speaker. Inferences are also derived in pre-sequence. Here, an utterance is used to prepare for a later one – “Have you finished with that telephone?” could be a simple enquiry, or an indication that the speaker wants to use it after you. The person holding the phone may then say something like – “Give me ten minutes”. This is an interesting example for a number of reasons: (a) The enquirer used a pre-request to check the situation, which influences the future conversation. If the other agent firmly indicates that they will be using the phone for a long time, the pre-request has revealed the conditions of your using the phone soon, to fail. (b) The person on the phone gave a reply to an implicit request to use the phone – there was no explicit statement that the other person wanted to use it. This is based on his knowledge of pre-sequence structure and the inference that the question was more than an idle request about his using the phone. (c) This person then responded in an equally implicit manner that he would be finished quite soon – it is hardly likely that he will time himself to be finished in exactly ten minutes, nor that the other agent will time him rigorously to that effect. The process could continue, in theory in a series of ad infinitum nested layers of implicit meaning and social convention. It has been noted that various physical cues, such as gestures or expressions, are in play during orthodox face-to-face exchanges, and these are obviously lacking in a telephone conversation. Since humans are so adept at speaking over the phone, it is easy to conclude that the cues are not as important as once imagined – we manage without them so well, after all. However, this argument does not take into account the cues one picks up from the voice – it is quite easy to detect if somebody is confident, or nervous on the phone, from the words they use, the pauses, the tone and pronunciations of the words. In short, we may be able to substitute these auditory cues for more conventional physical cues, and then empathise with the other person. This way, we could be visualising, or at least imagining with a fair degree of accuracy, how the person is feeling, and gaining cues that way. It may be more possible to follow turn-taking using a phone than a walkie-talkie, simply because the voices on a phone are so much clearer – these auditory cues cannot work as well over walkie-talkies, necessitating conventions such as saying “Over” when you have finished your utterance. With the promise of widespread computer-mediated communication, such cues are entirely lacking. It is significant that there is a Unix function, ‘write’ which recommends that users follow a similar code to the above walkie-talkie one – typing -o- to indicate the finishing of a statement. More evidence comes from the phenomenon of ‘flame wars’, where users communicate over vast networks in abusive ways, to complete strangers, that they would not dream of in conventional discourse. The lack of cues provides anonymity, and this is a powerful protection indeed. Conversational analysis, then, gives a fascinating insight into the implicit communicative rules which guide our social interactions. It is interesting to speculate how conversation may evolve in the future – with virtual meetings and chatting in ‘cyberspace’ destroying many of the implicit rules of traditional communication, conversational analysts may have much to write about in the future. BIBLIOGRAPHY Atkinson, J. – Order in Court: The Organisation of Verbal Interaction in Judicial & Drew, P. Settings (1979), Macmillan Press. Sacks, H., Schegloff, E & – A Simplest Systematics for the Organisation of Turn-Taking in Jefferson, G. Coversation, (1974) Goffman, E – ‘Replies and Responses’ in Language and Society,(1976), vol 5. Levinson, S. – Pragmatics, (1983), Cambridge Press.