CONVERSATION ANALYSIS PRELIMINARY ISSUES
In its most basic sense, CA is the study of talk. More particularly, it is 'the systematic analysis of the talk produced in everyday situations of human interaction: talk-in-interaction' (Hutchby and Wooffitt, 1998:13). Conversation as a discourse type has been defined by Cook (1989) in the following way:
It is not primarily necessitated by a practical task.
Any unequal power of participants is partially suspended.
The number of participants is small.
Turns are quite short.
Talk is primarily for the participants and not for an outside audience.
Although the field has adopted the name 'conversation analysis', practitioners do not engage solely in the analysis of everyday conversations. The range of forms of talk-in-interaction that have been the subject to study within CA is far larger than the term 'conversation' alone would imply. Here are some important issues that CA addresses
I. WHAT CA DOES
1. CA is the study of recorded, naturally occurring talk-in-interaction.
That is, the most distinctive methodological trait of CA is that research is based on transcribed tape-recordings of actual interactions. What is recorded is 'naturally occurring' interaction, i.e. the activities that are recorded are situated in the ordinary unfolding of people's lives, as opposed to being set up or pre-arranged. 858j98i
2. CA is only marginally interested in language as such, but first and foremost in language as a practical social accomplishment.
That is, words used in talk are not studied as semantic, syntactic or morphologic units, but as objects used in terms of the activities being negotiated in the talk: as requests, proposals, accusations, complaints, etc.
3. Its object of study is the interactional organization of social activities.
In other words, CA aims at discovering how participants understand and respond to one another in their turns at talk, with a central focus on how sequences of actions are generated. Throughout the course of a conversation or talk-in-interaction, speakers display in the 'next' turns an understanding of what the 'prior' turn was about. That understanding may turn out to be what the prior speaker intended, or not. This is described as next-turn proof procedure and it is the most basic tool used in CA to ensure that analyses explicate the way in which the participants themselves orient to talk, not based on the assumptions of the analyst.
EXAMPLE: Consider the following utterance, which is from an exchange between mother and her son, about a forthcoming parent-teachers' association meeting (Based on Schegloff, 1988):
Mother: Do you know who's going to that meeting?
Mother's question in 1 can be interpreted as doing one of two types of action. It could represent a genuine request for information about who is attending the meeting; or she could be using it as a 'pre-announcement' (a preliminary to some information she wishes to announce about who is going. In the first case the response would be an answer to the question, in the second, the response would provide the opportunity for the news to be announced.
Now look at the whole interaction and comment on how participants display their understanding of what is going on.
1.Mother: Do you know who's going to that meeting?
Rus: Who?
Mother: I don't know!
Oh, probably Mr. Murphy and Dad said Mrs. Timpte an' some of the teachers.
In the next turn, line 2, Russ responds with 'Who?', thereby displaying that his interpretation of Mother's first utterance is as a pre-announcement. But Mother's next turn, 'I don't know', displays that Russ's inference was in fact incorrect.: she was actually asking an information-seeking question. Notice that following this turn, Russ responds with the information his mother was seeking, thereby displaying even more powerfully that he interpreted line 1 as a pre-announcement, because he in fact knows quite a lot about who's going to the meeting. (Example and interpretation from Hutcby and Wooffitt, 1998).
I. ASSUMPTIONS ABOUT TALK
Order at all points
CA
emerged in the pioneering researches of Harvey Sacks into the structural
organization of everyday language use, carried out at the
He observed that, in the majority of the cases, if the person taking the call within the organization started off by giving their name, then the 'suicidal' person who was calling would be likely to give their name in reply. But in one particular call, Sack noticed that the caller (B) seemed to have trouble with the name of the answerer:
A: This is Mr. Smith, may I help you?
B: I can't hear you.
A: This is Mr. Smith.
B: Smith.
Sacks, 1995, Lectures on conversation)
Sacks observed that, for the rest of the conversation, the agent taking the call had great difficulty in getting the caller to give a name. His question then was: 'where in the course of the conversation could you tell that somebody would not give their name'?
Sacks noted that, on the one hand, it appears that if the name is not forthcoming at the start it may prove problematic to get. On the other hand, overt requests for it may be resisted. Then he remarked that it is possible that the caller'' declared problem in hearing is a methodical way of avoiding giving one's name in response to the other's having done so. In his analysis, Sacks shows that by 'not hearing', the caller is able to set up a sequential trajectory in which the agent finds less and less opportunity to establish the caller's name without explicitly asking for it. Thereby, the caller is able to begin the conversation by avoiding giving a name without actually refusing to do so.
The main concern of CA is to show how conversational devices exhibit general features and function in essentially the same ways across varying contexts.
Social activities are observable- ethnomethodology
Working within the same field as Sacks, Garfinkel (1967) developed a form of sociology which became known as ethnomethodology (see also Hutchby&Wooffitt, 1998:30-37). For Garfinkel, members of society are capable of rationally understanding and accounting for their own actions. The aim of the researcher would be then to describe methods that people use for accounting for their own actions and those of others. These are the 'ethno-methods', or the commonsense knowledge that members are using.
One of the problems that ethnomethodology encounters is that of how to gain analytic access to the level of the commonsense knowledge which it seeks to study. Garfinkel's earliest research consisted of 'breaching' experiments in which taken-for-granted routines of ordinary life were intentionally disrupted in order to observe how people dealt with their sudden lack of certainty. For instance, he would instruct his student 'experimenters' to engage others in interaction and then to repeatedly request the subject of the experiment to clarify whatever he or she said. Thus, on being asked 'How are you', the experimenter would ignore the routine or expected use of the question, and respond instead:
S: How are you?
E: How am I in regard to what? My health, my finances, my school work, my peace of mind, my.
S: (Red in the face and suddenly out of control) Look! I was just trying to be polite. Frankly, I don't give a damn how you are.
(Garfinkel, 1967:44)
Basically, by his experiments Garfinkel observed that subjects would find some way of accounting for the 'strange' behaviour of the experimenters. They did this by treating the experimenters as rational agents who had actively chosen to behave in this way, or who had some
underlying reason for doing so. The main points the ethnomethodologists wanted to make were that:
Conversational talk was not incoherent or irregular and was rule-governed
Such rules are 'people's rules', rather than linguists'.
II. BASIC NOTIONS
Turn-taking mechanism:
The starting point is the observation that conversation involves turn-taking and that the end of one speaker's turn and the beginning of the next latch on to each other with almost perfect precision. Overlap of turns (when two or more participants talk at the same time) occurs in about 5% of cases and this suggests that speakers know how, when and where to enter. They signal that one turn has come to an end and another should begin.
The turn-taking model has two components:
a) turn construction units
Turns at talk can be seen as constructed out of units which broadly correspond to linguistic categories such as sentences, clauses, single words (e.g., 'Hey!', 'What ?') or phrases.
Features of turn-construction units
projectability - it is possible for participants to project, in the course of a turn-construction unit, what sort of unit it is and at what point it is likely to end.
Transition relevance place - at the end of each unit there is the possibility for legitimate transition between speakers.
b) Turn distribution (e.g. who dominates the conversation in terms of number of turns taken, length of turns)
There is no strict limit to turn size, given the extendable nature of syntactic turn-constructional units;
There is no exclusion of parties;
The number of parties can change.
The rules operating for turn units (see Sacks, Schegloff, and
a) if C (current speaker) selects N (next speaker) in current turn, then C must stop speaking, and N must speak next.
c) if C does not select N, then any other party self-selects, first speaker gaining rights to the next turn
d) if C has not selected N, and no other party self-selects, then C may (but need not) continue.
Where, despite the rules, overlapping talk occurs, studies revealed the operation of a system:
one speaker drops out rapidly
as soon as one speaker thus 'gets into the clear', he typically recycles precisely the part of the turn obscured by the overlap.
If one speaker does not immediately drop out, there is available a competitive allocation system, whereby the speaker who 'upgrades' most, wins the floor. (uppgrading = increased amplitude, slowing tempo, lengthened vowels, etc.)
EXAMPLE: how do you explain the overlap in the following conversation?
1. Rose: Why don't you come and see me some/times
2. Bea: / I would
like to
4. Rose: I would like you to
(Hutchby and Wooffitt, 1998)
Bea is able here to recognise Rose's utterance as a form of invitation, and to respond to it with an acceptance before it has actually finished (line 2). By starting to talk when she does (not waiting for instance to end of the unit which might have been different, - e.g. sometime this week - triggering a different kind of response), Bea not only projects the end of a particular turn-construction unit, but also displays an understanding of what kind of invitation that unit represents.
Basic turn-type: adjacency pairs and 'preference'
Adjacency pairs
One of the most noticeable things about conversation is that certain classes of utterances conventionally come in pairs.
Example:
Question/answer
Greeting/greeting
Invitation/acceptance(declination)
Offer/acceptance (refusal)
These sequences are called adjacency pairs because, ideally, the two parts should be produced next to each other. The point is that some classes of utterances are conventionally paired such that, on the production of a first pair part, the second becomes relevant and remains so, even if it is not produced in the next turn. The next turn in an adjacency pair 'sequence' is a relevant second pair part. But that need not be the next turn in the series of turns making up some particular conversation.
Example of an insertion sequence: (Levinson 1983)
1.
A: Can I have a bottle of
B: Are you over twenty-one? Ins 1
A: No. Ins.2
B: No. A1
The reason this is an insertion sequence is because the question in line 2 does not ignore or propose not to answer the question in line 1. Rather, it serves to defer the answer until further relevant information (in this case, whether the speaker A is old enough to buy beer) has been obtained. As we see, speaker A orients to that deferral by answering the inserted question in line 3, rather than, for example, asking his initial question again, or complaining that it has not been answered. Once the insertion sequence is completed, B shows that he is still orienting to the relevance of the original adjacency pair by moving on, in line 4, to provide the relevant part.
This example illustrates a further aspect of adjacency pairs. The participants, by 'orienting' to the relevance of the adjacency pair, display to one another their understandings of what each utterance is aiming at accomplishing. The, participants can use adjacency pair mechanisms to display to one another, and to the analyst also, their ongoing understanding and sense-making of one another's talk.
The absence of a second pair part is most often treated participants as a noticeable absence, and the speaker of the first part may infer a reason for the absence.
Example in a question/answer sequence:
Child: Have to cut these, Mummy. (1.3) Won't we
Mummy.
Child: Won't we.
Mother: Yes
(Hutchby & Wooffitt, 1998:42)
The child asks his mother to confirm her observation that they will 'have to cut these', then getting no response in the 1.5-secon pause in line 3, makes an issue out of the absence of an answer by repeating the question.
This shows that talk-in-interaction is not just a matter of turn-taking but also a matter of accomplishing 'actions'. Within this framework, failure (or perceived failure) to take a turn in an appropriate place can itself be interpreted as accomplishing some type of action.
Preference
Another inferential aspect of adjacency pairs stems from the fact that certain first pair parts make alternative actions relevant in second position. In some adjacency pairs there is a choice of two likely responses, of which one is termed preferred response (because it occurs more frequently), and the other dispreferred (because it is less common).
Examples:
1. Offer A: Like some coffee?
-acceptance (preferred) B: I'd love some.
-refusal (dispreferred) B: Thanks, but I'm
waiting for my friend
2. Compliment A: That's a nice skirt.
-acceptance (preferred) B: Thanks
-rejection (dispreferred) B: Well, it's quite old
-agreement (preferred) B: It's nice, isn't it?
-shift B: My friend found it for me
-return B: Thanks, I like
yours too.
3. Blame A: You broke the plate.
- denial (preferred) B: I didn't do it.
- admission (dispreferred) B: Sorry, I didn't see
it.
4. Complaint A: You ate the cake I
left in the fridge.
- apology (preferred) B: Sorry
- denial (dispreferred) B: No, I didn't. It
must have been somebody else.
- excuse B: You shouldn't
have left it there.
- challenge B: So what!
A dispreferred response is usually marked by: a slight pause, a preface (like 'Well', 'You see') or by an explanation or justification.
The concept of preference is not intended to refer to psychological motives of individuals, but rather to structural features of the design of turns associated with particular activities, by which participants can draw inferences about the kinds of action a turn is performing. So, initial actions can be designed to invite particular kind of response. For example, the phrase 'isn't it?' might be appended to an assessment, thereby inviting recipient's agreement.
EXAMPLE:
Jo: T's-it's a beautiful day out isn't it?
Lee: Yeh it's gorgeous.
(Hutchby & Wooffitt, 1998:44)
Speakers may design first parts in particular ways in order to get certain social actions done.
In general, CA work has focused more on micro-structured issues rather than on the larger, macro-structures of conversations, i.e., a lot of work has been done on adjacency pairs and preference organization of talk. From this perspective, the major drawback is that, while focus on small excerpts of talk has been responsible for CA's discoveries about conversation, it is limited in its ability to deal with sustained interactions (cf. Eggins and Slade, 1997:32).
III. SUMMARY
CA is the study of recorded, naturally occurring talk-in-interaction.
CA is only marginally interested in language as such, but first and foremost in language as a practical social accomplishment.
Its object of study is the interactional organization of social activities.
In other words, CA aims at discovering how participants understand and respond to one another in their turns at talk, with a central focus on how sequences of actions are generated. Throughout the course of a conversation or talk-in-interaction, speakers display in the 'next' turns an understanding of what the 'prior' turn was about. That understanding may turn out to be what the prior speaker intended, or not. This is described as next-turn proof procedure and it is the most basic tool used in CA to ensure that analyses explicate the way in which the participants themselves orient to talk, not based on the assumptions of the analyst.
Basic notions
1. Turn-taking mechanism:
The turn-taking model has two components:
a. turn construction units
Turns at talk can be seen as constructed out of units which broadly correspond to linguistic categories such as sentences, clauses, single words (e.g., 'Hey!', 'What ?') or phrases.
Features of turn-construction units
projectability - it is possible for participants to project, in the course of a turn-construction unit, what sort of unit it is and at what point it is likely to end.
transition relevance place - at the end of each unit there is the possibility for legitimate transition between speakers.
b. Turn distribution
There is no strict limit to turn size, given the extendable nature of syntactic turn-constructional units;
There is no exclusion of parties;
The number of parties can change.
2. Adjacency pairs are utterances that are conventionally paired so that, on the production of a first pair part, the second becomes relevant and remains so, even if it is not produced in the next turn. The next turn in an adjacency pair 'sequence' is a relevant second pair part.
3. Preference organization
In some adjacency pairs there is a choice of two likely responses, of which one is termed preferred response (because it occurs more frequently), and the other dispreferred (because it is less common).
A dispreferred response is usually marked by: a slight pause, a preface (like ('Well', 'You see') or by an explanation or justification.
V. TASKS
Let's consider a facsimile of the typescript of the words of a conversation actually produced between a hearing woman and her deaf daughter-in-law, Niki. Identify who says what in this typed conversation. This will give you some idea of the problem of recognizing when a speaker's turn ends (Langford, 1994:73-74):
HELLO HI MOM HERE IS NIKI TODAY IM FINE MY ARM IS SORE YES THE DOCTOR SAID IT WOULD BE HOW IS YOUR TUMMY I THINK IT IS FINE ARE YOU STILL IN PAIN NO I DON'T HAVE PAIN THAT IS GOOD TELL IAN TO BRING OVER THAT BILL YOU GOT FROM THE BANK
AND AN OLD ONE SO I CAN SEE HOW MUCH DIFFERENCE YOU HAVE TO PAY MORE I HAVE TO KNOW HOW MUCH MORE YOU ARE GOING TO HAVE TO PAY THEM SO I KNOW WHAT IM TALKING ABOUT WHEN I CALL THEM YES MY MUM JUST TOLD ME THAT U WANTED HIM TO BRING THE BILL I'LL TELL HIM WE HAD XX HAVE TO PAY 16$ MORE OK I WILL CALL AND FIND OUT MORE THAT TOLD YOU IT WOULD GO UP A FEW DOLLARS RIGHT YES BUT THEY GAVE US MORE I KNOW I WILL CALL AND SEE WHAT THEY CAN DO WAIT TO SEE WHAT HAPPENS ANYWAY OK WHEN I DRINK TEA AND IT TASTES FUNNY IN MY STOMACH I THINK I HAVE BLEED IN MY THROAT OH WELL THE DOCTOR SAID IT WOULD FEEL FUNNY
Father (on the phone to university): So I think I'll be in tomorrow, when Peter is a little better. And if you could tell the ethics committee.
(in a loud voice): HEY STOP THAT RIGHT AWAY
Secretary: You want me to stop WHAT
Father: Sorry I was talking to the cat -
hold on
Secretary:
Father: The damn cat was fixin' to sit on the
baby's face.
Tourist: Is there a toilet around here?
Attendant: You want to use?
Tourist: Sure I do.
Attendant: Go down the steps.
(Two secretaries meet in the hallway of their common office):
A: Would you like a piece of apple cake?
B: Have you got some?
BIBLIOGRAPHY
Lazaraton, A., 1997, Peference organization in oral proficiency interviews: the case of language ability assessments, in Research on Language and Social Interaction, 30(1): 53-72
Schegloff, E., A., 1980, Preliminaries to preliminaries: 'Can I ask you a question?' in Sociological Inquiry, 50 (3-4): 104-52.
1988, Presequences and indirectness, in
Journal of Pragmatics,
|
