Harvey Sacks’ Conversation Analysis

This article is part 9 of the collection of articles on discursive cultural psychology. Go to the collection table of contents to read other articles. 

-- Previous -- Next --

Another element of discourse as joint activity is action sequence. Joint activity can be seen as an overarching project. This overarching project can be compared to A. N. Leontiev’s (1981) activity or to van Dijk’s (inter-)action macrostructure (van Dijk, 1980) in which an individual interlocutor’s action occurs in a certain place in the sequence of actions with certain goal that is derived from the goal of the overarching project itself. 

These actions are taken in sequence since one action would become a condition for the next. The later action is conditional on the undefective completion of the former. 

Deriving from telephone conversations, Clark provides an example of three broad actions in sequence:
I. A and B open the conversation
II. A and B exchange information
III. A and B close the conversation

These broad actions can be broken down further into smaller action sequences or projects, the minimum of which would consist of two or three actions. These minimum projects can be conceived of as building blocks of the overarching project. 

There has been many conversational analysis studies since initiated first by Harvey Sacks (Jefferson, 1989) that discover various such minimum action sequences in various activity contexts. Examples of such mini-mum sequences are question-answer (QA), summon-answer (SA), greetings (Hallo-hallo), pre-sequences, request-grant/rejection, invitation-acceptance/refusal, offer- acceptance/refusal (Silverman, 1998; Heritage, 1984). 

We shall now however discuss only one of such sequences that may be relevant for analyzing problem solving discourse, i.e. the question-answer sequence. The point is not that QAs comprise a large proportion of conversational activity in problem solving groups. What is intended is to show the form of analysis directed to these sequences.
Probably the most frequently discussed of these minimum action sequences is the adjacency pair of question-answer, usually abbreviated QA (Schegloff, 1972: Schegloff & Sacks, 1973; Sacks, 1989; Searle, 1979, 1992). 

Conversation in Transcript 9.1 below provides an example of such a QA sequence. What constitutes QA as an adjacency pair unit is what Schegloff (1972) termed ‘conditional relevance.’ The occurrence of the assertion 9.1.3 is conceived as an answer which is conditionally relevant to the question 9.1.2. This means that if A occurs, it occurs as a relevant response to Q; and if it does not occur, its non-occurrence will be conceived as an event. 

Transcript 9.1 Speech acts sequence in work-group discussion

Utt	Spk	Original utterances	English Translation
9.1.1	B	Ich habe da noch so eine generelle Frage.	I still have one general question
9.1.2		Geht’s nur um Produktion fürs Inland, oder geht’s auch um Export?	Is it about production only for local market, or also for export?
9.1.3	L	Im Moment geht das nur um Inlandprodukte..	Right now, only for local market
9.1.4		Aber das reicht	But that would be enough
9.1.5	C	Ja, genau.	Yes, exactly
9.1.6		Wir müssen uns erst mal hochbringen.	We must first of all bring ourselves higher

To illustrate this, let’s imagine, that after the speaker B utters the question 9.1.2 which is addressed to L, L does not utter anything. L’s silence would create questions as to why L does such an ‘inappropriate action.’ 

The QA adjacency pair may be explained by speech act theory (Searle &Vanderveken, 1985; Searle, 1979, 1992). In Searle’s speech act taxonomy, a question is categorized as a directive which is an action that attempts to get the addressee to perform a speech act (usually assertives). The answer which is an assertive conducted by the addressee would become a condition of satisfaction of the question. A question without answer then is an unsatisfied or unsuccessful speech act. 



-- Previous -- Next --

Discourse as conversational joint activity

This article is part 8 of the collection of articles on discursive cultural psychology. Go to the collection table of contents to read other articles. 

 -- Previous -- Next --

In what follows, we shall discuss discourse conceived mainly as conversational joint activity. The framework is mainly derived from studies in microsociology in general, especially in conversation analysis (Sacks, 1989; Garfinkel, 1967; Clark, 1994). We shall now try to identify elements involved in conversational joint activity. Clark (1994) has pointed out that there are four elements of discourse as joint activity: personnel, accumulating common ground, action sequences, and contribution. For our purpose however discussions about common ground and contribution will be conducted in one part, since the two are closely related to each other.

Personnel and Participation Roles

In order for a conversation to occur, it requires at least two participants that change their par-ticipation roles from one action to the next along the conversation. At one time, a participant may become a speaker and the other an addressee; and at other times the reverse. In conversa-tions involving more than two participants, one would have more participation roles than just speaker and addressees. Schematically, participation roles are described in concentric regions as follows:

In terms of this scheme, the interlocutors of work-groups above (including the experiment instructor L) play roles as speakers, addressees, or side participants at one time or the other; whereas discourse analysts who analyze the conversation through video-tape recordings or transcripts would play the role of eavesdroppers.

Accumulating Common Ground and Contribution

How is joint activity coordinated? For any joint activity to be coordinated, the participants must have rather strong assumptions about each other. When these assumptions fail, it would lead to breakdowns. 

In conversations this may lead the addressees to ask what the speaker means by his utterance; or to bring the conversation to directions unintended by the speaker against which he may utter corrections. Participants take for granted or presuppose that they share certain knowledge, beliefs or desires, and they each presuppose that they presuppose them. 

To these presuppositions one important element may also be added, that is what Searle (1990; 1992) has termed we-intentions or collective intentionality. This totality of presuppositions is called common ground. There are two parts of common ground. The first is communal common ground that represents all the assumptions that are supposed to be shared universally among members of particular community. In contrast, the second is personal common ground that represents the assumptions that the participants have inferred from personal experience with each other. 

As presuppositions, common ground normally is not made explicit in conversations. It may come up to foreground however when there are breakdowns that lead the participants to reground their shared assumptions to recover the conversation; or when it is used for arriving at some conclusions. Transcript 8.1 below is an example of communal common ground about year-seasons that is made explicit and used for arriving at a decision.

The common ground knowledge about year-seasons and its relations to the production of mantels is made explicit in utterance 8.1.2-.5, and used for arriving at decision in 8.1.6.

The importance of common ground for comprehending a sequence of utterances has been shown also by Garfinkel (1967). He has pointed out that in conversations there are many matters that the interlocutors understand are understood on the basis not only of what is actually said, but on what is left unspoken. 

To show his point, Garfinkel provided an example below. The left side column consists of what was actually said; and the right side is what the inter-locutors understood from the conversation, written by themselves. 

Have we, as eavesdroppers not read the right side column that shows the unspoken back-ground, the wife’s utterances in 8.2.2 may seem to be incomprehensible and seem to be just like a discourse topic switch. Taken without its background, utterance 8.2.2 may seem to be irrelevant response to the husband’s utterance 8.2.1.

The conversation above illustrates also what Clark (1994) has termed accumulating common ground. Clark has pointed out that along the sequence of utterances more presuppositions are added on top of the initial common ground with every single action. 

In the husband-and-wife conversation above, it starts with an initial personal common ground story about their son who could not yet put a penny in the parking meter by himself. First, on top of this common background the husband adds that now he is already able to do it (8.2.1). 

Using their common knowledge about the matter, the wife comes to utter a question (8.2.2) to confirm her interpretation. Again the husband’s answer (8.2.3) adds more to the common ground story. So in each phase of the action sequence the common ground accumulates.

Accumulating common ground is made possible through what has been called the grounding process (Clark, 1994; Clark & Brennan, 1991). Grounding is a process whereby the partici-pants endeavor to get their knowledge, or beliefs shared among them. Common ground is said to have been accomplished if it has reached the grounding criterion. 

This criterion consists of mutual beliefs that the participants have understood  what the speaker meant to a point sufficient for current purposes. The grounding process is undertaken through what Clark has termed contribution to the conversation. In contributing normally a speaker begins with presenting a utterance for his addressees to understand; and then they would in turn give him evidence that they have understood enough for current purposes. 

The first is called a presentation phase; and the second is an acceptance phase. In fact all meaningful conversations would involve contribution process such that they are inconceivable without it. 

In the husband-and-wife conversation above, the husband’s utterance (8.2.1) is his presentation, asserting that their son succeeded in putting a penny in a meter without being picked up. The wife’s question (8.2.2) is her acceptance, providing a positive evidence that she has understood his utterance. The evidence lies in the fact that her utterance is not questioned and considered as relevant response, or in schegloff’s (1972) term, a relevant next turn. 

Simultaneously the wife’s question (8.2.2) can also be considered as a presentation phase, to which the husband’s answer (8.2.3) is the relevant next turn that gives her positive evidence that he has understood her question. This process goes on as long as each utterance is considered relevant and understood. 

There are cases however that an utterance is only partly understood. In such a case, interlocutor in the acceptance phase would give negative evidence of common ground. 

The negative evidence in transcript 8.3 can be seen in utterance 8.3.2. The fact that in response to utterance 8.3.1 L utters question 8.3.2, instead of uttering an answer, indicates that common ground criterion has not been reached yet on the L side. Question 8.2.2 then functions to ask A to uncover more of what he means or to make further account.

-- Previous -- Next --