On a rainy sunday morning, I walked back with a colleague to the dormitories on a US campus. We had just attended the final discussion of an ethnomethodology conference. After some rather angry critical remarks by one of the participants, the discussion had developed into a stalemate repetition of position statements which did not seem to express a willingness to understand or appreciate the position taken by the other. Nobody seemed to be able to break this endless cycle. Harvey Sacks was right!, my companion said, in saying 'I only want to discuss in the presence of data(2). This statement seems to me to be a sensible characterization of the ethos of Conversation Analysis (CA), as it has developed in Sacks' footsteps. I want to combat, however, the impression that the CA position is one of bland 'empiricism'. On the contrary, I want to argue that it can be seen to be an example of 'empirical philosophy', to use a concept developed by Annemarie Mol (1994). Therefore, I will, in this paper, try to characterize CA in terms of such an enterprise, by discussing some images of CA, but also by quoting statements made by conversation analysts concerning their enterprise.
So, contrary to many research styles in the social sciences, CA puts 'data' rather than 'theory' in first position. Furthermore, the data it uses are of a very special kind: (transcripts of) recordings of 'natural' interactions. This means, on the one hand, that the favored data of other traditions of research--survey results, experiments, interviews, documents and even ethnographic observation--are expressly excluded (cf. Heritage, Atkinson, 1984), although some of the most qualitative ones are sometimes allowed as supportive 'ethnographic particulars'. On the other hand, the data CA-ists prefer, are largely ignored by most other traditions. And these data are analyzed in minute detail, which baffles those in other traditions, which tend to treat data in a summarized and summarizing fashion. This is related to the function data have in the analysis. In most traditions of research, data are used as 'indicators' for a theoretically conceived 'reality' beyond the data at hand, such as the 'underlying attitude' behind a statement, the life situation reported on, the political preferences of a population measured by a survey. Pertti Alasuutari (1995) has characterized this usage as belonging to a 'factist' perspective, which he contrasts with what he call a 'specimen perspective'. In the latter, data are treated as specimens of their kind. So, for instance, a transcript of an interview may be analyzed as an example of an interview, rather than for what it tells us about the life and opinions of the interviewee (as in Mazeland & Ten Have, 1996). It is clear that CA treats its data as 'specimens' of its phenomenon, talk-in-interaction in Schegloff's lucky phrase. This fact, that CA analyzes its data in and for themselves, rather than as indicators of, or testimonies about, an 'other'reality, means that a large set of traditional methodological considerations is no longer applicable.
So the issue to be considered is how CA conceives of its practice in terms of 'theory' and 'data' and what the rationale for this practice is, in relation to established practices and their rationales, which have their roots in different other traditions. In a recent booklet, George Psathas has provided a summary characterization of CA, which I quote below:
Conversation analysis studies the order/organization/orderliness of social action, particularly those social actions that are located in everyday interaction, in discursive practices, in the sayings/tellings/doings of members of society.
Its basic assumptions are:
1. Order is a produced orderliness
2. Order is produced by the parties in situ; that is, it is situated and occasioned.
3. The parties orient to that order themselves; that is, this order is not an analyst's conception, not the result of the use of some preformed or preformulated theoretical conceptions concerning what action should/must/ought to be, or based on generalizing or summarizing statements about what action generally/frequently/often is.
4. Order is repeatable and recurrent.
5. The discovery, description, and analysis of that produced orderliness is the task of the analyst.
6. Issues of how frequently, how widely, or how often particular phenomena occur are to be set aside in the interest of discovering, describing, and analyzing the structures, the machinery, the organized practices, the formal procedures, the ways in which order is produced.
7. Structures of social action, once so discerned, can be described in formal, that is, structural, organizational, logical, atopically contentless, consistent, and abstract, terms.
Elsewhere in his booklet, Psathas (1995: 45 etc.) elaborates on the above characterization in terms of a methodology for CA (cf. also Psathas, 1990). I will summarize his argument. Conversation analysts try to discover, describe and analyze structures of social action, or: the organizational features of various, naturally occurring, interactional phenomena. The existence of an 'interactional order' is presupposed. The task is to inquire into its emergence and persistence, which is not seen as based on specific properties of the participants or the situations they find themselves in, but rather on the 'methods' used in the order's accomplishment. On the issues of 'theory' and 'data', Psathas writes:
The variety of interactional phenomena available for study are not selected on the basis of some preformulated theorizing, which may specify matters of greater or lesser significance. Rather the first stages of research have been characterized as unmotivated looking. Data may be obtained from any available source, the only requirements being that these should be naturally occurring...He stresses the use of audio or video recordings, because these can be repeatedly played and transcribed in detail. It is in and by this process of repeated listening, looking and transcribing that the phenomena of interest can be noticed. In this process, neither the analyst's intuitions nor his or her theoretical presuppositions should be decisive, but only the orientation of the participants as available in their local activities.
... A major requirement is that the matters selected for study are those that persons in the setting are themselves demonstrably aware of and/or oriented to in the course of their action.
No assumptions are made regarding the participants' motivations, intentions, or purposes; nor about their ideas, thoughts, or understandings; nor their moods, emotions, or feelings; except insofar as these can be demonstrably shown to be matters that participants themselves are noticing, attending to, or orienting to in the course of their interaction. ...
Social research, in simplest terms, involves a dialogue between ideas and evidence. Ideas help social researchers make sense of evidence, and researchers use evidence to extend, revise, and test ideas. The end result of this dialogue is a representation of social life--evidence that has been shaped and reshaped by ideas, presented along with the thinking that guided the construction of the representation.
Ideas and evidence interact through images and analytic frames, (...) Think of analytic frames as a detailed sketch or outline of an idea about some phenomenon. Ideas are elaborated through analytic frames. Frames constitute ways of seeing the things they elaborate. (...)Images, by contrast, are built up from evidence. (...) To construct images, researchers synthesize evidence--they connect different parts or elements of the things they study in order to create more complete portraits based on some idea of how these parts are or could be related. Initial images suggest new data collection paths.
While in many traditions the starting point for any project is to deduce an analytic framework from the general repertoire of ideas, ideally codified in a systematic theory, such deductions are treated with suspicion in CA. This does not mean that CA does not have any 'ideas' or 'analytic frames', but rather that these are produced in a different, one could say more 'inductive' manner. It makes sense, in this respect, to differentiate the early phase of CA, leading to a coherent research tradition, and its later, secondary, elaboration. It is clearly the work of Harvey Sacks and his co-workers, as documented in his Lectures on Conversation (1992), to have built CA's analytic frameworks, centered around basic concepts like turn-taking, sequencing, repair, preference, etc. Against the background of a wide-ranging reading in a number of disciplines, confronting the details of the evidence, Sacks at al have not only constructed images of the interactions studied, but also the analytic instruments with which to make these constructions. Within a decade, a conceptual repertoire was formed that is still used today. Although Sacks and Schegloff were primary in developing this repertoire, their students, including especiallyJefferson and Pomerantz, have also made important contributions, as have members of later generations.
In the terms introduced by Thomas Kuhn in his The structure of scientific revolutions (1962), Sacks and his co-workers have produced a 'scientific revolution' by developing a relatively new style of doing social research. On the basis of their 'exemplary' work, others have followed doing similar studies, in this way making doing CA into a 'normal science', solving concrete puzzles within an established analytic framework.
Although the full intellectual history of CA still has to be written, one can learn from the existing sources, especially Harvey Sacks' two volumes of Lectures and Emanuel Schegloff's Introductions therein, that CA has emerged in a situation with the following properties:
CA, then, is not without 'ideas' or 'analytic frames', but these are mostly concerned with, on the one hand, basic issues of the organization and understanding of human conduct, and on the other the formulation of conceptual instruments to analyze various organizational phenomena in talk-in-interaction. The common ground of ethnomethodology and CA seems to be that both are interested in studying the membership competencies involved in the constitution of local social order. In established social research, general ideas are considered to 'underlie' or 'precede' 'theories'. The latter tend to take the form of 'phenomenon X is related in a certain (quantitative or causal) way to phenomenon Y'(6). Neither ethnomethodology, nor CA, have an interest in such 'theories'. I would say that only in that restricted way is CA indeed 'a-theoretical'.
Such recordings are not used on their own. however, by just repeated listenings or viewings. What has been recorded is 'transcribed' using a set of conventions developed by Gail Jefferson. Transcripts are unavoidably incomplete, selective renderings of the recordings, focussing at first on the text of the verbal stream, and adding various kinds of particularities of the ways in which the words were spoken later. The purpose of CA transcription is to make what was said, and how it was said, available for analytic consideration, at first for the analyst who does the transcribing, and later for others, colleagues and audiences. Transcribing recordings gives the analyst a 'feel' for what has been recorded, it helps to notice phenomena that may be later considered in detail. And, of course, transcriptions are used to communicate the evidence to an audience. In so doing, such an audience acquires a kind of 'independent access' to the data being analyzed. But even when a recording is being played as well, transcripts help in noticing specific phenomena and creating a 'shared focus' among audience and analyst.
When one compares the various conventions for transcribing verbal interaction(7), it is very clear that each system has its own theoretical and methodological 'bias'. For CA transcripts based on the Jefferson system, the prime function is to note sequential phenomena in much more detail than is necessary for other kinds of approaches. The basic technique is to visualize on paper the time line of the interactional stream, and to place each participant's contribution in relation to those of others. The space taken by the letters in the printed words is taken as a visual image of the length it took to produce the corresponding sounds during the interaction. And the details added to the textual transcript serve to make this picture more complete and exact, noting pauses, overlapping, slower, faster, latched or stretched speech. As Jefferson (1985) makes clear regarding one particular kind of phenomena (laughter), the system evolved in response to the emerging analytic needs and insights. For the vocal part, most CA-ists follow the Jefferson conventions. Details not covered by that system, especially visual ones where video recordings are used, can be added to that vocal 'base line' according to the needs of the analyst (Heath & Luff, 1993). In this way, CA's transcription system, while recognizing Gail Jefferson's foundational contribution, has become a kind of collective property, a 'language' that has evolved from the collectivity's experience (cf. also Psathas & Anderson, 1990).
In terms of Ragin's overall model of research, one could say that transcripts function as a kind of mediation between the raw data, the recordings, and the to be constructed images. For the latter, no exact equivalent is available in CA, since the data are not combined or summarized, before they are confronted with the analytic frames. What comes closest to Ragin's images are the interpreted transcripts that are sometimes produced, with a column reserved for interpretive glosses accompanying the transcript data themselves.
In Ragin's model, the ultimate analysis is characterized as a process of double fitting of the images and the analytic frame, leading to a representation of (an aspect of) social life. In CA, the analysis results in a collection of relatively abstract statements on the 'procedures' or 'devices' that participants have used in the specific cases analyzed. This can be done on the basis of one case, in a single case analysis, or on the basis of a larger set of cases, in a collection study, but one can also use various kinds of combinations, such as one focal case and a series of less fully analyzed supportive or contrastive cases(8).
Ragin suggests that, especially in qualitative research, the process of double fitting involves changes in both the images and the analytic frame. For the analytic frameworks of CA, I have suggested above that these have developed, and continue to be developed, over the cumulative series of CA studies, rather than in just the individual studies on their own.
In CA, 'instances' have a different function than imagined examples in philosophy or in most theoretical linguistics. They are the object of analysis, i.e. an effort to explicate the methods that were used to produce them, which would not make sense if they were only imagined. And, as I mentioned, the lived reality is more complex in its detail than one could ever imagine. An imagined example can only demonstrate the imaginability of what one proposes, rather than its possibility of existence. But, of course, most imagined examples are only used as pedagogical devices anyway. CA does not, however, collect its instances to count their variation, to prove its theories or to illustrate its aperçus about members' experiences. The purpose is, rather, to explicate the endogenous logic that provides for the sense of the (inter-)actions on display in those instances, as instances of a lived moral-practical order (Coulter, 1991: 38-9; Jayyusi, 1991: 242-3).
In phenomenology, which through ethnomethodology may have indirectly inspired CA, although this is not at all clearly apparent, the empirical ground for the abstract ideas is mostly constituted by a personal reflection on one's own experience, which is assumed to be analogous to that of others. The basic problematic concerns the possibility of knowledge, intersubjective understanding, etc. And the answers are sought in the structured functioning of individual consciousness. and especially the individual's pre-reflective experience. Ethnomethodology and CA differ from their philosophical precursors in two respects.; first in their empirical orientation, and second in their problematic, which is a more sociological one. i.e. a variant of 'how is society possible?'(9). It should be said, of course that ethnomethodology is, at times, much more explicitly philosophical than is CA. While philosophers, according to Mol, tend to hide their empirical inspirations, CA-ists hide their philosophical ones.
I want to illustrate this argument by discussing one of these themes, intersubjectivity(10). I take my materials from the CA-ist who has been seen as representing the ultimate scientism of CA, Emanuel Schegloff. Here is the text he uses as a self-description on the UCLA Web-site:
EMANUEL A. SCHEGLOFF, Professor (Ph.D. University of California, Berkeley) - Affiliation: EPOS, Social Psychologyfrom: WWW: The UCLA Department of Sociology: Faculty Members, Research Interests and Publications [http://www.sscnet.ucla.edu/soc/fac_list.htm].For me, direct interaction between persons is the primordial site of sociality. I am interested in exploring what we can learn about any of social science's traditional concerns through the detailed naturalistic study of interaction. In the course of pursuing this goal through the close study of (audio and/or video) recorded episodes of all manner of naturally occurring interaction, it has turned out that we can also discover previously unrecognized concerns for social science, and ones which appear to be central to the organization of conduct in interaction and of persons' experience of it. This mode of studying interaction ends up as an instrument for studying a broad range of topics in sociology and related disciplines.
In a way, CA demonstrates how a classical philosophical problem like intersubjectivity, which can hardly be solved within the classical framework, is continuously solved, or at least handled, in the practice of everyday life. One does never know, indeed, for sure, what another person thinks and feels, but you can see that people in their daily life do understand each other sufficiently to manage their everyday affairs, and what is more important, one can study how they achieve a sense of mutual understanding(11)
A recent paper by Emanuel Schegloff (1992) seems an excellent illustration of these points. As the paper's title demonstrates, a typical technical CA topic is framed in terms of a classical philosophical problem: Repair after next turn: The last structurally provided defense of intersubjectivity in conversation. Schegloff starts his argument with saying that 'intersubjectivity' could be considered to be a basic problem for sociology, but that it has been largely ignored. The classical solution of the problem consists of the presupposition of a common culture, which accounts for the experience that peoples' perspectives, although separate, are in fact similar. Schutz has put the issue again on the agenda, constructing a solution involving shared presuppositions of the reciprocity of perspectives and using a shared corpus of knowledge. In Garfinkel's writings, the issue gets a procedural switch, as the following quote demonstrates:
"Shared agreement" refers to various social methods for accomplishing the member's recognition that something was said-according-to-a-rule and not the demonstrable matching of substantive matters. The appropriate image of a common understanding is therefore an operation rather than a common intersection of overlapping sets. [Italics in Garfinkel's original text; PtH]
An alternative, or supplementary, stance might take the conduct of other social actors as ... together with interpretive procedures, coshaping an appreciated grasp of the world. Such a view would allow for the intervention by the accountable authors of conduct in what come to be stabilized as the effective understanding of that conduct. Intersubjectivity would not, then, be merely convergence between multiple interpreters of the world (whether understood substantively or procedurally) but potentially convergence between "doers" of an action or bit of conduct and its recipients, or as coproducers of an increment of interactional and social reality.In the context of such a stance, intersubjectivity is not a matter of a generalized intersection of beliefs or knowledge, or procedures for generating them. Nor does it arise as "a problem of intersubjectivity." Rather, particular aspects of particular bits of conduct that compose the warp and weft of ordinary social life provide occasions and resources for understanding, which can also issue in problematic understandings. And it is this situating of intersubjectivity that will be of interest here.
Empirie en filosofie incorporeren elkaar. Het een zit in het ander, het ander in het een. Maar soms, vaak, verborgen. Onuitgewerkt. Stilletjes. Ondergeschikt.Empirische filosofie probeert beschrijvingen van de wereld te bevrijden van de theoretische stellingen die hen aan zich onderschikken. En emancipeert, omgekeerd, filosofische vondsten van de empirie waarin ze verzonken zijn. Ze democratiseert de verhouding tussen theorie en werkelijkheid. Althans: in de tekst. Empirische filosofie is tekstpolitiek.
The empirical and the philosophical incorporate each other. One is included in the other, the other in the one. But sometimes, often, hidden. Sketchy. Quietly. Subordinate.Empirical philosophy tries to liberate descriptions of the world from the theoretical positions which make the first subordinate to the latter, and emancipates, reversely, philosophical discoveries from the empirical in which they are submerged. It democratizes the relation between theory and reality. At least: in the text. Empirical philosophy is a politics of the text.
3. See also: Psathas (1990: 3, 15, 24-5, note 3).
4. Within CA, 'speech act theory' is often used as the chosen opponent (cf. Heritage, Atkinson, 1984: 5); in related work we find similar contrasts (cf. Button at al, 1995; Suchman, 1987).
5. This is but a short and simplifying statement. A fuller treatment would require a paper in inself.
6. Ragin (1994) has argued that this general picture of the role of 'theory' in social research is inadequate for many types of research, especially qualitative ones. While quantitative research is, indeed, focussed on 'variables and relationships among variables in an effort to identify general patterns of covariation', qualitative research 'usually involves in-depth examination of a relatively small number of cases', focussing on 'commonalities', i.e. the common properties of cases belonging to a category.
7. A number of non-CA systems are explained in Edwards & Lampert (1993).
8. The issue of 'sampling', that might be raised in this context, is rarely dealt with in CA. This is defensible in the light of the 'exemplary' use of cases in CA, corresponding to the 'specimen perspective', discussed earlier. Consult scattered remarks in Sacks (1992), and Benson & Hughes (1991), especially 126-32, for a discussion concerning the irrelevance of established methodological considerations for ethnomethodological studies.
9. It should be stressed that this is only a very rough treatment of the subject of CA's (and ethnomethodology's) relation to phenomenology; see Heritage (1984) for a more scholarly treatment.
10. The theme of CA as an approach to 'intersubjectivity' has been earlier elaborated by John Heritage (1984: 254-60), who has, for instance, written: 'Linked actions (...) are the basic building-blocs of intersubjectivity' (256).
11. In a similar fashion, ethnomethodology has been able to take up the classical problems of the social science traditions, like the problem of social order, in a different fashion. Those problems don't have to be solved in the abstract, i.e. 'how is social order possible theoretically', you can right away investigate how people create and maintain a sense of social structure empirically. This seems to be the sense of Garfinkel's project of respecification (1991; see also the other contributions in the volume, i.e. Button, 1991).
12. C.f. The critiques of Bogen (1992),
Lynch (1993: 203-64), Lynch, Bogen (1994).
Benson, D., J. Hughes (1991) 'Method: evidence and inference evidence and inference for ethnomethodology'. In: Button, G., ed., Ethnomethodology and the human sciences. Cambridge: Cambridge University Press: 109-36
Bogen, D. (1992) 'The organization of talk', Qualitative Sociology 15: 273-96
Button, G., ed. (1991) Ethnomethodology and the human sciences. Cambridge: Cambridge University Press
Button, B., J. Coulter, J.R.E. Lee, W. Sarrock (1995) Computers, minds and conduct. Cambridge: Polity Press
Coulter, J. (1991) 'Logic: ethnomethodolohy and the logic of language'. In: G. Button, ed. Ethnomethodology and the human sciences. Cambridge: Cambridge University Press: 20-50
Edwards, J.A., M.D. Lampert, eds. (1993) Talking data: transcription and coding in discourse research. Hillsdale, N.J.: Lawrence Erlbaum
Garfinkel, H. (1967) Studies in ethnomethodology. Englewood Cliffs, N.J.: PrenticeHall
Garfinkel, H. (1991) 'Respecification: evidence for locally produced, naturally accountable phenomena of order*, logic, reason, meaning, method, etc. in and as of the essential haecceity of immortal ordinary society (I) an announcement of studies'. In: Button, ed. Ethnomethodology and the human sciences. Cambridge: Cambridge University Press: 10-19
Have, P. ten (1990) 'Methodological issues in conversation analysis', Bulletin de Méthodologie Sociologique, Nr. 27 (June): 23-51
Heath, C., P. Luff (1993) 'Explicating face-to-face interaction'. In: Gilbert, N., ed. Researching social life. London: Sage: 306-26
Heritage, J. (1984) Garfinkel and Ethnomethodology. Cambridge: Polity Press
Heritage, J. (1995) 'Conversation analysis: methodological aspects'. In: U.M. Quasthoff, ed., Aspects of oral communication. Berlin/New York: Walter de Gruyter: 391-418
Heritage, J., J.M. Atkinson (1984) 'Introduction'. In: J.M. Atkinson, J. Heritage, eds. Structures of Social Action: Studies in Conversation Analysis. Cambridge: Cambridge University Press: 1-15
Jayyusi, L. (1991) 'Values and moral judgement: communicative praxis as moral order'. In: G. Button, ed. Ethnomethodology and the human sciences. Cambridge: Cambridge University Press: 227-51
Jefferson, G. (1985) 'An exercise in the transcription and analysis of laughter'. In: Dijk, T.A. van, Handbook of discourse analysis. London: Academic PressVol. 3: 25-34
Kuhn, T.S. (1962) The structure of scientific revolutions. Chicago: University of Chicago Press
Lynch, M. (1993) Scientific practice and ordinary action: ethnomethodology and social studies of science. New York: Cambridge University Press
Lynch, M., D. Bogen (1994) 'Harvey Sacks' primitive natural science', Theory, Culture & Society 11: 65-104
Mazeland, H., P. ten Have (1996) 'Essential tensions in (semi-)open research interviews'. In: I. Maso, F. Wester, eds. The deliberate dialogue: qualitative perspectives on the interview. Brussels: VUB University Press: 87-113
Mol, A. (1994) 'Ondertonen en Boventonen, over empirische filosofie'. In: D. Pels, G. de Vries, red. Burgers en vreemdelingen, opstellen over filosofie en politiek. Amsterdam: Van Gennep: 77-84
Psathas, G. (1990) 'Introduction: Methodological issues and recent developments in the study of naturally occurring interaction'. In: Psathas, G., ed., Interactional Competence. Washington: University Press of America: 1-30
Psathas, G. (1995) Conversation analysis: the study of Talk-in-Interaction. Thousand Oaks: Sage
Psathas, G, T. Anderson (1990) 'The 'practices' of transcription in conversation analysis', Semiotica 78 (1990): 75-99
Ragin, C.C. (1994) Constructing social research: the unity and diversity of method. Thousand Oaks: Pine Forge Press
Sacks, H. (1992) Lectures on conversation. 2 vols. Edited by Gail Jefferson with introductions by Emanuel A. Schegloff. Oxford: Basil Blackwell
Schegloff, E.A. (1992) 'Repair after next turn: The last structurally provided defense of intersubjectivity in conversation', American Journal of Sociology, 98:1295-1345
Schwartz, H., J. Jacobs (1979) Qualitative sociology: a method to the madness. New York: Free Press
Suchman, L. (1987) Plans and situated action: the problem of humanmachine communication. Cambridge: Cambridge University Press