® A COMPANION TO QUALITATIVE RfSEARCH concepts have to prove themselves, (cl. Williams 1988: 71). Hie final ob|ective ol this both Inductive and deductive procedure Is a formal analytical language that will make it possible to describe the field of face-to-face interaction. Goffman achieves the highest level of formalization in his 'frame analysis'. There he succeeded in developing a 'meta-schema' foi the analytical description of the interaction order which also substantially incorporated his earlier conceptual apparatus. This meta-schema and its precursors in Cioffman's work are, as a sociological 'map' and as a throreticaUanalytieal programme, rather closer to ffarsons's sociology than Is generally believed. Goff man's critical distance and even opposition to Parsons cannot hide the fact that his approach deserves the title of 'struciural-functionallsm' that Is normally associated with Parsons. And even Parsons's formalism finds, In Ooffman's sociology, not an opponent buť'Hrther ^n emulation. FURTHER READING Burns, T. (1992) Erving Goffman. London; Rout ledge. Manning, Ph. (1992) Erving Goffman and Modern Sociology. Cambridge: Polity Press. Williams. R. (1988) 'Understanding Goffman's methods', In P. Drew and A. Wootton (eds), Erving Goffman - Exploring the Interaction Order. Cambridge: Polity Press, pp. 64-88. in »>i 2.3 Harold Garfinkel and Harvey Sacks Jörg R. Bergmann 1 Scientific and historical background 2 Development of the research programme 3 Research practice A Garfinkel, Sacks and qualitative social research ■■■ i 30 3d Harold Garfinkel (b. 1917) is widely known today as the founder of ethnomethodology. He #ave this research approach its name, and in his early work, which appeared in his !967 collection Studies in Ethnomethodology, he created the theoretical, conceptual and methodological foundations of the approach. The subject of ethnomethodology, according to Garfinkel, is practical everyday action in situations. Its goal is to determine the practices and procedures (or methods) that are taken for granted, and by means of which members of a society (or eth-nos), in their actions, make their own behaviour perceptible and recognizable, and structure and order meaningfully the reality that surrounds them. Unlike the work of Erving Goffman (sec 2.2), which dates from about the same period, Garfinkel's works are much more cumbersome and inaccessible: they are basic in their demands, thoroughly programmatic in character, and for these reasons are often very opaque, ■n spile of this, or perhaps even because of this, Garfinkel has attracted a large number of followers who made 'ethnomethodology' into a school of its own. In the 1960s and 1970s con-versaiion analysis (see 5.17) developed out of ethnomethodology, as an independent research orientation that concentrates on identifying the structural mechanisms of linguistic and non-linguistic interaction. In conversation analysis the work of Harvey Sacks (193S-1975), in particular his lectures (1992), was of fundamental importance. For reasons that will be explained below, both Garfinkel and Sacks were very reserved in explaining and setting out the methods of their procedure. It will therefore be all the more revealing to examine the research style of these two scientists more clowlv, 1 SCIENTIFIC AND HISTORICAL BACKGROUND Garflnkcl's decision to place everyday action at the centre "I social scientific interest was not due to a fascination with the exotic nature of trivial matters. It is based, rather, on a theoretical consideration with many underlying assumptions. Garfinkel's starting point is a theme that is known in sociology as the Hobbesian Problem, and relates to the question of how social order is possible when human beings pursue egoistical goals and are therefore constantly- in conflict with one another. Garfinkef began with the reflections of Talcott Parsons (1937), his doctoral supervisor, who had set out In his theory of social action a general framework for sociology, and who dominated International sociological debate at lhat time. Parsons v.w: tin: -.oliitmn ;\ ::■,. problem Ol social order not In utilitarian models of society, but In a way already landmaiked by Durkhelm and Freud: social order, he claimed, results from the collective adoption and internalization of 30 ) A COMPANION TO QUALITATIVE RESEARCH commonly shared values and norms, and this not only restrains the egotistical tendencies oi Individuals but also exerts Cultural control over the objects of their desires. Garflnkcl. however, In his dissertation Vie Perception of the Other: A Sliuly in Social Order, made Ihls kind ol solution the target of a theoretically developed critique, supported by an empirical interview-based study. (On the relationship between Parsons and Garflnkei cf.. In particular. Heritage 1984.) In his criticism of Parsons, Garflnkei relies essentially on the works of Alfred Schlitz, who was already - in the course ol a correspondence with Parsons (Schütz 1978) in the early 1940s -expressing reservations abojit Parsons's failure in his work to clarify* the subjective perspective. This is where Garflnkei also begins. His criticism is that the specific adaptations, interpretations, translations and decisions made by actors are glossed over as irrelevant or neutralized by the model of scientific-rational action. He argues that the solution to the problem of social order can only be found in the elementary processes of the everyday constitution of meaning, that is, by invesligatlng how actors, In their day-to-day .ii ilvltie\, transmit ■. iiilural norms and valu« W a situation, agree with others and make them relevant lo their actions. Because the Interest of ethnomethodology is based, from a theoretical point of view, only In the situational practices of everyday life, it Is not surprising that Garflnkei (1991: 11) -in one of his later texts on Parsons's (1937) Structure of Social Action - claims that 'Ethnomethodology has Its origins in this wonderful book. Its earliest initiatives were taken from these texts.' Although in preparing his dissertation Garfinkel was looking primarily at Parsons's theory of action and the subsequent development of ethnomethodology was not yet in sight, we already find, in this early work, at least the germ of many concepts and aspects that characterize the style of his later work: the sharp distinction he draws between scientific and everyday railonallty; the iransfer of meaning constitution from a transcendental or psychological frame of reference to the social events of everyday life; the Idea of sociaborder not as a flxable, almost material fact, but as a continuous creation over lime; the centring of research interes! on the adaptable situational practices of actors; the uncompromising refusal to accepl general schemata to explain social action; the bold 'empirical' readings of theoretical texts. To these characteristics a sharp and sometimes polemical confrontation with conventional 'formal analysis' (Garfinkel 1996) war added in Garfinkel's later work. In the mid-1950s, Harold Garfinkel and Harvey Sacks met for the first time (at a seminar led by Parsons at Harvard). At that time Sacks was studying law at Yale University, but he was less interested In practising law as an attorney than in discovering how law functioned as an institution (Schegloff, in Sacks 1992). For Sacks's further intellectual development the first thing that was of decisive significance was the continuing interchange with Garfinkel, and the second was the environment in the University of California at Berkeley, where he moved at the end of the 19SOs. It was here that, in the following years, Erving Goffman in particular had a strong influence on Sacks and on other later eihnomethodologists (David Sudnow) and conversation analysts (Emanuel Schegloff). Goffman's first l»ok (The Presentation of Self in Everyday Life) appeared in 1959 and turned the investigation of face-to-face communication into an Independent area of sludy. Prom the Lectures thai Sacks began to give from 1964, and which he had recorded, transcribed and circulated, we may conclude that, in addition to the development of his research interests, he attached great significance to the late work of Wittgenstein, classical philosophy and logic, the ethnographies of the Chicago School and cultural anthropology, generative grammar and the work of Freud. But he used all of these works rather as thought-stimuli and resources and -without being particularly faithful to the original - turned them to his own interests. On the other hand, what remained of central importance to him was Garfinkel's attempt to make the methodical nature of everyday action In Its situational practices into the primary subject of investigation. 2 DEVELOPMENT OF THE RESEARCH PROGRAMME Nowhere is a more accessible and convincing representation of the cthnomethodological research programme to be found than in one episode reported by Garfinkel himself under the title 'Shils' complaint' (Garfinkel et al. 1981: 133). In 1954, Fred Strodtbeck of the University of Chicago law School was appointed to analyse tyt- tape-recordings that had been secretly made of the deliberations of jurors. When Strodtbeck suggested using Bales's categories of interaction process analysis for this analysis, Edward Mill'. warned that 'By using Bales' Interaction Process Analysis I'm sure we'll learn what about a jury's deliberations makes them a small group. But we want to know what about their deliberations makes them a jury.' The fact that Strodtbeck countered this by claiming that Shils would ask the wrong question and Shils then agreed with this claim, was seen by Garfinkel as part of the moral of this story: Garfinkel was convinced that Shils had in fact asked the right question, but that the social sciences are not equipped with appropriate concepts and methods to translate Shils's criticism Into investigable phenomena. Shils's question is an exemplary formulation of what ethnomethodology is trying to achieve in its research programme: not to subsume a social phenomenon under a familiar sociological category, but to work out by what practical methods 'something' becomes 'something'. Garfinkel prescribes for ethnomethodology a constitutive-analytical programme and criticizes traditional sociology and social research lor ignoring the question of how a social phenomenon Is constituted in the situational practices of actors, and for using - without further clarification - everyday knowledge and common-sense practices as resources, instead of making these into its subject of study. In the early phase of ethnomethodology this was a recurrent theme that Garfinkel reflects in his papers on a range of different phenomena. An example of this is the subject of 'suicide', which is suitable as an illustration of the new-style ethnomethodo-logical way of looking at things because Durkheim had contributed a classical study to this topic that was important for the establishment of sociology as a discipline. In Durkhelm's work, the ethnomethodologists argue, we find everyday knowledge about 'suicide' and the use of this category without any more precise clarification. What Sacks holds against this practice (1963: 8) is that 'till we have described the category, suicide. I.e. produced a description of the procedure employed for assembling cases of the class, the category is not even potentially part of the sociological apparatus'. Garfinkel (1967a: 11-18) himself also subsequently showed, in a participant observation, the situational practical pioccdures used by a coroner to 'confirm' a HAROLD GAfiFINKEL AMD HARVEY SACKS (31 suicide and to construct, for a discovered body, an 'account of how a death míWy-for-all-practical-purposes-happened', using particular identifiable clues. Since ethnomethodology makes Into Its ob|ect of Investigation whatever was used In traditional sociology and social research as an unquestioned resource and precondition, Its procedures could not simply rely on the established methods of data collection, data processing and theory construction. In the first place, the object of ellinomethodology - the situational practices for generating reality - would be eliminated if social events are methodically processed by coding and numerical-statistical transformation. And in the second place, these practices cannot simply l>c accessed through an interview. They are, in Garfinkel's words, 'seen but unnoticed' in everyday life, judgements as to whether everyday actors' statements or declarations are appropriate, relevant, meaningful and so on are always practical judgements, because they are assessed und accepted with the help of situational procedures in respect of practical goals and needs. For this reason, ethnomethodology adopts an attitude of 'indifference' towards them (Garfinkel and Sacks 1970: 344ff.). This implies that ethnomethodology. In realizing Its constHutlonal-analyiical programme, cannot simply depend on interview responses (unless there Is some enquiry into what, in the behaviour of parties concerned, makes an interview into an interview). With its focus on the local practices and unvarnished details that constitute a social phenomenon, ethnomethodology seeks, in its own investigations, to collect data in which the events it is looking at are conserved. This obliges the discipline to use a conserving mode of data recording (Bergmann 1985), where social events are preserved in their raw form. Irrespective of plausibility or expectations of behaviour. This is the background for the interest of ethno-methodo/)glsts, which started In the 1960s, m tape- and video-recordings or social interactions in 'natural' or unmanaged contexts, and in the development of transcription conventions thai made It possible to fix a conversation In wtiiing without either orthographic 'normalization" or reduction. Of course, this creation ol a method out of an cthnomcthodoionical perspective Is not without problems. By its own admissions, genera) non-object-dependent categories and rules - and methods are in principle nothing © A COMPANION TO QUALITATIVE RESEARCH more than this - should not he used in the investigation of a social phenomenon, because otherwise there is a risk that the specific generative procedures and 'identifying features' (Garfinkel) of this phenomenon will be lost or prejudiced. It is therefore only consistent that Garfinkel's study of a transsexual person, or the practice in conversation analysis of processing and presenting data by means of transcripts, should themselves be ethnomethodologically deconstructed (cf. Anderson and Sharrock 1984; Rogers 1992). In the formative years of ethnomethodology and conversation analysis, (Jarfinkel and Sacks, when dealing with the objectives that characterized Garfinkel's texts In-' particular, firmly refused every request to make the procedural rules of their approach explicit and to make them binding in the sense of a school of method. (This is perhaps one of the reasons why Garfinkel and Sacks published comparatively little and had more influence through oral forms of academic communication - and here one might also think of Wittgenstein.) For methods of investigation Garfinkel postulated a 'unique adequacy requirement', which means that methods must be so fashioned that they are uniquely suited to their object - but this can only be decided after information about the object of investigation has successfully been obtained, which therefore makes any formalization impossible. Prescriptions and canonizations of particular methods were subsequently developed - particularly in conversation analysis -and even today these still attract criticism from many elhnomethodologists. The representatives of the two positions are, however, united in the conviction that methods should never come before the object of study and in doubtful cases must even be sacrificed. 3 RESEARCH PRACTICE If one wishes to characterize ethnomethodology and conversation analysis as qualitative approaches, on the basis of the wotk of Garfinkel and Sacks, and in addition to describe Ihe methods in a systematic way, then on the basis of the properties of these approaches that we have outlined, a very diffuse picture emerges. In data collection ethnographic methods (see 5.5) are used, and in particular methods of data recording; Sacks and other researchers frequently also rely on process-generated dala - such as tape-recordings oi telephone conversations with a 'Suicide Prevention Center*. What indecisive Is that social events are documented in their 'natural' context and in their real chronological sequence. Garfinkel, in the 'studies of work' that he inaugurated (see 3.2), required ihe researcher to become familiar with the specific competences of the workplace being Investigated - an extreme requirement that can only be met in exceptional cases. Conversation analysis, on the other hand, was frequently satisfied with data consisting of simple tape-recordings of conversations, without any requirement for a more profound knowledge of the conversational context. In accordance with their constitutional-analytical approach, Garfinkel and Sacks viewed questions merely as a global theoretical tool; they only take on their particular relevance when they are faced with the material that has been collected. In interpreting data both ethnomethodology and conversation analysis normally begin with familiar social scenes and intuitively intelligible communicative utterances, and attempt to discover analytically from these the formal procedures by means of which the structures and events of the social world are constituted in the behaviour of the actors. It is mastery of such procedures that makes up the interpretative and Interactive competences of the actors, and it is only through them that they become members of a society. Since these competences are largely a matter of routine, however, they do not normally attract attention and are difficult for the researcher to access. But in order to make visible the products of social reality that are concealed in the same social reality, Garfinkel, Sacks and other ethnomethodologists and conversation analysts have always employed certain tricks that are designed to assist in exposing the opaque nature of the everyday world. Three of these tricks will now be outlined: 1 In his dissertation (Garfinkel 1963: 187) Garfinkel already followed the strategy of asking, on the basis of a stable system of action, what one would need to do to create disorder. The motivation behind this was that the same operations that are necessary to evoke anomie and disorganization could also provide the key to understanding how social structures are maintained. The crisis experiments carried out by Garfinkel and his students indeed managed to bring about confusion and annoyance in the parties concerned. Ultimately, however, they were not very significant from an analytical or revelatory viewpoint, and served rather as a way of demonstrating to non-ethnomethodologists that the everyday world contains hidden structural features that had previously been taken for granted. It should be noted that Garfinkel never abandoned the idea of making heuristic use of critical occurrences where social order breaks down. In his 'studies of work', for example, he therefore examines the presence and effect of 'procedural troublemakers', that is, persons who are blind, confined to wheelchairs, or who suffer from other handicaps, because 'with these "troublemakers", work's incarnate social organizational details are revealed by overcoming their transparency' (Garfinkel 1996: 12). A further example is provided by Garfinkel's student Robillard (1999), who made use of his own disability - he suffers from progressive paralysis and is dependent on technological support to enable him to communicate - to gain insight into the practices that enable us to perceive everyday phenomena as normal and natural (cf. further examples in Schwartz and Jacobs 1979). 2 An opposing and apparently paradoxical movement is characteristic of the methods of ethnomethodology and conversation analysis. This consists of approaching as closely as possible the social event being investigated during the research process but at the same time distancing oneself from it. In conversation analysis this approaching consists of overcoming the fleeting nature of the observed social events by making audiovisual recordings. These are then fixed by precise transcription (see 5.9) in a written form in their smallest and apparently most insignificant details, and the representation becomes progressively more fine-grained and richer in nuance through repeated listening and viewing of the social events being investigated. Conversation analysts therefore put a social object under the microscope and examine it in a way that is not possible in the normal haste of everyday life and which is alien to current practice in the social sciences. At the same time, however, they distance themselves from the social object of their investigation by avoiding the normal everyday practice of making social events intelligible by hurriedly attributing motives to them. In addition they do not HAROLD GARFINKEL AND HAflvEY SACKS (íl) replace the recorded utterances and behavioural .sequences with condensing and inľerpretative paraphrases, and they admit knowledge of the context of a social interaction into an analysis only in a highly controlled and measured form. The point of this attitude of conversation analysis, Simultaneously approaching and distancing itself from its object of investigation, is to focus the analysts' attention completely on the interaction order of social behaviour and its creation by the participants. The aim is to reconstruct the constructive achievements of the interacting partners, and also to observe their observations, to interpret their interpretations and to find the methods in their (ethno-)methods. 3 Both ethnomethodologists and conversation analysts are committed to using, in their research work, procedures and' methods the analysis of which they have selected as the theme of their investigation. In the interpretation of an action or an utterance they have no other choice than to make continual use of their competence as members of society* and to employ their intuitive understanding. But ethnomethodology and conversation analysis both seek not simply to use intuition but to take a step back from the analyst's own intuition and to analyse the underlying generative mechanisms of this intuition, f n this way Harvey Sacks (1972), in a paper that has now become well known, analyses a story told by a three-year-old girl: The baby cried. The mommy picked it up.' First he presents his own intuitive understanding of this story, that the mother who picks up the baby is the mother of this particular baby, even though there is no explicit personal pronoun to mark this kind of relationship. His paper then turns to the problem of reconstructing that led him - and presumably most other people who hear this story - to the intuitive understanding that he describes, (for another ethnomethodological study of the same kind, with a paradigmatic character, cf. Smith 1978.) To make»easier this rather difficult distancing from ones own intuitive understanding. Sacks made use of a trick. One sees a person and intuitively notes that this person is 'angry'. Bui what is It in the behaviour of this person thai evokes ihe intuition of 'angry'? Sacks directs attention to these fundamental production practices by placing before the intuitively perceived marker of person the phrase "doing being'. So 'angry' becomes '[doing beingl angry', and a person intuitively perceived as being a policeman © A COMPANION TO QUALITATIVE RESEARCH becomes a '[doing l>eing| a policeman". Now there is a possibility of breaking clown Individual Intuition into observable production practices {Sacks 1984). This recourse to individual intuitive understanding is by no means a rule for Garfinkel and Sacks. On the contrary, for both of them it is not Intuition and spontaneous understanding but observation that is of* fundamental methodological importance. For Garfinkel and Sacks intuitive understanding does not play ihe role of the final piece of evidence; it is not explanatory but rather, as something created, has to be explained, ľor this reason tl«ose observable states of order in social behaviouiflhal go against Intuition must also be investigated to establish the meaningful nahne of their production. A GARFINKEL, SACKS AND QUALITATIVE SOCIAL RESEARCH Garfinkel and Sacks have made no explicit pronouncements on questions of qualitative social research. But dom their criticism of quantitative research and from their own research practice -their interpretative approach, their orientation towards the subjective perspective of actors, the tendency to use case studies, and so on - it is absolutely clear that in their own minds they associated themselves with qualitative social research, because it gives a better guarantee of preserving the integrity of data. From this it also becomes clear what their specific contribution to qualitative social research consists of. Their work shows that the construction of social reality can be observed in the communicative ill« ■ '..■ '. llhl siul.llioil.il pi u lit« '- Ol ■■'.':..1.1\ life; they draw attention to the fact that research must analyse its social objects within the tlmescale In which life lakes place; they licmoiistuť Ihe enormous gain thai can í" made for sociology in considcriiigoapparcntly insignificant details; and they encourage mistrust both towards individual common-sense interpretations and towards the scientific categories that scientists all too gladly use in handling data. Helmuth Plessner (1974: 146) once wrote of Husserl's phenomenology that it was characterized by 'the tendency to abolish philosophical theories and "-Isms", viewpoints and principles, to dispense with the systematic unit as opposed to the surging wealth of concrete themes, by the will to work and openness to the public, respect for the small, patience with the partial, modesty in face of the immeasurable'. It is this attitude which - mediated by Alfred Schütz -also characterizes the research style of Garfinkel and Sacks. Their unconditional orientation towards the matter in hand and the secondary role of method are perhaps the most important -if rather ambivalent - contribution that ethno-methodology and conversation analysis, by their own example, will make to a more general methodology of qualitative social research. FURTHER READING Garfinkel. H. (1967) Studies in Ethno-methodohgy. Englewood Cliffs, NJ: Prentke Hall. Heritage, J. C. (1984} Garfinkel and Ethno-methodology. Cambridge; Polity Press. Sacks, H. (1992) Lectures on Conversation, vols I and II (edited by G. Jefferson with introductions by E. A. Schegloff). Oxford; Blackwell.