Article The idea of constitutive order in ethnomethodology European Journal of Social Theory 2014, Vol. 17(4) 479–496 ª The Author(s) 2014 Reprints and permission: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1368431013516057 est.sagepub.com Andrei Korbut National Research University Higher School of Economics, Moscow, Russia Abstract Despite its frequent appearances in sociological textbooks, dictionaries and theoretical opuses, ethnomethodology is still one of the most misunderstood and undervalued domains of sociological inquiry. This is particularly evident in the case of the central sociological question: social order. Harold Garfinkel, the founder of ethnomethodology, provided a unique answer to the question of order. His answer emphasized a contingent, situated character of constitutive practices of local order production. Initially a response to Talcott Parsons’ question about the conditions of the stability of social order, Garfinkel’s conception of constitutive order was later radicalized and used as the foundation of the programme of empirical ethnomethodological studies. To properly understand the radical character of the conception and programme, it is necessary to reveal the core elements of it and to separate them from the historically changed components. Keywords ethnomethodology, Garfinkel, Parsons, Schütz, social order The problem of social order is one of the defining problems of sociology. Its various solutions form one of the underpinnings of sociology’s development as a discipline. According to Percy S. Cohen, ‘sociological theory centers around the problem of social order’ (1968: 18). Alexander (1982), Skidmore (1979), and Wrong (1994) show that every sociological tradition offers its own answer to the question of the nature of social order and implements a certain vision of the order in its methodological and research apparatus. Ethnomethodology, which initial principles were formulated by Harold Garfinkel and his associates in the United States in the 1950 and 1960s, proposed its own Corresponding author: Andrei Korbut, Centre for Fundamental Sociology, National Research University Higher School of Economics, Office 402, 12 Petrovka Street, Moscow, Russia 107031. Email: korbut.andrei@gmail.com 480 European Journal of Social Theory 17(4) conception of the social order, which may be called ‘constitutive’.1 This conception provided an alternative to the approaches to social order that had dominated the sociology of the time by highlighting the contingent, situated character of social action. Nevertheless, the ethnomethodological idea of social order was not recognized as a real alternative, so that even today its meaning for sociological theorizing and studies remains mostly undeveloped. There are two reasons for the lack of attention to the ethnomethodological idea of social order. First, many critics of ethnomethodology associate it with one of the wellknown sociological traditions (most often interactionism, phenomenological sociology, and microsociology) and therefore ascribe alien theoretical principles and assumptions to it. Second, the practice of theorizing has a marginal position inside ethnomethodology because ethnomethodologists claim to profess a radically empirical stance towards social phenomena. However, the problem of order has always been of fundamental interest for Garfinkel, whose works throughout his career contain not only a collection of concrete empirical studies, but also a network of concepts that are used to prove the necessity of radical empirical studies of social order and to show the critical shortcomings of the existing sociological enterprise. After the 1967 publication of the most famous of Garfinkel’s works, Studies in Ethnomethodology, the ethnomethodological conception of order immediately became the subject of analysis by social scientists, but in most cases, this analysis was based on a misunderstanding of the basic principles underlying the conception. The most common was the interpretation of it as subjectivistic or individualistic. In the critical works by Gouldner (1970: 390–5), Bauman (1973), Coser (1975), Gellner (1975), and Alexander (1987: 257–80), ethnomethodology is considered one of the ways to study the inner meanings that actors impose on other people, current events, and surrounding objects. However, such an individualistic interpretation is inconsistent with the ethnomethodological interest in the practical orderliness of ordinary actions which cannot be reduced to the interpretive activities of actors. The attempts to conceptualize the idea of social order shared by ethnomethodologists inside ethnomethodology itself are also mostly unsatisfactory. In review works by Heritage (1984) and Sharrock and Anderson (1986), the problem of order, though recognized for its paramount importance, receives little attention. Zimmerman and Wieder (1970), Mehan and Wood (1975), Leiter (1980), and Benson and Hughes (1983) treat the ethnomethodological notion of social order in terms of Schütz’s phenomenological sociology and Cicourel’s cognitive sociology, arguing that, for ethnomethodology, social order consists of practices of creating the appearance or sense of order. This understanding, however, appeals to the structures of everyday consciousness—an interest that can only be found in some of the early works of Garfinkel. In terms of an ethnomethodological explication of the conception of constitutive order, the most productive are the works of Eric Livingston (1987: 12–18; 2008: 123–30) and Anne Rawls (1987, 1989a, 1989b, 2009, 2010, 2011). Livingston distinguishes two kinds of sociological approaches to social order: ‘sociologies of the hidden social order’ and ‘sociologies of the witnessable social order’ (2008: 124). In the first case, social order is considered to be hidden behind the observable details of members’ actions, while, in the second, it consists of actions themselves and therefore can be Korbut 481 directly observed. Sociologies of the hidden order—that is, the majority of sociological approaches—inevitably face the problem of method, since access to the hidden order can be obtained only through special procedures. Order is not available to ordinary members of a society, but only to professional sociologists. Sociologies of the witnessable order, including ethnomethodology, are based, instead, on the idea that orderliness of behaviour is available not only to sociologists, but also to ordinary actors. Moreover, sociologists can only access it initially as ordinary actors and only then as professionals. With this distinction Livingston captures a very important element of the ethnomethodological conception of order, though he highlights only one aspect. There are others, for example, the procedural character of order production. Livingston’s idea can serve as a starting point for the conceptualization of the ethnomethodological notion of order, but should be expanded and refined. A more sophisticated and coherent conceptualization of social order in ethnomethodology can be found in the works of Anne Rawls. Rawls argues that the ethnomethodological notion of social order covers two forms of order: ‘one of these corresponding with the constitutive nature of face-to-face interaction and the other with the accountable and rule-‘‘governed’’ nature of institutions’ (1989b: 147). There is constitutive order and there is institutional order. The former is based on the constitutive practices of producing meanings. The latter is based on ‘retrospective institutional accountability frameworks’ (Rawls, 1989a: 15). Social institutions cannot be reduced to sets of interactions, but at the same time interaction cannot be reduced to the enactment of institutional norms and values. Rather, according to Rawls, locally occurring interactions and institutional realities limit each other. Rawls’ notion of two social orders suggests that ethnomethodology does not take any side in the agency/structure, micro/macro, private/public debates. Ethnomethodology takes a ‘middle ground’ (Rawls, 1989a: 5), i.e. it demonstrates how institutional structures are reproduced in the immediate interactions and how immediate interactions are institutionalized and structured. Later Rawls put the idea of the two orders in a much broader context than just the studies of everyday activities. In a sense, Rawls is turning back to Hobbes’ socio-political treatment of order and claims that these two types of order correspond to two types of social organization: pre-modern and modern. From her point of view, modern societies are characterized by the strengthening of constitutive orders and the weakening of institutional orders. As a result of this expanded interpretation of two orders, Rawls comes to the conclusion that constitutive order is a condition of the possibility of modern societies based on the division of labour and democratic values:2 ‘the idea of constitutive interaction order and of self and sensemaking resting on a working consensus offer[s] a new way of understanding social order, social facts, social persons, and their relationship to social institutions in modern differentiated societies that hope to be democratic’ (2010: 118). For Rawls, Garfinkel’s ethnomethodology (as well as Sacks’ conversation analysis and Goffman’s interaction sociology) elaborates the concept of constitutive order as an order of immediate interactions, whose structuring principle lies inside interactions themselves and is not imposed on them by the institutions. Institutions only offer and establish a variety of ways of accounting, explaining, and justifying locally emerging orders and thus of limiting them. 482 European Journal of Social Theory 17(4) Despite being the first who offered a rather coherent vision of the ethnomethodological conception of constitutive order, Rawls developed this vision in directions that restrict its applicability in the analysis of the ethnomethodological views of order. First, the very separation of two types of order contradicts the basic intuition of ethnomethodology. When Garfinkel writes about the ‘accountability’ of social actions, it seems that he does not imply that there is some institutional order that imposes restrictions on naturally occurring interactions. After all, how can institutional order limit a constitutive one? Reifying institutional order, Rawls undermines the idea of social order’s locality or situatedness.3 Second, seeing no differences between ‘rules, expectations or preferred orders of action’ (Rawls, 2009: 510), Rawls actually supports an early, interpretativistic view of the ethnomethodological conception of order because order is related here to unobservable expectations and rules that ‘govern’ behaviour and constitute a stock of everyday knowledge. Finally, Rawls treats social order as order of a certain type of society. Local actions of direct meaning negotiation are considered an example of practice characteristic of modern social relations. Constitutive order turns out to be not only a specific order of interaction, but simultaneously a public order or a societal order. In other words, Rawls tries to return a sociological understanding of the problem of order, formulated by Parsons and radicalized by Garfinkel, to the mainstream socio-political understanding of the possibility of a certain type of society. The above analysis shows that there is a lacuna in sociological literature concerning the ethnomethodological conception of order. The lacuna is widened by the fact that this conception is very different from other sociological conceptions. The ethnomethodological ‘solution’ to the problem of order is, according to Michael Lynch (2000: 59), rather its ‘dissolution’. Ethnomethodology proposes considering any topic of order as a phenomenon of order. Consequently, a theoretical discussion of the conception of social order in ethnomethodology can be justified only by the need to show through theoretical argument—in addition to tutorials and empirical studies—that sociology’s task is to describe the phenomenal properties of social organization. Of course, the vision of social order in ethnomethodology did not remain constant throughout its development. Although this vision was and still based on the fundamental intuition of constitutive order, some elements of the conception did change. In general, there are two distinguishable stages in the formulation of this conception in ethnomethodology: (1) the stage at which the problem of perceived normality as a basis of the order was outlined; and (2) the stage at which social order was studied in its concrete local organizations. By revealing the changes in ethnomethodological understandings of order, we can evaluate the stable core of these understandings. I will show how the idea of constitutive order has emerged at the first stage of the development of ethnomethodology and which elements of this idea remained unchanged and which were transformed at the second stage. To do so, I will rely primarily on the works of Harold Garfinkel as the most consistent attempt to conceptualize social order in ethnomethodology.4 Of course, given ethnomethodology’s radical empirical orientation, it would be, perhaps, more consistent to first of all analyze actual ethnomethodological studies. Although fruitful and a necessary step in the analysis of the ethnomethodological view of order, this step cannot, however, replace the theoretical and historical reconstruction of Garfinkel’s conceptualizations of order, since his ideas form the common background to all empirical Korbut 483 ethnomethodological work. Empirical investigations in ethnomethodology should be faithful to the phenomena studied, but the reasons for this phenomenal faithfulness were suggested by Garfinkel. Perceived normality as a constitutive feature of order The idea of constitutive order was originally inspired by two sociological traditions: Talcott Parsons’ theory of social systems and Alfred Schütz’s phenomenological sociology. As Parsons’ graduate student, Garfinkel inherited his interest in the definitive properties of stable social order. At the same time, being heavily influenced by Schütz, with whom he corresponded and met a number of times, Garfinkel was trying to use the phenomenological interest in the ordinary properties of social action as a starting point for the correction and development of Parsons’ conception. This ‘development’, however, has produced a conception which in some respects has diverged from the Parsonian line. Therefore, we have to consider what Garfinkel borrowed from the two theorists and why both approaches were superseded subsequently by the different interests. In his early works,5 Garfinkel (1952, 1962, 1963) focused on the constitutive properties of social actions or, as he called it, the ‘constitutive order of events’ (Garfinkel, 1963). His interest in this problem was stimulated primarily by the phenomenological tradition in philosophy and sociology. In this regard, it is necessary to refer to another important figure—phenomenologist Aron Gurwitsch, whose works were of fundamental importance to the development of the notion of constitutive order. Gurwitsch, Schütz’s lifelong friend, elaborated on what he called ‘constitutive phenomenology’ (Gurwitsch, 2009a, 2010). The main analytical focus of constitutive phenomenology is how the world, in which humans live, is constituted and, therefore, how the knowing subject relates to the object of his/her knowledge. The basic idea of constitutive phenomenology is: It is in acts of consciousness that the object unfolds and discloses itself for what it is to the subject who becomes conscious of it. Such are the acts of consciousness which, by dint of conferring on the object its nature, structure, and sense of being, have a constitutive function in relation to it. We may thus regard the object as the correlate of a group of acts corresponding to it, or, reciprocally, we can consider that group of acts as the equivalent of consciousness of the object. (Gurwitsch, 2009b: 309–10) Garfinkel borrows the principle of equivalence between object and acts of consciousness directed at it from Gurwitsch, but gives this principle a sociological treatment, replacing ‘object’ with ‘social order’ and ‘acts of consciousness’ with ‘everyday social actions’. That is how the first fundamental thesis of the ethnomethodological conception of constitutive order appeared: Social order consists in methods of its production. The term ‘ethnomethodology’ was devised initially, according to Garfinkel (Hill and Crittenden, 1968: 5–11), to name a study of ‘methodologies’6 used by juries. As a 484 European Journal of Social Theory 17(4) participant of the project on juries’ decision-making in 1953–54, Garfinkel found that they use, and require from each other, the specific ways of evaluating and describing the correctness, adequacy, impartiality, consistency, or validity of the actions accomplished in the jury room. Therefore, the emerging order is not simply a result of acts committed by the participants, but consists of exhibition and analysis of specifically ordered properties of one’s own and others’ acts. These properties are characterized by two features: (1) they are everyday, common sense, and taken for granted; and (2) they are stable, repeatable, and reproducible. Garfinkel discovered the formulation of these features and the primary efforts to describe and explain them in the works of Schütz and Parsons, whose theoretical solutions focus, respectively, on the issue of everyday life and the issue of stability. Schütz’s initial point is that, unlike the physical world, the social world has a particular meaning and relevance structure for the human beings living, thinking, and acting therein. They have preselected and preinterpreted this world by a series of commonsense constructs of the reality of daily life and it is these thought objects which determine their behavior, define the goal of their action, the means available for attaining them . . . (1953: 8) Social scientists have to deal with an already mastered and understood world, and therefore their constructs are ‘constructs of the second degree’ based on everyday ways of understanding and acting. These second-degree constructs are created in accordance with the rules of scientific procedure, but they are founded on a routinized world taken for granted by its participants. In his works, Schütz describes various aspects of this world, which in one way or another relate to the essential problem: the constitution of common, shared understandings regardless of the unique experiences and biographical situation of each individual. Schütz solves this problem by devising what he calls the ‘general thesis of reciprocal perspectives’ (1953: 8). This thesis consists of two components: the idealization of the interchangeability of the standpoints: I take it for granted—and assume my fellow-man does the same—that if I change places with him so that his ‘here’ becomes mine, I would be at the same distance from things and see them in the same typicality as he actually does; moreover, the same things would be in my reach which are actually in his. (All this vice versa.) (Schütz, 1953: 8) and the idealization of the congruency of the system of relevances: Until counter-evidence I take it for granted—and assume my fellow-man does the same— that the differences in perspectives originating in my and his unique biographical situations are irrelevant for the purpose at hand of either of us and that he and I, that ‘We’ assume that both of us have selected and interpreted the actually or potentially common objects and their features in an identical manner or at least an ‘empirically identical’ manner, namely, sufficient for all practical purposes. (Schütz, 1953: 8) Garfinkel used the thesis of reciprocal perspectives as the key to understanding the constitutive properties of social order, as this thesis refers to the conditions of social Korbut 485 order that are not dependent on its ‘content’: reciprocal perspectives can be discovered in any activity. Moreover, this thesis allowed Garfinkel to explain the stability of the social order, which, though formulated as a problem by Parsons, had not received a satisfactory explanation in Parsons’ theory (or rather, had been only partially explained). Parsons’ views of the social order, despite overall changes in his theoretical framework, remained relatively stable throughout his career. Starting with The Structure of Social Action ([1937] 1949) and extending to his later meta-theoretical works, Parsons grounded his understanding of social order on the notion of ‘normative orientation’. In The Structure of Social Action, he described the units of action systems which included: ‘actor’, ‘end’, ‘situation’ and ‘normative orientation’ (Parsons, [1937] 1949: 44). The ‘normative orientation’ is a culturally provided basis for selecting one of the alternative means of achieving the end. In this regard, Parsons suggests ([1937] 1949: 91–2) that stable social order is possible only where there is also a normative order in addition to the factual order of human actions. Later, in The Social System (1951), Parsons clarified the additional features of the ‘normative orientation’: It is inherent in an action system that action is, to use one phrase, ‘normatively oriented’. This follows, as was shown, from the concept of expectations and its place in action theory, especially in the ‘active’ phase in which the actor pursues goals. Expectations, then, in combination with the ‘double contingency’ of the process of interaction as it has been called, create a crucially imperative problem of order. Two aspects of this problem of order may in turn be distinguished, order in the symbolic systems which make communication possible, and order in the mutuality of motivational orientation to the normative aspect of expectations, the ‘Hobbesian’ problem of order. The problem of order, and thus of the nature of the integration of stable systems of social interaction, that is, of social structure, thus focuses on the integration of the motivation of actors with the normative cultural standards which integrate the action system . . . (1951: 36) This passage suggests that stable features of action are determined by the motivational orientation to the social standards. Social structure, as a stable system of social interactions, implies people’s acceptance of such standards and their use of them as conditions for the reproduction of the action system. Garfinkel borrows the general framework of the problem of social order, as connected to the question of the stability of social actions, from Parsons. However, Garfinkel shows that Parsons’ solution is deficient since it is not clear what the conditions of the ‘motivation of actors’ are, i.e. what constitutes norms as norms in the actors’ eyes. Later Parsons recognized this problem, indicating that we ‘depend tremendously on the kind of mutuality of expectations and trust that is involved in the operation of these generalized mechanisms [i.e. state]’ (1968: 384). Here, references to ‘mutuality of expectations’ and ‘trust’ are obviously stimulated by Garfinkel who studied expectations and trust in his famous ‘Trust’ paper (1963). In this paper, containing the basic principles of the conception of constitutive order as it is formulated at the first stage of the development of ethnomethodology, Garfinkel attempted to apply Schütz’s thesis of reciprocal perspectives in order to discover the conditions of the reproducibility of social action. 486 European Journal of Social Theory 17(4) The starting point of the ‘Trust’ paper is the analysis of games and, in particular, chess. Garfinkel distinguishes, along with the basic rules of the game which define any given game as a particular kind of game, a set of preference rules7 and a series of ‘constitutive expectancies’. These constitutive expectancies are: 1. From the standpoint of a player, from alternative territories of play, numbers of players, sequences of moves, and the like, they frame a set that the player expects to choose regardless of his desires, circumstances, plans, interests, or consequences of choice either to himself or to others. 2. The player expects that the same set of required alternatives are binding upon the other players as are binding upon him. 3. The player expects that, as he expects the above of the other person, the other person expects it of him (Garfinkel, 1963: 190). As we can see, the ‘constitutive expectances’ are largely restatements of Schütz’s thesis of reciprocal perspectives, but, unlike Schütz, Garfinkel says that they are assigned to a series of interrelated events which he calls ‘the constitutive order of events of the game’ (1963: 191). It is these constitutive expectancies that provide the stability of order and the motivated agreement with it. Thus, Garfinkel finds in Schütz the solution to the Parsonian problem of social order. However, when he transfers this solution from the domain of games to the domain of everyday life, it entails a full redefinition of the very idea of order. Parsons viewed stable social order as primarily normative, but Garfinkel cannot appeal to normative orientation. Garfinkel cannot do this for two reasons. The first is that norms themselves need to be explained. The second reason is stated by Garfinkel in the following manner: ‘current conceptions of the conditions of social order stress in common as a critical condition of a stable social order the extent to which rules are sacredly regarded’ (1963: 198). ‘Current conceptions’ seem to include, first of all and most importantly, Parsons’ conception,8 but not Schütz’s, since Schütz shows the way out of the impasse of sacralization; this way is connected with the fact that constitutive expectancies, as a basis of social order, make it possible to consider order as something taken for granted, something ‘perceivedly normal’ (Garfinkel, 1963: 188). Social order exists not because society’s members sacralize norms but because they make their own and others’ actions visibly normal. This evident normality, in turn, presupposes the ordinary methodological availability of actions, the possibility for members of a society to analyze them as consistent, coherent, adequate, reasonable, or logical. The constitutive conception of order is illustrated in the ‘Trust’ paper with a series of ‘experiments’ that made Garfinkel famous, though at the same time significantly complicated a proper understanding of the conception. Garfinkel himself, or with the help of his students, carried out a number of ‘interventions’ into everyday situations, designed to breach constitutive expectancies. For example, he asked his students to engage their relatives or friends in conversation while asking for clarification of the most trivial remarks, such as ‘How are you?’ (‘‘What do you mean by ‘‘How are you?’’?’’). Commentators have interpreted these experiments as proving that social interactions are based on tacit rules or norms whose violation causes confusion, bewilderment, and anger. However, Garfinkel himself regarded his experiments as a proof of the importance Korbut 487 of constitutive expectancies for the reproduction of stable social order involving the reproduction of the observable normality of the actions. The focus on the perceived normality allows the formulation of the second main principle of the ethnomethodological conception of constitutive order: Social order is observable. People perceive their own and others’ actions not ‘through’ the norms (which allegedly govern their behaviour and make it meaningful) but from within these very actions. The resulting social order is available to participants in the details of the accomplished acts. People orient not to the norms or values, but to the observable features of activities. Garfinkel differs in this regard from Gurwitsch. Gurwitsch assumed that the object and acts of consciousness directed at it cannot be identical because (1) in each of these acts, the object appears in a certain respect, but at the same time we perceive it in its totality; (2) these acts can be numerous, but the object they constitute is one and self-identified; and (3) the object’s physical temporality does not correspond to the phenomenal temporality of its experiences (Garfinkel, 2009b: 310–11). Garfinkel suggests the identity of the social order and of the actions of its production. Social order is not hidden behind the actions, as we would have to admit if we understood it primarily as a normative order. There is order in actions themselves. The constitutive observability of social order undermines the claims of sociologists that only scientific sociological methods can reveal regular features of people’s activities. Instead, Garfinkel emphasizes that it is members who are entitled both to produce and to recognize order in everyday situations. Members cannot but make their actions understandable ‘at a glance’; that is, they have to produce them reflexively, so that their descriptions can be a part of the settings they describe. This reflexivity means that in observing ordinary scenes, people see them as already ordered, and in ordered ways. Members do not need to infer from ‘the raw data’ that there is order in everyday settings. It is there, available for description, conversation, evaluation, analysis, and display, both by sociologists and by members of a society. Thus, the first two fundamental principles of the conception of constitutive order in ethnomethodology have been formulated as a result of attempts to solve the problem of the stability of social order by revealing the mechanisms of everyday social action. In doing so, Garfinkel had to overcome the limitations not only of Parsons’ approach, but of Schütz’s approach too (though Parsons himself considered Garfinkel a ‘follower’ of Schütz (Grathoff, 1978: 123)). If the overcoming of the idea of normative regulation of actions can already be found in the ‘Trust’ paper, the rejection of the thesis of reciprocal perspectives as the basis of stable order was determined by more complex problems with Schütz’s arguments. Schütz believed that the world of everyday life is one of the many worlds. These worlds, being the modifications of the ‘natural attitude’ of everyday life, are different from the ordinary world and constitute separate ‘finite provinces of meaning’ (Schütz, 1945). Garfinkel, however, suggesting that ‘constitutive properties extend to everyday events’ (1963: 198), proposes studying the everyday constitutive order of all activities, including ones that Schütz believed to be non-everyday (for example, scientific theorizing). The second problem with Schütz’s conception is that 488 European Journal of Social Theory 17(4) he tried to solve the problem of order primarily in a theoretical fashion. The theoretical status of the solution means that it presupposes a universal answer, i.e. a conception of social order which applies to any social action. But Garfinkel discards this universalizing attitude, thereby rejecting the possibility of a purely theoretical analysis of social order. Third, Schütz’s approach leads to the impasse of subjectivism. Garfinkel, trying to continue Schütz’s line, turns to how actors react to each other’s activities in specific situations and what kind of expectations they bring to these situations. As a result, one can get an impression that he, like Schütz, analyzes the ‘world seen from an actor’s perspective’. But if this is applicable to Schütz’s conception,9 this is not applicable to Garfinkel’s studies. Garfinkel focuses on the details of specific actions, rather than on the actors. These problems create three serious difficulties for Garfinkel as a researcher of social order. They may lead to the interpretation of his conception as limited by the segment of face-to-face social interactions, as proposing a system of abstract concepts to describe real social structures, and as reconstructing actors’ interpretive activities. All three interpretations miss the initial intuition that underlies the conception of constitutive order. In response to these difficulties, Garfinkel added to the ethnomethodological conception of constitutive order three supplementary principles which can be put under the rubric of ‘order spelled with an asterisk’.10 Order spelled with an asterisk The notion of ‘order spelled with an asterisk’, or ‘order*’, appears in Garfinkel’s later works (1988, 2002) as an illustration of what he calls ‘tendentious’ use (2002: 99). Garfinkel adds an asterisk to the words that he thinks should ‘correct’ a reader’s understanding, that is, tell him/her something different from what he/she expects to hear. This can be familiar vernacular words or some technical terms, but in either case they gain their intelligibility from the studies of which the word spelled with an asterisk is a part. Put another way, the meaning of the word spelled with an asterisk should be clarified not on the basis of the reader’s prior understandings (even if the reader is a fellow ethnomethodologist), but on the basis of the research context where the asterisked word is used and made intelligible. In this regard, the term ‘order*’ should be read as referring both to the entire corpus of ethnomethodological studies, since all ethnomethodological studies deal with social order, and to the specific studies that describe a particular order of concrete practices. In relation to the entire body of ethnomethodological investigations, ‘order*’ serves as a ‘collector and a proxy . . . for any topic of reason, logic, meaning, proof, uniformity, generalization, universal, comparability, clarity, consistency, coherence, objectivity, objective knowledge, observation, detail, structure, and the rest’ (Garfinkel, 2002: 118). These ‘clarifications’ point to the problems that may become a subject of interest for those who want to study order. However, in relation to particular ethnomethodological studies ‘order*’ refers to these problems not as topics for analysis, but as an observable phenomena for investigation. Garfinkel states that ‘every topic of order* offers to Ethnomethodological study its candidacy to a search for a phenomenon of order* as an achievement in and as of practical action’ (2002: 170). This phrase of Garfinkel, like many other passages in his later works, requires some ‘deciphering’. It means that each topic of order, formulated as such, can be considered ethnomethodologically as a Korbut 489 theme that glosses an area where one can search not for the illustrations of the topic, but for the phenomena of order produced in practice. For example, if we take the topic of ‘logic’, its ethnomethodological study would raise the question of what the logic is as a practical achievement of particular ‘members’. How do people evaluate the logicality or illogicality of certain actions? Who are these people? What methods do they use to evaluate and describe actions as logical or illogical? ‘Logic’ here is not a general property of any action, but an observable phenomenon.11 The same questions can be asked in relation to any conventional topic of order*. This phenomenal orientation confirms the proximity between ethnomethodology and phenomenology (as represented by Aaron Gurwitsch and Maurice Merleau-Ponty), but in ethnomethodology the notion of phenomenality undergoes a certain transformation. For Garfinkel, the problem of social order cannot be connected to the problem of the constitution of objects in the acts of consciousness of the subjects who communicate to each other on the ground of implicit assumptions and typifications. The phenomena of order are produced in common, but the property of ‘commonality’ is not a constitutive one. This is a fundamental break with the socio-phenomenological tradition of Schütz, adopted by Garfinkel earlier. If Schütz postulated a thesis of reciprocal perspectives to refer to such a commonality as a defining feature of the social world, Garfinkel now considers the issue of commonality as a practical question of the concerted activities of those who produce the phenomena of social order. This raises the question: if concertedness is not based on the mutual expectations (as for Schütz) or normative orientations (as for Parsons), how is it possible? In other words, what are the grounds of the ‘social’? The concept of ‘order*’ implies that these grounds may be only situational, i.e. order acquires the properties of stability and of reproducibility only to the extent that particular situations (rather than typifications or action systems) are produced and reproduced. Therefore, the next principle of the ethnomethodological conception of constitutive order follows: Social order is a situated order. Order’s situatedness means that every single ordered action is related directly to other actions within a current situation, and not to suprasituational ‘norms’, ‘rules’, ‘symbols’, ‘typifications’, ‘interpretations’, ‘codes’, or ‘signs’. Order emerges as a phenomenon of action sequences. What is happening is made understandable and reportable (accountable) through the co-ordination of current action with previous and subsequent activities, and not through the use of an interpretation scheme (such as frame, tacit knowledge, habitus, and so on). Here is the source of ethnomethodology’s demand for detailed analysis. Details are constitutive features of the situation that participants orient to. Any situation acquires orderliness in details of its accomplishment. This situation, however, should not be understood as an ‘objective environment’ of social action. Rather, it is more similar to the ‘workplace’ of the action since action unfolds in the situation, by the situation and as the situation. The situation should be understood as an arrangement of details with no external principle of connecting them. Garfinkel uses the vocabulary of Gestalt theory, adopted by Gurwitsch and Merleau-Ponty, to describe social order as a ‘phenomenal field’ and a ‘figuration of details’. There is no space in this article to describe the Gestalt properties of constitutive order as Garfinkel views them, however, 490 European Journal of Social Theory 17(4) it should be noted that Gestalt theory is of great relevance for the ethnomethodological conception of order since it shows how people can exercise a ‘direct understanding’ (Köhler, 1947: 246) of current orderliness without applying any principles external to this order. Emerging as an effect of the sequences of action, every situation has an inherently temporal structure. This temporal structure develops in time as people make attempts to concert their actions. Rawls calls Garfinkel’s conception of time a ‘sequential time’ where ‘sequential time’ is ‘an intrinsic ordering principle that creates a relationship between the parts of ongoing interaction moving forward’ (Rawls, 2005: 171). This means that the ethnomethodological view of social order presupposes time as a constitutive feature of ordinary actions. Order appears only as time, i.e. the ordering of actions, goes by. Social situations are not enclosed episodes, but temporally unfolding enactments of the participants. The situational character of social order also means that the phenomena of order can be described only when and where they are produced, that is, they are specific to a particular domain. Thus: Social order is domain-specific. Social order is inseparable from the practices by which it is produced. It can be said about chess that while playing chess, players do not create an order-in-general, but an order particular to a game of chess. This order constitutes a game of chess specifically as a game of chess, and not as an example of broader social order mediated by some type of cultural symbols. What’s more, these domain-specific phenomena of order can be discovered everywhere: ‘inquiries of every imaginable kind, from divination to theoretical physics, claim our interest as socially organized artful practices’ (Garfinkel, 1967: 32). In contrast to Schütz, for whom the difference between divination and theoretical physics was so profound that he admitted the possibility of experiencing a shock when switching from one to other, Garfinkel says that we have to explore any practice as an ordinary, taken-for-granted, skilful way of producing social order. This observation poses the question of who produces the phenomena of situational order. Most (if not all) sociological traditions view actors in terms of the place they occupy in the social structure. Whatever the postulated source of the actors’ identifying characteristics (we can say that actors obey social rules or we can say that actors orient toward the reactions of the immediately present others), the phenomena are considered as derivatives of the actors’ characteristics determined by a broader social context. The order, in this sense, is enacted through the actors, because it is not observable. Even if social scientists notice any apparent forms of order (for example, the orderliness of traffic on a highway), these forms always require a further sociological explanation as they do not speak for themselves. In ethnomethodology, the relationship between actors and the phenomena of order produced by them is reversed. According to Garfinkel, ‘the phenomenon exhibits its staff as a population’ (2002: 185). This statement can be ‘translated’ into the following principle: Social order specifies its production staff. Korbut 491 From an ethnomethodological point of view, producers of social order acquire identifying characteristics depending on their involvement in its production. They do not have these properties a priori, i.e. one cannot say that drivers create a particular order (‘traffic’). A ‘driver’ becomes a ‘driver’ insofar as he/she exhibits for other ‘drivers’ his/her ability to methodically produce and evaluate the phenomena of local order. The accomplishment of these phenomena in observable actions makes it possible to identify some population as this population producing this phenomenon. As a result, certain people can identify themselves as a cohort of ‘drivers’, but always in a situational sense, as a phenomenal population that produces the observable order in the traffic here and now. In other words, people can be described as ‘drivers’ (or ‘diviners’, or ‘sociologists’, or ‘theoretical physicists’) only in relation to the phenomena of order produced by them and not in relation to the norms and rules that they allegedly follow. These phenomena of order provide the stable features of social action because, with the change of the staff of the particular phenomenon, order continues to be produced and displayed, exhibiting, among other things, its staff as a population. Drivers change, traffic remains. It may be seen that the transition from the first formulation of the conception of constitutive order, built on the opposition of ‘normative’ and ‘normal’, to the second formulation, expressed in the formula ‘order spelled with an asterisk’, presupposes a radicalization of the empirical claims of ethnomethodology. These five principles allow one to speak of the ethnomethodological conception of social order not as a theory in the strict sense but as a kind of instruction. Instead of being another way of describing social phenomena, it delineates the principles of their investigation. Although the conception of constitutive order is still a conception, i.e. a network of ideas and concepts, its every topic should be re-specified as an observable phenomenon. Conclusion I have examined the main principles of the idea of constitutive order formulated in ethnomethodology. I have attempted to show that, though the views of Harold Garfinkel, as the most consistent theorist of the social order in ethnomethodology, evolved with time, the initial intuition remained the same. As Parsons’ student, Garfinkel inherited the latter’s interest in the problem of the stability of social order, but was not completely satisfied with the decision that Parsons proposed—the conception of normative order. In the attempt to enhance Parsons’ solution, Garfinkel turned to Alfred Schütz, whose ideas helped Garfinkel to formulate the first two principles of the conception. Garfinkel replaced the notion of normative orientation as the basis of social order with a notion of perceived normality based on common-sense constitutive expectancies. However, further articulation of the idea of constitutive order made it necessary to part with Schütz’s ideas because Schütz considered everyday life as a separate domain of meaning and reduced the problem of concerted actions to the problem of actors’ interpretations of each other’s behaviours. As a result of this shift, Garfinkel suggested considering order as something that can only be formulated as a phenomenon for empirical investigation. Conventional sociology views social order as a series of topics, while ethnomethodology describes a local phenomena of order. 492 European Journal of Social Theory 17(4) I have concentrated on the connections between the ethnomethodological idea of constitutive order and the circle of ideas formulated by Parsons, Schütz, and phenomenological philosophers. I have ignored other related approaches, for example, Wittgenstein’s analytic philosophy and Goffman’s interactional sociology. This does not mean that I find these traditions unimportant for the development and current state of the ethnomethodological conception of order. The parallels and divergences between ethnomethodology and these approaches can be significant, but the goal was to analyze the development and main principles of the idea of constitutive order as they were informed by Garfinkel’s discussions and readings of a number of scholars who left their mark on his works. Parsons, Schütz and phenomenologists, I think, left the deepest marks, but, of course, there are other marks too and they are discoverable. I have presented my argument historically. I have reviewed the basic principles of the ethnomethodological conception of constitutive order as formulated over the course of Garfinkel’s intellectual evolution. But these principles are important as facilitators of sociological perception, and in this function they can be very useful for current social studies. It seems that the ethnomethodological conception of constitutive order contains unique resources that can contribute a lot to contemporary sociology. Acknowledgements I would like to express my deepest gratitude to Anne Rawls for providing invaluable insights into Garfinkel’s ideas. Of course, I take responsibility for the exposition and arguments expressed in this article. In fact, it is possible that she will disagree with every point made here. I also want to thank Svetlana Bankovskaya for being a pilot in the ocean of sociological theory. Support from the Basic Research Program of the National Research University Higher School of Economics is gratefully acknowledged. Finally, I would like to express my gratitude to anonymous reviewers who helped sharpen my arguments and faultlessly detected its weaknesses. Notes 1. The term ‘constitutive order’ was introduced by Garfinkel in his paper, ‘A conception of, and experiments with, ‘‘trust’’ as a condition of stable concerted actions’ (1963). Later Anne Rawls (2009, 2010) turned this term into a proxy of the ethnomethodological conception of social order as such. 2. Kim (2003) sustains a different vision of modern society, but he, too, suggests that Garfinkel described an order of modernity. 3. Sharrock also criticizes the idea of two orders, but on the different grounds (1999: 132–4). 4. Moreover, initially the problem of constitutive order was formulated solely by Garfinkel. 5. Michael Lynch (1993) calls them ‘protoethnomethodological’. 6. The term ‘methodology’ can also be partially traced back to phenomenological tradition. One of the works that Garfinkel heavily draws on in this first stage was phenomenologist Felix Kaufmann’s methodology of the social sciences, in which ‘methods’ were viewed as a ‘habits of thought’ (1958: 43). 7. Although there is no corresponding reference in Garfinkel’s paper, the distinction between ‘basic’ rules and ‘preference’ rules was first proposed by Kaufmann (Kaufmann, 1958: 44). 8. It must be admitted here that Garfinkel, perhaps, would not agree that Parsons’ conception presupposes the sacred character of social rules since in the ‘Trust’ paper there is no clear Korbut 493 distinction between ‘normativity’ and ‘normality’. I think, however, that this distinction is in operation there, and it is so critical that it makes impossible to reconcile Parsons’ and Garfinkel’s conceptions. Moreover, it is necessary to note that the normative view of social order was developed not by Parsons alone. Parsons’ approach is the quintessence of the widespread belief in the normative character of stable social actions. For instance, such a completely different researcher as Erving Goffman says: ‘Briefly, a social order may be defined as the consequence of any set of moral norms that regulates the way in which persons pursue objectives’ (1963: 8). As we can infer, Goffman not only supports Parsons’ normative view of social order, but also considers norms in a Parsonian way, as a culturally determined choice among alternative means of achieving ends. 9. At least, passages like ‘social sciences have to deal with human conduct and its common-sense interpretation in the social reality . . . Such an analysis refers by necessity to the subjective point of view, namely, to the interpretation of the action and its settings in terms of the actor’ (Schütz, 1953: 27) allow us to admit that such an interpretation is justifiable. 10. The apparent shift in Garfinkel’s thinking in the late 1960s has not gone unnoticed among students of ethnomethodology. D. Wilson (1978), Ruggerone (1996), T. Wilson (2003, 2012), and Arminen (2008) have recognized a significant change in Garfinkel’s problematics and ways of analysis. Arminen, for example, distinguishes two different stages of ethnomethodology which he calls ‘scientific’ ethnomethodology and ‘radical’ ethnomethodology. Radical ethnomethodology pretends to break with conventional sociology, but this leads it into an impasse: ‘By claiming complete independence of social sciences, ethnomethodology loses its grip on a mundane world and becomes a self-sufficient, empty realm’ (2008: 174). Early ethnomethodology and conversation analysis are, for Arminen, much more appropriate approaches in satisfying the basic principle of scientific practice—the reproducibility of findings. T. Wilson suggests exactly the same argument when he states that the early studies of Garfinkel ‘constitute a coherent program of research in their own right, what I call ‘‘classical ethnomethodology,’’ that is fundamentally incompatible with the later radical program but fits coherently with conversation analysis’ (2012: 207). The problem with Arminen’s and T. Wilson’s reconstructions is that they show a lack of understanding of the difference between the two stages in Garfinkel’s evolution. The difference consists not in the abandonment of scientific principles, but in ethnomethodology’s abstinence, at the later stage, from any sociological assessment of completeness, consistency, efficiency, and effectiveness of studied practices. In fact, Arminen and Wilson make an existence of order in the observable actions dependent on procedures used by researcher, which is exactly what Garfinkel was protesting against. For him, a reproducibility that ethnomethodologists must achieve is a reproducibility of everyday phenomena and not of sociological findings. 11. The informative ethnomethodological study of logic may be found in Dušan Bjelić’s unpublished PhD thesis (1989). References Alexander J C (1982) Theoretical Logic in Sociology. Vol. I: Positivism, Presuppositions, and Current Controversies. Berkeley: University of California Press. Alexander J C (1987) Twenty Lectures: Sociological Theory Since World War II. New York: Columbia University Press. 494 European Journal of Social Theory 17(4) Arminen I (2008) Scientific and ‘radical’ ethnomethodology: From incompatible paradigms to ethnomethodological sociology. Philosophy of the Social Sciences 38(2): 167–91. Bauman Z (1973) On the philosophical status of ethnomethodology. Sociological Review 21(2): 5–23. Benson D and Hughes J A (1983) The Perspective of Ethnomethodology. London: Longman. Bjelić D I (1989) On the social origin of logic. PhD thesis, Boston University. Cohen P S (1968) Modern Social Theory. London: Heinemann. Coser L A (1975) Presidential address: Two methods in search of a substance. American Sociological Review 40(6): 691–700. Garfinkel H (1952) The perception of the other: A study in social order. PhD thesis, Harvard University. Garfinkel H (1962) Parsons’ primer: ‘Ad hoc uses’. Unpublished manuscript. Garfinkel H (1963) A conception of, and experiments with, ‘trust’ as a condition of stable concerted actions. In: Harvey O J (ed.) Motivation and Social Interaction: Cognitive Determinants. New York: Ronald Press, pp. 187–238. Garfinkel H (1967) Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall. Garfinkel H (1988) Evidence for locally produced, naturally accountable phenomena of order*, logic, reason, meaning, method, etc. in and as of the essential quiddity of immortal ordinary society (I of IV): An announcement of studies. Sociological Theory 6(1): 103–9. Garfinkel H (2002) Ethnomethodology’s Program: Working Out Durkheim’s Aphorism. Lanham, MD: Rowman & Littlefield. Gellner E (1975) Ethnomethodology: The re-enchantment industry, or the Californian way of subjectivity. Philosophy of the Social Sciences 5(4): 431–50. Goffman E (1963) Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press. Gouldner A W (1970) The Coming Crisis of Western Sociology. New York: Basic Books. Grathoff R (ed.) (1978) The Theory of Social Action: The Correspondence of Alfred Schütz and Talcott Parsons. Bloomington: Indiana University Press. Gurwitsch A (2009a) The Collected Works of Aron Gurwitsch (1901–1973). Vol. I: Constitutive Phenomenology in Historical Perspective. Berlin: Springer. Gurwitsch A (2009b ) Some fundamental principles of constitutive phenomenology. In: Gurwitsch A, The Collected Works of Aron Gurwitsch (1901–1973). Vol. I: Constitutive Phenomenology in Historical Perspective. Berlin: Springer, pp. 307–29. Gurwitsch A (2010) The Collected Works of Aron Gurwitsch (1901–1973). Vol. III: The Field of Consciousness: Phenomenology of Theme, Thematic Field, and Marginal Consciousness. Berlin: Springer. Heritage J (1984) Garfinkel and Ethnomethodology. Cambridge: Polity Press. Hill R J and Crittenden K S (eds) (1968) Proceedings of the Purdue Symposium on Ethnomethodology. Lafayette, IN: Purdue University Press. Kaufmann F (1958) Methodology of the Social Sciences. New York: Humanities Press. Kim K-K (2003) Order and Agency in Modernity: Talcott Parsons, Erving Goffman and Harold Garfinkel. Albany: State University of New York Press. Köhler W (1947) Gestalt Psychology: An Introduction to New Concepts in Modern Psychology. New York: Liveright. Leiter K (1980) A Primer on Ethnomethodology. New York: Oxford University Press. Korbut 495 Livingston E (1987) Making Sense of Ethnomethodology. London: Routledge & Kegan Paul. Livingston E (2008) Ethnographies of Reason. Aldershot: Ashgate. Lynch M (1993) Scientific Practice and Ordinary Action: Ethnomethodology and Social Studies of Science. Cambridge: Cambridge University Press. Lynch M (2000) The ethnomethodological foundations of conversation analysis. Text 20(4): 517–32. Mehan H and Wood H (1975) The Reality of Ethnomethodology. New York: Wiley. Parsons T ([1937] 1949) The Structure of Social Action: A Study in Social Theory with Special Reference to a Group of Recent European Writers. Glencoe, IL: Free Press. Parsons T (1951) The Social System. Glencoe, IL: Free Press. Parsons T (1968) Order as a social problem. In: Kuntz P G (ed.) The Concept of Order. Seattle: University of Washington Press, pp. 373–84. Rawls A W (1987) The interaction order sui generis: Goffman’s contribution to social theory. Sociological Theory 5(2): 136–49. Rawls A W (1989a) An ethnomethodological perspective on social theory. In: Helm D T et al. (eds) Interactional Order: New Directions in the Study of Social Order. New York: Irvington, pp. 4–20. Rawls A W (1989b) Language, self and social order: A reformulation of Goffman and Sacks. Human Studies 12 (1–2): 147–72. Rawls A W (2005) Garfinkel’s conception of time. Time & Society 14 (2–3): 163–90. Rawls A W (2009) An essay on two concepts of social order: Constitutive orders of action, objects and identities vs aggregated orders of individual action. Journal of Classical Sociology 9(4): 500–20. Rawls A W (2010) Social order as moral order. In: Hitlin S and Vaisey S (eds) Handbook of the Sociology of Morality. New York: Springer, pp. 95–121. Rawls A W (2011) Wittgenstein, Durkheim, Garfinkel and Winch: Constitutive orders of sensemaking. Journal for the Theory of Social Behaviour 41(4): 396–418. Ruggerone L (1996) The reflexive order of language and activities: Second thoughts on Garfinkel’s ethnomethodology. International Journal of Sociology and Social Policy 16(4): 91–102. Schütz A (1945) On multiple realities. Philosophy and Phenomenological Research 5(4): 533–76. Schütz A (1953) Common-sense and scientific interpretation of human action. Philosophy and Phenomenological Research 14(1): 1–38. Sharrock W W (1999) The omnipotence of the actor: Erving Goffman on ‘the definition of the situation’. In: Smith G (ed.) Goffman and Social Organization: Studies in a Sociological Legacy. London: Routledge, pp. 119–37. Sharrock W W and Anderson R J (1986) The Ethnomethodologists. Chichester: Ellis Horwood. Skidmore W (1979) Theoretical Thinking in Sociology. Cambridge: Cambridge University Press. Wilson D N (1978) An interpretation of the work of Harold Garfinkel: A developmental, epistemological and speculative inquiry. PhD thesis, Boston University. Wilson T P (2003) Garfinkel’s radical program. Research on Language and Social Interaction 36(4): 487–94. Wilson T P (2012) Classical ethnomethodology, the radical program and conversation analysis. In: Nasu H and Waksler F C (eds) Interaction and Everyday Life: Phenomenological and Ethnomethodological Essays in Honor of George Psathas. Lanham, MD: Lexington Books, pp. 207–38. 496 European Journal of Social Theory 17(4) Wrong D H (1994) The Problem of Order: What Unites and Divides Society. New York: Free Press. Zimmerman D H and Wieder D L (1970) Ethnomethodology and the problem of order: Comment on Denzin. In: Douglas J D (ed.) Understanding Everyday Life: Towards a Reconstruction of Sociological Knowledge. Chicago: Aldine Publishing Company, pp. 285–98. Author biography Andrei Korbut is a research fellow of the Centre for Fundamental Sociology at the National Research University Higher School of Economics, Moscow, Russia. He is also a lecturer at the Moscow School of Social and Economic Sciences and a member of the editorial board of the Russian-language online journal Sociologicheskoye Obozreniye (Russian Sociological Review) (www.sociologica.hse.ru). His additional publications include translations of the works of Harold Garfinkel and other ethnomethodology authors into Russian. His areas of interest are ethnomethodology, conversation analysis, and studies of learning practices.