The Sociology of Order in Interaction

ERVING GOFFMAN

The writings of the Canadian-born sociologist, Erving Goffman, are undoubtedly among the most influential theoretical sources for the study of spoken interaction, but he is not ordinarily ranked among the major social theorists of the past century. When asked for his specific contribution, one reply will consist of a brief list of "powerful" concepts (i.e. interaction order, frame, footing and face) but it is equally true that the relevance of Goffman's work for discourse analysis is still in more than one respect left to be explored (for instance, his distinction between the front and back regions of institutional action, cast as a distinction between formal public performance (front region) and more informal back region activity "where the impression fostered by the performance is knowingly contradicted as a matter of course" (Goffman 1959:114 )).

Some of Goffman's writings can be situated in the context of a sociology of psychiatric (and other "total") institutions but conducted against the background of action-oriented descriptions of participant orientations in "normal interactive behaviour" (note that this particular combination is echoed in Sack's early conversation analytic work on suicide helpline calls and in Garfinkel's work on psychiatric clinics). Note in this respect Goffman (1967b:47-48):

Data for the paper are drawn chiefly from a brief observational study of mental patients in a modern research hospital. I use these data on the assumption that a logical place to learn about personal proprieties is among persons who have been locked up for spectacularly failing to maintain them. Their infractions of propriety occur in the confines of the ward, but the rules broken are quite general ones, leading us outward from the ward to a general study of our Anglo-American society.

Thus, one will find in Goffman's writings not only a critique of pyschiatry as moving far too easily from social delicts to mental symptons, resulting in a failure to assess the impropriety of acts that bring a person under the attention of the psychiatric care (see especially Goffman 1967c: 137ff. ), such a critique is also couched in a programmatic enquiry into the construction and maintenance of the self in the rituals of face-of-face interaction. Goffman's sociological method has been influenced by phenomonology. It borrowed in particular from the work of A. Schutz on interactive relations, commonsense understanding via types and the situational character of relevance. Together with Harold Garfinkel, Goffman is at the basis of ethnomethodology and its further spin-off conversation analysis.

Interaction order. According to Anthony Giddens, Goffman's detailed enquiries into the micro-forms of practical interactional behaviour are informed by a set of systematic social theoretical assumptions. Goffman's sociology is one which centers on physical co-presence rather than on social groups. While groups continue to exist when their members are not together, encounters, by definition, only exist when the parties to them are physically in each other's presence (Giddens 1988 ). This helps explain Goffman's interest in

• particular settings (e.g. talk between strangers on a platform waiting for a train, how entering an elevator affects talk, etc.)
• forms of self-maintaining behaviour such as 'civil inattention' or the display of 'focused interaction' (e.g. forming a 'huddle' during a reception)
• conduct in public in situations involving embarrassment, face-saving behaviour and/or public displays of competence, for instance response cries (incl. 'spill cries' such as Oops); Goffman interprets response cries as an other-oriented form of self-talk in which "we display evidence of our alignment we take to events" (Goffman 1981:100 )
• the role of temporal and spatial activity boundaries which result in inclusion (e.g. how to display intense involvement in talk) and exclusion (e.g. how the exclusion from talk of others in physical proximity is maintained simultaneously by the included and the excluded; see also some the distinctions made under the heading of footing below)

It is this concern with situated activity systems which are manifest in behaviour under conditions of co-presence which are unrelated to transsituational groups or institutional forms of membership that can be captured under the heading of an interaction order (Goffman 1983:2). 

My concern over the years has been to promote acceptance of this face-to-face domain as an analytically viable one - a domain which might be titled, for want of any happy name, the interaction order - a domain whose preferred method of study is micro analysis.

For Goffman, talk is the basic medium of encounters, but talk isn't all: the state of co-presence is a physical state and it draws attention to the body, its disposition and display. For Goffman, "the body is not simply used as an 'adjunct' to communication in situations of co-presence; it is the anchor of the communicative skills which can be transferred to disembodied types of messages." (Giddens 1988:257 ) In this respect, Goodwin (2000:1491 ) talks about the need to investigate "the public visibility of the body as a dynamically unfolding, interactively organised locus for the production and display of meaning and action." Not surprisingly, such a programme will be critical of mentalist versions of meaning and cognition. According to Collins (1988: 51-52), the aim is to arrive at the social ecology which is at the basis of any conversational situation. For Goffman, 

the basis of language is not a primal intersubjectivity, a meeting of minds, but rather a common focus on a physical scene of action. [...] Goffman is looking at human beings the way a biologist would look at birds or mammals who are in range of each other. Goffman asserts that people must always pay attention to other human beings in their presence; each one needs to check out the others, if only to see if it is safe to ignore them.

 

One weakness in Goffman's concept of an interaction order lies in the passive claim to its universality and autonomy. Giddens (1988:279 ) talks about the need "to think rather in terms of the intersection of varying contexts of co-presence, knit together by the paths that individuals trace out through the locales in which they live their day-to-day lives" so as to shed light on the "modes in which everyday social activity is implicated in very broad patterns of institutional reproduction".

 

Goffman's frame analysis is essentially about how social actors organise their experience in terms of recognisable activities (e.g. a game of chess, a conference talk, a hold-up, etc.). A discussion of the concept can usefully begin with a reference to what Schutz called "frames of reference", i.e. a set of connections among objects, events, behaviours, etc. constituted as an anonymous and recognisable structure of relevancies. Goffman's development and use of the concept particularly brings out the multidimensional character and layeredness of such frameworks - frame built upon frame. He draws our attention to the interplay between primary frameworks and the constructed frameworks of social relationships (e.g. the availability of an open-air sound system with amplification which brings a single speaker into the earshot of a mass audience as a primary framework onto which the activity of making a speech during a political rally in achored; another example would be an analysis of the activity of flirting as layered upon the frame space of physical movements in performing a dance, which, in its turn, is an activity patterned upon the primary framework of a particular type of music being played). It is in such moves as the stand-up comedian's comments which draw the late arrival of a spectator into the anecdote being told on the stage that one can see how interactants attend to the layeredness and interdependence of frames and play with the boundaries. Goffman's frame analysis seeks to draw attention to the ease with which people handle multiple, interdependent realities. (CLICK HERE for example analyses of how actors establish frames, manage multiplicity and potential disruptions).

Layeredness is also presupposed by the attendant notion of key(ing). Key refers here to a set of conventions by which a given activity (which is already meaningful in terms of a primary framework) is transformed into something patterned on this activity but seen by the participants to be something quite else (Goffman 1974:43-4 ). Here we see again Goffman's self-declared interest in "game play", in con games, deceit, fraud, shows, etc. Keying brings the frame laminations of scripted situations of make-believe for purposes of entertainment into the scope of analysis, but note that keying is equally central to the occurrence (and meaningful recognition) of playful, ironic moves in 'ordinary' interaction (e.g. an overtly exaggerated and therefore insincere love declaration between two friends acted out in front of a larger company).

Frame analysis reveals the complexity of mundane social activities and it brings out the arbitrary nature of any fixed, social-domain or activity-based dichotomy between what is "staged" and what is "real". It brings out the reality-constructing capacities of what is staged, but also the staged nature of the everyday tangibly real. Note in this respect for instance that mass-media communication - including especially the solidly real called "news broadcasting" - is saturated by frame laminations which are deliberately and purposefully staged. What's more, an understanding of media communication is rather hard to arrive at, unless one comes to terms with the constructed pretense of an absence of mediation and the audiences' routine submission to an illusion of direct communication - even in situations where such a pretense becomes extremely hard to sustain (a well-known example in this respect is the live coverage of UN-troups landing on the shores of Somalia in 1994; this coverage was made possible by the primary frameworks provided by the various television crews which had landed ahead of the troops, they were waiting for the troops ready with camera, spotlights, sattelite connections, etc.). More generally we can observe with Collins (1988:61) that 

Framing permeates the level of ordinary social action. We live in a world of social relationships, in which roles are acted out, with various keyings and deceptions played upon them. This is the core of practical activities and occupations, of power and stratification. Here again, Goffman leads us to the brink of seeing the micro-reality upon which macro-structures are based, though he shies away from the theoretical implication.

Thus, in social-theoretical terms, Goffman's frames occupy a middle ground which avoids both the extremes of the total relativism of sheer intersubjectivity (whatever an actor appears to construct at the time) and that of the objective determinism of a reality which is pre-given and external to the actor (as if it would be enough for students and lecturer to enter the lecture hall for a lecture to take place). Frame analysis precisely invites attention to how the pre-given and the locally-constructed interrelate. Its particular analytical purchase lies in how interactants attend to the simultaneity of multiple realities, how they adjust constructions, manage disruptions, etc. Hence, not surprisingly, Goffman's interest in, on the one hand, impostors, spies, theatrical performances, etc. and, on the other hand, the acting out of sociability in talk, i.e. phenomena which reveal the transformation of ordinary action into things seen in a different light. Thus, we read in the opening paragraphs of the introduction to Goffman (1981:2-4) 

In what follows, then, I make no large literary claim that social life is but a stage, only a small technical one: that deeply incorporated into the nature of talk are the fundamental requirements of theatricality.

"Frame space" offers a more precise perspective on the nature of norms in interaction. In this view, norms are not learned rules which speakers carry around in their heads, but they are ways in which situations unfold, so that participants feel they have to behave in a particular ways or make amends for not doing so (cf. Collins 1988:57 ).

Additionally, one can interpret Goffman's concern with how activities are layered upon primary frameworks as an invitation to study in greater detail how technology-dependent "channels of communication" constrain discourse practice, in ways which take researchers beyond essentialist distinctions in "medium" such as the one between "speech" and "writing" (for instance, how else can one assess what the condition of talking over the telephone does to interaction?; cf. Slembrouck (1995 ); at the same time, the invitation is for us to move beyond taken-for-granted assumptions about what counts as a relevant constraining frame. Should the latter be restricted to "channels of communication" or simply entail any conditions of framing - irrespective their nature? For instance, how does talk unfold differently if it goes together with an on-line engagement with textual attributes or the handling of objects while performing a particular professional task?

Finally, striking a more critical chord, frame analysis also exposes the limitations of an autonomous conception of the interaction order. One can think here of the "framing" capacities of code selection, switching and slippage, including the boundaries which are imposed as result on the participation framework(s) which apply. For instance, code selection has primary frame function in establishing the secondary framework of a language class and, in such contexts, code switching will often mark the transition from one activity to another or bring out a clash of frames - CLICK HERE for an example of the latter). Considerations of this kind take one beyond the conditions of physical co-presence, venturing instead into the role of transsituational processes of identity-formation which connect to groups definable under conditions of "virtual co-presence" or even "absence". 

The concept of footing is closely connected to that of frame and the taxonomies it has given rise to. It reiterates Goffman's interest in boundaries, centre, margin, focus and (dis)engagement in interaction under conditions where talkers and hearers share a physical space, while not necessarily engaged in one-and-the-same conversation or activity. A social situation is thus defined in Goffman (1981:84) in terms of a gathering and a zone of behaviour and action earmarked as the "full physical arena":

[…] by a social situation I mean any physical area anywhere within which two or more persons find themselves in visual or aural range of one another. The term "gathering" can be used to refer to the bodies that are thus present. No restriction is implied about the relationship of those in the situation: they may all be involved in the same conversational encounter, in the sense of being ratified participants of the same state of talk; some may be in an encounter while others are not, or are, but in a different one; or no talk may be occurring. Some, all, or none of those present may be definable as together in terms of social participation, that is, in a "with".

Footing then stands for a speaker's and hearer's shifting alignments in relation to the events at hand, as a combination of production/reception format and participation status. A change in footing 

[...] implies a change in the alignment we take up ourselves and the others present as expressed in the way we manage the production or reception of an utterance. A change in footing is another way of talking about a change in our frame for events. This paper [i.e. the essay on "footing"] is largely concerned with pointing out that participants over the course of their speaking constantly change their footing, these changes being a persistent feature of natural talk.

Thus, 'footing' brings out the need to distinguish between various speaker roles, for instance, when dealing with phenomena such as speech report, being a messenger and other types of situations where a person speaks on behalf of someone else. Goffman suggests, the 'speaker' can be replaced by a 'production format' which consists of three components:

• animator: the person who makes the sounds, not a social role in the full sense, but a functional node in a communication system (language at the level of the "body" producing sound).
• author: the person who selected the sentiments expressed and the words in which they were put (language at the level of "mind").
• principal: the person whose position is established by the words which are spoken and who is committed to the beliefs expressed (language at the level of a person active in a social role or identity).

These distinctions are not only instrumental in explaining some of the differences and shifts in speakerhood one commonly comes across when one is acting as, say, a spokesperson speaking on behalf of someone else, a reporter who is repeating what someone else has said, etc. or when a person shifts between such positions within a single stretch of discourse. An analysis of production format is thus of primordal importance when one is facing situations where discourse production depends on coordinated team work with a particular division of labour which is to lead to a finalised product, say, a television commerical, where a voice-over takes on the role of 'animator' for a message scripted by an advertising agency ('author') which expresses the position of the manufacturer ('pincipal').

Goffman proposes similar distinctions following a deconstruction of the category of hearer which reveals a range of participation statuses. This way, Gofmann distinguishes between the primary addressee (a ratified hearer) and an overhearer (typically a non-ratified hearer, for instance, an accidental bystander). Of course, ratified participants may be turned into bystanders (e.g. when the talk becomes so centred on two interactants and chills the involvement of others present in the frame). The reverse may also occur, for instanse, when a response cry - typically aimed potential overhearers (though not a specific hearer), becomes the starting point for a conversation. (CLICK HERE for a schematic representation of additional distinctions relevant for the study of the dynamics of participation frameworks).

In many respects the essay on 'footing' can be read as a challenge to unwarranted theorising which limits a communicative situation to the presence of a speaker and a hearer engaged in a conversation. Essentializing speaker and hearer suggests that only sound is at isssue, whereas in fact sight is organisationally very significant, too - sometimes also touch (e.g. when at a convivial dinner table with competing and quite freely shifting participation frameworks, speakers feel it necessary to police listenership and rely on pitch raising, interruption, bodily orientation and touch to "bring back strays and encourage incipient joiners", 1981, p. 135). Similarly, the development of the category of overhearer/bystander (an original contribution by Goffman) is the communicative 'anomaly' which forces us to consider the facts of interaction as relative to a 'gathering' rather than an 'encounter': "[...] in dealing with the notion of 'bystanders', a shift was tacitly made from the encounter as a point of reference to something somewhat wider, namely 'the social situation,' defining this as the full physical arena in which persons present are in sight and sound of one another. (Goffman 1981: 136 ). The sections in the essay titled 'Footing' which deal with podium events (one-to-many) and the one which enquires into activities where non-linguistic activities provide the organising context for utterances (e.g. talk connected to an extended joint task such as repairing a car) can be read in the same way.

Footing offers a more precise perspective on the nature of role behaviour, which, on finer examination, as Collins (1988:57 ) points out, really consists of multiple voices and a way in which changes in footing are managed. The concept has been developed further in pragmatics (e.g. Levinson 1983, 1988 , Thomas 1986 ) and within linguistic anthropology - for instance, Hanks (1996:219-10), who captures its importance for enquiries into language use well: 

For our purposes, the details of Goffman's typology are less important than that it pushes beyond the simple dyad and opens up the possibility of a differentiated approach to multi-party talk. [...] it became inevitable that the speaker-addressee dyad would lose its place as the measure of all talk. Thus it is not accidental that the decomposition of the roles into animator, author, principal, addressee and overhearer was part of a broader push toward the study of different types of interaction, including multi-party talk. Once the boundaries of the dyad were breached, the inadequacy of its two parts became all too obvious. It is also clear in this framework that the relation between an individual and the language he or she speaks is mediated by social roles. [...] You simply cannot make inferences from utterance forms to human experience without working through the intermediate level of participant frameworks.

In a seminal article, Goffrman defines face as (1967a:5)  

the positive social value which a person effectively claims for himself by the line others assume he has taken during a particular contact. Face is an image of self delineated in terms of approved social attributes.

The term line here refers to a pattern of verbal and nonverbal acts by which someone expresses their view of the situation and its participants, especially him/herself. Quite importantly, such a line is seen in terms of effects which are ascribed to an individual as wilfully intended: 

Regardless of whether a person intends to take a line, he will find that he had done so in effect. The other participants will assume that he has more or less wilfully taken a stand, so that if he is to deal with their response to him he must take into consideration the impression they have possibly formed of him.

Interactants expect each other to behave in a way which is consistent with this image (to be in face or maintain face). Participants may find themselves in wrong face when information is revealed which cannot be integrated into the line adopted thus far or they may find themselves out of face when participating in a contact without having ready a line of the kind participants are expected to take in such situations.The latter two correspond to the everyday expression "to lose face" in Anglo-American society, but Goffman als draws attention to other culture-specific usages such as "to give face" which occurs when one speaker arranges for another to take a better line than he would otherwise have been able. Goffman includes social valuables such as dignity, honour, self-respect and contempt within the scope of an analysis of face-work: "[W]hile his social face can be his most personal possession, and the centre of his security and pleasure, it is only on loan from society; it will be withdrawn unless he conducts himself in a way that is worthy of it" (1955:213 ). In a sense then, Goffman's essay takes us in more than one respect beyond the concerns of Brown & Levinson's programmatic formulation of an enquiry into the mutual satisfaction of face wants in relation to (verbal) politeness strategies adopted in the production of individual utterances. The intermediary notion of "line" continues to remain undeveloped - it raises issues of qualitative diversity in the occurrence of ratifiable situated personae with corresponding face work strategies. Nor has the thematic lead into the moral character of the interaction order been taken up in linguistic politeness enquiry. One can also think here of the sections which Goffman devotes to the "agressive use of face work" which occurs when (1967a:24) 

the purpose of the game is to preserve everyone's line from an inexcusable contradiction, while scoring as many points as possible against one's adversaries and making as many gains as possible for oneself. An audience to the struggle is almost a necesssity.

In conclusion, note that Goffman's essay formed part of an enquiry in to the ritual roles of the self, where the self is seen both as a kind of image pieced together from the flow of events and as a kind of player in a ritual game. It leads to a particular view on human nature as a construct - for instance, by linking perception of what persons are "really" like with a repertoire of face-saving devices as situationally-called for. In more than one respect, one can trace echoes in this of Foucault's concept of individual subjectivity as a by-product of and as a function of discourse regimes (1967a:44-5): 

Throughout this paper, it has been implied that underneath their differences in culture, people everywhere are the same. If persons have a universal human nature, they themselves are not to be looked to for an explanation of it. One must look rather to the fact that societies everywhere, if they are to be societies, must mobilise their members as self-regulating participants in social encounters. One way of mobilising the individual for this purpose is through ritual: he is taught to be perceptive, to have feelings attached to self and a self expressed through face, to have pride, honour, and dignity, to have considerateness, to have tact and a certain amount of poise. These are some of the elements of behaviour which must be built into the person if practical use is to be made of him as an interactant, and it is these elements that are referred to in part when one speaks of universal human nature.

CONVERSATION ANALYSIS

 According to Deirdre Boden , conversation analysts are sociologists who have turned the problem of social order upside down. For them, the crucial question is not how people respond to a social order and its normative constraints, but rather how that order is brought about in a specific situation, through activities in quite specific time and place. To understand the orderliness of social life, one does not need abstraction and aggregation, but instead one must turn to the finegrained details of moment-to-moment existence and their sequential organisation.  

In other words, conversation analysts can be seen as sociologists who assume that everyday social structure is a skilled accomplishment by competent actors (cf. ethnomethodology). Conversation is one such type of action and one which is particularly salient in social terms. Moreoever, conversation can be recorded and described in detail, with transcriptions providing a yardstick for the replicability of social-scientific analysis. Thus, one version of the birth of conversation analysis - voiced in Harvey Sacks's methodological notes (1984:26) - links the technological advances of the 1960s in an accidental fashion to the fulfilment of a particular methodological project (critical of Weber, critical of Durkheim).  

When I started to do research in sociology I figured that sociology could not be an actual science unless it was able to handle the details of actual events, handle them formally, and in the first instance be informative about them in the direct ways in which primitive sciences tend to be informative, that is, that someone else can go and see what was said is so. [...] It was not from any large interest in language or from some theoretical formulation of what should be studied that I started with tape-recorded conversation, but simply because I could get my hands on it and I could study it again and again, and also, consequentially, because others could look at what I had studied and make of it what they could, if, for example, they wanted to be able to disagree with me.

At what point exactly this preference for tape-recorded data became invested with an actively-developed suspicion towards the use of 'unsatisfactory data sources' in language description (e.g. interview data, observational data obtained through field notes, invented examples and experimental elicitation) is a matter of hindsight interpretation. For its practitioners, conversation analysis is a stake within the methodology debate of sociology (where does one situate one's sociological object of enquiry, i.e. society), with a clear preference for the formal properties of social action, and, in many cases, a suspicion towards any kind of pre-analytical theorising. For instance, conversation analysts often state their reluctance to allow categories to enter the analysis other than those entertained by the participants or revealed in an analysis of the sequential flow of interaction. This point is often captured through the ironicising image of rejecting a "bucket"-theory of context which is contrasted with the preferred view of context as a project and a product of interaction. Note in this respect also the following proviso which John Heritage (1997:168) builds into a discussion of the structural organisation of turn-taking: 

Overall structural organisation, in short, is not a framework - fixed once and for all - to fit data into. Rather it is something that we're looking for and looking at to the extent that the parties orient to it in organising their talk.

This second, etnomethodological trait within conversation analysis is distinctly phenomonological (compare with the influence of Schutz's writings on Goffman). It marks a concern with the competences (seen as participant orientations) which underlie ordinary social activities. Note also how this trait receives a radical - and ultimately untenable - reading in the context of the inductivist positivism which characterises Sacks's formulation of the ideal of 'primitive science': for instance, Boden will have it that a description of the social-order-produced-in-context is a member's construct, not an analyst's construct.

• One central concept within conversation analysis is the speaking turn. According to Sacks et al. (1974 ), it takes two turns to have a conversation. However, turn taking is more than just a defining property of conversational activity. The study of its patterns allows one to describe contextual variation (examining, for instance, the structural organisation of turns, how speakers manage sequences as well as the internal design of turns). At the same time, the principle of taking turns in speech is claimed to be general enough to be universal to talk and it is something that speakers (normatively) attend to in interaction.
• A second central concept is that of the adjacency pair. The basic idea is that turns minimally come in pairs and the first of a pair creates certain expectations which constrain the possibilities for a second. Examples of adjacency pairs are question/answer, complaint/apology, greeting/greeting, accusation/denial, etc. Adjacency pairs can further be characterised by the occurrence of preferred or dispreferred seconds (CLICK HERE for examples). A frequently-used term for this is preference organisation.
• The occurrence of adjacency pairs in talk also forms the basis for the concept of sequential implicativeness: each move in a conversation is essentially a reponse to the preceding talk and an anticipation of the kind of talk which is to follow. In formulating their present turn, speakers show their understanding of the previous turn and reveal their expectations about the next turn to come. This is often singled out as conversation analysis's most important insight, viz. that actors, in the course of interaction, display to each other their understanding of what they are doing - an insight which can be traced to phenomenology's belief that actors maintain an awareness of their own actions, and it is this awareness which is displayed to the other party (see Goodwin 1990 , Hanks 1996:218 ).

/ The major strength of conversation analysis lies in the idea that an important area of interactional meaning is revealed in the sequence. Its most powerful idea is undoubtedly that human interactants continually display to each other, in the course of interaction, their own understanding of what they are doing. This, among other things, creates room for a much more dynamic, interactional view on speech acts than is enabled by analytical philosophy and mainstream pragmatics (Click HERE for a contrastive analysis of two exchanges in this light). Yet, note in one and the same breath, that there is a problem over the the kind of participant outlook that tends to be presupposed in conversation analysis. Margaret Wetherell (1998:402) captures this well when she conducts the "debate" between conversation analysis and post-structructuralism:  

If the problem with post-structuralist analysis is that they rarely focus on actual social interaction, then the problem with conversation analysis is that they rarely raise their eyes from the next turn in the conversation, and, further, this is not an entire conversation or sizeable slice of social life but usually a tiny fragment. [...]

In short, the problem might be what in practice is regarded as a sequence and why the "sequentially implicated" cannot also include linkage in the form of a display of uptake of what was said during a previous, related ocassion of talk (e.g. through the use of reported speech, speakers may simultaneously display their understanding of the immediately preceding turn in the conversation into which they are engaged and their understanding of what occurred during a previous conversation or textual experience). One may also take a here lead from recent research into complex technological work environments, for instance, a railway control room (Hindmarsh & Heath 2000:76), in which  

personnel have, as a matter of their daily activities, to make reference to, and mutually constitute, the sense and signficance of a continually changing range of 'objects' displayed on screens and in documents [and] talk and bodily conduct is used within organisational activities to produce coherent and sequentially relevant objects and scenes

Such research suggests that sequence in discourse may be understood in an extended sense, consisting of various "tracks" involving different modalities and characterised by the occurrence of selectively shifting attention producing moments of divergence and convergence with the conversational track(s) which the conversation analyst habitually attends to. By the same token, a shared television experience involving two viewers before a screen (say, watching a football game) may be understood in terms of conversational turns being implicated in both the sequence which is produced between the two talkers and the sequence of images and words coming from the television set and which the talkers' turns show their particular understanding of. Note how this point in one sense at least can amount to closing some of the gaps which over the years have separated conversation analysis from Goffman's early work. So, while conversation analysis originated in Goffman's project, it is also true that it resulted in a narrowing of relevant concerns. Giddens (1988:266) captures both. One half of the observation marks the departure from more narrowly 'linguistic' preoccupations:  

Talk is the basic medium of focused encounters and the conversation is the prototype of the exchange of utterances involved in talk. Using the word 'talk' rather than 'language' is of the first importance to the analysis which Goffman seeks to provide. 'Language' suggests a formal system of signs and rules. 'Talk' carries more the flavour of the situated nature of utterances and gestures embedded within the routine enactment of encounters. In speaking of verbal 'conversations', rather than of 'speech', Goffman stresses that the meaning of what is said must be interpreted in terms of a temporal sequence of utterances. Talk is not something which is just 'used' in circumstances of interaction. [...] In his early work, Goffman both anticipated and helped shape the development of what has subsequently come to be called 'conversation analysis'; in his later writings he has drawn upon it in developing hiw own discussions of talk and interaction.

The second half of the observation follows on the next page (Giddens 1988:267), where we read that  

Goffman is able to show some of the limitations involved in thinking of talk in terms of statements calling forth replies. Many moves do seem to invoke rejoinders, but there are a variety of ways in which individuals can express intentions, provide approval or disapproval, or otherwise make their views known, without directly committing themselves to turn-taking within the conversation. A key aspect of all talk in situations of interaction is that both speakers and listeners depend upon a saturated physical and social context for making sense of what is said.

Coming to terms with this idea of the "saturated physical and social context" constitutes one of the biggest challenges facing the discourse analyst. This is not only a matter of what is included (a range of observable phenomena such as talk, gesture, posture, objects which frame action, etc.) but also how we understand their manifestations and relevance. One promising possibility in this respect lies in a dynamic understanding of context, i.e. context itself as "sequentially implicated" but without conversation analysis's overt restriction of "context" to the "surrounding talk". As Goodwin (2000:1519-1520) suggests, context, in such a view,  

is not simply a set of features presupposed or invoked by a strip of talk, but a dynamic, temporally unfolding process accomplished through the ongoing rearrangement of structures in the talk, participants' bodies, relevant artefacts, spaces, and features of the material surround that are the focus of the participants' scrutiny. Crucial to this process is the way in which the detailed structure of talk, as articulated through sequential organisation, provides for the continuous updating and rearrangements of contexts for the production and interpretation of action. [my emphasis]

However, note that such a concern with a dynamic understanding of the processes through which social and physical context are actively constructed by participants (i.e. as displayed to one another) still says very little about the analyst's active role in the construction of "contexts-as-researched" (compare with natural histories of discourse).

/ A one-sided priority on a participant's understanding of what goes on in interaction and what that interaction is about also constitutes a pitfall in its own right, if it means that common sense categories and understandings of interactional purposes, goals and orderliness are mistaken for exhaustive explanations of why discourse displays the properties it does. While it is true that conversation analysis quite rightfully warns against the risks inherent in a macro leap which erases the participant from the picture of analysis and reduces the signficance of talk to a mere reflection of an a priori societal and/or cultural context, the converse holds equally in the case of a failure to recognise contextual relevance beyond the purview of the participant's "local talk"-orientations in discourse as well for assuming automatically that the categories revealed in the ways that speakers can be shown to attend to a local conversational sequence are necessarily also members' categories or exclusively members' categories. Wetherell (1998:402-3) observes:  

Schegloff argues that analysts should not import their own categories into participants' discourse but should focus instead on participant orientations. Further, analytic claims should be demonstrable. Schegloff's notion of analytic concepts uncontaminated by theorists' categories does not entail, however, that no analytic concepts will be applied [...] Rather, concepts such as conditional relevance, for example, or the notion of accountability, or preferred or dispreferred responses are used to identify patterns in talk and to create an ordered sense of what is going on. Presumably Schegloff would argue that this does not count as imposing theorists' categories on participants' orientations since such concepts are intensely empirical, grounded in analysis and built up from previous descriptive studies of talk. [...] the advantage for Schegloff of such an approach is that it gives scholarly criteria for correctness and grounds academic disputes, allowing appeals to data, and it closes down the infinity of contexts which could be potentially relevant to something demonstrable - what the participants take as relevant. [... However,] one problem from a critical perspective is that Schegloff's sense of participant orientation may be unacceptably narrow. [...] in practice for Schegloff, participant orientation seems to mean only what is relevant for the participants in this particular conversational moment. Ironically, of course, it is the conversation analyst in selecting for analysis part of a conversation or continuing interaction who defines this relevance for the participant. In restricing the analyst's gaze to this fragment, previous conversations, even previous turns in the same continuing conversation become irrelevant for the analyst but also, by dictat, for the participants. We do not seem to have escaped, therefore, from the imposition of theorists' categories and concerns.

The reluctance to admit the presence of 'pre-existing' categories in conversation analysis also brings with it a number of methodological uncertainties. For instance, with reference to a short telephone exchange in which a school employee rings a mother whose son may be a truant from school, John Heritage (1998:163) argues: 

The assumption is that it is fundamentally through interaction that context is built, invoked and managed and that it is through interaction that institutional imperatives originating from outside the interaction are evidenced and are made real and enforceable for the participants.

The question can be raised whether such 'imperatives originating from outside' are to be identified prior to the analysis (cf. "I am looking at a particular type of institutional routine or exchange") or should only be recognised as existing because they emerge from an analysis of interactive data (cf. "the data of the exchange tell me that I am dealing with a particular type of routine"). Is it at all feasible to separate these two moments of 'categorisation'? Even when such issues are understood in phenomenological rather than in positivist terms, a recognition of an "ethnographic moment" is in place here.

ETHNOMETHODOLOGY

Social reality is, for ethnomethodology, an intersubjective accomplishment. In the words of Harold Garfinkel (1972:309): 

I use the term ethnomethodology to refer to various policies, methods, results, risks, and lunacies with which to locate and accomplish the study of the rational properties of practical actions as contingent ongoing accomplishments or organised artful practices of everyday life.

Heritage (1984:4,8) paraphrases Garfinkel's subject matter as "mundane knowledge-in-action" and as "institutionalized conduct". Ethnomethodology's subject matter is: 

[...] the body of common sense knowledge and the range of procedures and considerations by means of which the ordinary members of societies make sense of, find their way about in, and act on the circumstances in which they find themselves.

For Francis and Hester (2004:23 ) ethnomethodology thus focuses on ordinary observational competencies: competent participation in a social setting demands of those involved that they pay attention to and make sense of what is happening around them. In this sense, observation is not so much a sociological technique in as much as it is a inevitable and necessary part of competent participation in everyday life. Consequently, ethnomethodology is viewed as an approach which takes seriously the implications of the routine observability of social activities (the moment-to-moment orientations of actors) both as a potential for inquiry and as a precondition for the construction of any type of knowledge. This yields a reflexive research programme in which the analysis of the observable features of social life in terms of situated productions by those who are party to it applies equally to sociological endeavour itself. It is the latter aspect of ethnomethodological enquiry which has led Aaron Cicourel to systematically develop a socio-cognitive programme of enquiry which is oriented to understanding the construction of professional, institutional domain-specific and scientific forms of knowledge as situated productions and which, echoing Schutz (1943 ), entails a view of rationality which is relative to "typified conceptions of practical adequacy" (1973:23): 

Whatever is "rational" about administrative activities, law statutes, household budgeting, or fertility behavior, always occurs in the context of everyday experience and typifications based on past conduct. Schutz suggests that the actor makes sense of his environment by creating loose equivalence classes (as opposed to clear-cut true-false categories). The typifications in the actor's stock of knowledge interacts with emergent meanings from a setting to concretize what is likely or unlikely. [...] The practical ratoinality exhibited by the actor is an emergent phenomenon and is not directed by fixed modes of orientation; instead, the actor invokes rules in creating accounts of "what happened".

Included within the scope of such a programme is any enquiry into "meaning" when language or interactional data is used for the purposes of qualitative sociological research (e.g. Cicourel's (1973 ) critique of questionnaires and interview data as intailing a set of interactional imperatives) or when language/inderactional data is examined in its own right (e.g. for the purposes of developing a theory of semantics, an argumentation developed in Cicourel (1974 ). That the study of how everyday practical reasoning is constitutive of all human activity, for Cicourel, comes with a number of basic considerations about the meaning of everyday talk (1974:1563): 

[O]ur ability to construct machines or develop complex logical systems always presupposes a necessary reliance on the presuppositions of practical or mundane reasoning with its constraints of indexicality and reflexivity that are inherent in the development and in all uses of the sciences of the artificial. [...] If we hope to construct a theory of meaning that enables us to understand how we assign sense to our everyday worlds and establish reference, then we cannot assume that oral language syntax is the basic ingredient of a theory of meaning. The interactional context, as reflexively experienced over an exchange, or as imagined or invented when the scene is displaced or is known through a text, remains at the heart of a general theory of meaning.

While ethnomethodology clearly has affinities with ethnography (e.g. it prioritises observation, immersion, contextual dependency and insists on explicating the methods used by members to produce "from within"), it also shares a set of phenomenological assumptions with conversation analysis, even if the question of text and talk comes with different emphases. First. Although it does not exclude detailed, turn-per-turn sequential analysis of talk, ethnomethodology is much more orientated towards examining the role of text and talk in the daily accomplishment of institutional goals and actions, i.e. what Rod Watson calls "texts as active social phenomena". Let me recall the impressively informative list at the beginning of Watson (1987:80): 

Tattoos, bus tickets, pay slips, street signs, time indications on watch faces, chalked information on blackboards, computer VDU displays, car dashboards, company logos, contracts, railway timetables, television programme titles, teletexts, T-shirt epigrams, 'on'/'off' switches, £10 notes and other bank notes, passports and identity cards, cheques and payslips, the Bible, receipts, newspapers and magazines, road markings, computer keyboards, medical prescriptions, birthday cards, billboard advertisements, maps, Hansard, graffiti on walls, music scores, church liturgies, drivers' licences, birth, marriage and death certificates, voting slips, degree certificates, book-keepers' accounts, stock inventories, cricket scoreboards, credit cards - these and countless other items that involve written language and diagrammatic forms indicate the immensely pervasive, widespread and institutionalised place of texts in our society. This list also indicates the extraordinary diversity in the work done by texts - contractual commitments, ratifying work, facilitating work, record-keeping, persuasive work, identity-establishing work, and so on. In fact, one might suggest that virtually every recognisable activity in our society has its textual aspects, involving and incorporating people's monitoring of written or textual 'signs' - texts that, in a wide variety of ways, help us to orientate ourselves to that activity, occasion or setting and to make sense of it.

Secondly, ethnomethodology is more about the moral-performative dimensions of text and/or talk, e.g. reporting as a persuasive display of professional competence which renders members' actions accountable. Again, in the words of Garfinkel (1972:323, see also Garfinkel 1967:vi ), the assumption is indeed that: 

Any setting organises its activities to make its properties as an organised environment of practical activies detectable, countable, recordable, reportable, tell-a-story-aboutable, analysable - in short, accountable.

Thirdly, and finally, ethnomethodology's focus on understanding the practical rationalities of members' moment-to-moment orientations-in-action is much more sensitive to a programme which seeks to elucidate field-specific frameworks of meaning-making. Such an analytical stance is typically captured through the phrase that utterance 'a' or interactional move 'b' will be "heard" as indexing 'x' (i.e. is viewed as 'existentially' tied to a set of contextual considerations) and this takes us beyond conversation analysis' interest in how the interactional sequence per se produces its own context (compare 'hearing a social worker's utterance as interactionally communicating a diagnosis' and 'hearing a social worker's utterance as interactionally establishing a diagnosis which is specific to a child protection procedure').

[continue to LIST OF REFERENCES]

[continue to ACKNOWLEDGEMENTS]