Interaction Analysis has been used to analyze video records for many different purposes and on a large variety of topics. In the last few years, increasingly, Interaction-Analytic case studies have appeared and there are now some collections that contain such studies, often mixed with analyses that use related conversation analysis or ethnomethodological approaches (e.g., Button, 1993; Chaiklin & Lave, 1993; Engeström & Middleton, 1993; Fisher & Todd, 1983; Greenbaum & Kyng, 1991; Hopper, 1991). However, discussions of the specific methods used for analyzing videotapes do not yet exist. This is due, at least in part, to the fact that practitioners are still much more concerned with the practice of doing Interaction Analysis than with the process of describing it. As is to be expected for an emerging field, much of the expertise in and results of Interaction Analysis have circulated informally within a community of practitioners, made available in the doing, not through reading. Nevertheless, we believe that something can be said at this point about what makes Interaction Analysis unique and differentiates it from other kinds of video analysis. In the following sections, we attempt to specify some of the practical expertise, the accumulated body of wisdom regarding productive and not-so-productive ways of looking at tapes that has grown up within our community. We talk about these orientations, these ways-into-a-tape, as "analytic foci," in intended distinction from "analytic categories" or "coding categories." Analytic foci are simply ways of looking that are quite consistently employed in Interaction Analysis because they have turned out to be relevant again and again in our practice. The following sections highlight a limited number of them. Again, we rely on extended citing of examples from our own and our colleaguesâ analysis work.

6.1 The Structure of Events

Chronological time provides analysts with a standardized time line for the activities they observe on tapes. Yet, people's experience is of time bunched into "events." Events are stretches of interaction that cohere in some manner that is meaningful to the participants. Some events have names and constitute recognizable, culturally significant tokens in social intercourse. In this culture, for example, we recognize meals, tutoring sessions, bedtime stories, medical consultations, design sessions, and so on. In the course of analysis, smaller units of coherent interaction within events are identified, such as "setting the table" or "serving the coffee" in a meal, or "history-taking" and "advice-giving" in a medical consultation. In our practice, we have often called such easily identifiable behavioral units "ethnographic chunks." Identifying ethnographic chunks is a possible first step towards analysis and may often overlap with content logging.

Bamberger and Schön (1991) describe this process as they try to understand the videotaped activities of two adults who are engaged in constructing a tune. They write:

The most powerful strategy we found as a starting point for our analysis was chunking the protocol. This involved looking for what seemed important boundaries that articulated observable phases or organic chunks within the continuing course of participants' work (p. 187).

In the first few passes they simply tried to mark when "something new" happened, searching for boundaries without trying to be explicit about the criteria they were using or exactly what sorts of behaviors they thought were signaling the boundaries they found. Later they went back and tried to identify the criteria they had spontaneously used.

In the process of identifying events and other ethnographic chunks, analysts clearly draw upon their own cultural knowledge. For tapes from settings with which analysts are not intimately familiar, more extensive and more careful ethnographic fieldwork is necessary. It may also be useful to involve local experts in the analysis. As we have noted, one way of doing this is through video review sessions with the people on the tape (see section 2.6).

6.1.1 Beginnings and Endings

Events always have a structure. Minimally, they have beginnings and endings, but generally a more complex structure can be discerned.²⁰ Frequently, there are "official" beginnings and endings. For example, a meal may start with a prayer, ladling out of soup, "bon apetit," the guest of honor picking up her or his knife. Work may start with punching in or with a series of rituals that clear the desk and get a cup of coffee on it. A lesson may start with a verbal announcement by the teacher that gets the class' attention and then again the actual start may be unmarked.²¹ Official beginnings, however, are themselves preceded by participants' verbal and nonverbal preparatory activities and, after the event is officially over, there is some period of time during which people disengage. We always want to observe the starting up and winding down process since significant interactions tend to happen at these junctures.²² Beginnings and endings are often marked by re-arrangements of artifacts. Tracking what is turned on, brought in, taken out, or re-arranged prior to the official start reveals what sorts of props and technologies are thought to be necessary for carrying off the event.

Beginnings and endings, though often perceived as externally imposed, are in fact collaboratively achieved by participants. For example, in a school setting, the bell officially ends the class period. But video analysis reveals a much more complex picture (CAN.90.02.19.IAL):

The de facto ending of class here is not necessarily signaled by the bell, nor simply imposed by the teacher. Rather, the bell portends an impending closing up of activities. But when the teacher indicates he needs more time, the students refrain from putting away their tools. Instead they remain attentive until after he has finished. The teacher, at the same time, has acknowledged the official ending ("sorry to hold you up") and has kept his overtime to a minimum. In other school situations, of course, the teacher may barely be able to speak after the bell rings and may have to write the homework assignment on the blackboard. In either case, the closing is a collaborative achievement.²³

6.1.2 Segmentation

Events of any duration are always segmented in some way. They have an internal structure that is recognized and maintained by participants. As analysts we are interested in the ways in which participants make that structure visible to themselves and each other, how they "announce" in some sense the fact that they have reached a segment boundary in the work and that the next stretch of interaction will be of a different character.

Kendon (1985) points out that spatial orientation serves as a means of negotiating transitions from one segment to a next. People test out each other's alignments to a given interpretive frame as a means of finding out if the other is willing to change to a new one. Small maneuvers in the direction of a new position are often observable as pre-closings. Finishing food or drink, stubbing out a cigarette, moving into a bodily position for exit, stepping back from the conversational circle are announcements of readiness and proposals to change the frame of interaction.

The negotiation of seamless transitions is particularly important in complex work settings. For example, in the operations room at an airline hub, the data show operators getting ready for arriving airplanes. As a comparatively slow period comes to an end, we see them square themselves up to their desks. One of them closes a magazine she has been reading, but she doesn't just close it casually. She closes it with deliberation and emphasis. The talk in the room begins to take on a different tone. Communication technologies are activated. It almost looks like a reaction to somebody entering the room, though nobody did. Rather, what we see is the group getting itself ready for a new work segment. They indicate to themselves and the rest of the team that they are indeed about to tackle another round of tasks. It is by observing and noting these kinds of changes that the analyst arrives at a statement of the structure of the event as participants experience it.

Sometimes we find that transitions between segments are done badly, in the sense that the transition is not smooth, not seamless. Something happens so that some participants are no longer able to project the ending of one segment and the beginning of a next. For example, in a videotaped episode from an elementary school classroom (CAN.90.02.19.IAL)²⁴, a teacher tells a student to come up to the blackboard, but when the student does so, the teacher tells her to go sit down again. On other occasions, she cuts a child's answer off with a new idea of her own, though she herself had solicited the child's participation with a question. Progression here is not seamless, not orderly, not projectable. New beginnings, segmentations, and endings of action sequences are not foreseeable, which leaves the students (and the teacher for that matter) confused. As a consequence, kids talk over her voice. Without successful negotiation of turn-taking, productive interaction is impaired.

Analytically, transitions from one segment of an event to another are often indicated by shifts in activity, heralded by changes in personnel, movement of participants in space, or the introduction and manipulation of new objects. For example, spatially separated design teams collaborating through individual visits do considerable work to organize and profit from the segmentation resulting from unexpected visits (Minneman, 1991). Similarly, in a meal, the transition from main course to dessert may be marked by pronounced changes in participants' body posture, replacement of dishes on the table, an increase in the volume of talk, and the like. Scientists' collaborative working sessions may be segmented into sedentary periods of discussion, alternating with stretches of writing on the whiteboard. In a study of two graphic designers' collaborative work, it was found that their working sessions were divided into clearly differentiated segments during which different projects were discussed and worked on. One way in which the designers mark transitions between these segments is by closing one project folder and opening the next, accompanied by topic changes. In a palpable way, the work stack in this setting provides the agenda. The visible pile of folders to be worked through corresponds to the number of new topics that has to be discussed (Linde, 1991).

Similarly, the gaggle of kids collecting around a teacher between classes constitutes an agenda, that is to say, structures the available time. The teacher will have to deal with all of the students in the space of the break, a fact which provides a constraint but also a resource for the mutually visible need to do this in an expedient and speedy manner.

Many learning and work activities involve a known (or at least discoverable) projectable sequence of events. Students may know that they have a given number of problems to solve and if that work is to be accomplished together, the transition from one problem to the next must also be achieved together. Similarly, in collaborative work situations, segment transitions and hand-offs between participants are significant issues that are sometimes explicitly recognized (as in procedures for shift changes) but more often are informally achieved. In every case, participants need to make clear to themselves and each other that something is finished and something new is starting. The ability to achieve such transitions in a seamless way is one of the ways in which membership in a community of practice is displayed.

6.2. The Temporal Organization of Activity

6.2.1 The macro level

There exists an extensive sociological literature dealing with the temporal organization of activities on a macro-level. Zerubavel (1981), for example, investigated the hidden rhythms of everyday life and the temporality of professional work in medical settings (Zerubavel, 1979). Dubinskas (1988) has edited a volume of ethnographies of high-technology organizations that deal with the social construction of time within North American professional culture: the world of medical technologists, particle physicists, design engineers, and laboratory biologists whose careers and daily practices are ordered by calendars and business plans, career cycles, and research protocols. Anthropologists have given us extensive culturally and ecologically based analyses of how natural and cultural periodicities, such as seasons, floods, and migrations of humans and animals, shape the lives of people in other societies. They have also provided insight into entirely different ways of conceptualizing time that contrast with our western linear chronologies (Bohannan, 1967; Fabian, 1983; Leach, 1971).

While social theorists and social historians concentrate on macro-scale temporal patterns in the activities under study, Interaction Analysis examines the temporal organization of moment-to-moment, real-time interaction. Interaction Analysis provides a focus on the shape of an event, its high and low points, the relaxed and frenzied segments, and the temporal ordering of talk and nonverbal activity. Above all it gives access to the ways in which participants experience and make visible the temporal orderliness and projectability of the events they construct.

Interaction Analysis also allows us to see the ways in which externally imposed time-tables organize the activities of many settings. In the modern world, where many enterprises are schedule-driven, things often have to happen at particular points in time (or at particular points in certain action sequences), in order to allow for the finely-tuned coordination necessary for complex production systems, from factories to school operations to the scheduling of surgery and transportation systems. In such settings the selection, spacing, pacing, and completion of activities is sensitive to their positioning vis-a-vis anticipated schedule events. The temporal patterning of externally generated demands makes many of the routine activities that occur in such settings time sensitive or time critical; that is to say, for some activities failure of on-time performance will lead to serious disruption of the ongoing work.

6.2.2 Rhythm and periodicity

Some measure of rhythmicity or periodicity is a feature of many, if not most, human activities. Action sequences that involve repetitively executed tasks or segments have a number of analytic interests. For one thing, one might expect that participants experience boredom and wandering of attention if they have to do the same task over and over again. On the other, however, repetitive sequences allow the development of stable routines as a durable infrastructure against which trouble can be managed and predictability and projectability of action sequences can emerge. Interaction Analysis looks both for the repetitive, routinizing aspects of activity sequences and for their variability.

In a given environment, when there is a string of "same" activities, questions that arise include: In what sense are the repetitive segments identical? How much variability is allowed before a sequence is no longer "the same" and becomes something else for participants? How is such segmentation achieved? For example, a tape of family evening activities focused on a toddler who has been putting spoon to mouth for some period of time, contains the following sequence:

Father: Is she still eating?

Mother: ......... No, she's just fiddling around.

The participants in this scene have just partitioned repetitive spoon-to-mouth activity into "eating" and "fiddling" segments which, of course, has implications as to further actions that might be taken.

Many, if not most activities people engage in have some sort of periodicity to them, involving some sort of repetition. Some of these periodicities are biologically driven. The necessity to take in a breath affects speech patterns. The appearance of uterine contractions during labor dramatically affects patterns of interaction during birth. In the latter case, there is not only periodicity but also an increasing intensity that builds until the baby is born. Talk and the execution of tasks in labor settings are observably responsive to that periodicity and are organized to take account of it (Grosjean, 1993; Jordan, 1992c).

Rhythm may be an aspect of interaction that is significant in more situations than has been recognized. In social settings where some kind of mutual engagement is ongoing, newcomers to the action must find a break in the ongoing stream of verbal and nonverbal activity in order to gain entry. As Erickson (1991) points out, social participation as management of access and transition appears to be facilitated by members' participation in a shared rhythmic framework.

Similarly, Kendon (1985) has shown that moving into synchrony with another person is one of the devices by which a person can indicate to the other that he or she wishes to establish "an action exchange system" without making an explicit request. By simply picking up on the rhythm of another's movements or talk, people establish a connection which at the same time does not commit them to an explicit initiation. Such co-calibration becomes visible on the tape.

Some rhythms are driven by technology, as for example the activities of installers working on an assembly line. In other cases, the periodicity is provided by the nature of the task as when a group of students has a series of problems to solve. Many workplaces are tightly organized around more or less rigid schedules that impose repetitive activities. For example, at hub airports, one may find interaction in airlines operations rooms organized around "complexes," periods of time during which a flock of connecting planes come in, exchange passengers and baggage, take on fuel and food, and take off again. During these complexes, activity in the operations room is highly energized, only to slow down to a more leisurely level in between.

For workers in ground operations, reiteration of tasks also occurs within complexes, since while a complex is in progress, plane after plane has to be brought in, serviced, and sent out again. For each plane, paper documents have to be assembled, fuel and food services have to be provided, passengers have to be deplaned and boarded, baggage has to be loaded, etc. Our recordings show that there is a large number of repetitive activities that are routinely done, and done over and over again, in response to the periodic demands of complexes or individual planes. There is then a kind of nested periodicity, a visible rhythm to the workday, a predictable recurrence of tasks and events, that itself furnishes resources for the management of contingencies as they arise.

One consequence of periodicity is that it provides for slack times, delivers breathing spaces during which some "other" set of activities is accomplished. These in-between activities are often seen as in some way extraneous, even superfluous. Yet, on analysis, it is usually the case that they constitute a crucial aspect of the undertaking in question. In the operations room, one might occasionally see an operator reading a magazine but there are also numerous job-related activities performed in the space between complexes, such as catching up on paper work, ordering food for later delivery, telling job-related "war stories," and the like.

In some parts of the economy there are jobs that require at least the appearance of continuous engagement in active work, regardless of any natural structuring of demands. This is typical for positions dealing with the public, such as receptionists, but is also true to some extent for pupils in schools. Policies that require the appearance of constant task orientation are interesting in that they invite ways of getting around them, so that we see students reading comic books under desks, and airline personnel ducking behind a counter to take a drink of water.

In schools, class periods provide major temporal structuring, but important work also goes on in the space in between class periods A videotape of the transition between classes shows the teacher fulfilling bureaucratic record keeping requirements: he deals with detentions, signs field trip permission slips, and gives information on students' current standing by consulting a grade book -- all of these are activities without which the enterprise of teaching and learning could not proceed.

An important outcome of understanding the temporal organization of activities in the workplace is the identification of resources and constraints for the design of work environments that support learning on the job. A classical issue for training has been that training should happen embedded in the actual activity. Analysis of video tapes shows that the time structure of these activities has a lot to do with their capacity to incorporate newcomers. If the work includes natural periods of relaxation, when workers can pay attention not only to the objects and tasks they have to monitor, but also to the newcomers who need to learn about these, stepwise incorporation through "legitimate peripheral participation" (Lave & Wenger, 1991) is much more easily accomplished. It is important that people be able prospectively to prepare for what is coming up and retrospectively to reflect on what has just happened. They have to be on-line and off-line at various times and able to switch back and forth. An uneven temporality, especially predictable periodicity, is more conducive to learning, more hospitable to newcomers (and, incidentally, ethnographers) than one that has a steady, high-level demand on participants' attention.

While much can be learned about schedules from interviewing and participant observation, the finer details of the temporal organization of activity are recoverable only from videotape analysis (Suchman & Trigg, 1993). We would suggest that the detailed investigation of what Erickson (1991) calls "the shared temporal ecology of social action" in work and learning settings should constitute one of the major foci for Interaction Analysis.

6.3 Turn-taking

Subsequent to a seminal paper by Sacks, Schegloff and Jefferson (1974), much has been written about the complex organization of turn-taking in conversation. For Interaction Analysis the situation is ever more complicated because an Interaction-Analytic turn-taking system has to take into account more than talk: it encompasses the whole range of behaviors through which people can "take a turn," that is, participate in an interactional exchange system. Not only "turns at talk" must be considered, but also "turns with bodies" and "turns with artifacts." For example, a videotape of kids working on a problem in front of a computer screen shows them taking turns with the mouse. Analysis reveals that grabbing and relinquishing the mouse constitute significant moves in collaborative problem solving.²⁵ The requirement that both talk and physical action must be taken into account if we are to understand complex activity, has come to be widely recognized by conversation analysts as well. For example, Schegloff (1987b) speaks of "talk-in-interaction" and Moerman (1990) has suggested that the distinction between verbal and nonverbal communication is fallacious altogether. He points out that "communication by means of pure language, without context, without body, without time," simply doesn't exist (p. 9).

We can make the complex relationship between talk and physical activity a bit clearer by thinking of social interactions as falling into two distinguishable categories that form the end points of a continuum. ²⁶ In some situations, whatever it is that is to be accomplished in the interaction is primarily accomplished in the talking. The interaction is constituted in and by the talk. This tends to be the case in business meetings, police interrogations, conversations (face-to-face or over the telephone), interviews, lectures, story-telling sessions, etc., where talking is what the event is principally about. We can contrast such "talk-driven interaction" with "instrumental interaction," where activities are motivated by the requirements of a physical task that has to be done. This is the case when, for example, surgery is performed, a car is repaired, homework corrections are made, a plane is loaded, and so on -- activities that crucially involve the manipulation of physical objects. These objects may be natural objects or human-made artifacts and technologies; in either case, the central business of the activity cannot be achieved by talk alone. In the course of such instrumental interaction, talk may and usually does occur, but it is not central as in talk-driven activities. It is usually ancillary to, supportive of, and sometimes even coincidental to the main business at hand. Thus the conversation of two surgeons about the morning's difficult operation is an instance of talk-driven interaction; but the actual surgery (including the talk that happens in its course) is instrumental interaction. For talk-driven interaction, the most relevant non-talk activities are gesturing and gazing, both of which coordinate the talk. Where interaction is instrumental, the nature of production tools, display spaces, and other aspects of the material environment significantly enter into the interaction and become an important part of the analysis.²⁷

Examination of the organization of talk-cum-activity in work or learning environments where physical activities have to be performed, adds to the findings of turn-taking research focused strictly on talk. In situations where a physical task has to be accomplished, talk and physical activity are complexly intertwined in the turn-taking system. We find that action turns may get taken in response to verbal turns and vice versa; that is, somebody may verbally ask for something and, in response, get not a verbal answer, but rather an activity performed. For example, in one of the Workplace Project tapes, when an operator announces that a seat cover on an incoming plane needs replacing, another operator responds by picking up the radio and calling cabin service, without giving a verbal response to the speaker. Conversely, a verbal reply may arise in response to an activity, rather than in response to a prior piece of talk. It seems to be the case quite generally that in most human activities turns are not only made up of turns at talk but incorporate other activities as well. For example, designers at work fluidly mix turns composed of drawing, talk and gesturing activities (Tang, 1989; 1991) even when using electronic media (Bly & Minneman, 1990; Minneman, 1991; Tang & Minneman, 1991).

We have found a series of differences in the organization of talk in instrumental interaction compared with talk as it happens in talk-driven (conversational) situations. For example, in instrumental interaction topics tend to stay alive much longer than in "pure" conversation; often, they are taken up a considerable stretch downstream. The requirement for talking "on topic" may be suspended for long stretches. Pauses show different characteristics. In complex work settings with multiple overlapping activities, documents, monitors, or other informational resources often need to be consulted before a turn-at-talk can be taken. Thus lengthy gaps and pauses are generated.²⁸ Another reason for differences with talk-driven interaction is the interruptability of many action sequences by peremptory inputs of various kinds. When a radio call comes in, it is lobbed into an ongoing activity and conversational environment, regardless of how it fits. A phone ring's onset is not negotiable -- the phone must be answered and takes precedence over most other activities. A filing task may be abandoned because a beeper beeps, an ongoing conversation may be interrupted by a co-worker requesting information, an operator may not respond to her supervisor because unbeknownst to him she is listening to the radio -- all these are ways in which routinized turn-taking sequences are continuously liable to be disrupted, only to be resumed when the interruption disappears.

Interruptability due to the coordination requirements of tasks that must be accomplished in parallel structure turn-taking in instrumental interaction, as does the nonnegotiable onset of technology-based communications. As new communication technologies are introduced in a workplace or educational setting, participants begin to develop a set of expectations about how the technology fits into their ongoing interactions. Our data show actors orienting to the demands of the technologies in their environment in ingenious and inventive ways that minimize the interruption. For example, they close up an on-going conversation just-in-time, rather than simply cutting it off. Again and again we observe the artful production of a sequence as finished, a conversation not so much interrupted as closed down in the nick of time.

In formal educational settings, the rules for turn-taking tend to be highly stylized and ritualized. Officially, the teacher is in charge of turn allocation. The teacher speaks (explains, lectures, demonstrates), and then specifically assigns turns to students by calling on them. Student self-selection is frowned upon, since it is considered disruptive to the sequence of activities planned by the teacher. Against this standard, our data reveal many deviations, some grounded in school philosophy, but also frequently visible as the specific accomplishment of students actively involved in the learning process. Teaching situations are dominated by talk-driven interaction though some physical objects also figure prominently in classroom turn-taking: copybooks have to be handed in, chalk has to be picked up in order to write on the blackboard, models and pictures are brought in by teacher or students -- but note that the activities of which these objects are a part are generally initiated by the teacher. In most school situations students play a passive role.

6.4 Participation Structures

Another topic of inherent interest for Interaction Analysis concerns the extent to which co-present individuals share a common task orientation and attentional focus. Mutual availability and alignment become visible in "participation frameworks"²⁹ -- fluid structures of mutual engagement and disengagement characterized by bodily alignment (usually face-to-face), patterned eye-contact, situation-appropriate tone of voice, and other resources the situation may afford. Students of interaction, from Goffman (1963; 1981), to Erickson (1982a; 1982b; 1991), Heath (1986), Kendon (1985; 1990), Goodwin and Goodwin (1992; in press), Goodwin (in press), Goodwin and Heritage (1990), Suchman (in press) and others, have noted that the social "work" that is done through participation frameworks provides the interactional infrastructure for the achievement of coordination and collaboration among co-present individuals. The resources for the production and maintenance of such social structures lie in the mutual visual and auditory availability of participants' bodily activity. Relevant issues for Interaction Analysis, then, revolve around such questions as: how do interactants make their engagement (or lack thereof) visible to each other; what strategies do people employ to gain entry; how do artifacts and technologies support or constrain particular participation structures; and the like.

A study of interaction in a hospital labor room (Jordan, 1992c) provides an example of two parallel but separate participation structures co-occurring in the same setting. On the tapes covering a particular labor and birth, the birthing woman interacts almost exclusively with her husband and the attending nurse but is excluded from the professional participation structures of the labor room within which physicians, nurses, medical students and other official personnel accomplish the delivery of the baby. The staff interacts as a team of which the physician is the focal member. No input is solicited from the woman; talk is not produced for her overhearing or participation. The staff does the business of examining her and preparing her for the delivery amongst themselves, without engaging her in consequential talk or activity. In this setting, the woman is not an active participant but simply the object to be delivered.³⁰

In a study of interaction in an airlines operations room, reported in the same paper, participation structures are found to be less exclusionary and more overlapping than those of the labor room. As contingencies arise in projected action sequences and are taken up for notice or action by co-workers, new alignments are constantly created and recreated. Multiple participation structures are generated, maintained, and disassembled in response to the requirements of the business at hand. Here there is no principled exclusion of individuals. All co-workers participate fairly equally, i.e. without structurally provided restrictions, not only in the flow of communication directly related to the work but also in the informal kinds of exchanges that appear in the interstices between tasks and when things slow down between complexes. Stories and jokes involve all those present as tellers, recipients, and commentators, without exclusion.

The notion of participation frameworks was originally developed to describe face-to-face interaction. Work and learning environments that are structured like the operations room provide an opportunity to extend that notion to situations where significant exchanges routinely and necessarily take place with persons in technologically connected remote work spaces. As is the case in many high-technology work situations, people in the airline ops room spend much time maintaining extended linkages and exchanges with co-workers through radios, telephones, and computer networks, only to turn back to interaction with their colleagues in the operations room as they conclude an externally-oriented exchange. Especially during high workload periods, the default activity for operations workers is preoccupation with, and orientation to, their workstation and thereby to the remote co-workers to whom they are connected. This primary involvement provides the background against which interactions with physically co-present colleagues take on a certain time-out character. This becomes visible in the many instances where cross-room communicants assume torque positions (Kendon, 1990; Schegloff, 1990) i.e., turning head and torso towards a co-worker without swiveling around completely, thereby indicating to themselves and their co-workers that they imminently intend to go back to, and indeed are still involved in, their prior activity. In this way, interaction with physically co-present co-workers is often displayed as an interlude in the ongoing work with remotely located co-participants.

The analysis of participation structures is also essential to understanding interaction in formal school settings. To what extent do teacher and students sustain different kinds of participation structures in group work or in lecture format? How do computers, workbooks, table arrangements, and other kinds of artifacts support or destroy such structures? Issues around distributed participation become important for classrooms, as more and more students are networked. How does the orientation to a computer change gaze and body orientation as compared to group interaction around a table?

Analyzing participation structures as they shift in the course of working and learning activities is one of the ways in which Interaction Analysis contributes to the study of "the C-Issues": cooperation, conflict, conviviality, competition, collaboration, commitment, caution, control, coercion, coordination, co-optation, combat, and so on. It is precisely because social life is rarely convivially univalent (Kling, 1991), that the complexity of the interrelationship between these forms of interaction needs to be better understood.

6.5 Trouble and Repair

Another major focus of analytic interest for Interaction Analysis is the occurrence of "trouble" in a particular activity sphere. Anthropologists have known for a long time to pay particular attention when the normal stream of activity is broken in some way. Careful analysis of the breach can often reveal the unspoken rules by which people organize their lives. As a matter of fact, the analysis of visible breaches of the local rules for social interaction is one of the best methods for coming to an understanding of what the world looks like from somebody else's point of view. Analysis of hitches in interaction may also reveal some of the constraints in the material world that routinely cause trouble.

Much work has been done in conversation analysis on how troubles that arise in talk are repaired (Jordan & Fuller, 1975; Sacks, et al., 1974; Schegloff, 1979, 1987a, 1987b; Schegloff, Jefferson, & Sacks, 1977). For Interaction Analysis the situation is complicated because we need to take into account not only the verbal aspects of repair, but also the ways in which participants draw on their bodily, artifactual, spatial and social resources to mend infractions of projected sequences.

For example, a recording of a couple of high school students struggling to understand how to operate the Envisioning Machine (Roschelle, 1991), shows them resolving hitches in their understanding by recourse to the social and material resources of this environment (Roschelle & Clancey, 1991). Sitting in front of a computer screen, their task is to match the movements of a particle under their control to the movements of a target particle. Just as in many real-world work environments, verbal communication alone is insufficient here to effect the intimate co-alignment to the screen and to the screen-borne representations through which an initial discrepancy in understanding becomes resolved. As Roschelle and Clancey point out, the two students, Gerry and Hal, use a combination of talk, gesture, and screen object manipulation to resolve their difficulties. The pinpointing of the trouble as well as the mutual alignment of the two students occurs in a shared visual and manipulative space. They point to objects on the screen to make references clear, gesture to act out concepts, and synchronize their talk with events on the computer screen (p. 341).

It is obvious that an audio tape of this transaction (or a transcript thereof) alone would be inadequate for doing a full analysis of how they repaired the troubles that had arisen.

In stable situations, people learn by experience which kinds of troubles tend to recur and what range of resources can be assembled and held available for their solution. Practitioners begin to recognize problems for their typicality, recycle them in stories -- thereby making them talk-aboutable (Orr, 1991) -- and otherwise domesticate them, so that their solutions become available on a next occasion. Much of this knowledge is not written down anywhere, nor does it lend itself to writing down, but resides in the community of practice that forms up around new technologies.³¹

Hitches in interaction are often invisible to the casual observer because participants are very good at fixing them on the fly, without missing a beat. Troubles in ordinary talk, such as mishearing or lack of understanding, are commonly repaired without participants being aware of what they do. Consequently often neither participants nor observers consciously experience the trouble. Yet, during detailed analysis of videotapes, participants' corrections of misunderstandings, of misalignments, of intrusions, and other types of rendings of the social fabric are routinely visible.

A parallel and increasingly interesting area of investigation is the repair of trouble in the interaction between humans and machines (Frohlich, Drew, & Monk, in press; Suchman, 1987). In their interactions, people pretty much assume that they share rules of interpretation. This assumption becomes problematic when computers and other complex machines are involved. Trouble occurs when there is a "miss-match" between the rules and procedures employed by the user and the computer in interpreting meaning from a sequence of symbols. It is precisely through the observable ways in which human-machine interaction runs into difficulties that people discover in the first place how the machine interprets human input. A successful repair of such trouble constitutes learning on the part of the user, if not on the part of the machine. This is one of the many ways in which human/machine interaction differs visibly and observably from communication between people.

6.6 The Spatial Organization of Activity

A trivial observation about human existence is that people occupy space. And they occupy space in characteristic ways that differ from the ways in which, say, birds or artifacts inhabit their space. Human beings' size, their sensorimotor abilities, and their shared ways of orienting to a social and material world facilitate certain uses of the space around them and make difficult, or prohibit, others. Within these constraints, many variations are possible and different social groups have developed particular ways of being in each others' presence (Hall, 1959). What is considered appropriate body distance or body attitude, how far one's gestures can intrude into another's personal space, how public spaces are used in contrast to private spaces, to what extent a shared focus of activity (for example in joint work on an object) can override otherwise expected behavior -- issues of this sort are resolved differently by different social groups and for different activity systems. They are of central interest to Interaction Analysis.

As we work with video tapes, we note that the physical co-presence of persons is always managed by socially recognized (though often unstated) expectations regarding occupancy of space, interaction with others, use of objects and resources, display of physical presence, and voice. People make use of these mutual expectations as resources for structuring their interaction with others and for accomplishing the business at hand. As Kendon (1985) has put it, spatial and orientational positioning serve as devices by which expectation and intention can be conveyed. This is a collaborative enterprise in the course of which the grounds on which people take certain actions, occupy space, impinge upon each other, make apologies, and otherwise produce repairs to infractions of expectations, become visible not only to participants, but also to analysts examining the videotape.

Certain activities require, or are customarily done in, settings with particular spatial configurations. Work practices then develop that take these spatial constraints and opportunities into account. In the airlines operations room, for example, the four operators' workstations are placed so that as pairs of operators orient themselves to their video screens, telephones, radios, etc., they have their backs to one another. As a consequence, much interpersonal interaction in the room requires leaning back into the common space while turning the head; often we observe individual operators scooting backwards on their chairs to come into closer physical proximity to a co-worker with whom they need to coordinate. The fact that workstations are centripetally arranged around the perimeter of the room significantly affects the nature of interaction occurring there.

A similar centripetal organization is visible in the workspace of bet takers at cock fights in Northern Luzon.³² Five or six bet takers face spectators from the cockfighting ring, a kind of focal pit in which they rhythmically prance forward and backward, shouting the name of the owner of the cock for which they are taking bets. At the same time, they wave their hands above their heads towards the audience, exhorting them to place their bets. Bet takers' bodies and faces are oriented to the audience. But on the tape they can be seen to be monitoring not only the spectators for betting action but also each other, if only not to collide if they step back at the same time. In contrast to the division of labor apparent in the operations room, workers here are doing parallel versions of the same activity, taking bets. In operations, there is a periphery, an outside, from which salient information is expected to come through a variety of communication technologies. For bet takers at the cockfight, salient information also comes from the periphery, a spectator indicating a bet by hand movement, but here it is face-to-face interaction, without mediation through technology. In either case, the management of space is an important, though implicit, concern for actors.

A crucial point to consider for Interaction Analysis is that in any given environment some spaces provide more interactional resources and others less. For example, during a committee meeting the position at the end of a table allows the maximum number of eyes to focus on its occupant as he or she speaks. By contrast, the position next to the main speaker is often visibly experienced as difficult by its occupants, because it is under the constant scrutiny of the speaker's audience. Interaction Analysis of such situations allows us to formulate what otherwise might be seen as intrapsychic phenomena (the person is nervous or uncomfortable) in terms of the interactional resources that become available (or not) depending on observable spatial positioning.

In the operations room we studied, there is a particular space in the center of the room which, by virtue of its centrality, is occupied by the supervisor whenever trouble arises in getting planes in and out. In this spot, he has all of the resources the environment provides at his finger tips: simultaneously, he can take in information by glancing at a bank of video monitors; he has a video camera's controls at arm's length enabling him to zoom in on the situation at a gate; he has easy access to phone and radio and is positioned in close physical proximity to his four operators (Suchman, 1993).

Actors often signal their intentions by the ways in which they occupy the available space. In a tape of a committee meeting between two departments vying for hegemony, the chairman of the weaker department was observed to seat himself in the "gallery row" of the meeting room, on one of the chairs arranged against the wall for an expected overflow crowd. The chairman of the stronger department took his place at the head of the table. In many ways, the outcome of that meeting was predetermined.

Similarly, we have noted that of a pair of researchers doing work at a whiteboard,³³ it is the person nearest the board who consistently gets up to write on the board, so that one can see the question of where they choose to sit as speaking to their initial intentions in regard to doing writing on the board. However, once the two researchers have their positions vis-a-vis the board established, writing at the board becomes a much more effortful move for the one who is comfortably ensconced in a chair at some remove, with his feet up on the table. His getting up to write would probably be marked as a significant move in which he actively takes the floor.

The series of videotapes documenting parents and toddlers baking muffins together, shows that the specific physical arrangement of the kitchens is quite varied (CAN91.02.05IAL).³⁴ Parents' attitude towards their child's participation and competence is quite varied as well. What is common across parents and different physical arrangements is the requirement to somehow get the child up to the parent's habitual work space. Typically, the preparation of muffin batter takes place at a table or a counter and all families have developed standardized ways of compensating for the short stature of the child. They may pull up a certain chair, or lift the child to sit on the counter, or some other habitual solution to this problem may become evident.

There are further issues related to their spatial configuration, in particular where the mother places herself vis-a-vis the child. In some sequences she stands behind the child, almost cradling him or her within the circle of her arms, a posture that produces a different kind of access to the relevant artifacts than for those cases where she positions the child opposite herself. In the first case, mother and child share the same perspective on the baking operation and they have the salient artifacts equally within reach. Whenever the mother does something, the child sees that action from the same angle as the mother. And the child directly experiences the mother's bodily mobilization for action, as she, say, reaches out for a spoon. By contrast, the second position, where mother and child face each other, invites the child to be an onlooker, an observer, who, when he or she decides to get into the scene, must do it on his or her own.

Another issue that is often relevant in Interaction Analysis is who owns the territory on which interaction takes place. Interestingly, ownership of territory affects the mobility of participants -- whether they can move around at will or have to ask for permission. It also affects rights to structure the event, to initiate the beginning and end, and probably other aspects as well.

For example, Eric Bredo's comparative study of different types of schools shows remarkable differences in the mobility of students. In some schools, they are more or less confined to their desks and other well-defined workplaces; in others they move around rather freely, consulting with other students, the teacher, and the material resources in the classroom. Differential mobility often indicates asymmetrical power relations, as exemplified by the hospital patient who is confined to bed, the guest who has to ask the host's permission to enter certain parts of the house, the child who is not allowed into the parent's study, or the worker who stops in the entrance to the supervisor's office.

Physical arrangements, the spatial layout of a setting, the arrangement of furniture, the open spaces, walkways, coffee niches, doors to the outside, and so on, have an important influence on structuring interaction. Of particular interest to Interaction Analysis is how these physical set-ups affect possible participation structures, that is to say, how they encourage or hinder certain kinds of interaction between people in the scene. Furniture and technology can have a major effect. A group of students arranging themselves for collaborative problem solving in front of a single computer differs notably from one positioned around a flat work table.

Interaction Analysis thus considers to what extent the spatial layout of the setting is fixed or allows choices; that is to say, to what extent physical configurations and spatial arrangements are imposed and to what extent they are under the immediate control of participants. Facilities layout and technology design always provide specific constraints on what kind of interaction is possible within a given setting and what kinds of activities and interactions particular material objects engender and support. Interaction Analysis investigates how those constraints influence what participants actually do and how what gets done is negotiated.

6.7 Artifacts and Documents

Artifacts are ubiquitously present in all human endeavors. They structure interaction, generate problems, and provide resources for the solution of difficulties as they arise. Sometimes they constitute the focus of an interaction, as when two people work together over a blueprint or at a computer terminal. Sometimes they are co-incidental to it, as when a pencil is used for gesturing and tapping in the course of talk during a working session. For Interaction Analysis, the basic premise is that artifacts and technologies set up a social field within which certain activities become very likely, others possible, and still others very improbable or impossible. One of our central interests lies in understanding what kinds of activities and interactions particular material objects engender and support and how these change as different artifacts and technologies are introduced.

Looking at the role of artifacts from a community-of-practice perspective, Lave and Wenger (1991) note that becoming a full participant involves not only partaking in the social relations of the community but also includes engaging with the technologies of everyday practice and the community's production processes. Lave and Wenger point out that the degree of engagement with technology can be extremely varied and elaborately progressively staged, depending on the form of participation it enables. Interaction Analysis attempts to include material objects as special kinds of participants in its analytic endeavors.

In our analyses of working and learning situations, we have found again and again that it is important to track where people's eyes are, when and how gaze moves between objects, from persons to objects, and back again, sustaining or shifting the focus of attention as the salience of particular objects or displays changes. Gaze clearly plays an important role not only in coordinating conversational interaction (a topic that has been studied extensively)³⁵ but also in carrying out physical tasks. Although the role of gaze in the production of talk has been apparent for quite some time, little analytic attention has been paid to the social coordination necessary for people to get their work done, as they manipulate objects and move around in space. Workplace studies are only now beginning to explore this important feature.

Given the ubiquity of artifacts in any human environment, it is often difficult to determine where and how to start an analysis. One place is with an inventory of material objects that tracks those objects in the environment that enter into the interaction in significant ways, addressing such questions as: What is their trajectory? How do they get into and out of the scene? Who are their human hosts? Are they uniformly or hierarchically distributed? How do they function in structuring interaction? For example, for students working in front of a computer in small groups, the mouse is an important object because by grabbing or relinquishing the mouse turns at doing work at the screen are taken. It is thus important to understand what kinds of activity ownership of the mouse generates. For a Quality Improvement Team that uses an electronic "liveboard" for consensually producing its minutes in real time, an electrical failure has a debilitating impact. In other groups where records are kept on paper, backup materials are more easily found. The choice of artifacts and technologies to support particular kinds of activities thus deserves considerable attention.

In general, the process of jointly constructing, completing, and revising "inscribed artifacts" -- documents in the widest sense -- provides resources for participants to monitor whether they are in agreement or not. For example, Rogers Hall analyzed pairs of teachers working together at a whiteboard as they were designing instructions for solving algebra problems. One pair constructed a shared drawing that decomposed "rate" into coordinated scenes in a thoroughly collaborative manner -- one would start a scene and narrative, then the other would finish the drawing as the original narrative proceeded, thereby indicating and holding up for verification her understanding of her partner's sense. Except when disagreements erupted, there was very little need to discuss explicitly the quite complex entailments of the drawings being produced. Instead, finishing the other's work synchronized and displayed their mutually shared understanding without much explicit talk.

Objects also frequently serve as territory markers. The marking of workspaces with shifting occupants is typically accomplished with personal items. Who is responsible for (or has rights to) re-arrangements? How temporary or stable are particular arrangements? Are they left in place across shifts? Considering current trends to dedicate office and work space less to individuals than to tasks and projects, what happens when you remove the possibility to personalize the workspace? OrCENTER>

Some types of artifacts are culturally marked for special functions, but it is also important to look for the ways in which unmarked objects work. A gavel, for example, is recognized as the instrument by which certain kinds of meetings are officially opened and closed. But careful analysis of videotapes shows that unmarked objects may play a similar, though less recognized role. For example, on a videotape of a Ph.D. dissertation defense, the candidate puts on the table the document that has to be signed at the end to certify that she has passed. In the course of the afternoon, this piece of paper wanders through the committee. Sometimes it just lies on the table, sometimes somebody picks it up and studies it idly, only to lay it down again. Though nobody appears to focus on the trajectory of the document, as the meeting draws to a close, and without a word being exchanged, various people collaborate in getting it to the chair, so that it ends up in front of him at the appropriate time.

Sometimes objects are salient not because of their instrumental value but because of their symbolic significance. A stethoscope, for example, has a "use value" as a tool that allows monitoring of patients' heart beat. It also has, however, symbolic value in the sense that, hanging out of the pocket of a white coat in a hospital, it identifies the owner as a clinician who does actual patient care.

An important issue in regard to artifacts and tools is their "ownership." It is often possible to tell from a tape who owns an object because the owner has rights to touch, to manipulate, to display, which are not shared by other participants. For example, in a hospital, patients are not permitted to handle the medical tools of the trade, and in general make no attempts to do so. In medical interactions, then, the use of a major proportion of artifacts available in the setting is restricted to one class of participants. Since knowledge is often tied to particular tools and technologies, this has important consequences for the distribution of knowledge within such settings.

Questions of ownership, interestingly, may arise also in regard to "non-material objects." For example, as designers draw symbols, graphics, and text on a whiteboard, they create objects that re-present the past interaction. These iconographic artifacts then become available as resources for further discussion, as something participants could be pointing to, playing around with, referring to, modifying, or erasing. Thus, for a given work setting or learning environment, one might become interested in questions such as: Who constructs these objects? Who uses them? How are they related to talk? What role do they play in constructing arguments? Under what circumstances are they given names (which make them easy to refer to, in addition to being easy to point to)? Who has ownership of the things a person has constructed in a shared work space? Can somebody else erase them? Can they modify them?

We have been particularly interested in artifacts and documents that function as public display spaces. They are common both in educational settings (in the form of blackboards, overhead projectors, and tack boards) and in work environments (in the form of video monitors of various kinds, papers tacked up on walls, bulletin boards, whiteboards, and computer screens). Such displays often provide a crucial focal point for marshaling a group's attention. They also serve the important function of supporting the public availability of the information they display, as well as of the practices and reasonings that are developed and warranted within a particular community of practice and which systematically inform the work and interaction of participants (Heath & Luff, 1991).

Public information displays compellingly structure interaction. We find it useful to make a distinction between restricted displays, which can be seen only by one or two persons at a time, and unrestricted displays that are available to a whole group. On videotapes of a survey interview, for example, the survey instrument, a hefty paper document from which the interviewer reads questions and into which she writes answers, significantly affects interaction between interviewer and respondent. It requires attending, pausing, turning over of pages, and other kinds of managing that interrupt the normal flow of talk. More significantly, however, since the interview schedule is visible only to the interviewer, much of the information available to her is hidden from the respondent who has no idea of what permissible answer categories are, what the branching structure of the interview is, or even how long the interview is to take in the first place. Manipulating the interview schedule in such a way that the respondent cannot see it (supported by the fact that the respondent makes no attempt to, say, take the schedule out of the interviewer's hands and write down the answers herself) constitutes one of the ways in which participants make clear to each other that they respect the interview as an impartial recording device.³⁶

Compare this way of doing work with another situation where access to the information display is not restricted. In industrial process control rooms, large public information displays often not only disseminate information but also provide the resources for making that information available for discussion. While the survey interviewer has to reconcile discrepancies by herself, anomalies that become visible on large public displays tend to generate conversations and thereby draw multiple expertise into the process of explanation and resolution. Understanding that changes in the structure of information displays almost always have consequences in social relations is crucial for the design of new display technologies and for the redesign of work processes (Kukla, Clemens, Morse, & Cash, 1992). While shared display of patient information in the consultation room has never been advocated, we might nevertheless speculate that information systems that display patient records on a large screen in the doctor's office would radically alter patient/physician interaction. Where now the physical characteristics of the patient chart give the physician privileged access to the patient's condition, a public display of the record would make it a common resource in a collaborative discussion of the case.