In the 1950s the study of communication became immensely popular, but there was not yet any clarity about levels of organization and the nature of communicational phenomena, Communicational processes were being located variously in mind, brain, the reflex area, the group, and in the electronic devices of mass media, and also according to the classical methods in which a variety of disciplines were used to examine the subject. The result was a number of conflicting theories of communication.

In 1956 a multidisciplinary group established a systems methodology specifically designed to study communication.¹ They operationally defined communication in behavioral terms so that visible and audible phenomena could be studied directly. The principles of natural history method were adopted as the premises of observation, but the recordings were made on audiovisual media. Behavior was analyzed by structural methods adapted from structural linguistics but applied to all types of communicational behavior. And the unit forms of behavior were studied by newer systems approaches to synthesis and integration at the social level. This procedure was named context analysis.

Context analysis is not to be confused with content analysis — a method which evolved in the 1940s to study the content of language. The method of context analysis, then, does not fall within the traditions either of clinical-subjectivist methods or of the experimental-statistical tradition of American psychology.

The procedures of context analysis have evolved and been further explicated since 1956. Similar methodologies also developed in the 1950s. So the basic approaches of context analysis are no longer unique and operations such as these I describe here are currently employed in a number of human and animal research projects.²

This account of context analysis overviews the full scope of operations which can be used for the detailed multidiciplinary analysis of a transaction. Many researchers do not care to use all of these operations. Some, for instance, may not be interested in cross-cultural sampling, while others may be interested only in comparing certain forms of behavior cross-culturally. Neither of these may wish to study the total structure of any given transaction. Accordingly, those readers who work within traditional disciplines may choose merely to scan this broad methodological statement, returning to those operations of specialized interest for more careful reading. Furthermore, because the criteria of context analysis may be unnecessarily stringent for many researchers’ purposes even within the scope of a specialized usage, this account of the method is somewhat idealized.

On the other hand, the methodology does cross older disciplinary lines and an overview of it may encourage broader studies in human communication than the compartmentalized ones we have had in the past. Five major operations are:

OPERATION I. OBTAINING AN AUDIOVISUAL RECORD OF A TRANSACTION

In context analysis we favor first-hand, naturalistic, and nonparticipant observation. Procedurally, we plan a sample, then make film or videotape recordings of specific examples.³

Step I-A. Planning an Observation

Film development is expensive and a context analysis is time consuming, so we cannot afford random and unrepresentative shooting. We select a subject transaction, choose a vantage point for observation, and try to anticipate disruptions.

1. Choosing a Type of Subject Matter

Since communication occurs among members of a group, our subject must be interactions or transactions, not individuals.

In order to examine all of the modalities of behavior in communication, we choose transactions of small or face-to-face groups. So far we have not studied electronic communication at a distance. We could not get all of the participants in our picture and they would use a contrived code derived from the natural systems of behavior we are trying to understand. We choose some communicational event which illustrates a point or is representative of a class of such transactions. In the latter case we want any example we study to be representative of the class.

For this purpose we prefer a customary, uncomplicated transaction. We try to locate a usual activity which people are used to enacting together. We seek participants who are native to the tradition of that transaction and experienced in taking part. And we make our observations under usual and favorable conditions. We prefer to study the transaction at sites where it usually occurs under customary conditions.

We have to learn something about the situation to make such decisions. We read the literature about that kind of transaction and talk to colleagues who have worked in that area. Then we interview subjects who are experienced in that kind of transaction. We visit sites where it usually occurs and make preliminary observations.

2. Choosing a Direct, Naturalistic Approach

We decide to observe the transaction directly. We have found on interviewing that participants do not know much about nonlanguage behavior. And they cannot see much of their own behavior in communication. So we decide to use interviewing as an adjunctive method of collecting data, but we will observe the events first hand.

We start with naturally occurring communicational processes before we attempt any experimental manipulation; otherwise we do not know what we are experimenting upon and what changes we have induced. So we will not bring strangers together in the laboratory and give them contrived tasks. We will go to sites where that transaction normally occurs on the occasions at which it would happen anyway. We prefer experienced, native participants who know each other. And we will take all possible measures to avoid being obtrusive and manipulative.

We will not participate directly in the transaction we are to study. In part nonparticipant observation will reduce our obtrusiveness and in part will give us a more useful perspective.

As group members we are constrained by etiquette to concentrate our attention on the speaker of the moment and we must make the effort to take a meaningful part in this action. As observers we can watch any participant or any grouping and we are free to concentrate on research. When we are members of the group we cannot see ourselves, and we may be too close to the others to see more than one or two of them at a time. So we stand back from the subject group far enough that we can see all of the ongoing behaviors at the same time. This will provide us the Einstellung from which to see behavioral relations at the social level.

The reader will notice that these premises of observation are classical for natural history methods.⁴

Step I-B. Making the Film Recording

Our techniques for film recording have been described elsewhere (VanVlack 1966). We have also written about the relative advantages of videotape and motion pictures (Scheflen, Schaeffer, and Kendon, 1970A), so I will merely mention a few requirements for obtaining filmed records which are desirable for a context analysis.

1. Preparing for the Filmed Observation

We contract with our subjects in advance and prepare them for filming. If we are to film a relatively small group in a situation that is at all private we explain our purposes and procedures to the participants in advance and seek their permission. In this case we co-sign a written agreement with the subjects. We usually pay them a fee and we promise that the filmed documents and other data will be confidential. We keep the promise.⁵

We set up the equipment in advance of the transaction, choosing a camera location which is inconspicuous but at an appropriate distance from the scene. Then we clear the site of personnel and extra equipment. There is no need for the process to be obtrusive.

The camera should be located, loaded, and turned on before the transaction begins. There is no excuse for technicians or staff members to be visible or audible at the scene. All equipment can be operated from an adjoining room. Nothing is visible, then, to the participants except one or two small cameras and microphones which do not move, make noise, or give off light.⁶

We do not, on the other hand, try to conceal the cameras and microphones. It is our experience that people search endlessly for these devices if they cannot see them. It helps to allow participants to examine the equipment before a transaction. When we are recording in fixed locations such as the office or home, we try to locate the cameras there for days or weeks in advance to facilitate the adaptation of the subjects.

Contingencies can occur after our cameras have been turned on and these can cause the transaction to be unusual. To some extent we can guard against filming distorted occurrences even at the scene of observation: (1) We can postpone the filming if untoward circumstances threaten the event. In an outdoor filming for instance, we may not film if a storm is threatening or the temperature is extreme. If known deviants, aliens, or novitiates attend that day, we may wait for another occasion: (Z) The research team can try to protect the scene. We stay out of the transaction and we prepare the participants so our own presence is not unduly disruptive. We station ourselves at points of entry to ward off intruders, rubberneckers, or incidental traffic, and we try to prevent undue noise.

Other interferences can be minimized. We can use fast film rather than intense lighting. We can use wide angle lens so the participants are not jammed together.

2. Recording the Complete Transaction

When feasible we film the entire transaction. First of all we record it from the beginning to the end. We turn the camera on before the transaction begins so we can record the initial encounter of the participants. In the early minutes they often negotiate their roles and establish the structure and format of the transaction. We are at a loss if we do not have this data on the record. When possible we also want to film the entire duration for it may progress in steps and we cannot generalize about later phases from the nature of the early ones.

We try to include all of the visible and audible behavior of all participants. We do not, therefore, manipulate the camera. It is placed to cover the whole scene and it is left in this position. If we want to zoom for close-ups, we use a second or mobile camera. And similarly we do not pan from one participant to the next and thus perpetuate the myth that people behave one at a time. We do not cut ourselves off from information about bodily communication by filming only the heads and upper torsos of our participants. We are not making an artistic film or studying facial expressions out of context,

3. Obtaining Information About the Context of the Scene

We make notes about events which occurred off camera during the transaction. Often we photograph the surrounding areas and neighborhood. Later we show the film to the subject participants and obtain a record of their thoughts and feelings, for these are regarded as a context of their visible and audible behavior.

Step I-C. Developing a Stratified Sample

Since one filmed example does not prove sufficient for generalization, we locate the transaction we filmed in categories of transactions and build a stratified sample.

1. Locating Our Example in a Class of Transactions

If a transaction is carried out in a number of cultural traditions, there are likely to be major variants in each of these. We must therefore either plan a cross-cultural sample or else avoid making generalizations about communication in general. At least we must identify the cultures and subcultures we sample.

Even within a single tradition there may be major institutional variants. Performances may also vary markedly among people of given gender, age groups, social position, and so on. Thus we try to identify certain categories of transactional type and collect further examples of them.

2. Determining the Representativeness of Each Example

We make every effort to pick examples that are representative of any dass we are studying. To some extent we have done this by exploring the distribution of occurrences and filming usual transactions at usual sites and occasions. We also try to assess how representative the examples are after we have filmed them. We show them to experienced informants of that tradition and ask them if we have captured typical examples.

3. Obtaining a Sufficient Number of Examples

We rarely know in advance how many film recordings we need to picture a transactional type and its major variants. In some cases transactions are highly variable from one occurrence to the next. We may never obtain a rounded and comprehensive sample. In other cases a transaction is highly ritualized. One enactment will be much like the next. If a number of experienced informants review the films and tell us that our examples are typical, we can proceed as though we had a significantly larger sample.⁷ If a transaction is highly customary and standard, we do not need a large number of examples.

4. Stratifying the Sample

In practice we can rarely follow the classical principles of research design. We cannot usually know the adequate size and distribution in advance. As we study a given example we learn of other variants and we come across unexpected and atypical occurrences. In the end, we are using a stratified sample, recalibrating our original plans and adding categories as we discover their significance.

OPERATION II. MAPPING THE BEHAVIORAL EVENTS

When we examine a communicational event we find a good many interesting dimensions for study and deduction. In fact each of these is the province of a classical discipline. Thus we might study the characteristics of the participants. We might make a variety of inferences about their feelings and motives or make deduction about neurological or cognitive processes in general. Or we might study the relationships of group members and the communicational networks in which they operate.

But we are not going to focus on these concepts. In fact we are not going to study physical systems like the group, the individual or the nervous system per se. We are going to study the behaviors of these systems. In the analysis of a group we find that individuals are the constituents, but the elements of analysis in a process like communication are behaviors.

Of the various behaviors we could study we will focus on those that are visible and audible. These are the only ones we can see on the film, and they are the ones which other participants in a transaction can actually perceive. Cognitive behaviors may mediate the processes of participation, but they are not the media or code of communication.

The subject of our analysis, then, is patterns of movement and sound — musculoskeletal and motor activities (and the sounds they produce). These are plotted on a time graph to map the events of the transaction.

Step II-A. Recording Speech Behavior

Speech behavior is formed by ordering a number of traditional sounds. These are made by altering the column of vibrating air with various positionings of the larynx, glottis, tongue, teeth, and lips. We cannot see these motor activities directly, but we can use a coding system that has developed in structural linguistic s.

In a detailed analysis of speech we record an analysis of the phonemic and morphemic structure of language (Harris 19 51; Gleason 19 55). We can code the nonlanguage sounds and vocal qualities by a paralinguistic coding system like that of Trager (1958). In a less-detailed analysis we may be satisfied merely to mark off the syntactic sentences and transcribe the lexical ‘content,’ noting as well the nonlanguage sounds and the gross speech qualifiers.

We insist upon one elaboration of the usual structural analysis of speech. We place the linguistic occurrences on an accurate time graph.

Step II-B. Making a Topography of the Visible Nonlanguage Behavior

We can distinguish and transcribe several types of nonlanguage behavior. In some transactions, the activity centers upon a task that involves physical objects — materials, tools, and the like. Such behavior may appear in any transaction including a conversation. M. Harris (1964) has called this type of behavior ‘actonic’ and he has advanced a method for its analysis which does not differ in principle from the context analysis I am describing.

If the participants speak to or otherwise service each other they also use a coded system of metacommunicative signals which Birdwhistell (1952) called ‘kinesic behavior.’ He has developed a coding system for this type of nonlanguage behavior. In either case the participants will employ a system of postural locations, orientations, and distances (Scheflen 1964, Hall 1963).

1. Setting Up for the Transaction

Our task now is to study the shadows of positioning and movement that are recorded on the film media, order these, and represent them on a graph. Remember that our representations will be descriptions of form — descriptions of the orientation, duration, and excursion of movement. They will not be abstracted qualities, inferences, notations about effort or style, or the like.

We have a copy of the motion picture or videotape and equipment to project the media. The motion picture has been developed by double exposure with a frame-numbered print, so that a number appears on every motion picture frame. And the videotape has been exposed with a digital clock video picture so that clock images appear in divisions of a thirtieth of a second on each video scan. Thus we can tell exactly where we are in the temporal sequences of the transaction and we can measure all intervals. We also have projectors or videotape recorders that will screen the media at normal and at slow speeds and play forward or backward at the touch of a control.

We need, in addition, a media on which to record the notations. This will have to be graphic, so we can see at a glance the location of any behavior. And it will have to be marked off in exact intervals of time.

A great many tapes of media have been tried by various workers in the field. Long sheets of graph or electroencephalograph (EEG) paper are probably the most useful. In the past I have used a metal board lined like graph paper and hung on the wall. Magnets of various shapes and colors are placed on the board to represent various bodily positions and movement patterns. We have also used representative 35mm or Polaroid still shots as mockups on a time graph.

2. Making the Transcription

The graphic media is therefore marked off in time by a series of equally spaced vertical lines. Each column then is a fixed interval of time, maybe eighth seconds for a microanalysis and one- or three- second intervals for a more gross analysis.

The graph is also divided into rows by horizontal lines. In each row we record the exact position (at that instant of time) of bodily region. We delineate a bodily region on the basis of our experience with segmental movement. If some bodily region can be moved separately it will rate a row in our topography. Thus we may have a row for the head as a whole, one for the brows, one for the eyelids, one for the eyes, and so on to the mouth, upper torso, arms, hands, pelvis, legs, and feet.

Now we screen the filmed media. We take one participant at a time and concentrate on the bodily part which we will describe on the first row. We notice that it is held for a while in a certain position and orientation, then it is moved to another place, then another. By showing the film forward and backward first at normal, then at slow speeds, we can record the exact dimensions of these patterns of movement and hold. We note exactly how long the region is held, where it is moved to, the excursion and range of the movement, and so forth.

Then we examine the next region and then the next, and record the information on the topography. Actually we place this data on the same time graph which we used for language behavior so that later we can note the co-occurrence of both language and nonlanguage behavior.

Step II-C. Adding the Contextual Data

We add notes about certain events that are not recorded on the filmed media. Maybe a door slammed or someone called on the telephone or the temperature dropped appreciably. These events, too, we note in a special row on the time graph. And we may have interviewed our subjects by showing them the filmed record. Their comments can be placed in yet other rows in the topography.

We now have a time-segmented, simultaneous record of all of the behavior we can see and hear. The recorded material is laid out on a single graphic record. We will use this record from now on in the analysis. We can carry it home to test out certain hypotheses we have about simultaneous events. We will mark on this same record the unit segmentations as we delineate these in Operation III. We can note the location of any given behavior in the larger picture of behavior which constitutes its context.

So far, you may notice, we have proceeded in the natural history tradition. We have recorded what happened in time. We have not isolated apriori any particular kind of behavior at the expense of the others.

OPERATION III. DELINEATING THE UNITS OF COMMUNICATIVE BEHAVIOR

We do not yet have a realistic picture of the form of behavior. The topography is laid out in rows, one for each bodily region, and we have acted so far as though there are continuous separate streams of activity in each of these regions. In fact, of course, activity is discontinuous. What seem like streams of behavior consist of discrete units which involve a number of bodily regions moved and held in concert.

These units of behavior have a customary or traditional form. The forms are species-specific for animals, and, in the case of man, they are also culturally specific.

Among nonhuman animals those of a given caste and species use the same forms of behavior in common. In man there are additional elaborations of form according to ethic group, class, and institutional tradition. Thus people of the same age group, gender, and social position in any subcultural category will use much the same system of behavioral forms when they are engaged in a given kind of activity or transaction. Thus we assume that behavioral forms have evolved and are transmitted genetically and culturally from one generation to the next. Consequently, children and novitiates in a given category learn to move and position themselves in the same general way.

The behavior of any acculturated person, then, is potentially communicative. It is a medium for communication when it is recognized or related to in concert.⁸

In Operation III, then, our task is to identify and describe these customary units of behavior as they appear in our sample of transactions. Methods for doing this have recently evolved in the sciences of behavior. They can be called ‘structural’ approaches.⁹ The idea of a unit form will take us beyond the diachronic natural history observation we have been using to a synchronic description of recurring, customary behaviors.

To make comparison of unit replications we will begin as we would in a statistical analysis, but we will go beyond isolation of variables techniques. We will develop descriptions of forms in context.

In this section we will still operate at the organismic level, defining the separate communicative contributions of one participant at a time. We do not reach the social level and communicational structure until Operation IV.

Step III-A. Delineation of a Sample Unit

In practice there are two general ways to make use of filmed data and the topography of a transaction. We can study all of the unit forms as a first step to analyzing the structure of the transaction (Operations III and IV). Or, more simply, we can focus on but one type of behavior in order to enlarge a sample of its occurrences. Eventually (Appendix B) I describe the more extensive procedure, but in this section I proceed as though we were studying but one unit type. This makes it simpler to convey an idea of the methodology. Then in Step III-B I will explain how all units are studied.

Here, in summary, are the steps employed to study one unit form: We sketch out its apparent boundaries on the topography, noting when the movement began and ended and how many bodily parts seemed to be associated in its enactment. Then we search the record for similar forms which may be recurrences or replications. We compare each of these by a series of criteria to determine if they are indeed regular replicates. Then we determine whether or not they are customary in some traditional distribution and whether or not they are communicative in a transaction.

A-1. The Preliminary Identification

If we are already interested in some given act we search the record for each of its appearances and we circle each of these, including all doubtful cases. If we have no preconceived idea we can select any movement at all for study.

Marking Off the Segment of Movement. In either case we scan the topography and notice precisely when the movement first appeared and when it ended. The movement started from some given position of that bodily part and ended at the same or at some other position. This is the segment we mark off — the movement or change from the initial to the terminal position. Ordinarily we have no difficulty in identifying this segment of movement. A body part is moved and then brought to rest at its baseline position or at some other comfortable or useful place. Many actions, once initiated, almost have to be completed in a reasonable interval of time. We must put out the match after we have struck it and it is embarrassing to hold our arms out for long periods, frozen midway in the completion of a gesture.

A sequence of movements is not the irreducible segment of action; several subelements are necessary to complete an act. To smoke a pipe we must take it out, fill it, light it, put out the match, and puff. To make a gesture we have to raise the arm, carry out a sequence of hand and finger movements, and eventually lower the arm. In doubtful cases we can leave the record of that particular occurrence incomplete at this point. We can identify exactly the usual segmentation when we compare many enactments of that unit type.

In practice we often find that our topography is inadequate for this task. We discover that we forgot to note when a shift occurred from one position to another or that we did not code exactly some dimension of the movement. In the end, we go back to the original film again and again to correct and amplify our record.

Identifying the Associated Segments: Now we look over each row of the topography to discover movements of other bodily regions which accompany this segment. We may find several kinds of relations here. The adjoining bodily regions may be involved as a pattern of activity *spreads’ over the face or successively involves one, and then the other hand. Or the associated movement may occur at a widely removed area. Some speakers, for instance, shake a foot when they speak. Or regional movements may occur ‘within’ each other. Thus the arm is moved, then the hand, then the fingers, and the fingers are stopped, the hand is held, and then the arm is positioned at rest. Later I will introduce a model for integrating such occurrences.

The identification of associated movements does not have to be highly precise at this point in the analysis. We can mark the doubtful cases and leave questions of coincidence open, for we are now making only a preliminary assessment. We will not know exactly what elements are combined in a given unit type until we have compared a number of recurrences.

Since the complex of movements we have identified is a tentative one, we will not waste time measuring it or describing it carefully. We only note its characteristics sufficiently so that we can recognize other examples when we scan the record. In order that we can find this instance again, we do mark it on the topography putting a dotted line around the probable configuration and numbering it.

A-2. Determining the Regularity of This Complex

Collecting a Sample of Possible Recurrences. We now scan our total record and look for other, similar complexes of movement which may be replications of the same unit form.

There are technical aids we can use in this comparison. Ekman (1969) cuts segments of videotape on which comparable forms appear and splices these together for easier comparison. I often make a short description and a series of representative still photographs of each type of segment and mount these on a chart so the forms will be fresh in my memory as I search for apparent replications.

If we do not find repetitions in our film there is no way we can go on with the analysis. We have to take other films and search these for examples. To determine the regularities of a unit form we must compare a number of examples.

Ordinarily we do not have any difficulty collecting tens or even hundreds of apparent recurrences. In a half-hour transaction the average participant uses possibly a dozen different hand gestures, a similar number of facial sets, and even fewer kinds of head-eye orientations, head movements, or brow placements; he uses the same forms over and over.

How many replications we need to examine depends, of course, on the variability. In some cases each recurrence is so alike in form that we have little doubt about the identity of the set, but in other cases the forms are highly variable and may have to be compared with hundreds of examples and carefully tested by the criteria described later in this Appendix.

We now align and compare all examples of our apparent replicates using any additional techniques we wish: any measurements of means and standard durations, rows of photographs, judges, and so forth. We put probable members of the set to the following comparative tests.

Test 1. The Interdependency of Elements. We compare the complexes we have marked off to see if each one is made up of the same constituent segments. For example: In courtship women simultaneously cock their heads, widen their palpebral fissures, smile, and finger their hair with their palm exhibited toward their partner.

If these segmental acts invariably appear together, if no one appears without the others, then the elements of the complex are interdependent and the complex is a regular unit of behavior. The composition of this entity is not in such a case, incidental, but an irreducible unit at the lowest level of integration.¹⁰

But this is not the case in the example I gave above. Women sometimes preen their hair and present the palm without smiling and head cocking. In this case the hair preening is the lowest unit of interdependent behavioral elements. The whole constellation is a more complex unit at a higher level of integration.

Test 2. The Identity of Form. The same set of bodily regions could be interdependent but still be assembled differently in various cases. In order to claim an identity of replicates we require that the same elements of behavior be combined in each instance to form the same general morphology or Gestalt. The replicates of a form must fall within a range of excursion, rate, rhythm, and duration, for instance. We can measure each of these dimensions accurately, since each frame of the film is numbered, and we can also make spacial measurements from film.

Test 3. Regularities in the Context of Each Occurrence. Because behavioral units are ordered in a transaction by physical and by traditional constraints which correspond to the rules of syntax in a language, the replicates of a unit type occur in the same context or contexts. Each occurs after particular antecedents and before certain successors in a sequence of events. Each occurs in given situations (see Operation V).

Test 4. Contrastibility to Other Forms. The possible replicates of a unit form will contrast obviously or noticeably from the replicates of some other unit forms, even though the two types may be similar to superficial inspection, and even though overlap occurs between extreme variants.

Replicates of any unit type, then, must have (1) the same form, (2) the same conditions of occurrence, and (3) distinguishably different form than some other class of behavioral events. If behavior were not regular in this way it could not be the basis of communication. The forms of behavior in a code must be recognizable and identifiable at a glance. And each must be easily distinguishable from others or there would be too much ambiguity to allow recognition.

A-3. Determining Whether the Unit is Traditional and Communicative

If the same unit form is used by a number of participants in some cultural or subcultural group it is traditional.

If an informant of that culture recognizes the unit we have depicted, we can assume that it is communicative. If he does not, then either we have wasted time defining an artificial configuration or else the unit is one not consciously recognized in that culture. In this case we check its relation to the performances of other people in the actual transaction (see Operation IV).

If we now have assurance that we have depicted a unit with cultural and communicative reality, we invest the time for a careful description. Now we can make careful measurements and do a statistical analysis. We describe and measure the component subunits and depict these with photographs or drawings. We systematically describe the major variations and identify the range of recognizable variants. And we describe the contexts of usual occurrence.

Step III-B. Delineating All of the Units in a Transaction

If we are to study the over-all structure of the transaction and visualize communicational units, we must now repeat the procedures for unit delineation for all of the behaviors we have recorded on film. If we plan in advance to do this, we carry out Step III-A more efficiently using alternative steps I review here.

B-1. Marking Off the Known Units

Many of the unit forms of the behavioral streams on our topography are already well known in common culture or they have been carefully described in previous context analyses. The structural linguists have already identified the forms of English speech and researchers in behavior have described actonic, kinesic, and postural forms in certain cultures. And in many cases we or our informants can immediately recognize some customary forms and describe them in detail.

There is no point repeating the analysis v f these units unless we are uncertain about their form. After all, the biochemist does not repeat the history of chemistry by resynthesizing each component molecule each time he works with a macromolecule.

So we draw a clearly visible line around the rows and columns on the topography in which we have recorded a known and clear unit. Then we turn our attention to the undefined area of activity between these blocks.

В-2. Preparing a Complementary Distribution

When we are to analyze all of the units in each participant’s contribution, we save time by going through the whole topography and tentatively marking off each of the elemental complexes of movement. Then we line up all those which seem alike and compare this set to other sets of similar complexes.

Each complex type may be represented by a symbol. I have used a small magnet of a certain shape and color for each type and plotted these on a large steel board hung on the wall. Then we can sit back and spot patterns of repetition at a glance. If the variants are numerous and complicated we take representative still photographs of each one and annotate these with descriptions and measurements. These become mock ups on a large chart. The apparent replicates of Type A are placed in one column, the apparent replicates of В in the next, and so forth.

Such a comparative representation of types and variants is called a complementary distribution (Z. Harris 1951).

B-3. Cataloguing Unit Types

If a type of unit does not occur often enough in a given transaction to do a structural analysis, we collect occurrences from a number of transactions. In this case we may scan all of the videotapes and motion pictures in our library, and sometimes we borrow the films of other investigators. We photograph and describe each replicate we find and file this data to build a sample. We call such data collection cataloguing.

OPERATION IV. DETERMINING THE UNITS OF TRANSACTION OR COMMUNICATION

So far we have developed a picture of the customary or communicative units that each person has contributed to the transaction, and we have plotted these on a time graph. Hence we know when and where each unit occurred in relation to the others. But we do not yet know how these units were ordered in the part that each participant took; nor have we shown how parts were related in communication. We are ready to take these steps now.

In visualizing communication we can no longer work only with the behavior of each individual separately, but we can exercise an option of priority in our steps toward such a view.

1. We can show the integration of all communicative units in one person’s total part in the transaction and then study how the parts were related in the over-all transactions.

2. We can take each unit we have now delineated and examine its relation to the units of other participants at that moment.

Actually we elect to move back and forth between these options of procedures, refusing to adhere rigidly to either one, and we will do this because of the way behavior is usually integrated in a transaction. At any given moment in a transaction some bodily region of a participant may be more closely related to someone else’s activity than it is to the activities of the rest of his own body. So the unit performance of a participant does not necessarily follow the anatomical integrity of the body and we lose perspective if we insist upon seeing the total contributions of each person as a whole. Similarly the part of a participant may change character as the transaction proceeds, so we do not want to conceptualize it as an entity without respect to the transactional or communicational relations of the transaction.

We, therefore, proceed as follows: We delineate units of movement in a communicative modality. A modality is defined as all of the regional movements which a participant uses interdependently at some moment of time.

Then we study the relations between units of behavior in the various modalities of each participant, thus defining the smallest unit of communicational relation. As a final step, we study the relation between these units until we can visualize the structure of the entire transaction.

Step IV-A. Delineating a Modality

We observe all of the communicative units in the performance of each participant at any moment of time and we draw a line about those which occur together. Sometimes the entire body is moved as a unit: when a participant rises to leave, for instance, sometimes the entire upper body is moved in concert. The participant turns from speaking to A and addresses B, for instance, but he turns from the waist and does not change the positioning and orientation of his pelvis and legs« And sometimes a participant moves only his eyes, looking, for example, from A or В without otherwise moving.

This pattern of movement has a duration. The bodily regions are moved to a new location, held there a while, and then moved back again or moved to another spot. We mark these off collectively as a complex segment of activity through time. Since all component bodily regions sometimes are not moved at exactly the same time, we overview the complex and include shifts that seem to be associated. As in the case of smaller units, there are cases in which we are not sure that we are delineating an entity, but we make an educated guess that we will test later.

Now we repeat the procedures I outlined in Operation III. We scan the entire record (and maybe additional films) to find like complexes which involve the co-occurring change and positioning of multiple bodily regions. We test each example for isomorphism of structure, regularity of context, contrast to similar forms, and customary usage. Then we describe, photograph, and mark off each regularly occurring unit type on the topography.¹¹

Hierarchical structure of this unit. As we examine these units a feature of behavior structure becomes increasingly apparent. Units of this complexity are made up of customary subunits, which are units in their own right when we focus on them one at a time. For example, a movement of the entire upper body occurs as a unit. This unit is composed of subunits of headeye shift, turning of the upper torso, and positioning of the hand. In other circumstances one of these subunits — for instance, a shift in head-eye orientation from one speaker to another — may be an independent unit in its own right. It could have occurred without any shift in the torso or hands. And in still other instances the eye shift alone is a unit of behavior, occurring without any perceptible movement of the head or the body.

In the act of behavior, we conclude, complex patterned movements are composed by integrating some number of patterned and customary subunits in an orderly way. The process of integration is crudely analogous to the way a builder puts together customary units of material to make standard rooms, apartments, modules, and housing projects. In systems theory such integrations are said to be hierarchical and we would hold that both material and behavioral systems are so constituted.¹² In such terminology each stage or degree of complexity is called a level.

We have visualized a behavioral integration by managing our conception of time and space in a systematic way. That duration of time which is required to complete a segment of behavior is visualized as an operational ‘now’ and we conceptualize all sequential events within this ‘now’ as a unit of structure, which we name or represent with a symbol or a picture. We similarly manipulated our conception of space. We declared all bodily regions that are interdependently related in an activity as members of the same ‘here’. Having described their relations as an entity we no longer enumerate them separately. To visualize a hierarchy of units we enlarge our definition of ‘here’ and ‘now’ progressively level by level.¹³

We can now explicate the principle of data reduction in a structural method like context analysis. We do not reduce the complexity of a behavioral experience by disregarding most of the variables or abstracting some feature of the whole. Instead, we discover the form and composition of a customary pattern of behavior; then we can visualize a hundred or a thousand elements of behavior as a single entity or unit. A unit seen as a whole is less complex than its subunits. Thus we recognize, code, and convey with a single word an image of integrations as complex as football, marriage, or a culture.

We might at this point wish to construct the total part which some participant took in the transaction. Spacially a part would consist of some configuration of modalities which corresponded to the multiple simultaneous activities which one participant carried out. Temporarily the part would embrace a sequence of modal constellations. How the part could be analyzed is obvious and needs no further documentation here. Relationships, too, are to be studied by principles which are now familiar. Furthermore, we are now dealing with a relative, small finite number of units, because by now we have already unitized thousands of microbehaviors. Accordingly we can now move rapidly in the analysis of the transaction.

Step IV-B. Studying the Relations Between Modalities

Imagine now that we have marked off on the topograph each unit of modality behavior for each participant. We inspect the topography to determine how each unit of this type is related to each of the other such units.

Some units at this level of integration are immediately familiar. There are relations of tactile contacting, synchronous movements, conversational interchange, tête-à-tête orientations, and so on. We can mark these off with a dotted line and then check all recurrences as we did on Operation III.

We see at first glance two kinds of relations:

1. Enactments in a modality often co-occur

At approximately the same time two or more participants may do something together. They may enact an isomorphic unit, addressing each other in the same head-eye orientation, for instance, or they may perform coterminous units which are very different in form. One may speak, for instance, while the other lowers his head and taps his foot.

2. Enactments in a modality may occur alternately in a sequence

One participant may speak, then the other speaks, and so forth, or the participants may take turns in enacting some subunit of a task or making moves in a game. This alternation is usually called interaction and sometimes a participant ‘goes around the room’ addressing each of the other persons or lighting each of the other’s cigarette.

Procedurally, we do not care whether the subunits of a relation co-occur or alternate. In either case they constitute a communicational unit if they occur regularly in customary integration.¹⁴

In fact, units of communicational relation are at once alternative and co-occurring. For example, participants may take turns speaking while they assume the same posture, and nod their heads rhythmically, and any customary sequence of joint participation appears as a co-occurring relation once it is unitized at the next level of integration.

In a large group we usually discover that multiple simultaneous relations occur among various modalities in various subgroupings. For example, there may be a conversation involving three people but one or two others are dissociated. These two may be engaged in a unit of holding hands and gazing. Two of the conversante may be exchanging glances related to the comments the third is making, and exactly the same postures may relate these two.

Thus at this level multiple simultaneous communicational units occur. Some number of communicative performances may appear that are not visibly related to any other enactments. And we may also find activities which we cannot even demonstrate to be customary or communicative.

Step IV-C. Studying the Structure of the Transaction

Ordinarily a configuration of multiple communicational relations like this will prevail for a few minutes, and there will then be a shift in the type of activity. These constellations of conjoint participation disappear and a very different arrangement is assembled. When we have delineated the units of relation for the entire film recording, we are likely to note that two, three, or a half dozen such constellations succeed each other in the course of the transaction.

When the configuration of modality relations differ at each stage of a transaction we must regard each stage as a unit of communicational behavior. This unit occurs at an intermediate level between the relation of modalities and the transaction as a whole. I have called this unit a relation of positions in other writings. Now our task is to delineate these units which appear as stages of the transaction.

C-1. Enlarging our Sample

We cannot tackle this job with the data of one transaction. Our methodology depends upon the ability to compare a number of unit occurrences. Because a unit as large as a stage occurs but once or twice in a given transaction, we must enlarge our sample and film other transactions.

We must collect this additional data systematically to be certain that we have other examples of the same kind of transaction. Knowing that transactions are culturally and situationally specific, we can expect people of the same background to enact the same kind of transactions, provided they come together at the same kind of site and occasion. So we will return to the same sites to film again.

The easiest approach is to film the same subjects again. In our present project we place video cameras in the home and leave them there for six or eight weeks, filming a number of repetitions of the same transactional types to begin with. And we leave the cameras in site while we study transactions. Then we can go back and enlarge our sample by simply pressing a recording switch.

We also record transactions in an analogous site where people of the same type congregate. To do this we relocate our cameras in another household where a family of the same social, institutional, and ethnic background lives.

By sampling in this way we hold cultural tradition, social position, and the physical ecology relatively constant. We can then allow personality type, mood, and other organismic factors to vary and we can vary the occasion and situational contingencies by selecting when to record.

We carry out a structural analysis of these additional recordings using the methods I have already outlined. Then we compare our data on each of the transactions studied.

Certain common features in each transaction are immediately apparent. Certain kinds of people take the specialized parts. They form given kinds of relations, terminate these, and assume others in a customary progression. Given tasks or subjects of conversation appear in each phase.

If we do not find such regularities we have failed to record a like series of transactions. Maybe our subjects are unusual, or we have not found an analogous site. In any event we will have to go back and make further recordings until we do have a comparable sample.

C-2. Identifying the Stages in the Transaction

When we compare like transactions we ordinarily find clear steps toward a point of completion, or at least a specific constellation of subtasks will appear at the end of which the transaction is terminated. In the case of transactions which feature a product or a physical task performance, the stages to completion are usually apparent. In conversational activities there is usually a sequence of topics and a number of speaker opportunities. In courtship, for example, progression by customary stages is evident. In planned social engagement there may be stages of initial contacting, some number of separate conversate conversations, a group conversation, the serving of food.

There may be no progression of stages in some given enactment of a transaction, but we can often identify a clear reason for this. Someone from the outside may interrupt, for example, or some participant does not take this expected part and the others leave off their performances to deal with his behavior.

С-3. Studying Regularities and Variations

We can now examine the regular and variable features of a type of transaction.

Specifying the invariable unit performances of a transaction. If a task is to be completed there are certain stages or subtasks which must be performed in order to satisfy the physical requirements for completion and to maintain the traditional expectancies of the others. There are, then, certain minimal, necessary unit performances in a transaction. We can identify these when we know the progression of stages, for the activity does not move on toward its goal until these fundamental steps are enacted. We next abstract these units of behavior that occur in every example in which the transaction progressed to completion.

Discovering the allowable alternatives. Except in the case of highly ritualized transactions, however, there is leeway in the composition of necessary steps. Some range of allowable alternatives occurs. The progression is forwarded by any one of some set of substitutable unit enactments or allomorphs. For example, in a usual supper it may not matter what edible food is served or it may be that it has to be cooked food. Otherwise the eating will not continue. But on Christmas the participants may not be willing to eat meat other than turkey.

In a given kind of conversation certain topics may be forbidden but there will be some number of allowable topics. All of those which do occur without disrupting the progression constitute a set of allowable alternatives. We sometimes say that these are allomorphic or equivalent at this level.

Our task, then, is to discover the allowable or at least the usual range of alternatives. We do this at each of the larger levels for the transaction we are studying; i. e. , we find the usual, alternative stages in the progression; the usual, alternative relations within a stage, the usual alternative modalities within a relation, and so on. After we have described the necessary units and alternatives, we can abstract the customary program of that type of transaction (Scheflen 1968).

Defining the normal variants in performance style. Members of a culture know the usual units and allomorphs of familiar transactions. Although they may not be very conscious of this knowledge, they can quickly spot an incongruent or deviant performance, and they signal such occurrences by comment, kinesic act, or a refusal to continue. So we can also discover the variations in form which are acceptible or allowable in a given population.

I will call those variations in unit form which pass in a transaction stylistic variants. They are often associated with participant traits, such as age, health, status or skill, mood, personality type, region of origin, or previous cultural membership.

Identifying various types of intercalated behavior. There also appear in most transactional performances supplementary or intercalated behaviors; behaviors that are not ordinarily or nесessarily unit elements in the program of that transaction. I have so far studied three kinds of these in my research on transaction.

1. Unit performances appear in a transaction which belongs to the programs of other types of transactions. Such behaviors can often be explained as attempts by a participant to change the definition of the situation (Goffman 19 57), or the consequences of some misunderstanding about what transaction is called for. We might infer ‘motives’ in the case of these supplementary performances.

2. Certain actonic sequences may be intercalated to maintain or alter the ecology of the situation. Someone may rearrange the furniture, for example, or turn up the heat.

3. Certain intercalated kinesic acts or lexical comments refer to the ongoing performances and relations and serve to correct, regulate, control, or reduce the ambiguity of these communicational processes. Following Bateson (1955) I will call such behavioral intercalations, metacommunicative behaviors and metabehavior. These behaviors will further concern us in Operation V.

С-4. Identifying Instances of Communicated Variation

Now that we have identified the regular or traditional forms of a transaction and the recurring variations, we carry out a final task in the definition of transactional structure — we study relations between commonly occurring variations. We examine the common allomorphs, stylistic variations, and intercalated behaviors to see which of these are repeatedly associated. We use the now-familiar procedures of context analysis to examine the social context of each variation. In some cases we will find that a variation occurs in the performance of one participant alone. There is no evidence at the time of a spread or interaction of variation among the participants.

In other cases the occurrence of variation in the performance of one participant is associated with the occurrence of variation in the activities of others. In fact, a flurry of variant performanees may appear at times in the transaction and sometimes these seem to disturb the progression of the transaction or even break up the group.

When variations in multiple performances are associated we can assume that an interaction or a communication of variant or deviant performances has occurred. We note such instances prominantly on our record, but we must be careful how we explain them. It may be that Smith and Jones are not reacting to each other, but rather performing variantly in the face of events in other contexts.

OPERATION V. THE CONTEXTUAL ANALYSIS OF THE BEHAVIORAL PROCESSES

Bear in mind that we now have acquired a structural view of one kind of customary transaction type. We know the fundamental units of communicative behavior and the necessary relations among these which replicate the program. And we know the recurring variations that occur red in our sample.

There is a good deal more we would like to know about the communicational behavior we have described. In presystems terms we would like to know the determinants of the events. Or conversely we would like to know what consequences, effects, function, or meaning these events have. In systems terms we would like to know the dynamics or systematic s of communicational processes in context analysis. We explain communicational events by relating them to larger systems of events or contexts.¹⁵

We are not satisfied to explain with simple determinisms and linear causation. As we do not settle for a motivational or a social or a cultural explanation, we must deal with all the contexts from which these classical constructs are derived. We will not rest with finding first causes of behaviors, but will repeat the examination of unit-context relations — this time relating unit occurrences within our filmed record to events in other systems of behavior.

Elsewhere I have described a model of interrelation (Scheflen 1970) so here I will merely describe methodological steps in the study of contexts. Starting with a look at yet other aspects of the immediate transaction (Step V-A), we then move to ways in which less immediate contexts can be scanned and sampled (Step V-B).

Step V-A. Studying the Immediate Contexts of the Transaction

By the time we have visited a number of sites and made a number of filmed records of a given type of transaction we will have abstracted the main features of the context of occurrence. As I already pointed out a given transaction occurs at usual sites on usual occasions and ordinarily involves particular kinds of people. With more careful observation we can also identify the stable features of the immediate context — those that remain unchanged throughout its enactment. In contrast, then, contextual changes which take place during the performance can often be associated with variations in relations and in the communicative behavior of the transaction.

Visible Elements of the Immediate Context

Two aspects of the immediate context may be represented on the films we have made.

1. Visible elements of the immediate ecology. Our films will show walls, furniture, props, and tools if these are used in the transaction. In addition the actonic behaviors of the transaction will show direct relations to these objects, and the territorial behaviors of the participants will be interdependent with the physical boundaries and obstacles at the site. In many cases we will also have data about climate at the site and here again the participants’ behaviors may reflect these ecological states and changes.

2. Perceivable indicators of suborganismic states and changes. Participants will provide indicators of their suborganismic states by autonomic signs, utterances about these matters, and by the qualities of behavior and demeanor.

Subjectivist Data about the Immediate Context

We can also interview subjects about their sensations, affects, thoughts, and perceptions during the transaction. We can also ask them about immediate contextual situations. At present we show participants the video or filmed records of the transaction asking them in detail about their usual and variant performances.

Step V-B. The Study of Broader Contexts

We also obtain what information we can about the larger contexts of the transaction and try to relate the features of or changes in these larger contexts to the events of the performance. To some extent we can gather such information by further study of the transaction itself, but we also have to enlarge our sample, thereby studying still other occurrences.

The Representation of Broader Contexts in the Transaction Itself

Many of the behaviors of the transaction itself make reference to broader contexts. Elements of decor and dress, statements made in the transaction, indications of approval and disapproval, and subjectivist comments on interviewing may provide information about the traditional practices, rules, and values of the culture in which that transaction ordinarily occurs. These behaviors may tell us about the social organizations under whose auspices the transaction is held.

Often one or more participants belong to somewhat different traditions or to different social networks. They thus have different commitments in the transaction which may relate to variations or deviations in their behavior. In these cases we may have to seek data about discrepancies between the memberships and identifications of the participants. I have elsewhere claimed that performance styles and predilections for certain allomorphs and parts discernible in a transaction provide information about the background of each participant and thus allow us to predict the qualities of his performance. I also suggested that these features of behavior be called paracommunicative (Scheflen 1971).

But because these kinds of information depend upon inference, we next will make direct observations of the broader context.

Strategic Sampling as a Quasi-Experimental Approach

We can expand our sampling in four directions to account for most of the variants we have been studying: (1) The same transaction type can be examined in other ecosystems. (2) The same transaction can be studied with people of other social positions and relationships. (3) The same participants can be studied in other kinds of transactions. (4) Analogous transactions can be examined in other cultures.¹⁶

These directions provide a view of contextual factors that are known to influence the form of behavior:

1. Ecological conditions of the transaction. The sites, occasions, and ecological conditions at the time of meeting will determine the type of transaction. In some instances transaction type is ritualized and the same communicational events occur again and again at prescribed sites and occasions. In other cases given transactions are held whenever certain contingencies ‘call for’ their enactment.

2. Social relationships and social organization. The durable and pre-existing relationships among people will determine their relations and parts of the transaction.

3. Traits of the participants. In addition to cultural origins and social position, participants vary in physical and affective state, age, gender, skill, etc.

4. Culture. We know that the form of behavior and relations are shaped by the cultural and subcultural memberships of participants, including ethnic background, social class, region, and institutional tradition.

By planning to make observations under such conditions we can in effect hold three kinds of contexts relatively constant while we observe variations in the fourth. For example, if we can find analogous transactions among participants of roughly equivalent position which occur under comparable physical conditions, we can observe a number of cultures and subcultural origins, and gain a rough idea about the variations which are incident to tradition.

We can observe the same subjects engaged in other transactions at the same site to identify the influence of the program of transaction, and so forth.

If we select a sufficient number of carefully selected exampies, we can achieve a quasi-experimental result without creating the artificial relationships and tasks of a laboratory experiment.¹⁷ Such methods have been called naturalistic experiments.

The Scanning of Existing Film Records

There are already a great many films and videotapes of human behavior we can use to scan broader contexts. In some instances one can use commercial films or videotapes, but often these are fragmented into overly short sequences and are replete with close-up shots which preclude the study of communication. But the network of researchers in the audiovisual field has now become large enough that it is possible to borrow hundreds of research films or tapes made in a large variety of cultures and situations.

We scan these rapidly. When we spot a familiar unit or relation we check the context quickly to make sure we are not dealing with a morphologically similar form which actually occurs in different cultural contexts and therefore has a different significance.

To record these observations we can make still photographs from the movie or videotape and describe the contexts with other photographs or written descriptions. If we are to study a form more carefully we can request the film owner’s permission to copy a segment of the medium.

By this time, however, we are much less dependent on film recordings to make our observations. We are more experienced observers and we have baselines in our mind with which we can make a rapid comparison. We can now make some of our observations in everyday life without film recording.

The Final Product of a Context Analysis

The outcome of a context analysis is a set of systematic descriptions implemented as much as possible with segments of representative movies, videotape, or still photographs. The following types of descriptions accrue:

The structure of a type of transaction and the usual, programmatic alternatives are depicted hierarchically.

Common, repetitive variants are also described and related to ecological contexts and social structures.

Certain variants related to traits of the performers and certain cross cultural variants are described.

Particular unit types of communicative behavior and relations may be singled out for more careful description. Common variants of these in various types of context are also described. From the analysis of their relations to context we may also derive the functions and meanings of these forms.

In general we can claim that customary units are enacted to maintain transactions, institutions, and cultures, and that variants restore dynamic equilibrium and adapt organismic, social, and ecological systems. In this framework we can try to formulate the functions of unit performances.

Many units also have meaning in the traditional linguistic and cognitive systems of a culture. Variants in style and the choice of allomorphs related to individual and cultural differences are also meaningful in deriving information about the participants, and such assessments on the part of participants may influence the transactional performance itself. The research may formulate the meaning of units by the study of subjectivist contexts in both of these dimensions.