Chapter 4: The Cognitive Analysis of Behavior in Activity
Compositors: Michael Cole, Lois Holtzman, Ray McDermott, Ken Trauppman

In the constant talk about extrapolating from the experiment to the ‘real’ or the ‘social’ world, we must not forget that the experiment is itself a part of that real and that social world. ––Friedman, 1967

A recurring theme of the work described in the previous chapters is our disenchantment with the use of standardized psychological tasks, whether used as psychometric or experimental techniques, to draw meaningful conclusions about cultural variations in cognitive abilities. Whether dealing with cultural differences in remembering and logical problem solving, or trying to design an experiment that adequately models Vai writing activities, we had come fully to agree with the epigram at the head of this chapter. The challenge was to work out the implications of an alternative, more adequate way of understanding psychological tasks as a particular form of activity and be able to “locate” the experimental events with respect to their socio-cultural context.

However, we were well aware that such a commitment in no way solves the problem of how to carry out such analyses in a way that makes methodological contact with the procedures of cognitive psychology on the one hand, and cognitive anthropology on the other.

The approach that we adopted, like the earlier work in Liberia and the Yucatan, began so to speak from each of the two disciplinary poles: experimental psychology and social anthropology. On the one hand, we worked “outward” from the experiment as the embodiment of what cognitive psychologists define as a psychological task: a goal and constraints on achieving it. On the other hand, we worked “inward” from the socio-cultural context to understand the interactional dynamics of the test situation, its constraints and prescribed goals. By combining insights from the two converging modes of analysis, we hoped we could open up the examination of the ways in which standardized tests are social events, whose relation, as social events, to other events in people’s lives we badly wanted to understand.

Cole and Scribner (1974) quote a Kpelle proverb in their reflection of this dual orientation: “I know how to begin the old mat pattern, but I do not know how to begin the new.” They might, perhaps, just have written, “easier said than done.” Criticism is the easy part. Arriving at deeper understanding of psychological tasks in relation to both the individuals engaged in them and their social contexts is a whole lot more difficult.

Addressing the need for a deeper understanding of the ways in which psychological tests are social achievements, LCHC was fortunate to be able to attract a number of scholars who specialized in the study of social interaction: conversational analyst Emmanual Schegloff, ethnographer Ray McDermott, microsociologist David RothLois Holtzman, developmental psycholinguist, joined this group in 1976. In parallel, stopping in for seminars and shorter visits we began to engage in active exchanges with Courtney Cazden and Bud Mehan, who were conducting an interactional study in a classroom with Cazden as the teacher. This network, which grew to include colleagues at Georgetown and the Graduate Center, was an important resource for us as we began to move into uncharted territory, where a new, coherent if not perfect, methodology could be created.

These interactions across disciplines and institutions, came to be formulated around an issue in psychology that had disquieted its development since its founding as an empirical science in the late 19th century: the problem of ecological validity. Ecological validity refers to the extent to which the psychological processes observed and recorded in a particular psychological experiment reflect the processes that actually occur in other, naturally occurring settings. Ecological validity is also used to address the problem of the “generalizability”: to what extent is it legitimate to apply the lessons learned in the experimental setting to the “real world” of events not deliberately set up for purposes of psychological testing?

We referred to this research as the MCS Project because our partner in the work was the “Manhattan Country School,” located at 96th St and Madison Avenue on the border between black/white, impoverished/affluent populations in the city. The school was deliberately highly diverse, with children representing a broad spectrum of social class and ethnic backgrounds.

The core project participants (Mike Cole, Lois Holtzman, Ray McDermott, and Ken Traupmann) decided to undertake a direct analysis of the relationship between cognitive psychological tasks, classroom lessons, and more informal, less restricted forms of activity where cognitive processes might be studied “in action.” We could then answer the essential question about cognitive consequences of schooling: Can cognitive difference be seen to operate in everyday settings that are not controlled by a researcher or a teacher? More generally, we would achieve a much better understanding of all of the social constraints that come together when children are taking formal cognitive tests, participating in classroom activities which are presumably rich in cognitive activity, and in other forms of group activity.

Begun in the fall of 1976, this research involved 17 children who were 8-10 years old and from a variety of ethnic and social class backgrounds. Roughly half were boys and half were girls. Our approach was based on the idea that if psychological tasks organized by psychologists actually measure the sorts of things that also take place outside of experiment interactions, then we should be able to see people engaging in such tasks in a variety of settings and be able to say something about how they are doing it.

We wanted to conduct this work in such a way that we could record the interactional dynamics that converged to enable recognizable psychological tasks to occur in each of three kinds of social settings: psychological tests, schoolroom lessons, and afterschool enrichment activities. On the one hand, the configuration of standardized test interactions served as the starting point of analysis in a search for “everyday life” settings that produced similar forms of events. On the other side we began with an analysis of the social interactions in the less constrained settings that converged to produce social interactions that mapped onto the formal testing situation.

The series of tests we selected were meant to be representative of tests used to evaluate scholastic aptitude or cognitive development. So, for example, a free recall task has the goal of committing to memory as many items as you can with the constraint that you do not write them down or “cheat” in any way, that is, achieve the goal but break the constraints. These procedures highlight the subject’s ability to carry out the psychological task without overtly interacting with the environment except in the closely prescribed manner.

The tasks were administered by a professional tester who did not know the purpose of the study. Our battery included: (1) the similarities subtest of the WISC, which was modified so that, in cases in which the child experienced difficulty, a third item was added to the original pair (e.g., we added onion to turnip -carrot); (2) a mediated memory test first developed by Leontiev (1932) and Luria (1928) that allowed children to use one set of pictures to help them recall another set; (3) a figure-matching task of the sort used to assess impulsiveness; (4) a syllogistic reasoning task; and (5) a classification task involving common cooking and eating utensils.

We reasoned that because psychological tests model school practices we should be able to “see” psychological processes at work in both a test and a classroom. Once identified, we could then determine if the child performs equally well in both settings. But what would happen if the social situation became even less constrained, yet saturated with potential cognitive tasks?

To address this question we also engaged the children in weekly activities such as “Nature Club” and “Cooking Club” that further to reduced the frequency of school-like, and thus test-like. It was an open question whether tasks occurring in the clubs or the test situation arose in the clubs, and if so, could we analyze them usefully?

Read more…

Repeated observations over several iterations of the clubs revealed that the conceptual apparatus of cognitive psychology could, when there was the right combination of circumstances, provide a plausible explanation of differences between behavior in laboratory and what we shall call “everyday life” settings (represented in our case by the clubs). Sometimes these tasks arose spontaneously. Sometimes we tweaked the social organization of the activities at the club as a way of making the cognitive work being done by members of the group more visible and public.

An example in which example is one in which an arithmetic/memory task arises in an unpremeditated manner as the children are gossiping during a lull in the activities. They begin to discuss how many rooms the children have in their apartments. Dolores initiates the task of remembering by asking Jackie how many rooms she has and then asserting that she thinks that the correct number is 13.
Dolores: (pointing to Jackie) Well she has thirteen
(rooms). No you don't.
Jackie: Uhhuh, I just made, I, we just made a bathroom.
Mike: They cut one (room).
Dolores: (to Jackie) Count them all (Dolores holds up pinky).
Jackie: Three bathrooms (holds up three fingers)
Dolores: That's not a room.
Reggie: They're rooms. Bath rooms.
Jackie: No, I'm counting bathrooms. For my (holds our arms) whole house we got thirteen rooms. There are two (inaudible).
Dolores: (holds up 3 fingers) Three, two (adds 2
more fingers).
Jackie: Five (holds up 5 fingers also; Dolores adds
one more). 0.K. (Jackie holds up six too).
Dolores: No five (holds up five).
Jackie: 0.K. (holds up 5 too). My' um' mother's and father's room. Six (both hold up six). Then we have (Pause) um.
Dolores: Your room.
Jackie: My room and my brother's room (both hold up
eight) and then we have um.
Dolores: The living room (9 fingers).
Jackie: The living room (9 fingers).
Dolores: The kitchen (10 fingers).
Jackie: The kitchen (10 fingers). And the place where we eat. The place where we eat. 11. And then we have, then when you walk (moves hands indicating direction) (inaudible) and you walk here. You know (pause) well we made a new room.
Dolores: Oh, well you counted that already.
Jackie: No, I didn't count the new room.
Dolores: Yes, you did.
Reggie: Yes, you did count the living room.
Dolores: No, the new room. You counted the new room. (Discussion continues)

Here both deliberate remembering and arithmetic reasoning are called for, and the problem is solved. However, the process is not clearly attributable to any of the participants, many of whom contribute in ways that are relevant to completing the task before the group. The task, has, in a visible sense, been distributed throughout the group. For example, we can see Dolores storing part of the information for Jackie (with her fingers), actively joining in the recall process, supplying the necessary information (line 13, line 1), and contributing by acting as a mnemonic cue for Jackie (line 15).

That this task (and many others like it that we identified in our video tapes) was distributed throughout the group makes visible the socialness of cognitive activity. The individual person (and her/his head) is not solving the task. Therefore, it is not valid to take the individual as the unit of analysis of cognitive activity. We were observing children and adults continuously arranging their environments in such ways that they did not need to engage in cognitive activity without environmental support. “Non-internalized thinking” in which cognition resides in the environment as much as in the individual was more common than we might have thought. We began to move in a direction of what we hoped would be a more ecologically valid psychology, with the unit of analysis being the person-environment interface (environment, in this context, encompassing the cultural and social surround, including of course other people).

We were able to collect a number of similar examples from our many hours of recording. Although a great deal of the time, it was difficult to say that particular children were confronting “a task” or to specify how they dealt with it. So long as one observed the behaviors of the children through the lens of cognitive psychological task analyses, the club behaviors seemed to reflect a standing rule of social interaction; get past the difficulties by changing the overall division of labor in the group to minimize the effort of the task. From this perspective, cognitive tasks encountered in everyday life appear to be “easier” because they can be strategically re-assembled and minimized.

At the same time, when one closely observed the video tapes, it became clear that in the less formal, more interesting, environment of the after school club, the social environment was not always coordinated and supportive. Children virtually never entered any of the activities with only one task on their minds; we had learned that lesson on day one. What we were unprepared for was the constant complexity of the intertwined social relations among the children, and the ways in which they entered into the joint club activities with their identities constantly on the line.

As we were in the process of coming to grips with the ways in which organization of the club activities changed the “cognitive division of labor” involved in solving “the task” an unexpected event occurred which sharpened our focus on the intricate interplay between children’s abilities, including academic abilities, and the way in which they deploy them to pass as full members of the social group. The case concerned a child who had been diagnosed as learning disabled.

Some relevant background: In order to avoid “contaminating” his ideas about the children, Mike had avoided participating in the observations at the school – he was one of the “club guys.” After a few club sessions had taken place, it seemed an opportune time to visit the school and compare notes with the classroom teacher whose lessons were being observed and recorded. As they began to chat, the teacher asked Mike if Archie was proving a problem in the clubs. Mike was surprised at the idea. Thus far, Archie had seemed eager to participate in the activities; he was generally helpful, compliant. By contrast, one of Archie's companions, Reggie, was a big problem for Mike because Reggie kept getting into interpersonal difficulties that tended to spill into the hallways and disturb our colleagues.

“Didn’t you know,” she inquired. “Archie is learning disabled.” The teacher’s news was a shock to the entire group. We had been through several sessions of the Club by this time, and we had observed in the classroom, but nothing in either setting had alerted us to this possibility. How could Archie be learning disabled at school and yet a perfectly normal participant in the Club where lots of reading and other forms of learning were pervasive?

Primed by the notion that Archie had difficulty with text processing at school, we began to focus carefully on moments in the flow of activity when Archie had to deal with print or some specific language issued. When we went back to check, we found that Archie indeed had experienced some distinctive difficulties in the formal testing procedure, especially on language-related tasks which are implicated in the notion of a “reading disability.” We also noticed from video recordings in class that he was a master at avoiding being called on to answer a question while seeming to be an ardent participant.

Our focused interest on moments when Archie was encountering print during the Club periods coincided with a realization that we needed to pay far more attention to our own tendencies to fill in and help out. Everyone, researchers included, shared the common goal of getting the cake baked before the bus came to take children home. We wanted the kids to have a good slice of cake just as much as they did. And we were acutely aware of the time constraints. We decided to ease back on our own “filling in” actions to the extent possible. The children would have to do more of the work. In the process of stressing the children’s abilities, we hoped, we would be able to create a “slow motion” version of their cognitive processes at work.

This strategy worked particularly well. Some of the most interesting observations from our subsequent work came from two occasions when Archie worked with Reggie. There was some tension in the relationship between the two boys because, in prior weeks, Archie had shunned Reggie in order to work with Tony, another child in the class. On the day in question, two deviations from the normal routine coincided. By prior arrangement, Ken told the children that I was not available and that he would be busy with some paperwork over in the corner. He had already gone over the routine with them. They were on their own.

We will summarize the subsequent interactions to save the reader the labor of transcript reading and hopefully provide an adequate account. Keep in mind that while Archie is a poor reader, but a good organizer of others in behalf of his need to know what the printed word said, Reggie was a very able reader who had a great deal of trouble keeping his mind on the task at hand. His flightiness made it difficult for others to cooperate with him and he stood at the bottom of the groups’ choice of a partner to work with.

On this occasion, Tony is absent, so by default (the girls have already gone ahead to organize their activities), Archie and Reggie. The tape reveals the pathos of Archie's situation and his remarkable socio-cognitive skills in dealing with it.

At the beginning of the session, Reggie very deliberately refuses to help Archie get the information they need to bake a banana cake. Archie attempts to get the information from Ken who is sitting on the floor against the wall, putatively doing important record keeping work. Ken, it should be noted, is annoyed with Archie because Archie was not paying attention when he had gone over the instructions with the whole group. After helping Archie decode a few items in the recipe, Ken loses patience and tells Archie to figure it out by himself. Reggie picks this moment to go to the bathroom, leaving Archie to figure things out on his own. He tries to elicit help from the girls, but they are engrossed in their ongoing projects and don’t have much time for a boy. In the interaction, they make it clear that Archie has failed to distinguish teaspoons from tablespoons and baking power from baking soda. Archie retreats to his corner of the table, bent over in confusion about what to do.

Reggie arrives back on the scene. He sees that Archie is upset. “You upset, Archie?” he asks. Archie brushes his hand away and says, “Come on Reggie.” The two boys focus together on the instructions and Reggie begins to read, as Archie begins to carry out the actions that he now has access to via Reggie’s ability to read. Reggie, in the meantime, is being a model student, under the control of his joint cake baking activity with Archie. The cake got baked. The following week Reggie begins by telling Archie that he will be cooperative, and the cognitive division of labor seen previously is replicated without special drama.

Our conclusion at the end of this work was that by and large, the use of standardized cognitive-psychological experimental procedures implies that a closed analytic system is being successfully imposed upon a more open behavioral system. To the degree that behavior conforms to the pre-scripted analytic categories, one achieves ecological validity in the classical sense. Yet our work, as well as studies by others showed that even psychological tests and other presumably "closed system" cognitive tasks were permeable and negotiable. In so far as the psychologist's closed system does not capture veridically the elements of the open system it is presumed to model, experimental results systematically misrepresent the life process from which they are derived. The issue of ecological validity then becomes a question of the violence done to the phenomenon of interest owing to the analytic procedures employed.

In this sense, we argued, ecological invalidity is built directly into the standardized test procedures themselves.

We had originally set out to answer a number of questions: Are the cognitive tasks that have been studied in the laboratory actually encountered in various classroom and club settings? Could we show similarity or differences in the behavior of individual children for tasks encountered in the different settings? Granting that the exact form of a given task would differ according to the context in which it occurred, could we specify how the context influenced the particular form of the task and the child's response to it?

Any initial optimism that we could easily identify cognitive tasks outside the laboratory and classroom and answer these questions was clearly wrongheaded. We could, with a lot of labor, identify cognitive tasks going on, but the idea that we could make within-subject comparisons across settings was clearly a formidable task. We managed this accomplishment in the case of Archie and his pal Reggie to a limited degree, but the amount of labor was enormous and the results, while illuminating, pointed to great difficulties in making a general practice of such comparisons. At the same time, we had made a great deal of headway on specifying why it is so difficult to generalize between laboratory experiments and psychological tests and everyday life. Their different social structures rendered such comparison implausible.

In trying to come to terms, theoretically, with what we observing, we found ourselves increasingly interested in the writings of Lev Vygotsky which we had viewed with extreme skepticism. At the same time, we had to come to terms with the theories growing out of the study of social interaction and discourse, which had been beyond our reach in the cross-cultural research. We deal with these two closely lines of theoretical development in Chapter 5, both of which remained signature features of Lab research in the years to come.

comment.JPGVisit the Chapter 4 Discussion Page

If you have additional changes or suggestions for the wiki content please send to

Chapter 4 References and Additional Resources

Next →
Continue reading with Chapter 5.
Or, move ahead to Section 3, LCHC at UCSD.