Reconsidering group cognition: From conceptual confusion to a boundary area between cognitive and socio-cultural perspectives?

Various strands of research in educational, social and organizational psychology focus on structures of collectively created meaning that emerge in and coordinate activities of groups. Despite expanding, this field still lacks conceptual clarity, enhanced by the multi- tude of terms used, such as common ground, shared understanding, collective mind, team mental models, and distributed cognition. We conducted a review of the conceptual frameworks being used in empirical studies, focusing on the premises of the conceptu- alizations. Therefore, we connected these conceptualizations to either cognitive or socio-cultural perspectives on the social nature of cognition. Some studies are identified as representing initial ways of boundary crossing between these perspectives. To conclude, we explore ways for boundary crossing and cross-fertilization in future research. p. 39

Fundamentally, both the working teams in organizations as the learning groups in schools are con- fronted with the same issues; “they are faced with challenges of establishing common frames of reference, resolving discrepancies in understanding, negotiating issues of individual and collective action, and coming to joint understand- ing” (Barron, 2000, pp. 403–404). p. 39

literature there is an identification of the need to study group learning as truly collaborative (Crook, 1998). Researchers became convinced that an individual approach will not fully grasp the phenomenon of group-work and group-learning (Thompson & Fine, 1999). Several cognitive constructs, such as mental models, which have traditionally been consid- ered at the individual-level of analysis, now become recognized as group-level phenomena (Klimoski & Mohammed, 1994). In trying to grasp and understand this collaborative effort, new research interests focus on the ways groups are creating meaning and are acting upon collectively developed cognition (Thompson, 1998). p. 40

multitude of terms that can be found in the literature, such as common ground, team mental models, shared understanding, distributed cognition and collective mind p. 40

Group cognition is argued to provide a basis for the coordination of individual actions as well as for future communication and activity of the group (Clark & Brennan, 1991; Hutchins, 1995). p. 40

The development of group cognition is a process of negotiating and interrelating diverse views of group members. This process enables group members to learn from others’ preferences and viewpoints by facing different viewpoints and by accepting the existence of them as legitimate (Engestro ̈m, Engestro ̈m, & Ka ̈rkka ̈inen, 1995). Moreover, this process is argued to lead to rich argumentations and creative problem solutions (Homan, 2001; Matusov, 1996), as well as to the members experiencing ownership of the activity and of group products (Fiol, 1994; Mohammed & Ringseis, 2001). p. 40

Carley (1997) examined the “team mental model” by looking at the similarity of declarative and procedural knowledge about the task. A different study is that of Levesque, Wilson, and Wholey (2001), who examined under the label “shared mental model” the similarity of cognition about the team processes and expertise in the team. Yet another study of Yoo and Kanawattanachai (2001) focused on the ‘collective mind’ of groups, hereby referring to the social cognitive system in which individuals heedfully interrelate their actions. p. 40

To address the question of how to frame the diversity of conceptualizations of group cognition, this article draws on Valsiner and Van der Veer (2000), who clearly identified and distinguished two socio-genetic views, one underly- ing cognitive and one underlying socio-cultural perspectives which entail basically different ideas about the social nature of cognition. p. 41

Valsiner and Van der Veer distinguished two socio-genetic views that theorize this relation differently. They state “there is the axiomatic preference for fusion (of person and the social environment) or inclusive separation (i.e., the person is viewed as distinguished from the environment, yet interdependent with it) bases for socio-genetic models” (p. 6). Whereas the former socio-genetic view perceives the person and the social environment as one whole and relies on terms like participation and adaptation, the latter socio-genetic view perceives the person and the social environment as separate units that are related to each other and uses terms such as internalization and externalization. p. 42

8). Salomon and Perkins (1998) noted that the different ways of understanding social contributions to learning are the result from different ideas about where information processes lie (within the individual mind versus within social interaction) and about what entity these processes serve (the individual or a social entity). It is the first dimension, where information processes lie, that distinguishes the two socio-genetic views. Combining these distinctions, the one view on socio-genesis is the “inclusive separation” view, which is related to the “acquisition metaphor” and connected to “learning within the mind”. The other view on socio-genesis is the “fusion” view, which is related to the “participation metaphor” and connected to “learning within social interaction”. p. 42

These two socio-genetic views also lay behind a previous discussion between researchers from the cognitive per- spective and those from the situated perspectives as found in psychological theories (Anderson, Reder, & Simon, 1996, 1997; Cobb & Bowers, 1999; Greeno, 1997). Also, several attempts have been made to create initial bridges between the two perspectives (Billet, 1996; Gauvain, 2004; Glick, 2004; Greeno, 1998; Saxe & Simonde, 2004) p. 42

In this article, we have chosen to use the general term “cognitive perspectives” which is associated to the inclusive separation view of socio-genesis and in a general sense the term “socio-cultural perspectives” to refer to the fusion view of socio-genesis. p. 42

In the one view, underlying cognitive perspectives, the individual is seen as an autonomous agent, an active person who constructs personal understanding of the world surrounding him or her. This understanding is reflected in a mental network of internal constructs of meaning stored in memory. The social world surrounding the individual is seen as a set of social contexts in which the person acts. These contexts are considered important and although complex, can be analyzed through its components. The individual moves through these contexts, responds to them and is affected by them. These notions indicate that the social is certainly not denied by these perspectives, but that the social is understood through its residence in the mind of the individual. p. 42

In the fusion view of socio-genesis, underlying the socio-cultural perspectives, the individual is seen as a participant in social practices, in which he/she is interacting with others and with material and representational systems. Participating in social and cultural practices contributes to the construction of the participants’ dispositions to agree with certain propositions or routine practices, and as such to his or her identity. The socio-cultural perspectives do not deny the p. 42

As for the conceptualization of cognition, the studies connected to the cognitive perspectives refer to the knowledge of the individual team members. More specifically, these studies rely mostly on the construct of mental models (Klimoski & Mohammed, 1994). This construct assumes that individual people organize knowledge into structured, meaningful patterns and store them in their memory (Johnson-Laird, 1983; Rouse & Morris, 1986). p. 45

The most elaborated categorization of the types of mental models and their knowledge content is proposed by Cannon-Bowers et al. (1993). They discern four types of mental models that can be shared in teams: o Task model (e.g., group cognition regarding task strategies, environmental constraints); o Team interaction model (e.g., group cognition regarding interactions patterns, roles/responsibilities); o Team model (e.g., group cognition in terms of awareness of team-mates’ knowledge, skills); o Equipmentmodel(e.g.,groupcognitionregardingasharedideaaboutequipmentfunctioning,operatingprocedures). p. 45

We found three studies (Derry et al., 1998; Banks & Millward, 2000; Yoo & Kanawattanachai, 2001) that we consider to be boundary crossing literature, because they were able to pursue a mixed discourse of both cogni- tive as well as socio-cultural perspectives on mind throughout the whole study (see Table 2). p. 48

These three studies focus both on a stable cognitive map (or memory structure representing information and form), and on dynamic, situated cognitions representing coordination of information and actions. Derry et al. distinguished the individual long-term memories and the thoughts shared by the individuals during the group work. Similarly, Banks and Millward (2000) made a distinction between on the one hand a mental model form, representing a map of elements and their relations and on the other hand mental model states, representing the dynamic configuration of the aspects of the model that can be changed when running the model. Yoo and Kanawattanachai (2001) distinguished transactive memory, referring to information that is encoded, stored, and retrieved from the memory of the individual mem- bers and the individual minds at work during group activity. These three studies were all concerned with the ideas and information in the teams relevant for conducting the task, or task models, and how these were created in the teams. p. 48

Cannon-Bowers and Salas (2001b) provide an overview of four interpretations that are outlined in the literature pertaining to the cognitive perspective (Cooke, Salas, Cannon-Bowers, & Stout, 2000; Mohammed & Dumville, 2001; Cannon-Bowers et al., 1993). The first refers to “overlap” of the individual cognitions: team members have a part of their knowledge base that is communal. In the second, the authors interpret group cognition as ‘similar’ or ‘identical’ between the team members: they hold identical knowledge. In the third p. 48

group cognition is considered as ‘complementary’: knowledge or team-members does not need to be identi- cal, but leads to the same expectations for the task at hand. In the fourth, group cognition is conceptualized in terms of ‘distributed’: the knowledge necessary for the task is dispersed among the different members of the group. p. 49

all studies from cognitive perspectives conceptualize group cognition as a similarity or overlap between individual mental structures. They are interested in the communality of the cognitions between team members. Cannon-Bowers and Salas (2001a) refer to this group of studies as literature on shared cognition and define them as follows: “They all encompass the notion that team members hold some knowledge that is similar or overlapping” (p. 87). Group cognition is conceptualized as a group-level construct, but therefore they rely heavily on the individual as entity. To share becomes nothing more than a cross-section of those individual entities. And ‘what’ they share reflects organized knowledge; individuals store concepts and relations between these concepts. Group cognition is defined in terms of similarity in these concepts and relations. What underlies this conceptualization is that cognition is housed in the mind, as an individual possession. This is also reflected in the use of concepts that are derived from cognitive psychology conceptualizing individual cognition. p. 49

The socio-cultural perspective rather perceives group cognition as something constituted by the group as an entity in itself. As such, cognition resides in the active mind, as a phenomenon situated in the group interaction. The concept group cognition is then defined as a process of coordination of participants’ contributions in joint activity (Matusov, 1996). p. 50

Group cognition is as such always process-like and situated within group work. They imply that it is not possible to “measure” it as a given state, or as an end state. Although group cognition in socio-cultural perspectives is not measurable separate from a certain process of activity, studies do differ in referring to group cognition concepts in three sequential moments (Matusov, 1996). According to Matusov (1996), group cognition can be perceived as a common ground (a shared background) between the group members that is activated in the group collaboration, as common engagement (shared activity), or as updated common ground (a shared experience) that is built during the collaboration. p. 50

Accordingly, in the three boundary crossing studies the authors interpret the concept of shared in terms of “distributed”: the knowledge necessary for the task is dispersed among the different members of the group (see Table 2). Derry et al. (1998) noted that a group distributed working memory would represent those thoughts from distributed long-term memory that becomes active within any individual’s attention during the group discussion. Connecting this to the notion of group cognition, they write: “Obviously, only ideas that are shared (discussed) by one individual (and attended to by others) have the potential to modify a group’s collective long-term memory” (p. 30). Similar to this, Banks and Millward (2000) define the central group cognition concept, shared mental model, as “a distributed system which runs a model collectively through the propagation of representational states across representational media”. By that definition they point on the one hand to the stable form of mental models. p. 51

The third boundary crossing study (Yoo & Kanawattanachai, 2001) uses the concepts of transactive memory and collective mind to combine cognitive and socio-cultural perspectives on group cognition. With transactive memory, following Wegner, Erber, and Raymond (1991), they imply a “shared system for encoding, storing, and retrieving information”. With collective mind they, following Weick and Roberts (1993), refer to a social cognitive system in which individuals heedfully interrelate their actions. Whereas transactive memory reflects the cognitive perspectives on group cognition, the concept collective mind reflects the socio-cultural perspectives on group cognition. The former refers to cognition within the minds of the group members, the latter points to cognition within group interaction. They draw upon these two concepts to examine how teams coordinate and interrelate their knowledge and actions in order to perform their tasks. In short, all three studies integrated elements of the cognitive perspectives (concepts like information processing, long-term memory, mental model form, transactive memory) and elements of the socio-cultural perspective (concepts such as situated cognition, distributed working memory, mental model state, collective mind) into an integrated approach centralized around the term distributed cognition. p. 51

Along with these different conceptualizations, the studies connected to both perspectives use different key terms and even different discourses. The studies in the cognitive perspectives use academic terms such as information, knowledge, mental models, knowledge structures, building models, while the studies in the socio-cultural perspectives use academic terms such as participation, activity, interaction, processes and coordination. Furthermore, in the cognitive perspectives input-process-output models are used in studying processes, whereas in the studies with socio-cultural perspectives group cognition is considered to be a process-like and situated phenomenon, which continuously needs to be re-established. p. 52

As Greeno (1998) described, cognitive perspectives assume the decomposability of complex systems into a set of subsystems. Because of this assumption, the cognitive strategy is able to learn about the properties of each of these subsystems (an individuals’ subjectivity) separately from the other subsystems, in order to build an understanding of the whole complex system. In relation to group cognition this means that one focuses on individuals’ subjectivity (in terms of ones knowledge, mental model, or ones perspective), and then looks at the similarity between the individual’s subjectivities. Socio-cultural perspectives assume systems of activity as “intact multiperson, human-technology systems” (Greeno, 1998). Hence, the socio-cultural tradition is able to learn about the intrinsic personality that complex systems (like groups or teams) endow, independent from those of its subsystems (individual members). p. 53

The drawback of each of these perspectives on group cognition seems to be exactly what is essentially highlighted by the other perspective. The cognitive tradition looses sight of the intrinsic nature of the complex system as a whole, and can only learn about it through the aggregation of the properties of subsystems. The shortcoming of this is that, in the end, groups are understood as some sort of sum of its members. In contrast, the socio-cultural tradition looses sight of the intrinsic nature of the subsystems, and learns about them only through the perspective of the system of which they are part. The drawback here is that subsystems (e.g., individuals, actions) are understood to be a function of their relations with other subsystems. Individuals are always participants of multiple social contexts. Anderson, Reder, and Simon (1997) noted that in such understandings, individuals tend to get ‘absorbed’ in the collectives of which they are part. p. 53

But, what is questionable in terms of socio-cultural thinking is treating cognition itself as decomposable. Distributed cognition is, at least in the three studies discussed here, treated as dividing up the cognitive processes taking place within the heads of individuals (individual cognition in terms of mental models, memory) and the processes taking place in the interaction (situated cognition). The situated cognition in fact only refers to the information that is conversationally shared between these individuals. And with the term distributed cognition the studies signify that not all information relevant for a specific situation needs to be conversationally shared between all individuals, in order to build up the informational structure that is needed to coordinate the collaborative work well. Although choosing the term distributed cognition instead of situated cognition, the emphasis in these studies remains on exactly these informational structures, whereby information is divided amongst the individuals. To understand the collaboration processes it becomes most relevant to p. 53

Combining the theories, or taking each theory and working towards overcoming the "shortcomings" p. 55