Reconceptualizing Activity Theory in distributed settingsFebruary 27, 2007
The learning environments cease to be primarly tutor-defined and well controlled spaces, thus, the user-selected variability in the learning environments has to be considered as part of the learning-design. For instasnce, Kirschner et al., (2004) suggest a new affordance based and learner-centered interaction design model. New competencies would become important both for the students and the educators, such as the ability of finding and choosing the appropriate learning tools and systems among social software, prognosing the educational affordances of these systems, composing the distributed learning space from the set of tools, collaborating in distributed spaces for sharing knowledge and regulating one’s learning with the remote others who might perceive the different concept of the how this learning environment might favour their activities.
The Activity Theory model has recently been of continuous interest in explaining the processes taking place in the learning communites. According to Leontjev (1975/78) and Kuuti (1995), any shared activity can been defined through the shared motives that form the object of the activity for the community, the activity is performed by the subjects of the community; the activity consists of goal-directed actions conducted by the community members, the goals of the actions are realised by performing certain operations using the tools (material- or knowledge-tools, language) of the community as the mediating devices. For Kuuti (1995) the learning environments can serve at the same time as the mediating tools enabling the manipulations with the objects, aiding knowledge construction and communication, but also as the object of the activity that turns implicit community knowledge externally observable. The latter is especially characteristic to new Web 2.0 software (e.g. social bookmarking, blogosphere etc.). The object of the activity can be realised as the outcome of the activity, knowledge artifact, community practice etc.
For defining the elements of collaborative activities in social-software supported learning environments at least two levels of the activity should be considered: the Activity System level (Engeström, 1987), and the operation level (Kuuti, 1995). The Activity System model (Engeström, 1987) describes general information flows at the community level within or between the communities. According to Engeström (1987), the object of the community brings forth the division of labour and the alignment to certain rules that constrain the availability and the use of possible operations and tools. In any Activity System the contradictions emerge between the subjects about the object, the division of labour, the alignment to the community rules, and the application of tools. The Activity System model suggests ‚Object-activity‘, ‚Rule-producing activity‘, ‚Subject-producing activity‘, ‚Instrument-producing activity‘, and ‚Central activity‘ to take place in the system.
Operating in the Activity System evokes different patterns of coordination emerging in groups of individuals engaged in joint action with material and informational systems in their environment (Greeno, 2006). Engeström, Engeström and Vähäaho (1999) distinguish the activity systems operating in teams or networks, and the activity systems emerging in the work communities. Teams and networks are typically understood as relatively stable structures, often assigned by instructors, sharing the same objectives and mediation tools. The subjects in work communities however are involved in workflows that consist of combinations of people, tasks and tools that get connected at relatively short duration. Yet in their basic pattern, these workflows are continuously repeated and evolving. Engeström et al. (1999) describe these temporal trajectories of successive task-oriented combinations of people and artifacts emerging within or between activity systems as knotworking situations. The notion of knot refers to a rapidly pulsating, distributed and partially improvised orchestration of collaborative performance between otherwise loosely connected actors and activity systems. Knotworking situations rely on fast accomplishment of intersubjective understanding, distributed control and coordinated action between actors. Since now the e-learning in Web 1.0 has been exploiting the activity patterns that form around shared objectives and joint medium in which learning was centrally controlled by facilitators. However, the e-learning in Web 2.0 would rather be knotworking between loosely connected learners using distributed tools. Engeström et al. (1999) assume that in knotworking the tying and dissolution of a knot of collaborative work is not reducible to any specific individual or fixed organisational entity as the center of control, the locus of initiative changes from moment to moment – the center does not hold. They suggest bringing the analysis of activity systems at knot level. In knots the subjects get connected via mediation tools and perform actions upon certain shared objects using or creating artefacts.
How do the subjects perceive themselves and the others in object-directed interaction in knots, could be described by the notion of „affordances“. Gibson (1979) defined affordances as the opportunities for action for the observer provided by an environment. However, as assumed by Gaver (1996) affordances are primarily facts about action and interaction, not perception. This contrasts with the common impression that affordances refer to—approximately—situations in which one can see what to do (Gibson, 1979). The mainstream view to the affordances in educational technology settings considers them the objective properties of the tools perceptable in the frames of some activities, suggesting that tools have concrete technological affordances for certain performances that can be brought into learner’s perception with specific instructions (Norman, 1988; Gaver, 1996). Neisser (1994) elaborated Gibson‘s affordances distinguishing three preceptual modules: i) Direct perception/action, which enables us to perceive and act effectively on the local environment, ii) Interpresonal perception/reactivity, which underlies our immediate social interactions with other human beings, and iii) Representation/recognition, by which we identify and respond appropriately to familiar objects and situations. Kreijns, Kirschner, and Jochems (2002) have defined social affordances as the “properties of a collaborative learning environment that act as social-contextual facilitators relevant for the learner’s social interaction. Neisser’s interpretation, however enables to consider also the interpersonal perception between subjects in action as the source of affordances at social and regulative domains. Kirschner (2002) defined educational affordances as those characteristics of an artifact that determine if and how a particular learning behavior could possibly be enacted within a given context. The contextual aspect of affordances relates with the artifacts and meanings. Thus, instead of relating affordances objectively with the software they should be related with the knots of the Activity System where subjects must realise how they perform joint actions with artifacts and tools in order to accomplish their shared object. Cook and Brown (1999) assumed that affordances are dynamic – the ongoing interactions with the environment and objects, where our previous knowledge applied during the activity helps us to evoke noticing of certain aspects, affordances, and knowing how these affordances could support the activity. This assumption supports the Engeström et al. (1999) view of the dynamic nature of knots in the Activity system.
Greeno (2006) has suggested that two components: an interaction analysis of the conversation including close attention to its turn taking, responses and contributions, and the semiotic structures of information that they constructed in the conversation should be taken into account when analysing Activity Systems. He suggests to apply both representations and representational practices in the same analysis of activities, assuming that when researchers shift the analysis of knowledge construction to the level of the Activity System, they include explanations about the various participants in the activity, and they analyse ways that individuals are positioned in the participant structures of interaction, and how that positioning contributes to generation of information structures.
The contradictions in activity systems arise and must be worked out at operation level in order to realise the object of the community (Kuuti, 1995). Thus the action and operation level elements of activities must be defined in order to explain the functioning at the activity system level. To work on shared learning objectives, activities in the communities must be dialogue-mediated for creating intersubjective knowledge, and consist of actions where learners jointly manipulate tools and artifacts or create shared knowledge artifacts. Laurillard (1993) has developed the conversational model between the teacher and the student where she distinguishes dialogue about the content, the manipulable actions, and the dialogue about the manipulable actions.
Laurillard framework was modified in order to model actions in the distributed activity-space for the collaborative object-oriented virtual activity involving two subjects (e.g. facilitator and the student, or n number of students).
The actions and operations in distributed activity space are conducted via the mediation of different tools. Typical actions what might take place while interacting with the tools, content and other groupmates are communicative (generating content; regulating content-relevant actions, regulating activity processes, regulating social atmosphere) and manipulative (operating with artifacts and tools). The operations conducted at each action, combined with the subjects and artifacts involved in that action of the activity enable to suggest which are the affordances of that knot at the certain time.
Engeström Activity system is defined for modelling face-to-face work settings, where people communicate directly, and do not need mediating devices for communicatively discussed activities related with the common objects, rules, tools and subjects in the work community. For modelling the Activity System for distributed virtual settings it is reasonable to modify Engeström (1999) model. The Distributed Activity System model distinguishes the direct manipulative actions performed with tools and material artifacts from the indirect communicative acts in, which people develop their common objects verbally and publish them at the same time via the mediating devices by involving some material artifacts into this process if necessary. The separation of the direct actions with manipulable objects and indirect mediated communicative actions in the Distributed Activity System model would reveal the mediated regulative dimension, in which the community must come to the common ground in their rules, distribution of labour, understanding what is planned to do to realise the shared object in the form of knowledge artifact, and receive feedback about the object-relevancy of their actions. The actions described in the Laurillard (1993) framework modified for distributed systems take place in each knot of the activity and enable to predict the affordances of the knot. The dynamically changing affordances at the Activity System level could be used for the analysis of the development and effectiveness of functioning of the system.
The question is how do the affordances of the Activity system develop if „the center does not hold“, and the learners must realise common goals in Web 2.0.
Secondly, it is of interest, how do the affordances of the Activity system, predicted by the instructor, differ of the actually emerging affordances in different learning groups, and how should the instructional designs tackle with this problem.