Towards an Affordance-Based Theory of Collaborative Action (CoAct)

Introduction

In complex environments, where not all variables and relationships are known, i.e., where data are uncertain and incomplete, humans create rather than discover their future (Nosek, 2005). They create the future by perceiving affordances within their environment and acting. The subsequent actions, including probing of the environment, lead to changes in the environment that must provide meaningful affordances for other actors, both human and non-human. These behaviors can be immediately perceived by other actors or they can modify the environment, such as the creation of a report, to provide affordances at some later time. Human and non-human actors must be attuned to relevant affordances, to act based on them, and to probe for additional relevant affordances. The more important the action, the more dynamic, equivocal the task, the more unreliable the data, the more important group sensemaking (Gephart, 1993; Weick, 1979) to the emergence of socially-constructed capacities to act, sufficiently coordinated to engender effective action (Nosek, 2005).

The Theory of Collaboration Action (CoAct) has grown out of the desire to provide more theory-based direction in information system development, especially in development of collaboration technology to enable creative solutions to wicked problems (Farooq, Carroll, & Ganoe, 2008). Relying on the assumption of the existence of idiosyncratic, intermediary internal representations, such as mental models, to filter sense-data limits practical, theory-based guidance. Extending this assumption so that collaboration technology must support the creation and maintenance of shared mental models within teams exacerbates the problem. “This all-absorbing concern for the internal, mental model unfortunately led to a neglect of other aspects, of which the most important was the flexibility and variability of human performance (Hollnagel & Woods, 2005, p. 41).” Hollnagel and Woods question derivative folk models, which are based on consensus and “privileged knowledge’ about how the mind works (Morick in Hollnagel & Woods, 2005, p. 51).” Examples of folk models include fatigue; workload (Stassen, Johannsen & Moray, 1990) and situation awareness (Endsley, 1995, 2001). Many folk models purport to measure intervening variables representing intermediate mental states rather than performance. Folk models may not be incorrect but are hard to disprove, i.e., they are not falsifiable. Others (Bloor, 1983; Brand, 1979; Heft, 2001; Wilson, 1998) have argued that there are no intermediary internal representations, i.e., no “proverbial little men in the mind,” such as mental models (Cannon-Bowers, Sala, & Converse, 1993; Gentner & Stevens, 1983; Johnson-Laird, 1983), beliefs, cognitive constructs (Adams-Webber, 1979; Fransella & Bannister, 1977), or scripts that are invoked to take-in sensory information and process it. They argue that there is no mind/body dichotomy (Heft, 2001). This view is supported by recent findings in brain research (Yufik & Georgopoulos, 2002).

The purpose of this paper is not to disparage folk models, but to introduce a theory of collaborative action that does not rely on sharing internal, mental models and extend Gibson’s ecological theory of affordance to provide guidance in understanding collaborative acts and developing technology that supports them. The paper provides the following: background for the development of an affordance-based theory of collaboration; explication of Individual Action (IAct) and Collaborative Action (CoAct) Models; propositions and research questions based on these models; results of an initial experiment to test some of these propositions and research questions; followed by examples of how CoAct can be used in design of information systems.