Getting our labels straight.

19/10/2009 Marek Introducing enaction, 0

The word “enaction” has seen a much greater success and more widespread appearance in the literature than the enactive approach itself. This serves as a source of potential confusion for people getting acquainted with these ideas in the first place, as they read one renowned author discuss their enactive account of perception (Noë, 2005), before reading someone else claim that that first author’s theory isn’t really enactive (Di Paolo, Rohde & De Jaegher, in press; McGann, in press), and then coming across a third theory which is termed enactive but seems to have little interaction with the first two (Newton, 1996). To make things even more interesting, there are a few theorists out there who seem to think in almost exactly the same way, but don’t like the label “enactive” and never use it, preferring to avoid niche descriptions (Hurley, 1998; Hurley & Noë, 2003). And finally, to top it all off, there are researchers in closely related disciplines using the term in a way that has very little relation indeed to any of the above.

Getting over this confusion will take time, as some kind of consensus is reached over years, and the various uses of the different terms converge or distinguish themselves more clearly. Any attempt to enforce a discipline in their use is a doomed venture, but here’s the way I think about them:

• Enaction: I reserve the use of the terms “enaction” and “enactive” for those theories that take as their starting the point the autonomy of the individual agent. There is a whole slew of associated concepts (I’ve taken to calling the group of them the “autonomy cascade”), but the key point is that cognition arises from the operation of an agent trying to maintain itself in the midst of a host of conflicting tensions. This version of enaction (what I suppose I think of as the “real” enaction) has a genuine theory of meaning associated with it – the idea of a system “sense-making”, adapting its own activity to coordinate the environment and itself as part of the process of maintaining itself. This underlying theory of meaning is what really sets the enactive apart from other related approaches.
• Dynamic sensorimotor: Theories which emphasise the skilled, dynamic interaction between an embodied agent and its environment are often termed “enactive”, but it is possible to hold this point of view while still assuming a computational or representational system at the core of cognition – thus, a dynamic sensorimotor view shares many claims of the enactive approach, but resists any commitment to particular theory of meaning. The enactive approach includes a dynamic sensorimotor view, but because it has the extra demand of starting with autonomous systems (rather than, for instance, attaching sensorimotor dynamics to the periphery of a computational system), it is distinct as an approach to explaining cognition.
• Embodied: The term embodiment has really taken a central place amongst cognitive science concepts in the past twenty years or so. It means many things to many people, and a few have tried to sort out just what the differences between different kinds of embodiment mean. The basic claim is that the physical characteristics of the cognitive system affect cognitive processes. Aside from that one principle though – that the body matters – embodied cognitive science has little to say by way of theories of meaning, consistent explanations of cognition or the just what the implications of the body for cognition actually are. As such, there are a great many embodied theorists who have very little in common with one another, and might strongly disagree with one another, when it comes to the question of just what the relationship between the body and cognition is.

Below is a table of these different streams within the “embodied” or “active” turn in modern Cognitive Science, and some theorists associated with each. If I’ve made any clear errors, or if you have other suggestions, let me know.

Enaction theorists	Dynamic sensorimotor theorists	Embodiment theorists
Francisco Varela	Alva Noë	Andy Clark
Evan Thompson	Susan Hurley	Natika Newton
Ezequiel Di Paolo	Kevin O’Regan	Larry Barsalou
		Antonio Damasio

Something a little less strictured would probably be better here. Outside of the enactive approach, where the question of autonomy is fairly well entrenched, instead of definitive streams of research we get something more like trends or tendencies in people’s thinking. Some will have more of a dynamic bent, some less, some seeing embodiment as central to the organisation of cognition, some more see it as a constraint on otherwise abstract computational systems.

–

Relativity versus selectivity in the enactive approach Text creole?