There has been much recent effort within the field(s) of enaction to both reintroduce and re-appraise the notion of habit as it relates to the discourse surrounding cognition, robotics etc. Speaking recently with somebody on this issue, that person asked me if I could answer the following question, why habit? In other words, why the reintroduction of this age old concept as part of the enactive project, and why now? Unable to come up with even a good deflection, let alone anything like a reasonable answer, I have taken it upon myself - having being instructed to do so - to do precisely that. I hope that what I offer in the following is something more like an plausible answer (though a brief one) and less like a piteous deflection.
The notion of a habit can be traced back to Aristotle. First appearing in book II of his Nicomachen Ethics, habits or ‘hexes’ for Aristotle, can be said to be something like character dispositions brought about through training, that in turn engender temporally extended regularities in the actions of an individual.
Indeed, since Aristotle many philosophers and thinkers have had something to say on the issue, including heavyweights like Aquinas, Kant, and Heidegger. In fact it would likely be easier to name those, at least prior to the 1950’s, who have not had something to say on the issue.
For many, the notion of habit was indeed one most central. Take one of the founders of modern psychology William James on the issue,“when we look at living creatures from an outward point of view, one of the first things that strikes us is that they are bundles of habits”, or fellow american pragmatist John Dewey, “concrete habits do all the perceiving, recognising, imagining recalling judging, imagining, conceiving, and reasoning … consciousness expresses functions of habits, phenomenaa of their formation, operation, their interruption and reorganisation”.
However, in the 20th century the notion of habit sustained what might be seen as a couple of significant setbacks. By way of behaviourism a habit was rendered little more than a “probability correlation between stimulus response”. As Barandiaran suggests “under the influence of logical-positivism (Watson, Skinner etc.) the notion of habit is purified of any internal/physiological component”. However, it was with the advent of cognitivism that habits fell off the map altogether, what some authors have called the ‘cognitivist blackout’. Somewhere around the 1960’s (this can be seen graphed here (p.128) this does not, however, assume any causal relation between the two) use of the term habit was usurped with the idea of representation, from which point it had a steady decline until about the late 1990’s, and is - as mentioned in the introduction - only making somewhat of a comeback in recent years, most notably in the work of the enactivists (Di Paolo, Egbert, Noe, Barandarian etc.).
But what does this re-appraisal entail, and more importantly for our purposes herein, why bother? A habit, from the view of enactionists, might be most easily described as a ‘self-sustaining pattern of behaviour’, wherein the ’stability of a particular behaviour is coupled with the stability of the mechanisms generating it’. This in turn is built into a bundle of habits that is thought to sustain itself as a whole through various interactions in an ongoing process know as equilibration.
Equilibration is a term borrowed from the work of developmental psychologist Piaget that subsumes two processes: assimilation and accommodation. These process basically allow for the integration and modification of environmental variations into the ever increasingly complex construction of an ‘organised network of habits’.
Under this conception the cognitive agent is seen simply as a ‘systematic and adaptive organisation of habits’. Habits here are conceived of analogously to an idea central to the enactive project, that of the autonomous agent. Without going into the details, an agent is seen as autonomous in that it maintains its own structure through its functional organisation and its interactions with the environment in which it exists.
In enaction this notion is applied broadly and is said to characterise many forms of life, from the cell to the society. Thusly, the habit can be characterised as having the same essential variables as the metabolic or biological organism that is performing the behaviour.
Consequently, ‘a habit’ becomes the behavioural equivalent - theoretically - of an individual autonomous agent.
So, why habits? Unlike a biological cell, or a human being, a habit can be modelled; and so, the habit becomes a unit of theoretical and explanatory currency for enactivists, one that is amenable to robotic modelling in a way that the activity of the cell, or the society, is not. Moreover, it brings the relationship between a robot and its environment much closer to the the conception of a living system, in which the world that exists for the robot is not divorced from its internal organisation. For Di Poalo, broadly speaky, this affords for the robot, not a life as such, but a way of life (a meaningful one at that) that will be sought to be maintained.