Madary notes, more or less in passing, that the PP vision of “organism-relative perceptual content” bears a close resemblance to views that have been defended under the broad banner of “enactivism”. I want to pick up on this hint, and suggest that the PP account actually sets the scene for peace to be declared between the once-warring camps of representationalism and enactivism. Thus consider the mysterious-sounding notion of “enacting a world”, as that notion appears in Varela et al. (1991)[2]. Varela et al. write that:
The overall concern of an enactive approach to perception is not to determine how some perceiver-independent world is to be recovered; it is, rather, to determine the common principles or lawful linkages between sensory and motor systems that explain how action can be perceptually-guided in a perceiver-dependent world. (1991, p. 173)
This kind of relation is described by Varela et al. as one of “structural coupling” in which “the species brings forth and specifies its own domain of problems” (1991, p. 198) and in that sense “enacts” or brings forth (1991, p. 205) its own world. In discussing these matters, Varela et al. are also concerned to stress that the relevant histories of structural coupling may select what they describe as “non-optimal” features, traits, and behaviors: ones that involve “satisficing” (see Simon 1956) where that means settling for whatever “good enough” solution or structure “has sufficient integrity to persist” (Varela et al. 1991, p. 196). PP, I will now suggest, has the resources to cash these enactivist cheques, depicting the organism and the organism-salient world as bound together in a process of mutual specification in which the simplest approximations apt to support a history of viable interaction are the ones that are learnt, selected, and maintained.
The simplest way in which a PP-style organism might be said to actively construct its world is by sampling. Action, as Madary noted, serves perception by moving the body and sense-organs around in ways that aim to “serve up” predicted sequences of high-reliability, task-relevant information. In this way, different organisms and individuals may selectively sample in ways that both actively construct and continuously confirm the existence of different “worlds”. It is in this sense that, as Friston, Adams, and Montague (Friston et al. 2012, p. 22) comment, our implicit and explicit models might be said to “create their own data”.[3] Furthermore, the PP framework depicts perception and action as a single (neurally distributed) process whose goal is the reduction of salient prediction-error. To be sure, “sensory” and “motor” systems specialize in different predictions. But the old image of sensory information IN and motor output OUT is here abandoned. Instead, there is a unified sensorimotor system aiming to predict the full range of sensory inputs – inputs that are often at least partially self-selected and that include exteroceptive, proprioceptive (action-determining), and interoceptive elements. Nor is it just the sensorimotor system that is here in play. Instead, the whole embodied organism (as Madary notes) is treated as a prediction-error minimizing device.
The task of the generative model in all these settings is (as noted in Clark this collection) to capture the simplest approximations that will support the actions required to do the job. And that means taking into account whatever work can be done by a creature’s morphology, physical actions, and socio-technological surroundings. Such approximations are constrained to “provide the simplest (most parsimonious) explanations for sampled outcomes” (Friston et al. 2012, p. 22). This respects the enactivist’s stress on biological frugality, satisficing, and the ubiquity of simple but adequate solutions that make the most of brain, body, and world. At this point, all the positive enactivist cheques mentioned above have been cashed.
But one outstanding debt remains. To broker real and lasting peace, we must tiptoe bravely back into some muddy and contested territory: the smoking battleground of the Representation wars.