Predictive Processing and the Representation Wars

Abstract

Clark has recently suggested that predictive processing advances a theory of neural function with the resources to put an ecumenical end to the ''representation wars'' of recent cognitive science. In this paper I defend and develop this suggestion. First, I broaden the representation wars to include three foundational challenges to representational cognitive science. Second, I articulate three features of predictive processing's account of internal representation that distinguish it from more orthodox representationalist frameworks. Specifically, I argue that it posits a resemblance-based representational architecture with organism-relative contents that functions in the service of pragmatic success, not veridical representation. Finally, I argue that internal representation so understood is either impervious to the three anti-representationalist challenges I outline or can actively embrace them.

Pragmatism and the predictive mind

“…In this section, I argue that predictive processing is as well (cf. Anderson 2017;Barrett 2016;Bruineberg et al 2016;Clark 2016;Seth 2015;Williams 2017).…”

From symbols to icons: the return of resemblance in the cognitive neuroscience revolution

“…Other philosophers have recently conducted excellent work on just this topic (cf. Churchland 2012; Gładziejewski and Miłkowski 2017;Kiefer and Hohwy 2017;Ryder 2004;Shea 2014;Williams 2017). Any exhaustive overview of such research-or of research into cognitive representation more generally in cognitive neuroscience-is impossible in a paper of this scope.…”

“…Third, and most importantly for our purposes, at the centre of predictive processing is a commitment to a thoroughly iconic account of cognitive representation, whereby cortical dynamics come to recapitulate regularity structures in the body and environment and thereby realise a rich, hierarchically structured generative model with which to anticipate the incoming signal (Gładziejewski 2015;Kiefer and Hohwy 2017;Williams 2017). Specifically, it is only by inducing and updating a hierarchically structured dynamical model of the bodily and environmental causes of their sensory input that predictive brains can fulfil their regulative function (Seth and Friston 2016).…”

“…7 As noted above, predictive processing capitalizes on research in machine learning in which unsupervised learning is driven by the exploitation of a generative model of environmental causes (Kiefer and Hohwy 2017). The punchline of this research is that the only way in which such a system can successfully predict or anticipate the sensory input generated by the world is by effectively becoming that world-that is, by inverting and thereby recapitulating the process by which sensory input is generated in its structure and dynamics (Kiefer and Hohwy 2017;Williams 2017). The link between prediction and iconic representation in this context is thus not incidental.…”

“…In this paper we identify one important aspect of this revolution: a dramatic shift away from thinking of cognitive representations as arbitrary symbols towards thinking of them as icons that replicate structural characteristics of their targets. This shift has received increasing attention in the philosophical literature in recent years (e.g., Churchland 2012;Cummins 1989;Grush 2004;Gładziejewski and Miłkowski 2017;O'Brien and Opie 2015b;Ryder 2004;Williams 2017). We aim to clarify what it consists in, and explain why it has occurred.…”