The Radically Embodied Conscious Cybernetic Bayesian Brain: Towards solving the Hard Problem, grounding intentionality, and rescuing free will by resurrecting homunculi and Cartesian theaters Adam Safron (Northwestern University, Chicago, IL ) C13
There is near universal agreement that cognitive science needs to divorce itself from the last vestiges of Cartesian thinking (Damasio, 1995; D. Dennett, 1992; Dolega & Dewhurst, 2015; Hobson & Friston, 2016). The only place where this agreement is lacking seems to be with respect to which aspects of Cartesian thinking are most egregiously mistaken. The charges are as follows: 1) Separating the body and mind as separate orders of being (i.e., the mind-body problem); 2) Describing perception in terms of the re-presentation of sensations to an inner experiencer (i.e., the theater fallacy); 3) Not realizing the inadequacy of inner experiencers as explanations, since these would require further experiencers to explain their experiences, resulting in an infinite regress (i.e., the homunculus fallacy). Many argue that the goal of cognitive science should be to explain away these mind-like processes in terms of non-mental computational and mechanistic processes (Deacon, 2011; D. C. Dennett, 2017). Enactivists further (and differently) argue that cognitive science will only be thoroughly cleansed of its Cartesian origins once we eliminate concepts such as representation from our explanatory frameworks (Varela, Thompson, & Rosch, 1992). Yet the overwhelming consensus is clear: the mind sciences must rid themselves of Descartes' errors. In this presentation, I will suggest that this overwhelming consensus is mistaken. Indeed, I will argue that each and every one of Descartes' errors actually point to invaluable perspectives. In brief: 1) Minds are thoroughly embodied, embedded, enacted, and extended (Clark & Chalmers, 1998; Rowlands, 2010; Rudrauf, Lutz, Cosmelli, Lachaux, & Le Van Quyen, 2003), but there are functionally important aspects of mind (e.g. integrative processes supporting consciousness) that do not extend into bodies, nor even throughout the entire brain; 2) The brain not only infers mental spaces, but it populates these spaces with representations that form the basis for simulations at various degrees of detail and abstraction. 3) Not only are experiences re-presented to inner experiencers, but these experiencers take the form of embodied person-models with degrees of agency, and even more, these quasi-homunculi form a necessary scaffolding for nearly all aspects of mind. I intend to justify these claims, and in the process, show how attention, imagination, and goal-oriented behavior may be explained using predictive coding frameworks and computational models of consciousness. Moreover, I hope to begin to integrate existing theories of consciousness to an extent that the "Hard problem" may begin to appear more tractable (Chalmers, 1995). In brief, this presentation is an attempt at unification in cognitive science, endeavoring to show how a radically embodied cybernetic Bayesian brain may create foundations for intelligence, consciousness, and will.