Jake Spinella: Dissertation Defense August 13, 2026
Jake successfully defended his dissertation on August 13, 2026
Jake successfully defended his dissertation "The Belief-Independence of Perception: An Essay on the Mind's Architecture", on August 13, 2026.
Abstract:
This dissertation argues that perception is belief-independent, that this property distinguishes perception from cognition within the mind's architecture, and that it helps explain how perception manages to be both computationally tractable and semantically anchoring.
Chapter 1 makes the notion of belief-independence precise. Beginning from familiar cases of perceptual recalcitrance in the face of what we know, I argue that a psychological process is belief-independent when cognitive states like beliefs, desires, and thoughts are dispensable to its computational explanation. Cognitive states are dispensable to the computational explanation of a psychological process when the contents of cognitive states do not determine the contents of the psychological process' states in a semantically traceable way. On this view, belief-independence amounts to inferential isolation from cognition, and perception, if belief-independent, is free from certain sorts of cognitive influence. In virtue of this, perception marks a natural joint in human mental architecture and is architecturally distinct from cognition, rather than a form of it.
Chapter 2 asks how a system could be implemented in such a way that guarantees belief-independence. I claim that one way to implement belief-independence is by making psychological systems informationally encapsulated. I then address cognitive penetration cases that prima facie threaten the idea that perception is inferentially isolated and therefore belief-independent. Against the cognitive penetration literature's leading cases—memory-color effects, high-level property perception, and feature-based attention—I argue, first, that equally good or better wholly perceptual explanations of the evidence are available, and second, that even granting the cognitive interpretation of the evidence, the effects exhibit no semantic tracing from perceptual to cognitive content, and so leave perception's belief-independence and architectural distinctness intact.
Chapters 3 and 4 go on to explain how belief-independence helps perception carry out two of its fundamental roles in human thought. The first of these chapters argues that perception's belief-independence makes it more computationally tractable than it otherwise would be, which helps perception succeed in producing quick and accurate outputs in a rapidly changing environment. In particular, I argue that the informationally encapsulated character of perception secures tractability not only by increasing the speed of computation—something commonly argued in the philosophy of perception—but also improves perception's accuracy. Perception does veritistically better for being insulated from cognitive influence.
Chapter 4 then examines perception's role in grounding the very possibility of determinate thought. Following Russell and Strawson, I argue that any nonderivatively referential representational system must be semantically anchored—that is, it must have a way of generating relationally fixed representations, if it is to avoid the systematic reference failure exhibited in massive reduplication scenarios. I claim that perception is the primary and original semantic anchor in creatures like us, in virtue of supplying us with relational representations of the world outside of us. It does so by way of the FINST-and-object-file structure underlying certain forms of visual tracking; this form of tracking constitutes an acquaintance relation to objects, and acquaintance is fundamentally relational.
I subsequently argue that perception can be semantically anchoring only if the acquaintance-fixing component of it is belief-independent, since belief-infected reference-fixing would be vulnerable to referential failure from massive reduplication. Perception's belief-independence is thus necessary for its being semantically anchoring and consequently necessary for the very possibility of determinate--viz. nonderivatively referential--thought. What the empirical and a priori arguments of the dissertation ultimately show is that an architectural distinction between perception and cognition must be posited if we are to adequately explain our peculiarly human form of mindedness.