Study notes: The Grain of the Hard Problem

The hard problem asks why physical processing is accompanied by experience. The meta-problem asks why we judge that there is a hard problem. The third asks whether the hard problem is hard or only seems so. If our judgment of hardness is fully explained by how our concepts work, with nothing in the world answering to the apparent omission, the problem is dissolved: shown to have no subject matter, not solved.

Suppose character is exhausted by discrimination, as on Yu and Carruthers’s view that a quality is a position in a similarity space constituted by discriminative capacities. Discriminative capacity is functional, so a functional duplicate has the same character and a zombie lacks nothing. The zombie’s conceivability then comes entirely from concept isolation, and the judgment that a physical story omits something is about nothing. So either experience outruns discrimination, or the hard problem has no subject matter. This is a dilemma and not a refutation: a functionalist can take the second horn, and the argument shows the price.

It shows that character outruns pairwise discrimination. Goodman’s criterion answers it: two qualities are identical only if they match exactly the same other qualities, and neighboring shades differ in which further shades they match. A quality space built from the whole matching profile registers those differences. The hardness argument needs character that outruns the whole profile.

Types and tokens. A matching profile belongs to a repeatable quality type, compared across occasions. Experience is a sequence of occurrences whose exact character depends on context. Counterfactual comparisons each supply a new context, and their outcome can depend on order, so there may be no fact about what an occurrence would match.

Realization. On the identity proposal, character is the organization of the vehicle at the grain at which its regulation corrects it. Two organizations can share every discriminative disposition and differ there, as when a recurrent network is unfolded into a feedforward one. Nothing in the profile settles whether experience is preserved.

The objection equivocates on which system discriminates. The regulator must register fine differences; the subject need not. Access is a narrow, many-to-one channel, and to keep what passes it accurate, regulation must use information the channel discards. “Discrimination” means what the subject can tell apart for its own purposes. The second qualification: the excess must be present in experience, not hidden in the implementation as Drescher’s gensym features are. A duplicate that lacked only implementation detail would lack nothing experiential.

Our conceptual frame cannot express grain dependency. A phenomenal concept is recognitional and carries at most what the subject can discriminate, while its referent lives at a finer grain, so the concept cannot encode the grain of what it refers to. Physical concepts specify grains and carry no phenomenal mode of presentation. No concept says “experience is the organization at grain G”, so subtracting experience from a physical description goes through in thought. The usual strategy takes the isolation of phenomenal concepts as basic; here it follows from the overflow. The intuition is partly right: a functional description at the grain of cognition does underdetermine experience.

It slides from a difference in experience to its absence: a description that fails to fix character need not fail to fix that there is experience. And it extends the verdict from functional descriptions to physics, because our frame presents both as third-person structure. The essay accepts that structure and dynamics entail only structure and dynamics. It holds that experience is structure and dynamics at its own grain. What structural description cannot deliver is the recognitional mode of presentation, a limit of concepts and not a gap in the world.

Cognition (the recipients of access), experience (the organization of the vehicles), physics (the grain at which observers like us formulate law). Hardness needs only that experience draws some distinctions cognition does not; the two may cross-cut. The solution needs physics to refine the grain of experience, so that a physical duplicate fixes it.

Two systems alike at the grain of their physics could differ in experience, though only below that grain, within a causal structure fixed above it. Their physics would be incomplete with respect to experience. Their physicalism, the claim that experience is part of the structure of reality with no extra ingredients, would still be true. Their hard problem would be unsolvable by physics without being a problem for physicalism. For us the identity of experience with vehicle organization is necessary and known a posteriori, and the a posteriori part is the discovered ordering of grains. The necessity is therefore not brute.

Shared: physics conveys only structure. The Russellian adds that structure must be the structure of something, and takes experience as the intrinsic nature beneath physical structure. The essay rejects the added premise, with ontic structuralism. It also explains the premise’s appeal: the one alleged instance of an intrinsic nature is experience, and experience seems unstructured because presence is finer than recognition. For the Russellian experience lies beneath physics. For the essay it is structure at a grain coarser than that of physics.

The photograph reproduces the cause and not the experience. The route bypasses her understanding: garbling her physics would change nothing. It hands the disputed step to her sensory apparatus. A route that does not cheat: the reasoner reads the description, derives the organization at the grain of experience, configures its own vehicles to match and brings them into its regulation, then undergoes and recognizes the result. It needs a path from its reasoning into its experiential carrier at that grain, a substrate general enough to realize the target, and regulation that takes the constructed vehicle in. The gap is a gap in analytic derivation, and Mary’s predicament is a fact about human architecture.

• The opening argument is a dilemma: a functionalist can accept that the hard problem has no subject matter.
• Whether a functional duplicate could lack experience altogether is “at best open”; whether unfolding preserves experience is contested.
• Chalmers can reply that knowledge by construction is acquaintance and so not a priori in his sense. The essay does not dispute the word.
• Whether the ordering of grains is a coincidence is open. A theory of observers might show why a system that experiences must extract a physics finer than its own vehicles.

Michał’s cases (two computers with swapped bits, pain that leaves no memory, eyes that render colors differently) decoupled phenomenality from cognition, and Michał drew the conclusion: “it’s easier to begin to suspect that it’s a physical process with fine substrate-dependence.” Thomas developed it: “Cognition buffers differences that phenomenality preserves.” That is the lower end of the ordering. Refinements: the essay’s overflow is not a difference with zero consequence, since it makes a difference to regulation within the occurrence and fails only to reach cognition. And the claim is not substrate-dependence: experience is a causal pattern any substrate could realize, at a grain finer than cognitive function.

He uses “cognition” more broadly (“All phenomenal actuality is computed by a kind of cognition”), where Part I means only the recipients of access. And Thomas rejects observer-relative coarse-graining, on which the claim that physics could have been coarser relies; the essay reads Wolfram’s observers as a case of interaction-relativity. The side taken: the essay fixes the grain of experience by operational individuation alone, while Michał holds that for phenomenal consciousness “the ontological aspect is paramount”.