Bridging the Explanatory Gap: A Physicalist's Challenge — Epoche B2
Where the Gap Closes and Where It Does Not: Functional Reduction and the Price of Physicalism Whose gap is it, and what does it say? The phrase "explanatory gap" is Joseph Levine's [1] , from a 1983 paper on materialism and qualia. His point was modest in form and hard to shake: an identity statement such as "pain is the firing of certain fibres" leaves something unexplained that "water is a certain compound" does not. Knowing the chemistry tells you why water behaves as water does. Knowing the neurophysiology does not seem to tell you why that state feels the way it feels. David Chalmers's contribution [2] , a decade later, was to name a related difficulty the hard problem and to argue that it is different in kind from the tractable ones. Explaining discrimination, memory, attention and report is a matter of finding mechanisms. Explaining why any of that is accompanied by experience is not, because a mechanism explains a function and experience is not a function. The two claims are often merged, and they are worth keeping apart: Levine's is about the explanatory shortfall of psychophysical identities, Chalmers's is about what kind of problem consciousness poses [3] . Does the shortfall show that physicalism is false? The reply physicalists reach for first is that an epistemic gap need not be an ontological one. The identity of water with a chemical compound was discovered, not deduced from the concept of water; likewise, on this view, phenomenal states are identical with physical states, and the identity is knowable only empirically. Chalmers labels this type-B materialism — the taxonomy of type-A [4] , type-B and type-C positions is his, not Ned Block's, though Block is among its best-known occupants — and it is important that its identities are held to be necessary . Kripke's argument that true identities hold of necessity is common ground; what type-B claims is that necessity here is not knowable in advance, not that the identity is contingent. The trouble is that the water analogy does not carry across, and the reason is the machinery Chalmers built for the purpose. In the water case there is a gap between how we pick the stuff out and what it is: we identify it by its appearance and its role, and that leaves room for a possible situation in which something else fills the role. That is why we can conceive of watery stuff that is not the compound; we are conceiving of the appearance being realised differently. Phenomenal concepts have no such structure. We pick out an experience by how it feels, and how it feels is not an appearance of something further. There is no gap between the appearance of pain and pain for the strategy to exploit. A physicalist who answers Chalmers by invoking a posteriori identity is using the case his opponent analysed in order to block that very move. What does Kim's model offer instead? Jaegwon Kim's Physicalism [5] , or Something Near Enough (2005) proposes a different reply, and its interest is that it does not treat the gap as one thing. Kim's model of reduction is functional. To reduce a property you first functionalise it, defining it by the causal role it occupies; then you identify what occupies the role; then you explain how the occupant does the work. This is how reduction proceeds where it has succeeded. The gene was functionalised as whatever carries hereditary information before anyone knew what the occupant was. On this model reducibility is not a global question but a property-by-property one, and the criterion is functionalisability. Kim argues that intentional and cognitive properties pass: believing, desiring, remembering and perceiving can be specified by their roles in producing behaviour and further mental states, and the explanatory gap closes for them. Qualia do not pass. The relational structure of experience does — the ordering of colours by similarity, the fact that this shade lies between those two, is a matter of discriminative role and is functionalisable. What is left over is the intrinsic qualitative character, and Kim holds that no role definition captures it. His title states the conclusion. Physicalism is true of almost everything and there is a residue. He accepts the consequence his own causal exclusion argument forces: an irreducible property does not get to do causal work, because the physical realiser is already doing it, so the residue is causally idle. So the answer is that the gap is not one gap? That is the position worth defending here. The honest physicalist reply to the explanatory gap is not that it is merely epistemic. It is that reduction succeeds where a property can be given a role definition and fails where it cannot, and that this criterion, applied to the mental, closes most of the gap and leaves a specific remainder. This is a claim someone can deny, and it is denied from both directions. Chalmers holds that a partial physicalism is not a stable resting place. Type-B materialists reject functionalisability as the test, holding that identities can be true without any role definition mediating them. And many readers take Kim's "near enough" to be a surrender dressed as a result. Has this argument been had before? In a different vocabulary, and with a rigour that repays attention, it has. The Four-Seven Debate occupied Korean Confucian philosophy for centuries [6] , beginning in the exchange between Yi Hwang, known as T'oegye, and the younger Ki Taesŭng in the 1560s, and continued between Yi I, known as Yulgok, and Sŏng Hon. The question was whether the Four Beginnings — the moral responses of commiseration, shame, deference, and the sense of right and wrong — arise from a different source than the Seven Emotions of ordinary feeling. T'oegye's mature answer was that they do: i , pattern or principle, issues and gi , psychophysical stuff, follows it, while in the ordinary case gi issues and i rides upon it. Yulgok rejected the first half. His formula is that gi issues and i rides upon it, and that there is