"Do you really believe raw-feel conclusions are conceptually necessitated by raw-material premises?" No, not really. I don’t even believe unmarried-male conclusions are conceptually necessitated by bachelor premises: I argue elsewhere (1996a) that we can discover, empirically, that none of the bachelors are male, and that none of them are unmarried—rhetoric about analytic conceptual connections notwithstanding. This matters because, even if gap-theorists are right to deny the existence of conceptual connections between nonphenomenal material and phenomenal feels—contrary to my stopgap argument—this lack might be blamable on the fishiness of "conceptual connections" rather than on the fishiness of "phenomenal feels." It is therefore important to a gap-theorist’s overall argument to find conceptual necessitation at work in explanations elsewhere in science. But if we consider their cases, I think, the proper conclusion to draw is this: even if there is a home for conceptual necessity, somehow and somewhere, it is too strong a requirement on fully satisfying scientific explanation.

Gap-theorists hold that scientific explanations of consciousness are doomed to incompleteness because of the following fact: there are conceivable worlds that are identical microphysically but differ in the distribution of conscious experience. If this is enough to doom explanations of consciousness, then the following fact should be enough to doom biological explanations of life: there are conceivable worlds that are identical microphysically but differ in the distribution of life—perhaps because they differ in the distribution of the "vital spirit" posited by vitalists or the "breath of life" described in Genesis. Even if vitalism and the Genesis account are false, if they are so much as conceptually possible this would mean that biology fails to live up to the high standards set by gap-theorists. And if vitalism is conceivable (even if false) for life, presumably vitalism-like positions are conceivable (even if false) for a wide range of phenomena studied in science: a "hydral spirit" explaining water, a "breath of electricity," etc. Chalmers considers and rejects the resulting threat to his position, but seems to miss its point:

[I]t is straightforwardly inconceivable that there could be a physical replica of a living creature [with its history physically replicated, as well] that was not itself alive. … Perhaps some ultrastrong vitalist would deny even this, claiming that something is left out by a functional account of life—the vital spirit, perhaps. But the obvious rejoinder is that unlike experience, the vital spirit is not something we have independent reason to believe in. Insofar as there ever was any reason to believe in it, it was as an explanatory construct—"We must have such a thing in order to be able to do such amazing stuff [adaptive behavior, reproduction, and the like]." (1996, pp. 108-109)

The mere fact that there is no great reason to believe in the vital spirit does not show that the existence of vital spirit is conceptually impossible. But its mere conceptual possibility—even in the absence of positive grounds for belief—would be enough to render biology unsatisfying, if gap-theorists are right.<2>

The conceivability of "hydral spirit" also vitiates Levine’s attempt to restrict the explanatory gap to consciousness:

[W]e justify the claim that water is H₂O by tracing the causal responsibility of, and the explicability of, the various superficial properties by which we identify water—its liquidity at room temperature, its freezing and boiling points etc.—to H₂O. But suppose someone pressed further, asking why being causally responsible for this particular syndrome of superficial properties should be so crucial. … I think we have to recognize an a priori element in our justification. That is, … our very concept of water is of a substance that plays such-and-such a causal role. (1993, p. 131)

Levine’s claims are too bold. We have weak or strong beliefs that water is liquid at room temperature, freezes and boils, etc. But none of this is built into our concept of water. There is, first, the point that none of these roles is individually conceptually necessary—as far as conceivability is concerned, water could be a fine powder (like toner) rather than a liquid, or it could contain miniature explosives that produce the mere appearance of boiling, etc. But also, there is no a priori guarantee that water is (if anything) something that plays "enough" of these roles. It is conceivable that water is that which is infused with hydral spirit, a spirit that gives water a strong desire to hang around something that plays these roles—say, H₂O. We do not need to stipulate "a priori" that this is conceptually impossible, in order to justify disbelief that it is true in fact. That would be typical philosophical overreaching. Rather, the denial in fact of conceptually possible hydral spirit can be justified on the usual domain-general grounds of scientific theory choice (simplicity, conservatism, etc.).<3>

Chalmers offers another example (besides life) of the alleged conceptual dependence of the biological on the physical:

What kind of world could be identical to ours in every last microphysical respect but be biologically distinct? Say a wombat has had two children in our world. The physical facts about our world will include facts about the distribution of every particle in the spatiotemporal hunk corresponding to the wombat, and its children, and their environments, and their evolutionary histories. If a world shared those physical facts with ours, but was not a world in which the wombat had two children, what could that difference consist in? Such a world seems quite inconceivable. (1996, p. 73)

On the contrary, it seems easy to conceive of a physically identical but biologically distinct world. Imagine a world in which physical creatures sometimes have as their immediate descendants purely spiritual creatures, and spiritual creatures sometimes have as their immediate descendants (via their physical parents’ bodies) purely physical creatures. In such a world, a wombat might have only a single child—an inhabitant of the spiritual realm—that in turn has two children—inhabitants of the physical realm. The physical creatures that develop within the wombat would then be its grandchildren rather than its children. This conceptual possibility does not in the slightest hinder science from explaining the production of children in our world. Biology has a fully satisfying account of the production of children—involving the material fertilization of material eggs in material environments, etc.—even though the operation of these material processes fails to necessitate the conclusion that children are produced.<4>

Perhaps biological categories are not the best examples for gap-theorists, since like most psychological categories they are functional categories whose members can be spiritual rather than physical. Perhaps nonfunctional macrophysical facts have a better chance of being necessitated by microphysical facts. (Of course, this outcome would not help gap-theorists in their attempt to argue that explanations of conscious experience are less satisfying than explanations elsewhere in biology and psychology.) Chalmers considers the "sorts of things relevant to something’s being a table," including the condition "that it have a flat top and be supported by legs." He calls this "a structural condition: that is, a condition on the intrinsic physical structure of the object," and states that "[s]tructural properties are clearly entailed by microphysical facts" (1996, p. 79). On this view, microphysical facts are sufficient conceptually to entail facts about whether something is a flat top or a leg. But they are not.

First, consider a conceptual possibility opened up by Hilary Putnam’s (1975) "division of linguistic labor," or the strong tendency we have to defer to our linguistic community (particularly, acknowledged experts) in categorization. We are quite confident about our ability to detect table legs, but we are ready to defer to expert carpenters. Suppose we dutifully seek out the best carpenter in the world, with a question about whether something is a leg. Luckily, he also turns out to be the world’s foremost microphysics expert, and in many other ways impresses us with his knowledge, intelligence, and honesty. Unbeknownst to us, this carpenter is an omniscient and partly spiritual being, whom we know merely by the name on his business card, "Jesus." In this case, unbeknownst to us, which things are legs depends not only on microphysical facts but also on nonphysical facts about Jesus. Just as on reflection we defer to experts who deny that certain seeming fish are really fish, so following Jesus we may withhold "leg" from items that seemed—due to their microphysical structure—to be legs. There is a similar problem with Chalmers’ claim that "there does not seem to be a conceivable world that is naturally identical to ours but morally distinct" (1996, p. 83). Hoards of people think moral facts depend on God’s will, which illustrates perfectly how the linguistic division of labor can cross the physical/spiritual boundary. For hoards of conceivers, if God considers two microphysically identical worlds and deems one good and the other bad, then—by golly!—one is good and the other is bad. (Chalmers seems to confuse the arguable supervenience of the moral on the nonmoral with the doubtful supervenience of the moral on the natural.) I suspect it would not be difficult to get such conceivers to defer similarly about legs and flat tops. I would not wish to infer, merely on these grounds, that they lack (or otherwise "violate") the concept of legs and flat tops, or of good and bad.

There are other ways for kinds we think are "structural" to be really nonstructural. This is not a matter settled by our concepts. Consider a variant of Rogers Albritton’s famous example in which pencils turn out to be reproducing organisms (credited by Putnam, 1975, pp. 242-43). In this variant, it turns out, legs are reproducing organisms, but they reproduce via the spiritual dimension (like wombats might). Contrary to appearances, legs form a biological kind—perhaps a theoretically very important species, a "missing link" between chimpanzee and man, say, whose discovery would overturn many theories of evolution. (It is also possible to build into the example that we discover some of this, storing the information in our nonphysical souls, but I will not pursue this possibility.) Biological species are in part determined by evolutionary history—history in the physical or nonphysical realms, as the case may be. They therefore give rise to the possibility of fakes—entities that are structurally similar but lack the right (physical or spiritual) evolutionary history. Now suppose all this is true of our world. Then a second world could be microphysically identical to ours and yet at some places in which our world contains genuine legs the other world could contain mere leg-fakes. Both legs and leg-fakes would support things, of course, but (as it turns out) this is not what makes something a leg. This is "conceptually possible," because our concepts do not determine which kinds are structural and which kinds are not. At best, they set up weak or strong expectations that certain kinds are structural. These expectations can go wrong.

Chalmers mentions and dismisses, in an abstract way, the threat I am pressing, that "purported conceptual truths are always subject to revision in the face of sufficient empirical evidence":

This is true for many purported conceptual truths, but it does not apply to the supervenience conditionals that we are considering, which have the form "If the low-level facts turn out like this, then the high-level facts will be like that." The facts specified in the antecedent of this conditional effectively include all relevant empirical factors. Empirical evidence could show us that the antecedent of the conditional is false, but not that the conditional is false. (1996, p. 55)

The problem with this response lies in the second-to-last quoted sentence. The "low-level facts" (in the antecedent of the conditional) do indeed include all relevant physical empirical factors, but they do not include all (or even any) relevant spiritual empirical factors.<5>

Conceptual necessity is hard to come by. I would not be surprised, then, if the presence of inner perception, together with my stopgap "analysis," does not literally entail the existence of conscious experience. One can always posit "phenomenal spirit" or the "breath of consciousness," or (as Chalmers might prefer to call them) fundamental phenomenophysical laws of nature. These are conceptually possible, I suppose, just as vital spirit (or fundamental biophysical laws) and hydral spirit (or fundamental hydrophysical laws) are conceptually possible. As in these cases, it is enough if inner perception and the stopgap provide excellent evidence of conscious experience (together with simplicity, conservatism, and other domain-general grounds for scientific theory choice).

Recap

In this paper I have argued that there is no explanatory gap between nonphenomenal material and phenomenal feels, or at least no gap that is not also present between nonvital material and vital life, nonhydral material and hydral water, etc. As near as matters for fully satisfying scientific explanation, we can deduce the existence of conscious experience from certain scientific facts—e.g., about inner perception. Roughly, if one innerly perceives one’s mental states as being some way, and this inner perception helps constitute the states themselves, then it follows that there is something it is like to have the states—at least if "that’s all," i.e., barring "phenomenal spirit" or the "breath of consciousness." If philosophers’ nonphenomenal zombies (exact physical replicas of phenomenal beings, but without conscious experience) are "possible" in any sense, this is in the same sense in which the following are "possible": nonvital zombies (exact physical replicas of vital beings, but without life), nonhydral zombies (exact physical replicas of hydral stuff, but without waterhood), nonfilial zombies (exact physical replicas of one’s filial beings, that are not one’s children), and so on. None of this threatens scientific explanation of consciousness any more than it does scientific explanation of life, water, and children.

Lormand, Eric. 1996a. "How to be a meaning holist," Journal of Philosophy, 93: 51-73.

Putnam, Hilary. 1975. "The Meaning of ‘Meaning,’" in Mind, Language, and Reality, Cambridge: Cambridge University Press: 215-272.