Six
Problems of Dualism and Materialism and Their Common Root

As I discussed in the introduction, there is widespread agreement that both dualism and materialism are inadequate. Of course, materialists have always found insuperable problems in dualism, and dualists have always found equally insuperable problems in materialism. But now members of each camp are admitting deep problems in their own positions, holding these positions not as adequate solutions but only as the least inadequate of the options. This attitude raises the question as to whether another option, perhaps more adequate than either, has been overlooked. I discuss that question in the next chapter. Here, I prepare the way for discussing that option—an option against which the modern mind has been so biased that much preparation is necessary—by pointing out that the problems of dualism and materialism are both rooted in the same source: the Cartesian intuition about matter.

The phrase "Cartesian intuition" has generally been used with regard to Descartes's view of the mind, as if it were the basic source of the mind-body problem. His view of the mind is indeed problematic (as I will discuss in chapter 8), but even more problematic is his view of matter. By this I mean his view, not unique to Descartes but associated primarily with him, that matter is completely different in kind from mind. Matter is spatially extended, mind is not. Mind has temporal duration, matter does not (in the sense that it can exist at an "instant," not requiring any temporal duration to be what it is).^[*] Mind has an "inside," consisting of thoughts, desires, feelings, and volitions, and thereby has intrinsic value; it is pour soi, something for itself. Matter is all "outside" and is therefore devoid of any

[*] See footnote, p. 49, below.

value for itself; it is only en soi. (Descartes expressed this idea by saying that [spatial] extension is the only essential attribute of matter; but other thinkers, while rejecting this reduction of matter to extension, have agreed that matter has only an outside, no inner reality, such as feelings.) Matter exerts causal efficacy only by efficient causation (for Descartes, only by impact; Newtonians would disagree); mind exercises final causation or self-determination.

Searle is right: To be willing to entertain a radically new view of the mind-body relation, we need to see that dualism and materialism, widely supposed to be the only real options, are both false (RM, 2–3). To develop an alternative that really solves the problem, however, we need to see what the root cause of the problem has been. Searle moves in the right direction by questioning the "conceptual dualism" inherited from Descartes, according to which if something is "physical" it cannot also be "mental" and vice versa (RM, 14, 26, 54). But he does not take this insight far enough. He ends up, accordingly, with simply one more problematic version of materialism, because he continues to assume that most physical things are not also mental. It is precisely this assumption, I will argue, that creates the insuperable problems of the various dualisms and materialisms alike.

Although I have thus far, like most writers, referred simply to "dualism" and "materialism," as if these terms were unambiguous, I now need to specify more exactly how I use them. "Dualism" I always use in the sense of ontological (or Cartesian ) dualism . This doctrine contains a double thesis: (1) that the mind is an actuality numerically distinct from the brain (the quantitative or numerical thesis) and (2) that it is ontologically different in kind from the entities of which the brain consists (the qualitative or ontological thesis).

Many writers use "dualism" as a synonym for dualistic interactionism . There are, however, parallelist forms of ontological dualism, which say that the interaction of mind and brain is merely apparent. Some parallelists have explained the appearance by reference to God (Malebranche, Geulincx), whereas those not able or willing to call on supernatural assistance have left the synchronization an even greater miracle. Because most thinkers today employ regulative principles ruling out miracles, whether explicitly supernatural or not, few dualists today are parallelists. Another form of dualism that does not affirm interactionism is full-fledged epiphenomenalism, according to which the mind is a semi-actuality numerically distinct from the brain. It receives causal influence from the brain, and may even be capable of determining some of its own states, but it cannot exert any downward causation back on the brain (which is why it can be considered only a semi -actuality). However, although Keith Campbell, quoted in the introduction, is one, there are few confessing epiphenomenalists in this sense

today. So for most practical purposes, "dualism" as I intend it can be equated with dualistic interactionism, and I always use it with this meaning unless indicating otherwise.

A widespread practice that should be avoided, however, is the use of "dualism" to refer to all positions that affirm interactionism or, which amounts to the same thing, the use of "interactionism" as a synonym for dualistic interactionism . The problem with these practices is that they imply that there could not be a nondualistic interactionism, thereby serving to restrict our thinking about alternatives. To call a position "dualistic" simply because it posits a distinction between brain and mind (the numerical thesis) creates the impression that any such position would necessarily have all the problems of Cartesian (ontological) dualism, and that might not be true.

Finally, dualism in the ontological sense is often called "substance dualism" (to distinguish it from so-called property dualism, which is generally best classified as a form of materialism). I do not use this term, however, partly because the term "substance" suggests its Cartesian meaning of "requiring nothing but itself in order to exist," whereas many ontological dualists today see the mind as dependent on the body for its existence. Also, the term "substance" suggests an entity of long duration, whereas a contemporary ontological dualist might well speak of a dualism of types of "events" or "processes."

By "materialism" I mean materialistic monism, which contains the double thesis (1) that there is only one kind of actual entities, namely, material or physical ones (the qualitative or ontological thesis), and (2) that what we call the "mind" is somehow numerically identical with the brain (the quantitative or numerical thesis), so that there is no interaction between mind and brain. (The other possible quantitative meaning of "monism," that there is in reality only one actuality, as Spinoza suggested, is not in view. In this respect, contemporary materialism is a form of pluralistic monism, being monistic qualitatively but pluralistic quantitatively.) Because this is how virtually everyone today uses the term, the simple use of "materialism" (unlike the simple use of "dualism") occasions no problems. The widespread practice of sometimes using "monism" as a synonym for materialistic monism, however, is problematic, because it suggests that there are no nonmaterialistic forms of (pluralistic) monism, and this is false. In any case, by "materialism" I mean the twofold thesis that (1) all actual things are material and (2) there is no mind or soul in the sense of an actuality numerically distinct from brain. In fact, it is a threefold thesis, because the statement that "all actual things are material" must be specified to mean that at least most actual things, certainly the most fundamental ones, are devoid of any experience. For example, Strawson says that materialists "believe that there was once no experiential reality on earth but plenty of nonexperiential reality, and that experiential reality came to exist as life evolved" (MR, 66).

Because of this usage, I classify as materialists some who are sometimes regarded as questioning full-fledged materialism, such as Strawson, Searle, and Nagel (although if Nagel were to turn his flirtation with panpsychism into full-fledged embrace, he might be a proponent of one of the nonmaterialistic monisms to which I referred above).

With these clarifications, I will now summarize the major problems of dualism and materialism, showing how all of these problems are rooted in the Cartesian intuition about matter. I will divide this discussion into those problems that are unique to dualism, those that are unique to materialism, and those that are common to both dualism and materialism.

I. Problems Unique to Dualism

1. The chief problem of dualism has always been to understand how two totally different types of things could causally influence each other . How could that which is spatially extended and embodies physical energy but is devoid of any duration,^[*] therefore of any "inside," therefore of any feelings and desires, and therefore of any intrinsic value, be capable of exerting causal influence on a nonphysical mind? As John Passmore says, according to dualism a "body can only push" (PRE, 55). How could a body exert efficient causation on that which takes up no space (at least not in an impenetrable way) and embodies no physical energy (which is one side of the problem suggested by speaking of the mind as the "ghost in the machine")? What would such a body have to offer something that lives in terms of values? Likewise, as Passmore says, according to dualism "the only force the mind has at its disposal is spiritual force, the power of rational persuasion" (PRE, 55). How could it exert causal efficacy on something that is constituted so as to be affected by other pushy things (which is the other half of the ghostin-the-machine problem)? Things with final causation inwardly and the

[*] It may seem that matter as conceived in modern thought is not devoid of duration, because it endures through time. What is meant, however (as the ensuing discussion will gradually make clear), is that matter as usually conceived is thought not to require any lapse of time in order to exist, which means that it can exist "at an instant" (with "instant" understood to be a durationless slice in time). In matter thus conceived, in Whitehead's words, "the lapse of time is an accident, rather than of the essence, of the material. The material is fully itself in any sub-period however short. Thus the transition of time has nothing to do with the character of the material. The material is equally itself at an instant of time" (SMW, 50). "Matter," accordingly, "involves nothing more than spatiality" (MT, 132). Whitehead's contrasting view is that what we call the physical world is made up of events, that "an event in realizing itself displays a pattern," and that "the pattern requires a duration involving a definite lapse of time, and not merely an instantaneous moment" (SMW, 124). Put otherwise, each primordial element is "an organized system of vibratory streaming of energy," and such a system is, like a note of music, "nothing at an instant, but . . . requires its whole period in which to manifest itself" (SMW, 35).

power of persuasion outwardly, on the one hand, and things with no final causation inwardly and the capacity outwardly to cause and be caused only by pushing power, on the other, are ill-suited for the kind of interaction at which our minds and bodies seem quite good. We can, accordingly, sympathize with Descartes when he finally said, in response to Princess Elisabeth's persistent questioning as to how mind and body could interact, that the human mind is not capable of conceiving this very distinctly.^[1] We can also understand why other thinkers, given their unquestioning belief in supernaturalistic theism, would assign the interaction, whether real (Reid) or apparent (Malebranche, Geulincx), to God.^[2] We can understand, further, why a contemporary dualist reluctant to call on a deus ex machina, such as Popper, would return (as cited in the introduction) to Descartes's confession of ignorance. Popper does try to mitigate the problem by saying that quantum physics has superseded Descartes's idea of causation, according to which bodies push each other around (SAB, 483, 499, 510). But then, in explaining what he means by affirming the "ghost in the machine," he says, "I think that the self in a sense plays on the brain, as a pianist plays a piano" (SAB, 494f.). How a physical-physical (finger-piano key) relation can be used as an analogy for the psychical-physical (mind-brain) relation is not cleared up by telling us that physical-physical relations are not as pushy as Descartes had thought.

Some dualists have tried to finesse the problem of causal interaction between ontologically unlike things by appeal to Hume's contention that only experience can tell us what in fact can or cannot cause what—in other words, that we have no basis for a priori claims that physical events cannot be the cause of mental events and vice versa. However, Hume's contention presupposed his analysis of (efficient) causation, according to which it refers merely to regularity of sequence ("constant conjunction"), and this understanding of causation was ruled out in the substantive regulative principles: Efficient causation is real influence, not mere constant conjunction. Furthermore, if one returns to a realistic view of causation as real influence, one cannot—philosophers such as C.J. Ducasse and H. D. Lewis to the contrary—validly appeal to Hume's rejection of all a priori ideas about causation.^[*] So the problem of the interaction of ontologically unlike things remains.

2. Dualism also violates the principle of continuity . The idea that somewhere

[*] C.J. Ducasse, with an appeal to Hume, says, "The Causality relation is wholly neutral as to whether the cause-event and the effect-event are both physical, or both psychical, or either of them physical and the other psychical" (MMB, 85). And yet Ducasse rejects as "patently invalid" Hume's view of causality as merely empirical regularity of sequence. The term causation means, Ducasse insists, that the event called the cause etiologically necessitates the effect (MMB, 83). A similar position has been taken by H. D. Lewis (EM, 26–99, 123, 173). Neither seems to see that Hume's first point presupposes the second.

late in the evolution of the universe an entirely new type of actuality has sprung into existence would seem either to require a supernatural cause or to constitute an even more incredible miracle. In truth, this problem could be counted as one of the problems shared with materialism, because materialists also affirm an essential discontinuity; and, indeed, I will raise this point below. But it is proper to include the problem here because dualism, in affirming the emergence of a new type of actuality, generally with a new type of causal power (that of final causation or self-determination), violates the principle of continuity in a particularly egregious way.

3. A third charge against dualistic interactionism is that it apparently violates the principle of the conservation of energy . Eccles has regarded this charge as the problem to which he as a dualist needed to provide an answer (HS, 23, 79, 140, 168). This problem, however, has been greatly exaggerated. For one thing, we do not know the absolute truth of this principle. Do we know, for example, that the energy of the universe has remained constant from the big bang, assuming that there was one, to the present? Do we know that the principle, insofar as it holds absolutely in some domains or contexts, does so in all? We certainly have less reason to be confident of the absolute, universal truth of this principle than we do of the mutual influence of mind and body. So, if forced to choose between them, one should give up the absolute truth of the principle of conservation. But perhaps the dualist need not choose. W. D. Hart has recently argued, in defending dualism, that energy may be conserved in the interactions between mind and body. That is, the principle of the conservation of energy depends on understanding energy as a quantity that is conserved while being converted into various forms, such as mechanical, electromagnetic, thermodynamic, and chemical (EOS, 62–64). Perhaps, Hart suggests, the idea of "psychic energy" should be taken literally, so that it would be one more form that would be interconvertible with the other forms (EOS, 127, 149, 152, 178, 186n). It may be, accordingly, that Eccles (tragically) devoted much of his life to the solution of a pseudoproblem.

In any case, although Hart properly describes his position as Cartesian dualism, insofar as he speaks of "two basic or fundemental sorts of things" (EOS, 1, 8), his resolution, in ascribing energy to both mind and matter, actually moves away from Cartesian (ontological) dualism toward a form of nondualistic interactionism. However, he does not go all the way, so his position still suffers, like that of Eccles and other dualists, from the problems of discontinuity and causal interaction between ontologically unlike things.

Whereas materialists typically use these problems unique to dualism (plus some others, to be mentioned below, that are in fact shared by materialism) to rule out dualism, materialism has even more problems unique to it .

II. Problems Unique to Materialism

1. One of materialism's problems is that of accounting for our unity of experience (as discussed in chapter 5). The idea that there is no mind over and above the brain, which is composed of 100 billion or more neurons, each of which is in turn composed of myriad particles (in which all agency is said finally to be lodged), creates a great puzzle as to why we should enjoy the kind of unified conscious experience we normally do. Dennett tries to mitigate this problem by saying that the unity is a mere appearance (CE, 23, 458). But if in reality there are billions of "miniagents and microagents (with no single Boss)" and "that's all that's going on" (CE, 458, 459), the very appearance of unity is utterly mysterious. Searle is more forthright: After illustrating the unity of consciousness—"I have my experiences of the rose, the couch, and the toothache all as experiences that are part of one and the same conscious event"—he adds, "We have little understanding of how the brain achieves this unity" (RM, 130). Dualism is more intuitively adequate here, because it can attribute our experienced unity to the actual unity of the mind. Eccles, for example, says that "a key component of the [dualist] hypothesis is that the unity of conscious experience is provided by the self-conscious mind, not by the neural machinery" (HS, 22). Indeed, dualism could identify the mind with the unification of various influences from the brain—if only it could explain how the mind, being ontologically different from the brain cells, could receive influences from them.

2. The mirror image of this problem is that of accounting for the unity of our bodily behavior . How can I pat my head with my hand while beating time to music with my foot while smiling at my wife while thinking about the mind-body problem while . . . ? If there is in no sense a "single Boss," so that all of our behavior, inner and outer, is produced by an aggregate, how can there be such remarkable coordination? Dualism, with its distinction between the (more or less) unified mind and the aggregational brain, again seems better able to handle the phenomena—if only it were not unable to explain how the mind can affect the body at all.

3. Closely related is the problem of freedom . If the "mind" is just the brain, or some aspect or function thereof, then it would seem impossible to ascribe self-determining freedom to us. Even if quantum indeterminacy is taken to qualify the older notion of absolute causal determinism, how the indeterminacy of trillions of particles could account for our sense of freedom would be far from clear. In any case, the indeterminacy of the individual particles or events is generally canceled out in aggregates by the "law of large numbers." And, indeed, at least virtually all materialists do deny freedom (as illustrated in chapter 5). In Elbow Room: The Varieties of Free Will Worth Wanting, Dennett tries to convince us that freedom in this sense is

not "worth wanting" anyway, ascribing to us only kinds of freedom that are compatible with causal determinism. McGinn, while rejecting Dennett's eliminativism with respect to consciousness, joins him with respect to free will (PC, 17n).

Even Searle, after berating most materialists for denying obvious facts of experience, articulates a theoretical position that denies freedom, although he agrees (as seen in chapter 5) that it is "obvious" in the sense of being one of our hard-core commonsense notions. After noting that Howard Gardner's comprehensive summary of cognitive science in The Mind's New Science does not contain a chapter or even an index entry on consciousness, Searle comments sarcastically, "Clearly the mind's new science can do without consciousness" (RM, 249). But Searle's book The Rediscovery of the Mind has nary a chapter or even an index entry on freedom. Clearly, one could retort, the rediscovered mind can do without freedom!

There is no doubt that Searle still denies freedom. (I speak here of his theoretical, philosophical position; as I reported in chapter 5, Searle pointed out in his earlier book [MBS ] that in practice he cannot give up his conviction of freedom.) Although he lists a dozen structural features of consciousness (RM, 127–41), he does not include freedom. When Searle does finally mention "free will" in a list of features of consciousness, he puts after it the qualification "(if there is such a thing)" (RM, 227). This implicit denial to conscious experience of any capacity for self-determining freedom is made explicit in a passage in which Searle explains in what sense consciousness is and is not an emergent feature:

A feature F is emergent2 if F is emergent1 and F has causal powers that cannot be explained by the causal interactions of a,b,c. . . . If consciousness were emergent2, then consciousness could cause things that could not be explained by the causal behavior of the neurons. The naive idea here is that consciousness gets squirted out by the behavior of the neurons in the brain, but once it has been squirted out, it then has a life of its own. (RM, 112)

Searle then adds, "On my view, consciousness is emergent1 but not emergent2" (RM, 112). In other words, consciousness is totally produced by the brain, and, once produced, it has no partially autonomous power by which it could determine some of its own states, such as its desires, attitudes, and volitions, which then might influence the brain with something that had not been simply produced by it. His position is then the same as Sperry's: Nothing goes down that had not first been sent up. Searle, in fact, had explicitly said this in his earlier book: "Top-down causation only works because the top level is already caused by and realized in the bottom levels" (MBS, 94).

In short, the datum of freedom, like the data of the unity of experience

and the unity of our bodily behavior, favors dualism over materialism—or at least would if the problems of discontinuity and dualistic interaction could be ignored.

4. Whereas the three preceding problems should be felt as real problems by materialists, they have received relatively little treatment. The problem that has received by far the most attention is, not surprisingly, the most obvious one, which is what it can mean to say that we, with our conscious experience, are wholly material (or physical) beings. Conscious experience, with all its features, evidently has to be portrayed as somehow reducible to, or somehow resolvable into, purely material stuff, entities, processes, structures, functions, or something that can in principle be adequately described from a wholly externalist, third-person perspective. The basic problem of all of these programs is the simple point raised by Nagel: If the world contains experiences, points of view, beings that it is like something to be—and we know that it does—then every purely externalist description will necessarily leave something out. Most of those materialists who have recognized this point have abandoned the various forms of identity theory and moved to eliminative materialism. They simply eliminate consciousness, experience, points of view, from the list of realities to which a theory must be adequate. As Paul Churchland has said, "If we do give up hope of reduction, then elimination emerges as the only coherent alternative."^[3] (If you can't join 'em, eliminate one of 'em.) This is the process that Searle has described as a long pattern of neurotic-like behavior and Seager as "an orderly retreat becoming a rout" (MC, 32).

Whereas Seager, however, agrees with Nagel, Robinson, McGinn, and Strawson that no solution is on the horizon, Searle believes that he has produced one. Although he criticizes materialism, his position is still within the materialistic framework, as I have defined it. That is, Searle says that the mind is not an actuality numerically distinct from the brain, that reality is "entirely physical" (RM, 54), and that the fundamental units of the world are "material" or "physical" in the sense of being devoid of any experience, any "point of view." (Although Searle says "we do not know at present how far down the evolutionary scale consciousness [by which he seems to mean experience of any sort] extends," he is "not inclined to ascribe any consciousness" to amoebas [RM, 89, 74]; he asks, "How do unconscious [by which he seems to mean totally nonexperiencing] bits of matter produce consciousness?" [RM, 55]; and he says, of objects about which we have a point of view, "the objects themselves have no point of view" [RM, 31].) He is critical of hitherto dominant forms of materialism insofar as they assume that if something is "physical" it cannot also be "mental," or, more precisely, that "the same phenomenon under the same aspects cannot literally satisfy both terms" (RM, 10, 14). He sees this notion, which he criticizes, to be

closely related to the assumption that "reality is objective (meaning accessible to all competent observers)" (RM, 16), which he sees in turn as closely related to the assumption that "the objectivity of science requires that the phenomena studied be completely objective," which is in turn "based on a confusion between the epistemological sense of the subjective-objective distinction, and the ontological sense" (RM, 10, 19). I think Searle is right about all this (except the assumption that the same thing under the same aspects can be both physical and mental). But the way that he develops his alternative position is (partly because of that faulty assumption) as problematic as the theories he has rejected.

According to Searle's revisionist materialism, "materialists [can] cheerfully embrace consciousness as just another material property among others" (RM, 55). According to Searle's "biological naturalism," more precisely, "the mental state of consciousness is just an ordinary biological (that is, physical) feature of the brain" (RM, 1, 13). What can Searle mean in referring to consciousness as "ordinary," as "just another material property among others"? He emphatically does not mean that it, like all others, can be described "objectively" in the sense of "from a third-person point of view." His realization that this is impossible is the basis for his protest against eliminative materialism and most forms of reductionism. He insists that "the actual ontology of mental states is a first-person ontology," that "the real world contains . . . elements that are irreducibly subjective" (RM, 16, 19). So what does it mean to say that these irreducibly subjective states are entirely physical?

One of the root problems in Searle's book is that although he tells us what he does not mean by "physical," he never tells us what he does mean. One common meaning is "devoid of experience"; but Searle is denying that this is a necessary feature of the physical, because he thinks that some physical things (brains) do have experience. Nagel, who does accept this definition of the physical, puzzles over the meaning of Searle's position: "But however great the variety of physical phenomena may be, ontological objectivity is one of their central defining characteristics; and as we have seen Searle insists that consciousness is ontologically subjective" (NYR, 40). Another candidate for the meaning of "physical" is "publicly observable"; but Searle also rejects this definition, because our own subjective experience is not publicly observable. If it means neither of these, however, what does it mean? If he were to say, for example, that to be physical is simply to embody energy, he could affirm a truly nonmaterialistic pluralistic monism, in which all units have mental as well as physical features, which would open the way to a resolution of the mind-body problem quite different from the position he has hitherto taken.

In any case, we still need to explore the intelligibility of Searle's claim

that conscious states are just another, ordinary physical property, analogous to others. To make this case, Searle offers several analogies. Two of his favorites are liquidity and solidity:

Consciousness is a higher-level or emergent property of the brain in the utterly harmless sense of 'higher-level' or 'emergent' in which solidity is a higher-level emergent property of H₂ O molecules when they are in a lattice structure (ice), and liquidity is similarly a higher-level emergent property of H₂ O molecules when they are, roughly speaking, rolling around on each other (water). Consciousness is a mental, and therefore physical, property of the brain in the sense in which liquidity is a property of systems of molecules. (RM, 14)

Lying behind Searle's analogy is the notion that at least many features of big systems can be causally explained by the behavior of the little systems of which they are composed: The macrophenomena are explicable in terms of microphenomena (RM, 87). He is claiming, then, that consciousness is a surface phenomenon caused by microphenomena in the brain, just as the solidity of ice and the liquidity of water are surface phenomena caused by the microphenomena beneath the surface. He uses the language of "supervenience"^[*] to express this relation: Consciousness is supervenient on neurophysiological states just as liquidity is supervenient on certain molecular states.

Is there a genuine analogy here? Seager, in response to an earlier version of Searle's position, pointed out one reason why this type of analogy breaks down. Seager distinguishes between constitutive supervenience, which holds in ordinary physical explanations such as those of solidity and liquidity, and merely correlative supervenience, which consciousness seems to exemplify in relation to the brain. In the former, the lower states are somehow constitutive of the supervenient states in a way that allows us to understand why that supervenient state should emerge:

Roughly speaking, in cases of constitutive supervenience the dual evidence provided by a knowledge of a system's basic components and their link to its behavior is decisive for ascription of the supervenient property. . . . [I]t makes credible the idea that the joint activity of the various components,

[*] For a thorough discussion of the meaning(s) of "supervenience," see Kim's Supervenience and Mind . According to the evidently dominant usage, the core meaning is that the observable, macroproperties of an entity are fully determined by its microproperties—although perhaps not reducible to them (SM, 275f., 339f.). As Searle's use of "supervenience" in place of "emergence" suggests, the two terms can be considered virtnally synonymous. In fact, as Kim points out, one of the early evolutionary emergentists, Lloyd Morgan, used "supervenience" as a stylistic variant of "emergence," and some nonreductive physicalists today speak of their higherlevel, supervenient properties as "emergent" (SM, 134, 348). I will continue to use primarily the language of "emergence" until discussing Kim's position in chapter 10.

through their own causality, could reasonably be claimed to produce the system's overall behavior. (MC, 179)

However, Seager points out, the relation of consciousness to brain states is not like that: Nothing about the behavior and property of the neurons as studied from without would lead us to expect some brain states to produce consciousness. The supervenience remains brute, merely correlative. Seager approvingly cites Simon Blackburn's statement that supervenience of the psychological on the physical is part of the problem, not the solution (MC, 180).^[*]

In response, Searle agrees that the kind of supervenience that is relevant to the mind-body problem is causal supervenience, not constitutive supervenience (RM, 125). But then, it would seem, all of the analogies he employs—photosynthesis, digestion, mitosis, electromagnetism, and transparency (RM, 90, 104) as well as solidity and liquidity—are of no intuitive help in understanding how consciousness is related to the brain, because they are all cases of constitutive supervenience. The only thing that causal and constitutive supervenience have in common is that in both the higher-level property is said to be "caused" by the lower,^[**] and in the case of consciousness we have been given no clue as to how this "property" could have emerged out of its causes. Strawson agrees. With regard to the idea that experiential properties are "reducible to nonexperiential physical properties in a way that is ultimately similar to the way in which the property of liquidity is held to be reducible to van de Waals molecular-interaction properties," he says: "This reduction is very hard—impossible—to imagine" (MR, 68). Searle's analogies may at first glance give the appearance of providing a glimmer of intuitive understanding, but on closer examination they are more confusing than helpful. As Nagel has said, "much obscurity has been shed on the [mind-body] problem by faulty analogies" (MQ, 202).

[*] Kim agrees, saying that supervenience "is a 'phenomenological' claim, not a theoretical explanation. Mind-body supervenience, therefore, does not state a solution to the mind-body problem; rather it states the problem itself." Kim adds that although the research strategy of explaining psychological functions in terms of the interactions of subsystems conforms to the generally accepted scientific approach, "whether such microstructural explanations really 'explain' mentality in the sense of making mentality, in particular consciousness, intelligible—something that the emergentists despaired of ever attaining—may be another question" (SM, 168).

[**] As Kim portrays it, supervenience is a relation of dependency but not of causal dependency. The macropropeties are said to be determined by the microproperties without being caused by them. The difference is that the causal relation is temporal—the cause is antecedent to the effect—whereas in supervenience, the determining and the determined are simultaneous (SM, 354, 359). As Kim points out, however, this distinction has little import (SM, 359), so the language of "causal supervenience" seems to do no violence to anything essential to the idea of supervenience.

Searle's position is not really as different from previous forms of materialism as he suggests. In spite of his emphatic rejection of "property dualism," his position, as Nagel points out (NYR, 40), seems to be another variant of it. I will return to Searle's position when I consider the problem of the emergence of experience out of nonexperiencing entities, which materialists and dualists alike share; at that time I will consider Searle's claim that his position is really that of "property polyism" (MBWP, 228).

5. A fifth problem for materialism is that it implies an epistemology that does not allow us to know various hard-core commonsense notions (which we all do know). I refer here to an epistemology based on the sensationist view of perception (along with assumptions, of course, that prevent us from knowing these things a priori, or through divine revelation, or otherwise transcendentally). This problem could be included under the problems that materialism and dualism share, because many dualists in modern times (beginning at least with Locke) have accepted a sensationist theory of perception (although Locke himself could still appeal to supernatural revelation to explain how he knew things he should otherwise not have known). However, dualists need not accept a sensationist view of perception, because a mind distinct from the brain might be able to perceive some things directly, without the aid of the brain and its sensory organs. This problem, therefore, is best included under problems unique to materialism.^[*]

Although each of the five examples I will give would be worthy of an entire chapter, if not a book, I can mention them only briefly. Two of them are most closely associated with Hume: knowledge of efficient causation and knowledge of a world external to the mind . Although there have been enormous efforts to get around Hume's demonstrations, they remain valid: If sensory perception, understood as perception providing only sensory data, were the only mode of perception we have, we should not know what efficient causation (in a non-Humean sense) is or that there is a real world to which the sensory data may refer. Likewise, as Santayana showed in an extension of Hume's argument, if this were our only mode of perception, we should not know (as distinct from taking on "animal faith") the reality of the past . With no knowledge of the past, furthermore, we would have no knowledge of time . Finally, our inexorable presuppositions about objective norms, such as truth, goodness, and beauty, are notoriously difficult to ex-

[*] Many modern philosophers simply assume the equation of perception with sensory perception. For example, in the editor's introduction to an anthology entitled Perceptual Knowledge (Oxford: Oxford University Press, 1988), Jonathan Dancy simply asserts on the first page that perceptual knowledge is the sort of knowledge we get by "using our senses." Some philosophers, however, state the sensationist position explicitly. Willard Quine, for example, says, "The stimulation of his sensory receptors is all the evidence anybody has to go on, ultimately, in arriving at his picture of the world" (Ontological Relativity and Other Essays [New York: Columbia University Press, 1969], 75–76).

plain from within a materialistic, sensationist perspective. (The implicit contradiction that results is illustrated by those philosophers and scientists who support their preference for a materialistic worldview in terms of aesthetic criteria, such as "simplicity" and "beauty," ignoring the fact that such a worldview denies the possible knowledge and even existence of such norms.) A position that could fully allow the genuineness of our knowledge of all these things and that could explain it naturally, without forcing, would be in this respect more adequate than any of the materialistic positions.

6. A related problem for materialists, given the virtual necessity of their restricting perception to that which occurs through the physical sensory organs, is the impressive evidence for extrasensory perception, in the sense of telepathy and clairvoyance.^[4] D. M. Armstrong, for example, has referred to psychical research as a "small black cloud on the horizon of a Materialist Theory of Mind." Herbert Feigl and Keith Campbell have both said that if the alleged evidence for extrasensory perception were authentic, materialism could not handle it.^[5] And many dualists point to the data from psychical research as decisive evidence in favor of dualism.^[6] Of course, these alleged data are rejected out of hand by most materialists, such as those just mentioned and McGinn (PC, 53), Humphrey (HM, 11), and Churchland (MAC, 17). However, these rejections generally demonstrate little if any familiarity with the evidence and discussions of it by otherwise reputable philosophers, such as William James, C. D. Broad, C. J. Ducasse, and H. H. Price. (By contrast, Seager, who shows more familiarity with the evidence, is more open to it [MC, 188, 241].) These rejections seem to be based more on paradigmatic and wishful-and-fearful thinking, in other words, than on considered judgments, after serious investigation, that good evidence is really lacking.^[*]

[*] While doing the final revision of this manuscript, I learned that Humphrey had recently written a book, Soul Searching: Human Nature and Supernatural Belief (London: Chatto & Windus, 1995), in which parapsychology is extensively discussed. Humphrey does indicate that he has become at least somewhat familiar with the available evidence and knows that some of it is good. He says that "the evidence that has accumulated over the ages [for ESP] is undeniably impressive" (115), and he refers to "the hundreds of experimental studies that over the last century have made strong claims to demonstrate the reality of PK or ESP" (138). He even says that in many cases one who is trying to explain the phenomena in normal terms may be "baffled" (115). However, he studiously avoids an examination of the actual evidence, which, he says, "would be tedious" (116). He instead constructs purely a priori arguments against the reality of extrasensory perception, concluding that, because the phenomena do not meet his "advance expectations" of how ESP would work if it were genuine, be need not take any of the evidence seriously (138). He is thereby able to conclude that parapsychology provides no credible evidence against a "materialism of the strictest order" (36). A more empirical thinker, I suspect, would not escape so unscathed. In any case, Humphrey's book does not provide a counterexample to my statement that few philosophers have evidently arrived at "considered judgments, after serious investigation, that good evidence is really lacking." He in effect admitted that the evidence was too good to be challenged on its own terms.

To the extent that this evidence is studied and is found convincing, materialism has another count against it.

7. Insofar as materialism rules out the possibility that our mind—our feelings, thoughts, desires, decisions—can have any power to exert causal influence other than that of the brain, it would also not be able to accommodate psychokinesis, the direct production by the psyche of changes in the extrasomatic world. And yet good evidence for psychokinesis exists.^[7] Insofar as psychokinesis is accepted, this will be another count against materialism's numerical identification of mind and brain. Of course, the inability of materialism here can be regarded as simply a part of its more general inadequacy to the fact, which we all presuppose in practice, that the mind has power of its own with which it acts back on the world (usually first of all its own body). But psychokinesis, when accepted, provides particularly dramatic evidence of this power, more difficult to interpret otherwise than more ordinary phenomena.

To summarize the problems unique to dualism and materialism: Those unique to dualism are all clearly due to the fact that the matter of which the body is composed is said to be different in kind from the mind. The problems unique to materialism can be seen to be due to the fact that, finding dualistic interactionism impossible, materialists gave up interactionism altogether, so that the mind is no longer thought to be a numerically distinct actuality (which could perhaps account for our unity of experience and of action and therefore our freedom, and which might be capable of the various forms of nonsensory perception arguably implied by various data and might even be capable of direct extrasomatic effects on the world). The problems unique to materialism are finally, therefore, traceable to the Cartesian intuition about matter, because it was the resulting problems of dualism that led to materialism's numerical equation of mind and brain. I turn now to problems, originating from the same root, that dualism and materialism have in common.

III. Problems Common to Dualism and Materialism

1. One problem shared by materialists and dualists alike—Campbell also lists this as one of the "embarrassing questions" his epiphenomenalism cannot answer (BM, 135)—is exactly where to draw the line between experiencing and nonexperiencing things. Given the continuity suggested by the evolutionary perspective and the increasing discovery of continuity where gaps once seemed to exist, any place that is chosen will seem arbitrary. Descartes's decision to draw the line between the human mind and the rest of nature, so that even his dog was said to be an insentient machine, has

seemed arbitrary to most. The most popular position has been to draw the line between those animals with and without a central nervous system, but where exactly is that? Having a central nervous system in borderline cases is a matter of degree.

This point is related to the problem of continuity raised in relation to dualism, but it is different. There the point was the philosophical one that any essential dualism seems to contradict the principle of continuity. Here I am raising the additional point that once this essential discontinuity has been accepted (whether by dualists or materialists), then there is the empirical problem of specifying a nonarbitrary place where that discontinuity allegedly occurs. Dualists and materialists typically avoid this problem: Because they "know" that the discontinuity must have happened somewhere in the evolutionary process, they are content to indicate only vaguely where that might have been. Searle, as I pointed out, thinks amoebas do not experience, but he says that he has no idea whether fleas, grasshoppers, crabs, and snails do (RM, 74, 81). Flanagan, while recognizing that scallops and paramecia receive information from and respond to their environments, seems confident that they do not feel or experience anything (CR, 35). McGinn supposes that consciousness must have arisen "when some of the fancier models of mollusc took up residence in the oceans, or when fish began to roam the depths. . . . [S]entience reverberated in the seas: awareness was born, quite late in the game" (PC, 44). But if these thinkers would try to specify exactly where this magical line is, the arbitrariness of the spot and the disagreement about it would make the problematic nature of the supposition more apparent. For example, evidence has been provided suggesting that bacteria make decisions (of a primitive sort) on the basis of memories (of a primitive sort).^[8] If our (soft-core) intuition is that experience and "life" go together, do we attribute experience to prokaryotic as well as eukaryotic life? Furthermore, the virus has some but not all of the features generally regarded as characteristic of living things. If we have included bacteria, do we attribute experience to viruses too? And if we do, then what about the remarkable DNA and RNA macromolecules, which certainly manifest organismic characteristics? And so on. On what basis could one claim to know, for example, that the apparent lack of complete determinism at the level of quantum physics does not reflect an element of spontaneity and thereby experience in subatomic events? Of course, Nagel is sociologically right: "If one travels too far down the phylogenetic tree, people gradually shed their faith that there is experience there at all" (MQ, 168). But that there actually is such a place does not belong to our hard-core common sense. It belongs at most to our merely soft-core common sense, the type empirical facts and rational reflection can modify.

2. A second problem is that of the Great Exception . This problem is generally raised by materialists against dualists. In fact, Michael E. Levin makes

this point "the main positive evidence for materialism," even more important than the problem of dualistic interaction. Appealing to the regulative principle of simplicity or parsimony, Levin says,

So far as we know, everything, except possibly the psychological states of sentient beings, is physical. . . . [Against dualism] it is simply more reasonable to think that the properties expressed by psychological predicates will turn out to be physical. Given that most of the universe is explicable physicalistically, the view which least multiplies independent principles is that the entire universe is explicable physicalistically. (MMBP, 87)

This point, however, can be turned back against materialism,^[*] especially in light of the agreement by Nagel, McGinn, Seager, and Searle, not to mention countless dualists, that conscious experience is not "explicable physicalistically." That is, would it not be strange that if human beings and many other animals (wherever one draws the line) are not fully explicable physicalistically (that is, in completely externalist categories), the rest of the universe would be? Because we know that we are not, the only way to avoid the problem of the Great Exception is to rethink the rest of the universe.

3. Yet another problem for dualists and materialists, because they both assume that experience arose at some late date in cosmic evolution, is to explain how the evolutionary process could have had the time—literally—to have gotten to the point at which time is said to have emerged . That is, as most of those who have thought much about it, such as the archmaterialist Adolf Grünbaum, have realized, time, in the real sense, presupposes experience: Without experience there would be no "now," and without a "now" there would be no distinction between past and future.^[9] (Incidentally, Newton's "absolute time" is usually thought to have been an exception to all relative views of time, according to which temporality is dependent on some kind of actual events; but, as Milic^[*] Capek^[*] has pointed out, this interpretation fails to recognize that Newton, quite a heretic, held a temporalistic view of deity: So-called absolute time reflected God's temporal experience.^[10] ) Assuming this necessary connection between time and experience, those who believe that experience arose historically must also hold that time arose at some point in the evolutionary process. This is the position of J. T. Fraser, who has probably (along with Capek) thought and written about time more than any other thinker in history. This position is reflected in the title of one of

[*] The fact that materialism has the problem of the Great Exception is illustrated by Herbert Feigl, one of the first formulators of materialistic identism. After speaking of the "identity" of matter and experience, Feigl clarified that "nothing in the least like a psyche is ascribed to lifeless matter." This implies that the language of psychology is applicable "only to an extremely small part of the world" ("Mind-Body, Not a Pseudoproblem," in Dimensions of Mind, ed. Sydney Hook [New York: New York University Press, 1960], 32, 33).

Fraser's books, The Genesis and Evolution of Time . The problem with this position, of course, is that it is circular, because evolution itself presupposes the existence of time. The paradox is expressed in Fraser's assertion that, although we cannot help thinking of several billion years passing between the big bang and the rise of life (which is where Fraser thinks time in the real sense of the word arose, evidently because that is where he dates the emergence of experience), we must say that all the events "prior" to the emergence of life were in truth all "contiguous with the instant of Creation" (GET, 132). Of course, the immediate response of most dualists and materialists will be that there must be something wrong with the assumption that time requires experience. (Indeed, the dominant position has been that time originated when there was sufficient organization of matter for entropic processes to begin; but that would only push the problem back: How would there have been time for cosmic evolution to proceed even to that stage?) To those who so respond, I would recommend reading Fraser and Grünbaum. If the argument is found convincing, what adjustment should be made? If you are a dualist or a materialist, are you more certain of the reality of time throughout the history of our universe or of the doctrine that experience as such emerged out of nonexperiencing things?

4. Although all of these problems are serious, perhaps the most serious problem shared by dualism and materialism is how the emergence of experience out of nonexperiencing entities is conceivable . I touched on this problem above in pointing to the problem of discontinuity for dualism and the problem of the meaning of the materialist's assertion that we, with our conscious experience, are entirely physical beings. I broached, especially with regard to Searle's position, the problem of the disanalogy between the alleged emergence of experience and the forms of emergence with which it is commonly compared. But the problem of this disanalogy runs even deeper. I will introduce this problem by citing a provocative statement by the materialist J. J. c. Smart: "How could a non-physical property or entity suddenly arise in the course of animal evolution? . . . [W] hat sort of chemical process could lead to the springing into existence of something non-physical? No enzyme can catalyze the production of a spook!" (M, 168f.). Although Smart directs this comment against dualism, it call be directed back at his own materialism. One can equally well ask, How can an enzyme catalyze the production of even the appearance of a spook? Regardless of whether one refers to a mind as a distinct actuality or does not (which is the only difference between dualists and materialists once the latter agree that mind, with its point of view and other subjective features, is irreducible to purely externalist description)—that is, whether it is called a real spook or only an apparent spook—how a mind could emerge out of enzymes or anything else assumed to have only external features is equally mysterious.

Nagel states the issue clearly in "What Is It Like to Be a Bat?" "One can-

not derive a pour soi from an en soi . . . . This gap is logically unbridgeable. If a bodiless god wanted to create a conscious being, he could not expect to do it by combining together in organic form a lot of particles with none but physical properties" (MQ, 189). A pour soi is something that exists for itself, being a subject of experience with feelings; an en soi, which is said to exist in itself but not for itself (being what Whitehead calls a "vacuous actuality"), is thought to have "none but physical [meaning purely external] properties." An en soi has only an "outside," having no features beyond those that are perceivable in principle by others and describable in externalistic language; it is hence nothing but an object (for others). A pour soi, by contrast, has an "inside," having features that are not externally perceivable by others and describable in externalistic terms; it is thus a subject (for itself). A subject or a pour soi, in other words, is something about which we can intelligibly ask, "What is it like to be one of those?" Strawson expresses this point by speaking of "the 'what-it's-like-ness' characteristics of experience" (MR, 45). Nagel is saying that it is inconceivable that a subject, something that it is like something to be, could arise out of mere objects naturally. This type of alleged emergence violates the principle of continuity and thereby implies a violation of naturalism.

Many philosophers, both dualists and materialists, have not yet seen the distinction between this alleged emergence and the nonproblematic types and thereby continue to think of them as analogous. Searle's analogies of experience with emergent properties such as solidity, transparency, and liquidity, considered earlier, provide an example. Searle, as I mentioned then, has described his position not as property dualism but as "property polyism," explaining that "there are lots of different kinds of higher-level properties of systems, and mental properties are among them" (MBWP, 228). This fits, of course, with Searle's attempt to portray "mental properties" as fully "ordinary,"just one more example of ordinary physical properties.

We have here a prime example of a category mistake . (I use this term, associated with Gilbert Ryle, to articulate a point quite opposed to Ryle's own philosophical behaviorism.)^[*] The alleged emergence of subjectivity out of pure objectivity has been said to be analogous to examples of emergence that are different in kind. All of the unproblematic forms of emergence refer to externalistic features, features of things as perceived from without, features of objects for subjects . But the alleged emergence of experience is not

[*] I refer here to Ryle's position in The Concept of Mind as it has usually been interpreted. As either a clarfication or a modification of his position in this early book, Ryle later eschewed behaviorism. For example, while still rejecting Cartesianism, he also said that "no account of Thinking of a Behaviourist coloration will do" (Gilbert Ryle, Collected Papers, vol. II: Collected Essays 1929–1968 [London: Hutchison, 1971], viii).

simply one more example of such emergence. It involves instead the alleged emergence of an "inside" from things that have only outsides. It does not involve the emergence of one more objective property for subjectivity to view, but the alleged emergence of subjectivity itself. Liquidity, solidity, and transparency are properties of things as experienced through our sensory organs, hence properties for others. Experience is not what we are for others but what we are for ourselves . Experience cannot be listed as one more "property" in a property polyism. It is in a category by itself. To suggest any analogy between experience itself and properties of other things as known through sensory experience is a category mistake of the most egregious kind.

In describing this confusion, which can be called the emergence category mistake, I am simply trying to drive home Nagel's point about faulty analogies. Although most contemporary commentators on the mind-body problem have accepted Nagel's point about the indispensability of including points of view in our world-pictures, many have evidently not seen the full force of his point about faulty analogies. Nagel says,

Every reductionist has his favorite analogy from modern science. It is most unlikely that any of these unrelated examples of successful reduction will shed light on the relation of mind to brain. But philosophers share the general human weakness for explanations of what is incomprehensible in terms suited for what is familiar and well understood, though entirely different. (MQ, 166)

In a passage only parenthetically cited earlier, Nagel says that "much obscurity has been shed on the [mind-body] problem by faulty analogies between the mental-physical relation and relations between the physical and other objective aspects of reality" (MQ, 202). Although Nagel does not use the term "category mistake," the thought is there.

Once alerted to the emergence category mistake, we can see it committed all over the place, by dualists and materialists alike. For example, Eccles, from a dualist standpoint, says, "Just as in biology there are new emergent properties of matter, so at the extreme level of organized complexity of the cerebral cortex, there arises still further emergence, namely the property of being associated with conscious experience" (FR, 173). Conscious experience is spoken of as simply one "further emergence," no different in kind from all others. Popper, coauthor with Eccles of The Self and Its Brain, almost shows recognition of the difference in kind, saying that the "incredible" invention of consciousness (out of wholly insentient neurons) "is much more incredible than, for example, the invention of flight" (SAB, 560). This recognition of incredibility does not, however, lead Popper to be incredulous.

The recognition that we have no real analogies for the alleged emergence of experience out of wholly insentient entities is what lies behind the

recent turn by McGinn to pessimism about ever finding a solution to the mind-body problem. McGinn is more explicit than most writers about the principle that no solution can require supernatural causation, even implicitly: "It is a condition of adequacy upon any account of the mind-body relation that it avoid assuming theism" (PC, 17n). But he also sees that no naturalistic explanation of the emergence of conscious experience out of "insensate matter" is possible (PC, 46). He is certain that this emergence is not "inherently miraculous" (PC, 2). He is equally certain, however, that no one can show that it is not.^[*] "The difficulty," says McGinn, "is one of principle: we have no understanding of how consciousness could emerge from an aggregation of non-conscious elements" (PC, 212). In that passage, he is discussing the possibility of strong AI;^[**] but he makes the same point with regard to the brain, saying that we also cannot "see how an entity constructed naturally from mere matter can be conscious" (PC, 205). Supernaturalism, accordingly, cannot be shown to be unnecessary.

McGinn compares the fact of conscious experience with the fact of complex organisms. Both could at one time have reasonably been pointed to as evidence for supernaturalism. The theory of evolution, however, "undermined the theism required by the creationist thesis." But no such explanation for consciousness is available or even on the horizon: "In the case of consciousness the appearance of miracle might also tempt us in the 'creationist' direction, with God required to perform the alchemy necessary to transform matter into experience. . . . We cannot, I think, refute this argument in the way we can the original creationist argument, namely by actually producing a non-miraculous explanatory theory" (PC, 17n). McGinn goes on to say, "But we can refute it by arguing that such a naturalistic theory must exist ." This, of course, is simply a confession of faith, not a real argument.

Another passage in which McGinn continues this comparison states the problem so clearly that it is worth quoting in full, partly because it embodies the recognition that it would be a category mistake to regard the emergence of sentience as analogous to the emergence of life, as if no greater violation of the principle of continuity were involved.

A basic continuity between the inorganic and the organic can . . . be demonstrated. No miraculous jump in the fortunes of the universe need be ruefully

[*] Strawson expresses a similar view: "Insofar as we are committed to naturalistic no-miracles materialism, we seem obliged to hold that the appearance of radical disconnection between experiential properties and nonexperiential properties is a kind of illusion. . . . But it won't go away, and it constitutes a vivid proof of the limitations on our understanding of reality" (MR, 75).

[**] "Strong AI," for Strong Artificial Intelligence, is a term for the position of those who believe that the "artificial intelligence" of computers could involve consciousness.

accepted. The Soft Rustle could therefore have occurred without the benefit of God's intervention. But in the case of consciousness we have no such understanding: we do not know how consciousness might have arisen by natural processes from antecedently existing material things. Somehow or other sentierice sprang from pulpy matter, giving matter an inner aspect, but we have no idea how this leap was propelled. . . . One is tempted, however reluctantly, to turn to divine assistance: for only a kind of miracle could produce this from that . It would take a supernatural magician to extract consciousness from matter, even living matter. Consciousness appears to introduce a sharp break in the natural order—a point at which scientific naturalism runs out of steam. (PC, 45)

McGinn adds in a note:

I do not know if anyone has ever tried to exploit consciousness to prove the existence of God,^[*] along the lines of the traditional Argument from Design, but in this post-Darwinian era it is an argument with more force than the usual one, through lack of an alternative theory. It is indeed difficult to see how consciousness could have arisen from insentient matter; it seems to need an injection from outside the physical realm. (PC, 45n)

This recognition that the alleged emergence of experience is sui generis, without analogy to anything else, is the basis for McGinn's agnostic physicalism. He compares his position to that of Locke's "idea that our Godgiven faculties do not equip us to fathom the deep truth about reality. Locke held . . . [that] only divine revelation could enable us to understand how 'perceptions' are produced in our minds by material objects" (PC, 4n).

Not all physicalists agree with McGinn's agnosticism, of course. Flanagan, in fact, offers his book as a response to McGinn's view that consciousness is "terminally mysterious" (CR, xi–xii). "The gap between the subjective and the objective," Flanagan argues, "is an epistemic gap, not an ontological gap" (CR, 221). McGinn, of course, agrees with that. But Flanagan goes on to argue, against Nagel and McGinn, that "we do understand how physicalism can be true":

It can be true if every mental event is realized in the brain. Those of us who believe that all mental events . . . are tokened in the brain do not believe that the theory that eventually explains how they are tokened will capture "the true character of the experiences" as experiences. The whole idea that the qualitative feel of some experience should reveal itself in a theoretical de-

[*] The answer to McGinn's query is Yes. His former Oxford colleague Richard Swinburne, for example, argues in chapter 9 of The Existence of God (Oxford: Clarendon, 1979) that God is needed to explain consciousness. Swinburne summarizes this argument in The Evolution of the Soul (Oxford: Clarendon, 1986): "The ability of God's action to explain the otherwise mysterious mind-body connection is just one more reason for postulating his existence. . . . God, being omnipotent, would have the power to produce a soul thus interacting" (198–99).

scription of how that experience is realized fails to acknowledge the abstract relation between any theory and the phenomena it accounts for. (CR, 93)

No language, Flanagan says, can "convey phenomenal feel!" (CR, 99). But surely McGinn and Nagel understand that . Their point is quite a different one—that a theory couched entirely in externalistic (third-person) language is the wrong kind of theory for referring to conscious states. Of course, to describe my wife, or my neighbor's cat, as "feeling pain" does not convey the phenomenal feel of their experience. But it at least indicates that I think there is some phenomenal feel (some "what-it's-like-ness") there, whereas the language of chemical processes or neuron firings does not. Flanagan has simply missed the main point. That this is so is shown by the fact that, even after reading Nagel and McGinn, he proceeds to commit the emergence category mistake:

If we operate with more sensible standards of intelligibility, several credible stories can already be told to explain how such things as sensory qualia supervene on certain patterns of neural activity. Just as ordinary water is H₂ O and is caused by H₂ O, so too are experiences of colors, tastes, and smells identical to and caused by activity patterns in certain brain pathways. (CR, 221)

His solution, in other words, is essentially the same as Searle's. It does not provide any basis for considering McGinn's agnosticism baseless. Strawson sees this point clearly:

As an acting materialist, I . . . assume that experiential phenomena are realized in the brain. . . . But this assumption doesn't solve any problems for materialists. . . . [W]hen we consider the brain as current physics and neurophysiology presents it to us, we are obliged to admit that we do not know how experience—experiential what-it's-like-ness—is or even could be realized in the brain. (MR, 81)

One temptation, whenever there is an alleged gulf between ontological unlikes to be spanned, is to assume that unlike things can be connected if we posit a sufficient number of intermediaries in between, so that the gulf need not be spanned in one leap. The transition can be made gradually. For example, many theological worldviews, positing an absolute difference between the ultimate divine reality (e.g., as totally timeless) and the world as we know it (e.g., as fully temporal), portray a hierarchy of intermediate realities. The one at the top is virtually godlike, sharing many divine attributes (e.g., being virtually timeless) and having only a hint of worldliness (e.g., being slightly temporal). The next one down has a little less divinity and a little more worldliness, and so on, until the nature of the lowest intermediary is so like that of the world that interaction between them does not seem too miraculous to swallow. That lowest intermediary can likewise

interact with its immediate superior, and so on, so that, finally, interaction between the divine and the world seems possible, in spite of their absolute heterogeneity.

Contemporary physicalists, being as human and therefore as unresistant to temptation as their forebears who operated out of a theological paradigm, sometimes employ a similar strategy. Dennett provides a good example. In his chapter "The Evolution of Consciousness," be begins by stating clearly the assumption that at one time there was no experience: "In the beginning, there were no reasons [i.e., final causes]; there were only [efficient] causes. Nothing had a purpose . . .; there was no teleology in the world at all. There was nothing that had interests" (CE, 173). Dennett's task is to get from that starting point to us, with our reasons, purposes, and interests. And, like McGinn, he makes clear that the explanation must not require, even implicitly, a supernatural injection somewhere in the process. He states repeatedly, even to the point of including a cartoon, that one of the steps in the explanation cannot say, in effect, "then a miracle occurs" (CE, 38, 239, 255, 455). Unlike McGinn, however, he thinks that an explanation devoid of miracle can be given. Let us see.

After millennia, Dennett says, simple replicatots emerged. Did they have purposes? Dennett's answer is ambiguous: "While they had no inkling of their interests, and perhaps properly speaking had no interests, we . . . can nonarbitrarily assign them certain interests—generated by their defining 'interest' in self-duplication" (CE, 173). Has he said that things with final causes (reasons, interests, purposes) have (miraculously) emerged out of things that operate mechanically, by efficient causes alone? No, at least not yet: In spite of his weasel word "perhaps," he seems to grant that the replicators "properly speaking had no interests," and he uses scare quotes in speaking of their assigned "defining 'interest' in self-duplication." Dennett explains his meaning, furthermore, by saying, "If these simple replicators are to survive and replicate . . ., their environment must meet certain conditions" (CE, 173). No miracle, evidently, has occurred. But then he continues:

Put more anthropomorphically, if these simple replicators want to continue to replicate, they should hope and strive for various things; they should avoid the 'bad' things and seek the 'good' things. When an entity arrives on the scene capable of behavior that staves off, however primitively, its own dissolution and decomposition, it brings with it into the world its 'good.' That is to say, it creates a point of view . (CE, 173–74; emphases added)

Although Dennett begins this statement with the disarming assurance that it is "put anthropomorphically," which seems to mean that it is a purely as-if account, by the end it seems that the miracle has really occurred: Replicators really do, "however primitively," have a point of view and hence interests,

purposes. Teleology has emerged out of pure mechanism. Once Dennett has, thanks to his quiet little miracle, crossed this divide, he can account for human consciousness in terms of greatly increased complexity alone: "The point of view of conscious observers is . . . a sophisticated descendant of the primordial points of view of the first replicators who divided their worlds into good and bad" (CE, 176). But even if the miracle occurs long before the rise of human consciousness, it is no less a miracle (unless, of course, Dennett is giving a purely externalist, and thereby wholly inadequate, account of a "point of view").

Humphrey provides a similar explanation based on tiny steps. He says, "Before life emerged . . . there were presumably no minds of any kind at all. It follows that four thousand million years ago the world was totally unexperienced. . . . [E]verything . . . in nature was insentient" (HM, 16, 217). Humphrey's task is to get from there to the existence of animals, for whom, "whether at the level of an amoeba or an elephant," there is a boundary wall between "me" and "not-me" (HM, 18). His story describes "how present-day sensory activities could have developed step-by-step from primitive beginnings, starting with a local 'wriggle of acceptance or rejection' in response to stimulation at the body surface" (HM, 180f.). (Note how externalist language ["wriggle"] and internalist language ["acceptance or rejection"] are combined, just as in theological accounts the intermediaries embodied both divine and worldly features.) What occurred first, according to Humphrey, was a slight degree of "sensitivity" (HM, 18), which at first could be understood in purely externalist categories. But then natural selection selected for this sensitivity, so that it became more sophisticated. Pretty soon, an inner, phenomenal life of sorts emerged: "The phenomenology of sensory experiences came first. Before there were any other kinds of phenomena there were 'raw sensations'—tastes, smells, tickles, pains, sensations of warmth, of light, of sound and so on" (HM, 21). At some point, then, "certain events were being responded to as good or bad . . ., as of significance to 'me'" (HM, 19). Subjectivity had emerged out of pure objectivity.

Humphrey thinks this explanation has answered McGinn's problem: "A seeming miracle? No, as close to a real miracle as anything that ever happened. The twist may be that it takes only a relatively simple scientific theory to explain it" (HM, 219). The theory is indeed simple; but it is deceptively simple. The deception is that an infinite step has been made to seem a tiny one.^[*] No matter how tiny, far back, and innocuous one tries to make

[*] After writing this, I learned that McGinn had addressed this issue in an earlier book, The Character of Mind . There he argues that although life admits of borderline cases, such as bacteria and organic molecules, which can be considered partly living and partly not, this is not the case with experience: Either there is "something 'inner,' some way the world appears to the creature," or there is not. Accordingly, whereas with life "we have to do with a gradual transition from the plainly inanimate to the indisputably living," with regard to experience "we cannot take such a gradualist view, admitting the existence of intermediate stages." The emergence of experience "must rather be compared to a sudden switching on of a light, narrow as the original shaft must have been." We must, therefore, think of the lowliest minds "as consisting in (so to speak) a small speck of [experience] quite definitely possessed, not in the partial possession of something admitting of degrees" (CM, 14). Although McGinn couches his argument in terms of "consciousness," I have phrased it in terms of "experience," because I do not, unlike McGinn, equate the two, and because I, thinking of consciousness as a very high-level type of experience, do believe that it can emerge gradually. With regard to the fimdamental issue, however, which is that an entity must either have an inner aspect or not—regardless of what term is used for this inner aspect—he seems absolutely right. The problem that McGinn's clarity on this point creates for a materialist, of course, is precisely the problem that Dennett and Humphrey are trying to circumvent: If an inner aspect suddenly emerges in things that previously had been completely devoid of any such aspect, the miraculous character of this alleged emergence, and thus its seeming impossibility within a naturalistic framework, is driven home.

it, however, it is still a miracle. For McGinn, the task is "to take the magic out of the link between consciousness and the brain" (PC, 2). This is because, for him, the question "How is it possible for conscious states to depend on brain states?" is assumed to be identical (see my list of problems at the outset of chapter 1) with the question "How could the aggregation of millions of individually insentient neurons generate subjective awareness?" (PC, 1). But if we assume that experience emerged earlier, perhaps with single-celled organisms such as amoebas, we would have the "experience-organelle problem," that is, the problem of how "to take the magic out of the link between" an amoeba's experience and its organelles. In other words, magic is implicit regardless of the level at which the first emergence of experience is said to occur.

This concern about magic, incidentally, is not new. At a conference on evolution in 1974, the great evolutionist Sewall Wright, a panexperientialist, declared in his paper, "Emergence of mind from no mind at all is sheer magic" (PS, 82). Theodosius Dobzhansky, an equally great evolutionist, replied, "Then I believe in magic!" Dobzhansky's reply, delivered in his thick Russian accent, brought the house down; but it also brought out the apparent trilemma: panexperientialism, supernatural intervention, or (Dobzhansky's choice) natural magic. McGinn, however, believes that there is another alternative. But this alternative is nothing but physicalist piety.

McGinn's "nonconstructive solution" (PC, 31) to the mind-body problem is simply to assert that there must be something purely natural about consciousness and therefore the brain that accounts for the relation between them, but that we are so constructed that we will never be able to know what this is. This must be the truth, he says, or else naturalism would be threatened: "The radical emergence of mind from matter has to be

epistemic only, on pain of accepting inexplicable miracles in the world" (PC, 6). Then, having thereby confessed that his position is based on paradigmatic and wishful-and-fearful thinking, he proceeds to speak as if he knows it to be true. For example, having stated that we cannot refute a creationist account of consciousness "by actually producing a non-miraculous explanatory theory," he says that "we can refute it by arguing that such a naturalistic theory must exist " (PC, 17n). How does this response differ from that of the traditional Christian theologian who, speaking "from faith to faith" to fellow sharers in the orthodox paradigm, argues that although we cannot actually refute the charge of self-contradictory nonsense by producing an account of how Jesus was fully divine and yet fully human, we can refute the charge by arguing that such an account must exist? McGinn begins his book by telling us that his solution has given him a great sense of relief, one that he tries to induce in his readership (PC, vii). But this sense of relief rests entirely on faith. "The philosophical problem," he says, "arises from the sense that we are compelled to accept that nature contains miracles." But, he continues, "we do not need to accept this: we can rest secure in the knowledge that some (unknowable) property of the brain makes everything fall into place" (PC, 18). (Believers commonly gain a sense of security by thinking of their faith as knowledge.) McGinn says that a proposal is adequate to the degree that it avoids the seeming necessity to choose between eliminativism and accepting a miracle, and proclaims his own proposal a success because "it allows us to resist the postulation of miracles" (PC, 18n). All it really does, however, is declare that no miracle occurred while giving not a hint as to how the feat was accomplished without one. Having pointed out that we do not know how to reduce consciousness to the brain, McGinn explains: "We need to distinguish being able to give a reduction from knowing that a reduction exists " (PC, 31n). But McGinn's position here is similar to that of reductionists who base their present confidence in physicalism on promissory notes about the future glories of neuroscience—except that McGinn gives no such note, saying instead that we will never be able actually to give a reductionistic explanation.

Indeed, in calling his position "agnostic realism" (PC, 120), McGinn recognizes that he is similar to an "agnostic theist" (PC, 119). In answering the question as to how he can so enthusiastically believe in something about whose nature he can say nothing, he claims that we "can know that something exists without knowing its nature. We can assert that a gap is filled without being able to say how it is filled" (PC, 119). His comparison with agnostic theism is apt. One of the main reasons for rejecting theism is the problem of evil: Why, critics ask, if God is both omnipotent and good, is there such horrendous evil in the world? One of the most popular answers today, employing a combination of claimed knowledge and agnosticism analogous to McGinn's, goes something like this: "Because we know that

God exists, and because we know that God is perfect in wisdom and goodness, we can know that God must have a perfectly good reason for allowing all these evils. We can know this without having the slightest inkling as to what that reason might be. In fact, we can see that it is perfectly reasonable, given our finite natures, that we should not be able to figure out what that reason is." This kind of argument, of course, is viciously circular: The socalled knowledge is based entirely on faith—a faith, in fact, that is insulated from the possibility of falsification.

In spite of all his talk of knowledge, McGinn does finally admit that it is something less. Of his naturalism, which he describes as "the thesis, metaphysical in character, that nothing that happens in nature is inherently anomalous, God-driven, an abrogation of basic laws," he says: "It is, I suppose, an article of metaphysical faith" (PC, 87).^[*] One can sympathize with Flanagan and others for not finding the relief that McGinn had hoped his "solution" would provide. Blind faith is blind faith whether it be of the supernaturalistic or the naturalistic variety, and blind faith does not provide a secure resting place for most intellectuals.

I am not at all, given my regulative principles, objecting to McGinn's naturalistic faith; I share it myself. But it should be a faith seeking understanding. Just as I hold that a theistic faith is not reasonable unless it can provide an intelligible and persuasive theodicy, which for me means rejecting supernaturalism in favor of a version of naturalistic theism,^[11] so any form of naturalism is precarious insofar as it cannot provide an intelligible and persuasive account of the mind-body relation. This account must be more than simply a restatement of faith.

McGinn's position is meant to be based, to be sure, not simply on faith but also on an argument. This argument, however, involves a fallacy, which can be dubbed the actual-possible fallacy . The best way to show that something is possible is to show that it is actual. Actually going to the moon settled all the doubts as to whether it was possible. That part of the argument is unassailable. Insofar as we take a particular paradigm for granted, however, we are apt to insinuate theoretical inferences into our statement of empirical fact. Some theists, for example, have answered the question as to whether an individual could be both fully human and fully divine by pointing to Jesus: Because it happened, it is obviously possible.^[**] Likewise,

[*] Strawson sees this point more consistently. "Belief in the truth of materialism is a matter of faith," he says, adding: "My faith, like that of many other materialists, consists in a bundle of connected and unverifiable beliefs" (MR, 43, 105).

[**] Millard J. Erickson provides an example: "We sometimes approach the incarnation the wrong way. We define deity and humanity abstractly and then say, 'They could not possibly fit together.' . . . If, however, we begin with the reality of the incarnation in Jesus Christ, we . . . recognize that whatever [the two natures] are, they are not incompatible, for they once did coexist in one person. And what is actual is of course possible" (Christian Theology [Grand Rapids, Mich.: Baker, 1985], 737).

in response to questions as to whether it is possible for a purely spiritual being to act upon a material world (an aspect of the mind-body problem writ large), one would commit the actual-possible fallacy if one pointed to the world as God's creation as proof. To a physicalist, these theistic examples of the fallacy are obvious, but physicalists are then likely to commit their own versions. For example, McGinn says, "It is just as hard to see how an entity constructed naturally from mere matter can be conscious as it is to see how an intentionally created material object can be. But we know that the former is possible, because we have seen it done" (PC, 205).

This argument may seem unobjectionable until we recall the theoretical idea packed into the term "mere matter," which entails that the "grey matter" of the brain is composed entirely of "individually insentient neurons" (PC, 1). It is one thing to say that we know that it is possible for conscious states to arise out of a brain, because it actually occurs. It is something entirely different to say that we know that it is possible for conscious states to arise out of a brain composed of neurons that are individually insentient, because it has actually happened. This we do not know; it is pure supposition. This actual-possible fallacy, I suspect, provides the basis for McGinn's twofold conclusion that the relation between the brain and consciousness is a purely naturalistic relation, meaning one consistent with physicalism, but that we will never be able to understand this relation.

The actual-possible fallacy with regard to our question is not at all rare; in fact, it seems to be committed by a good share of dualists and materialists alike. It goes back to Descartes, who was persistently questioned by Princess Elisabeth as to how mind and body, if they were totally different, could interact. After it was clear that reference to the pineal gland at best answered only the question of where, not that of how, Descartes finally admitted that he did not know how. He added that this is not very important, however, because it is empirically obvious that they do interact.^[12] It evidently did not occur to him that this empirical fact counted against his own characterization of the body, or the mind, or both. This same fallacy is committed by the contemporary Cartesian H. D. Lewis. In response to John Passmore's question, cited earlier, as to how minds, capable of only rational persuasion, can affect bodies, capable only of pushing and being pushed, Lewis says,

It seems quite evident that my putposing to put on my spectacles or wave my hand is not itself a physical state or process. . . . But if we find that this nonspatial purpose does in fact normally bring about a physical change, how can we possibly question this just because it defies further explanation? . . . [W]e must accept [the world] however remarkable it may seem to be in some respects. (ES, 34)

Lewis, like Descartes, fails to distinguish between the "evident" facts and his theoretical construal of them. In response to Bernard Williams, Lewis says,

We may indeed admit that there is "something deeply mysterious about the interaction which Descartes's theory required between two items of totally disparate natures." . . . But there is a limit to explanation and a point where we just have to accept things as we find them to be. . . . [W]ould it not be better for philosophers, rather than trying to explain away or discredit extraordinary facts of experience, to stop and wonder at them and their possible further implications? (ES, 38f.)

Lewis, having so deeply internalized the dualistic paradigm, evidently fails to consider that there may be nothing "extra-ordinary" about the facts of experience at issue, but that it is only the dualistic construal that makes them seem so, and that what at least some critics are doing is discrediting that construal, not the facts. Nor does he evidently consider that, having accepted interaction of mind and body, the "possible further implications" are that the Cartesian construal of mind or body or both is erroneous.

I return now to materialistic examples of the actual-possible fallacy, beginning with one provided by Dennett: "How could a complicated slew of electrochemical interactions between billions of neurons amount to conscious experience? And yet we readily imagine human beings to be conscious, even if we can't imagine how this could be" (CE, 433). Ergo, the conclusion is, consciousness has arisen out of insentient neurons. Humphrey makes a similar move, contrasting the problem of turning water into wine, which he considers unsolvable in principle, with the problem "of getting consciousness into the brain." The "interesting difference" between them, he says, is that "while the former has never been known to occur, the latter occurs all the time" (HM, 6). The latter part of that statement is certainly true, but Humphrey assumes that this fact carries with it the fact that consciousness has arisen out of a "foam of insensate matter." (It is, incidentally, puzzling that both Dennett and Humphrey, after having given an evolutionary sketch according to which an elementary form of sentience is present in amoebas, which are single-celled organisms, persist in thinking of neurons, which are single-celled organisms, as insentient.) Humphrey does not know, however, that consciousness has arisen out of brains composed of insentient neurons; he only believes this. What he knows, at best, is that conscious experience has arisen out of brains.

By recognizing the actual-possible fallacy for what it is, we can overcome the widespread assumption that we know that conscious experience has somehow emerged out of insentient matter, an assumption on the basis of which thinkers, both philosophic and scientific, set themselves the impossible task of trying to figure out where and how this occurred. We can

thereby realize more clearly that the problem is simply that of trying to figure out how conscious experience arose out of brains, a problem whose solution may require us to modify our inherited assumptions about the components of those brains.

The main purpose of this chapter has not been simply to rehearse the problems of both dualism and materialism; those are known well enough (although it is usually not appreciated how many problems they have in common). The main purpose has been to show that all of those problems are rooted in the same intuition: the Cartesian intuition about matter. As Whitehead points out, almost all schools of thought have "admitted the Cartesian analysis of the ultimate elements of nature" (SMW, 145). This was an analysis that completely excluded mind from nature. What is suggested, accordingly, is that a solution to the mind-body problem may need to be based on a philosophy that would regard mind as fully natural, thereby rejecting the Cartesian intuition about bodies. This is the topic of the next chapter.