The imitation man can know sea and sky are alike in color, and even call them "blue"...When he looks at sea or sky he forms the belief that what he is looking at has the color which he has been taught to call "blue". Yet the imiatation man does not see sea or sky as blue. He is not able to enjoy their color, for they do not appear as colored to him. Similarly, he can tell when his finger is burned or crushed, and have a powerful drive to eliminate the condition by which he knows this. Yet he cannot suffer. --Keith Campbell

The causal doctrine covers well the description of mentality by one observing and explaining his fellow men. But the theory leaves out, to put it briefly, what waking life is like to him who is living it. -- Keith Campbell

In his paper "A Critique of Central-State Materialism", Keith Campbell unambiguously proclaims his belief in what he calls a new epiphenomenalism or, equivalently, Central-State Materialism Plus. That is to say, while observable behavior in humans must be caused by strictly physical properties of certain mental states (which must be physical in nature), he postulates the existence of certain phenomenal properties of these mental states that constitute the way they seem to us as we are experiencing them. These properties are non-physical, which means they cannot be described in physical terms. Knowing everything physical there is to know about mental states in general would not tell you what it is like to actually see red, or feel pain, or taste a lemon. One knows what it is like to feel pain through what Campbell calls an "imperfect apprehension" (Campbell 336) of the non-physical, phenomenal property that goes with the mental state of being in pain. These phenomenal properties are also known as qualia in the literature. They are the characteristic and subjective way that the universe seems to be to those who perceive it. In fact, it could be said that many, if not most, modern objections to physicalism boil down to some kind of argument involving qualia, or the belief that science, as we know it today, cannot explain the subjective first person perspective. Any serious defense of physicalism must meet these objections. As a small step in this direction, I will attempt to show that the arguments presented in Campbell's Critique suffer from a form of "begging the question" and an incorrect but extremely common misunderstanding of familiar terms such as "feel" and "experience."

Central-State Materialism, according to Campbell, is the theory that "descriptions of mental events, states, and processes are descriptions of inner conditions insofar as they are, directly or indirectly, causally efficacious in the behavior of an organism" (Campbell 332). That is to say, the mind has an "inner" existence as something physical in the organism and is the cause of observable behavior in that organism. It is this central role of mind as the cause of human behavior that leads Campbell to dub the theory the "Causal Theory of Mind" throughout his paper. In his explanation, he gives us a useful example of the type of analysis that central-state materialists are wont to do :

"I am aware that my finger has been burned" is analyzed as "As a result of having been burned on the finger, I have entered a new inner state apt to produce behavior wherein I discriminate the burned finger from others which are not burned." In the discriminating behavior I not only favor the correct finger, I favor it in the burn-soothing way. That is, I give verbal and active expression to the belief that my finger has been burned (Campbell 333)

Campbell is critical of this analysis, and believes it leaves something essential out of its explanations. To illustrate his point, he asks us to imagine a creature that distinguished different sensations on the basis of throb frequencies. Fast throbs could mean a burning pain, slow throbs could mean a sour taste, and so on. This creature would still be aware of different sensations, he would still be able to tell when his finger is burned or not burned, but he wouldn't really be in pain (assuming, in this case, that throbs are not painful) and "the entire episode would not be one which hurt in the slightest" (Campbell 333). This creature, when it burns its finger, doesn't feel pain, but has mental states that are comparable to those of a man who, when he burns his finger, does feel pain. Hence, central-state materialism leaves something out of the analysis.

Campbell's argument certainly seems convincing on the face of it, but its flaws become evident when one invokes the concept of functionalism. According to "Functionalism, Qualia, and Intentionality" by the Churchlands, functionalism is "the thesis that the essence of our psychological states resides in the abstract causal roles they play in a complex economy of internal states mediating environmental inputs and behavioral outputs" (Churchland 349). In other words, understanding the mental state of pain means getting a handle on its causal role. So what does pain do for creatures who have the capacity to detect it? What is pain, functionally defined? At least one aspect of pain seems to be glaringly missing from Campbell's analysis, namely the aspect of unpleasantness. Pain, almost by definition, is a sensation that creatures wish to be rid of. To put it in terms of central-state analysis, the sensation of pain causes a mental state corresponding to the awareness of that pain, which begets another mental state corresponding to the desire that the awareness should cease at once. To look at it another way, part of what it means to be in pain, is to have a desire that the state of being in pain should go away. Under this analysis, so long as the proper causal chain is in place, from the awareness of pain to the desire that the awareness should cease, the above creature, when he senses a burning finger via his throbs, is in as much pain as anyone.

Although effective, this is not, perhaps, a very original argument. In fact, Campbell anticipates it later in his paper:

Pains are unpleasant. We prefer not to have them. We often think that we prefer not to have them because they are hurtful. But perhaps this is a mistake. Perhaps their hurtfulness is precisely that we desire to be rid of them (Campbell 334).

In response, he asks us to imagine two people (let us call them Jack and Jill, though Campbell does not) for whom "crushing pains" and "burning pains" are switched but who are otherwise behaviorally identical. Both these people exclaim "Ouch! I burned my finger!" when their fingers touch something hot, and they both exclaim "Ouch! My finger has been crushed!" when their finger gets trapped under something heavy. Furthermore, whether their fingers are being crushed or burned, both people enter a state of pain; that is, they both become aware that their fingers are being crushed or burned and they both gain an intense desire to make that awareness cease as soon as possible. The difference is that when Jack burns his finger, he feels as Jill would feel when her fingers are being crushed, and vice versa. The two people would be functionally equivalent, in the sense that their pains perform identical functions, but their qualia, the subjective nature of their pain, would differ (would, in fact, be inverted). Hence functionalism (and physicalism as well, inasmuch as functionalism is a form of physicalism) leaves something out.

At the risk of sidestepping the real issue, it can be argued that physicalism, cleverly disguised as functionalism, does admit the possibility, perhaps even the likelihood, of inverted or shuffled mental state implementations (note that I did not use the term "qualia".) Functionalism has never denied that a particular mental state could be instantiated in vastly different ways for different people. This is, in fact, the main idea behind functionalism. Mental states should be defined functionally, in terms of what they do, instead of physically, in terms of what they are made of. In the above example, Jack's sensation of burning pain simply has a different physical structure than Jill's sensation of burning pain. This position is outlined very plausibly, for instance, in the Churchland essay "Functionalism, Qualia and Intentionality":

Given the physiological and chemical variety we find in the nervous systems of the many animals that feel pain, it appears very unlikely that their pain states have a common physical nature underlying their common functional nature (Churchland 351).

This discussion, of course, sidesteps the issue that Campbell was really trying to get at, namely that the subjective character of each person's burning pain sensation was different for both, and that functionalism overlooked this important fact. I suspect that Campbell would be the first to acknowledge the possibility that different subjective experiences of burning one's finger could, in fact, be tied to the different ways in which the physical pain response is set up in individual people. But explaining how those physical pain responses lead to the subjective experiencing of pain is something the Churchland essay glosses over. In fact, the Churchlands seem, at times, to support Campbell's view that there really is an intrinsic way that things seem to be to a conscious being. In their discussion on the "opaque discrimination" of a 60 Hz spiking frequency in neural pathways, for example, they say the following:

The spiking frequency of the impulse in a certain neural pathway need not prompt the non- inferential belief, "My pain has a spiking frequency of 60Hz"; it may prompt only the belief, "My pain has a searing quality." But withal, the property you opaquely distinguish as "searingness" may be precisely the property of having 60Hz as a spiking frequency (Churchland 354).

It seems that the Churchlands are, in fact, capitulating to Campbell's contention that non-physical phenomenal properties, the property of the way things seem to us, have a real existence. Is functionalism and physicalism, therefore, necessarily false?

The answer is no. The problem is that Campbell, and even the Churchlands, show a misunderstanding of what it means to "feel" a burning finger. A burning pain in Jack's finger does indeed feel painful and burning, but not because the sensation possesses some intrinsic, non-physical, "burning" phenomenal property that we, as humans, intuit. Rather, it feels painful because, as argued before, the awareness of the sensation sets up a desire that the awareness should cease, and, more to the point, it feels burning because it is the sensation Jack has learned to pick out and associate with the act of putting his finger on something hot. Most importantly, Jack's pain will always feel as burning as ever, even though it may have a physical basis identical to Jill's crushing pain mechanism, precisely because it is the sensation he has learned, from birth, to associate with putting his finger on a flame or stove. The same argument goes for Jill, and her burning pain. Her burning pain may be implemented with the physiology that implements the crushing pain in Jack, but that is completely irrelevant. She has learned to associate the sensation she feels with the act of putting her finger on something hot. To say this another way, Jack's burning pain feels exactly as burning as Jill's burning pain, though they are implemented in opposite fashions. The functionalist analysis has left nothing out.

It may help to make the example more complicated. Imagine that when Jack puts his finger on something hot, the sensation not only sets up a state of mind where he is aware of the burning sensation and where he wants it to stop, but it also sets up a whole cascade of effects that do justice to the true complexity of a human being. Putting his finger on a flame subconsciously reminds Jack of the time he burned his finger when he was six and his mother took him to the doctor, of the time when he was just a baby and his father screamed at him for crawling to close to the stove, and of the hundreds of other times that he has put his finger onto something hot. Far from becoming aware of some mystical phenomenal property of the pain state, Jack merely, according to Dennet, enters into "various discriminative states, underlying a host of innate dispositions and learned habits of varying complexity" (Dennet 143). This is the experience of pain that Jack goes through.

Suppose that Jack undergoes surgery to switch his burning pain detection physiology with his crushing pain detection physiology. Such a maneuver is possible, at least in principle. Jack now feels as if his finger is being crushed every time he burns it and, unsurprisingly, it still hurts. We can still perform a causal analysis of the painful burning sensation, as we did before. An epiphenomenalist, like Campbell, might suppose that this causal analysis might miss the difference between Jack's pre-operation pain and Jack's post-operation pain (one is burning, the other crushing). Is this really the case? Is the functional analysis of Jack's burning pain the same in both cases? No, it isn't. The previously mentioned cascade of effects will be completely different. Whenever Jack now burns his finger he will no longer subconsciously remember the time when his father yelled at him for crawling too close to the stove but, instead, he will remember the time, when he was eight, that his sister closed the car door on his hand. Notice that the concept of qualia doesn't enter into the picture at all, and yet functionalism (with a physicalist underpinning) is still able to detect the difference between the crushing pains and the burning pains.

Suppose, furthermore, that Jack underwent another operation to change all his memories of which sensation constituted a burning pain and which sensation constituted a crushing pain. After such an operation, Jack would act and feel exactly the same as he did all his life. Is he still sensing crushing pains when he is supposed to be sensing burning pains? Absolutely not. His burning pain is precisely as burning as it was before the first operation, even though it is now implemented with crushing pain physiology. He would not detect any difference at all. The functionalist analysis would say that post-operation Jack is functionally equivalent to pre-operation Jack. It might be tempting to point out that Jack now somehow experiences his burning pains differently than how he did before, and that the functionalist analysis leaves that fact out. Looked at from a certain angle, this statement is partially true. By its nature, functionalism is fairly indifferent to physiological considerations. A physicalist analysis, however, would root out the difference. What is not true is the notion that Jack now experiences a different kind of burning pain than he did before. Jack would remember the pain feeling exactly as it always felt. Note again that, in this entire discussion, the concept of qualia was never brought up. If anything, the above discussion provides some evidence that phenomenal properties aren't real properties after all. Using traditional language, the mental state of Jack's that used to posses a phenomenal property of crushing pain now possesses a phenomenal property of burning pain. So what happened? Where did the "crushing pain" go? Where did the "burning pain" come from? These are tough questions to answer if one accepts the existence of qualia in the literal sense.

Dennet, in his paper "Instead of Qualia", insists that the above type of analysis is enough to explain everything there is to know about human experience usually attributed to qualia, and the illusion of why indeed it seems as if mental states have some sort of non-physical, inexpressible properties to which we alone are privy. In his own words:

The dispositional properties of those discriminative states already suffice to explain all the effects: the effects on both peripheral behavior (saying "Red!", stepping the brake, etc.) and "internal" behavior (judging "Red!", seeing something as red, reacting with uneasiness or displeasure if, say, red things upset one) (Dennet 143).

In other words, we don't need qualia to explain the "experience" of burning pain, whether it's the external behavior of saying "Ouch!" or the internal behavior of knowing you're in pain. This, of course, does not prove that phenomenal properties don't exist, but it does show that postulating their existence is a superfluous step. By the principle of Occum's Razor which says, essentially, that one should not make things more complicated then they need to be, the concept of phenomenal properties become useless.

The above analysis settles the issue of "inverted qualia", or "juggled phenomenal properties". Simply put, there are no phenomenal properties, as Campbell uses the term. As it happens, the above discussion also produces a convenient way to fight the notion of "absent qualia", as embodied in Campbell's story of the imitation man. Such a creature is behaviorally indistinguishable from what can presumably be called a genuine man, but his mental states do not posses any so-called phenomenal properties. That is to say, such a creature does not truly feel the hurtfulness of pain (even though he may say "Ouch!" when his finger is burned), and does not truly taste the sourness of lemon (even though he may pucker his lips when a piece of the fruit is placed in his mouth). The imitation man can tell the difference between different sensations but the beliefs "just pop into his head" (Campbell 333). He does not experience the phenomenal properties associated with them.

Campbell intended his imitation man to be a form of modal argument. In other words, he thought that such a being was logically possible, that it was possible to conceive of two creatures completely alike in every physical respect, except that the mental states of one possessed phenomenal properties which were "imperfectly apprehended" by the creature doing the experiencing, while the mental states of the other had no phenomenal properties at all. Since such a creature is logically possible, so the argument goes, then physicalism has to leave something out. One very easy way to dismiss this claim is by simply denying the existence of phenomenal properties in the first place. The imitation man is a logical possibility only if one already believes in the existence of phenomenal properties. The imitation man argument, therefore, suffers from fallacious reasoning. It begs the question. As discussed before, it is entirely possible to explain everything there is to explain about external and internal human behaviors by the dispositional properties of discriminative states, as explained by Dennet.

If one does not believe that phenomenal properties are real, then it suddenly becomes fairly obvious that there is no difference whatsoever between, say, an imitation Jack and a genuine Jack. Insofar as they are physically identical, they are one in the same. This can be further illustrated by thinking, once again, of a burning sensation. Suppose that genuine Jack burns his finger (again). How does Jack know that he has burned his finger? How does he know that it is painful in consequence? The answer, quite simply, is that he "just does". To put it in terms of the causal theory of mind, the hot object causes Jack's mind to enter into a mental state corresponding to the awareness that his finger has been burned, which in turns causes a whole host of other physical effects. If one does not believe in phenomenal properties, the story pretty much ends there. More to the point, imitation Jack learns about his burns is precisely the same way. As Campbell points out: "[The imitation man] spontaneously gains a new belief, it just 'pops into his head' that he has burned his finger or broken his toe, as the case may be" (Campbell 333). If one were to try and argue that imitation Jack wouldn't have the same experience of finger burning as genuine Jack, then I have to ask what is meant by the term "experience". When genuine Jack burns his finger, he is reminded of the time his father screamed at him for crawling to close to the stove. Imitation Jack is reminded of exactly the same thing. Genuine Jack yelps "Ouch!" and sucks his finger, and so does imitation Jack. Remembering that there are no such thing as phenomenal properties, what part of the episode is genuine Jack experiencing, that imitation Jack is not? The answer, of course, is no part. If imitation Jack were not going through the same mental states as genuine Jack, then he wouldn't be a very good imitation. That is to say, imitation Jack would be functionally different from genuine Jack, and hence they would be acting differently for functionalist or physicalist reasons.

In conclusion, the arguments presented by Campbell, although convincing on the face, do not show that central-state materialism in its current form is incomplete. It is possible to explain all the effects that human beings experience when they perceive red or feel pain or taste lemon, including the internal judgments they express, simply by appealing to physical effects and physical causes. Specifically, the ideas of inverted phenomenal properties (told through the story of Jack and Jill) and absent phenomenal properties (told through the story of the imitation man) only work if you assume the existence of phenomenal properties in the first place. As I showed, however, neither qualia or phenomenal properties are necessary in the explanation of human experience, though, of course, one cannot prove with absolute certainty that they don't exist. By the principle of Occum's Razor, we should throw them out of our explanations entirely, or at least realize that when we talk of them, we do not mean literal, real, properties of mental states that are, on principal, irreducible to physics, but rather the set of internal dispositional states that external stimulus are apt to produce in humans.


Campbell, Keith. "A Critique of Central-State Materialism." The Place of Mind. Ed. BrianCooney. Belmont: Wadsworth, 2000

Churchland, Patricia and Paul M. Churchland. "Functionalism, Qualia, and Intentionality" The Place of Mind. Ed. Brian Cooney. Belmont: Wadsworth, 2000

Dennet, Daniel. "Instead of Qualia" Brianchildren. Ed. Daniel Dennet. Cambridge: MIT Press, 1998