No, not those zombies! They aren’t merely possible, they are actual, as you can see.
Rather, our topic today is philosophical zombies, beings that are physically identical to us but without conscious experience. Thus, a zombie version of yourself, for instance, would be atom-for-atom identical to you. It would share all of your behavioral dispositions: it would walk like you, talk like you, have the same tendencies to be angry, happy, or sad as you, report any information that you are able to report, and perform any tasks that you are able to perform. It would also remember everything that you can remember, know everything that you know, and it would have all the same politics and cultural attitudes and biases as you. At least, it would do all these things as near as we could tell. It would be behaviorally and neurophysiologically identical to you. There would be no way for another person to tell that your zombie twin was not you merely by comparing you with it, no matter what tests he might arrange. Nevertheless, your zombie twin would not be the same as you, because it would not have consciousness. That is, it would not have subjective experience. In the phrase widely adopted from Thomas Nagel’s well-known essay, “What Is It Like to Be a Bat?”, there would be nothing it is like to be your zombie twin.
Philosophical zombies are a way to make vivid an old philosophical argument, going back at least to Descartes, known as the Conceivability Argument. In the present version, the idea is that it is conceivable that the world could be just as it is physically down to the last elementary particle, but without conscious experience. That is, we could have had the Big Bang and the evolution of the universe just as it is, including the evolution of life on earth and the human species, and physically everything would be just as it is today, including all of us having discussions just like this one, although none of us nor any other beings have conscious experience. By conceivable, I mean there is no contradiction in this scenario and no reason that science can discover, whether physical, psychological, or otherwise, why it could not have happened. If this scenario is conceivable, then it seems we must conclude that consciousness is epiphenomenal: it is nonphysical, it cannot be explained by the physical, and its presence or absence makes no difference to the causal, functional, or physical order of nature.
Although he did not originate the idea of a philosophical zombie, present-day discussion of zombies is due mainly to David Chalmers, who describes them in his 1996 book, The Conscious Mind. Here is a web page by Chalmers that briefly introduces philosophical zombies and provides references to some older literature (i.e., 1990s). Here is an excellent brief write-up, very short. (It is included in Rosen et al., eds., The Norton Introduction to Philosophy, 2nd Edition, if anyone happens to have that lying around). My discussion in this post is based on the idea of a zombie as deployed by Chalmers.
Why should we accept that zombies are conceivable? Consider the example of bat echolocation employed by Nagel in his above-mentioned essay. Bats can fly in the dark, avoiding walls and trees and finding insects to eat by emitting high pitched squeaks that are reflected from the surrounding objects and picked up by the bat’s hearing. These reflected sounds are of sufficiently high frequency to permit the bat’s auditory cortex to form a cognitive map of its immediate environment good enough for flight navigation. Thus, the bat can literally hear the layout of its environment well enough to negotiate tree branches and snag bugs out of the air! And we understand perfectly well how this is done, and we could in principle map the process down to the last synapse in the bat’s brain. But doing so would evidently be no help to us in knowing what it’s like to hear the layout of our environment the way the bat does. It would be like seeing with our ears. It is telling that even the most detailed scientific understanding of echolocation would yield no insight into what it’s like. More importantly, even the most detailed scientific understanding makes no reference to what it’s like. The scientific account of echolocation refers to sound waves, their frequency, their reflection, the structure and dynamics of the bat’s ears and auditory receptors, the bat’s neural pathways, patterns of activation, motor control, and so forth. The physics of sound and the anatomy and physiology of the bat are sufficient to explain echolocation. There is no need—apparently, there is no place—to discuss what it’s like for the bat to use echolocation, even if the physiology could give us some insight into what it’s like, which evidently it can’t.
Therefore, if there were nothing it’s like for a bat to use echolocation, this would make no difference. It is conceivable that a bat could be a zombie. And if this is true of the bat, it is true of us, too. What goes for echolocation goes equally well for vision and, apparently, every other human faculty. We may not notice this, because we know what it’s like to employ vision and all our other faculties, and we take for granted that, say, what it’s like to see things is important to our use of vision. But if space aliens—who did not have vision—were to come to Earth and study us as we study bats, they would come to the analogous conclusion: human vision is explained entirely by the physics of light and by human anatomy and physiology. There would be no need, and apparently no place, to discuss what vision is like for humans. Therefore, human zombies are conceivable.
This is a remarkable conclusion! I don’t know about you, but I feel very strongly that if there were nothing it was like for me to see things, I would notice!
Chalmers accounts for the difficulty of explaining consciousness by contrasting it with the usual method of explaining abilities and functions. This usual method is to identify the underlying mechanism by which the ability or function is realized. For example, the concept of genes was introduced long before DNA was discovered. Genes were proposed to be the means by which an organism’s traits are encoded and this information passed from parent to offspring in reproduction. However, when this concept was introduced, people had no idea how this was done or what structures in the body corresponded to genes. Later, it was discovered that the encoding and transmission of traits takes place in DNA molecules. In this way, it was discovered that genes are DNA.
What enabled this discovery was the existence of a functional definition of the gene. Given a functional definition, people could search for what mechanism realizes that function. When they find that mechanism, they will have found the gene. It is essentially the same procedure by which we discovered that heat is mean molecular kinetic energy and that lightning is electrostatic discharge, and by which we will eventually discover what memory is in the brain, and face recognition, and speech perception, etc.
But apparently this procedure will not work for consciousness. The reason is that we lack a functional specification of consciousness. Everything that we can functionally specify is a mental ability, like echolocation, vision, memory, speech, etc. And it seems that we can explain these completely, as in the examples of echolocation and vision, without reference to consciousness—to what they’re like or to their being like anything at all. After we have given a complete explanation of any mental ability you like in physiological terms, it seems we will always still be able to ask, “yes, but why is this accompanied by consciousness?” This implies that consciousness is not an ability. And this implies in turn that consciousness is beyond the reach of scientific explanation. If so, then—since all our mental abilities can be explained without reference to consciousness—zombies are conceivable.
Now, I find it difficult to believe that there could be a zombie version of me. I don’t think I could function without consciousness, and therefore neither could zombie-me. But it is not obvious how to reply to the arguments just presented.
One indication of trouble for Chalmers’s position is the evident contradiction between his claims, on the one hand, to have direct knowledge of his conscious experiences, and on the other, that a zombie would think it has conscious experiences. For instance, the second sentence of his well-known article, “Facing Up to the Problem of Consciousness,” reads, “There is nothing that we know more intimately than conscious experience.” Yet in his first published discussion of zombies, “Self-Ascription without Consciousness: A Case Study,” he writes, “qualia [i.e., qualities of conscious experiences] don’t seem to play a primary role in the process by which we ascribe qualia to ourselves! ([A zombie-me,] after all, ascribes to himself the same qualia; it’s just that he’s wrong about it.)” Thus, according to Chalmers, zombie-me attributes the same consciousness to himself that I do. And zombie-me’s attributions must be sincere, since we are functionally identical. Evidently, then, zombie-me doesn’t know he’s a zombie. How, then, do I know I’m not? It seems that, by Chalmers’s own admission, we do not know intimately our conscious experience.
However, I doubt that Chalmers would be willing to give up the claim that we are each acquainted with our own conscious experience. In order for there to be “problem of consciousness” to face up to, consciousness must be a datum. But this implies that there is something wrong with the zombie idea. If we know as a matter of acquaintance that we are conscious, then a zombie should know in the same way that it isn’t. Therefore, what I and zombie-me know, think, say, etc., will be different, which is contrary to the conception of a zombie. Functionally identical zombies do not seem to be conceivable, after all.
Can we put our finger on what has gone wrong? What justifies the alleged conceivability of functionally identical zombies is the claim that consciousness is not a function. That’s why a being could be physically identical to me, yet lack consciousness, and still be functionally identical to me. It’s also what justifies the claim that consciousness cannot be scientifically explained. Consciousness plays no functional or causal role; and we cannot find the mechanism that realizes a role that does not exist. What we must do, then, is identify the role that consciousness plays in our mental life. What does consciousness do? What is it for?
It is worth pausing for a moment to emphasize the seeming intractability of this problem. It can seem inevitable that consciousness must have no functional role. For, once any functional role is specified, it will seem necessary that it be performed by a physical mechanism or structure of some kind, in accordance with physical laws, and therefore with no need of consciousness. Physics is a closed system, in the sense that there can be no nonphysical causes without violating the conservation of mass–energy. And physical causation has nothing to do with consciousness. Therefore, any functional role must be performed by a physical mechanism to which consciousness is, at best, epiphenomenal.
Can we avoid this conclusion? I have an idea, which I will attempt to explain in what remains of this post.
The problem is to find a functional role for consciousness that is not a matter of physical causation as ordinarily conceived. I suggest that the presentation of intrinsic properties is such a role. By “intrinsic properties,” I mean properties that are not relational in any essential way. For example, being larger than, being a parent, and being heavy are all relational: they consist essentially in relations between things. By contrast, being a particular shade of red is not essentially relational. To have a certain color is something an object can do all by itself. It is a property it has in itself. Therefore, it is intrinsic.
In sense-perception, such as seeing, we perceive objects having intrinsic properties and standing in certain relations. For example, I see a cat that is on the mat or a tree that is larger than the shrub next to it. The relations we perceive are real, at least to a first approximation. The cat really is on the mat, the tree is larger than the shrub, and so forth. But the intrinsic properties are not real. They do not exist at the objects in the way they appear to. Thus, the cat exhibits certain colors, the leaves of the tree look to be green, etc., but these objects do not really possess these qualities—so far as we know. To be clear, I am speaking of the color qualities as they appear to us, the color qualities you can close your eyes and imagine. These are intrinsic properties. They do not consist in relations, including causal relations, and this is the fundamental reason we do not perceive them—meaning we do not detect them by means of our sense organs. To be perceived, they would have to have causal effects. But they do not have causal effects.
To see this more clearly, think of the scientific account of color perception. The colors of object surfaces are normally perceived when ambient light is reflected from those surfaces and stimulates the cone receptors in our retinas. The object surfaces reflect certain wavelengths and absorb others, and likewise our cone receptors are stimulated by certain wavelengths much more strongly than by others. Thus stimulated, the cone receptors send electro-chemical pulses down their axons and stimulate a series of other neurons across the synapses between them in a complex chain that results eventually in color experience. Now, none of this has anything to do with the color qualities that we experience at the end of the chain. The object surfaces need have no color at all. Whether they do or not—regardless of what they are intrinsically like—their role in color perception is only to reflect certain wavelengths and absorb others. This reflectance property is a causal disposition, not an intrinsic color quality. Nor do the reflected wavelengths have color. Light is electromagnetic radiation, the same as in your microwave oven or an x-ray machine. It does not have color. Even if it did, it wouldn’t matter, because the only role of the light is to differentially stimulate the three types of cones in our retinas. Nor, obviously, do the neurons or their electro-chemical pulses or the ion exchanges across synapses have color. Color qualities are not involved in any part of this story until the very end, when—by some unknown means—they are finally experienced. Therefore, color perception is not a matter of detecting color qualities by their presence in the physical world. Rather, the color qualities are attributed somehow to perceived object surfaces by our own visual system.
What goes for color goes for all other intrinsic sensory qualities, such as sounds, tastes, smells, and felt hot and cold. They have no causal properties, and therefore they are not detected by our sense organs. Rather, they are qualities that we experience as part of our way of perceiving the world despite not being detected by our sense organs.
There is a general lesson here worth emphasizing. What we are able to know about the world outside our minds is limited to its structural and causal properties. It does not include the intrinsic properties of things. This is evident from science. The properties that science discovers are all causal and structural (for example, spatial). They are the properties represented by “the bound variables in our scientific theories.” And what science is able to tell us about them is exhausted by the laws in which they appear. For instance, what do we know about mass? We know that it confers momentum on the things that have it, as well as inertia and the acceleration of gravity. We now know that it is convertible with energy in a certain way. This more or less concludes what we know about mass. What is mass “in itself”? What is its “categorical ground”? We have no idea and will never have. Such facts, whatever they may amount to, are forever inaccessible. The same goes for electric charge, frequency, and every other scientific property.
What goes for the properties referred to by scientific theories goes for everything else we can know about the world outside our minds. We can learn of things outside our minds only by their effects on us. What has no effects we cannot learn about. But that doesn’t mean it doesn’t exist! In fact, there must be more to the world than causation and structure. It can’t be relations all the way down. Relations require relata. There ultimately have to be things in themselves to be causally and structurally related. Only we will never know what they are. There is thus a great deal of physical reality, including physical properties of objects all around us, of which we are entirely and necessarily ignorant.
This is why, in sense-perception, the relations we perceive between objects may be real, but the intrinsic properties cannot be. Causal and structural properties make a difference to events in the world, and so our sense organs can detect them. Intrinsic properties do not make a difference to events in the world, and so our sense organs cannot detect them. But there must be some intrinsic properties! And since our sensory systems cannot detect them, they must supply them. Just as reality cannot be relations all the way down, neither can our perceptions. We cannot perceive only relations. We must perceive relations between intrinsic objects. Thus, the intrinsic properties of experience exist, in Simon Blackburn’s handy phrase, to fill in space. We live in a sensory world of real structure and dynamics but of illusory intrinsic qualities.
This is the role of consciousness. Consciousness is “what it’s like” to be you and do the things you do. In the case of sense-perception, what it’s like to perceive, say the cat on the mat, consists primarily of the intrinsic qualities, such as the colors, that are attributed in perception to the cat, the mat, and so forth. Likewise, when it comes to all other mental events, such as sensations, emotions, feelings, and thoughts, what it’s like to undergo these events consists in the first instance of intrinsic qualities. Consciousness is the faculty that presents intrinsic qualities.
If this is right, then zombies are impossible. Perception requires intrinsic objects, and consciousness is the presentation of intrinsic properties. I am right to think that if I didn’t have consciousness, I would notice. (And Chalmers is right about this, too.) If I did not have consciousness, I could not have sense-perceptions. Since zombie-me has no consciousness, it cannot have sense-perceptions. Therefore, it cannot be functionally identical to me. So, there is no zombie-me.
I don’t mean to suggest that any sort of “perception” in the sense of detecting and responding to properties in the environment requires consciousness. A thermostat detects the temperature in a room in the sense that it has a state that tracks the temperature. A self-driving car likewise detects road conditions. Maybe even bat echolocation similarly lacks consciousness (although this seems unlikely). But human perception is not like this. Human perception presents intrinsic objects; this requires consciousness. Consciousness is the presentation of intrinsic objects (and consequently of relations between them).
I cannot say why human perception (and mental life generally) employs consciousness. Perhaps it has to do with intentionality, by which I mean the representation of something other than what does the representing. Your thought of your mother, perception of a tree, and memory of what you did on your last birthday are all intentional states. Notice that the thermostat doesn’t actually represent anything. It has a state that tracks (i.e., covaries with) temperature and controls a heater or air conditioner accordingly. But the state that tracks temperature is not about the temperature or anything else. (More sophisticated proposals which, however, still run along essentially similar lines, such as this one, do not solve this problem.) Of course, we may interpret it as being about the temperature. But this is a character we impute to it. In itself, it is a “blind” causal mechanism. By contrast, your perception of a tree, by itself, really is about the tree, as is your thought of your mother or of the color red. The original representation of anything requires presenting it as something distinct from what does the representing. Perhaps this requires presenting an intrinsic nature. But this of course is entirely speculative.
I also cannot say how this account of consciousness avoids the problem of the closed system of physical causation. My account does not require intrinsic properties to have causal powers. If consciousness presents intrinsic properties, then it is consciousness that exercises causal powers; the intrinsic properties may be merely the objects of conscious awareness. But their presence or absence on this scheme still makes a causal difference in the world. This means that certain actions of consciousness will vary depending on factors (intrinsic properties) that are not recognized in the book of physics.
Unlike Chalmers’s own ideas about consciousness, this doesn’t require us to introduce anything nonphysical. This is not dualism or epiphenomenalism. Intrinsic properties are physical properties, and we have independent reason to believe they exist. But they are not part of current physical science, and, other than obliquely, they never can be.
- Thomas Nagel’s early article.
- “What Is It Like to Be a Bat?”, Philosophical Review, 83 (1974): 435–450.
- This article, published just as functionalism was becoming the dominant philosophy of mind, received attention for its catchy way of showing what functionalism leaves out. However, the problem of consciousness didn’t really gain traction until the work of David Chalmers.
- “What Is It Like to Be a Bat?”, Philosophical Review, 83 (1974): 435–450.
- David Chalmers’s work on consciousness.
- The Conscious Mind, Oxford U.P., 1996.
- “Self-Ascription Without Qualia: A Case Study,” Behavioral and Brain Sciences, 16 (1993): 35–36.
- This is the earliest published writing by Chalmers on zombies that I know of. It was a comment on a target article by Alvin Goldman.
- “Facing Up to the Problem of Consciousness,” Journal of Consciousness Studies, 2 (1995): 200–219.
- This ubiquitous article summarizes Chalmers’s view nicely, and you only have to read the first 20% to get the main argument (sec. 1–3). Unfortunately, it doesn’t talk about zombies.
- “Consciousness and Its Place in Nature,” in D. Chalmers (ed.), Philosophy of Mind: Classical and Contemporary Readings, Oxford U.P., 2002.
- The best essay-length statement of Chalmers’s views on consciousness.
- General article on zombies for those who want a broader view.
- Robert Kirk, “Zombies,” The Stanford Encyclopedia of Philosophy, 2019.
- The idea that all we can know of the world outside our minds is its structural and causal relations, not its intrinsic properties, is called structural realism. There is a large literature, but the following are noteworthy.