Preprint version: not for scholarly citation.
“Or we could live in sin, except that I'm not alive.” (Rachel to Deckard)
“You’re so perfect!” (J. R. Sebastian to Roy Baty) 
Zombies are undead. Something animates them, but whatever it is —whether voodoo or science run amok—it’s not their departed souls or, as philosophers (not wanting to prejudge the theological question of souls) would have it, their conscious selves or minds. Zombies enter philosophy where horror fantasy and science fiction meet dawning scientific fact. Robots, many believe, would be unconscious “soulless automata.” Many further believe robots will remain unconscious – not to mention soulless – no matter how convincingly human their appearances and behavior become. Even robots able to fool the most careful unaided observer (like Blade Runner androids) would still be unconscious. Soulless automata. Zombies! Wouldn’t they? Many think so.
That zombies enter philosophy in this vicinity partly explains the peculiar way in which philosophy zombies are undead. Much like computers (according to the general supposition), typical philosophy zombies are unfeeling, but not necessarily unthinking. This is quite unlike the stereotypical zombies of George Romero’s Living Dead series which, if not unthinking, are, at least, severely cognitively diminished; but not necessarily unfeeling. Maybe they have sensations – maybe brains taste like something to them. Who knows? On the other hand, they are certainly intellectually challenged. You can’t reason with a Living Dead zombie.
In Romero’s Land of the Dead zombies develop rudimentary communication and tool-use skills. In philosophy zombies’ intellectual abilities are imagined to be even further developed; in the usual case, as well developed as yours and mine. “As human as human” is the rule with zombies in philosophy: zombies in philosophy are more akin to the “wives” in Stepford Wives or the pod “people” in Invasion of the Body Snatchers than to the brain-munching ghouls of Night of the Living Dead. Still, with Stepford “wives” and pod “people” there’s something discernibly, albeit subtly, amiss. Philosophy zombies resemble normal humans more perfectly still. In philosophy, zombies are typically, outwardly, indistinguishable from normal humans, at least to the naked eye; “virtually identical to a human” (Prologue), like the Replicants in Blade Runner. A Replicant can only be detected by a special test measuring “capillary dilation of the so called blush response, fluctuation of the pupil, involuntary dilation of the iris” and other involuntary emotional responses. They call it “Voigt-Kampff,” for short. Some philosophers have even imagined zombies literally identical to humans; some philosophy zombies are humans. We’ll return to that.
Zombies that haunt philosophers’ imaginings these days are fully humanoid agents, information processors, and even organisms. Outwardly, or objectively, philosophy zombies appear to be typical individuals. Inwardly, or subjectively, however, there’s no conscious “light” behind their eyes, however much they seem to shine. Objectively philosophy zombies are perfect human replicas, as human as human; only, subjectively, there’s “nobody home.”
Despite their seeming intelligence – or rather precisely because of it – philosophy zombies do eat brains … in a sense. And herein lays their horror for materialistic philosophies of mind and, consequently, for scientific psychology. In the philosophy of mind, materialist theories maintain that mental functions can all be accounted for in physical terms, or identified with material processes, without recourse to immaterial entities like souls. In philosophical thought experiments, however, zombies seem to “devour” everything about brains that various materialist theories might suppose thoughts and sensations to be. Philosophy zombies “eat brains” by seeming to be counterexamples to every attempt to identify thoughts and sensations with the things about brains materialists propose to identify them with. Behaviorists propose to identify mental states with adaptive behavioral output. Functionalists propose to identify mental processes with computations that produce such output. Mind-brain identity theorists propose to identify minds with the underlying neurophysiology that, perhaps, implements such computations. All these alternatives seem to be ruled out by zombies. In philosophy zombies eat behavior, programs, and brains.
“Is that a real snake?” (Deckard to Zhora)
The first reported sighting of what philosophers have since come to call “zombies,” however, was not of humanoid zombies, but of zombie animals. In the wake of William Harvey’s discovery that the heart is a pump, René Descartes proposed, in 1637, that every life function could be explained materialistically, on mechanical principles. Even adaptive intelligent-seeming behavior, such as birds’ nest-building and navigational abilities, Descartes argued, could be explained as the activities of unreasoning, unconscious mechanisms. Proposing to regard animal bodies as machines, Descartes concluded that all animals are “automata,” furry (or scaly or feathery) robots, totally lacking reason and consciousness. Shades of Pet Semetary … except in Pet Semetary, the resurrected pet corpses – like Romeran zombies, unlike philosophers’ zombies – are manifestly not themselves. They’ve changed; turned vicious. Zombie animals, as conceived by Descartes, on the other hand, are indistinguishable from our pets, wildlife, and cattle; because they are our pets, wildlife, and cattle!
Descartes, notably, did seem to glimpse humanoid zombies once. Having introspectively assured himself of his own consciousness, still, looking out his window to see other “men crossing the square” he wondered whether what he saw were “any more than hats and coats which could conceal automatons.” What do I really know about anyone else’s inner life – or lack thereof – except what their words and acts seem to reveal? Zombies are us! Oh no! Or, rather, since I know from my own first person experience that I am conscious (or so Descartes taught) zombies might be you. The “other minds problem” philosophers call it. Descartes, however, dismissed the possibility of humanoid zombies. The variety and flexibility of human actions “guided by the will,” he reasoned, could not be mechanically replicated. Since the infinite variety of conscious reason outstrips the necessarily finite resources of material devices, Descartes concluded no machine could ever use words or other signs to “produce different arrangements of words so as to give a meaningful answer to whatever is said in its presence”; neither could a machine “act in all the contingencies of life in the way in which our reason makes us act.” Whereas “reason is a universal instrument which can be used in all kinds of situations,” Descartes reckoned such scope of application and flexibility of operation would be “impossible for a machine.”  Only the presence of an immaterial principle or soul in us (a theory called “dualism”) could explain the virtually infinite variety of human behavior. Whereas life-processes could be mechanized, thought processes could not.
Descartes, it now seems, dismissed the possibility of humanoid zombies too soon. He underestimated the potential of machines. In 1937 Alan Turing demonstrated the theoretical possibility of fully programmable devices, or “universal machines,” paving the way for the development of digital computers and showing how the infinite variety of behaviors “guided by will,” such as speech, might be, in principle, explained as mechanical effects of computation. Much as the heart is a pump, and biological inheritance is DNA replication, Turing-inspired “functionalists” (or “cognitivists”) hypothesize that the brain is a computer. Thought is computation.
“Have you ever ‘retired’ a human by mistake?” (Rachel to Deckard)
Just when everything seemed to be going materialism’s way – just when it was finally dawning what sort of mechanisms minds might be and what sort of material processes thought might be – on came zombies. At first materialists were in denial. They wanted to say, “A thought experiment – the mere possibility of zombies – proves nothing. Only if there really were NEXUS 6 units like Rachel, and they really weren’t conscious, would it refute behaviorism; only if there really were NEXUS 7 units (say) that computed the same subfunctions as brains, yet unconscious, would it refute functionalism; only if there really were NEXUS 8 units (say) with brains physiologically indistinguishable from human brains, yet unconscious, would it refute mind-brain identity theory. That such units conceivably might not be conscious proves nothing.” It was here that zombies in philosophy turned ugly. Modern materialism says that everything mental is identical (and scientifically identifiable) with something physical, just as water is identical (and scientifically identifiable) with H2O. Such scientific identification, however, entails that water is not just actually H2O (“in this possible world,” philosophers say); water is necessarily H2O (“in all possible worlds,” philosophers say). It has to be H2O, or else it wouldn’t be water. Essential scientific identifications – such as water with H2O – when true, are not possibly or conceivably otherwise. So the best philosophical opinion, following Saul Kripke, now has it.
But to imagine the presence of adaptive behavior, or underlying computational processes, or even brain processes as such, is not ipso facto to imagine conscious experience. Zombie thought experiments seem variously to imagine beings with humanoid behavior, programs, and brains, without subjective experiences. Zombies, so conceived, seem possible in ways inconsistent with the truth of all would-be materialist identifications. So the story goes.
“And since subjectivity can’t be identified with any sort of material processes – and I am directly aware in my own case that it exists – it must be essentially separate and immaterial; confirming dualism. On the main point, Descartes was right after all!” So, it seems, the story would continue — if zombies really are conceivable. Alas, it seems they too readily are.
To evoke zombies, John Searle instructs, “always think of [the thought experiment] from the first person point of view”; you have to imagine yourself as the zombie. As Searle sets the stage, suppose that doctors gradually replace your brain with silicon chips, perhaps to remedy its progressive deterioration. From here, you may conjure yourself a zombie by imagining as follows:
as the silicon is progressively implanted into your dwindling brain, you find that the area of your conscious experience is shrinking, but that this shows no effect on your external behavior. You find, to your total amazement, that you are indeed losing control of your external behavior. … [You’ve gone blind but] you hear your voice saying in a way that is completely out of your control, “I see a red object in front of me.” … [I]magine that your conscious experience slowly shrinks to nothing while your externally observable behavior remains the same.
You have imagined yourself gradually becoming a zombie. And no sooner are such zombies conjured, of course, than they’re off on their rampage. Against behaviorism, “we imagined … the behavior was unaffected, but the mental states disappeared.” And there’s no way to stop them before they transmogrify. Suppose the replacement chips implement the same programs, maintaining all the same internal functions as the brain cells replaced. The replacement chips perform exactly the same computations as the brain. There goes the functionalist idea that thought is computation – zombies ate its program.
Here Searle, himself hoping to spare some vestige of materialism leaves off. But zombies are not so easily stopped. Suppose silicon chip replacement therapy is unavailable. The medical doctors are powerless. You have heard of a certain witch doctor. Desperate, you fly to a remote isle, voodoo rites are enacted, and voila. The deterioration of your brain is magically reversed. To the amazed medical doctors back at the clinic, your brain is indistinguishable from your very own predeteriorated brain. But wait! As your brain is being magically restored … it’s just as before. Your conscious experience slowly shrinks to nothing. You’re a zombie. There goes mind-brain identity. Zombies ate its brain. Though some (including Searle himself) deny that such fully human bio-zombies really are conceivable, none have shown the contradiction in it. Indeed, it is just this that possibility that philosophers who have raised the “other minds problem” seem to be conceiving.
“`More human than human’” is our motto. (Tyrell to Deckard)
Since they do seem to be conceivable, bad brain-eating zombies must be stopped before they destroy civilized philosophy of mind and scientific psychology as we know it. I have a plan. Much as Earth enlisted Godzilla to battle Gidrah (in Gidrah the Three Headed Monster), I propose to create new breeds of zombies to battle their evil cousins. Familiar zombies – conjured, as it were, from the second person familiar perspective – increase our attachment: there’s I and thou and thou art zombie. Super smart zombies are not just as smart as us, but smarter. Being in these ways superlatively human, such zombies resist dementalization. Their imagined or stipulated lack of subjective experiences seems not to impugn the genuineness of their apparent intellectual endowments. These zombies can battle their evil cousins to a standoff, at least, and perhaps even defeat them.
To conjure familiar zombies, put yourself in Deckard’s shoes. Imagine your own true love to be a zombie. Imagine your beloved has no subjective experience of sensations, no inwardly felt experiences (what philosophers call “qualia) whatsoever. Yet your beloved, we’re imagining, behaves in every way just as your real-life beloved (who, presumably, is not a zombie) really behaves: smiles as warm; kisses as sweet; loyalty as steadfast; words as tender; love as true. Or so it seems. Only the “light” isn’t on. No qualia. I submit that you should not deny her (or his) cognitive abilities and attainments. She (or he) still wants you to prosper; still knows your preferences; still prefers scotch to bourbon, classical to jazz; and so on. To strengthen the intuition, extend the fantasy to your mother, your father, your children, all your friends, siblings, colleagues, teachers, everyone you know. They’re all zombies! Should you conclude that your beloved and the rest don’t think, that they know nothing at all and don’t understand English? It was from Mother, Father, and the rest, in particular, that you got your English, and such words as “think,” “know,” and “understand.” I think you should conclude, “How odd! I alone have these peculiar subjective experiences besides the wants, beliefs, etc. that others have.” The thought of mental life without subjective experiences may be horrible, but it’s conceivable. Familiar zombies show this.
As for super smartness … since the NEXUS 6 Replicants were “at least equal in intelligence to the genetic engineers who created him” (Prologue), let us suppose, off world, they undertake genetically reengineering themselves or their descendents. Suppose these descendents, NEXUS 9s, are decidedly more intelligent than the human genetic engineers who created their forebears; except, still no qualia. Suppose these NEXUS 9s return to earth. They show us how to make our microwaves synthesize food out of thin air and how to turn our Ford Tempos into time machines. They show us how to achieve peace on earth, with liberty and justice for all. What should we say? That our NEXUS 9 benefactors didn’t really know how to turn Tempos into time machines? That they didn’t really understand the revolutionary physical principles involved, but now we do? I think we’re not as ungrateful and conceited as that. We’d say our subjective-experience-bereft benefactors knew how to turn Tempos into time machines.
Finally, imagine zombies both smart and familiar; or, rather, venerated. Imagine it is discovered that many, most, or even all of the leading contributors to our human intellectual heritage(s) were zombies. Descartes, especially, included. Nevertheless, I submit, we should not deny their mental attainments. Despite not meeting the would-be dualistic essential condition for thought – despite being bereft of subjective experiences or “qualia” – these famously smart zombies remain paradigm thinkers on the strength of their achievements. Intelligent is as intelligence does, absence of itches, aches, tingles, visual images, and such, notwithstanding. Indeed the surpassing greatness of their intellectual attainments makes super smart venerable zombies especially easy to imagine, especially given the well-known antipathy between thought and feeling. No one ever solved an equation or proved a theorem in the throes of agony or orgasm.
“And if the machine doesn’t work?” (Deckard to Bryant)
It seems that intuitions about thoughts (cognitive states like belief and inference) and feelings (sensations such as itches and afterimages) diverge under different zombie-thought-experimental conditions. Brain-eating zombies like Searle’s undermine materialist accounts of sensation; qualia-eating zombies like mine undermine dualistic accounts of thought. Furthermore, the supposition that thought somehow requires feeling – besides being contrary to the well-known antipathy just noted – would be indecisive in its upshot. Even assuming this no-thought-without-feeling “Connection Principle” (as Searle dubs it), the question remains: should we conclude that intelligent acting androids are not really thinking (given their imagined lack of feelings); or should we rather conclude (given their evidently thoughtful behavior) that they have feelings after all?
Is it a standoff, then? Does dualism rule the experiential realm and materialism the intellectual? Yet, even this partial “triumph” of dualism would seem to be strangely empty and inconsequential. As pure sensations or experiences, shorn of every concomitant physical (behavioral, functional, neurophysiological) element, qualia are “saved” from materialistic identification precisely by being conceived as ineffectual; as events lacking further effects, what philosophers call “epiphenomena.” This is suspicious. Not everything thought to be conceivable really is so. Materialists who were in denial thought they could conceive of water not being H2O; but they were mistaken. Perhaps it is the same with the supposed conception of beings “which are physically and functionally identical [to us], but which lack [subjective] experience”: perhaps such “phenomenal zombies,” as David Chalmers calls them, likewise, only seem to be conceivable, but really are not. Chalmers notes parenthetically, “It is not surprising that phenomenal zombies have not been popular in Hollywood, as there would be obvious problems with their depiction” “Problems” to say the least! They look and act exactly like you and me. In the case of NEXUS 8 bio-zombies (imagined above), neither CAT scans nor MRIs nor any other conceivable objective tests could distinguish phenomenal zombies from ordinary human beings. This being the case, the “problems with their depiction” that Chalmers notes, are likewise problems with their conception. What exactly are we supposed to be imagining when we imagine beings “physically identical to me (or to any other conscious being), but lacking conscious experiences altogether”?
Here we are in a position to appreciate Searle’s observation that the conception of phenomenal zombies has to be done “from the first person point of view”; but who is that first person? Whose point of view is it? It must be the zombie’s, but the zombie, “lacking conscious experiences altogether,” is supposed to lack a “first person point of view”! Like mad-scientists in the movies who are destroyed by their own creations, it seems the zombie spawning thought experiments destroy the experimenters themselves, and with them the experiments. If qualia are required for a first-person point of view, the experiment abolishes the viewpoint on which it depends. It seems that zombies, then, are not coherently conceivable after all.
So, is that the end? Have zombies been defeated? Has Plan A prevailed after all? Has thoroughgoing materialism been saved? I wouldn’t be too sure. First, it seems we can imagine the process of zombification (as in Searle’s evocation) even though we can’t quite see our way to the end. So long as the subject’s awareness hasn’t shrunk to zero, there’s a first-person point of view that can be imaginatively taken. Perhaps, where imagination leaves off, we extrapolate (“and so on”) our way to the end. Furthermore, the incoherence of a first person narrative minus the first person only arises where qualia are totally absent: for so-called “absent qualia” scenarios. “Inverted qualia” (I-see-red-where-you-see-green type) scenarios would still seem conceivable, and pose similar challenges to materialism. If either green-experiences or red-experiences might conceivably “supervene” (as philosophers say) on the very same state of the brain, then that state cannot be identified with either experience.
Fortunately, while good, qualia-eating zombies, do not provide full immunity, they do provide a measure of protection, like flu vaccine: you still get zombies, but a milder case. It is only so far as the mental states in question depend on qualia for their existence, only so far as they are inconceivable without qualia, that bad, brain-eating, zombies show that the mental states cannot be identical with brain states. Good zombies show that this is not that far for mental states involved in cognitive thought. For a good portion of our mental life (the whole cognitive part, it seems), qualia are inessential; so, good zombies seem to show.
As for sensations (itches, aches, afterimages, tastes, and the like), though prospects for materialistic identifications of these remain in jeopardy, even here, good zombies diminish the havoc brain-eaters would otherwise wreak. Computers equipped with sensors, for instance, are “sentient” in a sense. They get information about their surroundings from ambient light (as in vision), vibration (as in hearing), and chemistry (as in taste) even if not in the sense of having subjective visual, auditory, and taste experiences. They can still be said (I think unequivocally) to “see” things they visually detect, “hear” sounds they aurally discern, and “taste” flavors they chemically differentiate.
What about consciousness? How can zombies behave intelligently if they’re unconscious? Good zombies are conscious in the sense of being cognizant of things (registering their presence) or being cognizant that certain things are or aren’t the case (representing them as so being) despite not being “phenomenally conscious” (possessed of qualia). Conceivably, your imagined zombie lover is in this manner aware of your presence: she registers it, and responds just as your real lover. Conceivably NEXUS 9 astro-zombies are aware that Tempos are gas-powered: they represent that fact and respond accordingly. This much is cognitive. It’s the specifically phenomenal (subjectively felt), not the cognitive (rational representational) aspect of consciousness that’s supposed to be lacking.
As for self-consciousness and subjectivity … good zombies, it seems, can even be “self-aware” in the sense of having access to their own internal states. Some computer programs, for instance, maintain state variables. Good zombies, it seems, can have points of view both literally (the loci of their visual sensors) and figuratively (in the form of unique overall representations of reality).
As for the soul … being pure sensations or “raw feels,” qualia seem very far removed from that and from actions “guided by the will.” Even if they do lack raw feels (as we have been imagining), Replicants still realize they are all-too mortal, fear death, and “want more life”; or so they say. And they say it with feeling; or so it seems.
“How can it not know what it is?” (Deckard to Tyrell)
Much as philosophy zombies are supposed to copy human beings in every way except their feelings, Blade Runner Replicants were “designed to copy human beings in every way except their emotions.” However, “the designers reckoned that in a few years they might develop their own emotional responses. Hate, love, fear, anger, envy” (Bryant to Deckard); and in the film, it seems their designers were right. From hot-headed
In Blade Runner it takes Deckard more than a hundred questions to determine that Rachel is inhuman. Being inhuman, she is denied moral standing. She can be summarily “retired” without trial or justification. To be morally disenfranchised on the basis of barely discernible differences in involuntary emotional responses – “capillary dilation of the so-called blush response, fluctuation of the pupil, involuntary dilation of the iris” – seems like picky grading, unless these involuntary emotional response differences are indicative of some deeper lack. But how would we know that?
We do have evidence, in the inhuman way Pris convulses when shot through the solar plexus, that Blade Runner Replicants differ from humans in their internal wiring, their subfunctional architectures. Plausibly this makes a difference in the way it feels to be them. But how would we know that? And why should it be supposed to be the difference between having feelings and having none? Perhaps Rachel’s loving feeling is merely different than ours, not absent.
We have further evidence in the way
And if she has no feelings whatever – if she really is a philosophy zombie – what then? “I love you,” she says. And if she not only talks the talk, but walks the walk ever after, if she wishes him well, and wants to be with him, is it all a lie for lack of some subjective amatory itch? And if it was no subjective feeling that moved her to pull the trigger, killing
But wait! Even if qualia are required for true love, and Rachel has none, her profession of love is no lie. She will have spoken untruly, but she will not have knowingly done so. What she thinks – including what she thinks about her feelings – is supposed to be unchanged. Good zombies, among other things, enforce this supposition. Lack them though she may, she still thinks she has feelings; as does Deckard. As do I. Conceivably, I too am deluded! Based on Descartes intuition of infallible self-awareness (the first person point of view) as the zombie thought-experiments are, yet seeming to undermine that very intuition, philosophy zombies may not be coherently conceivable after all. Plan B has led us back to plan A.
 Philip K. Dick, Do Androids Dream of Electric Sheep, chap. 17.
 Ridley Scott (dir.), Blade Runner. Quotations and references to characters and situations below refer to this film adaptation of Dick’s novel.
 The first to actually call the unconscious or soulless automata philosophers have imagined “zombies,” was Robert Kirk, in 1974 (“Zombies v. Materialists,” Aristotelian Society Supplement 48, 135-163).
 Meditations on First Philosophy, Discourse on Method, Med. 2, translated by J. Cottingham in Cottingham, Stoothoff, and Murdoch (trans.), The Philosophical Writings of Descartes, Vol. 2 (New York: Cambridge University Press, 1984).
Discourse on Method, Part 5, translated by R. Stoothoff in Cottingham, Stoothoff, and Murdoch (trans.), The Philosophical Writings of Descartes, Vol. 1 (New York: Cambridge University Press, 1985).
 Naming and Necessity (Cambridge, MA: Harvard University Press, 1972). To imagine the stuff filling our lakes and streams to be other than H2O, Kripke pointed out, is not to imagine water not being H2O; it is imagining the stuff so described (and so called) not being water! To imagine the presence of H2O is ipso facto to imagine the presence of water, and vice-versa.
 The Rediscovery of the Mind (Cambridge, MA: MIT Press 1992), 66-9.
 The Conscious Mind: In Search of a Fundamental Theory (Oxford: Oxford University Press, 1996), 94-6.
 And if she has some such feeling, but wishes him ill, wants him gone, and betrays and despises him ever after, has she spoken truly?