1. What Are Zombies?
Zombies are stipulated to be creatures that are in some way identical to human beings-and thus, in some sense, indistinguishable from human beings-but which lack consciousness. Zombies are at least behaviorally identical to human beings or other conscious creatures, and they may also be like us in other ways.
That there might be creatures indistinguishable from human beings but that lack minds or consciousness is an idea that appears repeatedly in philosophical considerations. In one traditional form, the worry arises as an epistemic problem about our evidence for taking other things to have minds. You may recall Descartes' concern that the hats and coats that he judges to be worn by people in the square might "conceal automatons." And Hilary Putnam reminds us that William James' entertained the possibility of what he called an "automatic sweetheart." In the middle part of the 20th Century, it seems, this "problem of other minds" was a major problem, perhaps the major problem, in Anglo-American philosophy of mind.
From these epistemic concerns spring zombies. The inventor of philosophical zombies, Robert Kirk, separated the metaphysical issues countenanced from their epistemological vehicle, skepticism about other minds (1974a)1. Kirk created a salient device for investigating the metaphysical intuitions that underlie the epistemic worries. If the skeptical challenge about other minds is even coherent, then it must be at least logically possible that certain kinds of creatures could exist:
But-here's the rub-the very possibility of zombies is at odds with at least some materialist (physicalist, naturalist) theories of mind-and perhaps with any version of materialism. For materialists (physicalists, naturalists) hold that the material (physical, natural) facts fix the psychological facts, including facts about conscious states. Thus, even taking seriously the skeptical worry involves, prima facie, rejecting materialism (physicalism, naturalism). In that case it might seem that a central problem in philosophy of mind has built into it the rejection of one supposed solution. Hence the problem of zombies.
Zombies are now in vogue2, but they are not the only denizens of recent philosophy of mind who press these issues. Zombies, like the unfortunate victims absent, inverted, alien, and dancing qualia, are just one way of pushing questions about consciousness (cf., Block 1980a, Shoemaker 1982, Chalmers 1996). All these characters serve the purpose of putting questions about the nature and causal efficacy of consciousness in a particularly salient form. By reifying the entities that our theories describe or license, thought experiments force us to think carefully about the consequences of various theories of mind and consciousness. They draw out the commitments, conditions, and caveats of differing views about mind and consciousness. Are zombies possible? The answer depends on the details of how the zombies are stipulated, what kind of possibility is in question, and what sort of theory of consciousness you hold. And it depends on where you begin your considerations, what you take to be given. As Güven Güzeldere writes, "playing with the idea of zombies could turn into playing with philosophical fire. But precisely for that reason, it is important to pay attention to the particulars in using zombies as a tool of imagination in thought experiments" (1995: 327).
Güzeldere provides a taxonomy of zombies based on the question of how zombies are identical to conscious creatures: Are they supposed to be behaviorally, functionally, or physically3 identical to conscious beings? (1995) Behaviorally identical zombies make all the overt movements and utterances that conscious creatures do, but they may have any internal structure and may be composed of whatever material.4 Behaviorally identical zombies needn't be hollow shells, they could be quite sophisticated; however, in considering behaviorally identical zombies their internal organization is left unspecified. Functionally identical zombies not only make the movements that conscious creatures do, they also have in some sense the same internal organization that conscious creatures do. Physically identical zombies are identical to conscious creatures cell for cell, molecule for molecule, atom for atom. These three ways in which zombies could be stipulated as identical to conscious creatures parallel three families of theories: behaviorism, functionalism, and Identity Theory.5 These are some of the theories that zombies can help us explore.
Owen Flanagan and I have drawn also a modal distinction regarding zombies, as has David Chalmers.6 We distinguish not only the sort of similarity that zombies might bear to human beings, but also the sense in which zombies are supposed to be possible: Are zombies logically, metaphysically, or naturally possible? (Flanagan and Polger 1995; Chalmers 1996). Logically possible, I take it, just means not contradictory.7 There is some question about whether there is an even weaker sort of possibility, something like conceivability (Horgan 1987) or epistemic possibility (Kripke 1972). Naturally possible I take to be something like compatible with all and only the actual substances and laws of nature.8 There is no general agreement about what metaphysical possibility is. Whether or not metaphysical possibility is connected to conceivability is a point of contention; although there is de facto consensus that conceivability (or imaginability, if that is different) is our only guide to metaphysical possibility.
Figure 1. Zombie Scorecard.
Taken together these distinctions form the axes of what I call the Zombie Scorecard (Polger 2000). Crudely put, the zombies that we are asked to conceive increase in strength (i.e., how much you are required to imagine) from left to right and from bottom to top of the Zombie Scorecard. The strongest claim is of the natural possibility of physically identical zombies. The weakest claim is of the logical possibility of behaviorally identical zombies. The Zombie Scorecard is a way of organizing questions about zombies. For example, in box (7) we ask the question:
(Q7) Is it logically possible that there be zombies that are behaviorally identical to human beings?9
Can we answer this question? Although no claim about zombies is entirely uncontroversial, most philosophers these days would regard as harmless the logical possibility of behaviorally identical zombies. Nevertheless, someone who is inclined towards analytic behaviorism, for example, would not accept this possibility; and Dennett rejects (7) on the grounds that behavioral identity would require functional identity, and he denies the coherence of functionally identical zombies (1995; see my "Zombies Explained").
2. Getting to Know Zombies
Below I provide a tour of the zombie questions, but it is not my purpose to recount every reason for which any thinker accepts or rejects each of the nine possibilities. My goal, rather, is to illuminate the structure of the problem and the surrounding debates. So before we look at the zombies themselves, let us attend to some nuances of the debate.
1. Comparative Anatomy of Qualia Mutants
If the zombie questions are asked in the form, "Is it y-ly possible that there be creatures that are x-ly identical to human beings but which lack consciousness?" (where y is a mode of possibility and x is a degree or kind of identity) then someone might object that this begs the question as to whether human beings are conscious. To avoid any appearance of impropriety, the questions should be rephrased, "Is it y-ly possible that there be two creatures that are x-ly identical to one another but differ in that one is conscious and the other is not?"
This formulation has the additional advantage of making transparent how the form of the zombie construct is related to absent, inverted, alien, and dancing qualia thought experiments:
Zombies and Absent Qualia: Is it y-ly possible that there be two creatures that are x-ly identical to one another but differ in that one is conscious and the other is not?
Inverted Qualia: Is it y-ly possible that there be two creatures that are x-ly identical to one another but differ in that one's conscious states are "inverted" with respect to the other's?
Alien Qualia: Is it y-ly possible that there be two creatures that are x-ly identical to one another but differ in that one has conscious states that are entirely different in quality from those had by the other?
Dancing Qualia: Is it y-ly possible that there be two creatures that are x-ly identical to one another but differ in that one is always conscious and the other sometimes has the same sort of consciousness as the first and other times has a different sort (inverted or alien qualia) or none at all (absent qualia)?
These questions are all phrased in terms of interpersonal comparisons. If you think that the interpersonal cases are always or sometimes ill-defined (what Shoemaker (1982) calls the Frege-Schlick view), you can still ask all of these questions in their intrapersonal form. To do so, simply consider the two creatures in the above formulations as two creature-stages; that is, "Is it y-ly possible that a creature at time t2 be x-ly identical to the same creature at t1 but differ in that…?" It is a short step from the intrapersonal form to the first-person form, "Is it y-ly possible that I at time t2 be x-ly identical to myself at t1 but differ in that…?" Both of the intrapersonal and first personal variations are envisioned by Robert Kirk (1974a).
2. Zombies v. Swampman: Physical and Functional Identity
If the "functional organization" of a thing is thought of in a way that is both narrow (local, system internal) and ahistorical (synchronic, not dependent on the system's history) then physical identity entails functional identity. (Similarly, physical identity entails behavioral identity, but that is trivial on a thin notion of behavior.) It will be useful, in the discussion that follows, to exploit the asymmetric connection between physical and functional identity. The trouble with this strategy is that there are available notions of function, some of which are used by philosophers of mind, that do not fit these constraints-they are neither narrow nor ahistorical. For example, synchronic organismic physical identity does not entail functional identity according to "wide content" or teleological versions of functionalism. The functional organization of a thing at a time may depend on the history of the thing and its ancestors. In this way, functional identity can differ from behavioral and physical identity, which are narrowly individuated. This makes things very messy.
To avoid the complication presented by wide or historical notions of function, we must be willing to expand the scope of the stipulated functional identity to whole worlds or whole world-histories rather than individual creatures. (Ned Block makes a similar move in his "Inverted Earth".) The question then becomes: "Is it y-ly possible that our world be diachronically (through its entire history) x-ly identical to how it is in fact but that some creature in the world differ with respect to its consciousness?" This may seem an ad hoc move, but there is some support for it. The zombie problem admits the indexical formulation, "Could it be the case that this thing, right here, right now, have a different state of consciousness than it actually does, viz., none at all, while remaining identical in x way?" This version of the question licenses the whole-world identity formulation of the zombie problem.
Another reason for allowing the whole-world formulation is that denying the possibility of physically identical zombies is supposed to be a benchmark for materialism; it ought not be the case that any obviously materialist theory is grouped with non-materialist theories due to a technicality in how the zombie question is put. So for the purposes of the zombie question, physical identity entails functional identity.
By broadening the temporal and spatial scope considered in framing the stipulated identity of zombies, the zombie problem is kept distinct from its intentional cousin, Swampman. Questions about Swampman are specifically questions about two things that are currently identical but that have distinct histories. If the zombie question is interpreted with the widest temporal and spatial scope, the zombie question is about whether there are possible differences between two things which are currently identical and have identical histories. Thus, on this reading, Fred Dretske (1995), for example, could answer differently to questions about the natural possibility of physically identical zombies (no) and the possibility of Swampman (yes). For teleological functionalists and teleological representationalists about qualia, the only difference between zombies and swamp-people will be the temporal scope over which the modal possibilities are to range. In other ways, they are the same question. If you reject the wide historical scope formulation of the zombie question then some wide-content theorists (e.g., Dretske) will have to admit that physically identical zombies are naturally possible. In this case it will turn-out that the natural possibility of physically identical zombies is no challenge to materialism-for wide-content representationalists like Dretske count as materialists, if anyone does!
3. Zombies and Conceivability
Those who (like David Chalmers) begin with the intuition that zombies are possible usually regard this as showing that some or all versions of materialism (physicalism, naturalism) in philosophy of mind are inadequate. Robert Kirk (1974) introduced the logical or metaphysical possibility10 of functionally or physically identical11 zombies as prima facie reason for rejecting materialism. Likewise, Chalmers' (1996) takes the possibility of functionally and physically identical zombies to show the inadequacy of physicalist theories of consciousness. Kirk and Chalmers offer conceivability arguments: We can conceive of zombies of thus-and-such sort, therefore such zombies are possible, therefore such-and-such consequences follow for what are metaphysically viable theories of consciousness.
In contrast, those who (like Daniel Dennett) are committed to materialism (physicalism, naturalism) in philosophy of mind usually regard this stance as foreclosing the possibility of some or all varieties of zombies. Dennett denies the natural possibility of any sort of zombies because of his dedication to a loose behavioral functionalism and his wariness of the notion of consciousness that might be needed by the friend of zombies (1991, 1995). Fred Dretske, as mentioned, takes himself to be committed to the natural possibility of physically identical zombies on the basis of a prior commitment to a teleofunctional theory of consciousness (1995).12 Robert Kirk now finds zombies impossible because he thinks that their description is incompatible with our own grasp of psychological concepts (1999). And I have argued that an Identity Theory permits the natural possibility of functionally or behaviorally identical zombies (2000).
The point is that not all zombie friends and foes are running conceivability arguments.13 This is important for at least two reasons. First, we need to see what sort of argument we are faced with if we are to know how to respond to it. No progress is to be found from a clash of intuitions in which each side simply insists that it is better at conceiving or imagining. Both sides bear the burden of at least accounting for the opposing intuition; and if someone offers a theory that permits/limits the possibilities of zombies, then it is no good to simply reiterate one's intuition. Second, those who advocate zombies on the basis of conceivability arguments tend to hold that if behavioral zombies are possible then so are functional zombies, and if functional zombies then so too physical zombies. This makes the types of zombies seem like an all-or-nothing possibility.14 But those of us whose belief in zombies is a consequence of prior metaphysical commitments are able to discriminate between the possibilities more readily. We are able to do this because our preferred theories provide us with ways of justifying principles for making distinctions whereas the zombie conceivability advocates cannot appeal to such principles-for if they do, they are revealed as having prior theoretical commitments after all.
4. Zombies and Skepticism
Zombies arose from skepticism about other minds. Yet this is not the only way that some kind of skepticism figures into claims about zombies. To sort out the connections is beyond the scope of this Guide; but I'll indicate some considerations. At various points, some zombie debaters appear to be arguing from the position of skepticism about certain kinds of scientific explanation. Chalmers, who supports the possibility of zombies, doubts that any possible physicalist explanation can yield the conceptual necessity between body and mind that is required of a theory of consciousness. But Dennett's anti-zombie argument also has a skeptical component. He sees the claim of identity over-and-above correlation, for example, as unwarranted and philosophically extravagant because he is skeptical that non-functional differences can make a difference in the world. I do not hope to offer any wisdom concerning what we should think about these moves other than to note that the skeptical element in debates about zombies and qualia in general has not been much attended to.
3. A Guide to Species of Zombies
Talking about zombies just so, if we are not quite careful about how we take the details to be filled out, will only lead to confusion. But if we are careful, then zombies can be the useful tool that Kirk intended for exploring what we would like out of a theory of consciousness.
For some years now I have defended the natural possibility of functionally identical zombies. This is not because I have strong intuitions about zombies, it is because I take this possibility to be an acceptable consequence of the metaphysical view about consciousness that I find independently attractive. That is, my commitment (such as it is) to functionally identical zombies does not result from a conceivability argument. Of course a great deal of what passes for philosophy of mind depends on one's intuitions; I mean only to indicate that the intuitions with which I begin are not intuitions about zombies.
Descartes, I hazard to say, did not have intuitions about zombies. But we can be confident that if you are Descartes, then you are committed to the logical possibility of behavioral, functional, and physical zombies. Whether you would also take all zombies to be naturally or metaphysically possible depends on whether you take the soul-body union to have any necessity; perhaps it is anachronistic to expect this detail from you, Msr. Descartes.
This is how the non-conceivability use of zombies proceeds. Once you have a theory of consciousness in hand, you can determine whether or which zombies your theory allows for. Dualisms are the easy cases; the particular details tend to matter more for versions of materialism (physicalism, naturalism) because they deliberately distinguish themselves according to notions of identity and modal strength. Here then are some suggestions concerning how the advocates of some theories of consciousness should regard the various permutations of zombies. (If you are beginning with a conceivability argument to the effect that some or other zombies are possible then this will tell you which sorts of theories are available to you, and which your intuitions may render problematic.)
Questions (Q7)-(Q9). Is it logically possible that there be two creatures that are behaviorally, functionally, or physically identical to one another but differ in that one is conscious and the other is not?
Yes, all three kinds of zombies are logically possible. Two creatures that are physically identical are also functionally and behaviorally identical. So if physically identical zombies are logically possible then so too are functionally (remember the caveats!) and behaviorally identical zombies. Unless dualism is not even logically possible, physically identical zombies are logically possible.
The troubles with dualism are famous; but it has yet to be decisively shown false as a matter of logic alone, to be logically impossible.15 So for the momentwe cannot be certain that dualism is logically impossible. We must conclude that all three kinds of zombies are logically possible.
There are other ways to secure the logical possibility of functionally and behaviorally identical zombies even if it turns-out that dualism is not logically possible. But this brief sketch is sufficient for present purposes. Logical possibility (non-contradiction) is a rather weak claim. The fact that various kinds of zombies are logically possible has nary an implication for the metaphysical theories of consciousness.16
Questions (Q1), (Q4), and (Q7). Is it logically, metaphysically, or naturally possible that there be two creatures that are behaviorally identical to one another but differ in that one is conscious and the other is not?
Yes, all three kinds of zombies are possible.17 We can take advantage of the relation between the modes of possibility to once again answer several questions at once. If it is naturally possible that there be two creatures that are behaviorally identical to one another but differ in that one is conscious and the other is not, then it must also be metaphysically and logically possible.
And, as noted earlier, there is little doubt that behavioral zombies are naturally possible.18 Remember that it is irrelevant how the effect is achieved. So, for example, a life-size robotic puppet remotely controlled by a giant super-computer, or by my nervous system, would do the trick. When we see what a weak claim is being made, it is striking. It's not merely that a human behavioral duplicate is highly improbable but would violate no laws of nature. Rather, such a thing is not so improbable at all. We should all be rather surprised if a future George Lucas or Steven Spielberg or Rodney Brooks didn't actually manufacture some such thing in the not too distant future.
The case is stronger still if we remind ourselves that the zombie question can be considered with respect not only to humans but to any putatively conscious creature, including conscious animals. You may doubt that perfect behavioral duplicate human puppets are close at hand; it's more difficult to ignore the likelihood of robotic dogs or cats, say. Don't put RoboPuppy on lay-away yet-but it would be a tough road to argue that such a thing is precluded by laws of nature.19
What is naturally possible is also metaphysically and logically possible; so the answers to (Q1), (Q4), and (Q7) are each "yes."
Question (Q3). Is it naturally possible that there be two creatures that are physically identical to one another but differ in that one is conscious and the other is not?
If dualism is false then there cannot be two creatures that are physically identical but that differ with respect to whether or not they are conscious.20
The analytic or metaphysical behaviorist is committed to any behaviorally identical organisms being ipso facto mentally identical. Adding "more" identity makes no difference as long as physically identical things are also behaviorally identical. So the analytic or metaphysical behaviorist is committed to the natural impossibility of physically identical zombies. Functionalists are committed to the view that two things that are functionally identical are identical in consciousness; and two things that are physically identical are also functionally identical. So functionalists deny the possibility of physically identical zombies. Finally, for the Identity Theorist mental states are identical to physical states, so two things with the same physical states necessarily have the same mental states-physically identical zombies are impossible.
An exception to the negative answer to question (Q3) might be some sorts of anomalism, dual aspect theory or property dualism. But in order to accept the natural possibility of physically identical zombies one would have to deny any lawful or lawlike connection between the two aspects or kinds of properties-even a one-way connection.
David Chalmers' double-aspect view of information, for example, requires that there will be novel universal bridge principles connecting the physical properties and the conscious (informational) properties in any possible world. If natural possibility covers the expanded physical theory (physics*, or whatever the new expanded physical theory is to be called) that Chalmers advocates, then he may answer negatively to (Q3) (1995, 1996).21 But if Chalmers' view does not fall within the bounds of naturalism, then physically identical zombies will be naturally (but not naturally*) possible for him.22
Questions (Q5) and (Q6). Is it metaphysically possible that there be two creatures that are functionally or physically identical to one another but differ in that one is conscious and the other is not?
I don't know whether functionally or physically identical zombies are metaphysically possible. The trouble, mentioned above, is that there is no agreement about the nature of metaphysical possibility. I will leave questions (Q5) and (Q6) as officially unanswerable at this time. Even so, perhaps I can gingerly offer a few thoughts on how one might approach these questions.
On its most natural reading, Identity Theory denies that (Q6) is possible. That is, if sensations are in actuality natural kinds that are identical to natural kinds of brain processes, then there are no metaphysically possible worlds in which sensations and brain processes are not identical. This is exactly parallel to Kripke's (1972) widely accepted denial that there could be any metaphysically possible worlds in which water is not H20, given that water is in fact H20 in the actual world. Whatever metaphysical possibility comes to, it will be metaphysically impossible for things that are in fact identical to be non-identical. Thus, the Identity Theorist denies the metaphysical possibility of physically identical zombies.
But even with this standard example there are troubles, for if metaphysical possibility is (contra Kripke) more like conceivability23 then the answer to (Q6) may be, "yes, physically identical zombies are metaphysically possible." David Chalmers sometimes interprets the zombie question in this way (1996). Question (Q6) then becomes, "Might we have discovered that it is metaphysically possible for two things to be physically identical to one another but differ in that one is conscious and the other is not?" On Kripke's reading, that question has the same answer as question (Q3), whether physically identical zombies are logically possible; and they are indeed logically possible so long as dualism is logically possible.
The third question in the metaphysical possibility row of the Zombie Scorecard, (Q4), regarding behavioral zombies, received a positive answer thanks to a reasonably uncontroversial answer to the modally stronger question, (Q1) that behaviorally identical zombies are naturally possible. Can a similar strategy be employed to answer (Q6)? Suppose we assume a negative answer to (Q3): physically identical zombies are not naturally possible. Still, the negative answer to the question of the metaphysical possibility of physically identical zombies does not follow. The impossibility of the stronger claim, (Q3), tells us nothing about the possibility of the weaker claim at stake in (Q6). It takes a stronger claim than the denial of (Q3) to secure a negative answer to (Q6). It would take the force of necessity, as would flow from the claim of type-identity or logical supervenience with respect to consciousness.
As indicated, in view of the confusion over what exactly it is for something to be metaphysically possible, it seems best to set questions (Q5) and (Q6) aside for the time being. This is not particularly satisfying, but it's just as well. For Güzeldere (1995) has argued that the real zombie debate these days is over the remaining question, whether functionally identical zombies are naturally possible.
The Seventh Inning Stretch
All this wrangling over the details makes clear that, when it comes to zombies, the details matter. Imagining zombies is just a vivid way of forcing ourselves to face the consequences of views that we already hold. Some of the consequences of a view, or the caveats necessary to maintain it, may not be palatable to all philosophers. But those consequences and caveats don't come from the notion of a zombie, they are merely highlighted by asking the zombie questions. Güzeldere writes,
Belief in zombies has become a litmus test for intuitions in recent philosophy of mind.… The set of answers one chooses to give to questions of this sort is usually a good indicator of where one stands with respect to a variety of issues regarding consciousness-its ontology, nature, function, evolutionary role, and so on. (1995: 326-327)
Before going on, let's take stock of where we've been. If you have been following along with your Zombie Scorecard it should now look like this:
Figure 2. Zombie scorecard for (Q1), (Q3)-(Q8).
The Scorecard shows that all kinds of zombies are logically possible, behaviorally identical zombies are possible under all modes of possibility, physically identical zombies are not naturally possible, and we have no idea what to say about the metaphysical possibility of functionally or physically identical zombies. Whereas a large consensus allowed us to answer with some confidence questions (Q1), (Q4), and (Q7) through (Q9) with little difficulty, it became evident with (Q3), (Q5) and (Q6) that we must not so much expect to answer questions about zombies as to use the zombies to sort our intuitions and views. Answering these questions required us to take stances on controversial theories.
One position remains unexamined.
Questions (Q2). Is it naturally possible that there be two creatures that are functionally identical to one another but differ in that one is conscious and the other is not?
I'm not here going to defend the view that functionally identical zombies are naturally possible.24 To show that functionally identical zombies are naturally possible I would have to provide either an argument for Identity Theory or dualism (either of which entail the natural possibility of functionally identical zombies), or an argument against functionalism (which entails the natural impossibility of functionally identical zombies.)
As a matter of fact I am attracted to an Identity Theory. So on my view there might be some-many, innumerably many-systems that are (in some sense) functionally identical to human beings but that would lack consciousness. Although the ranks of Identity Theorists are growing, we are still the minority by far. So most philosophers of mind who fall roughly in the functionalist tradition (saying just what that is, of course, is a difficult matter) will reject the possibility of functionally identical zombies.
This is what is at issue between Dennett and myself, and there is no sense in recounting it here. But I will note that even functionalists ought to admit that there are some notions of function and functional equivalence according to which two things may be (in that sense) functionally equivalent but differ with respect to consciousness. Until they tell us which notion or notions they prefer, it is difficult to assess their claims with respect to the possibility of zombies. My suspicion is that many functionalists are helping themselves to a very broad notion of function-one that includes all causal relations, a view better called "mechanism" (Brandon 1996). In that case, when they deny the possibility of "functionally" identical zombies they are really denying the possibility of physically (i.e., causally) identical zombies. And that I can agree to, though I arrive at the view by a different route. Further exploring this suggestion is beyond the scope of the current Guide.
4. The Current Status of Zombies
For better or for worse, zombies are a hot topic in philosophy of mind. There are two particularly lively discourses. (Of course they are related.) The better-known strand stars David Chalmers and his numerous critics and supporters. It has been carried out in Chalmers' book The Conscious Mind, in The Journal of Consciousness Studies, and recently in Philosophy and Phenomenological Research. The lesser known strand, of which I am rather fond myself, stars Daniel Dennett. It began in The Journal of Consciousness Studies but has developed in a series of talks and book chapters.
The Chalmers strand of the zombie debates involves the logical and metaphysical possibility of physically identical zombies, and therefore the viability of materialism itself. It is fair to say that this strand is the direct descendent of Kirk's argument; indeed he has recently weighed-in again (Kirk 1999). Chalmers argues that physically identical zombies are metaphysically possible on the basis of a conceivability argument. His critics and supporters concern themselves with whether such zombies are possible, and also with whether that possibility is an obstacle to any form of materialism (physicalism, naturalism) as Chalmers argues.
The Dennett strand of the debates concerns the natural (and perhaps metaphysical) possibility of functionally identical zombies, and therefore the viability of certain forms of materialism (physicalism, naturalism). Dennett argues that if functionally identical zombies are naturally possible then consciousness must have no functional role in psychology, and must therefore be epiphenomenal or systematically mysterious. Dennett regards this consequence as a reductio ad absurdum of the very idea of zombies-if conceiving of zombies requires that consciousness is epiphenomenal, then the idea of zombies is to be rejected altogether. One reason that Dennett thinks this is that he takes all zombie supporters to be arguing in Chalmers' way, running conceivability arguments against materialism. Counter to this line of reasoning, I have maintained that views that allow for the natural possibility of functionally identical zombies do not render consciousness epiphenomenal; I offer Identity Theory as an example of one such view.
5. Timeline: The Evolution of Zombies
Here are a few milestones in the short life of zombies in philosophy of mind.
1975 Robert Kirk invents zombies
[Zombies in hibernation]
1991 Daniel Dennett discusses zombies in his Consciousness Explained.
[Later I argue that, far from being mere curiosities, zombies play a central role in Dennett's defense of his Multiple Drafts Model.]
1994 Todd Moody publishes "Conversations with Zombies," also in The Journal of Consciousness Studies.
1995 David Chalmers summarizes his arguments in a Precis to The Conscious Mind in The Journal of Consciousness Studies. A symposium on Moody's paper gives zombies some momentum, and launches the disagreement between Dennett and myself (with Owen Flanagan).
1996 Chalmers' The Conscious Mind is published. Zombies are a hot topic at the Tucson II conference. Chalmers mentions them in his keynote talk. Robert Kirk chairs a session including myself, Güven Güzeldere, and NNNN.
1998 Dennett's response to Flanagan and I, "The Unimagined Preposterousness of Zombies," appears in his Brainchildren anthology, garnering the attention of Scientific American and Salon, among others. I reply to Dennett with my talk, "Zombies Explained," at a conference in Dennett's honor in Newfoundland; Dennett responds.
1999 Dennett gives Royal Institute talk on zombies. A symposium on Chalmers' The Conscious Mind in Philosophy and Phenomenological Research discusses zombies.
2000-2001 John Perry's discussion of zombies in his latest book, no doubt many zombies at Tucson IV.
* Portions of this Guide are drawn from my "Zombies Explained" (2000).
1_C. B. Martin and M. Deutscher used the term in passing in a related way (1966), but clearly did not intend the consequences that Kirk explicates.
2_It is worth noting that Kirk, though he did not dwell on the permutations of zombie thought experiments, described or mentioned nearly all of the varieties that now have currency. More recent innovations are Dennett's "zimboes" (1991) and Chalmers' "dancing qualia" (1996).
3_Güzeldere considers behavioral, functional, and physiological zombies. I shall consider zombies that are physically identical to any arbitrary degree of specificity. The purpose of stipulating zombies in this way is to ensure that if the distinction between the physical and the functional can be maintained at all (and several have argued that it cannot; e.g., Lycan 1987, Hill 1991) then physically identical zombies are a distinct construct from functionally identical zombies. Specifically, the distinction holds even if the biological (physiological) structure of organisms is essentially functional in nature (as per Millikan 1989, 1993; Lycan 1987, 1996; Neander 1991).
4_This is supposed to be neutral about whether behavioral zombies have psychological states in general, and in particular about whether their movements qualify as actions.
5_For the present purposes behaviorism, functionalism, and Identity Theories, as well as other views mentioned in the discussion to follow, are considered as theories of consciousness. Although functionalism, for example, is often held as a theory of mind (cognition and intentionality, say) but not of consciousness (e.g., Ned Block 1980a), what is at stake here are just those functionalist theories that are intended to explain consciousness (e.g., William Lycan 1987, 1996).
6_See also Don Locke (1976).
7_Some philosophers, including some in this debate, take logical possibility to be a more robust notion akin to what I am calling metaphysical possibility.
8_This is what Terence Horgan calls physical possibility, which he distinguishes from a slightly weaker nomological possibility (1987). Complications about the locality of laws in space-time can be avoided; see remarks regarding Chalmers with respect to (Q3) (note ##, below).
9_Güzeldere suggests that an equivalent, and more convenient, way of talking about the ways that zombies could be identical is by considering them as distinct kinds of zombies: behavioral zombies, functional zombies, and physical zombies (1995). I will sometimes use this way of talking. So when I ask, say, whether behaviorally identical zombies are naturally possible, this is shorthand for question (Q1), "Is it naturally possible that there be zombies that are behaviorally identical to human beings?"
10_Kirk says he is using logical possibility in a loose sense, so it is more like what I am calling metaphysical possibility.
11_Kirk sees physical states as causally individuated. The theories he mentions as representing "materialism" are Lewis' and Armstrong's causal-specification theories. Whether these are to be classified as functionalist or Identity theories is an issue I set aside for the moment. (Lewis now says that he does not know if he is a functionalist; see his entry in Guttenplan 1994.)
12_Keep in mind that Dretske, once we make the caveat that I suggested, could deny that he allows for physically identical zombies.
13_It is a criticism that I level at Dennett that he persistently treats any zombie advocate as running a conceivability argument.
14_Chalmers has pushed me on this in correspondence.
15_But see Douglas Long (1977) for a contemporary argument against the coherence of disembodied minds. Hilary Putnam (1997) seems to suggest that while dualism may make sense in a religious context, it is not given a meaning in the scientific-cum-philosophical context.
16_It is very revealing, however, if what is at stake are certain views of semantics and reference, such as logical or analytical behaviorism, or the (capital-F) 'Functionalism' that Block (1980a) attributes to, for example, David Lewis. Putnam's X-worlders (super-super-Spartans) in his classic critique of logical behaviorism, "Brains and Behavior," are forerunners of the zombies (1968).
17_Notice that there is an overlap with the previous set of questions. Even if dualism turns out to not be logically possible, a "yes" answer to (Q7) can be secured by this second line of reasoning.
18_Adherents of some versions of behaviorism, however, would reject the possibility of behaviorally identical zombies. The logical or analytic behaviorist holds that mentalistic terms, such as 'consciousness', refer only to behaviors and dispositions to behave. An analytic behaviorist would therefore deny that behaviorally identical zombies are logically possible-and thus also deny that they are metaphysically or naturally possible. A metaphysical behaviorist might allow that behaviorally identical zombies are logically possible, but deny that they are metaphysically or naturally possible on the grounds that consciousness simply is behavior. These views do not figure prominently in the zombie debates.
20_This answer also depends on the caveat made to accommodate wide-content and historical theories.
21_Even if those fundamental bridge principles are local to time-space, Chalmers could at least deny that physically identical zombies are naturally possible at a location l and time t. If we allow comparison across space and time, so that the laws might change, then Chalmers view might allow for such a possibility as (Q3). But it is reasonable to construe the zombie debate as restricting the possibilities by time and space. For each imagined comparative question, the point could just as well be made by asking, "Could this thing here and now, such as myself, have different consciousness than it does in actuality have, while remaining identical in x way?"
This time-space restriction that allows proponents of views like Chalmers' to deny the natural possibility of physically identical zombies is compatible with the time-space permissiveness discussed with respect to Dretske's wide content view, above. In the wide-content case, we are considering how far into the past a things' history is relevant to our considerations. By including an organisms complete evolutionary history, we ensure that the wide-content materialist can deny the natural possibility of zombies. But in order for Chalmers to deny the natural possibility of zombies we need yet more-for it is compatible with his view that two things that are physically identical through their entire history differ with respect to consciousness if the fundamental bridge principles joining matter and consciousness were different. To counter this possibility we can either (i) insist that the fundamental bridge principles are universal in time and space for any given world, or (ii) make the comparison only between things that are identical both in their complete physical and nomological histories (including bridge principles.) The latter is accomplished, for example, by asking the zombie question in the first personal form. So the same form of the question that permits the Dretske-type wide-content qualia theorist to deny the natural possibility of physically identical zombies also allows Chalmers-type dual-aspect theorist to deny the natural possibility of physically identical zombies, by accounting for any effects of intra-world variation in fundamental bridge principles.
22_We need to be careful about how such cases are described. Chalmers sometimes denies that physically identical zombies are naturally possible. But in those cases he is invoking his new bridge laws, so it is really naturally* that he is discussing.
23_Kripke's (1972) "epistemic" possibility; Horgan's (1987) vague conceivability-but Horgan denies a connection between that kind of conceivability and metaphysical possibility.
24_Against Daniel Dennett I argue that the possibility of functionally identical zombies neither requires nor entails an epiphenomenalist conception of consciousness.