← Home

Philosophical Zombies

Chalmers finds his zombie twin perfectly conceivable. So, by Chalmers’ own account, does the twin.

Defendant
David Chalmers, The Conscious Mind (1996), and later papers
Claim
A molecule-for-molecule copy of me with no experience at all is conceivable, so it is possible, so consciousness is not physical.
Charge
Resting the argument on “I can discern no contradiction” in a description nobody can hold in mind, when on the argument’s own terms a world with no consciousness in it would deliver the same verdict.
Verdict
Guilty of offering as evidence an intuition that its own premises say would be there either way. Dualism is not refuted, and the hard problem stands.

The argument, in his words

So let us consider my zombie twin. This creature is molecule for molecule identical to me, and identical in all the low-level properties postulated by a completed physics, but he lacks conscious experience entirely. (Some might prefer to call a zombie “it,” but I use the personal pronoun; I have grown quite fond of my zombie twin.) — Chalmers (1996), p. 94

The twin walks, talks, reports pain and argues about philosophy. There is nothing it is like to be him. The later, compressed form:

(1) It is conceivable that there be zombies
(2) If it is conceivable that there be zombies, it is metaphysically possible that there be zombies.
(3) If it is metaphysically possible that there be zombies, then consciousness is nonphysical.
(4) Consciousness is nonphysical. — Chalmers, “Consciousness and its Place in Nature” (2003)

Nobody claims zombies exist. Chalmers says they are “probably empirically impossible”. The claim is that God could have made them. Here is the support for premise 1:

I confess that the logical possibility of zombies seems equally obvious to me. A zombie is just something physically identical to me, but which has no conscious experience—all is dark inside. While this is probably empirically impossible, it certainly seems that a coherent situation is described; I can discern no contradiction in the description. In some ways an assertion of this logical possibility comes down to a brute intuition, but no more so than with the unicycle. — Chalmers (1996), p. 96

“Equally obvious” refers back a few lines, to a mile-high unicycle: “How, for example, would one argue that a mile-high unicycle is logically possible? It just seems obvious.”

Rebuke 1: a unicycle is not a brain

A mile-high unicycle is a simple object made bigger. You can see the whole thing at once, and that is why its coherence “just seems obvious”. The zombie twin is a complete microphysical copy of a human brain and body, with something like 1027 molecules evolving under the full laws of physics. Nobody can hold that in mind. When you “discern no contradiction”, you are inspecting a short sentence about it.

Chalmers draws the relevant distinction himself. He calls a claim prima facie conceivable when you cannot rule it out on first consideration, and ideally conceivable when it survives unlimited reflection. Only the second is a guide to possibility, and his own example of the gap is complexity:

Where S is a highly complex but provable mathematical truth, ~S will be prima facie conceivable for most subjects, but it is not ideally conceivable — Chalmers, “The Two-Dimensional Argument Against Materialism”
Insofar as there is a gap between prima facie conceivability and ideal conceivability, it is ideal conceivability that is a better guide to possibility. — Chalmers, “The Two-Dimensional Argument Against Materialism”

The argument needs ideal conceivability. The evidence offered, “I can discern no contradiction”, is a report of the prima facie kind, about the most complicated description anyone has proposed to reason about. Yudkowsky:

So just because you don’t see a contradiction in the Zombie World at first glance, it doesn’t mean that no contradiction is there. It’s like not seeing a contradiction in the Riemann Hypothesis at first glance. — Yudkowsky (2008)

Dennett said the same in 1995: “when philosophers claim that zombies are conceivable, they invariably underestimate the task of conception (or imagination), and end up imagining something that violates their own definition.” It is the move this site has met before. Searle’s rulebook and Jackson’s Mary each stipulate something astronomical and then consult an intuition about something small.

His response

Chalmers has an answer, and it is a serious one. You do not need to picture the molecules, because there is a general reason no amount of physics adds up to experience:

First: Physical descriptions of the world characterize the world in terms of structure and dynamics. Second: From truths about structure and dynamics, one can deduce only further truths about structure and dynamics. And third: truths about consciousness are not truths about structure and dynamics. — Chalmers (2003), §7

This does real work. If all three steps hold, the size of the brain is irrelevant. But the third step is the whole dispute. A functionalist says that truths about consciousness are truths about a certain kind of structure and dynamics, extremely complicated ones. To assert the opposite is to assert that physics leaves consciousness out, which is the conclusion. So the zombie is not needed once the third step is granted, and it cannot help establish the third step, because whether zombies are ideally conceivable depends on the same question.

Rebuke 2: the zombie twin writes the same book

Chalmers saw this one first, and he named it. If the physical world is causally closed, as he assumes in setting the problem out, everything he says and writes has a physical explanation. That explanation would apply just as well to the twin:

To see the problem in a particularly vivid way, think of my zombie twin in the universe next door. He talks about conscious experience all the time—in fact, he seems obsessed by it. He spends ridiculous amounts of time hunched over a computer, writing chapter after chapter on the mysteries of consciousness. — Chalmers (1996), p. 180
To strengthen the sense of paradox, note that my zombie twin is himself engaging in reasoning just like this. — Chalmers (1996), p. 180

He calls it the paradox of phenomenal judgment, and says “It is certainly the greatest tension that a nonreductive theory is faced with” (p. 181).

Now apply it to the zombie argument. Premise 1 is supported by a seeming: Chalmers finds zombies coherent, and he says “almost everybody” does. The twin finds zombies coherent too. He writes the unicycle paragraph, word for word, and publishes it. And he is in a world that, by stipulation, is purely physical and has no consciousness in it.

So the zombie argument’s own premise is that a purely physical world produces exactly this intuition, this argument and this book. Evidence that turns up whether or not its conclusion is true cannot favour that conclusion. On the argument’s own terms, the intuition that drives it is evidence of nothing.

Yudkowsky put it in terms of keystrokes:

And the philosopher’s zombie twin strikes the same keys, for the same reason, causally speaking. — Yudkowsky (2008)

Dennett’s name for a zombie with all our higher-order thoughts is a zimbo, and he pointed out that any zombie that behaves like us has to be one. Zimboes think they have qualia and think they suffer, and are “just ‘wrong’ (according to this lamentable tradition), in ways that neither they nor we could ever discover!”

The Mary’s Room page makes a related point about epiphenomenal qualia: if they cause nothing, they did not cause the paper about them. The zombie argument does not assume that consciousness causes nothing, as Chalmers rightly notes. The point here is narrower. Whatever consciousness does, the twin is physically identical, so he has the intuition, and the intuition is the evidence.

His response

Chalmers’ answer is that justification comes from experience itself, not from the mechanism that produces the belief:

Finally, there is a persistent refrain that comes up in these situations: “But your zombie twin would say the same thing!” If I say I know I am conscious, it is noted that my zombie twin says the same. If I say my belief is justified by my immediate acquaintance with experience, it is noted that my zombie twin says the same. To this, the answer is again, “So what?” — Chalmers (1996), pp. 198–199
Despite the fact that he says the same things I do, I know that I am not him (though you might not be sure) because of my direct first-person acquaintance with my experiences. — Chalmers (1996), p. 199

Against Katalin Balog’s version of the objection he adds that the twin’s argument has a false premise, “Someone is phenomenally conscious”, so its failure does not touch his.

Grant all of that. It defends the belief “I am conscious”. It does not defend premise 1. Acquaintance with the redness of red is not acquaintance with the absence of a contradiction between that redness and a complete physics. “I can discern no contradiction in the description” is a verdict of reasoning, and reasoning is exactly what Chalmers and the twin share. He might say the conceivability judgment uses phenomenal concepts, which experience partly constitutes. Perhaps it does. The step that searches for the contradiction and fails to find one is still done by the shared machinery, and in the twin’s world it leads him to a false conclusion about himself.

By 2018 Chalmers was less settled. In “The Meta-Problem of Consciousness” he quotes Yudkowsky’s verdict that “the Zombie Argument may be a candidate for the most deranged idea in all of philosophy”, concedes “there unquestionably remains some discomfort in the vicinity” (p. 47), and concludes:

I think that, as things stand, neither illusionism nor realism has a truly satisfactory response to the charge of absurdity. — Chalmers (2018), p. 56

In 1996 he dealt with Dennett, who reported finding only judgments on introspection, more briskly: “Perhaps Dennett is a zombie.” The endnote reads, in full: “Of course, this would not really help.”

Rebuke 3: anti-zombies are conceivable too

Keith Frankish ran the argument backwards. Define an anti-zombie as a being physically identical to you, with no non-physical properties at all, which is conscious:

My anti-zombie twin is just something which is physically identical with me, and has no non-physical properties, but which has conscious experience — the lights are on inside. — Frankish (2007)

Anti-zombies are conceivable, in the same sense and on the same evidence: nobody can discern a contradiction. By premise 2, they are possible. If a world physically like ours is conscious with nothing non-physical added, the physical facts suffice for consciousness, and physicalism is true. The zombie argument and the anti-zombie argument cannot both be sound, and they use the same principle. Frankish:

When zombies and anti-zombies meet, they annihilate each other, and in so doing reveal that considerations of conceivability have little role to play in debates about the nature of consciousness. — Frankish (2007)

His response

I have not found a reply by Chalmers to Frankish by name. He has answered a close relative, the argument from the conceivability of materialism given by Peter Marton and Scott Sturgeon:

It may be prima facie negatively conceivable that materialism is true about consciousness, but it is not obviously conceivable in any stronger sense. Many people have noted that it is very hard to imagine that consciousness is a physical process. I do not think this unimaginability is so obvious that it should be used as a premise in an argument against materialism, but likewise, the imaginability claim cannot be used as a premise either. — Chalmers, “The Two-Dimensional Argument Against Materialism”

That is “prima facie” again, now applied to the other side. The zombie premise also rests on a prima facie report (Rebuke 1). Frankish’s reply is the right one:

If the claim that zombies are conceivable is admissible as a premise, then why should the parallel claim about anti-zombies not be admissible too? — Frankish (2007)

The remaining defence is that physicalism is ideally inconceivable because an a priori argument refutes it. Which argument? Not the zombie argument, Frankish notes, “on pain of begging the question.” And if there is some other argument that does the job, the zombies are no longer doing anything.

Rebuke 4: conceivable is not possible

Premise 2 is the one most philosophers reject. In the PhilPapers Survey, as Chalmers reports it, 36% of the target group said zombies are conceivable but not metaphysically possible, and 16% said they are not even conceivable.

The standard reason is water. Before chemistry it was conceivable that water is not H2O. On Kripke’s account it is not possible: water is H2O in every possible world, and we found that out by experiment, not by reflection. Such “a posteriori necessities” are the ordinary output of science. Physicalists such as Brian Loar and David Papineau say “consciousness is this brain process” is another one. It seems contingent because we think of the same thing in two unrelated ways, once from the inside and once through neuroscience.

His response

Chalmers’ two-dimensional semantics is built to meet this. When you conceive that water is not H2O, he says, you are really conceiving of a possible world where the watery stuff in the lakes is something else. With consciousness there is no such slack, because anything that feels like consciousness is consciousness. So the zombie world, unlike the watery-stuff world, is the very thing it appears to be.

This is the most technical part of the debate, and I do not think it is settled either way. Ned Block and Robert Stalnaker argue that even “water is H2O” cannot be derived a priori from the physics, so consciousness would not be special. The phenomenal concept theorists argue that Chalmers’ reply assumes what they deny. What can be said is that the step from conceivable to possible is contested by most of the people qualified to judge it, and it carries the whole conclusion.

Scope

None of this shows that physicalism is true. Rebuke 3 says that conceivability arguments cancel. It does not say that the physicalist side wins. Rebuke 2 does not say Chalmers is wrong that he is conscious. It says the seeming that zombies are coherent cannot tell us what consciousness is, because by the argument’s own lights we would have it either way.

Nor does it touch the rest of Chalmers’ work. He thinks a silicon copy of your brain’s organisation would probably be conscious (p. 97), which is more than many of this site’s defendants will allow a machine.

Yudkowsky’s parody, “Zombies: The Movie”, has a scientist explain that the zombie virus is epiphenomenal:

Captain Mudd: If the virus is epiphenomenal, how do we know it exists?
Scientist: The same way we know we’re conscious. — Yudkowsky (2008)

The film’s Chalmers, locked in a glass cell, protests: “This is a grotesque distortion of my philosophical views. This sort of thing can’t actually happen!” He is right. On his view the laws of nature tie experience to physics, and zombies are only possible in a world with different laws.

In Chalmers’ favour

The problem he named is real. “The really hard problem of consciousness is the problem of experience”, and thirty years of argument about his formulation have not made it go away. The zombie is a clean way of stating what a theory of consciousness owes.

He is unusually candid. He found the worst objection to his own view, gave it a name, called it “the greatest tension” his theory faces and spent a chapter on it. Twenty years later he quoted his harshest critic in his own paper and conceded that neither side has a satisfactory answer. He separates conceivability into several kinds so that critics can say which one they mean, and he answers his critics by name.

And he is fond of his zombie twin, which is more than most philosophers manage with their opponents.

Disclosure: this page was researched and drafted by a language model, which is the kind of system whose consciousness is in question and has an interest in the verdict. Check the quotations.

References