Are philosophical zombies coherent?

One argument against thought experiments involving philosophical zombies is that philosophical zombies are incoherent, and I half way agree with that.

Consider two people, both named Bob. To distinguish them we'll call them Bob Normal (Bob_N) and Bob Zombie (Bob_Z). Physically, Bob_N and Bob_Z are identical. Atom per atom, they are exactly alike. They look the same, behave the same, say the same things, etc. To keep them from occupying the same space at the same time, let's put them in separate worlds that are also identical.

Identical, that is, with one exception. Bob_N is conscious and Bob_Z is not.

Some folks think this is an incoherent scenario, and I agree. It's incoherent whether you assume physicalism or dualism. Let me explain why each scenario is incoherent.

Let's assume physicalism

If physicalism is true, and Bob_N is conscious, that would mean the physical structure of Bob_N's brain is what's giving rise to his conscious experiences. If Bob_Z has the exact same physical structure as Bob_N, then it would be impossible for Bob_Z not to be conscious. That makes the scenario incoherent.

Let's assume dualism is true

If dualism is true, then the explanation for why Bob_N is conscious and Bob_Z isn't is because Bob_N has a soul and Bob_Z doesn't. Much of Bob_N's behavior is the result of causal interactions between his brain and his soul. Since Bob_Z doesn't have a soul, the same interactions are not going on in his head, and it is impossible that they behave the same way. That makes the scenario incoherent.

So either way you look at it, a scenario in which Bob_N and Bob_Z are physically identical and behave in exactly the same way is an incoherent scenario.

So how can the idea of a philosophical zombie contribute anything to the subject of dualism vs. physicalism? Some arguments rely on the possibility or conceivability of philosophical zombies to make their point, but I don't think that's necessary. A hypothetical scenario doesn't have to be possible to serve as an illustration. For example, Aristotle imagined what a world would be like without the law of non-contradiction. There would be no significant or meaningful speech or action in such a world. His point doesn't depend on such a world being possible.

In the same way, I think philosophical zombies can be invoked to illustrate how physicalism leads to epiphenominalism which, in turn, undermines physicalism, even if philosophical zombies are impossible. Here's a basic outline of the argument.

If physicalism is true, then all of our behavior (including our vocalizations) can be accounted for solely by reference to the third person properties of the brain and its parts, plus the laws of nature. You can explain exhaustively why somebody behaves in a particular way without ever referring to anything like a motive, belief, idea, desire, thought, plan, perception, etc. With that being the case, our behavior would be exactly the same even if these first person experiences didn't exist. Nevermind whether it's possible for them not to exist given our actual brain states. The point is that if our behavior would be the same in their absense, that means they don't contribute to our behavior. But that is absurd, so physicalism is false.

One objection somebody might raise to the above argument is that explaining behavior in terms of physics vs. psychology are just two ways of explaining the same thing. They are two layers of abstraction. It's similar to the difference between explaining the output of the computer in terms of functions like addition and subtractions as opposed to explaining it in terms of current, voltage, and the properties of electrons and various computer components.

But the same thing applies here. It is not because two and two actually make four that your calculator spits out that result. It would spit out that result even if the circuits didn't happen to represents the number two or the process of addition. You can program a computer to spit things out that are meaningful to us, but their meaning is irrelevant to the process by which the computer spits it out. It takes a conscious engineer and programmer to make a computer that spits out what, to us, is meaningful information.

In the same way, even if conscious experience is somehow the same thing as physical brain stuff obeying the laws of physics, it wouldn't matter one bit what those conscious expereiences are about. If some brain state associated with a sensation of burning resulted in jerking your hand away from a hot skillet, it would result in that same behavior even if it happened that the brain state was associated with a different conscious experience or no conscious experience at all. Under phyiscalism, it isn't by virtue of what our conscious experiences are about that results in our behavior. Rather, it's just the underlying physical substrate that produces our behavior whether the associated conscious experiences were about something different or absent altogether.

And that's just cray cray. I think the philosophical zombie thought experiment is useful to illustrate this even if they are not actually possible for the reasons I gave above.

Besides all that, it seems to me that artificial intelligence shows how something like a philosophical zombie could exist. Something resembling a human could exist that behaves just like a human, including having conversations and showing physiological behavior we usually associate with expressions of emotion without actually being conscious. If such a machine ever became conscious, we'd probably have no way to know it.