In the previous post, I laid out some of the terrain in the philosophy of mind. (If you have not read that post, it is important that you at least skim it before you read this one, unless you already have a background in the topic. Otherwise this post will not make sense.) In this post, I will present some arguments for property dualism. You will notice that similar argumentative moves take place in each one. I intend the zombie argument to be the main argument; the other two are, in a way, subsidiary arguments meant to provide further support for the zombie argument’s crucial premise.

The Zombie Argument

It used to be that basically all philosophers were at least property dualists. In the twentieth century, naturalism became the rage, and for awhile, basically all philosophers were physicalists. But in the past few decades, property dualism has experienced a strong resurgence; while it remains a minority position, there is no question that it must be taken seriously. This is largely due to the work of David Chalmers, who may be the single most prominent and talented philosopher of mind working today. (He is also an atheist, so you don’t have to worry about a hidden religious agenda.) Chalmers’ most famous argument for property dualism is essentially the one I will present in this section.

Here’s the argument. Consider that conceivability seems to be our best guide to possibility. I think that it was possible, in the broadest sense of the term, for the Netherlands to have beaten the US in the Women’s World Cup Final earlier this year. (To borrow a term from analytic metaphysics, I think there is a “possible world” where this happens.) So do you. Why? Well, we can coherently conceive of it. We can think of a series of events that leads to the Netherlands scoring more goals throughout the course of the game without running into any impossibilities. On the other hand, while I think that the Netherland’s top scorer, Vivianne Miedema, could have been a member of the winning team, I think she couldn’t have been a prime number. Why not? Well, it’s inconceivable. What would it be for her to be a prime number? It isn’t just that I happen to be unable to conceive of it, as I might be unable to conceive of all the details of some very complicated state of affairs. It seems inconceivable in principle. So, if something is conceivable, that seems to be evidence of its possibility; if something is inconceivable, that seems to show that it’s not possible, at least if it is inconceivable in principle, and not merely as a result of our contingent limitations. We’ll return to this shortly, but set it aside for now.

Now try something: conceive of a zombie world. Physically speaking, the zombie world is exactly identical to ours. There is a creature just like me, sitting at a computer just like mine, pressing just the same keys as I am. This creature is in exactly the same neurological states as me; everything lights up the same if he’s put in an MRI. But there is one very important difference between the two of us. He does not have any subjective experiences, any qualia. He is a philosophical zombie. There is nothing it is like to be him. And the same is true, we can add, for your zombie analogue, and my cats’, and everyone else’s, too. In the zombie world, there is nothing it is like to be anyone. The lights are on, but no one’s home. ^[1]

Now, ask yourself: is what I just described conceivable? Did some part of it sound contradictory or incoherent? I think it’s perfectly conceivable, and I think you will think it is, too, unless you have a prior commitment to physicalism and can see where I’m going with this. We understand perfectly well what it would be for the zombie world to be actual. We don’t seem to run into any impossibilities in thinking of the scenario; there don’t seem to be any entailments between the physical properties and the mental properties which are violated by supposing that the former are instantiated without the latter. Of course, we have not imagined all the nitty-gritty physical details of the zombie world, both because we don’t know all of them and because there are too many for us to keep track of all of them. But of course, we didn’t conceive of all the nitty-gritty details of the hypothetical game where the Netherlands beat the US, either. This wasn’t a problem because it’s pretty clear that those details could be filled in without introducing any incoherence into the scenario. And I think the same is true for zombie world. Filling in the little physical details is just not the kind of thing that will introduce incoherence into the scenario, given our understanding of what physical and mental properties are. (I’ll say more about this later.)

Suppose we grant that the zombie world is conceivable. And suppose, on this basis, we conclude that it’s possible. What follows? The falsity of physicalism. For if all the physical phenomena could obtain without any of the qualia obtaining, it follows that qualia are not physical phenomena. (If you think of something–rain, let’s say–which is an uncontroversially physical phenomenon, it will immediately be obvious that there could not be a world where the physical facts were the same but it never rained.) Phenomenal consciousness must be something over and above the physical.

Some Common Objections

People sometimes mistakenly think the zombie argument begs the question against physicalism. After all, if we assume physicalism is true, reasoning parallel to that above shows that zombies are impossible, so isn’t the argument just assuming what it needs to prove? This is misguided. The suggestion here is not that we start with the claim that physicalism is false, then reason our way to the possibility of zombies, and then use the possibility of zombies to circularly conclude that physicalism is false. The claim is instead that, whatever we start out thinking about physicalism, we are able to independently see (through conception) that zombies are possible, and therefore gain justification for thinking physicalism is false. (Suppose someone advanced the view that the property “winning the 2019 Women’s World Cup” just is the property “being the 2019 US Women’s World Cup team.” Clearly it wouldn’t beg the question to respond by noting that we can see that the Netherlands might have won instead, so that it’s possible to have one property without the other, so that they can’t be identical.) To refute the zombie argument, it is not enough for the physicalist to note that the premises are inconsistent with their position: it is a feature of any deductively valid argument that the conjunction of its premises are inconsistent with the negation of its conclusion. To defuse the argument, the physicalist instead needs to show what is wrong with the premises–to show either that zombies are not conceivable after all, or else that the inference from conceivability to possibility is somehow problematic here.

The second option seems to me to be more promising for the physicalist. Even if we grant that zombies are conceivable, you may be aware that conceivability does not entail possibility. To borrow a famous example from Saul Kripke: it’s possible to conceive of water being something other than H2O. For instance, some people used to think it was one of four fundamental elements. This is why we had to wait for science to figure things out in order to know what water actually was; otherwise, we could have just noted that every option except H2O was inconceivable. Yet, water is in fact just the same thing as H2O, and things could not fail to be the same thing as themselves, so it turns out that it is actually impossible for water to be something other than H2O, even though we can conceive of such a situation.

An opponent of the zombie argument might claim that the zombie world is like this: conceivable, but not really possible. What to say about this? One option would just be to say that conceivability is still evidence for possibility, even if it doesn’t entail it, and be content with that. Perhaps there are just limits to how confident we can be about modal judgments, or perhaps we can bring factors to bear other than conceivability to help decide them.

But I think we can go further. After all, sometimes conceivability really does seem to entail possibility–or at least to provide conclusive evidence for it. Consider: at present, forty-four individuals have been president of the United States, and all have been men. Many people (including me) think this is partly the result of sexism, and that, all else being equal, it would be good to elect a woman president. Suppose someone suggested that, for all we know, working towards this goal would be pointless, since it might be metaphysically impossible for anyone except a man to get elected. They grant that we can conceive of a woman being elected. But maybe there is some hidden impossibility in the scenario, just like in the scenario where water is something other than H2O–just like, they might add if they are a physicalist, in the scenario where there are philosophical zombies.

Well, that would be nuts. But why? The answer seems to be that we can conceive of a woman’s election in some kind of extra special way that rules out a hidden impossibility, or at least makes it really unlikely. So the question becomes what this extra special way is, and how we can know when we’ve achieved it. Proposals quickly become complicated. I don’t claim that I can provide the full answer here. But I’ll make a few remarks.

Our susceptibility to conceiving of things which are actually impossible–to failing to see hidden impossibilities–correlates with our ignorance of the nature of the things in question. Back in the olden days, we were in a position to know that we were ignorant about the nature of H2O. Indeed, while we could conceive of water’s being identical to H2O, we could also conceive of its being identical to something else, and we knew those couldn’t both really be possible. Now that we know water is H2O, we cannot coherently conceive of its being something else, provided we build that fact into the scenario we are conceiving. (We can, in some sense, conceive of water being something else, but it requires discarding our scientific knowledge. I could imagine, for instance, that a conspiracy of scientists tricked us all into thinking water was H2O when it was really something else. But my confidence that something like this isn’t the case–my confidence that we really do know the nature of H2O–is proportional to my confidence that scenarios where water is something else are impossible.)

On the other hand, our understandings of what it is to be a woman, and of what it is to be elected president, are clear in enough of the right ways for us to be able to see that one does not rule out the other. This is not to say that we have a perfect understanding of either. (Indeed, there is currently a big cultural debate over whether “woman” is fundamentally a biological category, or a psychological one, or a social one, or something else. And even people within each camp have many disagreements. So many people are to some extent incorrect about what it is to be a woman.) It is just to say that, again, however the details are filled out, we are in a position to see that they will not introduce some impossibility into the scenario.

I think our position with respect to the zombie scenario is more like this. We know perfectly well what kind of thing pain (say) is, being directly acquainted with it in our experience. We do not have a complete understanding of the physical world, but we do know enough about what kind of thing physical properties are (“structural dynamic” properties, in Chalmers’ words) that I think we are able to see that they don’t entail phenomenal properties in the way that would be needed to introduce an impossibility into the zombie scenario.

But maybe you’re still not convinced. Let me try two other moves. They might provide a kind of glide path into accepting the zombie argument.

The Inverted Spectrum Argument

Here is a slightly less extreme thought experiment. When you were a kid, you may have wondered whether the qualitative character of your experiences was the same as those of other people. People sometimes say things like “what if your red is my green?,” i.e., what if the qualitative experience I have when I look at something like grass is the same as the qualitative experience you have when you look at something like strawberries, and vice versa? (The reason people wonder about this is, of course, that we would still agree in calling grass green and strawberries red, unaware that we experienced different things when looking at them. This will be important in a minute.)

Thoughts like this help inspire inverted spectra scenarios. Start out by considering a world which, like the zombie world, is physically identical to ours, with creatures who look just like us. Happily, unlike the zombies, these creatures have phenomenal experiences. But their color qualia are correlated with physical states in a way which differs from ours, perhaps in one of the ways described here. (The link has some pictures which may be helpful.) The zombie argument asked us to consider a world where the correlations between physical states and conscious states don’t exist; this argument asks us only to consider a world where they are different. But if it is possible for these correlations to be different, then, again, qualia are not just physical states: if they were, it would not be possible for the physical states to be the same while the qualia were different. The possibility of zombies doesn’t immediately follow, but once one grants that an inverted spectrum is possible (and thus that qualia are non-physical), it isn’t obvious why one wouldn’t also grant the possibility of zombies. ^[2]

Furthermore, think back to that footnote about functionalism which I told you to read in the last post. I suggested that, if qualia are to be physically reduced, the best bet might be trying to identify them with functional properties. However, to show that qualia are not functional properties, we don’t need to stipulate that the creatures with inverted spectra are physically identical to us, only that they are functionally identical to us (even if their neurology is very different). And it does seem clearly possible that a creature–perhaps with a very different kind of brain than ours–could be in the same functional states as us while having different qualia. As Jaegwon Kim writes:

Pain as a sensory quale is not a functional property. In general, qualia are not functional properties… [this can be seen by considering] the qualia inversion hypothesis, namely the possibility of creatures… whose quality space is inverted with respect to ours—who, for example, when they look at mounds of lettuce, experience a color quale of the kind we experience when we look at ripe tomatoes, and who, when they look at ripe tomatoes, sense the color that we sense when we look at lettuce. Such spectrum-inverted people would be as adept as we are in picking tomatoes out of mounds of lettuce and obeying traffic signals, and in general they would do just as well as we do with any other tasks requiring discrimination of red from green. If this is the case, color qualia do not supervene on behavior; two perceivers who behave identically with respect to input applied to their sensory receptors can have different sensory experiences. If that is true, qualia are not functionally definable.

But if functionalizing qualia is the physicalist’s only bet, and qualia cannot be functionalized, so much the worse for physicalism. Kim therefore continues:

So qualia are not functionalizable, and hence physically irreducible… This means that global physicalism is untenable. It is not the case that all phenomena of the world are physical phenomena; nor is it the case that physical facts imply all the facts. There is a possible world that is like this world in all respects except for the fact that in that world qualia are distributed differently. I don’t think we can show it to be otherwise.

The Homunculi Argument

Here is a final argument; I include it partly on the grounds that it’s kind of fun. The exact use made of it here is again Chalmers’, though it involves a thought experiment most famously associated with Ned Block. Suppose some villain cuts my brain out of my skull and throws it in the trash. They replace it with a special transmitter. Meanwhile, they get together billions and billions of people and give each one a radio. Each person is given a role corresponding to a particular neuron in my (now trashed) brain. They make calls to each other on their radios to simulate neurons firing, transmitting information to one another just as my neurons would have. When something harms my body and nerve signals indicating this reach my head, the special transmitter sends a message to the people standing in for my sensory neurons; a pattern of “neuronal” activation plays out which is exactly like what would happen in my brain if the same nerve signals were received; the people standing in for my motor neurons then transmit signals to my head, and the transmitter sends out nerve signals exactly like those my brain would have sent. And so on for everything else, so that my body goes around doing exactly the things I would have done (such as writing this post).

Now: is this whole system, with all these billions of people, conscious? When my body is damaged, is there an experience of pain, even though none of the particular individuals in the “neuron” network are feeling any pain? Most people think not. But even if you’re not sure about that, perhaps you’ll grant that it’s at least possible that this system is not conscious. But if that’s right, qualia are not functionalizable. The people with the radios are functionally equivalent to my brain: when they get the same inputs, they produce the same outputs. So if it is possible that the system is not conscious, there is a possible difference in qualia without a difference in function.

Further, not only are the people with radios functionally equivalent to my brain, they are also mimicking its internal structure. The only difference is that, while my brain is made of neurons, the replacement is made of people with radios. While this is quite a difference, it doesn’t seem like the sort of difference that could explain why consciousness is metaphysically necessary in the one case and not the other. As Chalmers notes:

…given that it is conceptually coherent that the group-mind set-up… could lack conscious experience, it follows that my zombie twin is an equally coherent possibility. For it is clear that there is no more of a conceptual entailment from biochemistry to consciousness than there is… from a group of homunculi… Nothing in this substitution [of people with radios for neurons] could force experience into the conception; these implementational differences are simply not the sort of thing that could be conceptually relevant to experience. So consciousness fails to logically supervene on the physical.

This seems quite plausible to me. If it is really possible that the system in the above thought experiment isn’t conscious, the zombie world is also possible, and from that follows the robust property dualism I was after.

So: that’s it for property dualism. Phenomenal consciousness is not reducible to the physical. But what am I, the subject of phenomenal consciousness: am I some material object, like a body or a brain, or am I an immaterial soul? I address this in the next post.

Footnotes