Sunday, August 06, 2006

The Problem of Other Minds

We all know ourselves to be conscious. But what of everybody else? They might be very complex automata (or "zombies") and we'd never know it. The obvious response is to argue from analogy: I'm conscious, you're relevantly similar to me, so you're (probably) conscious too. Filling this out, we posit psycho-physical natural laws, or universal regularities, which link physical and phenomenal properties (or brain states and mental states). Given the natural laws, zombies become strictly impossible: anything physically just like me must likewise be mentally identical.* But I wonder: is this solution open to substance dualists?
* = Zombies remain metaphysically possible because the natural laws could have been different. If you value your inner lights, be thankful they aren't!

It's not entirely clear what substance dualism involves. But here's one possible view in the vicinity: brains and other physical stuff are (even nomologically) insufficient for consciousness. To be conscious, one needs special soul-stuff, bestowed perhaps by a supernatural santa deity. If scientists constructed a synthetic atom-for-atom replica of a human, it would not have a mind (unless God intervened to give it one). Zombies are thus nomologically possible.

Now, that view sounds crazy. But it also sounds a lot like the sort of thing the average theist in the street might say. (Right? Or am I being uncharitable here? Average street-going theists have been known to say many a crazy thing...) Anyway, the particular problem I want to highlight is that this kind of substance dualist cannot respond to skepticism about other minds by appeal to the argument from analogy I offered above. The "relevant similarity" is no longer merely physical/behavioural. These latter properties offer the dualist no evidence whatsoever that there is anything mental going on in addition, since the evidence is equally consistent with zombiehood, which is a possibility they're just as happy to grant.

Is there any escaping this problem for the dualist? If we grant the general (nomological) possibility of zombies, then it's hard to see any basis for ruling out the possibility in particular cases (excepting our own, of course). Perhaps you could appeal to a benevolent God -- but wouldn't he only grant consciousness to lives that are worth experiencing? The theist should then expect that people apparently suffering outrageous hardship are in fact zombies. (Have I hit upon a novel solution to the problem of evil? Solipsism!)

But perhaps substance dualists are not committed to the nomological possibility of zombies? I'd need an explanation of what their view involves instead, then. (I have trouble making sense of the claim that physical substance could nomologically necessitate the existence of corresponding mental substance. Would they really be two distinct substances in that case?) Do tell...


  1. "If scientists constructed a synthetic atom-for-atom replica of a human, it would not have a mind (unless God intervened to give it one). Zombies are thus nomologically possible."

    I'd think that most theists and substance dualists would think that an atom-for-atom replica would simply be a body and bodies aren't zombies. Bodies don't engage in any activities but laying about, no walking, acting human like, etc. Dead bodies are basically what you'd described. Certainly dead bodies are nomologically possible, but they aren't zombies. At least I think they say something like this. So I think substance dualists are not committed to the nomological possibility of zombies

  2. Yeah, either a soulless body would be dead or it would act differently from a normal, ensouled body. It wouldn't pass the Turing test, and it wouldn't qualify as a philosphical zombie. At least that's what I'd expect to hear from a "theist on the street." Paul Bloom has more to say about people's dualist intuitions.

  3. Maybe we can distinguish mechanical behavior and behavior of free will.
    I guess substance dualist would have to accept that any willful behavior in certain and limited (in time, space, possibilities etc..) circumstances can be copied in mechanical behavior. I agree with Blar and Matthew that substance dualist (or street theist) probably won't believe that the body IS such mechanism which can produce/copy the same behavior, but let's say that maybe they do...
    To me it seems that in such cases the dualist will have to say that the distinction is not in the behavior itself, but in how that behavior came to be, or so to say that the distinction is in the cause of the behavior. In the case of the conscious humans, it will be the will (the second substance), while in the other (zombie) case it will be merely a mechanical behavior, without will... just a mechanical system undergoing lawful changes. But this "spiritual-substance" which can't be result of behavior then must have somehow appear as second to the material-substance, and by being spiritual-substance, I guess it can't be caused by material-substance.
    It seems to me, that the dualist could try to defend from other-minds skepticism, in this moving from analogy of behavior to analogy of causes. So, he could say... it is not that we all behave the same, but that we all came to the world in same way, that it is plausible that we all are part of the same causal logic.

    I guess this argument might have ended little confused, but after all I'm not dualist, so I did my best.

  4. > The theist should then expect that people apparently suffering outrageous hardship are in fact zombies.

    there could be somthing to this...
    It seems to agree with a reasonable amount of real life experience.

  5. Richard, your solution to the problem of evil is excellent; God could presumably even remove consciousness from conscious subjects when bad stuff goes down, and return it afterwards! Sounds like fun to flesh out that view....

    I'm not sure the substance dualist really has any special problem here though. First, the argument from analogy has usually been presented as an inductive inference. But if we knew that mental states nomologically necessarily supervene on physical states we wouldn't need induction at all. To put it the point another way; why suppose that nomologically necessary supervenience (or knowledge thereof) is necessary for the inference to confer knowledge?

    Secondly, arguing from analogy is often offered to justify, e.g., my everyday belief that those around me are minded. So put, it takes as the relevant similarity not that they have the same physical make up as me, but that they behave in the same ways I do. But to claim that there mental states nomologically necessarily supervene on brain states doesn't yet give you a link between behaviour and mental state. So if you've raised a problem for the substance dualist, it might also be a problem for some non-substance dualists (although not guys how accept the nomologically necessary supervenience of the mental on behaviour).

    And thirdly, the dualist could just reject the need for argument from analogy in the first place. Here's one way: how do I know that the guy outside my window is minded? Well, I see that he's happy, and if you're happy, you're minded. And of course, I couldn't see that if he were a zombie (i'm ignoring issues about whether being minded is sufficient for being conscious).

  6. "Dead bodies are basically what you'd described."

    Interesting... I guess you're right that folk would initially want to say something like that. But I meant to stipulate that all the same physical (hence chemical and biological) processes are going on as in the "original" person. So the synthetic body does move and all that: the mechanism/behaviour is indistinguishable from a normal person.

    If the "street theist" insists that vital functioning cannot be physically explained, then they would seem to be suffering from mere empirical ignorance -- cf. the discussion of vitalism here. So perhaps I should set them aside for now. Besides, as Tanasije points out, it's hard to see how the possibility of sufficiently complex mechanical behaviour could plausibly be denied (even if they don't think the human body is itself such a mechanism).

    So, assuming that physical mechanisms/automata *can* exhibit complex behavioural functioning (on a par with our own), what is this more informed substance dualist going to say about the problem of other minds?

    Tanasije proposes to analogize from origins. But all we know about others is that their body has a similar origin to ours (i.e. conception, etc.). But presumably theistic dualists hold that there is a second aspect to *our* origins, which is God's granting us a soul. Wouldn't it beg the question to assume that others share this aspect of origination with us? Whether they have souls or not is precisely the issue in question...

  7. Hi Russ, I take the positing of natural laws here to be an inference to the best explanation of why we are conscious. So the analogy is less direct: once the laws are posited, the behaviour of others suggests to us that they probably have the right sorts of internal brain states to figure in the laws that give rise to mental states. Or something like that.

    Your last point sounds awfully fishy. The qualitative visual input you receive is not affected by whether the guy is a zombie. So if "seeing that he is happy" implies that he's minded/conscious, then we have no evidence that you *truly do* see that he is happy. Perhaps you merely see that he is acting as if he were happy (but it's just an act, because he's a zombie). You think you see his happiness, but this visual representation is non-veridical. There is no happiness there to see -- merely behaviour.

  8. Hmmm; I think I see the force of the charge a bit better... the non-substance dualist postulates the natural laws (by inference to best explanation) to get him from the physical evidence to the mental states, and you're questioning whether the substance dualist has anything... in which case why can't the substance dualist just postulate some modal connection between the physical evidence and mental states that falls short of nomological necessity? Such a connection is going to be entailed by the non-substance dualist's position, so he can't deny it's true. Is the thought that one isn't justified in such a postulation unless one goes all the way to nomological necessity?

    Re the last point, I'm not sure it's so fishy :) In the case where you see an unhappy (non-happy?) zombie acting as if he's happy, you can't see that he's happy, but it might be to you as if you see that he is happy. But I don't see why that stops you seeing that non-zombies are happy. It seems to me to be a datum that we do sometimes see that people are happy; at the very least, it's highly revisionary of our pre-theoretical views to deny this. I'm not really sure what to say about whether the qualitative visual input in both cases is the same: I guess the same distribution of photons is reflected from zombie-guy and non-zombie guy... but if visual input is a matter of what one sees (or sees to be the case), then the visual input is different.

  9. Richard,

    I don't see how this problem (if it's that) is only a problem for substance dualism. (But maybe you meant to pose it as any issue for all forms of dualism—including emergent types—since the fourth paragraph of your entry asks, "Is there any escaping this problem for the dualist?" However, since you posed it earlier as a problem for substance dualists specifically, it isn't exactly clear to me at whom you really meant to aim the issue.) If you're assuming, rightfully, that there's a difference between being conscious and being a zombie then, it seems to me, knowing for certain whether anyone else is not simply a zombie (or complex automata) is an issue for any dualistic position. Even emergent, non-reductive views hold, if I'm correct, that a computer isn't conscious.

    I'm not sure about the "special soul stuff" you say is required of (substance?) dualism. But if so, then I guess emergent-type views require "special complex stuff" if one is to claim that a computer isn't complex enough to be conscious, yet a brain is. (Personally, it seems to make more sense to say the physical and the mental are separate substances than to say that one can rearrange the physical just right and "Walla!"—mental; but that's a different subject.) I agree with the last paragraph of your entry. I just don't think that what you have in parenthesis there describes substance dualism.

    The "other minds" issue seems to be an epistemological, not a metaphysical, issue. So it wouldn't be a problem specifically for substance dualists since that is a metaphysical stance. In fact, it would be a problem, it seems, for everybody (other than solipsists and eliminativists).

    I have to confess, Richard, that I don't understand how your "natural laws" postulation has any relevance. Are you suggesting that in a non-substance dualist world complex automata, as automata, can't possibly exist? If not then I don't see the point. The non-substance dualist would still be left with the same possibility as the substance dualist—namely, that all other similar beings are actually complex automata. (The only escape from such an epistemological jam is one of two ways: one, by way of the solipsist, who insists that the only consciousness that exists is one's own—thus there are no other minds; two, by way of the eliminativist, who claims that consciousness simply doesn't exist—thus there are no minds whatsoever—but rather that all that exists is complex automata.)

  10. I take it that the substance dualist does not (usually) claim that life, i.e., ordinary vital functioning, cannot be completely physically explained, but that rational consciousness can't; and that rational behavior (not functioning) suggests rational consciousness as its explanation. In that sense, I don't see that the substance dualist has much of a problem. The analogical argument is that my acting as I do is due to mind, and his acting similarly probably has a similar cause; and it doesn't appear that this requires anything but the legitimacy of causal analogies in general. It isn't clear that zombies are nomologically possible for most substance dualists -- the fact that something acts like it has a mind is simply taken as evidence that it has one, so it isn't clear what could possibly be evidence that a thing acting like it has a mind doesn't (unless it did something that involved acting like it didn't have a mind, i.e., gave itself away).

    The Other Minds Skeptic brings up a case of perfect simulation; but, like Don, I don't think this is a problem only for the substance dualist, since it would be an issue in any case where analogical reasoning is involved. After all, if you and I are sitting across the table, how do you know that I don't have a completely hollow head, and am doing what I do without a brain at all? Well, you know it by (very good) analogical reasoning from what you know of brains and behavior. And likewise if we were engaged in an online chat. The fact that a perfect simulation is possible, however (e.g., that you are chatting with a cleverly designed program rather than someone with a brain and mind) doesn't change the fact that your inference is a very good one -- all it shows is that your inference is defeasible. And that's not a problem.

  11. "It's not entirely clear what substance dualism involves. But here's one possible view in the vicinity: brains and other physical stuff are (even nomologically) insufficient for consciousness. To be conscious, one needs special soul-stuff, bestowed perhaps by a supernatural santa deity. If scientists constructed a synthetic atom-for-atom replica of a human, it would not have a mind (unless God intervened to give it one). Zombies are thus nomologically possible."

    What about the following substance dualist view though, the mind is a seperate, non spatial, with the power to influence the brain thing but the brain automatically creates the mind, God knows why, something to do with it's strucutre perhaps? Then the synthetically created zombie would be nomologically but not conceptually impossible.

    Of course the view is ontologically messy, but the dualist will say that the need to match the evidence trumps ontological parsimony.

  12. Do substance dualists really endorse vitalism? I think it would be interesting to do a survey...

  13. Just as a note, Alvin Plantinga actually sees the possibility of other minds as evidence for the existence of God (See God and Other Minds). This is developed in his position of Reformed Epistemology. Plantinga argues that it is just as rational to believe in the existence of God as it is of other minds, while neither can be proven.

    Also, another possibility that I argue for is not dualism (physical stuff and soul stuff) but trichotomy: soul, spirit, and body. The positing of a third substance allows one to avoid the phenomenological implication that all knowledge is proceeded by observation of some sort.

  14. I have a friend who would reject the first move in this argument - We all know ourselves to be conscious - he is a cyber artist by the name of STELARC. It is an interesting question, what precisely do we think we know when we 'know' that we are conscious?

  15. "what precisely do we think we know when we 'know' that we are conscious?"

    How about Descartes' "I think, therefore I am."

  16. Don/Brandon, my thought was that the substance dualist has an extra point of vulnerability compared to others.

    Basically I see the anti-skeptical position to involve a two-step inference: first, we infer that anyone displaying similar behaviour to us has a similar physical constitution (brain states, etc), and then we further infer that anything with our kind of brain states also has our kind of mental states.

    It's true that everyone is on equal footing for the first step. (Like Brandon says, I can't be certain that his head ain't hollow. It does seem a pretty reasonable inference though!)

    The second step is easy only for those of us who deny the nomological possibility of zombies. We think that, given the laws of nature, this step can't go wrong. There's no way you could have anyone physically the same as me but mentally different.

    It's not so clear to me how the substance dualist can make this second step. Everything they observe would be explained just as well by positing a zombie. What does epiphenomal consciousness add to the mix? How can one make an inference to the best explanation when your conclusion doesn't actually help explain anything?

  17. Well, Cartesians, for instance, universally denied the possibility of zombies. (Not directly, of course; but it was the Cartesian view that some behavior of bodies was explicable only by minds. So they would deny that everything we observe can be explained just as well by a zombie, since they would hold that such would be positing an effect without an adequate cause.) So I don't think they would have a problem here. For Cartesians zombies are not rationally conceivable (Arnauld, for instance, argues that the existence of other minds is more certain than the existence of bodies).

    But even setting this aside, one could see the analogy as primarily valuable in classification. We know of one case where such-and-such behavior of such-and-such type of body is caused by a mind (namely, myself); therefore, one might argue, the natural thing to do is classify this same type of behavior in this different body of the same type as (probably) caused by mind. The zombie doesn't classify things (and perhaps doesn't even explain things) 'just as well' because it posits a merely possible case, whereas the analogy starts with a case we actually know. We know that if mind is a substance it does explanatory work for at least one instance of bodily substance (our own); but this is more than we can say for the posited zombie case (where we are being vague about the the actual explanation of the behavior without having a clear actual example we can point to for further study).

    For substance dualists like the Cartesians, of course, consciousness is not epiphenomenal at all, but an explanatory cause. What's really in danger of becoming epiphenomenal (cf. Arnauld, and in an even stronger sense idealists) is the body.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)