Sunday, April 06, 2008

Zombie Rationality

'Zombie' writes:
On Chalmers' view, wherein the 'psychophysical laws' are contingent, it seems that across possible worlds most brains like ours will be zombies or at least have 'associated' qualia that don't 'match' the information processing in the brain. So sophisticated brains proceeding according to ordinary standards of rationality should zombie-conclude that they probably are not conscious (as they don't have access to any non-material qualia), despite their zombie-perceptions of being conscious (shared by both zombie and non-zombie brains). Yet Chalmers thinks that in our actual world the psychophysical laws lead to conscious experience mirroring the information processing in the brain. So, upon hearing the argument, shouldn't Chalmers' brain zombie-conclude that it is probably a zombie brain, and 'phenomenal Chalmers' consciously think the same?

No. Conclusions are drawn by people, not brains. Standards of rationality likewise apply to agents and their beliefs, not to their physical components (brains and neural states) in isolation.

On my view, beliefs are partly constituted by phenomenal properties -- that's what gives them their representational content. Zombies don't have beliefs like we do. They exhibit all the same behaviour, and make all the same noises, but there's no meaning in it. It's not really about anything.

One might define a 'z-belief' as the functional (physical, dispositional) component of a belief. It's not so clear how to assign pseudo-contents to these z-beliefs, but I guess a reductionist may offer a stipulation of some kind: S has a z-belief that P iff S has such-and-such physical dispositions [e.g. 'S behaves as though P were true', or 'S has a brain state which covaries with evidence of P', or some such. See my essay 'What Behaviour is About' for a more sophisticated empirical approach to attributing "content".]

Presumably we're to suppose that whenever I really have the belief that P, my brain has the z-belief that P. But I doubt whether any such reduction can be given that perfectly mirrors my actual belief contents. (If epiphenomenalism is true, and qualia are partly determinative of belief content, then the physical facts underdetermine what it is that I believe. My inverted-spectrum duplicate has the same brain -- hence z-beliefs -- as me, but our phenomenal beliefs are very different. My 'red' is his 'blue', or whatever.)

There's a more fundamental problem, even if we grant the reductionist his impossibly fine-grained z-content. Let's grant - per impossibile - that my brain (and zombie twin) "z-believes that P" iff I believe that P. However, my brain (understood as a purely physical system, i.e. excluding its phenomenal properties) is in possession of only a subset of my total evidence. Qualia - the contents of experience - are among my evidence if anything is. But these phenomenal properties are not causally accessible to my neural processes. So the conclusion 'I am conscious' follows from my evidence, but not from the "information" available to my brain. One can be a rational person, or have a "rational" brain, but not both.

Now, it's pretty obvious that being a rational person is better than having a "rational brain" (insofar as the latter attribution is even meaningful). Brains are parts of people, and like any body part we really only care about it for how it can serve the whole person. If quick feet didn't make for a quick person, we wouldn't much care for the former. Similarly, a rationally desirable brain is one that makes for a rational person, with justified beliefs.

One could imagine a brain that is instead built in such a way that it tends to produce "z-justified" z-beliefs. What this means is that it tends to end up in physical states such that a conscious person in that physical state would have beliefs in line with the physically accessible subset of their evidence. When put like that, it becomes clearer that what we've really described here is a defective brain. Let's call it "z-rational", and reserve the term 'rational' for brains that give rise to rational people -- people whose beliefs are in line with their total evidence.

Here are two implications:
(1) A z-rational brain can be expected to have more true z-beliefs (across all possible worlds).
(2) A rational brain can be expected to yield more true beliefs.

Fortunately, my brain is rational rather than z-rational. Hopefully yours is too (otherwise, you're a defective agent). One might try to argue that there's something "wrong" with a brain that isn't z-rational, but I don't think that'll work. For one thing, since you're really just describing a physical state it's not clear that brains or z-beliefs are even open to this sort of normative assessment. Norms apply primarily to people, and to our organs only derivatively. What a well-functioning agent really needs is a brain that will make them rational, not z-rational. As suggested above, a z-rational brain is defective from the standpoint of contributing to the functioning of the whole person (which is the relevant standpoint against which to assess brains). Further, when you stop to think about what it really means to have 'z-rational z-beliefs', you see that there's not really anything significant (worth caring about) there.


  1. My brain ... is in possession of only a subset of my total evidence. What we've really described here is a defective brain. Let's ... reserve the term 'rational' for brains that give rise to rational people -- people whose beliefs are in line with their total evidence.

    But by assumption this subset must be sufficient to produce all of the beliefs that the brain might physically express, such as in words or writing.

  2. Hmm.. If Qualia don't effect the universe then how can qualia be used as evidence?

    I.e. zombie Richard is 'right now' arguing that qualia are, if anything, part of his evidence that he is not a zombie.

    If they have no causal relationship to the words coming out of your mouth or mine, or for that matter on the concept of qualia which we share (as a zombie would mean* qualia) then it pretty much by definition is an irrational position (which may happen to be correct).

    *Mean is a neutered concept here - but not too neutered. Zombie richard can explain to me in detail philosophy and also the concept of qualia and what they are like. So he has a structure in his head that defines the term.

    and I guess I'm repeating things here but it is a little surprising that real Richard might have an experience additional to zombie Richard that is not in anyway recorded even in the most obscure form in his brain.

  3. Two questions :

    a) first, it would be helpful if you could clarify the above point Genius makes. In a universe identical to this, but without qualia would Richard2 belief that he is a Zombie? As far as I can follow your reasoning, he would act in every way as the original Richard would, writing the exact same post as above, but all this philosophy talk would now be meaningless (I shall resist the urge to make a cruel joke ;) ), as as a Zombie he can't believe anything?

    And whether something is meaningful or not, has absolutely no effect on the rest of the physical universe, or else its not epiphenomenal, right?

    (Strictly speaking, if whether we have qualia or not has no effect on the rest of the universe, doesn't that imply even if this debate is one day concluded it'll make absolutely no difference to anything else... )

    b) I read your other post again, and I still don't see why beliefs need phenomenal content - could you clarify, provide some arguments for this position?

    YMMV on this point - I would be prepared to say that my computer has a belief, say, that I have 10 emails unread. It has something real, encoded into information, and it uses that information to make useful and accurate preditions and actions affecting the real world. I'm not sure what more you want in a belief?

  4. Richard,

    1. If you remove possibility of z-beliefs or something of a sort (mental pattern), there would be no explanation about why you behave as you behave. (I guess as epiphenomenalist you agree that you your behavior should be fully accounted for in physical terms?)

    2.I mentioned something about this in the discussion on your other post (and wrote more on it here ), but epiphenomenalist should have account for this z-epiphenomenalist behavior and those z-beliefs. You said that you are not in worse position at all than physicalists, but you are as they have abstract entity in their physical world, that you don't - the awareness that one is conscious. Epiphenomenalists push that one outside of the physical world, so they should provide different explanation for z-epiphenomenalist-behavior, and z-epiphenomenalist-beliefs.

    2. Wouldn't z-beliefs (or other NCCs), actually give rise to the beliefs, given the psycho-physical laws? If that is so, in the actual world, such things as inverted qualia can't happen, and you can't use the inverted-qualia example to argue that given same z-beliefs (or something of a kind) you could have different beliefs. It would so appear, that whatever you mean by "partially constituted", it is fully determined by a)physical world and b)psychophysical world, so the beliefs are fully dependent on z-beliefs or something of a kind.

  5. Right, I'm perfectly happy to grant some kind of z-belief/ pseudocontent for its explanatory/predictive power (see my linked essay, which is very empirically oriented). I just don't think such empirically attributable contents can be so fine-grained as real mental content (as the inverted qualia case demonstrates -- N.B. the nomological impossibility of inverts and zombies is irrelevant to this point).

    Jonathan - my acquaintance with meaning, like my acquaintance with qualia, is essentially first-personal. I see them as bound very closely together. I know what it is for me to have a thought that is about something, and part of this (it seems to me) comes from the phenomenal feel of my thoughts. Strip away my qualia, and you've also taken away my thoughts in the fully-fledged sense. Cognitive processing would continue to occur, of course, and outsiders could tell no difference -- my mere z-beliefs suffice for their purposes -- but 'from the inside' things are very different.

    (I don't expect hardened materialists to be persuaded by any of this. My argument is more defensive: assuming epiphenomenalism, these other views may reasonably be held, which in turn show why various materialist objections are not insuperable for the epiphenomenalist.)

  6. You say... "I just don't think such empirically attributable contents [let's call them EAC] can be so fine-grained as real mental content."

    But different mental content (in epiphenomenalist view) will be result of different EAC, no? So you have to have as fine grained EAC as fine grained your phenomenology is for this to happen.

    So to say, if we see psycho-physical laws as function, that take the physical state, and return the mental state - the function's domain needs to be as fine grained as its range.

  7. Different mental content will be a result of different EAC plus the psychophysical bridging laws. EAC alone does not determine mental content, because there is a possible (invert) world where the EAC is the same but the mental content differs. EAC alone is thus coarser grained.

  8. It seems to me not entirely natural that one could deny that a zombie perspective exists. I don't have any major issues with a table perspective existing or a 'richard&elizer' perspective or a 'american' perspective existing.
    So it seems to be zombies would have a perspective no matter what you do - and if that is an intelligence empowered perspective (like a human) your getting close to being a human even if you strip out all particular qualia.

    Also just thinking... I wonder if one could chop qualia into sub parts. For example red - and cut it down into a thousand parts none of which having any meaning by themselves (but still have something special to them) but all of which remaining epiphenomenal - as a way to study them.

  9. Okay, so it sounds like your view of beliefs is roughly what I mean by belief with the addition of qualia. belief = z-belief + qualia. Is that fair?

    If so, how does the phenomenal aspect have any affect on our rational calculation? I'm just not sure what rational / informational content qualia have - in fact of course, they can't have effect or else we would no longer be in epiphenomalist territory. They're purely a bonus, a free gift on top as it were.

    I must admit, I think I've got a bit confused somewhere along the line with your exact point. When you talk about a person being rational, rather than just a brain, you seem to be implying that there's something else doing the rational stuff, thinking or whatever, other than just the brain. If you're not saying that, I'm not sure why you're saying that the brain in itself isn't rational. But if you are saying that, the thinking needs to take place somewhere. So you're explicitly arguing for the existence of (for the lack of a better word) a soul then?

  10. Jonathan - I granted for sake of argument that z-belief = belief - qualia, but this isn't really possible: empirically attributed content won't always correspond to mental content for reasons explained above.

    "If so, how does the phenomenal aspect have any affect on our rational calculation? I'm just not sure what rational / informational content qualia have"

    I have no idea what you're asking here. Qualia don't influence how our brain functions ('calculates'), if that's what you mean. But since they are partly determinative of what it is that we believe or conclude in our reasoning (and also feature in our evidence), they thereby affect the normative fact whether the conclusion of our reasoning is rationally supported by our evidence.

  11. Oh, Jonathan, I know what might help: see my post on The Problem of Other Minds -- there I contrast my view with substance dualism (as I understand it).


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.)