Saturday, April 15, 2006

Misusing Kripke; Misdescribing Worlds

Too often, philosophers (including my past self) misunderstand Kripke as having shown that the space of metaphysically possible worlds does not include all logically/conceptually possible worlds. This leads to sloppy defences of physicalism. We may fail to take seriously the challenge of logically possible zombie worlds, prefering to dismiss them with a wave of the hand and some vague muttering about Kripke's necessary a posteriori. Or consider my own past sloppy defence of ethical naturalism, where I dismissed Moore's open question argument on the basis of an analogy with 'salt = NaCl'. Such swift dismissals are unwarranted -- let me now explain why.

The crucial point arose in my introduction to two-dimensionalism. The 2-D framework shows us that we need only one space of possible worlds, and the difference between 'conceptual' and 'metaphysical' necessity is merely a matter of how we consider the worlds in question (i.e. as actual, or as counterfactual). The point is made nicely by Frank Jackson in his From Metaphysics to Ethics, pp.77-78:
What convinced us [that 'water = H2O' is necessarily true, albeit a posteriori] were the arguments of Kripke and Putnam about how to describe certain possibilities, rather than arguments about what is possible per se. They convinced us that a world where XYZ is the watery stuff of our acquaintance did not warrant the description 'world where water is XYZ', and the stuff correctly described as water in a counterfactual world is the stuff - H2O - which is the watery stuff of our acquaintance in the actual world be it watery or not in the counterfactual world.

The key point is that the right way to describe a counterfactual world sometimes depends in part on how the actual world is, and not solely on how the counterfactual world is in itself. The point is not one about the space of possible worlds in some newly recognized sense of 'possible', but instead one about the role of the actual world in determining the correct way to describe certain counterfactual possible worlds.

So the Kripke/Putnam cases do not really shrink the space of possibilities. Nobody thinks that the 'Twin Earth' world Putnam describes is an impossible one. Rather, we still grant the possibility of the world itself, but merely re-assess how best to describe it. The necessity of 'water = H2O' says less about the space of possibilities than it does about our use of the term 'water'. All the same worlds are possible as before; it's just that now we refuse to use our term 'water' to describe any of the non-H2O stuff in any of those worlds.

It should now be obvious why appeals to the Kripkean necessary a posteriori do nothing to answer the aforementioned challenges. Consider the zombie world. We are conceiving of a genuine possible world, just as in the 'Twin Earth' case. The only question is how best to describe it. Perhaps our term 'consciousness' is, like 'water', a rigid designator. But who cares about the words? Twin Earth still contains watery stuff, even if we refuse to call it 'water', and the Zombie World still lacks phenomenal stuff (qualia), even if we stipulate that our term 'consciousness' refers to some neurophysical property (and so is guaranteed to exist in this physically identical world). Redescribing the Zombie world won't make the physicalist's problems go away. There's no denying that the world itself is possible, physically identical to our own, and yet lacking something of a mentalistic sort, however we end up describing it. And that's enough to refute physicalism (excepting Russellian monism, which allows us to deny that the zombie world is physically identical to our own).

We can clarify the point by appeal to the apparatus of 2-D semantics. Kripkean sentences which are necessary a posteriori have contingent primary intensions but necessary secondary intensions. But this cannot be the case for truths about consciousness, for our qualitative concepts are 'semantically neutral' in the sense that their primary and secondary intensions coincide.

Kripke himself noticed something along these lines. While we can imagine a world where watery stuff isn't truly water, it's incoherent to imagine a world where "painy" stuff isn't truly pain. To feel painful is to be painful. Whether something counts as 'conscious' does not depend on features of the actual world. We should come to the same judgments about consciousness regardless of whether we consider a world 'as actual' or 'as counterfactual'.

See also Dave Chalmers on 'The Two-Dimensional Argument Against Materialism'.


  1. Yup, I understand the zombie world as having different natural laws. Not different physical laws, though. So from the "outside", third personal perspective of science, the world would be indistinguishable. (No "feedback effects" here.) But there is nothing phenomenal in the zombie world, it's "all dark inside", with nothing it is like to be our zombie twins.

    The idea is that once the physical nature of a world is fixed, something more remains to be added. You could just leave it at that, and have a zombie world which unfolds exactly like ours except that nobody is truly conscious. It would be a world lacking in subjectivity, or first-personal experience. But our world has something more, thankfully. There are psycho-physical natural laws which govern nomologically necessary (but metaphysically contingent, since these laws could be different or even entirely absent) correlations between certain physical states and certain phenomenal states. These laws guarantee that zombies are nomologically impossible, that certain brain states give rise to certain conscious experiences. But the laws are contingent. So the physical does not entail the phenomenal, contra physicalism.

    A bit rough, but I hope that makes things clearer. (If you're interested in this stuff, let me recommend Dave's book: The Conscious Mind: In Search of a Fundamental Theory.)

  2. Regarding that something extra - (maybe this os one of your named theories)

    Can one have an otherwise ideantical world that was not identical in some individual aspect?

    (for the moment I will assume a comprehensive multiverse of all valid possibilities)

    I imagine a sort of world (for example) which is identical except gravity doesnt exist (specifically in that it doesnt hold you to the earth because it doesnt exist) being "impossible" in that gravity also does all sorts of other things - how can they remain despite it not existing? (maybe that is possible?)

    Related to that -is there an infinite number of possible rules? if so you might get around it. But just like 1+1=2 2+1=3 etc maybe there is no world where 1+1=3 has any meaning there may not be an infinite amount of rules or an infinite amount of ways they can interact.

    Now another back stop - are there an infinite set of outcomes (including every outcome) for every set of "rules". Could you have a universe with our rules but which behaved by chance almost identicaly to a universe that had a totally different set of rules?

    If not maybe there is no zombie world. I.e. now way a new rule could have no effect.

    If there is such a world it raises the question whether we are saying conciousness genuinely never has an effect (ie it changes no decisions uses no energy etc). Which raises a question of what exactly we are measuring.

  3. G., similar issues are discussed in my post Are Laws Internal to Worlds?. But I don't think they're too relevant to the present point, since phenomenal consciousness is plausibly epiphenomenal or causally impotent. You can't "measure" consciousness. (Like I said, the zombie world is scientifically indistinguisable from our own.) Its nature is essentially first-personal. The only access one can possibly have to phenomenal properties is the direct access of the experiences we find ourselves presented with.

    Epiphenomenalism raises interesting questions about how we know that we're not zombies. Such concerns are addressed in section 4.3 of Chalmers' article here.

  4. Let me add that if the world were to remain physically indiscernible despite lacking gravity, there would have to be some kind of "as if" gravity to replace it. (That is, something has to be pulling masses together, and hence "playing the gravity role".)

    But consciousness is obviously completely different. It is easy to imagine a physically indiscernible world which entirely lacks consciousness. This doesn't require any "as if" consciousness, because there is no causal/functional role which needs filling here. That's the whole point of consciousness; it isn't a causal or functional property, it is a phenomenal property, and this is ontologically independent.

  5. Richard, it seems that coming to work with Chalmers has made you a full-fledged dualist! What happened to the Richard that admired Dennett...?

    Now seriously, are you so sure that we can concieve of a zombie world? I found this article a pretty convincing case that we cannot. At least it makes one rethink matters that seem very easy when reading Chalmers.

    (I am also inclined to think that even if we can concieve zombies nothing important follows from that, but that would require a much longer argument).

  6. Is being able to concieve of somthing sufficient to say it is possible?

    Take for example God. It is easiy to concieve of a god (any random god) but I would expect you to suggest it is not possible.

  7. If, as a matter of scientific fact, consciousness in our world is instantiated by physical properties, then there is no world anywhere which has the same physical laws, and the same physical things, and no consciousness. That, it turns out, just is impossible, just as a world which has the same physical laws as ours can't contain H2O without containing water.

    Now, there could be a possible world which is physically very similar to ours, but in which mental activity is manifested by brain stuff which just isn't quite good enough to do all the jobs our brains do, but where there is a little bit of extra soul stuff that fills in the gaps, in which people are conscious. Call it David's world. And there could be a physical duplicate of that world in which people all acted the same but where they lacked consciousness because they lacked the little bit of soul stuff. But, although that would be a zombie world of sorts, it wouldn't be the world needed to refute physicalism, as it wouldn't be a physical duplicate of our world, only of David's world.

    Of course, we don't know for certain that we don't live in a Dave world; perhaps we're wrong about the scientific facts. If we do live in a Dave world, then there is a zombie world which is an exact physical duplicate of ours. But we didn't need that possibility to show us that physicalism would be false if we lived in a Dave world; obviously physicalism would be false if we lived in a Dave world.

    Oh, and the traditional philosopher's God, a person who is omnipotent, omniscient, and omnibenevolent, is clearly both inconceivable and impossible.

  8. Also if conciousness is associated with the physical universe (ie it exists in certain places with certain physical properties - which is required for it to be useful). Doesn't it defy some pretty basic principles for that to be true and yet it to have no effect at all on the universe? (maybe there is a way around that?)

  9. G., it's ideal conceivability that's relevant here. See Chalmers' paper: Does Conceivability Entail Possibility? The challenge for type-A materialists (who think consciousness is a priori reducible to the physical) is to show that there is some hidden contradiction in the notion of a zombie, which means that it isn't really conceivable (on ideal rational reflection).

    But note that my present post assumes that zombies are conceivable (as most philosophers accept). I don't mean to be offering any argument for that here. Rather, my point is to show how the 2-D framework undercuts "type-B materialism", i.e. the view that zombies are ideally conceivable but not really possible. That's the view which appeals to a posteriori necessities.

    (I don't know what you mean by phenomenal consciousness being "useful". As already noted, it is ontologically independent from causal/functional facts, so it's not like evolution could select for it or anything. Zombies would be just as biologically "fit" as us. Natural selection merely selects for causal/functional abilities, so it couldn't us apart from our zombie twins. We evolved for our physical abilities. It just so happens that the natural laws add phenomenology to beings of our causal/functional structure. So the reason we're not zombies is because of the psycho-physical natural laws, not because evolution found phenomenology to be "useful". But perhaps you meant something different?)

    Alejandro - I still like Dennett, but I was never convinced that his account explained qualia. Indeed, I've found property dualism unobjectionable ever since I realized that it respects the nomological necessity of physical-phenomenal correlations. I'm certainly no substance dualist! Re: conceivability of zombies, see my response to G. above. Perhaps I'll revisit the question in a future post.

    Protagoras - who said anything about "soul stuff"? Water just is H2O, which is why you can't have one without the other (and this is so regardless of whether the physical laws differ. For example, there could be other worlds where water/H2O isn't wet). But it isn't plausible that consciousness just is some physical property. If you say that, you're changing the subject, because I mean to be talking about phenomenal properties here.

    Now, presumably what you mean to say is that physical properties give rise to consciousness (or phenomenal properties). And I completely agree. As it happens, our universe is governed by natural laws which ensure that our brains suffice for consciousness. But the laws could have been different. We can imagine a world which has the same physical laws but lacking the psycho-physical laws* which connect the physical with the phenomenal. This world really could be physically identical to ours -- atoms in all the same places, whose movements are governed by the same mechanistic principles -- but lacking in phenomenal consciousness.

    * = Perhaps you were mistakenly thinking that these count as purely "physical" laws. But that would be a mistake. The point is that a physical duplicate of our world could have lacked consciousness. Once all the third-personal stuff studied in physics was fixed, this wasn't enough to guarantee a universe containing first-personal experience. Something more needs to be added. Not any nonsensical supernatural "soul stuff", of course. What's lacking isn't a substance. It's rather a rule, or natural law, which specifies that (and how) the physical properties give rise to phenomenal properties. This is all perfectly naturalistic. But such psycho-physical laws are not part of physics (a discipline which is solely concerned with third-personal phenomena). They are fundamental natural laws in their own right, which serve to bridge the gap between the physical and the phenomenal.

    Another way to put the point is that no amount of third-personal physical knowledge suffices to entail the existence of first-personal experience -- let alone its character. We can coherently conceive of all the former facts obtaining without the latter, as when we imagine the zombie world. (And even if you don't like zombies, the argument still goes through on more moderate differences, say "inverted spectrum" cases where someone physically identical to you sees red as blue and vice versa.) Physics by itself has nothing to say about consciousness. It isn't part of the third-personal picture. It needs to be supplemented by a science of consciousness.

  10. Richard,

    How confident are we that, say, chemistry reduces to particle physics of nuclei and electrons?

    I would say that we are highly confident of this reduction because of the correlations we have observed between states in the two fields of observation.

    If a precise and predictable correlation is found between first-person states and third-person (physical) states, why should we consider this a lesser form of reduction than in the case of chemistry and particle physics?

  11. Hi Doc., our confidence in such scientific reductions rests in more than mere correlation. It is not even conceivable (at least, given sufficient reflection) that a world could be physically identical to our own and yet differ in its chemical properties. (Or its biological properties. Or, indeed, almost any macroscopic properties at all, excepting those which depend on the phenomenological sphere.)

    Strict correlations allow us to draw nomological conclusions, as I've discussed at length above. But to draw ontological conclusions (about metaphysical and not merely nomological necessities) requires conceptual analysis.

    See also Chalmers and Jackson on Conceptual Analysis and Reductive Explanation. (They say all that needs saying far better than I ever could!)

  12. I did mean to say that consciousness just is some physical property, or at least at this world that seems likely given everything we know.

    Oh, I said that it's impossible for there to be an H2O world without water in that way for a reason. I think it's just a matter of arbitrary convention whether XYZ is water, and so in some contexts I'd say it's possible for there to be water without H2O. But definitely no H2O without water, unless physical laws change.


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.)

Note: only a member of this blog may post a comment.