Sunday, May 10, 2009

Understanding (Zombie) Conceivability Arguments: Part I

The zombie argument for dualism is commonly misunderstood. [For a broad overview and assessment of the argument, see my 'Zombie Review'.] In particular, misunderstanding the precise role conceivability plays in the argument often leads to overly hasty dismissals. In this post, I want to set out and correct three such misunderstandings. Let me begin on a conciliatory note by emphasizing that many conceivability arguments are no good. Each of the common mistakes discussed below begins by correctly noting a certain way that conceivability arguments can go wrong. But in each case, I will show, the zombie argument -- properly understood -- avoids the identified pitfall.

(1) In ordinary language, to call something 'conceivable' is just to say that you're not certain it's impossible. But such mere prima facie conceivability does not entail metaphysical possibility. There can be non-obvious or unknown necessities -- just look at mathematics. So clearly 'conceivability' in this sense doesn't prove anything much at all. That's true enough, but the zombie argument does not invoke 'conceivability' in this loose sense. Instead, it invokes the technical notion of ideal conceivability, or what can be conceived without (even implicit) contradiction. This stricter sense of conceivability more plausibly entails possibility. (The flip side of this is that it makes the premise [zombies are conceivable] more controversial!)

(2) Another common objection is that Kripke's discovery of the necessary a posteriori shows that some claims (e.g. "water is not H2O") can be ideally conceivable without being metaphysically possible. That's true, but a proper understanding of the Kripkean necessary a posteriori reveals that it is limited in scope. Kripkean complications arise only for concepts (like 'water') where the primary and secondary intensions diverge. Importantly, we can tell by conceptual analysis whether this is the case. We can tell, for example, that our water concept rigidly designates "the substance ___, whatever it is, that is the actual watery stuff of our world." But then to run a valid conceivability argument we simply need to take care to avoid those problematic concepts, and employ only their 'semantically neutral' (purely descriptive/qualitative) analogues: e.g. the qualitative term 'watery stuff' in place of the kind term 'water'. (Most everyone agrees that it's metaphysically possible that watery stuff be other than H2O: Twin Earth itself is an example!) Since the zombie argument can likewise be stated using semantically neutral terms, this suffices to defang the standard Kripkean objection.

[Since this post is getting overly long, I've shifted the third section to a new post.]


  1. It's nice that you've made explicit the respect in which the zombie argument rests on some controversial claims about 2-dimensional semantics.

    For instance, a lot of people doubt whether primary intensions are always to be had. Casually suggesting that we can talk about the "watery stuff" makes the task of finding primary intensions seem easier than it is--It's not obvious that twin earth cases couldn't be constructed for "liquid", "potable", "clear", and the like (at least, for some of the concepts that you need to use in spelling out the "watery role"), just as much as they could for "water."

    I'm inclined to think that it would make a lot of philosophical projects easier if we there were always primary intensions for our concepts, but there's not much reason to expect that there are. See e.g., Byrne and Pryor's "Bad Intensions," and Stalnaker's "Assertion Revisited: On the Interpretation of Two-Dimensional Modal Semantics."

  2. In elaborating the effects of his 2d view, Chalmers claims that it is enough for his zombie argument that zombies be conceivable "in the sense that it is conceivable that water is not H2O" (The Conscious Mind 98), that is, in the sense that it is conceivable that the watery stuff in a world not be H2O. I'm not certain that this analogy works, because I suspect an asymmetry. Is it conceivable that H2O is not watery stuff? It seems to me that it is probably not, and that seems to be the better parallel for brains not being conscious.

  3. Aaron - right, I don't think the example there is meant to be doing any positive work. (Dualists certainly aren't committed to thinking it conceivability that H2O be non-watery!) It's just to note that the Kripkean necessary a posteriori isn't really any threat to the conceivability argument. If zombies are conceivable in the sense that the physical P is compatible with there not being any phenomenal stuff, that'll do.

    (If you don't think zombies are conceivable at all, then that's a very different kind of objection. The objections considered in this first post grant conceivability but dispute the inference to possibility.)

    Daniel - thanks, I'll have to check out those references.

  4. Well, of course I don't grant conceivability. What I'm always doing in these discussions is trying to undermine the intuition that conceivability is a small thing, which should be readily granted. It isn't, and shouldn't (at all, in fact, but certainly not readily). The water example was just another opportunity to point out that not everything can be made conceivable by playing word games, no matter how many cute dimensions you have in your semantics.


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.)