Could you pass me the salt?

Jennifer Saul is organizing a very interesting workshop here in Sheffield, on semantics and pragmatics. The programme looks very exciting. I will be helping her a little bit with the registration and stuff on Saturday, but I hope that, in spite of that, it will be a success! ;-)

Apart from that, I have been quite busy recently, sending papers to journals (for the first time ever!), and applying for jobs. So, now, and after waiting for so long, I can say, with respect to all these things: Been there/done that. (Hopefully, one day I'll be able to say that with respect to being published and getting a job. Fingers crossed!).

The weak, the strong, the local and the global

This has been a bit quiet recently. I have been busy, working on, among other things, my first, introductory chapter. These days I have been drafting the section on physicalism (see some earlier discussion on this here). I have introduced the main notions of supervenience that are invoked on the debate on physicalism. So far, I have talked about weak, strong, local and global supervenience, characterised as follows:

(Weak supervenience): A-properties supervene on B-properties if and only if, for any individuals x and y in the actual world, if x and y are identical concerning B-properties, they are also identical concerning A-properties.

(Strong Supervenience): A-properties supervene on B-properties if and only if, if anything has property F in A, then there is at least one property G in B such that that thing has G, and necessarily everything that has G has F.

(Local Supervenience): A-properties supervene on B-properties if and only if, for any individuals x and y and any possible worlds v and w, if x at v and y at w are identical concerning B-properties, they are also identical concerning A-properties.

(Global Supervenience): A-properties supervene on B-properties if and only if, for any possible worlds v and w, if v and w are identical concerning B-properties, they are also identical concerning A-properties.

Well, there is enough food for the brain here. There has been a lot of discussion on the relation among all these notions. I am not an expert at all, but at any rate, these are my reactions (any comments or criticisms are more than welcome!). It seems to be widely held that weak supervenience is too weak for the purposes of defining physicalism, and I agree with that. Many people hold that local supervenience is stronger than global supervenience, and in particular, that there are truths that do not supervene locally on the physical, but supervene globally. I guess this depends on how you characterise the physical properties of individuals x and y (when these are the B-properties, in the schema of the definition of local supervenience above). If we allow the physical properties of x and y be broad enough, then any property P that supervenes globally will supervene locally too.
I was also wondering whether strong and local supervenience are equivalent or not. At first sight, they do seem the same thing to me, but I do not know whether this assumes something that might be controversial (you can never be too careful...). So I would like to know what other people think.

In any case, it's a relieve to think that I can be neutral on all this, for the purposes of my thesis. It does not really matter what of these notions you use to define physicalism (excluding weak supervenience, obviously): the other three resulting definitions of physicalism seem to be committed to global supervenience. And this is what conceivability arguments seek to refute. On my view, unsuccessfully. But that's another day's story...

Acting like a Zombie

Daniel Stoljar has a new paper, "Two Conceivability Arguments Compared", forthcoming in PAS. There, he compares these two arguments:

The Zombie Argument:

Z1. Zombies (physical duplicates of us, with different phenomenal properties) are conceivable
Z2. If zombies are conceivable, zombies are possible
Z3. Ergo, zombies are possible

The Actor Argument:

A1. Actors (behavioural duplicates of us, with different phenomenal properties) are conceivable
A2. If actors are conceivable, actors are possible
A3. Ergo, actors are possible

The Zombie Argument (ZA) is used to defeat physicalism. The Actor Arguments (AA) is used to defeat behaviourism, or so Stoljar claims. He argues that reflection on the similarities among these arguments can pose problems for the phenomenal concept strategy.

Stoljar assumes that AA is a good argument against behaviourism. He points out that AA is usually presented to undergraduate students as a sound refutation of behaviourism. He also assumes that AA and ZA are of the same kind, that is, they are both concerned with the inference from conceivability to possibility, and the relation between phenomenal and (some or all) physical truths.

Using these two assumptions, Stoljar argues that the phenomenal concept strategy is committed to rejecting AA. The PC strategy claims that Z1 does not entail Z3, because Z1 has an alternative explanation which does not entail Z3. Stoljar argues that if AA and ZA are of the same kind, then the PC strategy would have to say that A1 has an alternative explanation which does not entail A3, so that AA would not be sound. But AA is sound, he assumes. So the PC strategy is incorrect.

I am working on a response to this argument. My thought is that we can disambiguate AA in two ways:

• AA*

A1*: Actors are conceivable
A2*: If S is conceivable, S is possible
A3*: If actors are conceivable, then actors are possible
A4*: Ergo, actors are possible

• AA'

A1': Actors are conceivable
A2': If actors are conceivable, actors are possible
A3': Ergo, actors are possible

I think that Stoljar's argument equivocates on these two readings of the actor argument. For there is no single reading of it that makes both of his assumptions plausible:

Assumption 1 (AA and CA are of the same kind) is plausible only under reading AA*. That is, advocates of ZA take it that the conceivability of zombies entails the possibility of zombies because conceivability entails possibility in general.
Assumption 2 (AA is sound) is plausible only under reading AA'. AA' will be sound if the argument is valid and all the premises are true. Under reading AA' it is uncontroversial that AA' is sound, since it is uncontroversial that A1' and A3' are true. And since A2' is just a material conditional, if the consequent is true, the conditional is true. So it is uncontroversial that AA' is sound. But it is not uncontroversial that AA* is sound: premise A2* is precisely what is at issue in these debates.

Could Stoljar understand ZA in a similar way to AA', so that both assumptions involve the actor argument in the sense of AA'? Let's see:

ZA*:
Z1*: Zombies are conceivable
Z2*: If S is conceivable, S is possible
Z3*: If zombies are conceivable, zombies are possible
Z4*: Ergo, zombies are possible

ZA':
Z1': Zombies are conceivable
Z2': If zombies are conceivable, zombies are possible
Z3': Ergo, zombies are possible

If we consider ZA' and AA', then we can safely assume that the zombie argument and the actor argument are of the same kind, and we can also assume that the actor argument is sound. Does this pose a problem for the PC strategy?
Well, not really. Because the PC strategy is not committed to saying that AA' is unsound, even if they do claim that ZA' is unsound. The strategy entails that conceivability of zombies is not a reliable guide to the possibility of zombies. It also entails that the conceivability of actors is not a reliable guide to the possibility of actors. But, of course, this is compatible with the claim that actors are both conceivable and possible.

Putting the point in a different way: when we compare ZA* and AA*, if we assume that AA* is sound, then the PC strategy is in trouble because their response against ZA* entails that AA* is unsound too: they will deny Z2* and A2* (which are the same). But, of course, it is very controversial to claim that AA* is sound. In my paper, I plan to argue that Putnam and Block's arguments against behaviourism are not of the form AA*.