philosophy is for everyone
and not just philosophers

philosophers should know lots
of things besides philosophy

Philosophical Connections

Electronic Philosopher

Feature Articles

University of London BA

Philosophy Lovers Gallery

PhiloSophos Home

International Society for Philosophers

Hilary Putnam and brains in vats


To: Stuart B.
From: Geoffrey Klempner
Subject: Hilary Putnam and brains in vats
Date: 12 March 2007 13:31

Dear Stuart,

Thank you for your email of 6 March with your latest University of London essay, on Hilary Putnam and the brain in a vat, and the rewrite our your previous essay on the coherentist theory of justification.

You needn't have pulled your hair out over my remark about tracking truth. It is a common complaint levelled against the coherence theory of truth that a maximally coherent set of propositions can still be false. But a coherence theory of justification is not the same as a coherence theory of truth. (I'm aware that I'm not telling you anything you don't already know!)

One's first reaction to the question (and this is something I should have said right at the beginning but didn't) is why should the possibility that a coherent set be false be an objection to coherentism about knowledge, if on the standard tripartite theory one happily accepts the fallibilist concession that the justification which we offer from a proposition might be false? For the coherentist, a coherent set is not logically equivalent to truth, but it is the best way to track truth. That's just how knowledge grows, by adding propositions to a coherent set.

My response to your essay might seem a little oblique, but you will see the point in a minute.

The first point I would make is that Putnam does himself a disservice with the motto, 'meanings are not in the head'. The first question one has to ask is why do we have natural kind terms? As Putnam recognizes, two things are required for meaning, an intention and the requisite causal link. But the reason why a causal link is required is because of the particular nature of the intention that we form when we coin a natural kind term. We are seeking a maximally explanatory theory, attempting to 'cut reality at the joints' to use Plato's phrase. That means that our intention is to capture a salient aspect of the world and link the term to THAT rather than to limit its meaning to whatever set of attributes we can think up.

That is why white gold is still gold, why graphite, diamond and charcoal are all carbon etc. etc.

The interesting question to ask is, Could our intentions be different? Why couldn't we deliberately 'pull in our horns' and refuse to coin any natural kind terms, confining language to nominal concepts? Our capacity for understanding the world would be greatly impoverished. But suppose the world was a lot more recalcitrant to understanding than the actual world is. Maybe in that alternative world attempting to divide things into natural kinds just doesn't get us anywhere.

(As a passing remark, for many, many years Aristotle's natural kinds theory was looked on with scorn by analytical philosophers, as dubious metaphysics - which indeed it was.)

A counter-argument would be that even if we give up natural kind terms, we still have to interact with objects. Giving up a causal theory of reference - or a more sophisticated theory incorporating a causal aspect as in Gareth Evans 'Varieties of Reference' - leaves one teetering on the edge of idealism.

However, the brain in a vat is not an idealist evil demon scenario. The vat-inhabitants believe that there exists a physical world in space and there is a physical world in space. They coin proper names and concepts with the intention of setting up causal links but the intention fails, or rather, catches on to different 'objects' from the ones they intended to catch onto.

In the real world, brains and vats are natural kinds (at least, vats are made of some metal which is a natural kind) but in the virtual world of the vat-dwellers, the objects they identify as 'brains' and 'vats' are not natural kinds. Anything can happen in virtual reality (as computer games amply demonstrate). Even if it doesn't, it might at any moment even though the vat-dwellers do not know this.

I think this is sufficient to show that vat-dwellers do not have false beliefs about brains and vats. With luck, their conditioning has set them up to be capable of conversing about brains and vats were they to be set free and put in living human bodies (this relates to one of your points), so they can truly say, 'We once had false beliefs about brains and vats,' but that is not sufficient to establish that they are were already conversing about brains and vats before they were set free. So I am partly agreeing and partly disagreeing with Putnam.

Your first objection is mistaken, in my view. The first premiss is not the premise of an or-elimination but rather a hypothesis set up for reductio. Assume I am a BIV. Now let's see what follows. What follows (or so Putnam claims) is a contradiction. Therefore, I am not a BIV.

However, your second objection had me stumped for a while. Is it really so easy to rejig the BIV scenario so as to be immune from Putnam's argument? I've already considered the possibility of freeing a vat-dweller. They would soon get along fine with the rest of us. But what about vat-imprisonment? I have my causally acquired knowledge of what brains and vats are, and now continue applying these not realizing that I am no longer in causal contact with the world. For reasons which I have already given, however, I don't think this works. My semantic intentions fail, they do not succeed. I don't have false beliefs about brains and vats. I don't have beliefs, period. I seem to express thoughts but do not.

To meet your third objection, it is sufficient to point out that accidental coincidence is not enough. In the vat-world, a tree can morph into a vat and vice versa at any time, even though for accidental reasons this has never happened so far. Ignorance of that possibility wrecks the semantic intentions of the vat-dwellers.

Having gone to such lengths to defend Putnam, however, I agree with you that his argument does not work as an argument against scepticism. Forget about vat-ish, let's assume a stronger claim than Putnam makes, that vat-dwellers do not 'speak' any 'language' although it seems to them as if they do. They do not express thoughts with a sense and a reference. This is in fact my view.

Now, as a sceptic I say, 'Maybe I'm a brain in a vat.' This is possibly true. If it is true, then I am not saying anything because I can't 'use a language'. So what? That merely makes my sceptical conclusion all the more extreme. Not only do I not know anything, I cannot be certain that my words mean anything. There is no cure for this except a leap of faith. Or, in the words of Wittgenstein, I am not shutting my eyes to doubt: 'They are shut.'

All the best,