To: Soo Chuen T.
From: Geoffrey Klempner
Subject: Kant's Refutation of Idealism
Date: 11 May 2001 09:01

Dear Soo Chuen,

Thank you for your e-mail of 27 April, with your second essay for the Metaphysics program, in response to the question, 'Give a careful account of the argument of Kant's second "Refutation of Idealism".'

This is an excellent piece of work. I don't recall receiving a better essay on this topic.

In a relatively short space - and using less words than Kant uses! - you succeed in giving an of Kant's argument which is both persuasive and true to the text. You also make some pertinent criticisms.

Because your account is so clear, I have been prompted to rethink my position on this issue, which has remained more or less static since I studied Kant's 'Critique of Pure Reason' as a graduate student under the supervision of P.F. Strawson (author of 'The Bounds of Sense' - still arguably the best book on Kant). I formed the view then that the upshot of Kant's 'Refutation' was, in effect, an elaborate defence of egocentric idealism - even though this was certainly not Kant's intention - and that there was a crucial gap between what Kant's 'Refutation' establishes, and the conclusion established by Wittgenstein's argument against a private language.

I also thought then, but do not think now, that the private language argument defeats idealism. It certainly defeats egocentric idealism. As I argue later in the program, however, it is possible to defend a non-egocentric idealism - Kant's theory of phenomena and noumena is perhaps the best example - while fully conforming to the strictures of the private language argument. Some further argument is needed. This is where the claim about the primacy of physical agency comes in. (I recall, however, when I suggested this point to Strawson, his responding, 'How is that different from Dr Johnson kicking the stone? I still do not have an adequate answer to that challenge.)

One virtue of your account is that you make it clear that Kant is arguing dialectically. The only philosopher we are concerned with is the Cartesian, who claims that 'there is only one empirical assertion that is indubitably certain, namely that "I am".' The argument has no effect whatsoever on the sceptic who is happy to give up this one piece of Cartesian certainty.

But what is the argument? Let's look at premise P1:

I am conscious of my own existence as determined in time.

It is a familiar point about the concept of perception that, e.g., "I can see a stretch limo parked outside my front door" can be understood in two distinct senses:

A. There is a stretch limo parked outside my front door and I can see it.

B. There appears to me that there is stretch limo parked outside my front door.

B. would be true and A. false in a situation where, for example, there were two identical cars parked bumper to bumper, but the middle was hidden from view by an obstruction.

So, in a similar way, we can distinguish between:

P1A. I am actually aware of my own existence as determined in time.

P1B. I appear to be aware of my own existence as determined in time.

In both cases, I have beliefs about my subjective states and their objective time order. Only in P1A is there an implication that those beliefs are true.

Descartes, notoriously, argued that P1A follows from P1B. The question for us, however, is whether Kant is arguing against P1A or P1B.

If Kant is arguing against P1B, then the conclusion would be that it is impossible to have beliefs about one's subjective states in the absence of beliefs about an external, spatial world.

If Kant is arguing against P1A, then the conclusion would be that the truth of my belief concerning 'my own existence as determined in time' implies the truth of my belief that there exists a world of objects in space outside me which I perceive.

Of course, we still have to settle what it would mean for it to be 'true' that there exists a world of objects in space outside me.

On the interpretation which I give in the program, what is required is that I have a continually updated *theory* of an objective world, whose empirical support consists in the continuing stream of 'intuitions' or data which are immediately subsumed under concepts of external objects. There is no way to describe the data except in terms of the theory.

However, suppose we ask, in terms of this picture, what is it for one particular theory to be true? For example, what does the truth of the statement, 'There is a stretch limo parked outside my front door' consist in?

I said that the theory is continually 'updated'. We can imagine a sequence of experiences which would lead us to form, first one theory, then a different theory, then go back to the first theory again, and so on. It is always possible that in the light of the 'criteria of all real experience' I should have to revise my beliefs concerning what is real and what is not. For example, I might wake up the very next moment to find myself on the demon scientist's operating table, with wires attached to my scalp.

However, if one presses the question about truth, then one is forced to acknowledge the idea of a *correspondence* between my 'theory of an objective world' and that in virtue of which my theory is true, or false, as the case may be. There is therefore a case for saying that in Kant's view the idea of the 'transcendental object' or the 'noumenal' world is inseparable from the idea of the 'empirical reality' of the world of objects in space - on the grounds that truth requires correspondence - and this is the result which the second 'Refutation' establishes, or is meant to establish.

Does it establish this? According to you, all that Kant has in fact done is achieve the 'demonstration that calling something subjective presupposes our having the concept of the objective'. In other words, one can be an egocentric idealist and still accept Kant's point that in order to have a concept of my 'self' it is necessary for me to have a theory about a world of objects existing in a world in which I am spatially located. As you can see, however, everything hinges on how hard one presses the question of truth. If my belief that I exist is true, then my belief in the existence of an external world must be true. If that belief is true, then there must exist *something* - which does not belong within 'my spatio-temporal world of phenomena' - which my theory about that world ultimately corresponds to.

Finally, you raise the interesting question what happens 'if our minds change'. On Kant's theory, there is indeed room for the idea that there might be creatures whose 'forms of perception' were different from ours, although this is a very obscure aspect of his philosophy. I would be tempted to argue that noumenal reality must have a *structure* which different forms of perception mirror in different, structure preserving ways. In these terms, it does not seem possible that there could be a form of inner sense which allowed it to be true that 7 plus 5 equals 13.

All the best,

Geoffrey