To: David Y.
From: Geoffrey Klempner
Subject: Kant's 2nd refutation of idealism
Date: 27 November 2006 12:03

Dear David,

Thank you for your email of 20 November, 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'.

You are right to start with Descartes. As Kant makes clear in his preliminary discussion, Descartes is his intended target, the 'problematic idealist' who seeks proof of an external world (by contrast with the 'dogmatic idealist' who regards matter and space as something 'utterly impossible').

I think Kant is a little bit off the mark here. As I argue in the program, Kant's argument does meet the challenge of one kind of 'dogmatic' idealist - the naive subjectivist who refuses to understand what could be meant by 'external objects' because all I can possibly know is my own subjective experience.

The person Kant had in mind as being 'dogmatic', however, was Berkeley. Here, it could be argued that Kant is much closer to Berkeley than he wants us to think (the Refutation is from the Second Edition to the Critique of Pure Reason where Kant was concerned to rebut allegations that he was an 'idealist' like Berkeley). Both Kant and Berkeley reject the idea that space and matter are something 'in themselves'. They differ only in that Berkeley attempted to explain the nature of the noumenal - as 'archetypes in the mind of God'.

But lets stick with Descartes. What we would expect - and what we in fact find - is an argument which goes along the following lines:

1. I know I exist in time as a thinking subject.

2. If I do not perceive objects in space then I cannot have knowledge of my own identity through time as a thinking subject.

Therefore, 3. I do perceive objects in space.

Note that there is no mention, either in my summary or in the original text, of 'other subjects'. This just isn't a topic for Kant. He assumes that anyone can run this argument for themself, and when they do, the upshot is all that is required to escape the clutches of idealism.

Your example of perceiving a Chimera on an unexplored planet would do better as an illustration of what Kant goes on to establish after the refutation of idealism: that all experiential knowledge presupposes the law of causality. In the refutation, causality is implied in the idea of a world whose changes are predictable, but not explicitly alluded to.

I disagree that Kant calls this thing 'space'. By 'thing' Kant means what we would mean, things, like tree or a house. (Later, along with the argument for the law of causality, Kant makes a stronger claim for the permanence of 'substance' - the stuff of which trees and houses etc are ultimately made of.)

(By the way, I don't agree with Mautner's definition of 'noumenon'. To my knowledge, nowhere in the Critique does Kant claim that a noumenon is an 'object of awareness' as such. He denies that we can have any knowledge of noumena, direct or indirect. All we can know is that they must exist. If you like, you can call this an 'intellectual acknowledgement' of the necessity for noumena.)

How exactly does Kant's refutation of idealism work?

This is the hard part. I have a particular interpretation in mind (which goes along similar lines to that of Strawson in his book 'The Bounds of Sense') but it is not necessarily the only possible account of the structure of Kant's argument.

On this interpretation, 'awareness of my own existence as determined in time' should be understood as 'awareness of my own identity through time'. In other words, I am aware that yesterday I did X and the day before yesterday I experienced Y, and so on. The claim is not merely that X was done or that Y was experienced but that I did and experienced these things. This is, after all, what Descartes is claiming when he says in the Discourse, 'I think therefore I am'.

(Why can't the egocentric subjectivist pull in his horns and say, 'X was done and Y was experienced'? Why not just give up the I? The problem is that the identification of a subject is required in order to distinguish between true and false memory.)

Now comes the controversial step: in order for there to be a genuine distinction between what I seem to remember and what actually happened there must be room for a clash between what I predict on the basis of my memory, and my present experiences. There might never be such a clash in practice if the world is perfectly well behaved and never trips me up; the point, however, is that we have to allow logical room for such a clash. For example, I remember digging a hole, but when I visit the spot there is no hole to be seen. Either the hole was filled in, or I have forgotten where I dug the hole, or maybe I just dreamed that I dug a hole.

If my experience is not 'as of' a world of objects in space, any clash between what I seem to remember and my present experiences can be fixed any way I like. What the 'theory' of objects in space supplies is the necessary empirical constraint which prevents the 'fixing' from being merely arbitrary.

There is still the worry why Kant wasn't more explicit about the role of the distinction between 'true' and 'false' memories, but perhaps he just thought that this point was too obvious to need labouring.

As I said before, I don't think that 'other subjects' play a role here. The argument would work perfectly well for a solitary individual who lived his entire life unaware that any other conscious beings existed. Wittgenstein's private language argument gives reasons for possibly questioning that assumption, although this again is controversial.

All the best,

Geoffrey