To: Joel S.
From: Geoffrey Klempner
Subject: Sense datum theory and the reality principle
Date: 21 January 2005 10:59

Dear Joel,

Thank you for your email of 8 January, with your first essay for the Metaphysics program, in response to the question, 'What are the attractions of the idea of a 'sense datum'? Can the theory of sense data be defended against criticisms based on the reality principle?'

Thanks also for your second email of 8 January with your extended question about the reality principle.

Sense datum theory

I will discuss the issue about 'questioning the validity of the reality principle' below. I have no disagreements with your account of the argument against the SDT based on the assumption of the reality principle.

Two main issues. The first concerns your characterization of SDT itself. You could have raised the question whether there is such a thing as 'the' sense datum theory. Instead, you assume a particular version of the theory which is not shared by all sense datum theorists.

Your version of SDT involves the causal theory of perception. However, not all SDT theorists accept the causal theory. For example, A.J. Ayer in his earlier days (when he wrote Language, Truth and Logic) argued for a version of phenomenalism, according to which statements about an 'external world' are logically reducible to statements about sense data. Critics of the classic version of the causal theory of perception point out that there is no possible way that one could make a valid inference, from the occurrence of sense data, to an external cause.

Sense datum theorists do hold that sense data have properties, such as colour, sound, odour etc. This distinguishes SDT from Kant's theory of 'concepts' and 'intuition'. Hegel's critique in the Phenomenology of Mind is in fact directed against Kant's notion of intuition, rather than at sense data as understood by the SDT. Hegel was hostile to the very idea that there is any component of reality which is not conceptual.

According to Kant's Refutation of Idealism (2nd Edition) it is a necessary condition for the possibility of experience that I apply objective concepts - concepts involving material objects in space - to what is (indescribably) given. Going by this argument alone (i.e. ignoring what Kant says about the need to recognize the existence of a noumenal world) it would be possible, as I argue in the program, to construct a transcendental subjectivist theory according to which all that exists is the indescribable 'given' of intuition and the concepts applied to it. It is possible to mount an argument against this theory based on the reality principle, but the argument is different from that against the SDT.

The other issue concerns your first argument against the SDT. 'If we can be certain only of what we find in our senses, how can we be certain of the SDT? One way to understand this is as a critique of the causal theory of perception. In which case the SDT survives in its phenomenalist guise. If you then went on to ask, On what basis is the phenomenalist claim held to be true?' then the sense datum theorist can resort to a doubly hypothetical claim: If it is true that there is a tree outside, then the truth of 'There is a tree outside' must consist (by logical inference from the assumption that all we know to exist are sense data) in hypothetical statements about sense data.

However it can also be argued (this is a line taken by Russell at one time) that the SDT bears the same kind of relation to beliefs about an external world as any other set of data to a theory constructed on the basis of that data. The sense datum theorist does not just say, 'this is the data', but rather puts forward a theory which gives the best explanation of the data. The theory is not held to be true, merely useful. (However, Russell also held that external objects are 'logical constructs' out of sense data which you might think is inconsistent with the claim about explanation.)

The Reality Principle

Later on in the program, we will be defending a version of 'anti-realism' against the reality principle. According to the anti-realist about truth, there is no such thing as cognitively inaccessible 'truth'. Common sense believes that the world is packed full of cognitively inaccessible truths. (E.g. shake a coin in your hand and shake it again: the question whether the first shake produced heads or tails can never be answered, no matter how far human knowledge progresses.)

Later still, both realism and anti-realism will be rejected along with the very notion of 'truth' as traditionally understood. The resulting theory is still (I will claim) fully consistent with the reality principle.

There is no ultimate proof of the reality principle. It belongs to a long line of metaphysical assumptions (like Descartes 'I exist' or McTaggart's 'something exists') which define an investigation.

'What if truth is merely a world for some kind of internal consistency we can witness in a conceptual structure?' What is consistency? If there is consistency then there is also inconsistency. How can any proposition be inconsistent P with any other proposition Q? 'At this moment in time, I just feel that I cannot hold P and Q,' 'It seems to me now that it cannot be the case that both P and Q.' Those are claims whose validity is itself up for question. Maybe it only seems to me now that it seems to me now that P is inconsistent with Q.

Well then, how about, 'I can only judge what I judge.' That sounds similar to, 'I can only see what I see.' E.g. I see a pink elephant. I may believe there is no pink elephant there, but I can't avoid the vivid impression of a pink elephant.

But judgement makes a claim. If you take the reality principle away from judgement, then there cannot be such a thing as making a claim. There can be no such thing as 'structure' because a structure binds, restricts, defines. - I don't know what else to say.

All the best,

Geoffrey