To: Matthew M.
From: Geoffrey Klempner
Subject: Can truth be defined?
Date: 27 June 2007 10:46

Dear Matthew,

Thank you for your email of 26 June with your third essay for Possible World Machine, which you first attempted to send on 19 June, in response to the question, 'Define truth.'

The original question for units 6-9 was, 'Can truth be defined? - If you think that it can, give a definition and explain its philosophical significance. If you think that it cannot, what conclusions should the philosopher draw from that?' I'm mentioning this because a possible response to the statement, 'Define truth' is to say, 'Sorry, it can't be done!'

You've said a lot of things about truth, but haven't in fact offered any definition apart from mentioning the correspondence theory in paragraph one. You go on to state that there is a difference between 'the idea of truth' and 'the valuation of proper ways of judging'. I'm not sure whether what you have in mind here is the difference between a definition of truth, which states what truth IS, and a criterion of truth, which gives a recipe for determining whether a given judgement is true. It is generally accepted that the latter would be extremely difficult to do, although there can be methodologies for particular branches of inquiry which help us to discover truth.

Let's stick to the correspondence theory for the moment. Clearly, it looks a lot easier to say what we mean by 'correspondence' when we are talking about ordinary empirical, contingent truths such as 'There's a turtle dove in the garden.' On one side is the statement or proposition, on the other side how things are in the world, by virtue of which that statement is true (if it 'corresponds') or false (if it fails to 'correspond').

With logical truth on the other hand, for example, 'If P then P', it is harder to see what could be meant by 'correspondence' since the statement is true no matter how things are in the world. That is one reason which led Wittgenstein in the Tractatus to deny that logical truths, or 'tautologies' say anything at all.

Kant proposed a third variety of truth, which he called 'synthetic a priori' which is concerned, as you rightly state, with 'the way that all humans have to view the world'. Actually, Kant's argument is not just intended to work for humans but for any beings who share our 'forms of sensibility', i.e. gain access to a spatial world through sensory perception. Very briefly, the argument is that there can be no experience at all unless we can interpret that experience as perception of a world of objects in space. Causality and the permanence of 'substance' are two synthetic a priori principles which follow (or which Kant believed to follow) from this. Here the 'truth' in question seems more meaty than mere logical truth. There is a sense, perhaps, in which one can say that the world of objects in space and time 'necessarily corresponds' to the a priori law of determinism (as Kant held).

In the unit on truth, quite a lot is made of the question of 'philosophical truth'. There are many forms of philosophical argument: transcendental argument (as Kant described his method of proof), philosophical analysis, phenomenological description and Hegelian dialectic - each claiming to establish some kind of necessary 'truth'. How are we to conceive of this truth? Is it out there waiting to be discovered, like ordinary contingent facts? Or is it invented? Would this truth still exist even in a world whose inhabitants were too busy with practical and technological matters to bother with philosophy?

It would be an interesting exercise to determine to what extent, if any, the notion of 'correspondence' gives a useful insight into these various (alleged) forms of truth. It could be argued that the notion of correspondence is really just a way of attempting to capture the 'realist' notion of a 'truth out there waiting to be discovered', by contrast with the 'anti-realist' view that truth is not discovered but in some sense created or brought into being as we question and probe.

Faith is another category of truth which, again, seems fundamentally different from contingent, logical or philosophical truth. Although one sometimes talks of 'faith' in things that are presupposed and never questioned - e.g. like my faith that the floor will not collapse under me, or my faith that induction is a valid way of acquiring knowledge - as you state, there is another important category of 'faith' where the subject makes a decision, takes an action to embrace a belief, which is not (as is normally the case with belief) seemingly forced on the subject by his/her perception of the facts but rather seems a matter of free choice and commitment. This was a very important idea for Kierkegaard, who was hostile to any attempt to 'prove' or find a factual basis for religious belief.

So what about truth? Can it be defined? In my view, the only possible definition of truth is that it is the predicate which satisfies the following 'Tarski schema':

'P' is T if and only if P

Where for the variable P one may substitute any sentence you like - contingent, logical, synthetic a priori or whatever. In other words, stating that 'P' is true, for any P, is simply a way of removing the quotation marks. To state a proposition IS to state that it is true. Of course that doesn't tell us what we really want to know - what it is that 'makes' any proposition true. But why should we assume that there could ever be an answer to that question?

All the best,

Geoffrey