To: David S.
From: Geoffrey Klempner
Subject: Essay on 'snow is white'
Date: 28 April 2004 11:08

Dear David,

Thank you for your email of 17 April, with your second essay for the Metaphysics program, in response to the question, ''Snow is white' is true if and only if snow is white' - Discuss.

You start of by stating that 'this discussion concerns statements such as 'snow is white' from the point of view of a realist metaphysics.' It is not clear to me whether this is a stipulation, or alternatively your interpretation of the question.

If it is a stipulation, i.e. if you are simply proposing to discuss realism and its implications then one set of responses would apply. If, on the other hand, you are offering a commentary on the statement about ''Snow is white' is true if and only if snow is white' (SST) then I would respond slightly differently. Let's try both ways:

1. You say, 'Realism is the common sense viewpoint that reality consists of objects that exist independently of our experience of them and which have properties which are independent of the language with which we describe them.'

What view of reality do these statements describe? One way to find out is to consider 'how things might be otherwise'.

Consider a universe in which 'objects come into existence as we probe' (a phrase from Michael Dummett's original essay on 'Truth'). This might be the world created by a lazy deity who paints in just as much as is needed at any time to make the picture *look* complete to beings who only have a limited view. Or consider a universe in which objects have properties which are dependent on the language with which we describe them. This might be a 'world' existing in someone's imagination, where objects gain substance and form only though the process of being described (but cf. the doubts raised about this in the Sophie's World thought experiment).

Neither of these visions, I would argue, is an essential part of anti-realism. Using the apparatus of possible worlds, the anti-realist generalizes from Aristotle's view of 'future contingency' -- the idea that there are multiple future worlds. So, for example, in '5000 years ago to this day it snowed at the North Pole', in each possible 5000 year ago North Pole world (a world where there is such a thing as 'the North Pole'), conceived to be determinate in every detail, it either snows or does not snow. The anti-realist asserts that no single North Pole world is metaphysically designated to the 'actual' by contrast with all the other 'possible but not actual' North Pole worlds.

(Incidentally, Brian Tee tells me that 'It is too cold to snow at the North Pole.' I don't know whether this is true or not, but it sounds plausible.)

This is not intended as a defence of anti-realism, merely an exposition of one version of that theory. The argument for anti-realism consists in a challenge to any self-professed 'realist' to state a definition of realism which cannot be understood in anti-realist terms, as I have sought to do with your statements about 'realism'.

2. The redundancy theory of truth traces back to the British philosopher Frank Ramsay, who claimed that to say that 'P is true' is the same as asserting that P.

Arguably, Tarski's definition of truth in his famous essay goes beyond this simple redundancy claim. In Tarski's account, axioms are laid down which describe the assignments of objects to names, a, b, c... in the language, 'satisfaction conditions' for predicates F(x), G(x), H(x)... dyadic relational expressions P(x,y), Q(x,y), R(x,y)... etc (and so on) which generate theorems of the form:

'Fa' is T if and only if Fa...

'Gb' is T if and only if Gb...

'Qbc' is T if and only if Qbc...

(and so on)

The 'Tarski schema' is laid down as a condition for an adequate truth theory. The theory is adequate if and only if the axioms suffice to generate the T-sentence for every possible sentence of the language.

Tarski claimed that his account was a vindication of the correspondence theory of truth. (Here is where your 'true propositions fit reality like a glove' comes in.) It certainly *looks* as though we are describing the perfect 'fit' between sentences/ propositions and reality.

But consider vague statements. The theory is not designed for these. Attempts have been made to adapt the theory, to give a 'theories' of vagueness, but none has been completely satisfactory. The fact is, ordinary discourse deals in a large part with vague concepts, concepts with a fuzzy borderline whose 'satisfaction conditions' cannot be laid down in black and white.

You propose that vague propositions are 'not statements about truth [but do] have a natural meaning'. The idea, then, is to get rid of all vague expressions, and consider what we would say in a precise language with all vagueness removed. Perhaps this would save realism.

However, even if this could be done (and there are strong reasons to say that it couldn't be done), the anti-realist is waiting in the wings to take over all that the would-be 'realist' says about a fully determinate reality fully describable in precise terms, and apply it to a world of possible worlds as I have done above. Such an anti-realist is perfectly happy with the idea of a 'truth theory' or the Tarski schema.

All the best,

Geoffrey