16 December 1997
Thank you for your letter of 11 December, in response to unit 8 of The Possible World Machine, on the topic of Truth.
Can truth be defined? Yes it can. Absolutely and precisely? Yes. Here is a definition of the terms, 'It is true that' and 'is true': If you take any statement S and prefix it with the phrase, 'It is true that' then the result will be true just in the case where the original was true, and false just in the case when the original was false. In other words, apart from rhetorical emphasis, in saying 'It is true that S' you are saying the same thing as S. The same applies if you take any quoted statement, 'S' and append the phrase 'is true'. 'S' is true if, and only if S, for any statement substituted for 'S'.
Obviously, that does not solve the problem of truth. But why? What is missing?
The realist and the anti-realist agree that the definition given above correctly identifies the concept of truth. There is only one concept of truth. If you try to replace the term 'true' in 'it is true that' or 'is true' with any other term (e.g. 'verified', 'believed with the highest possible degree of certainty', 'written on the oracle stone' or whatever) then the result is that it will not always or necessarily be the case that it is 'true' that S if and only if S, or that 'S' is 'true' if and only if S.
Possibly one exception to the statement I have just made (apart from the trivial point that you could use another word for 'true', e.g. 'vrai') is the phrase 'corresponds with the facts'. A crude way of expressing the difference between the realist and the anti-realist is that the realist believes in the existence of 'facts out there' whereas the anti-realist does not (in somewhat the same way as the materialist believes in a world of objects 'external' to experience whereas the Berkeleian idealist or immaterialist denies this).
The trouble is that 'corresponds with the facts' is itself a very slippery idea. If 'the facts' was defined as necessarily identical with the diary of the Recording Angel (supposing that we believe existence of a Recording Angel) then we have to recognise the logical possibility that the Recording Angel might lie. The same applies to any physical object like an oracle stone. On the other hand, there is a benign reading of 'corresponds with the facts' according to which when we say that a statement corresponds with the facts we are simply saying that things are the way the statement states, which is a roundabout way of saying that 'S' is true if and only if S.
In short, while we can precisely identify the concept of truth, the concept identified cannot be further defined any interesting or non-circular way. That would be the end of the matter if it were not for the fact that there is, apparently, an area where it is possible to hold radically different views of the nature of truth, consistently with the definition given. The realist holds that, even in cases where we cannot ever know the truth about some particular matter, nevertheless the statements we consider as pure speculations either possess the property of truth or possess the property of falsehood. The anti-realist denies this; or, rather, refuses to understand what talk of a statement's 'possessing truth' means in this context, other than a vehement way of saying that a statement 'is true'.
Both realist and anti-realist can agree, e.g., that 'Either Pilate studied philosophy or not', but they seem to understand this 'either-or' in radically different ways. I say, 'seem to understand': the problem is not so much finding a means to resolve the dispute between the realist and anti-realist as discovering just what it is that is being disputed. Every attempt you make to characterise the bone of contention turns out to be circular!
If this problem leaves you cold, I would not scold you for failing to take a 'properly philosophical view'. It was a question that hounded my during my days as a graduate student at University College, Oxford. At the time (1978-82) the main focus of debate was on the 'construction of a theory of meaning' for the language that we speak, and whether such a theory should employ the concept of truth or the concept of verification as the 'central concept'. The Oxford philosopher of language Michael Dummett's views dominated the debate (see, e.g. his collection of articles Truth and Other Enigmas Duckworth 1978), as did the work of my own supervisor John McDowell.
I took the view that the problem of realism and anti-realism could not be pinned down to a dispute about language. The risk of doing this was that all the bad things that have been said in the past about metaphysics that metaphysical disputes are empty, or meaningless attacks to which Dummett's re-definition of metaphysics as the philosophy of language was intended to fend off, would be thrown back in my face. And they were! Even today, it is an issue that causes me deep anxiety. In The Ultimate Nature of Things, I come as close as I can to a resolution of sorts, basically rejecting both 'realism' and 'anti-realism' in favour of a pragmatic theory of truth.
I do understand why so many are non-plussed by the problem. However, the metaphysical poet John Donne is on my side. 'Tell me where all past yeares are' (from his poem entitled simply 'Song') seems to me a clear expression of perplexity concerning the realist view of statements about the past. For once, the metaphysics is more than a mere 'conceit'!