. Truth is something we all approve of, and aim, or should aim, to achieve. But just what is it? What we say may be criticized in many ways, only one of which is for failing to state the truth (see below). But failure to do that seems somehow to be the most basic kind of failure. What then is it that we want to achieve? Is truth a relation between what we say and the subject-matter we say it about? Or between different things we do or could say? Or is to call something true simply a way of repeating it? Or perhaps of expressing a certain kind of approval of it or willingness to endorse it? Or is something true if, in a certain sense, it works or produces satisfaction? Various theories exist corresponding to such views.
The correspondence theory is perhaps the commonest theory of truth, partly because ‘correspondence’ can be interpreted strictly or loosely. In its strictest form, primarily associated with Moore and Russell, this theory involves a relation between two things, that which is true (a proposition, belief, judgment, etc.) and that which makes it true (a fact, or perhaps a state of affairs or event). The fact has a structure which the proposition, etc., copies or pictures. But finding pairs of things which correspond in this way is difficult, especially since the sort of structure that a proposition might have, involving the relations between things like nouns and verbs, or subjects and predicates, seems entirely different from any features of the outer world. Subjects and predicates are different (linguistic) things in a way that objects and properties are not; and how would one distinguish the correlates of ‘Snow is white’ and ‘For snow to be white’? Similar difficulties confront correspondence, or picture, theories of how sentences or propositions have MEANING. Also if all we know is propositions, and propositions picture the world, how can we compare the propositions with the world itself to see if they picture it accurately? (Cf. PERCEPTION on the difficulty of knowing about an external world if we must begin from ‘pure experience’.)
There is a less strict form of the theory, which no longer requires that each part of the proposition should correspond to a part of reality. On this view something is true if it can be correlated with a fact and false if not, or if its negation can be correlated with a fact. A still looser form calls something true if it simply ‘says things as they are’. This leads naturally to F.P.Ramsey’s redundancy (also occasionally called no truth) theory, whereby to call something true is simply to repeat what it says. On this view truth is not a property of anything, but the use of ‘true’ provides a shorthand way of referring to things said. A development of this is Strawson’s performative (occasionally, ditto) theory, which is a SPEECH ACT theory. Here when we call something ‘true’ we perform an act of agreeing to, repeating, or conceding it—we say ‘ditto’ to it. These theories, which are often called deflationary, have been criticized as unable to cover all senses of ‘true’, and loose correspondence theories have been returned to and defended by Austin and, in a modified form, Hamlyn. Recently deflationism has been revived by Horwich’s minimal theory.
A recent variant of the correspondence theory is Tarski’s semantic theory. This is primarily designed for, though not confined to, artificial languages, whose elements or ‘words’ may be infinitely many but are of definite kinds and have definite roles. Sentences in a given language, L, are called ‘true-in-L’ (not to be confused with ‘L-true’, Carnap’s symbol for ‘logically true’: see ANALYTIC, VALID) when their elements are so combined as to state what is the case, e.g. ‘Snow is white’ is true in English if and only if snow is white. This avoids treating propositions and facts as entities, and is easy to develop in a rigorous and formal way. But there are limits to how far it can be applied without great complication, even to formal languages, for reasons connected with the LIAR PARADOX and GÖDEL’S THEOREMS.
The correspondence theory suits philosophers who make a sharp distinction between knower and known. Those who refuse to do this, notably IDEALISTS, often favour the coherence theory. The basic idea of this is that something is true if it coheres or is logically consistent with a wider system than any of its rivals cohere with. This theory presupposes that CONSISTENCY can be defined independently of truth, and also that one of the various possible consistent systems of propositions is wider than any of the others. Idealists who hold this theory, however, say that strictly only the system taken as a whole is true. Single propositions in it give only partial approximations to the truth. Can we really understand ‘Caesar crossed the Rubicon’, it might be asked, and so be in a position to call it true, until we appreciate all its causes and consequences, even to the end of time? The theory is therefore sometimes called a degrees of truth theory. Another reason for thinking truth has degrees is that we cannot know whether something belongs to the widest coherent system, if there is one, without examining every proposition, so in practice we must be satisfied with something less. (See also below.) Moreover, in advanced sciences like cosmology it is often difficult to decide between different but consistent ways of describing the universe. All the known celestial movements can be explained on a geocentric theory, if it is sufficiently complex. The coherence theory can be supplemented by devices like adopting the simplest out of several competing hypotheses. It is then attractive in those regions of science where immediate verification is impossible. Perhaps partly for this reason it was favoured by some positivists (Neurath, Hempel). Cf. also CONVENTIONALISM, INSTRUMENTALISM, POSITIVISM.
This sort of consideration also lies behind the pragmatic theory, for which see PRAGMATISM.
What sort of things can be true? Is ‘true’ properly or primarily applied to mental acts and states such as acts of utterance, judgments, beliefs, or to linguistic things, such as indicative sentences, or to certain abstract things such as propositions (see SENTENCES)? ‘Statement’ can be a convenient non-committal word. How we answer this will affect what we say about how truth is related to time and tense: can something, like the sentence ‘I am hot’, become true?
Is everything which is of the right kind to be true or false actually true or false, as one form of the law of EXCLUDED MIDDLE asserts? Is ‘My wife is asleep’, said by a bachelor, false (Russell) or neither true nor false (Strawson)? Strawson’s view claims that there are ‘truth-value gaps’, i.e. an absence of TRUTH-VALUES where one might expect them. Some have said certain statements about the future are neither true nor false, since the future is not yet determined. Cf. FREEWILL (for logical determinism).
Compare also cases in fiction and mythology, like ‘Unicorns are vegetarian’, where the legends give no evidence. There are other things which seem to be of the right kind to be true or false but are sometimes thought to be neither, or to be not really of that kind. Examples include value statements, laws of nature and counterfactual CONDITIONALS. It seems, then, that some things fail to be true or false because they are not of the right kind, while others may be not false but wrong for some other reason (‘My wife is asleep’, said by a bachelor, ‘You are a wog/Canuck’ said to a foreigner/Canadian). Metaphors and comparisons, etc., raise again the question of degrees of truth, e.g. should exaggerations be described as nearly true, or fairly true, or containing some truth? The HEAP paradox has also been thought to call for degrees of truth.
Special problems arise in certain cases. What is it for a logical or mathematical statement to be true? How is ‘true’ related to ‘provable’? (Cf. GÖDEL’S THEOREMS, mathematical INTUITIONISM.)
A distinction is often drawn, though also often ignored, between the meaning of ‘true’ and the CRITERIA of truth. It is sometimes unclear to which of these a given theory is intended to apply.
Another use of ‘true’ and ‘false’ is that whereby predicates can be true or false of subjects. Linguistic PHILOSOPHERS sometimes ask what significance there is in such facts as that artificial teeth, but not artificial silk, are called false, while no teeth are true. See also SATISFY, TRUTH-VALUE, FACTS, VALID, LIAR PARADOX, philosophy of MATHEMATICS.
*M.Black, ‘The semantic definition of truth’, Analysis, vol. 8, 1948, reprinted in M.Macdonald (ed.), Philosophy and Analysis, Blackwell, 1954. (On Tarski. Easier than Tarski though it presupposes elementary knowledge of logic.)
D.Davidson, ‘The structure and content of truth’, Journal of Philosophy, 1990. (Discusses Tarskian truth, and rejects both deflationist and correspondence theories, and also antirealist theories (such as the coherence theory and pragmatism), changing some of his earlier views, and finally proposing a further theory of his own.)
D.W.Hamlyn, ‘The correspondence theory of truth’, Philosophy Quarterly, 1962. (Fairly difficult defence of loose correspondence theory.)
P.Horwich, Truth, Blackwell, 1990. (States a deflationary theory, calling it ‘minimal’, though distinguishing it from redundancy theory at pp. 39–41; then defends it against thirty-nine objections.)
H.Joachim, The Nature of Truth, 1905. (Coherence.)
D.H.Mellor, Real Time, Cambridge UP, 1981. (See chapter 2 on relations between truth and time: is truth timeless in all cases?)
G.E.Moore, Some Main Problems of Philosophy, Allen and Unwin, 1953, Collier Books, 1962, written much earlier.
G.Pitcher (ed.), Truth, Prentice-Hall, 1964. (Contains papers by J. L.Austin, P.F.Strawson, G.J.Warnock, M.Dummett, with excerpt from F.P.Ramsey. On Dummett cf. bibliography to TRUTH-VALUE.)
*B.Russell, The Problems of Philosophy, Home University Library, 1912, chapter 12. (Cf. also Russell in bibliography to LOGICAL ATOMISM.)
R.M.Sainsbury, ‘Degrees of belief and degrees of truth’, Philosophical Papers, 1986. (Defends degrees of truth in connection with heap paradox.)
S.Soames, ‘What is a theory of truth?’, Journal of Philosophy, 1984. (See esp. p. 411 for brief statement of three things a theory of truth has tried to do.)
P.F.Strawson, Logico Linguistic Papers, Methuen, 1971. (Includes several relevant items. For truth-value gaps see chapters 1, 4.)
A.Tarski, ‘The semantic conception of truth’, Philosophy and Phenomenological Research, 1944, reprinted in H.Feigl and W. Sellars (eds), Readings in Philosophical Analysis, Appleton-Century-Crofts, 1949.
This is the complete article, containing 1,720 words
(approx. 6 pages at 300 words per page).
