if 'the Snark was a Boojum' is a sentence of the object language under consideration, we must define 'true'
'The Snark was a Boojum' is true if and only if the Snark was a Boojum.
But what does this tell us? If it tells us anything (and a strong case can be made out for the view that it doesn't), what it dos is to convey the information that the sentences on the left and right of the 'if and only if' are somehow equivalent: that, e.g. anyone who accepts the one is committed to accepting the other. This may be an important fact, if it is a fact, about the usage of 'true', but it is hardly the sort of fact that the correspondence theorist thought that he was pointing out.
Putnam – 'Do true assertions correspond to reality?' in Mind, Language and Reality – Philosophical Papers vol. 3. – 1975 (1960), p. 71.
we have been mislead into thinking that Tarski's semantics is semantics by the usual inductive definitions of satisfaction and truth in an interpretation. There it seems that we have something like the world, and denotation, and that we are establishing a relation between sentences and the world. It appears however that this is only a mirage, and the truth is part of the morphology of language.
Tarski's approach through (T) undermines his work on truth both as an attempt to capture the classical conception of truth and as an attempt to formulate a semantic account of truth.
Chateaubriand – Logical Forms, 2001, pp. 230, 214.
The reason for my uneasiness concerning the notion of truth was, of course, that this notion had been for some time attacked by some philosophers, and with good arguments. It was not so much the antinomy of liar which frightened me, but the difficult of explaining the correspondence theory: what could the correspondence of a statement to the facts be?
Tarski's theory, as all you know, and he stressed first, is a rehabilitation and an elaboration of the classical theory that truth is correspondence to the facts.
Popper – Objective Knowledge, 1972 (196?), pp. 320, 323.
convention T is expressed in terms of the object language L and the metalanguage ML, but it states something about the relations between language L and the (rest of) the world. This hold even of the example (5) [(5) 'snow is white' is true if and only if snow is white], which is based upon contingent facts about the use of the terms 'snow' and 'white' in English. It is a tautology to state that 'snow' means (the same as) 'snow', but it is a fact of English that 'snow' means snow.
Niiniluoto – Tarskian Truth as Correspondence – Replies to some objections, 1999, p. 97.
