Page 222 - Encyclopedia of Philosophy of Language

Truth and Meaning
is the proposition that p' and 'Einstein's claim is true.' For if truth is not a property, then we can no longer account for the inference by invoking the principle that if X is identical with Y then any property of X is a property of Y and vice versa. Thus the redun- dancy/performative theory, by identifying rather than merely correlating the contents of 'The proposition that p is true' and 'p,' precludes the prospect of a good explanation of one of truth's most significant and useful characteristics. So it is better to restrict our claim to the weak, equivalence schema: The prop- osition that p is true if and only if p.
Support for deflationism depends upon the possi- bility of showing that its axioms—instances of the equivalence schema—unsupplemented by any further analysis, will suffice to explain all the central facts about truth; for example, that the verification of a proposition indicates its truth, and that true beliefs have a practical value. The first of these facts follows trivially from the deflationary axioms. For given our a priori knowledge of the equivalence of 'p' and "The proposition that p is true,' any reason to believe 'that p' becomes an equally good reason to believe that the proposition that p is true. The second fact can also be explained in terms of the deflationary axioms, but not quite so easily. Consider, to begin with, beliefs of the form:
If I perform act A, then S will happen. (2) The psychological effect of such a belief is, roughly,
to cause the performance of A if S is desired. That is: If S is desired, then A is performed.
Also, when the belief is true, then, given the deflation- ary axioms, the performance of A will in fact be fol- lowed by S. That is:
If (2) is true, then if A is performed, S will happen. Therefore:
If (2) is true and if S is desired, then S will happen.
In other words, when the belief is true, then the agent will get S if he or she wants it. So it is quite reasonable to value the truth of beliefs of that form. But such beliefs are derived by inference from other beliefs and can be expected to be true if those other beliefs are true. So it is reasonable to value the truth of any belief that might be used in such an inference.
To the extent that such accounts can be given of all the facts involving truth, then the explanatory demands on a theory of truth will be met by the collection of all statements like 'The proposition that snow is white is true if and only if snow is white,' and the idea that some deep analysis of truth is needed will be diminished.
However, there are several strongly felt objections to deflationism. One reason for dissatisfaction is that
the theory has an infinite number of axioms, and there- fore cannot be completely written down. This alleged defect has led some philosophers to develop theories which show, first, how the truth of any proposition derives from the referential properties of its con- stituents; and, second, how the referential properties of primitiveconstituents are determined (Tarski 1943; Davidson 1969). However, it remains controversial to assume that all propositions—including belief attri- butions, laws of nature, and counterfactual con- ditionals—depend for their truth values on what their constituents refer to. And there is no immediate pros- pect of a decent, finite theory of reference. So it is far from clear that the infinite, list-like character of deflationism can be avoided.
Another source of dissatisfaction with this theory is that certain instances of the schema are clearly false. Consider (3)below:
THE PROPOSITION EXPRESSED BY THE SENTENCE IN (3) SMALL CAPITALS IS NOT TRUE.
Substituting this into the deflationary schema, one obtains a version of the 'liar' paradox, namely:
The proposition that the proposition expressed by the (4) sentence in small capitals is not true is true if and
only if the proposition expressed by the sentence in small capitals is not true.
from which a contradiction is easily derivable. (Given (4), the supposition that (3) is true implies that (3) is not true, and the supposition that it is not true implies that it is.) Consequently, not every instance of the equivalence schema can be included in our theory of truth; but it is no simple matter to specify the ones to be excluded (see Kripke 1975; Gupta 1982).
A third objection to the deflationary theory con- cerns its reliance on propositions as the basic vehicles of truth. It is widely felt that the notion of proposition is defective and that it should not be employed in semantics. If this is accepted, then the natural deflationary reaction is to attempt a reformulation that would appeal only to sentences: for example:
'p' is true if and only if p.
But this so-called 'disquotational theory of truth' comes to grief over indexicals, demonstratives, and other terms whose reference varies with the context of use. It is not the case, for example, that every instance of / am hungry is true if and only if I am hungry. And there is no obvious way of modifying the deflationary schema for sentences to accommodate this problem (see, however, Horwich 1990; Quine 1990).
3. The Role of Truth in Semantics
It is commonly assumed, following Davidson (1967), that a sentence is given meaning by associating it
200