1. From my survey of the literature, there are four main types of truth theory being discussed: substantive theories, nihilist (for want of a better label) theories, deflationary theories, and identity theories. Let me say just a little about the first two main types and then move on to deflationism. The Commenter (William Woking) will be sure to disagree with me about deflationism, which is good: by abrasion the pearl (of wisdom) is formed. Or as I read on a T-shirt at a road race recently: No pressure, no diamonds.
2. Substantive theories maintain that truth is (i) a metaphysically substantive item, presumably a property or relation, (ii) susceptible of non-trivial analysis or explication. Correspondence, coherence, and pragmatic theories count as substantive theories. Such theories purport to analyze truth in terms of other, presumably more basic, terms such as a relation of correspondence or adequation to 'reality' or to facts as in Veritas est adequatio intellectus ad rem. Or in terms of coherence of truth-bearers (beliefs, propositions, etc.) among themselves. Or in terms of conduciveness to human flourishing as in William James' "the true is the good by way of belief." Or in terms of broadly epistemic notions such as rational acceptability or warranted asseribility as in the Putnamian-Peircean 'Truth is rational acceptability at the ideal limit of inquiry.'
The latter is not a good proposal for reasons I won't go into now, but it illustrates the project of giving a substantive theory of truth. One tries to analyze truth in more basic terms. One tries to give an informative, noncircular answer to the question, What is truth? The sunbstantive approach is in the Granbd Tradition deriving from Plato wherein one asks What is X? for many values of 'X.'
The substantive approach to truth can be summed up in three propositions:
A. The facts about truth are not exhausted by the substitution-instances of the equivalence schemata 'p' is true iff p and *p* is true iff p.
B. There is a substantive property of truth common to all and only truths.
C. This substantive property is analyzable.
3. The 'nihilist' as he is known in the truth literature rejects substantive theories, not because they are substantive, but because they are theories. He may grant that truth is a deep, substantial, metaphysically loaded, ontologically thick, topic. But he denies that one can have a theory about it, that one can account for it in more basic terms: truth is just too basic to be explained in more fundamental terms. The nihilist accepts (A) and (B) above but denies (C).
4. The deflationist, like the nihilist, rejects substantive theories of truth. The difference is that the deflationist holds that an account of truth is possible albeit in very 'thin' terms, while the nihilist denies that any account is possible thick or thin: truth is too basic to be accountable. Nihilism allows truth to be a thick (metaphysical) topic. Deflationism disallows this. Deflationists deny (A), (B), and (C).
5. The deflationist makes a big deal out of certain perfectly obvious equivalences and he tries to squeeze a lot of anti-metaphysical mileage out of them. Here are two examples, one involving a declarative sentence, the other involving a proposition. Note that asterisks around a sentence, or around a placeholder for a sentence, form a name of the proposition expressed by the sentence.
E1. 'Grass is green' is true iff grass is green.
E2. *Grass is green* is true iff grass is green.
Note that such biconditionals express logical, not material, equivalences: they are not just true but true across all metaphysically (broadly logically) possible worlds. With respect to such biconditionals, there is no possible situation in which the RHS is true and the LHS false, or vice versa. If asked for the ground of this necessity, I would say it resides in the mere logic of the truth predicate. Saying this, I do not concede that there is nothing more to truth than the merely syntactic role played by 'true ' in equivalences like the above.
Now let us assume something which, though false, will simplify our discussion. Let us assume that there is no other type of use of the truth predicate other than the uses illustrated in logical equivalences like the foregoing. (Thus I am proposing that we ignore such uses as the one illustrated by 'Everything Percy says is true.')
The deflationist thesis can now be formulated as follows: There is nothing more to truth than what is expressed by such truisms as the foregoing equivalences. Thus there is no metaphysically substantive property of truth that the LHS predicates of 'Grass is green' or of *Grass is green.* The content on both sides is exactly the same: 'is true' adds no new content. 'Is true' plays a merely syntactic role. In terms of Quine's disquotationalism (which is a version of the deflationary approach), 'is true' is merely a device of disquotation. 'Is true' has no semantic dimension: it neither expresses a substantive property, nor does it refer to anything. Truth drops out as a topic of philosophical inquiry. There is no such property susceptible of informative explication in terms of correspondence, coherence, rational acceptability, or whatnot. The question What is truth? gets answered by saying that there is no such 'thing' as truth: there are truths, and every such truth reduces via the equivalence schema to a sentence or proposition in which the truth predicate does not appear. Accordingly, there is nothing all truths have in common in virtue of which they are truths. There is only a multiplicity of disparate truths. But even this says too much since each 'truth' reduces to a sentence or proposition in which 'true' does not appear.
6. Now for my misgivings about deflationism. But first three preliminary points.
a. Equivalence is symmetrical (commutative); if p is equivalent to q, then q is equivalent to p. But explanation is asymmetrical: if p explains q, then q does not explain p. From ' p iff q' one cannot infer 'p because q' or 'q because p.' 'p iff q' is consistent with both. Connected with the asymmetry of explanation is that equivalences do not sanction reductions. Triangularity and trilaterality are logically equivalent properties, but it doesn't follow that either reduces to the other.
b. If two items are equivalent, then both are propositions or sentences. There cannot be equivalence between a sentence or proposition and something that is neither.
c. To define equivalence we need to recur to truth. To say that p, q are logically equivalent is to say that there is no possible situation in which p is true and q false, or q true, and p false.
Now what is the deflationist saying? His thesis is negative: there is nothing to truth except what is captured in the the equivalence schemata and their substitution-instances. Consider
E. *p* is true iff p.
First Misgiving: The truth of the biconditional is not in question. But equivalences don't sanction reductions. From (E) one cannot infer that the LHS reduces to the RHS, or vice versa. But the deflationist is saying that the LHS reduces to, and is explained by, the RHS. But what is his justification for saying this? Why not the other way around? Why not say that p because *p* is true?
Second Misgiving: For an equivalence to hold, both sides must be true (or false). Suppose both sides are true. Then, although the predicate 'true' does not appear on the RHS, the RHS must be true. So, far from dispensing with truth, the equivalence schemata and their instances presuppose it!
You don't get it, do you? Let me try an analogy with existence. A deflationist about existence might offer this equivalence schema:
F. Fs exist iff something is an F. (E.g., 'Cats exist iff something is a cat.')
Squeezing this triviality hard, our deflationist announces that 'exist' plays a merely syntactic role and that there is no substantive property of existence. But is it not obvious that if something is an F, then that thing must exist? Are you quantifying over a domain of nonexistents? If yes, then the equivalence fails. But if you are quantifying over a domain of existents, then the existence of those existents is being presupposed. So, even though 'exist' does not occur on the RHS of (F), existence is along for the ride. Same with (E). Even though 'true' does not occur on the RHS of (E), truth is along for the ride. In both cases, existence and truth in meaty substantive senses are being presupposed.
Third Misgiving. 'Grass is green' and 'It is true that grass is green' have exactly the same content. That is perfectly obvious and denied by no one. 'Is true' adds no new content. But how is it supposed to follow that truth is not a substantive property? What follows is that truth is not a content property. How do our deflationist pals get from 'Truth is not a content property' to 'Truth is not a substantive property'? Isn't it obvious that truth refers us outside the content of the proposition or sentence?
Compare existence. A thing and the same thing existing have exactly the same quidditative content. The fastest runner and the existing fastest runner are numerically the same individual. Does it follow that existence is not a property? No, what follows it that existence is not a quidditative property. Same with truth. There is no difference in content between p and true p. But it makes a world of difference whether p is true or false just as it makes a world of difference whether an individual exists or not.
Fourth Misgiving. If p and q are equivalent, then both are propositions. The instances of (E) therefore do not get us outside the 'circle of propositions.' But isn't it obvious that whether or not a sentence or a proposition or a belief (or any truthbearer) is true or false depends on matters external to the truthbearer?
>>Isn't it obvious that truth refers us outside the content of the proposition or sentence?
Yes. 'S says (or states, or asserts) that p' is merely about the content of the sentences itself. 'It is the case that p' is not merely about the content. 'It is the case...' or 'it is true ..' put forward or propose or assert the content itself.
Or: the truth of 'S thinks that p' depends only on the state of someone's mind. 'It is the case that p' depends on something more.
This is consistent with deflationism, nonetheless.
Posted by: William | Wednesday, June 16, 2010 at 12:25 AM
>>If yes, then the equivalence fails. But if you are quantifying over a domain of existents, then the existence of those existents is being presupposed.
Rather, if you are quantifying over a domain of things, then you are quantifying over a domain of things. What you want to say above is already given by the word 'thing'.
Posted by: William | Wednesday, June 16, 2010 at 12:32 AM
>>Or: the truth of 'S thinks that p' depends only on the state of someone's mind. 'It is the case that p' depends on something more.<< That is certainly true.
>>This is consistent with deflationism, nonetheless.<<
No! It is clear that I haven't got through to you.
Posted by: Bill Vallicella | Wednesday, June 16, 2010 at 07:55 PM
>>No! It is clear that I haven't got through to you.
I knew you would say that! But why?
Posted by: William | Thursday, June 17, 2010 at 12:33 AM