Convention T and the Redundancy Theory of Truth
A sentence is context-dependent if different utterances of it in different contexts have different truth values. A common kind of context-dependence is contingency. For instance, 'there are unicorns' is true when uttered in a world that contains unicorns, and false otherwise. Now look at Convention T:
'p' is true iff p.
When 'p' is context-dependent, it doesn't really make sense just to call it true. However, Convention T certainly isn't meant to apply only to non-contingent (and otherwise non-context-dependent) sentences. So what shall we make of it? Two possibilites come to mind:
1) 'p', uttered in the present context, is true iff p.
2) For every context c, an utterance of 'p' in c is true iff p in c.
(These possibilities roughly correspond to the material vs. strict reading of "iff" when the only context-dependence in question is contingency.)
It seems to me that (1) is far too weak to be of much interest: In ordinary cases, 'p' will not be uttered in the present context, that is, the context in which Convention T is stated (depending on how fine-grained contexts are individuated, the only exception may be if 'p' is convention T itself). Then (1) says nothing whatsoever on the truth of 'p'. We should at least strengthen it to
1') If 'p' were uttered in the present context, this utterance would be true iff p.
I find this still too weak, and now also too vague. Let's switch to (2). Here we face a similar problem, for consider
For every context c, an utterance of 'there is no utterance' in c is true iff there is no utterance in c.
If there really is no utterance in c, then by this principle (right to left), an utterance of 'there is no utterance' in c is true. Contradiction. So if the principle is true, there trivially is an utterance in every possible context. Whether this problem is serious depends on what we want the principle to do. If it is meant to be a semantical rule giving the meaning of sentences, I would regard it as serious, and propose to change (2) to something like:
2') For any context c, 'p' is true relative to c iff p in c.
Truth has now become relational, but perhaps that's not too weird, since we can stipulate that 'true (simpliciter)' means 'true relative to the present context'.
There is a further problem: 'p in c' will not always have the required meaning. One reason for this is that some indexicals, such as 'I', cannot be made to shift their reference by pre- or postfixing 'in c'. Example: Suppose somebody called 'Bond' says 'my name is 'Bond''. Let c be this context of utterance. What Bond said in c was true. So according to principle (2), my name is 'Bond' in c. But that's false: My name is not 'Bond' in c. I don't know of any satisfactory repair.
Anyway, if there is some principle that could plausibly replace Convention T, I doubt that it would serve the case of the 'redundancy theory of truth', as the old principle is said to do. The old principle at least provided a rule to translate any sentence of the form 'x is true' into an equivalent sentence not containing 'true'. But with context-dependent sentences this doesn't work. How could I avoid 'true' in my statement that Bond's utterance was true? 'My name is 'Bond'' will not do, neither will 'in the context of Bond's utterance, my name is 'Bond'', and neither will 'If Bond had just uttered 'My name is 'Bond'', my name would have been 'Bond'".
Well, there are ways to avoid 'true' even in such cases. I could for instance say 'Bond's name is 'Bond''. All the redundancy theory really needs is that for any term t, there is some sentence q, not conaining 'true', such that an utterance of 't is true' is equivalent (in some suitable sense) to an utterance of q. There needn't be a single simple rule like Convention T.