Davidson's guiding thought is that we can reverse the Tarskian procedure so that rather than define truth we instead presuppose it and use the theory of truth to deliver an account of meaning.

A simple reconstruction of the train of thought might proceed as follows. A theory of meaning might enable us to form clauses of the following form:

1. (i) S means that P

2. That is, on the left-hand side we have a name of a sentence in the language while on the right-hand side we have a clause specifying its content. Now, one way of thinking of what a theory of meaning is about is to think of it as about some relation that can replace the means that relation while preserving the pairings of sentences with content specifications. Now the content specification can, undoubtedly, be used to specify the truth-condition of the sentence. So if (i) holds then the following clause,

3. (ii) S is true iff P

is guaranteed to be correct. But what we want is rather the reverse. We want to be able to move from the specification of truth-conditions (from a clause of the form (ii)) to a specification of meaning (to a clause of the form (i)). So what we need to do is to place conditions on the specification of truth-conditions in clauses of the form (ii), which will entitle us to treat these effectively as meaning specifications.