If the previous chapter is right, the concept of truth is an appendix to our practice of saying how things are. But what is it to say how things are? More generally, what is it to say something meaningful? What is it for words or sentences to have meaning? These are the central problems of philosophical theories of meaning.

I start by sketching Quine's treatment of these questions (section 1). This sets the stage for Davidson's truth-conditional theory of meaning. Unlike Quine, Davidson rarely tackles the question of what linguistic meaning is head-on. But the assumptions behind his approach, notably his antipathy towards ‘meanings’, his third-person perspective and his holism, derive from Quine. Davidson's departure consists in his focus on the question of what form a theory of meaning for natural languages should take. Section 2 presents the requirements Davidson imposes on such a theory, and his argument that they are met by a Tarskian truth-theory. The core of that argument is the idea that the meaning of a sentence is determined by its truth-conditions. Section 3 discusses various problems with that idea. The final section turns to some of the recalcitrant idioms that pose a threat to the application of truth-theories to natural languages.

Quine on reference, meaning and use

Quine radically challenges some of the assumptions about meaning that guided previous analytic philosophy.