Metasemantics: on the limits of semantic theory

METASEMANTICS is a wake-up call for semantic theory: It reveals that some semantic questions have no adequate answer. (This is meant to be the epistemic point that certain semantic questions cannot be settled--not a metaphysical point about whether there is a fact-of-the-matter.) METASEMANTICS thus checks our default optimism that any well-formed semantic question can be settled (at least in principle). Chapter One argues that relative to certain assumptions, a question like What does 'Pollux' denote? has no adequate answer. If an answer is to be non-circular, then any answer ultimately depends on an uninterpreted term--meaning that this term occurs absent an answer to what it denotes. This, I argue, makes the answer uninformative in certain crucial respects. The lesson here essentially vindicates Quine's thesis of ontological relativity (though not his behaviorism or semantic nihilism). Chapter Two and Three build on this pessimism in considering ontic-idioms, such as 'exist', 'actual', etc. If Chapter One entails there is no saying what an ontic-idiom's extension is, these chapters show there is no saying what their intension is. Any attempt, I claim, will be equivocal. As corollaries, I show that a univocal statement of Realism about x is impossible--as well as a criterion of ontological commitment. Chapter Four considers truth-conditional semantics, generally speaking. After elaborating Davidson's claim about the folly of defining truth, I counter-balance his pessimism by showing that an informative analysis of 'true' is still possible (though only for certain translational purposes). Finally, Chapter Five evaluates a pessimistic argument concerning mental content. I argue that under externalism, a priori knowledge of content is impossible, at least for knowing whether a concept is about H2O versus XYZ. But this limit on the a priori should be unsurprising; I argue, moreover, that for other purposes we indeed know a priori what we think