In recent years there has been increasing interest in philosophical theories which downplay the importance of the idea that our words and thoughts represent aspects of the world. The best-known example of these non-representational theories is metaethical expressivism, the view that ethical language and thought is best understood not as representing or describing ethical features of the world, but as expressing our attitudes towards it. Other theories apply similar ideas to other kinds of language and thought, and global versions apply it to all kinds. Non-representationalism has undergone a major shift in the last few decades, and lack of clarity about what it now involves has led some to worry that it is either unintelligible, or else indistinguishable from its representationalist rivals. In the first part of my thesis, I offer a novel reading of the new kind of non-representationalism. I argue that this reading, for the first time, makes the view both intelligible and distinct from representationalism. However I also show that this reading collapses one of the major debates in the recent literature – the debate between global and local non-representationalists. This debate turns out to be empty: properly understood, the disputants already agree with each other.

Many writers think that non-representationalism threatens metaphysics, particularly theories which purport to say what makes statements of given kinds true, and to what various kinds of terms refer. Some take this to be an advantage of the view, others a disadvantage. In the second part of my thesis I argue that this common view is deeply mistaken – nonrepresentationalism does not undermine metaphysics. I respond to a number of recent arguments, showing that neither global nor local forms of non-representationalism undermine metaphysics. I argue that non-representationalism is compatible with metaphysics, and that this is not a problem for the view.