Abstract

In the wake of Kolodny (2005) and Raz (2005), the normativity of rationality has become the topic of an intricate debate: what normative reasons are there, if any, to be rational? This article explains what kinds of ‘reasons to be rational’ there are. It then argues that, while we often have reason to be rationally disposed, rationality is neither itself normative nor necessarily underwritten by normative reasons – at any rate not when construed as conformance with coherence requirements. Allegedly omnipresent reasons to be coherent are shown to be ‘ghost-like’: hard to disprove, and yet irrational to believe in. This conclusion is rendered persuasive with scenarios featuring rewards for violating coherence constraints. The article also deals with the long-standing objection that such ‘irrationality rewards’ would be reasons ‘of the wrong kind’. Furthermore, Kolodny’s explanation of why rationality has the appearance of being normative is replaced with a more credible ‘error theory’.