<p>Expressivists aim to account for our ordinary discourse about moral reasons, which involves the distinction between merely having a pro tanto reason to do something and having all-things-considered reasons for or against it. However, they typically hold that reasons-talk lacks cognitive content and instead functions to express non-cognitive commendations of ‘doing something’ with our reasons—whether that means acting in accordance with them, being inclined to do so, or merely entertaining them. My main contention is that this is implausible, since we don’t typically approve of being moved, to any significant extent, by outweighed normative reasons. Those are situations where a subject believes that there is pro tanto reason for ϕ-ing while simultaneously believing that, overall, ϕ-ing ought not to be pursued. To demonstrate this, I apply the argument, in turn, to the accounts of Simon Blackburn, Neil Sinclair, and the Preferentialist framework outlined by David Sobel and Daan Evers. I ultimately conclude that contemporary expressivist proposals fall short of accommodating our moral reasons-talk.</p>