We conceive of ourselves as capable of acting in response to normative reasons. Given that normative reasons are facts, this self-conception entails that we are capable of acting in response to facts. Arguments from error cases might seem to force us to deflate this self-conception, for they seem to show that to act in light of a fact must simply be a way of acting in light of a belief. The starting point of this paper is the rejection of this deflationary view. In order to reject the argument from error cases, we should adopt a disjunctive view of motivating reasons. According to this view there are two distinct ways of acting in light of a consideration: acting in light of a fact and acting in light of a belief. Disjunctivism about motivating reasons, however, is the target of a skeptical challenge grounded in a cognitivist account of the mind. According to this account, cognition is to explained in terms of the manipulation of representations and there is no meaningful difference between acting in light of a fact and acting in light of a belief: in both cases, one decides in light of a representation. The goal of this paper is to defend disjunctivism from this objection. In order to do so I appeal to Gibson’s ecological psychology. I argue that the ecological account of perception allows us to hold that in the case of actions guided by perception we act in light of the facts themselves (not representations of those facts) and that the same approach can be extended to cover cases in which we act in light of sensorily unavailable facts.