In my Practical Reality I argued that the reasons for which we act are not to be conceived of as psychological states of ourselves, but as real (or maybe only supposed) states of the world. The main reason for saying this was that only thus can we make sense of the idea that it is possible to act for a good reason. The good reasons we have for doing this action rather than that one consist mainly of features of the situations in which we find ourselves; they do not consist in our believing certain things about those situations. For instance, the reason for my helping that person is that she is in trouble and I am the only person around. It is not that I believe both that she is in trouble and that I am the only person around. Give that the (good) reason to help is that she is in trouble etc., it must be possible for my reason for helping to be just that, if it is indeed possible for one to act for a good reason. In fact, this sort of thing must be the normal arrangement. The reasons why we act, therefore, that is, our reasons for doing what we do, are not standardly to be conceived as states of ourselves, but as features of our situations.

I further argued that the explanations of action that we give when we specify the agent's reasons for acting—sometimes called rational or rationalizing explanations—are unusual in being non-factive.