(Which facts? Typically, facts including some kind of evaluation of the desired object as, for example, pleasant, interesting, advantageous, stature-enhancing, decent, and the like.)
Sometimes those facts include our “subjective reactions”. For example, “a large part of the point of eating ice cream or taking a vacation is doing something that I will enjoy, so one's “subjective reactions” are obviously of prime significance to the reasons one has for doing these things one way rather than another.” (1998: 42). But notice that these subjective conditions are not facts about a person's desires but facts about what a person enjoys.
Other times subjective reactions have no importance and they should be corrected to fit the worth of their object; so for example the fact that I have a reason to prevent the destruction of some great building, it has nothing to do with how I feel about it. (1998: 42)
There might be cases in which it appears that desires provide additional reasons for actions. Scanlon argues that, appearances notwithstanding, this is mistaken. There are at least two roles that desire can play, which are worth mentioning (none of which corresponds to being an independent source of reasons for actions):
(1) desires provide evidence of future enjoyment (44-45);
(2) there are states - that we may vulgarly call “desire” - which involve, in addition to the recognition of a certain fact as a good reason, the agent's decision to take those facts as grounds for action. (p. 46)
The latter fact expresses a desire or intention, understood as a provisional adoption of a plan of action. Facts about the agent's intentions (understood in this way) make a difference to what the agent has most reasons to do, since at every moment there are many goals that are worth considering, and some plans of action must be selected. Following Bratman, Scanlon holds that there is a reason to carry out a plan that has been already adopted unless there is some reason to reconsider it. But such plans are not independent sources of reasons, since their adoption is justified, at the end of the chain of justifications only by reasons of the more usual kind (1998: 46).