Abstract
On my own view likewise, if we look for a straight analysis of the concept of a reason, we will find none that works. I mean the concept of a reason to do something, as when I say that the health effects and addictiveness of smoking are reasons not to take it up. I don't mean the psychological concept of an “operative reason”, as Scanlon calls it, a person's reason for doing what she does, as in “Her reason for taking up smoking was that it would freak out her parents;” saying this doesn't commit us to thinking that that's any reason at all to take up smoking. Now for this normative concept of a reason to do something, I agree with Scanlon: we can't do better than he does by way of straight analysis; we can only say, somewhat uninformatively, that to be a reason to do something is to count in favor of doing it. On much the grounds he lays out, for instance, we can't translate talk of reasons into psychological claims about preferences or desires.