Currently a large part of what I do is what Al Mele has (amusingly) called ‘Reasonology’. I’m interested in the way in which what we think reasons are and do (that is, I’m interested in the way what reasons are and do) has an effect on the debate between particularists and generalists, particularly in the moral domain. One thing I’m trying to work out is how we might get an argument for particularism – that would be really good – or just an outline of what a particularist view would look like if holism about reasons were false. Now, I’m inclined to say that holism isn’t false. But there are certain sorts of considerations that prevent me from fully endorsing that view. For that reason, I think it’s very likely that some hybrid view is correct; a view according to which some reasons behave atomistically and some holistically. But it depends on what reasons are, I think.
The following is a list of desiderata (I think) for a theory of reasons, and by ‘a theory of reasons’, I mean an account of ‘What it is to be a reason’ (What’s it like to be a reason?):
- A theory of reasons ought not to contradict the way we talk about reasons in our everyday lives.
- A theory of reasons ought to respect the distinction between reasons, values, and oughts.
- On any theory of reasons, reasons ought to be able to work in opposite directions – that is, it ought to be possible for there to be a reason for acting and a reason against acting, all in the same situation.
- It ought not to be crazy, according to the view, for whatever it is that constitutes a reason both to be a reason for and a reason against the same action.
- Reasons ought to be stackable. That is, the more reasons there are for acting, the harder it is to justify not acting.
There’s got to be more, right?
I think 1 might be problematic. In one sense, of course requiring we revamp our entire way of speaking seems troubling. But sometimes we just get things wrong and it takes advances in theory (more often in science) to help us get it right. So, maybe respecting the way we talk about reasons is asking for more conservatism in our theorizing than is necessary.
Regarding 2, this seems kind of right. Reasons and value do seem to come apart. For instance, WD Ross famously argued against utilitarianism by appealing to special relationships. Suppose I have something of value and am trying to decide whether to give it to my friend Ann or to give it to a stranger. By giving it to Ann, I would promote 1000 units of value, whereas if I gave it to the stranger I would promote 1001 units of value. According to the (simple) utilitarian, I ought to give it to the stranger because 1001 > 1000. According to Ross, apart from the value promoted, there is a reason to give it to Ann – viz. that she is my friend. So, there are reasons that are not directly related to value. Naturally, the response is that the value of the friendship is what accounts for the reason to give it to her. Then, the response to that is that the value sums 1001 and 1000 are meant already to have taken account of the value promoted by the fact of our friendship; I should still give it to her.
3, 4, and 5, all seem pretty much right, to me. 3 and 5 seem most obviously true, whereas 4 might require an example to show how it works. Stephen Kearns and Daniel Star employ a nice example in a paper on their view according to which reasons are evidence of oughts. The example, if memory serves correctly, goes something like this: Suppose we’re in a footrace and we’re both near the finish line. The fact that if I trip you I will win is both a reason to trip you and a reason not to trip you. It’s a reason to trip you because, well, then I’ll win. It’s a reason not to trip you because to do so would be to win by cheating and winning by cheating isn’t good. So, the fact that if I trip you, I will win is both a reason for and a reason against tripping you.
So, at least 3, 4, and 5, it seems, are things we want in a theory about what reasons are and do. But what else is there? The conception of a practical reason that I’m trying to work up probably does not fulfill, probably, any of the items 1-5. This troubles me.