I really like this paper of John Hyman's. Go read it.
Welcome back. After hemming and hawing for a while, I've come around to the idea that you can't act for the reason that p unless you know p. Previously, I had argued that various sorts of Gettier cases caused trouble for the idea. I know think that I'm wrong. I don't want to dwell on that.
Hyman offers an account of knowledge according to which it's a species of ability, not belief. Knowledge is the ability to do things, or refrain from doing things, or believe, or want, or doubt things, for reasons that are facts (441). While this strikes me as entirely correct, a weaker claim would do for my purposes. Let's suppose knowing p entails having the ability to X for the reason that p.
In previous work, I've grappled with some of Fantl and McGrath's suggestions about justification and reasons. Their view is that to justifiably believe p, one must have the right to treat p as if it's a reason. It wasn't always clear to my mind whether they thought that to justifiably believe p, p must be a reason that you can treat as such. In places they seemed to like this idea. In others, it wasn't clear. As far as I can tell, their current view is that motivating reasons can consist of falsehoods, so they might endorse both the authority idea and the ability idea:
Authority: To justifiably believe p, one must have justification to treat p as if it's a reason.
Ability: To justifiably believe p, one must have the ability to treat p as a reason for X-ing (for some appropriate X).
It's also clear, I think, that if they were convinced that to X for the reason that p, p has to be true, they'd drop Ability and retain Authority. Now, I think splitting these two up is a rather strange idea. If we want to understand the point of belief, surely the point of belief is to provide one with reasons that one can then reason from and treat as reasons. If any belief doesn't do that, it seems to violate the fundamental norm of belief. Norm violations can be excused, but they don't count as justified when there's no further norm that would require the belief.
That's a big picture sort of argument. I don't expect it to persuade anyone unless I add in lots and lots of detail that I won't repeat here. Instead, let me offer an alternative line of argument. On standard accounts of doxastic justification, doxastic justification is propositional justification plus proper basing. To justifiably believe p, one must have a justification to believe p and that has to be the reason for which one believes. With this much in place, we can easily establish this much: to justifiably believe p, there must be something known that serves as the basis for one's belief.
In some cases, it's clear that there's no further reason apart from the fact believed that's eligible as a basis for belief. If so, the distinction between justification and knowledge should collapse. What about the other cases? It will be interesting to see if the argument can generalise. (I think it can, but doing so will wait for later. It just repeats some arguments connecting the factivity of evidence to the factivity of justification discussed in the book.) Anyway, surely this is controversial enough. There are only a handful of people who think that to justifiably believe with non-inferential justification one must know.
I'd be curious to know whether there's any principled reason for pulling apart assessments of authority from assessments of ability. I think people think that justification is a normative concept, knowledge might be a concept that's tied to abilities, and assume that there's a kind of independence here. Maybe there's no good reason to think that. Maybe claims about abilities have some bearing on claims about normative authority.