Wednesday, June 9, 2010

Warrant and action

Had a really fantastic time in Edinburgh and wanted to write up a thought on an exchange I had concerning the warrant action principle.

The dialectic is a bit complicated, but Hawthorne and Stanley defend the idea that you shouldn't treat p as a reason for action and they think we can reasonably and rightly say things like this:

(1) You shouldn't have acted on the assumption that p since you didn't know that p.

Gerken suggested (plausibly, I think) that it's not reasonable to complain that someone acted on something they didn't know if the reason they didn't know is that their belief was Gettiered. He seemed to want to say something stronger, which is that really the only time it is proper to say that someone shouldn't have acted on an assumption is when they weren't warranted in believing that it's true. I agree, but we disagree on what it takes to believe with (sufficient) warrant or justification. Me: to believe with warrant/sufficient justification just is to believe in such a way that it's not improper to put that assumption into deliberation. Him: warrant/justification is something in the neighborhood of what orthodox views say.

Here's the problem case. I think on his warrant principle, this never comes out right:

(2) Although you were warranted in believing p, you shouldn't have acted on the assumption that p since you didn't know p.

The reason this won't come out right is that warrant for believing p rules out that you shouldn't have acted on the assumption.

Here's the worry: on orthodox views of warrant, you can have warranted false beliefs. Suppose that 'p' stands for 'I ought to push the green button' and suppose that that's false because you oughtn't push the green button.

(2g) Although you were warranted in believing that you should have pushed the green button and it was okay to push the green button, you shouldn't have acted on the assumption that you should push that button because you shouldn't have pushed that button.

If warranted belief suffices for permissibly acting on an assumption, if (2g) is true, I think this is true as well:

(3) Although you were warranted in believing that you should have pushed the green button, it's okay to push the green button on the assumption that you should but not okay to push the green button.

If, as seems plausible, 'You shouldn't A' entails 'You shouldn't A on any assumptions at all', (3) can't be right. So, I think you have to deny that there can be false, warranted beliefs about what you ought to do if you want to understand the warrant principle in the way that I think Gerken does. Now, he suggests that this might be because there's a special kind of 'outcome' ought. Maybe, but the case doesn't say anything about outcomes. The mistake might have nothing to do with hidden consequences, it might have to do with a mistake about the comparative weight of reasons where those reasons are all accessible to you.

It's worth noting that if you sever the connection between warrant and action to avoid the apparent counterexample, we can ask questions like this: Given that you were warranted in thinking that you should have pushed the green button, what were you supposed to do? If you say 'Something else', then the mark of the warranted is not the reasonable. If you say 'Push the green button', you aren't avoiding the counterexample. The problem with the first option is the tendency to think of the reasonable as the mark of the warranted. If reasonable just entails warranted, the objection stands.

No comments: