Tuesday, August 11, 2009

Mine shafted

I've been thinking some about one of the discussions of mine shaft cases from RoME. In particular, I've been trying to make sense of the "second grade of immoral involvement". The suggestion was that in a certain version of a mine shaft case where the subject had calculated that the best option was A, forgot whether it was A or B, remembered that one of these would be disastrous, the 'good agent' might just do C knowing that either A or B is the right thing to do. That seems fine, I suppose, but what distinguishes the "first" from "second" grade of immoral involvement is that the agent knowingly chooses an option that is the option she shouldn't choose and seems not to be morally blameworthy for so doing. Or so the suggestion went.

The question I had went something like this. While some of us are willing to swallow the idea that the method that a morally responsible agent can diverge from the principles that determine whether the agent does what she should do or acts rightly, there's got to be some way to describe the psychology of the agent who does C because she knows that she doesn't know whether it is A or B that she should do but knows that it is either A or B that she should do.

Intuitively, it seems that the agent who follows the method knowing that it will lead her to do what she oughtn't is nevertheless rational and morally responsible for doing C. But, and here's the puzzle, it seems that she can't judge rationally:

(1) I should do C.

She can't because she also knows.

(2) It is either A or B that I should do but I don't know which and one of these is quite bad.

Obviously, you can't rationally believe:

(3) (1) & (2).

What about the intention to do C? I take it that the morally responsible agent intends to do C. Does she intend to do so rationally? It seems so, but then it looks like this is her psychology:

(4) I intend to do C but I should do A or B.

If that 'should' is the should of full practical rationality, I don't see how (4) could be a rational attitude since among the things the agent would seem to think if (4) is her attitude is:

(5) I intend to do C but I shouldn't, I should do something else.

That's weird because that seems as irrational as (3).


Joanne said...

Hey there! I think the morally responsible agent is in a confused psychology :) and when one is confused, and the syllogism unclear, it's hard to choose rationally.

I enjoy going through your blog!

Mike Almeida said...

(1) I should do C.
(2) It is either A or B that I should do but I don't know which and one of these is quite bad.
Obviously, you can't rationally believe:(3) (1) & (2).

You can believe someting like that if (C) is a gamble over (1) and (2). As the case is described, it seems like a mistake to take the pure strategy of just choosing A or just choosing B. So, the 'should' in (2) cannot be recommending a pure strategy. You could be doing something terribly bad, and in any case (even if you choose right) it is wrong to take such a risk. But certainly there is some distibution D of risk over A & B such that D = the value of C.

Clayton said...


I think that's probably right! The problem is that I think (but I'm getting bad at guessing what people's attitudes are) that others think that he's not confused at all. And that's confusing to me.

Hey Mike,

As the case was described (but not by me, sorry about that), C was a choice to do neither A nor B but to pursue some option that is not as bad as the worst or as good as the best. That's among the reasons that the case strikes me as weird.

I guess the way to describe it is as follows. Either A or B is a gin and tonic. That's a favorite, but one is laced with arsenic and you forgot which. C, however, is a vodka tonic so you go for it. Kagan wanted to say that in such a case you know that you shouldn't do C because you know that you should do A or B. But, when I asked, he seemed to concede that the intention to C is perfectly rational. There's a sense in which I guess that's right, but I can imagine someone arguing as follows:
You can't rationally judge that you sholdn't C and intend to C. You can't but rationally judge that you shouldn't C when you know that you ought to A or B. So, _really_ what you ought to do is go for C.

That's the conclusion he wanted to reject.

Alright, I'm off to Big Sur!