<body class="blog" onload="">

Ratiocination

A Little Paper


I'm in the midst of writing a paper on the Kantian Maxim that 'ought' implies 'can.' I got distracted by a number of side-issues related to this debate. Instead of dropping these projects altogether, my goal is to turn each of them into a little paper, perhaps suitable for submission as a discussion note or whatnot. Here's such a little paper.

A brief summary: Frances Howard-Snyder has recently come to the defense of one formulation of the maxim that 'ought' implies 'can.' In this short paper, I argue that the formulation Howard-Snyder favors is false. But there is a fix, once we countenance the addition of a "tracing" condition to our formulation of the maxim.

6 Comments:

Add your own


Anonymous at 7:30 PM  said... A typo: in your second paragraph, you state that 's' (lowercase) will range over subjects, but in the remainder of the paper you use 'S' (uppercase). Certainly not a meaningful error, but it looks sloppy.



Andrew Bailey at 7:46 PM  said... Fixed, thanks. =)



Anonymous at 4:49 AM  said... No problem.

What can I say, I'm pedantic.



Noumena at 6:47 AM  said... What do you mean when you write that Terry is `constitutionally incapable' of refusing a drink? I assume you're saying something about Terry's dispositions or habits -- she will always or almost always accept a drink when one is offered to her. And by using `constitutionally' to describe this disposition, you seem to be saying that it's very, very deeply ingrained. Perhaps even so deeply ingrained that it can never be removed?

Of course, if it can never be removed, six months or six years in AA won't make a difference: she'll still have the disposition, and still not refuse the drinks at the party, and still break the promise. Tracing Maxim doesn't seem to do any better than the other two maxims in this case.

So, both because we want Tracing to do something and because we want to be optimistic about alcoholics getting better, we say that this disposition is not so deeply ingrained that it can never be removed, and that if Terry had made the decision to enter AA however long ago, then she would be able to refuse the drinks at this party right now. But then her actual intemperance seems to fall back into the cases of self-imposed inability that Time Maxim handles quite nicely: however long ago, if Tracy had made this decision instead of that one, she would now be able to follow through on her promise. Tracing Maxim is superfluous here.

I think the problem is this: You want to be able to treat scenarios where agents (a) have an obligation, (b) are unable to follow through on it, and (c) are not responsible for their having that inability. But the point of ought->can is that, when (b) and (c), then not (a).

Alternatively, Tracing Maxim just does a more perspicacious job of articulating what's meant by `S can at t* do a at t'. Then we quibble about whether or not it's a different `formulation', which will not doubt devolve into questions about the relationship between sentences, statements, and propositions, and I start to need a couple of drinks myself.



Andrew Bailey at 7:08 AM  said... If the disposition to accept drinks is deeply ingrained (such that it can never be removed), then there is no action available to Terry such that had she performed it, she would have been able to refuse the drink. Tracing Maxim gives the right result (and indeed, so do the other two maxims); Terry gets excused of her duty.

But suppose the disposition isn't that deep. You say that Time Maxim can handle the case too, since "however long ago, if Tracy had made this decision instead of that one, she would now be able to follow through on her promise." But this doesn't account for the difference between being able to perform action *a* at some time in the past and any being able to perform some action *b* (b need not be the same action as the one we're supposing Terry has a duty to perform, ie, a). I gave an argument saying that Terry wasn't able to perform *a* at *any* time. And yet it is still true that she ought to perform a. Time Maxim can't deal with the cases where doing something *different* than the action a in question way back in the past would have made the agent able to perform a now.

You finally suggest that Tracing Maxim just does a more perspicacious job of articulating what's meant by `S can at t* do a at t'. I'm inclined to agree, and just want to make this all explicit. In natural language, ability-talk often seems to presuppose an ability that goes beyond a capacity or opportunity to perform an act at a time. We rather have in mind something Dan Speak calls "Deep Capability," which is a capacity or opportunity *or* a tracing condition.



Noumena at 5:22 PM  said... Look at this example, which you quote from Howard-Snyder: `Beatrice promises to meet Zoe at 6:00, and then drives to a beach in the opposite direction. At 5:45, Beatrice cannot reach the meeting place by 6:00.'

Here the action a is meeting Zoe at 6:00. Because the action is time-indexed, taken literally, it is impossible for Beatrice to perform action a at any other time. So, given the choices she has made, there is no time t* at which Beatrice could ever perform action a. This seems to be a case `where doing something different than the action a in question way back in the past would have made the agent able to perform a now'. Perhaps not really *way* back, but that doesn't seem critical.

If you're right, and Howard-Snyder can't give an account of Terry's situation, then she can't give an account of Beatrice's either. But, though I haven't read her paper, I suspect that she gives this example as a paradigm case of the sorts of situations Time Maxim was designed to solve. That makes me further suspect that Tracing Maxim and Time Maxim aren't actually radically different. Even if they are radically different in some other way (maybe Tracing makes a distinction between acts of commission and acts of omission, but Time doesn't?), then I would suspect that Howard-Snyder's analysis of Beatrice's case could be adapted to fit Terry's as well.