Thursday, October 07, 2004
Williams on reasons
In this post I offer an objection to Bernard Williams' internalist view of reasons. A useful summary of Williams' view (given in a response to certain criticisms offered by John McDowell) is the following:
The central idea is that if B can say truly of A that A has reason to F, then (leaving aside the qualifications needed because it may not be his strongest reason) there must be a sound deliberative route to F-ing which starts from A's existing motivations. It follows that what an agent has a reason to do will be a function of what I called his 'S' – that is to say, the existing set of his motivational states…I want to focus on a very small part of Williams' position, the notion of a sound deliberative route. In what follows I'll attempt to show that this notion can't do the work demanded of it.
Williams contrasts his account with externalism, a view which he says would hold that for an agent to have a reason to F, "F-ing will have to be an action that an agent could rationally decide to do as a result of deliberation whatever his S might be." What externalism is supposed to amount to is the view that there is some kind of independent Reason with a capital R to which an agent's deliberations must conform in order for the agent to be counted as rational. One might put it another way, saying that externalism holds that there exists an agent independent rational order which exemplifies the standards of rationality against which the deliberations of particular agents may be judged.
It is sometimes thought that one strike against externalism is that it seems, as Mackie might say, metaphysically queer. What sort of thing is this agent independent rational order, and how is it that beings like us are supposed to have access to it? Such questions have force but don't settle the issue. This is because the idea of an agent independent rational order is compelling in other realms, particularly the mathematical realm, even though such a realm is no less queer. The point here is that the mere fact that externalism posits a set of abstract objects – i.e. Reasons - doesn't indicate a metaphysical indulgence beyond those we are already inclined to tolerate.
In any case, Williams' objection to the externalist picture is different. Postulating an independent rational order, Williams argues, does no work in explaining how it is that reasons motivate us. Such an explanation can only be established, he says, by tying reasons to the agent's subjective motivational set. Since an independent rational order is, by definition, independent of the agent's subjective motivational set, it follows by simple argument that it cannot play the right explanatory role.
Let us grant all of this, and also the further premise that agents do not have the kind of intuition into the independent rational order which would make it possible to expand the agent's subjective motivational set through an act of pure rational deliberation.
Consider Melman, who must decide whether he will have fried chicken for dinner. Melman is not already disposed one way or the other, and doesn't see himself as having a reason to be disposed. So he must deliberate. We'll leave him to those deliberations for a moment while we take a look at his situation from Williams' point of view.
Either Melman has a reason or he does not. If he does, then this is because there is "a sound deliberative route to F-ing which starts from [Melman's] existing motivations." Let us assume, for convenience, that our philosophical omniscience extends widely enough to give us a view of Melman's subjective motivational set. Then we should be able to say whether there is a sound deliberative route from that set to eating fried chicken, and, hence, whether Melman has reason to have fried chicken for dinner. And let us say that he does, by our lights, have such a reason.
Now let us return to Melman. He has finished his deliberations but has concluded, wrongly as we see it, that he has no reason to eat fried chicken. And let us suppose that since we are philosophers and long to be midwives to wisdom that we point this out to him. We show him the sound deliberative route from his subjective motivational set to the eating of fried chicken and offer to drive him to the nearest KFC.
What shall we do if Melman examines – with full understanding - our account of his reasons and yet demurs?
I assume that the problem here is clear. Once we have demonstrated the sound deliberative route to Melman and he has rejected it we can no longer have a grip on the question of whether he does or does not have a reason. The important thing to see here is that Melman's rejection of our explanation means that our standards for sound deliberation are external standards with regard to him. As such, if we say that he does, in fact, have a reason then we alienate his reasons from his motivations in precisely the way that Williams says would deny explanatory power to those reasons. If, on the other hand, we say that Melman doesn't have a reason then his actions will no longer be rationally intelligible to us. In admitting that Melman doesn't have a reason we are saying, in effect, that it is his prerogative to set his own standards for what counts as a sound deliberative route for him. And once this is admitted we can never be in position to offer any rational criticism of him at all.
Let me touch on two responses that don't work. First, it might be said that the assumption that an agent either does or does not have a reason is too strong. This is actually one of a family of responses, each of which looks to claim that the Melman case is peculiar. None of these responses is adequate because all that the objection needs to go through is the possibility of a Melman case along the lines of the one I have suggested. Second, it might be said that Melman couldn't have rejected our account of his reasons without making a mistake by his own lights. The thought here is that Melman's standards for sound deliberation are part of his subjective motivational set, meaning that omniscient philosophers like us would have applied his own standards for sound deliberation in determining whether he had a reason or not. The problem with this sort of reply is that it leads to the same worries about rational intelligibility that were considered above.