Hawthorne and others take knowledge to be central to practical reasoning, in the sense that if one employs premises that one does not know to be true, one has done something wrong. I wonder the extent to which this claim is dependent on refusing to think that knowledge is explicable in terms of some of the standard approaches to the nature of knowledge, such as undefeated justified true belief. One way to approach this question is to ask what sense of ‘wrong’ might be involved in various failures of these conditions.
Suppose we distinguish senses of ‘wrong’ in terms of things that are wrong because the person is justly criticizable or blameworthy from things that are wrong in some other way. Consider, then, a case where you reason through premises that you don’t believe. Paradigm cases of this occur when the subject is depressed or experiencing emotional interference at the opposite end of the emotional spectrum. In such cases, the subject is justly criticized for employing those premises: consider a guy at the craps table who, giddy because of a winning streak, lets hope outrun conviction and bets the farm on the next roll. “You’re pathetic” comes to mind…
So, too, if the person isn’t letting hope outrun conviction, but lets conviction outrun evidence (as any normal human would have to in order to be convinced that they were going to win on the next toss). We shake our heads in dismay as we watch the folly unfold, and our actions constitute justified criticism of the reasoning process leading to the action.
Compare these reactions to those when the premises turn out to be false, but where the belief in question is epistemically justified. Part of being justified in the sense required for knowledge is that one’s evidence confirms for one further inquiry would turn up only misleading evidence against the belief in question, so the cases we are thinking about are not cases involving reasoning through lottery propositions, such as the claim that my ticket is a loser. In such cases, there is, perhaps, some sense in which things go wrong when one reasons through false premises. Here’s one such sense: such reasoning will not lead to securing the goals of practical reasoning as often or as well as reasoning through true premises, or so it’s reasonable for us to think.
In that sense, such reasoning is faulty. In the more important sense, however, there is nothing wrong with such reasoning–no one should criticize you for it and you are not blameworthy for having reasoned in this way. You might rue your plight at having been duped, but your reaction is much different than what you’d experience when you risk the farm at the craps table. Self-castigation results in the latter case when the passion of the moment has dissipated because you knew better than to reason and behave that way.
When one takes knowledge to be a norm of practical reasoning, one glosses over this distinction, thereby committing oneself to the view that the distinction between the two kinds of cases is not especially instructive or important in the context of practical reasoning. But that’s a mistake: one kind of wrong you’re responsible for, and the other kind you’re not. Sounds too important to ignore to me.
This result suggests that the important way in which practical reasoning can go wrong is by reasoning through premises that one is not epistemically justified in believing. This result comports well with similar results for careful assessments on the issues of the norms of assertion and belief (though it may be that the norms of assertion beyond sincerity are inherited from the norms of belief, as Kent Bach argues).
It is interesting to note that those who are most attracted to the idea that knowledge is the appropriate constraint on assertion, belief, and practical reasoning also think of knowledge as unanalyzable. Of course, nothing I’ve said here implies that knowledge is analyzable, if one has in mind some sort of semantic constraint on the account. I think, however, that those of this persuasion think of knowledge as not being explicable in terms of parts at all. The result would be, I think, that they view the above analysis as utterly foreign to their way of thinking, so foreign that its status as a competitor is hard to appreciate. It is worth noting a point that Lycan has made recently, however. Even if knowledge is not a composite thing constituted by its parts, it still has necessary conditions and that is the only point needed to sustain the above approach to norms of practical reasoning.