Question Is Kant's Categorical imperative overly dependent on empirical considerations? I think it is since judging the morality of an action by asking what would happen if everybody did the same thing means that the morality of an action is dependent on the contingent features of the world that produce that effect. If everyone did a certain thing then there would be chaos so that is not good Kant seems to say. Well that chaos of course depends less on the nature of the action and it underlying intentions than on the world that action took place in. If everyone stole then society would fall apart but that seems to have more to do with principles of sociology than something that pertains to ethics.

Accepted:
February 2, 2012

Comments

You suggest that Kant's criterion of wrong conduct turns on this question: "If everyone acted the way I am proposing to act, would this have undesirable consequences?"

I think Kant's actual question differs in two respects. Kant is not asking whether the agent would like some fictional world (find it desirable), but whether the agent can will it and her own proposed conduct in it. And the world Kant envisioned is not one in which all act the way the agent is proposing to act, but one in which all are permitted (and take themselves to be permitted) so to act. So Kant's question is: "Can I will the action I am considering along with its universal permission?" The basic idea here is that I should not permit myself an action that I cannot permit all others at the same time.

Let's see how this plays out in Kant's promising example. The agent considers extricating himself from financial difficulty by making a false (lying) promise. He then asks himself whether, in a world in which all took themselves to be permitted to make such promises, he could still will to act in this way. Kant's answer is no: in that fictional world, such promises would not be believed and therefore refused; and agents could thus not will to offer them because they would be useless for their intended purpose.

Now your objection survives this clarification. Suppose the world were such that some nice fairy fulfilled any promises that the promisor fails to fulfill. In that world, it would seem, making false promises would be permissible. For in that world, even if all took themselves to be permitted to make false promises, such promises would (not be believed but) still be accepted. In that world, then, the agent can will his proposed action alongside its universal permission. So it would seem that, as you say, the permissibility of making a lying promise turns on a contingent empirical fact, namely on whether there is some third party ready to step in to ensure that even lying promises are fulfilled.

I am sure Kant and orthodox Kantians would not want morality to be like this. But ask yourself in conclusion whether such responsiveness to basic empirical facts about the world isn't actually an advantage in morality. Would it really be wrong falsely to promise repayment if such false promises were to hurt no one? And is it really implausible to hold (to give another example) that the question whether one is duty-bound to procreate depends on whether enough children would be born even without such a duty?