Kant’s distinction in Groundwork II between perfect and imperfect duties rests on a distinction between two types of contradiction (Kant 1997, 33/4:424). These have been called “contradiction in conception” and “contradiction in willing.” Controversy remains over how best to understand these contradictions.

Logical and practical contradictions. A logical contradiction refers to an impossible, inconceivable object or circumstance, like a married bachelor. In the Groundwork, Kant can be read as suggesting that maxims of action fail the test of the categorical imperative because their universalization would imply a logical contradiction (see Galvin, 1991). A maxim of owning slaves cannot be universal because this would result in the logical contradiction of every slave’s owning slaves. But not every case of a maxims’ failing the universalization test rests on a contradiction like this. In the example of the lying promise, the expected result of its universalization is that no one will lend money to people in need, which itself is no contradiction. A borrower would nevertheless contradict herself by willing both: (a) that I borrow money on a false promise when I need it; and (b) that everyone borrow money on a false promise when they need it. In willing both (a) and (b) she would will the contradiction that she borrows money when it is impossible for her to do so. But in the view of some commentators, a contradiction like this is better described not as logical but as practical (Korsgaard 1996, 92-101). It is practical because in this example one wills something that makes it impossible to achieve the end of her action. She ends up defeating her own purpose by willing both (a) and (b).

Contradictions in willing. According to Kant, some maxims are impossible to will as universal laws, even though the result of their universalization is conceivable. The clearest example of this is the maxim never to assist others in need. There could be a world in which this maxim functions as a law, Kant says; for a world can exist where no one ever helps anyone else. But it is impossible to will to be a part of it—because then one would will not to have any needed assistance; and it is irrational to will not to have what one needs. So the maxim passes the test of the conceptual possibility of its universalization, but it fails the test of its volitional possibility. Maxims that fail the universalization test in this way are thought to generate imperfect duties. (It is noteworthy that the explanation of the contradiction in willing just given seems to coincide exactly with the explanation given above for the contradiction in the case of the lying promise; although the former example is supposed to present a contradiction in conception, while this one is said to present a contradiction in willing. But see below.)

A unified account of the contradictions. It would be best if the two forms of contradiction that can result from maxims’ failing the universalization test—contradictions in conception and in willing—had something in common; something in virtue of which they are both forms of the same thing: contradiction. The practical interpretation of the contradictions mentioned above offers the advantage of this unification. The logical interpretation does not; or at least so far those who favor reading the contradictions in failing maxims as logical have not been able to explain contradictions in willing. The practical interpretation does better by saying that every maxim failing the universalization test involves some form of the same thing: self-defeatingness. It says that the difference between the two types of contradictions Kant points to is this: in the so-called contradiction-in-willing cases the agent’s end is an end we are supposed to have necessarily, as rational agents; in the contradiction-in-conception cases, the end the agent thwarts by his own action is contingent, or arbitrary. The agent in the example of assisting the needy thwarts his own, necessary end of well-being or happiness in universalizing his maxim. The agent in the lying-promise example thwarts his own, contingent end of getting the money required to get out of his present financial bind. (This therefore addresses the problem raised at the end of the previous section.) The practical-contradiction interpretation also allows that the maxim failing the universalization test may involve a logical contradiction (Korsgaard 1996, 97). But in such cases, that means that the agent’s willing her maxim as a universal law will require her doing something logically contradictory in order to achieve her end, which leads to self-defeat. The slave-owning maxim mentioned above is a good example of this. If the way you choose to accomplish your end requires that every slave own slaves, you will not succeed.

The problem of natural actions. It turns out that both the logical and practical interpretations of the contradictions have a fatal flaw. They are incapable of detecting a contradiction in the universalized maxims of actions, or of ends, in wholly natural situations. These are typically cases involving homicide or bodily harm. Kant’s suicide example presents such a case, and here neither the logical nor practical interpretations can find a contradiction in the maxim’s universalization. Absent from the examples of wholly natural actions that are supposed to fail the universalization test is some kind of convention, practice, or background agreement. Something like this is always required for a logical contradiction’s arising from a universalized maxim, or from a maxim whose universalization results in self-defeat. In a maxim of stealing, for example, the background convention of private property is required in order to get a contradiction from its universalization: if stealing were universal, there would be no convention of property, and then it would not be possible to steal. If everyone lied, there would be no background expectation of learning anything from what anyone says, so no one would listen to anyone, and consequently there would be no speech, and no language, and so no lying. Korsgaard, who supports the practical interpretation, claims to find a contradiction in the maxim for the natural action of killing someone in order to get his job (Korsgaard 1996, 98). Killing is a natural action; but by introducing the idea of employment she unwittingly converts the example to a conventional action. Universalization of the maxim of killing anyone who has some natural object I want leads to neither a logical nor a practical contradiction, though it does create a world I would not want to live in, and so generates a contradiction in willing. But that is not the kind of contradiction we should expect from a maxim of convenience killing. It could generate nothing stronger than an imperfect duty not to kill (see Herman 1993, 116-18).

Teleology and the system of nature. An alternative to the logical and practical interpretations of the contradictions is known as the “teleological interpretation” (see Paton 1948, 149-57). Though described in different ways, the basic idea is that nature can be conceived as a system of laws describing means-end relations, and therefore maxims, as means-ends principles, can be tested by the coherence of their universal forms with the system of nature (see Stapleford 2007). Maxims like those in the suicide example, or in the example of convenience killing, would result in an incoherent system of nature if they were elevated to the status of universal laws, since they both prioritize human happiness over human life. As Kant attempted to establish in an earlier argument regarding the end of practical reason, nature would not have made us rational if happiness were the ultimate end of our existence. So this implies that maxims of action destructive of human life or bodily functioning, for purposes of happiness, are inconsistent with the background system of nature they require in order to be effective. In a sense, then, the teleological interpretation is able to rely on the idea of systematic nature in the same way the logical and practical interpretations rely on background conventions. But the teleological interpretation can also recognize social conventions as background conditions for probably most human actions; for these can be imagined to emerge in society in response to natural needs. Since it is natural that human beings can articulate and recognize patterns of sounds, and natural that they can advance their interests through communication, the convention of honest, truthful conversation emerges quite naturally. Since it is more convenient for human life that tools and other implements are ready to hand, the convention of property emerges naturally as well. Maxims that would thwart these conventions when universalized can therefore be recognized as introducing forms of contradiction in the system of nature: presumably, both contradictions in conception and contradictions in willing. At the present time, however, considerably more interpretive work is required in order to shore up the credibility of the promising teleological interpretation.