Kant’s first example of how to apply the principle of duty derived from the preceding analysis is that of a person hoping to escape financial difficulty by borrowing money on the pretense that he will repay it (Kant 1997, 15/4:402-03). Of all his examples, this one seems to be the most persuasive; but, of course, it has not escaped
How it works. The principle to be applied is: I ought never to act in such a way that I could not also will that my maxim should become a universal law. The maxim, in this example, can be stated as follows: “When in need of money I’ll lie to a lender in order to get what I need.” The universal-law form of this maxim would thus be that “Everyone in need of money lies to lenders in order to get what they need.” There is obviously something incoherent about this maxim’s universal form. There would be no money lenders if no one ever intended to repay money lent to them—not even the lenders themselves. For this reason I would exhibit a similar incoherence if I willed my maxim to become a universal law. I would will a condition where there are no money lenders, which would make it impossible for me to get the money I need. The idea, then, is that as a rational being I cannot will this incoherence, which is why the lying-promise maxim fails what has come to be called the “universalization test.”
Doctoring the maxim. If that maxim fails the universalization test, another maxim for the same action might not fail, and so a lying promise might be shown to be permissible after all. One way to concoct a lying-promise maxim that (at least seems to) pass the universalization test is to specify the circumstances of action so narrowly that the lender’s not being repaid will not have the expected overall effect on the practice of money-lending. An example of a maxim like the following is sometimes offered to illustrate:
I’ll make a lying promise on Tuesday, August 21 to a person named Hildreth Milton Flitcraft in order to get some money I need
(Wood 1999, 102).
It has been assumed that this maxim’s universalization would not affect the practice of borrowing and lending overall; so it should still be possible get the needed money with a lying promise. The assumption has been that the maxim could hold as a universal law, and so long as Flitcraft doesn’t know about this, it would be possible to act on the maxim. And really, how could he know about such a qualified universal law? On the other hand, it does seem impossible to hoodwink poor Flitcraft in this way if we consider that the maxim’s holding as universal law means that it would have to be a maxim by which everyone acts, not just a maxim for me. The point is that my maxim’s universalization should imply that Flitcraft will be just as wary of lending his money on the date in question as I am keen on trying to “borrow” it. Remember the point about the lying-promise maxim made above. When that maxim holds as a universal law, even the money lenders know that, if they were in my situation, they too would try to get by with a lying promise. In effect, then, a maxim’s holding as a universal law implies that it is everyone’s maxim. That is why even the Flitcraft maxim leads to contradiction when universalized. So doctoring the maxim with specific details of action will not help it pass the test; because its holding universally implies everyone’s knowing those details also!
False Positives. In the extensive literature on the universalization test for maxims, examples like the one preceding are called “false positives.” Their maxims purport to pass the test (return a positive result) when their actions are obviously wrong, and should not pass. In general, since there are many different kinds of maxims, there are multiple forms of false positives. If one way of doctoring the Flitcraft maxim actually fails to generate a false positive, as just seen, another way of doing so might succeed. The literature includes purported examples of false positives that specify unique details of the agent, instead of, or in addition to, details of the circumstances. For example:
If I am red-haired and named Ignatz MacGillicuddy, I’ll rob any bank northeast of my house open at 5 p.m. on any Thursday (Guyer 2007, 139).
Supposing my name is Ignatz MacGillicuddy, and that I have red hair, it might seem that I could permissibly rob banks on that maxim. There could only be a few people like this, and so the banking system will not collapse even if all of us act on the maxim. But here’s the problem with examples like this. A maxim should not be formulated to contain identifying information about the agent. There is really no point in doing so, as can be seen from considering the nature or function of maxims. Since one always has one’s identity, one would always act on any maxim that specified one’s identity as one of its conditions; and this does not make sense. The point of specifying a condition for action in a maxim is to indicate when the person would act as indicated in the maxim and when she would not. If the condition could not ever be false, or unfulfilled, then there could never be a time when one would not act on the maxim. Specifying one’s identity as a condition in a maxim is like specifying a tautology as a condition; for example:
If a=a I’ll rob any bank northeast of my house.
The maxim is the same whether the tautology is included as the condition of action or not. So for the same reason it is pointless to include identity conditions of the agent as conditions of action for any maxim. That is why examples like the MacGilluddy maxim can be dismissed: because they are not genuine maxims they are not reliable for testing the moral permissibility of actions. Just to drive the point home: the maxim’s being Ignatz MacGilluddy’s maxim is the same as its calling for him to act under the conditions it specifies. So there is no reason to include being Ignatz MacGilluddy in the maxim as one of the action’s conditions.
False Negatives. It also seems possible for a maxim to fail the universalization test (return a negative result) when it should not, because the action is morally permissible. A variety of familiar examples of maxims like this are found in the literature. One particularly bothersome type of example involves a compound course of action where doing one of a pair of actions depends on someone else’s doing the other. If the maxim includes my refusing to do the other action, then when this refusal is universalized I will be unable to do the first action. For instance:
I’ll buy clockwork trains and never sell them (Nell 1975, 76-77).
Everyone else’s refusing to sell them means that I can never buy them. So the universalized form of the maxim is incoherent, and thus, apparently, the compound action is shown to be wrong. Yet what could be wrong with buying clockwork trains without ever selling them? Nothing, of course. Here’s the solution to this puzzle. We must test the maxim for each component-action separately; then, if each of these passes, the maxim for the compound action will pass. If everyone can buy clockwork trains, and if everyone can omit selling them, then the maxim above passes the test, and its compound course of action is shown to be morally permissible—as we know it is. The reason this testing procedure works, and is not merely ad hoc, is that permissibility is directly analogous to possibility. In fact, Kant often referred to morally permissible actions as morally possible. For any set of conditions, each of which is possible, their conjunction is possible. So also, for any set of morally permissible actions, their conjunction is permissible. There are important exceptions to these rules, however. Two or more possible conditions are not jointly possible if they are contradictory. It is possible that the person who wrote “Happy Birthday” is alive, and it is possible that he or she is dead. But it is not possible that that person is both alive and dead. This type of exception in the case of possibility, when transferred to permissibility, will mean that sometimes two morally permissible actions should not be jointly permissible. The reason why will depend upon yet another action. Suppose you have done this: you have required that if I buy trains from you, I must also sell them. My buying trains from you is permissible. My never selling trains is permissible. But the conjunction of my buying trains from you and never selling trains is not permissible—because in order to buy trains from you I have to agree to sell them. This is an example of the permissibilty-analog to the exception to the rule for joint possibility. A person cannot be both alive and dead even though each condition is possible—but that is because of the relation between life and death. Likewise, I may not be permitted to buy clockwork trains and not sell them even though each of these actions is permissible—but that would be because of the relation between these actions you would introduce by refusing to sell trains to me unless I would also sell them. I would then be wrong to buy trains from you and never sell them, because I would break my promise. Or I would be wrong to buy trains from you without intending to sell them, because this would be a lying
promise—and we know why that would be wrong.