Kant's categorical imperative: can morality rest on a single principle?

One of the most quoted sentences in moral philosophy belongs to Kant: "Act only according to that maxim whereby you can at the same time will that it should become a universal law."

Why "categorical"?#

Kant distinguishes two kinds of imperative. A hypothetical imperative depends on an end: "If you want to be healthy, sleep regularly." Drop the end and the imperative falls with it. A categorical imperative depends on no condition at all; its moral force is independent of any result you want.

The three formulations#

1. The formula of universal law — Can you make the principle of your action into a law that everyone follows? Universalise lying promises and the institution of promising collapses; the principle cancels itself.
2. The formula of humanity — Never treat a person merely as a means, but always at the same time as an end.
3. The formula of the kingdom of ends — Think of yourself as a member of a community in which everyone is both lawgiver and subject to the law.

The classic objection#

The best-known criticism grows out of Kant's own example: must you tell the murderer at your door where the person hiding inside has gone? Kant's strict reading says one may not lie, and the consequentialist tradition offers that as the price of blind attachment to a principle.

The Kantian defence is that the categorical imperative is not a list of actions but a method of testing: the question to ask is what principle you are in fact acting on.

Its present relevance#

Kant's criterion has acquired an unexpected currency in the ethics of artificial intelligence: the argument over whether systems should "follow fixed principles or weigh consequences case by case" is the old line between deontology and utilitarianism translated into the language of engineering.