The Categorical Imperative

Kant begins Chapter 2 with a framework for thinking about practical reasoning. Every rational action, he says, is guided by some principle. When you decide to go to the store, you are acting on a principle like "if I want groceries, I should go to the store." Even when we do not articulate these principles explicitly, they are implicit in what we do.

For Kant, an imperative is a principle that tells a rational being what they ought to do. Imperatives come in two fundamentally different kinds.

A hypothetical imperative tells you what to do if you want something. It has the form "If you want X, then you should do Y." Examples: "If you want to pass this course, you should do the reading." "If you want to be healthy, you should exercise." "If you want to be a good carpenter, you should learn to measure carefully." Hypothetical imperatives are conditional: their force depends on the fact that you actually want X. If you do not want X, the imperative does not apply to you. Someone who does not care about being healthy is not irrational to skip exercise.

A categorical imperative tells you what to do unconditionally, regardless of what you happen to want. It has the form "You should do Y." No "if," no "when," no conditions. Examples (proposed): "You should not lie." "You should not use other people merely as tools." "You should keep your promises." Categorical imperatives, if there are any, bind you regardless of your preferences or goals.

Kant's claim is that genuine moral rules are categorical, not hypothetical. If morality told us only what to do to achieve certain ends, then anyone who did not share those ends would be exempt. But we do not think a person who lacks moral feelings is thereby exempt from moral duties. We think lying is wrong for everyone, full stop. That implies morality is categorical.

This sets up Kant's central question: what could a categorical imperative look like? If an imperative is supposed to hold unconditionally, without appeal to any particular desire or goal, what content could it possibly have? How could reason alone generate such a rule?

Kant's answer is ingenious. A categorical imperative cannot specify particular content (since that would make it hypothetical, dependent on wanting that content), so the only thing it can require is a certain form. It must require that our principles have the form of universal law.

He gives us the most famous formulation:

Unpack this carefully. A maxim is the principle on which you are actually acting, stated in general terms. If you are thinking about borrowing money you cannot repay, your maxim might be "when I need money and cannot get it otherwise, I will borrow it with a false promise to repay." Your maxim is the general rule that describes what you are doing.

The categorical imperative asks you: could you will that everyone act on that maxim? That is, could you consistently will a world in which the maxim held as a universal law?

Think about the lying-promise case. Your maxim: "when I need money, I will make a false promise to get it." Imagine this as a universal law: every person who needs money makes a false promise to get it. Now notice what happens. If this were a universal law, promises would lose their meaning entirely. Everyone would know that anyone asking for a loan is lying, and no one would accept any promises. The institution of promising would collapse. But without the institution of promising, you could not make a false promise in the first place. The maxim, universalized, destroys the very conditions of its own operation. It is self-defeating.

Kant claims this is not just a practical problem. It is a logical problem: the universalized maxim is actually contradictory. You cannot rationally will a world in which the institution of promising exists (so you can make a false promise) and simultaneously will a world in which the institution of promising has collapsed (because everyone makes false promises). The two conditions are incompatible. So this maxim fails the test of universalization, and the action is morally forbidden.

Kant gives the categorical imperative several formulations in Chapter 2. He claims they are all equivalent in content, just different ways of looking at the same principle. Whether they really are equivalent is disputed, but most readers find the second formulation much more intuitive than the first, and much easier to apply to concrete cases.

This is the Humanity formulation. It claims that every rational being has a special kind of worth, which Kant calls dignity, and that this worth generates moral constraints on how we can treat each other.

The core distinction is between treating someone as a means and treating them merely as a means. We treat people as means all the time: when you order coffee, the barista is a means to your getting coffee; when you hire a plumber, the plumber is a means to getting your sink fixed. This is not morally problematic. What is forbidden, Kant says, is treating someone merely as a means, as if their worth consisted entirely in their usefulness to you.

Consider the lying-promise case again, now through this second formulation. When I make a false promise to get money, I am using the other person as a tool in my scheme. I am treating them as a means to my financial goals. But I am also doing something worse: I am deceiving them into serving a goal they would not consent to serve. If they knew what I was actually doing, they would refuse. So the question Kant puts to us is: could the other person, fully informed, consent to be treated this way? If the answer is no, you are treating them merely as a means.

Notice what this formulation accomplishes. It gives us a vivid way to think about moral wrongs, one that matches many people's intuitions well. Slavery treats slaves merely as means. Exploitation treats workers merely as means. Manipulation treats the manipulated person merely as a means. Even using someone's labor or body in a way they have not consented to, however small, runs afoul of this principle.

The Humanity formulation also underwrites much of modern thinking about human rights. Each person has a worth that cannot be reduced to their utility to others, and that worth demands to be respected. This is a profoundly Kantian idea, and it has proven more durable in moral and political thought than the details of Kant's system.

Kant applies his formulations to four cases, meant to illustrate how the categorical imperative generates a range of duties. You will find these in the middle of Chapter 2.

The matrix is not accidental. Kant wants to show that the categorical imperative covers all four combinations: duties to ourselves and to others, and two kinds of duty (perfect and imperfect).

A perfect duty is one we must always observe, in every relevant situation. Do not make false promises. Do not commit suicide (on Kant's view). These are strict requirements.

An imperfect duty is one we must observe in general, but with some latitude about how and when. We have an imperfect duty to help others: we cannot simply ignore human need, but we have discretion about whom to help, how much, and on what occasions. We have an imperfect duty to develop our talents: we cannot let all of our capacities atrophy, but we have discretion about which to cultivate.

The cases illustrate how the universalization test works. A maxim of false promising generates a contradiction in conception (promising collapses). A maxim of refusing all help generates a contradiction in the will (we could not consistently will such a world, given what we ourselves need). And they illustrate how the humanity formulation works. The suicide case treats oneself merely as a means to escape pain. The false promise treats the other person merely as a means. The talents case and the help case both fail to treat the humanity in oneself or others as an end.

Kant's examples have been debated for over two centuries. Critics have pointed out that careful restatement of the maxims can often pass the universalization test even for clearly immoral actions. Defenders have responded that correct application requires identifying the right level of generality, which is harder than it looks. For our purposes, the cases are most useful as illustrations of how Kant thinks the test is supposed to work in practice, regardless of whether every case turns out as he says.

Watch this lecture after you have worked through Chapter 2 and this page. It will help reinforce the distinctions between the formulations and the structure of Kant's examples.