The categorical imperative is, in Immanuel Kant's moral philosophy, the supreme principle of morality: an unconditional command of reason binding on every rational agent regardless of their desires, ends, or circumstances.
Immanuel Kant (1724 to 1804) built a moral theory that locates the worth of an action not in what it brings about but in the principle from which it is done. Where a consequentialist asks what result an action produces, and a virtue theorist asks what a good character would do, Kant asks whether the action is required by duty. His central claim is that morality rests on a single principle of practical reason, the categorical imperative, which each rational agent can discover and is bound to obey. This account, worked out chiefly in the Groundwork of the Metaphysics of Morals and the Critique of Practical Reason, is the founding statement of deontological ethics.[1]
The theory is demanding because it removes the usual props of moral motivation. Sympathy, self-interest, and the wish to be happy may all accompany a right action, but on Kant's view none of them can make it moral. Only reason, legislating to itself, can supply a law that holds without exception. NoteKant does not deny that inclinations exist or matter; he denies that they can be the ground of moral obligation.
The good will and duty
The Groundwork opens with a striking claim: nothing can be conceived as good without qualification except a good will. Intelligence, courage, wealth, and even happiness can be put to bad use, but a will that wills rightly is good in itself. What makes a will good is that it acts from duty, meaning it acts because a course of action is right, not because the agent happens to want the outcome.[2]
Kant draws a careful distinction between acting from duty and acting merely in accordance with duty. A shopkeeper who deals honestly because honesty is good for business acts in accordance with duty, but the moral worth of the act depends on whether the shopkeeper would deal honestly even when dishonesty would pay. Only the person who tells the truth because it is required, whatever the cost, acts from duty and displays genuine moral worth.
Categorical and hypothetical imperatives
Reason issues commands that Kant calls imperatives. A hypothetical imperative is conditional: it tells you what to do if you want a certain end. If you want to pass the exam, study. Its force depends entirely on your having adopted the goal; abandon the goal and the command lapses. A categorical imperative, by contrast, commands unconditionally. It does not say do this in order to gain something; it says do this, full stop, because you are a rational agent. Moral requirements, Kant argues, are categorical in exactly this sense, which is why they cannot be reduced to prudence or desire.
Act only according to that maxim whereby you can at the same time will that it should become a universal law. after Kant, Groundwork of the Metaphysics of Morals
The formulas
Kant holds that there is one categorical imperative, but he states it in several formulas that he regards as equivalent expressions of the same law. A maxim is the subjective principle on which a person actually acts, the personal rule behind a choice. The formulas are tests a maxim must pass.
Formula of universal law
The first formula asks whether the maxim of your action could be willed as a universal law that everyone follows. Kant applies two tests. Some maxims cannot even be conceived as universal without contradiction: a policy of making false promises destroys the practice of promising it relies on, a contradiction in conception. Other maxims can be conceived but cannot be rationally willed, such as a refusal to help anyone in need, since a rational agent cannot consistently will a world in which aid is never available to them, a contradiction in will. These tests generate a distinction between perfect duties, which admit no exception, and imperfect duties, which leave latitude in how they are met.[3]
Formula of humanity
The second formula requires that you treat humanity, whether in your own person or in that of another, always as an end and never merely as a means. Using another person is not itself wrong; buyers and sellers use one another constantly. The wrong lies in treating someone merely as a means, bypassing their rational agency through deception or coercion so that they cannot share in the end for which they are being used. NoteThis formula is often the most intuitive entry point to Kant, since it captures a widely shared sense that persons have a dignity that forbids manipulation.
Formula of the kingdom of ends
The third formula asks each agent to act as a legislating member of a kingdom of ends, an ideal community in which every rational being is at once author and subject of the moral law. This introduces Kant's notion of autonomy: a will is autonomous when it gives the law to itself rather than receiving it from desire or external authority, which he calls heteronomy. Autonomy, for Kant, is the ground of the dignity of rational beings.
Objections
Kant's ethics has drawn sustained criticism. One charge is rigorism: because perfect duties admit no exception, the theory seems to forbid lying even to a murderer asking where an intended victim is hiding, a case Kant himself notoriously addressed. A second objection is that the universal law test is empty or too permissive, since a suitably narrow maxim can be crafted to pass. A third points to conflicts of duty that the theory gives no clear procedure to resolve. Defenders reply that many objections target crude readings of the formulas and that the humanity formula supplies substantive moral content the universal law test alone does not.[4]
Kant and AI ethics
Kant's framework has become a common reference in debates over artificial intelligence. The humanity formula is invoked against systems that manipulate users or treat people merely as data sources, and the language of autonomy shapes arguments about consent and algorithmic influence. Some researchers have explored whether a categorical imperative could be operationalized as a constraint for automated moral agents, while others argue that machines, lacking a rational will of their own, can at most model Kantian rules rather than genuinely act from duty. The theory thus serves less as an algorithm than as a standard against which the treatment of persons by automated systems can be judged.
For the wider setting of these questions, see the entry on meta-ethics, normative, and applied ethics, and compare Kant's rule-based approach with the case-driven reasoning of the trolley problem. A full list of entries is on the home page, and the project's method is described on the about page.