Kant's Categorical Imperative

Kant said Hume woke him from a dogmatic slumber, and the shock was specific. Hume had shown that causation cannot be read off experience. But Newtonian science, which Kant regarded as unquestionably successful, treats causation as universal and necessary. Both could not be right, and abandoning either was unthinkable.

His solution is the strangest and most consequential move in modern philosophy. What if the necessity is not out there in the world, but in us? What if space, time and causation are not features we discover but conditions our minds impose on anything that could count as experience at all?

He called this his Copernican turn. Copernicus explained the motion of the heavens by moving the observer; Kant explains necessary knowledge by making objects conform to our cognition rather than the other way round. If the mind structures all experience, then the structure holds for every possible experience, with certainty, in advance. That is how substantive knowledge can be had before you look.

The price is severe and Kant pays it openly. If everything you experience already carries the structure your mind imposes, you can never step outside your own apparatus to check the unfiltered version. Things as they appear (phenomena) are knowable completely. Things as they are in themselves are unknowable in principle, not merely undiscovered.

That sounds like pure loss until you see what it buys. Science gets the world of appearances, causally closed and certain. Freedom, which strict causal law would otherwise crush, operates in a space appearances cannot reach. The two claims Kant refused to give up can both stand, because they are about different things.

The ethics follows the same logic. A moral requirement must bind whatever you happen to want, or it is not a moral requirement but a piece of advice about getting what you want. So it cannot be grounded in any desire, or any consequence, or any authority. Strip all of that away and what remains is the bare form of law itself: universality.

Hence the categorical imperative. Act only on a maxim you could will to become a universal law. And the test is more precise than it sounds. Consider promising to repay a loan you know you cannot repay. Now universalise it: everyone in that position makes the same promise. The result is not merely unpleasant. Promises of that kind stop working, so the practice you were relying on to make your promise effective could not exist. The maxim destroys its own precondition. That is a contradiction in conception, not a bad forecast.

The second formulation is the one that entered public life. Rational beings set their own ends, which gives them a worth that is not exchangeable for anything: dignity rather than price. So: treat humanity, in yourself or anyone else, always as an end and never merely as a means. The word merely carries the weight. Paying a taxi driver uses them as a means and is entirely fine, because it does not treat them as only that.

Two objections have real force. Hegel argued that universalisability is empty, since a maxim described narrowly enough will always pass. And there is the murderer at the door: asked where your friend is hiding, an absolute ban on lying seems to require you to tell. Kant took up that case in a late essay and accepted the conclusion, which almost nobody has been able to defend since.

But look at what survived. Human dignity as a non-negotiable in constitutions. Informed consent in medicine. The refusal to accept that a person can be traded off for a larger total. All of that is Kant, and all of it is a working refusal of the idea that people are resources.