The concept connected by Kant with the phrase “categorical imperative” seems easy to explain. It means a law valid for the will of every rational being and therefore valid unconditionally. This is in no need of interpretation: it can hardly be misinterpreted. Kant has stated in the clearest and most intelligible terms. Understanding becomes much more difficult if we are concerned, not merely with the verbal definition, but with the content of the categorical command and the interferences to be drawn from it.

Even specialists have fallen into confusion about these questions. But the doctrine of a categorical imperative inherent in the will of man himself appears at present to meet with most unexpected and most unwanted repercussions in the common opinion. To lovers of humanity it may seem a lofty and worthy aim that doctrines elaborated by science and claiming to determine the conduct of every man should gradually spread wider and wider until at last every cottage.

The law formulated by Kant in his categorical imperative is not one by which any principle whatever to which a many may find himself drawn under conditions of experience—such as obedience to superiors or abstention from pleasure of life—can be imposed on him categorically. Kant’s law is rather a way of expressing the conditions under which alone a principle can have the character of a categorical demand. The categorical imperative is thus conceived as the fundamental principle determining which possible principles can be objectively valid for the decisions of our will as such.

When we say it is our duty to do something or to refrain from doing it, we manifestly have in mind such a categorical demand or such an objectively valid principle. Hence we can also say that on Kant’s view the categorical imperative contains nothing but the concept of being under a possible moral obligation as such. If he was wrong in maintaining that such a command is binding upon our will, it is not to be inferred that duty must be determined by some other law; what it would mean is that there are no universally valid demands on human behavior, so far as this depends on our will, and consequently that nothing whatever can be our duty and that we are entirely free to do whatever we may happen to want.

We might indeed by acting in this way get entangled in all sorts of disagreeable consequences or we might be astute enough to find means of escaping these disagreeable consequences. But what Kant maintains is this: the sum total of these means could never have the character of a system of precepts for the will such that men would be under an objective obligation to obey them. Nor could a necessary harmony of these means be discovered (independently of a categorical imperative) such that it would be free from all possible conflict of the will with itself and with the will of others.

We can sum all this up in the proposition that the categorical imperative determines the concept of duty solely as regards its form. It states only what duty as such is and consequently what all duties have in common. But it contains nothing to show how the particular duties determined by it are materially different from each other. Yet we ought not to imagine that this purely formal character of the categorical imperative as the law of duty is identical with the property envisaged by those who describe as formal both the principle of Kant’s ethical theory and consequently this theory itself in distinction from other systems of morality.

Here we find the first misinterpretation of the categorical imperative—the view that this necessarily confines moral philosophy to stating what the concept of duty is simply as regards to its form and makes impossible the articulation of particular duties that are materially different. Every doctrine of duties, if it is based on any principle at all, must begin by stating in what duty consists—i.e. by stating what is the concept of duty simply as regards to its form. If we say, for example, that duty consists in performing the actions required to realize the happiness of the greatest number, we have determined the concept of duty simply as regards to its form. This special characteristic may be described—negatively—as the assertion that if we are to discover the concept of duty, we must abstract, not merely from all the matter of duty, as is obvious, but also—as it not so obvious—from all the matter of the will; that is from all purposes or ends.

Naturally, if anyone intends to realize something tat he has made the end of his action, it is necessary that he should adopt the means adequate to realize it. But such necessity is a conditioned one (hypothetical, not categorical); and it contains no immediate law for his will since he must already will something definite before he can become subject to this necessity. “Yes,” it may be replied, “we must naturally being with some final end. This is already stated by Aristotle. And just as naturally such an end must possess some kind of necessity for the human will.”

“We are not to infer,” John Stuart Mill says, “that it is acceptance or rejection must depend on blind impulse or arbitrary choice.” (U. 1) But if, with Aristotle and Mil, we take happiness as the final end on which we may hope to find all men agreed, can we then say that there is for man an unconditioned command to make his own happiness his end? Even if he may in fact always make it his end, no one can say that this inevitably to found in experience is the necessity of characterizing a demand in virtue of which man has the duty of adopting this end. And is it not at least conceivable that it may be both possible and necessary for the will of man to subordinate the end of his own happiness to a higher condition of his actions?

Suppose one answers, “Yes, admittedly there must be an end necessary in and for itself—that is, necessary for the will of every rational being independently of all subjective conditions—and end which alone supply the ground for an unconditioned demand that it be adopted.” The question then arises: in what way is it possible for us to have such an end? If we assume that man by nature can have no ends conflicting with the end that is necessary in itself, then clearly he can be subject to no command at all about the decision of his will: his will must always of itself have this highest good as its object. In that case there can be no question of duty.

If we admit, on the other hand, that he can decide to act contrary to this end which is necessary in and for itself, there must be, even beyond this highest end, a motive which can move him to do what he does not already do of himself—namely to make this end his own. He would then need to have above this highest end, a still prior end by which he would be required o take as his end what was good in and for itself. But such a prior end contains a contradiction; for all possible ends must be naturally subordinated to the end which is by definition the highest. Hence along this line we could produce neither a categorical nor a hypothetical imperative for adopting the end on which the whole necessity of human action is supposed to be dependent. We should be left merely with a choice between a will that was infallibly good and a will that could not even be asked to be good at all.

It is purely analytic considerations like these that ought to be adduced if we want to understand the real motive which led Kant to what constitutes the strictly formal character of his ethics. This character is to be found in the fact that in place of some previously given end—that is, in place of some matter of the will—Kant asserts that the mere form of maxims (that is, of the subjective principles of our arbitrary choice), so far as this fits them for the making of a universal law, is what determines duty as regards to its form.

After what has been said, there should be no need to explain further why in this statement the word “form” occurs twice. The form which fits maxims for the making of universal law—whatever may be their matter, that is their end—is here identified with that demand by which alone an imperative can have the character of being categorical, or in other words, by which duty is determined as regards to its form. If in the language of this abbreviation we call this imperative “formal,” this does not mean that it has no content. It has precisely the content we have just adduced; and we cannot infer conversely that because the categorical imperative has in fact a content, therefore Kant must be in error when he maintains that only the form which fits maxims for the making of laws can supply a universally valid ground for determining our will.

Apart from objections of this kind, which rest merely on a play of words, the fact that Kant has calmly built up a whole system of commands and prohibitions on the basis of his moral law has been looked on with the utmost suspicion and indeed with open scorn. “When he begins,” says John Stuart Mill, “to deduce from this precept any of the actual duties of morality, he fails, almost grotesquely, to show that there would be any contradiction, any logical impossibility, in the adoption by all rational beings of the most outrageously immoral rules of conduct.

This charge has its history. It is already present in Hegel’s assertion that when Kant deduces particular duties, he goes continually round in circles since in order to show a contradiction between a maxim and the possibility of willing it as a universal law he has always to presuppose as necessary the very volition whose necessity it is his business to prove. By this method it would be open to us to put forward any kind of arbitrary conduct as a demand of duty. This seems to confirm one view of the history of moral evaluation.

Yet we cannot say in Kant’s sense that atrocities have been commanded “at some time and some place” by conscience and duty, or are still commanded, unless we have first shown that such conduct is contained in the demand of his moral law. Obviously in the opinion of the author the formalism of Kant’s ethics consists in this—that his concept of duty can be applied to any arbitrary behavior, so far as man believes himself constrained to it by any will beyond which, for some subjective reason, he recognizes no higher.

We have at least establishes this—that the obedience required by the categorical imperative is the direct opposite of an obedience by which a man could subject himself at random to any arbitrary power. If anybody is looking for definite precepts contained in the categorical imperative, he can put this down as the first—that such subjection is forbidden. The ban admittedly is only another way of saying that the categorical imperative is in fact nothing but the law of the will’s autonomy (it’s making of its own laws). But with this view of it are we not plunged into a difficulty of the opposite kind?

Instead of inferring that our will must be unconditionally fettered to the will of another are we not bound to infer that our own arbitrary will must be unconditionally unfettered? Must not a will subjected solely to its own legislation be able to pass any law it likes? Must it not be able in virtue of its autonomy “to will as a law” any arbitrary maxim it says to use in regulating its choice of ends?

At any rate Kant—as we have already heard—fails “almost grotesquely” in the eyes of his critics when he tries to persuade us that we must will any particular maxim as a law or that we cannot will it as a law. This failure is obviously called “grotesque” because, in the examples he gives of such necessities and impossibilities, it looks as if he casually introduces, in order to reach a decision, that very consideration of the agent’s personal advantage which the categorical imperative professes to set aside as a principle of moral evaluation.

Suppose, for example, he wants to show that nobody can will as a law the maxim of not bothering about the distress of others: “For,” he says, “a will which decided in this way would be in conflict with itself, since many a situation might arise in which the man needed love and sympathy from others, and in which by such a law of nature sprung from his own will, he would rob himself of all hope of the help he wants for himself. A clear case, it might seem.