If now we attend to ourselves on occasion of any transgression of duty, we shall find that we in fact do not will that our maxim should be a universal law, for that is impossible for us; on the contrary, we will that the opposite should remain a universal law, only we assume the liberty of making an exception in our own favour or (just for this time only) in favour of our inclination. Consequently if we considered all cases from one and the same point of view, namely, that of reason, we should find a contradiction in our own will, namely, that a certain principle should be objectively necessary as a universal law, and yet subjectively should not be universal, but admit of exceptions. As however we at one moment regard our action from the point of view of a will wholly conformed to reason, and then again look at the same action from the point of view of a will affected by inclination, there is not really any contradiction, but an antagonism of inclination to the precept of reason, whereby the universality of the principle is changed into a mere generality, so that the practical principle of reason shall meet the maxim half way. Now, although this cannot be justified in our own impartial judgement, yet it proves that we do really recognise the validity of the categorical imperative and (with all respect for it) only allow ourselves a few exceptions, which we think unimportant and forced from us.
We have thus established at least this much, that if duty is a conception which is to have any import and real legislative authority for our actions, it can only be expressed in categorical and not at all in hypothetical imperatives. We have also, which is of great importance, exhibited clearly and definitely for every practical application the content of the categorical imperative, which must contain the principle of all duty if there is such a thing at all. We have not yet, however, advanced so far as to prove a priori that there actually is such an imperative, that there is a practical law which commands absolutely of itself and without any other impulse, and that the following of this law is duty.
With the view of attaining to this, it is of extreme importance to remember that we must not allow ourselves to think of deducing the reality of this principle from the particular attributes of human nature. For duty is to be a practical, unconditional necessity of action; it must therefore hold for all rational beings (to whom an imperative can apply at all), and for this reason only be also a law for all human wills. On the contrary, whatever is deduced from the particular natural characteristics of humanity, from certain feelings and propensions, nay, even, if possible, from any particular tendency proper to human reason, and which need not necessarily hold for the will of every rational being; this may indeed supply us with a maxim, but not with a law; with a subjective principle on which we may have a propension and inclination to act, but not with an objective principle on which we should be enjoined to act, even though all our propensions, inclinations, and natural dispositions were opposed to it. In fact, the sublimity and intrinsic dignity of the command in duty are so much the more evident, the less the subjective impulses favour it and the more they oppose it, without being able in the slightest degree to weaken the obligation of the law or to diminish its validity.
Here then we see philosophy brought to a critical position, since it has to be firmly fixed, notwithstanding that it has nothing to support it in heaven or earth. Here it must show its purity as absolute director of its own laws, not the herald of those which are whispered to it by an implanted sense or who knows what tutelary nature. Although these may be better than nothing, yet they can never afford principles dictated by reason, which must have their source wholly a priori and thence their commanding authority, expecting everything from the supremacy of the law and the due respect for it, nothing from inclination, or else condemning the man to self-contempt and inward abhorrence.
Thus every empirical element is not only quite incapable of being an aid to the principle of morality, but is even highly prejudicial to the purity of morals, for the proper and inestimable worth of an absolutely good will consists just in this, that the principle of action is free from all influence of contingent grounds, which alone experience can furnish. We cannot too much or too often repeat our warning against this lax and even mean habit of thought which seeks for its principle amongst empirical motives and laws; for human reason in its weariness is glad to rest on this pillow, and in a dream of sweet illusions (in which, instead of Juno, it embraces a cloud) it substitutes for morality a bastard patched up from limbs of various derivation, which looks like anything one chooses to see in it, only not like virtue to one who has once beheld her in her true form.11
The question then is this: "Is it a necessary law for all rational beings that they should always judge of their actions by maxims of which they can themselves will that they should serve as universal laws?" If it is so, then it must be connected (altogether a priori) with the very conception of the will of a rational being generally. But in order to discover this connexion we must, however reluctantly, take a step into metaphysic, although into a domain of it which is distinct from speculative philosophy, namely, the metaphysic of morals. In a practical philosophy, where it is not the reasons of what happens that we have to ascertain, but the laws of what ought to happen, even although it never does, i.e., objective practical laws, there it is not necessary to inquire into the reasons why anything pleases or displeases, how the pleasure of mere sensation differs from taste, and whether the latter is distinct from a general satisfaction of reason; on what the feeling of pleasure or pain rests, and how from it desires and inclinations arise, and from these again maxims by the co-operation of reason: for all this belongs to an empirical psychology, which would constitute the second part of physics, if we regard physics as the philosophy of nature, so far as it is based on empirical laws. But here we are concerned with objective practical laws and, consequently, with the relation of the will to itself so far as it is determined by reason alone, in which case whatever has reference to anything empirical is necessarily excluded; since if reason of itself alone determines the conduct (and it is the possibility of this that we are now investigating), it must necessarily do so a priori.
The will is conceived as a faculty of determining oneself to action in accordance with the conception of certain laws. And such a faculty can be found only in rational beings. Now that which serves the will as the objective ground of its self-determination is the end, and, if this is assigned by reason alone, it must hold for all rational beings. On the other hand, that which merely contains the ground of possibility of the action of which the effect is the end, this is called the means. The subjective ground of the desire is the spring, the objective ground of the volition is the motive; hence the distinction between subjective ends which rest on springs, and objective ends which depend on motives valid for every rational being. Practical principles are formal when they abstract from all subjective ends; they are material when they assume these, and therefore particular springs of action. The ends which a rational being proposes to himself at pleasure as effects of his actions (material ends) are all only relative, for it is only their relation to the particular desires of the subject that gives them their worth, which therefore cannot furnish principles universal and necessary for all rational beings and for every volition, that is to say practical laws. Hence all these relative ends can give rise only to hypothetical imperatives.
Supposing, however, that there were something whose existence has in itself an absolute worth, something which, being an end in itself, could be a source of definite laws; then in this and this alone would lie the source of a possible categorical imperative, i.e., a practical law.
Now I say: man and generally any rational being exists as an end in himself, not merely as a means to be arbitrarily used by this or that will, but in all his actions, whether they concern himself or other rational beings, must be always regarded at the same time as an end. All objects of the inclinations have only a conditional worth, for if the inclinations and the wants founded on them did not exist, then their object would be without value. But the inclinations, themselves being sources of want, are so far from having an absolute worth for which they should be desired that on the contrary