The great influence at present exercised by Kant's teaching makes it worth while to state briefly the arguments by which he attempts to establish the duty of promoting the happiness of others, and the reasons why I am unable to regard these arguments as cogent. In some passages he attempts to exhibit this duty as an immediate deduction from his fundamental formula---``act from a maxim that thou canst will to be universal law''---when considered in combination with the desire for the kind services of others which (as he assumes) the exigencies of life must arouse in every man. The maxim, he says, ``that each should be left to take care of himself without either aid or interference'', is one that we might indeed conceive existing as a universal law : but it would be impossible for us to will it to be such. ``A will that resolved this would be inconsistent with itself, for many cases may arise in which the individual thus willing needs the benevolence and sympathy of others'' (Grundlegung, p. 50 [Rosenkrantz]). Similarly elsewhere (Metaph. Anfangsgr. d. Tugendlehre, Einleit. §8 and §30) he explains at more length that the Self-love which necessarily exists in every one involves the desire of being loved by others and receiving aid from them in case of need. We thus necessarily constitute ourselves an end for others, and claim that they shall contribute to our happiness: and so, according to Kant's fundamental principle, we must recognise the duty of making their happiness our end.

Now I cannot regard this reasoning as strictly cogent. In the first place, that every man in need wishes for the aid of others is an empirical proposition which Kant cannot know a priori. We can certainly conceive a man in whom the spirit of independence and the distaste for incurring obligations would be so strong that he would choose to endure any privations rather than receive aid from others. But even granting that every one, in the actual moment of distress, must necessarily wish for the assistance of others; still a strong man, after balancing the chances of life, may easily think that he and such as he have more to gain, on the whole, by the general adoption of the egoistic maxim; benevolence being likely to bring them more trouble than profit.

In other passages, however, Kant reaches the same conclusion by an apparently different line of argument. He lays down that, as all action of rational beings is done for some end, there must be some absolute end, corresponding to the absolute rule before given, that imposes on our maxims the form of universal law. This absolute end, prescribed by Reason necessarily and a priori for all rational beings as such, can be nothing but Reason itself, or the Universe of Rationals; for what the rule inculcates is, in fact, that we should act as rational units in a universe of rational beings (and therefore on principles conceived and embraced as universally applicable). Or again, we may reach the same result negatively. For all particular ends at which men aim are constituted such by the existence of impulses directed towards some particular object. Now we cannot tell a priori that any one of these special impulses forms part of the constitution of all men: and therefore we cannot state it as an absolute dictate of Reason that we should aim at any such special object. If, then, we thus exclude all particular empirical ends, there remains only the principle that ``all Rational beings as such are ends to each'': or, as Kant sometimes puts it, that ``humanity exists as an end in itself''.

Now, says Kant, so long as I confine myself to mere non-interference with others, I do not positively make Humanity my end; my aims remain selfish, though restricted by this condition of non-interference with others. My action, therefore, is not truly virtuous; for Virtue is exhibited and consists in the effort to realise the end of Reason in opposition to mere selfish impulses. Therefore ``the ends of the subject, which is itself an end, must of necessity be my ends, if the representation of Humanity as an end in itself is to have its full weight with me'' (Grundlegung, p. 59), and my action is to be truly rational and virtuous.

Here, again, I cannot accept the form of Kant's argument, The conception of ``humanity as an end in itself'' is perplexing: because by an End we commonly mean something to be realised, whereas ``humanity'' is, as Kant says, ``a self-subsistent end'': moreover, there seems to be a sort of paralogism in the deduction of the principle of Benevolence by means of this conception. For the humanity which Kant maintains to be an end in itself is Man (or the aggregate of men) in so far as rational. But the subjective ends of other men, which Benevolence directs us to take as our own ends, would seem, according to Kant's own view, to depend upon and correspond to their non-rational impulses---their empirical desires and aversions. It is hard to see why, if man as a rational being is an absolute end to other rational beings, they must therefore adopt his subjective aims as determined by his non-rational impulses,