In Section III of the Groundwork , Kant attempts to prove that the categorical imperative, derived in Section II by the analysis of the concept of free and rational beings in general, actually puts us under an obligation by proving that we are indeed free and rational beings. In his terminology, he wants to show that it is not merely an analytic but a synthetic a priori proposition that our wills are constrained by this imperative. Both the interpretation and the assessment of the arguments by which he proposes to accomplish this remain controversial.
The first claim that Kant makes is that ‘every being who cannot act except under the idea of freedom is just on that account really free in a practical respect, that is, all laws that are inseparably bound up with freedom are valid for it just as if its will were really declared to be free in itself and in theoretical philosophy’, and that every being with a will must indeed act under the idea of freedom (4: 448) (see Will, the ). This seems to mean that agents who conceive of themselves as choosing their own actions, whether or not they conceive of themselves as subject to determinism, do not or perhaps even cannot consider any antecedent determinants of their actions in deciding what to do, but only what now seems most rational to do; thus they must govern their actions by rational and therefore moral laws. This seems right for agents considering their own future actions, but leaves unclear how we are to assess the freedom of the actions of others or even our own past actions.
However, Kant goes on to offer what seems to be a theoretical and therefore general proof of the existence of human freedom. He argues that theoretical philosophy has shown that we must distinguish between considering ourselves as phenomena and noumena, or members of the sensible and the intelligible worlds. From the first point of view, we must consider our actions to be governed by the causality of nature, while in the second, since we cannot consider our actions there to be governed by no law at all, we must consider them to be governed by another kind of causality, namely causality in accord with laws of reason (4: 451-3 ). Thus while our actions appear to be determined by natural causes, in reality they not only can but in fact must accord with laws of reason, hence with the categorical imperative.
There are two problems with this argument. First, it flouts transcendental idealism by assuming positive knowledge about things in themselves. Second, as Henry Sidgwick was to object a century later, it precludes moral responsibility for wrong- doing: if the real laws of our behaviour are necessarily rational and hence moral, any wrong-doing could only show that an agent is not rational, and therefore not responsible, at all.
Whether consciously aware of such objections or not, Kant began to alter his argument for freedom of the will in the Critique of Practical Reason. Here he does not argue from a theoretical proof of our freedom to the fact of our obligation under the moral law, but conversely from our consciousness of that obligation – the ‘fact of reason’ – to our freedom as the necessary condition of our ability and responsibility to fulfil it (5: 29-31). This argument first assumes that transcendental idealism has left open at least the theoretical possibility of freedom of the will, and then depends upon the famous principle ‘ought implies can’ (‘Theory and Practice’, 8: 287 ). Transcendental idealism, of course, seems problematic to many; and although the ‘ought implies can’ principle seems an intuitive principle of fairness, Kant does not actually argue for it. Nevertheless, since this argument assumes only that ‘ought’ implies ‘can‘, it does not imply that any agent who is obliged under the moral law necessarily will act in accordance with it, and thus avoids Sidgwick’s problem about the very possibility of wrong-doing.
Kant depends upon this result in his next major treatment of freedom, in Religion within the Limits of Reason Alone, although there he seems to go too far in the other direction by assuming that evil-doing is not just possible but even necessary. Kant begins this discussion with an elegant account of wrong-doing, arguing that because no human being is simply unaware of the demand of morality – that is implied by the ‘fact of reason’ – acting immorally never comes from mere ignorance of the moral law, but rather from deciding to exempt oneself from this obligation. This position is compatible with the argument for freedom in the second Critique, although not with that of the Groundwork . However, Kant goes on to argue that an evil rather than virtuous choice of fundamental maxim, or ‘radical evil’, is not only possible but inevitable, to be escaped from only by a moral conversion. This doctrine hardly follows from Kant’s previous argument, and seems instead to rest on an odd mixture of empirical evidence and the lingering grip of the Christian doctrine of original sin.
The reality of freedom is only the first of Kant’s three ‘postulates of pure practical reason’; the other two are the existence of God and the immortality of the soul. Again Kant’s argument is that, as the first Critique showed, neither of these can be proven by theoretical metaphysics, but they can nevertheless be postulated as necessary conditions of something essential to morality. In this case, however, they are conditions not of our obligation under the categorical imperative but for the realization of the ‘highest good.’ This is another complex and controversial concept. Kant typically defines it as happiness in proportion to virtue, which is worthiness to be happy (5: 110), but suggests different grounds for the necessity of this conjunction. In the Critique of Practical Reason, Kant sometimes treats happiness and virtue as two separate ends of human beings, one our natural end and the other our moral end, which we simply seek to combine (5: 110). In other places, however, beginning with the ‘Canon of Pure Reason’ in the first Critique, he holds that since what virtue does is precisely to coordinate our mutual pursuit of ends, and happiness arises from the realization of ends, maximal happiness would inevitably follow maximal virtue under ideal circumstances (A 809/B 837 ). Of course, circumstances are not always ideal for morality: as far as we can see, no one achieves perfect virtue in a normal lifespan, and such virtue as is attained is hardly always rewarded with happiness. To counter this, Kant holds that we may postulate immortality, in which to perfect our virtue, and the existence of God, who can legislate a nature in which the ends of virtue are achieved.
This theory has seemed to many to be Kant’s vain attempt to save his personal faith from his own scathing critique of metaphysics. Before such a claim could even be discussed, we would have to know what Kant really means by a postulate of practical reason. Kant gives several hints about this which have not been adequately explored. In the first Critique, he discusses the practical postulates in a section where he considers readiness to bet as a measure of belief, thus suggesting that what he actually has in mind is Pascal’s wager (see Pascal, B.): since there is no theoretical disproof of these postulates, and nothing to lose if they are false, but their value to happiness is great, it is rational to act as if they were true. In a later essay, a draft on the ‘Real Progress of Metaphysics from the Time of Leibniz and Wolff’ from the early 1790s (posthumously published), Kant makes an even more striking suggestion. There he says that in the assumption of the practical postulates ‘the human being is authorized to grant influence on his actions to an idea which he, in accord with moral principles, has made himself, just as if he had derived it from a given object’ (20: 305 ). Here the suggestion is that the practical postulates are nothing less than another expression of human autonomy: not theoretical beliefs at all, let alone religious dogmas, but ideas which we construct for ourselves solely to increase our own efforts at virtue. This idea, that God is in fact nothing but an idea of our own making for use within our moral practice, is a thought Kant repeatedly expressed in his very last years (see §14).