One very common way to read Kant’s moral philosophy is with a heavy emphasis on the self-legislative activity of the subject. Championed here is the anti-dogmatist tenor of Kant’s thought: the idea that normative authority can no longer be seen to simply rest in God or nature but must be grounded in a radically self-authorising subject who neither needs these external props nor can in good conscience accept them if they are to heed the Enlightenment call of ‘sapere aude!’ — to dare to know, to step into the maturity announced by the use of one’s own powers of reason.
As attractive as this picture can be in some respects, it often risks underplaying the formal elements in Kant’s project that are crucial to understanding the role of law and so why it is self-legislation that is invoked. To get a grasp on what work the word ‘formal’ is doing in this case, consider the following maxims (taken from Plato’s Republic as presented by Korsgaard):
(a) I will keep my weapon, because I want it for myself.
(b) I will refuse to return your weapon, because I want it for myself.
(c)I will refuse to return your weapon, because you have gone mad and may hurt someone.
They can be seen to be composed of two parts: an action and a purpose. Insofar as we are inclined to think that maxims (a) and (c) are good ones and maxim (b) a bad one, it seems that this cannot merely be a function of the actions or purposes. This is because the purposes of the first two are the same (acting because I want the weapon) and the actions of the second two are also the same (refusing to return the weapon). (a) and (c) both share parts — certainly different ones, but in combination exhaustive — with (b), so it cannot be merely in having a certain part that the maxim becomes a good one. Rather, it is the way that actions and purposes are combined that make the two maxims good ones. But form is not merely the way in which certain matter is combined. We said that the two maxims were good because of their combinations of actions and purpose; and form includes this directedness: it is the functional arrangement that enables something to do what it does, being an exemplum of its kind. (Obviously there are both relatively anodyne and much more controversial ways of understanding the ontological implications of this.)
Kant is concerned with what it is that distinguishes good maxims — the ones that are acceptable to act on — from bad ones — those that we ought not to act on. As we have seen, for him this cannot merely consist in identifying either certain actions or purposes that are good in-themselves. It requires us to examine the functional arrangement of actions and purposes. Kant’s suggestion as to the property of maxims that gives shape to this functional arrangement, that, as it were, locates what function that the arrangement should serve if the maxim is to be a good one, is being law-like. That is, those maxims that can be acted upon while it would be possible to consistently will them to be a natural law such that everyone would follow its dictates.
The way that I think Kant’s thought should be taken is as a thesis about the universality of reason, which implies its authority for and over us irrespective of our desire to heed its call. We can think of this as follows. For Kant, as for Socrates, to deliberate is to decide what ends are valuable, and insofar as we do decide to do something then we take this end as valuable. (Although akrasic factors might intervene before performing the action.) In performing an action pursuant of such an end, as distinct from a reflex reaction occurring or engaging in incoherent behaviour, we must be able to take ourselves to be acting for a reason: this is constitutive of actions (at least for Kant, if not Freud, say). So, given all the details of the particular situation that we find ourselves in, acting at all requires of us to be able to be able to produce reasons as to why we acted as we did in this situation.
The status of these reasons can then be questioned; the reasons in question being justificatory reasons, that is, not explanatory ones — we want to know whether although you took such-and-such to be a good reason to act as you did, was it really such a good reason? In assessing this, especially in the case of a morally dubious action, we may be tempted to point to extenuating circumstances: ‘Yes, I stole her wallet but I live in grinding poverty and have a starving family to feed.’ In a sense this is fine, for all the details of a particular situation are open to be appealed to as potentially relevant factors. However, it will not do to simply appeal to the fact that it is oneself who is acting, that the standards that are applied to others do not apply to me. For there is no metaphysically positive property that distinguishes you as exceptional and at the same time would be normatively relevant here. This means that there is no rational justification for free-riding that could not also be appealed to by anyone else with equal vehemence. We might put the point like this: for Kant there is no greater sin than ontological narcissism.
The formal structure of a law goes some way to excluding this temptation to make exceptions for oneself, for laws are universal. This means that, as Onora O’Neill puts it, they “prescribe for all cases in their scope.” A formal law is attached to fixed criteria, and these provide a measure of impartiality to it — it is not directed specifically at me. Contrast this with commands, which, stricto sensu, are particular, e.g. ‘Shut the door!’ or ‘You three are coming with me.’ If we find that our maxims do not have the formal structure of laws, that to be a subjective normative force for us they must be construed as commands, then considered as contributing reasons for action then they will be found wanting. This is because a brute command will find no rational warrant, whereas a law will be valid for everyone — it will contribute ‘public reasons’, so to speak, in support of the action.
So, to briefly return to the beginning, what is important in self-legislation is not the misleading picture of the agent that somehow binds or authorises itself, but rather the fact that the agent is following maxims with the formal structure of laws, stripped of all partiality. In following these rationalised maxims, ones where no unjustifiable exceptions are made simply because it is I who is acting, we act in a way proper to a being whose nature is itself infused with rationality, with the laws thus no longer being alien impositions: this is the self-legislative component.