The previous posting on Chapter 9 of Allison's commentary on the Groundwork looked at his treatment of the various formulas of the categorical imperative and the type of defence he offers for the view that they are all "equivalent", at least in terms of extensionality. In this posting, by contrast, I'm going to look at how Allison treats Kant's contrast between autonomous and heteronomous principles of morality.
Allison opens his treatment of the contrast between autonomous and heteronomous principles of morality by looking at the three different ways Kant views autonomy. On Allison's reading Kant relates to the principle of autonomy as the principle of morality, as a formula of the categorical imperative and and as a property of the will, taking these to be three distinct characterisations. However, the key point for Allison is that it is because autonomy is a property of the will that it is the principle of morality and not because it is presented as one of the formulas of the categorical imperative. The reason for making this claim is that the property of autonomy is what enables it to be said that the will can bind itself without referring thereby to anything "objectively" required beyond its own willing. It is because, however, that autonomy is the supreme principle of morality that this principle gives unconditioned authority to the categorical imperative. Once these two characterisations are brought together, however, it is less obvious than Allison suggests that we can claim that the reference to the categorical imperative is not part of what makes the principle of autonomy the supreme principle of morality. After all, in the very name of the "categorical imperative" we have reference to the notion that it is unconditionally binding, something that supports my contrary suggestion that there is an intimate connection between the notion of the categorical imperative and the idea of autonomy.
Now, in turning directly to the contrast with heteronomy, it follows easily from Allison's description that the notion of heteronomy can be captured as requiring that the will is not the source of the moral law but is rather dependent for its relation to the moral law on something beyond itself. However, many readings of Kant's description of heteronomy view this as requiring that heteronomous principles of morality essentially depend on a hedonistic kind of motivation (something that prevents, however, the recognition of intellectualist forms of principle being seen as heteronomous). There are quite a few problems with this, even when the question is restricted to considering what is going on for Kant when reference to sensations is paramount in the choice of moral principle. For example, as Allison points out, the benevolent philanthropist of Groundwork I feels pleasure in being benevolent but this does not necessarily mean that it is only for the sake of the pleasure that accompanies the benevolent acts that they are being benevolent. After all, if they did not think the acts they performed were genuinely addressing cases of need, they would not perform them.
The basic problem with heteronomous principles is not found in reference to a presumed hedonistic motivation but rather in a false conception of the relationship between the law and the agent. In adopting a heteronomous principle the assumption is made that the decision to will the law is insufficient for actions governed by it to have moral worth, perhaps on the grounds of an assumed "anarchic" conception of such autonomous willing. Whatever the rationale in any case, in adopting a heteronomous principle of morality, the agent assumes that worth is given to action by something intrinsically external to the act of willing itself and thus cannot find the property that is distinctive to morality within willing. There are, as Allison goes on to point out, two distinct types of ground for heteronomy: empirical and rational. Empirical theories are not all reducible to hedonism since sentimentalist accounts of morality can take there to be delight in the adoption of maxims that are virtuous. However they are classified by Kant in relation to hedonism in a different way, namely, in viewing the increase of "happiness" in a general sense (even separately from reference to the one willing) as the central matter.
Allison's discussion of the contrast between autonomy and heteronomy is brief and concentrates primarily on making the view of heteronomy more complicated than many other interpreters have suggested. However the conclusion of the chapter on autonomy does suggest a basic problem with Allison's overall approach which rests on a failure to relate the unconditionally binding character of autonomy to the unconditional requirement that is stated in the categorical imperative clearly enough. This is not contingent to Allison's presentation, however, but essential to it given that his interpretation cannot recognise the claim that the principle of autonomy is bound up with the very notion of the categorical imperative.