2. The independence of Recht

Rawls understands the difference between a political conception of justice and a comprehensive doctrine to be a ‘matter of scope’ (Rawls Reference Rawls1993: I.2.2, 13). A political conception affirms specifically political values, while a comprehensive doctrine makes reference to a broad range of values, both political and non-political. A comprehensive doctrine ‘includes conceptions of what is of value in human life, and ideals of personal character, as well as ideals of friendship and of familial and associational relationships, and much else that is to inform our conduct, and in the limit to our life as a whole’ (Rawls Reference Rawls1993: I.2.2, 13). Kant’s doctrine of right is thus political and not comprehensive if it excludes appeals to such more general values, and appeals only to political values. In fact, as Pogge stresses (Reference Pogge and Ellis2012: 88), Kant’s doctrine of right is likely more narrowly political even than Rawls’, because it relies only on purely juridical concepts, concepts derived only from the nature of coercive law and its essential relation to external freedom.

Kant begins his analysis of right by declaring: ‘The sum of those laws for which an external lawgiving is possible is called the Doctrine of Right’ (DR Introduction §A, 6: 229).Footnote ² An empirical doctrine of right would study the particular laws enacted by particular sovereigns (DR, 6: 229), but a pure or philosophical doctrine of right seeks ‘to establish the basis for any possible giving of positive laws’ (DR Introduction §B, 6: 230). Kant’s concern is thus with the very idea of coercive law: the ability of some person or set of persons to prescribe rules for the conduct of others, which those others can then be coerced to obey. What pure concept of right must be presupposed if such lawgiving is to be possible? Kant makes clear immediately that this most basic juridical concept is in fact a ‘moral concept’, because ‘it is related to an obligation corresponding to it’ (DR Introduction §B, 6: 230). This moral concept is already implicit in the nature of positive law, since any sovereign who makes positive laws takes their subjects to be obligated to obey those laws.Footnote ³

In this most basic sense of demanding legal obligation, all positive law pays homage to the pure concept of right as having moral content (a point that Kant had already exploited in ‘To perpetual peace’). Still, it would seem that this moral content might well come from outside the conditions of external lawgiving: for instance, from the separate moral value of freedom or autonomy. But Kant’s argument is that freedom and its value are also internal to the conditions of positive law. In demanding legal obedience, in expecting all subjects to recognise submission to the law as obligatory, a sovereign is claiming the right to regulate the external freedom of subjects (to tell them what they must do, no matter what they think of those actions) who at the same time are recognised to be internally free (because they must also accept the obligation to obey, even if they do not themselves independently want to do what is legally demanded of them). The concept of freedom is thus essential to the pure, moral concept of right, because law is always addressed to free persons, in a way that presumes a basic distinction between external and internal freedom.Footnote ⁴ The main division of the Metaphysics of Morals, the distinction between right and virtue, between legality and ethics, thus follows immediately from the basic idea of a juridical subject, a free person who is expected to understand and (externally) obey a legal rule, whether or not they (internally) agree with that rule. The point also explains why, in the Metaphysics of Morals, the Doctrine of Right has to come first, before the Doctrine of Virtue. That there is a distinct realm of virtue itself follows from the concept of a juridical subject, which belongs to a prior doctrine of right. Juridical subjects are taken to be obligated to obey the law, but since they are required only to obey, their reasons for obeying are not legally relevant. But if they do obey for moral reasons – which they always can do – then that internal form of obedience belongs to virtue and not to right.

The concept of freedom, including the distinction between internal and external freedom, is thus presupposed as a condition of external lawgiving and thus belongs to the pure moral concept of right. But to say that freedom is a condition of external lawgiving is not yet to say that freedom, internal or external, is somehow the purpose or value of external lawgiving. Certainly, external lawgiving has purposes that can and do go beyond the nature of a juridical subject as free. The law can mandate the building of roads and bridges, even if the law can do that only if there are juridical subjects who are legally obligated to pay for and carry out their construction. Kant addresses this point by distinguishing between the matter of the law (the particular purposes the law wants to achieve, like the building of roads and bridges) and its form (the law’s authority to command free subjects to carry out a particular purpose, whether a road or a bridge or anything else). He then argues that only the form of law is relevant to the moral concept of right because juridical obligation cannot derive from the matter of the law. A road or a bridge is something the sovereign wants or something the sovereign thinks their subjects want. But how can I be obliged to do what someone else wants, or they be obliged to do something I want? Such a framework of obligation presupposes a kind of contract to which the citizen or sovereign are parties, since for Kant, a contract just is the structure in which one person’s choice is able to determine the choice of another (DR §18, 6: 271). But in Kant’s view, even the idea of a valid contract presupposes a prior notion of right, which concerns only the form and not the matter of the parties’ choices. To determine whether a contract is juridically enforceable, ‘it is not asked, for example, whether someone who buys goods from me for his own commercial use will gain by the transaction or not. All that is in question is the form in the relation of choice on the part of both, insofar as the choice is regarded merely as free’ (DR Introduction §B, 6: 230). The contract is valid not because anyone gets what they want, but only if both parties have freely agreed to it, which means they have coordinated their freedom according to a common rule.

This external coordination of agents’ choices is the real value of right, the only thing that could give rise to an obligation to comply with the wishes of another, in complying with the provisions of a valid contract or with the dictates of the law. There is thus, Kant is saying, a basic juridical relation, in which internally free beings can be related to law, solely in terms of their external freedom, and solely in terms of the form of their choices. This basic juridical relation is implied in every conception of positive law, in every sovereign’s presumption that they can command their subjects to do particular things, no matter what those subjects think of those particular things. But since this claim is independent of the identity of any particular sovereign and any particular purposes that a particular sovereign wants to advance, the basic juridical relation is really a possible relation between human beings more generally, a relation under which it is possible for any particular human being legally to coerce any other. Though he is certainly concerned with the question of what entitles the state to coerce its citizens, Kant understands this question to belong to a more general and fundamental juridical inquiry into what entitles any person to coerce any other. Any legal system is ultimately an attempt to resolve this more general and fundamental problem. Kant assumes, reasonably enough, that any realised legal system will include a conception of public right, in which there are established authorities who can make legal rules, and a conception of private right, through which individuals can make legal claims against one another, if they are injured in ways that the law recognises as wrong. What the concept of external lawgiving thus promises is a complete determination of the form of the choices of every individual with respect to every other: for any particular action, one should be able to say, according to a clearly specified rule, whether one is legally wronging anyone else; and if not, one can say that they are legally free to act. Contrast this rightful condition with the state of nature, in which one, by using any object or even by occupying a particular position in space and time, is thereby limiting the external freedom and thus potentially wronging everyone else.Footnote ⁵ What the concept of right does is to declare that our external freedom can and should be subject to complete juridical determination, so that one can always say whether or not they are legally free to act as they specifically choose. This juridical determination is not simply about one’s relation to the sovereign; it is about one’s relation to everyone else.Footnote ⁶ If you are legally free to do something, that means you have not wronged anyone. ‘Right is therefore the sum of the conditions under which the choice of one can be united with the choice of another in accordance with a universal law’ (DR Introduction §B, 6: 230).

In Rawlsian terms, this pure, moral concept of right is a strictly political concept, because it relies only on essentially juridical concepts, concepts that have an essential relation to the nature of external lawgiving. Those concepts, we have now seen, are just the concepts of (external) freedom, force (or coercion), and (universal) law, related in a particular way: as the subjection of our external freedom to a universal coercive law. Without any of these, or if they are related in any other way, there is no rightful or proper juridical condition at all. Hence, the famous passage from the Anthropology distinguishing a republic (force with freedom and law) from anarchy (law and freedom without force), from despotism (law and force without freedom), and from barbarism (force without freedom and law) (Anth, 7: 330–1; for a fuller discussion, see Varden Reference Varden2022). As the reference to republicanism should suggest, Kant wants to extract significant moral content from the basic juridical notion of a person subject to a legal rule. It is not just that this notion has moral content in that legal rules assume an obligation to comply. It is also that this obligation assumes a broader moral conception of human beings as free beings who stand in an ordered relationship with one another, who are subject to a set of legal rules in order that they can assert the moral claim that they are equally free to act as they do (DR Introduction, Appendix: 6: 237–238; Ripstein Reference Ripstein2009: 31–42; Reference Ripstein and Ellis2012: 45–48). For a legal system to function, everyone, both sovereigns and subjects, must understand the law as serving this basic moral purpose. Hence Kant’s argument about property: any legal system will have some conception of property rights since it will say what individuals are or are not free to do with objects in space, and any conception of property rights implies the idea of a general will, a collective permission of all that particular individuals can exercise their particular rights to property (DR §8, 6: 255–6; Ripstein Reference Ripstein2009: 94–96; Reference Ripstein and Ellis2012: 48–51; Krasnoff Reference Krasnoff, Krasnoff, Sanchez Madrid and Satne2018b: 110–12). To have a right to a thing means that all others should take themselves to be excluded from using it. Such an obligation can only be understood as the product of a rule we have equally imposed on ourselves to regulate our freedom with respect to all others. The ultimate trajectory of Kant’s political philosophy is that the obligation implied in any juridical relationship ultimately implies the idea of the general will (see Krasnoff Reference Krasnoff2018a) and a commitment to republicanism (TP, 8: 289–96). The sovereign is always the guardian of the general will, no matter what particular interests they have, and so it is always legitimate for any person to publicly criticise any particular law as inconsistent with the equal freedom of all juridical subjects. Hence, the argument of ‘To perpetual peace’: rulers may say that they need to give no consideration to the moral objections of ‘philosophers’, but in practice, they are always subject to their influence, since their claims to rule always rest on the idea that they do so on behalf of the public interest, which already implies the moral idea of the general will (see especially TPP, 8: 375–6 and TP, 8: 304).Footnote ⁷

All of these specific claims require much more elaboration. Here my concern is just to describe the basic structure of the Rechtslehre, which is to begin with the moral content already implicit in the bare concept of juridical obligation and then to trace its moral consequences. Since this inquiry is moral throughout, it is quite misleading for Pogge to declare that ‘Kant’s political philosophy is independent from his morality’ even if only in the sense that his political philosophy does not imply his moral philosophy, while his moral philosophy is intended to imply his political philosophy (see Pogge Reference Pogge and Ellis2012: 89–90). Rather, we should say, as the structure of the Metaphysics of Morals clearly insists, that Kant’s political philosophy is an essential part of his moral philosophy, though a distinct, independent, and indeed primary part.Footnote ⁸ What moral claims are implied in Kant’s purely juridical account of right as the conditions of external lawgiving? Again, Kant understands any notion of positive law to already imply an obligation to obey, an obligation that only makes sense if we see any law as fundamentally committed to the value of relating all human beings to one another in a basic juridical way. What this basic juridical relation does is to subject the external freedom of every human being to a universal law, such that they can will particular things with assurance that they are not wronging anyone. Hence the moral claims implied in this universal principle of right are really quite simple. First, a rightful condition, one in which human beings are related in this fundamentally juridical way, is absolutely preferable to its alternative, the state of nature, in which anyone can be said to be wronging (or equally, not to be wronging) everyone else by acting in any way, without constraint. And second, to have this absolute preference is to endorse the equal value of everyone’s external freedom.

If we focus our attention on the independent commitment of Kant’s political liberalism to these two basic moral claims, then we can better understand the way that they function as premises, not conclusions or implications, of Kant’s more comprehensive practical philosophy. In the argument above, I have not sought to derive the two basic moral claims from the Groundwork’s (or any other) version of the categorical imperative (as in Gregor Reference Gregor1963 or Tomassini Reference Tomassini2023), or from some prior concept of freedom or its value (as in Guyer Reference Guyer and Timmons2002). Nor have I suggested that the notion of juridical obligation poses any special metaphysical problems, which first need to be addressed (as in Flikschuh Reference Flikschuh2000, although it should be noted that the moral account of innate right as legal accountability in Flikschuh Reference Flikschuh2021 is quite consistent with everything I have said here). In this sense, I am rejecting any form of derivationism in favour of an independence reading. I take Kant’s political philosophy to imply moral claims from the start, but I also take those claims to be independent of the ‘pure’ moral philosophy of the Groundwork and the Critique of Practical Reason. The deeper challenge facing any independence reading, of course, is then to explain what purposes these works have, if not to ground the moral claims of Kant’s political philosophy. My contention is that Rawls’ concept of comprehensive liberalism, properly understood, can help us to answer this question, and so put independence readings like mine on firmer ground.

3. The function of a Kantian comprehensive liberal doctrine

What theoretical purpose would a Kantian comprehensive liberalism serve? Pogge suggests a particular answer to this question, which unfortunately turns out to be neither particularly Kantian nor particularly Rawlsian. For Pogge, the problem to be solved is essentially motivational. On his reading, the idea of right as the complete juridical determination of the form of our choices means that our external freedom is thoroughly subject to rules so that we can always say what we are and are not free to do. Pogge thus describes Kant’s theory of right as a kind of game, a thoroughly rule-bound sphere of activity in which we can always say, clearly and definitely, what counts as a permissible move (Pogge Reference Pogge and Ellis2012: 84–5). Within this game, we can always say what is juridically right and wrong, but if we are to be motivated to play the game, we need to have an independent interest in playing it. Exploiting Kant’s point that the law demands the performance or non-performance of certain actions, as opposed to any particular reason or motivation for those actions or restraints, Pogge argues that different comprehensive doctrines might justify obedience to the rules of right in different ways:

Since I can have moral or prudential reasons for obeying any particular law, Pogge suggests that we might equally have moral or prudential reasons for preferring lawful conduct, in general, and thus for ‘playing the Rechtslehre game’. Particular comprehensive doctrines would then spell out these moral or prudential reasons in clear and coherent fashion. Kant’s comprehensive moral philosophy, in particular, would supply us with moral reasons for caring about right, and thoroughly explain the source of those reasons, presumably in the categorical imperative and its essential connection to our autonomy or transcendental freedom. A different, empirically grounded comprehensive doctrine, by contrast, would supply us with prudential reasons for caring about right, presumably through Hobbesian arguments for the self-defeating character of purely self-interested behaviour. To support his claims about the coherence of the second possibility, Pogge cites the notorious ‘people of devils’ passage from ‘To perpetual peace’:

For Pogge, the devils suggest the possibility of a non-Kantian comprehensive argument for an interest in Kantian right, ‘because it is, as the quote shows, just as true that selfishness gives its immoral adherents self-interested reasons for supporting Recht and a republican constitution in particular’ (Pogge Reference Pogge and Ellis2012: 89).

In fact, however, the quote shows no such thing. Pogge needs his hypothetical prudential comprehensive doctrine to supply the devils with reasons to accept the moral demands of right. This does not necessarily mean that those reasons will themselves be moral, but it does mean that the reasons will give the devils a stable commitment to acting rightly. But as Kant’s formulation suggests, the devils have not been rationally convinced to act rightly, since they remain secretly inclined to exempt themselves. In this passage, Kant is talking not about comprehensive theorists supplying the devils with general reasons to support right, but about external mechanisms being organised so that the devils’ self-interested motives collide in such a way as to ‘cancel one another’, producing the same large-scale behavioural outcomes that moral observers would hope for – which is very different from the devils supporting or accepting the rational force of right. The motivational perspective Pogge wants comprehensive doctrines to take up is first-personal: why should I rationally commit myself to caring about right? But the perspective of this passage is third-personal: how can sovereigns get the devils to care about things that cause them to act rightly, even though (by hypothesis) they have no stable, rational commitment to right? The larger perspective Kant takes in ‘To perpetual peace’, and especially in the context of the ‘people of devils’ passage, is not that of the individual, moral or otherwise, who needs to have reasons to be morally committed to right. Rather, Kant’s focus is on the perspective of an individual who is already morally committed to right, and who is then concerned with the subsequent question of whether a rightful, republican condition will actually prevail in the world, given that all our moral commitments are so wavering and precarious, against our many selfish inclinations. Such morally concerned individuals have a special interest in empirical mechanisms which might force self-seeking individuals to act rightly. But the teleological question of the realisation of right is quite different from Pogge’s motivational question of what reasons we have to commit ourselves to right and to the republicanism it implies.

Because Kant takes right and republicanism already to have moral force, Pogge’s motivational view is a bad candidate for an answer to the question of what a Kantian comprehensive doctrine might be expected to do. But more generally, it is also a bad fit for Rawls’ idea of what any comprehensive doctrine should be expected to do, which is to relate political and non-political values in some rationally coherent way. Rawls does not expect any comprehensive doctrine to explain why political values really are valuable, and why we have good reasons to accept them. It is the task of a political conception itself to articulate its own values in a compelling way. A comprehensive doctrine’s job is instead to relate political values to non-political values in a coherent and rationally compelling way so that individuals can rationally assure themselves that they can stably integrate their commitment to political values into the rest of their lives.Footnote ⁹ Recall that what makes a doctrine comprehensive is its scope: it applies to the full range of possible human values, as opposed to political values in particular. For this reason, a successful comprehensive doctrine must take up and give a rationally satisfying answer to a general theoretical question that a political conception does not have to face: what is the nature or source of value in the first place, such that political and other values can be rationally compared and integrated into a coherent form of life? This general theoretical question obviously has a justificatory aim, but we should be careful not to infer that the justificatory project of a comprehensive doctrine will necessarily include the aim of justifying every particular kind of value, including political value, in a deeper or more fundamental sense. A successful comprehensive doctrine should be able to explain why political commitments have value, in a way that can be compared to other kinds of value, but that kind of explanation does not have to imply that the source of political value lies in something deeper, in something external to the political itself. Notably, Rawls includes among his examples of possible comprehensive doctrine a pluralist view, grounded in a fundamentally negative claim that no unified account of the sources of value is possible (Rawls Reference Rawls1993: IV.3.2, 145–6). Such a view is clearly not committed to grounding the value of the political in some basic source of value. All that a pluralist or any other comprehensive doctrine must do is explain why it makes rational sense for an individual to be stably committed to political values, given the claims that other values make on us. This justificatory work will require general theoretical claims about the nature and sources of value, but these claims do not themselves have to imply an independent justification of political value.

Unlike pluralism, a Kantian comprehensive liberal doctrine will be committed to a unified account of the nature and sources of value, which it will attempt to explain in a fundamentally liberal way. But this commitment need not imply that political liberalism rests on some deeper moral value. In my reading of the independence of right, Kant is arguing that political liberalism has moral value in itself. The task of his more comprehensive practical philosophy is then to provide a more general account of value that explains how the moral value of right fits into the rest of our lives, given that right demands only obedience to law, not anything deeper about how we best ought to live. If Kant’s comprehensive practical philosophy is fundamentally liberal, that means that his more general account of value is fundamentally informed by liberalism itself, by his independent account of the moral value of right. We can then use the fundamentally political claims of this account to develop a more general account of the nature and source of value. In the next section, I will describe the specific way that the ‘pure’ moral philosophy of the Groundwork attempts to carry out this theoretical project. But in my account, the concept of juridical obligation and its two basic moral implications (an absolute preference for a rightful condition over the state of nature, and the equal value of everyone’s external freedom) will be taken as premises that support the theoretical argument, not as conclusions that the theoretical argument somehow establishes.

If it seems that this logic is undermined by the chronology of Kant’s works – after all, the Groundwork was published in the mid-1780s, while the Rechtslehre did not appear until the late 1790s – we should keep in mind that even in the 1780s, Kant was also presenting his account of right independently, as evidenced by the Feyerabend lectures on natural right. Even more decisively, Kant had already declared in the Critique of Pure Reason that ‘[w]ithout doubt, the concept of right that is used by the healthy understanding contains the very same things that the most subtle speculation can evolve out of it’ (CPR, A43/B61). Later in the work, when he introduces the ideas of reason, Kant defends their practical reality by asserting a version of the Rechtslehre’s principle of right: ‘A constitution providing for the greatest human freedom according to laws that permit the freedom of each to exist together with that of others … is at least a necessary idea, which one must make the ground not merely of the primary plan of a state’s constitution but of all the laws too’ (CPR, A316/B373). None of these claims seem to require the argument of the Groundwork. So the chronology actually points in favour of my reading. Still, chronological arguments can never be decisive here: what finally matters is the theoretical question of whether the arguments of the Rechtslehre require the arguments of the Groundwork. To the contrary, I will argue that we can understand the arguments of the Groundwork much better if we already assume the basic moral claims of the Rechtslehre.

4. What is a groundwork for the metaphysics of morals?

The aim of Kant’s best-known work in moral philosophy may seem obvious from its title’s reference to a ‘foundation’ for a metaphysics of morals (including a doctrine of right), and from Kant’s explicit conception of the work as, in Gregor’s translation, ‘nothing more than the search for and establishment of the supreme principle of morality’ (G, 4: 392). (Ellington, similarly, offers ‘to seek out and establish the supreme principle of morality’.Footnote ¹⁰ ) But the subsequent arguments of the work make these professed aims more complicated than they might first appear. In what sense do we need to search for or seek out the supreme moral principle, given that Kant later tells us that ordinary moral agents already understand and employ this principle all the time (G, 4: 492)? In what sense is the principle finally established, given that it rests on the claim that we are transcendentally free, which Kant says can never be finally proven (G, 4: 459)? In fact, the original German words Kant uses for his aims – nicht mehr, als die Aufsuchung und Festsetzung des obersten Prinzip der Moralität – are softer and less determinative than Gregor’s and Ellington’s translations suggest. The meaning of Aufsuchung is closer to ‘exploration’, which suggests that the supreme moral principle is something more like a territory that we have already found, though with features that still need to be mapped and clarified. (Aufsuchen means, among other things, ‘to visit’.) Festsetzung does mean ‘establishment’, but not in the sense of a proof (as in, ‘the mathematician established the theorem’), but rather more in the sense of a determination, as in the establishment or fixing of a price. So Kant is saying that the aim of the work is not so much to find and prove the supreme principle of morality but rather to explore and determine the nature of that principle.Footnote ¹¹ As the structure of the work then makes clear, Kant’s more specific version of this aim is to explore the nature of the principle that ordinary morality already takes to be supreme, and then to determine that it is a principle of pure practical reason.

In the first section of the Groundwork, Kant argues that ordinary uses of the word ‘good’ reserve a special, unconditional kind of praise for a particular kind of willing, for willing that aims to do what is morally required, merely because it is required. (Since my concern here is only with the structure and purpose of the work, detailed defences of these and other specific claims of the Groundwork are beyond the scope of this essay.) Though such a dutiful will is not itself part of right in the juridical sense – what is required by right is external conformity to law, not any particular reason for conforming to it – the idea of such a will is already implied by the idea of right in the juridical sense, because the moral content of right includes the very idea of external lawgiving as addressed to beings who morally ought to obey the law. As we have already seen, the idea of right itself implies the distinction between right and virtue. The content of virtue can and for Kant will extend well beyond ethical commitment to right; that is what the Doctrine of Virtue is meant to explain. But the content of virtue will include at least ethical commitment to right, since the idea of right itself implies that there is such a thing as ethical commitment, and that right is one of the things that one can and should be ethically committed to. In this sense, we can understand the ‘common human reason’ of Groundwork I as referring to the person who accepts the independent moral force of right, and who then takes this moral force to apply not just to juridical claims, but to all of their willing. Recall that the point of a Rawlsian comprehensive view is to relate political to non-political values. The representative agent in a Kantian liberal comprehensive doctrine is the person who takes liberal political values to have moral force and then understands the ethical realm to be exhaustively defined as just what is implied by the idea of a thoroughgoing ethical commitment to liberal political values, one that applies to all of life.

We can see this even more clearly once we understand that Kant’s specific derivation of the supreme moral principle itself depends on the basic arguments of the Doctrine of Right. At 4: 402, Kant famously argues that the moral law that applies to a dutiful and thus unconditionally good will can only be the principle of universal law:

Clearly, Kant is saying that every law has a formal, universal structure (‘Everyone must do X’) that is independent of its particular content (the action specified by ‘X’). The moral law that governs the dutiful will cannot have any particular content, because there is no particular reason for the dutiful will to do X rather than Y, given that the dutiful will is committed to obeying its special law merely because it is a law, not for any other reason. So Kant then concludes that the dutiful will can only be committed to the value of universal law, and thus to his supreme moral principle. But there is an obvious logical gap in this argument, one that will later be repeated in Groundwork II’s nearly parallel derivation of the principle of universal law from the bare idea of a categorical imperative (G, 4: 420–1).Footnote ¹² Universality is clearly a necessary condition of an unconditionally valid practical principle, but what Kant is saying is that because the bare ideas of a dutiful will and a categorical imperative imply no reference to any determinate content, we must also treat universality as a sufficient condition of practical justification. Universalizable maxims are really candidates for practical laws, not practical laws in themselves. Kant’s derivations are convincing arguments only if we add to the values of his representative agent an absolute preference for there being a law of willing over the absence of such a law – in other words, the basic moral commitment of the Doctrine of Right to prefer a juridical condition over the state of nature, now applied to the ethical and not just to the juridical realm. Kant says explicitly not that his supreme moral principle is a law, but that it must serve as one if there is going to be meaningful content to the idea of a dutiful will. In this reading, it is certainly not the case that the universal principle of right is derived from the categorical imperative (see also Ripstein Reference Ripstein2009: 11–13 and 355–88). Rather, it is that the categorical imperative expresses the idea of a thoroughgoing ethical commitment to the values implied in the universal principle of right (and only those values). Without a prior and independent commitment to right, treating the categorical imperative as the supreme principle of practical reason makes little sense.

Thus far, Kant’s comprehensive doctrine has explained how the specifically political value of right can fit into our lives, by our taking up an ethical stance that expresses this value in all of our willing. Does that mean that our lives are governed solely by this value? Certainly not, because the very idea of dutiful willing implies the commitment to moral values against possibly competing interests, which for human beings are all the heterogeneous desires that we gather together only under the indeterminately specified concept of happiness. So we still need to understand how to relate the values of duty and happiness in a single, rational conception that can govern our lives. This is what the argument of Groundwork II is intended to do, by developing a unified conception of rational agency that includes both hypothetical and categorical imperatives. It might seem that the turn to a theory of rational agency after Groundwork I is motivated by the thought that ‘common human reason’ is a mere artefact of socialisation, without any sort of rational justification. But what Kant explicitly says is that the motivation for the turn is not theoretical but practical; the demands of happiness are so pressing and fundamental that they constantly threaten the overriding claims of duty unless we can settle on a distinctive conception of rational agency that integrates both of them into our lives (G, 4: 405). Just as Rawls’ understanding of a comprehensive doctrine suggests, we need coherently to relate Kantian political and ethical values to other sorts of values.

The substantive argument of Groundwork II is that the common source of all value is the will, which Kant says is ‘nothing other than practical reason’ (G, 4: 412). Kant defines a practically rational being as one who ‘has the capacity to act according to the representation of laws’ and equates this capacity to having a will (G, 4: 412). To will something is to commit yourself to bringing about an end, which Groundwork I has already told us implies ‘the summoning of all means that are within our control’ (G, 4: 394). You are practically rational insofar as your actions make rational sense as means to ends you have chosen. In understanding your actions in this way, you are representing the rule-governed empirical connections between means and ends (A cannot happen in the world without B) as laws for you as an agent. We could call them laws for your will, but really Kant is saying that you cannot will anything without representing rules as laws for yourself, by committing yourself to projects which then have to be rationally carried out. To will an end is just to be rationally bound by the hypothetical imperatives it implies, which is why Kant can say that the concept of a hypothetical imperative is analytic of the concept of willing (G, 4: 417). The concept of a categorical imperative is not similarly analytic – you can clearly will ends or choose means to ends that violate the principle of universal law – but it does express the value of willing, in that obedience to this principle means counting everyone else’s willing of an end as equally important as your own. By willing only in ways that all others could also will, you express the value of willing itself, apart from the content of any specific end (G, 4: 428–9). So both hypothetical and categorical imperatives locate the source of value in the will, in our representing actions as laws for ourselves. When I act on a hypothetical imperative, I take my willing of an end to be sufficient to justify the value of my subsequent actions. When I act on the categorical imperative, I take my willing of an end to be sufficient to justify constraining my behaviour to give equal value to the potential willing of all others with the same capacity.

If both hypothetical and categorical imperatives have a common source of value, then there really is no inherent conflict between our commitment to duty and the claims that happiness makes on us. Or, in the more Rawlsian language that we have been using, there is really no inherent conflict between our ethical commitment to politically liberal values and our non-political interests in other sorts of values. If willing is the source of value, then both dutiful and happiness-seeking actions can be understood as part of a common conception of ourselves as rational agents. When I pursue my own happiness, I act to realise ends I have set for myself. When I act out of duty, I act to respect the capacity of myself and all others to set ends for themselves. There can of course be conflicts between duty and happiness, because my willing of an end might not be consistent with the capacity of everyone else to equally will that end. But we can also clearly understand why it makes rational sense that such conflicts should always be resolved in favour of duty, because if the source of the value of my happiness is my willing something as an agent, then it makes rational sense for me to equally value the agency of all others, and thus to subject my willing to moral constraints.

With this account in place, the question that remains for Kant is the theoretical issue we have already noted that all comprehensive doctrines must face, which is to give a rational defence of their general account of value. In what sense can we justify the claim that the will is the source of value, especially when this thesis, on its own, sounds dangerously voluntarist and arbitrary? The most familiar contemporary answer to this question is the constitutivist argument most notably developed by Christine Korsgaard (Reference Korsgaard1996, Reference Korsgaard2008, Reference Korsgaard2009): the conclusion that the will is the source of value is supposed to follow from our identities as rational agents. On this view, a basic rational identity is built into the deliberative structure of our agency, and we cannot consistently identify with the projects that result from our deliberations unless we also recognise a deeper identification with the basic deliberative identity that enables the choice of our particular projects. Beyond the many objections that have been levelled at the logic of this argument, it seems that Kant himself is rejecting it at the start of Groundwork III. He does argue that we must understand ourselves as agents who make rational choices for ourselves (G, 4: 448), but he immediately goes on to ask how and why we take an interest in this capacity (G, 4: 448–9), which suggests that he does not believe that our having to deliberate itself implies that we must identify with or value our deliberative nature. This thought is also supported by Kant’s various reflections on our having to decide what will make us happy, in which he consistently tells us that we might prefer the simple and clear directives of instinct to the burdens of deliberation (G, 4: 395–6; CB, 8: 111–5). It is in morality itself, Kant tells us after Groundwork III’s opening reflections on our rational agency, that we take an interest in the freedom that is the presupposition of deliberation (G, 4: 450).

This is not the place to enter into the many controversies over the text of Groundwork III and its relation to the argument of the Critique of Practical Reason. I take the two texts to be basically consistent, and recent scholarship to have more effectively supported this conclusion,Footnote ¹³ but here it is enough to note that even readings that find divergence typically hold that the claims of the second Critique are Kant’s more considered answer to the justificatory question he is raising in Groundwork III. That answer turns on Kant’s assertion of a fact of reason (CPrR, 5: 29–31), which cannot be given any further justification, and which is supposed to be revealed by one’s first-personal response to an example Kant proposes. I take it to be especially telling for my argument that this example is specifically juridical: a sovereign demands, under threat of death, that you make a false deposition against one of the sovereign’s political enemies.Footnote ¹⁴ Structurally, it would seem that Kant could have chosen any example in which a person took themselves to be morally bound to sacrifice their own life to avoid killing another, thereby violating a duty of both right and virtue; see, for instance, Kant’s example of the shipwreck at DR, 6: 235. But in the example Kant chooses here, the sovereign is demanding not just that you act wrongly, but to pervert the very structure of right. (Compare Kant’s moral horror at the formal execution of a monarch at DR, 6: 320–1.) The sovereign is using the juridical procedures of right to carry out an extra-juridical killing, by issuing an extra-juridical threat (not a legal command) to you to wrongly provide pseudo-juridical support for his act of wrong. If you resist this demand, this can only be explained by your ethical commitment to the moral ideals implied in right. In accepting that you should, and also can, sacrifice your life for this ethical commitment, even against your strongest physical desire, you take your ethical commitment to right over the claims of happiness to be self-justifying. In what sense can an ethical commitment to right be self-justifying? It simply means understanding juridical standards as laws that we have willingly, meaning freely and rationally, imposed on ourselves, and taking this conception of self-legislation to be theoretically sufficient for purposes of our practical commitments. All of the metaphysical work of Kant’s critical philosophy then plays the merely defensive role of resisting any theoretical challenge to this understanding of right, by showing that it is not inconsistent with or undermined by anything implied by or found in an empirical or scientific perspective on the world.

What, then, is the purpose of a groundwork for a metaphysics of morals? It is not to show that the moral ideal of right is true, or even that the categorical imperative is true so that the moral ideal of right can follow from that. The purpose of the work is rather to show that accepting the moral force of right makes rational sense for living our lives as individuals, without presupposing any other ethical ideals beyond those that are implied in the idea of right itself. That is, the purpose of the work is to affirm a comprehensive liberal doctrine. On this Kantian comprehensive doctrine, we take a form of political liberalism to have moral force and then explain that an ethical commitment to this political liberalism itself implies a conception of value that can explain and order all our actions under a unified conception of practical reasoning. This conception asserts that the source of all value is the will, a theoretical claim that Kant then defends by arguing that is rationally coherent to understand politically liberal ideals as something we have willed, as laws we have imposed on ourselves. This comprehensive moral argument does not rely on any positive theoretical claims beyond what is implied in an ethical commitment to political liberalism itself. That is why Kant’s comprehensive view is a comprehensive liberalism: it is a comprehensive view that justifies itself in a politically liberal way.