The Ambitions of Kantian Political Philosophy

I. Introduction

What do contemporary political philosophers stand to learn from Kant’s Doctrine of Right? For a long time, the answer to that question was thought to be very little. Everyone knew that some of our most influential political thinkers—including John Rawls, Ronald Dworkin, and Thomas Nagel—were Kantians, broadly speaking. But their theories owed more to the Groundwork than to Kant’s properly political writings. They were uneasy about aggregation, they took rights seriously, they viewed autonomy as paramount. But the ideas on which the main argument of the Doctrine of Right turns—the fundamental justificatory role of external freedom, say, or the provisional status of rights in the state of nature—were conspicuously absent from their works.

A different outlook emerged over the past two decades. Philosophers began exploring the details of Kant’s political writings and asking what lessons these might hold. They mined the texts for insights on topics ranging from territorial rights and colonialism to poverty, sexual orientation, and abortion.¹ Some even suggested that a complete political philosophy based entirely on the demands of freedom could be unearthed from the Doctrine of Right—one that, suitably worked out, could rival our most sophisticated contemporary theories. What had long been dismissed as the work of a great mind in decline appeared poised to shake political philosophy out of its post-Rawlsian torpor. For a moment, it looked like the answer to our opening question might turn out to be everything.

The high-water mark of this reassessment of Kant’s contribution to political philosophy was Arthur Ripstein’s Force and Freedom.² More than any other work, it succeeded in providing an interpretation of Kant’s starting point, the innate right to freedom, that many found intuitively compelling. It presented powerful reconstructions of Kant’s key arguments—concerning the necessity of property and contract rights, for instance, and the requirement to enter the civil condition. And it conveyed a sense of how a whole theory of justice might follow from Kant’s sparse premises, accounting for the state’s duty to support the poor, say, or for the demands of equality of opportunity. There was even the hint that Kantian principles should displace the “more robust ideas of material equality”³ embraced by many contemporary thinkers. After a good dose of Ripstein’s arguments, you might be tempted to think not only that some conclusions about distributive justice follow from Kant’s premises, but that all principles of justice should be derived in this way. Kantian political philosophy seemed as bold and ambitious a research program as political philosophy had seen in a generation.

Yet many remained skeptical. Could all matters of justice really be decided in terms of the demands of freedom? Must values such as welfare, equality, and fairness be set aside except insofar as they are part of the ideal of universal freedom? And could such an austere approach yield conclusions that liberal egalitarians would find palatable? No doubt some minimal conception of distributive justice can be derived from Kant’s premises, but won’t it be too minimal? For all the initial excitement, isn’t the Kantian framework too conservative to be taken seriously?⁴

In what follows, I argue that the standoff between Kantian political philosophers and their critics rests on a mistaken understanding of Kant’s contribution to political thought. Both sides accept what I’ll call an ambitious construal of this contribution, according to which Kant’s framework aims to answer all the main questions of political philosophy—including, crucially, the various questions about economic justice that animate the post-Rawlsian literature. I believe that this is a mistake. Throughout the Doctrine of Right, Kant’s focus is on a specific subset of political questions about what grounds the state’s exclusive right to make and enforce laws—questions that fall under the broad heading of legitimacy.⁵ Questions of substantive justice—concerning what the state should ideally do, how it should administer the economy, and so on—occupy at most a peripheral place in his thinking. Kant’s freedom-based outlook has strengths and limitations that align with this focus: it has much to contribute on questions of legitimacy, but it provides limited guidance when it comes to tackling issues of substantive justice. In keeping with this diagnosis, I want to defend a modest construal of the Kantian position, one that takes Kant’s argument for the state seriously while recognizing the need to appeal to considerations that go beyond his framework in determining what constitutes an ideally just state.⁶ In short, the Kantian view should form an important part of our overall political theory, but it cannot answer all our questions on its own.

I begin in Section II by introducing the ambitious construal of the Kantian position that Ripstein and others have defended, and I rehearse some grounds for doubting its viability. I then present the case for a more modest construal. In Sections III and IV, I articulate what I take to constitute the heart of Kant’s political thought: the argument that enforcing rights consistently with everyone’s right to freedom is possible only in a civil condition—that is, only when a state is in place. The argument turns on the idea that attempts to enforce rights in the state of nature give rise to what we might call a wrong kind of agent problem, because no agent in the state of nature has the agential structure required to enforce rights in a way that addresses everyone’s potential objections. Only by putting in place an agent whose structure is tailored to address these objections is it possible for rights enforcement to be compatible with the demands of freedom—and putting in place such an agent just amounts to establishing a state. With this argument in hand, I then ask, in Sections V and VI, whether Kant’s conception of the enforcement problem and of its solution commits us to a specific theory of substantive justice. I argue that it does not: Kant’s framework is more open-ended than has commonly been assumed. The different theories of justice developed by contemporary philosophers are therefore not necessarily in competition with Kant’s outlook; on the modest construal I advocate, they may instead turn out to be possible complements to it.

II. The Ambitious Construal

The reasoning that motivates the ambitious construal of Kant’s political philosophy is straightforward. It starts with the idea on which the whole Doctrine of Right rests: that the only right human beings have innately, solely by virtue of their rational nature, is the right to freedom—that is, the right to “independence from being constrained by another’s choice . . . insofar as it can coexist with the freedom of every other in accordance with a universal law.”⁷ Basically, each person’s external freedom—her independence from the choices of others—is protected so long as it remains consistent with everyone else’s.⁸ This has important implications for the justification of coercion. Since coercion is a restriction of freedom, it can only be justified to prevent violations of the right to freedom: coercion must be, in Kant’s words, the “hindering of a hindrance to freedom.”⁹ Insofar as rights are coercively enforceable, it would appear to follow that all rights—the right to a fair trial, the right to vote, the right to raise one’s own children, the right to social assistance, the right to equal treatment—must ultimately stem from this more basic right to freedom. From there, it’s a short step to the conclusion that our entire political outlook must be answerable to the right to freedom.

A lot turns on what answerability to the right to freedom amounts to exactly. Ripstein doesn’t shy away from a strong reading of the idea. In the opening pages of Force and Freedom, he explicitly writes that on the Kantian view freedom “carries the justificatory burden of the entire argument, from the prohibition of personal injury, through the minutiae of property and contract law, on to the details of the constitutional separation of powers”¹⁰—and on to distributive justice and criminal punishment, based on what he argues later in the book. This is very much the thought fueling the ambitious construal: freedom is to do all the justificatory work for political philosophy, determining not only what counts as a minimally legitimate state but also what counts as a fully just one. Freedom-based arguments may leave certain questions unresolved, of course; as we will see, Ripstein stresses that there remains considerable room for democratic decision-making and for the exercise of judgment by state officials. But the key idea still stands: any conclusion about what the state must do—about what is beyond the discretion that can rightfully be exercised by officials and the electorate—has to be grounded in the demands of freedom.

This looks like a recipe for a highly revisionist outlook. Yet proponents of the ambitious construal tend to be remarkably confident that their conclusions will remain within the mainstream of contemporary political thought. Ripstein is especially sanguine on this point. He admits that the requirement that the state’s activities be justified in terms of equal freedom places limits on what it may do, but he insists that the Kantian outlook “does not preclude most of the familiar activities of modern states.”¹¹ On the contrary, he argues, it strongly supports the key tenets of liberal egalitarianism, including poverty relief, universal health care, and equality of opportunity.¹²

One could debate whether the ambitious construal is accurate as a textual interpretation of Kant’s position, but that is not my present concern.¹³ As I said at the outset, I want to ask whether contemporary political philosophers should take the Kantian outlook seriously. In the remainder of this section, I argue that if we adopt the ambitious construal, we should answer this question in the negative: so interpreted, the Kantian position runs into too many difficulties to be a serious contender. I will explain why the modest construal that I favor deserves to be taken more seriously in later sections.

The ambitious construal gives rise to both substantive and methodological worries. Critics of the Kantian outlook have mainly focused on the former, expressing deep skepticism about whether conclusions that are not implausibly revisionist can be extracted from such a meager starting point as the right to freedom. I want to mention two specific worries arising on this front, both of which have received a characteristically vigorous articulation in Victor Tadros’s critical discussion of Force and Freedom.¹⁴ I will then consider some noteworthy methodological concerns.

A first substantive worry concerns the level of economic redistribution that the Kantian outlook can justify. Ripstein maintains that provision for the needy is required on the Kantian view, because the poor would otherwise be left to depend on charitable donations for their survival, in violation of the Kantian ideal of independence. More specifically, he argues that a state failing to provide assistance for the poor can’t be the object of a united will, since it leaves them “completely subject to the choice of those in more fortunate circumstances.”¹⁵ He even compares a beggar who relies on the kindness of others to meet his “most necessary natural needs” to a slave—the paradigmatic example of a person who is wholly dependent on the will of another.¹⁶ This certainly identifies one important respect in which deep poverty is troubling. As Tadros notes, however, what the argument implies for less extreme cases is far from obvious. Once minimal measures of poverty relief are in place—once everyone’s most basic needs have been satisfied—aren’t concerns about complete subjection to the choices of others put to rest? And if so, doesn’t the view fall well short of the principles of distributive justice that liberal egalitarians typically advocate?

A similar worry arises about publicly funded health care.¹⁷ The ideal of freedom as independence plausibly demands some minimal provision of health care. If you fall seriously ill, you may become unable to provide for yourself. And if your very life depends on obtaining care you can’t afford, you will become dependent on the goodwill of the more fortunate. Ripstein concludes somewhat hastily that “a state can act proactively to provide publicly funded universal health care.”¹⁸ But here too, as Tadros stresses, there are questions about how far the argument can take us. Can it justify anything like the extensive universal health insurance coverage that liberals generally favor (and that actually exists in many countries)? Why not restrict coverage to major illness if the goal is just to prevent complete dependence on others? Once again, if we stick to the logic of the Kantian framework, we risk ending up with uncomfortably minimal conclusions.

Proponents of the ambitious construal might retort that Kant’s theoretical framework was never intended to be answerable to the sort of commonsense claims that Tadros invokes.¹⁹ But this seems forced. The concerns that Tadros raises rest on claims that very few would reject. Besides, if the Kantian framework doesn’t support decently robust conclusions about distributive justice, then it can hardly provide a compelling alternative to theories such as Rawls’s or Dworkin’s—which is what’s at stake here. And note that it won’t do to insist that the Kantian ideal leaves room for ample discretion regarding its precise realization (a theme running throughout Force and Freedom to which I return below). It’s true that, as Ripstein emphasizes, state officials (or the electorate) may decide that an extensive universal health insurance scheme should be implemented because this is the most reliable way to ensure that all citizens retain their independence.²⁰ Yet that doesn’t dispel the central worry. On this interpretation of the Kantian position, the demands of justice still run out implausibly quickly; it’s just that those in charge may elect to be more generous than justice requires.

The substantive worries are quite serious, then. The methodological concerns I now want to raise have received less attention, but if anything they strike me as cutting deeper still. They pertain to the two fundamental ideas that Kantians tend to invoke to extract conclusions about distributive justice from the ideal of freedom as independence: that of a united (or general, or omnilateral) will, and that of the original contract.²¹ These are highly abstract ideas; Kant is not particularly explicit about the relation between them, but it is natural to interpret the first as referring to the minimal conditions for the existence of a civil condition, and the second as representing the ideal toward which a civil condition should be steered.²² The picture is then the following: you and I are in a civil condition only if we are ruled by an agent whose decisions represent neither my nor your particular will, but rather our united will. And if we are in a civil condition, then we—citizens, representatives, and state officials in general—have a duty to improve it by bringing it more in line with the idea of the original contract.

Based on this characterization of the two ideas, you would expect ambitious Kantians to rely heavily on the idea of the original contract: if anything will deliver full-blown principles of justice, it should be the ideal in light of which the civil condition must perfect itself. Strikingly, however, it is much easier to find arguments proceeding in terms of what could be the object of a united will. Take Ripstein’s argument for the duty to support the poor. He writes that the absence of appropriate welfare measures is incompatible with the idea of the original contract, but he goes on to specify that this is so because this lack of protection for the poor couldn’t be the object of a united will.²³ The argument is certainly valid: insofar as the idea of the original contract isn’t separate from that of a united will but rather something like its most fully realized version, showing that a measure can’t be the object of a united will plainly suffices to establish its incompatibility with the idea of the original contract.²⁴ But since the latter is the more demanding ideal, you would expect more ambitious conclusions to be derived directly from that idea, not from claims about what is incompatible with the comparatively minimal idea of a united will.

This leads to a deeper worry. It’s not just that, because the idea of the original contract is elusive, ambitious Kantians appeal to the wrong idea for their purposes. At a more basic level, we should question our ability to derive determinate conclusions from either fundamental idea: for all their pedigree, they are both highly imprecise—and very far from the level of detail at which contemporary theories of distributive justice aim to operate. A quick comparison with Rawls’s framework should make this evident. The two fundamental Kantian ideas are reminiscent of Rawls’s original position, which also conveys an ideal of hypothetical agreement. But remember how Rawls characterizes his choice situation: he tells us what information the parties in the original position have access to; he provides a conception of the considerations in terms of which they are to decide; and he specifies what forms of reasoning they may use.²⁵ There is simply no comparable structure to the idea of forming a united will, or of being party to the original contract.²⁶ This should make us doubtful of attempts to derive conclusions about distributive justice from these ideas: their imprecision makes it too easy to invoke them in support of whatever conclusions we are inclined to endorse. In a context where issues of distributive justice are the subject of deep controversy—as they are in contemporary democratic societies—the fundamental Kantian ideas simply cannot provide the kind of guidance we need.²⁷

The worry may seem at risk of overshooting. If the fundamental ideas on which Kant relies are imprecise, doesn’t that undermine not just the ambitious construal but the whole outlook? Not necessarily. For reasons I outline below (in Section IV), I believe that Kant’s approach remains well suited to addressing the questions for which it was intended: questions about state legitimacy and about the justification of basic rights (to property or contract, say). It is therefore important to establish whether we can accept this core theory without committing ourselves to the ambitious construal—that is, whether it’s possible for Kant’s theory to form just one part of our political philosophy, requiring supplementation when questions outside its main purview come into play.

To make progress on this front, we need a more detailed articulation of the core Kantian theory. This requires unpacking the central argument of the Doctrine of Right, a task to which I devote the next two sections. That will set the stage for the claim I want to defend in the remainder of the paper, namely, that Kant’s position allows for more flexibility in our approach to questions of distributive justice than has generally been allowed.

III. Kant’s Main Argument, I: The Enforcement Problem

I take the heart of Kant’s political philosophy to be his diagnosis of a problem that arises when individuals claim rights against one another in the state of nature, along with his articulation of the shape that a solution to that problem must take. The general thought is clear enough: Kant holds that what he calls “acquired rights” (including property rights and contract rights) are “provisional” in the state of nature, and that the state is needed to make these rights “conclusive.”²⁸ Since he thinks that individuals living together undermine each other’s freedom unless they have conclusive rights against one another, it follows that the state is required for freedom.

There are two crucial ideas here: first, that rights are necessarily provisional in the state of nature; second, that conclusive rights, which only the state can provide, are required for freedom. For present purposes, I propose that we simply grant the connection that Kant draws between conclusive rights and freedom, since the details of the argument for that thesis do not bear on our question.²⁹ By contrast, a lot turns on how exactly we understand the idea that rights are provisional in the state of nature. Let me say more about how I interpret this aspect of Kant’s position.³⁰

Why are acquired rights merely provisional in the state of nature? My view is that the fundamental problem concerns the enforcement of rights, which is necessarily defective outside the civil condition. In the state of nature, I can use force to back what I take to be my rights, but I can’t do this in a manner that is compatible with everyone’s right to freedom, and hence rightful in Kant’s sense. It is only when we leave the state of nature and enter the civil condition that rights enforcement can be rightful. This gives rise to two questions: why can’t the enforcement of rights in the state of nature be rightful? And how does the state solve the problem?

Imagine you and I are on a desert island together. I have picked a coconut from a tall tree in which, let’s assume, there are many left. I place it down for a moment. You quickly grab it and hold on to it, refusing to give it back. I argue with you: “I climbed all the way up there to get this coconut! You can do the same if you want one!” You remain unmoved. Convinced that the coconut is rightfully mine, I wrestle it from you. What should we say of this attempt at enforcing a property right in the state of nature? Is it rightful? If not, why not?

On one highly influential conception of property rights, usually traced back to John Locke’s Second Treatise, my action is entirely unobjectionable. By picking the coconut, I have modified it in one crucial respect: I have made it conveniently located. In doing so, I have invested my labor in the coconut (or “mixed” my labor with it, as Locke sometimes puts it).³¹ That suffices to turn it into my property, so long as I don’t make you worse off by appropriating this particular bit of nature—which I don’t in our example, as there are plenty of other coconuts to go around.³² The result is that I have a full-fledged property right in the coconut, a right that, like anyone else in the state of nature, I have the executive authority to enforce against those threatening to violate it.³³ On Locke’s view, it bears stressing, there is nothing problematic about this authority. This doesn’t mean that my having the authority is obvious (he concedes that “this will seem a very strange doctrine to some”³⁴). But there is no respect in which my action is anything less than ideal.

So the picture is the following. If you and I agree about the extent of our respective rights, then all is well and good. If we don’t agree—if we’re unable to decide together, and hence one must decide for both—then the authority to settle the matter belongs to whichever one of us is in the right. It is entirely legitimate for that person to stand up for her rights; she gets deputized by morality, so to speak.³⁵ On this view, if I use force against you to reclaim my coconut, and if I do so proportionately, then everything is as it should be (except that you shouldn’t have taken the coconut in the first place). Certainly, you have no valid complaint against my use of force. Locke grants that allowing individuals to enforce their own rights involves certain risks: they are likely to be ignorant and biased, and accordingly to take more than they should, use excessive force, and so on.³⁶ But these are merely practical problems. The crucial point is that when a person gets it right—when she enforces a right she actually has in a proportionate manner—then the fact that the enforcement takes place in the state of nature creates no legitimacy deficit whatsoever.³⁷

There is a simple and seductive thought at work here: if I truly have a right to an object, and if I use force proportionately to enforce this right, then there is nothing to criticize about my action. This is the reasoning that anchors Locke’s conception of natural rights. Importantly, it does not depend on his specific theory of property or on the role that labor plays in making acquisition possible in the state of nature: the claim concerns the conditions under which the enforcement of rights can be fully legitimate, not the specific content of rights or how they can be acquired. It can be taken on board by people whose views about rights are otherwise very different from Locke’s. Indeed, despite Locke’s worry that this might seem “a very strange doctrine,” my sense is that it is accepted by most contemporary political philosophers.

This apparently innocuous aspect of the Lockean position is what Kant opposes at the deepest level. To begin with a very broad characterization of his objection, you might say that he takes the relational character of rights more seriously than Locke does. Of course, Locke sees rights as relational in one sense: ultimately, they concern how we can rightfully treat each other. But consider how he would justify my use of force against you in our scenario. Basically, he would point to my relation to a certain coconut (which I picked and you didn’t) and to your relation to other nearby coconuts (which are still there for the picking). How you and I can justifiably relate to each other is a function of how each of us relates to other stuff.

This gives rise to a puzzle. How do we get from claims about relations that obtain between each of us and external objects (about my having picked a coconut, and about there being plenty of coconuts to go around) to a claim about my having the authority to use force against you, that is, to a claim about how I can (forcibly) relate to you? Locke seems to think that no further argument is required: there is no gap between my having a claim to an object and my having the authority to use force against you. On Kant’s view, by contrast, there is a crucial gap. Whether I have the authority to use force against you doesn’t depend only on whether I have a claim to a certain object that you lack; it also depends on whether my having the authority to enforce rights against you can be justified from your point of view. To put the point in more contemporary terms, it depends on what objections you have to my using force against you, and on how I can respond to those objections. Kant’s view is that so long as we remain in a state of nature, you have objections that can’t be fully answered. Short of entering the civil condition, there is no closing the gap.

What objections do you have to my using force against you in our desert-island scenario? What could be problematic about my using force to prevent you from taking what is surely my coconut? And why doesn’t it suffice to say, with Locke, that I don’t claim any special standing for myself, since you are entitled to do the same when your rights are at stake? Kant’s key thought is contained in a passage that strikes a somber, almost Hobbesian note:

[B]efore a public lawful condition is established individual human beings, peoples, and states can never be secure against violence from one another, since each has its own right to do what seems right and good to it and not to be dependent upon another’s opinion about this.³⁸

This is the heart of the enforcement problem. The point is not that everyone has a right to everything in the state of nature; not everything will seem “right and good” to you if you reason in good faith.³⁹ It’s that as between you and any other individual in the state of nature, you are entitled to follow your own judgment. You don’t have to give in to what another person thinks should be done. And, crucially, you can reject her alleged authority to decide for both of you.

Consider again our desert-island scenario. Suppose you ask me, as I am wrestling the coconut away from you, “What gives you the right to do this? Why should you get to determine who will end up with the coconut?” What can I reply? At the very least, I owe you an answer that isn’t just a restatement of my reasons for thinking that the coconut is mine—an answer that goes beyond the reasons that didn’t convince you the first time around. Arguing that morality (or Locke’s theory) is on my side won’t do in this regard: I can’t simply presume that I am entitled to decide the matter for both of us, based on my judgment of what morality requires. You and I disagree about what we are respectively entitled to do. If I am to do right by you as a self-governing being, I can’t just brush that disagreement under the rug.

It may seem odd to speak specifically of doing right by you as a self-governing being, as that idea also features prominently in Locke’s outlook. In his influential book on Locke’s theory of rights, A. John Simmons articulates the conception of natural freedom that undergirds Locke’s approach precisely in terms of what he calls a “right of self-government.”⁴⁰ But although Locke and Kant both start from a right of self-government, they construe it in strikingly different ways. On Locke’s view, as Simmons reconstructs it, we respect this right so long as we recognize that “[w]e all have . . . an equal right to freedom from harm and interference by others, provided only that our actions stay within the bounds of natural law.”⁴¹ In short, you must respect my right of self-government so long as I act in accordance with everyone’s rights. The moment I exceed the bounds of natural law, however, you acquire the authority to step in and stop me.

Kant favors a more robust understanding of what it means to respect a person’s standing as a self-governing being—or, in Ripstein’s terms, a person’s “entitlement to be his or her own master.”⁴² Doing so requires not just taking seriously that, by default, a person should decide for herself, but also recognizing that she has her own point of view on questions of right. It doesn’t follow that she is entitled to violate other people’s rights; Locke is correct on that count. But she can question the authority of any other person’s judgment about such matters. If I judge that you have exceeded the bounds of the law of nature and you disagree, you can ask why my point of view should prevail over yours. After all, we each have our point of view on the issue, and neither can persuade the other. On Kant’s view, neither of us can break this stalemate, because neither of us can use force in a way that addresses the objections arising from the other’s point of view. That is why our desert-island scenario raises a problem of legitimacy, a problem not about who should have the coconut—about the merits of our respective cases—but about who is entitled to settle our disagreement about who should have the coconut.

The Lockean may seem to have an obvious rejoinder: why should your mistaken beliefs about who owns the coconut restrict my claim to it?⁴³ However, the issue is not my claim per se but rather what I may do when it is contested. Perhaps I am right and you are mistaken about who has the better claim to the coconut; the problem is that this is not the right kind of fact to answer the objections you have against my use of force. There is a parallel here to an important point that Rawls makes to motivate his turn to political liberalism. He writes:

[T]hose who insist, when fundamental political questions are at stake, on what they take as true but others do not, seem to others simply to insist on their own beliefs when they have the political power to do so. Of course, those who do insist on their beliefs also insist that their beliefs alone are true: they impose their beliefs because, they say, their beliefs are true and not because they are their beliefs. But this is a claim that all equally could make; it is also a claim that cannot be made good by anyone to citizens generally.⁴⁴

Rawls is concerned in this passage with the wielding of political power in a civil condition, not with a state-of-nature scenario such as ours. But his claim about the limits of relying on the truth as one sees it in good faith in the context of political justification carries over. If I insist on imposing my view on you, and if I don’t have anything better to say than that I am correct (along with arguments you don’t find convincing for why I am correct), then from your point of view I am simply insisting on my own beliefs when I have the physical power to do so. You can complain that I am not doing enough to address your objections.

What we should make of all this plainly depends on what would count as doing enough. If Kant’s argument merely highlights inescapable aspects of the human condition, then it’s not clear how much we need to worry about it.⁴⁵ We must therefore ask whether it is possible to address the shortcomings of private rights enforcement. If it is, then we can do better by each other, which supports the claim that you have a valid complaint when we don’t. By contrast, if there is no feasible way to address your objections, then that casts doubt on whether they need to be taken seriously in the first place. So let us consider what a solution to Kant’s problem would look like.

IV. Kant’s Main Argument, II: Solving the Enforcement Problem

How can we resolve our disagreement in a manner that respects our equal standing as self-governing beings? The Kantian answer starts from a simple idea: to solve our problem, we need to bring into play an agent whose status differs from ours—an agent that is not just like you and me. As I see it, that is the whole point of the section of the Doctrine of Right in which Kant argues for the necessity of transitioning from a state of nature to a rightful condition—or, as he sometimes puts it, for the necessity of putting in place a properly public authority.⁴⁶

What does bringing into play an agent whose status differs from ours entail exactly? And where are we supposed to find such an agent? The short answer is, we need to invent one. We have to create an agent whose structure marks it out as the appropriate instance to do the deciding for both of us: an agent constituted to address the problem we encounter in the state of nature. The guiding thought is simple, at least in outline. You have objections to my using force against you—to my deciding who will have the coconut—precisely because I am not the right kind of agent to decide for both of us. I am just a person roaming about the state of nature, like you; there is nothing about me, about the way I am constituted as an agent, that makes it appropriate for me to take it upon myself forcibly to settle our dispute. If you ask why my judgment, rather than yours, should determine who gets the coconut, I have no compelling answer to give. I can explain why I think the coconut should be mine—why I believe my judgment to be correct. But that doesn’t answer your query. I’m telling you why I think I’m right about the substantive issue; what you want to know is why my judgment should prevail over yours given that we disagree about the substantive issue.

It may be tempting to think that I can break the stalemate by showing that I’m a more reliable judge of matters of right than you are. But that won’t do. For one thing, the move can’t help in a one-off interaction: what evidence could I possibly be going on? And even if it can be established that I am more reliable in general, if you believe that I am wrong in this specific instance, and if you take yourself to be sufficiently justified in that belief to discount my general reliability, then it remains unclear, from your point of view, why my judgment should prevail over yours.⁴⁷ You can still complain that not enough has been done to address your point of view in light of our disagreement.

Appealing to the considerations that legitimize enforcing rights on a Lockean view won’t help either. The key elements of Locke’s solution—the absence of hierarchy in the state of nature, the fact that all are equal in their right to use force against each other—are precisely the source of our problem. There is nothing that sets me apart from you, that grants me a different status, and hence nothing that makes me the right kind of agent to decide for both of us. That’s why I have nothing compelling to say when you object to my doing so. Only by bringing a different kind of agent into play, one structured so that its claim to decide—and to resolve our disagreement by force if necessary—is not on a par with ours, do we break the symmetry, and thereby the stalemate.

Before I say more about what this means concretely, let me explain briefly how my suggestion connects with Kant’s argument. Kant doesn’t speak of the need for a different kind of agent per se, but he does insist on the need for a different kind of will. This is how the idea of an omnilateral will enters the picture.⁴⁸ I view the difference as largely terminological: where I speak of an agent with a certain structure, you could just as well speak of an agent whose will has a certain structure; it just seems a more roundabout way of referring to the same thing. If that is correct—if the two ideas are essentially equivalent—then we can see why the methodological issues I raised above don’t apply to Kant’s main argument. As I reconstructed it, the argument doesn’t require that we derive conclusions by unpacking the idea of an omnilateral will. What does the heavy lifting is our conception of the problem that arises in the state of nature. The idea of a different kind of agent gets introduced as shorthand for the kind of agent that would solve that problem. In the same way, if we want to use the phrase “omnilateral will” responsibly, then we should treat it as a way to refer to the solution to our problem—that is, we should view it as referring to the best feasible solution to the enforcement problem, whatever that turns out to be.⁴⁹

On this construal, the worry that the central notions that Kant deploys are too imprecise to do any work doesn’t arise, since our starting point is not the idea of an omnilateral will (or of the original contract) but rather a conception of the problem that we face in the state of nature. Our aim is to articulate that problem as precisely as possible and to figure out how it can be solved. We give content to the idea of an omnilateral will simply by attending to these questions.⁵⁰ This explains why Kant’s theory provides guidance about questions of legitimacy in a way it doesn’t for broader questions of distributive justice. When we articulate the Kantian theory of legitimacy, we fill out the idea of an omnilateral will from within, in light of our conception of the problem to which the state provides the solution. We can’t do the same to develop a theory of distributive justice: that concerns what the state should ideally do once it exists as a legitimate entity, and thus once the problem we face in the state of nature has been solved. Tackling that question requires more than filling out the idea of an omnilateral will from within.

Let’s return to our main thread. What would an agent whose structure is tailored to solve the enforcement problem look like? For Kant’s answer to that question, we must turn to his theory of the state. He doesn’t go into as much detail on this point as one might have wished, but what he tells us fits nicely with my framing of his outlook. He paints an overall picture of the state structured around key elements of a solution to the enforcement problem. In that sense, his theory of the state is just what we should expect it to be: an outline of the kind of complex institutional structure that is required to solve the problem arising in the state of nature.

His guiding thought follows straightforwardly from the nature of the enforcement problem: rights enforcement shouldn’t be objectionably unilateral. It shouldn’t be the case that one person gets to decide for others; rather, in keeping with the well-known Rousseauian ideal, all should decide for all—or, as Ripstein sometimes puts it, the people should decide for itself through its institutions.⁵¹ This is the general idea that must be filled out from within. Kant does so by appealing to a cluster of familiar notions: the rule of law⁵²; what he calls “republican” government (including the requirement that the three branches of power—the legislative, the executive, and the judiciary—be kept suitably distinct⁵³); and the credible commitment, backed by force, to protect the basic rights of all who share a territory. Given the potentially radical implications of Rousseau’s ideal, these may seem like surprisingly pedestrian ideas to invoke in this context. But if we understand Kant’s characterization of the state as aiming to answer the different complaints that rational agents have against the private use of force, we see that he is headed in the right direction. Let me explain.

The most obvious complaint you can raise in our desert-island scenario is that when I wrestle the coconut away from you, my use of force is not principled—it is just what I want to do at that moment. Kant invokes two ideas that can be seen as jointly addressing this complaint: the rule of law, and his conception of republican government. An institutional structure conforming to the ideal of the rule of law is organized to ensure that any use of force will be governed by laws that apply to all, not by individual fiat. And the ideal of republican government demands that the laws themselves not be just a reflection of the whims of particular individuals. If our institutions live up to these two ideals, their use of force will be principled in the relevant sense. How these ideals can be realized concretely is a difficult question in institutional design, of course—and one that can be answered in different ways. But for our purposes, the key lesson is that they can be seen as addressing a valid complaint that individuals have against the private use of force.

A further objection you could present is that, from your point of view, my action is indistinguishable from a use of force that simply aims to advance my own personal interests. I claim to be enforcing my rights, but you reject that claim: as you see it, I am just imposing my beliefs on you. To respond to this objection, we need an institutional structure that credibly embodies a commitment to the aim of enforcing rights, as opposed to one person’s private interest—and not just my rights or your rights, but the rights of all who share a given territory and whose rights are therefore co-dependent. This requirement goes beyond the rule of law as traditionally conceived, since it rests on a substantive conception of the basic rights that individuals have. The exact content of the Kantian view will crucially depend on what specific rights are included. Property and contract rights, along with the right to one’s body, form the core of the outlook, but, as Ripstein argues, further rights are plausibly required for citizens to be independent from one another, including the right to basic sustenance and to essential health care.⁵⁴ For our purposes, the details of the arguments for specific rights matter less than their general form: in each case, the aim is to show that certain rights must be protected for individuals to live together without undermining one another’s right to freedom.⁵⁵ This may seem likely to yield an implausibly limited list of rights, but recall that, on the present interpretation, the outcome of the argument isn’t meant to be the full complement of rights that a state must recognize to be just—only the rights that it must protect to be minimally legitimate, since that is the question that the Kantian view allows us to answer.

A third, more Hobbesian complaint you can voice is that the private use of force leaves the question whether your (or anyone’s) rights will be respected objectionably dependent on the actual balance of power. As between you and me, maybe it’s obvious that I will always succeed in wrestling coconuts from you, while you have no hope of doing the same. Or maybe it’s unclear whether either of us can ever secure the coconuts to which he thinks he is entitled. Either way, we face a problem: we can’t have effective rights against one another. What is needed in response is an institutional structure that is not only committed to protecting the rights of all but also able to back that commitment with overwhelming power.⁵⁶

An important higher-level condition must also be met if our institutional structure is to solve our problem: in addition to being designed to address the different objections we have to the private enforcement of rights, it must be designed to do so in a way that we are able to assess. This sheds light on an idea that is central to the entire Kantian tradition in political philosophy: that a civil condition must be understood in terms of “an authority giving laws publicly.”⁵⁷ The motivation for this condition is straightforward. An institutional structure can constitute the new kind of agent we need, and address the objections that arise when rights are enforced by private agents in the state of nature, precisely because it is not just another person like you and me: it has a status that neither of us can claim. But for such a difference in status to do the necessary work, it must be grounded in institutional features that we can both ascertain—features that we can point to, as it were. That’s what allows me, when you object to having the coconut taken from you, to do more than just rehash the substantive considerations that I view as deciding the case. Once the right kind of public institutional structure is in place, I can point to that very structure and explain how it is designed to address the complaints that you have against my snatching the coconut from you. Even if we disagree about the substance of the case, we can still agree that this institutional structure has a claim to decide for both of us that neither of us, as individual agents, could possibly rival.⁵⁸ This is what breaks our stalemate.

Of course, that we can agree doesn’t mean that we will: disagreement is possible here as well.⁵⁹ But not all disagreement has the same significance. When you disagree with me about who rightfully owns the coconut and I use force against you, there are two distinct levels of disagreement. At the most basic level, we disagree about who should have the coconut—I think it should be me, you think it should be you. But we also disagree at a second level: despite our first-level disagreement, I take myself to be entitled to wrestle the coconut from you, whereas you regard my use of force as objectionable in the ways discussed above. What’s crucial on the Kantian view is that the validity of the objections you can present at this second level doesn’t depend on the merits of your case for having the coconut. You have a valid objection insofar as I am not the right kind of agent to use force in a situation in which we disagree; that remains true even if your first-level disagreement with me is unreasonable.⁶⁰

This is not to deny that there may be situations in which I cannot be expected to refrain from using force against you—because it’s just not possible to set up the right kind of institutional structure there and then, and because the stakes are high. That would be the case if you posed an imminent threat to my life by attacking me, or by appropriating all the coconuts. Yet even in such cases there remains a residual problem, stemming from the fact that I act as a private enforcer of rights. Not everything is as it should be. If I could bring into play an agent with the right structure to enforce rights, then that is what I should do. So long as that is not possible, we cannot relate to each other as we ideally should.⁶¹

Contrast this with the case in which we, all of us who share a territory, have put in place a state—an institutional structure aiming to address the various objections we have against the private enforcement of rights. Once we have done this, even if some people don’t agree with the state on this or that point, or question its very existence, that no longer has the same kind of significance. If we have done a good enough job in setting up a state, then dissenters have no valid complaint against the fact that it gets to decide: it is an institutional structure that, by its nature, addresses the different objections they might have, and that does so in a way that no individual agent could. The upshot is that the complaints that you have against my deciding who gets the coconut (and that I would have against your deciding) don’t apply. There may still be some who reject the authority of the state, of course. But that doesn’t mean that they have just as good a claim to decide as it does, or that nothing sets it apart from them. What sets the state apart is precisely that it is structured to answer objections against the private use of force—something that is not true of any individual agent. Because it is so structured, it has the standing to enforce rights consistently with everyone’s right to freedom, and hence to make our rights conclusive. It thus plays an essential role in enabling us to live side by side without undermining one another’s freedom. This results in a particularly robust justification of the state’s authority, one that rests on the very value that a person who asserts the right to decide for herself must invoke.⁶²

Dissenters can still insist that they are right and the state is mistaken. And they may have a valid complaint—not just about the substance of the state’s decision, but about the specific procedure through which it was reached. The point here is not that the state is always right, or that the procedures it relies on are beyond criticism, but that once a legitimate state has been established, the complaints against the use of force that arise in the state of nature no longer apply.⁶³ As between any given person and the state, so long as the state is doing a good enough job, there is simply no question who should get to decide. The state may still make questionable decisions; it may even act so unjustly as to undermine its own legitimacy—but only because it has legitimacy in the first place, grounded in the fact that its structure fundamentally sets it apart from individual agents like you and me.

None of this is to say that someone who rejects the authority of the state is necessarily making a worse mistake, from an epistemic standpoint, than you do when you refuse to acknowledge my claim to the coconut. That will depend on the specifics of the two cases. The crucial contrast turns on the nature of the objections that can be raised in each case, and on what can be done to address them. If you disagree that the state—and not you—should decide matters of rights, we can reply that, by putting in place a state, we have done enough to address your disagreement. You may or may not be convinced, but the point is that you have no residual claim to decide these matters, or to escape the state’s authority, that we need to take seriously. By contrast, in our desert-island scenario, I have no compelling answer to offer when you complain that I haven’t done enough to address your disagreement. The symmetry of our situations gives rise to objections that can only be answered by putting in place an institutional agent of the right kind. So long as we haven’t done that—so long as we don’t have a state—any use of force between us remains problematic.

V. The Case for a Modest Construal

I draw two lessons from the foregoing argument. First, the enforcement of rights between private persons in the state of nature can’t be fully rightful. Second, tackling this problem demands that the interactions of private agents be mediated by a special kind of agent—one structured to address, in a suitably public manner, the objections that arise against private rights enforcement. I now want to ask what implications these lessons carry for how we should approach questions of substantive justice. Does Kant’s main argument bring into play all the elements we need for a complete theory of justice, as proponents of the ambitious construal contend? Or is it better read as providing something more limited: a characterization of the kind of structure at issue in the political context, which must inform our reflection about what an ideal state would look like, but without determining its outcome?

Let me begin with a general reason for thinking of the Kantian contribution as open-ended. Kant agrees with Hobbes on a crucial point: solving the problem that arises in the state of nature requires introducing a new kind of agent. To be sure, the two thinkers construe the thought very differently. For Hobbes, the problem is fundamentally about security. What matters is that our new agent—this “artificial man”⁶⁴—have the overwhelming power required to impose his will on all. This suffices to break the symmetry of mutual vulnerability that gives rise to the war of all against all.⁶⁵ Kant is worried about a different kind of symmetry, stemming from each person’s entitlement to judge for herself. What characterizes the state on this view—what allows it to break the symmetry—is not just power per se, although as we saw power is necessary, but more generally the aspects of its structure that allow it to address the objections that arise when rights are enforced in the state of nature.

Even if the two pictures differ, however, their structural similarities are sufficient to suggest a straightforward way to flesh out the modest construal of Kant’s political philosophy. As Hobbes himself notes, states of various kinds can put an end to the war of all against all. He regards absolute monarchy as a particularly robust and stable solution, since it minimizes the risk of internal division within the sovereign, but democracy and aristocracy are also possible solutions.⁶⁶ Hobbes’s theory is thus importantly open-ended: by itself, it can’t tell us whether an ideal state should look more like Sweden or China, since they both successfully put an end to the war of all against all. How then are we to choose among all the possible institutional structures that can solve Hobbes’s problem? We need to appeal to considerations other than security. There are many options. We could ask which of the relevant institutional structures would best satisfy Rawls’s two principles of justice. Or we could invoke a different theory of justice. Either way, the lesson remains that even if Hobbes’s main argument is correct, it can’t be the whole story. It is at most one part of a more complex theory.

A similar move, I submit, is available within Kant’s theory: we can accept his view about the enforcement of rights in the state of nature, along with its consequences for how we should understand the state’s legitimacy, yet appeal to considerations of a different sort to choose among the various institutional structures that can solve the enforcement problem. The solution won’t be quite as open-ended in this case, since solving Kant’s problem calls for a more determinate kind of structure than the irresistible concentration of power required by Hobbes’s argument. We need an institutional structure that abides by the rule of law, protects property and contract rights, provides minimal poverty relief and health care services, and so on—ideas that have no direct equivalent in Hobbes’s framework. But these more specific requirements can also be satisfied by different institutional structures. How are we to choose among them, if not by appealing to considerations beyond the ones deployed in Kant’s main argument?

The analogy between the two pictures is admittedly imperfect. Hobbes starts from an ideal whose demands are inherently circumscribed. Provided that the horrible scenario he depicts is averted—provided that the state reliably puts an end to the war of all against all—the basic demands of security are met.⁶⁷ Kant’s ideal of universal freedom is normatively richer; insisting on realizing it as fully as possible isn’t plain mad in the way that insisting on maximizing security at the expense of all other values would be. You could therefore think that, properly interpreted, the ideal of universal freedom will be highly specific in its prescriptions. You might even worry that appealing to other values would be in tension with Kant’s claim that coercion can be justified only in terms of the “hindering of a hindrance to freedom.”⁶⁸ Why then expect the Kantian framework to be open-ended in the same way that the Hobbesian picture is?

The key once again is to focus on the nature of the problem that arises in the state of nature, and on what it takes to solve it. The right to freedom demands that individuals have conclusive rights against each other. We saw that this is possible only if a specific kind of agent is in charge of enforcing these rights—an institutional structure that abides by the rule of law, lives up to the ideal of republican government, protects basic rights for all, and so on. But now suppose that different feasible institutional structures can satisfy these requirements. Why would the right to freedom determine what considerations we must appeal to when choosing among these structures if they all satisfy its demands? It seems more plausible to say that, once the freedom-based complaints that make the state necessary have been met, we face a different set of questions—about what further complaints individuals have, given the kind of thing the state is and how it affects their lives; and about how the state can address these further complaints. And that takes us beyond the freedom-based complaints from which we started.

The ambitious Kantian might object that different institutional structures will never deliver equally on the demands of freedom: some will do better than others, and the right to freedom requires that we choose whichever structure maximally satisfies its demands, leaving aside all other considerations. There are two responses to this. The first is that the objection overestimates how determinate the demands of freedom plausibly are.⁶⁹ Insofar as the idea makes sense, we might grant that only structures that maximally realize the demands of freedom should be considered—that is, only structures that, given the stage of development of society and technology, are set up to maximally live up to the ideal of the rule of law, to maximally protect basic rights for all, and so on. Even then, given the nature of the ideas at play, there will inevitably be many possible versions of the relevant structures. We therefore still need to ask what considerations should be invoked to choose among them.

The second response is that, at a more fundamental level, the objection misrepresents how the demands of freedom operate in the Kantian framework. The demand that we can make of each other in the state of nature—and that inevitably remains unmet so long as we remain in that condition—is not that an institutional structure be put in place that maximally realizes or promotes the value of freedom.⁷⁰ It’s that an institutional structure be put in place that does enough to address the complaints to which the private enforcement of rights gives rise. A structure that does this allows us to have conclusive rights, and thus to live side by side without undermining one another’s freedom. The thought that motivates the modest construal is then simply that, when we choose among different possible institutional structures that achieve this—institutional structures that do enough to address our freedom-based complaints—we must do so based on considerations that go beyond the right to freedom. After all, from the point of view of that right, the relevant structures are all acceptable.⁷¹

It is worth noting that Ripstein also seems to view the demands of freedom as open-ended. From the beginning of his discussion, he is careful “to preempt the impression that everything a legitimate state does must be a direct application of Kant’s starting point in the idea of equal freedom.”⁷² Later on, he stresses that, once a rightful structure has been established, judgment on the part of public officials—taking into account “empirical and anthropological factors”⁷³—will be required to determine the specifics of the structure and how it should deal with particular cases. He also emphasizes the essential role of democratic politics in determining what level of social provision should prevail in a given society.⁷⁴ On his view, then, it seems clear that the right to freedom doesn’t crowd out all other considerations: important decisions remain to be made once the demands of freedom are fulfilled.

This goes in the right direction—just not far enough. We can’t rest with the thought that the actual workings of legitimate institutions determine what is rightful in specific cases, and that different choices could legitimately be made by state officials or by the people acting through its institutions. We need to ask what decisions should be made, either by the people or by state officials, once institutional structures meeting the demands of freedom have been put in place. Ripstein is not always ideally explicit on this point, but there are hints that he sees the need to confront this further question. This comes out notably in the “Roads to Freedom” chapter. After forcefully arguing that public roads are essential for freedom, he observes that countless possible networks of public roads would be compatible with the demands of freedom. How should public officials choose among them? At this point, Ripstein suggests, they may take into account considerations other than freedom—efficiency, consistency, and fairness, for instance.⁷⁵ He even entertains an intriguing idea: that the appeals to the interests of individuals that pervade contemporary political philosophy may be acceptable so long as they concern the specifics of a structure that is itself justified by the demands of freedom. Interests (or more specifically, non-freedom-based interests) couldn’t justify putting in place a coercive structure in the first place. But if we have an appropriate freedom-based justification for the structure, and if there are different ways of realizing it, then appealing to interests to make that choice is legitimate.⁷⁶ Ripstein goes so far as to say that, in such a context, taking considerations of fairness into account may be required, even if no argument for the state could proceed in terms of that value.⁷⁷

This closely parallels how we should develop Kant’s theory on the modest construal I advocate. Once we have justified coercive institutions by appealing to the main Kantian argument, we should ask what would count as an ideal version of these institutions—and we should keep an open mind about what considerations might be relevant to answering that question. To make progress on this point, however, we need to go beyond what Ripstein allows. He is right that we must take considerations of efficiency and fairness into account, but why stop there? Why not ask what other values are relevant? Why not ask how these different values relate to one another, and what priority relations hold among them? Once we get started, why stop anywhere short of a full-blown theory of justice?

VI. Beyond the Freedom-Based Approach

It may be tempting to object that your political philosophy is either grounded in the demands of freedom or not—you can’t have it both ways. But there is nothing mysterious about the thought that freedom plays a central role in answering some questions, and a more peripheral one in answering others. When we ask who should make the rules, we are asking what gives one person the right to decide for another. This engages freedom at the most fundamental level. Answering in terms of welfare or of other interests gives rise to familiar objections. To the claim that someone else should decide for you because your interests will be better served overall, you can retort that, as a competent adult, you are entitled to make your own decisions even if they are not always ideal.⁷⁸ But if the state’s making certain types of decisions for everyone is a condition of your being free, then it’s not obvious what your objection could be: as we saw above, the justification proceeds in terms of the very value on which your claim to decide for yourself rests.

A freedom-based approach thus appears particularly well suited to tackling basic questions of legitimacy. But once these have been answered—once we have established the need for an institutional structure that gets us out of the state of nature—we face further questions about how to choose among the different possible forms that such a structure can take. This is where the need to broaden our perspective becomes evident. After all, shouldn’t we consider all the different ways in which, through its political, legal, and economic components, an institutional structure will affect the lives of those who are subject to it? And shouldn’t we take into account the full range of objections that people can make against the workings of institutions of this kind?

This suggests an inquiry into the theory of justice taking the following form. First, you ascertain the kind of impact that the state (understood as the kind of institutional agent that the main Kantian argument calls for) has on the lives of its citizens—how “profound and pervasive”⁷⁹ the impact is, how much of a choice people have about being subject to it, and so on. Second, you consider the various objections that individuals have against a structure that affects their lives in these ways. Finally, you ask what principles would have to guide the state’s actions for these objections to be suitably addressed.

We find nothing of the sort in the Doctrine of Right, but it is very much the shape of Rawls’s inquiry in A Theory of Justice. Famously, Rawls begins by considering the impact of what he calls the “basic structure” on the lives of individuals, and then goes on to ask what principles that structure must satisfy to be justifiable to those subject to it.⁸⁰ Other contemporary theorists may be less explicit about their methodology, but they tend to follow Rawls on this point. On a modest Kantian outlook, it is natural to view the different theories of justice defended by contemporary philosophers in this light, as competing accounts of the broader considerations relevant to the choice among the different possible institutional structures that satisfy basic requirements of legitimacy.

I have claimed that shifting our focus from questions of legitimacy to questions of justice demands that we widen the scope of our investigation to consider the full range of ways in which political institutions affect the lives of those subject to them. If this is correct, it points to an important further reason to expect Kant’s arguments to require supplementation. The impact that political institutions have on the lives of those subject to them plainly depends on the level of development and sophistication of these institutions. At the time of Kant’s writing, states did not have the power to affect the lives of their citizens in the myriad ways that they do now. Their economic role was virtually non-existent by contemporary standards, and the economic structure was immeasurably simpler than it has become. Had Kant considered issues of economic justice, he would have confronted much more straightforward questions than we now face.⁸¹ As the political structure evolves and as its administrative capacities increase, our conception of what it means for such a structure to be justifiable to all must be further expanded. That doesn’t mean that the Kantian solution to the problem arising in the state of nature loses its relevance: it provides the outline that our theory of justice must fill in, since it articulates necessary conditions that have to be satisfied regardless of how the institutional structure evolves. But the content of our theory of justice will depend on how that evolution unfolds.

VII. Conclusion

I have argued that the answer to our opening question—what do contemporary political philosophers stand to learn from Kant’s Doctrine of Right?—isn’t, as the ambitious construal would have it, everything. But it is certainly something. Kant’s central argument yields a conception of what the state is that, if correct, must be at the heart of subsequent inquiries into the state’s duties and responsibilities. In closing, I want to highlight the significance of this point by flagging some important ways in which the modest Kantian outlook that I advocate could enrich contemporary theorizing.

A first contribution comes into focus when we notice that we don’t have to discard the claims about justice associated with the ambitious construal—only reinterpret them as conditions of legitimacy. The nuance matters. Take Ripstein’s contention that, on a Kantian view, no one should be left so destitute or ill that they find themselves dependent on the goodwill of others. We saw above that this claim is too minimal to anchor a plausible alternative to theories of justice such as Rawls’s or Dworkin’s. However, if we take the Kantian position to be that certain minimal distributive measures are required for a state to be legitimate, then not only does it appear more plausible, but we can also see how it makes a significant contribution.

Contemporary theories often remain hazy when it comes to explaining why some demands of distributive justice are also conditions of legitimacy. For the most part, they leave us with little to say beyond, “Well, if the state falls short of our principles of justice but not by much, then it’s unjust but still legitimate. If it falls way short, however, or if it’s not even trying, then it’s illegitimate.” Ripstein’s contention that minimal poverty relief is a precondition of freedom proves more illuminating. It entails that, by failing to provide minimally for the poor, the state violates not just demands of distributive justice but also the fundamental ideal of independence on which its legitimacy rests. Viewed in this light, the later chapters of Force and Freedom (chapters 8 and 9, in particular) take on a less revisionist character. Ripstein’s arguments in favor of redistribution, health care, and equal opportunity may not justify the full extent of social measures that a just society should adopt, but they point to a principled way of determining when a state’s failure to provide for its citizens threatens its legitimacy. This has the potential to sharpen our understanding of the boundary between justice and legitimacy by helping us distinguish between criticisms entailing that a state is not doing a great job, and criticisms entailing that its claim to rule—its very claim to be a state—is shaky.

A further important contribution of the Kantian framework is that it informs how we should approach questions of justice. I have maintained that the Doctrine of Right contains a distinctive and forceful argument for the conclusion that legitimacy requires a specific type of institutional structure. An obvious implication of this argument is that some institutional structures must be ruled out without regard to their substantive merits (those promising happiness for all but failing to abide by the rule of law, for instance). But there are also implications of a more theoretical nature. We have seen that the Kantian argument cannot answer all our questions—it cannot yield a complete theory of justice. But it is not silent regarding the kind of theory that can be brought into play to complete the picture. Kant’s argument establishes that institutional structures must meet certain conditions to enforce rights consistently with everyone’s right to freedom. When we ask what an ideally just version of a structure that meets these basic conditions would be, we can’t turn around and bring in a theory that is inconsistent with the conditions established by Kant’s argument, or that undermines the very possibility of a right to freedom. We must maintain a coherent outlook.

Consider classical utilitarianism, understood as the view that there is one single moral demand that applies to all individual actions and political institutions: to maximize overall happiness. A monistic view of this kind plainly leaves no room for the Kantian argument. It may allow for some of the conditions on legitimacy that Kant defends—perhaps it can be shown that institutions that conform to the rule of law, and that protect property and contract rights, are good tools to maximize human happiness in the aggregate. But the argument would have to be purely instrumental; the more robust Kantian account of these conditions, based in the value of freedom, is excluded outright by the monistic character of the theory. This is an extreme example, but it illustrates that there are distinct limits to the kind of theory that can be brought in to complement the Kantian argument. At the same time, we don’t want to overstate these limits. Classical utilitarianism is not a possibility, but a mixed view that includes some utilitarian considerations may be. Presumably, you could have a mixed theory recognizing the demands of freedom while holding that, among feasible structures that satisfy these demands, we should choose whichever one will maximize human happiness. Whether such a theory could be made to work remains to be seen, but it is not ruled out by the Kantian argument as I interpret it—even if other approaches to justice (Rawls’s or Dworkin’s, for instance) would seem a more natural fit for Kant’s starting point.⁸²

How then are we to choose among the different theories of justice that could complement the Kantian outlook? On the modest view I advocate, the choice clearly can’t proceed from Kant’s premises alone. But that doesn’t mean that it is independent of these premises—or that there is no right answer. At this point, we must appeal more explicitly to the justificatory framework that should, in my view, guide our entire inquiry: Rawls’s method of reflective equilibrium.⁸³ The question we need to ask is which theory of justice, when combined with the Kantian argument, makes for the overall theory that best accounts for our considered judgments. Combining the Kantian framework with our substantive theory of justice in this way has the advantage of giving us a more structured theory to work with. This is significant—not because we should care about structure for its own sake, of course, but because the Kantian theory renders our normative theorizing about the state more systematic. It elucidates the context in which questions of justice arise by telling us what the state is at the most fundamental level. And it explains why, when we consider how the kind of impact that the state has on our lives could be justified, we are not asking how to alleviate the effects of an unfortunate system we happen to be stuck with; we are asking how best to organize an institutional framework that is necessary for individuals living side by side to be free. This keeps our theorizing about justice firmly grounded in one of the deepest normative questions we face as rational agents sharing a physical world.

Acknowledgements

Earlier versions of the material that slowly morphed into this paper were presented at Princeton’s University Center for Human Values, at a workshop at Boston University, at a conference on Kant’s political philosophy at Merton College, Oxford, at the ECPR General Conference at the University of Oslo, at UNAM, and at La Sapienza. I am grateful to the audiences on all these occasions for extremely helpful questions and comments. I am particularly indebted to Luke Davies, who delivered thought-provoking comments during the Oxford workshop, and to Rafeeq Hasan and Arthur Ripstein, both of whom provided insightful comments and suggestions on the paper. Finally, I want to thank two reviewers for this journal whose detailed and challenging comments greatly improved the paper.

Competing Interests

The author has no competing interests to declare.