2. Preliminaries: moral cosmopolitanism

Before turning to a defense of a relational version of statism, let me begin by sharpening the challenge to a statist account and by clarifying the notion of moral cosmopolitanism with which I shall work.

Let us begin, then, by distinguishing two kinds of cosmopolitan views.Footnote ⁵ Institutional cosmopolitanism is a political view about the best way to structure our legal and political institutions. The institutional cosmopolitan recommends establishing a global political authority or world government. By contrast, moral cosmopolitanism is a normative view about the best way to justify normative concepts and practices. The moral cosmopolitan takes as fundamental the interests of individuals. Here is how Charles Beitz, an influential proponent of moral cosmopolitanism, puts the point:

These two views of cosmopolitanism differ in nature: one is a political thesis, the other a thesis about the structure of moral justification.

Distinguishing these two kinds of cosmopolitanism is important because the theses they involve are logically independent. One may be an institutional cosmopolitan without being committed to moral cosmopolitanism: one might maintain that a world government is of fundamental moral significance. And one may be a moral cosmopolitan without being committed to institutional cosmopolitanism: one might maintain, as Pogge and Beitz do, that the interests of individuals are of ultimate moral significance and that instituting a single world government is a bad idea.

Furthermore, isolating the moral cosmopolitan thesis is important because it offers a distinctive answer to our question about the relationship between human rights and state sovereignty. Since the moral cosmopolitan endorses the view that only individual interests are of fundamental moral significance, human rights and state sovereignty can matter only derivatively. Human rights matter insofar as they offer protections of individual interests. And state sovereignty matters only insofar as it secures, protects, or promotes human rights.

To appreciate the structure of the moral cosmopolitan answer, let us introduce a working notion of state sovereignty. Following Thomas Pogge, we may think of sovereignty as follows (Pogge Reference Pogge and Brown1994, 94). Let A be a governmental agent and B an individual (or corporate) person. A is sovereign over B if, and only if, A has the authority to (a) make rules constraining A’s conduct, (b) judge A’s compliance with rules, and (c) enforce said rules.

The moral cosmopolitan, then, appears to offer a straightforward account of limited state sovereignty. A has authority over B subject to the condition that A’s exercise of said authority furthers human rights and, ultimately, the interests of individuals. Take as an example Pol Pot’s policies of mass executions and forced labor leading to deaths, approximately, of 25% of Cambodia’s population. The moral cosmopolitan can argue straightforwardly that Pol Pot’s dictatorship lacked sovereign authority because Pol Pot’s policies clearly undermined the interests of individuals. So Pol Pot lacked domestic legitimacy and external sovereignty, opening his regime to sanctions of various kinds.

Moreover, the moral cosmopolitan can offer two further kinds of reasons for its view. The main competitor, a statist view, appears normatively opaque and faces a dilemma: either it is committed to unlimited sovereignty or it collapses into moral cosmopolitanism.

Begin with the first kind of reason and consider briefly the statist version defended by John Rawls in Law of Peoples. Rawls does not ground human rights in the interests of individuals. Rather, he justifies human rights in virtue of their particular inferential role in the interactions among peoples: domestic conformity to a minimal set of human rights is a necessary and sufficient condition to enjoy good standing in the international society of peoples.Footnote ⁶ A sovereign who violates the minimal human rights of its members loses good standing and opens itself to international intervention.

Rawls’s statism has been charged with normative opacity. Allen Buchanan argues that in refusing to ground human rights in human interests, Rawls’s account of human rights remains opaque about the normative basis of human rights (Buchanan Reference Buchanan2010, 31–49). Furthermore, the view that the state has fundamental moral significance itself appears normatively opaque. Thus, Charles Beitz argues that the ‘foundations of the view [that state sovereignty matters unconditionally and intrinsically] are obscure, ’ for it is not clear how we could defend the intrinsic importance of state sovereignty (Beitz Reference Beitz and Brown1994, 129). So the statist view appears normatively opaque: it is not clear exactly how a statist can justify the ultimate moral significance of states or of human rights without appealing to the ultimate significance of individual human interests.Footnote ⁷

To make matters worse, statism appears to face a dilemma. On the one hand, if we endorse the statist view that state sovereignty is of fundamental moral significance, we appear committed to an unlimited conception of sovereignty. If state sovereignty matters intrinsically, it becomes difficult to see how it could possibly be limited by external considerations, such as the importance of individual interests and rights. On the other hand, if state sovereignty is limited, then we need to explain the basis of such limit. A natural explanation of the limits of state sovereignty is instrumental: state sovereignty is justified to the extent that it tends to promote justice and the interests of individuals. But if so, then statism would collapse into moral cosmopolitanism. This dilemma suggests that moral cosmopolitanism is not only a plausible view about state sovereignty, it is also the only plausible view of limited state sovereignty. Moral cosmopolitanism appears inescapable.

These considerations suffice for giving shape to the challenge to a statist account of state sovereignty I shall consider in this article. Such an account must show (a) that moral cosmopolitanism is not the inescapable explanation of limited state sovereignty, (b) while avoiding normative opacity. And in so doing, it must explain (c) the normative basis of human rights and the rights of states. In what follows, I shall argue that a relational account can satisfy these demands. To anticipate: for the relational account I shall propose, human rights and state sovereignty matter non-instrumentally as requirements of a basic norm of reciprocal independence. If so, moral cosmopolitanism is not inescapable; statism need not be opaque; and a relational statist account can illuminate the normative basis and unity of human rights and the rights of states.

3. A relational account: reciprocity, authority, and legitimacy

This section introduces the three key organizing ideas of a relational account: Reciprocity, Relational Authority, and Relational Legitimacy. These ideas will enable us to see, in the next section, how a relational account can meet the challenge to a statist conception of state sovereignty.

Let us begin with the basic norm of the proposed relational account: Reciprocal Independence.Footnote ⁸

Reciprocal Independence names an irreducibly relational deontic norm of independence. Let me explain.

The right to independence entails a deontic prohibition on subordination and the requirement to interact with one another as moral equals. Roughly, I subordinate you when I treat you simply as a causal means for the pursuit of my purposes without your consent.Footnote ⁹ This prohibition abstracts away from the value of my purposes. Even if my purposes are morally worthy, such as promoting the common good and acting charitably, the quality of my purposes cannot override the prohibition on subordinating you.

Notice that, understood as a right to interact with others on a footing of equal independence and as a prohibition on subordination, Reciprocal Independence does not prohibit relations of dependence, for dependence as such need not amount to subordination. You may depend on farmers and fishermen for your food, but so long as farmers and fishermen do not treat you simply as a causal means to their purposes, they do not subordinate you. Similarly, you may depend on a doctor’s judgment to get better from a nasty flu, but so long as the doctor does not treat you simply as a causal means to her purposes, she does not subordinate you. Reciprocal Independence, then, is perfectly compatible with a variety of relations of dependence necessary to the fabric of social life.

Relational Independence is a deontic and irreducibly relational norm. It is deontic because it functions as a basic right against others and as a basic duty of respect owed to others. Independence, then, is not a good to be promoted; it is a constraint on how others can treat you. And Reciprocal Independence is irreducibly relational because it is not grounded in some further non-relational norm. Independence is a standing you have in relations to others, not a good you can enjoy even in the absence of others. This basic norm captures, then, a relational notion of equality, our equal standing as independent, rather than a non-relational notion of equality, an equal measure of some good, like autonomy or well-being (Anderson Reference Anderson1999; Darwall Reference Darwall2013, ch. 7).

In its emphasis on independence and non-subordination, Reciprocal Independence draws from the republican tradition of political thought. Yet, by representing a relational deontic norm, Reciprocal Independence departs from the currently popular forms of republicanism, which treat non-domination as a consequentialist good to be promoted.

Similarly, Reciprocal Independence helps us develop a republican understanding of the justification of public law and public authority. In a nutshell, public authority is not justified instrumentally as a helpful or indispensable means to securing some further good. Rather, public authority is justified non-instrumentally because it is partly constitutive of our independence. The state as a public authority is morally required because without it, we would remain exposed to the subordination of others. Let me unpack this argument.

What might it mean to say that a public authority is partly constitutive of our reciprocal independence? To see this, consider John Rawls’s early and illuminating contrast between two conceptions of rules (Rawls Reference Rawls and Freeman2001). Conceived as a summary, a rule is justified as a ‘rule of thumb, ’ as a principle that, in the past, has helped us attain an independently valuable good. Given our epistemic limitations, our inability to secure the good ‘effortlessly and flawlessly, ’ there is need of rules to save time and guide our action based on a summary of past performance (Rawls Reference Rawls and Freeman2001, 35). Thus, when at the supermarket you follow the rule Shop for flour X, you are deploying the summary conception of rules: in the past, you have found that flour X was best in terms of price and quality, and now you lack the time to ensure that this is still the case, so you continue purchasing flour X. By contrast, the rule of a practice, Rawls suggests, is constitutive of said practice: ‘the rules of practices are logically prior to particular cases’ (Rawls Reference Rawls and Freeman2001, 36). As Rawls illustrates, the specific action of stealing base is logically parasitic on the rules of the game of baseball. The action can only count as stealing base because it is constituted by the practice and its rules. A rule is constitutive, then, when it defines and makes possible a specific kind of action as an instance of the practice.

We may draw a parallel distinction when thinking about the structure of the state’s authority. For the moral cosmopolitan, the justification of public laws must take an instrumental form: laws are morally justified to the extent that they promote the independent good of individual interests. Public laws, then, take the form of summary rules. By contrast, on a relational account, the justification of public laws takes a non-instrumental form: laws are morally justified because they are partly constitutive of our reciprocal independence. They only partly constitute our independence because the fundamental idea of independence is interactional, pertaining to the relationship between you and me. But public law, I suggest, still plays a constitutive function in publicly entrenching and defining our standing as independent. On this picture, public laws take the form of constitutive rules: they define and make possible our status as independent. But what supports this constitutive justification?

The basic point is the republican idea that public law does not interfere with our freedom but makes it possible. Footnote ¹⁰ The broadly liberal idea is that public law is justified instrumentally, as a means for securing compliance and assuring compliers that others comply too. The republican idea is that public law is justified non-instrumentally, not only by securing and protecting our independence, but also by publicly entrenching it so as to eliminate relations of interpersonal subordination.

Imagine that we lived in a world without coercive public authorities where people are disposed to act morally and where resources are reasonably limited, so that some enjoy more wealth and power than others.Footnote ¹¹ You are wealthy, powerful, and disposed to respect the independence of others. Yet, whether you happen to respect the independence of others will depend on your continued good will. And this means that the poor and powerless would continue to enjoy their independence only on the condition of your continued good will. But this would mean that in a world without a coercive public authority, the poor and powerless would remain exposed to subordination to the powerful since their continued enjoyment of their independence would depend on the good will of the powerful.

The point, then, is not just that a coercive public authority would protect the weak from the strong, for that it would certainly do. The point is that a public authority would play a further role in conferring and constituting the status of all as free by publicly entrenching that status. This would mean that the independence of the weak, as Pettit puts it, would not be a gift for which they ought to be grateful, but an institutionally guaranteed claim (Pettit Reference Pettit2012, 183).

Consider an (admittedly limited) analogy. Your status as an independent car driver turns on the extent to which you can drive around free from subordination. Absent public laws and public rules, you may be said to be free to drive your car. But can we all enjoy our independent status as car drivers absent public rules? The problem is that not all of us could be independent drivers simultaneously. Some, the powerful and strong, may take advantage of their position by starting to impose unilaterally tolls, fees, and other kinds of lucrative measures, treating you simply as a means to their purpose of enrichment. Others may set up road extortion rackets: setting up roadblocks and subsequently demanding payment for their removal. More generally, the lesson is that without public rules, not all can be independent drivers: driving around becomes subject to the good grace and good will of others. Public rules, the relational republican argument goes, are required so that all can be equally independent drivers.

When a public authority institutes a system of roads, a system of rules, and an enforcement mechanism, it does not simply solve a technical problem of coordination (for which it may not be necessary); it solves a moral problem of subordination (Ripstein Reference Ripstein2009, ch. 8). Furthermore, the establishment of such rules also publicly entrenches and partly constitutes your status as an independent driver, for now you have standing under a public law, for instance, by having a driver’s license. Similarly, I am suggesting, a public authority partly constitutes you as an independent person by making possible your independence.

If correct, these reflections offer support for the following principle:

Relational Authority is a norm warranted by the thought that without a public authority, a system of independence would not be morally possible, as some would remain subject to subordination. This is a relational concept insofar as it represents the normative relationship between a public agent and its members, rather than some intrinsic normative property possessed by a public agent. And it is deontic. It represents the rights of the public agent to make, adjudicate, and enforce public rules correlative to the duties of private members to obey such rules.

At the same time, Relational Authority supports a second public principle:

The basic idea behind Relational Legitimacy is simple: Relational Authority cannot be a carte blanche on the public agent to act as it please. If that were so, Relational Authority could not solve the moral problem of subordination. Instead, it would become itself an agent of subordination. The relation of authority must then be complemented with a relation of legitimacy: a public agent’s activity will be legitimate only if it is constrained by the reciprocal independence of its members.

Later, I shall specify further this norm of legitimacy by introducing the notions of human rights and of a systematic violation of human rights. For now, this admittedly sketchy presentation of the three principles should suffice to show how a relational account of human rights and the rights of states can satisfy the explanatory demands we identified in the previous section.

4. Human rights and internal state sovereignty

This section shows how our three relational principles, Reciprocity, Relational Authority, and Relational Legitimacy, generate a distinctive conception of the normative basis of human rights and illuminate the unity of human rights and state sovereignty. The next shows how the relational conception can meet the challenge to a statist account.

Begin with the idea of human rights. The relational approach suggests thinking of human rights as necessary specifications of our equal right to independence. The thought may be formulated as follows:

As we have seen, Reciprocal Independence is a relational deontic norm: it represents the correlation of basic rights and duties to independence. Your independence thus represents a constraint on the conduct of others, rather than a good to be promoted.

Understood relationally, then, a human right is a necessary specification of independence in the following sense: it specifies a particular kind of relationship as a relational deontic requirement.Footnote ¹² To say that a human right is a specification of your right to independence is to say that a human right takes the same form (relational and deontic) as the master norm of reciprocity. Human rights, then, are not justified instrumentally as means to some further good, but as deontic constraints on others. To say that a human right is a necessary specification is to say that a human right is necessarily required by your status as independent.

For instance, to say that there is a right to life would then be to say the following. First, your right to life functions as a relational deontic constraint on others: in the minimal case, others have a duty not to murder you. Second, the right to life specifies a particular kind of possible subordination. The violation of this right involves treating your body simply as a means to the other’s purposes by murdering you. This is a specific form of subordination for the obvious reason that others may subordinate you in other ways that fall short of murder, such as by coercing you, deceiving you, and enslaving you. And third, the right to life is necessarily required by your status as independent. Since murder necessarily contravenes your independence, the necessary prohibition on murder becomes the object of a human right.

It is worth pausing here to consider how distinctive the relational account of human rights really is. While Relational Human Rights represents human rights as irreducibly relational deontic constraints, the dominant views justify human rights teleologically, as (indispensable) means for the production of independent goods. For instance, as Allen Buchanan puts it, the dominant view of human rights may be understood as the basic human interest conception: human rights are derivative norms justified by the fact that they promote the external and independently intelligible value of basic human interests (Buchanan Reference Buchanan2010, 116). In this way, human rights are grounded variously in the goods of normative agency (Griffin Reference Griffin2008), human functioning and capabilities (Nussbaum Reference Nussbaum2000; Sen Reference Sen1999), a minimally good and decent life (Tasioulas Reference Tasioulas and Pogge2007), or the satisfaction of urgent human interests (Beitz Reference Beitz2009; Buchanan Reference Buchanan2010; Nickel Reference Nickel2007; Pogge Reference Pogge2002; and Shue Reference Shue1996). Despite their differences in the characterization of the basic good to be promoted, these various conceptions share the same teleological justificatory structure and thereby regard human rights as morally derivative from more basic goods. By contrast, the relational view is irreducibly relational and deontic, thereby treating human rights as morally basic in our practical thought.

This difference between teleological and relational deontic strategies of justification is important for two reasons. First, the relational account seems particularly apt for a public justification of human rights.Footnote ¹³ Teleological views necessarily rely on some substantive conception of the underlying good to be promoted. By contrast, the relational account abstracts away from any particular view of the good and defends, instead, the basic thought that no one is entitled to subordinate others. Reciprocal independence may be endorsed, then, from a variety of normative perspectives, as one may come to believe in the prohibition on subordination for a variety of reasons, such as a command by God, a teleological demand of our human nature, and a constraint on democratic deliberation. Since Reciprocal Independence is morally basic in the order of justification, it does not rely on any substantive view about the human good. So this makes the formal idea of Reciprocal Independence particularly apt for a public justification.

The second reason why the distinctive character of the relational account matters is that it also suggests a distinctive picture of the justification of the state’s authority and legitimacy. Unlike teleological views, Relational Authority and Legitimacy do not ground the state’s authority and legitimacy instrumentally in terms of the good consequences to be obtained by living under a state. Rather, the relational view justifies the state’s authority and legitimacy in the fact that without it, our reciprocal independence would not be morally possible. This is to say, as we have seen, that without a state’s authority and legitimacy, while some of us might be able to enjoy their independence, perhaps the wealthy and powerful, not all of us would. But this would contravene the universal scope and egalitarian character of Reciprocal Independence. And so, a relational account justifies the state’s authority and legitimacy by virtue of the fact that these normative powers make our independence possible.

This is a significant argumentative upshot because it yields a distinctive picture of the justification of state sovereignty. Drawing from the previous four steps, we may infer the following principle:

(5) says that S is internally sovereign when it satisfies Relational Authority: S will then have the right to have authority over its members to make, adjudicate, and enforce public rules concerning the rights to independence of its members. But (5) also says that S will be internally sovereign so long as its authority remains legitimate, that is the exercise of its authority remains consistent with the basic right to independence of all its members. A tyrannic state like Pol Pot’s is a clear example of a state that subordinates its subjects by failing to respect their right to independence. Such a state not only would lack relational legitimacy, it would thereby lack authority. Losing legitimacy means losing the authority to make, adjudicate, and enforce rules.

Furthermore, the relational conception of human rights sketched in (4) allows us to specify further the notion of legitimacy. General compliance with human rights is a further necessary condition for the legitimacy of the state. This follows from (1–4). On the relational conception, human rights just are necessary specifications of a basic right to independence. If legitimacy is constrained by the basic right to independence and if human rights are necessarily entailed by such a right, it will follow that legitimacy will be constrained by human rights. This master argument for legitimacy would have to be specified further to understand the precise conditions of legitimacy. Yet, for now, this argument should suffice to solve the problem for statism we encountered earlier.

5. A relational defense of statism

Recall, a statist account faces the following explanatory challenges: (a) moral cosmopolitanism offers a natural explanation of limited state sovereignty; (b) by comparison, statism can appear normatively opaque; and (c) indeed, moral cosmopolitanism appears inescapable, offering the only viable account of limited state sovereignty. This section shows how a relational account can support statism. I address each of these challenges in reverse order.

First, statism appears to face a dilemma. In a nutshell: if state sovereignty is of fundamental moral significance, it must be unlimited, but if state sovereignty is limited, it must be justified only to the extent that it satisfies some independent notion of social justice. This makes it seem as if moral cosmopolitanism is inescapable.

But moral cosmopolitanism is not inescapable. That is because (5) Internal Sovereignty gives us the resources to deny the first disjunct. State sovereignty can be of fundamental moral significance and limited. We may see this by recalling the argument for Relational Authority and Legitimacy.

For the relational account, state sovereignty is of fundamental moral significance because it is morally required by Reciprocal Independence. Without a public authority, some of us would remain exposed to subordination and would remain subject to the good will of those more powerful, stronger, and wealthier. If Relational Authority obtains, the justification for state authority is not instrumental but partly constitutive: a public authority is morally required in order to make possible a life together in equal independence.

At the same time, granting Relational Authority does not commit us to the view that the state’s authority must be unlimited. If Relational Authority granted the state an unlimited right to make, enforce, and adjudicate public laws as it pleased, Relational Authority could not solve the moral problem it was introduced to solve. As republican thinkers have warned us, the state would then prevent one form of subordination, a horizontal one, by replacing it with another, a vertical one. If the state’s authority were unlimited, it would contravene Reciprocal Independence, as the grim example of Pol Pot’s regime illustrates. That is why, as we saw above, Relational Authority must be complemented with Relational Legitimacy. A necessary condition of a state’s legitimacy is that its actions be compatible with its members’ reciprocal independence. And so, illegitimacy is a sufficient condition for a given state to lose its authority. But this is just another way of saying that a state’s authority cannot be unlimited.

Furthermore, according to (4) Relational Human Rights, every human right is justified as a necessary specification of reciprocal independence. So general compliance with its members’ human rights must be a necessary condition of the state’s legitimacy. A relational account, then, gives us a way of seeing why a state’s authority must be limited by human rights.

If these reflections are correct, they show that the first disjunct of the moral cosmopolitan dilemma is false. The state’s authority can be of fundamental moral significance while remaining limited by human rights. And if so, moral cosmopolitanism is not the sole way of accommodating and justifying a limited notion of state sovereignty.

Let me sharpen this argument by considering another way in which one might support the truth of the first disjunct. One might insist that insofar as the proposed relational account defends a version of statism, it is committed to the view that states and their interests are of ultimate moral significance. And if so, as Fernando Tesón argues against any form of statism, statism would sanction sacrificing the interests of individuals for the sake of protecting the state’s interests (Tesón Reference Tesón1998, 1–5, 40). But if this is right, statism lacks the resources to block a commitment to a fairly unlimited view of state sovereignty.

Tesón is certainly correct in claiming that some versions of statism defend the ultimate moral significance of the state in such a way that they end up committed to the view that, in cases of conflict, the state’s interests trump those of individuals. Yet, it is hasty to generalize and claim that any version of statism must be committed to such a view.

In fact, as I have been arguing, when we justify the moral significance of the state relationally, the moral significance of the state is only intelligible against the backdrop of Reciprocal Independence. So Reciprocal Independence shows why we might think of the state as morally required while showing why we would also hold that the state’s authority is not a carte blanche. If this is correct, there is at least one version of statism – the relational one – for which it would be true to say both that the state’s authority is of fundamental moral significance and that said authority must be limited by the human rights of its members. If so, promoting the ‘interests of the state, ’ whatever these might be, would not warrant treating individuals simply as means for the state’s policies. Such a course of action would amount to a form of public subordination prohibited by Relational Legitimacy.

Now, even if moral cosmopolitanism turns out to not be inescapable, the moral cosmopolitan rightly worries about the normative basis of a statist account. The basic concern is that by decoupling human rights from considerations about the human good and human well-being, we end up taking the human out of human rights, in Allen Buchanan’s wonderful turn of phrase. As a result, we end up with an obscure picture of the normative grounds of human rights and of the rights of states.

For instance, as I mentioned earlier, Allen Buchanan voices this type of objection to John Rawls’s account of human rights in The Law of Peoples. Buchanan argues that Rawls develops a theory of human rights ‘without recourse to a conception of minimal human good and indeed without reliance on the idea that there are any morally fundamental characteristics that all human beings have’ (Buchanan Reference Buchanan2010, 32). Instead, Rawls apparently justifies human rights in virtue of their function in an international society of peoples (Rawls Reference Rawls1999, 80). One of Buchanan’s central concerns is that this functional justification is empty, for it lacks the resources to identify what rights should count as basic. And even if Rawls wished to ground human rights in the idea of reasonableness, this strategy would also fail. This is because, Buchanan argues, ‘there is understandable skepticism about the usefulness of the notion of what cannot unreasonably be rejected’ (Buchanan Reference Buchanan2010, 37).

Similarly, Charles Beitz has powerfully argued that the morality of states, and in particular the principle of state sovereignty, is normatively opaque and thereby arbitrary. Charles Beitz articulates a battery of arguments against the internationalist view and concludes that since none of the arguments for state independence work, we should endorse moral cosmopolitanism. Thus, Beitz considers and rejects the following arguments for the principle of state sovereignty. States ought to be treated as independent because (i) they protect (or are themselves a) free association (Beitz Reference Beitz1999, 72); (ii) because this principle is impartial ‘between competing conceptions of the good in international relations’ (Beitz Reference Beitz1999, 87); and (iii) because this principle is anti-paternalistic since ‘a state is more likely to know its own best interests than any other state’ (Beitz Reference Beitz1999, 84). Beitz finds each of these arguments unconvincing and concludes that the principle of state independence ‘lacks a coherent moral foundation’ (Beitz Reference Beitz1999, 121). So how forceful is the objection that a statist account of human rights and/or state sovereignty is normatively opaque?

The first three steps of the relational account alleviate a concern about normative opacity. Indeed, the relational justification of human rights and the rights of states is perfectly transparent. Human rights matter morally because they are required by our right to independence. Similarly, as we saw in (5), state sovereignty matters because it is morally required. It makes possible and thereby partly constitutes a system of reciprocal independence. Of course, one may not agree with the proposed relational account on other grounds, but it should be evident that there is nothing normatively opaque about it.

It is true: unlike Buchanan’s teleological account, the relational account does not have ‘recourse to a conception of minimal human good.’Footnote ¹⁴ But, naturally, this need not mean that the account is normatively opaque, for there are non-teleological ways of justifying moral norms. Indeed, as I have mentioned, it may be a virtue of the relational account that it requires no commitment even to a minimal notion of the human good, as this may render it more apt for public justification. Although deontological, the relational account does not tell us to respect human rights simply because moral rules matter. Instead, it tells us to respect human rights because our reciprocal independence is a plausible candidate for basic moral significance.

Furthermore, even if, like Buchanan, one were skeptical about the possibility of grounding human rights in the Rawlsian idea of reasonableness, one should bear in mind an important difference between that idea and Reciprocal Independence. Reciprocal Independence is both more austere and more robust than Rawlsian reasonableness. Rawls conceives of reasonableness in terms of the conditions of the possibility of a practice of fair cooperation. Reciprocal Independence is more austere because it does not presuppose the idea of such a practice, but simply focuses on a basic deontic norm of independence and respect. By the same token, Reciprocal Independence is more robust than reasonableness because it appeals explicitly to the ideal of independence, the right of each not to be subordinated to others. If these reflections are correct, the objection of normative opaqueness misses its mark when directed at the relational account.

But how about Charles Beitz’s objection that a statist account is normatively opaque because it lacks a coherent moral foundation? Notice that we can accept each of Beitz’s three main arguments against statism and still resist his conclusion. This is because the relational account does not appeal to any of the possible justifications Beitz considers and (let us assume) rightly rejects.

The relational account grounds the state’s sovereignty in a moral requirement of our independence. This is different from grounding state sovereignty in the (distributive or collective) right to free association. My argument was not that state sovereignty matters because it is a means of embodying our right to associate freely. Rather, my argument was that state sovereignty matters because it is partly constitutive of our independence. Similarly, the relational argument does not appeal to the state’s impartiality among competing conceptions of the good or in a form of international anti-paternalism. It emerges that even if Beitz’s arguments against these views are sound, they leave untouched the relational account. And so, I conclude that the objection of normative opaqueness is not forceful against the relational account.

Nevertheless, one’s concern with a relational account of statism may stem from the first reason in favor of moral cosmopolitanism, namely: (a) that moral cosmopolitanism just offers a natural and straightforward explanation of the state’s limited authority. A subtler concern, then, is whether the relational account offers us a sharp enough contrast with the moral cosmopolitan view. Could one not endorse a relational view about the justification of human rights and still deny that the state is of fundamental moral significance? If so, the relational justification of statism may not be so straightforward and may even collapse into moral cosmopolitanism.

To begin, it will be helpful to recall the main point of contrast between the relational and moral cosmopolitan accounts. The difference concerns the strategy for justifying the authority and legitimacy of the state. The moral cosmopolitan adopts an instrumentalist strategy: the state has only derivative moral significance as a means to the good of individual interests. The relational account adopts a non-instrumentalist strategy: the state has fundamental moral significance because it makes possible (and so partly constitutes) a system of reciprocal independence. But, one may continue to press, why could the relational account not be moral cosmopolitan?Footnote ¹⁵

The reason the relational account defended here cannot be moral cosmopolitan turns on the relational character of its basic norm: independence. Independence, as we have seen, departs from certain liberal understandings of freedom as essentially negative. Independence is not liberty from interference with our preferred choice from a given option-set; neither is it liberty from interference with our option-set simpliciter. Both of these conceptions share an essentially non-relational understanding of liberty. This difference is crucial because on these views, one might reach the highest pitch of freedom in perfect isolation from others, in a situation where one lacks interpersonal interference altogether. By contrast, independence is an irreducibly relational idea: you are free only when you relate to others on a footing of moral equality. Independence, then, is a way of relating to others, not a way of being left alone.

If the argument sketched earlier is correct, independence cannot be realized for all unless a public authority is in place. This suggests, as we have seen, that a public authority is morally basic, as the sole agent that can make our independence co-possible. But if that is so, then the relational account defended here cannot be moral cosmopolitan because its notion of independence commits it to the fundamental moral significance of the legitimate state, a view the moral cosmopolitan, by stipulation, must deny.

Indeed, pursuing this contrast further enables us formulate a challenge for a moral cosmopolitan view. To be clear: the argument that follows is a challenge, not a refutation. The challenge is meant simply to cast some doubt on the view that moral cosmopolitanism has a natural and straightforward explanation of the relationship between human rights and the rights of states.

The basic concern is the following. The relational account suggests that human rights cannot be fully specified without reference to the relation between a state and its subjects.Footnote ¹⁶ Human rights may well be partially specified without reference to the state. One might grant, for instance, that we can understand the concept of a human right to life or against slavery without the concept of a public authority. Still, if it were a constraint on one’s account that human rights be fully specified without any reference to the notion of a public authority, this would seem to result in an incomplete and arguably under-inclusive picture of human rights.

Take the human right to due process, for instance, a staple civil human right. This human right makes necessary reference to the notion of a public authority under whose judgment and adjudication of rules we must be treated as equals. But if the notion of human rights could not make reference to a public authority, we would lose our grip on the nature and content of this right. At least some human rights, then, seem to require reference to the notion of a public authority. However, the moral cosmopolitan cannot include a relation between a public authority and its members as part of its fundamental moral categories. If it did, it would collapse into some version of statism. And so, if at least some human rights make essential reference to the state, no moral cosmopolitan story could offer a complete and straightforward account of human rights.

Oddly enough, moral cosmopolitans tend to grant the basic premise of the argument I have just sketched. They tend to endorse the institutional view that human rights are not fully intelligible independently of the institution of the state.Footnote ¹⁷ The right to due process, as I have mentioned, ceases to make sense if it cannot make essential reference to a specific public authority constituted by public laws. Similarly, the human right to nationality may be understood as the right to be a citizen of some public authority or another. Some regard this as the most important human right of all. But the point that matters here is that this human right, if it is one, makes necessary reference to the concept of a public authority.

To repeat, for the argument to succeed, one need not claim that any human right must be understood as making reference to the state, although that may be true. The argument only requires that some human rights make necessary reference to a public authority as the bearer of duties to respect and protect said rights. If some human rights make essential reference to the relation between a state and its members, and if human rights are morally fundamental norms, then the relation to the state must also be morally fundamental. And if so, the moral cosmopolitan must renounce her core claim that a public authority is only of derivative moral significance, to be justified simply in instrumental terms.

We may sharpen this argument by framing it as a dilemma for moral cosmopolitanism. On the one hand, the moral cosmopolitan may well block this objection by denying that human rights are essentially institutional norms. She could then retreat into the pre-institutional view that human rights are fully and completely intelligible without any reference to the notion of a public authority. Yet, this move is anathema to some of the most prominent moral cosmopolitans, and for good reason, since (many) human rights appear to make necessary reference to the notion of a public authority. On the other hand, the moral cosmopolitan could grant that human rights make essential reference to the state. But then it becomes difficult to see how the moral cosmopolitan could block a collapse into statism: if human rights are morally fundamental and they make necessary reference to the state, the notion of the state must also be morally fundamental.

Let me clarify the dialectical significance of this argument. I do not present this argument as a fatal objection, much less a refutation of moral cosmopolitanism. Rather, I articulate it to cast some doubt on the thought that moral cosmopolitanism offers a natural and straightforward explanation of the relationship between human rights and the rights of states. This argument, then, is meant to shift the burden of proof back to the moral cosmopolitan and thereby to strengthen the relational version of a statist account.

Let us take stock. Earlier, in section one, we contemplated three main reasons in favor of moral cosmopolitanism and, accordingly, three explanatory challenges a statist conception of human rights must meet: (a) cosmopolitanism offers a natural justification of human rights and the authority of the state; (b) competing statist accounts appear normatively opaque; and (c) so long as we hold that state sovereignty is limited, moral cosmopolitanism appears inescapable. If successful, this section has shown that a relational account of human rights and the rights of states can satisfy these explanatory demands.

My argument has proceeded in reverse order. (c) Moral cosmopolitanism is not inescapable. A relational account promises to show how the authority of the state can be both morally fundamental and limited. (b) At least one version of statism, the relational one, is not normatively opaque, for it grounds human rights and the rights of states in the basic right to independence of individuals. And (a), upon reflection, there is some reason to doubt that moral cosmopolitanism offers a natural justification of the authority of the state and its relation to human rights. By contrast, a relational account itself offers a natural justification. Taken together, these arguments suggest that a relational account can illuminate the relationship between human rights and the rights of states.

6. External sovereignty

The proposed relational account can still appear vulnerable to an important objection: if the state’s internal sovereignty is of fundamental moral significance, does this not entail that a state’s external sovereignty must be unlimited? If so, this would return us to the regressive view that, from an international point of view, states must be seen as ‘moral black boxes.’

In this final section, I address this objection by sketching a relational account of external sovereignty.Footnote ¹⁸ Following an old analogy between the independence of persons and of states, a relational account establishes a basic norm of reciprocal independence at the international level: states owe each other a basic duty of respect. But just as internal sovereignty is limited by the human rights of its members, so too external sovereignty is limited by human rights. Internal illegitimacy implies a forfeiture of external sovereignty and triggers a right on other states to intervene. Developing in detail the conditions for the permissibility or obligatoriness of intervention lies beyond the scope of this article. Instead, my aim here will be simply to show how a relational account can support a limited conception of external sovereignty and thereby avoid the envisaged objection.

Recall that (5) Internal Sovereignty maintains that a state S is internally sovereign if, and only if, S satisfies Relational Authority and Relational Legitimacy. Now, suppose that there is not just one but a plurality of internally sovereign public authorities. How should we understand their moral relations? Following an old internationalist tradition variously articulated by Hugo Grotius, Samuel Pufendorf, and Immanuel Kant, we may develop an analogy between the independence of persons and that of states: just as in interpersonal relations, so too in international relations each state owes the other a basic duty of respect correlative to each state’s basic right to independence.Footnote ¹⁹ Call this Reciprocal State Independence:

Reciprocal State Independence follows from (5), Internal Sovereignty. If a state is internally sovereign, then that state enjoys a right to independence from other states in the international arena.Footnote ²⁰ But a state’s internal sovereignty is not a carte blanche on the exercise of authority, for a state’s authority is constrained by its legitimacy. And we traced Relational Legitimacy ultimately to the idea of Reciprocal Independence and human rights. So a state with rightful domestic authority (an internally sovereign state) is a state with a right to independence in the international realm. This view has two important normative corollaries.

The first implication of (6), which I cannot discuss in detail here, is that it effectively bans any aggressive or punitive wars, for any such war would violate the basic duty of respect owed by every internally sovereign state to any other.Footnote ²¹ This places an important restriction on the use of force internationally. In general, liberal states, for instance, may not go to war to ‘punish’ illiberal states, say, due to the other state’s illiberal practices (e.g. discriminating against minorities or establishing a theocratic government).Footnote ²² As Immanuel Kant acutely noticed, European states in the sixteenth and seventeenth centuries used this presumed power to punish the ‘barbaric’ practices of communities in the newly discovered Americas as a subterfuge for colonialism.Footnote ²³

More importantly, for our purposes, a second corollary grows out of the fact that, as formulated, (6) says that the international duty of respect and the right to independence are only keyed to internally sovereign states. Generally respecting and adequately protecting human rights is a necessary condition for internal legitimacy. Transposed now to the international domain, this thought suggests that just as sovereignty is limited internally, so too it is limited externally. A state that systematically violates the human rights of its members forfeits its international right to independence and opens itself up to intervention by foreign states.

We may formulate this normative implication as follows:

Let me explain three key concepts in (7): the notions of a systematic human rights violation, of a right to intervene, and of intervention.

The notion of a systematic violation is crucial to (7), though it is difficult to render precise. So far, when discussing the notion of legitimacy, I have been talking about the state’s general respect and protection of rights. We may now distinguish a state’s general from strict compliance with the human rights of its members. Strict compliance means that the state has no failures to respect or protect the rights of its members. Strict compliance would be too stringent a criterion for legitimacy, for most (if not all) states currently in existence commit one violation or another. But this need not be morally problematic so long as the state in question has the institutional resources to rectify the wrong and prevent similar wrongs from occurring in the future. This suggests that general compliance is sufficient for legitimacy. A generally compliant state may fail to respect or protect the rights of its members on occasion, but still has institutional resources to check the continued violation, to prevent it in the future, and to provide adequate reparations. For example, suppose a state has been found to have failed to respect adequately the human right to health of some of its members due to discriminatory public health policies. But suppose, further, that the violation was identified by an independent domestic commission and that, upon learning of this finding, the state makes adequate efforts to change its public policies. Such a case would signal the presence of domestic institutional checks. In spite of its violation, the state may be said to be in general compliance with the human rights of its members.

The notion of a systematic violation of human rights, as I shall understand it, appeals to the notion of generic compliance, not to that of strict compliance. A systematic violation is one where generic compliance fails. Suppose now that in its fight with a minoritarian radical group, the state begins to perform forced disappearances. If the state’s own agents are involved in disappearing dissidents, one may not turn to them for protection. Such cases typically lack domestic institutional checks, for the domestic institutions are involved in the violation of the right. So this would likely count not only as a case of strict non-compliance, but also as one of general non-compliance and thereby as a systematic violation of human rights.

Since a systematic violation would suffice for illegitimacy, it would thereby entail a loss of internal authority. Correspondingly, it would trigger a right to intervene on other states. This appears a sensible implication, but it is a controversial one. (7) does not say that systematic violation triggers a duty to intervene.

Let me bring this out by contrasting (7) with Responsibility to Protect, R2P, a doctrine currently endorsed by the United Nations. This doctrine involves the following claim as one of its three core principles: ‘If a State is manifestly failing to protect its populations, the international community must be prepared to take collective action to protect populations, in accordance with the Charter of the United Nations.’Footnote ²⁴ Two features of this claim are noteworthy.

First, although not mentioned in my quote, the duty to protect in this context makes reference exclusively to genocide. This limits the responsibility to protect to cases of genocide, not to systematic human rights violations as such. In this way, (7) is far more demanding than current U.N. doctrine and limits external sovereignty on a far wider basis.

Second, this claim presupposes that the duty in question is collective rather than distributive. This is to say that the bearer of the duty is the international community as a whole rather than individual states. If there is a duty to protect, then the collective route is sensible. Otherwise, if each state had a duty to protect, failure to protect would mean that all states – perhaps subject to a capability proviso – would wrong the victims of systematic human rights violations by failing to protect them. This may be too demanding. By contrast, (7) suggests a distributive (and perhaps collective) right to intervene without going the extra step of making all states responsible for protection. This seems a modest but plausible and feasible proposal.

To be clear: (7) does not entail that there could not or should not be a duty to protect. (7) may be developed further in that direction. All I want to show here is that (7) is enough to address the objection that a relational account is committed to an unlimited conception of external sovereignty.

Having discussed the notions of systematic violation and of a right to intervene, let us turn to that of intervention. By ‘intervention’ I mean any use of force by which one state interferes with the authority of another. This notion of coercion is scalar rather than polar: it comes in degrees. In order of decreasing degrees of coercion, one state may intervene in the affairs of another by various means: full-scale military intervention, economic sanctions, or diplomatic sanctions. Sanctions count as coercive measures because they imply the use of force to intervene in the affairs of another state to get it to change its ways.

The proposal implicit in (7), then, is that there should be a correlation of proportionality between the degree of the systematic violation and the degree of force authorized. While the arbitrary detention of political opponents may warrant only diplomatic or perhaps economic sanctions, full-scale military intervention is likely to be disproportional and so unwarranted.

Of course, (7) leaves many fundamental questions unanswered, for it is not meant to provide a complete normative theory on the limits of the international use of force. Sample questions include: Does (7) recommend a scalar or a polar conception of legitimacy? Are there important variations in the quality and kind of systematic violation, such that it would affect differently the degree to which legitimacy and thereby authority is forfeited? If indeed there are significant variations in the degree of systematic violation, are there also significant variations in kind, thereby warranting different kinds of international interference? For instance, should an equally systematic violation of the right to vote trigger the same kind of right to intervention than a systematic violation of the right to life, when the state commits genocide? Presumably not. But then, how are we going to account for various kinds of importance of human rights? Are we warranted in not only holding that other states are permitted to intervene, but also in holding that they are obliged to do so? And if so, is this duty triggered by any kind of human rights violation or only by violations of a particularly egregious kind?

These are important questions that a relational account would have to address when developed further. But for now, the key point is that, contrary to the envisaged objection, a relational account does not commit us to the view that external sovereignty must be unlimited. Rather, it commits us to the plausible view that external sovereignty is indeed limited. To be sure, as I mentioned in the introduction, the view that state sovereignty is limited domestically and internationally is not novel. It is the core of the recent shift in the ordinary, pre-philosophical conception of state authority. What is novel, I propose, is the relational explanation and justification of this idea, for this proposal denies the basic and pervasive moral cosmopolitan view that state sovereignty matters only instrumentally. The relational account, then, gives us a distinctive and compelling picture of the idea that state sovereignty is limited.

Before closing, let me try to bring this picture into slightly sharper focus by contrasting it with the statist view defended by John Rawls. Since Rawls’s account has attracted considerable criticism, showing three key points of divergence should help preempt these familiar criticisms.

In the Law of Peoples, John Rawls appears committed to the following claims:Footnote ²⁵

• (A) The list of human rights is to be minimalist in order to exclude human rights with strong liberal or democratic dimensions.

• (B) A people’s good standing in the international society of peoples is a polar matter. Good standing entails a duty of toleration.

• (C) The primary function of human rights is circumscribing the use of international force.

Let me show how the relational account does not commit us to any of A–C and so avoids the criticisms A–C are thought to attract.

First, a statist account like Rawls’s, as evidenced by A, has attracted the objection that it ends up committed to too narrow a conception and list of human rights.Footnote ²⁶ Yet, nothing in the relational account commits us to this view. As I have argued elsewhere, a relational account can generate a robust and comprehensive list of human rights, encompassing most (if not all) rights enshrined in the Universal Declaration of Human Rights. To be forceful, in any case, this objection would have to be tailored to some specific human right. As a general worry, it musters no force.

Second, Rawls’s account has attracted the criticism that it is too tolerant of illiberal regimes and that, by making a people’s good standing an on–off matter, it misses some of the most important and interesting forms of human rights protection and promotion: the criticism of human rights violators. As Tan puts it, toleration for Rawls ‘is not just non-intervention; it is, very importantly, also non-criticism’ (Tan Reference Tan, Martin and Reidy2006, 81). So Rawls seems committed to the implausible view that no state may criticize another state that violates some human rights below the threshold required to lose good standing. Similarly, Nickel argues persuasively that what he calls ‘jawboning, ’ the criticism and condemnation of other countries that is not accompanied by significant threats, amounts to ‘the most common means of promoting human rights across international borders’ (Nickel Reference Nickel, Martin and Reidy2006, 271). Rawls’s theory, Nickel argues, misses this important dimension.

The relational account can accommodate both points. (7) sets the threshold for the permission of force in the systematic domestic violation of human rights. The relational account can also support criticism and jawboning since neither of these forms of interference are coercive. So just as the degree of force permitted must be proportional to the seriousness of the systematic violation, so too criticism and jawboning may be the warranted, proportional forms of non-coercive interference warranted by violations of human rights that are short of systematic.

And third, although the relational account, with Rawls, maintains that one of the key functions of human rights is circumscribing the international use of force, it does not maintain that this is the only or the main function. Human rights set, as we have seen, conditions both for internal and external sovereignty. In so doing, they can also function as standards for guiding legislation, standards for education, standards of criticisms of governments by their own citizens and NGOs, or standards for evaluating the suitability of possible candidates for foreign aid. Human rights, then, do not just set conditions of political legitimacy. On the relational account, they articulate the conditions required for realizing a social world under which we can relate to one another as equally independent. So understood, human rights capture some of our basic ideals about life in society.

In sum, it should be evident that the account offered in this section stops far short of offering a complete picture of the function of human rights in the international arena. Still, the relational account seems to offer a novel and compelling account of the limited external sovereignty of states, one that also avoids some of the main criticisms of more familiar statist views.

Funding

This work was supported by the SSHRC [grant number 756-2013-0761].