Editor's note: Professor Dworkin died while this article was in the late stages of revision. The editors would like to thank Liam Murphy and Samuel Scheffler for their help in finalizing the article for publication on Professor Dworkin's behalf.
A New Philosophy for International Law
Article first published online: 11 APR 2013
© 2013 Wiley Periodicals, Inc.
Philosophy & Public Affairs
Volume 41, Issue 1, pages 2–30, Winter 2013
How to Cite
DWORKIN, R. (2013), A New Philosophy for International Law. Philosophy & Public Affairs, 41: 2–30. doi: 10.1111/papa.12008
- Issue published online: 11 APR 2013
- Article first published online: 11 APR 2013
When I was last instructed in international law—at Oxford in the 1950s—the first and most lively question, bound to appear on the examination paper together with tedious questions about navigable bays, was existential. Is there any such thing as international law? Or does the subject we were being asked to study actually not exist? Is what some academics and state department officials call international law not law at all but only common practices that no state is really legally bound to continue practicing?
You may well ask: Why does this matter? Isn't the only important point really whether there are rules that nations do follow in their dealings with one another? And that nations join in criticizing and if possible punishing other nations that do not follow those rules? Does it really matter whether we call these rules “law”? Or whether we say, instead, that in certain ways they are like the rules of more familiar national legal systems?
The question whether it matters is deeper and more difficult than may first appear. I reserve it for discussion later. In any case, the question whether there is international law seems no longer to trouble anyone. Almost everyone assumes that there is international law and also assumes that it includes, for example, the Charter of the United Nations and the Geneva Conventions—or at least some of them. But nothing has actually changed. The old grounds for challenge remain; they are only ignored.
The existential challenge remains important, however. Even though almost everyone agrees that “international law” is really law, and that the rules and principles set out in documents of that kind are part of it, the question of why these documents constitute some kind of legal system is crucial because how these rules and principles should be interpreted hinges on it. Interpretive issues are both controversial and dramatically important. Nations and lawyers disagree, for instance, about the legal status of associates of Al Qaeda and the Taliban under the Geneva Conventions, and whether there is such a thing as an enemy noncombatant who is not covered by those Conventions. I will later discuss another celebrated interpretive issue: whether the NATO intervention in Kosovo, without the consent of the Security Council of the United Nations, was a violation of international law.
First, however, we should notice why many people did doubt, half a century ago, that there was any such thing as international law. This was not because the rules and practices were very different from what they are now, but because a certain philosophical theory of what law is, called “legal positivism,” was more popular. This theory holds that whether a law exists is fundamentally a question of historical fact. Law exists only when some person or group has created that law. Legal philosophers who regard themselves as positivists have disagreed about who those law-making people are, and how they make law. Different answers to these questions have been influential in different times. John Austin, a nineteenth-century legal philosopher, answered by proposing a definition of law: law is by definition, he said, the command of an uncommanded commander, a sovereign with absolute power over some territory. If Parliament, with the Queen's consent, has unlimited power, the Queen in Parliament is the uncommanded commander in the United Kingdom. But since there is no such sovereign body commanding the parliaments of all the nations, it seems to follow from Austin's theory that there is no international law.
By the middle of the last century, however, another legal philosopher, H.L.A. Hart, had introduced a more sophisticated version of positivism. He denied that law always depends, as Austin had said it did, on the commands of an uncommanded commander. In the United States, Hart pointed out, no institution is such an absolute sovereign. He described, instead, a more general set of social facts that give rise to law. He said that law exists when the “bulk” of the officials of a political community have come to accept, as rules they have an obligation to follow, two kinds of rules: “secondary” rules, which stipulate how law is created, enforced, and identified, and “primary” rules, which are created and identified when those secondary rules are followed. Hart insisted that one secondary rule, which he called the community's “rule of recognition,” serves in any legal system as the fundamental test of all the rest of the secondary and primary rules of that system. His theory thus preserved the core principle of legal positivism: what the law of a community actually is depends on nothing more than a contingent aspect of its social and political history. Political or personal morality has nothing to do with it.
Notice that under any version of legal positivism the law of any particular political community is bound to have “gaps.” Written and spoken defamation are subject to different rules of civil damages, and no official may have declared, either way, whether false statements videotaped at a political rally should be treated as written or spoken. So those whose job it is to enforce the law—judges, for example—must have what legal positivists call “discretion” to fill in the gaps by legislating, retrospectively, themselves.
Hart himself raised the question whether so-called international law really counts as law on his new test. However, though he phrased that question in the traditional way, he actually changed the subject.1 He asked a question for social scientists: whether there is any system of practices that can sensibly and usefully be described, for their sociological or anthropological purposes, as international law. That is very different from the doctrinal question posed to the lawyers and judges who practice international law: the question, for instance, whether the intervention in Kosovo was legal under international law. Hart approached his sociological question by conceding, first, that the distinction between primary rules and secondary rules could not be made in the international realm. He found nothing comparable to what he understood as the familiar secondary rules of domestic law—general rules of legislation, for instance—in the international sphere.
That was nevertheless not decisive, he said, of the question whether it would be helpful for theoretical and practical purposes to include international law within the more general concept of law. He suggested, at least, that it might be. His analysis was therefore like the recent discussions among astronomers whether it would be sensible to continue to use the word “planet” in such a way as to make Pluto a planet.2 But interpretive doctrinal questions, such as whether the Kosovo intervention was legal under international law, cannot be answered by considering whether it would be useful to speak of an international law. For such doctrinal questions, we need an account that helps us decide not whether it would be useful to speak of international law as a sociological category, but what international law holds on particular issues.
Many contemporary international lawyers have tried to do what Hart did not: construct a doctrinal account of international law from his version of positivism.3 They assume that a sovereign state is subject to international law but, on the standard account, only so far as it has consented to be bound by that law, and they take that principle of consent to furnish an international rule of recognition. This is a firmly positivist view of international law because whether a state has consented to a particular rule is just a matter of history. Positivism in that version seems to be now generally accepted by practitioners and scholars of international law. Contemporary textbooks and manuals of the subject (at least those I have consulted) uniformly cite Article 38 (1) of the Statute of the International Court of Justice, established by the United Nations, which they take to state an international rule of recognition. This Article reads:
The Court, whose function is to decide in accordance with international law such disputes as are submitted to it, shall apply:
- international conventions, whether general or particular, establishing rules expressly recognized by the contesting states;
- international custom, as evidence of a general practice accepted as law;
- the general principles of law recognized by civilized nations;
- subject to the provisions of Article 59, judicial decisions and the teachings of the most highly qualified publicists of the various nations, as subsidiary means for the determination of rules of law.4
True, international lawyers also speak of what they call ius cogens, or “peremptory norms” that cannot be canceled by treaty or even by decisions of the United Nations. However, the Vienna Convention on the Law of Treaties brings these, too, under the umbrella of consent:
For the purposes of the present Convention, a peremptory norm of general international law is a norm accepted and recognized by the international community of States as a whole as a norm from which no derogation is permitted and which can be modified only by a subsequent norm of general international law having the same character.5
Law for nations, on this view, is grounded in what nations—or at least the vast bulk of those that others count as “civilized”—have consented to treat as law. Signatories to a treaty are assumed to have consented to its provisions being law for them just by their signature. States that have assumed, in their practices, that certain rules are law for them have in that way consented to the rules being law for them. If enough states to constitute “the international community of States” have recognized fundamental rules as preemptory and nonnegotiable, then these rules are preemptory and nonnegotiable for the whole international community. If there is evidence that a general practice is very widely accepted as law, or that it is recognized by all civilized nations, then it is law for all nations. The scheme has one apparent advantage. Since it bases law on consent, it solves an apparent paradox born of the modern state system. How can a sovereign state nevertheless be subject to law? It answers: because it (or at least almost every state) has accepted, in the exercise of its sovereignty, to be bound by that law.
But the scheme has several defects as a proposed rule of recognition that are finally fatal. First, it offers no priority among the different sources it recognizes. Must treaties yield to general practices? Or vice versa? More important, though it is founded on the idea of consent, it sometimes binds those who have not consented. It offers no explanation why states that have not accepted a rule or principle as law may nevertheless be subject to it because the bulk of other states, or of “civilized” states, have accepted it. It offers no standard for deciding how many states must accept a practice as legally required before the practice becomes “customary” and therefore binding on everyone. It offers no guidance as to which states are sufficiently civilized to participate in that essentially legislative power. Or which norms are peremptory. These latter difficulties stem from the scheme's perfectly understandable ambition to extend the ambit of international law beyond those communities that have explicitly consented to its principles to include those that have not. International law could not serve the purposes it must serve in the contemporary world—disciplining the threat some states offer to others, for example—unless it escaped the straitjacket of state-by-state consent. But yielding to that ambition seems to undermine the axiomatic place of consent in the scheme, and thus its assumed jurisprudential foundation.
However, I shall set that unsolvable problem aside for now to concentrate on difficulties that infect even the core of the scheme—the propositions that treaties create law for signatory nations and that the constraints that nations have accepted as law in their practices and statements are thereby made law for them. We should notice, first, that the interpretive strategies licensed by this jurisprudential core are particularly unhelpful. If a constraint is part of international law for particular nations only because they have consented to it in either of those ways, then the master interpretive question must be: what is it most reasonable to assume that these nations, whose consent made the principle law, understood that they were consenting to? That question may in many cases be answered satisfactorily by the plain meaning of the text (though interesting issues may arise about translation). But in many cases, the text will not be decisive.
Here is an example I mentioned earlier as an important question of interpretation. Article 2 (4) of the United Nations Charter provides that:
All members shall refrain in their international relations from the threat or use of force against the territorial integrity or political independence of any state, or in any other manner inconsistent with the Purposes of the United Nations.6
How should we understand the phrase “territorial integrity or political independence” in this provision? Does a humanitarian intervention undertaken by a group of states, as by NATO in Kosovo (or Libya), violate territorial integrity or political independence if its sole aim is to stop genocide or crimes against humanity without any change in boundaries or constitution? The most natural reading says yes: any invasion or bombing violates territorial integrity. But some prominent international lawyers have disagreed,7 and there is room for disagreement. Does dropping humanitarian aid packages over a disaster area violate the territorial integrity of a state whose government has not approved it? Drones carefully targeted to kill a terrorist chief without the government's permission? Does territorial integrity require an effective government in place over a defined territory? Perhaps there is no territorial integrity to be respected in Syria as I write, so that intervention would not violate Article 2 (4).
It seems very unlikely that all the states that created the United Nations in 1945, or that joined that organization since, shared answers to all these questions when they joined. It also seems unclear whose opinion, among the different officers or citizens of these states, counts as manifesting a state opinion. Nor has there been sufficient practice by nations or statements by their foreign ministries to provide a firm answer. Nor does there seem any disposition among states to accept, in the spirit of a positivist approach to law, that a body applying international law, like the International Court, should be deemed to have discretion to impose whichever answer it wishes.
If the theory that consent is the ultimate basis of international law were persuasive, then we would quickly come to an interpretive dead end on such questions. Fortunately it is not persuasive, even if we set aside the difficulty that sometimes nations are treated as bound by rules to which they have not consented. There are more fundamental problems. Consider, first, the proposition that international law is created for nations, without any formal treaty, when they accept that certain constraints on their acts and policies are required not just by decency or prudence but as a matter of law. This assumes that in some way nations decide for themselves whether some constraint they accept is imposed as a matter of law and not just decency. What principle—what “rule of recognition”—do they supposedly follow in making that discrimination? It won't do to say that they follow the principle that what they regard as law is law. They need some other standard to decide what they should regard as law.
Suppose we say: they accept the principle that what other nations accept as law is law. But then the other nations that each nation treats as making law for it need a test of what to treat as law for themselves. Our explanation must break out of the circle somewhere. Suppose we say: the requirement means that law is created by convention, by the fact that each nation accepts some constraint because and only so long as other nations do. But not all conventions generate obligations; there are many conventions of convenience that people are morally free to disregard when they wish so that the convention ends. When do conventions create legal obligations? The idea of customary law presupposes that there is some different, more basic principle at work, in the identification of international law, or at least that the subjects of international law think there is some such principle at work. We need to ask: what is that more basic principle? If we find an answer, it is that more basic principle, not the fact of consent, that provides or is thought to provide the grounds of international law.
Now consider the claim, even more fundamental for the consent thesis, that treaties create international law for the parties to those treaties. Treaties are signed at a particular time: the all-important United Nations Charter nearly seventy years ago. Nations change dramatically over such periods of time. Boundaries change, regimes and constitutional structures change. We personify states when we treat them, rather than their citizens, as the subjects of international law, and we might therefore be tempted to say that just as individual people are bound by promises long after they make them, so are states, in spite of all these changes. But the fiction of a continuing national person, as distinct from its structure of government and its individual citizens, cannot bear that weight. It seems unfair that people should suffer serious disadvantage only because politicians chosen by entirely different people under entirely different constitutions signed a document many generations ago. We cannot justify that disadvantage by any analogy to the law of contract: contracts cannot bind people not parties to them. True, the domestic law of some states makes treaties a continuing obligation of the state. The American constitution, for example, declares treaties part of “the Supreme Law of the Land.” But what domestic law creates it can destroy: a state would not be bound by international law if it were free, through its domestic legal processes, to unbind itself.
We need an explanation why the citizens of contemporary Ruritania have an obligation under international law that cannot be canceled by any new Ruritanian political process. It does not serve to declare that international law contains a more basic principle—pacta sunt servanda—that treaties must be respected over generations. What makes that more basic principle part of international law? It would, once again, be circular simply to reply that states consent to that principle when they sign treaties. Compare the familiar institutions of promising. As many philosophers have pointed out, there is mystery in the bare assumption that promising creates obligation. How can an individual change his moral situation just by speaking a runic phrase? If we want to explain why promises do create moral obligations, we must point to different, more basic moral principles that a promise invokes. Philosophers have suggested a variety of such principles.8 We must look for similar, more basic principles within international law.
I draw this conclusion: we cannot take the self-limiting consent of sovereign nations to be the basic ground of international law. The temptation to do so is understandable. It makes international law compatible, as I said, with the doctrine of state sovereignty. It also resonates with a very popular conception of political legitimacy: that coercive dominion can be justified only by the unanimous consent of those subject to that dominion. That conception of legitimacy generated the social contract tradition in political philosophy and the artificial conceptions of consent that were necessary to sustain that tradition. I have argued elsewhere that these accounts all fail and are anyway unnecessary because consent is neither a necessary nor a sufficient ground of legitimacy.9 We must locate the source of political obligation elsewhere: in my view, we must locate it in the more general phenomenon of associative obligation.
Return to the distinction I drew earlier between two concepts of law: a sociological concept, useful to social scientists deciding how most conveniently to classify law as one type of system of social control, and a doctrinal concept that figures within the operation of such systems by allowing people to invoke a special kind of right or obligation. We share the sociological concept as what I have called a criterial concept: we can sensibly agree or disagree in the application of such concepts because we share roughly the same criteria of application. The concept of a triangle is also a criterial concept: we share that concept because we use the same test—a three-sided figure—for deciding what is or is not a triangle. Some criterial concepts are vague, however. We share roughly the same tests of application for the concept of a book, but in some cases these shared tests are not decisive. We may, if this proves convenient either practically or theoretically, agree to stipulate a more precise sense of “book.” That is what astronomers did for the concept of a planet, and what Hart proposed to do for the concept of international law.
The doctrinal concept of law is very different. It is not a criterial but an interpretive concept: we share it not by agreeing about tests for application but by agreeing that something important turns on its application and then disagreeing, sometimes dramatically, about what tests are therefore appropriate to its use, given that its application has those consequences.10 Any theory about the correct analysis of an interpretive political concept must be a normative theory: a theory of political morality about the circumstances in which something ought or ought not to happen. Since the doctrinal conception of law is interpretive, we provide a theory of the grounds of law by posing and answering questions of political morality.
That, to my mind undeniable, fact poses the most general problem of jurisprudence. We know that there is a difference, often profound, between what the law is and what it ought to be. But if what the law is itself depends on a moral theory, then how can we make that distinction? In Justice for Hedgehogs, and earlier work, I offered this answer (which I here state roughly): we identify the law of a community by asking which rules its citizens or officials have a right they can demand be enforced by its coercive institutions without any further collective political decision. Americans have the right, on demand, to the benefits that past congressional legislation, properly interpreted, awards them. They may also have the right that future legislation improve those benefits, but they have no right that the coercive force of the state be used to secure those further benefits unless and until that future legislation is actually adopted. We count the former right a legal one; we count the latter, if it exists, a moral one. We articulate law, then, as part of political morality—but very much a distinct part. We ask: which political rights and obligations of people and officials are properly enforceable on demand through institutions like courts that have the power to direct coercive force? That is a moral question whose answer is a legal judgment.
This very abstract account of the relation between law and morality must not obscure how the distinction actually works in practice: in how judges and academic lawyers identify particular rules of law in concrete circumstances. They do not ask themselves basic questions of political philosophy about which rights are properly enforceable on demand. They begin in answers to those questions that they take to go without saying. They agree, in America and most other places, that only the political constitution, legislation pursuant to that constitution, and past judicial decisions can create rights enforceable on demand. There is often controversy about how particular constitutional clauses, statutes, and precedents should be interpreted, but that controversy does not challenge the sole authority of these sources.
The value of the abstract account lies in the possibility it provides of justifying—or challenging—these settled working assumptions. We justify them—if we can—through a political theory that combines an attractive conception of political legitimacy together with a convincing conception of the special political virtue of fairness, one that makes history, convention, and expectation particularly pertinent to the identification of rights that are enforceable on demand, and to the isolation of those rights from other political rights, including some that contradict them, that are not enforceable on demand. (Some purported legal systems cannot be justified in that way. Then we face the complex interpretive question whether such systems give rise to any genuine legal rights.)11 That is much better than simply taking the settled practices as brute facts or seeking some mythical social-fact rule of recognition they all supposedly exemplify.
It is better, among other ways, because these justifications cast an interpretive shadow. They encourage theorists to refine their theories of constitutional and legislative interpretation, for instance, by asking which interpretive methods best serve democracy and fairness so understood. I do not mean, to repeat, that judges interpreting statutes must explain, in each case or ever, why the best theories of democracy and fairness support their interpretive methodology. Their training and experience, supplemented, we might hope, by some academic curiosity, will form their working and largely unexamined methods. It does mean, however, that those judges who interpret the most critical constitutional clauses, particularly those who write books explaining their methods, should be more sensitive to these questions than they seem to be.12
How far can we construct an international jurisprudence on the same understanding? How far can we treat international law as a part, but a very distinct part, of what morality and decency require of states and other international bodies in their treatment of one another? We must abandon the positivistic, supposedly consent-based jurisprudence of international law: that is flawed beyond redemption. We should return to what I take to be a golden age of the subject, seventeenth-century European politics, to an at least partially moralized conception of international law. But we face a problem. We can draw that distinction easily for national legal systems because we find institutional structures there that provide an appropriate vocabulary. These structures broadly distinguish between courts, which have the responsibility and power to enforce rights and obligations on demand, and other sorts of political institutions, like legislatures, that do not. So we can helpfully frame our basic political question in institutional terms: we can ask what rights courts have the responsibility and right to enforce. But no such structure, in any but the most rudimentary form, is yet in place in the international domain, and none can be expected soon.
Here is my suggestion. Let us imagine (though initially not in much detail) an international court with jurisdiction over all the nations of the world. We imagine that cases can be brought before that court reasonably easily and that effective sanctions are available to enforce the court's rulings. Of course that is fantasy upon fantasy, at least for the far foreseeable future. But bear with me.
If we can imagine such a court, even as fantasy, then we can frame a tractable question of political morality. What tests or arguments should that hypothetical court adopt to determine the rights and obligations of states (and other international actors and organizations) that it would be appropriate for it to enforce coercively? This is a moral question but a special one because judicial institutions with compulsory jurisdiction and sanctions at their disposal are subject to special moral standards of legitimacy and fairness. They have no right to declare and enforce general standards of comity, decency, or wisdom. We can identify a general theory of what it would be appropriate for such an institution to enforce as the foundation of international law.
Is it a serious objection to this counterfactual exercise that there is not—and in indefinite foreseeable circumstances cannot be—a court of that character? No international court can now deploy effective coercion without the cooperation of powerful nations who would, as a practical matter, refuse to submit to a court of the power we are imagining. I offer the counterfactual exercise only as a way of providing a scheme for identifying international law, not, at least in the first instance, as a way of persuading anyone to accept that law. Of course, it would be an important part of the exercise whether a hypothetical court would be right in endorsing its own hypothetical authority. Perhaps there are sound reasons of political legitimacy why such a court should not exist. In the next section I will explain why I think it would be legitimate, and answer in a very general, abstract way the question of what standards and methods it would be appropriate for it to adopt if it did exist.
But, in advance, we should distinguish two issues about the practicality of a theory of international law. First, is it sensible to try to develop a theory of the grounds of international law that is unlikely to be accepted by powerful nations because their power would be limited if they accepted it? It seems so, for various reasons. First, even powerful nations now claim to defer to international law: they appeal to their conception of what that law requires or permits to justify their actions. The Bush administration repeatedly declared that international law permitted its treatment of terrorist suspects, for example. It would be important to undermine such claims by showing that a much more persuasive account of international law contradicts them. Second, a time may come, sooner than we suppose, when the need for an effective international law is more obvious to more politicians in more nations than it is now. Climate change, for example, may provoke that shift in opinion. It would be a shame if lawyers and philosophers had not improved the jurisprudential discussion of international law before that day arrived. If the standing theories of international law are radically defective, as I have suggested they are, we have at least an intellectual responsibility to propose a better one.
Second, does the fact that a legal theory is unlikely to be generally accepted soon show that it is not only impractical but wrong? There are two reasons why we might think so. The first is conceptual: if we accept the positivist account of international law, which bases law on consent, then of course a theory to which almost no one is likely to consent is obviously a false theory. But we reject that account. The second is internal: a better theory, which grounds international law on moral principle, may show that a particular claim of international law is unsound if there is no prospect of general endorsement. Later in this article I defend the importance of a principle of salience in international law. In some circumstances, this principle makes the authority of more concrete principles depend on the prospect of wide acceptance. The effect of this second kind of impact of acceptance on law is retail, not wholesale. We must wait and see.
Most conventional international law treatises begin by describing the traditional subjects of international law.13 The world is divided, in the conventional story, into sovereign states, each of which is in principle immune from interference by other states. None of these states is permitted to impose rulers on any of the others, or to dictate its religion or laws or policies. The sovereign power of each of these states can be limited only by the voluntary acts of its own institutions. This regime is often called the “Westphalian” system of international order; the name assumes that the system was initially established in a series of seventeenth-century European treaties grouped under that name. The Westphalian system was created out of the interests of hereditary monarchies intent on substituting economic competition for the bloody religious conflicts that had marked the previous century.
It is important to notice that the system balkanized not only sovereignty but political legitimacy as well. In and after the seventeenth century, political philosophers, statesmen, religious leaders, and revolutionaries asked fresh questions about political power. The old assumption that hereditary monarchs have an absolute right to govern, at least in the temporal sphere, was gradually replaced by a starkly different assumption: that coercive political power is consistent with the dignity of citizens only if it can be justified not just in pedigree but in substance—in the way it is exercised—as well. Competing theories of legitimacy were constructed and debated; these finally settled into theories about the best conceptions of democracy and of the rights of individual citizens in a democracy. But all these theories were confined to arrangements within sovereign states. John Rawls offered his influential theory of justice as limited to the basic structure of an individual state, for example.
However, the modern question—what justifies coercive political power?—arises not just within each of the sovereign states who are members of the Westphalian system but also about the system itself: that is, about each state's decision to respect the principles of that system. For those principles are not independent of but are actually part of the coercive system each of those states imposes on its citizens. It follows that the general obligation of each state to improve its political legitimacy includes an obligation to try to improve the overall international system. If a state can help to facilitate an international order in a way that would improve the legitimacy of its own coercive government, then it has a political obligation to do what it can in that direction. Of course that obligation demands only what, in the circumstances, is feasible. It does not require any state to ignore the division of the world into distinct states and suppose that it has the same responsibilities to citizens of other nations as it has to its own. But it does require a state to accept feasible and shared constraints on its own power. That requirement sets out, in my view, the true moral basis of international law. It therefore also states the basic interpretive principle that the hypothetical court I imagined should use in deciding what international law now requires.
We should now notice the different ways in which individual states fail their responsibilities to their own citizens when they collectively accept the benefits and burdens of the pure unrestricted sovereignty that the Westphalian system gives them: when they accept, that is, that states are forbidden to interfere in or with the conduct of other states, come what may. A coercive government is of course illegitimate if it violates the basic human rights of its own citizens. Any state, even one that has so far been just and benign, therefore improves its legitimacy when it promotes an effective international order that would prevent its own possible future degradation into tyranny. Germany had an obligation to facilitate an international order—a more effective League of Nations, perhaps—that could have prevented its Nazi future. So did the United States, which signally failed in that obligation.
A state is also defective in its legitimacy when it cannot protect those over whom it claims a monopoly of force from the invasions and pillage of other peoples. Any state therefore has a reason to work toward an international order which guarantees that the community of nations would help it to resist invasion or other pressure. There is a mirror-image problem in the Westphalian system that threatens legitimacy in a different way. People around the world believe they have—and they do have—a moral responsibility to help to protect people in other nations from war crimes, genocide, and other violations of human rights. Their government falls short of its duty to help them acquit their moral responsibilities when it accedes to definitions of sovereignty that prevent it from intervening to prevent such crimes or to ameliorate their disastrous effects. Together these duties call for a more complex regime governing collective intervention in the affairs of individual states that that system provides.
Governments fail their citizens' legitimate expectations in a third and less obvious way when they accept an international system that makes impossible or discourages the international cooperation that is often—and increasingly—essential to prevent economic, commercial, medical, or environmental disaster. People are subject to the constant risk of what philosophers call “prisoners' dilemmas”: circumstances in which it is rational for them one by one to do something—drop litter in the park—that ends in loss for them all—a park destroyed. These situations pose difficult challenges of coordination. Governments can and do respond to such challenges, when these can be solved locally, by adopting and enforcing laws, by making littering a crime, for instance. But some problems—overfishing of the seas, for example, and pollution of the atmosphere with carbon—cannot be met by governments each acting only for its own territory. People in the separate states need the protection that only a coordinated policy backed by all or nearly all governments can provide. But an unmitigated Westphalian system simply repeats the dilemma at the international level. If no state can be forced to cooperate, they will all have a reason not to participate.
The legitimacy of coercive government requires, fourth, that people play some genuine, even if minimal and indirect, role in their own government. Political theorists disagree about what kind of participation is essential in different forms of government, but it is generally understood (even if this is far from universally provided) that some form of widespread suffrage in the election of officials is both necessary and sufficient within a distinct political community. In a world of strong and increasing economic interdependencies, however, people's lives may be more affected by what happens in and among other countries than by what their own community decides. Dignity seems to require that people everywhere be permitted to participate in some way—even if only in some minimal way—in the enactment and administration of at least those policies that threaten the greatest impact on them. An unmitigated Westphalian system makes that impossible.
These are all ways in which the unchecked state sovereignty system impairs or threatens the legitimacy of the individual states that make up the system. But since each of those states derives its moral title to govern a particular territory from the arrangements that make up that international system, it therefore has the further, independent reason, different from those I just listed, for concern that the system on which its legitimacy depends in that more fundamental way is not itself illegitimate.
Coercive government (I include not just traditional “sovereign” states but also any institution or organization claiming coercive authority) has a standing duty to improve its own legitimacy. Each traditional state therefore has a duty to pursue available means to mitigate the failures and risks of the sovereign-state system. That duty of mitigation provides the most general structural principle and interpretive background of international law. But as it stands, it is not sufficiently determinative. In many circumstances, a number of very different regimes of international law would each serve to improve the legitimacy of the international system, were it enacted and enforced, and states may reasonably disagree about which would be best. That obvious fact explains a further fundamental structural principle. This is the principle of salience: If a significant number of states, encompassing a significant population, has developed an agreed code of practice, either by treaty or by other form of coordination, then other states have at least a prima facie duty to subscribe to that practice as well, with the important proviso that this duty holds only if a more general practice to that effect, expanded in that way, would improve the legitimacy of the subscribing state and the international order as a whole. If some humane set of principles limiting the justified occasions of war and means of waging war gains wide acceptance, for instance, then the officials of other pertinent nations have a duty to embrace and follow that set of principles. The Universal Declaration of Human Rights so declares in the final “whereas” of its preamble: “a common understanding of these rights and freedoms is of the greatest importance for the full realization of this pledge.”14 The salience principle has an obvious snowballing effect. As more nations recognize a duty to accept and follow widely accepted principles, those principles, thus even more widely accepted, have greater moral gravitational force.
In the seventeenth century, salience was provided by two traditions. The first was the political force of Christianity. The Westphalian system was European and Europe was Christian. Church teaching, so far as it was pertinent, could be treated as the spine of developing international law—of the rules of jus ad bellum and in bello, for example. The second was the idea, inherited from imperial Rome but now put to different uses, of a ius gentium of legal principles common to nations across the Westphalian system. Early international law reflected both influences. It incorporated the natural law tradition developed through Aquinas, and it recognized principles widely shared by domestic legal systems, a recognition fossilized now in paragraph (c) of Section 38 (1) of the International Court Statute I quoted.
The world that emerged from World War II was very different. There was no longer a dominant religious tradition across the world, and widespread secularism in Europe and North America would have negated the influence of any such tradition anyway. War and economics had made different nations with different legal cultures and traditions, including the Soviet Union and China, crucial to world order. Reliance on shared legal principle was no longer possible. The retreat from colonialism that left many new or newly independent nations behind made such reliance even less useful. Some new focus of salience was needed and was quickly provided in San Francisco.
The charter and institutions of the United Nations are best understood not as arrangements binding only through contract or on signatories but as an order all nations now have a moral obligation to treat as law. The obligation is created not by consent but by the moral force of salience as a route to a satisfactory international order. Indeed—more generally—multilateral agreements setting out conceptions of such an order, like the Charter, the Geneva Conventions, the genocide agreements, and the Treaty of Rome establishing the International Criminal Court, are made international law for all, not just their initial signatories, through that principle. It is therefore important to distinguish the force of such multinational treaties, and the appropriate interpretive strategies for them, from that of agreements creating international organizations, like the European Union and the WTO, that are designed from the start for only a club of signatory nations and members later expressly admitted, with institutional procedures that cannot sensibly be used outside that club.
In that way the salience principle explains the contemporary as well as the ancient role of ius gentium in international law. It also helps to explain the domestic use of that idea. It explains, for example, why the constitutional courts of separate nations are (and should be) drawn to notice and to attempt to achieve some integrity with the constitutional principles of other nations.15 The debate among Supreme Court justices in the United States about whether that court should cite foreign legal materials looks silly when the practice is defended by its proponents as simply providing helpful suggestions, as a law review article might, that judges are free to accept or disregard. Who could sensibly object to that? The practice becomes more consequential when the responsibility of individual nations to seek a ius gentium is noticed. International order is strengthened as the “general principles of law recognized by civilized nations” grow more uniform. Interaction between the international and the domestic laws of human rights is particularly important for that reason.
The principle of salience provides a better account of the sources of international law set out in Article 38 of the International Court Statute than the consent theory can offer. We noticed earlier that the consent theory made a good part of Article 38 circular and unhelpful. The salience principle provides much better support for the sources that are named in that Article, though with the important proviso I mentioned added. It supports them individually: it explains the demands of customary international law, for instance. The more general principle I called the principle of mitigation, from which the salience principle flows, itself explains the idea of a ius cogens.
Salience explains the great popularity of Article 38—the consensus among scholars that it correctly states the grounds of international law—in a further way as well. The article not only sets out the implications of the principle of salience but is itself a beneficiary of the principle. Its provisions are self-confirming: it contributes to international order to continue to treat those provisions as sources of international law. Snowballing works at that level as well. According to the positivist account that makes consent fundamental, these sources flow—imperfectly—from the very idea of law as based in consent. On the account I describe, they flow instead from the moral demands, on which the legitimacy of an international system depends. These are taken to be more fundamental than consent and not contingent on consent.
If the two jurisprudential accounts end in roughly the same view of the actual sources of international law, does it make any difference which we choose? Have I only marched you up the hill and then marched you down again? No, because, as I said, the major yield of any theory about the grounds of international law is an interpretive strategy for international law. The consent account, I said, yields no helpful strategy. But the legitimacy account does. We should interpret the documents and practices picked out by the principle of salience so as to advance the imputed purpose of mitigating the flaws and dangers of the Westphalian system. The correct interpretation of an international document, like the UN Charter, is the interpretation that makes the best sense of the text, given the underlying aim of international law, which is taken to be the creation of an international order that protects political communities from external aggression, protects citizens of those communities from domestic barbarism, facilitates coordination when this is essential, and provides some measure of participation by people in their own governance across the world. These goals must be interpreted together: they must be understood in such a way as to make them compatible.
I have now twice mentioned one example of an important interpretive question. It has been very widely assumed by distinguished international lawyers, including the late Tom Franck, that no humanitarian military intervention is legal under international law unless it has been approved by the Security Council. But the Security Council is often crippled by the power of each permanent member to veto even otherwise unanimous decisions. That power allows what should be an essentially legal decision—does a violation of human rights justify intervention?—to be distorted by considerations of political and economic advantage. A permanent member might, for example, seek favorable economic treatment in Africa by promising its veto in aid of dictatorial regimes.
On occasion, however, states or groups or international organizations have intervened in force without Security Council authorization. The Iraq invasion is a minatory example. The United States did claim that it had Security Council permission—as did the United Kingdom, in spite of its Attorney General's initial opinion to the contrary. But that claim was spurious. The invasion is almost unanimously condemned as illegal in the wider international community. But the intervention by NATO forces in Kosovo is, on the contrary, widely even if not unanimously approved. No one claimed Security Council authorization for that intervention. Tom Franck declared it illegal because it lacked that authorization. But he also declared the intervention morally necessary: he called it a morally mandatory act of international civil disobedience.16 That is a dangerous description, particularly from an eminent international lawyer. International law is fragile, still nascent and in critical condition. The proposition that a sense of moral duty can justify violations of international law threatens to strangle the child.
I mentioned another possibility: that the popular interpretation of Article 2 (4) of the United Nations Charter, on which Franck relied, is mistaken. Perhaps we should understand its prohibition to be limited to the use of military force aimed at territorial change or political dominion. We might argue to that conclusion in several ways. We might understand the “Purposes of the United Nations” cited in Article 2 (4) to be those that flow from the moral responsibility nations had to create that institution: the responsibility to protect people from the dangers of the insulated sovereignty of the Westphalian system. External aggression is one of those dangers, but internal terrorism is another, and we can sensibly attribute protection from both dangers as among the United Nations' purposes. That understanding is strengthened by the General Assembly's early (1950) Uniting for Peace Resolution (often called the Acheson plan). The General Assembly:
Resolves that if the Security Council, because of lack of unanimity of the permanent members, fails to exercise its primary responsibility for the maintenance of international peace and security in any case where there appears to be a threat to the peace, breach of the peace, or act of aggression, the General Assembly shall consider the matter immediately with a view to making appropriate recommendations to Members for collective measures, including in the case of a breach of the peace or act of aggression the use of armed force when necessary, to maintain or restore international peace and security.17
It is further strengthened by the international community's generally favorable reception to the Responsibility to Protect declaration of the International Commission on Intervention and State Sovereignty (ICISS) in 2001. The Commission's report stated:
We have made abundantly clear our view that the Security Council should be the first port of call on any matter relating to military intervention for human protection purposes. But the question remains whether it should be the last. In view of the Council's past inability or unwillingness to fulfill the role expected of it, if the Security Council expressly rejects a proposal for intervention where humanitarian or human rights issues are significantly at stake, or the Council fails to deal with such a proposal within a reasonable time, it is difficult to argue that alternative means of discharging the responsibility to protect can be entirely discounted.18
A “World Summit” of more than 170 nations in 2005 endorsed the spirit of the ICISS report in this language:
[W]e are prepared to take collective action, in a timely and decisive manner, through the Security Council, in accordance with the Charter, including Chapter VII, on a case-by-case basis and in cooperation with relevant regional organizations as appropriate, should peaceful means be inadequate and national authorities manifestly fail to protect their populations from genocide, war crimes, ethnic cleansing and crimes against humanity.19
This language did not recognize intervention that is not authorized by the Security Council, but it nevertheless bears on the interpretation of Article 2 (4) because it clearly supposes that a nation's sovereignty, or “territorial integrity,” does not protect it from legitimate intervention in its domestic affairs.
In this way, we can build a salience case for an interpretation of the UN Charter that would permit humanitarian intervention even if the Security Council failed to authorize that intervention because one or more of the permanent members exercised a veto. But recall the proviso embedded in the salience principle. We must now ask: would an interpretation of the Charter that would permit intervention unauthorized by the Council actually improve the legitimacy of coercive government in the world? There is a powerful argument that it would not: Iraq. Any doctrine that would allow powerful nations to justify aggressive war as a protection of basic human rights boils with the danger of abuse. (The humanitarian justification was not offered by American or British officials in advance in the Iraq invasion, but it has been suggested by former officials of both in retrospect, and could be expected to be offered much more often if established in international law.) This counterargument is powerful because it warns that permission to invade without Security Council authorization would prove massively divisive rather than a principle around which further consensus and salience might develop.
However, a safer though more ambitious interpretation of Article 2 (4) might yet be sustained. We might argue, not that the Article permits unauthorized humanitarian military action as it stands, but that it would permit General Assembly action that would have that effect subject to appropriate safeguards. Imagine that the General Assembly has adopted a resolution with the following substance: member states are forbidden, acting unilaterally or in groups or regional organizations, to threaten or use military force without the authorization of the Security Council, unless a majority of the Security Council has voted to authorize the intervention and the International Court, pursuant to its authority to issue advisory opinions upon the request of the General Assembly, declares that the actions of the regime against which force is proposed constitute crimes against humanity.20
This would be fresh legislation, of course, rather than an interpretation of the Charter as it stands. But would this General Assembly resolution be ultra vires the Charter or otherwise invalid? The argument that it would be valid legislation has two parts. The general argument I offered for a permissive interpretation of Article 2 (4) would support the resolution, and the counterargument I described, about the danger of unilateral action, would be removed by the resolution itself. Crimes against humanity have been sufficiently well defined in other documents and in international practice to provide a satisfactorily clear standard for the International Court to apply in its advisory opinions. In an influential discussion, Antonio Cassese summarizes the position under general international law as follows. Crimes against humanity are
[i] particularly odious offences in that they constitute a serious attack on human dignity or a grave humiliation or a degradation of one or more human beings; (ii) they are not isolated or sporadic events, but are part either of a government policy (although the perpetrators need not identify themselves with this policy) or of a wide practice of atrocities tolerated or condoned by a government or a de facto authority. However, murder, extermination, torture, rape, political, racial, or religious persecution and other inhumane acts reach the threshold of crimes against humanity only if they are part of a widespread or systematic practice. Isolated inhumane acts of this nature may constitute grave infringements of human rights or, depending on the circumstances, war crimes, but may fall short of meriting the stigma attached to the category of crimes under discussion.21
So, on the moralized approach to international law that I am now defending, the resolution I imagine would not be ultra vires and action taken pursuant to that resolution would not be illegal under international law. Any sensible resolution would be much more elaborate, of course. It might revise the procedures of the International Court so as to expedite the advisory decisions required, for example. Or—a more radical idea—it might bypass the International Court altogether and create a special international court for this ad hoc purpose. I shall not try to explore the appropriate character of any such court now. But any of these possibilities would establish a remarkable and important improvement in international law.
The third of the risks I mentioned, posed by the Westphalian system of full sovereign independence, is the lack of any international legislative body with sufficient jurisdiction to solve the grave coordination problems that every nation now confronts. We are already seized by devastating prisoners' dilemmas: about terrorism, climate change, Internet communication, and economic policy. If we had an entirely different form of international organization—a worldwide federal system, for instance, with a supreme parliament—we could attack those problems through comprehensive global legislation. The unmitigated Westphalian system allows no comparable opportunity. Each nation, I suggested, has a general responsibility to do what it can to improve the legitimacy of its own coercive government, and therefore a responsibility to attempt to improve the organization of states in which it functions as a government. What follows for international law?
Once again, we can make use of the hypothetical international court I imagined. Suppose the General Assembly of the United Nations enacted, by majority vote of the member states, a comprehensive plan to regulate carbon emissions, stipulating quotas for each nation. Would our imagined court be right to enforce those quotas through the international police force? No, not as things stand, for a variety of reasons. No plausible interpretation of the language of the UN Charter or of more than a half-century of practice under the Charter would justify assigning that kind of legislative power to the General Assembly. Stipulating emission quotas cannot be supposed to be among the cardinal purposes of the United Nations, as protecting populations from crimes against humanity can be.
There would be, I think, an even more basic objection. The overall argument I have proposed requires international law doctrine to be interpreted to improve the legitimacy of the international arrangement. It hardly improves legitimacy to allow the General Assembly as now composed—all member states are represented by a single vote even though some have exponentially greater populations than others—to dictate to the world. A world parliament cannot be manufactured out of the international structures agreed as compromises in San Francisco. But the bases we have identified for international law, which include the principle of salience, do provide procedures for establishing a more legitimate world parliament that I believe could be defended as now authorized.
I will only sketch one example: a four-majorities system of international legislation. Suppose an international conference is convened in which almost all nations, though perhaps not including all the permanent members of the Security Council, agreed to a General Legislative Convention. This Convention authorizes the United Nations General Assembly to adopt legislation addressed to global dangers requiring coordinated international action, and stipulates that such legislation would be enacted if it received votes representing states that hold a majority of the members' total populations, a majority of votes in the General Assembly, a majority of votes in the Security Council, and a majority of votes among the permanent members. Legislation so enacted would automatically be submitted to the International Court of Justice, which would apply principles of subsidiarity, as these have been developed within the European Union, for example, to determine whether such legislation was sensibly regarded as of international dimension or was a matter properly left to national determination.
I suggest this voting formula—four-majorities together with judicial review—only as an example of a new institutional structure that might be created by broadscale treaty making followed by salience. Other formulas can be constructed of equal or perhaps much greater merit: events and salience would pick out the first adequate one. Any such international voting system, coupled with International Court protection against international overreaching, would go some way toward mitigating the fourth impact of the Westphalian system on legitimacy I mentioned. This is often described as a democratic deficit. Of course it would be a stretch to say that a voting scheme like the four-majorities would provide worldwide democracy. Democracy does not mean just majority rule,22 but it does suppose a political community almost all of whose population participates as equals directly or indirectly in the broad range of political decisions affecting their lives. Still, some such formula would provide greater indirect participation in their own governance than most of the world's population now enjoys. A multiple-majorities formula is also sufficiently complex to protect a range of cultures and traditions from any deeply offensive authority.
If a genuine and effective world government had developed, in place of the Westphalian system of separate states, then demands of justice that are now local to particular states would be demands on the world as a whole. A coercive political community must respect the dignity of those over whom it exercises dominion by showing equal concern and respect for them all. Justice would therefore require a world government to show that equal concern for everyone alive. It is hard to predict the results of a world government accepting that responsibility. There are different conceptions of what equal concern requires. But though a global requirement of equal concern would certainly not generate equal wealth for everyone in the world, it would almost certainly produce much less inequality than the Westphalian system now generates. I must therefore add that, certainly to the disappointment of some readers, nothing in my argument supposes that the duty of nations to mitigate the risks imposed by an unmitigated Westphalian system includes a duty to form a world government or to assume each for itself the responsibilities for equality that a world government would have. Still, that space must be watched. I have been describing a future, so far imaginary, in which the duty to mitigate abetted by a principle of salience produces international legislation of greater and greater scope. If that future materializes, then at some point the question will be pressing whether there is then sufficient coercive authority in a world institution to engage some international responsibility of equal concern.
You may have been surprised by the freewheeling character of my arguments about the true grounds and therefore the contemporary content of international law. But remember that international law is very young: it was effectively reborn in 1945. The arguments of famous judges in the comparable formative period of the Anglo-American common law—of Baron Bramwell, Edward Coke, and George Jessel, and of some federal judges in the brave days of Swift v. Tyson in the United States—might strike contemporary lawyers as equally freewheeling.23 The constitutional arguments of John Marshall, who transformed a written constitution, and of Aharon Barak, who made a constitution for Israel without a written one to transform, surprised many of their colleagues. If law is understood as a special part of political morality, and if it serves its community well, its doctrines will crystallize over time. Its roots in political morality will grow less prominent—though will be available when needed—in ordinary legal argument. That progress from principle to doctrine will signal its success. But a rigid separation between legal and moral argument in the development of international law would be premature now and would accelerate its practical irrelevance. We must free the subject from the torpor of legal positivism. We need, now, to nourish the roots, not the twigs, of international law.
See the introduction to Justice in Robes (Cambridge, Mass.: Harvard University Press, 2006).,
See Justice for Hedgehogs (Cambridge, Mass.: Harvard University Press, 2011), pp. 165–166.,
See, e.g., Theorizing the Sources of International Law,” in The Philosophy of International Law, ed. Samantha Besson and John Tasioulas (Oxford: Oxford University Press, 2010)., “
See http://www.icj-cij.org/documents/index.php?p1=4&p2=2&p3=0. It would be nice to treat section (d) of this account as granting power to academic philosopher kings in international affairs. Law is what NYU's international law group says it is. I assume, however, that the clause is only meant to allow the International Court to appeal to academic interpretations of the other clauses, so we must concentrate on those.
See, e.g., The Responsibility to Protect: The U.N. World Summit and the Question of Unilateralism,” Yale Law Journal 115 (2006): 1157–1165; and and , “Humanitarian Intervention to Protect the Ibos,” in Humanitarian Intervention and the United Nations, ed. Robert B. Lillich (Charlottesville: University of Virginia Press, 1973), pp. 175–177., “
I discuss several of these in Justice for Hedgehogs, chap. 14.
This is only a crude statement of the character of interpretive concepts. See ibid., chap. 8, for a fuller and more accurate account.
See my discussion in Ronald Law's Empire (Cambridge, Mass.: Harvard University Press, 1986), chaps. 6–7; and , Justice for Hedgehogs, chap. 19.,
See, e.g., Reading Law: The Interpretation of Legal Texts (St. Paul, Minn.: Thomson/West, 2012).and ,
It is a further defect of the conventional approach that the “subjects” of international law—the entities to whom its rights and duties attach—must be defined exogenously. Commentators simply stipulate that only the kinds of political community contemplated in the Westphalian system count as subjects. That cannot be made true, without the circularity discussed in the last section, by the unanimous consent of the very states who are assumed to exhaust the pertinent population. Under the different approach I describe in the remainder of these remarks, the question of application is internal to the body of law developed. The imaginary court I describe can recognize as a sensible and justiciable question, for instance, whether multinational corporations are subject to its jurisdiction and rulings. They answer that question by exploring the principles they recognize as structuring the law they are charged with enforcing. See my brief discussion, provoked by a very helpful conversation with Sam Scheffler, at http://www.justiceforhedgehogs.com/human-rights/.
See Foreign Law and the Modern Ius Gentium,” Harvard Law Review 119 (2005): 129–147; and , “Partly Laws Common to All Mankind”: Foreign Laws in American Courts (New Haven, Conn.: Yale University Press, 2012), chaps. 2–3., “
Lessons of Kosovo,” American Journal of International Law 93 (1999): 857–860., “
I do not use a more general description, such as “violations of human rights,” because I believe that only certain violations of human rights justify military or even serious economic intervention or sanction. See Justice for Hedgehogs, chap. 15.,
Crimes against Humanity,” in The Rome Statute of the International Criminal Court: A Commentary, ed. Antonio Cassese , Paola Garta , and John R.W.D. Jones (Oxford: Oxford University Press, 2001), 1:353–377, at p. 360., “
See Justice for Hedgehogs, chap. 18.,
Swift v. Tyson 41 U.S. (16 Pet.) 1 (1842).