Public, stable rules

Formal legality describes the sense in which rules must be made such that individuals can distinguish between legal and illegal actions. The rules must be clear and public, they must be forward-looking, they must be written in language that is sufficiently specific and yet be designed for general categories of behaviour rather than particular incidents. A functionalist case for formal legality rests on the argument that law can only be a useful guide for individual behaviour if the citizen can be reasonably confident that she or he knows what is lawful and what is not. Unless the rules are public, prospective, and somewhat stable, the citizen cannot meaningfully take them into account when acting. As Raz says, ‘the law should be such that people will be able to be guided by it … . The law must be capable of being obeyed … . It must be such that they can find out what it is and act on it.’^{Footnote 13} This underpins Hayek’s point that predictability in the law is essential for its contribution to human liberty.^{Footnote 14} The ability to differentiate the lawful from the unlawful undergirds law’s capacity to influence decisions, to organise society, and to predict general patterns of mass behaviour.

The goal of fixed and known rules is in perpetual tension with the possibility of making changes in the law. Law should be stable, but must be also open to change and amendment – thus, for instance, when Joseph Raz discusses the stability of the law, he couches it as ‘relative stability’.^{Footnote 15} This is managed in the theory of the rule of law by constructing non-trivial institutional requirements that govern the process for changing laws: Parliamentary approval, for instance, or a plebiscite. These are distinct from rules designed to prevent the government from taking certain substantive decisions, as might be controlled by a constitution or a bill of rights which make it impossible for the state to seek certain ends. The procedural rules on law-making exist to make it more difficult for the state to change existing law, but their existence in some form is necessary to accommodate the possibility of change. They reconcile the government’s authority to make and remake rules with the rule-of-law requirement that rules be (relatively) fixed. Of course, this generates the very real danger that the capacity of the government to change the law will undo the benefits thought to come from ‘relatively’ stable law in the first place, and as a consequence it leads to the second component of the rule of law: subordination of the government itself to the body of law.

Government limited by laws

Knowability and stability of the law is not enough for the rule of law. The doctrine of the rule of law also insists on certain forms of equality. One of these is represented by what Tamanaha calls ‘government limited by law’, which is the requirement ‘the state and its officials are limited by the law’ just as are regular citizens.^{Footnote 16} (The second is described below as equality of application across cases.) The law is binding on all citizens, and all operate in the context of the law. In Dicey’s words, this means ‘not only that with us no man is above the law, but … that here, every man, whatever be his rank or condition, is subject to the ordinary law of the realm and amenable to the jurisdiction of the ordinary tribunals’.^{Footnote 17} This is the idea at the heart of Karen Alter’s suggestion that ‘diminishing the absolute power of governments is, of course, the objective of the rule of law’.^{Footnote 18} This provision subordinates the government to the law and equalises individuals and government as subjects of the law. Its absence in a legal system is often described as ‘rule by law’, in which ‘authoritarian rulers … capitalize on the regime-supporting roles that courts perform while minimizing their utility to the political opposition’.^{Footnote 19} As we shall see below, the application of this aspect of the rule of law to the international context is problematic due to the very different nature of governance in the interstate realm.

Rules applied consistently across cases

Finally, the rule of law requires that the law be applied across cases in a particular manner – that judges and government officials follow or apply the relevant body of rules to the situation before them in accordance with laws (or norms) of procedure.^{Footnote 20} This is usually seen as being in distinction to the arbitrary exercise of power of some individuals over others, or to decisions taken based on the particular character or identity of the parties. As Raz has noted, this does not prevent laws that treat different groups or people unequally – such as guaranteeing rights to landowners that are not given to others.^{Footnote 21} Inequalities created by law are common, perhaps inevitable.^{Footnote 22} The rule of law requires only that the terms and categories created by law be applied equally across the cases to which they apply. This element of the concept produces the independent judiciary. Courts, ‘with the duty of applying the law to cases brought before them and whose judgments and conclusions as to the legal merits of those cases are final’,^{Footnote 23} are a device to implement the dispassionate adjudication of disputes arising in and from law. Other institutions may well serve this purpose as well, or instead, but the independent judiciary is its leading institutional form in domestic societies today.

These three elements combine to constitute a conventional account of the rule of law as an amalgamated concept.^{Footnote 24} Its various components represent answers to different political problems and come from distinct historical processes of contestation.^{Footnote 25} These can be combined in different ways, giving rise to the comparative study of legal systems within the rule-of-law family. They are nevertheless usually seen as mutually reinforcing and as constituting together a distinct and coherent mode of governance, law, and society. It is this form of society that is often taken to characterise a ‘normal,’ modern state: this is the system that the historian Paul Johnson calls ‘the greatest public achievement of the second millennium’;^{Footnote 26} and that the United Nations and others strive to create in post-conflict societies and elsewhere;^{Footnote 27} and that conservatives see themselves as defending against the ‘creeping instrumentalism’ of law unmoored from a consensus over the common good.^{Footnote 28}

Two important consequences are often said to follow from this bundle and these give it its normative appeal: first, it may generate the obligation on individuals to obey the law, thus producing stability and social quiescence; and second, it may contribute to valuable substantive social outcomes, such as respect for individual rights and private property, and lower levels of government corruption.

On the first, the rule of law is sometimes said to be the source of citizens’ obligation to follow the law. This obligation can be identified as simultaneously a legal obligation, a political obligation, and a moral obligation. People are required to obey the law as a matter of law (that is, it is binding), and they are politically forced to obey (even accepting that some degree of non-compliance as unavoidable), and they are morally obligated to the extent that the rule-of-law criteria produce an ethical imperative – if the rules were produced legitimately, it is said, there is a moral obligation on the individual to comply.^{Footnote 29} Taken together, they lead to Raz’s conclusion that ‘people should obey the law and be ruled by it’.^{Footnote 30} The rule of law is thus intrinsically linked to the idea of compliance on the part of the subjects and with a particular theory of social order and modern governance. In a society organised by the rule of law, one is expected to comply with the rules and one should be able to trust that others will generally comply with the rules. This gives order and predictability to society and marks one kind of transformation from the state of nature to society.

The second effect that follows from the rule-of-law society relates to substantive outcomes: various good things are assumed to follow from having the legal–political arrangement described above. These might include respect for human rights and dignity, a free press, anti-corruption, private property and transactions, individual autonomy, the capacity to plan in advance, and more. These outcomes give the rule of law its political appeal, since it is often thought that it represents an institutional arrangement that is more likely to produce a good society than are other alternative arrangements. This is what Lon Fuller sought to capture with the claim that the rule of law has an ‘affinity with the good’.^{Footnote 31}

The relation between the rule of law and these substantive goals provides the basis for a key schism in the rule-of-law literature, between a thick and a thin version. In this article, I emphasise the rule of law in ‘thin’, institutional/procedural terms, as a set of rules and practices that organise a society. A competing view takes the position that a rule-of-law system cannot exist unless it is substantively devoted to goals like human rights, equality, or justice. This is a ‘substantive’ or ‘thick’ approach to the rule of law, and requires that the rules enshrine particular social ends, such as equality or liberty or justice. It arises in the form of a complaint that the ‘thin’ conception does not clearly rule out social evils (such as Nazism or torture) if these are permitted or required under the prevailing legal instruments. The thick version makes this conceptually impossible by marrying the idea of the rule of law to a substantive theory of social goods, so that a regime that fails to respect human rights (for instance) can at best be described as performing ‘rule by law’ but cannot be described as ‘rule of law’.^{Footnote 32} In the language of global constitutionalism, this is akin to suggesting that there exists a set of fundamental global norms – perhaps individual rights, or the regulation of warfare – that have constitutional status and against which formal laws can be judged for their constitutionality or lack therefore.^{Footnote 33} This view is common among scholars of world politics and law – it opens the door to the possibility (on the one hand) that states might sometimes rightly break international law to achieve these more important goals and (on the other) that some acts are inherently unlawful if they violate these norms.^{Footnote 34}

Arguing against this thick view, Joseph Raz warns that it is a mistake to equate ‘the rule of law with the rule of the good law’.^{Footnote 35} He is critical of the tendency to mix the theory of the rule of law with theories of substantive social and political goods, a tendency which is widespread among scholars and policymakers in international law and politics. For instance, the UN Secretary-General in 2004 defined the rule of law as ‘a principle of governance in which all persons … are accountable to laws that are [among other things] consistent with international human rights norms and standards’.^{Footnote 36} A commitment to human rights norms is thereby defined into the very concept of the rule of law. Similarly, Tom Bingham explains the contribution made by the rule of law by describing society as it would exist in its absence. He says ‘the hallmarks of a regime which flouts the rule of law are, alas, all too familiar: the midnight knock on the door, the sudden disappearance, the show trial, the subjection of prisoners to genetic experiment, the confession extracted by torture, the gulag and the concentration camp, the gas chamber, the practice of genocide or ethnic cleansing, the waging of aggressive war. The list is endless.’^{Footnote 37} To produce this list, Bingham assumes that the things he finds abhorrent are illegal, with the result that a perfectly faithful adherence to the rule of law will lead to a world in which these things do not happen. He constructs a model of the rule of law in which non-compliance with the law is the problem that needs addressing in the pursuit of greater human welfare. Compliance with the law is therefore the path to social goods, and non-compliance is a problem in search of a remedy.

I adopt a thin, formalist model of the rule of law here. This is agnostic about whether important social goods follow from the rule of law, but it insists that they cannot be constitutive of its existence. The ‘thick’ substantive model assumes too much: it suggests that the question to ask about a legal system is how well it institutionalises a particular set of goals. Not only does this subordinate the rule of law to a prior theory of what those goals are (and presumes prior agreement on that theory) but it also makes it hard to ask questions about how a legal system is different from other kinds of governance or morality or politics. A great deal of interesting politics takes place when the demands of the legal system do not map perfectly onto either individual interests or personal morality. The substantive approach presumes that justice and law align neatly with each other. But there are important phenomena to be studied in the instances and ways that they do not. International treaties are at best imperfect vehicles for justice, and at worst they may be complicit in giving governments impunity from responsibility for atrocities.^{Footnote 38} In such a world, even perfect compliance with one’s legal obligations goes not guarantee that one’s actions will be just, for the simple fact that the laws do not naturally or automatically reflect the demands of justice.

Can a society embody the three components of the rule of law but not protect (for instance) individual human rights? This is both an empirical and a conceptual question. Conceptually, it asks if such a place deserves to be called a ‘rule of law society’. This is a question about the definition of the rule of law, and many would deny such a place that label. Empirically, it asks whether the rule-of-law institutions (properly understood and instituted) necessarily lead to respect for individual human rights. The thin position on the rule of law takes law to exist in the rules and practices of a particular kind of governance and in the structure that upholds them. It presumes that this has interesting political causes and consequences, some of which may be related to achieving justice or other social goods. But not necessarily.

In this section, I have sketched what I take to be a conventional account of the rule of law as it is commonly applied in domestic legal theory, taking a formalist or ‘thin’ position as opposed to a substantive or ‘thick’ view of its essential components. In the next section, I consider the problems that arise when each of these components is brought to bear on the international setting of interstate relations. None can be directly translated to international affairs. The rule of law for world politics therefore must rest on foundations unique to the arrangement of political power and institutions that exist in the international domain.

Public, Stable Laws?

The first component of the domestic rule of law (that is, that rules be public and stable and forward-looking) has a simple international analogue in the form of the interstate treaty, and more generally in the codification movement from the nineteenth century through to the post-WWII period. The rise of legal positivism in international law, characterised by the belief that state consent is the ultimate source of legal obligation, produced a strong motivation for explicit treaties on a range of subjects, making the treaty the pre-eminent legal instrument. This roughly overlaps with what Koskenniemi has called the ‘heroic period’ for international law from 1870 to 1960.^{Footnote 39}

Treaties are thought to be valuable because they incorporate both the fixity that is presumed to be required for the rule of law and the consent of the subjects (i.e. states) through ratification by domestic political institutions. These two values are promoted by the legal positivist school into the very definition of international law: in the well-known words of the Permanent Court of International Justice in the Lotus case ‘International law governs relations between independent States. The rules of law binding upon States therefore emanate from their own free will.’^{Footnote 40} Consent to a formal text is, for Reinold and Zürn, the most direct means by which international law can achieve its ‘core objective’, which they summarise as to ‘stabilize actors’ normative expectations in an otherwise volatile world and shield them from … the vicissitudes of politics’.^{Footnote 41} To do so, it is required ‘that legal rules display certain features, such as transparency, clarity, non-retroactivity, etc’.^{Footnote 42} Treaties are widely assumed to embody these characteristics and therefore the codification of international law is often equated with its ‘progressive development’.^{Footnote 43}

The codification of international law can now be seen as a central project of twentieth-century international relations. From the opening decades of the twentieth century, it is evident in the Covenant of the League of Nations (which demanded ‘a scrupulous respect for all treaty obligations’), and in Woodrow Wilson’s Fourteen Points (whose first principle is ‘1. Open covenants, openly arrived at …’), and in the League’s Committee of Experts for the Progressive Codification of International Law and its 1930 Conference on Codification. Mid-century, the United Nations was founded with a commitment to ‘the progressive development of international law and its codification’ through the General Assembly (UN Charter, Article 13). This mandate produced the International Law Commission (ILC), which took over from the League’s Committee of Experts in 1947, as a permanent body of international-legal professionals committed to expanding the scope of international law. In the words of the ILC Constitution, its goal is to advance ‘the more precise formulation and systematization of rules of international law’ and ‘the preparation of draft conventions on subjects which have not yet been regulated by international law’.^{Footnote 44} The codification movement sees it as a success that the number of interstate treaties has grown exponentially. ‘Treaties,’ said the UN Secretary-General recently, ‘are a critical foundation to the rule of law.’^{Footnote 45}

The drive for codified law can be understood as the operationalisation of the first piece of the rule-of-law ideology: the requirement for clear and stable rules. Explicit well-written law is believed to make a particularly powerful contribution to international order. This is reflected in the modern ‘legalisation’ project among scholars of law and political science: Goldstein, Kahler, Keohane and Slaughter suggest that ‘fully legalized institutions bind states through law: their behavior is subject to scrutiny under general rules, procedures, and discourse of international law and, often, domestic law. Legalized institutions also demonstrate a high degree of precision, meaning that their rules unambiguously define the conduct they require, authorize, or proscribe.’^{Footnote 46} The stability and clarity of the rules are essential in the ‘move to law’ that these authors identify. These are the qualities that Thomas Franck identified as constituting the ‘determinacy’ of a rule: ‘that which makes its message clear’.^{Footnote 47} And while ‘some degree of indeterminacy is inevitable in any body of rules … indeterminacy also has its costs, which are paid in the coin of legitimacy’.^{Footnote 48} ‘Indeterminate normative standards make it harder to know what is expected.’^{Footnote 49} ‘The more determinate a standard, the most difficult it is to justify non-compliance.’^{Footnote 50} Legalisation changes the calculations of actors because it specifies whether an action is permitted or not. It therefore gives others a standard by which to judge the state as compliant or not, and gives a basis for states’ reputations regarding rule-following and rule-breaking. Clarity and stability are the essence of international law.

Against all of this scholarship is the fact that individual states have the capacity to change the legal status of their behaviour – from illegal to legal, from violation to compliance – by the exercise of their legal and political agency. This is unthinkable in the classic domestic rule of law where the legality of an act is set by the state not the actor. States themselves choose which obligations will apply to them, and they choose as individuals. This does not negate the international rule of law – instead it signals that the relationship between codification and the rule of law is different in the international context than it is in the domestic. A clear rule does not mean a clear obligation, and the presence of an obligation does not mean there is consensus over the meaning of compliance.

This is an implication that follows from the often noted fact that sovereign states are both the authors and the subjects of international law, and from the emphasis on consent that is the heart of legal positivism.^{Footnote 51} The existence of a treaty does not imply an obligation on the part of a state – codified rules of international law wait on state consent. The curious status of unratified treaties reflects the fact that it is state consent to the treaty, rather than the treaty itself, that is legally binding.^{Footnote 52} This is different in the domestic version where a law on the books is by definition compelling on citizens (and on their government). The residual autonomy of sovereign states means that they can tailor their legal obligations to suit their needs and interests: states decline to sign or ratify treaties they disagree with, and they write reservations and understandings which limit or explain their obligations under the treaties that they do ratify. They exit treaties that come to impinge on their decisions in ways that they dislike.^{Footnote 53} Each state constructs its own set of legal obligations and fine-tunes it through reservations. Domestic legal subjects cannot do this.

This is evident in the North Korea’s withdrawal from the Nuclear Nonproliferation Treaty (NPT) in 2003. As a non-nuclear weapon member of the NPT, North Korea’s nuclear programme in the 1990s was a clear violation of international law. In 2003, it withdrew from the treaty following the required three-month notification period. In doing so, it ended its obligations to the treaty. (It may still be in violation of other international obligations with respect to its behaviour, including Security Council resolutions binding through the UN Charter.) Canada made a similar move in 1994 when it revised its optional-clause declaration to the International Court of Justice (Article 36(2)). Canada had embarked on an activist position toward protecting North Atlantic fish stocks, and to pre-empt any case before the ICJ for its behaviour it revised its optional-clause declaration to exclude any ‘dispute arising out of or concerning conservation and management measures taken by Canada with respect to vessels fishing in the NAFO [Northwest Atlantic Fisheries Organization] Regulatory Area’.^{Footnote 54} With that, its conduct it removed the possibility of being found to be in violation of its obligations, and its position was affirmed shortly thereafter by the ICJ in the Etsai case.^{Footnote 55}

As a consequence, one cannot ask what international law is on a given topic and expect an answer that is generalisable across states. Instead, one must ask what the law is for a given state, and perhaps even in relation to a specific other state, and then look for the answer in the treaties, protocols, and custom that apply to that interstate relationship. For instance, the legal obligations taken on by Australia under the International Convention on the Regulation of Whaling (ICRW) are different than those taken on by Iceland (which has formally objected to certain provisions of the treaty), and both are different from the laws that apply to Turkey (which has not signed the Convention). The same act – for instance, taking a whale on the high seas in the Southern Ocean – has a different legal status depending on which of these states committed it. It is a violation of international law for Australia to permit an Australian vessel to hunt whales. It is not a violation for Iceland to allow Icelandic ships to do the same.^{Footnote 56} Turkey, because it has not signed the treaty, has an unlimited right to hunt whales.^{Footnote 57}

Much of international law is devoted to managing the degrees of freedom that states enjoy in tailoring their legal obligations. This includes the Vienna Convention on the Law of Treaties and its rules on denunciation of and withdrawing from treaties, adding or changing reservations, fundamental changes in circumstances, obligations and government succession, the relation of subsequent treaties to prior treaties on the same subject, etc.^{Footnote 58} These rules limit how states might use the autonomy that they possess to redefine their legal obligations – but they do not eliminate the autonomy of states to remove themselves from unwelcome legal constraints.^{Footnote 59}

In the domestic setting, Joseph Raz says ‘the rule of law is often rightly contrasted with arbitrary power. A government subjected to the rule of law is prevented from changing the law retroactively or abruptly or secretly whenever this would suit its purposes.’^{Footnote 60} In the interstate setting, this goal of a unified set of public rules that applies to all subjects cannot be achieved or even approximated because states retain the authority to accept, reject, or modify their legal obligations through treaty accession, reservations, and persistent objections. Each state could potentially have a unique set of legal obligations.^{Footnote 61} This is manifest in the nuanced way in which the UN Millennium Development Goals describe the commitment to the international rule of law: it says ‘We resolve therefore … to strengthen respect for the rule of law in international as well as national affairs and, in particular, to ensure compliance by Member States with the decisions of the International Court of Justice, in compliance with the Charter of the United Nations, in cases to which they are parties.’^{Footnote 62} The UN cannot make a straightforward claim that states should comply with all international law. Only that subset of law chosen by the state itself is binding. It is not an artefact of ‘insufficient’ progress toward an idealised version of the domestic rule of law – it is an essential implication of the institutional foundations of international law, and therefore is an insurmountable obstacle for the liberal project of applying the domestic rule-of-law model to international affairs.

In sum, the international legal system cannot satisfy the standard expectation in a rule-of-law society that law be clear and stable and known in advance. No matter how clearly it is written, it cannot indicate whether an action is legal or illegal independent of the identity of the actor-state in question. Because states are free to tailor their commitments to suit their needs and interests, the bundle of laws that attach to each state are unique to that state and the legality of an act may not be determinable as a general matter. What it means to ‘comply’ with international law depends on who is doing it, and what that state has done and said about the rule in the past and what it might do and say in the present case. The same act may be legal when committed by one state but illegal when committed by another; it may be legal when committed toward one state but illegal toward another. The legality of an act is endogenous to the choices of the state in question, rather than independent of it as it is in the domestic setting.

Government limited by law?

The second component of the domestic rule of law is the requirement that law apply to the government as well as to the citizens. This is a solution to the problem of despotic leaders. It implies that the legal system should be an instrument for limiting the government’s authority over citizens and perhaps subjecting it to the will of the people, somehow defined. It can be seen as preserving the autonomy of the legal sphere from the political,^{Footnote 63} or as ensuring that law is superordinate over government,^{Footnote 64} or as protecting individuals from the state.^{Footnote 65}

The international analogue to this is difficult to find: if the international rule of law is understood with reference to how the government is made accountable under law, then what is the international ‘government’?

The most popular answer may be to say that there is none, on the theory that the interstate political space is missing the central political institution and is therefore an anarchy in a very specific sense. This accords the standard model of ‘international anarchy’ that is represented in international relations theory in various forms by Kenneth Waltz, Hedley Bull, Alexander Wendt, and others. This is the ‘anarchy problematique’.^{Footnote 66} It is the premise of much international theory scholarship and suggests that the fundamental difference between domestic society and international society is the presence of government in the former and its absence in the latter.

This framing of the problem of international law may be sustainable if one takes the narrowest view of what constitutes ‘government’ – a centralised, authoritative, coercive institution kept in power by the twin forces of legitimacy and coercion. The institution of state sovereignty presumes that such a world government does not exist, and therefore the distinctive features of international law are often seen as a result of the attempts by states to manage their relations of interdependence in its absence.^{Footnote 67} This is the thought behind Kant’s Second Article for Perpetual Peace: ‘As nations, peoples can be regarded as single individuals who injure one another through their close proximity while living in the state of nature … For the sake of its own security, each nation can and should demand that the others enter into a contract resembling the civil one and guaranteeing the rights of each.’^{Footnote 68}

However, most scholars no longer subscribe to this absolutist view of international anarchy. Across the range of IR theory perspectives, it is widely accepted that various forms and degrees of governance exist in the interstate system and as a consequence scholarship has turned to examine the extent to which the functions of governance are performed in international society by agents other than states as understood in classical public international law. These might include quasi-public regulation by private actors,^{Footnote 69} contracts or coordination among legal equals,^{Footnote 70} Great Power leadership or dominion,^{Footnote 71} pooled sovereignty organisations,^{Footnote 72} transnational class factions,^{Footnote 73} and other legitimated international institutions.^{Footnote 74} While none of these meet a formalist’s definition of ‘government’, they provide evidence of disaggregated governance among states and sometimes exercise great influence over the processes of international (and perhaps domestic) life. Harold Koh suggests that we should think of state sovereignty in terms of the functions performed by a sovereign (rather than the centralised bureaucracy that often houses those functions), in which case these diverse institutions of international governance might be said to exercise sovereignty.^{Footnote 75} The international system is not an anarchy in the technical sense defined by Waltz, Bull, and others.

It follows therefore that ‘governance without government’ may exist at the international level,^{Footnote 76} and so the international rule of law may be expected to regulate its institutions. Two possibilities are worth considering as locations for international ‘government’ which might be regulated by the international rule of law: strong states and the UN system.

First, following Gerry Simpson and others, the international legal system might be seen as granting the most powerful states a governance role over the community of states, and in this case one can ask whether these agents of governance are subordinate to international law. It is often said that certain states (or the Great Powers as a class) have special responsibilities toward the international system, composed of rights and obligations for the maintenance of international order which do not attach to rank-and-file states. Claims about the importance of Great Power leadership, about hegemonic stability, and about collective security all rest on this premise.^{Footnote 77} In this view, Tamanaha’s category of ‘government limited by law’ might be satisfied in the international system by ensuring that the strongest states be bound by international law in their governance of the system. This they certainly are in a formal sense, but it is clear also that they have greater capacity to bear the costs of law-breaking and also greater influence in rule-making, both by dominating diplomacy where rules are negotiated and by the rules of recognition for customary law.

In practice, it is unlikely that the international rule of law can rest on the degree to which the international community or some ‘leading’ states are bound by the law as it is written. On the contrary, it is common to see customary law as changing due to changes in state practice, especially by the ‘leading’ state. This is precisely the conventional theory of how customary law develops but it is also evident in relation to treaties. The Security Council practice of treating abstentions by permanent members as something other than a veto arose in this way. This has been the consistent practice since the earliest days of the Council and yet it contradicts the plain meaning of the Charter clause that regulates voting in the Council. The first time the practice was used (in 1946) the irate ambassador from the Soviet Union stormed from the room in protest, but its repeated use and its practical advantages quickly extinguished any controversy.^{Footnote 78} Today, it is widely accepted as an informal revision of the Charter, and was affirmed by the ICJ in a 1971 opinion that rejected South Africa’s claim that a resolution against it was illegal because three permanent members had abstained when it was passed.^{Footnote 79}

A similar development has underwritten the concept of humanitarian intervention since 1990. The use of force by states is illegal under the UN Charter unless authorised by the Security Council, and no distinction is made in the Charter for war motivated by humanitarian rescue. The development of Responsibility to Protect and other doctrines has arguably legalised humanitarian intervention even without Council approval.^{Footnote 80} The case in favour of the legality of humanitarian intervention rests on recent developments in practice among states, international organisations, and others in favour of the practice. Together, these form the foundation of the argument that a progressive development in the law has taken place so that Council approval is not necessarily a requirement for legal intervention.

In relation to both Security Council abstention and humanitarian intervention, powerful players in world politics decided that the new interpretation was a desirable improvement over the plain language of existing treaties and that as a consequence the new practice should not be considered rule-breaking. This represents the collective legitimation of violation, ‘constructive non-compliance’, by which apparent violations are transformed into revisions of the law.^{Footnote 81}

One might alternately identify international ‘government’ with the United Nations or another formal international organisation. It would then be normal to ask how that body is limited by international law. This may amount to seeing the UN Charter as the constitution of the international system but it need not go that far.^{Footnote 82} The argument is particularly relevant for the UN Security Council with its decisive governing authority over the member states of the UN under the Charter. Its powers include the right to decide when to use force against member states, the right to act on behalf of all states when taking a decision, and the right to decide on its own sphere on competence. Together, these look very much like the Weberian definition of a state: a centralised agent that holds a monopoly on the legitimate use of force, and the Security Council may be a kind of world government.^{Footnote 83} Moreover, the Council undoubtedly has the authority to make law that is binding on all states, both in relation to specific threats to international peace and security and more generally in its recent ‘legislative’ mode, insisting on certain policies of all states.^{Footnote 84} If the Security Council (or the UN) is a form of international government, then it might follow that the Council or the UN must be situated within a setting of legal rules that limit its autonomy relative to its subjects (presumably: states).

The Council’s relationship to international law has long been debated, beginning with the organisation’s pre-history at San Francisco in 1945 when the US and others defeated a proposal to explicitly make the Council subordinate to international law all the way to Kadi in the Court of Justice of the European Union recently.^{Footnote 85} The UN Charter stakes a clear position on the matter – it affirms that the Council is a political rather than a legal organ and that it responds to ‘threats to international peace and security’ rather than to violations of international law. Subsequent practice has upheld the view that the Council does not decide legal issues.^{Footnote 86} For instance, Article 24(1) describes the Council as having ‘primary responsibility for the maintenance of international peace and security’ with no mention of international law, in contrast to the International Court of Justice which is described as the ‘principal judicial organ of the United Nations’ (Article 92).^{Footnote 87} When the Council does create legal obligations on UN members by using its powers in Chapter VII these must be understood as legal demands regarding political policy choices, rather than settlements of legal disputes themselves.

Whether the Council itself is required to follow international law is an open question. It is clear that it must abide by the Charter, but it is arguable whether the Council is superior or subordinate to the rest of international law. The argument for its superiority comes from the fact that the Charter authorises the Council to impose any solution it desires in response to a threat to international peace and security. It follows that it is not limited by international law in doing so. On this reading, the Council could not (for instance) undertake an enforcement action for any purpose other than ‘to maintain or restore international peace and security’ (Article 39), but it could demand that states seize the assets of foreign governments or foreign nationals in a manner which would normally violate international laws on expropriation. This points to an ‘imperial Council’, in which the Charter authorises the Council to stand above international law.^{Footnote 88} The contrary argument, for its subordination to international law, rests on the view that since it is constituted by international law through the treaty process it is therefore in no position to act outside it. The Kadi II decision of the CJEU asked but did not answer the key question of whether that court could invalidate a duly passed resolution of the Council that conflicts with European law.

A second-order question concerns whether there exists an international institution with the competence to review the legality of Council actions. The International Court of Justice is the most logical candidate for such a role, though this is in considerable doubt – the ICJ has had several opportunities to exercise judicial review of the Council and it has avoiding explicitly claiming that right.^{Footnote 89} Alvarez notes that ‘The World Court [ICJ] has never found any action taken by any UN organ to be illegal in a binding context’ and he suggests that doing so ‘would provoke a political crisis and appears unlikely given the limits on the ICJ’s jurisdiction, as well as gaps in international law’.^{Footnote 90} In both the Namibia and Lockerbie cases the Court was asked questions which implicated the limits of Council authority but in both it declined to provide a decisive interpretation either way.^{Footnote 91} Judge Schwebel said in the Lockerbie case (as summarised by the ICJ):

This does not lead to the conclusion that international organisations are in general not bound by international law. They very clearly are, with a very small number of possible exceptions: notably, the UN Security Council when acting under Chapter VII of the Charter and the ICJ when deciding a case ex aequo et bono under Article 38(2) of the ICJ Statute. Instead, it shows that the domestic ideal of a government subordinated under law does not apply unproblematically to the international context since it presumes answers to a host of deep (and perhaps unanswerable) questions about the nature and location of government or governance at the international level.

Rules applied consistently across cases?

The third element commonly cited in the domestic rule of law is that rules should be applied dispassionately and similarly across similar cases. This suggests that the content of the law should be constant, and be applied constantly, regardless of the identity of the actor. As with the other two elements, this idea meets decisive obstacles in the international realm in that international legal obligations are founded on precisely the opposite principle: that the identity of the state is an essential consideration in assessing the international lawfulness of an act.

At one level, the idea of legal equality is central to the self-understanding of international law in the sense that each sovereign state is said to begin as a legal person with equal rights and obligations to all other states. Article 2(1) of the UN Charter codifies this into what amounts to a legal premise for the interstate system: ‘The Organization is based on the principle of the sovereign equality of all its Members.’ Equality is enshrined in many other instruments and practices, including the ICJ Statute (Article 35), non-discrimination rules in the World Trade Organization, the law of treaties as embodied in the Vienna Convention on the Law of Treaties, and much more. The legal personality that underlies the existence of a sovereign state is usually understood to be a bundle of rights and obligations that is precisely the same for all states. With all of this weight placed behind sovereign equality, deviations from equality carry the burden of explanation – they stand in need of justification or defence or explanation.

In practice, however, this expression of sovereign equality moves the international legal–political system away from rather than toward the idea of legal equality as imagined in the domestic rule of law. As authors of their own legal obligations, for states the equal application of the law means something different than it does among individuals. First, states’ legal rights and obligations diverge as soon as some of them consent to treaties. Each acquires a distinctive portfolio of obligations, with the effect noted above that the lawfulness of a particular act is therefore variable and depends on which state is doing it. Second, the absence of compulsory jurisdiction means that court decisions are equally particularistic, since states possess the negative right to not be bound by the legal decisions of others. This is explicit in the ICJ Statute, which says ICJ decisions ‘have no binding force except between the parties and in respect of that particular case’ (Article 59, Statute of the ICJ). Third, it is obvious that states have unequal influence on making and remaking international rules – the statements, acts, and interests of some count for more in international legalisation, sometimes formally and sometimes informally.^{Footnote 93} This is equally true of customary international law as it is of treaties.^{Footnote 94} And while this is sometimes dismissed as merely a deviation from the ideal of equality, it is in fact well integrated into the practices and institutions of legalisation and legal interpretation – it has been ‘normalised’ as Simpson and others have shown.

In sum, the international legal system demands in various ways that the legal affairs of states be understood in particularist terms – as the expression of the specific circumstances of the parties involved. It cannot treat legal obligations as consistent across the identities of various states. In many areas of international law, there can be no general answer to the legality of an act: from torture to nuclear research to labour standards, to determine what actions are lawful for a state one must first know what rules it has accepted and what reservations or understandings it has imposed on those acceptances.

States possess more freedom toward particular rules of international law than is compatible with the rule of law as understood in domestic affairs. They have the legal right to accept or reject them, and some practical capacity to redefine them, as they wish. This is particularly striking for strong states, as these are given greater weight in the ongoing interpretation of rules in practice and have greater capacity to write, resist, and enforce rules as they wish. This is not to say that the rules are meaningless.

Each of the three pillars often thought to be essential to the domestic rule of law is contradicted by fundamental elements of the interstate system. Sovereign states do not accept that they are automatically bound by stable, shared rules that are applied equally and dispassionately horizontally across cases and vertically between agents and a ‘government’. As a consequence, it is clear that the international rule of law cannot be deduced from the domestic version.