Threats to state survival as emergencies in international law

Abstract

Does international law restrict the use of force by states in self-defense even when their survival is threatened? Should it? To answer these questions, I compare international law to domestic law and develop two ideal-types of emergency: in a ‘subject emergency’ law imposes absolute, justiciable limits on self-defense. In a ‘community emergency’ the sovereign, not law, determines what is necessary for the survival of the community and its legal system: sustaining the rule of law justifies its temporary retreat. I show that international law has elements of both ideal-types. It imposes some absolute limits on self-defense. However, international law also retreats, allowing the victim state to determine the (1) aims, (2) ad bellum proportionality, and (3) end of self-defense, as if armed threats triggered community emergencies. These three retreats serve the function of sustaining the rule of international law over the states at war. Retreats (1) and (3) also help sustain the rule of international law over the international community. That international law does and should not treat armed threats against states simply as subject emergencies, shows it can only sustain the rule of international law in an emergency by retreating. This is a negative litmus test for international law's ability to diffuse anarchy in International Relations.

Does international law restrict the use of force by states in self-defense even when their survival is threatened? Should it? ‘No’, realist scholars of International Relations (IR) would answer. Anarchy means that states must rely on self-help if their survival is threatened. As John Mearsheimer famously put it ‘[i]f a state gets into trouble in the international system, it can't dial 911 because there's nobody … to come to its rescue’.¹ Scholars of IR therefore consider it irrational for states to enter into legal agreements that concern their security. States will simply break such agreements when it suits their interests, certainly when they face an existential threat. Yet today, all 193 states have ratified the UN Charter, a treaty that concerns their security and ostensibly restricts their right of self-defense.

We expect international lawyers to answer both questions with ‘yes’. International law undoubtedly restricts states' right of self-defense. It would undermine the rule of law if states could impose unlimited costs on the international community in their struggle to survive. To uphold the rule of international law, laws that protect the interests of the international community should constrain how states respond to armed threats, even threats to their survival. Yet, in stark contrast to this mainstream position, the International Court of Justice (ICJ) found that it could not ‘reach a definitive conclusion as to the legality of the use of nuclear weapons by a State in an extreme circumstance of self-defense, in which its very survival would be at stake’.² International law, one judge held, is ‘not entitled to demand the self-abandonment, the suicide’³ of a state.

In this article, I argue that it is indeed ambiguous whether international law restricts states' right of self-defense when their survival is threatened. To show this, I compare international law to domestic law and develop two ideal-types of emergencies in domestic law: the ideal-type of domestic law when it regulates individual self-defense is a ‘subject emergency’. In a subject emergency law allows the threatened individual to use defensive force within absolute and justiciable legal limits, like proportionality and necessity. When domestic law regulates individual self-defense, it sometimes demands the self-abandonment of the threatened subject. In contrast, the ideal-type of domestic law when it regulates public emergencies is a ‘community emergency’. In a community emergency domestic law does not restrict self-defense with absolute justiciable limits. Instead law retreats to allow the sovereign to decide what force is necessary to ensure the survival of the community and with it the survival of the legal system. The logic of a community emergency is that it is justified for law to temporarily retreat when the life of the nation is threatened if this retreat secures the rule of law in the long run.

Does international law, when it regulates armed threats against states, resemble the ideal-type of domestic law when it regulates individual self-defense (subject emergency) or the ideal-type of domestic law when it regulates public emergencies (community emergency)? I will show that the answer is ‘both’: states are only allowed to use force in self-defense if the threat they face amounts to an armed attack. Moreover, International Humanitarian Law (IHL) limits how states can conduct defensive wars. These limits on the beginning and the in bello means of self-defense are absolute and justiciable as if armed threats against states triggered subject emergencies in international law. In principle, these legal limits could prevent a state from effectively defending itself. However, in practice, international law also retreats in three ways that make it unlikely that international law would ever demand the self-abandonment of a state: international law leaves it to the victim state to determine the (1) aims, (2) ad bellum proportionality, and (3) end of self-defense, as if armed threats triggered a community emergency for the victim state.

Should international law impose absolute limits on states' self-defense? Or can the three outlined retreats be justified with the normative aim that they sustain the rule of law in an emergency?⁴ The answer to this normative question is likewise more complicated than either ‘yes’ or ‘no’. I demonstrate that, in accordance with the logic of a community emergency, a retreat of some legal rules serves the function of facilitating the warring states' return to the rule of international law. Ultimately, however, the rule of international law depends on the survival of the international community, not the survival of the states at war. International law should therefore not retreat if this risks a conflict escalating to threaten the international community. I show that leaving the determination of ad bellum proportionality to the victim state risks conflict escalation. Only two of the three documented retreats therefore serve the normative aim of sustaining the rule of international law.

What do we learn from this comparison of international law to different ideal-types of emergencies in domestic law? I show that in an international legal order with robust enforcement mechanisms, armed threats to states would and should be treated as subject emergencies, plain and simple. The three retreats of international law would cease to serve the function of sustaining the rule of law over the states at war. Therefore, international law would not retreat. Moreover, the best way to uphold the rule of international law would be to constrain states' self-defense with absolute, justiciable limits. Therefore, international law should not retreat. The finding of the paper that international law does (three retreats) and should (two retreats) regulate threats to state survival partly like community emergencies, reveals that international law lacks the coercive capacity to treat states as subjects, fully subordinated to the rule of international law. That international law needs to retreat to sustain the rule of law in an emergency, I argue, is a negative litmus test for its ability to effectively defuse anarchy in IR.⁵

The argument proceeds in five sections. The first section develops the two ideal-types of emergencies in domestic law. The second section asks what we can learn from investigating which ideal-type international law resembles when it regulates armed threats against states. The third section interprets international law and shows that it has elements of both-ideal-types: it puts up some absolute justiciable limits, but also retreats in three ways, allowing the victim state to determine the (1) aims, (2) ad bellum proportionality, and (3) end of self-defense. The fourth section explains this ambiguity by demonstrating that the three retreats serve the function of allowing the states at war to return to the rule of international law. The final section turns to a normative evaluation of international law. It cautions that only two of the three retreats are justified because leaving the victim state to determine ad bellum proportionality risks conflict escalation and jeopardizes the rule of law over the international community.

Domestic law: two ideal-types of emergencies

Imagine a member of the legal community you inhabit threatens you with violence. Domestic law can treat this as an ordinary rule breach and expect you to seek redress before the courts. Alternatively, law can consider this an emergency for you, the threatened subject: a ‘subject emergency’. In response, law grants you an exceptional permission to use force in self-defense. The force must be necessary, and the permission to use force ends when the police step in. Whether there really was an emergency, which measures were necessary for self-defense, and when the emergency ended are justiciable questions. That means courts may answer them later and hold you to account for what you did in self-defense. In a subject emergency, law is available as a guide and arbiter.

Most domestic jurisdictions afford you a right of self-defense not only in response to threats to your survival, but also to lesser threats. Law defines the legitimate aims of self-defense, such as preserving your physical integrity. Even means that are necessary to achieve these legitimate aims can, however, be impermissible if they are disproportionate.⁶ You are not permitted to kill 50 innocent bystanders even if this is the only way to save yourself. To protect the interests of other subjects and the legal community, domestic law sometimes demands ‘self-abandonment’ of a threatened subject.⁷ Law imposes absolute limits on self-defense; absolute in the sense that even if the limit makes effective self-defense impossible, it does not yield. Even when violence among members of a legal community threatens more than one individual, domestic law tends to treat this as an emergency only for the threatened subjects and grants permissions to use defensive force that have absolute, justiciable limits, like necessity and proportionality.

If, however, violence among members of a legal community or an armed threat from the outside imperils the community itself, not only its subjects, domestic law tends not to create an emergency exception within the law, but an exception to the rule of law. In the famous words of Alexander Hamilton, when the life of a community is threatened, ‘no constitutional shackles can wisely be imposed on the power to which the care of it is committed’.⁸ In a public emergency, domestic law retreats because every claim that the law protects and even the rule of law itself is best served by maximizing the community's survival chances. The survival of the community is a precondition for the survival of the legal system and, eventually, a return to the rule of law. A long line of legal theorists therefore thinks of what I call a ‘community emergency’ as beyond the reach of law.⁹

What counts as a threat to the survival of a community? Here, a community emergency has three cumulative characteristics: an armed threat (1), which threatens the lives of many members of the community (2), and its cultural, political, and legal character (3).¹⁰ Ukraine experiences a threat to its survival, right now, as the Russian aggression imperils its population and its character. Russia might come to pose a similar threat to the Baltic states. In recent history, many states have faced community emergencies. France and Britain in 1940 and Israel between 1948 and 1973 experienced armed threats that, if successful, would have caused massive loss of life and erased the character of these communities. During the Cold War, both blocks assumed that a nuclear attack would constitute an emergency in this sense. Community emergencies do not require a threat by a state. The armed threat by the so-called Islamic State to various states across the Middle East could reach the level of gravity to create a community emergency.¹¹

The argument that it is justified to suspend domestic law when the state faces a community emergency seems to assume that the community prioritizes survival over everything else. If a community valued the rule of law or some values it normally protects with law more than its own survival, should law not govern even a community emergency? Legal rules could prescribe what weight a community accords to self-preservation compared to other values. To retain the restraining force of law, the rules governing a community emergency would have to be attached to procedures that determine when the life of the community is threatened, what measures are necessary to preserve it, and what measures remain illegal even if they are necessary for survival. In other words, the parameters of the community emergency would have to be justiciable.

Justiciable emergency laws would, however, fail to maximize the chances of sustaining the rule of law. They could also end up hurting the values the community seeks to uphold even at the expense of its survival. As soon as the community ceased to exist, these values would be in jeopardy. As Grant Lamond puts it, ‘the existence of a legal system is predicated upon the continued existence of the community whose system it is’.¹² The death of a community is the most definitive disruption of the rule of law over the social and political relations in that community. The best way to advance the rule of law and values normally protected by law is, therefore, to suspend domestic law when a state faces a community emergency.¹³

Even if the rule of law is suspended, could law not still ‘govern’ the community emergency in the sense that it permits its own suspension and the implementation of all measures that are necessary for survival? In John Locke's words, law could give the sovereign the ‘power to act according to discretion for the public good, without prescription of the law and sometimes against it’.¹⁴ To avoid that such a law attempts to preserve the legal system at the expense of a community's survival only to thereby jeopardize both, a legal permission to ‘do whatever it takes’ would have to relinquish all control over its application. In other words, the law that ‘governs’ a community emergency would have to put the determination of whether there is an emergency, what is necessary to address it, and how long it persists in the hands of the agent it empowers. Such a law would be unavailable as a guide or arbiter. This would be the opposite of the rule of law.

If we argue that a community emergency should not be governed by that community's law, we must recognize a form of authority in the state that is not constituted by that state's law. Theorists from Jean Jacques Rousseau to Carl Schmitt have cast sovereignty as ‘the highest, legally independent, underived power’.¹⁵ In this view, the sovereign, not law determines the beginning, permissible means, and end of a community emergency. This does not mean, however, that the emergency is unconnected to law. If the legal justification for why law should retreat is that sustaining the rule of law requires the survival of the community, then the survival of the community and a return to the rule of law is the only legitimate aim of using force. Even though the measures the sovereign takes in a community emergency are not subject to legal review, they are only legitimate if they aim to facilitate a return to the rule of law.¹⁶ The state of exception retains a relation to the rule of law by sustaining it in the long run.¹⁷

Before we conclude this discussion of domestic law, it is important to recall that the community emergency as a state of exception that is beyond the reach of law is an ideal-type. Many domestic jurisdictions have public emergency laws which suspend some laws while leaving others intact.¹⁸ Even though community emergencies in reality are not entirely devoid of law, they are easily distinguishable from the ideal-type of a subject emergency: if a threat triggers a subject emergency, law pre-ordains the permissible reaction. A subject must fit self-defense into the exception that law envisages (legal limits are absolute).¹⁹ Legal review determines whether she has done so (legal limits are justiciable). If a threat triggers a community emergency, in contrast, the sovereign reacts to the emergency by curtailing law as necessary (lack of absoluteness). The sovereign weakens or suspends statutory law or rules by executive order.²⁰ Whether these curtailments are necessary is not routinely subject to legal review (lack of justiciability). In short, a subject emergency features absolute and justiciable limits on the beginning and end of the exception and the means and aims of self-defense. A community emergency does not.

International law: the benefits of domestic analogies

When international law regulates armed threats against states, does it resemble the ideal-type of domestic law regulating individual self-defense (subject emergency) or the ideal-type of domestic law regulating public emergencies that threaten the life of the nation (community emergency)? International law does not depend on the survival of a threatened state. Unlike that state's domestic law, international law would not disappear if the state died. We might therefore expect that armed threats against states trigger something resembling a subject emergency in international law. The third section of this paper will show that this is mostly not the case. Before we interpret international law in detail, however, we need to discuss why we compare international law to ideal-types derived from domestic law in the first place. We could characterize states' right of self-defense as permissive or restrictive to understand how international law treats threats to state survival. What is the benefit of comparing it to domestic law?

Domestic law concepts are pervasive in international law and scholars use comparisons with domestic law, as a matter of course.²¹ Some scholars nevertheless caution that international law is too different to generate valid analogies.²² Most recently, scholars have argued that domestic analogies are fruitful only if we identify a ‘source domain’ in domestic law that is sufficiently similar to the international law we seek to interpret.²³ The purpose of this paper is not to argue that international law should be interpreted in a way to resemble a particular domestic law, however. Resemblance with domestic law is neither my interpretive nor my normative aim in this article. Instead, I use two different source domains in domestic law – individual self-defense and public emergencies – to interpret, explain, and evaluate international law. I ask which ideal-type international law resembles (third section), why it resembles both ideal-types to some extent (fourth section), and which one it should resemble if the normative aim were to sustain the rule of international law (fifth section).

Using competing domestic analogies in this way has a long tradition in international law scholarship, for instance in debates about whether state sovereignty resembles private property or trusteeship and whether international treaty-making resembles legislation or private contracting.²⁴ It allows us to understand whether and how international law meets the ‘needs’ that all legal orders have.²⁵ In this case, all legal orders need to maintain the rule of law when violence creates an emergency. Domestic law meets this need in radically different ways in a subject and a community emergency. In a subject emergency, law retains an iron grip on its subjects as a guide and arbiter and thereby sustains the rule of law. In a community emergency, law temporarily sacrifices the protection of its subjects and compliance with its rules. It can sustain the rule of law only by retreating. Comparing international law's treatment of armed threats against states to these two ideal-types reveals how international law sustains the rule of law in an emergency.

How international law sustains the rule of law in an emergency, in turn, indicates whether international law has developed the coercive capacity to treat states as subjects, fully subordinated to the rule of international law, even when they face armed threats. Or must international law retreat in an emergency, behind the sovereign prerogative of states as legal communities in their own right? The status of states as either subjects or sovereigns²⁶ is associated with two competing visions of international law. International lawyers have debated these competing visions in terms of a conceptual spectrum ‘from bilateralism to community interest’²⁷ or as international law being a law of ‘coexistence’ vs. as law of ‘cooperation’.²⁸ Scholars of IR have debated these competing visions in terms of the power of international law to constrain state behavior in IR. Let me explain how these two competing visions imply different expectations about which ideal-type of emergency international law resembles when it regulates armed threats against states. In turn, which ideal-type international law resembles is indicative of how international law operates now in IR.

In the traditional vision, international law is a web of mostly bilateral contracts and customs that allow states to coexist. Sovereign states partake in customs or contract into legal obligations, which they shed when law ceases to serve their interests.²⁹ Obligations that limit self-defense create the paradigm case of a clash between law and a state's interests. We should therefore expect armed threats to resemble community emergencies for the victim state. Law retreats behind the state's sovereign prerogative to determine what force is necessary to ensure survival. The status of states as sovereign legal communities trumps their status as subjects under international law. As states can always contract out of their obligations, a bilateralist legal system does not feature institutions that reliably enforce international law. The only way for states at war to return to the rule of international law is therefore to ‘fight it out’. Sustaining the rule of international law over states at war requires temporarily curtailing it. This vision evokes an understanding of international law long mainstream among realist and rational-institutionalist IR scholars: international law attracts compliance only if it does not clash with states' security interests.³⁰

According to the alternative vision, international law serves the interest of the international community. Community interests prevail over the interests of individual states. For the purposes of international law, states are not legal communities. They are subjects and this status obviates their sovereignty. In this view, it is uncontroversial that states have legal obligations that they cannot contract out of even if these obligations contradict their interests.³¹ We should therefore expect that international law treats threats to state survival like subject emergencies. As part of this ideal, the institutions of collective security enshrined in the United Nations act in the place of an international sovereign with a monopoly on the use of force. They ensure that states at war comply with international law. There is no reason to treat threats to state survival as anything other than subject emergencies. IR scholarship that argues that international law can change how states define their interests because legal norms become internalized, proffers an understanding of international law closer to this vision.³²

Many domains of international law now resemble the latter, more hierarchical vision where enforceable international laws serve the interests of the international community and states are mere subjects.³³ For instance, human right law is increasingly interpreted to deemphasize state sovereignty and to uphold universal enforceable standards. Similarly, IR scholars acknowledge that international institution-building over the last 75 years means international law now reliably constrains state behavior in many areas of IR. Armed threats to states, however, pose the greatest challenge for international law to maintain its grip over its subjects. Does international law (subject emergency) or the sovereign state (community emergency) decide in an emergency? The next section will reveal the answer.

A legal interpretation: self-defense between subject and community emergency

This section systematically interprets international law on the use of force to describe whether it restricts states' right of self-defense even when their survival is threatened. The following four subsections ask whether (a) the beginning of states' right of self-defense, (b) the aims of self-defense, (c) the means of self-defense, and (d) the end of the right to use defensive force are each subject to absolute, justiciable limits, as we would expect if armed threats against states resembled subject emergencies. Or does international law retreat to let the victim state determine the beginning and end of the exception and what is necessary to ensure state survival, which we would expect in a community emergency? Crucially, the survey of legal sources in the following subsections does not draw only on self-defense cases in which the victim state's survival was obviously threatened. Who determines whether an armed threat imperils state survival depends on whether international law resembles a subject emergency (international law) or a community emergency (the victim state). We must therefore review all applications of the international law of self-defense to answer the question which ideal-type international law resembles.

The beginning of the exception

If armed threats to states resembled subject emergencies, we would expect international law to establish a threshold for when the threat or use of force against a state warrants an emergency permission to use force. Customary international law and Article 2(4) of the UN Charter prohibit the threat or use of force by states.³⁴ States' right of self-defense is the exception.³⁵ Scholarship suggests that a threat of force is not enough to permit defensive force.³⁶ Moreover, not all uses of force on another state's territory trigger the right of self-defense.³⁷ The threshold for the emergency exception is ‘if an armed attack occurs’. A state that is a victim of lesser force must defend itself by non-forcible means.³⁸ State practice and Security Council Resolutions provide ample examples of states and institutions protesting resorts to force against lesser threats.³⁹ The question of whether an armed attack occurred is justiciable. Several cases before the ICJ confirm that international law does not leave it to states to decide whether they face a threat that warrants a response by force.⁴⁰

Today, states face external armed threats not only from other states, but also from non-state actors, such as armed groups. The question whether an armed attack by an external non-state actor triggers a state's right of self-defense has likewise played out before the ICJ.⁴¹ Monica Hakimi shows that defensive operations against ISIL in Syria have weakened the position in scholarship and practice that only states can commit armed attacks.⁴² Although international law remains unsettled, it is not in question that the use of defensive force is subject to a gravity threshold also when deployed against non-state actors.⁴³ Indeed, the intense controversy around when a state can resort to defensive force against non-state actors corroborates that international law, not the victim state, determines what triggers the right of self-defense.⁴⁴

The absolute and justiciable armed attack threshold suggests that armed threats against states in international law trigger a subject emergency. In principle, if a use of force below the threshold of an armed attack threatened a state's survival and that state could not defend itself by non-forcible means, international law would demand the ‘self-abandonment’ of its subject. In reality, it is hard to imagine that an armed threat is both existential and too insignificant to count as an armed attack.

Both subject and community emergencies begin when using force is necessary. In the case of a subject emergency, the question whether self-defense was necessary would be justiciable. In the case of a community emergency, the sovereign, not law, would determine whether defensive force is necessary. Even though Article 51 of the UN Charter does not mention necessity, resort to defensive force must be necessary as a matter of customary law.⁴⁵ The ICJ has interpreted necessity to mean that states can only resort to force in self-defense against an armed attack if there are no peaceful alternatives.⁴⁶ In the Oil Platforms Case, the Court called this requirement ‘strict and objective’, seemingly affirming the justiciability of the question whether resort to force in self-defense against an armed attack was necessary.⁴⁷

In practice, however, Akande and Liefländer show that ‘a state is not required to seek or use alternative means’ of defense before resorting to force against an armed attack.⁴⁸ The authors' examination of state practice demonstrates that even if a state ‘has a more realistic chance of achieving a cessation of the attack by other means, practice does not deny that the state has an “inherent” right to use force’.⁴⁹ Instead, Akande and Liefländer diagnose international law with an ‘almost irrebuttable presumption’⁵⁰ that the resort to force in response to an armed attack meets the ad bellum requirement of necessity. In its landmark cases on the use of force, the ICJ has indeed focused on the question whether the initial threat amounted to an armed attack.⁵¹ In the above-mentioned Oil Platforms Case, the Court failed to determine whether the USA had peaceful alternatives. Instead, the Court argued that the actions of the USA, specifically its failure to complain to Iran about the military activities on the oil platforms, did ‘not suggest that the targeting of the platforms was seen as a necessary act’ [emphasis added].⁵² Strikingly, this pronouncement implies that the trigger for the right of self-defense is the perception of the victim state that force is necessary, not an objective lack of peaceful alternatives.

Customary law also demands that resort to self-defense is proportionate.⁵³ What that means, however, is unclear. In Thomas Franck's interpretation, ad bellum proportionality is part of the determination of whether the initial armed threat meets the threshold for self-defense.⁵⁴ If defensive force is a proportionate response, we know that the threat amounts to an armed attack. Yoram Dinstein argues that once we conclude that the armed threat crosses the threshold of an armed attack, we should presume that resort to defensive force is proportionate.⁵⁵ Both Franck and Dinstein purport then that the gravity of the initial armed attack alone triggers the emergency. The principle of proportionality does not further limit access to the permission to use defensive force.

In sum, international law sets a justiciable gravity threshold for when the use of force against a state warrants forcible self-defense (an armed attack occurs), which suggests that armed threats to states in international law resemble subject emergencies. However, ad bellum necessity as a condition for resorting to self-defense turns out to be neither ‘objective’ nor ‘strict’. In practice, if an armed attack occurred, the determination of whether self-defense is necessary lies with the victim state. The proportionality of resorting to defensive force is assumed. Both ad bellum principles, necessity and proportionality, retreat behind the prerogative of the victim state to determine that it faces an emergency.

The legitimate aims of self-defense

The question of what triggers an emergency is closely related to the question of what aims defensive force should pursue. In a subject emergency, law determines the legitimate aims of self-defense; legal review determines whether the subject pursued a legitimate aim. In contrast, if armed threats resembled a community emergency, the legitimate aim of self-defense would be the survival of the state whereby not international law, but the sovereign state itself would determine what concrete end-state secures its survival. In the Nuclear Advisory Opinion, the ICJ implied that the use of force can amount to an armed attack even if it does not threaten the survival of the victim state.⁵⁶ Self-preservation is not the only legitimate aim of self-defense in international law. Beyond this, neither the ICJ nor Article 51 mentions the aims of self-defense.

Scholarly opinion is conflicted. One prevalent interpretation of Article 51 is that self-defense should seek to ‘halt and repel’ the armed attack.⁵⁷ David Kretzmer considers this too narrow and argues ‘that the legitimate aims of using force in self-defense may differ, depending, inter alia, on the nature and scale of the armed attack, the identity of those who carried it out, and the preceding relationship between the aggressors and the victim state’.⁵⁸ Deterrence and the prevention of future threats are often argued to constitute legitimate aims of self-defense.⁵⁹ Christian Tams has noted that states moreover sometimes seek retaliation against their attacker, which is rarely challenged as exceeding the permissible aims of self-defense.⁶⁰

Even scholars who agree that a state must do no more than to ‘halt and repel’ the armed attack admit that it is contested what that means. Many argue that it must be more than a permission to end an ongoing aggression.⁶¹ In reality, states at war often seek political end-states that do not resemble the status quo before an armed attack. Instead these end-states concretize the victim state's understanding of what it means to be free from the threat.⁶² Based on this practice, Georg Nolte argues that self-defense implies ‘the freedom to repel … an attack in a way which ensure[s] that the underlying immediate threat [is] removed’.⁶³ He conceives of the state of being without threat as ‘socially constructed’.⁶⁴ The argument that there is no objective benchmark for what it looks like when a state repels an armed attack, implies that the victim state determines the legitimate aims of self-defense. This evokes a community emergency for the victim state: according to the community emergency logic, the sole legitimate aim of self-defense is self-preservation. The sovereign state determines what that looks like. It is not stipulated by international law or adjudicated by courts.

The permissible means of self-defense

Does international law limit the means of self-defense? The ad bellum principle of necessity could be read to imply not only that there must be no peaceful alternative to resorting to force, as discussed above. It could also mean that the kind of armed violence a state uses in self-defense must be the least that is still effective in achieving the legitimate aims of self-defense.⁶⁵ As it is in practice the victim state that defines the legitimate aims of self-defense, it would also be the victim state rather than international law that determines what kind of defensive force is necessary. Ohlin and May, moreover, show that it is a long-standing view that defensive military campaigns should not be scrutinized for the kind and amount of violence they use.⁶⁶ Even if ad bellum necessity operated as an absolute limit on the means of self-defense, it would not imply that a state could be unable to secure its survival. Only the justiciability of necessity would indicate that armed threats resemble subject emergencies. Yet, the previous section revealed that the ICJ does not treat ad bellum necessity as an objective, justiciable condition of self-defense in concrete cases.

The ad bellum principle of proportionality as a limit on the means of self-defense would imply that even measures that are necessary for state survival could be impermissible if they were disproportionate. Scholars who consider ad bellum proportionality a threshold condition for resort to defensive force therefore deny that it limits the means of self-defense. Dinstein argues that ‘once war starts, [ad bellum] proportionality … does not interfere with the ability of the state to fight to victory’.⁶⁷ Other scholars object to the interpretation that ad bellum proportionality requires merely a one-off judgment when a state first resorts to self-defense. Instead proportionality requires an ongoing assessment of the means of self-defense. Two positions exist about what that ongoing assessment looks like.

First, some scholars argue that the means of self-defense must be proportionate to the initial attack.⁶⁸ In a recent examination, Eliav Lieblich shows that advocates of such a backward-looking application of proportionality (toward the armed attack) tend to accept that ‘as a practical matter, the role of ad bellum proportionality diminishes as a conflict intensifies’.⁶⁹ Robert Sloane, for instance, argues that ‘[r]ealistically, in any sustained conflict, the effect of ad bellum proportionality as a genuine constraint on force will diminish as the objectives of that force multiply and expand’.⁷⁰ In practice, this position comes closer to presuming the ad bellum proportionality of force in response to an armed attack than to implying concrete limitations on the means of self-defense.

ICJ judgments mostly feature this backward-looking understanding of proportionality.⁷¹ Although the court affirms in the abstract that the means of self-defense must be proportionate to the armed attack, it has not specified what measures count as proportionate in concrete cases. Christodoulidou and Chainoglou brand the approach ‘confusing, if not phobic’, given the court's ‘reluctance to define or even analyze the dimensions of proportionality’.⁷² Furthermore, several times, the ICJ has characterized proportionate measures as those that are ‘necessary’ to respond to an armed attack.⁷³ Collapsing proportionality into necessity precludes that measures that are necessary for self-defense could turn out to be illegal, ruling out that international law might ask for the ‘self-abandonment’ of a state.

The second position likewise demands a continuous assessment of the means of self-defense. However, it requires a forward-looking assessment that compares the means of self-defense to the value of the aims of self-defense.⁷⁴ Again, the possibility that measures that are necessary for effective self-defense could be disproportionate and thus impermissible makes this interpretation controversial. Some scholars who endorse forward-looking proportionality judgments therefore equate proportionality with measures that are necessary for achieving a state's legitimate defensive aims.⁷⁵ Even if forward-looking proportionality was not collapsed into necessity, the principle would not amount to an objective, and therefore potentially absolute and justiciable, limit on defensive force as long as the victim state determines the legitimate aims of self-defense.

The International Law Association in its report on aggression recommends a forward-looking proportionality judgment. It states that the ‘proportionality assessment should weigh the relative interests against each other, thereby assessing whether the harmful effects of the force taken in self-defense are outweighed by achieving the legitimate aims’.⁷⁶ Although the report defines the legitimate aims of self-defense as ‘halt[ing] any ongoing attack and prevent[ing] the continuation of further attacks’,⁷⁷ it offers no indication of whose relative interests we should take into account, what harmful effects count in the assessment of ad bellum proportionality, and in light of what values we should assess the importance of halting an armed attack. The lack of writing about the stakeholders, harms, and values relevant to ad bellum proportionality indicates that neither states nor international institutions routinely ask whether a particular means of self-defense is proportionate to the importance of achieving a particular war's aims.⁷⁸ This is, however, what we would expect if armed threats to states resembled subject emergencies in international law.

So far, the only limitation on states' right of self-defense that is determined by international law and that is subject to routine legal review in concrete cases is the threshold condition that ‘an armed attack occurred’. It suggests that armed threats to states trigger a subject emergency. However, as mentioned, this legal limit is unlikely to interfere with a threatened state's self-preservation because armed threats below the threshold are unlikely to be existential. At the same time, the victim state determines the legitimate aims of self-defense and the ad bellum proportionality of defensive force. These two retreats of international law behind the prerogative of the victim state suggest that, in international law, armed threats resemble community emergencies for the victim state.

Although, in practice, ad bellum proportionality and necessity do not limit the means of self-defense, IHL does just that. Many of IHL's customary restrictions on the conduct of hostilities are insensitive to military necessity. The prohibitions on directly attacking civilians,⁷⁹ on indiscriminate attacks,⁸⁰ on perfidy,⁸¹ on methods and means of warfare ‘which cannot be directed at a specific military objective’,⁸² or those which ‘cause widespread, long-term and severe damage to the natural environment’⁸³ are all absolute in the sense the term is employed here. These restrictions remain applicable even if they make effective self-defense impossible. Similarly, IHL's demand that each attack has to be proportionate indicates that attacks can be unlawful even if they are necessary to achieve a particular military advantage and perhaps effective self-defense.

Moreover, IHL's rules for the conduct of hostilities are insensitive to the ‘causes espoused by or aims attributed to’ states fighting a defensive war.⁸⁴ Whether a state is defending itself against a ‘minor’ armed attack or whether it struggles for survival has no effect on the implications of these restrictions. Each absolute restriction may therefore in extreme circumstances amount to a demand that a state abandons effective self-defense. Domestic and international courts and tribunals holding individuals to account for violations of the laws and customs of war show that these limits are justiciable. When international law regulates armed threats to states, it not only resembles a community emergency for the victim state. IHL's limits on the in bello means of self-defense evoke a subject emergency.

The end of the exception

In the domestic context, the intervention of law enforcement ends the subject emergency because we assume that it makes self-defense unnecessary. In contrast, the sovereign determines the end of a community emergency. In international law, Article 51 demands that ‘[m]easures taken by Members in the exercise of this right of self-defense shall be immediately reported to the Security Council’. As we would expect in a subject emergency, states remain accountable even after they have resorted to self-defense. The text further appears to indicate that international law ties the duration of the permission to use force to Security Council involvement: ‘nothing shall impair the inherent right of individual or collective self-defense … until the Security Council has taken measures [emphasis added]’. Is it the role of the Security Council to end states' right of self-defense as we would expect in a subject emergency?

As self-defense is described as an ‘inherent’ right, some scholars have held that the duty to report defensive measures does not affect a state's right to defend itself.⁸⁵ Majority opinion, in opposition, agrees that the Security Council matters for the question when defensive force must end.⁸⁶ According to William Lietzau, the drafters envisaged that the victim state's right to use force would end with the report of defensive measures to the Council: ‘The collective security aspects of Security Council authority were designed to be sufficient to meet most needs for self-defense’.⁸⁷ In reality, the Security Council is not a reliable law enforcement body. One Commentary concludes that ‘the system of collective security has been of little practical significance, (…) international legal practice since 1945, contrary to the intentions of the authors of the Charter, has continued to be determined by unilateral use of force by states’.⁸⁸

Putting aside its lack of practical relevance, is the Security Council even meant to be a law enforcement institution? If its involvement were designed to end the right of self-defense, we would expect Article 51 to mention rescuing the victim state. Instead, the provision empowers the Council to take all measures ‘to maintain or restore international peace and security’. Some scholars argue that this mandate coincides with the goal of enforcing international law.⁸⁹ Yet, the restoration of international peace and security may not always be inclusive of the defense of the victim state. Even if the drafters intended restoring peace and security to imply intervention on behalf of the victim state, we would need to know what kind of intervention by the Council would cause the state's defensive rights to lapse.

Does the right of self-defense end as soon as the Security Council becomes ‘seized of’ a matter, when it takes measures to respond to the armed attack, or only when those measures prove effective?⁹⁰ If we take the object and purpose of Article 51 to include the protection of states' rights, the right of self-defense should lapse only when the Council's measures render self-defense unnecessary. However, international law, as discussed above, does not determine the legitimate aims of self-defense. It thereby fails to determine for what defensive force has to be necessary. By implication, it fails to define for what Security Council measures would have to be sufficient. By logical implication, the victim state that defines the aims of self-defense would also determine whether it is necessary to continue to exercise self-defense after the Council has become involved.

In practice, other developments that might render self-defense unnecessary, for instance an opportunity to negotiate with the aggressor, are not understood to terminate a state's right to use force in self-defense either.⁹¹ Similarly, scholars who maintain that ad bellum proportionality requires a continued assessment, whether backward- or forward-looking, do not argue that states have to sue for peace when the overall amount of defensive force they have used becomes disproportionate.⁹² Instead, in practice, it is the victim state's assessment that the threat is neutralized that ends the right of self-defense. This evokes a retreat of law behind the prerogative of the victim state, the kind of retreat we would expect if in international law armed threats triggered a community emergency for the victim state.

In reality, a state will often cease to use defensive force, not because it deems the threat neutralized, but because it was defeated. The equality of belligerents under IHL means that IHL is agnostic toward who prevails, the aggressor or the state exercising self-defense. Belligerent equality does not formally affect the status of the resort to force or the responsibility of belligerent parties under international law. However, in its effect on when and how an armed conflict ends, belligerent equality amounts to a suspension of international law's asymmetrical estimation of the belligerents' respective aims. It is a temporary retreat of the rule of international law. International law by implication accepts the old adage that ‘war does not determine who is right, only who is left’.

Summary of the legal interpretation

The third section revealed that when international law regulates armed threats against states, it features elements of both ideal-types of emergency. When domestic law regulates individual self-defense (subject emergency) it sets a gravity threshold and restricts the means of self-defense. Similarly, Article 51 sets the threshold for when an armed threat warrants forcible self-defense (an armed attack occurs). Moreover, IHL restricts the means of self-defense regardless of whether this precludes survival. These absolute limits on the beginning and in bello means of self-defense are justiciable as we would expect in a subject emergency.

At the same time, international law retreats from defining three parameters of self-defense: first, in practice, the threatened state determines the aims of self-defense. Second, in practice, ad bellum proportionality is not applied to restrict a state's means of self-defense. Third, international armed conflicts, and with them the victim state's right of self-defense, end when states have ‘fought it out’. They do not end when the Security Council intervenes on behalf of the victim state or when the victim state runs out of proportionate force. These three retreats suggest that international law, when it regulates armed threats against states, also resembles the ideal-type of how domestic law regulates public emergencies that threaten the life of the nation, that is, community emergencies. International law on the use of force features elements of both ideal-types of emergency.

A functional explanation: why does international law retreat?

Can we explain why international law features elements of both ideal-types of emergencies? In particular, why does international law retreat from imposing absolute, justiciable limits on the (1) aims, (2) ad bellum means, and (3) end of self-defense? International law emerges from several sources. Monica Hakimi and Jacob Katz Cogan have argued that the agents behind these sources have divergent interests when it comes to the resort to force, which explains why there are ‘two codes’: international courts and institutions maintain a restrictive code whereas states that seek to preserve their freedom of action practice a more permissive code.⁹³ Do the three retreats emerge from certain sources?

No. State practice mostly, but not exclusively, evokes a community emergency for the victim state. Treaty law is ambiguous. Article 51 casts self-defense as subject to absolute, justiciable limits. IHL's codified rules on the conduct of hostilities likewise suggest that armed threats trigger subject emergencies, but its principle of belligerent equality does the opposite. The ICJ's abstract pronouncements on the law of self-defense evoke a subject emergency. When the ICJ applies the law to concrete cases, meanwhile, it suggests that the victim state determines the aims and ad bellum means of self-defense. Scholarly opinion is likewise conflicted. Put simply, each source of international law paints on its own an ambiguous picture. The three retreats are not tied to specific sources of law, instead they concern specific parameters of self-defense – the aims, ad bellum means, and end of the exception.

It is important to stress that the three retreats are not formal suspensions of the applicability of international law. Except belligerent equality, the three retreats stem from state practice and from the ICJ's not raising certain questions when adjudicating concrete cases. The retreats also stem from scholars, the Security Council, and other international institutions not interpreting Article 51 and customary law as implying absolute limits on state conduct, limits that these sources could be argued to imply. In other words, the three retreats are the result of interpretive choices that depart from a strict reading of the treaty text and from the abstract definition of the customary right of self-defense. Can we explain the interpretive choices that imply the three retreats? I will argue that these interpretive choices serve a function. That function is the need of the international order to sustain the rule of international law over the states at war. Let me explain.

International law suffers from what Prosper Weil called a ‘structural weakness’ due to ‘the inadequacy of its sanction mechanism’.⁹⁴ Unlike domestic law, which is associated with the coercive power of a sovereign, international law relies on states' beliefs in its legitimacy, their internalization of its demands, and their interests in compliance.⁹⁵ States that violate their obligations rarely encounter mandated force. They are sometimes shamed or incur economic sanctions. It depends on the political context whether states deploy these measures against each other and whether they are costly for the deviant state. International law often elicits compliance, relying on states' perceiving law as legitimate or on a convergence between the demands of law and how states want to behave anyway. However, without coercive capacity, international law can ill-afford to make a demand that states have a decisive interest in defying every time they encounter it. Restrictions on states' capacity to ensure their survival are the paradigm case of such a demand. Effective self-defense is where push comes to shove.

In the absence of a sovereign that enforces international law, the best way to sustain the rule of international law over states at war may therefore be to defer to these states' own determination of what end-state restores their safety. By the same token, the path toward a return to IR under the rule of international law may well be to allow states to ‘fight it out’ until, in their own estimation, the emergency has ended. It may be fruitless to insist that proportionality limits the means or spells the end of self-defense in the absence of an executive that can rescue the victim state or force it to sue for peace. The three retreats which temporarily curtail the rule of international law over the states at war hence serve the function of allowing these states to return to the rule of international law in a legal system in which no sovereign reliably enforces international law. The three retreats then follow the logic of a community emergency, that is, a temporary retreat of the rule of law secures the rule of law in the long run.

Functionalism can only explain a phenomenon – here, interpretive choices that imply the three retreats of international law – if we can identify the mechanism by which the phenomenon emerges.⁹⁶ Can we explain these interpretations? We can. The first mechanism is the goal of securing compliance even when states are at war. As shown in the third section, contestation over the principle of ad bellum proportionality in scholarship is fueled by concerns for whether the principle will attract compliance. Similarly, scholars frequently advance as a reason for upholding belligerent equality that it facilitates compliance with IHL.⁹⁷ The debate about the aims of self-defense is also strongly influenced by what limits scholars think states will accept.⁹⁸ In general, scholars often acknowledge that not undercutting international law's authority by making demands that have little chance of being followed is a goal of their interpretation.⁹⁹ In short, when international lawyers and scholars seek to secure compliance, a legal order without a proper sovereign ‘selects for’ interpretations that imply the three retreats.

Why do interpretations of the armed attack threshold and IHL's rules for the conduct of hostilities as absolute, justiciable limits on self-defense not raise the worry that they might undermine compliance? The consequences of these interpretations for the effectiveness of self-defense are less direct and less predictable. As mentioned, it is unlikely that legal review finds an armed threat to be below the threshold of an armed attack, but the threat proves existential.¹⁰⁰ Although it is possible that a state can only win a defensive war by violating IHL, it is not predictable in advance that IHL's putting a particular attack off limits means a state must abandon effective self-defense. Absolute limits on the beginning and in bello means of self-defense therefore stand a better chance of attracting compliance than absolute limits on the aims, ad bellum means, and end of self-defense. The goal to secure compliance favors interpretations of international law that defer to the victim state regarding those parameters of self-defense that directly and predictably affect a state's capacity to secure survival.

The second mechanism that explains interpretations giving rise to the three retreats of international law is pragmatism about the executive limitations of international institutions. When the ICJ affirms the applicability of ad bellum proportionality, in general, while not interpreting this principle as an absolute, justiciable marker for the end of concrete self-defense cases, it implicitly acknowledges that it would be unrealistic to demand that a state abandon the use of force half-way through a defensive war. It is unrealistic because the Security Council does not systematically end international armed conflicts by intervening on behalf of the victim state as we would expect in a subject emergency. The Council does not regularly pronounce on what it means to halt an armed attack in concrete cases either. Put differently, the Council fails to even attempt to enact what Chapter VII and Article 51 proffer. By default, it leaves it to the victim state to determine the aims of self-defense and when the emergency ends. The Security Council practices international law by retreating from these tasks. Other institutions, courts, and scholars, when practicing international law, reflect these retreats and thereby affirm them.

Were the coercive capacities of international law improved, for instance, if we imagined a Security Council that acted as a reliable law enforcement body, international legal practice would change. Neither the goal to secure compliance nor the goal to accommodate executive limitations of international institutions would lead to interpretations that give rise to the three retreats. The treatment of armed threats to states in international law would resemble the treatment of individual self-defense in domestic law: a subject emergency plain and simple. The findings of the third section then suggest that international law still lacks the coercive capacity to meet the need of all legal orders to sustain the rule of law in an emergency by holding the states at war to absolute, justiciable limits (subject emergency). Instead, it must retreat from limiting self-defense, at least along some parameters (community emergency). The status of threats to state survival as an emergency under international law is a litmus test for whether international law effectively defuses anarchy in IR. The three retreats uncovered above suggest that this test is negative.

In conclusion, this section explained why international law treats armed threats against states partly as community emergencies. Interpretations that imply that international law retreats to allow the victim state to determine the (1) aims, (2) ad bellum means, and (3) end of self-defense serve the function of sustaining the rule of international law over the states at war. When scholars, courts, and institutions seek to secure compliance with international law or reflect the executive limitations of international institutions, the weakness of international law's coercive capacity selects for interpretations that defer to the victim state who then determines the aims, ad bellum means and the end of self-defense. I showed that the beginning and in bello means of self-defense are subject to absolute, justiciable limits because these limits do not directly or predictably make effective self-defense impossible. These limits do not implicate state survival. If international law's coercive capacity improved, the three retreats would stop serving the function of allowing the states at war to return to the rule of law and international law would treat armed threats to states as subject emergencies, plain and simple.

A normative evaluation: should international law retreat?

Should international law restrict the use of force by states in self-defense even when their survival is threatened? Or can the three retreats be justified? We could answer this question by holding international law to a normative standard that is external to it, for instance a moral standard. Philosophers have contemplated whether moral considerations should ever prevent a state from effectively defending itself. Michael Walzer famously argued that a state could violate the ‘war convention’ in a ‘supreme emergency’.¹⁰¹ By that he meant that a state facing a threat to its survival could deliberately target civilians if this was necessary for survival.¹⁰² According to Thomas Nagel, threats to state survival can create a dilemma between the moral duty not to kill the innocent and the duty to protect a community from disappearance. He suggested that the latter was weightier.¹⁰³ For two reasons, we cannot easily evaluate the three retreats of international law with this moral standard.

First, both Walzer and Nagel attribute moral value to communities.¹⁰⁴ In this view, communities can violate categorical moral prohibitions to prevent their demise, even though individuals who are unjustifiably threatened with death are not morally permitted to deliberately kill innocent bystanders to save themselves. That is because, in this view, at least some communities have moral value beyond the value of the individuals that are their members.¹⁰⁵ If we assessed the international law of self-defense against this moral standard, we would therefore have to judge the moral value of the victim state. International law, however, does not account for the moral value of the communities that constitute states. The retreats of international law, described in the third section and explained in the fourth section, empower states regardless of the moral value of their communities.

A second difficulty in evaluating the three retreats of international law with a moral standard is that the retreats are not permissions to violate concrete prohibitions if, and only if, effective self-defense is otherwise impossible. Instead, international law defers to the victim state in all self-defense cases to determine how some legal limits are met. It depends on the circumstances whether this deferral leads to a violation of moral duties. Similarly, it is not possible to determine in general whether the three retreats have better or worse moral consequences than international law's attempting to set absolute, justiciable limits on all parameters of self-defense, that is, treat armed threats against states purely as subject emergencies. Again, our moral judgment depends on the moral value of community survival in a particular case and on what the victim state does with its prerogative to determine the aims, ad bellum means, and end of self-defense.

If not based on a moral standard, how can we evaluate the three retreats then? The normative argument introduced in the first section, that domestic law should not govern a community emergency, does not rely on a moral standard. The logic of a community emergency is that the normative aim of sustaining the rule of law justifies its temporary retreat. That retreat has costs. Curtailing the rule of law, even just for a while, weakens the values that a community normally protects with law, including the rule of law itself. Law temporarily sacrifices the protection of its subjects and compliance with its rules. The logic rests on the assumption that these costs are justified because the values are preserved for the long-term and members of a threatened community will continue to benefit from them as both the community and the rule of law survive. This logic is self-referential. It operates regardless of the substantive moral value of a community. In that sense it is a normative argument about what law should do (sustain itself in an emergency), but not a moral argument that holds law to a moral standard that is external to it.

What if we drew on this community emergency logic to evaluate the three retreats of international law? The states at war benefit from the retreats when they eventually return to the rule of international law, as envisaged by the community emergency logic. However, the domestic analogy gets complicated here. If the three retreats allow an armed conflict to escalate, they have costs also for the members of the international community. Moreover, sustaining the rule of international law over the states at war is not the same as sustaining it over the whole international community. The latter is the aim that a retreat of international law should serve according to the community emergency logic. After all, the rule of international law depends on the survival of the international community in the way that domestic laws depend on the survival of the state. It would turn the logic of a community emergency on its head if international law retreated to facilitate a return of the states at war to international law, but this retreat allowed a conflict to escalate and thereby jeopardize the international legal system. International law should preclude that a community emergency for the victim state becomes an ‘international community emergency’, that is, a threat to the survival of the international community.

What is a threat to the survival of the international community? According to the definition of a community emergency adopted in the first section, it would be an armed threat that imperils many members as well as the cultural, political, and legal character of the international community. Unlike community emergencies for individual states, we cannot easily point to many examples of international community emergencies, though scholars of IR have long warned that a nuclear war would threaten the survival of the international community. Recent threats by Russia to resort to a nuclear weapon in the context of its invasion of Ukraine have concretized this potential threat to the survival of the international community. Regardless of how close we think such a threat is in reality, if we take seriously that international law depends on the survival of the international community and the overriding normative aim is therefore sustaining the rule of international law over this community, then international law must prevent that armed conflicts escalate. Not all escalations of armed conflicts threaten the survival of the international community. However, the weakness of centralized enforcement powers that can reliably reign-in an escalating conflict suggests that international law should treat any escalation beyond the states at war as a first step toward an international community emergency.

It follows that according to the community emergency logic, international law should restrict the use of force by states facing an existential threat even if this means they abandon the rule of international law. It should do so if a restriction is apt to forestall conflict escalation to third states. Could international law's retreats from limiting the aims, ad bellum means, and end of self-defense escalate an armed conflict? Leaving the determination of the end of self-defense to the victim state might prolong rather than spread the conflict. If a state defined the aims of self-defense too expansively, a conflict could escalate in the sense of spreading beyond the states involved. Crucially, international law does not need to restrict the aims of self-defense to forestall escalation if it limits the ad bellum means of self-defense.

If ad bellum proportionality were applied in a forward-looking manner and not collapsed into necessity, it would require that any harm that a defensive war is expected to cause to the international community, including harms to third states, be proportionate to the value of a state's war aims. Let us imagine that the weightiest possible aim of self-defense is a state's self-preservation. Any means of self-defense that has the foreseeable cost of threatening the international community would reasonably be considered disproportionate even to an aim as important as state survival. After all, what is at stake is the survival of the international community and with it the rule of international law, not only over the warring states, but tout court. This understanding of ad bellum proportionality is not established law. This is how the principle would be applied if it were forward-looking and not collapsed into necessity.¹⁰⁶ This is how it should be applied if the overriding normative aim were sustaining the rule of international law.

One might ask why in bello proportionality is not sufficient to prevent the escalation of armed conflicts. Without doubt, it provides an absolute, justiciable limit on the means of self-defense. In bello proportionality requires a balance between the concrete and direct military advantage anticipated to arise from an attack and the expected incidental harm to civilians. Much is contested about the interpretation of this principle, but it is sensitive neither to the harms an attack causes to the international community,¹⁰⁷ nor to the value of a state's overall war aims.¹⁰⁸ Ad bellum proportionality is both, if applied in a forward-looking manner.¹⁰⁹ It is therefore the most suitable rule of international law to forestall escalation of an armed conflict beyond the states involved. It is the right tool to avoid the emergence of an international community emergency. It follows that the normative aim of sustaining the rule of international law does not justify the retreat of international law from limiting the ad bellum means of self-defense, even if the retreat facilitates that the states at war return to international law.

Would it really make a difference if scholars, courts, and international institutions started interpreting ad bellum proportionality as an absolute, justiciable limit on self-defense, one that required that the means of self-defense are weighed against the importance of a state's defensive war aims, and one that treated escalation risks as disproportionate even to the aim of state survival? A state at war would still have an incentive to violate proportionality. Even if interpreted in this way and routinely reviewed by courts, ad bellum proportionality would remain open-ended. What means of self-defense are disproportionate because they risk escalation would also be contestable. The finding that one of the three retreats of international law is not justified and that we should therefore change how we interpret ad bellum proportionality is nonetheless important.

International institutions have wrestled from states the capacity to limit and review the beginning of the emergency (when an armed attack occurs) and the in bello means of self-defense (IHL's rules for the conduct of hostilities). They could do so only because of legal interpretations that cast IHL and Article 51 as setting absolute limits that are justiciable. This is the first and a necessary step for a change in normative beliefs, internalization, and institution-building to potentially follow. In other words, the mechanisms by which non-coercive international law attracts compliance or develops coercive capacity can only start to operate when international law makes a demand in the first place and when it is consistently interpreted and practiced as making this demand, whatever its initial chances of compliance. The finding that ad bellum proportionality is the crucial means to forestall conflict escalation also implies a policy recommendation: whatever coercive capacity international institutions and states can collectively muster should be deployed to prevent states from violating ad bellum proportionality.

The international law of self-defense, including ad bellum proportionality, is unsettled. A lot is at stake in how this law evolves. In the above quoted Advisory Opinion, the ICJ balked at the idea that international law would prevent a state from securing its survival. Would the Court have advised differently if it had contemplated that the very retreat of international law that would allow a state to secure its own survival would threaten the survival of the international community? The use of nuclear weapons in self-defense is precisely what ad bellum proportionality – interpreted as proposed here – would prohibit, because a nuclear exchange likely escalates conflicts and threatens the international community. Given the lack of a sovereign, it may not generally be conducive to the rule of international law to ‘demand the self-abandonment’ of a subject. It is only the aim of preventing the emergence of a true international community emergency that means international law should, for its own sake, make such a demand.

Conclusion

Does international law restrict the use of force by states in self-defense even when their survival is threatened? Should it? This article has shown that neither question has a simple answer. Armed threats to states do not just resemble subject emergencies, where law puts absolute justiciable limits on a subject's self-defense which might make effective self-defense impossible. Rather in practice, international law retreats from imposing absolute, justiciable limits on the (1) aims, (2) ad bellum means, and (3) end of self-defense. I showed that interpretations that uphold these three retreats serve the function of allowing the states at war to return to the rule of international law. This evokes the logic of a community emergency according to which a temporary retreat of the rule of law can sustain it in the long run.

Only two of the three mentioned retreats, however, are justified if the normative aim is sustaining the rule of international law. Ad bellum proportionality can prevent a threat to state survival (a community emergency for the victim state) from becoming a threat to the survival of the international community (a true international community emergency). As the rule of international law is predicated on the survival of the international community, not the survival of the states at war, preventing conflict escalation is of overriding importance. If the normative aim is sustaining the rule of international law, international law should keep hold of ad bellum proportionality, while it should (as it does) retreat from limiting the aims and end of self-defense.

In short, this paper has shown that when international law regulates armed threats to states, it does (three retreats) and should (two retreats) partly resemble the ideal-type of domestic law when it regulates public emergencies, not simply the ideal-type of domestic law when it regulates individual self-defense. These findings have implications for how we think about the role of international law in IR. Which ideal-type international law resembles is indicative of the status of states in international law as either subjects, like individuals under domestic law, or sovereign legal communities in their own right. The analysis has shown that international law lacks the coercive capacity to treat states as mere subjects when they face armed threats. The traditional vision of international law which assumes that international law cannot ultimately constrain sovereign states' pursuing their security interests remains relevant for understanding the role of international law in IR today. How international law sustains the rule of law in an emergency – by keeping tight hold over its subjects or by retreating – is the litmus test for its ability to defuse anarchy. This test is negative. Over 75 years after states first agreed that international law should restrict their right of self-defense, when a state faces an emergency in the international system, international law still fails to decide on the exception.