THE ARMED ATTACK EXCEPTION TO NEUTRALITY IN INTERNATIONAL PEACE AND SECURITY LAW

Abstract

This article argues that the scope of the neutrality duties of non-assistance and prevention allows for an exception – a carve-out for assistance given to the victim State of an armed attack. Rather than weighing in on debates as to whether current State practice accepted as law suffices to establish this rule inductively, the article offers a different approach to grounding the argument for this exception in the methodology of the sources of international law, which thus far has been underexplored. The central argument of the article is that the exception or carve-out—and its contours—deductively flows from the structure of international law of peace and security and, in particular, the victim State's right to self-defence. The purpose of that right—enabling the effective termination of the armed attack—must not be undermined through prohibitions of military assistance and duties of prevention. These considerations define the scope of neutrality duties as revealed through systemic treaty interpretation. Such deductive reasoning equally determines the scope of customary neutrality duties, whether discerning that scope is framed as systemic interpretation or as identification of custom.

I. INTRODUCTION

Speaking in 1932 on the occasion of the entry into force of the 1928 Kellogg–Briand Pact, the US Secretary of State Henry Stimson declared that ‘[w]ar between nations … is no longer to be the source and subject of rights. It is no longer to be the principle around which the duties, the conduct, and the rights of nations revolve. It is an illegal thing.’¹ As Stimson implied, outlawing war was also to shift fundamentally the rights and duties of States that are not party to a war. Indeed, he stressed that:

[w]e no longer draw a circle around [States breaking the Kellogg–Briand Pact] and treat them with the punctilios of the duelist's code. Instead we denounce them as law-breakers. By that very act we have made obsolete many legal precedents and have given the legal profession the task of reexamining many of its codes and treatises.²

Among the areas most in need of ‘re-examining’ in Stimson's view was certainly the law of neutrality.³ The law of neutrality traditionally regulated the relationship between belligerent States and States not party to a war—ie third States. That body of law required that third States remain—and be kept—detached from a war. This is because war was, after all, a lawful way of settling disputes, even if nineteenth-century international law scholarship and practice may not have been as indifferent to the use of force as is sometimes assumed.⁴ The premise of war as a method of dispute settlement was challenged at its core by the emerging prohibition of the use of force. The prohibition fundamentally changed international law's stance regarding war and raised serious questions as to the position of the traditional law of neutrality within that new paradigm.

In 1941, as the United States (US) increasingly ramped up its military support to the Western Allies, the ‘re-examining’ of the law of neutrality that Stimson had called for was still in full swing. To justify the US's assistance, the Attorney General Robert Jackson famously proclaimed that ‘[n]o longer can it be argued that the civilized world must behave with rigid impartiality toward both an aggressor in violation of the treaty and the victims of unprovoked attack’.⁵ And yet it was highly controversial at the time whether such assistance could indeed be reconciled with the law of neutrality.⁶

Neutrality law has proved remarkably resilient. Today, almost a century after the Kellogg–Briand Pact, the ‘re-examining’ of how neutrality law fits with the current peace and security architecture appears yet to be concluded. When Russia launched its full-scale invasion of Ukraine in 2022, many Western States found themselves in a situation not dissimilar to that of the US prior to Germany's declaration of war in World War II, aiming to assist militarily the victim State of blatant aggression short of entering the war themselves as belligerent States. This war has once again raised the issue of squaring military support in reaction to aggression with neutrality duties of abstention and prevention.

International law scholarship has long grappled with whether such squaring is even possible. Much of the debate has turned on whether a separate status of ‘non-belligerency’ exists (in addition to or replacing that of neutrality), a status that would permit action in support of victim States of aggressive force.⁷ As little attention has been paid to how reconciling such assistance with the jus ad bellum would be grounded in the sources of international law, these debates have been deadlocked and have arguably overlooked important avenues. This article does not aim to resurrect these old debates. Instead, it makes a distinct contribution to the unresolved underlying methodological issue and puts forward an argument for how jus ad bellum considerations feed into neutrality duties from the perspective of the sources of international law.

The article makes the case that the scope of the neutrality duties of non-assistance and prevention today contains an exception in the form of a carve-out for assistance given to the victim State of an armed attack. Rather than weighing in on debates as to whether State practice accepted as law suffices to establish such a rule inductively, the central argument of this article is that the exception or carve-out—and its contours—deductively flows from the structure of international law of peace and security and, in particular, the victim State's right to self-defence. The purpose of that right—enabling the effective termination of the armed attack—must not be undermined through prohibitions of military assistance and duties of prevention. These considerations define the scope of neutrality duties as revealed through systemic treaty interpretation. Such systemic considerations equally determine the scope of customary neutrality duties, whether discerning that scope is framed as systemic interpretation or as identification of custom.

Carving out assistance to victim States of armed attack from the scope of neutrality duties avoids the practical obstacles that alternative avenues to constructing the legality of such action would face: there is no violation of neutrality law that would need to be justified as collective self-defence or whose wrongfulness would need to be precluded as a countermeasure. The solution to the neutrality problem would thus also be freed from the persisting general international law controversies of whether self-defence may justify non-forcible measures and whether collective countermeasures are permissible. Although the approach put forward here is not inductively built on State practice and opinio juris, the article demonstrates that the approach resonates better with recent practice and statements by States than these alternatives. It also helps overcome the conceptually inadequate and practically unhelpful status-framing of ‘non-belligerency’. The account put forward refines the contours of the exception or carve-out. This refined understanding, in turn, allows for delineating what remains of neutrality law, in practice, as it is harmonised with the jus ad bellum, a follow-up issue that has also rarely been examined.⁸ The article will show that sweeping suggestions that the exception renders neutrality law entirely irrelevant overlook the remaining (if residual) added value of various neutrality rules.

A clarification of terminology is warranted to avoid misunderstandings as to the legal nature of the argument. The notion that neutrality law does not prohibit assistance to the victim of an armed attack operates at the level of primary neutrality law obligations to limit their scope of application in a particular set of circumstances—a carve-out. On one account, such a primary-level carve-out is an ‘exception’, as distinct from a ‘defence’, which would presuppose a breach of the primary norm and operate as a secondary norm precluding wrongfulness (and thus responsibility).⁹ There is, however, no consistency in how international law scholarship uses the notion of ‘exception’. A different account conceives of exceptions as presupposing that the relevant rule is applicable in the first place and that the conduct at issue is inconsistent with the rule.¹⁰ On this account, the subject of this article is not an ‘exception’. Yet others use the heading ‘exception’ more broadly than either of the above approaches.¹¹ For the purposes of the argument put forward here, ‘exception’ is understood as a carve-out at the level of primary neutrality law obligations, excluding the scenario of assistance to victims of armed attack from the scope of application of these neutrality rules.

The article makes its argument as follows. Following this introduction, Section II highlights the continued existence of neutrality in current international law, despite it having been considered doomed to disappear for almost a century. Against that background, the section explains the basic challenge that neutrality law poses within today's international legal regulation of war, namely articulating its co-existence and interaction with the jus ad bellum. To address how the neutrality duty of non-assistance fits with a situation in which one party violates the jus ad bellum, Section III charts different possible legal avenues to construct the legality of such assistance with respect to the law of neutrality and discusses their conceptual and practical merits and drawbacks. Section IV then argues that an exception to the scope of neutrality duties flows from the central place of the victim's right to individual self-defence within the structure of international peace and security law. It situates the argument within the methodology of the sources of international law and shows how these structural considerations inform neutrality duties under treaty law as well as under customary international law. Section V then further specifies the legal contours of the exception, that is, its nature, scope of application, and limits. In light of these contours, the section also delimits the practical relevance remaining for neutrality law. Section VI concludes.

II. NEUTRALITY AND THE PROHIBITION OF THE USE OF FORCE: (NORM) CONFLICT, CO-EXISTENCE AND INTERACTION

Neutral States’ duties not to provide military assistance to belligerent States¹² and to prevent them from certain usages of neutral territory¹³ must be seen as stemming from neutrality's overarching aim of keeping third States detached from the conflict. As part of this aim, neutrality law, conversely, required belligerent States to respect the inviolability of neutral territory and to abstain from carrying out belligerent activities on such territory.¹⁴ Neutrality law also granted ‘belligerent rights’ permitting certain restrictive measures against neutrals, such as blockades or searches for and seizures of contraband on neutral vessels,¹⁵ as well as the use of self-help—including by force—to terminate another belligerent's breach of neutral territory should the neutral not terminate that breach.¹⁶

The rationale was to prevent the conflict from extending, as far as possible, to further States.¹⁷ Of course, as long as third States and belligerent States were relatively free to extend a war, respectively, by joining it or by attacking a third State or declaring war on it, neutrality's conflict-restraining function was naturally limited in its purchase.¹⁸ The underlying logic was that third States were to be kept out of the warring States’ ‘duel’. Indeed, war was considered a legal way of settling disputes by force and States’ right to such dispute settlement should not be hampered by third States affecting the outcome of the duel.¹⁹ Being kept out of war—while continuing trading with both sides—was also, of course, in the interest of third States themselves.²⁰

The prohibition of the use of force—established through the 1928 Kellogg–Briand Pact, Article 2(4) of the United Nations (UN) Charter²¹ and customary international law—has thus fundamentally changed the premise of international law's regulation of war, on which neutrality had been based.²² Accordingly, neutrality has repeatedly been declared anachronistic and obsolete over the past century.²³ Yet, all swansongs notwithstanding, the law of neutrality has persisted. This is reflected in references to this body of law by the International Court of Justice (ICJ),²⁴ the International Law Commission (ILC)²⁵ and the International Law Association (ILA).²⁶ Crucially, States themselves have not been willing to let go of neutrality entirely, even if it has rarely been at the forefront of international legal discourse since 1945. Many States continue to devote sections in their military manuals to neutrality law, including States that have updated their manuals in recent years.²⁷ Also, neutrality law has not been relegated to a dead letter in those handbooks but has instead been regularly referred to in major international armed conflicts.²⁸ Looking ahead, several States even explicitly consider neutrality law applicable to military operations in cyberspace in recent position papers on the subject.²⁹ As will be seen below in more detail, neutrality is best understood not as an optional regime that States choose to apply but as an automatic status for third States (ie those that are not a party) in (arguably all) international armed conflicts.³⁰

At the same time, neutrality's persistence in a radically transformed international legal order does, however, raise serious questions as to how neutrality law now co-exists and interacts with the jus ad bellum. Specifically, how can the duty not to assist either side in a war (and prevent each side from using one's territory) apply if, under the jus ad bellum, one side breaks the law by waging an aggressive war, while the other side lawfully defends itself against an armed attack?

Regarding support to the aggressor, the jus ad bellum does not stand in the way of applying the neutrality duties to refrain from assisting that party and prevent it from using neutral territory for the purposes of waging its aggressive war.³¹ In contrast, whether neutrality duties can today prohibit support to the victim of an armed attack remains controversial.³² The better view is that they cannot.

To be clear, international law does not positively require third States to take action in support of the victim of aggression or armed attack.³³ The duty to co-operate to terminate serious breaches of peremptory norms as reflected in Article 41(1) of the Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA)³⁴ can be fulfilled without necessarily violating neutrality law.³⁵ The power of the UN Security Council to impose such obligations has rarely been exercised.³⁶ Indeed, if it chooses to act it will usually merely authorise action or at most impose less specific duties that do not require deviations from neutrality law.³⁷

While third States may therefore abide by neutrality obligations vis-à-vis the aggressor, requiring them to do so appears inconsistent with the fundamental legal and normative change in the international legal regulation of war brought about by the prohibition of the use of force.³⁸ The jus ad bellum and the peace and security architecture around it aim to prevent and terminate aggressive uses of force effectively. That aim is institutionalised in the possibility of the Security Council authorising Member States to take effective action against threats and breaches of the peace under Article 39 of the UN Charter. But the aim is also reflected in a decentralised way in the permission for the affected State to exercise individual self-defence and for other States to come to its assistance in collective self-defence.³⁹ Prohibiting third States from contributing to the effective termination of an illegal use of force by assisting the attacked State in its exercise of self-defence would undermine this rationale, even if such assistance does not itself reach the threshold of force required to be justified as collective self-defence.

If one accepts this premise of a primacy of jus ad bellum effectiveness over neutrality duties, the question remains as to how that primacy is conceptualised and operationalised more specifically. It is to this question that the following sections now turn.

III. LEGAL AVENUES FOR ASSISTING VICTIM STATES OF AN ARMED ATTACK

There are different potential legal avenues for States willing to support a victim of an armed attack. This section will briefly explore these avenues and highlight their respective potential and drawbacks.

A. Security Council Authorisation

First, the Security Council, acting under Chapter VII of the UN Charter, can identify a threat to or breach of the peace or an act of aggression⁴⁰ and authorise Member States to take measures in support of the victim State of that act.⁴¹ States are then free to deviate from their neutrality obligations to the extent that this is necessary for carrying out the authorised measures.⁴² They bear obligations under Article 2(5) of the UN Charter to assist enforcement action authorised by the UN and to not assist the aggressor as the target State of such action. Under Article 103 of the UN Charter, obligations stemming from the UN Charter take precedence over obligations under any other international agreement, including States’ neutrality obligations, to the extent that the two sets of obligations conflict.⁴³ Thus, if the Security Council decides on non-forcible measures against the aggressor under Article 41 of the UN Charter, Member States’ Article 25 obligation to carry out the decision also prevails over conflicting neutrality obligations. It should be noted, however, that Article 41 measures need not (and often will not) be such as to require conduct that would be prohibited under the law of neutrality.

Centralised identification of the aggressor and authorisation of action is preferred in today's peace and security system. The obvious practical shortcoming of this centralisation is that the wider geopolitical polarisation of many inter-State conflicts frequently leads to a veto-induced deadlock of the Security Council.

B. Collective Self-Defence

Second, if the victim State so requests, third States would be entitled to invoke collective self-defence under Article 51 of the UN Charter and customary international law in support of the victim of an armed attack. Relying on collective self-defence, third States could use such force as is necessary and proportionate to repel the armed attack. Using force would, however, terminate third States’ neutral status and make them parties to the international armed conflict. While military assistance, such as providing arms, may be a violation of neutrality law, it does not amount to a use of force⁴⁴ (nor would the potential violation of neutrality law make third States parties to the conflict).⁴⁵

It could be argued that the permission to use force in collective self-defence must also, a fortiori, contain a permission to use a lesser means of support—such as delivering arms or allowing the use of one's territory for the launch of military operations by the victim State. Collective self-defence would thus, on this account, legalise such conduct short of force under the law of neutrality.⁴⁶

It should be noted, however, that this argument would be based solely on Article 51 and the corresponding right under customary international law. The circumstance precluding wrongfulness of self-defence, enshrined in Article 21 of ARSIWA, arguably cannot independently justify non-forcible measures by neutral States.⁴⁷ According to the ILC, that rule is intended to ‘justify non-performance of certain obligations other than that under Article 2, paragraph 4, of the Charter of the United Nations, provided that such non-performance is related to the breach of that provision’.⁴⁸ The rationale of Article 21 of ARSIWA is that a State using force in self-defence should not thereby be considered to violate other rules of international law that it contravenes in taking measures of self-defence.⁴⁹ Drawing on the ICJ's judicial practice, Crawford gives the examples of trespassing on the territory of the aggressor, interfering in its internal affairs, and disrupting trade contrary to the provisions of a commercial treaty.⁵⁰ To the extent that the self-defending State's use of force remains within the limits of self-defence, its action should not be rendered illegal by ‘collateral’ breaches of other obligations owed to the aggressor through lawful self-defence.⁵¹ The scenario envisaged by Article 21 of ARSIWA thus differs from that of neutral States providing military assistance. In the case of non-forcible assistance by neutral States, neutrality duties are not infringed incidentally through a lawful use of force by the assisting States—as noted above, their assistance does not constitute force; instead, neutral States’ obligations are the very rules to which self-defence, as a primary rule under Article 51 of the UN Charter, would constitute an exception.

Whether one accepts collective self-defence as a ground for deviating from neutrality law depends on the stance one takes on the controversial general question of whether self-defence—individual or collective—covers non-forcible measures. The orthodox position considers self-defence as a specific exception to the prohibition of the use of force only. On the orthodox account, the right to self-defence thus merely circumscribes that prohibition by adding ‘except in self-defence’ and has nothing to say on non-forcible measures.⁵²

Rejecting that narrow conception, Buchan has made the case that self-defence is a broader, general right of States that flows from a long-standing notion that States were permitted to deviate from rules of international law if necessary for their self-preservation.⁵³ That ‘inherent’ right was already established under customary international law prior to the prohibition of the use of force—which Article 51 of the UN Charter explicitly does not constrain.⁵⁴ But it is also not inconsistent with Article 51's wording to interpret that provision as including the unmentioned ‘minus’ of non-forcible means, nor with the structure of the Charter, which positioned Article 51 as an exception to the collective security system in Chapter VII, rather than specifically to the prohibition of the use of force.⁵⁵ This account is, on Buchan's reading, also reflected in States’ practice of relying on self-defence to justify measures below the use of force.⁵⁶ Ultimately, it would appear paradoxical and undesirably escalatory if States were incentivised to use force to defend themselves, rather than non-forcible means, because those non-forcible means were subject to different, and potentially more restrictive, conditions under general international law.⁵⁷

In practice, however, fears of escalation also cut the other way. Indeed, if States were to rely on collective self-defence to provide non-forcible assistance to the victim States of an armed attack, Article 51 prescribes that States would then have to report these measures to the Security Council immediately. This formal requirement may deter third States from invoking collective self-defence if they fear an escalation of the conflict. This deterrent effect is hardly hindered by the legal nuance that the reporting requirement is merely a separate procedural obligation and an evidential indication, rather than a condition of the lawfulness of an act of self-defence.⁵⁸ Article 51 reports would formally signal that these States consider themselves as operating within that provision's war-paradigm to restore military balance.⁵⁹ As such, Article 51 reports can be characterised as a partial functional replacement for declarations of war, which have become rare today.⁶⁰ It may, in part, be for this reason, for example, that States providing military assistance to Ukraine against Russia's armed attack have refrained from invoking Article 51 and thus avoided reporting to the Security Council.⁶¹ As will be seen below in more detail, however, that practice can also be explained by the fact that States did not believe that their conduct infringed neutrality in any way so as to require a defence.

C. Third-Party Countermeasures

It is also conceivable to categorise non-forcible measures by the victim State of an armed attack as countermeasures under the customary international law of State responsibility to preclude the wrongfulness of deviations from neutrality. Conceptually, there is a question as to whether considering such responses as law enforcement—as countermeasures imply—can account for their aim of self-preservation in the face of violence as adequately as considering them as self-defence.⁶² More importantly for present purposes, however, unlike with self-defence, the individual and collective exercise of which is placed on equal footing by Article 51 of the UN Charter, it is controversial whether countermeasures can be taken by States other than the injured State. The orthodox view has rejected the permissibility of such third-party countermeasures and would thus also foreclose this avenue for third States’ assistance,⁶³ if on grounds other than self-defence.

There is, however, a reasonable case that in today's international legal order third States must be allowed to enforce obligations erga omnes, owed to the international community as a whole, by way of collective countermeasures against the wrongdoing State, and there is considerable State practice that may be read as supporting such an account.⁶⁴ The prohibition of the use of force and particularly the prohibition of aggression appear as prime cases of obligations erga omnes.⁶⁵ Accordingly, a third legal avenue for military assistance to victim States of such illegal uses of force could be third-party countermeasures as circumstances precluding the wrongfulness of the assistance regarding neutrality duties.⁶⁶

On this account, third States’ assistance to the victim of an armed attack would have to conform to the general conditions for lawful countermeasures to preclude the wrongfulness of a deviation from neutrality obligations. The assistance would thus notably have to meet the requirements of necessity and proportionality—that is, the measures must be limited to what is necessary to induce the attacker to comply with its obligations and proportionate in light of ‘the injury suffered’.⁶⁷

In practical terms, relying on collective countermeasures would have the advantage of avoiding Article 51's duty to report the measures taken in support of the victim State. It should be kept in mind, however, that Article 52(1)(b) of ARSIWA prescribes prior notification of the responsible State accompanied by an offer to negotiate. This procedural law-enforcement requirement⁶⁸ seems structurally ill-suited to respond to armed attacks effectively.⁶⁹ Third States may have an interest, for example, in not warning the aggressor about the exact nature, quantity and timing of their assistance to the victim State. The practical concern may be somewhat mitigated by Article 52(2) of ARSIWA's exception for ‘urgent countermeasures as are necessary to preserve [the injured State's] rights’. That exception has, however, been designed for measures that need to be taken quickly to prevent the wrongdoing State from immunising itself from the measure, for example by removing its assets to pre-empt an asset freeze⁷⁰—a rationale that may apply to some but not all military assistance in the course of a prolonged war of aggression.

More importantly, however, countermeasures’ temporary and law-enforcement nature makes them inapt for military assistance to the victim of an armed attack. That nature is not only expressed in the procedural requirements discussed above but also chiefly in the requirement that countermeasures must be reversible as far as possible.⁷¹ Military assistance to a State fending off an armed attack, by definition, cannot be reversed. Such support could thus only be taken as a countermeasure if no reversible effective alternatives exist, such as asset freezes.⁷² This law-enforcement logic establishes a hierarchy or preferred order of measures of assistance that is at odds with the thrust of the victim State's right to self-defence, which would demand all effective assistance to be given.⁷³ Accordingly, States providing military assistance to Ukraine have, indeed, not framed their assistance as temporary law enforcement designed to induce Russia to resume performing its obligations but rather as enabling Ukraine's effective self-defence.⁷⁴

D. Carve-Out to the Scope of Neutrality Duties vis-à-vis Victims of Armed Attacks

Arguably, however, third States assisting a victim State of an armed attack need not have recourse to any of the above discussed legal avenues. This is because a final legal avenue for such assistance exists within the (primary) rules of neutrality law themselves. If the scope of neutral States’ duties of non-assistance and prevention contains an armed attack exception, ie a carve-out regarding assistance to the victim State of an armed attack, then such assistance would not need to be authorised by the Security Council or justified in collective self-defence, nor would its wrongfulness need to be precluded by way of a countermeasure.⁷⁵ For States, this solution has the advantage of avoiding the practical difficulties of the three other legal avenues: the risks of deadlock in the Security Council (foreclosing a Security Council authorisation); or of conflict escalation through reports to the Security Council (of measures taken in collective self-defence); as well as the irreversible character of the assistance that sits badly with the legal requirements for (collective) countermeasures. The exception to the scope of neutrality duties also avoids the doctrinal quagmires of collective self-defence for non-forcible measures and collective countermeasures. However, similar to the latter two avenues being pitted against orthodox wisdom, the armed attack exception is not in line with the traditional understanding of neutrality obligations. Nevertheless, as the following section argues, it best reflects the scope of the neutrality obligations of non-assistance and prevention and can be grounded in the methodology of the sources of international law.

IV. GROUNDING THE EXCEPTION IN INTERNATIONAL LAW

The idea of locating the legal avenue for third States’ assistance in the law of neutrality itself dates back to the inter-war period.⁷⁶ Emblematically, in its 1934 ‘Budapest Articles of Interpretation’ of the 1928 Kellogg–Briand Pact, the ILA considered that, ‘[i]n the event of a violation of the Pact by a resort to armed force or war by one signatory State against another, the other States may’, among other things, ‘[d]ecline to observe towards the State violating the Pact the duties prescribed by International Law, apart from the Pact, for a neutral in relation to a belligerent’ and ‘[s]upply the State attacked with financial or material assistance, including munitions of war’.⁷⁷ Although the Articles were criticised at the time as not representing the views of States,⁷⁸ the notion of neutral discrimination against the aggressor was borne out by State practice in the Second Italo-Abyssinian war in 1935–1936 and the Soviet–Finnish winter war in 1939–1940 and during World War II by States invoking a status of ‘non-belligerency’ to justify their deviations from neutrality obligations of non-assistance.⁷⁹

From the very outset, the debate has focused so heavily on whether the law of neutrality can treat support to the aggressor differently from support to the victim of aggression that little attention has been paid to how exactly that solution within the law of neutrality would be grounded methodologically within the sources of international law. The debates have, to a considerable extent, been framed in terms of whether or not there is sufficient practice to bear out a new status of ‘non-belligerency’.⁸⁰ Section V will show why that ‘status’ framing is unhelpful. At present, however, the key point is that the debate has rarely explicitly been grounded in the methodology of the sources of international law.⁸¹ Implicitly, for the most part, the debate has turned on whether there exists a special customary rule permitting deviations from neutrality obligations vis-à-vis the victim of aggression. Conventional wisdom has held that such a rule is not (yet) backed by sufficient practice and opinio juris to have crystallised into a rule of customary international law.⁸² Based on a thorough reassessment of State practice and opinio juris since the Kellogg–Briand Pact in 1928, Talmon has recently called that orthodoxy into question and suggested that the non-applicability of neutrality law to assistance to the victim State of an act of aggression was already established prior to Russia's war against Ukraine.⁸³

A special rule of customary international law to that effect is, however, but one possibility for establishing that neutrality duties can provide for this exception. Instead of further weighing in on the debate as to whether a limitation to the scope of neutrality duties has emerged as a specific rule of customary international law from practice accepted as law, this article shows that a separate, inductively discerned customary rule is not necessary for the notion that assistance to the victim State of an armed attack is grounded in the sources of international law. This is because a carve-out is today contained in the respective neutrality duties themselves.

Regarding the non-assistance and prevention duties as codified in Hague Conventions V and XIII, establishing the scope of these duties requires interpreting these treaty duties; and by interpreting them, the exception for assistance to victims of armed attacks can be discerned. Neither the ‘ordinary meaning to be given to the terms’ of these treaties nor their ‘object and purpose’ under Article 31(1) of the Vienna Convention of the Law of Treaties (VCLT)⁸⁴ provides a useful basis for the carve-out.⁸⁵ Subsequent State practice is of course relevant to this interpretation. Even if practice pointing towards an exception to neutrality duties may not reach the threshold of an ‘agreement’ in the sense of Article 31(3)(b) of the VCLT, instances of practice may at least be considered as supplementary means of interpretation under Article 32 of the VCLT,⁸⁶ to be relied on when the meaning of provisions would otherwise be unclear or obscure.⁸⁷ Arguably, however, unclear or obscure results can be avoided by relying on a systemic interpretation of neutrality duties. While the ‘context’ of the terms under Article 31(1) of the VCLT in the sense of a reference to other provisions of those same Hague Conventions is not helpful in that regard, a systemic interpretation in light of ‘other relevant rules of international law applicable in the relations between the parties’ under Article 31(3)(c) of the VCLT will prove key.

Indeed, the prohibition of the use of force and the right to self-defence under Articles 2(4) and 51 of the UN Charter and their customary law equivalents are key rules that embody the notion of international law as a system.⁸⁸ The system-building character of the prohibition of the use of force entails that the prohibition informs the limits of other rules. For example (as has been noted above) countermeasures must not—unlike reprisals as previously understood—involve the use of force.⁸⁹ Accordingly, the prohibition of the use of force is commonly labelled a cornerstone of the international legal order.⁹⁰

Given the UN Charter's near-universal membership and customary law status, the prohibition of the use of force and the right to self-defence are ‘rules of international law applicable in the relations between the parties’ to Hague Conventions V and XIII, as required by Article 31(3)(c) of the VCLT. These jus ad bellum rules are also ‘relevant’ in the sense of Article 31(3)(c) of the VCLT. To clarify how they are ‘relevant’ and what implications they may have on neutrality treaties by way of systemic interpretation, the purpose of the jus ad bellum as contained in the prohibition of the use of force and the right to self-defence is key.⁹¹ The purpose of the jus ad bellum is to prevent and repress uses of force. If an illegal use of force does occur, the jus ad bellum aims at allowing the effective termination of that use of force.⁹² This aim is chiefly reflected in the right to self-defence, which exceptionally allows for a decentralised decision on resorting to force to ensure that an armed attack can be warded off. Prohibiting third States from assisting others in their exercise of individual self-defence would hamper the effectiveness of such self-defence. Neutrality duties of non-assistance and prevention, if applied to a victim of an armed attack, would thus conflict with the effective exercise of the right to self-defence. To avoid that conflict, systemic considerations require reading these neutrality duties in light of the jus ad bellum and understanding their scope as containing a carve-out for this scenario.

There are several methodological challenges in applying this systemic interpretation. As will be seen, however, all of these challenges can be overcome. The first challenge is that the proposed systemic interpretation entails reading the older neutrality treaties in light of the more recent jus ad bellum rules (chiefly the right to self-defence). Such a purported evolutionary interpretation thus raises the issue of intertemporality.⁹³ While finer questions around intertemporality remain unsettled, there seems to be some common ground in that such an evolutionary interpretation is permissible to the extent that it can be grounded in the parties’ intention.⁹⁴ There is no hint in the wording of Hague Conventions V and XIII that the parties intended an evolutionary interpretation of the Conventions’ terms. Yet, in the event that a jus cogens rule subsequently emerges, the parties to a treaty must be presumed not to have intended a contradiction with that rule.⁹⁵ Indeed, as the ILC notes in its recent draft conclusions on peremptory norms, ‘[w]here it appears that there may be a conflict between a peremptory norm of general international law (jus cogens) and another rule of international law, the latter is, as far as possible, to be interpreted and applied so as to be consistent with the former’.⁹⁶

While there remain uncertainties as to the precise extent that the jus ad bellum constitutes jus cogens,⁹⁷ the ILC considers the prohibition of aggression as jus cogens.⁹⁸ As Special Rapporteur Tladi has explained, the reference to aggressive force also ‘caters for the right to use force in self-defence as part of the jus cogens norm’.⁹⁹ A conflict in the above sense may be understood ‘as the situation where two rules of international law cannot both be simultaneously applied without infringing on, or impairing, the other’.¹⁰⁰ In this broad sense, norm conflicts can also arise between permissions of certain conduct and prohibitions of that same conduct.¹⁰¹ In the present case, the duties of neutrality would thus impair the purpose of the right to self-defence if they were understood as prohibiting military assistance against an armed attack. As has been seen, that potential conflict can be avoided by understanding the latter as containing an exception for assistance to victims of an armed attack. There is thus no need for invalidating and terminating the neutrality treaty provisions—or indeed the entire treaties—as foreseen by the VCLT for conflicts with jus cogens norms,¹⁰² a legal consequence that should be reserved for conflicts that cannot be avoided by interpretation.¹⁰³ In sum, the armed attack exception to neutrality duties can be derived from an evolutionary interpretation mandated by the jus cogens character of the prohibition of aggression and the right to self-defence.

The second methodological challenge stems from the fact that the neutrality duties of non-assistance and prevention also exist in customary international law, which the 1907 Hague Conventions V and XIII sought to—partially—codify. Membership of these Conventions is far from universal. For States not parties to the Conventions, customary international law is the only basis for neutrality obligations. Methodologically, there is thus a question of whether and how the above systemic considerations can feed into establishing the content of the rules of customary international law on neutrality.

On the orthodox account, rules of customary international law cannot be interpreted because they lack written expression¹⁰⁴ and the content of a customary rule must be determined based on evidence of State practice and opinio juris.¹⁰⁵ On that account, custom also need not be interpreted because the process of custom identification suffices to discern the content of a rule.¹⁰⁶ A good case can be made, however, that interpretation—in the sense of ascribing or elucidating meaning—of customary rules is possible and, indeed, necessary.¹⁰⁷ From a jurisprudential perspective, if ‘interpretation is the elucidation of meaning’ then anything that ‘has meaning which is not trivially obvious can be interpreted’.¹⁰⁸ Indeed, the interpretation of unwritten instruments—such as oral agreements or unilateral acts—is well accepted in international law, as it is in domestic law.¹⁰⁹ Interpreting customary international law is also indispensable to establish the content of customary rules, since the identification process, based on general, ie ‘sufficiently widespread and representative, as well as consistent’ practice accepted as law¹¹⁰ would hardly be sufficiently precise to establish whether and how a rule applies to any of the infinite number of potential specific sets of facts.¹¹¹

If it is accepted in principle that customary international law can be interpreted, systemic interpretation will be just as important an interpretive method for custom as it is for treaties.¹¹² The jus cogens character of the prohibition of aggression and the right to self-defence would equally mandate that customary neutrality duties be applied consistently with these jus cogens rules. The ILC has clarified that although this interpretive rule ‘constitutes a concrete application of article 31, paragraph 3 (c), of the 1969 Vienna Convention, it does not apply only in relation to treaties but to the interpretation and application of all other rules of international law’.¹¹³ This is because ‘peremptory norms of general international law generate strong interpretative principles which will resolve all or most apparent conflicts’.¹¹⁴

Ultimately, however, it need not even be decisive whether or not the customary law duties of neutrality can be interpreted. This is because even in the process of identifying rules of customary international law and their content, deductive reasoning can play an important role. Of course, establishing the traditional two elements of custom, ie general practice and opinio juris, in principle requires inductive reasoning. This does not preclude, however, rules of customary international law being deduced from other existing rules.¹¹⁵ Indeed, the ICJ frequently resorts to deductive methods when induction yields no clear result—for example as a result of inconsistent practice—or to ‘confirm and strengthen results reached by induction’.¹¹⁶

Relying on deduction is not only pragmatic, but also doctrinally and conceptually sound. It is consistent with the requirement of States’ consent that such consent can also be expressed to a more general rule in which the deduced rule is implied.¹¹⁷ Inductive reasoning therefore does not necessarily confer greater ‘consent-based legitimacy’ to a rule than systemic reasoning.¹¹⁸ If international law—including customary international law—is considered to be, at least in some respects, a system, and not simply an entirely random collection of rules reflected in practice accepted as law, then it follows that systemic considerations may matter in establishing the content of a rule.¹¹⁹ As Crawford has explained, establishing a rule of international law ‘requires either a sufficiently general consensus on the existence of the rule … together with some agreement on key aspects of its formulation’ or that such a rule can be deduced ‘by recognized methods of reasoning from other clearly established rules’. Accordingly:

[r]ules can thus be ‘isolated’ or ‘positive’, or they can be structural or systematic, deriving part at least of their validity from the assumption that international law is a system, not merely a set of primary norms.¹²⁰

It is true that the systematicity of international law has inherent limitations, particularly given that States may deliberately accept conflicting obligations.¹²¹ However, the issue at hand is not a case of deliberately accepted conflicting obligations. By accepting the jus cogens rules prohibiting aggression and permitting self-defence, States have consented to the systemic effects of these rules. States have thus accepted that norm conflicts will be avoided or resolved in a way that gives effect to these rules.¹²² In other words, it is by virtue of States’ consent that international law can—and, indeed, must—be treated as a system as regards neutrality duties vis-à-vis the victim of an armed attack.

In sum, considering that disagreement persists as to whether the existence of a special customary rule regarding the non-applicability of neutrality duties to the victim of an armed attack can at present be induced from general practice accepted as law, it is crucial to note that such a rule still flows deductively from structural jus ad bellum rules. Even if one considers the rule as inducible, deductive reasoning would at least further buttress the induced rule. But induction and deduction could also be intertwined more closely here. Indeed, the ICJ's approach seems to be that deductive reasoning reduces the burden of proof or standard of evidence for the inductive method, such that a smaller amount of practice and opinio juris, which would otherwise be deemed inconclusive, can then suffice.¹²³ On this account, the exception to the scope of neutrality duties may be more readily induced from practice accepted as law because its deducibility from jus ad bellum rules has lowered the evidentiary standard for the requisite practice and opinio juris. The point here is not to take a view on whether the exception is inducible or not but simply to highlight the added value of deductive reasoning to that debate as a further methodological contribution, in addition to providing an alternative avenue for grounding the exception or carve-out.

In a final step, it is worth illustrating that recent practice seems to resonate with—and thus inductively confirms¹²⁴—the exception established by way of deductive reasoning. Most of the States providing military assistance to Ukraine in its self-defence against Russia's aggression have not specifically provided any justification for why their conduct does not violate the law of neutrality.¹²⁵ Assisting States do, however, assert the lawfulness of their military support to Ukraine.¹²⁶ That view is also confirmed by States that do not themselves provide such assistance.¹²⁷ Russia thus seems to be isolated in claiming that States militarily supporting Ukraine violate the law of neutrality.¹²⁸ Russia even specifically rejected the notion that potential justifications could cover these deviations from neutrality duties. It argued that collective self-defence was not reported to the Security Council, the assistance failed the proportionality test for countermeasures, and ‘so-called qualified neutrality’ to the detriment of the aggressor fails because the General Assembly lacks the competence to qualify Russia authoritatively as the aggressor.¹²⁹ Without further assessing these arguments, it is interesting to note that Russia does not seem to deny the existence of these legal avenues in principle.¹³⁰ Russia's position is, however, in contrast to that taken by many other States, albeit without legal reasoning specifically on neutrality law.¹³¹

Absent such reasoning on neutrality on the part of most States, firm conclusions are difficult to draw. Yet it is noticeable that no State reported to the Security Council that it had taken measures in (collective) self-defence. And while some of the military assistance provided to Ukraine has been publicly announced at some point,¹³² this was often done ex post facto (probably to ensure military effectiveness) and rather with a view to transparency for domestic audiences than specifically to inform Russia and enable it to cease its breach. The prior notification requirement for countermeasures as reflected in Article 52(1)(b) of ARSIWA therefore does not appear to have been widely complied with for military assistance (the case may be different for the economic sanctions imposed against Russia).

At a general level, moreover, the absence of legal justifications explicitly made regarding neutrality law may suggest that States did not believe they needed a defence regarding neutrality law. Their silence is more consistent with the notion that neutrality's non-assistance and prevention duties do not apply to the assistance to Ukraine—and therefore need not be addressed—than with a reliance on the defences of collective self-defence or third-party countermeasures.¹³³

Regarding collective self-defence, some States have even explicitly rejected the need to rely on this justification. Questioned by an opposition member of parliament why Germany had neither invoked the right to collective self-defence nor reported its weapons transfers to the Security Council, a Foreign Office State Secretary¹³⁴ replied:

The Federal Republic of Germany and its partners support Ukraine in exercising its individual right of self-defence against Russia's war of aggression, which is contrary to international law, by supplying weapons. These support measures, which are in conformity with international law, do not exceed the threshold of a collective exercise of the right of self-defence.¹³⁵

Such explicit rejections of collective self-defence as the legal basis for military assistance have remained the exception. States have, however, generally been careful to frame as the core—and virtually the only—legal justification for their assistance that Ukraine lawfully exercised its right to individual self-defence under the UN Charter.¹³⁶ Albania's statement at the Security Council that ‘Article 51 of the Charter provides the legal basis for individual States to offer whatever assistance to a country exercising its inherent right to self-defence’¹³⁷ has been identified as a potential exception to the general pattern.¹³⁸ Yet even this statement does not unequivocally suggest that States assisting Ukraine would be relying on their own right to collective self-defence, and not merely on Ukraine's right to individual self-defence.

The fact that the assistance was framed as assistance to Ukraine's lawful self-defence struggle and not as law enforcement to induce Russia's compliance with international law also sits badly with considering countermeasures as the legal basis relied upon by States. The statements by supporting States seem to suggest that the legality of Ukraine's conduct, as a victim of an armed attack (and aggression), ipso jure legalised assistance, including under the law of neutrality, without the need to have recourse to an additional legal concept authorising the assistance to ground that assistance in law, whether collective self-defence or collective countermeasures. This understanding is also hinted at in a rare official statement addressing the compatibility of assistance to Ukraine with neutrality law specifically, made by the German Federal Minister of Justice to Parliament:

What are the legal consequences of the arms deliveries to Ukraine? … [S]ince the UN Charter has been in force, the old neutrality requirement, as it is still called, has been overridden to some extent. War is in principle forbidden. The only legitimate form of war is defensive war. This is expressly stipulated in the UN Charter.¹³⁹

That line of reasoning is most consistent with and can best be explained through the lens of an exception to the neutrality duties of non-assistance and prevention for assistance to the victim of an armed attack.

To be clear, it is not argued here that these recent instances of practice suffice for an ‘agreement’ as a matter of subsequent practice under Article 31(3)(b) of the VCLT, nor that they would be sufficiently widespread, representative, and consistent as well as accepted as law to fulfil the elements of an induced rule of customary international law—indeed, the recent instances of practice and opinio juris mostly stem from a particular group of States.¹⁴⁰ While those who believed that practice and opinio juris had already established an exception to neutrality duties before Russia's full-scale invasion of Ukraine will see that rule further consolidated, those who were sceptical will hardly see this practice as the decisive and unequivocal tipping point. As the deductive approach advocated here is not built on this practice but on systemic considerations, it does not favour particular States’ positions as such. This approach avoids these serious inductive methodological difficulties and may thus also help overcome the deadlocked debates on neutrality and aggression. If nothing else, the fact that recent practice can best be explained through the lens of this approach, without being necessary to sustain it, adds to its plausibility.

Having examined how the exception to the scope of neutrality duties can be grounded in international law, the following section aims to provide some basic contours to the armed attack exception.

V. CONTOURING THE EXCEPTION

The previous section developed the rationale for the carve-out regarding assistance to a victim of an armed attack from the obligations on neutrality. The contours of that exception must also be derived from that rationale. Accordingly, the exception must be confined to what is necessary to give due effect to the jus ad bellum.

A. No Separate Status and No ‘Optional’ Neutrality

The notion that third States are permitted to discriminate against the aggressor is often associated with the existence of a separate status. As this section will show, however, this is by no means a necessary connection.

The question of whether a separate status exists ‘in between’ or simply next to¹⁴¹ neutrality and belligerency is closely linked to the question of whether neutrality is a mandatory, automatic status for third States or an optional status that States may choose to adopt.¹⁴² Historically, neutrality could be conceived of as optional in the sense that States had the alternative option of picking a side and joining the war. In practice, this amounted to an ‘opt-out’ possibility (by joining the war) for third States, which made neutrality the default status for third States.¹⁴³ Today, joining an international armed conflict as a party will, as a matter of fact, generally involve force,¹⁴⁴ which can only lawfully be used within the confines of the limited jus ad bellum exceptions. The element of choice between belligerency and neutrality is thus greatly restricted. Today, conceiving of neutrality as an optional status therefore presupposes that there is an additional status, something else that States may freely choose as an alternative to neutrality and party status—even if that third status is not explicitly articulated and may simply mean that neither neutrality law nor the rules applicable to parties to the conflict applies. Although advocates of optional neutrality tend to position their view against that of non-belligerency as a third status, both approaches share the notion that neutrality is a status adopted by choice. The optional approach seems to assume that the choice of being a neutral or not is somewhat more long term and less a matter to be decided in an ad hoc manner for a particular conflict, thus again implicitly suggesting that neutrality should mainly be reduced to permanent neutrality.¹⁴⁵

At first sight, the 1977 Additional Protocol I to the 1949 Geneva Conventions¹⁴⁶ seems to presuppose the existence of a third status, in addition to that of the parties to the conflict and neutrals, since it refers to ‘neutral or other States not Parties to the conflict’ several times.¹⁴⁷ The genesis of these provisions reveals, however, that the choice of wording was merely a drafting compromise between those advocating the use of the phrase ‘not engaged in the conflict’ as a descriptive label for third States throughout the treaty for the sake of avoiding any misunderstanding, and those—particularly permanently neutral States—insisting that the word ‘neutral’ be specifically used.¹⁴⁸

Practice in support of an exception discussed above has rarely taken the form of claims of a specific ‘status’ as a justification for deviating from neutrality obligations.¹⁴⁹ Some domestic judicial pronouncements have even explicitly rejected this notion.¹⁵⁰ Recent statements by States considering that neutrality law ‘applies’ to international armed conflicts in cyberspace (as distinct from the notion that it merely ‘can be applied’ when a third State so chooses) subtly hint towards neutrality as an automatic, rather than an optional, status.¹⁵¹ There is also some domestic case law that supports this automatic conception of neutral status.¹⁵²

Conceptually, a separate status is not warranted. It has been seen that the exception is justified because—and to the extent that—it gives effect to the victim State's right to self-defence under the jus ad bellum. Under this rationale, third States only have limited freedom to deviate from their neutrality obligations, namely only in so far as is necessary to fulfil that aim. In that respect, third States, of course, have an actual choice, since they are neither prohibited from providing assistance to the victim of an armed attack (given the exception to neutrality obligations’ scope for this setting), nor are they under a positive obligation to provide such assistance.¹⁵³ Making neutrality an entirely optional status would, however, go beyond that: an optional status seems to suggest that third States could free themselves of their neutrality obligations in all respects, regardless of whether assistance to a victim of an armed attack is at stake or not.¹⁵⁴

Beyond what is required to give effect to the purpose of the jus ad bellum, there is no reason not to give effect to the wider, conflict-constraining function of the law of neutrality,¹⁵⁵ which would be undermined if neutrality were an entirely optional status. As a matter of legal policy, Schmid argues to the contrary that optional neutrality makes those States that opt for neutrality more ‘predictable and credible interlocutors who can facilitate negotiations for humanitarian access’.¹⁵⁶ This argument mainly seems to have the specific situation of permanently neutral States in mind, to which the present article cannot fully do justice. It should be noted, however, that the argument made here does not require any State, and a fortiori not permanently neutral States, to give up neutrality against their will.¹⁵⁷ Whether a neutral State is perceived as a credible and reliable interlocutor probably depends less on the mere possibility to make use of a carve-out from neutrality duties than on the mechanisms by which it cements its neutrality (into permanent neutrality). It is arguably these mechanisms and a State's neutrality policy that determine the perception of stability.¹⁵⁸ These elements are not inherent to the notion of optional neutrality as such, which reveals little about how easily a choice can be undone.

If the status of ‘non-belligerency’ or ‘qualified neutrality’ is to be confined to the above limits of giving effect to the victim's self-defence,¹⁵⁹ there is no point in conceiving of the legal position of third States as a separate ‘status’. What is more, doing so risks obscuring that the modified duties of third States flow from—and are thus circumscribed by—a specific interaction between the jus ad bellum and the law of neutrality, and instead suggests that rights and duties flow from a separate ‘status’. Ultimately, it would also seem strange that the jus ad bellum, of all things, should give rise to a separate ‘status’ regarding war, given that this body of law applies to individual acts and not to status-type categories.

B. Scope of Application and Limits of the Exception

Being tied to the victim State's right to self-defence, the exception only applies to assistance to a State that has become the victim of an armed attack. Accordingly, the exception does not apply to assistance provided to a State suffering uses of force below the threshold of an armed attack.

There is no need for an authoritative or collective identification of the aggressor for the exception to apply. Requiring such an identification by the Security Council would effectively undermine the very point of the exception to ensure the victim State's effective self-defence by way of decentralised measures when the Security Council is unable to take centralised action. There are, of course, undeniable institutional risks inherent in permitting individual third States to identify who the aggressor is in a particular international armed conflict.¹⁶⁰ Such identifications will often be contested, as frequently both sides claim the lawfulness of their actions under the jus ad bellum, leaving the international community, in many conflicts, divided on this matter. While not formally required for assistance to fall within neutrality's exception, UN General Assembly designations of the aggressor (pursuant to the Uniting for Peace Resolution in the case of a deadlock in the Security Council¹⁶¹) as in the case of Russia's aggression against Ukraine may therefore lend crucial political legitimacy to military support measures taken by individual States.¹⁶²

The limitations of the exception, therefore, only stem from the victim State's right to self-defence itself, to which the exception is coupled. That is, the assistance only falls within the exception if the assisted State's action remains within the limits of necessity and proportionality in defending itself. It could thus be said that the assistance itself must be necessary and proportionate to fend off the armed attack. Accordingly, assistance to a warring State exceeding the limits of its individual self-defence—ie third-State assistance to an unlawful use of force—may not only violate the prohibition of the use of force in and of itself, or at least make the assisting State complicit in the use of force¹⁶³ but will also violate the law of neutrality. Structurally, it would be incoherent to derive a carve-out to the law of neutrality from the jus ad bellum that is more permissive than that latter body of law itself. Just as the exception itself, its limits also flow from a systemic understanding of the law of neutrality in light of the jus ad bellum. This coupling of the neutrality duties of non-assistance and prevention to legality assessments under the jus ad bellum thus reveals a qualitative change of the structure of these key neutrality rules, which blends complicity-type structures into neutrality law.

C. What Remains of the Law of Neutrality?

To refine the contours of the exception outlined in this article, it is useful to set it against what remains of the neutrality duties if the exception is applied. Does the exception ultimately swallow the rule¹⁶⁴ and thus render the law of neutrality practically irrelevant?¹⁶⁵ This question can be answered in three steps. First, it needs to be clarified whether the exception applies to all conflicts to which neutrality applies or whether there are conflicts in which it is of no relevance. Second, when the exception applies, does it render neutrality duties practically irrelevant? And third, are there still other elements of neutrality law that are practically relevant, besides neutrality duties?

In practice, the exception will apply to most conflicts to which the law of neutrality applies. Of course, technically speaking, an international armed conflict to which neutrality law applies does not as such presuppose the existence of an armed attack (which triggers the exception outlined in this article). In practice, however, it is difficult to conceive of conflicts to which neutrality law applies that do not involve an armed attack. There is debate as to whether neutrality law applies to all international armed conflicts¹⁶⁶—understood as resort to armed force between States¹⁶⁷—or only to conflicts of a certain duration and intensity.¹⁶⁸ If the latter view is followed, the intensity of the hostilities required for applying neutrality law would probably involve acts amounting to an armed attack (based on the predominant view that an armed attack under Article 51 of the UN Charter requires acts of greater gravity than other uses of force under Article 2(4) of the UN Charter).¹⁶⁹ If the former view is followed, there may be room for applying neutrality in low-intensity international armed conflicts in which no armed attack occurs, though that room appears rather theoretical. At any rate, conflicts in which third-State assistance and neutrality law play a role tend to be prolonged conflicts of heightened intensity. Those conflicts in practice involve armed attacks, such that the exception outlined in this article will apply.

When the exception applies, it certainly takes away core aspects of the traditional scope of neutrality duties. It does not, however, render these duties entirely irrelevant. As noted above, the exception has certain limits. It covers only assistance to necessary and proportionate measures of self-defence. Accordingly, there remains scope for neutrality duties vis-à-vis the victim State of an armed attack, as soon as that State oversteps these limits. Regarding assistance to the aggressor, neutrality duties still apply. Assisting the aggressor is, of course, also prohibited by rules prohibiting complicity in internationally wrongful acts. Under Article 3(f) of the Definition of Aggression, certain acts of assistance to an aggressor qualify as aggression, namely placing one's territory at the disposal of another State and allowing it to be used by that other State for perpetrating aggression.¹⁷⁰ That rule is reflected in customary international law, and it has been argued that the same is true of other acts of assistance in aggression, such as supplying weapons.¹⁷¹ At any rate, such acts of assistance may qualify as complicity under the rule of customary international law reflected in Article 16 of ARSIWA. If certain conduct by neutral States qualifies as (complicity in) aggression, either under the specific rule of Article 3(f) or under the general complicity rule, the fact that the conduct also violates neutrality obligations no longer has added value.¹⁷²

Crucially, however, these complicity rules have certain requirements attached to them that neutrality rules do not have. For one, complicity rules require a certain nexus of the assistance to the wrongful act, that is, the assistance must ‘contribute significantly’¹⁷³ to the commission of the wrongful act or materially facilitate it.¹⁷⁴ No such requirement exists to establish a violation of neutrality duties. Perhaps even more importantly, complicity requires a subjective element. Article 3(f)'s wording (‘allowing’, ‘placed’) should be understood as requiring at least knowledge of the circumstances constituting the aggression.¹⁷⁵ While Article 16 of ARSIWA requires ‘knowledge of the circumstances of the internationally wrongful act’, the ILC commentary suggests a standard of wrongful intent.¹⁷⁶ Whatever the exact applicable subjective standard(s),¹⁷⁷ the subjective requirement clearly sets complicity rules apart from neutrality duties and thus leaves scope for a persisting practical relevance of neutrality duties. Accordingly, an assisting State that errs about the fact that the assisted State uses the assistance for committing acts of aggression would not be complicit in that aggression but would still violate its neutrality duties of non-assistance.

The situation is similar regarding the neutrality duty to prevent parties to the conflict from using neutral territory for certain belligerent activities. Here, a similar function could be taken over by the positive duty under general international law ‘not to allow knowingly its territory to be used for acts contrary to the rights of other States’.¹⁷⁸ This ‘no-harm rule’ has in practice, however, only been used regarding acts of non-State actors that cannot be attributed to the territorial State;¹⁷⁹ its application to acts conducted by other States on the territory of the State in question is conceivable but has yet to be tested in practice. In any event, however, despite being similar, the ‘no-harm rule’ is not identical to the neutrality duty of prevention.¹⁸⁰ For example, the ‘no-harm rule’ requires knowledge that one's territory is used for a violation of the rights of others (thus in the present scenario for committing aggression) while the neutrality duty of prevention at most requires knowledge that one's territory is used by a belligerent.¹⁸¹ Such knowledge should be easier to establish since knowledge of whether the territory is used for lawful or unlawful acts is not required.

Moving beyond the duties of neutrals, other elements of the law of neutrality may still hold relevance. Here, too, the picture is nuanced. Certain other elements of neutrality law are equally circumscribed by the jus ad bellum. This is notably the case of traditional ‘belligerent rights’, including the right to impose blockades or search for and seize contraband. These rights can no longer grant permissions to use force that would extend beyond what the jus ad bellum would permit.¹⁸² Less problematic is the protection of neutral States’ territory under the law of neutrality.¹⁸³ Of course, every State's territorial integrity is protected under general international law. That protection is less precarious than the protection of neutral territory under the law of neutrality,¹⁸⁴ which was subject to belligerent reprisals (that is, belligerents could terminate each other's breaches of neutral territory by entering neutral territory themselves and use force if the neutral State proved unwilling or unable to prevent or terminate those breaches).¹⁸⁵

Yet, the protection of neutral territory also came with certain prerogatives for neutral States that may still be relevant in that they are more specific than the general protection of territorial integrity and more far-reaching than territorial States’ prerogatives under other rules of international law. For example, neutral States may restrict or even totally prohibit passage through their territorial waters by the parties to an international armed conflict.¹⁸⁶ This prerogative goes beyond coastal States’ prerogative to restrict innocent passage (which in general also applies to warships), which is subject to strict conditions under Article 25 of the UN Convention on the Law of the Sea.¹⁸⁷ Going further, neutral protection may also be useful as it is not confined to territory.¹⁸⁸ In particular, it may also play a role in cyberspace. It remains controversial whether sovereignty is a self-standing rule of general international law that protects against cyber intrusions that would not meet the coercion criterion required to be covered by the prohibition of intervention.¹⁸⁹ Given this persistent uncertainty, the protection of neutral infrastructure from interference by the parties to an international armed conflict—including by way of cyber-attacks—has the potential to add legal certainty to the qualification of a cyber-attack as unlawful, if only within the scope of application of neutrality law.¹⁹⁰

In sum, it would be an overstatement to say that the law of neutrality has been rendered entirely irrelevant by the exception for assistance against armed attacks, even if this exception is considered together with other rules of international law that have taken over certain functions of the law of neutrality today. Given the different scope and requirements of these rules, neutrality law retains potential practical uses. It is certainly true, however, that the interplay with the jus ad bellum has rendered the law of neutrality very much a residual regime. The practical relevance of its rules has to be examined carefully in each specific case and the extent to which it retains relevance is contingent on the extent to which other rules leave space. As has been seen, however, they do appear to leave some such space. In addition, where the scope of other rules is unsettled, the application of neutrality law may also add legal certainty. Overall, neutrality can thus be envisioned as a deeper layer of soil that has, for the most part, been covered by other layers over time, but which does occasionally come to the surface.

VI. CONCLUSION

The persistence of neutrality law presents international law with a puzzle. Neutrality law has survived profound systemic changes of the international legal order brought about by the prohibition of the use of force. Yet it can only play a role in today's legal regulation of war if it is harmonised with the jus ad bellum. This article has shown that international law allows for a methodologically sound harmonisation by way of systemic reasoning. It flows from the structure and purpose of peace and security law that the scope of the neutrality duties of non-assistance and prevention contains a carve-out regarding military assistance to the victim State of an armed attack. This exception is justified and limited by the attacked State's right to individual self-defence. By thus integrating the law of neutrality within the current regulation of war, international law balances States’ desire to maintain that old body of law with the need for the effectiveness of the jus ad bellum around which those same States have reorganised international law.