Self-Defence against Non-state Actors: The Interaction between Self-Defence as a Primary Rule and Self-Defence as a Secondary Rule

Abstract

This article examines the law of self-defence as applied to non-state attacks in light of the coalition air strikes against ISIL in Syria. It critiques the two current interpretations of the law of self-defence – one based on attribution and the other on the ‘unable or unwilling’ test – for failing to address adequately the security threat posed by non-state actors or for not addressing convincingly the legal issues arising from the fact that the self-defence action unfolds on the territory of another state. For this reason, it proposes an alternative framework which combines the primary rule of self-defence to justify the use of defensive force against non-state actors, with the secondary rule of self-defence to excuse the incidental breach of the territorial state's sovereignty.

1. Introduction

In recent years, ISIL has emerged as the most powerful and brutal jihadist group posing ‘a global and unprecedented threat to international peace and security’. ¹ In contrast to al-Qaeda or other terrorist groups, ISIL has a territorial basis, having seized large swathes of Iraqi and Syrian territory, from where it can plan and organize its nefarious activities and attack states, including Syria and Iraq. In response to such attacks, Iraq requested external assistance ² and a US-led coalition of states launched air strikes against ISIL in both Iraq and Syria. Although the strikes against ISIL in Iraq were conducted with the consent of the Iraqi government, ³ Syria did not consent to the US-led strikes; but the majority of states involved in the air campaign against ISIL in Syria invoked their right of individual and/or collective self-defence. According to the US: ⁴

ISIL and other terrorist groups in Syria are a threat not only to Iraq, but also to many other countries, including the US and our partners in the region and beyond. States must be able to defend themselves, in accordance with the inherent right of individual and collective self-defense, as reflected in Article 51 of the Charter of the United Nations, when, as is the case here, the government of the State where the threat is located is unwilling or unable to prevent the use of its territory for such attacks. The Syrian regime has shown that it cannot and will not confront these safe-havens effectively itself. Accordingly, the US has initiated necessary and proportionate military actions in Syria in order to eliminate the ongoing ISIL threat to Iraq.

Similarly, Turkey invoked its inherent right of individual and collective self-defence since ‘the regime in Syria is neither capable of nor willing to prevent these [ISIL] threats emanating from its territory which clearly imperil the security of Turkey and safety of its nationals’. ⁵ In the same vein, Australia stated that its action is ‘in support of the collective self-defence of Iraq’ and that ‘States must be able to act in self-defence when the Government of the State where the threat is located is unwilling or unable to prevent attacks originating from its territory.’ ⁶

The UK Parliament initially authorized strikes only in Iraq in response to the request by the Iraqi government but refused to authorize strikes within Syria. ⁷ However, the Government's view was that the collective self-defence of Iraq can justify action inside Syria and that the UK can exercise its ‘inherent right of self-defence’ against specific threats emanating from Syria as when it targeted two British citizens in Syria. ⁸ In November 2015, the UK Parliament authorized strikes against ISIL in Syria ruling out at the same time any deployment of troops. ⁹

France justified its action by invoking its right to self-defence ‘in response to attacks carried out by ISIL from the territory of the Syrian Arab Republic’, whereas after the Paris attacks of 13 November 2015, it labelled its previous action as collective self-defence and its subsequent action against ISIL inside Syria as individual self-defence. ¹⁰

The international reaction to the strikes was rather muted with only a handful of critical voices. Russia condemned the strikes because, in her view, they were carried out without Security Council authorization or approval by the Syrian government. ¹¹ Yet it should be recalled that self-defence does not require Security Council authorization or host state consent. Russia was later involved in military action inside Syria apparently with the consent of the Syrian government. ¹²

Syria complained to the UN Secretary-General about the French, British and Australian strikes. Syria claimed that the self-defence justification ‘distorted . . . the intention of Article 51 of the Charter of the United Nations [and] is blatantly inconsistent with the Charter and the resolutions of the Security Council’. ¹³ Yet it did not explain why and how the acting states’ interpretation of Article 51 is distorted. The Syrian government also declared that:

[i]f any State invokes the excuse of counter-terrorism in order to be present on Syrian territory without the consent of the Syrian Government, whether on the country's land or in its airspace or territorial waters, its actions shall be considered a violation of Syrian sovereignty

and that ‘States must respect the unity, sovereignty and territorial integrity of the Syrian Arab Republic.’ ¹⁴ It transpires that Syria did not denounce the coalition action as violating Article 2(4) of the Charter prohibiting the use of force but as violating its sovereignty.

From the preceding overview of state justifications, self-defence (individual or collective) emerges as the main justification for the air strikes against ISIL in Syria. Yet this justification is not without its problems, mainly because ISIL is a non-state actor and because the strikes unfold on the territory of Syria, a sovereign state, not itself implicated in the attacks.

In this article, I will first present the two most prominent approaches to the use of defensive force against non-state attacks and analyse their reasoning. The first relies on attribution that is, the attribution of the non-state attack to a state that subsequently becomes the target of the self-defence action, whereas the second relies on the inability or unwillingness of the host state to suppress the non-state attack. In the author's view these approaches fail to address the full gamut of legal and security issues that non-state attacks give rise to. I will therefore put forward an alternative framework based on the interaction between self-defence as a primary rule and self-defence as a secondary rule. ¹⁵ To explain, whereas primary rules contain substantive rights and obligations or, to put it slightly differently, prescribe or proscribe certain conduct, secondary rules establish the conditions under which a primary rule is breached and the consequences that flow from such a breach. The law of state responsibility is, for instance, a regime of secondary rules which apply generally and uniformly to all breaches of primary rules. Self-defence as a primary rule is contained in Article 51 of the UN Charter and in customary law according to which the defensive use of force in response to an armed attack is lawful per se; there is no wrongfulness and no question of responsibility arises provided that the self-defence action adheres to the conditions set by law. Self-defence as a secondary rule is contained in Article 21 of the ILC Articles on Responsibility of States for Internationally Wrongful Acts (ASR) according to which self-defence can exonerate breaches of certain international obligations. ¹⁶ Consequently, whereas the primary rule of self-defence can justify the use of force against non-state attacks, the secondary rule of self-defence can exonerate incidental breaches of obligations owned to the state on whose territory the action takes place provided that they are committed in the course of self-defence. This framework, it is hoped, provides a more systematic conceptual and legal treatment of the use of defensive force against non-state attacks.

2. Self-defence on the basis of attribution

The first approach to self-defence against non-state attacks relies heavily on the law of state responsibility in order to identify the state responsible for the armed attack. ¹⁷ Self-defence according to this approach is an inter-state affair involving the state that suffers an armed attack and the state responsible for the attack. The ‘responsibilization’ of self-defence is performed through the device of attribution as formulated in the law of state responsibility. This approach is most evident in the ICJ's Wall in the Occupied Palestinian Territory Advisory Opinion where the Court opined that the inherent right of self-defence exists ‘in the case of armed attack by one State against another State’ but since the attacks against Israel were not ‘imputable to a foreign State’, self-defence was not relevant. ¹⁸ Conversely, if an attack is not attributed to a state, the victim state cannot use defensive force against another state or on its territory unless the use of force is authorized by the Security Council ¹⁹ or is requested by the territorial state. ²⁰

Attribution takes place on the basis of an institutional, a functional and an agency test. ²¹ The institutional test is contained in Article 4 ASR according to which an attack will be attributed to a state if it has been committed by a de jure or a de facto organ of that state. ²² Following the functional test, an attack will be attributed to a state if it has been committed by an entity that is empowered by that state to exercise governmental authority or is committed by an organ of another state that has been placed at the disposal of the first state. ²³ According to the agency test as formulated in Article 8 ASR, there needs to be an ad hoc relationship between a state and the non-state actor that commits the attack which is established when the state instructs or directs the non-state actor to attack ²⁴ or when the state exercises ‘effective control’ over the specific non-state attack. ²⁵

It becomes apparent then that the attribution criteria in the law of state responsibility require very close links between a state and a non-state actor in order to hold states responsible for non-state acts. Yet, non-state actors may collaborate with states in more subtle ways than the ones envisaged by the existing attribution tests or they may have the resources to act independently. Furthermore, non-state actors may operate from failed or failing states in which case the attribution criteria become almost redundant. This is the case for example with ISIL. Its attacks cannot be attributed to Syria or to any other state because ISIL is not a de jure or de facto organ of Syria or of any other state, it does not exercise governmental authority over parts of Syria on behalf of the Syrian Government and does not act under the instructions, direction or control of any state. Moreover, ISIL operates from areas that are not controlled by the Syrian Government.

It thus transpires that applying the attribution tests of the law of state responsibility to non-state attacks creates a void which non-state actors, either independently or in collusion with states, can exploit to attack with impunity other states, whereas victim states are left with no lawful means of defence. Such legal incapacitation may delegitimize states to the extent that defence and security are a state's primary responsibility but may also delegitimize international law because it would permit non-state actors and colluding states to infringe interests and rights protected by international law.

For this reason, attempts have been made to either ease or expand the attribution criteria whilst maintaining at the same time the state-centred reading of self-defence.

First, it has been suggested that with regard to organized groups the requisite level of state control over non-state actors who commit armed attacks should be lowered from effective to overall control. ²⁶ Overall control is about the general influence that a state may exert over an organized group, shaping its actions, but does not require proof of state involvement in specific acts as it is the case with effective control. As explained by the ICTY, a state ‘wields overall control over the group, not only by equipping and financing the group, but also by coordinating or helping in the general planning of its military activity’ and, added that ‘it is not necessary that, in addition, the State should also issue, either to the head or to members of the group, instructions for the commission of specific acts contrary to international law’. ²⁷

Secondly, state complicity in the activities of non-state actors has been promulgated as an additional attribution criterion. ²⁸ State complicity includes active but also passive support in the form of harbouring or tolerating non-state actors and their activities. The US, for example, justified its self-defence action against Afghanistan following the ‘9/11’ attacks because the attacks ‘have been made possible by the decision of the Taliban regime to allow the parts of Afghanistan it controls to be used by this organization as a base of operation’ and ‘despite every effort by the United States and the international community, the Taliban regime has refused to change its policy’. ²⁹ The Security Council endorsed this argument in Resolutions 1368 (2001) and 1373 (2001) by affirming the US’ inherent right to self-defence. In the same vein, the OAS condemned the ‘9/11’ attacks and declared that ‘those responsible for aiding, supporting, or harboring the perpetrators, organizers, and sponsors of these acts are equally complicit in these acts’. ³⁰

The immediate question is what is the legal status of the ‘overall control’ and complicity standard? It may be contended that they constitute lex specialis that is, special attribution criteria of the use of force regime. ³¹ It should be recalled that whereas the ICJ rejected in the Bosnia Genocide Case the ‘overall control’ test, it went on to say that ‘logic does not require the same test to be adopted in resolving the two issues which are very different in nature’. ³² The Court's dictum implies that there is legal space for the development of special attribution rules, yet, whether the aforementioned standards have thus been established has been challenged by commentators. ³³ That having been said, even if they were to apply to ISIL attacks, they could not be attributed to Syria or Iraq since neither state provides the required level of support.

It thus becomes apparent that the attribution approach to self-defence either in its narrow or in its more expansive formulation does not solve the security problems posed by non-state actors such as ISIL.

This is not the only flaw of this approach. Its most important flaw is conceptual because it conflates the law of state responsibility with the law on the use of force; two legal regimes with different rationales, content and exigencies. ³⁴ The use of force regime is a regime of primary rules which set out the circumstances and conditions under which force can be lawfully used in international relations. For example, Article 51 of the UN Charter establishes a legal entitlement to use force when an armed attack occurs irrespective of its author or of issues of responsibility. The law of state responsibility instead sets out the conditions and methods for holding states responsible for violations of their international obligations. Attribution in the law of state responsibility is thus the mechanism according to which non-state acts are transformed into state acts or, to put it in different terms, non-state acts are ‘subjectivised’ for purposes of responsibility. ³⁵ It is for this reason that the law of state responsibility requires compelling state input into non-state acts or non-state actors namely, in order to distinguish private from public (state) acts. ³⁶

Because of the different content and rationale of the two regimes, questions arise as to the function and propriety of such inter-systemic transfer. More specifically, questions arise as to why secondary rules of attribution should determine the content and scope of the primary rules on the use of force. With regard to the law of state responsibility, questions arise about the possible effects on the coherence of the law of state responsibility of the emergence of differentiated attribution standards. Such standards may metastasise to the law of state responsibility challenging the whole edifice of a unitary and common system of secondary rules which underpins the institution of international responsibility. If that is to happen, the law of state responsibility may gradually and steadily extend beyond its current codification but as the ICJ warned with regard to the use of the ‘overall control’ standard, it would ‘stretch too far, almost to a breaking point, the connection which must exist between the conduct of a State's organs and its international responsibility’. ³⁷ By rejecting this standard, the ICJ contained the attempted expansion of the law of state responsibility and reassured states.

3. Self-defence on the basis of the ‘unable or unwilling’ test

Ever so often in recent years, states rely on the ‘unable or unwilling’ ³⁸ test to justify the use of defensive force on the territory of a state against non-state attacks. Indeed, the US, Australia and Turkey among others relied on this test to justify their action against ISIL in Syria, whereas the UK and France relied on self-defence without mentioning the ‘unable or unwilling’ test, at least in official documents, although they alluded to that test. For example, the British Prime Minister justified the targeted killing of two British nationals in Syria under the rubric of self-defence:

because there was no alternative. In this area, there is no Government we can work with; we have no military on the ground to detain those preparing plots; and there was nothing to suggest that Reyaad Khan would ever leave Syria or desist from his desire to murder us at home, so we had no way of preventing his planned attacks on our country without taking direct action [. . .] and given the prevailing circumstances in Syria, the airstrike was the only feasible means of effectively disrupting the attacks that had been planned and directed. It was therefore necessary and proportionate for the individual self-defence of the United Kingdom. ³⁹

Interestingly, the UN Secretary-General said in relation to the strikes in Syria:

I am aware that today's strikes were not carried out at the direct request of the Syrian Government, but I note that the Government was informed beforehand. I also note that the strikes took place in areas no longer under the effective control of that Government. I think it is undeniable – and the subject of broad international consensus – that these extremist groups pose an immediate threat to international peace and security. ⁴⁰

In contrast to the attribution approach discussed in the previous section, the ‘unable or unwilling’ test moves away from attribution and recognizes non-state actors as independent authors of armed attacks – and direct targets of self-defence – even if such action takes place on the territory of the host state. Its rationale is the following: states have the primary responsibility to prevent and suppress non-state attacks from within their territory but when they are unable or unwilling to fulfil that obligation, the victim state can take self-defence action against the non-state actor.

Although this approach to self-defence addresses the security concerns of states, it is not without its problems. The first question to ask concerns the meaning of inability and unwillingness; the second question is more fundamental and concerns the nature of the ‘unable or unwilling’ test; whereas the third question concerns the available justifications for infringing the territorial state's sovereignty.

With regard to the first question, Ashley Deeks identified a number of factors that should be taken into consideration when assessing whether a state is unable or unwilling. Such factors are: the territorial state's consent or co-operation in suppressing or preventing the non-state action; the nature of the threat posed by the non-state actor; prior requests to address the threat; reasonable assessment of the territorial state's control over its territory and of its capacity to act; proposed means to suppress the threat posed by the non-state actor; and prior interactions with the territorial state. ⁴¹

These factors are not, however, without complications. First, with regard to consent, Deeks opines that ‘if the territorial state gives the victim state consent, the latter need not perform an “unwilling or unable” analysis’. ⁴² Questions may be asked as to who should grant such consent and whether it should always be granted by the government in power irrespective of its legitimacy. For example, the Syrian government invited the US to coordinate their actions against ISIL ⁴³ but, would the invitation of a government that is engaged in serious violations of international law be valid? Would cooperation with such a regime amount to aiding and abetting in the commission of crimes? What would happen if the host state grants consent to certain states but not to others? Moreover, from a legal perspective, the requirement of prior consent seems to make defensive force subsidiary to consensual intervention ⁴⁴ but self-defence and consensual intervention ⁴⁵ are independent bases for the use of force in international law.

Second, concerning requests to address the threat posed by non-state actors as one of the factors taken into consideration when assessing state inability or unwillingness, there is always a very thin line between permissible requests and unlawful intervention in that some requests may amount to coercion. ⁴⁶

Third, it is not clear how a state's capacity and effectiveness to deal with non-state actors can be assessed. Would the fact that a state deals effectively with the threat non-state actors pose but at the expense of human rights or of other international law guarantees preclude the operation of this test? How would the debacle of one state claiming that it did all that was required and another state disputing such a claim be settled?

With regard to the second question concerning the nature of the ‘unable or unwilling’ test, it is not clear whether it complements the attribution approach or, instead, whether it is the only ground for using defensive force against non-state attacks. It appears that the ‘unable or unwilling’ test is often projected as if it were the only ground for using defensive force against non-state attacks. ⁴⁷ Whether this is the case in law is very much debated ⁴⁸ but, the most important obstacle is the fact that, according to the law, self-defence as a right becomes available when an armed attack occurs and not when a state is ‘unable or unwilling’. Consequently, the ‘unable or unwilling’ test cannot determine the availability of the right to self-defence. It can perhaps condition the exercise of this right as part of the necessity calculus but this is a completely different thing. ⁴⁹ In other words, it can answer the question of whether force is the only effective option available to the victim state when faced with a non-state attack launched from another state but in this case it is not an autonomous test, nor the only consideration in the necessity calculus. Still it is not clear what is inability and whether it makes self-defence automatically necessary. In sum, the nature of this test and its place in international law or in the self-defence matrix is not clear.

Related to the above is the third question, namely, how the violation of the territorial state's sovereignty can be justified under this test? This has not received adequate consideration in the literature or in official pronouncements, but one can glean a number of assumptions. One assumption is that the violation is part of the ‘unwilling or unable’ test as an autonomous test or of the necessity calculus as explained above. In other words, the violation of the host state's sovereignty is necessary in order for the victim state to be able to exercise its right to self-defence. ⁵⁰ Although this may be correct when the territorial state is the author of the attack, when the territorial state is not the author of the attack, a different justification is needed for trespassing its territory because in that case the necessity of self-defence justifies the action against the non-state actor and not against the territorial state which is a third party in the self-defence relationship. Put another way, the ‘unable or unwilling’ test as part of the necessity calculus of self-defence can explain why the use of force against a non-state actor is required but cannot justify the violation of the territorial state's sovereignty. That necessity is different from the state of necessity in Article 25 ASR.

Secondly, if the territorial state is ‘unable’ because it has lost control over parts of its territory, as is the case with Syria, there is the assumption that no violation has occurred because the territorial state's sovereignty has receded. ⁵¹ This is perhaps what the UN Secretary-General meant when he said that the US strikes ‘took place in areas no longer under the effective control of the government’ ⁵² and what SC Res 2249 (2015) perhaps alluded to when it called upon states to take all necessary measures on the territory under ISIL control. ⁵³ The British Prime Minister also said with regard to the UK strikes against ISIL in Syria that ISIL operates from an ungoverned space and that the objective of the UK action is not to attack the Syrian regime. ⁵⁴ Likewise, Israel claimed with regard to its 2006 action in Lebanon against Hizbollah that its action was not against Lebanon ⁵⁵ since its Government had lost control of south Lebanon to Hizbollah, something that was recognized by most states and by the Security Council. ⁵⁶ The Institut de Droit International also recognized the right of self-defence against non-state actors when the attack ‘is launched from an area beyond the jurisdiction of any state’. ⁵⁷

The problem with such an assumption is that, in international law, actual or effective sovereignty is not conterminous with the legal institution of sovereignty, and thus a state's sovereignty is violated even if the action affects areas not controlled by that state.

Thirdly, when a state is unwilling, there is the implicit assumption of fault, in that the territorial state allows knowingly ‘its territory to be used for acts contrary to the rights of other states’ ⁵⁸ which can justify action on its territory. For example, Russia invoked Georgia's failure to live up to its sovereignty responsibilities in order to justify its self-defence action inside Georgia against Chechen fighters. ⁵⁹ Yet even if the duty of due diligence is a corollary to sovereignty and to non-intervention, there is no rule in international law that permits forcible intervention or self-defence action when a state breaches its duty of due diligence. ⁶⁰ Related to this is another argument that relies on the law of neutrality according to which a belligerent can take self-defence action on the territory of a neutral state if the latter allows its territory to be used by another belligerent in violation of its duties as a neutral state. ⁶¹ The immediate question is whether such a rule can be transposed to the use of force regime and, if that is possible, whether the use of force regime has recognized such a rule which is what is debated as far as this test is concerned.

Yoram Dinstein uses a different term to describe the cross-border force against non-state actors when the territorial state is ‘unable or unwilling’ to act. For him, it is ‘extraterritorial law enforcement’ in that the acting state enforces international law within the territory of the host state as a form of self-defence. ⁶²

What transpires from the preceding discussion is that the ‘unable or unwilling’ approach leaves much unexplained. That having been said, it is apparent that the ‘unable or unwilling’ approach to self-defence operates within a context of ‘responsibilization’. ⁶³ It is premised on the view that the territorial state is responsible for not preventing or suppressing non-state attacks and that self-defence is complementary to state action. To explain, not only does the unwilling state fail its primary obligation to prevent or suppress non-state attacks but is also complicit therein; it thus bears responsibility for its complicit acts or omissions although not for the actual armed attack. ⁶⁴ Complicity in this case is not an attribution criterion as it is in the attribution approach discussed previously but still establishes some form of responsibility of the territorial state. The unable state on the other hand is a state that cannot fulfil its obligations and defaults on its responsibility. Self-defence then becomes a complementary means of enforcing international law. The ‘unable or unwilling’ test is in other words a jurisdictional test of who has primary and who has secondary jurisdiction to enforce international law ⁶⁵ and in essence it is similar to Dinstein's extraterritorial law enforcement theory. As Dinstein put it, a state is ‘entitled to enforce international law extra-territorially’ if another state is ‘unable or unwilling’ to prevent an armed attack. ⁶⁶ Yet, as was said above and will be developed further in the sections that follow, self-defence as a primary right is not premised on a prior violation of international law but on an occurrence – an armed attack – and, moreover, treating self-defence as a means of enforcing international law is not only contrary to the nature of self-defence which is about defence and protection from attacks but also conflates self-defence with the law of state responsibility and in particular with the institution of countermeasures ⁶⁷ which are decentralized means of enforcing international obligations when a state is unable or unwilling to address or redress wrongfulness. ⁶⁸

4. Self-defence against non-state actors: the interaction between self-defence as a primary rule and self-defence as a secondary rule

In view of the issues raised in the preceding sections, in this section, I will put forward an alternative framework of analysis of self-defence against non-state attacks which is based on the interaction between self-defence as a primary rule and self-defence as a secondary rule. For this reason, I will first explain the scope self-defence as primary rule before explaining the scope of its operation as a secondary rule.

4.1 Self-defence as a primary rule

As was said, self-defence is recognized as a primary rule in customary law and in Article 51 of the UN Charter which recognizes self-defence as an inherent right. As a right, it empowers states to use force and such force is lawful per se; it is not a prima facie violation of the prohibition of the use of force enshrined in Article 2(4) of the UN Charter which is subsequently exonerated. ⁶⁹

According to Article 51 of the Charter, the right to self-defence is activated by an armed attack. An armed attack is defined as such not because of its author but because of its physical attributes. ⁷⁰ Article 51 for instance does not define the provenance or the author of the armed attack. Consequently, both states and non-state actors can commit an armed attack and activate a state's right to self-defence which will be directed against the attacking state or the attacking non-state actor.

The immediate question is when do states become the author of an armed attack? A state becomes the actual author of an armed attack if the attack is committed by its organs for example by its regular forces. A state may, however, use proxies to commit an armed attack. In relation to this, the ICJ relied on the General Assembly's Definition of Aggression ⁷¹ and in particular on Article 3(g) to say that state authorship of an armed attack also includes:

the sending by or on behalf of a State of armed bands, groups, irregulars or mercenaries, which carry out acts of armed force against another State of such gravity as to amount to an actual armed attack conducted by regular forces, or its substantial involvement therein. ⁷²

The first alternative of ‘sending’ refers to situations where groups are part of the state apparatus as de facto organs. ⁷³ In the same vein, the International Fact Finding Mission on the Conflict in Georgia equated ‘sending’ with de facto organs in order to determine whether Russia had committed an armed attack against Georgia by sending groups to Georgia. ⁷⁴

The second alternative, ‘on behalf’ may refer to de facto organs, as the ICJ opined in the Nicaragua case, ⁷⁵ but can also encompass cases where non-state actors are prompted, discharged, instigated, instructed or controlled by a state. ⁷⁶ This resembles the attribution standard found in Article 8 ASR, which in its previous iteration spoke of actions on behalf of a state. ⁷⁷ In the same vein, the International Fact Finding Mission on the Conflict in Georgia equated effective control over non-state actors with the term ‘on behalf’. ⁷⁸

What transpires is that there is substantive correlation with the attribution tests found in the law of state responsibility, in particular with the institutional and agency test. The rationale is however completely different. Here it is about state authorship of an attack ⁷⁹ whereas in the attribution approach as discussed previously, it is about the ‘subjectivisation’ of the non-state attack through attribution. ⁸⁰ It is for this reason that certain authors speak of primary rules of attribution in contrast to the secondary rules of attribution contained in the law of state responsibility. ⁸¹ The difference between the two approaches (attribution vs authorship) becomes even more pronounced in the case of ‘substantial involvement’, which does not correspond to any of the attribution tests promulgated in the law of state responsibility and is neither an attribution nor a derivative standard but establishes direct, albeit constructive, authorship of an armed attack. ⁸² Put another way, it is about the state's involvement in the attack by virtue of which the state becomes its author.

The critical question is what constitutes ‘substantial involvement’. International jurisprudence is not particularly helpful in this regard; it has not clarified the issue and often equates substantial involvement with attribution. In the Nicaragua case for example, the ICJ required high degree and specific – not general – involvement, effectively amounting to agency. ⁸³ The Court excluded ‘assistance . . . in the form of the provision of weapons, logistical support or other support’ from the definition. ⁸⁴ Judge Jennings and Judge Schwebel were very critical in their dissenting opinions of the Court's interpretation of ‘substantial involvement’. As Judge Jennings said:

the mere provision of arms cannot be said to amount to an armed attack . . . [but it] may nevertheless be an important element in what might be thought to amount to an armed attack where it is coupled with other kinds of involvement.

He went on to say that:

Logistical support may itself be crucial. . . . [It] covers the ‘art of moving, lodging, and supplying troops and equipment’ . . . If there is added to all this ‘other support’, it becomes difficult to understand what it is, short of direct attack by a State's own forces, that may not be done apparently without a lawful response in the form of . . . self defence. ⁸⁵

Moreover, ‘substantial involvement’ has gradually lost significance in the Court's jurisprudence. In the Armed Activities case, for example, the ICJ relied on the ‘sending’ and ‘on behalf’ criterion and completely ignored the substantial involvement standard; although it duly mentioned it as part of the definition of armed attack. ⁸⁶ Similarly, in the Wall in the Occupied Palestinian Territory Advisory Opinion the Court spoke of imputation of armed attacks and did not even use its own definition of an armed attack developed in previous cases, such as in the Nicaragua case. ⁸⁷ Likewise, the International Fact Finding Mission on the Conflict in Georgia ignored the ‘substantial involvement’ criterion.

That said, if ‘substantial involvement’ is an autonomous criterion for constructing state authorship of an armed attack, it should be given full effect: and for this reason it should not only be distinguished from ‘sending’ and ‘on behalf’, but its content needs to be defined.

It is submitted that ‘substantial involvement’ includes any aid or assistance in the form of acts or omissions given by a state to a non-state actor that substantially contributes to the commission of the armed attack, provided that the state knows that the non-state actor is willing to commit attacks and that the aid or assistance facilitates them. Such assistance may include, for example: material support, planning and preparations; selection of targets; intelligence sharing for particular attacks; provision of technical advice for specific attacks; provision of ‘safe havens’ or sanctuary; ⁸⁸ and training, but also more general support which, over time, may amount to substantial involvement. ⁸⁹

In sum, substantial involvement is about state input that contributes qualitatively and/or quantitatively in a non-incidental manner to the armed attack but, and this is the difference from ‘sending’ and ‘on behalf’, does not require direct effectuation of the attack, as the ICJ seemed to require in the Nicaragua case, ⁹⁰ neither does it require any kind of effective control over the non-state actor. ⁹¹ Substantial involvement amounting to constructive authorship does not in other words envisage the state as the dominant power behind the non-state actor but as a facilitator. ⁹²

Even if ‘substantial involvement’ is quite broad as far as means and methods are concerned and takes a macro view of state actions and/or omissions, ⁹³ it is narrowed down by the knowledge requirement. Knowledge does not include knowledge of the specific attack, because in that case it would transform knowledge into purpose, but requires knowledge of the non-state actor's willingness to commit attacks and of the contribution thereto of such assistance.

The inclusion of omissions within the scope of ‘substantial involvement’ may appear to contradict the ICJ's interpretation of aiding and assisting as requiring positive action. ⁹⁴ It has also been claimed that toleration and harbouring falls below the threshold of ‘substantial involvement’. ⁹⁵ In the opinion of the author, there should be a case-by-case but also contextual assessment of whether a specific omission amounts to substantive involvement. If, for example, a state tolerates a non-state actor who uses its territory for training or recruiting purposes, for acquiring resources or as a base of their operations and such omissions enable that non-state actor to mount an armed attack on another state, this can amount to substantial involvement provided that the state was aware of the willingness of the non-state actor to commit attacks and that its omission contributed thereto. ⁹⁶ Under different circumstances, toleration or harbouring may fall below the threshold of substantial involvement and constitute a violation of a state's general duty of due diligence or of a state's treaty or customary law obligations to prevent certain activities from its territory.

One may then ask when does an omission constitute constructive authorship and when does it constitute dereliction of a state's duty of due diligence? ⁹⁷ The difference lies in the fact that due diligence is an obligation of conduct ⁹⁸ and as such it is dependent on state capacity, whereas, for substantial involvement, capacity is irrelevant. What matters is the level of contribution the omission makes to the non-state attack. Secondly, substantial involvement translates into constructive authorship of the attack whereas according to the obligation of due diligence, a state is responsible for its own failure and not for its contribution to the acts of non-state actors. Third, whereas a state will evade responsibility if it meets its due diligence obligation even if the impugned act occurs, the occurrence of a non-state attack combined with a state's substantial involvement will make the latter the author of the attack and the target of the self-defence action. Finally, due diligence requires knowledge or constructive knowledge of wrongful activities, whereas substantial involvement requires knowledge of the willingness to commit attacks and of the contribution thereto.

The inclusion of omissions in the form of toleration and harbouring in the constructive authorship of armed attacks can also be supported by the General Assembly Declaration on the Inadmissibility of Intervention and the General Assembly Friendly Relations Declaration, which include a provision to the effect that:

every State has the duty to refrain from organizing, instigating, assisting or participating in acts of civil strife or terrorist acts in another State or acquiescing in organized activities within its territory directed towards the commission of such acts, when the acts referred to in the present paragraph involve a threat or use of force. ⁹⁹

These resolutions were relied upon by the ICJ in the Nicaragua case ¹⁰⁰ and Armed Activities case ¹⁰¹ in order to determine and indeed expand the meaning of ‘prohibited force’, whereas the Definition of Aggression was used to determine and indeed expand the meaning of ‘armed attack’. All these resolutions have a common denominator: they are concerned with the use of force and with forms of state involvement therein. ¹⁰² The Court then went on to distinguish between the different types of force on the basis of gravity with grave uses of force being categorized as armed attacks triggering self-defence action. ¹⁰³ If the use of force ‘topographies’ ¹⁰⁴ adumbrated in the aforementioned resolutions are to have any rational coherence, the types of force described in the Declaration on Non-Intervention and the Declaration on Friendly Relations can be used to interpret the ‘involvement’ criterion in Article 3(g) of the Definition of Aggression. Accordingly, state involvement in the form of organizing, instigating, assisting, participating or acquiescing in the use of force by non-state actors described in the Declaration on Friendly Relations will amount to an armed attack and to constructive authorship if the particular state involvement is substantial, the non-state actor commits an armed attack instead of a use of force and there is knowledge by the state. Otherwise there is no logical consistency in saying that a state becomes the author of a use of force when it instigates, organizes, assists, participates or acquiesces to the non-state use of force but does not become the author of an armed attack when it substantially acts in the same way vis-à-vis groups that go on to commit an armed attack.

The inclusion of such activities within the scope of self-defence is also corroborated by the African Union's Non-Aggression and Common Defence Pact, which expressly qualifies the harbouring of terrorists, as well as any provision of support for them, as an act of aggression ¹⁰⁵ that leads to common defence; ¹⁰⁶ and in Security Council Resolution 1378, which mentioned the Friendly Relations Resolution in the context of self-defence.

Equally important is Article 3(f) of the Definition of Aggression, according to which an act of aggression is committed if a state allows its territory to be used by another state to attack a third state. The essence of this provision is to transform an otherwise act of assistance into an act of aggression by the assisting state. The ICJ did not refer to this provision when defining an ‘armed attack’ but it should be noted that the ICJ did not engage in any comprehensive interpretation of the Definition of Aggression, neither does the resolution define an armed attack as such. ¹⁰⁷ In other words, it is open to interpretation and there is nothing to preclude the use of this provision to establish state authorship of an armed attack. ¹⁰⁸ Even if Article 3(f) refers to states, in view of the aims of the declaration and the fact that non-state actors mount attacks from state territories, it stands to reason to apply it by analogy to non-state actors. ¹⁰⁹

In summary, what constitutes state authorship of an armed attack is quite broad and includes actual as well as constructive authorship under the label of substantial involvement. In all these cases, the victim state can use force by way of self-defence against the perpetrator state. If, however, the assistance provided to non-state actors does not amount to substantial involvement, the assisting state does not become the author of the armed attack but it may be held responsible for violating assorted international law obligations arising from treaties, customary law, Security Council resolutions or from the duty of due diligence.

If a non-state actor commits an armed attack independently from states or when there is insubstantial state involvement, the non-state actor becomes the author of the attack and consequently the target of self-defence, according to self-defence as a primary rule. In this case, because the self-defence action is carried out on the territory of a state that is not the author of the armed attack, it violates obligations owned to that state. It is at this juncture that self-defence as a secondary rule becomes operative.

4.2 Self-defence as a secondary rule

It was James Crawford, the last Special Rapporteur on the topic of state responsibility, who made explicit the distinction between self-defence as a primary rule codified in Article 51 of the UN Charter and in customary law and self-defence as a secondary rule. ¹¹⁰ Roberto Ago, who is credited with that distinction, treated self-defence exclusively as a secondary rule. According to Ago, all violations of international law give rise to international responsibility unless they can be justified. ¹¹¹ Thus, any use of force in principle violates the primary international law obligation prohibiting the use of force unless such use of force can be justified by the existence of a justificatory circumstance as it is self-defence. As Ago opined: ‘Acting in self-defence means responding by force to forcible wrongful action carried out by another; and the only reason why such a response is not itself wrongful is that the action which provoked it was wrongful’. ¹¹²

For this reason, he also confined self-defence to an armed attack by another state, and excluded an attack by private individuals, because only states are bound by the primary rule prohibiting the use of force. ¹¹³

Following the last Rapporteur's approach to self-defence, Article 21 ASR codified self-defence as a circumstance precluding wrongfulness (CPW). ¹¹⁴ The role of CPW is to relieve states from responsibility in certain unusual circumstances in view of the fact that the law of state responsibility as codified by the ILC is not based on fault. CPW thus refer to an event or situation whose occurrence precludes the wrongfulness of the violation by the affected state of certain of its obligations, leading to non-responsibility.

Before identifying the specific obligations whose violation is exonerated by self-defence, it should be stressed that the scope of self-defence as a secondary norm is limited to the exigencies of the specific state of affairs created by self-defence. Article 21 ASR thus refers to violations of certain obligations committed in the course of self-defence which are strictly occasioned by it and are incidental to the exercise of self-defence. ¹¹⁵ As the last Special Rapporteur put it, ‘in the course of self-defence, a State may violate other obligations towards the aggressor. For example, it may trespass on its territory, interfere in its internal affairs, disrupt its trade contrary to the provisions of a commercial treaty, etc’. ¹¹⁶

If self-defence as a CPW were to apply to any violation of international law committed in the course of self-defence, then it would not only exceed its exigencies but its scope would become so broad that could potentially destabilize the international legal order. Consequently, any action by the defending state that is not related to the use of defensive force and is not strictly occasioned by it, for example, the suspension or termination of treaties or the adoption of other measures against the attacking state, needs to be justified by other rules of international law in order to be lawful. Such actions may, for example, be justified by the law of treaties and in particular by the rule on changing circumstances (rebus sic stantibus); ¹¹⁷ they may additionally be justified as countermeasures. ¹¹⁸

Which are then those international law obligations whose incidental breach in the course of self-defence can be exonerated by the application of Article 21 ASR? As the aforementioned statement by the Special Rapporteur indicates, they refer to the obligation of respect of sovereignty and the obligation of non-intervention but, does Article 21 ASR also cover the obligation not to use force? It should be recalled that in international law they form a concentric circle of obligations protecting states and the ICJ has treated them as being separate. Moreover, the ICJ has consistently held that, even if a specific conduct does not breach the norm on the non-use of force, it can still violate the non-intervention norm or the norm of respect of sovereignty. ¹¹⁹

Here, a distinction needs to be made between the state that is the author of the armed attack and the state that is not the author of the armed attack but on whose territory the self-defence action against the non-state author of the attack takes place.

With regard to the former, Article 21 ASR does not apply to the aforementioned obligations because the use of force is lawful per se according to the primary rule of self-defence ¹²⁰ which also justifies the breach of the attacking state's sovereignty and intervention. In the same vein, the ICJ, if there is a claim of self-defence, only deals with the legality of the action according to the criteria contained in the primary rule and has never enquired whether the use of defensive force is a prima facie violation of the norm prohibiting the use of force or a violation of the target state's sovereignty or of the non-intervention norm. With regard to the state that authored the armed attack, Article 21 ASR can instead cover other obligations related to and incidentally breached by the use of defensive force, for example, a treaty obligation the breach of which was incidental to the defensive action as was the Treaty of Amity between the USA and Iran, which became a cause of contention in the Oil Platforms case. ¹²¹

Article 21 ASR however acquires full meaning in cases where the self-defence action unfolds on the territory of a state that has not authored the armed attack. The self-defence action in that case gives rise to two sets of relations: the first set concerns the relation between the defending state and the attacking non-state actor, which falls under the primary rule of self-defence; whereas the second set concerns the relation between the defending state and the territorial state, where action on its territory is incidental to the self-defence action but may violate obligations owed to that state. In this case, the territorial state is a third party in the self-defence duel.

The ILC left the question of whether Article 21 ASR extends to third states open ¹²² but it is submitted here that self-defence as a secondary rule in the law of state responsibility is critical in such a situation because the primary rule of self-defence relates to the non-state author of the attack and cannot justify the trespassing of the territorial state's sovereignty as it does when the territorial state is the author of the attack. The application of Article 21 ASR to third states is also supported by a number of other considerations. One is historical and relates to the context from which this rule emerged which concerns the laws of war and the rights of third states. ¹²³ The second concerns certain modern practices, namely the use of self-defence to justify violations of obligations towards third states in the context of maritime exclusion zones ¹²⁴ or, in the context of forcible interdiction at sea. ¹²⁵

With regard to the territorial state, Article 21 ASR does not apply to the obligation not to use force because the employed force does not fall within the terms of Article 2(4) of the UN Charter. To explain, it falls below the qualifications contained in Article 2(4) and is not intended to coerce the territorial state, ¹²⁶ instead, it is limited and targeted and its aim is to defend against attacks emanating from the territory of that state. It does not also constitute unlawful intervention because it lacks a coercive element. ¹²⁷ State practice corroborates this view. States taking self-defence action against non-state actors on the territory of another state go to great lengths to confirm that the action is not against the host state (or its government) but against the non-state actor. ¹²⁸ For example, the British Prime Minister declared that the purpose of the action against ISIL in Syria ‘would not be to attack the Syrian regime’. ¹²⁹ With regard to the 2006 action in Lebanon, Israel claimed that its action was against Hizbollah and not against Lebanon ¹³⁰ and made the same claim in relation to its 1982 action against PLO in Lebanon ¹³¹ and in relation to its bombardment of the PLO headquarters in Tunisia. ¹³²

Instead, such use of force can amount to a violation of the territorial state's sovereignty. This is corroborated by state practice. For example, Syria, as mentioned at the beginning of this article, claimed that the allied action violated its sovereignty and not the non-use of force norm. Iraq condemned the 2007–2008 Turkish incursions into its territory against the PKK as a violation of its sovereignty but not as a use of force ¹³³ and used the same language with regard to previous Turkish incursions. ¹³⁴ With regard to Colombia's operation in Ecuador against FARC, Ecuador condemned the action as a ‘violation of Ecuador's territorial integrity’ and reserved its right to self-defence only against the rebels. ¹³⁵ Similarly, the OAS Foreign Ministers condemned the operation as ‘a violation of the sovereignty and territorial integrity of Ecuador and of principles of international law’. ¹³⁶ Pakistan condemned the US operations in its northern territory as violations of its territorial integrity. ¹³⁷ Finally, in relation to the 1998 US action in Sudan and Afghanistan against al-Qaeda, the Arab League condemned the action against Sudan as ‘a blatant violation of the sovereignty of a State member of the League of Arab States, and of its territorial integrity’. ¹³⁸

It is to such incidental violation of the territorial state's sovereignty that Article 21 ASR applies and exonerates the breach, provided that the self-defence action is otherwise lawful. ¹³⁹ To explain, the action needs to be a reaction to an armed attack and should satisfy the conditions of necessity and proportionality. ¹⁴⁰ The necessity condition requires, among others, that the action targets the non-state actor and that only incidentally affects assets or persons belonging to the territorial state if engaging them is necessary for the effective exercise of this right against the non-state actor. ¹⁴¹ If the self-defence action does not comply with these requirements or the state continues to use force against a non-state actor after the conditions of self-defence have elapsed, this would constitute a breach of the primary norm of self-defence whereas vis-à-vis the territorial state, it will constitute a breach of the prohibition of the use of force and of the obligation to respect sovereignty. Article 21 ASR would not apply in these cases. ¹⁴²

The immediate question is what legal consequences flow from the application of Article 21 ASR to such incidental breaches of state sovereignty? These depend on the legal meaning of ‘circumstance precluding wrongfulness’. ¹⁴³ The ILC is rather ambivalent in this regard and uses almost interchangeably the language of ‘justification’ and ‘excuse’. ¹⁴⁴ Justification alludes to a legally non-objectionable act, in other words, to a lawful act, whereas excuse alludes to an unlawful act that is excused because of certain special circumstances. Such an excuse may refer to the act itself or to the ensuing responsibility. If the excuse refers to the act, excusing its wrongfulness, it acts as justification. If the excuse refers to the responsibility of the author of the act, it means that the act is unlawful but responsibility is excused because of the intervening special circumstance. The latter approach is in line with the rationale of the law of state responsibility and its distinction between primary and secondary norms. Whereas justifications relate to substantive rules that is, to primary norms, excuses relate to secondary rules concerning the consequences arising from violations of primary rules. Self-defence as CPW is thus an excuse and as such it excuses the responsibility of the defending state for the incidental breach of obligations occasioned by the defensive use of force. Although it recognizes that the intrusion is in principle a breach of the host state's sovereignty, responsibility is mitigated because of the state of affairs created by the non-state armed attack. This approach is corroborated by state practice. Although territorial states such as Syria branded the actions as violation of their sovereignty, they did not take any measures to enforce their responsibility.

Moreover, the Security Council also seems to have exonerated states from responsibility in resolution 2249 (2015) by calling upon:

Member States that have the capacity to do so to take all necessary measures, in compliance with international law, in particular with the United Nations Charter . . .. on the territory under the control of ISIL also known as Da'esh, in Syria and Iraq.

To the extent that most states invoked self-defence to justify their actions and such actions incidentally breached Syria's sovereignty, the Security Council's approval may be interpreted as exonerating states, at least politically, from any responsibility. It is also interesting to note that the General Assembly has not criticized such actions. The reaction of these two bodies is important because they often play the role of world-wide ‘juries’ of the propriety of particular actions. ¹⁴⁵

Even if responsibility is excused by virtue of Article 21 ASR, the issue of compensation remains open. ¹⁴⁶ The host state may thus request compensation for any damage caused in the course of the self-defence action. ¹⁴⁷ This is fair and proper because otherwise the territorial state will be unnecessarily disadvantaged.

5. Conclusion

Because the attribution and the ‘unwilling or unable’ approaches to self-defence against non-state attacks suffer from practical and normative weaknesses, the author put forward an alternative framework to deal with non-state attacks based on self-defence as a primary rule that justifies the use of force against the non-state author of an armed attack; and self-defence as a secondary rule in the law of state responsibility which excuses responsibility for the incidental breach of the territorial state's sovereignty in the course of self-defence. This framework provides a conceptually coherent reading of self-defence as applied to non-state attacks.