1. Introduction
The right to individual or collective self-defence in Article 51 of the Charter of the United Nations (UN Charter)Footnote 1 continues to generate extensive debate amongst scholars.Footnote 2 One of the bones of contention relates to the increasing invocation of the right to self-defence in response to armed attacks conducted by armed groups located in the territory of another state, regardless of the extent to which such groups have received assistance from the territorial (or another) state.Footnote 3 Invoking the right to self-defence in these situations seems at odds with the principles of attribution applicable to the jus ad bellum, as endorsed by the International Court of Justice (ICJ) since its Nicaragua decision of 1986.Footnote 4 In particular, the manner in which states have invoked the right to self-defence in recent years seems to challenge the high threshold that the ICJ has set for attribution of conduct of non-state actors to states within the self-defence paradigm. In addition, it seems to challenge the ICJ’s position that armed attacks can only be executed by states.Footnote 5
It is worth recalling that Article 51 of the UN Charter merely states that the right to self-defence is triggered by an ‘armed attack’ against a member state of the UN. It does not explicate who the author of such an armed attack is. Yet, states are regarded as the traditional authors of armed attacks through their land, air, and maritime forces. This is illustrated by Article 3 of the 1974 UN General Assembly’s Definition of Aggression,Footnote 6 which informed the ICJ’s understanding of ‘armed attack’ in its Nicaragua decision, as well as in that of Democratic Republic of Congo (DRC) v. Uganda.Footnote 7 Article 3(g) of the Definition of Aggression also acknowledges indirect acts of aggression (that is, indirect armed attacks) by states acting through non-state actors such as armed groups. It determines that an indirect act of aggression would exist in case of:
[t]he 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 the acts listed above, or its substantial involvement therein.Footnote 8
According to the ICJ in the Nicaragua decision, the relationship (nexus) between the state and the non-state actor had to be very close for the state to be regarded as the author of an armed attack.Footnote 9 Such a group either had to be sent and controlled by the state itself (as a de facto state organ), or had to act under its direct instruction or control.Footnote 10 Furthermore, the understanding of the majority of the ICJ of ‘substantial involvement’ in an armed attack also implied a high threshold for imputing such an attack to a state.Footnote 11 State assistance to armed bands in the form of weapons, funds, training or logistical support, while violating the prohibition of the use of force, would not amount to an armed attack that triggered the right of individual or collective self-defence.Footnote 12 In fact, this understanding raises the question as to whether any form of assistance, falling short of direct instruction, could amount to the type of ‘substantial involvement’ that crossed the threshold of an armed attack. After all, if the support of non-state actors by means of weapons, training, and/or intelligence did not amount to substantial involvement, it is difficult to see how the term ‘substantial involvement’ could in practice be anything but synonymous to direct instruction.Footnote 13
While the restrictive interpretation of the ICJ majority was in line with the drafting history of the Definition of Aggression, dissenting judges and scholars have criticized it for not sufficiently taking into account the different ways in which states can incite forcible measures (amounting to armed attacks) against other states.Footnote 14 These can vary from the active provision of (a combination of) arms, training, financial, and logistical support to armed bands perpetrating attacks against third states, to passive support in the form of safe havens or tolerating such groups.Footnote 15 The result is a responsibility vacuum in situations where a state that shares a common purpose with a non-state actor provides the latter with the opportunity to carry out an armed attack, without incurring responsibility for the attack as such.Footnote 16
The subsequent analysis will explore whether the invocation of the right to self-defence by states in recent times constitutes a response to this criticism, in the sense that it lowered the threshold requirements for indirect armed attacks. It will also examine whether armed attacks at times are directly attributable to non-state actors. In so doing, it first provides an overview of situations in which states have invoked the right to self-defence (Section 2), followed by one concerning the reactions of other states to such invocations (Section 3). While non-exhaustive, this account of the (reactions to) invocations of self-defence across regions since the 1980s provides the necessary factual (empirical) basis for assessing the threshold requirements for indirect armed attacks (Section 4.1), as well as for examining whether armed attacks can be directly attributable to armed groups (Section 4.2). The analysis of the threshold requirements for indirect armed attacks focuses on the meaning of ‘substantial involvement’ in an armed attack, the ‘harbouring’ doctrine, the ‘unwilling’ prong of the ‘unwilling or unable’ doctrine, as well as the potential inter-relationship of these terms. The assessment of the attribution of armed attacks directly to non-state actors includes an examination of the implications of the ‘unable’ prong of the ‘unwilling or unable’ doctrine for the invocation of Article 51 of the UN Charter. This analysis places particular emphasis on the meaning of the necessity requirement within the self-defence paradigm and supports an interpretation of this requirement in accordance with Article 25(1)(a) of the Articles on Responsibility of States for Internationally Wrongful Acts (ASR).Footnote 17
2. The invocation of the right to self-defence in state practice
In 1981 Israel invoked the ‘unwilling or unable’ doctrine in justifying the use of force in Lebanon against Hezbollah and the Palestine Liberation Organization.Footnote 18 During a debate in the United Nations Security Council (UNSC), Israel asserted that ‘under international law, if a state is unwilling or unable to prevent the use of its territory to attack another state, the latter state is entitled to take all necessary measures in its own defence’. In October 2003, Israel attacked alleged Palestinian terrorist targets in Syria on the basis of self-defence.Footnote 19 In July 2006, it also retaliated in self-defence after Hezbollah had kidnapped two Israeli soldiers and continued indiscriminate firing of rockets into Northern Israel.Footnote 20 At the time, Lebanon did not have effective control over its territory (notably Southern Lebanon).Footnote 21 Israel inter alia carried out air strikes against targets in Lebanon, including Beirut airport, followed by extensive military operations in Lebanon that lasted a month.Footnote 22
In 1985 President Reagan described the provision of ‘material support for terrorist groups which attack US citizens’ as ‘armed aggression against the United States under established principles of international law, just as if it had used its own armed forces’.Footnote 23 The ‘harbouring’ and ‘unwilling or unable’ doctrines gained momentum subsequent to the 9/11 attacks in the United States. While these attacks were instructed by Al Qaida, the United States regarded the Taliban regime in Afghanistan as co-responsible for allowing Al-Qaida to use Afghan territory as its operational base.Footnote 24 Relying explicitly on the right of individual and collective self-defence, the United States informed the UNSC on 7 October 2001 that it had initiated military measures in response to these attacks.Footnote 25 Shortly afterwards, in 2002, Russia used force in self-defence in Georgia against Chechen rebels who, operating from Georgia, had conducted violent attacks in Russia. According to Russia, Georgia was ‘unwilling or unable’ to suppress the activities of the rebel groups.Footnote 26 In the wake of the Beslan school siege in 2004, Russia described the incident as an attack on the country.Footnote 27 It held Chechen rebels backed by Muslim militants responsible and vowed to strike terrorist bases ‘anywhere in the world’.Footnote 28
In the following year, the African Union (AU) adopted the AU Non-Aggression and Common Defence Pact of 31 January 2005, in which it defined aggression as ‘the encouragement, support, harbouring or provision of any assistance for the commission of terrorist acts and other violent trans-nationally organized crimes against a member state’.Footnote 29 This broad definition (which is in force in respect of the 22 signatories)Footnote 30 suggests that any involvement in the use of force by non-state actors as described in the Friendly Relations DeclarationFootnote 31 will amount to aggression, that is, an armed attack.Footnote 32 A similarly progressive understanding of the type of assistance that amounts to an armed attack can be witnessed in the Gulf region, in relation to the current conflict in Yemen.Footnote 33 In his letter to the Gulf Cooperation Council (GCC) of 24 March 2015, President Hadi of Yemen invoked the collective self-defence provisions enshrined in Article 51 of the UN Charter, in the Arab League Charter,Footnote 34 and in the Arab Treaty of Joint Defence.Footnote 35 Referring to aggression by the Houthis, who were also supported by ‘regional powers’, he urged the six GCC countries to provide immediate support and take the necessary measures, including military intervention, to protect Yemen and its people.Footnote 36 While President Hadi’s official request for military assistance in terms of Article 51 of the UN Charter did not specify the regional powers supporting the Houthi rebels, other statements by himself and Yemeni officials at the time pointed to Iran.Footnote 37 With the exception of Oman, the members of the GCC engaged in military operations under the leadership of Saudi-Arabia.Footnote 38 Saudi-Arabia also subsequently identified Iran as the country arming the Houthi rebels. In addition, it submitted that the Houthis were engaging in cross-border attacks against Saudi-Arabian territory.Footnote 39 Thus far, the military coalition has limited its operations to the territory of Yemen and has not engaged in air strikes against Iran.
Turkey, for its part, has frequently invoked the right to self-defence since the 1990s to justify its incursions into Northern Iraq and the use of force against the Kurdish Workers’ Party (PKK), despite strong protest from the Iraqi government.Footnote 40 Turkey emphasized that the PKK operated from areas not under the control of the Iraqi government.Footnote 41 Turkish operations inter alia included attacks on weapons stock piles of militants associated with the PKK in 2008, as well as incursions against the PKK subsequent to the outbreak of the Syrian conflict.Footnote 42 However, Turkey did not explicitly base these claims on Article 51 of the UN Charter, nor did it consistently report its operations to the UNSC as required by this Article.Footnote 43 Iran, for its part, also engaged in cross-border operations in Iraq against Kurdish fighters in the 1990s. Iran did report these incursions to the UNSC, basing them on Article 51 of the UN Charter. It accused Iraq of supporting terrorist groups through military training and financial, logistical and intelligence support, as well as providing them with a sanctuary from where they prepared their attacks.Footnote 44 On one occasion, Iran also claimed that it had been forced to take defensive measures, as Iraq did not have control over its territory.Footnote 45
The United States-led air strikes against Islamic State (IS) in Syria commenced in September 2014. The letter by the United States to the UNSC underscored the fact that Iraq had requested the United States to lead an international coalition against attacks perpetrated by IS from Syrian territory.Footnote 46 The United States further emphasized the right to individual and collective self-defence, as IS posed a threat not only to Iraq, but also to other states, while the Syrian government was ‘unwilling or unable’ to prevent the use of its territory for such attacks.Footnote 47 It is worth noting that, unlike Russia and Iran, which are militarily supporting the Syrian regime with its consent,Footnote 48 the United States explicitly did not request Syria’s consent to engage in air strikes against IS,Footnote 49 although it gave the Syrian regime advance notice of its impending military operation against IS.Footnote 50
Several other (also non-Western) states, have since joined the United States-led military coalition against IS in Iraq and Syria.Footnote 51 These states tend to rely on a mixture of individual and collective self-defence in justifying their respective military operations, sometimes also invoking the ‘unwilling or unable’ doctrine. For example, Turkey emphasized the threat posed by IS to itself, while invoking the right to individual and collective self-defence. It also emphasized that Syria was ‘neither capable nor willing’ to prevent attacks by IS.Footnote 52 Similar reactions were forthcoming from Australia and Canada. Footnote 53 In its 2018 cross-border military operations into the Afrin region in Syria, Turkey invoked the right to self-defence under Article 51 of the UN Charter (without specifying whether in its individual or collective manifestation). It described Turkey as being under direct threat from PKK-affiliate groups in Syria (Yekîneyên Parastina Gel (YPG/PYD)), as well as Daesh (IS). On this occasion, Turkey did not explicitly invoke the ‘unwilling or unable’ doctrine.Footnote 54 Instead, it stated that the Afrin region was under control of the YPG/PYD from where rockets were fired at Turkish provinces.Footnote 55
Even before the Paris attacks of November 2015, France referred to the direct threat which IS also posed to itself, invoking Article 51 of the UN Charter.Footnote 56 The United Kingdom in 2014 initially emphasized the right to collective self-defence, as IS was engaged in an ongoing armed attack against Iraq.Footnote 57 Since September 2015, the United Kingdom has also been invoking the right to individual self-defence in accordance with Article 51 of the UN Charter, in light of planned attacks against the United Kingdom.Footnote 58 However, France and the United Kingdom have not (yet) referred to the ‘unwilling or unable’ doctrine in their official invocation of Article 51 of the UN Charter.Footnote 59 Other states invoking Article 51 of the UN Charter against IS for perpetrating attacks against third states from Syrian territory, subsequent to the Paris attacks, included Germany, Denmark, Norway, and Belgium.Footnote 60 In these instances there also was no explicit reference to the ‘unwilling or unable’ doctrine. Germany and Belgium instead underscored that IS occupied a certain part of Syrian territory over which the Syrian government did not at the time exercise effective control.Footnote 61
3. Reactions of other states
The reaction of other states in instances where the right to self-defence was invoked is an important consideration in assessing whether the scope of this right has evolved under Article 51 of the UN Charter and under its customary counterpart. Such reaction constitutes subsequent practice which is relevant for treaty interpretation in accordance with Article 31(3)(b) of the Vienna Convention on the Law of Treaties of 1969.Footnote 62 It can contribute to clarifying the meaning of a treaty, including the widening of its interpretation.Footnote 63 The weight of any subsequent practice will be influenced by whether and how it is repeated.Footnote 64 While no particular frequency is required, the subsequent practice must reflect a common understanding of the parties regarding the (evolution of the) meaning of the treaty terms.Footnote 65 Reactions of other states can further amount to opinio juris, i.e., conduct ‘undertaken with a sense of legal right or obligation’, for the purpose of identifying customary international law.Footnote 66 Both in terms of subsequent practice and opinio juris, inaction by third states (in addition to oral and written statements) can be an indication that they accept certain conduct as law. Whether such toleration by (a) state(s) amounts to acceptance of or acquiescence in the legality of the conduct in question will depend on the circumstances of the case. In relation to subsequent practice one would have to determine whether reaction by other states was called for under what is likely to be very facts-specific circumstances.Footnote 67 When determining opinio juris, the relevance of inaction would inter alia hinge on whether the silent state was deemed to have had knowledge of the conduct at the time, as well as whether its interests were affected.Footnote 68
When considering the reactions of third states to the above-mentioned invocations of the right to individual or collective self-defence, it is difficult to identify an unambiguous pattern of explicit endorsement or acquiescence. In fact, the military response of the United States subsequent to the 9/11 attacks seems to be the only instance in which the invocation of the harbouring doctrine met with virtual unanimous support. At the time, it received explicit support from inter alia the United Kingdom,Footnote 69 the European Union (EU),Footnote 70 members of the North-Atlantic Treaty Organization (NATO),Footnote 71 and the Organization of American States (OAS).Footnote 72 However, this response occurred against the backdrop of UNSC Resolution 1368 of 12 September 2001 and UNSC Resolution 1373 of 28 September 2001. Both resolutions were adopted under Chapter VII of the UN Charter and referred to the right to individual and collective self-defence in their Preambles. This may have convinced some states that the UNSC endorsed the military measures against Al-Qaida and the Taliban under Chapter VII of the UN Charter.Footnote 73
No similar resolutions were adopted in relation to the other incidents of self-defence referred to above, and the reaction of third states is not as clear-cut. For example, explicit condemnation of Russia’s invocation of Article 51 of the UN Charter in relation to Georgia in 2002 came from the United States and the Council of Europe’s Parliamentary Assembly.Footnote 74 Yet, reactions of other states remained muted. The broad definition of aggression in the AU Non-Aggression and Common Defence Pact, while not yet invoked in practice, has not met with any protest by third states. The Israeli air strikes in Syria in 2003 were explicitly rejected by almost all UNSC members at the time, with the exception of Russia and the United States.Footnote 75 However, in 2006 the objection to the Israeli intervention in Lebanon was not so much directed at the invocation of the right to self-defence as to the scope (disproportionality) of the Israeli response.Footnote 76
International reaction to Turkey’s incursions into Iraq received support from the United States and the United Kingdom on the basis that Iraq was ‘unable or unwilling’ to prevent the use of its territory for attacks against Turkey.Footnote 77 The EU expressed concern about the proportionality of the larger-scale operations but did not condemn these as illegal.Footnote 78 Nonetheless, the incursions in the mid-1990s were criticized by the League of Arab States and the Non-Aligned Movement (NAM) as a violation of Iraqi territorial integrity and the rules of international law.Footnote 79 Iraq also protested against the Turkish operations on its territory in 2008 as a violation of its sovereignty.Footnote 80 While the United States was supportive of the Turkish military incursions into Iraq, it claimed that similar incursions by Iran had no justification.Footnote 81 This statement comes across as contradictory, given the United States’ long-standing support for the ‘harbouring’ and ‘willing or unable’ doctrine.Footnote 82
In the case of Yemen, Iran rejected allegations of direct involvement in the supply of arms to the Houthis. It underscored that the allegations were not corroborated by concrete evidence.Footnote 83 When Yemeni officials intercepted a shipment of arms originating from Iran and destined for Houthi rebels in January 2013, Iran emphasized that Yemen had not provided evidence of actual Iranian involvement in the shipment.Footnote 84 Allegations of Iranian involvement in arms supplies nonetheless persisted, while the Houthis themselves claimed extensive financial support from Iran.Footnote 85
After the commencement of the air strikes by the United States and its allies against IS in Syria in September 2014, the Assad regime initially did not protest against the air strikes. At the time, its protests were primarily directed at the military support provided by Turkey, Israel, Qatar, Jordan, Saudi-Arabia, and numerous Western states to (other) rebel groups in Syria.Footnote 86 However, as of 17 September 2015 this position changed.Footnote 87 Syria explicitly objected to the fact that the military action by the United States, along with that by Australia, Canada, France, and the United Kingdom, had not occurred in co-ordination with the Syrian government. It underscored on several occasions that it was in continuous military confrontations with armed groups, including IS, and that it was also ready to co-operate with the United Nations in this regard.Footnote 88 Syria rejected their invocation of Article 51 of the UN Charter in relation to the fight against IS as justification, claiming that this amounted to a distortion of the UN Charter.Footnote 89 In a subsequent letter, the Syrian government repeated this position. The letter continued by stating that in order to conform to the requirements of Article 51 of the UN Charter, there should be an ongoing act of aggression by an armed force against a member state; that the response should be temporary and that it should respect the authority and responsibility of the UNSC. These conditions were not met in Syria by states such as the United Kingdom.Footnote 90
The US-led air strikes against IS in Syria were also condemned as illegal by Russia and Iran.Footnote 91 China did not explicitly condemn the attacks. Instead, it noted that China supported counter-terrorism efforts but that international law, sovereignty, independence, and territorial integrity of the relevant countries should be respected.Footnote 92 Explicit support for and/or participation in the air strikes against IS were forthcoming from five Arab states, including Saudi-Arabia.Footnote 93 While these states did not explicitly link their support to the right to self-defence or to the ‘unwilling or unable’ doctrine,Footnote 94 they also did not criticize the invocation of these doctrines. Given their knowledge of the air strikes and that their own (regional) security interests were also affected, it is not unreasonable to interpret this lack of criticism as acquiescence in the manner in which these doctrines were invoked.Footnote 95 Similarly, there was a muted reaction from the 75 states belonging to the Global Coalition against Daesh. This coalition was formed in September 2014 with the purpose of eliminating IS from Iraq and Syria, including through military means ‘in accordance with the principles of international law, including the United Nations Charter’.Footnote 96 The fact that these states chose to belong to this coalition may be interpreted as an affirmation of the relevance of the military operations in Syria for their own security interests. In addition, through their membership of this coalition, they may be deemed to have knowledge of the manner in which other coalition members invoke the right to self-defence in Syria. Their lack of protest in this regard, therefore, may be interpreted as tacitly accepting such invocation to be in accordance with the UN Charter.
In 2016 the Non-Aligned Movement vaguely stated that ‘consistent with the practice of the UN and international law, as pronounced by the ICJ, Article 51 of the UN Charter is restrictive and should not be rewritten or re-interpreted’.Footnote 97 It is uncertain whether this statement has any direct bearing on any of the conflicts in which self-defence was invoked at the time, given its general nature. Furthermore, several members of the NAM support a broad notion of aggression also by non-state actors. These include Jordan, Bahrain, and the United Arab Emirates, which are participating in the air strikes against IS in Syria,Footnote 98 as well as the 22 African states that ratified the AU Non-Aggression and Common Defence Pact. This makes it difficult to determine what is meant by a ‘restrictive’ interpretation of Article 51 of the UN Charter, or whether the statement would constitute a criticism against the ‘unwilling or unable’ doctrine.Footnote 99
As far as the 2018 Turkish incursion into Afrin was concerned, Syria repeatedly objected to the Turkish military presence, which it described respectively as an occupation, a violation of the UN Charter and aggression towards Syria.Footnote 100 However, reaction of third states was either not forthcoming or nuanced. This can be exemplified by the German reaction only two months after the commencement of the incursion. On the one hand, it condemned the high number of civilian casualties and impact on the civilian population, and also expressed concern about the potential illegality of long-term Turkish presence in the region.Footnote 101 Germany, therefore, was critical of the disproportionality of the operation and its humanitarian consequences. On the other hand, Germany acknowledged the justified security interests of Turkey and stopped short of depicting the military operation as illegal.Footnote 102 Similarly, the European Parliament criticized the humanitarian implications of the military operation as such, as well as its political repercussions for a negotiated settlement in Syria.Footnote 103
On the whole, it is fair to conclude that there is growing state support for applying the right to self-defence outside the narrow parameters proclaimed by the ICJ in the Nicaragua decision.Footnote 104 This broader application also includes support for the invocation of the ‘harbouring’ doctrine, as well as the ‘unwilling or unable’ doctrine. Whether this practice is already clear and consistent enough to reflect a widened common understanding of Article 51 UN Charter is debatable, given the vagueness or ambiguity plaguing many reactions. Similarly, there may be diverging views as to whether the (acceptance of) the broader application of the right to self-defence is sufficiently wide-spread and consistent to reflect customary international law.Footnote 105 For example, the protests of the territorial (and thereby specifically affected) states such as Syria may cast doubt in this regard.Footnote 106 On the other hand, protests by third states were at times delegitimized by the fact that they had themselves invoked the ‘harbouring’ or ‘unwilling or unable doctrine’ in connection with the right to self-defence.Footnote 107 Moreover, the criticism often was more concerned with the application of the right to self-defence in a particular context, such as the lack of the proportionality of the defensive measures, or of conclusive evidence about the author of the armed attack.Footnote 108 It is therefore plausible to argue that states are increasingly supporting a broader invocation of the right to self-defence outside the narrow parameters proclaimed by the ICJ in the Nicaragua decision.
If one accepts this reasoning, the question arises as to how this impacts the principles of attribution of conduct within the self-defence paradigm. The subsequent sections argue that these developments imply a lowering of the threshold requirements for indirect armed attacks. In particular, they imply a more flexible interpretation of ‘substantial involvement’ in an armed attack, as well as a potential inter-relationship between the notions of ‘substantial involvement’, ‘harbouring’, and the ‘unwilling’ prong of the ‘unwilling or unable’ doctrine. In addition, the above-mentioned developments support the conclusion that armed attacks may be directly attributed to non-state actors in states that are ‘unable’ to prevent attacks by non-state actors located in their territory.
4. Implications for attribution of conduct within the self-defence paradigm
4.1 Expansive attribution to states of armed attacks
The above overview of state practice suggests that the current threshold of an indirect armed attack in terms of Article 51 of the UN Charter is lower than that established in the Nicaragua decision.Footnote 109 Similarly, this threshold is more flexible than the customary international law standard for attribution of conduct of private actors to states, according to which a close nexus between the state and the non-state actor is required.Footnote 110 State practice indicates that third state support to armed groups, which the ICJ in the Nicaragua decision qualified as ‘mere’ violations of the prohibition of the use of force, nowadays could trigger the right of individual and collective self-defence of the victim state against the state(s) that supported the non-state actor.Footnote 111 This implies that state support in the form of weapons, finances, training, logistical support – or any combination of these – to non-state actors attacking another state could amount to the type of ‘substantial involvement’ in an armed attack that would make such an attack against a state attributable to the supporting state.Footnote 112 While states are yet to describe such measures of support as constituting ‘substantial involvement’ in explicit terms, it is difficult to see how else to reconcile this requirement with, for example, Article 1(c)(xi) of the AU Non-Aggression and Defence Pact.
Furthermore, the right to self-defence could also be triggered by a state’s violation of its due diligence obligation under international law, where it failed to act against a non-state actor engaged in attacks against another state. This, in turn, suggests that the ‘harbouring’ doctrine and the ‘unwillingness’ prong of the ‘unwilling or unable’ doctrine constitute particular concretizations of the violation of the due diligence obligation in the jus ad bellum context. Both concepts may be understood as referring to the inaction of the territorial state vis-à-vis a non-state actor, despite knowledge on the part of the territorial state of the presence of the non-state actor and its activities.Footnote 113 Inaction under these circumstances can trigger the right of individual or collective self-defence of the victim state.Footnote 114 Moreover, it is also plausible to regard ‘unwillingness’ and/or ‘harbouring’ as a form of ‘substantial involvement’ through omission.Footnote 115 This is because of the fact that the territorial state has knowledge of the circumstance that the presence of the non-state actor on its territory will facilitate an armed attack against another state.Footnote 116 Read in this way, ‘unwillingness’ and/or ‘harbouring’ will constitute the flipside of ‘substantial involvement’.
At this juncture, one has to point out that state practice is yet to explicitly endorse any inter-relationship between ‘unwillingness’, ‘harbouring’, and ‘substantial involvement’. In fact, it is not even certain that those states invoking the ‘unwilling or unable’ doctrine actually depart from a clear common denominator, as they either do not explain what they mean or use vague and, to some extent, different terminology.Footnote 117 For example, in the ongoing Syrian conflict, the United States have indicated that Syria cannot and will not confront IS safe havens in Syria ‘effectively’.Footnote 118 Australia described the ‘failure to constrain’ attacks by IS in Iraq as a demonstration of unwillingness or inability.Footnote 119 Canada and Turkey, for their part, merely claimed that the (Syrian) government was not able (capable) or willing to ‘prevent attacks emanating from its territory’. Footnote 120
That being said, it is possible to read state practice as implicitly confirming that the ‘harbouring’ and ‘unwillingness’ doctrines are manifestations of violations of due diligence, as well as that they constitute ‘substantial involvement’ through omission. If state practice indeed were reflective (or to become reflective) of a close relationship between ‘unwillingness’, ‘harbouring’, and ‘substantial involvement’, an armed attack would be attributable to a state either through its substantial involvement in the form of active support of non-state actors attacking the victim state, or in the form of passive involvement through failing to act against the presence of such a non-state actor on its territory. In these instances, the exception to the prohibition of the use of force against the territorial integrity of a state contained in Article 51 of the UN Charter is triggered, allowing the victim state (and its allies) to use defensive force in the territory of (the) states(s) which is/are the author(s) of the armed attack.Footnote 121
Where Article 51 of the UN Charter is triggered in such a way, all forcible measures exercised by the victim state and its allies will have to conform to the principles of necessity and proportionality embedded in the right to self-defence.Footnote 122 The necessity requirement first of all relates to the question of whether, in the circumstances, less severe means of redress are available. A response in the form of self-defence, therefore, should be a last resort.Footnote 123 In making this assessment, the issue of the consent of the territorial state becomes an important consideration. The victim state will bear the onus to indicate why it did not first attempt to obtain the co-operation of the ‘unwilling’ territorial state, including consent for executing military measures in its territory.Footnote 124 If the victim state claimed to have concerns about the collusion between the territorial state and the non-state actor, it would have to present credible evidence to this effect.Footnote 125 The necessity requirement further implies that the defensive response has to target military objects that are the actual source of the attack.Footnote 126 This would include those non-state actors involved in the execution of the armed attack along with their infrastructure. It can also include those facilities and personnel of the state to which the armed attack is attributed, in as far as they directly supported or facilitated the armed attack.Footnote 127
The issue of military targeting, however, can also have bearing on the proportionality of the defensive measures. The proportionality requirement concerns the overall scope and scale of the defensive measures. It thus concerns the relationship (that is, the need for correlation) between the gravity of the armed attack and the intensity of the response.Footnote 128 This at times can have consequences for military targeting practices, suggesting a close intertwinement of necessity and proportionality.Footnote 129 For example, the targeting practices of the Saudi-led coalition against the Houthis in Yemen has faced severe criticism, for lacking in both necessity and proportionality.Footnote 130 Both principles would also have to be considered before expanding the defensive operations to targets in Iran (which, according to Yemen and Saudi Arabia, sponsored the attacks). Thus far, the coalition has limited the military operation to the territory of Yemen.
4.2 Limited attribution to non-state actors of armed attacks
While the above analysis illuminated the lowering of the threshold requirements for attribution of armed attacks to states, this section focuses on the attribution of armed attacks to non-state actors. One situation in which this seems to be occurring is where a state has lost effective control over the area in its territory where the non-state actor is located. As a result, the territorial state is unable to prevent the use of the respective part of its territory by non-state actors in the preparation of an armed attack against other states.Footnote 131 This, in turn, suggests that the loss of territorial control constitutes ‘inability’ in line with the ‘unable’ prong of the ‘unwilling or unable’ doctrine.
For example, in Syria the incumbent government did not have effective control over the areas from which IS operated, despite attempts by the government to eradicate it.Footnote 132 This lack of control was also explicitly mentioned by Germany and Belgium when invoking Article 51 of the UN Charter against IS.Footnote 133 Similarly, Turkey claimed that the Afrin region was under the control of the YPG/PYD, implying that the Syrian government did not have control over the region.Footnote 134 Interestingly, none of these countries explicitly invoked the ‘unwilling or unable doctrine’ in these respective situations. Therefore, they have not yet explicitly linked the ‘unable’ prong of the doctrine to a lack of territorial control. Even so, it is arguable that an inability would exist in situations where the non-state actor has gained and stabilized its control over a territory within a state.Footnote 135 Also, it seems that in such situations states have attributed the armed attack directly to the non-state actor, allowing the victim state(s) to use self-defence directly against it.Footnote 136 For example, Belgium and Germany explicitly made IS responsible for armed attacks and underscored the fact that their military measures were directed against IS only and not against the Syrian state.Footnote 137 In relation to the situation in Afrin, Turkey suggested the same by stating that its military operations would ‘target only terrorists and their hideouts, shelters, emplacements, weapons, vehicles and equipment’.Footnote 138
Furthermore, when invoking Article 51 of the UN Charter in relation to IS, Germany, Belgium, Denmark, Norway, and the United Kingdom made reference to UNSC Resolution 2249 (2015) which called upon:
[m]ember 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 … in Syria and Iraq. Footnote 139
This suggests that they interpreted the Resolution as allowing them to invoke the right to individual and collective self-defence directly against non-state actors. In fact, DenmarkFootnote 140 and GermanyFootnote 141 explicitly confirmed this interpretation. However, it is worth noting that, while UNSC Resolution 2249 (2015) refers to the use of ‘all necessary means, in compliance with international law’, it was not adopted under Chapter VII of the UN Charter, nor does it explicitly refer to Article 51 of the UN Charter. Therefore, it does not in and of itself authorize force against IS as a non-state actor as a matter of self-defence.Footnote 142 Instead, it merely encourages states to use force within the IS-controlled areas in as far as international law allows for this in the absence of a UNSC Chapter VII authorization.Footnote 143
Stated differently, while the Resolution may be read as implicitly encouraging states to consider whether Article 51 of the UN Charter can be invoked directly in relation to IS in Syria, it would not seem to give a clear answer as to whether this indeed is the case.Footnote 144 Nonetheless, despite the vagueness of this Resolution, there does seem to be growing support among states for attributing armed attacks exclusively to non-state actors in situations where they operate from areas over which the territorial state does not have control.Footnote 145 While the current military operation in Syria may not in itself be sufficiently widespread evidence of this development, one should not overlook Article 1(c)(xi) of the AU Non-Aggression and Common Defence Pact. This Article, which is already in force for 22 signatories in Africa, explicitly acknowledges that an act of aggression can be committed by a non-state actor.Footnote 146 These developments in state practice seem to vindicate the views of individual ICJ judges in the DRC v. Uganda decision, according to which cross-border attacks carried out by irregular forces from territories which were not under governmental control nonetheless amounted to armed attacks, despite not being attributable to the territorial state.Footnote 147
If one accepted that an armed attack can be attributed directly to a non-state actor located in the territory of an ‘unable state’ that has lost control over parts of its territory, the logical consequence would be that measures of self-defence would be allowed against the non-state actor as such. This, in turn, would imply an intrusion into the territorial integrity and sovereignty of the territorial state, despite the fact that it was not responsible for the armed attack. This raises the question of whether such an intrusion could be accommodated within the UN Charter Article 51 paradigm. After all, the exception to the prohibition of the use of force against the territorial integrity of a state contained in this article is only triggered as of right, where the armed attack is attributable to that state.Footnote 148
It is arguable that defensive measures against a non-state actor on the territory of an ‘unable’ state can be accommodated within the Article 51 paradigm through the necessity principle. Section 4.1 above indicated that necessity is also applicable where self-defence is invoked in relation to armed attacks attributable to states. However, when invoked in relation to ‘unable’ states, the necessity principle requires the function of a circumstance precluding wrongfulness. It allows for the victim state and its allies to exercise measures in self-defence in the ‘unable’ state, without engaging state responsibility for violating the territorial integrity and sovereignty of the latter.Footnote 149 This reasoning essentially regards the right to self-defence as both an inherent right and a justification (or excuse). Stated differently, it supports the view that Article 51 of the UN Charter functions as a primary and secondary norm.Footnote 150 Such an interpretation is likely to be opposed by those supporting a strict separation between primary and secondary norms. In line with a strict separation, the right to invoke self-defence against the perpetrator of an armed attack in Article 51 of the UN Charter represents a primary norm, while justification for intrusion into the territory of an (unable) state has to be found in a secondary norm.Footnote 151 However, the propriety of such a strict separation between primary and secondary is questionable, as the ILC introduced the distinction for organizational purposes and not as a conceptual tool.Footnote 152
Moreover, accepting a more flexible relationship between primary and secondary norms would facilitate a reading of Article 51 of the UN Charter that is informed by Article 25(1)(a) of the ASR.Footnote 153 This article, which allows for the invocation of the principle of necessity if it is ‘the only way for the state to safeguard an essential interest against a grave and imminent peril’,Footnote 154 constitutes one of the six circumstances precluding wrongfulness in Chapter V of the ILC Commentary to the ASR.Footnote 155 Such a reading would be quite novel, as the state of necessity referred to in Article 25 of the ASR is a separate construct from the customary principle of necessity that governs the right to self-defence in Article 51 of the UN Charter. Also, there is nothing in the work of the ILC pertaining to state responsibility (nor as of yet in state practice) that explicitly refers to a link between these different manifestations of necessity.Footnote 156 On the other hand, there is nothing in the work of the ILC (nor in state practice) that necessarily excludes such a link. Accepting a reading of Article 51 of the UN Charter that is informed by Article 25(1)(a) of the ASR would infuse the imprecise customary standard of necessity in the self-defence context with more exact benchmarks.Footnote 157 This seems desirable in light of the severe impact of defensive military operations on the sovereignty of the ‘unable’ territorial state.Footnote 158
According to the wording of Article 25(1)(a) of the ASR, the peril must be both grave and imminent, must be clearly established on the basis of evidence reasonably available at the time, while the measures for its eradication must be the only means available for safeguarding the interests at stake.Footnote 159 When applying this reasoning to the invocation of necessity in the context of Article 51 of the UN Charter, one may argue that the inability of a state to control territory from where armed attacks by non-state actors emanate can result in a grave and imminent peril to the national security of other states, which is an essential interest.Footnote 160 However, the victim state and its allies would further have to argue that the invocation of self-defence was the only way for safeguarding the essential interest. This places a strong onus on the victim state and its allies to establish why less severe measures were not feasible under the circumstances. Among others, they will have to demonstrate why military measures were unavoidable and, if so, why this did not occur with the consent and in co-ordination with the government of the ‘unable’ state. The states invoking self-defence will further have to limit its defensive measures very strictly to the source of the attack, which in the case of an ‘unable’ state would only include (the infrastructure of) the non-state actor(s) responsible for the armed attacks.Footnote 161
If, by way of illustration, one applies this reasoning to the manner in which the right to self-defence was invoked against IS in Syria, it is questionable whether the exacting standards of Article 25(1)(a) would be met. On the one hand, the Assad regime did not at the time have effective control over the parts of its territory from which IS operated. The activities of IS also constituted a grave and imminent peril to the national security of the states invoking self-defence. On the other hand, the Assad regime was willing to co-ordinate its military operations against IS with those of other States.Footnote 162 The United States-led coalition nonetheless insisted on conducting its own, separate military campaign against IS on the basis of self-defence. In line with the necessity principle as described above, these states would have to show that there was no other way of safeguarding their national security. One plausible argument can be that this military strategy was the only way to eliminate the grave threat posed by IS, without becoming complicit in the international crimes of the Syrian regime.Footnote 163 Stated differently, accepting the Assad regime’s conditions for military co-operation would necessarily implicate the members of the international coalition in the widespread violations of human rights and humanitarian law by the Syrian regime.Footnote 164 At the same time, the threat emanating from IS posed a grave and imminent peril that justified a military response.
However, for this argument to meet the high threshold of Article 25 of ASR(1)(a), the respective states would have to present it in clear and convincing terms. In particular, they would have to explain why their justified concerns for complicity in war crimes would necessarily exclude every form of co-operation with the Syrian regime in relation to the fight against IS. Moreover, similar justifications would be required in situations in Syria where government forces had regained territorial control but could not yet entirely prevent IS from forming a grave peril to other states. In such a situation, the inability would not so much relate to a lack of territorial control as to limited military capacity of the government forces. Thus far, such detailed explanations have not been forthcoming from the states invoking self-defence against IS in Syria. As a result, it is doubtful whether they comply with the stringent requirements of Article 25(1)(a) of the ASR, assuming that one accepted the relevance of this article for the interpretation of the necessity principle within the jus ad bellum paradigm.
5. Conclusion
The standards for attribution of armed attacks executed by non-state actors have evolved since the 1986 Nicaragua decision. First, the threshold requirements for attributing armed attacks to states have been lowered. Second, it seems that armed attacks can now be attributed directly to non-state actors. While these developments may have been necessitated by a responsibility vacuum that previously existed within the jus ad bellum paradigm, its eroding impact on collective security should not be underestimated. It is no exaggeration to claim that any significant expansion of the scope of the right to individual or collective self-defence has the potential to eradicate the UN Charter system for the centralization of the use of force in its entirety.Footnote 165
In an attempt to find a balance between evolution and erosion of the UN Charter, the above analysis proposed an interpretation of the ‘harbouring’ and ‘unwilling or unable’ doctrines that would anchor them more firmly in principles of state responsibility. As far as the threshold requirements for attribution of armed attacks to states are concerned, the article makes a case for interpreting ‘harbouring’ and the ‘unwillingness’ prong of the ‘unwilling or unable’ doctrine as manifestations of due diligence and as the potential flipside of ‘substantial involvement’. By anchoring these concepts in due diligence, it becomes more difficult for states to infuse them with entirely subjective and arbitrary standards. It would further mean that even if the invocation of the ‘unwilling or unable’ doctrine were, in time, to eclipse that of ‘harbouring’ in practice, it need not necessarily mean that the threshold requirements for attribution were lowered even further. At least as far as the ‘unwillingness’ prong is concerned the benchmarks of due diligence would remain relevant for attributing the armed attack to the state where the non-state actor was located.
The attribution of armed attacks directly to non-state actors would be most plausible where non-state actors operated from an area over which the territorial state has lost control. Such a situation constitutes the clearest example of ‘inability’, that is, of the ‘unable’ prong of the ‘unwilling or unable’ doctrine. However, as the invocation of self-defence in such a context has drastic consequences for the ‘unable’ state’s sovereignty, the necessity requirement attached to the right to self-defence must be strictly interpreted. In order for this to happen, the imprecise necessity principle within the law of self-defence needs to be infused with more exacting standards. One possibility for doing so is to interpret it in line with the requirements underpinning Article 25(1)(a) of the ASR.
While the above analysis developed this argument specifically in relation to ‘unable’ states, it is worth considering the application of such a stringent interpretation of necessity in all instances where the right to self-defence is invoked. Concepts such as ‘harbouring’, ‘unwilling’, and ‘substantial involvement’ suffer from various degrees of imprecision which can make them open to abuse. In addition, the lines between them can be fluid. For example, victim states and territorial states might disagree over whether ineffective measures by the territorial state against an armed group within the territory are a result of a lack of capacity (inability) or lack of political will (unwillingness). Moreover, they may also disagree as to what constitutes ‘effective measures’ in a particular context. In such circumstances, the application of an exacting standard of necessity, along the lines of Article 25 of the ASR, can assist in preventing an abusive or pre-textual invocation of the right to self-defence.
