There is controversy about whether and in what circumstances a State may act in self-defence in response to armed attacks carried out by non-State actors. Through an examination of State practice and ICJ decisions, this article examines the requirement that an armed attack must be attributable to the State against which self-defence is exercised. The author argues that there is confusion in the way in which the topic has been dealt with, and seeks to clarify some important conceptual issues. Ultimately, it is argued that the previously accepted 'effective control' attribution threshold permitting self-defence has been altered by the military response in the wake of the 11 September 2001 terrorist attacks, to a test of 'sanctuary and support'.
If country X, within its borders, is openly tolerating or incapable of managing a location where people are consistently attacking a neighbour, is it sufficient to say, 'Well, it's within their sovereign territory, nobody can do anything about it'? I think that's not true and 1 think there's a serious question about whether that's what the law ought to be. (1)
The phenomenon of modern terrorism has exposed a serious international legal problem affecting global peace and security. The difficulty can be described as follows. Non-State entities, such as terrorists, carry out attacks on a State ('the victim State'), but operate from or take sanctuary in another State ('the sanctuary State'). The victim State wishes to quell the attackers residing within the sanctuary State's borders. The sanctuary State did not actually carry out or direct the attacks, although it may be idle in preventing or removing the presence of the hostile actors from its soil. Can the victim State lawfully take cross-border action to neutralise the non-State attackers?
In this situation, the elimination of hostile actors or terrorists within the territory of another nation collides with two fundamental principles of international law: territorial sovereignty and the prohibition on the use of force prescribed in article 2(4) of the United Nations Charter ('UN Charter'). The only possible exception to a violation of both of these principles is the right of self-defence set out in article 51 of the UN Charter. (2) Under that provision, all States have a right of self-defence in response to an 'armed attack'. But what level of culpability, if any, is required on the part of the sanctuary State to permit the victim State to exercise its right of self-defence? This question has given rise to an incredible amount of controversy since it emerged as a topic of major global importance when the United States (US) carried out military activities in Afghanistan in the aftermath of the 11 September 2001 ('September 11') terrorist attacks. The issue is fundamentally that of the link required between the sanctuary State and the non-State entity.
In answering this question, this article advances three ultimate propositions. First, that the current position of international law is that an armed attack must be attributable to the State against which self-defence is exercised. Second, the threshold test for that attribution has changed from a State having 'effective control' over the attackers, to providing the attackers with 'sanctuary and support'. Third, this test of attribution is justified as the best principle for protecting international peace and security, and strikes a reasonable balance between self-defence and territorial sovereignty. Additionally, this article attempts to clarify some confusing conceptual problems that have arisen in discussions on this topic.
The article begins by examining the traditional requirements for a valid exercise of self-defence: Sections 1 and 2 discuss the requirements of 'armed attack', as well as 'necessity' and 'proportionality'. Section 3 examines the requirement that an armed attack must be attributable to the State against which self-defence is sought to be exercised. The decisions of the International Court of Justice ('ICJ') and relevant State practice are examined, and a conclusion on the current state of the law is reached. Section 4 provides a normative evaluation of whether the new, 'sanctuary and support' attribution threshold test is defensible.
1. Armed attack
Under article 51 of the UN Charter, all States have a right of self-defence only in response to an 'armed attack'. (3) The term 'armed attack' is not defined in the Charter, but the ICJ in the Nicaragua case clarified the concept. The Court effectively set out a test of sufficient gravity, distinguishing between the gravest uses of force constituting armed attacks, and less grave forms. (4) It held that armed attacks were not limited to actions taken by a State's armies across international borders, but also encompassed the sending of irregulars into another State, where the irregulars carried out attacks of similar gravity to regular armed forces. The touchstone of an armed attack is, therefore, the gravity of the attack. (5)
Two concepts are often elided when discussing self-defence, which can cause confusion, especially in considering the attribution requirement discussed later in this article. One concept is that of whether an 'armed attack' has taken place; and the other concept is that of which entity actually carried out the attack. There has been a tendency to speak of whether or not non-State actor terrorists can carry out armed attacks, or whether a government's involvement was sufficiently linked to a non-State actor so as to constitute an armed attack. For example, in Nicaragua, the ICJ held that 'assistance to rebels in the form of the provisions of weapons or logistical support' could not constitute an armed attack. (6) In the passages surrounding this quote, the Court was attempting to convey two ideas. First, not all forms of force are sufficiently grave to constitute an armed attack. The world would be a very unstable place if mere border incidents or lesser forms of force triggered the gears of self-defence. Second, States will not be responsible for attacks in which they have no involvement. Hence, the Court held that for the attack to be attributed to the State, the State must have 'effective control' over the individuals carrying out the attacks. (7)
However, these are two logically distinct legal ideas and should be treated as such. Any entity is theoretically capable of carrying out an 'armed attack'. Armed attacks do not necessarily emanate only from States. If terrorists carry out continuing attacks of sufficient gravity and take shelter on the high seas, a State should be able to rely on its right of self-defence. In such a scenario, no State would be involved in any possible way, but it does not follow that an armed attack has not occurred. To mingle the concept of the attribution or culpability of a sanctuary State with the concept of the seriousness of the harm inflicted upon the victim State only creates legal confusion.
Accordingly, it is important to be very clear about the two different questions involved when discussing armed attacks. One question is whether the attack was sufficiently grave so as to constitute an armed attack, and the other question is which entity carried out the attack. It is this latter aspect, the question of attribution, which is the focus of this article. (8)
2. Necessity and proportionality
The law requires any exercise of self-defence to be both necessary and proportionate. (9) These requirements can be traced to the well-known Caroline incident, although they were not universally accepted until after the UN Charier was adopted. (10) In 1837, British personnel infiltrated a US port and destroyed the SS Caroline because it was being used (by non-State actors) for carrying out attacks in Canadian territory. Britain claimed it was acting in self-defence and the US Secretary of State replied through diplomatic correspondence that self-defence could only justify conduct where it was necessary and where the response was not unreasonable or excessive. (11) Britain accepted this.
Proportionality has not been precisely defined in the context of self-defence and the ICJ has tended to simply review the facts and assert a conclusion that the action was or was not proportionate in the circumstances. The few cases in which proportionality has been explicitly addressed have involved fairly obvious disproportionate exercises of self-defence. In these cases, the Court's method has been to compare the level of force used in the attack to the level of force used in self-defence. For example, in the Congo v Uganda decision of 2005, the ICJ held that Uganda acted disproportionately in seizing airports and towns hundreds of kilometres from Uganda's border. (12) In the Oil Platforms case, the US had bombed Iranian oil refineries and naval vessels in a clearly disproportionate response to the damaging of a US frigate caused by a mine and the firing of a missile of disputed origin that hit a US oil tanker.
Currently, the most significant issue in the law of self-defence is whether or not an armed attack must be attributable to the State that is the prospective target of an exercise of self-defence, and if so, what the test for that attribution requirement might be. Three primary views have emerged. (13) First, there is the view that the attribution test set out by the ICJ in the Nicaragua case must be satisfied: the State subjected to an exercise of self-defence must have had 'effective control' over the conduct of the individuals who carried out the attacks. Second, some argue that no attribution is required. (14) That is, States may exercise self-defence against any State and within any State's borders, regardless of whether or not that State is in any way responsible for the attacks. Third, there is the view that, whilst attribution is required, the test for attribution is not so high as to require 'effective control'. Rather, a lower threshold of support, such as acquiescence or harbouring, is said to be sufficient. (15) There are persuasive arguments in support of all three positions.
An examination of the decisions of the ICJ, State practice and opinio juris will reveal that the most accurate view is that the attribution threshold is now the provision to the attackers of sanctuary and support.
A. Decisions of the International Court of Justice
(i) Nicaragua case
The starting point for any consideration of this topic is the 1986 decision of the ICJ in the Nicaragua case. The Court set out a number of important principles governing the law of self-defence and, most importantly, the link required with non-State actors.
Nicaragua brought claims against the US for acts of force and violations of sovereignty, including the mining of Nicaraguan ports, operations against bases and installations, and supporting contras to fight the Nicaraguan Government. The Court held that these actions were violations of international law. In its defence, the US pleaded collective self-defence, arguing that Nicaragua had been actively supporting armed groups operating in El Salvador through the provision of weapons. However, the ICJ rejected the US's argument, holding that assistance to rebels in the form of the provision of weapons or logistical support could not constitute an armed attack. (16) At the most, such assistance may constitute a threat or use of force, or amount to a violation of the prohibition on intervention. Accordingly, the US could not rely on the right of collective self-defence. The same was also true for Nicaragua, in responding to the contras aided by the US. Both countries had been supplying arms and providing support, but the Court held that 'the provision of arms to the opposition in another State does not constitute an armed attack on that State'. (17)
It may be doubted whether the Court recognised that the threshold test for establishing attribution was significant. The Court did refer to a requirement of 'effective control' over the persons or group in question as the standard for attributing acts to States. (18) However, this did not form the basis of the Court's rejection of self-defence. What the Court said was the following:
the Court does not believe that the concept of 'armed attack' includes not only acts by armed bands where such acts occur on a significant scale but also assistance to rebels in the form of the provision of weapons or logistical or other support. (19)
A common interpretation of this passage is that the Court was qualifying the test required for attribution; namely, provision of arms or logistical support is insufficient for an armed attack to be attributed to a State. One then reads in the 'effective control' test referred to elsewhere in the judgment (20) and arrives at the conclusion that the Court was, in the above paragraph, merely applying that test of attribution. (21) However, a reading of a later passage in the judgment shows that this interpretation is wrong and shows that the Court was not there addressing the issue of attribution:
Even assuming that the supply of arms to the opposition in El Salvador could be treated as imputable to the Government of Nicaragua, to justify invocation of the right of collective self-defence in customary international law, it would have to be equated with an armed attack by Nicaragua on El Salvador. As stated above, the Court is unable to consider that, in customary international law, the provision of arms to the opposition in another State constitutes an armed attack on that State. (22)
In other words, even assuming that attribution is made out, the conduct attributed does not constitute an armed attack. Hence, when dealing with the supply of arms earlier, the Court was not saying that the supply of arms was insufficient to …