SELF-DEFENCE AGAINST TERRORISM IN THE POST-9/11 WORLD

In 1986 the then United States Secretary of State George Shultz asserted that:'It is absurd to argue that international law prohibits us from capturing terrorists in international waters or airspace; from attacking them on the soil of other nations, even for the purpose of rescuing hostages; or from using force against states that support, train and harbor terrorists or guerrillas'. At that time the United States’ claim of a right to use military force in self-defence against terrorism2 received little support from other states.3 The predominant view then was that terrorist attacks committed by private or non-state actors were a form of criminal activity to be combated through domestic and international criminal justice mechanisms.4 The notion that such terrorist acts should be treated as ‘armed attacks’ triggering a victim state’s right of self-defence was not accepted by the majority of states. To suggest, as Shultz had done, that a state not directly responsible for terrorist acts could have its territorial integrity violated by military action targeting terrorists located within that state, was a controversial proposition in 1986. However, some fifteen years later, when the United States and a coalition of allies launched a military campaign in Afghanistan following the 11 September 2001(hereafter ‘9/11’) terrorist attacks, there was virtually unanimous international support for the use of force.


I INTRODUCTION
In 1986 the then United States Secretary of State George Shultz asserted that: It is absurd to argue that international law prohibits us from capturing terrorists in international waters or airspace; from attacking them on the soil of other nations, even for the purpose of rescuing hostages; or from using force against states that support, train and harbor terrorists or guerrillas. 1 At that time the United States' claim of a right to use military force in self-defence against terrorism 2 received little support from other states. 3 The predominant view then was that terrorist attacks committed by private or non-state actors were a form of criminal activity to be combated through domestic and international criminal justice mechanisms. 4 The notion that such terrorist acts should be treated as 'armed attacks' triggering a victim state's right of self-defence was not accepted by the majority of states. To suggest, as Shultz had done, that a state not directly responsible for terrorist acts could have its territorial integrity violated by military action targeting terrorists located within that state, was a controversial proposition in 1986. However, some fifteen years later, when the United States and a coalition of allies launched a military campaign in Afghanistan following the 11 September 2001 (hereafter '9/11') terrorist attacks, there was virtually unanimous international support for the use of force.
This development raises a number of important questions. What impact has 9/11 and the subsequent war in Afghanistan had on the right of self-defence? 5 Are states now permitted to use force in response to terrorism committed by non-state actors? If so, how exactly has international law changed to accommodate this development? What limits are there on the right to use force against terrorism? Is recognition of such a right a necessary response to the emergence of new threats in the international system, or a dangerous development which undermines international stability?
This article addresses these issues. 6 Part II of this article begins with an overview of the right of self-defence as it has traditionally been understood, and then discusses the difficulties in accommodating responses to non-state terrorism within this traditional, inter-state framework. Part III examines state practice involving the use of force against terrorism. It argues that the almost unanimous international endorsement of the United States-led war in Afghanistan has extended the right of self-defence to permit the use of military force against states that support or harbour non-state terrorist organisations that have already committed serious terrorist attacks. Part IV of this article then analyses how the law has changed to accommodate this extension in the scope of self-defence. It argues that the requirement that an 'armed attack' be attributable to a state has been retained, but that the threshold for attribution has been lowered. Tolerating or hosting a non-state terrorist organisation now appears to be a sufficient basis for characterising non-state terrorist acts as an 'armed attack' by a host state. Finally, Part V considers the possible consequences of an extended right of self-defence. It argues that although unilateral action in self-defence may be required in exceptional circumstances, the potential for states to abuse an extended right of self-defence means that it is preferable for military responses to non-state terrorism to be conducted under the umbrella of United Nations (UN) Security Council authorisation.

II THE RIGHT OF SELF-DEFENCE A
The Traditional Requirements of Self-Defence The right of self-defence is one of only two exceptions 7 to the general prohibition on the use of force contained in Article 2(4) of the UN Charter. 8 It is derived from two sources Since the Charter's inception in 1945 the precise scope of the right of self-defence and the relationship between its two sources has been controversial. On the one hand, the restrictive view held by the majority of scholars and states is that the right of selfdefence, under both Article 51 and customary international law, is dependent on an actual 'armed attack' occurring. 9 According to this position, a broader interpretation of self-defence is contrary to the wording of Article 51 and incompatible with the Charter's primary purpose of restricting the unilateral use of force. 10 On the other hand, some writers 11 and a small number of states -most notably the United States and Israelhave consistently argued in favour of a more expansive right of self-defence that encompasses anticipatory self-defence, 12 the protection of nationals abroad, 13 and the use of force against terrorism. 14 They rely on the words 'the inherent right' in Article 51, as evidence that the less restrictive, pre-Charter customary international law right of self-defence has been preserved, rather than superseded by Article 51. 15 In The second key question in relation to the 'armed attack' requirement is who can commit such an attack? Although Article 51 does not explicitly state that an 'armed attack' must be committed by a state, the Charter framework was intended to govern relations between states and hence, the traditional assumption was that such attacks had to emanate from a state. 30 However, this has never meant that an 'armed attack' could only be carried out by the regular armed forces of a state. 31 On the contrary, international law recognises that the conduct of a non-state actor will be attributable to a state if there is a sufficiently close relationship between the two entities. 32 Consequently, grave attacks by non-state actors can qualify as 'armed attacks', provided such conduct is attributable to a state. In this way, the inter-state dimension underpinning the law of self-defence is retained.
Prior to 9/11, the test for attribution required a high degree of cooperation between a state and a non-state actor. An armed attack must be understood as including not merely action by regular armed forces across an international border, but also 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 (inter alia) an actual armed attack conducted by regular forces, or its substantial involvement therein (emphasis added The remaining conditions governing the right of self-defence are the requirements of necessity and proportionality. 38 A state that has suffered an 'armed attack' triggering its right of self-defence is not given carte blanche to respond with whatever degree of force it wishes. First, a state can only use force where it is necessary to respond to the 'armed attack'. This implies that recourse to force is a last resort where there are no other means of resolving a situation. Second, the degree of force used in self-defence must be proportionate to the 'armed attack'. 39 This means a state cannot use greater force than is needed to repel the armed attack. These two requirements help to distinguish lawful self-defence from unlawful reprisals 40 -the latter being forcible responses for retaliatory or punitive purposes, rather than for defensive reasons. B

Self-Defence Against Terrorism
In principle at least, it has always been possible to accommodate the use of force in response to terrorism within the traditional, inter-state self-defence paradigm outlined above. 41 For this to occur, three conditions need to be satisfied. First, a state must suffer a terrorist attack which meets the gravity threshold of an 'armed attack'. Second, those terrorist attacks must be attributable to another state. Third, the use of force in self-defence must be necessary and proportionate.
In practice, however, the nature of terrorist acts and non-state terrorist organisations means that it is often difficult to satisfy these conditions. 42 There are two major difficulties. The first concerns the gravity threshold for an 'armed attack', while the second relates to the attribution requirement.
Prior to 9/11, few individual terrorist attacks were serious enough to meet the ICJ's high threshold for an 'armed attack'. 43 The majority of international terrorist acts consisted of relatively minor attacks on nationals abroad, rather than large-scale attacks on the actual territory of the victim state. It was therefore difficult to equate terrorist acts committed by non-state actors with conventional attacks by a state. As such, it was rare for an individual act of terrorism to qualify as an ' In an attempt to overcome this difficulty, the United States and Israel have adopted the 'cumulative effect' argument. 45 Under this approach, rather than measuring the gravity of each individual terrorist attack according to the threshold for an 'armed attack', consideration is given to the cumulative effect of a series of attacks. Hence, the argument is that an ongoing campaign of terrorist acts, viewed collectively, can amount to an 'armed attack'. This approach, which appears to have been accepted by the ICJ, 46 has played a major role in attempts by the United States and Israel to establish that terrorist acts should be treated as 'armed attacks'.
The second major difficulty with non-state terrorism is satisfying the attribution requirement. 47 As discussed above, to qualify as an 'armed attack', terrorist acts by non-state groups must be attributable to a state. However, in the case of non-state terrorist groups it is often difficult to fulfil the 'effective control' test or other applicable bases for attribution. 48 Such groups may receive varying levels of support from a host state. This support can range from the state having complete operational control over the non-state actor, to the provision of weapons and funding, to mere acquiescence or toleration of the group's presence on a state's territory. 49 Only in the first scenario is the connection between the two entities sufficiently close to satisfy the 'effective control' test. In circumstances where support falls below this level, a state is not considered responsible for the acts of a non-state terrorist group.
The United States and Israel have long argued in favour of a less stringent test than the 'effective control' standard. 50 In their view, any form of assistance, or even mere acquiescence or toleration of a terrorist group, should make a host state responsible for the conduct of that group. 51 However, prior to 9/11, this assertion received little support from other states. The proposition that a state not directly responsible for terrorist attacks could have its territorial integrity violated by military action targeting non-state actors based on its territory, remained highly controversial at that time. 45 For more on the 'cumulative effect' argument see Garwood-Gowers, above n 19, 251. For more on the types of support a state may provide see Cassese, above n 41, 597-600. 50 See Sofaer, above n 11, 98-105. 51 Ibid 103. This is essentially the same argument that the United States made in the aftermath of 9/11, when it declared that 'we will make no distinction between the terrorists who committed these acts and those who harbor them '

III STATE PRACTICE INVOLVING THE USE OF FORCE AGAINST TERRORISM A Incidents Prior to 9/11
States using force in response to attacks by non-state actors is by no means a recent phenomenon. The famous Caroline incident of 1837, which led to the formulation of the necessity and proportionality criteria governing self-defence, is one such example. 52 More recently, from the 1960s onwards Israel and apartheid South Africa used military force on numerous occasions to respond to cross-border attacks by non-state groups based in neighbouring states. 53 These responses were generally condemned by the Security Council and the international community. 54 In addition, since 1991 Turkey and Iran have made a number of incursions into Iraq to target Kurdish military groups operating there. 55 In the context of self-defence against terrorism, there were four major incidents prior to 9/11. The first occurred in October 1985, when Israel bombed the Palestinian Liberation Organisation (PLO) headquarters in Tunis, Tunisia, in response to a series of terrorist attacks on Israeli citizens. Israel's argument that Tunisia's toleration of the PLO's presence on Tunisian territory had made it necessary for Israel to act in selfdefence to prevent further attacks, was not accepted by the international community. The Security Council condemned Israel's response as an 'act of armed aggression perpetrated by Israel against Tunisian territory in flagrant violation of the Charter of the United Nations, international law and norms of conduct'. 56 The second major incident occurred in April 1986 when the United States attacked Libyan government facilities in Tripoli following a terrorist attack on a Berlin nightclub which targeted United States service personnel. 57 Despite the United States' claims that the Libyan government had been directly involved in planning the nightclub attack, the international community condemned the United States' response. 58 The United States again used force to respond to terrorism when it fired cruise missiles at Iraq in 1993, following the discovery of a plot to assassinate former President George H. Bush while in Kuwait. 59 Although the plot was foiled, the United States claimed that it still amounted to an 'armed attack' triggering the right to use force in self-defence. 60 The response of the international community to this incident was mixed. While the United States' action was supported by the United Kingdom and Russia, there was 52 For more on the Caroline incident see above n 38. 53 For discussion of incidents involving Israel see generally O'Brien, above n 3. For a summary of incidents involving South Africa see Gray, above n 13, 99-102. 54 Ibid. 55 See Gray, above n 13, 103. Note that Turkey justified its actions on the basis of a state of necessity, rather than as self-defence. For more on necessity see Laursen, above n 6; Romano, above n 6. criticism from China and the Arab states. 61 On the whole, however, the international community's reaction indicated a growing acceptance among states of the need to use military force against terrorism, although this could not be considered as clear endorsement of the legality of the United States' action. 62 The final significant incident prior to 9/11 occurred in 1998 when the United States attacked an Al-Qaida terrorist training camp in Afghanistan and an alleged chemical weapons factory in Sudan. 63 This was a response to the bombing of the United States' embassies in Kenya and Tanzania, which the United States blamed on Al-Qaida. The United States' claim that it had acted in self-defence to 'prevent these attacks from continuing', 64 received support from several states. 65 However, there were condemnations from Russia, Pakistan and the Arab states, although neither the Security Council nor the General Assembly took any formal action. 66 At least part of the concern over the United States' action stemmed from the fact that the territorial integrity of Sudan and Afghanistan had been violated not in an attempt to target those states themselves, but rather to attack non-state terrorists located there. 67 The international community's response to these incidents indicates that even prior to 9/11 there was some acceptance by states of the need to use military force against terrorism. 68 However, it is clear that in the cases discussed above, there were doubts as to whether the particular factual circumstances satisfied the traditional requirements of self-defence.

B The 9/11 Attacks and the Afghanistan War
The international community's reaction to the 9/11 attacks and the subsequent United States-led military action in Afghanistan was markedly different from its responses to previous incidents involving the use of force against terrorism. In the aftermath of the 9/11 attacks, the right of self-defence was explicitly recognised by the Security Council in resolutions 1368 and 1373. 69 In addition, on 12 September, 61 Kritsiotis, above n 59, 163-4. 62 Ibid 175. 2001 NATO determined that the terrorist attacks amounted to an 'armed attack' and invoked Article V of the 1949 Washington Treaty for the first time. 70 This was followed two days later by invocation of the equivalent provision in the ANZUS Treaty. 71 When the United States-led military action in Afghanistan began on 5 October 2001, there was almost unanimous international support for the use of force. 72 Many states, including NATO members plus Japan, Australia, Russia and Pakistan, provided assistance during the military campaign. 73 Even states like Malaysia, which opposed the use of military force on political grounds, recognised that such force was 'a legitimate course of action as an act of self-defense '. 74 This international acceptance of the use of force against Afghanistan can be explained largely, but not entirely, by factual differences between the 9/11 attacks and previous incidents involving military responses to terrorism. 75 In several key respects the Afghanistan war was easier to accommodate within the traditional self-defence paradigm than were previous incidents involving the use of force against terrorism. 76 For a start, the 9/11 attacks were clearly serious enough to meet the gravity threshold for an 'armed attack', whereas previous terrorist attacks such as the foiled assassination plot against former President Bush and the Berlin nightclub bombing, were less likely to satisfy this threshold. 77 In addition, the fact that the 9/11 attacks occurred on the territory of the victim state, rather than being attacks on overseas targets, made it easier 70

NATO Press Release, NATO's Contribution to the Fight Against Terrorism <www.nato.int/terrorism> at 20 July 2004. Article V of the Washington Treaty reads:
The Parties agree that an armed attack against one or more of them in Europe or North America shall be considered an attack against them all and consequently they agree that, if such an armed attack occurs, each of them, in exercise of the right of individual or collective self-defence recognised by Article 51 of the Charter of the United Nations, will assist the Party or Parties so attacked by taking forthwith, individually and in concert with the other Parties, such action as it deems necessary, including the use of armed force, to restore and maintain the security of the North Atlantic area.
Any such armed attack and all measures taken as a result thereof shall immediately be reported to the Security Council. Such measures shall be terminated when the Security Council has taken the measures necessary to restore and maintain international peace and security . Each Party recognizes that an armed attack in the Pacific Area on any of the Parties would be dangerous to its own peace and safety and declares that it would act to meet the common danger in accordance with its constitutional processes. Any such armed attack and all measures taken as a result thereof shall be immediately reported to the Security Council of the United Nations. Such measures shall be terminated when the Security Council has taken the measures necessary to restore and maintain international peace and security.
Article V of the Treaty provides that: For the purpose of Article IV, an armed attack on any of the Parties is deemed to include an armed attack on the metropolitan territory of any of the Parties, or on the island territories under its jurisdiction in the Pacific or on its armed forces, public vessels or aircraft in the Pacific. Above n 72. 75 Stahn, above n 5, 36. 76 Ibid. 77 Beard, above n 5, 574-5. to classify them as an 'armed attack' against the United States. 78 Furthermore, after 9/11 the United States went to significant lengths to provide evidence linking Al-Qaida to the terrorist attacks; evidence which was accepted by other states. 79 This prevented the types of evidentiary doubts which undermined previous incidents, most notably the United States' 1998 strike on an alleged chemical weapons factory in Sudan. 80 These three factual differences between 9/11 and earlier instances of self-defence against terrorism contributed significantly to securing widespread international support for the use of force in Afghanistan.
However, the factual circumstances of 9/11 cannot provide a complete explanation of the international community's new acceptance of the use of force against terrorism. The requirement that an 'armed attack' must be attributable to a state appears not to have been satisfied after 9/11 -at least not according to the previously recognised test of 'effective control'. 81 Hence, there appears to have been a sudden and significant change in the international community's interpretation of the 'armed attack' requirement. As Cassese notes: it would thus seem that in a matter of a few days, practically all states … have come to assimilate a terrorist attack by a terrorist organization to an armed aggression by a state, entitling the victim state to resort to individual self-defence and third states to act in collective self-defence. 82 While customary international law usually takes a considerable period of time to develop, the notion of 'instant custom' provides a basis for new principles to emerge more rapidly. 83 Thus, it can be argued that the response of the international community in the aftermath of 9/11 has given rise to a new principle of customary international law that permits the use of force in self-defence in response to 'armed attacks' committed by non-state terrorist organisations. 84 The key question, discussed in the next section, is whether this change is the result of a lowering of the threshold for attributing terrorist acts to a state, or whether the attribution requirement has now been discarded altogether. 85 78 Stahn, above n 5, 36. 79 For more on the evidence provided by the United States, see Beard, above n 5, 576-8. 80 Ibid 576. 81 For similar conclusions see Arai-Takahashi, above n 5, 1098 ('[I]t is difficult and only faintly possible to consider the September 11 th attacks per se to be the act of the Taliban government'); G Ulfstein, 'Terrorism and the Use of Force' (2003) 34 Security Dialogue 153, 163 ('Even though in the case of Afghanistan there was close contact between the authorities and Al-Qaeda, there is little to suggest that the terrorists who attacked the United States were sent by the authorities, that they were acting on behalf of the those authorities or that the authorities were substantially involved in sending them in the manner required in the Nicaragua judgement [sic]'). 82 Cassese, above n 69, 996-7. 83 For more on 'instant custom' in the aftermath of 9/11 see Arai- Takahashi For a similar conclusion see Arai-Takahashi, above n 5, 1095. 85 Stahn, above n 5, 42.

IV THE POST-9/11 RIGHT OF SELF-DEFENCE AGAINST TERRORISM A The Attribution Requirement
It is clear from 9/11 and the war in Afghanistan that the 'effective control' test for attribution no longer applies to international terrorist acts in the context of the right of self-defence. 86 The more difficult question is what has replaced it. 87 There are two possible answers to this question. On the one hand, the threshold for attribution may have been lowered, such that any level of support, or even merely hosting or tolerating non-state terrorist groups, is now sufficient to make a state responsible for the conduct of those groups. 88 On the other hand, the more radical possible change is that the attribution requirement has now been removed altogether, thus permitting the conduct of non-state groups to qualify as an 'armed attack' even where there is no connection at all with a state. 89 This latter position appears to have been adopted by several of the judges who gave separate opinions in the Legal Consequences of the Construction of a Wall in the Occupied Palestinian Territory advisory opinion. 90 The better view, it is suggested, is that the attribution requirement remains part of the concept of 'armed attack', but the threshold for attribution has been lowered. This conclusion is supported by two main factors. First, it accords with the legal arguments made by the United States in the aftermath of 9/11. Several hours after the attacks the United States asserted that 'it would make no distinction between the terrorists who committed these acts and those who harbor them'. 91 This was an attempt by the United States to engage the responsibility of the Taliban, and thus of the state of Afghanistan, for the terrorist conduct of Al-Qaida. 92 The international community's acceptance of this argument points to the retention of the attribution requirement, albeit with a lower threshold, rather than its entire removal.
A second factor supporting the retention of the attribution requirement is that even since 9/11 the right of self-defence remains primarily a state-centric principle. Where force is used to target non-state actors located within a state, this will amount to a breach of that 86 Ibid 37. 87 For a detailed study of this question see generally D Jinks, 'State Responsibility for the Acts of Private Armed Groups' (2003) 4 Chicago Journal of International Law 83. 88 Authors that favour this interpretation include Byers, above n 5, 409-10 ('actively support or willingly harbour terrorist groups'); Travalio and Altenburg, above n 4, 111 ('the standard for state responsibility is one of sanctuary or support'); R Wolfrum, 'The Attack of September 11, 2001, the Wars against the Taliban and Iraq: Is There a Need to Reconsider International Law on the Recourse to Force and the Rules in Armed Conflict' (2003) 7 Max Planck Yearbook of United Nations Law 1, 34 ('Therefore a given action of a non-state actor is attributable to the respective subject [ie. a state] of international law supporting it, if that subject deliberately created a situation which was a necessary for a later event, provided the happening of that event was not beyond reasonable probability and constituted a breach of international law'). 89 See Stahn, above n 5, 42 ('It may be of greater consequence to admit openly that the requirement of attributability does not play a role in the definition of armed attack'). For more on the United States' legal strategy following 9/11 see Byers, above n 5, 408-10. state's territorial integrity. 93 To allow such a breach in circumstances where a state is not responsible for the conduct of non-state actors, is, as Schachter put it, 'too simplistic in a world in which territorial sovereignty of states is a dominant principle'. 94 While the notion of territorial sovereignty cannot present an absolute barrier to a victim state's right to defend itself against attacks by non-state actors, it is suggested that international law since 9/11 does not permit the use of force on the territory of states that are themselves free of any complicity in terrorist attacks committed by non-state actors. Instead, there remains a requirement that the conduct of non-state actors must be attributable to a state. 95 Thus, the inter-state dimension to self-defence has been retained.
The new post-9/11 threshold for attribution is considerably lower than the previous 'effective control' standard. It now appears that any level of support, or even willingly hosting, tolerating or harbouring terrorists, will be sufficient to make a state responsible for the conduct of those non-state actors. 96 This new standard will cover most situations where a non-state actor operates on the territory of another state. However, one scenario that would not satisfy the new attribution threshold involves weak or failed states that are unable to prevent terrorist groups from operating on their territory. 97 Here a host state may be willing but unable, due to lack of control or resources, to prevent the terrorist activities of that group. In such circumstances that host state is not willingly hosting or tolerating the terrorist group and thus, even according to the new, less stringent attribution threshold, it cannot be considered responsible for the conduct of that group. 98 While this anomaly would rule out self-defence as a legal basis for military action on the territory of that host state, it does not render a victim state powerless. It may be possible for a victim state to obtain the host state's consent to intervene on its territory in order to target a non-state terrorist group located there. 99 93 The view that some forcible incursions for limited purposes do not breach a state's territorial integrity, is not accepted by the present writer. <http://www.icjcij.org/icjwww/idocket/imwp/imwp_advisory_opinion/imwp_advisory_opinion_20040709.htm> at 20 July 2004 ('Article 51 of the Charter thus recognizes the existence of an inherent right of selfdefence in the case of armed attack by one State against another State. However, Israel does not claim that the attacks against it are imputable to a foreign State') (emphasis added). Note that this passage from the ICJ's judgment has received some criticism. See above n 90, and L Piggott, 'Has the ICJ Managed to Tie our Hands in the War Against Terrorism?' On Line Opinion (15 July 2004) <http://www.onlineopinion.com.au/view.asp?article=2371> at 20 July 2004. However, it is suggested that much of this criticism is misplaced. The law as expressed by the ICJ is capable of accommodating responses to attacks by non-state actors. 96 See various explanations of this threshold outlined in note 88, above. Note that there has been some criticism that the notion of harbouring or hosting terrorists is 'both too broad and too illdefined'. See Travalio and Altenburg, above n 4, 117. The recent Australian intervention in the Solomon Islands is an example of intervention by invitation, albeit in the context of restoring law and order, rather than targeting terrorist organisations.
Hence, it is suggested that this weakness in the new attribution requirement is unlikely to pose significant problems in practice.
A final issue in relation to the new attribution threshold is how widely such a standard applies. In this regard, it is suggested that the lower threshold applies only in the context of international terrorism and the right of self-defence. Thus, it is lex specialis, a special rule which forms an exception to the general law of attribution. 100 The general tests for attribution contained in the Articles on Responsibility of States for Internationally Wrongful Acts 101 continue to apply to all other situations.

B Limits on the Post 9/11 Right of Self-Defence Against Terrorism
The post-9/11 right to use force in self-defence against terrorism is subject to a number of limitations. One such pre-condition, the attribution requirement, has been discussed above. The others are derived from the general conditions governing self-defence, namely the gravity threshold for an 'armed attack', the need for an actual attack to have occurred, and the principles of necessity and proportionality. 102 The first of these limits, the gravity threshold, means that a state is only entitled to respond with force in self-defence if it has been the victim of serious or sustained terrorist attacks that are comparable to an 'armed attack' by a state. The 9/11 attacks provide an example of terrorist acts that satisfy the gravity requirement. On the other hand, isolated or minor terrorist attacks should continue to be considered criminal acts, rather than 'armed attacks'. 103 In such circumstances, a victim state is limited to responding through law enforcement procedures, criminal justice mechanisms and other non-forcible measures. 104 Allowing the use of force in response to minor terrorist attacks would amount to the 'militarization of crime' and would severely undermine international law's general prohibition on the use of force. 105 Thus, the extended right of self-defence against terrorism remains subject to the gravity requirement.
A second limitation relates to the need for an actual terrorist attack to have occurred before a state's right of self-defence can be exercised. The new right to use force against states that host or harbour non-state terrorist organisations only applies where there has already been an attack by those non-state actors. 106 The events of 9/11 and the subsequent war in Afghanistan do not provide any basis for the use of force against Above n 32. The Articles largely codify customary international law relating to state responsibility. For more on the Articles see generally Crawford, above n 24. 102 In Part II of this article these requirements were discussed in general terms. This section discusses their specific application in the context of self-defence against terrorism. 103 Ulfstein, above n 81, 168. 104 For more on non-forcible measures that states can use to combat terrorism see O'Connell, above n 97, 904-8. Note that some authors have argued that there is an emerging customary international law right of self-defence that permits states to use force in response to attacks that fall short of an 'armed attack'. See 105 Stahn, above n 5, 45. 106 Byers, above n 5, 409. potential threats posed by terrorist groups and their host states, although the United States did make such a claim in its National Security Strategy of 2002. 107 Rejection of that claim by most of the international community confirms that, for now at least, the right to use force against terrorism is limited to responding to actual 'armed attacks'. 108 Whilst there may be exceptional circumstances in which it is justifiable for a state to take anticipatory action where a terrorist attack is truly imminent, such action would not fall within the expanded right of self-defence that has crystallised since 9/11 and the Afghanistan war.
The remaining limitations on the post-9/11 right of self-defence against terrorism are derived from the principles of necessity and proportionality. These dictate the types of targets and the degree of force that a victim state is permitted to use in self-defence. 109 The necessity principle means that a state can only respond to terrorist attacks with military force if such force is the only reasonable way of removing the terrorist threat. 110 Non-forcible means of averting the threat should be exhausted before military force is used.
The principle of proportionality provides that the degree of force a victim state may use in self-defence depends on what action is needed to defend against the terrorist threat. 111 Where the threat can be removed by limited military force targeting the non-state terrorist actors themselves, this should be the extent of the victim state's response in self-defence. If, however, the relationship between a host state and a terrorist organisation is particularly close, removal of the terrorist threat may require more extensive military force against both the terrorist group and the host state itself. 112 Thus, the precise degree of force permitted in self-defence will vary depending on the particular factual circumstances of a situation.
While in principle, the limits to the extended right of self-defence discussed above are relatively easy to identify, in practice there may be significant difficulties in their application. Three main areas of concern are outlined in the next section. 113

V CONSEQUENCES OF AN EXTENDED RIGHT OF SELF-DEFENCE A Imprecise Conditions and Uncertainties
The post-9/11 right to use force in self-defence against states that host or harbour terrorist organizations that have already committed serious terrorist attacks raises three main concerns. The first relates to the range of targets that may be the subject of force 107 See The White House, The National Security Strategy of the United States of America (Washington, September 2002) <http://www.whitehouse.gov/nsc/nss.pdf>. For a discussion of the United States' doctrine of pre-emptive action see generally Garwood-Gowers, above n 12; Greenwood, above n 12. 108 For more on the international reaction to the United States' doctrine of pre-emptive self-defence see Garwood-Gowers, above n 12, 53-6. 109 For a detailed discussion of the various targets of the use of force in self-defence against terrorism see Stahn, above n 5, 46-51. 110 Travalio and Altenburg, above n 4, 113-4. 111 Stahn, above n 5, 50. 112 The war in Afghanistan against both Al-Qaida and the Taliban regime can be viewed as an example of this type of situation. 113 For more on these potential problems see Cassese, above n 69, 996-8.
in self-defence. 114 Given that transnational terrorist organisations such as Al-Qaida may have cells in many different states, the post-9/11 right to use force against states that host or harbour such groups could potentially allow each of these states to be targeted by military action. 115 At present, the doctrine of harbouring or hosting appears to make no distinction between a state that hosts or harbours one terrorist who poses little ongoing threat, and another state that hosts an entire terrorist network which continues to commit acts of violence. 116 To avoid the uncertainty of having multiple targets it is necessary to refine this doctrine, so that it provides a more concrete basis for identifying those states whose assistance or acquiescence in terrorist activity makes it truly necessary to use military force in self-defence. 117 Unless this is done, there will remain a danger that states may use the new right of self-defence against terrorism as means of removing unfriendly governments or pursuing their own strategic interests. 118 Closely linked to this first concern are potential problems relating to the duration of military action taken in self-defence. 119 With traditional self-defence against an 'armed attack' by another state, the right of self-defence ended when the 'armed attack' was repelled. 120 This was relatively straightforward to identify. However, in the context of international terrorist threats it may be difficult to determine when action in self-defence is complete. This is illustrated by the United States' admission that its so-called 'war against terrorism' may take years. 121 Such uncertainties over the duration of selfdefence create the potential for conflict between a victim state's right to defend itself against ongoing threats and the Security Council's role in maintaining international peace and security. 122 The third and most significant concern about the extended right of self-defence against terrorism relates to evidence provision and the authority of states to make unilateral decisions to respond with force. 123 Unless states are required to present clear and convincing evidence 124 of the need to use military force in self-defence against a state that harbours or hosts terrorist groups, there is a danger that this new right of selfdefence may be abused. 125 To avoid such a problem, Cassese argues that only: 114 Ibid. 115 Ibid 997. 116 Travalio and Altenburg, above n 4, 117. 117 Ibid.

118
Although the United States-led war against Iraq in 2003 involved a slightly different context (ie. the notion of taking pre-emptive action to avert a threat), it nevertheless provides an illustration of this danger. 119 Cassese, above n 69, 998. 120 Ibid. 121 Ibid. 122 Stahn, above n 5, 41. The 2003 Iraq war can be viewed an example of a conflict between the United State's claim to a right of pre-emptive self-defence and the UN Security Council's desire to continue weapons inspections as part of its role in maintaining international peace and security. 123 Cassese, above n 69, 1000. Establishing a Security Council-based mechanism of this nature is vital to ensuring the legitimacy of further military responses to terrorism. While the right of a victim state to act in self-defence does not require prior approval from the Security Council, it is clearly preferable for military action against terrorism to be conducted under the umbrella of Security Council authorisation. 127 A state's right of self-defence should therefore only be relied on in exceptional circumstances where it is impossible or impractical to gain such authorisation.
B Diminishing Role for the Security Council?
The development of a broader right of self-defence in the aftermath of 9/11 may also have consequences for the Security Council's role as the primary body responsible for maintaining international peace and security. 128 The less stringent requirements of selfdefence may mean that states now have an incentive to by-pass the Council and instead take unilateral action in self-defence. 129 This was the case after 9/11, when the United States made a strategic decision to rely on an expanded right of self-defence as the basis for military action in Afghanistan, rather than use the existing mechanism of collective security. 130 Unfortunately, the precedent created by the Afghanistan war's status as an exercise in self-defence may encourage other states to follow a similar course of action. Indeed, in the months that followed 9/11, Israel relied on the extended right of selfdefence to justify its incursions into Palestinian-administered territory, 131 while India threatened to take military action against Pakistan after the bombing of its parliament by a Pakistani-based terrorist group. 132 The United States-led war in Afghanistan, NATO's intervention in Kosovo in 1999 and the recent Iraq war are three recent examples in which military action has been taken by groups of states acting outside the UN collective security framework. These incidents may be evidence of a trend in which multilateral coalitions are increasingly replacing the UN as the body responsible for maintaining international peace and security. 133 Given that these coalitions may be used by stronger states to further their own interests, this may prove to be an undesirable development. Thus, in order to avoid the potential for such abuse, and to maximise the moral authority and legitimacy of any military response to terrorism, it is preferable for such action to be authorised by the Security Council.

VI CONCLUSION
The events of 9/11 and its aftermath have had an immense impact on the international system generally and, to a lesser degree, on the law of self-defence. While the emergence of non-state actors capable of carrying out large-scale attacks comparable to those committed by states posed some difficulties for the traditional, inter-state selfdefence paradigm, international law has demonstrated that it has the flexibility to accommodate such actors within its legal framework for the use of force. 134 By accepting the legality of using force in self-defence against the Taliban and Al-Qaida following 9/11, the international community recognised the existence of an expanded right of self-defence that permits the use of military force against states that host or harbour non-state terrorist groups that have already committed serious attacks. This new right still requires the actions of non-state terrorist organizations to be attributable to a state, but the threshold for attribution is now significantly lower.
Whether this expanded right of self-defence will ultimately prove to be a necessary and desirable development in response to new threats in the international system remains to be seen. However, it is clear that the current notion of self-defence is now a broader, more uncertain concept that it was previously. 135 While the development of an expanded right of self-defence against terrorism is less troubling than the doctrine of pre-emptive self-defence proposed by the United States in 2002, it carries a similar risk of abuse by states. 136 Military force alone will not defeat terrorism, but where such force is clearly necessary it is preferable for military action to be taken by way of Security Council authorisation. Wars of self-defence mounted by individual states or so-called 'coalitions of the willing' cannot provide the same legitimacy as action conducted under the banner of the UN. Thus, the right to take unilateral action in selfdefence against terrorism should be relied on only in exceptional circumstances. 133 Stahn, above n 5, 41. 134 Ibid 38. 135 Cassese, above n 69, 997. 136 Arai-Takahashi, above n 5, 1102.