While force used by a state in self-defence must meet the demands of proportionality there is confusion over the meaning of the term in this, jus ad bellum, context. One source of confusion lies in the existence of two competing tests of proportionality, the ‘tit for tat’ and the ‘means-end’ tests. Since the legality of unilateral use of force by a state depends on the legitimacy of its aim – self-defence against an armed attack – the ‘means-end’ test would seem more appropriate. However, there is no agreement over the legitimate ends of force employed to achieve this aim. Is the defending state limited to halting and repelling the attack that has occurred, or may it protect itself against future attacks by the same enemy? May a state that has been attacked use force in order to deter the attacker from mounting further attacks? The ‘means-end’ test of proportionality rests primarily on the necessity of the means used to achieve legitimate ends. Disagreements over proportionality are in this context usually really disagreements over those ends. While the appropriate test in this context is generally the ‘means-end’ test, in some cases, such as use of force in response to a limited armed attack, the ‘tit for tat’ test of proportionality might be more appropriate. Finally, I show that little attention has been paid in the jus ad bellum context to the ‘narrow proportionality’ test, which assesses whether the harm caused by the force outweighs the benefits to the state using that force. The apparent reason for this is the assumption that this question is only relevant in jus in bello. I argue that while necessity of the force used is indeed the main issue in jus ad bellum, there is still place for assessing narrow proportionality.
In the Newsletter of the American Society of International Law (ASIL) published in September/October 2006 the President of the Society, Professor José Alvarez, wrote an editorial entitled ‘The Guns of August’ in which he discussed legal aspects of the 2006 Israeli military campaign against the Hezbollah in Lebanon, and related specifically to questions of proportionality, mainly in jus in bello.1 In the same Newsletter five international lawyers, all members of the Executive Council of the ASIL, were asked how they would analyse ‘the recent conflict between Israel and Hezbollah in terms of the jus ad bellum and jus in bello rules requiring necessity and proportionality’.2 One of the respondents thought that in jus ad bellum proportionality played a part only in precluding ‘the legality of one state’s destroying or wholly occupying another by reason of a real or imagined minor infraction, such as a trivial border raid’.3 This did ‘not preclude Israel, under the circumstances, from attacking Hezbollah wherever it is to be found (including not only Lebanon but also Syria) in response to its string of attacks and incursions against Israel and its armed forces’.4 Another respondent opined that the question was whether ‘Israel’s response to the capture of two Israeli soldiers and the killing of eight’ was proportionate. His answer was that Israel’s response ‘wasn’t even close’.5 A third respondent considered that the question was whether the force used was ‘required to deter and protect against further attacks’. Applying this test he concluded that Israel’s actions were clearly legitimate acts of self-defence that met the demands of proportionality.6
All are agreed that the proportionality principle plays a central role both in jus in bello and jus ad bellum. In jus in bello the meaning of the principle itself is quite clear; it involves assessing whether the expected collateral damage to civilians and civilian objects of an attack on a legitimate military target is excessive in relation to the concrete and direct military advantage anticipated.7 It is admittedly notoriously difficult to apply this test,8 but the difficulty lies in evaluating and comparing factors that are not quantifiable rather than in the meaning of the principle itself. On the other had, as the above exchange so clearly reveals, in jus ad bellum the very meaning of the principle is shrouded in uncertainty.
Under just war theories proportionality has long been regarded as one of the elements for determining whether resort to war is justified.9 In traditional just war theory proportionality was based on an assessment of the anticipated goods of waging war in relation to its harms.10 The views of the international lawyers cited above reflect the confusion amongst contemporary moral philosophers and international lawyers, who, while acknowledging that states resorting to force must comply with the demands of proportionality, display little, if any, agreement on what this implies. Some take the view that it is the very decision to resort to force that must meet a proportionality test.11 Others adopt what has been termed a ‘tit for tat’ approach, under which the amount of force used by a state as a countermeasure against B must be proportionate to the force previously used by B.12 Yet others, probably the majority, argue that proportionality must be judged against the legitimate ends of using force13 or in relation to the threat.14 Antonio Cassese argues that the force used must be judged both against the legitimate ends and the attack against which it is responding.15 Finally, Yoram Dinstein suggests that ‘[i]t is perhaps best to consider the demand for proportionality in the province of self-defence as a standard of reasonableness in response to force by counter-force’.16
The very right of states to use of force in international relations and the para meters of that right are obviously highly loaded questions. States that are themselves faced with armed attacks or threats of such attacks are inevitably going to have a different perspective from uninvolved states. The perspective of the latter is likely to change radically once they too are faced with an attack. The bias of involved states is self-evident; that of uninvolved states may be less so. Yet experience tends to show that uninvolved states and outside observers will often be highly selective in deciding whether use of force was both justified and proportionate. When force is used in situations in which they have sympathy for the victim state, and little or no sympathy for the state or group that provoked the use of force by that state, they are not likely to be critical of the force used, provided it is not obviously incompatible with jus in bello. However, when similar force is used by a state to which they are either unsympathetic or outwardly hostile, or when they actually identify with some or all of the goals of the state or group whose actions provoked the use of force, they are likely to condemn that use of force as disproportionate.
The bias that will affect the attitudes of both involved and uninvolved states and consequently of inter-state organizations is facilitated by the confusion amongst international lawyers over the meaning of proportionality in the jus ad bellum context. While conceding the obvious political dimensions of the issue, my object in this article is to explore that meaning and hopefully thereby to reduce the confusion.
My analysis rests on the following assumptions and propositions:
a. The term ‘proportionality’ is used in various legal contexts. But it is used to mean two radically different things. Sometimes the term refers to the relationship between an act and the legitimate response to that act (‘just desserts’, ‘eye for an eye’, or ‘tit for tat’ proportionality). The response must be proportionate to the act that provoked it. This is the way the term is used in judging criminal sanctions: the punishment must fit the crime.17 In other contexts, proportionality relates to an assessment of the harm caused by means used to further legitimate ends (‘means-ends proportionality’). That harm must not be disproportionate to the expected benefits of achieving those ends. In human rights law, for example, proportionality judges the harm caused by restrictions on a protected liberty when weighed against the legitimate ends those restrictions are meant to serve.18
b. In jus ad bellum proportionality has traditionally been used in both of the above senses.19 When judging armed reprisals that were once regarded as legitimate the accepted meaning referred to weighing the force used in the reprisal against the unlawful act that provoked the reprisal.20 On the other hand, when placed in the context of a state defending itself against an armed attack, proportionality relates to whether the force used (the means) is proportionate to the legitimate ends of using that force (self-defence).21
c. In the post-Charter era unilateral use of force by states is limited to the exercise of their inherent right to self-defence, recognized in Article 51 of the UN Charter. Proportionality should therefore seemingly be based on an assessment of the force used in relation to that end.22 The problem is that there is no consensus on what that end may be. All accept that a state acting in self-defence may halt and repel an ongoing armed attack, but there is a singular lack of agreement on whether it may also act to prevent or deter further armed attacks from the same enemy. What ends are legitimate becomes especially acute when the response in self-defence takes place after the attack has been carried out and completed, and there is no longer an attack to halt or repel, or when the armed attack has not yet occurred but is imminent.
d. The legality of force used in self-defence depends, inter alia, on necessity and proportionality.23 Necessity is generally taken to refer to the resort to force, rather than to non-forcible measures, while proportionality assesses the force used.24 However, the term ‘necessity’ is also used to assess whether the force used was necessary to achieve legitimate ends of self-defence. When used in this sense there is an obvious affinity between necessity and proportionality. Means can only be proportionate when they are necessary to achieve the legitimate ends.25
e. The first stage in assessing proportionality is to define the legitimate ends of using force in the specific case. The second stage will involve assessing whether the forcible means used by the state acting in self-defence were necessary to achieve those ends. In this context, ‘necessary’ may have one of two meanings: that there is a rational connection between the means and the ends, or that there were no less drastic means available to achieve those ends.
f. In other contexts in which proportionality is assessed according to a means-ends test, after establishing whether the means used were necessary in both of the above senses, they are then subjected to a ‘narrow proportionality’ test.26 This requires assessing whether the harm caused by those necessary means outweighs the expected benefits. There is a difference of opinion whether this narrow proportionality test is relevant in jus ad bellum. Many experts assume that it is only relevant in jus in bello. If we adopt their view, proportionality in jus ad bellum boils down to assessing whether the forcible means used were necessary in light of the legitimate ends of self-defence in the particular case. Differences of opinion on whether force was proportionate reflect disagreement on the legitimate ends of force in the case under discussion.
g. All actions of a state that exercises its right to use force in self-defence must comply with jus in bello.27 Hence all its actions must be compatible with the principle of proportionality in IHL. The demand for proportionality in jus ad bellum is an independent demand that is divorced from the question whether the defending state complies with jus in bello. Use of force may be disproportionate under jus ad bellum even if all forcible measures are compatible with jus in bello in general, and the proportionality principle in jus in bello in particular.
h. Proportionality arises in this, and other, contexts only when the aim or ends pursued are legitimate. When it comes to state liability, if those ends are illegitimate all forcible measures used will ipse facto be illegitimate, whether they are proportionate or not. Thus, for example, the Iraqi invasion of Kuwait in 1990 involved unlawful use of force and no question of proportionality in jus ad bellum arose. Consequently the UN Security Council ‘reaffirmed that Iraq is liable under international law for any direct loss, damage, or injury to individuals, governments, and business organizations resulting from Iraq’s invasion and occupation of Kuwait starting on Aug. 2, 1990’.28
My argument in this article is that the legitimate ends of using force in self-defence may differ, depending, inter alia, on the nature and scale of the armed attack, the identity of those who carried it out, and the preceding relationship between the aggressors and the victim state. Proportionality will usually, but not exclusively, involve a means-end test. Whether such a test is employed and, if it is, whether force used will be regarded as proportionate will both depend on the legitimate ends of force in the concrete case.29
Assessing proportionality requires exploring the scope of the right of a state to use force in self-defence. The second section of this article is devoted to exploration of this question. In the third section I examine various theories regarding the legitimate ends of using force in self-defence. In the final section I draw the conclusions regarding the test of proportionality.
2 Use of Force in International Law
A Charter Principles
To what extent international law prohibited unilateral use of force in the pre-Charter era of the 20th century is a matter of debate.30 Whatever the position may have been before the adoption of the UN Charter, the principles on use of force since the adoption of the Charter have been fairly clear. The Charter set out to ban war between states. To achieve this aim it adopted a policy of collective security to be guaranteed by the Security Council, which is responsible for maintaining and restoring international peace and security. Under Article 2(4) of the Charter states are prohibited from ‘the threat or use of force against the territorial integrity or political independence of any state, or in any other manner inconsistent with the Purposes of the United Nations’. The only concession made in the Charter for the unilateral use of force by states is the recognition in Article 51 of their inherent right of individual or collective self-defence when an armed attack occurs, and even then only ‘until the Security Council has taken measures necessary to maintain international peace and security’. The Charter prohibition on unilateral use of force and its exception in the case of self-defence against an armed attack are regarded as part of customary international law,31 and have the status of jus cogens.32
Restricting the right of a state to use force to the case of self-defence would seem to be uncontroversial. The problems with, criticisms of, and attempts to modify the Charter regime do not relate to the principle itself, but to the parameters of the right to use force in self-defence defined in Article 51. Discussion of proportionality in the use of force in self-defence requires consideration of these parameters. While there have been experts who have argued that states may have a right to use force in self-defence which does not meet the requirement of Article 51, their views have not gained much credence.33 My assumption in this article shall be that the only basis for unilateral use of force against another state or against non-state actors in its territory is self-defence under Article 51 of the Charter.34
B The Boundaries of Article 51
Restricting the right of a state to resort to force in self-defence solely to the confines of Article 51 of the Charter requires us to examine these confines. It is accepted that exercise of the right to self-defence must meet the demands of immediacy, necessity, and proportionality. 35
Controversy has arisen over a number of issues that relate both to the interpretation of Article 51 and to the question whether state practice and the attitude of the UN Security Council to the use of force have led to a widening of the right to resort to force not contemplated in the text itself.36 I shall briefly discuss three controversial issues that are relevant to the discussion of proportionality: (1) the scale of force required for the use of force to be considered an armed attack; (2) whether an attack by non-state actors may be the kind of attack contemplated by Article 51; (3) whether preemptive use of force may ever be lawful.
1 Armed Attack – The Scale and Effects of Force Required
The definition of the term ‘armed attack’ that appears in Article 51 must take into account a number of factors: the scale of the force; the target of the attack; the identity of the attacker; the military nature of the attack; and the attribution of the attack to the state against which force in self-defence is to be employed.37 In this section I confine myself to the scale of force.
There is an obvious disparity between the language of Articles 2(4) and 51 of the Charter. While the former speaks of ‘the threat or use of force’ (‘à la menace ou à l’emploi de la force’ in the French version), the latter refers to ‘an armed attack’ (‘une agression armée’ in the French version). The clear implication is that while every use of force against the territorial integrity or political independence of another state is prohibited, not every such use of force will constitute an armed attack.38 In the Nicaragua case, the ICJ stated that ‘it will be necessary to distinguish the most grave forms of the use of force (those constituting an armed attack) from other less grave forms’.39 As an example of use of force that would not be of the ‘scale and effects’ to warrant being termed an armed attack the Court mentioned ‘a mere frontier’ incident.40 The Eritrea–Ethiopia Claims Commission opined that ‘[l]ocalized border encounters between small infantry units, even those involving the loss of life, do not constitute an armed attack for purposes of the Charter’.41
The demand for the force used to meet a threshold of ‘scale and effects’, or gravity of harm in order for it to be regarded as an armed attack for the purposes of Article 51 has not been universally accepted. Yoram Dinstein concedes that not every use of force will amount to an armed attack, but rejects the view that frontier incidents are not necessarily armed attacks. He argues that the gap between ‘use of force’ under Article 2(4) and ‘an armed attack’ under Article 51 ought to be quite narrow. In his view any use of force causing human casualties or serious damage to property constitutes an armed attack. The authors of the Chatham House Principles of International Law on Use of Force in Self-Defence take the view that ‘[a]n armed attack means any use of armed force, and does not need to cross some threshold of intensity’.42 This has also been the view taken by the US.43
While there is a lack of consensus on whether the force has to meet a threshold of intensity or not, and what that threshold should be if required,44 even according to those who demand a threshold it is not very high.45 On the contrary, excluded are only ‘mere frontier incidents’ or, at the very most, ‘localized border encounters between small infantry units’. Hence an armed attack that serves as the trigger for exercise by the victim state of its right to use force in self-defence may range from a fairly restricted use of force, such as a border raid causing limited loss or damage, to a full-scale invasion of its territory.
In justifying the use of force in response to cross-border attacks or bombings Israel has consistently relied on the ‘accumulation of events theory’.46 So have other states faced with a series of low-scale attacks.47 This theory has implications not only for deciding whether an armed attack has taken place at all, but whether the victim state may defend itself not only against the use of force which triggered its forcible response in self-defence, but against the threat arising from the whole series.48
The ‘accumulation of events theory’ has received a cold reception in the UN Security Council.49 On the other hand, while the International Court of Justice has never expressly endorsed the theory, preferring always to judge whether specific attacks amounted to an armed attack, in the Oil Platforms case it used language that suggested that the cumulative nature of a series of forcible actions could possibly turn them into an ‘armed attack’.50
The accumulation of events theory has not gained general acceptance in the international community. There are, however, signs that with the growing awareness that transnational terrorist attacks present states with a serious problem, it is not as widely rejected as it was in the past.51 As Christian Tams puts it, ‘states seem to have shown a new willingness to accept the “accumulation of events” doctrine which previously had received little support’.52
2 Attacks by Non-state Actors
May an attack by non-state actors be the kind of armed attack that allows a state to exercise its right of self-defence by using armed force that would ordinarily be regarded as a violation of Article 2(4)? If the attack may be imputed to a state, the answer is obviously positive.53 But what is the situation if the attack cannot be imputed to a state, or at least not to the state from whose territory those non-state actors are operating and in which the victim state wishes to use force? There is hardly an issue that has given rise to more controversy than this. It raises the question whether the rules of international law could be such as to prevent states from taking action necessary to defend their citizens and residents from attacks, merely because no state is responsible for them.
In its Advisory Opinion on the Legal Consequences of Construction of a Wall, the International Court of Justice opined that Israel could not defend the legality of the separation barrier it was building on the West Bank on the basis of its right to self-defence under Article 51 since it did not claim that the attacks which the barrier was designed to prevent were imputable to another state.54 The implied assumption was that under Article 51 only an armed attack by a state triggers the right to use force in self-defence.55 Three judges on the Court saw fit to disassociate themselves from this view,56 which has been subjected to scathing academic criticism.57
The view that an attack by non-state actors that is not imputable to a state cannot constitute an ‘armed attack’ has been rejected by the vast majority of publicists,58 who base rejection of this view on a variety of arguments: that Article 51 does not refer to an armed attack by a state; that the inherent right to self-defence to which Article 51 refers has always included the right to use force in defence against an attack by non-state actors; that Security Council Resolutions 1368 and 1373, passed after the 9/11 attack on the US by non-state actors, refers within the context of this attack to the inherent right of individual and collective self-defence in accordance with the Charter; that state practice confirms the right to use force in international waters and in the territory of another state in defence against an attack by non-state actors; and that denying that an attack on a state by a group of non-state actors may ever be an armed attack under Article 51 places the victim state in an impossible position and is therefore unrealistic. States have a duty and a right to protect their citizens and residents from attack. When the persons planning and executing attacks are operating from outside its borders, if neither the state from which those non-state actors operate nor the international community takes effective measures to stop the attacks the victim state must have the right to use force in self-defence. The support, either passive or active, given by large sections of the international community to major instances of a state using force against non-state actors in the territory of another state in such circumstances lends support to the argument that state practice today accepts that this right exists.59
Notwithstanding these arguments, some scholars still stick to the view that Article 51 refers solely to armed attacks by states. Their main argument is that the prohibition on use of force in Article 2(4) of the Charter refers expressly to the use of force by one state against the territorial integrity or political independence of another. As Article 51 is an exception to this prohibition, it obviously refers only to the kind of force that is the subject of the prohibition, namely force by one state against another.60
It should be stressed that the debate does not refer to the question whether a state attacked by non-state actors may take measures to defend itself. Obviously it may. Rather the question is whether those measures may include the use of force against the non-state actors in the territory of another state that cannot be held responsible for the attack by the non-state actors.
Notwithstanding the statements of the ICJ in Legal Consequences of the Construction of a Wall, and the analysis of some experts,61 for the purposes of my discussion of proportionality I shall accept the view of the majority of scholars that an attack by a group of non-state actors may constitute an armed attack. If the attack may be imputed to the state from which the non-state actors are operating, the victim state may use force in self-defence both against the host state and against the non-state actors present in that state’s territory. If, on the other hand, the host state cannot be held responsible for the attack by the non-state actors on the victim state, but does not take effective action to prevent the non-state actors from carrying out their attacks, the victim state must restrict its use of force to the non-state actors themselves.62 What proportionality requires in this case will obviously depend on which form of self-defence is being exercised: self-defence against the state and the non-state actors in its territory, or self-defence solely against the non-state actors.63
3 Pre-emptive Use of Force
Article 51 of the Charter expressly limits the right to use self-defence to the case in which an armed attack occurs, thus implying not only that states may not use force to prevent or deter a future attack, but that they may not even use force to thwart an imminent attack which, if it takes place, could have catastrophic consequences for the victim state or afford the attacking state a significant military advantage.64 There is evidence that in drawing up the Charter many states assumed that the inherent right to self-defence includes the right to use force against an imminent attack if the conditions of the Caroline test are met,65 namely that a state is faced with the threat of an armed attack which presents a ‘necessity of self-defense, instant, overwhelming, leaving no choice of means and no moment of deliberation’.66 Nevertheless, it must be conceded that the literal reading of Article 51 favoured by some scholars is not devoid of logic, and may be regarded as consistent with the collective security policy adopted in the Charter.67 Under this policy a state facing an imminent attack should not pre-empt the attack by using force but should run to the Security Council for help. Furthermore, the actual occurrence of an armed attack is usually evident and leaves little room for abuse by states, whereas the fear of an imminent attack rests by its very nature on subjective assessments of likely developments in the future.68 Allowing states to act on the basis of such assessment would open up the right to use force in self-defence to abuse.
The arguments for recognizing a right to anticipatory use of force in the face of an imminent attack rely not only, or not mainly, on the meaning of the ‘inherent right to self-defence’ recognized in Article 51, nor on the drafting history of this provision, but on the reality both of modern warfare and international politics.69 With the development of new weapons of mass destruction it would border on the perverse to maintain that a state facing an imminent attack by an enemy armed with such weapons would have to sit by idly and wait for the attack to start before it could defend itself.70 Similar arguments have been made in light of the threat from terrorist attacks.71 Given the rather poor record of the Security Council in preventing acts of aggression, demanding that a state rely on the Council to protect it is not likely to be an appealing option for a state that is convinced that it is threatened with an imminent attack that could have catastrophic consequences.
Not surprisingly, therefore, while it is almost universally accepted that a state may not use force in order to prevent or deter future attacks,72 it is widely (but certainly not universally) acknowledged that it may do so to thwart an imminent attack. The High Level Panel of Experts appointed by UN Secretary-General Kofi Annan to examine UN reform, after having cited the ‘restrictive’ language of Article 51, took a clear stand on this issue when it wrote:
However, a threatened State, according to long established international law, can take military action as long as the threatened attack is imminent, no other means would deflect it and the action is proportionate.73
The Secretary-General adopted the same position in a report he published a year after the release of the High Level Panel’s report.74 Many experts accept the High Level Panel’s view as reflective of the inherent right to self-defence recognized in Article 51.75
A few experts, bothered by the apparent contradiction between the express demand in Article 51 that ‘an armed attack occurs’ and recognition of any right to use force before it has occurred, have sought alternative explanations for recognizing the right to use force in the face of an imminent attack. Thomas Franck argued that state practice, as reflected in the decisions of the UN Security Council and other UN organs, altered the original confines of Article 51 by acknowledging response to a clearly imminent attack as equivalent to response to an armed attack that has occurred.76 Dinstein accepts that Article 51 excludes any use of force until an armed attack has actually occurred, but widens the time-frame for deciding when such an attack begins by equating a situation in which one state ‘embarks on an irreversible course of action’ to an actual attack on the victim state.77
The question whether a state may respond forcibly to thwart an imminent attack is clearly relevant when addressing the issue of proportionality and the related issue of necessity. In his report on state responsibility Roberto Ago mentioned that the duty to try non-forcible measures before resorting to force (generally thought of as the question of necessity, but regarded by some as an issue of proportionality) is especially relevant if force may be used before the armed attack has actually started.78 Assessing proportionality between the force used and an attack that might (and might not) have taken place if the force had not been used poses tricky questions.
Notwithstanding the arguments presented by purists who maintain that Article 51 means what it says, and only what it says,79 my assumption in discussing proportionality will be that the view of the majority of experts is the preferred position. When a state is faced with the threat of an armed attack that creates a ‘necessity of self-defense, instant, overwhelming, leaving no choice of means and no moment of deliberation’80 it may use force to remove the threat. The international community will have to review the matter after the event and offer a ‘second opinion’ on whether that state’s assessment that an attack was imminent was well-founded.81
3 The Meaning of Self-defence
The assumption of Article 51 is that when an armed attack actually occurs the victim state obviously cannot be expected to remain passive while it waits for the Security Council to act. It may therefore use force in self-defence as an interim measure until the Security Council takes measures to restore peace and security.82 The primary object of using force in self-defence in such a situation seems pretty clear: it is to halt and repel the attack. If the enemy has invaded the victim state’s territory self-defence may include using force to expel it.
But does the matter stop there? What if the attack is over before the victim state is able to respond? Assuming that the aggressor has not invaded the victim state’s territory, or even if it has that it no longer retains any presence there, what exactly is the victim state defending itself against? Any use of force by that state will not be to defend itself against the armed attack that occurred, but against the threat of further attacks by the state or non-state actor which carried out the armed attack. The object of using force in such a case may be preventive, punitive, deterrent, or a combination of these. In some cases use of force by the victim state may be based on a credible assessment that another attack is imminent; in others there may not be any credible evidence on which such an assessment could be based, and to the extent that the use of force is forward-looking, rather than purely punitive, it may appear to be an example of preventive use of force.
In this section I review the fundamental assumptions behind the right to self-defence recognized in Article 51, and the conclusions that states and international bodies have drawn from these assumptions. As it is widely accepted that armed reprisals are prohibited under the Charter regime, in an attempt to draw conclusions regarding the legitimate aims of self-defence I shall first discuss the legality of armed reprisals and the distinction between such reprisals and actions in self-defence. Following this discussion I shall examine conceptions of the aims and scope of force that may be used in self-defence when an armed attack occurs or is imminent.
A Armed Reprisals
In the pre-Charter era armed reprisals were accepted as a lawful response to the use of force against a state.83 The essence of such reprisals is that they involve the use of force that would normally be regarded as unlawful, but is permitted when carried out in response to an unlawful act by another state in an attempt to achieve redress by compensation and/or to prevent or deter repetition of the unlawful act in the future. In the famous Naulilaa arbitration between Portugal and Germany that followed an attack by Germany on Portuguese territory in Africa, the arbitrators held that in order to be regarded as lawful an armed reprisal must meet three conditions: an unlawful act by the state that is the object of the reprisal; a prior non-forcible unsatisfied demand for redress; and proportionality of the reprisal to the unlawful act which provoked it.84
In the post-Charter era the rhetoric of unaffected states, international bodies, and experts in international law stresses that armed reprisals are incompatible with Article 51 and are therefore unlawful.85 This rhetoric appears in numerous resolutions of the Security Council and General Assembly, the commentaries of the International Law Commission, and the writings of a host of highly qualified publicists.86 In a 1964 Resolution relating to a complaint by Yemen regarding a British air attack on Yemeni territory, the Security Council declared that it ‘[c]ondemns reprisals as incompatible with the purposes and principles of the United Nations’.87 The Security Council included an identical condemnation in resolutions relating to the use of force by Israel in the territory of Arab countries.88 In its 1970 Declaration on Principles of International Law Concerning Friendly Relations the General Assembly declared that ‘States have a duty to refrain from acts of reprisal involving the use of force’.89 The General Assembly has included a similar statement in other resolutions.90 In its Articles on State Responsibility approved by the General Assembly, the International Law Commission states that countermeasures incompatible with the rules on use of force under the UN Charter are outlawed.91 This in itself does not necessarily imply that armed reprisals are prohibited.92 However, in its commentaries on the Draft Articles the Commission cites the above statement in the General Assembly Declaration on Friendly Relations with approval, and opines that it is consistent ‘with the prevailing doctrine as well as a number of authoritative pronouncements of international judicial and other bodies’.93
Despite the wide degree of consensus that armed reprisals are unlawful, such reprisals have not always met with disapproval and condemnation by the international community and international organs, especially the Security Council. In an article published in 1972 Derek Bowett attempted to show that there was a clear disparity between the rhetoric condemning reprisals in principle and the failure of the Security Council to condemn all specific reprisals as unlawful ipse se, preferring on occasion to stress that disproportionate force had been used, that the provocation had not been severe enough to justify the use of force, or that the use of force had been premeditated rather than a response to an unlawful act.94 Bowett regarded the total exclusion of reprisals as unrealistic, and suggested conditions under which an armed reprisal should be regarded as legitimate. In a 1990 follow-up to Bowett’s seminal article, William V. O’Brien showed that since Bowett’s article had been published the Security Council had continued to condemn Israel’s use of force against targets of the PLO and other organizations in Lebanon and other Arab countries as unlawful reprisals.95 O’Brien also called for reconsideration of the prohibition on reprisals when it comes to counter-terror measures.96 Since then the debate on the desirability of the prohibition of armed reprisals has continued.97
Although armed reprisals are regarded as unlawful, it is far from clear what the difference is between such reprisals and legitimate forcible acts of self-defence. Hence it is not that clear why armed reprisals are regarded as prohibited under the Charter regime.98 In the various Security Council debates on use of force that was condemned as unlawful, often with a sentence included in the Resolution adopted recalling the prohibition on armed reprisals, states participating in the debates gave different reasons for their objections to the particular use of force, and the Security Council itself did not take a position on the issue.99
There are three conceivable reasons why the use of force in the territory of another state might be regarded as an armed reprisal rather than a legitimate use of force:
a. Absence of an armed attack. One of the elements of an armed reprisal is that it is a response to a delinquent act by a state. However, even if that delinquent act involves use of force that is incompatible with Article 2(4), it does not necessarily amount to an armed attack which gives rise to the right to use force in self-defence. One of the possible grounds for regarding use of force as an illegitimate ‘armed reprisal’ is that the use of force which served as provocation for the armed action was not regarded as an armed attack that could be attributed to the state in whose territory the armed action was carried out. Following this line of thought Arend and Beck define a forcible reprisal as ‘a quick, limited, forcible response by one state against a prior action by another state that did not amount to the level of an armed attack’.100
The lack of an armed attack could be due to an assessment that the scale of force used did not reach the gravity required for an armed attack, that the state which was reacting to the force used against it was relying on the ‘accumulation of events’ theory that has been rejected by the Security Council, or that it was reacting to an attack by non-state actors which the Security Council was not prepared to impute to the state in whose territory the reprisal was conducted. One can find strong support for all of the above grounds in the SC debates on Israeli military operations in the territory of Arab countries where the issue of reprisals was often raised.101
b. The punitive nature of the action. This is the most widely cited basis for regarding the use of force as an armed reprisal rather than a legitimate act in self-defence.102 The notion here is that the aim of force used in self-defence must be protection against an ongoing or imminent attack, whereas the aim of a reprisal is to punish for the harm done, either out of a feeling of justice or outrage, or to deter the delinquent state from using force against the victim state in the future. As described by Jean Combacau, the state carrying out the reprisal attempts to dissuade the other state ‘by a “punitive” action, either from persisting [in the delinquent behaviour] or from reverting to it in the future; the aim is therefore entirely foreign to self-defence’.103 The assumption behind this view is, of course, that the primary goal of using force in self-defence must be to ward off an ongoing or imminent armed attack.104 I return to this below in the discussion of the distinction between the retributive and deterrent functions of punitive measures.
c. The timing of the action. Reprisals are carried out after the delinquent act to which they are responding has been completed and they are unconnected with protection against that act. Obviously, this has some affinity with the punitive ration ale, since punitive action is by its nature carried out after the event that is said to justify it. However, while every punitive action is carried out after the event, action carried out after the event will not necessarily have a punitive aim. It may be aimed to prevent, rather than to deter, future attacks against the victim state. If such an attack is imminent the use of force may be regarded as defensive.
Many writers combine both the punitive and timing rationales as explanations for the incompatibility of armed reprisals with the notion of self-defence recognized under Article 51 of the Charter.105 This was essentially the position taken by Roberto Ago in the seminal paper he wrote for the International Law Commission on the issue of state responsibility. In explaining the difference between defensive action and reprisals Ago wrote:
‘Self-defence’ and ‘sanction’ are reactions relevant to different moments and, above all, are distinct in logic. Besides, action in a situation of self-defence is, as its name indicates, action taken by a State in order to defend its territorial integrity or its independence against violent attack; it is action whereby ‘defensive’ use of force is opposed to an ‘offensive’ use of compar able force, with the object – and this is the core of the matter – of preventing another’s wrongful action from proceeding, succeeding and achieving its purpose. Action taking the form of a sanction on the other hand involves the application ex post facto to the State committing the international wrong of one of the possible consequences that international law attaches to the commission of an act of this nature. The peculiarity of a sanction is that its object is essentially punitive or repressive; this punitive purpose may in its turn be exclusive and as such represent an objective per se, or else it may be accompanied by the intention to give a warning against a possible repetition of conduct like that which is being punished, or again it might constitute a means of exerting pressure in order to obtain compensation for a prejudice suffered.106
The significance of the grounds for the prohibition on armed reprisals for the definition of ‘self-defence’, and consequently for the test of proportionality in jus ad bellum, is self-evident. If the grounds for the prohibition are solely the lack of an armed attack that may be imputed to the state in whose territory the use of force takes place, when such an armed attack has taken place it may be lawful for the victim state to respond with a use of force the dominant aim of which is punitive, and which takes place after the armed attack has been completed. And indeed Yoram Dinstein, the one major publicist who supports a limited right to what he terms ‘defensive armed reprisals’, stresses that such reprisals can be justified only if they are in response to an armed attack.107
On the other hand, if the grounds for the prohibition on armed reprisals relate to their aim and/or timing, the assumption must be that the right to use force in self-defence is protective in nature. This fits well with the limitation in Article 51 that this right may be exercised only ‘until the Security Council has taken measures necessary to maintain international peace and security’, a reflection of the policy of collective security adopted by the Charter.108 For the purposes of proportionality, the implications are that it must be judged by the means-end rather than the just desserts test. This was indeed the view advanced by Roberto Ago, who, as we have seen, based the objections to reprisals on their punitive purpose and timing. Ago regarded it as ‘mistaken … to think that there must be proportionality between the conduct constituting the armed attack and the opposing conduct’.109
The differing perspectives on the distinction between illegitimate armed reprisals and lawful use of force in self-defence were clearly articulated in the pleadings of the parties in the Oil Platforms case. That case involved a claim by Iran against the United States following US attacks on Iranian oil platforms. The US argued that these attacks were actions in self-defence taken in response to a series of attacks on US and neutral shipping which the US attributed to Iran.110 More specifically the US argued that attacks on two ships amounted to armed attacks against the US by Iran. While the US denied that its attacks on the oil platforms had a punitive intent, it contested the Iranian argument that the use of force in self-defence is limited to the time the attack is in progress. It stated that the Iranian attacks on the ships had lasted only a few seconds and that the ‘status quo ante could not be restored simply by driving an attacking force back across the border from whence they came’.111 Mindful of the objection to the use of force the only object of which is punitive or deterrent, the US claimed that the oil platforms had been used to identify and target vessels for attack. In its reply Iran contested the US claim that it had been subjected to an armed attack, arguing not only that the US had not proved that the attacks on the US ships could be attributed to Iran, but that an armed attack had to be a specific attack and not a series of events. More importantly for the present discussion, Iran argued that ‘self-defence is limited to that use of force which is necessary to repel an attack’, and that ‘once an attack is over, as was the case here, there is no need to repel it, and any counter-force no longer constitutes self-defence. Instead it is an unlawful armed reprisal or a punitive action.’112 It added that use of force to deter future attacks is not included in the right to self-defence, but constitutes unlawful pre-emptive use of force.113
In deciding the case the Court did not expressly choose between the two opposing views on the nature of forcible action allowed in self-defence, nor did it opine on the difference between such action and an armed reprisal. Rather it held that the US had not lifted the burden of proof to show that the attacks on its ships were carried out by Iran. Hence any use of force by the US against Iranian installations was unlawful. The Court did intimate, however, that in judging the response in self-defence to a limited armed attack, such as an attack on one vessel, the test of proportionality would be one that assessed the scale of the response in relation to that of the armed attack.114
A number of attempts have been made in recent years to revive armed reprisals as a legitimate use of force in certain circumstances. I have already mentioned Dinstein’s view that ‘defensive armed reprisals’ may be a lawful response to a limited armed attack. Dinstein’s argument rests on his position that a rather low threshold of armed force is required for the use of force to qualify as an armed attack, and that in exercising its right to self-defence a range of possibilities is therefore available to the victim state. These include both ‘measures short of war’ and outright war.115 The former may be divided into limited on-the-spot reactions to a small-scale localized armed attack and ‘defensive armed reprisals’ that take place after a small-scale armed attack has been completed.116 Such reprisals may not be based on ‘purely punitive, non-defensive, motives’, although some element of retribution is likely to be present. Their dominant motive must be forward-looking, in the sense that they are based on an assessment that the aggressor is likely to repeat the attack in the future and that their object is to deter it from doing so. Dinstein stresses that all defensive armed reprisals must meet the demands of immediacy, necessity, and proportionality. As mentioned above, given their nature as deterrent measures, proportionality here is one of ‘just desserts’ and not of means-end. As Dinstein puts it:
the responding State must adapt the magnitude of its counter-measures to the ‘scale and effects’ of armed attack. A calculus of force, introducing some symmetry or approximation between the dimensions of the lawful counter-force and the original (unlawful) use of force, is imperative.117
While some other writers have also expressed support for the idea that armed reprisals may, in certain circumstances, be legitimate measures of self-defence under Article 51,118 Dinstein himself concedes that his view does not reflect that of the majority of scholars.119 The majority of writers take it as settled law that use of force following an armed attack that has been completed, and the only object of which is deterrent rather than protective, will amount to an armed reprisal that is not a legitimate exercise of the right to self-defence recognized under Article 51 of the Charter.120
While the prevailing view is clearly that armed reprisals are incompatible with the Charter principles on use of force, military operations which look very much like armed reprisals are still fairly prevalent. States that carry out such operations invariably present them as exercises of their right to self-defence following an armed attack. They are careful to stress that the objects of the operations are not punitive, or at least not predominantly so, but preventive. I have already mentioned the US arguments in the Oil Platforms case, in which the argument was that the platforms served as a base for military attacks. Another case in point is the attack by the US on targets in Libya, following a series of attacks on US targets abroad that culminated in a bomb attack on a Berlin discothèque frequented by US military personnel.121 While initial statements by President Reagan and senior officials gave the impression that the object of the attack had been retaliation and deterrence,122 when the matter reached the UN Security Council the US Permanent Representative was careful to stress only the preventive aspects of the attack. In summing up the US case the Permanent Representative stated:
In light of this reprehensible act of violence – only the latest in an ongoing pattern of attacks by Libya – and clear evidence that Libya is planning a multitude of future attacks, the United States was compelled to exercise its rights of self-defense.123
This attempt to underplay the punitive and deterrent motives of the US attack failed to convince many observers. A number of scholars regarded the attack as an unlawful armed reprisal.124 There was a similar reaction to the 1998 US attacks on targets in Sudan and Afghanistan following Al Qaeda attacks on US embassies in Africa.125
It is not only the involved states themselves who try to present armed reprisals as actions of preventive self-defence. In some cases the reaction of the international community towards armed reprisals has been surprisingly restrained and has failed to call the child by its name. Thus, for example, the US attack on the buildings of army intelligence in Baghdad in 1993, carried out in response to the attempted assassination of former President Bush that was attributed to the Iraqi intelligence, was not strongly condemned by the international community. It is difficult to regard this action as anything other than an armed reprisal the aims of which were largely retribution and deterrence.126
The disparity between the formal rules and the actual practice of states which are faced with limited attacks inevitably leads those states to deny that their armed actions are aimed at punishing or deterring the parties responsible for those attacks, and to claim that they are defensive in character. This lends credence to Dinstein’s view that ‘defensive armed reprisals’, which in his view have a clear deterrent motive, are in reality still, or again, an accepted form of responding to a limited armed attack. I shall argue below that while the use of force the motive of which is clearly retributive is still universally condemned as an unlawful armed reprisal, there is often a more tolerant attitude towards the use of force as a deterrent against further attacks by the parties responsible for the armed attack.
B Protective Self-defence
Under traditional laws of war, once a war had started each party could carry on fighting until victory (whatever that may mean) was achieved. Proportionality was relevant in assessing how one fought (jus in bello), but not whether one continued to fight until victory. Yoram Dinstein is the foremost proponent of the view that this doctrine is still valid.127 As we have seen above, Dinstein distinguishes between armed attacks and armed attacks. Minor or localized armed attacks will justify only on-the-spot responses or ‘defensive armed reprisals’. However, more serious attacks that either themselves amount to war, or are of such a critical nature as to demand a massive response, justify the state under attack engaging in or resorting to war. Under this view, proportionality comes into play only in judging whether resort to war in response to the armed attack is justified.128 If it is justified, proportionality (in the jus ad bellum sense) is no longer relevant. In Dinstein’s words, ‘Once a war of self-defence is legitimately started, whether as a counter-war or in response to an isolated armed attack, it can be fought to the finish (despite any ultimate lack of proportionality).’129 In making this statement Dinstein’s assumption seems to be that proportionality necessarily implies a ‘just desserts’ rather than a ‘means-end’ test, for as support for his approach he cites the view of Ago, quoted above, that in self-defence it is mistaken to think that the force used must be proportionate to the force of the armed attack.130
Dinstein may be correct in stating that when the armed attack itself takes the form of ‘war’, such as would be the case when there is a full-scale invasion by one state of the territory of another, the response of the victim state may be full-scale war which is fought until the enemy state is vanquished. It is much more problematical to take his approach when the armed attack is indeed fairly large-scale, but is isolated and has been completed before the victim state has had the chance to respond. It is also more problematical when the attack was carried out by a group of non-state actors. And these are the more common scenarios in the present world. The question must be what self-defence against such attacks implies and, on the basis of the answer to that question, what place proportionality has to play in assessing the response of the victim state. These were the questions that arose both in relation to Israel’s conflict with the Hezbollah in the summer of 2006 and its conflict with the Hamas in Gaza in 2008–2009. They would also seem to be the questions that arise when assessing the response to the 9/11 attack on the US.131
If one takes the text of Article 51 seriously, the right to use unilateral force in self-defence would appear to be a function of the ‘emergency situation’ in which a state finds itself. An armed attack occurs and the victim state must obviously have the right to defend itself against that attack until such time as the Security Council takes measures to secure international peace and security. This implies that the victim state can do anything compatible with jus in bello that is needed to halt and to repel the attack. However, if the attack has taken place and the damage has been done, and the attacker is no longer attacking nor present on the territory of the victim state, the victim state may have a valid complaint about a severe violation of international law, but in what sense can it defend itself against the armed attack, as opposed to the armed attacker?132 If it takes forcible measures in this situation it will be doing so for one or more of a number of reasons: to prevent the enemy from attacking again in the near or not-so-near future, to punish the enemy for its attack, or to deter it from attacking again.133 Which, if any, of these are legitimate aims for using force following an armed attack? Ostensibly all of them present difficulties. States do not have the right to use pre-emptive force unless they are faced with an imminent attack. If there is no evidence that the aggressor intends launching another attack imminently, should the fact that the victim state has been subjected to an armed attack allow it to use pre-emptive force? Punitive motives for the use of force are regarded as illegitimate and incompatible with the very notion of self-defence. If the response of the victim state is purely punitive it will be regarded (by most, at least) as an unlawful armed reprisal. And what about deterrence? Can one distinguish between it and punishment? In an age in which it is the duty of the Security Council to maintain and restore international peace and security surely deterrence should be in its hands? It has the powers to adopt both non-forcible and forcible measures to prevent and deter states from violating their international obligations. If the victim state is not faced with the threat of an imminent attack, surely the regime perceived under the Charter requires that it refrain from the unilateral use of force and that it place the matter in the lap of the international community?
C The Aims of Self-defence
On the basis of the fundamental assumptions of Article 51, that a state may exercise the right to self-defence when, and only when, an armed attack occurs or is imminent, what does the right to self-defence imply? One can discern a number of possible answers to this question:
1 Halting and Repelling the Attack
In his report on state responsibility Roberto Ago took issue with the ‘tit for tat’ test of proportionality and made it clear that proportionality was to be judged by the purpose of using force. Ago wrote:
The requirement of the proportionality of the action taken in self-defence, as we have said, concerns the relationship between that action and its purpose, namely – and this can never be repeated too often – that of halting and repelling the attack or even, in so far as preventive self-defence is recognized, of preventing it from occurring.134
In her separate opinion in the Nuclear Weapons case, Judge Rosalyn Higgins cited Ago’s view with approval.135
The notion that the object of using force in self-defence is restricted to halting and repelling the armed attack enjoys fairly wide support among academic writers. In her monograph on necessity and proportionality in the use of force, Judith Gardam points out quite correctly that the first step in examining proportionality is to determine the legitimate aim of self-defence. She distinguishes between response to an armed attack that has occurred and the use of force to thwart an imminent attack. In relation to the former she has this to say:
In the case of self-defence against an armed attack that has already occurred, it is the repulsing of the attack giving rise to the right that is the criterion against which the response is measured. Repulsion of the attack in this context encompasses not only resistance to an ongoing armed attack but the expulsion of an invader and the restoration of the territorial status quo ante bellum.136
Having said this Gardam concedes that it may make a difference whether the defence is against an isolated attack ‘or whether there is an ongoing state of armed conflict’.137 She also mentions that state practice and commentators differ as ‘to the extent to which the destruction of the enemy is justified in order to repulse the attack’, and she asks whether ‘the requirements of proportionality in self-defence under the Charter proscribe action to remove a continuing threat’.138 Gardam then poses the question that must be addressed by proponents of the ‘halting and repelling’ concept of self-defence: if this is indeed the object ‘is it proportionate to take action that is designed to prevent such an attack occurring again and restore the security of the State?’139 This is not really a question of proportionality, but one that goes back to the very aim of using force in self-defence. And, as Gardam herself admits, one cannot discuss proportionality without defining the legitimate aims of self-defence. Acceptance that a state that has been attacked may take action to prevent such an attack occurring again necessarily rejects the ‘halting and repelling’ theory.
The ‘halting and repelling’ theory is supported by Cassese,140 Corten,141 and Cannizzaro.142 While Cassese seems to accept that this means what it says, namely that the Charter norms and the corresponding norms of general international law ‘do not authorize or condone any military action above mere opposition to, and repelling of, aggression’,143 like Gardam, Corten displays an ambivalence in accepting that the victim state can go no further than that. He sticks to his guns in supporting the ‘halting and repelling’ test, but concedes that in some cases a state may be entitled to respond to an attack that has been completed, such as a missile attack on a state. Arguing otherwise, in Corten’s view, ‘would be a plainly absurd and unreasonable interpretation in respect of the very objective of self-defence’.144 Unfortunately he does not offer an explanation of what the objective of self-defence becomes once the attack has been completed and can no longer be halted or repelled. Cannizzaro is emphatic that the use of force ‘must necessarily be commensurate with the concrete need to repel the current attack, and not with the need to produce the level of security sought by the attacked state’.145 In his view ‘the forcible removal of threatening situations and the creation of permanent conditions of security seem to have been reserved by the international community as tasks to be performed collectively’.146
The claim that the sole aim of using force in self-defence is to halt and repel the armed attack that triggered the right to use force seems removed from reality. It may reflect the rhetoric of states that have not themselves been subject to a fairly large-scale attack or series of attacks, but it certainly does not reflect the practice of states that have been the victims of such attacks. The failure of the international community to provide effective protection for states that have been subject to repeated attacks is hardly an encouragement to those states to rely on that community to remove the threatening conditions and provide them with permanent conditions of security.
Support for the ‘halting and repelling theory’ is strongest amongst those, such as Cassese, Corten, and Cannizzaro, who are strong supporters of the philosophy behind Article 51 of the UN Charter, for it is the theory that is most compatible with that philosophy. Licence is given to states to use only the minimum force necessary to defend themselves against the armed attack that has occurred until such time as the Security Council fulfils its function of restoring and maintaining international peace and security.
Unfortunately, as many commentators have pointed out time and again, the Charter philosophy has not been implemented in practice.147 The UN SC has not proved itself to be capable of restoring and maintaining international peace and security and providing effective protection for vulnerable states, especially when those states are not popular amongst UN member states, or when one or more of the permanent members of the UN SC sides with the aggressor state. In these conditions it does not seem reasonable to demand that the victim state restrict its response to halting and repelling the attack, even when it has well-founded fears that the aggressor may well mount another attack in the future. Furthermore, the halting and repelling theory stacks the cards in favour of the aggressor. The most the aggressor stands to lose is that its attack will be halted and repelled. If the armed attack is successful it will have attained its goals; if not, it will merely be returned to the ante bellum situation. It may then prepare for another attack at a time of its choosing.
2 The Trigger Theory
Proponents of the ‘trigger theory’ accept that states may not use force unless they have been subject to an armed attack Once such an attack has occurred, however, the victim state may defend itself not only against that attack, but against threats posed by the aggressor, whether imminent or not. In this situation the victim state may use a show of force in order to deter the aggressor from repeating its attack in the future or to destroy the military potential of that state so that it will not be able to mount another attack in the near future.148 This theory has an obvious affinity with Dinstein’s view, presented above, that when a state has been subjected to a large-scale armed attack it may resort to war that may be pursued until the victim state has achieved its aims. The difference between Dinstein’s theory and the trigger theory is that the latter does not rest on a distinction between different levels of armed attack. The original proponent of the idea that Article 51 is based on the trigger theory maintained that the right of self-defence in that Article is the right to resort to war, and that this right may be exercised as long as an armed attack has occurred, whatever its scale.149
The trigger theory is not likely to appeal to those who are attached to the collective security philosophy behind the Charter. It widens the scope for the legitimate use of force by states, rather than containing it to the minimum necessary to allow a state to defend itself until the UN SC fulfils its duty to restore international peace and security. Proponents of the theory may reply that the collective security philosophy behind the Charter is reflected in the powers of the SC under Chapter VII to take measures to maintain international peace and security.150 In exercising these powers the SC may impose a cease-fire between the parties or subject them to other demands, such as withdrawal from territory taken by force.151 Such a decision will bind the states involved (although, if the conflict involves a non-state actor such as Al Qaeda, it will not necessarily bind all the parties involved in the conflict).
When it comes to inter-state use of force, it is generally accepted that the threshold for an armed attack is not high. Only excluded, according to the ICJ, are mere frontier incidents, and even this exclusion has been criticized. It does not seem reasonable that even a low-level armed attack should by itself allow the attacked state to respond by destroying the military capacity of the aggressor. Thus Dinstein’s distinction between different levels of armed attacks and the response each level justifies has more appeal and logic than the ‘pure’ trigger theory. The trigger for an all-out military campaign to destroy the military potential of the aggressor and to deter it from further attacks in the future may only be a large-scale attack, or at least a series of attacks culminating in the attack which is the final straw that triggers the response. As we have seen, the SC and uninvolved states have been highly reluctant to accept an ‘accumulation of events’ test for an armed attack. There are some signs, however, that when an event which constitutes an armed attack has occurred they may be willing to consider prior events in assessing the legitimacy of the force used in response to the attack.152
The trigger theory may not appear to be the kind of theory that serves the purpose of reducing the scope of armed conflicts. It is impossible to draw a clear line between armed attacks that justify a limited response and those that trigger the right to use massive force to destroy the enemy. States may use a fairly low level attack as an excuse to pursue aims that are unconnected with that attack. It would seem that one of the grounds for the general perception that Israel’s use of force in response to the Hezbollah attack of 12 July 2006 was disproportionate was exactly the feeling that that attack served merely as a pretext for Israel to pursue aims that had been defined beforehand.153 On the pro- side, the trigger theory sends a clear message to potential aggressors that they will not be able to determine the level of force used in response to an armed attack. The knowledge that any armed attack on another state could trigger a massive response could potentially serve as a deterrent against launching an attack.154
Under the trigger theory, once a large-scale armed attack has been launched jus ad bellum proportionality no longer plays a part. The victim state is constrained by the norms of jus in bello and possibly by the notion of military necessity, namely that it may use only such force as is necessary to achieve its military objectives.155 But given the acceptance that those objectives may be extremely wide, it may be difficult to regard as unnecessary a concrete military action that is compatible with the rules of jus in bello.
3 The Future Attack Theory
As stressed above, it is widely accepted that a state may use force to thwart an imminent attack. It would seem illogical to argue that if a state has not yet been attacked it may use force to thwart an imminent armed attack, but that if it has already been attacked it may not do so. Even some scholars who reject in principle the notion that a state that has not been subject to an armed attack may mount an anticipatory attack accept that if a state has been attacked it may use force to prevent further attacks.156
Is a state that has been attacked in the same situation as a state that has not yet been attacked and wishes to pre-empt a future attack? Obviously, as long as the fighting with the aggressor continues there is not much point in trying to find a line between imminent and non-imminent future attacks. But the question has arisen in recent years in relation to 9/11 type attacks, which are over and done with before the victim state has had a chance to respond. Some commentators maintain that even in such a situation the victim state may use force only if it has evidence that further attacks are imminent.157 Others accept that, having been attacked, the victim state may use force not only to repel that attack but to prevent future attacks too, without requiring that such attacks be imminent.158 Michael Schmitt speaks of a response which is ‘no more than necessary to defeat the armed attack and remove the threat of reasonably foreseeable future attacks’.159
Defence against future attacks seems to be the fairly standard argument advanced by states in justifying their response to an armed attack that has been completed. Sometimes states expressly refer to imminent attacks. More often, however, they simply refer to future planned attacks. Examples are the statement of the US representative in the SC debate on the US attacks on targets in Libya and the arguments of the US before the International Court of Justice in the Oil Platforms case. This was also the argument made by both the US160 and the UK161 when justifying their resort to force against Afghanistan following the attacks of 9/11. Aligned with this approach is the reliance of states on a pattern of attacks as evidence that further use of force against them in the future is anticipated.162 As I noted above, while non-involved states and UN bodies have been reluctant to accept a definition of an armed attack that is based on the ‘accumulation of events’ theory, once there has been a single attack that reaches the scale and effects required for an armed attack, they seem more open to accepting that in judging the forcible response previous uses of force against the victim state should be relevant.163 This view was originally presented by Roberto Ago in his report on state responsibility, in which he wrote:
If … a state suffers a series of successive and different acts of armed attack from another state, the requirement of proportionality will certainly not mean that the victim state is not free to undertake a single armed action on a much larger scale in order to put an end to this escalating succession of attacks.’164
Allowing states to act to prevent or deter future attacks, whether imminent or not, would seem to give them wide leeway, and in some respects resemble the trigger theory. Maybe it is this fear of opening the door to massive use of deterrent force in the face of an amorphous fear of future attacks that led one well-respected expert to interpret the above US and UK explanations for their military campaign against Al Qaida and the Taliban in Afghanistan as referring to imminent attacks even though the explanations themselves do not include this description.165
Obviously it will usually be extremely difficult, if not impossible, to gauge whether attacks, or further attacks, are really imminent. Restricting the right of states that have been attacked to halt and repel the attack and, if the attack has been repelled or has been completed before the victim state may respond, to prevent imminent attacks therefore has an air of artificiality about it. It might change the rhetoric of states so that their standard justification for use of force in response to a completed attack would be that future attacks were imminent, but it would be unlikely to change their practice. While a state that has not been attacked is likely to meet a sceptical reaction when it claims that it used force to stop an imminent attack, states that have already been subject to an armed attack, especially one nearing the scale and effects of the 9/11 attack, would probably find other states and international organs more understanding of their claim that they were acting to prevent future attacks.166 Michael Schmitt argues that the same test of imminence will apply whether the state has been attacked or not, but the mere fact that a state has been attacked already will make it easier to conclude that it will be attacked again. Furthermore, ‘it may also be reasonable to conclude that the first attack was part of an overall campaign that in itself constitutes a single extended armed attack’.167
Licence to states that have been the victims of an armed attack to use force in anticipation of future attacks will in practice provide such states with the justification for using the force needed to weaken the military potential of the aggressor. This was in essence one of the central aims of Israel’s military campaign in response to the Hezbollah attack of 2006 July.168 Even many states that accepted Israel’s right to use force in self-defence in this case were apparently not persuaded that this was a legitimate aim, since they criticized Israel’s use of force as disproportionate.169 It must be conceded, however, that this is not the only possible explanation for their criticism. They may well have felt that even if the aim itself was legitimate the damage caused by the means used to achieve that aim was excessive in relation to the that aim.170
4 Self-defence and Proportionality
The diverse types of situation in which a state might use force in self-defence, and the different ends that might be legitimate in those situations, lead to the conclusion that differing tests of proportionality might be appropriate in different cases. This is not a case of ‘one size fits all’.
Several variables may affect the ends of the force used in self-defence, and hence the test for assessing its proportionality; whether the attack is ongoing, completed, or imminent; whether it was carried out by non-state actors and, if so, whether it can be imputed to a state; and, finally, the scale and effects of the attack, when judged in the context of the relations between those responsible for the attack and the victim state. In the analysis that follows I shall attempt to show how these variables will affect the issue of proportionality. Before I do so it is important to dwell on the distinction between various end-goals behind the use of force and on the dynamics of force and counter-force.
A Motives of Force in Self-defence
When faced with an ongoing attack the primary goal of the victim state will be halting and repelling that attack. However, even when that is the initial goal of the forcible response the matter may not end there. Unless the armed attack is limited and localized the situation is likely to be dynamic and could deteriorate rapidly into a wider armed conflict. As David Rodin mentions, when an armed conflict begins ‘the scale of force is intrinsically open-ended on both sides and open to escalation’.171 In such a situation each side may do what is necessary to weaken the military capacity of its enemy, constrained only by the norms of jus in bello.172 According to some views, at this stage proportionality is judged solely by those norms and has no relevance in jus ad bellum.173 These views would seem to be based on the assumption that assessment of proportionality in jus ad bellum is a once-and-for-all decision, made in advance when the decision to use force in response to the armed attack is made.
The point in time at which evaluation of proportionality is supposed to be made by the state using force is far more complicated in jus ad bellum than it is in jus in bello. In the latter context evaluation of the anticipated military advantages and expected damage to civilians must obviously be made before the attack. When deciding to use force in self-defence decision-makers in the state involved should ideally define the aims of such force and assess the scope of the force and the means necessary to achieve those aims. But once force has been used in response to an armed attack the dynamics of the situation are likely to require ongoing decision-making. War aims may have to be redefined and consideration given to the necessity of using further force as well as to the harm that has been caused, is still being caused, and is going to be caused by pursuing those aims. Hence, examining the proportionality of the force may become an ongoing process that must pay heed to the harm that has already been caused, and not only to the expected harm from further use of force.174
While halting and repelling an attack is accepted as a legitimate goal of self-defence, the matter becomes more complicated when the armed attack is completed before the victim state can respond, or when an armed attack is imminent. Any use of force by the victim state in such circumstances will necessarily be forward-looking and will involve one or more of three possible motives: punishing the aggressor, preventing further attacks, and deterrence. It is important to distinguish between these potential motives for using force.
By the term ‘punishing the aggressor’ we are referring to a notion of retribution or just desserts. In reality, as Yoram Dinstein so aptly remarks, some element of retribution may be present in many, if not all, cases in which a state responds to an armed attack.175 There may be strong political reasons for this, since the governments of victim states may perceive (rightly or wrongly) that public opinion demands a response to the armed attack even when it recognizes that such a response will serve no purpose other than retribution, and may even be counter-productive. Be this as it may, in our discussion of armed reprisals we saw that if there is one thing that is universally accepted it is that the notion of self-defence does not allow for the use of force the motive of which is purely retributive.
Prevention entails force the object of which is to weaken the military capacity of the enemy, thereby reducing the chances that it will be capable of mounting another attack. In the short term, at least, it is clear that the more damage done to that military capacity the less chance there will be of a further attack by the same enemy.
As opposed to prevention, deterrence involves trying to influence the willingness or motivation of potential enemies to use force against the victim state.176 A distinction must be made here between general deterrence and specific deterrence. The former is designed to send a warning message to all potential aggressors regarding the likely costs to them of attacking the victim state; the latter to affect only the willingness of the state or armed group responsible for the armed attack from mounting another attack. As general deterrence is not directed to defence against the armed attack that occurred, nor even solely to defence against further attacks by the party responsible for that attack, it cannot be regarded as a legitimate aim of the use of force in self-defence. But what of specific deterrence?
Deterrence is an inherent part of punitive measures, and it may be argued that, since these are not considered compatible with the right to self-defence, deterrence itself cannot be recognized as a legitimate motive for using force. The counter-argument differentiates between the retributive and deterrent motives of punitive measures, and holds that only the former are never legitimate, while the latter may conceivably be legitimate in certain circumstances. Taking the latter view in the context of force against terrorists, Oscar Schachter wrote that self-defence may include force that is ‘sufficient to cause the terrorist to change his expectations about the costs and benefits so that he would cease terrorist activity’.177 Michael Schmitt takes the same position.178 It would seem to me that their view is reflective of state practice, especially in cases of armed attacks by non-state actors. In many such cases specific deterrence may not be the sole motive for the use of force, but it is highly unrealistic to believe that it will not be one of the motives.179
Assuming that specific deterrence may in some circumstances, and to a certain degree, be a legitimate motive for the use of force in self-defence, the big question in the present context is by which test of proportionality it should be judged: by the means-end or ‘tit for tat’ test? The argument could be that the force used as a deterrent should be judged by its potential of achieving its ends, namely reducing the willingness of the aggressor to mount a further attack. Under this test it would seem that the greater the force used the more effective it is likely to be.180 Contrarily, as deterrence is generally regarded as part and parcel of the aims of punishment, the argument could be that it should be judged according to the ‘just desserts’ notion of proportionality. Any other notion would lead to results which are incompatible with the fundamental philosophy behind the UN Charter and its principles on the use of force by states – that the use of unilateral force should be contained and restricted as far as possible – since even a small-scale armed attack could justify a massive deterrent response. It seems to me that this latter argument should have the upper hand. To the extent that specific deterrence is a legitimate motive for using force in response to an armed attack, the proportionality of the force used must be measured in relation to the scale and effects of the armed attack. This does not imply that the victim state can use no greater force than was used against it, but that it cannot justify force that vastly exceeds the force used against it on the strength of the argument that the greater the counter-force the more effective it is likely to be as a deterrent.
B The Variables of Self-defence, Legitimate Motives and Proportionality
1 Ongoing Attacks
When the aim of forcible measures is to halt and repel an ongoing armed attack, the test of proportionality is a clear means-end test. Anything necessary to achieve this aim that is compatible with norms of jus in bello will be proportionate for the purposes of jus ad bellum. However, as we have seen above, despite the prevalence of the ‘halting and repelling’ school of thought on the purpose of self-defence the matter is not as simple as all that. Halting and repelling as the sole end of self-defence is appropriate only when the ongoing attack is of a limited scale and does not fit into a wider picture of hostility between the aggressor and the victim state. When there is a wide-scale attack, or an attack that is part of an ‘accumulation of events’ that reveals a clearly hostile intent towards the victim state, and consequently makes the chance of further attacks a real possibility, many experts agree that the victim state may also use force to remove the immediate threat from that same aggressor. Some go further and accept that the victim state may even use force to eliminate future threats that are reasonably foreseeable, or that it may use measures ‘to restore the security of the State after an armed attack’.181 The most extreme view holds that if the armed attack was wide-scale the victim state may go to war and fight until victory.
I shall take the middle road and accept that unless the armed attack is limited, localized, and unconnected to a previous ‘accumulation of events’ or war-threatening situation, the victim state may use force to reduce reasonably foreseeable future threats. While the primary aim of such force must be to damage the military capacity of the aggressor, it would be naïve to pretend that such force may not also be aimed at affecting the willingness of the aggressor to mount another attack. In other words, in such a case the legitimate ends of using force in self-defence might be halting and repelling the ongoing attack and reducing the threat of further attacks by prevention and deterrence. The proportionality of the force used will have to be assessed in light of these ends.
This poses a problem since, while the tests both of halting and repelling and prevention of threats are means-ends tests that rely first and foremost on gauging whether the means are necessary to achieve the ends, the test for deterrence is a ‘tit for tat’ test that must consider the proportionality of the force used in self-defence against the scale and effects of the armed attack. I maintain, however, that while deterrence may in practice be an additional end, it should not affect the proportionality test used in these cases. Proportionality will be based on the necessity of the means used to halt and repel the attack and to prevent further attacks by harming the military capacity of the enemy. Deterrence will have to be a function of the force used to achieve these ends.
2 Imminent Attacks
What is the situation when an armed attack has not yet occurred but is imminent? As seen above, according to most experts, including the High Level Panel appointed by the UN Secretary General and indeed the previous Secretary General himself, in this situation the threatened state does not have to sit and wait until the armed attack occurs but may act to thwart the pending attack. What are the legitimate objects of force in this situation? Are they restricted to thwarting the imminent attack, or may the threatened state go one step further and act to prevent reasonably foreseeable but non-imminent attacks, or even to deter the ‘imminent aggressor’ from attacking in the future?
Finding the appropriate test of proportionality when a state that has been subjected to an armed attack acts to prevent future attacks is a tricky enough matter. It is even trickier when the state has not yet been attacked and uses force based on its assessment both that an armed attack is imminent and that the anticipated scale of the attack is such as to demand the anticipatory use of force. Seemingly the state should not be permitted to do anything beyond what is needed to thwart the imminent attack. But how is one to gauge whether the force used was indeed necessary to do this? As long as the enemy retains some military potential it may have the capability of mounting another attack in the immediate future. Whether another attack is imminent depends, of course, not only on the capacity to carry out an attack, but on the motivation and intention to do so.182 Is there any objective standard for assessing whether such motivation and intention have been destroyed by the counter-force? Or are we talking about use of force as a deterrent against further attacks? If deterrence, the test of proportionality will be the ‘tit for tat’ test.
Even if we accept that both prevention of non-imminent future attacks and deterrence are legitimate objectives after an armed attack has occurred, when the victim state has used force on the basis of its assessment that an armed attack was imminent there will almost inevitably be a problem of evidence and credibility. If the potential victim state succeeds in frustrating an attack that was indeed about to take place, it may afterwards meet a great deal of scepticism over its claim that the attack was imminent. The degree of that scepticism will probably be influenced by the previous relationship between the victim state and the alleged aggressor.183 In all events, while the victim state will have to act on the basis of its own assessment of the likelihood and scope of the pending attack, its assessment will be subject to a second opinion by the international community after the event.184 States making use of their inherent right to self-defence in the face of an imminent attack must be prepared after the event to present evidence which supports the assessment they made of imminence.
The aim of using force in face of an imminent attack will be largely dependent on the context.185 Judith Gardam maintains that force ‘must be limited to countering the threatened attack and no more’.186 At the same time she recognizes that ‘the scale and mode of the response will be dictated by the nature and magnitude of the anticipated armed attack’.187 It seems to me that the expected scale of the imminent attack affects not only the scale of force that may be used, but its very purpose. If the pending attack is expected to be isolated and limited, the purpose of the force used in self-defence should be restricted to stopping that attack from occurring. If, however, the expected attack is one of a massive scale, amounting to war, an invasion of the victim state’s territory or wide-scale bombardment of its territory, or is part of an ‘accumulation of events’, the victim state cannot be expected to restrict its anticipatory use of force to thwarting the imminent attack alone. It may act to remove the wider threat of reasonable foreseeable future attacks by the ‘imminent aggressor’.
The legitimate ends of using force to thwart an imminent attack will of course dictate how proportionality is to be gauged. Once again the primary test will be whether the force used was necessary to achieve those ends.
3 Completed Attacks
As mentioned above, when the attack has been completed before the victim state can respond it is meaningless to speak of halting and repelling it. Neither especially affected states nor the international community have accepted the argument that this implies that the attacked state may therefore no longer exercise its right to self-defence, and must rely solely on non-forcible measures or action by the SC.
In responding to a completed attack it is self-evident that any force used by the victim state will necessarily relate to punishing the attacker or to preventing or deterring further attacks. Punishing the aggressor by use of force is not regarded as a legitimate exercise of the right to self-defence. We are left with force as a preventive or deterrent measure.
Use of force to pre-empt future attacks is not regarded as legitimate unless the attacks are imminent. As I showed above, however, once a state has been attacked the demand that forcible action to prevent further attacks relate to imminent attacks is severely weakened. The aggressive intentions of those responsible for the armed attack have been revealed and are not purely a matter for surmise. Still, there must be a ‘sound basis for believing that further attacks will be mounted’,188 and the force used must be necessary to ‘prevent any further reasonably foreseeable attacks’.189
In this case much must also depend on the scale of the completed armed attack and whether it may reasonably be seen as a localized one-time attack, or part of a wider pattern of aggression towards the victim state. This is where Dinstein’s distinction between ‘defensive armed reprisals’ and wider use of force is most relevant. If the armed attack is limited and localized, any forcible reaction by the victim state will have to be limited in scope too. Such force should be directed only against the military capacity of the state or group that carried out the attack and should be limited in both scope and location. While nobody except Dinstein seems prepared to call such use of force by its name – an armed reprisal – that is in essence what it will be, and its proportionality should be gauged accordingly. It is not surprising that the main criticism in many cases of forcible responses that look very much like armed reprisals is that they were disproportionate. In such cases the criticism would appear to rely on the ‘tit for tat’ test of proportionality.
On the other hand, when the completed armed attack is wide-scale, or part of a pattern of events, the legitimate ends of the force used in response will be both to prevent and to deter further reasonably foreseeable attacks. While in practice deterrence is likely to be one of the ends of such action, it should not be included as an end when gauging proportionality. The test here is once again whether the force used was necessary to damage the military capacity of the aggressor and thereby reduce the chances of future attacks.
4 Non-state Actors
A great deal has been written in recent years about extraterritorial use of force against terrorists or other non-state actors.190 As mentioned above, there are still some experts who maintain that unless a cross-border attack by non-state actors may be attributed to another state, the victim state may not use extra-territorial force against those non-state actors. This is no longer the majority view. I have proceeded on the basis of the well-accepted approach, according to which if the host state does not act to curb the activities of the non-state actors in its territory, a state which has been victim to an armed attack by those non-state actors may use force against them in that territory. In doing so it will be exercising its inherent right to self-defence recognized in Article 51 of the Charter. It will therefore obviously be bound by all the limitations that apply to the exercise of that right, including necessity and proportionality. Necessity in this context implies that the victim state has unsuccessfully tried non-forcible measures to persuade the host state to stop the activities directed against it by the non-state actors acting in that state’s territory, and that, given the pattern of attacks, it is left with no other effective way to defend itself.191 In order to understand what proportionality implies in this context we have once again first to consider what the legitimate ends of self-defence are in this case.
Whether the armed attack by non-state actors may be attributed to the host state or not will affect the ends of the force used. When the armed attack may be attributed to the host state, the victim state may use force in self-defence, both against that state and against the non-state actors who carried out the attack. The aims of such force will have to be either preventive or deterrent.
Preventive force against the host state will be relevant only when the nature of the connection between that state and the group of non-state actors is reflected in military support. Attacking the capacity of the state to provide such support could then be regarded as preventive in nature. A more radical preventive approach, adopted by the US in Afghanistan, would involve toppling the regime in the host state and replacing it with a regime that would no longer give support to the non-state actors. This is a highly problematical approach that does not seem compatible with the limited self-defence doctrine that lies behind Article 51 of the Charter.192 It is unlikely to find much support in the international community unless there are cogent reasons for believing that regime change will serve wider purposes, such as preventing crimes against humanity.
It is in this context, more than any other, that the legitimacy of specific deterrence as a motive for the use of force must be addressed. The distinction between prevention and deterrence is not purely semantic. The former involves destroying or harming the capacity to mount future attacks; the latter trying to affect the willingness or motivation to do so. Could affecting the willingness of the host state to continue its support of the non-state actors be a legitimate aim of force when the armed attack by those actors may be attributed to that state? While it seems that victim states will most probably have this aim in mind when responding with force against the host state, given the prevailing view that deterrence on its own is not a legitimate aim of self-defence they will usually claim that the real purpose of their action was preventive. The case of the 1986 US attack on targets in Libya in response to terrorist attacks by a group of non-state actors for which the US held Libya responsible is a case in point.193 The ambivalent attitude of the government of Israel whether its war aims in the 2006 campaign in Lebanon were directed solely against the Hezbollah or were also directed against the government of Lebanon is another.194
What about transnational attacks by non-state actors that may not be attributed to the host state? They will usually be of two kinds. In some cases, there is a pattern of small-scale attacks, some or all of which in isolation may not constitute an armed attack. These may or may not culminate in the ‘last straw’ attack, which may or may not be of the scale and effects to constitute an armed attack. In other cases there may be a dramatic once-and-for-all large-scale attack such as the 9/11 attack that in itself certainly triggers the right of the victim state to use force in self-defence. Common to both types of cases is that in responding with force the victim state will not be defending itself against the armed attack that is occurring, but against further attacks from the same group of non-state actors. The difference between the two types of cases is that when there is a pattern of attacks, the victim state has fairly strong evidence that there are likely to be further attacks if it does not react.195 On the other hand, when subjected to an isolated attack the attack itself does not necessarily indicate that future attacks are in the offing. Use of force against the non-state actors will therefore be based on notions of retribution for the attack, deterrence against further attacks, or preemption of further attacks on the basis of intelligence information, statements by heads of the group which carried out the attack, or surmise that such attacks are planned or being planned.
What will be the legitimate ends of using force in such cases? States know only too well that retribution on its own is not regarded as an aim that is compatible with the notion of self-defence, and that forcible action that is motivated solely by the desire for retribution will be regarded as an illegitimate armed reprisal. Hence they are seldom, if ever, likely to admit that their use of force is based on motives of retribution. Prevention of further attacks is regarded as a legitimate aim, although there is disagreement on how imminent those attacks must be when an armed attack has already occurred. Michael Schmitt argues that, given the difficulty in locating and tracking terrorists in the territory of another state, when it comes to terrorists the imminence requirement for use of pre-emptive force is not judged by the imminence in time of the expected terrorist attack but ‘by the extent to which the self-defense occurred during the last window of opportunity’.196 In such circumstances the prime purpose of using force must be preventing future attacks.
What about deterrence? Schmitt rules out general deterrence of terrorists as a primary aim of the use of force, but includes deterrence of the group of terrorists who carried out the armed attack as part and parcel of the preventive purpose. As we saw above, in using force against terrorists Oscar Schachter considered the latter type of deterrence to be legitimate. It seems to me that this reflects the way that states involved in responding to terrorist attacks do in fact regard the purpose of using force.197
As in all other cases, the proportionality test for the force used will depend on the legitimate ends of using that force. The problem here is that we have two conceivable ends that lead to different and competing tests of proportionality. As far as the preventive aim is concerned – damaging the capacity of the non-state actors to attack again – the test will be an instrumental one: is the force used needed to damage that capacity? If, on the other hand, the purpose is to deter the terrorists from further attacks, the scale of the armed attack will be relevant in assessing the proportionality of the force used. As we have seen, many experts and certainly especially affected states argue that in making this assessment the ‘accumulation of events’, and not only the armed attack which triggered the forcible response, is relevant in assessing the proportionality of the response.
In its National Security Strategy, first published in 2002, the Bush administration argued that traditional concepts of deterrence are inadequate in fighting against terrorists committed to wanton destruction and martyrdom. It therefore declared that it would expand its use of pre-emptive force beyond traditional ideas of an imminent attack. As declared in the Strategy:
The greater the threat, the greater is the risk of inaction—and the more compelling the case for taking anticipatory action to defend ourselves, even if uncertainty remains as to the time and place of the enemy’s attack. To forestall or prevent such hostile acts by our adversaries, the United States will, if necessary, act preemptively.198
Some writers have also suggested a different model of self-defence for the struggle against terrorism.199 This would widen the goals of using force against suspected terrorists, thereby necessarily widening the notions of proportionality.
The Bush administration’s declaration was widely criticized and was not repeated in the National Security Strategy published in 2010 by the Obama administration.200 Other suggestions to widen the scope of self-defence when it comes to terrorism are not likely to gain recognition by states that are not especially affected by terrorist attacks. It seems to me that the present position, when taken to include the right to use force both against imminent attacks and against non-state actors whose host state is unwilling or unable to take action to curb their activities, is probably still the best balance we can find between the need to contain the use of force by states while allowing some recourse to the unilateral use of force in self-defence.
C Assessing Proportionality
The discussion so far has been based on the premise that the first step in gauging proportionality must be to determine the legitimate ends of using force in the particular case. The force used must then be judged by whether it was necessary to achieve those ends. We can certainly conclude that force that was not necessary to achieve legitimate ends will be regarded as disproportionate. As Olivier Corten writes:
What is a disproportionate measure if not a measure that goes beyond what its purpose requires, that is, which is not necessary for the pursuit of that same purpose? 201
This does not imply, however, that the use of force that is necessary to achieve those ends is ipse facto proportionate. Two questions arise in relation to such force. The first relates to the meaning of ‘necessary’ in this context; the second to weighing the harm caused by necessary measures against their concededly legitimate ends.
The means-end test of proportionality is used in other legal contexts, among which are the legitimacy of restrictions on protected liberties and the legality of administrative action that affects the interests of the individual. In these contexts proportionality is widely reviewed by the three-pronged test that was first developed in German administrative law and was later adopted in other legal systems, including those of Canada and Israel, as the test for examining the proportionality of limitations on protected rights.202 When used in this last context the test involves three questions:
a. whether the restriction serves to achieve the legitimate ends;
b. whether those ends could be achieved by less restrictive means; and
c. whether the harm to the protected right caused by the restriction outweighs its benefits.
The first two questions both relate to aspects of what may be regarded as ‘necessary’, but it is important to appreciate the difference between them. The first question relates to the functional aspect of the restriction: was it necessary in the sense that it could achieve the legitimate ends? As often phrased in jurisprudence on human rights: was there a rational connection between the means and the ends, or were those means suited to the ends? The second question relates to the comparative aspect of the restriction: was it necessary in the sense that no less drastic means were available for achieving the same ends?
In jus ad bellum it seems to me that the term ‘necessary’ is used in both ways. When asking whether the very resort to self-defence was necessary, the question is whether there were non-forcible means of dealing with the armed attack.203 Once there has been a resort to force, many, but not all, experts argue that a second necessity test arises: were the means used by the state acting in self-defence necessary to achieve the legitimate ends of self-defence in the specific context?204 This question is usually answered by the first test of what is necessary, namely by asking whether there is a rational connection between the force used and the legitimate ends of its use.205 There are, however, many cases in which reference is made to use of ‘excessive force’.206 It is not clear whether the implications are that the same legitimate ends could have been achieved by less force, or whether the reference is to the final peg of the three-pronged test. I turn to that peg now.
Surprising as it may seem, while the question whether the harm caused was excessive in relation to the benefits is the very essence of the proportionality means-end test,207 this aspect of proportionality in jus ad bellum has received very little attention. Discussions of proportionality in this context have dwelt more on the necessity issues discussed above, and especially on the first of these issues, namely whether the force used had a rational connection to the legitimate ends of using force. There are probably a number of reasons for this. Assessing whether the costs outweighed the benefits, sometimes referred to as ‘narrow proportionality’, is possibly the most difficult question to answer, as it involves comparing values that are not quantifiable. This is an inherent problem in gauging proportionality, which has been widely discussed in the context of proportionality in jus in bello. Both the ICJ and commentators may well have found it easier to deal with the necessity issues than get to grips with this inherently difficult issue.
As proportionality in the narrow sense plays an important role in jus in bello, it was perhaps considered unnecessary to discuss the question in the context of jus ad bellum. Some commentators hold that the damage caused by the use of force is only relevant in jus in bello.208 It should be appreciated, however, that the questions in the two contexts are somewhat different. While the question in jus in bello relates to attacks on specific targets, in jus ad bellum the question relates to the whole picture. Use of force could conceivably be disproportionate under jus ad bellum even if all specific attacks met the demands of proportionality in jus in bello. The jus in bello test refers to collateral damage to civilians or civilian objects which are not in themselves legitimate targets, whereas the jus ad bellum test includes (but is certainly not confined to) damage to combatants and military objects.209 Finally, in jus in bello the question relates to the expected collateral damage and the anticipated military advantage. In jus ad bellum the question refers to the actual damage caused by the means used to pursue the legitimate ends of military force.210 It is therefore necessarily based on an approach that calls for constant assessment of the marginal benefits and costs of force used to pursue those ends.211
Issues of ‘narrow proportionality’ in jus ad bellum have not been subjected to much academic analysis, and the impression is that many experts assume that whether the means were necessary to achieve the legitimate ends is the be-all and end-all of proportionality in jus ad bellum.212 However, in the reaction of states and commentators to specific cases of force used in self-defence, the damage caused plays a major role in descriptions of the force as disproportionate. Thus, in their criticism of Israel’s use of force in Lebanon in the summer of 2006 as being disproportionate, many states referred to the extensive damage caused to civilians and to infrastructure.213 In his analysis of this same case, Enzo Cannizzaro also mentions the threat to and harm sustained by civilians as one of the three factors which explained why Israel’s use of force was regarded as disproportionate.214 The problem in such statements is not the consideration of the damage caused as a factor in assessing proportionality, but the absence of a serious analysis of the other side of the coin: the necessity of the force which caused the damage in advancing the legitimate ends of self-defence. It is perhaps inevitable that such an analysis is likely to be biased. The states using force will invariably tend to give undue weight to the contribution of the force used to achieving their ‘war aims’, while outside observers will tend to see the concrete damage caused as the determining factor. Courts and other decision-making bodies do not seem equipped to decide between the conflicting perspectives.
D Proportionality and Israel’s Campaign in Lebanon
I began this article by citing the widely differing approaches to whether Israel’s campaign in Lebanon was compatible with proportionality in jus ad bellum. I have no intention of answering that specific question here. But I shall end by using four ana lyses of that question to illustrate the arguments employed here.
In his assessment of Israel’s military campaign, Enzo Cannizzaro mentions three reasons for his conclusion that the campaign did not meet the demands of proportionality in jus ad bellum: the scale of the action, which exceeded the force necessary to repel the attack; that attacks were made on infrastructures hundreds of miles from the area in which the armed attack took place, and ‘were therefore unrelated to the defensive objective of the action’; and ‘the threat and harm sustained by civilians’.215
The first two reasons provided by Cannizzaro are clearly determined by his view that the only legitimate ends of using force were to repel the specific armed attack that took place on 12 July. The force used by Israel was not necessary to achieve these ends and was therefore disproportionate. This view is problematical on at least three levels.216 In the first place, by the time Israel could respond the armed attack was over. Repelling that attack was no longer on the cards, and use of force could only have been forward-looking. Secondly, as we have seen above, the notion that halting and repelling an armed attack is the only legitimate aim of using force in self-defence is not widely accepted. Most experts concede that the victim state may use force to prevent reasonably foreseeable future attacks. Thus the legitimate ‘defensive objective of the action’ was not dealing solely with the armed attack that had taken place on 12 July, but with the reasonably foreseeable threat posed by Hezbollah. If this were a legitimate objective, it is difficult to see why destroying long-range missiles of the Hezbollah aimed at various points in Israel and control centres of the Hezbollah was unrelated to the ‘defensive objective of the action’. Finally, Cannizzaro ignores the dynamics of the situation. As Wrachford, whose analysis is reviewed below, stresses, the Israeli forcible response to the attack resulted in a counter-response that involved massive bombardment of Israeli territory with rockets and missiles. This cannot be ignored when assessing the necessity of action taken to weaken the military capacity of Hezbollah.
Cannizzaro’s last point really does raise a question of ‘narrow proportionality’, which, as we have seen, is regarded by some as being a question only for jus in bello. If one accepts that ‘narrow proportionality’ is indeed part of the calculation in jus ad bellum, Cannizzaro’s argument cannot be dismissed that easily. The problem is, however, that it provides only one side of the equation. Just as mentioning the military benefits of action cannot answer the question of narrow proportionality unless account is also taken of the damage caused by that action, citing the damage caused cannot by itself answer the question. We have to know what the gains were in terms of the legitimate ends of using force – in this case reducing the threat of future attacks by Hezbollah. Given his approach that the only legitimate aim was halting and repelling the initial armed attack, it is not surprising that Cannizzaro simply ignores this issue.
In his analysis of the same military campaign Andreas Zimmermann is far less sure that Israel’s use of force was disproportionate. Zimmermann’s starting point is that proportionality in jus ad bellum is determined by the nature and scope of the armed attack and how in the specific circumstances that attack could be repelled. In judging that question Zimmermann takes into account the dynamics of the conflict that ensued and opines that much depends on factual questions, such as the control and command structure of the aggressor. Applying this approach, Zimmermann concedes that attacking Hezbollah control centres in the Lebanese hinterland met the test of proportionality even if the armed attack that provoked the Israeli response ‘only originated from a limited territory adjacent to the territory of the attacked state’.217 He also accepts that the longer the attacks by Hezbollah on Israel continued, the wider the measures of self-defence could be, provided that such wider measures were necessary to stop those attacks. In the context of jus ad bellum Zimmermann does not mention the damage caused to civilians and infrastructure, which he regards as matters to be considered as part of jus in bello.
In her discussion of the Lebanon campaign Yäel Ronen quite rightly points out that proportionality depends on whether the right to self-defence is restricted to repelling an ongoing attack or includes the prevention of future attacks.218 She discusses the requirements of proportionality under both theories, and after analysing the various goals Israeli decision-makers mentioned, voices the opinion that even if Israel was entitled to use force to prevent future attacks, it could not use such force as a general deterrent to enforce SC Resolution 1559 that called for the disbandment of all militias in Lebanon, or to prevent Hezbollah from ‘establishing itself as a regional provocateur’. To the extent that force was used to further these objectives it was disproportionate. Like Zimmermann, Ronen tends to think that the question of the damage caused to civilians and to infrastructure is a jus in bello rather than a jus ad bellum question.
Jason S. Wrachford takes a different line. He argues that as Israel’s initial response to the armed attack was met by a massive Hezbollah rocket and missile bombardment of Israel, the situation deteriorated into a large-scale armed conflict. Following Dinstein’s approach which was discussed above, Wrachford holds that in such a situation proportionality has no place in jus ad bellum and should only be judged by norms of jus in bello. Nevertheless, Wrachford opines that the proportionality of force was problematical for two reasons. In the first place, it seems that Israel’s intent was both punitive and deterrent, and while
proportionality may very well have some notion of deterrence, and possibly even some aspect of punishment to it, deterrence and punishment should not be the ultimate purpose for continued use of armed force. This would turn the military action into more of a reprisal, rather than an act of self-defense.219
Secondly, while Israel was justified in taking some action against Lebanon, its ‘extensive actions against Lebanon itself were simply not proportional in response to Lebanon’s failure to control Hezbollah’.220 It is not at all clear on what Wrachford bases this statement. Given his view that Lebanon could not be held responsible for the Hezbollah attack on Israel, the conclusion should have been that any action against Lebanon itself (as opposed to action directed solely at Hezbollah) was unlawful. Hence proportionality should have been irrelevant. But, even assuming that some notion of proportionality was indeed relevant in assessing Israel’s actions against Lebanon itself, what test of proportionality is Wrachford using here? That the same result could have been achieved by less force? Or some kind of ‘just desserts’ notion of proportionality?
Examining the above analyses reveals that, when it comes down to it, the real division of opinion on proportionality in jus ad bellum relates to the legitimate ends of using force in given circumstances. Common to all the analyses is that the forcible measures used in self-defence must serve legitimate ends. Use of force that seems to be largely punitive, the primary motive of which is deterrence or which has wide political aims will be regarded as disproportionate. Beyond that consensus, the differing views reflect the various theories on the legitimate ends of using force in self-defence, and their applicability in this specific case. Cannizzaro, a strict ‘halting and repelling’ proponent, ignores both the dynamics of the situation and threats of future attacks by the same enemy, and assesses whether the force used by the victim state was necessary to halt and repel the original armed attack that occurred. Zimmermann, taking a wider view of the ‘halting and repelling’ theory, considers both the nature of the enemy’s control structure and the dynamics of the conflict as it develops, but still examines the force in the light of armed attacks which have occurred, and not in the light of future threats. Ronen accepts that preventing reasonably foreseeable future attacks is a legitimate end, and that force that was necessary to pursue this end was proportionate. She stresses that wider political aims are not legitimate and that force to achieve such aims was disproportionate. Wrachford’s analysis reflects the Dinstein version of the trigger theory: once a wide-scale armed conflict ensues, proportionality in jus ad bellum is no longer relevant.
It is worth noting that only one of the four experts considers that the damage caused is a jus ad bellum question. According to the majority view among these commentators proportionality in jus ad bellum is not really proportionality at all, but purely a question of whether the force was necessary to achieve the legitimate ends of using force in self-defence.
5 Concluding Comments
Proportionality is invariably an elusive concept. It becomes even more so when used in a context which is as highly loaded as the right of a state to use force in self-defence. Nevertheless, to paraphrase the Duchess in Alice’s Adventures in Wonderland, everything has a meaning, ‘if only you can find it’.221
One source of confusion over the meaning of the principle of proportionality in jus ad bellum lies in the natural and popular tendency to assume that proportionality must imply some notion of ‘tit for tat’. I have argued, however, that the main source of disagreement and confusion flows from the lack of consensus over the legitimate ends of force employed by a state that is exercising its inherent right to self-defence. Even when it is accepted that the appropriate test of proportionality is a ‘means-end’ test, in the absence of agreement on these ends it is obviously impossible to agree on the necessary means to achieve them. It is also impossible to apply the ‘narrow proportionality’ test that can enter the picture only once a determination has been made that the means used were indeed necessary to achieve a legitimate war aim.
In stressing the lack of consensus on the legitimate ends of force in self-defence it may seem to the reader that I have merely replaced one area of uncertainty and indeterminacy with another. Rather than over whether force used in response to an armed attack was proportionate, the argument will relate to the legitimate ends of using that force and whether the means used were indeed necessary to achieve those ends. That may well be the case. But at least we will then know what it is we are arguing about.