The issue of whether a state is entitled to respond in self-defence to attacks committed by non-state actors, even if the state on whose territory force is used is not substantially involved in these attacks, has been the object of a large controversy in the legal scholarship in the aftermath of the US military operation (‘Enduring Freedom’) in Afghanistan in reaction to the 11 September 2001 attacks on the United States.Footnote 1 It is no doubt too early to draw any firm conclusion from this practice. Future practice will be decisive. In this respect, it is of the utmost importance to note that recent years have seen the launching of significant military actions in response to attacks by non-state actors in which no state could be considered as being substantially involved: in July 2006 Israel invaded parts of Lebanon to put an end to the firing of rockets by Hezbollah; in February 2008 Turkey launched a major operation into Iraq in order to stop attacks being carried out from there by the Kurdistan Workers’ Party (PKK); in December 2008 and until January 2009, Israeli military forces invaded Gaza with the aim of destroying the Palestinians’ capacity to fire missiles and rockets into southern Israeli territory.
This article seeks to determine the lessons that can be learned from this state practice about the evolution of the law of self-defence. The main and most significant conclusion that will be asserted in this respect is that recent state practice confirms the tendency evidenced by the operation Enduring Freedom towards allowing states to respond in self-defence to private armed attacks – that is, attacks which are committed by non-state actors only. It will be submitted that this practice, as considered together with the wide support given to the US operation in Afghanistan, strongly suggests that private armed attacks may nowadays be regarded as amounting to an armed attack under Article 51 of the UN Charter. Besides this critical conclusion concerning the nature of the authors whose attack may trigger the right of self-defence, other conclusions will also be put forward with regard to some other particular conditions of the law of self-defence, since they played a significant role in the assessment of the legality of the recent state practice. It will be argued in this respect that this practice confirms that any armed attack must reach some level of gravity – which may be evaluated by accumulating minor uses of force – in order to trigger the right of self-defence, and that proportionality of the action taken in self-defence must be assessed in quantitative terms as a means of making a prima facie judgement about the necessity of this action.
The article is divided into four main sections. Section 1 is devoted to a preliminary analysis, whose purpose is to establish the framework of the debate. It will consist in analysing the sources of the law of self-defence in order to identify the general conditions under which this law may evolve through state practice. The results of this preliminary analysis will lead us to address two core issues: first, whether the three aforementioned cases constitute proper applications of the law of self-defence and, second, whether the reaction of the international community to these cases may be interpreted as leading to an evolution of this law. These questions will be discussed in sections 2 and 3, respectively, the second question receiving particular attention as it constitutes the crucial part of the paper. The answers given to these questions will eventually allow the formulation of concluding remarks in section 4.
1. Preliminary analysis: the sources and evolution of the law of self-defence
Before analysing the three aforementioned cases, one must identify the conditions under which the law of self-defence may evolve through state practice, which implies an analysis of the sources of this law. It is generally upheld in this respect that the right to act in self-defence is regulated by both customary and conventional norms.Footnote 2 Would that mean that parts of the law of self-defence may evolve differently according to the source on which these parts rest, since the evolution of customary and conventional law is regulated by different rules? It is submitted here that the evolutions of customary and of conventional aspects of the law of self-defence merge in some ways.
First, it is difficult to identify which of these two sources, customary or conventional law, are invoked by states when they refer to the right of self-defence and, therefore, to differentiate between the evolution of each of these sources in practice. Indeed, states frequently do not indicate whether the right of self-defence to which they refer is based on customary and/or conventional law. Moreover, when Article 51 of the UN Charter is expressly invoked, as usually happens in practice, this does not necessarily mean that only conventional law is concerned, since Article 51 itself refers to customary law.Footnote 3 By invoking this article, states may only have intended to refer to the customary right of self-defence as recognized by Article 51.Footnote 4 In the same way, customary law and, more particularly, state practice pre-dating the UN Charter, is sometimes used by states merely as a means of interpreting Article 51 of the Charter.Footnote 5 Finally, it is interesting to note that the conditions of necessity and proportionality, being normally of a customary nature since they are not expressly provided by Article 51 of the UN Charter, are sometimes presented as imposed by this article.Footnote 6 In fact, state practice evidences that Article 51 tends to attract all the aspects of the right of self-defence without formally giving to these aspects a conventional nature. It is actually the provision with which the entire law of self-defence is generally associated. In brief, although this law is legally composed of customary as well as conventional aspects, the distinction between these aspects and the respective evolutions thereof are nearly impossible to establish in practice.
Moreover, the conditions under which customary and conventional law may evolve through state practice are not so dissimilar. State practice may lead to the creation, interpretation, or modification of a customary rule only if this practice is associated with a state's belief of creating or interpreting or modifying such a rule and if the practice is general and constant. Similarly, according to Article 31 of the Vienna Convention on the Law of Treaties, interpretation of a treaty can be based on subsequent state practice if this practice is an application of the treaty and establishes the agreement of the parties regarding this interpretation as to what actually implies, for such an agreement to be clearly identifiable, that state practice be repeated over time and approved by the other parties.
Yet it could be objected that interpretation may not formally be considered as amounting to a true evolution of the rule interpreted, since it presumably reveals the original meaning of this rule and has no retroactive effect. It could also be objected that the Vienna Convention on the Law of Treaties provides that subsequent state practice has only an interpreting and not a modifying effect on treaties. These objections may, however, be easily countered. Indeed, on the one hand, interpretation can be considered a means by which a rule may evolve to the extent that it reveals the meaning of a text that was originally not foreseen. On the other hand, as has been correctly underlined in the legal literature,Footnote 7 treaty modification based on subsequent state practice is generally admitted in jurisprudence, scholarship, and state practice. The general condition agreed on is that the practice of states parties to a treaty must clearly evidence an agreement between these states to modify the treaty.
One may therefore conclude from a general point of view that the evolution of the law of self-defence through state practice merely requires that this practice be followed in application of this law and be constant and general.Footnote 8 As a result, this article will now analyse whether the recent above-mentioned state practice constitutes a pertinent application of the law of self-defence, before scrutinizing the international community's reaction to this practice in order to identify whether it has been shared by the other states and has led to an evolution of the law of self-defence.
2. Does recent state practice constitute an application of the law of self-defence?
Generally speaking, it is far from easy to establish in all cases the legal ground on which a state has undertaken an armed action abroad. States do not often justify such action in clear and unambiguous legal terms. They frequently invoke only political or moral considerations, or voluntarily use a variety of arguments in such a way that the legal ones are hardly distinguishable from the others. Sometimes no justification at all is given, particularly when states resort to quick or minor uses of force.
It is fortunately not the case as regards the Israeli invasion of Lebanon in 2006. Indeed, Israel reported its military operation to the UN Security Council as required by Article 51 of the UN Charter and elaborately justified it by relying on its right of self-defence.Footnote 9 Moreover, all the other states assessed the legality of this operation in the light of the conditions imposed by the law of self-defence.Footnote 10 It seems therefore clear that the Israeli invasion of Lebanon was an application of the law of self-defence. However, the two other cases – the Turkish incursion into Iraq (subsection 2.1) and the Israeli intervention in Gaza (subsection 2.2) – raise more problematic issues regarding the legal ground on which both these operations are to be regarded as legally founded.
2.1. The Turkish incursion into Iraq
It is true that Turkey did not report its military action to the UN Security Council and did not offer any eloborated legal justification for it. However, several elements suggest that the Turkish operation in Iraq in February 2008 was an application of the law of self-defence.
First, the Turkish prime minister, Recep Erdogan, explicitly stated that the operation was justified under ‘international laws governing self-defence’. Although very brief, these declarations were repeated several times and more specifically in October 2007, in order to justify the adoption by the Turkish parliament of a plan for an incursion into Northern Iraq,Footnote 11 and in February 2008, during the military campaign.Footnote 12 Besides, even if most international reactions remained vague regarding the legal justification of this campaign, some states expressly tested it against the law of self-defence.Footnote 13
Second, if one looks at the attitude that Turkey adopted in the past with regard to its earlier incursions into Iraq in response to the PKK's attacks, one must notice that, while not generally giving a full legal justification for these interventions, Turkey came close to asserting that it was acting in self-defence in letters addressed to the UN Security Council.Footnote 14 The prime minister even expressly declared, in relation to some of these interventions, that any Turkish action in Iraq in response to the PKK's attacks would be justified under the law of self-defence.Footnote 15 In the same way, several states viewed these Turkish interventions as (legal or illegal) applications of the law of self-defence.Footnote 16
Third, and more fundamentally, the law of self-defence offers the most appropriate justification for the 2008 Turkish incursion into Iraq. Indeed, Turkey could rely neither on the consent of the Iraqi authoritiesFootnote 17 nor on any authorization by the UN Security Council, two generally recognized exceptions to the prohibition on the use of force. Besides, Turkey does not seem to have argued seriously for recognizing a new (possible) exception to this prohibition, such as one founded on humanitarian or pro-democratic grounds or on the concept of necessity as a circumstance precluding wrongfulness.Footnote 18
2.2. The Israeli intervention in Gaza
Identifying the legal ground against which the legality of the Israeli intervention in Gaza should be tested is particularly difficult. Self-defence is meaningful only if it serves to justify a resort to force that would be otherwise unlawful under the prohibition on the interstate use of force. For example, it is uncontested that self-defence is not relevant to justifying a use of force if the latter is resorted to on the territory of the state acting in self-defence and is not directed against any foreign military base. Indeed, this action would not amount to any violation of the prohibition on the use of force between states. Such an action, having a policing nature, should actually be evaluated under the criminal law rules of the state concerned.
It is true that the Israeli intervention in Gaza has been reported to the UN Security Council as an application of the law of self-defenceFootnote 19 and that some other states have assessed it in the light of this law.Footnote 20 However, the territory in which the Israeli force was used – Gaza – has a very particular status which is currently highly controversial. Israel contends that Gaza is no longer an occupied territory since it withdrew its forces from this territory in 2005.Footnote 21 This position is opposed by many other states and organizations, such as the United Nations,Footnote 22 as well as international institutions,Footnote 23 which consider that Israel continues to be an occupying power because it retains effective control over this territory – that is, of its airspace, seashore and land borders. The solution to this debate actually depends upon the way the notion of occupation, provided in Article 42 of the 1907 Hague Regulations,Footnote 24 is construed with respect to all the relevant facts, such as – and especially – the Israeli blockade of Gaza. But, whatever position is defended, the Israeli intervention in Gaza does not seem to be a proper application of the law of self-defence.
Indeed, if it is argued that Gaza is still occupied by Israel, one may doubt whether the law of self-defence is applicable to any use of force by Israel in this territory: once a territory is occupied, conducts between the occupying power and the occupied territory are normally regulated by the law of occupation. According to this law, the occupying power can and even must ‘take all the measures in his power to restore and ensure, as far as possible, public order’.Footnote 25 It is therefore in the light of these police powers and not on the basis of the law of self-defence that the Israeli intervention in Gaza should be assessed in this first hypothesis.
If one assumes in a second hypothesis that Gaza is no longer an occupied territory, it is also doubtful whether the law of the interstate use of force, including the law of self-defence, is applicable to all the forcible measures taken by Israel against the Palestinians in Gaza. Indeed, although Gaza is definitely not a part of Israeli territory, neither it is considered to be a (foreign) state. Israeli resort to force in Gaza could not formally be regarded as a violation of the prohibition on the use of force between states, and it is consequently irrelevant for Israel to invoke the law of self-defence to justify it. This does not, however, mean that such a use of force would be legal. It is generally acknowledged that the Palestinians are people enjoying a right to self-determination.Footnote 26 As a result, any use of force between Israel and the Palestinians in Gaza should be assessed in the light of the – often ambiguous – rules regulating the right of people to self-determination and not on the basis of the law of interstate use of force. It is interesting to note in this respect that, when invoking the right of self-defence under Article 51 of the UN Charter, Israel indirectly contributes to reinforcing the idea that Gaza is a state.
The Advisory Opinion of the International Court of Justice (ICJ) in the Wall case may be interpreted as supporting both these conclusions – that is, the law of self-defence is not relevant to justifying a use of force when it is resorted to in a territory controlled by the state acting in self-defence or when it is not directed against a foreign state. Indeed, in the case at issue, the Court based its dismissal of the Israeli argument that the construction of the wall (in the West Bank) is justifiable under Article 51 of the UN Charter on two main considerations: Israel did not claim that the attacks to which it responded by building the wall originated outside the territory in which it exercised controlFootnote 27 and neither did Israel argue that these attacks emanated from a foreign state.Footnote 28 The Court therefore concluded that Article 51 of the Charter had no relevance in this case.Footnote 29 The words used by the Court – ‘no relevance’ – are of the utmost importance in interpreting the conclusions of the judgment. The Court did not merely mean that some conditions of the law of self-defence were not met. It expressly asserted that this law was not applicable at all in the two aforementioned situations.
Having said that, statements from Israel and other states in relation to this law can hardly be simply ignored. They actually express these states’ opinio juris regarding the application of the right of self-defence in response to attacks by non-state actors even if, from a legal point of view, no exercise of this right may be considered as having taken place. These statements should actually be given a weight similar to the recognition of the general declarations pronounced by the states in the frame of the adoption of the UN General Assembly (GA) resolutions on use of force issues (including the right of self-defence), such as Resolution 2625 (XXV) and Resolution 3314 (XXIX). They must therefore be taken into account in a study devoted to the evolution of this right in the light of recent state practice.
3. Does recent state practice lead to an evolution of the law of self-defence?
As demonstrated above, all the three aforementioned cases may – fully or partly – be considered relevant state practice from which lessons can be drawn on the law of self-defence. One may therefore turn now towards analysing the international community reaction to these cases in order to determine whether the law of self-defence has recently evolved. Recent state practice seems to give helpful indications, especially with respect to two issues: first, and most significantly, whether the right of self-defence may now be exercised in response to attacks by non-state actors even if no state is substantially involved in these attacks (subsection 3.1) and, second, how to interpret some of the (other) particular conditions under which this right is to be exercised (subsection 3.2). The condition of necessity should normally be given particular – that is, separate – attention, since it is a very, perhaps the most, important condition of the law of self-defence and since it played an important role in the assessment of the legality of recent state practice. However, as will be explained below, this role is indirect in the sense that the condition of necessity is intrinsically linked to the fulfilment of other, more particular, conditions. As a result, the condition of necessity will not be analysed as such in a separate section of this article. It will rather be addressed in relation to the other, more particular, conditions to which it is linked.Footnote 30
3.1. Armed attack
There are different ways in which recent state practice may be interpreted in relation to the question of whether attacks by non-state actors may trigger the victim state's right of self-defence. One may consider that the law of self-defence has remained unmodified (subsection 3.1.1) or, on the contrary, has evolved through this practice. In the latter case, evolution of the law of self-defence could consist either of extending or of adapting the state attribution rules (subsection 3.1.2), or of recognizing that private armed attacks may themselves amount to an armed attack and trigger the right of self-defence (subsection 3.1.3). It will be shown that only this last interpretation seems to be correct.
3.1.1. Maintaining the traditional requirement that a state be substantially involved in the attacks committed by non-state actors
One may first consider that no evolution of the law of self-defence can be drawn from the above-mentioned cases. This view has been supported in the legal literature with respect to some of them.Footnote 31 According to such an interpretation, the law of self-defence is still to be conceived in its classical meaning, that is, allowing states to act in self-defence only in response to an armed attack by another state, what implies either an attack by the military forces of this state or attacks by non-state actors in which this state is, at least, substantially involved. The argument advanced to support such a view is that recent state practice does not actually meet the conditions under which the law of self-defence may evolve – that is, repetition, as well as explicit and general approval of the practice.
It is true that both of the Israeli interventions in Lebanon and Gaza have been condemned by the international community. It is also true that the Turkish incursion into Iraq has been expressly approved by a few states only.Footnote 32 International reaction to these cases must, however, be analysed in more depth. Such an analysis actually reveals an implicit but real approval of the right to respond in self-defence to attacks by non-state actors, even if no state is substantially involved in such attacks.
Indeed, as far as the Israeli interventions in Lebanon and Gaza are concerned, one must first note that these interventions were condemned for reasons completely alien to the nature of the authors of the attacks to which Israel responded. States condemned Israel mainly because of the disproportion between the attacks committed by the non-state actors and the Israeli military reaction.Footnote 33 In other words, they criticized the way in which Israel exercised its right of self-defence, while not contesting that the rocket attacks from Lebanon and Gaza could allow a legal response under the law of self-defence, even if no state could be considered as substantially involved in these attacks.
Many states expressly recognized and, therefore, acknowledged as a matter of principle that Israel enjoyed a right to respond to the attacks by Hezbollah and Hamas under the law of self-defence.Footnote 34 Such a recognition is very remarkable. First, it sharply contrasts with the traditional responses of the international community to past Israeli interventions, which were generally limited to condemning Israel for an unnecessary and/or disproportionate military response. Second, the declarations in which states recognized the right of Israel to defend itself against attacks launched from Lebanon or Gaza were unambiguous. Finally, these declarations were expressed unexpectedly by a large number of states, especially in the case of the Israeli intervention in Lebanon in 2006.
As far as the Turkish incursion into Iraq is concerned, one must also note that, if only a few states explicitly approved the operation,Footnote 35 even fewer states condemned it,Footnote 36 these condemnations being moreover particularly insubstantial, some coming very late and being formulated in general terms, while none was directly concerned with a denial of the right of Turkey to respond in self-defence against the PKK's attacks. In addition, many states, although not expressly approving the Turkish operation, did not condemn it and merely recommended that Turkey resort to proportionate force in fighting the PKK in Iraq.Footnote 37 More generally, states refrained from giving any comment on the Turkish operation. This operation, in contrast to most of the major military operations, was not in fact reported to the Security Council and was not therefore discussed at the UN level. This silence is surprising, given the scale of the incursion and the wide-ranging opportunities that states now enjoy to expressing their opinion on the behaviour of other states.Footnote 38 Taken together, international reaction to the Turkish operation strongly suggests that states in fact condoned this operation and thus did not oppose Turkey's right to act in self-defence against the attacks of the PKK.
It is true that approval of the aforementioned state practice is not absolutely clear and explicit. International reaction does, however, evidence some particularity, which allows us to conclude that there is at least a tendency towards authorizing a state to act in self-defence in response to attacks committed by non-state actors, even when no state is substantially involved in it. As will be explained below, such a conclusion may be asserted even more strongly if this evolution actually goes towards allowing a state to act in self-defence in response to an attack which is committed by a non-state actor only (infra, section 3.1.3.1).
3.1.2. Extending or adapting the state attribution rules
A second way of interpreting recent state practice is to consider that it does not alter the classical conception of the right of self-defence, according to which it can only be exercised in reaction to an armed attack by a state, but implies that situations in which attacks committed by non-state actors can be attributable to a state have been extended or adapted. This position, which has been held by most of the legal scholarship to explain the general support given by states to the US operation Enduring Freedom,Footnote 39 does not seem well founded.
First, it does not reflect the declarations made by most of the states in the cases under consideration. Indeed, in all these cases, self-defence was invoked in order to respond to attacks presented as being committed by non-state actors only. No state has been accused of being the author of these attacks, either directly or indirectly. Lebanon and Iraq were merely criticized by Israel and Turkey respectively as being unwilling or unable to put an end to the attacks launched from their territory.Footnote 40 No states contested the fact that the attacks were committed only by non-state actors and some even acknowledged it.Footnote 41
Second, the classical state attribution rules can hardly be stretched as far as to allow the attribution of hostile activities conducted by non-state actors to states which are merely unable or unwilling to stop these activities on their territory.Footnote 42 In any case, these rules do not seem to fit this kind of situation.Footnote 43
Third, and more fundamentally, the yardstick for assessing whether attacks by non-state actors may be considered attacks committed by a state is not properly to be found in the state attribution rules. Two main reasons support this opinion.
From a conceptual point of view, an attack committed by a state ‘using’ non-state actors – that is, an indirect armed attack – actually consists in this state involvement in the activities of the non-state actors. It does not result from the non-state actor's attack first being qualified as an armed attack and then attributed to a state. In other words, what primarily counts in determining the existence of an indirect armed attack is not the attacks by the non-state actors in which a state is more or less involved, but the involvement itself of this state in these attacks. If one attributes attacks committed by non-state actors to a state in order to qualify it as committed by this state, it logically implies that such attacks are to be considered as an ‘armed attack’ by a state under Article 51 of the UN Charter and, therefore, as the proper casus foederis of the right of self-defence that could possibly be exercised in response. Yet again one must insist that for an ‘armed attack’ to be taken into account under Article 51 in order to justify exercising the right of self-defence in reaction to an indirect armed attack, it is not the attacks themselves by the non-state actors which count but the state involvement in such attacks.
Moreover, the rationale underlying the state responsibility rules is radically different from that underlying the law of self-defence. Although the latter does not prevent the victim from making accountability claims and asking for reparations for the damages that have been suffered, its primary goal lies elsewhere. Its purpose is first and foremost to protect a state against an armed attack, such a situation being per definitionem so urgent that it does not logically allow a state victim of an armed attack to wait to raise questions of responsibility.Footnote 44 In fact, the law of state responsibility and the law of self-defence are two different branches of international law pursuing different goals. Criteria used in the former in order to attribute activities to a state may logically be different from criteria used in the latter to assess whether a state involvement in attacks by a non-state actor may be considered an (indirect) armed attack by this state.
This consideration may find indirect support in the jurisprudence of the ICJ. Indeed, in the Bosnia Genocide case, the Court held that criteria set forth by the state attribution rules do not necessarily have to coincide with those used for assessing whether a state involvement in activities of non-state actors in an internal armed conflict internationalizes this conflict. For, according to the Court, these criteria apply to the pursuit of different goals and may themselves accordingly be different.Footnote 45 This reasoning may similarly be applied to the distinction between the criteria for attributing private acts to a state and those to which one has to refer in order to define an indirect armed attack.
Moreover, in both the Nicaragua Footnote 46 and the Armed Activities Footnote 47 cases, the Court did not apply the state attribution rules when addressing the issue of whether the involvement of Nicaragua and the Democratic Republic of the Congo (DRC) respectively in rebel activities in El Salvador and Uganda could be qualified as an armed attack under Article 51 of the UN Charter. Rather, it relied on Article 3(g) of GA Resolution 3314 (XXIX), which defines the notion of aggression. The Court referred to the state attribution rules only when examining whether the United States and Uganda had breached the prohibition on use of force by supporting the rebel activities in response to the alleged aggression by Nicaragua and the DRC respectively.
In accordance with the ICJ jurisprudence, UNGA Resolution 3314 (XXIX) seems to provide the most appropriate criterion for assessing the existence of an indirect armed attack. This is confirmed by the fact that the resolution, drafted during the Cold War period – when wars were generally covert – was clearly meant to deal with aggressive acts committed by states by indirect means – that is, non-state actors.Footnote 48
According to Article 3(g) of the resolution, a state shall be considered as having committed an indirect armed attack if it has sent non-state actors to commit attacks abroad or has been substantially involved in this sending. It is clear from the text as well as from the preparatory works of the resolutionFootnote 49 that the mere inability or unwillingness of a state to put an end to the activities of non-state actors on its territory cannot amount to an armed attack by this state under Resolution 3314. One must therefore conclude in the light of this resolution that the above-mentioned cases, in which a state was only considered as being at the most unable or unwilling to stop attacks by non-state actors, concern the situation of action taken in self-defence in response to attacks committed by non-state actors only, and not by a state.
3.1.3. Having private armed attacks considered as amounting to an armed attack under Article 51 of the UN Charter and applying the condition of necessity
A third and final way of interpreting recent state practice is to consider that this practice has led to the recognition that attacks by non-state actors may themselves amount to an armed attack under Article 51 of the UN Charter and trigger the right of self-defence of the victim state without imposing the need to demonstrate that a state has been substantially involved in these attacks or that the latter are attributable to a state. As will be demonstrated below, such an interpretation seems to be the most appropriate for several reasons (see section 3.1.3.1).
This interpretation nevertheless leaves one critical problem unaddressed. Although private armed attacks seem to be enough to trigger the right of self-defence in the light of such an interpretation, recent state practice also evidences that it is still necessary to prove some link between the state acting in self-defence and the state on whose territory the defensive action takes place. Such a requirement may, however, be easily met; indeed, it will be demonstrated further that this proof, far from being imposed by any rule pertaining to state responsibility, is simply required by the condition of necessity of the law of self-defence (see section 3.1.3.2).
3.1.3.1. Private armed attacks as an armed attack under Article 51
The interpretation of recent state practice according to which attacks by non-state actors may amount to an armed attack under Article 51 of the UN Charter correctly reflects the tendency, emphasized above, towards allowing a state to respond in self-defence to attacks by non-state actors even when no state may be identified as substantially involved in these attacks (see supra, section 3.1.1). Such an interpretation, moreover, avoids the numerous difficulties raised when the issue is addressed in the light of state attribution rules; it is also perfectly in line with all the declarations made in recent practice, when states unambiguously considered that the attacks against which self-defence was resorted to were attacks committed by non-state actors only and not by a state (see supra, section 3.1.2).
Besides, although it is true that this practice has not been the object of clear and explicit approval by other states, one must admit that the conditions that the law of self-defence must fulfil in order to evolve through state practice, and, more particularly, repetition, as well as general and explicit approval of the practice, must not be construed too strictly, at least if this evolution is to go towards recognizing a right to respond in self-defence not only to an armed attack by a state but also to attacks by non-state actors. Such a right has never been excluded, either in the legal texts, in ICJ jurisprudence, or in past state practice.
Indeed, one must first acknowledge that, although not expressly requiring that an armed attack be committed by a state in order to trigger the right of self-defence, legal texts, including Article 51 of the UN Charter and regional defensive treaties, were classically considered as envisaging the exercise of the right of self-defence in response to an armed attack by a state.Footnote 50 However, preparatory works of the UN CharterFootnote 51 and an analysis of the numerous regional defensive treatiesFootnote 52 clearly evidence that states were originally unconcerned with the (state or non-state) nature of the author of the attacks which could be responded to in self-defence. In other words, military defensive reactions to attacks committed only by non-state actors were not excluded. They were simply not envisaged at the time the legal texts were drafted.
The same conclusion may be reached as far as the ICJ jurisprudence is concerned. It is true that, in both the Nicaragua Footnote 53 and the Armed Activities Footnote 54 cases, the Court considered that the plaintiff states’ involvement in the rebel activities committed abroad was not serious enough to amount to an armed attack by these states. It therefore dismissed the self-defence argument put forward by the defendants to justify their military reactions to the rebel activities. What may be surprising is that the Court did not consider whether these activities could themselves amount to an armed attack and justify the defendants responding in self-defence. This may suggest that, in the Court's view, the right of self-defence can only be exercised to counteract an armed attack by a state. But the reason why the Court did not actually address this situation is simply because it was not asked to do so. States before the Court only envisaged the hypothesis of a defensive action undertaken in response to an armed attack by a state; it does not mean that another hypothesis was impossible.
Such another hypothesis – that is, the right to act in self-defence in response to attacks committed by non-state actors only – has not been excluded by previous state practice. Even though most of the actions undertaken in self-defence to counter attacks committed by non-state actors have been condemned in the past, these condemnations were due to considerations entirely other than the (state or non-state) nature of the attackers. These considerations were often related to the necessity and proportionality of the action undertaken in self-defence.Footnote 55 They also resulted from the lack of proof that the targets of the defensive action were linked to the hostile activities that this action was supposed to fight.Footnote 56
As a result, recognizing a right to act in self-defence in response to attacks by non-state actors and not only to an armed attack by a state should not be seen as a modification of the law of self-defence, but at most as an interpretation of this law which had never been excluded. Such an evolution of the right of self-defence should not therefore require state practice to reach the same level of explicit approval and repetition as would be required for a modification of this right.
One may thus conclude from the international reaction to the recent state practice, as analysed in the light of some general considerations related to the conditions under which the law of self-defence may evolve, that this practice evidences a clear tendency towards allowing a state to act in self-defence in response to a attacks, even if these attacks are committed by non-state actors only.
3.1.3.2. The application of the condition of necessity
One question remains to be resolved in order to complete the interpretation of the recent aforementioned cases. Indeed, in these cases and, more particularly, in the Israeli intervention in Lebanon and the Turkish operation in Iraq, the states acting in self-defence, although responding to attacks committed by non-state actors only, still insisted on a link between these actors and the state in whose territory their military action was taking place. More precisely, they emphasized that this state had failed to abide by its obligation to prevent attacks from being committed from its territory and was therefore responsible for these attacks.Footnote 57 Such an argument is not new. It has almost always been invoked by states when justifying their military responses to attacks by non-state actors.Footnote 58 This practice, as well as some states’ reactions to such responses,Footnote 59 suggests that it is required to establish a link between the non-state actors and the state in whose territory the self-defence action takes place, and that such a requirement consists in establishing the responsibility of this state for the attack committed by the non-state actor. This seems to be consistent with the indisputable fact that, even when only directed against non-state actors, any self-defence action affects the territorial integrity of another state.Footnote 60 One may nevertheless wonder on which legal basis this requirement should be considered grounded.
As explained above, it cannot be regarded as imposed by the state attribution rules. Should it therefore be assumed that it merely derives from state practice and that, as a result, there exists an independent rule according to which, as far as self-defence is exercised in response to an attack committed only by non-state actors, the state in whose territory the defensive action takes place must at least be proved responsible for the private attacks? It is submitted that such responsibility does not necessarily have to be established. This paper more precisely argues that it is only required to prove the existence of a link between the non-state actors and the ‘harbouring’ state, consisting in at least the inability or unwillingness of the latter to stop the activities of the former, and that such requirement does not stem from the responsibility rules but from the condition of necessity of the law of self-defence.
This opinion may find some support in the recent aforementioned state practice. Indeed, while expressly approving as a matter of principle that Israel had the right to defend itself against Hezbollah attacks, many states did not seem to share Israeli opinion that Lebanon's inability to stop these attacks amounted to wrongful conduct.Footnote 61 According to their opinion, Lebanon could not be held responsible for the attacks emanating from its territory. Likewise, although having insisted in the past on Iraq's obligation to prevent attacks emanating from its territory,Footnote 62 Turkey also seems to have acknowledged at this time that Iraq was merely unable to fulfil this obligation and could not be held responsible for the PKK's attacks.Footnote 63
Yet it is true that these declarations may be called into question. In other words, one may wonder whether Lebanon and Iraq, contrary to many states’ declarations, should not actually be regarded as responsible for the attacks committed by the non-state actors in the light of the relevant rules and the relevant facts.
Indeed, it may be difficult at first glance to imagine that a state could not be held responsible for the attacks directed against another state when these attacks are committed from its territory, given the existence of a general obligation imposed on states to prevent attacks from being committed abroad from their territory.Footnote 64 Such a situation cannot, however, be excluded, since this general obligation has a particular nature. Indeed, this obligation is one of means and not of results.Footnote 65 This implies that states must use all the reasonable means at their disposal to prevent or stop attacks from being committed abroad from their territory and that they cannot be held responsible for the mere commission of such attacks. This responsibility would require an in-depth analysis of whether the ‘accused’ state behaved according to the general standards applying to any state finding itself in this particular situation. The results of such an analysis are particularly difficult to establish. They imply both the knowledge and a careful assessment of all the relevant facts, these facts being generally unknown to the public (including the legal scholars) and the definitive evaluation thereof being normally for the judges. Moreover, the distinction between the unwillingness and the inability of states to stop hostile activities on their territory, the former necessarily amounting to wrongful conduct in contrast to the latter, is often very difficult to draw. A state's apparent unwillingness may be the result of national political pressures which prevent it from any effective action.
As a result, it is difficult to assert with certainty that Lebanon and Iraq, as suggested by many states, were not responsible for the attacks emanating from their territory. Having said that, it must be acknowledged that recent state practice, especially the Israeli intervention in Lebanon, evidences that, at least in states’ view, the state in whose territory self-defence is exercised in response to the attacks by non-state actors must not necessarily be held responsible for these attacks. Moreover, this practice makes it clear that what actually counts in order to justify such a response in self-defence is that the harbouring state must at least have failed to stop the attacks from its territory. And, as explained above, this failure does not necessarily imply a wrongful act.Footnote 66
More fundamentally, the establishment of any responsibility does not seem relevant. As already mentioned, the law of state responsibility and the law of self-defence are two different branches of international law and rest on distinct rationales. The primary purpose for which a state, victim of attacks by non-state actors, invokes the responsibility of the state where the response has taken place is certainly not to obtain any reparation for the damage caused by these attacks. It is first and foremost to emphasize that the harbouring state is unwilling, or at least unable, to stop the attacks from its territory and, as a result, that the responding state cannot rely on the police force of this state to protect it. It actually intends to show that no other means are available for it to defend itself against the non-state actors. In other words, the harbouring state's responsibility is invoked in order to conform with the condition of necessity of the law of self-defence, which indeed requires – in one of its numerous aspectsFootnote 67 – that any action in self-defence be taken only as a last resort – that is, after all the available alternatives have been exhausted. The existence of a link between the non-state actors targeted in self-defence and the harbouring state – that is, one consisting in the unwillingness or at least the inability of the latter to stop attacks committed by the former – is thus ultimately required by the condition of necessity of self-defence and not by any independent (customary) rule.Footnote 68
3.2. Other conditions
As demonstrated above, recent state practice is certainly of the utmost importance for assessing whether the right of self-defence may now be exercised in response to attacks by non-state actors, even when no state is substantially involved in these attacks. But one should not fail to notice that this practice also sheds some light on two other fundamental features of the law of self-defence which have greatly influenced its evaluation – that is, the gravity of the attacks against which there was a response in self-defence (subsection 3.2.1) and the proportionality of this military response (subsection 3.2.2).
3.2.1. The gravity of the attacks
In the aforementioned state practice, the attacks committed by non-state actors, while not consisting of a massive invasion of the territory of another state, was still considered to be serious enough to trigger the victim state's right of self-defence. Indeed, states recognized as a matter of principle that Israel enjoyed the right to respond in self-defence to the Hezbollah and Hamas attacks, even though these consisted only of a succession of rocket firings. Similarly, the Turkish incursion into Iraq has been condoned by third states although it was responding to a series of low-intensity attacks by the PKK.
Two main conclusions may be inferred about the gravity criterion from this practice. First, it seems to confirm that the gravity of an armed action against which there is a response in self-defence should be assessed against the background of a number of attacks which closely succeed one another, provided that they are linked by the same aggressive intention. The Israeli and the Turkish operations were evaluated against the continual firing of rockets and the repeated attacks by the PKK respectively. In other words, recent state practice seems to support the theory qualified by the literature as the ‘accumulative theory’, a theory which has been implicitly endorsed by the ICJFootnote 69 and never condemned by the Security Council.Footnote 70
Second, this practice clearly confirms that an armed action must reach a certain level of gravity in order to trigger the right of self-defence. Indeed, the attacks from Lebanon, Gaza, and Iraq were not seen as mere private (transnational) criminal acts allowing the victim state to take only (transnational) police actions. They were considered, mainly because of their level of gravity – evaluated on the basis of aggregated facts – to be true armed attacks against which the law of self-defence could be exercised. In other words, their gravity evidenced that they were not to be assessed merely in the light of the criminal law rules but more seriously, in the light of the rules of international law governing the use of force.
Having said that, it is difficult to ascertain whether recent state practice supports the traditional opinion that any armed action must reach a still higher level of gravity than is required in order to be considered more than a criminal act, that is, a use of force in the meaning of international law. According to this traditional opinion, there would be a gap between the notion of use of force, as provided in Article 2(4) of the UN Charter, and the notion of armed attack, used in Article 51. The latter would require a higher level of gravity than the former.
Many scholars, following the ICJ jurisprudence,Footnote 71 strongly defend this traditional position, notably on the grounds that it would prevent a massive reaction to a minor incidentFootnote 72 or the interference by a third party in a minor conflict between two states (which could result from the application of the right of collective self-defence).Footnote 73 These scholars are nevertheless aware of the undesirable effect of such a position, which may prevent a state attacked by minor uses of force from responding, even when there is no other available means to defend itself. In an effort to attenuate this effect, some of them have argued that, if a minor use of force cannot trigger the right of self-defence, it can nevertheless allow the victim to take proportionate measures, implying some forcible action.
This position may be founded on some ambiguous statements by the ICJ in the Nicaragua case.Footnote 74 It has been strongly supported and developed by Judge Simma in his opinion attached to the ICJ judgment in the Oil Platforms case.Footnote 75 It may be implicitly inferred from the resolution adopted in October 2007 by the Institute of International Law on the law of self-defence.Footnote 76 More recently, similar considerations have been expressly made by authors with respect to the Israeli intervention in Gaza: although having explicitly argued that the firing of rockets from Gaza was not serious enough to amount to an armed attack and to trigger the Israeli right of self-defence, these authors conceded that Israel was nevertheless entitled to use proportionate force in response.Footnote 77
These considerations actually undermine the traditional position, according to which only a significant use of force, in the meaning of Article 2(4) of the UN Charter, may trigger the right of self-defence. Indeed, one may call into question the relevance of establishing a threshold under which no force can be resorted to in self-defence, if some force may still be used, although under another denomination, for the same objective – that is, to defend the victim state – and under the same conditions – that is, proportionality, necessity, and a report to the UN Security Council of the forcible response.Footnote 78
More fundamentally, the traditional conditions under which the right of self-defence may be exercised, namely proportionality and necessity, render such a threshold superfluous and the reasons supporting it unjustified. Indeed, what matters is not that the attack against which self-defence is exercised is sufficiently serious but that this exercise is proportionate to the gravity of the attackFootnote 79 and is resorted to only when there is no other available means for the victim state to protect itself, including minor responses such as police action.
In any case, the question of whether there is a conceptual gap between the notions of use of force and armed attack remains a scholarly question. Indeed, the recent state practice brings nothing new to the debate.
3.2.2. Proportionality of the self-defence action
As has been the case for most of the actions undertaken in self-defence in the past, proportionality proved to be crucial in the assessment of the legality of the Israeli and Turkish operations. The Israeli interventions in Lebanon and Gaza were unanimously condemned because of their disproportionate nature,Footnote 80 while most of the states commenting on the Turkish operation were willing to approve this operation provided only that Turkey would use proportionate force in countering the PKK rebels.Footnote 81
Also in line with past state practice, recent state practice clearly evidences that states usually support a quantitative conception of proportionality, that is, a conception requiring a balance between the damages caused and the military means used by the attackers and the damages caused and the military means used by the state acting in self-defence.Footnote 82 This conception, while apparently endorsed by the ICJ,Footnote 83 has been opposed by some scholars, who argue for a teleological conception of proportionality, that is, a conception requiring that the action in self-defence must be limited to what is necessary to achieve its objective – to defend the victim state.Footnote 84 In this conception, proportionality is actually confused with – one of the several aspects of the condition of – necessity.Footnote 85
In fact, the quantitative and teleological conceptions are not antinomic. Indeed, it is very difficult to evaluate in practice what is necessary for a state to defend itself, since such an evaluation may be founded on very subjective factors. Comparing two objective quantities and analysing, as states do, whether the action taken in self-defence was not manifestly disproportionate by comparison with the attacks having triggered it, actually provide a useful means for evaluating whether this action was limited to what was necessary for the victim state to protect itself. By emphasizing such a disproportion, states and the ICJ intend not to proceed to mere cynical and sophisticated calculations but to show that the self-defence action went beyond what was necessary for achieving its objective or, more precisely, was only necessary to achieve another goal, such as reprisals or exploitation of natural resources. In other words, quantitative proportionality – that is, proportionality based on objective elements – may be used as a tool for evaluating the respect of the teleological proportionality, that is, the condition of necessity, the assessment of which can normally be based only on a variety of subjective elements.
This conclusion does not mean that every manifest disproportion between the attacks and the action in self-defence implies that the latter must be judged unnecessary. As it is argued, teleological proportionality (or necessity) is the ultimate condition against which any exercise of self-defence must be tested, quantitative proportionality being only, as is suggested by state practice and the ICJ, one of the easiest tests for evaluation in this respect. In other words, while the latter allows us to make an easy prima facie judgement regarding the fulfilment of the former, this judgement may be rebutted by proving that the action in self-defence, although being manifestly quantitatively disproportionate, was actually necessary.
This position is well illustrated by recent state practice, especially the Israeli operations. Indeed, the manifest disproportion between the deaths caused to the Lebanese and the Palestinians by the Israeli forces and the deaths caused to the Israeli people by the rockets led most states to make a prima facie judgement on the legality of the operation – that is, that the operation was illegal. This disproportion actually evidenced, in the states’ view, that the Israeli operations exceeded what was necessary to protect Israeli citizens and, even more, that Israel was actually pursuing other objectives, such as the destruction of Lebanon's economic capacity or the elimination of Hamas. But Israel, supported by other states, worked hard precisely in order to convince that its action, although disproportionate, was necessary, since rockets continued to be launched from Lebanon and Gaza. In other words, it tried to reverse the judgement of illegality made by the international community originally based on the quantitative proportionality test by emphasizing the necessity of pursuing military action in order to end the rocket attacks.
4. Concluding remarks
Although they are not absolutely clear, international reactions to recent state practice contrast sharply with the condemnatory attitude generally adopted by states before Operation Enduring Freedom, regarding military actions undertaken in self-defence in response to attacks by non-state actors. The right of Israel to react in self-defence to the rocket attacks from Lebanon and Gaza has been widely recognized as a matter of principle, although no state was considered as being substantially involved in the carrying out of these attacks. Likewise, the international community refrained from condemning the Turkish operation in Northern Iraq, which was undertaken ‘under the laws of self-defence’ in response to the PKK attacks from Iraq. These cases evidence that a state need not necessarily be substantially involved in attacks by non-state actors in order for the victim state to respond in self-defence.
Should one therefore consider that recent state practice evidences that the state attribution rules have actually been extended or adapted so that attacks by non-state actors may now be attributed to a state, even if the latter is not substantially involved in these attacks? This paper argues that such an interpretation would be erroneous. Indeed, relying on state attribution rules raises many problems and does not seem consistent. In addition, it does not conform to the declarations pronounced by states in recent state practice, according to which attacks were committed only by non-state actors and not by a state.
The correct interpretation of recent state practice amounts to considering that the latter evidences a clear tendency towards allowing states to act in self-defence in response to attacks, even if these attacks are committed only by non-state actors. Such a tendency had already emerged in 2001 from the wide support given to Operation Enduring Freedom. As a result, one should conclude more precisely, if one does not analyse the recent state practice separately – that is, without taking into account Enduring Freedom – that this practice actually confirms the evolution of the law of self-defence. In other words, in view of recent state practice, which supplements Enduring Freedom, private armed attacks nowadays seem to be enough to trigger the right of self-defence of the victim state.
This conclusion does not mean that no link has to be established between the state acting in self-defence and the state in whose territory the action takes place. Such a link, clearly evidenced by recent state practice, is in fact required by the condition of necessity of the law of self-defence. It serves to demonstrate that the harbouring state is at least unable or unwilling to prevent attacks from being committed from its territory so that its police forces cannot be considered by the victim state an available means of protecting itself and that resorting to self-defence therefore becomes necessary.
These are the lessons that can be drawn from recent state practice regarding the right to respond in self-defence to attacks by non-state actors even when no state is substantially involved in these attacks. Recent state practice also provides some valuable indications concerning the level of gravity that an attack must reach in order to trigger the right of self-defence and the proportionality of the action undertaken in self-defence.
As far as the gravity criterion is concerned, it confirms that successive minor uses of force, linked by the same aggressive intention, may be accumulated and regarded as a single use of force. It also confirms that any use of force, particularly by non-state actors, must reach some level of gravity in order to be considered as being more than private transnational crimes – that is, amounting to a use of force in the meaning of international law. It does not, however, provide any clear indication on the issue concerning the ‘gravity’ gap between the notions of use of force and armed attack.
As far as the notion of proportionality is concerned, recent state practice confirms that the quantitative test, which consists in verifying whether the self-defence action was manifestly disproportionate to the attacks, is favoured. However, this practice also clearly supports the opinion that this test is in fact used in order to assess the fulfilment of the condition of necessity and that a judgment of illegality based on such a test could eventually be reversed by proving the necessity of the self-defence action in the case under consideration.
To sum up, recent state practice greatly contributes to the evolution of the law of self-defence, mainly to the extent that, analysed altogether with Operation Enduring Freedom, it implies the recognition of the right to respond in self-defence to private armed attacks, but also because it gives weight to some controversial interpretations of the classical conditions of this law, such as that according to which the gravity of the attacks must be assessed by aggregating a succession of minor – but linked – acts of force, or such as that according to which the condition of proportionality implies a quantitative assessment of the action taken in self-defence, but only as a means of making an easy prima facie judgement on the necessity of this action.
