1. Introduction
Use of force issues have always been a subject of divided opinion in legal scholarship. Two main camps are generally identified in that regard: those who argue for a rigid application of the relevant rules, i.e., the ‘restrictivists’, and those who advance a more flexible approach to the matter, i.e., the ‘expansionists’.Footnote 1 Recent developments in the legal doctrine demonstrate that the orthodoxy on the law on the use of force has dramatically switched from a restrictivist to an expansionist perspective.Footnote 2
This article seeks to analyse this recent shift, especially with respect to the law of self-defence, from an expansionist point of view. The main purpose is to determine how both USFootnote 3 and non-US expansionists have reacted to this shift, including their arguments against restrictivism, and, more generally, to examine the state of the argumentative landscape which currently exists on the expansionists’ side about the law of self-defence. The crucial question in that regard is whether this argumentative landscape has changed and, if so, what issues such changes concern and what the new arguments used by expansionists are.
The article is divided into two sections. The first part observes the recent shift in the literature on the law of self-defence. It is purposely descriptive, as scholars have never addressed in detail the issue dealt with in this article.Footnote 4 It will present the main aspects of the law of self-defence with respect to which changes can be identified in the recent argumentative landscape existing on the expansionists’ side (Section 2). The second part is devoted to analysing the changing nature of the arguments used by expansionists in support of such evolution of the right of self-defence and, in particular, the recent move from arguments based on state practice to those based more on policy and pragmatic considerations. It will call into question the increasing recourse to policy oriented arguments and demonstrate that those arguments alone cannot justify any evolution of the law of self-defence, while emphasizing that state practice remains central in that respect and explaining the different ways through which this practice may play such a role (Section 3).Footnote 5
The article confirms that the legal scholarship on the use of force has clearly moved towards a broader conception of the law of self-defence. It concludes in particular that expansionists have recently succeeded in imposing their views upon the restrictivist camp with respect to important issues of this law, which enables some of them to not only no longer struggle and argue on those issues but also, and more importantly, to devote their current detailed arguments to strengthening and assessing in more detail their wide conception of the law of self-defence. However, the article observes that those new positions are not based on the traditional means generally used in legal scholarship to support any evolution of the law, including state practice. The arguments developed in that regard are mainly based on policy reasons, which remain for the most part highly subjective and, therefore, debatable.
2. Observing the argumentative changes
There are two main issues with respect to which changes in the argumentative landscape on the expansionists’ side are clearly noticeable regarding the law of self-defence: the right of self-defence in response to attacks by non-state actors (Section 2.1) and anticipatory self-defence (Section 2.2).Footnote 6 As will be seen, both issues evidence similar changes.
2.1. The right of self-defence in response to attacks by non-state actors
It has long been accepted that an armed attack triggering the right of self-defence does not only mean an armed attack committed by the military forces of a state on the territory of another state, but also includes large scale attacks committed by non-state actors.Footnote 7 Yet, in this case, the non-state actors must have been sent by a state or this state must have been substantially involved in sending them. This amounts to what is generally qualified in legal scholarship as an ‘indirect’ armed attack or act of aggression. In other words, such an attack is still an attack committed by a state, but through indirect means, i.e., by using non-state actors to attack another state.Footnote 8 What has remained unsettled for years is the right to use force in self-defence in order to respond to an armed attack which is not directly or indirectly committed by a state, but originates from non-state actors,Footnote 9 in the sense that no state is substantially involved in it. Article 51 of the UN Charter has traditionally been interpreted as allowing a state to resort to self-defence in response to an armed attack by another state,Footnote 10 mainly because, according to the majority of authors, self-defence is an exception to the prohibition on the use of force, which only applies to interstate relations.Footnote 11 However, the US military response in Afghanistan with respect to the 9/11 terrorist attacks and, especially, the supportive reaction of the international community to such response, have significantly influenced these orthodox positions. The majority of scholars, including those who could be called ‘restrictivists’, now recognize the right of states to respond in self-defence to any armed attack, whatever the state or non-state origin of such attack.Footnote 12 Such relaxation regarding the orthodox conception of the right of self-defence inevitably involves some changes with respect to the arguments usually used by expansionists on that issue. Such changes are twofold: the decreasing use of arguments in order to demonstrate the existence of a right to react in self-defence to armed attacks by non-state actors (Section 2.1.1) and an increasing focus on the conditions under which such right may be exercised (Section 2.1.2).
2.1.1 Arguments about the existence of the right
It is clear that expansionists no longer struggle to demonstrate that a right to act in self-defence in response to armed attacks by non-state actors exists. Before the US intervention in Afghanistan in response to the 9/11 terrorist attacks, authors espousing a wide conception of self-defence generally tried to establish such a right by relying on a series of detailed arguments. The latter included references to the famous Caroline incident;Footnote 13 instances in which the UN Security Council qualified attacks committed by non-state actors as an armed attack;Footnote 14 recourse to legal realist considerations, according to which it would be senseless that a state be obliged ‘to endure painful blows, only because no sovereign State [is substantially involved in the attacks]’;Footnote 15 references to Article 51 of the UN Charter as not expressly mentioning that an armed attack must come from a state to trigger the right of self-defenceFootnote 16 and, finally, interpretation of state practice.Footnote 17 This latter argument has been particularly well developed by expansionists.Footnote 18
This argumentative landscape has changed significantly in recent years. Today, a much more limited number of arguments are usually cited by expansionists – and, generally, by most authors – in support of a right of self-defence in response to armed attacks by non-state actors. Two main arguments may be identified in that respect: the first consists of emphasizing that many scholars currently acknowledge the existence of such a rightFootnote 19 – and, conversely, that very few authors still oppose it;Footnote 20 the second consists of referring, without necessarily examining it in detail, the approbation by the international community, in particular by the UN Security Council in Resolutions 1368 and 1373 (2001), of the US intervention in Afghanistan in response to the 9/11 attacks.Footnote 21 These attacks have therefore played a pivotal role in changing the arguments on the use of force used by legal scholars, and in particular by the expansionists. Such a change is clearly apparent in the last edition of Dinstein's seminal book on the use of force.Footnote 22 While the previous editions devoted several pages to the issue of attacks by non-state actors, the most recent edition is much briefer on that subject with Dinstein limiting himself to developing the two above-mentioned arguments.Footnote 23
Yet, expansionists are aware that the current ICJ case law apparently does not fit with their opinion on the right to use force in self-defence in response to attacks committed by non-state actors. In the expansionists’ view, the most problematic assertions made by the Court are those upheld in the Wall and Armed activities cases. Indeed, in the former case, the Court stated that ‘Article 51 of the Charter . . . recognize[d] the existence of an inherent right of self-defence in the case of armed attack by one State against another State’,Footnote 24 while it asserted, in the Armed activities case, that there was not sufficient proof that attacks by the irregular forces located in DRC against Uganda were attributable to the DRC and concluded that Uganda could not therefore act in self-defence; this conclusion has been interpreted by scholarsFootnote 25 as requiring that an armed attack must be attributed to a state to amount to an armed attack under Article 51 of the UN Charter.Footnote 26 There are different ways through which expansionists have dealt with those assertions. The first is to acknowledge that the law upheld by the Court does not coincide with the law deriving from state practice and to implicitly consider that the latter is the only one reflecting international law.Footnote 27 Another way, which is the one most frequently used to address the problem, is to criticize the Court's assertionsFootnote 28, most often by referring to the strong criticism already raised in that regard by Judges Kooijmans,Footnote 29 BurgenthalFootnote 30 and HigginsFootnote 31 in their separate opinions. A third and final way is to interpret the Court's assertions in order to demonstrate that such assertions do not necessarily conflict with the view that attacks committed by non-state actors may trigger the right of self-defence. In this sense, Michael Wood argues, in relation to the ICJ assertion in the Wall case, that ‘it seems that the Court was merely reflecting the obvious point that unless an attack on a state is directed from outside that state's territory the question of self-defence does not arise’.Footnote 32 In the same way, Kimberley N. Trapp interprets the ICJ conclusion in the Armed Activities case as not
ruling out the legitimate use of defensive force against non-State actors unless the armed attacks of such non-State actors are attributable to a State [since] the Court's [decision] should be understood as requiring that armed attacks be attributable to a State if the State itself is to be the subject of defensive uses of force [and therefore not if the non-State actors are the only targets of such uses of force].Footnote 33
2.1.2. Arguments on the conditions for the exercise of the right
While the number of arguments used by expansionists to demonstrate the existence of a right of self-defence in response to armed attacks by non-state actors has significantly reduced, most expansionists now increasingly focus their arguments on the conditions for the exercise of such a right. Although some expansionists already addressed those conditions before the events of 9/11,Footnote 34 one can clearly observe an increasing focus on such conditions in the current expansionist legal scholarship. As rightly emphasized by one author: ‘The academic debate no longer seems to be preoccupied with determining if a right of self-defence against private actors exists, but rather how such a right is to be exercised’.Footnote 35 The recent discussions on that issue in the American Journal of International Law are very illustrative of such a shift. These discussions were initiated by Daniel Bethlehem, former Legal Adviser to the United Kingdom Foreign and Commonwealth Office, and were motivated by the fact that, in his opinion:
It is by now reasonably clear and accepted that states have a right of self-defense against attacks by non-state actors – as reflected, for example, in UN Security Council Resolutions 1368 and 1373 of 2001, adopted following the 9/11 attacks in the United States – [but there] is, however, a paucity of considered and authoritative guidance on the parameters and application of that right in the kinds of circumstances that states are now having to address.Footnote 36
Bethlehem therefore proposes a series of principles that should guide the exercise of the right of self-defence in response to private armed attacks.
These principles reflect a wide conception of such a right. In particular, principle 7 recognizes the right to act in self-defence not only against the non-state actors who directly perpetrated the armed attack but also against any of those: ‘taking a direct part in [these] attacks through the provision of material support essential to the attacks’.Footnote 37 This may cover a wide range of actors, including any state materially supporting the attackers though not necessarily hosting them. This principle stands in stark contrast with the majority view according to which, if a right of self-defence is allowed in case of private armed attacks, only the non-state actors may be targeted.Footnote 38 A particularly wide conception of self-defence is also noticeable in principle 12. This principle does not require seeking the consent of a state before targeting the non-state attackers located on its territory when it is established that this state is unable to effectively restrain the armed activities of those non-state actors and that:
seeking of consent would be likely to materially undermine the effectiveness of action in self-defense . . . or would increase the risk of armed attack, vulnerability to future attacks, or other developments that would give rise to an independent imperative to act in self-defense.Footnote 39
Yet seeking consent is generally considered one of the last attempts for the victim (or threatened) state to protect itself without resorting to self-defence and, therefore, as indirectly required by the condition of necessity which imposes that any action in self-defence be taken only as a last resort – that is, after all the available alternatives have been exhausted.Footnote 40 Some scholars nonetheless go even further on that issue and argue that the right of self-defence may be exercised on the territory of a state even without requiring that it be demonstrated that the state was unable or unwilling to stop the attacks.Footnote 41
Although both this view and the Bethlehem principles are claimed to be (partially) based on state practice, the former only refers to limited state practice, mainly the recent US counterterrorist interventions in Pakistan and Afghanistan,Footnote 42 whereas the Bethlehem principles are essentially drawn from concerns raised in some intra- and intergovernmental circles on the practicability and operational nature of the law of self-defence.Footnote 43 More generally, they are the result of taking the view ‘that the credibility of the law depends ultimately upon its ability to address effectively the realities of contemporary threats’.Footnote 44 In other words, expansionist views are mainly driven by pragmatic considerations rather than a detailed analysis of the whole state practice concerning the matter.
2.2. Anticipatory self-defence
The other major issue with respect to which changes in the argumentative landscape of the expansionists’ side are clearly noticeable is the right to act in self-defence before an armed attack occurs, i.e., anticipatory self-defence. Such a right has long been discussed in legal literature. The majority view was traditionally opposed to such a right, with few scholars supporting its legality.Footnote 45 Yet the academic debate has dramatically changed since the US published its National Security Strategy in 2002 (NSS).Footnote 46 This document was created by the Bush administration in the context of the US counter-terrorist measures adopted after the 9/11 attacks and just before the US invasion of Iraq in 2003. While providing that ‘[f]or centuries, international law recognized that nations need not suffer an attack before they can lawfully take action to defend themselves against forces that present an imminent danger of attack’, the NSS nonetheless emphasizes that such a legal position must be adapted ‘to the capabilities and objectives of today's adversaries’,Footnote 47 and that the United States may therefore use force in order to pre-empt ‘emerging’ (i.e., non-imminent) threats. In other words, the document makes a clear distinction between two kinds of anticipatory self-defence: one exercised in response to an imminent threat of armed attack, which it considers as well-established in international law and which may be qualified as pre-emptive self-defence; and the other exercised in order to prevent a non-imminent threat of armed attack, that the NSS seems to consider on a de lege ferenda basis and which may be qualified as preventive self-defence.Footnote 48 Such distinction and qualifications have been explicitly endorsed by the High Level Panel on Threats, Challenges and Change in its 2004 report. However, although that report acknowledges, as the NSS does, that ‘a threatened State, according to long established international law, can take military action as long as the threatened attack is imminent [that is pre-emptively]’,Footnote 49 it rejects, contrary to the US position, the right to act preventively in self-defence, i.e., against non-imminent threats, and recommends that the threatened state go before the UN Security Council in such situations.
With few exceptions scholars, including European academics, generally started to follow this view.Footnote 50 This may be surprising since, as already noted, legal literature was traditionally reluctant to admit any form of anticipatory self-defence. In fact, the debates surrounding both the NSS and the US invasion of Iraq, although evidencing a strong opposition to the US position for a right to self-defence against non-imminent threats,Footnote 51 nonetheless show a clear shift in the academic landscape, which has moved from a majority of authors traditionally opposed to any anticipatory self-defence towards a legal scholarship which now generally agrees that a specific type of anticipatory self-defence may be resorted to, i.e., the one exercised in response to imminent threats. In other words, as rightly emphasized by one author, ‘[w]hile the notion of [anticipatory] force against non-imminent threats has not been accepted, a by-product of the Bush doctrine appears to be greater explicit support for the more limited right of anticipatory self-defence in relation to imminent threats’.Footnote 52 Such a shift has brought major changes to the argumentative landscape existing on the expansionist side, both with respect to the existence of a right of self-defence against imminent threats (Section 2.2.1) and to the scope of the condition of imminence which is required for exercising such a right (Section 2.2.2).
2.2.1. Arguments on the right of self-defence against imminent threats
Prior to the academic debates on the Bush doctrine embodied in the 2002 NSS, scholars who supported a wide conception of the law of self-defence, in particular a right to take anticipatory measures in self-defence, used to develop many arguments to back up their view. The two most important series of arguments were based on an in-depth analysis of the Caroline incident, dating back to 1837,Footnote 53 as well as a detailed scrutiny of state practice since the adoption of the UN Charter.
The diplomatic letters exchanged about the sinking of the Caroline were viewed as confirming the existence of a right to take anticipatory measures in self-defence as well as providing the precise parameters for the exercise of such right,Footnote 54 that is, according to the formula used by Webster, the US Secretary of State, ‘a necessity of self-defence, instant, overwhelming, leaving no choice of means, and no moment for deliberation’.Footnote 55 More generally, it was argued that the agreement evidenced in those letters on the law applicable to such situations reflected the state of the customary law of self-defence of that period and that such customary law not only survived until the adoption of the UN Charter but also, and more importantly, remained unaffected by it, since it was claimed that those who drafted Article 51 of the UN Charter did not want to modify pre-existing law.Footnote 56 State practice post-UN Charter is also analysed and considered as confirming the continuing validity of this old customary law.Footnote 57 Several precedents are mentioned and scrutinized in that respect, including: the Cuban missile crisis (1962), the Israeli-Arab War (1967), and the Israeli attack against the Iraqi nuclear reactor (1981). In addition to those arguments, expansionists also base their view on pragmatic grounds, in particular that it would be unrealistic to expect a state ‘to be a sitting duck’ and wait until ‘the bombs are actually dropping on its soil’.Footnote 58
Such argumentative landscape has significantly changed since 2002. A relaxation of the restrictivist camp with respect to the right of anticipatory self-defence has naturally led to a significant decrease in the arguments that expansionists used to develop in order to support the existence of such a right, at least as far as the right to respond to imminent threats is concerned. The Caroline incident is still mentioned but most often only briefly and without elaborating its link to customary law.Footnote 59 It is mainly referred to when explaining the traditional conditions under which anticipatory self-defence may be exercised and as the starting point for a discussion on the imminence requirement.Footnote 60 Most scholars content themselves with referring to and quoting from the 2004 report of the High Level Panel on Threats, Challenges and Change as well as the 2005 UNSG report ‘In Larger Freedom’,Footnote 61 which endorses the conclusion of the former.Footnote 62 Basically, expansionists no longer seem preoccupied with establishing the existence of a right to respond in self-defence to imminent threats of armed attack since, as they claim, such a right is now (if not universally at least) widely accepted.Footnote 63 They focus their new arguments on the scope of the notion of imminence with a view to broadening its traditional meaning.
2.2.2. Arguments on the scope of the notion of imminence
The notion of ‘imminence’ of an armed attack, as required for the exercise of anticipatory self-defence, has not been the object of any in-depth discussion in legal literature before the 2002 US NSS. Its scope was generally construed in the light of the Caroline incident, as reflected in the above-mentioned Webster formula, which implies a very strict temporal requirement and makes ‘imminence’ synonymous with ‘immediacy’.Footnote 64 Expansionists were certainly satisfied with the growing consensus which resulted from the academic debate following the NSS document on the legality of a right to use force in self-defence against imminent threats. Most of them even expressly gave up the idea of arguing in favour of a right to preventive self-defence, that is, a right of self-defence against remote threats.Footnote 65 This is particularly interesting as it shows a pragmatic withdrawal by the expansionists from more extensive positions due to the recognition by orthodoxy of a slightly extensive one. That having been said, expansionists were not ready to endorse the traditional meaning of the imminence requirement. Most of the arguments which are still developed by expansionists on the matter are therefore essentially devoted to calling into question such traditional meaning.
Admittedly, some of those scholars are reluctant to question the temporal dimension of the imminence requirement. Yet they argue that the temporal test of the Webster formula must be broadly interpreted and that such ‘interpretation must not restrict measures to mere reaction or interception of an attack, but rather includes actions of a truly anticipatory character in the face of a clear and concrete threat of an attack within the foreseeable future’.Footnote 66 Moreover, only a limited number of expansionist scholars share such a view. Most of them criticize the pure temporal dimension of the imminence requirement and propose a series of non-temporal criteria for assessing the imminence of an attack or, more fundamentally, the legality of an anticipatory action in self-defence. The most cited criteria include the probability of the attack; the nature and magnitude of the threat involved and, as a result, the estimated gravity of the harm caused by the future attack; the availability and exhaustion of alternatives to using force and, in particular, the availability of possible windows of opportunity for preventing the attack; the capability and intent of the threatening actor to launch an attack; and a high level of evidence with respect to those issues.Footnote 67 Opinions diverge with respect to the legal basis of such criteria. Some merely consider that they are imposed by this new construed requirement of imminence, which departs from the pure temporal requirement expressed in the Caroline case;Footnote 68 others conceive those criteria as stemming from this new type of imminence requirement, but establish a link between the new conception and the traditional condition of necessity of the law of self-defence,Footnote 69 imminence being interpreted as included in this condition. Finally, others support a more extensive view on the matter by contesting the applicability of the imminence requirement to contemporary uses of anticipatory self-defence and by considering that the aforementioned criteria are basically imposed by the traditional condition of necessity.Footnote 70
The main reason underlying all the aforementioned positions comes close to that set out by the NSS: the need to adapt the traditional right of anticipatory self-defence to the new threats of today's world, being mainly threats of attacks by terrorist groups and the potential use by such actors or any (rogue) state of weapons of mass destruction, such as nuclear weapons.Footnote 71 Many arguments inspired by legal-realist and strategic considerations are developed on this basis. It is, for example, increasingly claimed that, because of the magnitude of the harm that may be caused by the new weapons, there are greater risks for a threatened state to await the (temporal) imminence of an attack since it would be no longer in a position to defend itself when the attack occurs.Footnote 72 Basically, all those arguments stem from the view that ‘[l]aws that impose unrealistic standards are likely to be violated and ultimately forgotten’.Footnote 73
3. Analysing the changing nature of the arguments
As already noted, the evolution of the argumentative landscape existing on the expansionist side with respect to important issues of the law of self-defence shows that expansionists tend to pay less attention to state practice and increasingly focus on policy as well as pragmatic considerations to support their (new) positions. One must therefore examine whether policy/realist considerations may constitute valid means through which the law of self-defence may evolve (Section 3.1.). As it will be concluded that it is not the case, one will then turn towards what may arguably be considered as the most relevant basis for assessing the evolution of the law of self-defence, i.e., state practice, and analyse the legal conditions under which this practice may play such a role (Section 3.2).
3.1. Evolution based on new facts and realities alone
Although expansionists have always relied on policy and pragmatic considerations to support their views, such considerations tend to become their major arguments. They increasingly invoke new facts or realities in isolation in order to widen their conception of the law of self-defence. One must first analyse whether such an approach is not the result of an (implicit) application of more traditional argumentative positions and could not therefore be considered as a well-founded approach to assess the evolution of the law of self-defence on such a basis.
The first traditional argumentative method to which the policy oriented approach could be linked is the interpretation of a norm in light of its object and purpose. Such a method, implying a ‘teleological interpretation’ of the norm, is expressly provided, with respect to treaties, under Article 31 of the Vienna Convention on the Law of Treaties. It could be argued that scholars who support the idea that the law of self-defence should be adapted given the emergence of new facts and realities apply such an argumentative method, as their opinion would imply that, without the adaptation of that law, its purpose for states, i.e., to defend themselves against an armed attack, could no longer be achieved. In other words, according to those scholars, the law of self-defence should evolve in order to remain able to fulfil its objective. This raises two comments.
First, such a method of interpretation is not expressly relied upon by scholars who adopt a policy-oriented approach, as there is no express reference to it and no mention of the Vienna Convention on the Law of Treaties. It is therefore doubtful that their intention is to base their approach on this traditional interpretative method. Such a silence contrasts with the attitude adopted by some other scholars, who explicitly refer to the ‘object and purpose of Article 51 of the Charter’ as a means of interpreting that Article in accordance with the Vienna Convention on the Law of Treaties and in order to support their restrictive view on the law of self-defence.Footnote 74 Second, relying only on this interpretative method does not seem to be sufficient to support any evolution of the law of self-defence. Indeed, the object and purpose of that law and, more generally, of the UN Charter, as well as the possible ways to achieve such a purpose are so malleable that they may be used in order to support wide as well as restrictive conceptions.Footnote 75 This is why scholars who expressly refer to this interpretative method take the trouble to confirm the result of this application and to strengthen their position by resorting to other methods, including the one based on state practice.Footnote 76
This last method, which will be addressed in more detail below (Section 3.2), is the second traditional one that may be tempting to consider as (implicitly) applied by scholars who follow a policy oriented approach. Although, as will be shown, state practice cannot lead to any evolution of the law of self-defence if it is not combined with some opinio juris, it is sometimes argued in legal literature that scholars espousing a wide conception of that law generally tend to infer its evolution from state practice alone (or mainly from it), as they would focus on material facts, in particular state material conduct, whereas those supporting a restrictive conception would give priority to the opinio juris of states since they privilege states’ legal declarations.Footnote 77 Relying only on new facts and realities in order to assess the evolution of the law of self-defence would therefore be in line with the traditional expansionist approach based on state practice. This again raises two main comments.
First, it is doubtful that the aforementioned description of the current state of legal scholarship, which refers to state practice for assessing the law of self-defence, is entirely correct. As a matter of principle, state practice should not only be equated with material facts and, in particular, state material conduct. As emphasized in the recent work of the International Law Commission on the identification of customary law, the notion of state practice ‘comprise[s] both physical and verbal (written and oral) conduct’.Footnote 78 Indeed, according to a majority view,Footnote 79 state practice must be viewed as a broad concept which encompasses any manifestation of state action linked to an international issue, including state declarations. As a matter of fact, scholars espousing a wide conception of the law of self-defence do not rely only on state material conduct – which is logical since such conduct is itself meaninglessFootnote 80 – to argue that there has been an evolution of that law. More precisely, in contrast with the restrictivist side, they are ready to take into account any state manifestation, including state – political, moral, historical, or even legal – declarations.Footnote 81 Their aim is to collect as much state practice as possible in order to identify the real opinio juris and not that which is officially in the ‘mind’ of the intervening states, and to avoid asserting a legal discourse that would not correspond to the actual conduct of states.Footnote 82 Basically, proponents of both approaches distinguish themselves by the fact that they focus on different types of state practice as a means for revealing the opinio juris of states, not because some focus on state material acts and others on opinio juris.Footnote 83 All of this shows that no scholar, even those espousing a wide conception of the law of self-defence, seems to have ever claimed that state practice alone, in particular without being coupled with the opinio juris of the relevant states, may lead to the evolution of that law. This therefore stands in contrast with the recent trend in legal scholarship to rely on new realities alone in order to extend the limits of the law of self-defence.
Second and more fundamentally, the concept of state practice must not be confused with the notion of (new) realities. The latter are indeed pure facts which, contrary to state practice, are not necessarily related to states or expressly linked to a specific issue of international law. Relying only on them therefore raises a range of new questions, including whether they can lead to an evolution of the law. Yet, as has been correctly pointed out by one author, new facts or realities alone cannot change law ‘if law is a rule, a norm in any sense of the word, if law is more than a simple description of what happens.’Footnote 84 It is true that, as emphasized by scholars,Footnote 85 the reason for giving such a role to facts is driven by the very idea that law should espouse, as much as possible, (new) realities; otherwise it would remain just a law on paper and would fail to regulate state behaviour. In other words, it would become ineffective. In this way, the restrictive approach to the regulation of interstate use of force should be softened in order to prevent it losing its regulatory force, leading to a legal vacuum on a matter that is so essential for international stability. Yet, it is not certain that the effectiveness of a norm depends upon its capacity to be consistent with reality. The effectiveness of a norm is a complex meta-legal notion.Footnote 86 As argued in detail elsewhere,Footnote 87 such a notion can be linked to the capacity of a norm to regulate the relationships between its addressees and to influence their behaviour. Such capacity can stem from the mere power of legitimation that a norm may offer to the party invoking it and international law, including regulations on the use of force, is far from being devoid of such power. The effectiveness of this regulation is therefore not a matter of strict conformity to reality.
In addition, policy and realist preoccupations, as well as the solutions proposed to meet those preoccupations, may vary greatly from one (expansionist) scholar to another. Indeed, they involve highly subjective assessments, whose process does not follow a clear legal framework – like the Vienna Convention on the Law of Treaties. Any new law asserted on the basis of such proposed solutions would therefore pay more attention to the opinions of scholars than the opinio juris of states. In fact, it seems that such solutions are generally presented from a de lege ferenda perspective. The objective is less to establish the current state of international law than to push (states) to create a new law which would be more adapted to the current realities.
3.2 Evolution through state practice
State practice remains central for the assessment of the evolution of the law of self-defence. The other traditional means for assessing such evolution do not seem satisfactory, especially if they are not accompanied by scrutiny of state practice. Like the object and purpose criteria, the ordinary meaning of the terms used in Article 51 of the UN Charter as well as the context surrounding this Article are so flexible that they may be invoked to justify restrictive as well as extensive conceptions of the law of self-defence.Footnote 88 In addition, recourse to the preparatory works may only shed light on the meaning of the law of self-defence at the time of the writing of the UN Charter. It does not therefore come as a surprise that one can hardly find a scholar, except those only resorting to a policy-oriented approach, who has dispensed with analysing state practice when discussing use of force issues.
According to the majority of scholars, both restrictivists and expansionists, state practice may lead to the evolution of the law of self-defence provided that some conditions are met.Footnote 89 It is generally acknowledged that the conditions under which a norm may evolve usually depend upon the formal source of this norm. Since it is widely admitted that the right of self-defence predated Article 51 of the UN Charter as a customary rightFootnote 90 and that such a right is now regulated by both customary and conventional norms,Footnote 91 this suggests that one should distinguish between the evolution of the customary versus conventional aspects of the law of self-defence through state practice. In addition, there may be different forms of evolution of a norm through state practice, namely, the evolution of that norm as such through its modification, and the evolution of its meaning through its interpretation.
Given these two possible distinctions, i.e., evolution of the ‘conventional’ versus ‘customary’ aspects of the law of self-defence (Section 3.2.1) and ‘interpretation’ versus ‘modification’ of (the meaning of) that law (Section 3.2.2), there are multiple ways in which state practice may be – and has been – claimed to lead to such evolution, including the emergence of new customary law, modifying the previous one,Footnote 92 or leading to a new interpretation of Article 51 of the UN Charter,Footnote 93 or through the modification or direct interpretation of this Article.Footnote 94 These distinctions must however be scrutinized in more detail in order to determine the conditions under which state practice may validly lead to the evolution of the law of self-defence.
3.2.1 Customary versus conventional evolution
One may wonder how to identify the evolution of the interrelated customary and conventional sources of the law of self-defence through state practice. Should one differentiate the evolution of the various aspects of that law according to the formal source upon which they rest? An affirmative answer would be particularly difficult to put into practice regarding the debated condition of an armed attack, which is generally considered as being both of a customary and a conventional nature.Footnote 95 It could first be claimed, in line with the famous paradox advocated by Baxter,Footnote 96 that it is currrently nearly impossible to establish the state of the law of self-defence outside of Article 51 of the UN Charter, since almost all states are party to the Charter and therefore the practice of UN non-member states, which is the relevant state practice for the identification of the customary law of self-defence, is no longer available. This problem could apparently be overcome by paying particular attention to the legal justifications invoked by states when they resort to force and assessing in each specific case whether their practice evidences their belief in acting only in relation to Article 51 or, more generally, in accordance with customary law, that is, following a terminology used in legal scholarship,Footnote 97 whether such practice is to be associated only with their opinio juris conventionalis (based on treaty) or, more broadly, with their opinio juris generalis (based on custom). Kolb asserts that, ‘[w]ith respect to [the problem of this] split of opiniones iuris (conventionalis and generalis)’ arising in cases of largely ratified treaties:
. . . [t]he opinion and practice of the treaty-parties will also merit scrutiny [in order to establish that customary law has developed along the lines of the treaty-regime], since the implementation of the treaty may well be linked with a conception that the course of action it prescribes is the most reasonable and convenient way to deal with a matter, [and that] if such statements are made, they may be taken as an expression of a general opinio juris.Footnote 98
In the same way, Meron argues that:
. . . [i]f it could be demonstrated that in acting in a particular way parties to a convention believed and recognized that their duty to conform to a particular norm was required not only by their contractual obligations but by customary or general international law as well . . ., such an opinio juris might and should be given probative weight for the formation of customary law . . . .Footnote 99
In other words, while, as already detailed above, state practice must always be associated with some opinio juris, in order to lead to any evolution of the law of self-defence, one should assess whether this opinio juris is of a particular (treaty) versus general (non-treaty) nature in order to assess whether this practice has possibly led to the evolution of the conventional versus customary aspects of that law.
However, this solution is hardly applicable with respect to the law of self-defence. It is, indeed, nearly impossible to establish in practice which of the two relevant sources, customary or conventional law, states are considering when they refer to the law of self-defence and, therefore, to differentiate between the evolution of each of these sources in practice. It is true that states usually invoke Article 51 of the UN Charter when referring to the right of self-defence. Yet, this does not necessarily mean that only conventional law is of relevance, as Article 51 itself refers to customary international law.Footnote 100 By invoking this Article, states may only have intended to refer to the customary right of self-defence as recognized by Article 51, that is, which predated (or survived after) the adoption of the UN Charter.Footnote 101 The same problem arises when references are made to UN General Assembly Resolutions in relation to the law of self-defence, such as Resolutions 2625(XXV) and 3314(XXIX), as they may be considered either an expression of customary international law,Footnote 102 or an (authoritative) interpretation of Article 51.Footnote 103 Similarly, although states sometimes invoke customary international law, mainly by referring to state practice predating the UN Charter, it may be that such practice is mentioned simply as a means of interpreting Article 51,Footnote 104 rather than as an instance of ‘proper’ customary practice. Finally, it is particularly difficult to identify the exact source to which states intend to refer precisely when such identification would be the most useful, that is, with regard to the controversial applications of the law of self-defence; in particular actions taken in self-defence against armed attacks by non-state actors or against imminent armed attacks, since such actions may be claimed to be based either on customary international law or on a specific interpretation of Article 51.
Although Article 51 is frequently invoked by states, in practice it seems to be used as a general umbrella under which all the aspects of the right of self-defence are considered, without formally giving to these aspects a conventional nature. Article 51 is simply the provision with which the entire law of self-defence is generally associated. It is interesting to note in this respect that the conditions of necessity and proportionality, being normally of a customary nature, are sometimes presented as being imposed by Article 51.Footnote 105 This does not, however, mean that one must dismiss the idea that the right of self-defence is regulated by both customary and conventional law. Actually, the difficulty in identifying the precise source of the different aspects of the law of self-defence in practice is not a matter of principle but a matter of proof and methodology. As Meron rightly argues (in the context of his attempted response to the Baxter paradox):Footnote 106
To be sure, it is difficult to demonstrate . . . that in acting in a particular way, parties to a convention believed and recognised that their duty to conform to a particular norm was required not only by their contractual obligations but also by customary or general international law, but this poses a question of proof rather than of principle.Footnote 107
This means that, although the law of self-defence is both of a customary and a conventional nature and is regulated by both sources, the assessment of its evolution through state practice should not distinguish between its conventional and customary aspects.
Moreover, the conditions upon which customary and conventional law may evolve through state practice do not seem that dissimilar. State practice may lead to the creation of a customary rule only if this practice is followed out of a belief in conforming to such a rule and if the practice is general and constant. Similarly, as detailed and argued below (Section 3.2.2), subsequent state practice may serve as a means for interpreting or modifying a treaty if this practice is followed in relation to that treaty and if it establishes the agreement of the parties regarding its interpretation or modification. This actually implies that for such an agreement to be clearly identifiable, state practice should be repeated over time and approved by the other parties, namely almost all the states, given that the treaty to be interpreted or modified is the UN Charter. As a result, it seems necessary to only show that three general conditions have been fulfilled in order to assert the evolution of the law of self-defence through state practice, namely that this practice is: relevant (i.e., followed in the application of the law of self-defence), general,Footnote 108 and constant, without seeking to differentiate between conventional and customary practice (i.e., to establish whether the practice is followed out of the belief that it is required by conventional or customary law).
3.2.2. Evolution through modification versus interpretation
Contrary to the distinction between the customary and the conventional evolution of the law of self-defence through state practice, the second distinction, i.e., interpretation versus modification of (the meaning) that law, seems to be relevant. As rightly pointed out by Ruys when considering the evolution of a norm:
[t]he key criterion [for distinguishing between interpretation and modification] is (in)compatibility: . . . [m]odification can arguably be defined as the situation where the new rule cannot be fit in any of the plausible meanings that could be given to the treaty text, nor into the special meaning which the parties intended to give to the text at the time of its adoption . . . Determining whether a situation of incompatibility exists is itself a matter of interpretation.Footnote 109
In accordance with such considerations, one may argue that any claimed evolution of the law of self-defence through state practice with regard to some issues, in particular the right to respond in self-defence to attacks by non-state actors and anticipatory self-defence, should be seen as an attempted interpretation rather than modification of that law. Indeed, regarding the right to react in self-defence to private armed attacks, although this right was not expressly acknowledged at the San Francisco Conference in 1945, it was not expressly excluded either, before or at the time of the drafting of the UN Charter.Footnote 110 Article 51 of the Charter does not expressly require that the armed attack triggering the right of self-defence has to be committed by a state. Even if it may be argued that the law of self-defence has been implicitly interpreted as entailing such a requirement,Footnote 111 this arguable position nevertheless remains a matter of interpretation. It does not assert the existence of a well-established rule, the evolution of which should be considered as a modification thereof. The same may be said about the issue of anticipatory self-defence, which may also arguably be considered a matter of interpretation – rather than of modification – of the law of self-defence, since the existence of a clear-cut prohibition on anticipatory self-defence at the time of the drafting of the UN Charter is far from clear.Footnote 112
That having been said, one must determine whether and under what conditions the law of self-defence may be interpreted or modified through state practice. The former form of evolution is largely acknowledged, provided that state practice is an application of the interpreted norm and that it establishes the agreement of the states regarding this interpretation. Its well-known legal basis regarding international conventions is Article 31(3)(b) of the Vienna Convention on the Law of Treaties. By contrast, modification through state practice is very much contested.Footnote 113 It is indicative that this Convention provides that subsequent state practice has only an interpretative and not a modifying effect on treaties. However, it is also indicative of the fact that using state practice as a means for modifying a treaty is open to challenge in the light of international case law, legal scholarship and state practice,Footnote 114 and that the reason for not having included such a means in the Vienna Convention was that practice leading to the modification of a norm was difficult to distinguish from practice infringing such a norm.Footnote 115 This logically suggests that, although state practice should not be excluded as a means of modifying a norm, one should accept that it produces such an effect only if the quality and quantity of state practice is particularly high in order for this practice to demonstrate a clear modification rather than the violation of a norm.
More specifically, the three general above-mentioned requirements, i.e., that state practice be relevant, general and constant, should be more or less narrowly construed depending upon whether this practice is intended to modify or interpret the law of self-defence. As a result, the modification of that law through state practice should imply: first, that the weight of this practice is high in the sense that it must be able to reveal the opinio juris of states in the least ambiguous manner; second, that in the case of international reactions to a specific practice, those reactions be as abundant as possible and consist of explicit – rather than only implicit – approbations when assessing the generality of such practice; and third, that the repetitions of the relevant practice over time be numerous and its uniformity well established. In contrast, it would be irrelevant to adopt such a restrictive approach with respect to interpretation, since the problem of distinguishing between interpretation and violation of a norm does not per definitionem exist. This seems to be the case with respect to the right to respond in self-defence to private armed attacks and anticipatory self-defence.
4. Conclusion
Legal scholarship on the use of force has undeniably moved towards a broader interpretation of the law of self-defence. This is clearly noticeable with respect to two important aspects of this law: the right of self-defence against armed attacks by non-state actors and the right to prevent imminent threats of armed attack in self-defence. The majority view has shifted from the rejection of those rights to their general acceptance. This is the result of some specific events, in particular the approval by the international community of the US military response to the 9/11 terrorist attacks and its reaction to the 2002 NSS. Those events sent real shockwaves throughout legal scholarship on the use of force and have impacted on the expansionist argumentative landscape. One may first observe that expansionists no longer develop any detailed arguments for establishing the right to resort to self-defence against armed attacks by non-state actors or the right to act in self-defence to prevent imminent threats of armed attacks. They content themselves with referring to the main events, practice or declarations evidencing such rights. However, expansionists do not stop there. They now embark upon new discussions in favour of an even wider interpretation of those rights. Basically, they keep struggling to broaden their flexible understanding of the law of self-defence with respect to those issues.
Yet, such extensive views are no longer mainly based on the traditionally accepted means invoked in legal scholarship for claiming any evolution of the law of self-defence, notably state practice, which must be sufficiently relevant, constant and uniform to lead to such evolution, depending upon whether this practice seeks to modify or merely interpret the law of self-defence. They result instead from policy and realistic considerations, the relevance of which and ability to lead to any evolution of that law is highly controversial, in addition to the fact that they are based on subjective and therefore debatable assessments. As a result, although legal scholarship on the use of force has moved towards a wider conception of the law of self-defence, it is clear that vigorous discussions will long remain between those favouring a restrictive (or less extensive) approach to such a law and those supporting an (even more) extensive understanding of it.
