Conclusion - Self-Defence against Non-State Actors – The Way Ahead

Summary

This Trialogue has discussed whether and – if yes – under which conditions international law as it stands allows for self-defence against non-State actors on the territory of a non-consenting State. Unsurprisingly, it has not come up with one clear answer. Rather, it has come up with three distinct answers – the contrast and interplay of which illuminate the facets and intricate details of one of the most pressing problems of international peace and security law. Dire Tladi advocates an inter-State reading of self-defence based on a thorough investigation of the UN Charter framework and recent State practice and thus concludes that self-defence against an ‘innocent’ State is unlawful. Christian Tams arrives at the opposite result. Employing – as Tladi does – a principally positivist method, his finding is that the better interpretation of the law is open for self-defence against non-State actors.

This Trialogue has discussed whether and – if yes – under which conditions international law as it stands allows for self-defence against non-State actors on the territory of a non-consenting State. Unsurprisingly, it has not come up with one clear answer. Rather, it has come up with three distinct answers – the contrast and interplay of which illuminate the facets and intricate details of one of the most pressing problems of international peace and security law. Dire Tladi advocates an inter-State reading of self-defence based on a thorough investigation of the UN Charter framework and recent State practice and thus concludes that self-defence against an ‘innocent’ State is unlawful. Christian Tams arrives at the opposite result. Employing – as Tladi does – a principally positivist method, his finding is that the better interpretation of the law is open for self-defence against non-State actors. Mary Ellen O’Connell, by contrast, advocates an originalist and value-imbued reading of the UN Charter. Based on such an understanding of the Charter, O’Connell concurs with Tladi and adduces an interpretation of the law that categorically prohibits self-defence against non-State actors.

I. Different Modes of Engaging with the International Law on Self-Defence

Pulling the strings of this Trialogue together, our first observation is that the chapters manifest different modes of engaging with international law. These modes unveil divergent expectations about the functions of law and of legal scholarship. The underlying mindset seems to inform how the authors tackle the more technical and specific aspects of legal interpretation. Dire Tladi and Mary Ellen O’Connell follow a more traditional legal approach in that they seek to determine what the law objectively provides for at a given moment in time, thus essentially raising the question as to what ‘international law is on a given day’.¹ The binarity of the law (related to its task of distinguishing lawful from unlawful behaviour) is crucial for both authors. Indeed, the benchmarking-function is normally viewed to be the essential job of law as a mode of governance. By painstakingly seeking to identify this benchmark, the two chapters are apt to provide political practice with clear criteria for the legality and illegality of potential courses of action.

Christian Tams, by contrast, is concerned with nuances in the substance of the law over time. He uses legal debates to illustrate historical trends and incremental shifts in perceptions and interpretations. Thus, Tams engages with the law not so much as a binary system that paints a black and white picture but treats the law as a flexible device to accommodate legal and political developments. This type of analysis does not need to search for the tipping point at which a legal rule changes (or ceases to exist and gives way to a new one). Tams’ method allows for a deep understanding of the matters, but does not strive for unequivocal answers to the complex questions surrounding self-defence. Depending on the perspective and interests of legal actors, notably practitioners, this approach can be either seen as suitable or not. It may be of little use in that it does not provide a clear answer to the question of what the law says at a concrete point in time, but it may be of some use for those (notably for practitioners such as legal advisers) who must develop legal arguments for justifying a certain course of action (e.g. of a government) and therefore study prior incidents seeking argumentative support in them.

Both perspectives (O’Connell’s and Tladi’s on the one hand, and Tams’ on the other) manifest or reveal diverging assumptions about the functioning of international law. For Dire Tladi and Mary Ellen O’Connell, international law provides a rather static framework. Once we have identified what the law objectively provided for, we carefully have to analyse whether and when the strict requirements for legal change are met. From Christian Tams’ perspective, the law provides a rather flexible framework. The emphasis is not on one objective content, but rather on the different legal positions that can be formulated within a legal discourse and which are neither a correct nor an incorrect reading of the law, but have ‘degrees of legal merit’.² In that sense the law provides an argumentative resource in which legal views and the related nuances are preserved even when they are not the dominant reading of the law – and they can more easily be activated than within the static framework.

II. Handling the Sources of Self-Defence

An important focus of this Trialogue rests on doctrinal analyses of the state of the law. A first question is which sources are most relevant for the rules on self-defence: custom, treaty (the UN Charter), or a combination of both? While this question is controversial in international legal debates in general,³ the authors of the Trialogue all – in one way or another – prefer to analyse self-defence under the UN Charter. According to Christian Tams, self-defence has, with the adoption of the UN-Charter, been shaped as a question of treaty law.⁴ Dire Tladi agrees that treaty law is crucial, but assumes – in accordance with the position espoused by the ICJ in the Nicaragua case – that both sources – treaty and custom – have a parallel existence, and Tladi therefore also attributes significance to custom. In the end, however, he holds that the substantive content of the rules flowing from both sources is ‘co-extensive and identical’.⁵ Mary Ellen O’Connell is less explicit about the sources of the obligation, but the UN Charter plays a crucial role in her overall argument.

For Dire Tladi and Christian Tams, the legal debate pivots around the interpretation of Article 51 based on the familiar canons of interpretation, as articulated in Articles 31 and 32 of the Vienna Convention on the Law of Treaties,⁶ and both authors focus their chapters on this point. This debate involves the usual questions of interpretation, addressing the ‘ordinary meaning’ of the terms ‘armed attack’, their context, object and purpose (Article 31(1) VCLT). Both authors then zoom in on ‘subsequent practice’, which must be taken into account when interpreting a treaty norm (Article 31(3) lit. b) VCLT). Although the examination of subsequent practice is mentioned in the VCLT as a means ‘for the purpose of … interpretation’ (Article 31(3) VCLT), such practice might – arguably – not only clarify but also modify the meaning of a treaty rule (in our case Article 51 of the UN Charter).⁷ Along this vein, Tams’ and Tladi’s chapters address the following questions. How dense does the practice need to be, i.e. how many States must perform acts in order to constitute subsequent practice? What is the significance of the silence of the majority of States that do not actively support the potentially law-shaping State practice of other States, but do not articulate protest either? Which weight shall we attribute to practice that is not accompanied by an explicit opinio iuris?

Dire Tladi and Christian Tams give strikingly different answers to these questions and arrive at opposite conclusions. Tams places more emphasis on the practice of States, arguing that it has reached a sufficient density to allow for self-defence against non-State actors. Tladi acknowledges existing practice, but argues that this practice is not accompanied by an agreement of the parties⁸ to the UN Charter and therefore not sufficient to allow for or even compel a broader interpretation of self-defence, especially since many States have not taken a position on the issue.⁹ The difference of opinion between both positions manifests two divergent overall understandings of international law. The first is a consent- and intent-oriented conceptualisation of international law which emphasises that all parties to a treaty need to agree to a reinterpretation (or silent evolution) of legal provisions. Tladi subscribes to that requirement by writing that ‘any expression of criticism or objection will most certainly inflict a deathblow’ to attempts to reinterpret the Charter.¹⁰ By contrast, the second view of international law – to which Tams leans – privileges the objectives and purposes of treaties over the original intent of the States concluding them. Treaty interpretation along that line tends to accord more importance to a treaty’s intrinsic telos (to some extent detached from the views of the original drafters). Correspondingly, this interpretive approach places more emphasis on the practice of States than on their verbal statements, because non-verbal practice is a more malleable sign of the parties’ ‘agreement’ (as required by Article 31(3) lit. a) and b) VCLT) than the States’ utterances.¹¹

Another question is whether the content of a formerly established rule may be said to have changed even if the exact contours of the presumable new rule have not taken shape yet. Postulating that the original rule allowed self-defence only against attacks led by States, can we assume that – in view of latest practice – self-defence against non-State actors has become permissible, although we do not (yet) know exactly under which legal conditions? Not even those States which have taken military action have espoused a consistent view on this point. As discussed in our Introduction, the interventions in Syria have been justified by some States with the ‘unwilling or unable’ standard, while other States, notably Belgium and Germany, have focused on the loss of effective control by Syria over parts of the State’s territory.¹² Tladi negates that under such circumstances subsequent practice would be capable of reinterpreting (or modifying) the law, because we do not have ‘a clear meeting of minds as to the interpretation of Article 51’.¹³

III. Moral Values and ius cogens

Mary Ellen O’Connell’s chapter is less concerned with the practice of States. She espouses an originalist reading of the UN Charter, whose centrepiece is certain fundamental moral values, above all the right to life of the individual and the quest for peace. O’Connell argues that these moral values form the nucleus of the concept of ius cogens. The peremptory quality of the prohibition on use of force affects – so she claims – the possibility of development of the legal rule.

This touches a further legal issue related to the development (evolution or change) of international law:¹⁴ how relevant is the concept of ius cogens for the involved rules’ potential for legal development? While it is widely believed that at least parts of the legal regime on the use of force in fact constitute ius cogens,¹⁵ the exact scope of this presumable ius cogens quality is a subject of controversy.¹⁶ It is often acknowledged that the (assumed) peremptory character of the prohibition on the use of force has some impact on the possibility, modalities and thresholds of the evolution of this legal norm. According to Article 53 of the VCLT, ‘a peremptory norm of general international law … can be modified only by a subsequent norm of general international law having the same [i.e., ius cogens] character.’¹⁷ Therefore, many observers who analyse the prohibition on the use of force as a customary rule (and who accept its peremptory quality) find that the threshold for its change is higher than for ordinary customary law.¹⁸ In contrast, if we analyse the prohibition on the use of force only as a treaty norm, the requirements for legal development¹⁹ differ, and the ius cogens character of a rule would not have any significant effect: the ‘general rule regarding the amendment of treaties’ is that ‘a treaty may be amended by agreement between the parties’ (Article 39 VCLT).²⁰ A novel interpretation of a treaty requires an ‘agreement of the parties’ regarding its interpretation (Article 31(3) lit a) and b) VCLT). This means that all parties need to agree on the amendment or reinterpretation of the law. This establishes a very high threshold, also for a ‘silent’ development of treaty norms (i.e. without a change of the treaty text).

Dire Tladi and Christian Tams, in their chapters, do not attribute significance to the issue of ius cogens in determining the lawfulness of self-defence against non-State actors. Their main argument is that even if the prohibition on the use of force might possess a ius cogens status, this status does not fix the exact substantive content or coverage of the norm. This coverage is determined by the interplay between the prohibition and the exception (self-defence).²¹ Figuratively speaking, the extension of the exception shrinks the scope of the prohibition. Thus, if an action is justified as self-defence, the prohibition does not cover this action. Such an adjustment of the coverage of the rule would, in that conceptualisation, not even be a ‘modification’, in the terms of Article 53 of the VCLT quoted above. Or, one might say that self-defence co-determines the prohibition, that hence it partakes in its normative ‘character’ as a peremptory rule, and that therefore any expansion of self-defence is apt to change the rule even under the conditions formulated by Article 53 of the VCLT. Key to this understanding is a perception of the ban on the use of force and self-defence as communicating vessels.

Mary Ellen O’Connell’s account departs from established positivist conceptions of ius cogens. She claims that ius cogens rules are ‘discerned … through natural law method’.²² In her view, peace is the ‘superior moral and legal norm’²³ as articulated in the ius cogens character of the prohibition on the use of force. Her conclusion is that peremptory norms can only develop in one direction, namely towards the greater realisation of the moral norm they aim to protect (‘principle of progression’).²⁴ Thus, O’Connell believes that contrary State practice may not dilute or undermine an established scope of the prohibition on the use of force. If self-defence against non-State actors was – as O’Connell holds – once illegal, it cannot become lawful, because this would create more legal opportunities for the use of force. This view implies but does not spell out a novel theory of legal change and raises difficult conceptual questions about the relationship between law and morality. On Martti Koskenniemi’s spectrum between apology and utopia it is far on the utopian side.

IV. The Indeterminacy of the Law on Self-Defence

The chapters demonstrate how a seemingly narrow legal question (such as whether and under which conditions self-defence against non-State armed attacks is lawful) may be and often is answered differently, both in the abstract and when deciding a concrete case. International law (in all its shapes, including notably treaties and customary norms) is notoriously blurry and gives plenty of leeway to those interpreting and applying its rules.²⁵ The well-known phenomenon has been described with many terms, as the law’s ‘indeterminacy’,²⁶ as ‘uncertainty’²⁷, as ‘vagueness’,²⁸ and as ‘ambiguity’.²⁹ We here use the term ‘indeterminacy’ to describe a characteristic feature of legal rules.³⁰ The term describes the fact that any person handling legal rules must make choices about which scope to give to the rule, and how to apply it to the given facts. Put differently, legal rules are indeterminate if they do not provide only one answer for deciding a controversial case but – on the contrary – various (not exactly predictable) answers can be given.³¹

The indeterminacy of the law and the resulting uncertainty in applying it have a number of causes. The two main causes are, firstly, the properties of the ordinary language (which is imprecise and malleable) that furnishes the technical language of the law and, secondly, the unforseeability of future situations to which the rules are designed to apply.

In the field of international relations, a frequent third cause of the indeterminacy of treaty provisions is deliberate compromises in the drafting stage that lead to a choice of wording which (due to its ambiguity or vagueness) can be understood differently by the negotiating parties. In this context, the treaty law’s indeterminacy is intended by the drafters. This technique might be called ‘constructive indeterminacy’, because it pushes fundamental disagreement from the drafting stage to the stage of applying and implementing the law, opening a window of opportunity for consensus in a later point in time.³²

A fourth cause of indeterminacy has to do with the ‘sources’ of international law, i.e. the shapes or forms in which international law comes (the traditional major ones being treaty and custom). We therefore face a ‘second-order indeterminacy’³³ in international law. Myres McDougal and Michael Reisman highlight that beyond Article 38 of the ICJ Statute, ‘all agreement among commentators [on sources] ends.’³⁴ With regard to the law on the use of force, Andrea Bianchi observes that ‘the interpretive community is currently divided and is no longer able to agree on the method that must be used for interpreting the law.’³⁵ Or, should there be some amount of agreement on method, it is not robust enough to provide much guidance in controversial cases.

Specifically on the norm of self-defence, Georg Nolte and Albrecht Randelzhofer write: ‘A significant amount of current disagreement over the proper interpretation of Article 51, both among States and among commentators, can ultimately be traced to underlying differences of opinion over the interpretation and application of the rules on the sources of international law.’³⁶ It is generally agreed that a customary norm on self-defence exists – the ‘inherent right’ as mentioned in Article 51 of the UN Charter. Unwritten rules, which are not fixed in verbal form, are even more indeterminate than written rules, because the starting points for discussions about their scope and application are even less agreed.³⁷ Additional insecurity flows from the possible interplay of the customary rule with the Charter provision and from its possible quality of the rule as ius cogens. For example, the current practice on self-defence against non-State actors seems to constitute both a means of interpreting the Charter law (as ‘subsequent practice’ in the sense of Article 31(3) lit. b) VCLT) and the objective element of the customary rule.³⁸ In fact, as the Trialogue manifests, there is no full agreement on how to handle the complex legal questions in this constellation. The three chapters’ analyses of these points diverge to some extent, although all focus on the UN Charter as opposed to custom.

Generally speaking, the methods for identifying and determining customary rules are undertheorised and lack rigour.³⁹ The poor understanding of the process of development of customary law and the paucity of guidance for the exact determination of the content of customary rules are highly relevant for the law of self-defence because of its (also) customary law quality. Methodological choices for norm-identification are, for example, whether to prioritise practice versus opinio iuris, and how to range ius cogens, as highlighted by Mary Ellen O’Connell. These choices co-determine the outcomes of legal reasoning, but they are themselves not entirely determined by law. Rather, they significantly depend on external factors, and they are informed by the theoretical, practical, geo-political and political background assumptions of those interpreting and applying the law. In this Trialogue, Dire Tladi makes this explicit. He acknowledges that he seeks to avoid ‘an interpretation of law that facilitates the “de-constraining”’ in the interest of those who enjoy much military power,⁴⁰ stressing that this position is not ‘merely a policy preference’, but one ‘which is grounded in the law’.⁴¹ O’Connell, on the other hand, builds her argument on strong normative convictions at the heart of which is a ‘presumption of peace’⁴² that is underpinned by a natural law theory.

The indeterminacy of the law and its application, including the methodological openness for reaching legal answers, is a problem both for legal practice and for legal scholarship. Because the legal findings presented by scholars or practitioners in the relevant legal debates can easily be called into question and refuted by opponents who favour a different substantive result,⁴³ such legal findings offer little practical guidance for the conduct of States and no stable knowledge-base on which further research can build.

Indeterminacy of the law is a matter of degree.⁴⁴ All legal rules are indeterminate to some extent. Problems arise when a rule is so extraordinarily indeterminate that it has only a weak capacity to coordinate, guide or ‘govern’ State behaviour. Extreme indeterminacy (or insufficient determinacy) corrodes the international rule of law because it leaves much space for arbitrary interpretations and applications of the rule. Moreover, insufficient determinacy saps a rule’s legitimacy which in turn impacts on its compliance pull.⁴⁵

The Trialogue has shown that the rules on self-defence are indeed extremely indeterminate. The law on the use of force is a prime example of methodological openness and indeterminacy on various levels (regarding the canons of interpretation, regarding the relevance of the sources and regarding the conceptualisation of legal development).⁴⁶ This is not meant to suggest a radical indeterminacy in the sense that no objective content of the international law of self-defence could be established and that therefore this body of rules could never constitute a yardstick for distinguishing lawful from unlawful employments of military force.⁴⁷ There is a core of certainty where one State defends against the ongoing armed attack of another State. This core, however, is surrounded by a ‘twilight zone’⁴⁸ of indeterminacy, in which the applicability of the rule is controversial, the case of self-defence against non-State actors being a focal point of contestation.

Although there is not one right answer to the question of whether and under which conditions self-defence against non-State actors is lawful, there are better and worse answers within the parameters of the legal discourse.⁴⁹ The better answer is not inevitably simply politics in disguise but can be a specifically legal answer.⁵⁰ The simple observation that legal discussants, such as the three participants of the Trialogue, in fact seriously struggle and disagree about a legal answer illustrates that there is a specifically legal sphere of the meaning of self-defence.

V. How Does the Law of Self-Defence Change?

A key issue brought to light in this Trialogue is legal change. All three contributions explicitly or implicitly address the widespread assumption that the law of self-defence seems to be changing or has changed. As a political matter, the support for the old orthodoxy of State-centred self-defence has weakened. Apparently, a significant number of States opine that they need to act in self-defence against non-State actors on the territory of a non-consenting State. They therefore seek to loosen the State-nexus of self-defence. We find ourselves in constant debates about whether the tipping point for the modification of a previously established norm on self-defence (or dissolution of the old norm and the emergence of a new norm) has been reached or not. The development (or evolution⁵¹) of the law of self-defence concerns both legal sources, the treaty law (Article 51 of the UN Charter) and the ‘inherent’ right mentioned in the Charter that is based on a customary rule.

The Trialogue authors address this matter in dissimilar terms and draw different conclusions. Their divergence shows that the process and procedures of the change of international law are still poorly conceptualised in legal terms. This stands in contrast to studies in the field of international relations, notably by social constructivists who have developed theories of norm change since the 1990s, such as the ‘cascade’ model by Kathryn Sikkink,⁵² the ‘cyclic theory’ of norm change by Wayne Sandholtz⁵³ and the ‘spiral’ model by Thomas Risse and others.⁵⁴

A. Change of the Charter Law

The textually unrevised provision of Article 51 of the UN Charter might have acquired or might be acquiring a new meaning. This could be conceptualised either as a progressive interpretation or as an (‘unwritten’ or ‘silent’) modification without changing the text of the treaty. Put differently, the question arises whether the understanding that an ‘armed attack’ might emanate from a non-State actor would still be a legitimate form of interpreting the Charter rule, or whether this would amount to a silent modification (an unwritten amendment of the UN Charter). If it amounts to a modification, it might be more prone to critique, notably as an illegitimate or even unlawful juridical operation.

Doctrinal international legal scholarship postulates a conceptual dividing line between interpretation ‘proper’ and silent modification ‘effected under the pretext of interpretation’ – although the writers agree that this line is blurry.⁵⁵ Simplistically speaking, the interpretation of a legal rule is considered to be somehow bounded (although it is not fully clear by what), whereas the modification of a rule is potentially unlimited. Also, an interpretation happens first of all for one single case (and lives on only through precedent), while a modification of a treaty provision will formally govern all future cases.⁵⁶

The dichotomy between the dynamic interpretation of a treaty and its ‘silent’ modification does not withstand the insights of legal theory that every text needs interpretation, and that every interpretation creates new meanings.⁵⁷ And actually, the continuum between interpretation and silent modification of a legal text does not pose a normative problem when it is performed by the identical actors, and when no third parties are affected by the development of the law. Put differently, it is not always necessary to determine whether the members to a given treaty (such as the UN Charter) interpret that treaty or silently modify it.⁵⁸

Nevertheless, the simplistic and perhaps even false dichotomy between the interpretation of a treaty and its amendment persists in practice. The Vienna Convention contains one section on ‘Interpretation of Treaties’ (in Part III) and another (Part IV) entitled ‘Amendment and Modification of Treaties’. Both the case-law and treaty texts postulate that the interpretation and the modification of legal rules are two conceptually and legally distinct operations. The leading case is an ICJ Advisory Opinion on treaty interpretation of 1950 in which the Court stated that ‘[i]t is the duty of the Court to interpret the Treaties, not to revise them.’⁵⁹ This dictum has been picked up by further judgments⁶⁰ and by numerous arbitral awards.⁶¹

Article 3(2) of the WTO Dispute Settlement Understanding, which prohibits the Dispute Settlement Body (DSB) to add or diminish rights and obligation provided in the WTO agreements,⁶² presupposes that the interpretation by the WTO Dispute Settlement Body of those treaties and their modification are two different things. The case-law and the mentioned legal provision bring us to the heart of the matter. The normative reason for upholding the dichotomy is a presumed separation of powers in the international legal process.⁶³ When States parties – and not international courts – are recognised as the primary law-makers and law-changers, it is necessary to draw a line between the operations of interpreting the law on the one hand and changing it on the other, because the operations might befit different actors. The normative question undergirding this separation of powers is who is entitled to shape the legal environment for third parties, notably for citizens who are subjected to the rules. The traditional answer, mainly motivated by the concern for State sovereignty, is that this job is assigned to the State parties. Arguably, democracy and the rule of law (comprising the principles of legal certainty and previsibility of the law) are also better safeguarded in treaty-making or amending processes than in judicial law-making.

The problem becomes apparent in a proceeding pending before the German Federal Constitutional Court which concerns self-defence actions against a non-State actor, namely the Islamic State of Iraq and Syria. The German parliamentary faction Die Linke in 2016 filed a complaint against the government’s deployment decision in the context of the anti-ISIS-operation ‘Inherent Resolve’. The applicant argues, inter alia, that the deployment is based on an over-extensive reading of Article 51 of the UN Charter that is no longer a legitimate interpretation of the UN Charter, but rather a modification. The constitutional law argument, then, is that this practice also oversteps the parliamentary statute approving of Germany’s accession to the United Nations in 1973.⁶⁴

The UN Charter, formally a treaty, may be interpreted (and re-reinterpreted) by ‘taking into account’ any ‘subsequent agreement between the parties regarding the interpretation of the treaty’ (Article 31(3) lit. a) VCLT) or by taking into account ‘any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation’ (Article 31(3) lit. b) VCLT). Past evolutions of the UN Charter without changing its text have arguably not amounted to modifications but have remained within the realms of progressive interpretation.⁶⁵ Indeed, it is often argued that the UN Charter is particularly prone to and in need of dynamic interpretation due to its constitutional character and due to the difficulty of formal amendment.⁶⁶

A true change of the UN Charter, notably of its Article 51, under the guise of interpretation would pose additional problems. From the perspective of the VCLT, the required agreement between the parties could be completely informal, for example through practice.⁶⁷ But the VCLT provides only residual rules; treaty amendment is first of all subject to the rules of revision in the given treaty itself.⁶⁸ The UN Charter itself contains specific rules of procedure for its revision and amendment (Articles 108–9). ‘Informal’ amendments risk undermining the UN Charter’s formal revision procedures which seek to safeguard institutional balance and legal clarity. The concern about circumventing procedures is exacerbated if the UN Charter is perceived as a constitutional document whose specific function as a repository of the most fundamental norms is put at risk by adding bits and pieces of ‘constitutional by-law’ whose extent and content is not clear and which cannot function as a guideline.⁶⁹ These reflections are in line with the chapters of both Mary Ellen O’Connell and Dire Tladi who find that the Charter law of self-defence has not been amended, and O’ Connell even implies that such a change would not be legally possible.

B. Change of the Customary International Law on Self-Defence

The customary law of self-defence may have evolved through a change of practice and of opinio iuris.⁷⁰ However, in the face of more than a few isolated incidences of military action against non-State actors and accompanying legal assertions, it is hard and probably impossible to keep apart practice, which simply breaches the outlasting ‘old’ norm, from the advent of a novel, norm-generating practice. Christian Tams’ chapter discusses this problem on the level of treaty law, but the similar difficulty exists with regard to the change of custom. So how to distinguish breaking from making the customary rule on self-defence?

Looking at the intensified military strikes against armed groups in the territory of a non-consenting State alone, we cannot tell whether the behaviour is aberrant, or whether the practice is ‘right’. In order to find out whether the old customary law is being breached or whether – conversely – the formation of a new customary rule is going on, we need to examine the second element of customary law, the opinio iuris of the States involved. However, at the point in time when the customary law changes, the opinion of the legal subjects that their behaviour (in our case, for example, the US and allies conducting air strikes in Syria) is lawful can only be in error. The States cannot be (rightfully) convinced that they are acting in conformity with the law, that their behaviour is permitted or required by law, because such a legal rule does not (yet) stand. An opinio iuris in its strict sense (in the sense of a conviction that a permissive legal rule allows such self-defence independent of any attribution of the non-State actor’s attack to the territorial State) cannot exist as a matter of logic.⁷¹

Legal scholars therefore concur that in this phase the States’ attitudes cannot be more than a ‘claim’⁷² or ‘signal’⁷³ to others and the acceptance of that claim. International Relations scholar Wayne Sandholtz, too, has highlighted that the norm change hinges on the justification of the new practice and on the reaction of other States.⁷⁴ A powerful State such as the United States might get away easier with breaking the law of self-defence. But if most other States condemn an action, a violation remains simply a violation, and the old rule is affirmed. Sandholtz argues that a violation can lead to new or modified rules in two scenarios. An only mild pro forma condemnation evidences that the old rule is weakening. If, in addition, apparent violation is followed by subsequent similar conduct, then the new pattern can be evidence of an emerging law: ‘[i]n that case, the initial non-compliant act would be seen as not just a violation but as the first step in defining a new rule. Of course, such judgements can only be made in retrospect.’⁷⁵

In legal terms, the failure of other States to condemn the initial violations evidences a shift of opinio iuris.⁷⁶ Initially, that opinio iuris may exist (only) as to the immediate future. It may then take ‘the form of a settled conviction as to what the law should be, and would be for the proclaiming state … One can without contradiction announce the intention to live by a certain rule, if one does live by it from that time.’⁷⁷

C. The Law in Transition

The difficulty for understanding and describing legal change is that – both for customary and for treaty law – the traditional doctrines on the sources of international law lack the vocabulary to capture the transitional period, during which the old law is still present, but during which a new rule has not yet taken shape.⁷⁸ The case of self-defence against non-State actors illustrates that during a protacted period of change an indeterminacy of the law reigns.

Such indeterminacy is often the result of a deliberate or welcome strategy of international actors. States, international organisations and other bodies carefully pursue strategies of indeterminacy that create and maintain a situation in which differences of legal interpretations can be upheld without one or the other view of the law being clearly endorsed or condemned. The legal meaning and value of States’ statements and acts and their (non-)reaction remains obscure because many States avoid taking a clear position. The Security Council, too, sometimes employs tactics of indeterminacy. Resolutions 1368 and 1373 of 2001 in relation to Afghanistan, and Resolution 2249 of 2015 on Syria, use language which is apt to provide arguments for both sides: for ‘restrictivists’ and for ‘expansionists’.⁷⁹ Finally, international courts also often pronounce judgments which contain vague and obscure passages, most likely due to compromises when reconciling different views on the bench during the judges’ deliberations. The ICJ’s landmark decisions dealing with self-defence, ranging from Oil Platforms to the Wall Advisory Opinion to the judgment in Congo v. Uganda, are written in a way that can be interpreted in different directions.⁸⁰

Our claim is that such indeterminacy fulfils an important function for international law. With the help of vagueness, with widespread silence, and through moderation in their international reactions, States (and other international legal persons with law-making power) are continuously creating and upholding a ‘regulatory sphere’ in which the rules may be reinterpreted or even remade until a norm has taken sufficient shape and the point has been reached that it becomes widely accepted as being the law. This ‘method’ for developing international law is not explicit and not explicated, and therefore deserves the label ‘method’ only in quotation marks. Rather, it is tacit, intuitive, unstructured, pragmatic and politicised. On the plus side, it allows for incrementalism, which is often needed in order to allow for change to happen. This perspective means abandoning the purely static view on the law. Observers should accept and welcome the fact that, in order to be functional, law needs to be a living instrument. Once this dynamic vision is espoused, we must acknowledge that there will be a phase of uncertainty during a certain period of time. The emergence of the new rule will only be identifiable in hindsight. The exact turning point of legal change cannot be pinned down with precision. The disagreement among the Trialogue authors confirms the impression that we currently find ourselves in this ‘grey’ period.⁸¹

The downside of this ‘method’ of legal change is that it does not embrace open exchange of arguments and public debate, but significantly relies on indirectness, implicitness and reluctance. Moreover, while potentially beneficial for the consolidation and development of international law when limited to specific issues, this ‘method’ has the potential to dilute and eventually undermine the ius contra bellum’s regulatory function.⁸² When States refrain from articulating their legal views, when they shy away from protesting against the use of doctrines such as the ‘unwilling or unable’ formula that lend themselves for arbitrary interpretations, we might witness a general dilution of the law’s regulatory power.

VI. Conclusion

The indeterminacy of the law on self-defence is both opportunity and risk. It is an opportunity because it is the precondition and environment in which the law may develop towards greater clarity and in which it may adapt to new security threats that are caused by non-State actors. At the same time, it carries significant risks of abuse and the danger of an overall erosion of the existing security architecture. Indeterminacy becomes ‘a deficit when it lends itself to arbitrariness’ and when it allows actors ‘to exempt their actions from the reason of law’.⁸³ The more indeterminate the law is, the more leeway it gives to political decisions and to those States that have the financial and military resources to act. For example, the ‘unwilling or unable‘ standard – for identifying States from whose territory terror attacks have been launched and against which self-defensive action should then be allowed – would benefit mainly or exclusively the powerful States which arrogate themselves the privilege to apply this standard against others. But fully generalising the standard would create a high risk of escalating military actions. It therefore seems as if the concept could ultimately not become a general legal rule because this would deeply erode or even destroy international order.⁸⁴

Unsurprisingly, proponents of Third-World approaches to international law have noted that indeterminacy in general ‘very rarely works in favor of Third World interests. Ambiguities and uncertainties are invariably resolved by resort to broader legal principles, policy goals or social contexts, all of which are often shaped by colonial views of the world.’⁸⁵

The danger of abuse of an extended reading of self-defence is particularly pressing when we take into account the broader context in which military action against non-State actors is taken. Mary Ellen O’Connell’s ‘three pernicious doctrines of expansive self-defence’⁸⁶ are important here. Each of these ‘doctrines’ in isolation seems manageable: terrorist acts might be reasonably qualified as armed attacks once a threshold of gravity is met; self-defence against imminent attacks might not pose a problem when it is based on strict and verifiable indicia for imminence as opposed to merely unsupported assertions of threats; and ultimately self-defence against non-State actors would – if sufficiently limited by legal criteria – appear containable, too.

It seems, however, as if exactly the interplay of all three ‘doctrines’ constitutes the danger O’Connell evokes. First, it lies in the nature of terrorist organisations that their armed attacks are clandestine and difficult to detect. Therefore, the type of objective evidence that can be furnished for upcoming State attacks is lacking here. States invoking terrorist attacks need to rely more often on intelligence sources which other States and the public cannot access and assess. Accepting, secondly, that such a lofty type of attack does not even have to be ongoing to trigger self-defence, but contending ourselves with an ‘imminent’ putative attack, gives all leeway to well-armed powerful States. So far we lack any standards for objectively establishing imminence, and such establishment again requires superior intelligence information. If the military reaction is, then, thirdly, directed against any State that has not countered this potentially quite fanciful ‘imminent attack’, these three doctrines in combination can essentially furnish a justification for each and every intervention. The result is that we currently witness the morphing of the battle against allegedly imminent terrorist attacks into a generalised toleration of pre-emptive action against remote threats. The Turkish operation ‘Olive Branch’ against the region of Afrin in Northern Syria in March 2018, which simply asserted ‘a clear and imminent threat of continuing attack from Daesh’⁸⁷ without proffering any evidence, illustrates this unwelcome trend.

The de iure acceptance of this practice as a genuine entitlement or allowance would – as Mary Ellen O’Connell and Dire Tladi warn – undermine the structure of multilateralism which forms the heart of the UN Charter. The danger is that a threat of terrorism is used as an argument for indefinite self-defence actions, thereby turning a special exception into a permanent authorisation. It is therefore necessary to develop criteria for operationalising notably the temporal dimension of self-defensive actions, including both their lawful starting point and the legal parameters demanding their termination.

So what could be the way ahead for the law on self-defence as applied to non-State actors? The best way which Christian Tams describes as the ‘obvious cure’⁸⁸ ─ would be to foster multilateral cooperation. We need multilateralism for the functioning of the system of collective security system and for allowing the Security Council to fulfil its mandate – this is the preferred way ahead for Mary Ellen O’Connell and Dire Tladi. Multilateralism would also be the ideal way to elucidate the substance of international law concerning self-defence. A definition of self-defence following the example of the General Assembly’s Definition of Aggression could provide clarification. If such a definition were unanimously adopted by the General Assembly, such a resolution could qualify as a subsequent agreement in the sense of article 31(3) lit. a) of the VCLT. It could thus provide an authoritative interpretation of self-defence.⁸⁹ The contemporary climate of world politics, with its emphasis on Realpolitik, ‘post-globalisation’, a renewed focus on the nation State, and the crisis of multilateralism seem adverse to such an exercise for the time being. However, one should remember that the period in which States attempted to define aggression was by no means more favourable than it is today. Quite to the contrary: the Cold War was gaining momentum, with deep cleavages between the two blocks showing up. It took almost twenty-five years from the first Soviet proposal of a definition of aggression tabled in 1950⁹⁰ to its actual adoption by the General Assembly in 1974. It is likely that such a timeframe would be needed to develop roughly shared positions on the law of self-defence and in particular on the issue of non-State actors. Hence, multilateral approaches will in any case, if they are at all feasible, not provide any short-term answers.

Therefore, the most likely scenario is the persistence of the indeterminacy of the law for some time to come. Along this line, Monica Hakimi states that ‘[e]fforts to clarify the law on the use of defensive force against non-State actors are premature’ because of the ‘ongoing struggle over the law’s proper content’.⁹¹

However, while the questions of whether and under which conditions self-defence against non-State actors situated in another State is lawful are likely to remain controversial for some time, it will be crucial to not lose sight of the ‘how question’. Inquiries into the requirements of necessity and proportionality of self-defence actions against non-State actors need to be intensified, as advocated by Christian Tams: ‘A more active, and more robust, debate about the necessity and proportionality of self-defence could help define the limits of forcible responses more clearly.’⁹²

Such a sharpening of existing principles (and their transfer to the constellation of a non-State attack) could be bolstered by attempts to improve the procedural side of self-defence, as advocated by Larissa van den Herik.⁹³ Van den Herik suggests tightening the requirements on reporting of self-defence actions as demanded by Article 51 of the UN Charter. This could be done, for example, by holding routine debates once Article 51 is invoked, by setting up a database of Article 51 letters, by developing best practices about when and how often such letters should be submitted and what they should contain, by creating a subsidiary body that collects and monitors the submission of Article 51 letters, and by installing expert panels tasked with collecting and examining information and making prima facie evaluations.⁹⁴ Christian Tams also points in that direction by suggesting that international actors should insist more on the provision of ‘credible evidence supporting self-defence claims’ which ‘could force reacting States into a public dialogue’.⁹⁵ Such a fortification of procedural requirements could in fact provide an institutional framework that would probably also contribute to the development of the substantive law.

It is hoped that the multiperspectivism of this volume contributes to a richer understanding not only of the law on self-defence but also of the dynamics of legal change in a pluralist world. Finally, the trialogical approach might be a pathway for teasing out universally acceptable legal answers which strike a fair balance in the tripolar tension between security demands of States, respect for all States’ territorial integrity and the objective of containing military violence.