1 Introduction

A century ago, Baron Descamps, who presided over the work of the Committee of Jurists that drafted the Statute of the Permanent Court of International Justice (PCIJ), observed that custom has shaped ‘the development and establishment of the law of nations’.Footnote ¹ The continuing relevance of custom as a source of international law is unquestionable to the present day. Despite efforts to codify rules and principles of customary international law (CIL) in multilateral conventions, often initiated by the International Law Commission (ILC), few of these instruments ‘have achieved universal or truly broad participation’.Footnote ² Numerous recent decisions of international courts and tribunals confirm that custom is not condemned to disintegrateFootnote ³ and certainly does more than simply fill gaps left by the existing treaty regimes, as has been suggested elsewhere.Footnote ⁴ International law has always been and remains principally a customary law.Footnote ⁵

It is perhaps unsurprising that custom, as a universal unwritten law binding upon sovereign states, has intrigued legal minds over the centuries, not least because it raises a host of fascinating and challenging theoretical and practical issues relating to the method and process of its formation.Footnote ⁶ To some observers it resembles ‘a riddle wrapped in a mystery inside an enigma’:Footnote ⁷ how can states act out of a sense of legal obligation in order to create a new customary norm, if the legal obligation does not exist until they have acted?Footnote ⁸ What is the balance of power in the married couple of state practice and opinio juris sive necessitatis? How does one go about ascertaining the required generality of practice in order to conclude the existence of custom? What role should be reserved to the practice of specially affected states, and how may these be identified? Many of these issues have been subject of or would merit a treatise on their own but, overall, they are indicative of what is often seen as the rather disorderly and chaotic nature of the process by which unwritten law develops in a horizontal and decentralised system of sovereign states which, whether we like it or not, remain the primary providers of custom.Footnote ⁹

In 1988, at a colloquium entitled ‘Change and Stability in International Law-Making’, Jimenez de Aréchaga, the former president of the International Court of Justice (ICJ or the Court), spoke of CIL as being ‘spontaneous, unintentional, unconscious in its origin, disorderly, uncertain in its form, slow in its establishment’.Footnote ¹⁰ Similarly, Henkin described ‘the process of making customary law [a]s informal, haphazard, not deliberate, even partly unintentional and fortuitous … unstructured and slow’.Footnote ¹¹ More recent inquiries into the method and process of CIL have called it ‘chaotic, unstructured, and politically charged [in which] the participants make and respond to competing claims on the law as they advance their own agendas’,Footnote ¹² an ‘inherently contingent and variable kind of law’Footnote ¹³ or even ‘hopelessly indeterminate’.Footnote ¹⁴ Others have decried custom as being unfit to accompany the rapid pace of developments in international relations nowadaysFootnote ¹⁵ and in a more heterogeneous international community.Footnote ¹⁶

Accordingly, no effort has been spared at the international level to analyse the process of custom formation and methodology for its identification.Footnote ¹⁷ The detailed inquiry into the subject conducted by the ILC under the leadership of Sir Michael Wood is the most recent and authoritative of such efforts.Footnote ¹⁸ The outcome of its work, in the form of Conclusions on Identification of Customary International Law (ILC Conclusions), sought to provide greater certainty as to the process of identification of CIL and to provide therefore practical guidance to judges and lawyers called to apply such law. Some of the ILC Conclusions will be discussed below.

The focus of this chapter, however, will be on the reasoning of international courts and tribunals and the ways in which they have identified rules of CIL and their content. Four recent studies have already examined various aspects of this practice, with a particular focus on the ICJ. This is to be expected as the Court is the only mechanism of general jurisdiction, has had the greatest number and variety of cases among international courts and tribunals where it has had to ascertain the existence of CIL, and in light of the authority it enjoys as the principal judicial organ of the United Nations. Talmon concludes that beyond inductive or deductive reasoning, the ICJ usually proceeds by asserting the existence of rules of CIL, or combining ‘a mixture of induction, deduction and assertion’.Footnote ¹⁹ Tams describes the role of the Court as clarifying the ‘meta-law’ on the identification of CIL, and has already identified similar ‘argumentative shortcuts’ to those that will be addressed in this chapter.Footnote ²⁰ Choi and Gulati argue that the ICJ has completely ignored the traditional methodology.Footnote ²¹ Petersen helpfully sets out a detailed classification of the Court’s approaches to CIL and factors that shape the ICJ’s decision-making in that context.Footnote ²² This chapter builds on some of these findings, and its added value is intended to lie in providing an up-to-date analysis of the ICJ’s practice as well as expanding the scope of the inquiry beyond the ICJ. In classifying the dominant shortcuts that courts and tribunals have adopted in their reasoning when identifying CIL, this chapter seeks to highlight the systemic issues they may raise in the foreseeable future.

The chapter proceeds as follows. Section 2 briefly sets out the traditional methodology for the identification of CIL as recognised in the jurisprudence of the ICJ and as reaffirmed in the ILC Conclusions. Section 3 analyses the recent practice of international courts and tribunals and specifically their use of a variety of shortcuts for the identification of CIL. Section 4 contains preliminary conclusions as to the challenges that employing such shortcuts in the reasoning of international courts and tribunals poses to the continuing validity of the methodology for the identification of CIL, the authority of the decisions rendered, and the perception of the role of the international judge not only as an idle scribe of CIL but as a lawmaking agent.

2 The Traditional Methodology for the Identification of CIL

Much ink has been spilled on the methodology for the identification of ‘international custom, as evidence of a general practice accepted as law’.Footnote ²³ This perhaps not fully felicitous wording of Article 38(1)(b) of the Statute of the ICJ,Footnote ²⁴ sets out two constituent elements of CIL, namely a general practice and its acceptance as law (also known as opinio juris sive necessitatis or opinio juris).Footnote ²⁵ To borrow the words of one arbitral tribunal, these two elements are the ‘guiding beacons’ of CIL, a law which is not frozen in time but continues to evolve in accordance with the realities of the international community.Footnote ²⁶

As noted by the ILC, the process of identifying CIL ‘is not always susceptible to exact formulations’.Footnote ²⁷ Indeed, the end-product of the Commission’s inquiry into the subject aimed to set out a ‘clear guidance without being overly prescriptive’.Footnote ²⁸ The ILC Conclusions and their commentaries contain a wealth of materials on almost every aspect of the two-element methodology, its theoretical and practical application.Footnote ²⁹ This chapter does not aim to revisit the methodology for the identification of CIL, which by and large is well-established and accepted by states,Footnote ³⁰ if not ‘set in stone’.Footnote ³¹ This chapter only aims to show that the demonstration that this methodology has been followed is often missing in practice and is frequently replaced by shortcuts in the reasoning of international courts and tribunals. It is in this context that it may be worth briefly recalling the gist of the traditional methodology for the identification of CIL, as it will set the scene for the subsequent analysis of international jurisprudence.

The ICJ’s North Sea Continental Shelf judgment remains a central reference point for any inquiry into the processes of formation and identification of CIL. It represents the fundamental mark that has been left by the Court on ‘shaping the meta-law of custom’.Footnote ³² In that case, the Court had to determine whether the rule of equidistance, as set out in Article 6 of the 1958 Geneva Convention on the Continental Shelf, was binding on Germany under CIL as Germany was not a party to the Geneva Convention. In holding that this was not the case, the ICJ set out its methodology for the identification of CIL, outlining a range of criteria that may be relevant to that process. The Court held that ‘two conditions must be fulfilled’, namely the existence of ‘a settled practice’ as well as ‘evidence of a belief that this practice is rendered obligatory by the existence of a rule of law requiring it’.Footnote ³³ In its more recent jurisprudence, the ICJ has reaffirmed that both elements of custom ‘are closely linked’.Footnote ³⁴ Thus, in ascertaining the existence of custom, one must look at what states do or do not do, and whether their conduct reflects the sense of a legal obligation. This is the crux of the traditional ‘two-element approach’, which according to the ILC ‘serves to ensure that the exercise of identifying rules of customary international law results in determining only such rules as actually exist’.Footnote ³⁵

As far as the required evidence of each element of custom is concerned, ILC Conclusion 3 illustrates the level of scrutiny that is ordinarily required from an adjudicator pronouncing on the customary nature of a given rule:

Two important aspects follow from this conclusion insofar as the practice of international courts and tribunals in identifying CIL is concerned. First, evidence of practice must be assessed considering the overall context, the nature of the rule and the circumstances in which the evidence is to be found. Second, there should be an independent demonstration of each of the two constituent elements in the reasoning of international courts and tribunals. In other words, two distinct inquiries must be carried out. On the one hand, the adjudicator must be satisfied that the relevant practice exists and is sufficiently widespread, representative and, most importantly, consistent.Footnote ³⁶ On the other hand, the adjudicator must ascertain that that practice is accompanied by the sense of legal right or obligation. While this approach is commonly accepted and sound in theory, the actual practice of international courts and tribunals is sometimes a rather different reality.

Judges of international courts and tribunals have been described as ‘technician[s] of the application of international law’.Footnote ³⁷ But it is no secret that they are much more than that, particularly when it comes to unwritten sources of international law. As the scribes of CIL, seeking to make sense of the unwritten practice of states in the reasoning of their decisions, judges speak with authority and expertise. They are certainly sophisticated scribes and not robotsFootnote ³⁸ for they do not follow a prescribed form of legal reasoning. It is thus unsurprising that every time an international court or tribunal, and especially the ICJ, renders a judgment or an advisory opinion, there is no shortage of opinions on what the Court did right or wrong (often beside the point of what was actually in dispute). Be that as it may, what a scholar expects from the reasoning in a judicial or arbitral decision is often quite different from what the parties to the case as well as judges, arbitrators or other external observers, including states, consider to be sufficient. In addition, a host of factors, many of which may not be visible to outside observers, influence the content of that reasoning. This may also explain why the content of such reasoning regarding the identification of CIL is so variable across different institutions and even from the same institution over time.

Institutional and practical constraints are particularly evident when a court or tribunal is called to pronounce on the existence and content of CIL. Institutionally, there is an expectation of efficiency and good administration of justice, which would not allow to undertake a comprehensive analysis of the practice and opinio juris of almost 200 states in every single case and in respect of every single rule of CIL that the parties may seek to rely on. In many instances, it would be a hopeless or non-manageable exercise; in others, there is simply no need to reinvent the wheel where the rule or principle in question is well-established. Practically, there are also several limitations that may prevent international courts and tribunals from making their legal reasoning and demonstration of CIL more comprehensive and consistent with the traditional two-element methodology, which are often overlooked in the existing scholarship. Three stand out in particular. First, a judgment or an advisory opinion is not an academic exercise; it aims to make the legal reasoning as succinct and clear as possible to dispose of the relevant issues. Second, it is not always feasible to arrive at a comprehensive and/or representative selection of state practice and opinio juris in the preparation of a decision. The issue of selectivity and unbalanced representation of practice (either because the practice of many states is simply unavailable, unreported or inaccessible) is often addressed by being less specific in order to secure a more convincing majority or bypass issues that could undermine the logical structure and coherence of a decision as a whole. Some courts and tribunals have expressly acknowledged these concerns as directly impacting how they present their reasoning on CIL.Footnote ³⁹ The third and final constraint, which applies to judges even more prominently than to academics, is that of language, and with it legal culture, and the influence it has on ‘how its speakers conceptualise, and therefore approach, legal reasoning’.Footnote ⁴⁰

Keeping the above constraints in mind, this chapter will demonstrate that there is a considerable variety of shortcuts in the reasoning of international courts and tribunals. These shortcuts show in turn that the methodology for the identification of CIL, as laid down in North Sea Continental Shelf, is often sacrificed for the sake of expediency. While it may be a matter of course in international adjudication, this may, in the long run, raise questions about the continuing validity of that methodology, the role that international judges play not only in the identification but also in the formation of CIL, and the coherence of CIL as a source of law made by states and for states.

3 Shortcuts in the Reasoning of International Courts and Tribunals on CIL

Having recalled the basic tenets of the traditional methodology for the identification of CIL and the challenges faced by international courts and tribunals, this section turns to their actual practice in recent years. It will quickly become apparent that international courts and tribunals have often found shortcuts in their reasoning to sidestep a full-fledged demonstration of the application of that methodology. The logic of ‘less is more’ is a unifying aspect of many of the decisions analysed below. This chapter will only discuss a few examples of cases without aspiring to be comprehensive. However, the set of cases analysed below is sufficiently representative to show that recourse to these shortcuts is on the rise. It is a phenomenon which is visible both across various international courts and tribunals, and within a single institution, as the example of the ICJ aptly demonstrates.

This review of the recent practice also shows that there is a variety of shortcuts that international courts and tribunals have followed when identifying CIL. In this author’s view, the following three approaches dominate, with apologies for the somewhat colloquial terminology: (1) ‘Check the written materials to find CIL’; (2) ‘It is CIL because the ICJ has said so’; and (3) ‘It is CIL because the ILC has said so’.

In addition to these three shortcuts which will be analysed in greater detail, there are also other ways of sidestepping the full-fledged demonstration of the traditional methodology.Footnote ⁴¹ Some of them have already been studied, at least in respect of the ICJ. There are several examples of mere assertions of CIL without any demonstration at all, but this phenomenon is not limited to the ICJ.Footnote ⁴² In some cases, the Court states that it has carefully examined the existing state practice and opinio juris, without however making any demonstration thereof.Footnote ⁴³ ‘Homework done but not demonstrated’ so to speak. It is open to question whether these relatively common instances of declaring that a given rule is or is not part of CIL constitute a shortcut to the existing methodology, or rather simply a way of presenting the conclusions without demonstrating the exact elements in support of those conclusions.Footnote ⁴⁴ Either way, these examples are methodologically unsatisfactory, because in law, just like in mathematics, the result, even a correct one, may not always withstand scrutiny without adequate demonstration.

Similarly, on several occasions the ICJ appears to have accepted the existence of an agreement of the parties to a dispute on the CIL status of a given rule as instrumental in reaching its conclusion on the subject, without any additional inquiry into state practice or opinio juris beyond those two states.Footnote ⁴⁵ There are instances of other international courts and tribunals taking the same shortcut.Footnote ⁴⁶ While such an approach may be justifiable in cases of local or regional custom, it is unclear as to why or how the purported agreement of the parties to a given dispute sheds light on the existence or absence of a particular rule of CIL on a universal scale. Although scholars have considered that this approach may allow the ICJ to signal impartiality due to the institutional constraints it faces,Footnote ⁴⁷ it may equally lead to expansive and not adequately supported conclusions.Footnote ⁴⁸

Other shortcuts are not sufficiently widespread to merit an in-depth discussion given the limited scope of this chapter. For instance, one decision has been identified where an investor-state arbitral tribunal held that it had to determine the content of a rule of CIL by looking into indirect evidence, such as judicial decisions or scholarly writings, because otherwise it would be compelled to declare non liquet.Footnote ⁴⁹ Customary international law thus conceived would be nothing but a means available to the adjudicator to fill the gaps of international law. Be that as it may, what all these and other approaches have in common is that they fuel this and other authors’ concerns about departing from any, even if minimal, demonstration of the application of the methodology for the identification of CIL and the impact it may have on the certainty and predictability of international law.

4 Conclusion

This chapter has shown that international courts and tribunals employ various shortcuts to the methodology for CIL identification. In their decisions, international courts and tribunals have often sidestepped an inductive analysis of the two elements, and have found comfort in indirect evidence such as written materials, prior judicial or arbitral decisions, or the work of the ILC. This is telling of the fact that beyond the dichotomy of ‘traditional’ and ‘modern’ CIL, as it has been discussed in scholarship, perhaps the time is ripe to speak of ‘postmodern’ approaches of international courts and tribunals to the identification of CIL. While these approaches do not expressly reject the traditional methodology, the reasoning employed is terser, more assertive, and often fails to provide any demonstration of state practice or opinio juris, in whichever order or form. Thus, such ‘shortcuts are just too appealing not to be taken’.Footnote ¹¹⁰

Of course, some of these shortcuts may be more or less justified in light of various factors, including the particular circumstances of the case, the subject-matter in which such determinations are being made, the level of institutional integration of the dispute settlement mechanism, the authority with which it is endowed, and the considerations of efficiency and economy of means. Incidentally, these approaches may preserve ‘the inherently flexible nature of this source of international law’.Footnote ¹¹¹ They may also be instrumental in obviating inherent concerns about selectivity or political expediency when embarking upon a more thorough demonstration of relevant state practice and opinio juris in the reasoning of any court or tribunal. As noted by Judge Tomka, former president of the ICJ, ‘the Court has never found it necessary to undertake such an inquiry for every rule claimed to be customary in a particular case and instead has made use of the best and most expedient evidence available to determine whether a customary rule of this sort exists’.Footnote ¹¹²

However, as the decisions referred to in this chapter show, the fundamental issue is that the legal analysis undertaken by international courts and tribunals too often fails in demonstrating even a minimal inquiry into those material elements of custom. Thus, although in principle many of the shortcuts could be justified in light of the various institutional and practical constraints referred to in the introduction, these shortcuts become a serious issue when they are the sole or the dominant element in the reasoning underlying the identification of CIL.

In the long run, the summary and flexible approach according to which the ICJ and other international courts and tribunals have gone about identifying CIL may lead to systemic issues. This author sees the potential for at least three. First, the more frequent use of shortcuts brings with it an increased risk that conclusions are being reached that are not fully supported by the practice of states and opinio juris, thus departing from or undermining the traditional methodology for the identification of CIL.Footnote ¹¹³ Second, and relatedly, judicial declarations of CIL may determine the direction of further development of state practice or, even worse, hamper the development of the law in a given area. The power of the court or tribunal to identify, or not, a given norm as part of CIL has an immeasurable impact on developing or, conversely, arresting ‘processes of growth without which the law will be atrophied’.Footnote ¹¹⁴ Once an international court or tribunal, particularly the ICJ, declares that a rule is part of CIL, states rarely if ever question the validity of that finding in their subsequent practice.Footnote ¹¹⁵ The same holds true for other international courts and tribunals, which rarely if ever question the validity of findings on CIL made by their international peers.Footnote ¹¹⁶ Domestic courts follow suit, and their restatements based on the pronouncements made by their international peers become the relevant state practice, thereby generating a vicious circle as far as the development of CIL is concerned.Footnote ¹¹⁷ In such circumstances, pronouncements by international courts and tribunals on CIL often are just short of a self-fulfilling prophecy of CIL, and domestic courts simply materialise that prophecy. Third, the increasing use of shortcuts in the identification of CIL may definitively cast doubt on a legal fiction, according to which ‘judges merely state, but never create – the law’.Footnote ¹¹⁸ This would have important flow-on consequences for the distribution of powers in the existing lawmaking framework in the international legal order, however imperfect and unsatisfactory it may be.

These potential systemic issues are not to be dismissed lightly. At the same time, they are not insurmountable, as there are several examples in the jurisprudence of much more satisfactory efforts, even if perhaps not perfectly comprehensive, at demonstrating the process and the evidence on the basis of which CIL is identified. The ILC Conclusions are certainly a useful reminder for judges and arbitrators to follow in ascertaining CIL.Footnote ¹¹⁹ In the interests of legal certainty and predictability, it is hoped that greater methodological rigour and formalism will prevail over the expediency offered by shortcuts in the reasoning of international courts and tribunals on CIL.

1 Introduction

A number of scholarly contributions on the theme have tackled the determinationFootnote ¹ of rules of customary international law (CIL) under the umbrella of the methodological dualism between induction and deduction.Footnote ² Induction indicates the method of extrapolating a general rule by observing specific instances of practice; deduction is instead the method whereby a specific rule can be inferred from generally accepted rules or principles.Footnote ³ ‘Filling lacunae’ by ascertaining rules of CIL is a canonical example of deduction. Accordingly, two main approaches have been described as underpinning the ascertainment of rules of CIL by interpreters. Pursuant to the former, a rule of CIL may be induced from patterns of state practice and opinio juris. This way of ascertaining rules of law proceeds from the observation of empirical facts and, via induction, finds rules of customary law which are created by the combination of the two constitutive elements.Footnote ⁴ As such, ‘lawyers move behind the law and cannot pretend to lead it’.Footnote ⁵ For the latter, instead, a rule of CIL may be (logically) deduced from the existence of axiomatic rules or principles of international law, for example the principle of sovereign equality between statesFootnote ⁶ or the principle of good faith.Footnote ⁷ This way of reasoning is based on the fundamental assumption that international law is a system of rules where claims to the existence of CIL rules draw justification from their coherence with other rules within the system.Footnote ⁸

However, the methodological dualism between induction and deduction is too ambitious and short-sighted at the same time. It is too ambitious, because it presumes that an extensive review of empirical elements would point to the existence of a legal rule presumably and incontrovertibly existing ‘out there’, ready to be singled out; and it is too short-sighted, because it disguises – as empirically or logically based – the argumentative nature of claims to existing rules of CIL and the role that judicial discretion plays therein. Interestingly, the methodological oscillation between induction and deduction may be portrayed as a struggle between a historical and a philosophical approach to the identification of rules. While the historical approach (induction) would point to the collection of facts as empirical evidence from which to extract a certain historical narrative, on the other hand the philosophical approach (deduction) would serve as an efficient short-cut to make a logically based descriptive claim of the law.Footnote ⁹ Importantly, both approaches strive to advance claims to scientific truths, thus leaving little space to the contestation of such findings.

In light of the foregoing, this chapter has a twofold aim. First, it recalibrates the debate surrounding the ascertainment of CIL towards an argumentative lens. Such a recalibration is conducive to illuminate the element of discretion involved in the ascertainment of rules of CIL, which remains controversially clothed in a method-focused debate. Importantly, this implies looking at potential rules of interpretationFootnote ¹⁰ of CIL not as a method to find the law ‘out there’, but rather as shared arguments to justify any claim to existing rules of CIL.

Secondly, this chapter clarifies an irony surrounding the determination of rules of CIL. If, on the one hand, illuminating the element of discretion defeats the idea of an entirely objective reality observable by courts; on the other hand, the authoritative verbalisation of such rules by courts is necessary for their materialisation and for their coming to fruition in the legal practice. In the absence of such authoritative verbalisation, there would hardly be any ‘rule’ of CIL; at best a rough idea of a metaphysical CIL. This is demonstrated by a number of casesFootnote ¹¹ in which, where applicable, courts have relied on prior judicial decisions ascertaining rules of CIL or of ‘soft law instruments’ codifying such rules qua written utterances on CIL.

This chapter is divided into four sections followed by a fifth conclusive one. Section 2 takes the cues from the recent work of the International Law Commission on the Identification of Customary International Law and considers the implications of the shift from a methodological to an argumentative lens for such identification. Section 3 presents a perusal of judicial decisions in the context of international criminal law illustrating the range of discretion exercised by judges in appraising evidentiary elements for the purposes of ascertaining rules of CIL. Section 4 reflects upon the role of courts for the materialisation of ‘rules’ of CIL and the correlated role that past judicial decisions play in the ascertainment of such rules. Finally, Section 5 draws conclusions.

2 Revisiting Old Myths: From Epistemological Methods to Argumentative Strategies

The work of the International Law Commission (ILC) on the identification of CILFootnote ¹² intervenes in the debate about the determination of CIL rules by tackling the long-standing question of the ‘methodology’ that interpreters ‘must’Footnote ¹³ apply to identify such rules. Indeed, the international law literature has repeatedly emphasised the difficulties linked to the determination of rules of CIL. One of such difficulties rests with the fact that evidence of state practice and of opinio juris may be interpreted differently by different courts, may be considered quantitatively insufficient to prove the existence of customary rules or to be regarded as conclusive of such an existence. Different types of practice may be taken into account, as well as different methods may be employed in this identification activity. This point was expressed by Judge Tanaka in his dissenting opinion in the seminal judgment in the North Sea Continental Shelf cases:

In the face of such difficulties, the ILC has laid down preliminary conclusions seeking ‘to offer practical guidance on how the existence of rules of customary international law, and their content, are to be determined’.Footnote ¹⁵

Two points are in order here. First, the ILC conclusions make reference to two types of activities: one ascertaining the existence of a rule of CIL, which, from a formal point of view, was created by state practice and opinio juris; the other determining the content of such an identified rule. Although both these activities are interpretive in character, they concern two ontologically different dimensions: that of law-ascertainment and that of content-determination.Footnote ¹⁶ While the former articulates itself along elements that are constitutive ingredients to a claim to an existing customary rule, the latter typically hinges on interpretive strategies such as the textualist, intentionalist and purposivist.Footnote ¹⁷ It is germane to acknowledge that while the ascertainment of rules of CIL is ingrained in a vigorous doctrinal convergence towards the two-pronged structure of state practice and opinio juris, albeit identifiable via different methods, the content-determination activity appears fuzzier and is indeed a dimension where the exercise of discretion by interpreters is left most unrestrained. This chapter primarily focuses on the law-ascertainment activity.

Secondly, by offering such preliminary conclusions, the ILC seemingly perpetuates two intrinsically entangled myths, namely the myth of a universal methodology to explore and assess state practiceFootnote ¹⁸ and opinio juris; and the myth of a hypothetical ‘out there’ where to identify already existing rules of CIL.Footnote ¹⁹ The idea of these being myths stems from a sceptical conception of interpretation, defined as an act consisting in ascribing, as a matter of choice, normative meaning to texts as well as in engaging in legal constructions, especially when no text to interpret in the former sense is available. Indeed, legal construction is particularly relevant in the context of ascertaining rules of CIL as, by definition, such rules are unwritten or, rectius, ‘unexpressed’, and are made expressed though the ascription of a normative meaning to empirical facts.Footnote ²⁰ Such definition of interpretation may be further reduced by accepting that also texts are no more than facts and therefore interpretation is no more than an act of legal construction of facts bearing a normative meaning. As a consequence, law is a set of interpretive practices in which judges play a central role in constructing the object to interpret.

Against this sceptical understanding of interpretation, the problématique of reiterating these legendary beliefs essentially rests with the normative view which produces the empirical facts upon which to substantiate the existence of a certain CIL rule. Indeed, state practice and opinio juris do not exist, under these labels, in the empirical world out there, but are an interpreter’s intellectual construction. As such, they are first identified, selected, assessed and categorised like relevant by the interpreter, as a reflection of his/her normative ideology.Footnote ²¹ In other words, the selection and assessment of practice and opinio juris are but the result of an exercise of discretion, which looms in every act of legal interpretation.

The ILC Report does not consider this stage of construction of relevant facts, but rather assumes that state practice and opinio juris are given, intelligible to interpreters in equal terms.Footnote ²² However, this position has largely displayed its limits,Footnote ²³ in that legal interpretation entails a subjective choice of the judge between different possible interpretive outcomes and, thus, it cannot be retained watertight to an interpreter’s own normative stance vis-à-vis international law as a legal order and its function.Footnote ²⁴ Once assumed that interpreters contribute themselves to construct the object of interpretation, professing that interpreters operate a finding exercise of legal rules appears a commitment of faith more than anything else. As such, questions pertaining to the law-ascertainment and the content-determination of rules of CIL are inescapably accompanied by rival ideologies about the ontology of interpretation in international law and, more broadly, about international law as a legal order.

In light of the foregoing, the ILC conclusions are worthy of reflection beyond the myth’s objectivity and ‘out-there-ness’ in the ascertainment of rules of CIL it seemingly reiterates. Rather, by moving away from understanding law-ascertainment and content-determination as a finding exercise, one could appreciate the ILC draft conclusions as directives constraining the interpreters’ range of discretion in the context of justification. In other words, evidence of state practice and opinio juris are used to justify the claim to existing rules of CIL, not to find them. Looking at induction and deduction as argumentative strategies entails that interpreters of international law lay down norm-descriptive statements about the law that require justification in order to be accepted as correct.

The implications of a recalibration from a methodological to an argumentative lens are manifold. First, it entails looking at opinio juris and state practice as corroborative or evidentiary elements, rather than truly constitutive or formative ones. Importantly, their persuasive strength rests on the fact that they are traditionally accepted as necessary ingredients to a claim to existing rules of CIL. As questions about the existence and content of CIL rules are addressed within an argumentative framework, it follows that, by way of legal justification, these findings need to persuade that they are correct.Footnote ²⁵ Secondly, understanding the ascertainment of rules of CIL as a finding exercise rather than an argumentative activity suggests that there is one objectively correct rule to which general practice and opinio juris point. Conversely, argumentation, as a process of justification, is premised upon the idea that potentially a range of different hypotheses about existing rules of CIL can be justified and regarded as correct in law.Footnote ²⁶ By admitting that different simultaneous plausible interpretations of facts and legal rules are possible, the argumentative lens emphasises the subjective element involved in the ascertainment of rules of CIL and, as such, it embraces rather than negating the diverse and competing normative views informing interpretation in international law. Thirdly, a recalibration from a methodological to an argumentative framework entails that criteria (or meta-rules) envisaged as a universal methodological roadmap to the ‘identification’ of rules of CIL – for example those proposed by the ILCFootnote ²⁷ – are instead arguments restraining the discretion of interpreters – with special regard to courts – that is, what it can be considered and how much weight shall be given to these elementsFootnote ²⁸ in determining the existence and the content of rules of CIL.Footnote ²⁹ Against this backdrop, the point is not to establish the appropriate method to identify customary international rules existing out there, but rather to establish the range of discretion which a court can possibly exercise in order for the ascertainment of rules of CIL to be reasonable and not to result in arbitrary adjudication.

3 Judicial Discretion in the Ascertainment of CIL: Clues from the Practice

The preceding sections have attempted to problematise the myth of epistemological methods reiterated in the scholarly debate on the determination of rules of CIL. In the wake of this, a twist to an argumentative lens is suggested to illuminate the element of discretion in legal interpretation, typically left in the background. Discretion, in the context of legal interpretation too, is not a concept of easy definition. One tentative definition has been provided by Cass R. Sustein as ‘the capacity to exercise official power as one chooses, by reference to such consideration as one wants to consider, weighted as one wants to weight them’.Footnote ³⁰ In Sustein’s view, ‘[a] legal system cannot avoid some degree of discretion, in the form of power to choose according to one’s moral or political convictions. … [T]he interpretation of seemingly rigid rules usually allows for discretion. But a legal system can certainly make choices about how much discretion it wants various people to have’.Footnote ³¹

Typically, in a legal order, courts are afforded some degree of interpretive discretion, enabling judges to make a choice between possible interpretive outcomes. The international legal order is no exception to this. For instance, Article 38(2) of the Statute of the International Court of Justice provides a useful illustration of the discretion vested in the court by state parties, in that it acknowledges the non-prejudiced ‘power of the Court to decide a case ex aequo et bono, if the parties agree thereto’. Likewise, in the Continental Shelf case, the International Court of Justice (ICJ) expressly recognised its power to discretionary choices: ‘when applying positive international law, a court may choose among several possible interpretations of the law the one which appears, in the light of the circumstances of the case, to be closest to the requirements of justice’.Footnote ³² Indeed, past judicial decisions on points of CIL are a good terrain to explore the way in which courts exercised discretion in the assessment of evidence of state practice and opinio juris. Qualities typically associated with rules of CIL such as repetition, generality, uniformity and duration, as well as the weight to allocate to opinio juris as compared to state practice were laid down and elaborated in judicial decisions. Arguably, these case-law-based criteria are an expression of how discretion is channelled into legal argumentation and enables the exercise of discretion by a judge to appear rationalised, rather than arbitrary, in that they offer a range of arguments that a court may put forward to justify a certain holding.

This section considers some judicial decisions, as well as separate opinions laid down by the International Criminal Tribunal for the Former Yugoslavia (ICTY) established by UN Security Council resolutions under Chapter VII.Footnote ³³ Looking at these decisions is particularly appropriate for the purposes of this contribution, given the tribunal’s mandate to apply rules that had, ‘beyond any doubt’, crystallised into CIL.Footnote ³⁴ The purpose of showcasing these judicial decisions is to illustrate, by reference to practice, the range of approaches exhibited by judges in the ascertainment of rules of CIL. Arguably, such a variation cannot be adequately explained by the methodological dualism between induction and deduction, as the evaluation of evidentiary elements supporting the existence of a rule of CIL is far from incontrovertible. After all, what judges do is to argue in favour of an interpretation rather than another based on certain elements of state practice and opinio juris.Footnote ³⁵ As such, statements about the existence of a particular rule of CIL are argumentative in nature and seek to persuade a certain audience of their correctness.

In the seminal Erdemović case, the ICTY Appeals Chamber was to consider whether, under CIL, duress would allow a complete defence to a soldier charged with the killing of civilians.Footnote ³⁶ To this purpose, national courts’ decisions and state legislations were examined. Yet, the threshold beyond which such evidence suffices to demonstrate the existence of a rule of CIL lies within the discretion of an interpreter. For instance, the joint separate opinion of Judges McDonald and Judge Vohrah, appended to the judgment is a good illustration of how elements of state practice and opinio juris are hardly incontrovertible and can be differently appraised by different interpreters.

This holding considered the defence’s survey, in its Notice of Appeal, of

Judges McDonald and Vohrah finally concluded that ‘[n]ot only is state practice on the question as to whether duress is a defence to murder far from consistent, this practice of States is not … underpinned by opinio juris’,Footnote ³⁹ since ‘the decisions of these tribunals [the post–World War Two military tribunals] or those of other national courts and military tribunals constitute consistent and uniform State practice underpinned by opinio juris sive necessitates’.Footnote ⁴⁰

The approach of Judges McDonald and Vohrah can be contrasted with the declaration of Judge Robinson to the Appeal Judgment in the Furundžija case,Footnote ⁴¹ in which the judge considered that ‘[a] global search, in the sense of an examination of the practice of every state, has never been a requirement in seeking to ascertain international custom, because what one is looking for is a sufficiently widespread practice of states accompanied by opinio juris. … [I]t is accepted that such [national] decisions may, if they are sufficiently uniform, provide evidence of international custom’.Footnote ⁴²

This strikes a significant discrepancy between the approach of Judges McDonald and Vohrah, in upholding an extensive empirical test, as formulated by the ICJ in the cited North Sea Continental Shelf cases, for ascertaining the existence of a rule of CIL, and Judge Robinson who instead submitted that a wide (‘global’) test has never been the requirement, but rather a sufficiently widespread practice. The threshold of empirical evidence demanded by the two approaches is expression of the range of discretion available to the interpreter when engaging in the ascertainment of rules of CIL.

Moreover, judges have granted a different weight to state practice and opinio juris for the purposes of establishing rules of CIL. One such illustration is offered by the Kupreskić caseFootnote ⁴³ in which the ICTY Trial Chamber acknowledged that opinio juris may play a primary evidentiary role at the expense of state practice.Footnote ⁴⁴

The ICTY Trial Chamber further elaborated on the formation of a rule of CIL prohibiting reprisals against civilians by reference to ‘widespread opinio necessitatis’ … ‘confirmed, first of all, by the adoption, by a vast majority, of a Resolution of the UN General Assembly in 1970 which stated that “civilian populations, or individual members thereof, should not be the object of reprisals”’ and by the high number of states that have ratified the First Protocol.Footnote ⁴⁶ The reference to manifold instruments such as the above mentioned UN General Assembly (UNGA) resolution of 1970, a Memorandum of the International Committee of the Red Cross (ICRC) of 7 May 1983, the pronouncement of ICTY Trial Chamber I in Martić, ‘substantially upholding such a rule’,Footnote ⁴⁷ shows the intention of the chamber to find ample corroboration to its claim to the existence of a rule of CIL. This overview, in the Kupreskić case, finally led the chamber to conclude that ‘the demands of humanity and the dictates of public conscience, as manifested in opinio necessitatis, have by now brought about the formation of a customary rule also binding upon those few States that at some stage did not intend to exclude the abstract legal possibility of resorting to the reprisals under discussion’.Footnote ⁴⁸

In the Furundžija case, the ICTY Trial Chamber was to establish the customary character of the prohibition of torture in time of armed conflict. The chamber found that ‘the broad convergence of international instruments and international jurisprudence demonstrates that there is now general acceptance of the main elements contained in the definition set out in article 1 of the Torture Convention’.Footnote ⁴⁹ In particular, indication of the customary character of the prohibition of torture in time of armed conflict was inferred from the number of ratification of relevant international treaties, as well as in the lack of opposing claims by states purporting the contrary.Footnote ⁵⁰ This finding was finally sealed by reference to relevant ICJ judicial decisions.Footnote ⁵¹

This overview of judicial pronouncements suggests that judges play a fundamental role in the ascertainment of CIL. In particular, judges’ verbalisation of ‘rules’ of CIL in judicial decisions appear a propaedeutic step for making such rules materialise in an authoritative form and bringing them to fruition in legal practice. Courts’ engagement in such verbalisation may also be determinant to assess the interpretive steps (meta-rules) claimed to have been adopted for the ascertainment of such rules and possibly challenge them. As recalled earlier, judges may engage in the formal ascertainment of rules of CIL, as well as in the determination of their substantive content.Footnote ⁵² While for the former, state practice and opinio juris occupy a prominent role in legal argumentation, for the latter courts are seemingly inclined to refer to existing written formulations as bearing normative value. In fact, reference to existing written formulations allows a court to articulate an interpretation of the content of existing rules of CIL in a more persuasive way.

4 The Materialisation of ‘Unexpressed’ Rules and the Role of Past Decisions

Based on the judicial decisions considered thus far, at least two factors have played a role in allowing the interpreter to modulate the range of discretion: first, the threshold of empirical evidence required for a claim to CIL; second, the more or less weight that an interpreter may attribute to state practice and opinio juris as evidentiary elements. In addition, one may consider factors which instead appeared to constrain a judicial exercise of discretion. For instance, the following examples show that prior written formulations of unexpressed rules – first and foremost, although not exclusively, judicial decisions – were typically relied upon in international adjudication.

In the recent Chagos Advisory Opinion,Footnote ⁵³ the ICJ was to determine ‘when the right to self-determination crystallised as a customary rule binding on all States’.Footnote ⁵⁴ After recalling the trite adage that ‘custom is constituted through general practice accepted as law’, the court turned to the UNGA resolutions to survey the evidence of state practice, which it considers relevant and determinant for sealing the customary nature of the right to self-determination, notably resolutions 637 (VII)/1952, 738 (VIII)/1953, 1188 (XII)/1957 and 1514 (XV)/1960. The court regarded this latter as ‘a defining moment in the consolidation on State practice on decolonization’ clarifying ‘the content and scope of the right to self-determination’.Footnote ⁵⁵ In ascertaining the customary character and the substantive contours of the right to self-determination, the court thus deferred to UNGA resolution 1514/1960 not only as declaratory of the existing customary right to self-determination,Footnote ⁵⁶ but also to determine ‘the content and scope of such a right’,Footnote ⁵⁷ namely to interpret such a right.Footnote ⁵⁸ Unsurprisingly, such material is used by the court to justify the claim of ascertained rules of CIL having a certain meaning.

In the Rwamakuba case,Footnote ⁵⁹ the International Criminal Tribunal for Rwanda (ICTR) Appeals Chamber was confronted with the question whether joint criminal enterprise was an existing mode of liability under CIL, whereby conviction of an individual was permissible. The chamber approached the question by reference to state practice and opinio juris, but instead of engaging with these elements, it upheld the finding in the Tadić Appeals Judgment pursuant to which the participation to a common plan to commit a crime against humanity was criminalised under CIL before 1992.Footnote ⁶⁰ The ICTY Appeals Chamber has placed similar reliance in other cases on proceedings held following World War II, including the proceedings before the International Military Tribunal and before tribunals operating under Allied Control Council Law No 10 (‘Control Council Law No 10’), as indicative of principles of CIL at that time.Footnote ⁶¹

Similarly, in the Kayishema & Ruzindana case, the Appeals Chamber considered the principle of the right to a fair trial as ‘part of customary international law … embodied in several international instruments, including Article 3 common to the Geneva Conventions [See Čelebeći Appeal Judgment, §§138 and 139]’.Footnote ⁶² In the Hadžihasanović et al case, the ICTY Appeals Chamber considered that ‘to hold that a principle was part of customary international law, it has to be satisfied that State practice recognised the principle on the basis of supporting opinio juris’.Footnote ⁶³ By reference to the ICJ judicial decisions concluded that ‘Article 3 common to the Geneva Conventions of 1949, which has long been accepted as having customary status [See Corfu Channel, Merits, I.C.J. Reports 1949, p. 22, and Military and Paramilitary Activities in and against Nicaragua, I.C.J. Reports 1986, pp. 112 and 114].’ In the same case, the Appeals Chamber found ‘that the customary international law rule embodied in Article 3(e) is applicable in all situations of armed conflict [international and non-international], and is not limited to occupied territory [Kordić Appeals Judgement, §78 (“[t]he prohibition of plunder is general in its application and not limited to occupied territories only”)]’,Footnote ⁶⁴ and that, as such, ‘violations of the prohibition against “plunder of public or private property” under Rule 3(e) entail, under customary law, the individual criminal responsibility of the person breaching the rule’.Footnote ⁶⁵ Similarly, in the Tadić Appeal Judgment, the ICTY Appeals Chamber found case law to be reflective of CIL.Footnote ⁶⁶

At a very first glance, the ascertainment of rules of CIL, more than any other ambit, seems to confirm the tenets of a legal realist approach to law. If law is fact, namely the law which is applied in practice by courts, what else than ‘finding’ rules of CIL can prove that such rules are brought to ‘reality’ through judicial pronouncements? Indeed, the ascertainment of ‘unwritten law deriving from practice accepted as law’Footnote ⁶⁷ entails important juristic and epistemological implications. From a juristic standpoint, the ascertainment of rules of CIL consists in an act of interpretation carrying with itself claims of formal and substantive validity. From an epistemological point of view, the act of ascertainment presupposes that rules of CIL exist ‘out there’ and that an interpreter may bring them to perceived ‘cognition’ or to ‘reality’, hence to fruition of actors in the international legal practice.

In relation to this, two entangled questions are in order. First, what kind of act is the act of ascertaining rules of CIL? It is argued that this is an act of legal construction that is adjudicative, not cognitive, in nature.Footnote ⁶⁸ Second, are interpretive utterances claiming the existence of CIL norm-descriptive or norm-expressing statements? In Alf Ross’ view, judicial decisions may be considered as norm-descriptive statements about the law, as opposed to deontic rules, which are norm-expressive statements of the law.Footnote ⁶⁹ More precisely, the written formulation of rules of CIL in judicial decisions provides these rules with an authoritative text constituted by the written utterances of what the court ascertained as existing rules of CIL and what it interpreted as their normative meaning. This owes to, among other things, the nature of international law, and law more generally, as a learned profession in which participants – including courts – articulate verbal/written expressions about the formal and substantive validity of the law.Footnote ⁷⁰ Importantly, such verbalisation stems from an evaluative process – entrenched in an exercise of discretion – channelled through the judges’ normative ideologyFootnote ⁷¹ about what they believe exists – or should exist – as a matter of legal rules, universally binding qua CIL. Within this learned profession, judicial decisions constitute authoritative statements on rules of CIL, embedding a standard of correctness.Footnote ⁷² As such, this actual formulation of rules of CIL in their form and content is necessary in order for ‘rules’ as such to materialise, as well as to formally and substantively challenge such rules on the basis of a cognised formulation. Even more so, if courts claim to have found rules of CIL based on state practice and opinio juris. Whether those verbal expressions truly reflect existing law is arguably irrelevant as long as those expressions are accepted as correct.

As such, judicial decisions verbalising rules of CIL fall short to be considered as purely norm-descriptive statements on the law, as they embed the (deontic) expression of rules of CIL. In other words, sentences which formulate unexpressed norms are ‘secretely prescriptive’,Footnote ⁷³ as they pretend to be describing existing law but are actually constructing new rules.

To illustrate this ambiguity, one may refer to the ILC Report on the identification of CIL mentioned above, whose proposed meta-rules are not laid down in a vacuum. Rather, they considerably draw from ICJ pronouncements determining the qualities of the constitutive elements of CIL, that is, the criteria necessary to claim the existence of a CIL rule. For instance, in the commentary to Draft Conclusion 2, the ILC maintains the same criteria for the identification of rules of customary law as those established by the ICJ in its judicial decisions:

The ample reliance on these judicial decisions suggests that criteria determined therein have been accepted as correct. In particular, criteria such as ‘settled practice’ or ‘consistent practice of the majority of the States’, found in judicial decisions inasmuch in the report of the ILC, stem from the discretion that a court enjoys in the adjudication of legal issues – that is, they are set forth according to the discretion which the court considers it is able to exercise – and have the power to limit or further enlarge the measure of discretion afforded to the judge in later cases. The ILC Report sanctions the criteria relevant for the ascertainment of rules of CIL that have been considered persuasive. Furthermore, the determination by the ILC that the test to ascertain the existence of a rule of CIL ‘must always be: is there general practice accepted as law?’ is eloquent for the constraint to interpretive discretion which the ILC conclusions, too, seek to place onto subsequent interpretive authorities.

The spurious nature of judicial decisions ascertaining rules of CIL as merely norm-descriptive statements is further exacerbated by the sceptical understanding of interpretation discussed above, looking at it as an argumentative art rather than an exact science. In fact, courts ascertaining rules of CIL operate an existential interpretationFootnote ⁷⁵ and may not be regarded as performing a merely declaratory function. Although this outlook bears the marks of legal realism,Footnote ⁷⁶ it is not limited to it. Admittedly, even Hans Kelsen argued that ‘the function of adjudication is constitutive through and through’ and ‘the judicial decision is itself an individual legal norm’.Footnote ⁷⁷

5 Conclusions

Qua unwritten by definition, CIL seems to appertain more to a metaphysical dimension than to the world of reality. In this scenario, the judge seemingly plays an intermediary role between the metaphysical dimension of intangible CIL and the world of reality in which rules materialise through the pronouncements of the judge. As such, courts may be seen as bringing CIL to real life – as opposed to a metaphysical dimension – drawing from a world of hypothetical rules of CIL. In ascertaining the existence of such rules, and formulating their content,Footnote ⁷⁸ courts lay down written utterances of otherwise unwritten ‘law’ presumably existing ‘out there’. In other words, the route from the metaphysical space to the world of reality channelled by courts enables the materialisation of rules (verbalised in written utterances), the scrutiny of the methods and criteria (meta-rules) used to ascertain such rules, as well as the evaluation of the evidence that a court considered.

Courts are in a special position to pronounce such statements because of the authority typically vested in them within a legal order. As argumentative strategies, induction and deduction enable courts to portray the ascertainment of CIL as an act of finding, which does not depend on an exercise of discretion, but rather sets the interpreter in the context of exploring an objective reality. This ascertainment confers to CIL an aura of objectification and divests it of the potential criticism as judge-made law. As such, interpretation – which entails a discretionary choice between possible interpretive outcomes – is perceived as an act of cognition rather than adjudication. Discretion not only lies in the power to make such a choice, but also in formulating a hypothesis about a presumably existing rule of CIL, as a reflection of, inter alia, the ideal of international legal order that a court seeks to realise, as well as in regarding certain principles of international law as axiomatic. Accordingly, a judge may do away with the principle of sovereign equality between states less easily than – say – with the principle of responsibility to protect, depending on which normative ideology he/she would present as axiomatic.

Against this background, this chapter has revisited the methodological dualism between induction and deduction as applied in the context of the ascertainment of rules of CIL. Revisiting such dualism came with suggesting embracing an argumentative lens. Like shifting lenses may entail empowering or disempowering one’s sight, similarly, twisting a methodological focus, which has featured the legal discourse on the identification of CIL, towards an argumentative lens may entail that elements which previously appeared obfuscated become more candid and vice versa.

It has been contended that while the methodological lens obscures the range of discretion exercised by the court in the ascertainment of rules of CIL, the argumentative lens sheds light on it, insofar as a claim of the existence such rules necessarily entails the selection and assessment of state practice and opinio juris which is far from being incontrovertible. The cursory survey of judicial decisions, primarily drawn from the field of international criminal law, has sought to show the different argumentative strategies whereby judges evaluated ‘evidentiary elements’ (state practice and opinio juris). Whether and how judges engage in the argumentative strategies of induction or deduction of existing rules of customary law is after all a discretionary choice. Yet, judicial decisions verbalising rules of CIL are necessary for the materialisation of such unexpressed rules in an authoritative form, as well as for the contestation of such rules, based on the arguably identified form and content. As such, courts play a fundamental role to nurture the myth of rules of CIL as an empirically based discovery rather than a discretion-centred activity.

The ample reference to prior judicial decisions corroborates the fundamental role played by courts in interpreting the world of facts bearing a normative significance (‘practice accepted as law’) and in verbalising ‘rules’ of CIL. In other words, courts are in a special position as interpreters, insofar as their pronouncements are understood as authoritative statements on the law embedding a standard of correctness, upon which actors in a legal field can rely, and which seemingly motivates actors to reiterate the myth of rules of CIL existing ‘out there’.

1 Introduction

The traditional definition of customary law follows the wording of Article 38 of the Statute of the International Court of Justice (ICJ), which sets forth ‘international custom, as evidence of a general practice accepted as law’ as a source of international law.Footnote ¹ This formulation has been read to reflect two elements constituting customary law: (i) a general practice (objective element) which is (ii) accepted as law, the so-called opinio juris requirement (subjective element). In its Draft Conclusions on the identification of customary international law (CIL), the International Law Commission (ILC) stated that ‘[t]o determine the existence and content of a rule of customary international law, it is necessary to ascertain whether there is a general practice that is accepted as law (opinio juris)’.Footnote ²

However, while in theory determining the existence of the two constitutive elements of CIL (practice and opinio juris) is the accepted methodology for the identification of a customary rule,Footnote ³ the practice of international tribunals does not always follow such methodology.Footnote ⁴ Although the ILC conclusions and commentaries provide clarifications on the theoretical underpinnings for the identification of a rule of customary law, in practice the determination of its existence is far less clear. At the same time, international case law can provide great clarification on the existence, content and scope of CIL.

Against this backdrop, this contribution examines the approach followed by the World Trade Organisation’s (WTO) Appellate Body (AB) on the identification of and resort to CIL. One should recall that the WTO dispute settlement mechanism (DSM) has one particularity: its jurisdiction is limited to ascertaining violations of WTO law. With this in mind, the aim of this article is twofold: first, to determine how the AB ascertains the existence and content of a customary rule. Second, to examine whether the AB recurs to this source of law for the interpretation of WTO provisions, or whether it directly or indirectly applies CIL.

To this end, Section 2 reviews the rules which have been considered CIL by the AB. It examines the method of identification employed by the adjudicators to qualify a given rule as ‘customary’. For the sake of clarity, ‘method of identification’ is here understood as the approach followed by adjudicators when ascertaining the existence of CIL. Section 2 also analyses the general approach by the AB towards the identification of and reliance on CIL. Section 3 studies the AB’s references to CIL in order to determine whether the adjudicators have referred to this source of law for interpretative purposes, or whether they have applied it as more than interpretative tools. While the relevance of the practice of panels is not dismissed, this contribution focuses on reports issued by the AB, as it is the permanent organ of WTO dispute settlement.

2 The AB’s Methodology of Identification of CIL

Although the AB’s practice reveals reference to several non-WTO sources and concepts of law,Footnote ⁵ only in few instances the adjudicators have declared the customary status of a rule. More specifically, such references cover only two ‘areas’ of international law: rules governing the law of treaties (described in Section 2.1) and the law of state responsibility (described in Section 2.2).Footnote ⁶ Section 2.3 describes the trends and draws general conclusions on the method employed by the AB, addressing in particular the question of what it considers to be CIL.

3 Use for Interpretation versus Application of CIL: Where Is the Line Drawn by the WTO AB?

World Trade Organisation dispute settlement has limited material jurisdiction, as it can only adjudicate WTO obligations. The difference between the applicable law (understood as the sources to which reference can be made) and the jurisdiction of WTO dispute settlement (understood as the sources which can be enforced) has been intensively debated by the scholarship.Footnote ⁴⁶ It seems well-settled that WTO dispute settlement only has jurisdiction over the so-called WTO covered agreements,Footnote ⁴⁷ while its applicable law can range further than that.Footnote ⁴⁸ Therefore, panel and AB reports cannot enforce obligations deriving from external sources.

This section enquires which role the AB has granted to CIL in settling disputes in the multilateral trading system. Section 3.1 proposes a working distinction between the concepts of interpretation and application of law, specifically for the purposes aimed at here, that is, to distinguish between the use of CIL for the interpretation of WTO provisions, as opposed to the application of customary law as a source of rights and obligations. Section 3.2 departs from this definition to examine the instances in which the AB has referred to CIL. The cases in which the AB has resorted to CIL can shed light on understanding the line between interpretation and application of non-WTO law in WTO case law.

4 Conclusions

Reference to CIL in the AB case law is very limited, both in scope and in methodology. The AB does not thoroughly follow the two-element approach to identify CIL, and it restricts its resort to this category of norms to codified secondary and meta-norms. Thus, in resorting to these sources the AB does not engage in a query of the constitutive elements for the identification of CIL. Instead, it bases the determination of customary law on reference to authoritative texts such as relevant scholarship, ICJ decisions and ILC commentaries. The very fact that the rules at stake have been codified may be another reason why the AB has referred to them. This gives the adjudicators not only an authoritative source to refer to when invoking such norms, but also allows them to resort to customary law without having to proceed to the identification of these sources.

The AB referred to what could be understood as opinio juris only once, when the adjudicators looked for instances outside the WTO system in order to confirm that the United States had recognised the proportionality principle as customary. Yet, the AB did not clarify whether it invoked the United States’ position as reflective of opinio juris. Moreover, in this case, the United States was the party being ‘affected’ by the reasoning flowing from resort to this principle. For this reason, reference to its recognition of the rule as customary seems to have been crafted to gauge legitimation for that specific finding.

Moreover, the AB adopts a cautious approach in determining which rules reflect CIL: it has only done so with respect to concepts that are operational and have no autonomous content. In fact, the adjudicators have refrained from determining the customary status of concepts which could be viewed as having autonomous substantive content and of creating rights and obligations not provided by the WTO legal system, such as the precautionary principle. This arguably also contributed to shelter the AB’s interpretative practices from claims of judicial activism (at least with respect to references to non-WTO sources of law).

From these considerations, it can be inferred that the AB has not been concerned in giving a contribution to public international law through the identification of customary rules as an authoritative international adjudicative organ. While it seems aware of the need to bridge the relationship between the trade law regime and general international law, its reference to CIL is instrumental and widely attentive to internal legitimacy questions.

1 Introduction

In orthodox international law, the formation of customary international law (CIL) takes root in the practice and opinio juris of states. The prevalence of this doctrine echoes international law’s state-centric tradition. The post-war international legal order, however, has deeply changed in many respects. Two of those changes are particularly pertinent to the present topic: first, there has been a proliferation of non-state actors playing an increasingly important role in this legal system; and second, in international humanitarian law (IHL), only since the adoption of the 1949 Geneva Conventions has non-international armed conflict (NIAC), representing the majority of contemporary armed conflicts and involving non-state armed groups (NSAGs), been placed under regular and systematic regulations of international law.Footnote ¹ Against this background, it is logical to call into question the traditional doctrine as to whether it has or, if not, should have evolved to confer a role upon non-state actors, including NSAGs, in the formative process of CIL.

As shown throughout this chapter, a number of prominent scholars have uttered their approval, with or without substantive reservations, for the proposition that the practice of NSAGs shall be incorporated into, and thus directly relevant to, the formation of customary IHL. This chapter serves as a critical appraisal of this notion. Section 2 reviews the relevance of the practice of NSAGs to customary IHL under lex lata. Section 3 analyses, from a lex ferenda perspective, the credibility of the proposed rationales for incorporating the practice of NSAGs. Admitting that this proposition is theoretically possible and, in some ways, desirable, Section 4 turns to examine which types of practice of which armed groups should be potentially absorbed into the corpus of customary IHL. Section 5 zeroes in on the legal implications of effectuating this proposition on the existing frameworks of CIL and IHL.

2 Lex Lata

An initial inquiry can be made as to whether the practice of NSAGs has been recognised as an element of CIL under lex lata. ‘International custom’ is defined in Article 38(1)(b) of the Statute of the International Court of Justice (ICJ) as ‘evidence of a general practice accepted as law’.Footnote ² The use of words ‘general practice’ has led some to argue that international custom is not restricted to the practice of states only,Footnote ³ as practice may emanate also from non-state actors.Footnote ⁴ While such an interpretation arguably runs against the drafters’ intention,Footnote ⁵ it is not ruled out by the textual meaning of the term ‘general practice’. For the future purpose at least, this interpretation retains its viability.

At any rate, given this inborn ambiguity of the law in books, international institutions have bred divergent understandings of the relationship between the practice of NSAGs and CIL. The most pertinent and specific elucidation is found in the International Committee of the Red Cross (ICRC)’s Customary IHL, in which the ICRC estimated the legal significance of the practice of NSAGs as ‘unclear’ and classified it under the heading of ‘other practice’, with a view that such practice may at best contain evidence of the acceptance of existing IHL rules.Footnote ⁶ The International Law Commission (ILC), in its comprehensive Draft Conclusions on Identification of CIL, asserted that the conduct of non-state actors other than international organisations ‘is not practice that contributes to the formation, or expression, of rules of [CIL]’,Footnote ⁷ although it ‘may have an indirect role in the identification of [CIL], by stimulating or recording the practice and acceptance as law (opinio juris) of States and international organizations’.Footnote ⁸

Hitherto, only a few international institutions have recognised the direct relevance of the practice of NSAGs to the formation of CIL. In the Tadić case, the International Criminal Tribunal for the former Yugoslavia (ICTY) considered the practice of NSAGs to be ‘instrumental in bringing about the formation of the customary rules at issue’.Footnote ⁹ In the same vein, the International Commission of Inquiry on Darfur observed that many rules of customary IHL originate from the practice of states, international organisations and armed groups.Footnote ¹⁰ Be that as it may, there is little other support for such a stance.Footnote ¹¹ It is also noteworthy that post-Tadić ICTY jurisprudence tended to revert to the elements of state practice and opinio juris.Footnote ¹²

It may be concluded that under lex lata, only states and international organisations have actually been entrusted with a law-making power.Footnote ¹³ While the practice of NSAGs plays an indirect role in the formation of CIL, its direct relevance to this process has yet been generally recognised. Accordingly, debates over this issue by and large dwell in the domain of lex ferenda instead of lex lata.

3 Rationales for Incorporating the Practice of NSAGs

Scholars advocating the incorporation of the practice of NSAGs into the formation of CIL have commonly built their argumentation upon two grounds, namely, curing the legitimacy deficit of customary rules binding NSAGs and enhancing NSAGs’ compliance with IHL.

4 Scope of Incorporating the Practice of NSAGs

Despite the flaws identified, it is admitted that the proposition of incorporating the practice of NSAGs is still theoretically possible and, in some ways, desirable. Going along with this line of thinking, it has to be decided which types of practice of which NSAGs should be absorbed into the corpus of customary IHL.

5 Legal Implications of Incorporating the Practice of NSAGs

The proposed direct relevance between the practice of NSAGs and the formation of customary IHL would have legal implications for not only the theory of CIL, but also the contents of IHL. Theoretical and practical difficulties arising therefrom would affect in return the acceptability of this project.

6 Concluding Remarks

Reviewing the seemingly attractive proposal of incorporating the practice of NSAGs into the formation of CIL, this chapter demonstrates that such a concept as it has been explored thus far may not effectively solve as many problems as it may cause. Consequently, while non-state participation in international law-making is often suggested and encouraged, there could be some reasonable hesitation when NSAGs come into sight. It is acknowledged that ‘non-state actors’ is a designation for entities of diverse origins. A general acceptance of giving a bigger role to them does not naturally guarantee a place for NSAGs. There are good reasons to isolate NSAGs from those non-states entities which are deemed inherently benign (e.g., international organisations, NGOs and judges). What is at the heart of this debate appears not to be the theoretical hurdles for NSAGs to be called law-creators, but the necessity or desirability of moving towards that direction. Indeed, should most NSAGs be able to create exquisite and sophisticated norms and strictly adhere to them, many doubts and objections to their custom-making capacity would fade away.

Such a consequentialist approach is helpful for assuring that international law is evolving on a progressive track. International humanitarian law is a body of law whose implementation greatly depends on voluntary action and goodwill of belligerents.Footnote ¹⁰⁰ More efforts to engage NSAGs are certainly desirable,Footnote ¹⁰¹ but law-making represents only one option of engagement. If, through their direct participation in custom-making, NSAGs’ compliance records are not expected to be enhanced and/or the application of IHL becomes more uncertain and less protective, it might be hard to convince the mainstream in academia to accept the direct relevance project by hinging on a simplistic call for more inclusiveness with a somewhat hollow legitimacy argumentation. Solutions to the challenges raised during this debate would partially rest with further extensive empirical surveys concerning NSAGs’ behaviour. At this stage, it is not unreasonable to stay sceptical about whether such admission of NSAGs marks one of the correct directions of future development for reckoning with the contemporary challenges to IHL.

From a policy-making perspective, it is even more challenging at present to imagine that the direct relevance proposal would be welcomed by states, which are often reluctant to do anything that may legitimise the armed groups with which they are in conflict.Footnote ¹⁰² In essence, states may have a keen interest in maintaining their exclusive or dominant role in law-making.Footnote ¹⁰³ Even if states, the ICRC or other authorities agree to put this proposition onto the agenda, there would be more questions, beyond what is discussed in this chapter, waiting for answers, such as the difficulty in discerning opinio juris of NSAGs as a result of NSAGs’ general lack of knowledge of the law.Footnote ¹⁰⁴ For the sake of theoretical completeness and practical utility, proponents of this project are invited to make further elaborations on it.

1 Introduction

In January 2019 the Dominican Republic went before the United Nations (UN) Human Rights Council (HRC) for its review during the third Universal Periodic Review (UPR) cycle. In the report it submitted to the council in advance of their review it was at pains to demonstrate the various legislative measures it had taken in respect of domestic violence.Footnote ¹ The country had been criticised by NGOs for systemically failing the survivors of domestic violence both in legislative terms – prior to 1997 it was not a crime in the country – and in relation to the training of law enforcement officials.Footnote ² As such the Dominican Republic was keen to demonstrate that, following on from its review in the second UPR cycle in February 2014, it had made changes to the law to reflect the recommendations made to them on the issue of domestic violence by states conducting the review at the HRC.Footnote ³ When considering the Dominican Republic’s report in 2019 the Australian representative to the HRC expressed concern that some of the domestic legislative reforms did not go far enough, leading it to offer a series of further recommendations on domestic violence and the strengthening of police accountability.Footnote ⁴ Throughout 2019 recommendations were offered to states undergoing the UPR on domestic violence, some going further than the standards set out by the Committee on the Convention of the Elimination of all forms of Discrimination Against Women (CEDAW Committee) others more general in nature, and some building on recommendations offered to states over the previous two review cycles. Universal Periodic Review recommendations are numerous and wide ranging but, when aggregated, can demonstrate certain trends in relation to human rights protection. Although some UPR recommendations are relatively trivial in nature and others concern matters relating to commitments under human rights treaties, there is a large class of recommendations framed in legal language, recommending a specific practice in relation to the protection of human rights to the state under review that is novel or relates to a particular interpretation of a widely acknowledged right. An analysis of these recommendations can show the emergence of customary international human rights law.

The HRC was created by UN General Assembly (GA) in 2006 and the UPR process was one of the most significant features of the new body, which replaced the UN Commission on Human Rights.Footnote ⁵ There has been a wide-ranging debate about the role of GA resolutions in the formation of custom. At the 1945 San Francisco Conference which founded the UN a proposal by the Philippines to give the GA the power to enact rules of international law which would become effective and binding upon members was defeated 26–1. Yet, from the beginning UNGA resolutions were often shaped in a way that suggested that they aimed to have some form of legal effect on states. In 1951 in the Advisory Opinion on Genocide Reservations Judge Alverez observed that GA resolutions had ‘not yet acquired a binding character’ but noted that if resolutions had the support of ‘public opinion’ they might be recognised as having some form of force over a state.Footnote ⁶ During the 1960s it became clear that certain declarations contained in GA resolutions were treated as quasi-legal statements of authority – in particular Resolution 1514 which called for the end of western colonialism which was often recited in subsequent resolutions.Footnote ⁷ Rosalyn Higgins concluded in a 1965 paper that the repeated practice of UN political organs was of ‘probative value as customary law’.Footnote ⁸ There are a number of issues that arise with any analysis of the legal status of GA resolutions, such as the status of opposition to resolutions, which makes their customary status contentious.Footnote ⁹ The importance of the debate over custom and UNGA resolutions is that it provides a useful comparison point for understanding how custom can be observed in UPR recommendations.

Understanding the status of UPR recommendations is important in the context of understanding customary international human rights law. Section 2 of this chapter shows that human rights law poses a number of problems for the traditional assumption that custom requires both state practice and opinio juris.Footnote ¹⁰ Due to the way that UPR recommendations shape state behaviour and because of the importance of the review recommendations, the remainder of this chapter argues that UPR recommendations can be a useful means for observing the formation of customary human rights rules. The International Court of Justice (ICJ) has held that certain features of GA resolutions such as the context of their emergence, the language of their substantive provision and the reaction to states can be evidence of a consensus supporting the emergence of a new customary rule.Footnote ¹¹ It is possible to trace similar features within a series of UPR recommendations on a particular norm and a framework for analysing UPR recommendations and identifying custom is set out in the final part of this chapter. But as the conclusion goes on to outline this raises wider questions about the rules for identifying custom.

2 Identifying Customary International Human Rights Law

The traditional concept of customary international law (CIL), which the International Law Commission’s (ILC) Draft Conclusions adopts, is justified on the basis that the two-element concept is necessary in order to maintain the ‘unity and coherence of international law’.Footnote ¹² The idea of different branches of international law having different rules in relation to law formation was rejected by the ILC even though, as Hugh Thirlway notes, there has been a widespread literature arguing that certain fields, such as international human rights law and international criminal law, ought to be treated differently with only the requirement for opinio juris to be present.Footnote ¹³ As Jean d’Aspremont also notes, exceptionalist thinking about the role of international human rights law, drawing on the non-reciprocity of rights and the importance of human rights for a constitutional framework of international law, led to a call for the ‘argumentative structures of general international law … not [to] apply’ to international human rights law or for them to be in some way ‘loosened’.Footnote ¹⁴ The seeming inflexibility of the rules surrounding custom has led other scholars to argue that the requirements of CIL are actively detrimental for the protection of human rights.Footnote ¹⁵ There is not really space here to re-examine the different schools of thought in this debate, but there are three important practical issues which arise with any attempt to identify customary international human rights law.

Firstly, the requirement to demonstrate consistency of state practice is hobbled by the basic reality that there are widespread human rights abuses perpetrated by states and human rights practice is often wildly inconsistent.Footnote ¹⁶ Realist critics, such as Goldsmith and Posner, have explained this by arguing that divergence of practice is simply a reflection of state interest as the ‘behavioural regularities’ require external incentives, such as coercion from more powerful states.Footnote ¹⁷ Other realist critics have taken this further arguing that state interest is in practice reducible to what is required for state survival and as a consequence custom is an inappropriate vehicle for human rights, when compared to treaty law which clearly binds states and defines their obligations.Footnote ¹⁸ The counter argument to this is that the practice requirement matters less in relation to international human rights law, and, as the ICJ has clarified, the search should be for consistency of practice rather than rigorous conformity.Footnote ¹⁹ A lot of this argument depends on what is recognised as state practice, as the implementation of rights is different from commitment to the protection of rights.Footnote ²⁰ Even though there are potentially good reasons for acknowledging that acceptance of human rights norms through instruments such as GA resolutions is important in altering the normative consensus that leads to the protection of human rights, this leads to uncertainty about what precisely constitutes state practice and where the intention of a state behind a practice can be distinguished from the intention to be bound by that practice.Footnote ²¹

Secondly, there is genuine debate about the nature of the prohibited practices involved in the protection of human rights, sometimes referred to as the secondary rules problem.Footnote ²² D’Amato frames the problem thus: ‘[w]hat are the parameters of torture? … Is the battering of wives “torture”? … what constitutes “inhuman treatment or punishment”?’.Footnote ²³ The existence of a strong general consensus over particular norms – such as the prohibition of torture – does not necessarily mean clarity over the practical implications of what they entail, even if the customary status of certain rights are citied and recited by courts.Footnote ²⁴ For example, Ghana is a party to the International Covenant on Civil and Political Rights (ICCPR) and the Human Rights Committee has since 1994 consistently interpreted the right to privacy under Article 17 as being incompatible with the criminalisation of same sex relations; yet Ghana not only ignores this but has also rejected UPR recommendations to decriminalise sexual orientation.Footnote ²⁵ This speaks to a third related problem in relation to customary international human rights law – the idea that some states do not regard human rights law or human rights practice as creating a legally binding obligation upon them. Even though almost every state in the world is party to at least one legally binding international human rights instrument, mechanisms such as reservations have allowed states to manage the scope of their obligations, and states have used a variety of arguments to maintain they are not bound by treaty obligations in relation to specific rights, relating to practices considered culturally sensitive, or in areas where they have security concerns.Footnote ²⁶ The persistent objector doctrine can also perform a similar function in that it allows states to define the limits of what should and what should not be considered human rights.Footnote ²⁷

In an attempt to refocus the debate surrounding the nature of the subjective element of custom, Brian Lepard has argued that a rule or principle ought to be considered customary law if it can be shown that ‘states generally believe that it is desirable now or in the near future’ to make a ‘rule or principle’ legally authoritative ‘for all members of the global community’.Footnote ²⁸ Lepard’s argument is that ‘state practice is importance of evidence of the belief that a norm should be universally binding’ but that that it is not an ‘essential independent requirement’; a position criticised by some scholars as fusing the two elements together.Footnote ²⁹ Yet, as Lepard goes on to argue, his reformulation clarifies the role of opinio juris, because it looks at what should or ought to be binding, rather than what is or may be perceived as binding.Footnote ³⁰ A case study of this approach to custom is the GA resolution containing the Declaration on the Granting of Independence to Colonial Countries and Peoples.Footnote ³¹ This recognised emerging practice from Western states to grant colonial independence and was recognised by many states (particularly those who were newly independent) as having a legal quality to it because it supported the desirable goal of independence, whereas others (European states with colonial territories) were far more reluctant to concede its legal status.Footnote ³² Although this is not a formula for replacing the two-element rule – it instead refines how the individual elements are identified and examined – it has come in for criticism as being a form of ‘norm entrepreneurship’ which dilutes the meaning of custom.Footnote ³³ Given the competing difficulties that the identification of customary human rights law poses, there is a need not so much for new rules but for a broader consideration of the materials used in the identification of the two elements of custom. An analysis of UPR recommendations, as the remainder of this chapter sets out, provides two clarifying functions for the identification of customary international human rights law. Firstly, it allows for a transparent and more democratic way of measuring the existence of a common consensus on a particular human rights norm. Secondly, the way the UPR process works allows for the contours of any norm to be defined, which is important when that norm is a secondary or interpretative norm about the scope of a particular practice which emerges outside of an agreed codification of a particular right in a treaty or one that is recognised elsewhere as jus cogens.

3 UPR Recommendations and Their Effect on State Behaviour

Under UPR rules every UN member has their human rights record reviewed around once every four years – known as UPR cycles.Footnote ³⁴ The review is conducted based on documentary evidence from the state under review, reports from treaty bodies to which the state under review is a party, stakeholder reports from civil society groups in the state under review, international NGOs and other consultative bodies. Every state has participated in the process since its inception in 2006 and the treatment of states as equal peers has been a significant attraction of the process, differentiating it from its predecessor the UN Commission on Human Rights. During the review there is first a documentary review, examining the state under review’s performance at treaty bodies, reports from stakeholders, and its own report. Then there is an interactive dialogue between the state under review and other members of the review panel. After that there is the opportunity for all states to issue recommendations to the state under review about changes in domestic law in order to improve their human rights practices. The UPR is a political process and was not intended to be law making but it involves scrutiny and discussion of a state’s human rights obligations which leads to it sometimes overlapping with other international legal processes.Footnote ³⁵ It was meant to complement and not duplicate the work of treaty bodies and in relation to some treaties the UPR has played a role in reinforcing obligations, by recommendations being cited by treaty bodies as evidence of state practice in relation to a particular norm.Footnote ³⁶

At the time of writing in October 2021 there have been nearly 79,000 recommendations issued to states in the 11 years that the review process has been in operation. Assessing recommendations is difficult because they have no real set form and are constructed by states acting individually rather than collectively or drafting them in concert with others, as is the case with UNGA resolutions. Edward McMahon has devised a system for categorising recommendations based on the nature and quality of action required of a state, ranking them from one to five.Footnote ³⁷ Category two recommendations concern general comments about what the state under review is currently doing and a request to continue an ongoing course of action. For example a recommendation to Brazil from Senegal during the second review cycle to ‘continue fighting violence against women’ simply asked the state under review (Brazil) to do nothing beyond what they were currently doing.Footnote ³⁸ There has been some criticism of these sorts of recommendation being little more than offerings of praise from the states conducting the review. Sometimes when the state under review was their political ally, states would praise-bargain hoping that favourable recommendations would lead to the state under review affording them similar treatment when it was their turn for review.Footnote ³⁹ In 2011, at the end of the first review cycle, a lot of emphasis was placed on the process of following up the implementation of recommendations accepted by states in subsequent review cycles.Footnote ⁴⁰ Follow up can occur in the portion of the review dedicated to interactive dialogue. Sometimes recommendations specifically cross-reference previous commitments; for example, the recommendation issued to India by Botswana during its third cycle review to ‘Ratify the Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment, as previously recommended’ was designed to reinforce earlier commitments.Footnote ⁴¹

Recommendations in action categories four and five are framed using language which requires positive action on the part of a state, and directly reflects the legal language of a treaty or declaration. Recommendations classed as category five are those using legal verbs such as ‘abolish’, ‘accede’, ‘adopt’, ‘amend’, ‘implement’, ‘enforce’, or ‘ratify’.Footnote ⁴² To illustrate this with another example from Brazil in its second cycle review; Spain issued a recommendation to ‘adopt Bill No. 2442 in order to guarantee the independence and autonomy of the members of the National Preventive Mechanism, in conformity with Brazil’s obligations under OPCAT’ which is both precise in its intent and framed with clear instructions to the state party on the course of action to take.Footnote ⁴³ After recommendations have been issued it is up to states to either ‘Support’ (UPR terminology for accept) or classify a reservation as ‘Noted’ (UPR terminology for reject). There has been some criticism of McMahon’s framework as being too narrow, and concern that its focus on the type of action required to implement it obscures the utility of the recommendation for the state under review.Footnote ⁴⁴ Yet, the linguistic framing of recommendations is vital for identifying and distinguishing what could appear to be the language of obligation and commitment from general descriptive language about human rights. Category five recommendations are the recommendations that are most likely to be rejected by the state under review – in the first UPR cycle which ran from 2008 to 2011, 60 per cent of category five recommendations were rejected. In the second cycle, which ran from 2012 to 2016, 55 per cent were rejected.Footnote ⁴⁵ Significantly, in spite of this rejection rate the number of category five recommendations has remained steady over successive review cycles – 35 per cent of all recommendations in the first cycle were category five, in the second cycle it was 37 per cent of recommendations, and of the data available so far for the third cycle, 38 per cent have been category five recommendations. It is, however, important to remember that the categories are analytical tools for understanding the framing of recommendations and not a formal part of the UPR process. Whilst shaping the wording of recommendations, as they author them, states are not necessarily conscious that they are making a recommendation of one particular category or the other. They are best understood as an instrument of measure; ascertaining both the quantity of recommendations and the relative severity of linguistic framing. Moreover, while they may serve as a useful proxy for state intention, they are not always definitive proof of it.

Once accepted, a recommendation does not create an obligation upon the state under review to implement its substance. Recommendations do however affect state behaviour; firstly, the process of follow up, or at least the expectation of follow up, and the deliberative nature of the review process does encourage states to make incremental change to their behaviour in respect to their laws and policies respecting human rights.Footnote ⁴⁶ Jane Cowan and Julie Billard’s observation that states under review can treat their review like an ‘exam’, with state delegations to the UPR concerned about proving they have met minimum standards or demonstrating the implementation of legal reforms, shows that the UPR process can change state behaviour, even if it does not meet some of the loftier objectives about the promotion of dialogue on human rights intended by the UPR’s creators.Footnote ⁴⁷ Secondly, in terms of implementation and delivery there is a correlation between states accepting recommendations and implementing changes to their law and policy surrounding human rights in response to those recommendations.Footnote ⁴⁸ Sometimes this relates to a course of action already decided upon by a state party and recommendations help reinforce this course of action.Footnote ⁴⁹ Yet on other occasions there are signs that recommendations act as drivers of reform independently – for example in relation to protection from human trafficking and maternal health there has been some research showing recommendations correlate with the adoption of higher standards on these issues in countries which have accepted recommendations.Footnote ⁵⁰ There was also an upsurge in the number of states signing up to and ratifying treaties in the wake of the first UPR cycle, again seemingly in response to a wave of recommendations in the first cycle relating to treaty provisions.Footnote ⁵¹

Regular patterns of state behaviour in accordance with the terms of some GA resolutions has been cited by some scholars as proof of their customary status, even though such behaviour might not be in conformity with all of the terms of a resolution.Footnote ⁵² Regardless of whether states consider the UPR to be a form of ritualised audit or approach the review as an opportunity to advance strategic interests, states’ behaviour towards the process indicates that issuing and accepting recommendations carries a degree of importance.Footnote ⁵³ A crucial distinction between UPR recommendations and GA resolutions however is that certain resolutions, such as those on outer space or environmental issues, were claimed to create instant custom, without state practice.Footnote ⁵⁴ Some of the most compelling arguments in this debate related to recommendations containing declarations which made normative pronouncements about what the law ought to be in a particular area, and which was then put in a codified text of a resolution and voted on.Footnote ⁵⁵ In 1977 this argument was made by the Group of 77 (a UN grouping principally consisting of newly independent states in Africa and Asia) who contended that a UN resolution on the seabed should be regarded as binding because it represented a true international consensus on what the law ought to be, had been expertly drafted, and the GA resolution containing it passed without any votes in opposition.Footnote ⁵⁶ Universal Periodic Review recommendations differ from GA resolutions as they are drafted by individual states and reflect their own interest and priorities. As Gujadhur and Limon note, one of the problems the UPR process faces is the sheer weight of recommendations, many of them of variable quality and relatively poorly constructed.Footnote ⁵⁷ Acceptance and rejection of individual recommendations can mean relatively little and can vary considerably from state to state. It is therefore more important to look at recommendations in aggregate on a particular issue to see the reflection of a normative consensus on any one human rights issue.

4 A Lens through Which to See Custom in UPR Recommendations

Custom’s formation is often described as being observed rather than generated. As Anthea Roberts notes, traditionally custom was ‘inductive’ in that it was derived from an observation of state practice, whereas modern custom is ‘deductive’ in that is deduced from international instruments, such as declarations and reservation.Footnote ⁵⁸ As Stefan Talmon notes, however, it is incorrect to think of this as a choice between the two methods and at the ICJ there have been situations where it was simply not possible to use an inductive method to identify custom.Footnote ⁵⁹ What is presented here therefore is a deductive framework to use for identifying the emergence of customary human rights norms in the UPR. This framework breaks into three parts and arguably provides greater clarity in the context of the UPR than the two-element approach.

4.2 Quantification: Accepted Recommendations Making a Rule?

The ICJ’s Advisory Opinion on Nuclear Weapons noted that a ‘series of [GA] resolutions may show the gradual evolution of the opinio juris required for the establishment of a new rule’.Footnote ⁷⁸ Empirical studies on the formation and identification of custom in court briefs and submissions before international tribunals have been relatively inconclusive in establishing trends of how custom is identified, yet this does not mean that thresholds of practice cannot be established.Footnote ⁷⁹ As Christopher Joyner noted in relation to GA resolutions, even though the law making competence of the assembly was qualified, when ‘delegates representing almost all the world’s national governments cast votes on a resolution, they are in effect providing a common confirmation (or rejection) of the presence and acceptance of that issue in international law’.Footnote ⁸⁰ In the case of UPR recommendations it would mean establishing a common linguistic framing, a common subject matter and a pattern of acceptance from states with a reasonably wide geographic spread – all of which is possible using a database such as UPR Info to track the emergence or existence of such a trend.Footnote ⁸¹ A constant series of recommendations all aimed at a particular practice, which are accepted and over the course of multiple cycles are adopted by states, could amount to what the ICJ describes as a ‘general recognition’ that a law or legal obligation is involved.Footnote ⁸² Quantifying recommendations helps to establish to what extent a consensus surrounding a particular norm actually exists, which is important for establishing the existence of practice in the formation of custom.Footnote ⁸³ Even adopting a theory of customary international human rights law, of the sort outlined by Lepard, there would need to be some acknowledgement of the scale of moral consensus surrounding a particular norm, to give weight to the claim that it ought to be universal.Footnote ⁸⁴ Because of the nature of recommendations, as outlined in the second section above, it is necessary to trace a particular norm through recommendations and quantify the use of certain words in a series of recommendations on that subject.

A good case study of how this process might work is the prohibition on corporal punishment. According to the Global Initiative to End Corporal Punishment of Children, at the time of writing around 140 states prohibit corporal punishment in the criminal justice system and 132 prohibit it in the education system.Footnote ⁸⁵ There are, however, far fewer states that have prohibited corporal punishment in the home or care system and in total only fifty-six states have a total prohibition on corporal punishment as a matter of law. The European Court of Human Rights has been clear that state sanctioned corporal punishment constitutes inhuman and degrading treatment.Footnote ⁸⁶ An advisory opinion of the Inter-American Court of Human Rights stated that the American Convention on Human Rights required state parties to take ‘positive measures … to ensure protection of children against mistreatment’ especially in ‘relations among individuals or with non-governmental entities’ but stopped short of formally requiring the prohibition of corporal punishment.Footnote ⁸⁷ Article 19 of the Convention on the Rights of the Child (CRC) requires state parties to take ‘appropriate legislative, administrative, social and educational measures to protect the child from all forms of physical or mental violence’ which does not explicitly prohibit corporal punishment.Footnote ⁸⁸ In General Comment 1, the Committee on the Rights of the Child noted that in the context of protecting the right to education it had previously made clear in its concluding observations that the use of corporal punishment did ‘not respect the inherent dignity of the child nor the strict limits on school discipline’ protected in the convention.Footnote ⁸⁹ In General Comment 8 the committee went further, arguing that corporal punishment was incompatible with the requirement to protect children from harm and that laws permitting corporal punishment in education needed to be repealed.Footnote ⁹⁰ But crucially it stopped short of recommending the prohibition on corporal punishment in the home.

There have been a large number of UPR recommendations submitted concerning corporal punishment and overall, 58 per cent of them have been accepted. The vast majority of accepted and rejected recommendations are in action category 4 and 5, containing the words ‘prohibit’, ‘end’, ‘ban’ or ‘eliminate’. For example, France in its second cycle review accepted a recommendation from Uruguay ‘to explicitly ban the corporal punishment of children in all settings, including the family, schools and institutions’.Footnote ⁹¹ Other recommendations can be more explicit in cross referencing the CRC and existing obligations in international law. For example, Uruguay recommended to Algeria in its second cycle review, after commending it during the interactive dialogue for introducing a prohibition on corporal punishment in schools, that it extend the prohibition to ‘home care institutions, penitentiary centres and any other settings, in conformity with Article 19 of CRC’.Footnote ⁹² Other states have, however, been wary of recommendations which include a prohibition that would entail them prohibiting corporal punishment in the home, potentially entailing the introducing of laws which might criminalise parents; Switzerland in the third cycle accepted one recommendation on the prohibition of corporal punishment but rejected another which specifically referenced prohibition ‘in all settings, including in the home’.Footnote ⁹³ Out of all rejected recommendations referring to corporal punishment, 25 per cent of them are category four or five recommendations referring to the ‘home’, ‘family’ or other term referring to prohibition on the domestic sphere.

When analysing recommendations, a basic three-part approach to quantification of recommendations would help identify the emergence of custom. Firstly (as detailed in Figure 15.1) there would need to be a quantification of both the practice, or the noun (i.e., ‘corporal punishment’) and the verb in connection to the noun (i.e., ‘prohibit’) because the commitment to the ‘doing’ or ‘enacting’ of a human rights norm is what differentiates a mere verbal statement from something which can be considered state practice. Given the scale of friendly recommendations and praise bargaining, there needs to be a standard to distinguish accepted recommendations which might be evidence of custom from accepted recommendations which are of largely political significance – hence recommendations looked for as evidence of custom considered would need to be in action categories four and five, as they involve an active commitment from a state party.Footnote ⁹⁴ Although it is conceivable that a state might take action over recommendations in categories one to three, their vague and open-ended wording, which often lacks any clear description of subject matter or action to be undertaken by the state, means that it would be difficult to treat these recommendations as evidence of customary law. Secondly, the framing of rejected recommendations needs to be analysed to see where in the rejected recommendations there appears to be limitations of rights. For example, in the case of corporal punishment the only consistent trend in rejected recommendations on corporal punishment appears to be scepticism about extension of the prohibition to the domestic sphere. This should be read in tandem with other sources on the practice outside the UPR process – such as court decisions – to see if this would constitute a ground for rejection that indicated a substantive objection to a specific right and hence accepting the limited nature of any customary norm that could be identified in recommendations.Footnote ⁹⁵ Thirdly, the recommendations analysed would need to be sufficiently numerous – in figure 1 above the number of accepted recommendations numbers over 200 – and across more than one cycle and geographic region, to demonstrate evidence of the widespread consensus on a particular issue.

Figure 15.1 Recommendations on banning corporal punishment across UPR cycles to date.

5 Conclusion

By way of conclusion, it is worth identifying two potential lines of criticism about the framework advanced here and what it means for the identification of CIL. Firstly it is open in adopting what Noora Arajärvi critically termed the ‘paradigm shift’ toward the ‘demands of humanity’ away from the more orthodox position in some of the literature, on the evidence of state practice required for custom to be identified.Footnote ¹¹⁰ Fernando Tesón almost pre-empted the argument advanced in the first part of Section 3 with his description of ‘the Ad Nauseam Fallacy technique’ whereby ‘profusely citing nonbinding resolutions’ is used to advance a ‘sense of normativity’ that is not actually present.Footnote ¹¹¹ Yet, this neglects the institutional framework of the UPR described above. As both the ILC and the ILA investigations into the source of custom highlight, interaction with an institution such as the UPR is a key part of what a state ‘does’ both in terms of the internal procedure and the effect it has on states and in terms of the commitments that states make. This also relates to the sense of obligation. One key criticism of the way that the ICJ has interpreted the legal effect of GA resolutions is that the court has focused more on their binding nature than on the way that they shape legal discourse.Footnote ¹¹² Individual UPR recommendations do not bind the states who accept them but collectively a series of accepted recommendations demonstrate the existence of an emerging consensus on a particular norm. This means that a series of recommendations could and, from the perspective of those seeking to defend human rights, probably should, have an authorising effect – in that they identify rights that states need to protect and highlight the legal obligation to protect those rights.Footnote ¹¹³

Secondly this theory involves adopting a constructivist interpretation of both elements of the two-element theory reasoning that state interests are shaped by the social conditions that surround them (in this case the rules of the UPR process).Footnote ¹¹⁴ Rationalist and realist critics would probably reject this description of state behaviour, and any analysis of UPR recommendations gravitates heavily towards an empiricist understanding of custom. However, an empiricist understanding of accepted recommendations would allow for the identification of a consensus on a particular norm, and importantly, would provide grounds for establishing a definitive explanation of which rights ought to be protected. The UPR was not meant to be a legal process but its effects have altered the way states act towards certain norms and the recommendations they accept. The mechanism outlined above helps identify where a norm contained in recommendations could have a customary status and addresses some of the criticisms surrounding the identification of customary norms.Footnote ¹¹⁵ Overall this theory puts a heavy institutional gloss on the identification of CIL, and questions of the practice and the binding nature of norms are answered technically with reference to the nature of the UPR process. In the context of a process which has mass buy-in from states, with every country in the world being subject to at least one review, a justification for a heavy institutional emphasis on the question of custom can be constructed on the lines that Lepard outlines above as it is possible to discern from the UPR process a strong sense of what rights ought to exist.Footnote ¹¹⁶ Whilst challenging some assumptions about what constitutes practice within the existing literature, identifying custom in UPR recommendations can help give coherence to the identification of customary international human rights norms.