Part III - The Practice of Customary International Law Across Various Fora

11 Customary International Law in the Reasoning of International Courts and Tribunals

Vladyslav Lanovoy ^*

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’.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’.² Numerous recent decisions of international courts and tribunals confirm that custom is not condemned to disintegrate³ and certainly does more than simply fill gaps left by the existing treaty regimes, as has been suggested elsewhere.⁴ International law has always been and remains principally a customary law.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.⁶ To some observers it resembles ‘a riddle wrapped in a mystery inside an enigma’: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?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.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’.¹⁰ Similarly, Henkin described ‘the process of making customary law [a]s informal, haphazard, not deliberate, even partly unintentional and fortuitous … unstructured and slow’.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’,¹² an ‘inherently contingent and variable kind of law’¹³ or even ‘hopelessly indeterminate’.¹⁴ Others have decried custom as being unfit to accompany the rapid pace of developments in international relations nowadays¹⁵ and in a more heterogeneous international community.¹⁶

Accordingly, no effort has been spared at the international level to analyse the process of custom formation and methodology for its identification.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.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’.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.²⁰ Choi and Gulati argue that the ICJ has completely ignored the traditional methodology.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.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’.²³ This perhaps not fully felicitous wording of Article 38(1)(b) of the Statute of the ICJ,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).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.²⁶

As noted by the ILC, the process of identifying CIL ‘is not always susceptible to exact formulations’.²⁷ Indeed, the end-product of the Commission’s inquiry into the subject aimed to set out a ‘clear guidance without being overly prescriptive’.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.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,³⁰ if not ‘set in stone’.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’.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’.³³ In its more recent jurisprudence, the ICJ has reaffirmed that both elements of custom ‘are closely linked’.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’.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:

Conclusion 3

1. 1. In assessing evidence for the purpose of ascertaining whether there is a general practice and whether that practice is accepted as law (opinio juris), regard must be had to the overall context, the nature of the rule and the particular circumstances in which the evidence in question is to be found.

2. 2. Each of the two constituent elements is to be separately ascertained. This requires an assessment of evidence for each element.

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.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’.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 robots³⁸ 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.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’.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.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.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.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.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.⁴⁵ There are instances of other international courts and tribunals taking the same shortcut.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,⁴⁷ it may equally lead to expansive and not adequately supported conclusions.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.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.

3.1 ‘Check the Written Materials to Find CIL’

The relationship between treaties and custom is longstanding and intertwined.⁵⁰ It is widely accepted that treaties may: (i) reflect pre-existing rules of CIL; (ii) generate a new rule and serve as evidence of the customary character of that rule; or (iii) have a crystallising effect for an emerging rule of CIL.⁵¹ The ICJ has long recognised that ‘multilateral conventions may have an important role to play in recording and defining rules deriving from custom, or indeed in developing them’.⁵²

At the same time, CIL has an ‘existence of its own’ even where an identical or similar rule may find expression in a treaty.⁵³ As such, it is perhaps surprising to see how often international courts and tribunals resort to treaties or other written materials to identify CIL. The practice is particularly prominent in the ICJ’s jurisprudence.⁵⁴ There are numerous examples in which the Court has, with little or no additional analysis, recognised the customary status of certain treaty provisions.⁵⁵ The examples below of two recent decisions rendered by the ICJ and one by an arbitral tribunal demonstrate some of the potential problems with this shortcut. In all three cases, in identifying CIL, recourse was had to written materials, namely treaties that were not even in force between the parties to the dispute, or resolutions of the General Assembly (GA).

In the Jurisdictional Immunities of the State case, the Court had to determine whether the so-called territorial tort exception to state immunity existed under CIL.⁵⁶ While the Court attempted to demonstrate the application of the two-element methodology in its reasoning, it focused its analysis rather disproportionally on Article 11 of the 1972 European Convention on State Immunity and Article 12 of the 2004 UN Convention on Jurisdictional Immunities of States and Their Property.⁵⁷Having acknowledged that neither convention was actually in force between Germany and Italy, the Court stressed that these instruments were therefore ‘relevant only in so far as their provisions and the process of their adoption and implementation shed light on the content of customary international law’.⁵⁸ However, the Court appeared to ascribe much weight to these instruments amidst its analysis of other aspects of the relevant state practice, including case law and legislation of various states, as well as in identifying opinio juris.⁵⁹ It would thus seem that this was a conscious shortcut on the part of the Court in the process of identifying CIL.

Similarly, the Enrica Lexie arbitral tribunal recently adopted the same shortcut in examining whether the ‘territorial tort’ exception to immunity from criminal jurisdiction was recognised under CIL. The tribunal noted that even though national courts in a relatively significant number of states look at the 2004 UN Convention on Jurisdictional Immunities of States and Their Property as a reflection of CIL, the ‘states that consider that there is immunity for foreign states before other states’ national courts do not accept the provisions of this convention, including Article 12’.⁶⁰ It went on, however, to analyse the criteria set out in Article 12 of that convention to conclude that ‘even if a “territorial tort” exception were recognised under CIL, the exception would not apply’ in the circumstances of that case, as the marines were on board the Enrica Lexie, and not on Indian territory.⁶¹ Even though the arbitral tribunal ultimately did not rule on whether such an exception exists under CIL, it is telling that its reasoning relied exclusively on an unratified treaty instrument rather than on any inquiry into the relevant state practice and opinio juris.

In the Chagos Advisory Opinion, the ICJ had to determine whether the right to self-determination existed as a customary norm at the time of events, that is, in the period between 1965 when the UK excised the Chagos archipelago from Mauritius, then a non-self-governing territory administered by the UK, and 1968 when Mauritius attained independence.⁶² In determining whether the right to self-determination was part of CIL at the time, the Court held that ‘State practice and opinio juris … are consolidated and confirmed gradually over time’,⁶³ and highlighted the ‘progressive consolidation of the law on self-determination’.⁶⁴

It is noteworthy that in identifying the existence, content and scope of the right to self-determination under CIL, the Court placed much emphasis on the GA resolution 1514 (XV) of 14 December 1960, and the circumstances in which it was adopted.⁶⁵ The Court saw a clear correlation between the acceleration of the decolonisation process (with eighteen countries in 1960 and additional twenty-eight non-self-governing territories during the 1960s exercising the right to self-determination) and the adoption of resolution 1514 (XV) which ‘clarifie[d] the content and scope’ of that right.⁶⁶ It considered the adoption of this resolution to be ‘a defining moment in the consolidation of State practice on decolonization’.⁶⁷ The weight that the Court ascribed to this and other resolutions of the GA in reaching its conclusion on the right to self-determination and its content under CIL⁶⁸ was more significant when compared to its earlier jurisprudence, which had taken account of resolutions as evidence of opinio juris.⁶⁹ It also allowed the Court to effectively dispose of the issue of an allegedly inconsistent practice underlying the obligation incumbent on administering powers to respect the boundaries of the non-self-governing territory.⁷⁰ That said, in the particular circumstances of this case, the Court’s reliance on the relevant resolutions as a shortcut for identifying CIL might be said to be justified by the conditions in which these resolutions were adopted, their normative value, and the absence of any genuine opposition among states to the existence and content of the right to self-determination.⁷¹

Beyond these and many other examples of this shortcut to the traditional methodology in practice, one area of increasing interaction between custom and treaty law has been in the context of investor-state arbitration. Several investor-state arbitral tribunals have resisted the temptation of automatically relying on hundreds of bilateral investment treaties (BITs) to inform their task of ascertaining whether certain standards of protection find expression in CIL. For example, in Glamis Gold, the tribunal rightly rejected the contention that Article 1105 of the North American Free Trade Agreement (NAFTA) was merely a ‘shorthand reference to customary international law’,⁷² having emphasised that the task of seeking a treaty interpretation of a given standard is fundamentally different from that of ascertaining CIL.⁷³ The tribunal held that ‘arbitral decisions that apply an autonomous standard provide no guidance inasmuch as the entire method of reasoning does not bear on an inquiry into custom’.⁷⁴

As a result, the tribunal rejected the claimant’s so-called convergence theory between CIL and specific treaty provisions in BITs, ruling that while ‘it is possible that some BITs converge with the requirements established by customary international law; there are, however, numerous BITs that have been interpreted as going beyond customary international law, and thereby requiring more than that to which the NAFTA State Parties have agreed’.⁷⁵ Similarly, the Cargill tribunal considered that ‘significant evidentiary weight should not be afforded to autonomous clauses inasmuch as it could be assumed that such clauses were adopted precisely because they set a standard other than that required by custom’.⁷⁶

Other investor-state tribunals have however reached their conclusions on the content of CIL by relying on specific treaty provisions.⁷⁷ The temptation of adjudicators to rely on written materials is strong, particularly where these are the culmination of a codification process or seem to crystallise an emerging rule of CIL.⁷⁸ However, investment treaty context is a particularly salient example of an organic mismatch between customary and treaty law. This is the case, for instance, of the evolution of the minimum standard of protection of aliens and their property, as opposed to the evolution of the standard of fair and equitable treatment in the treaty practices over the last couple of decades.⁷⁹ Hasty attempts at converging the two, albeit perhaps desirable in the interest of a greater and more uniform protection to be accorded to investors and investments, are not justifiable through the lenses of a proper methodology for the identification of CIL.⁸⁰

The above examples show how in identifying CIL international courts and tribunals have used written materials, including treaties that are not in force between the parties or GA resolutions. They have done so, at least in part, to circumvent the practical difficulties that may arise in demonstrating the two elements of custom. As such, the existence of codification in a particular area of law allows the courts and tribunals to consider whether the instances of practice support the written rule rather than induce that rule from specific instances of practice. Some authors have seen in such increasing reliance on written law in identifying CIL a departure from ‘traditional’ towards ‘modern’ custom, from a predominantly ‘inductive’ towards a ‘deductive’ process, from the examination of particular instances of practice towards general statements.⁸¹ While there is nothing per se problematic with such a shortcut,⁸² the absence of any detailed discussion on the actual evidence of state practice and opinio juris beyond the written materials themselves may lead to unconvincing or incomplete reasoning, which could have been easily remedied with an even minimal attempt at applying the traditional methodology for the identification of CIL.

3.2 ‘It Is CIL Because the ICJ Has Said So’

Another frequently used shortcut for identifying CIL is that of relying on previous decisions of the ICJ. This shortcut poses several normative issues.

First, it suggests that the dispute settlement mechanism has adopted a de facto system of binding precedent, whereby earlier decisions constitute authoritative pronouncements on CIL, even if that law may be susceptible to change over time. This is visible, for instance, in Jones et al v. United Kingdom, where the European Court of Human Rights, when addressing the so-called jus cogens exception to state immunity, turned directly to the ICJ judgment in the Jurisdictional Immunities of the State, considering it to be ‘authoritative as regards the content of customary international law’ and that no such exception had yet crystallised in CIL.⁸³ It did so without any additional demonstration in support of its conclusion. Similarly, in cases where the World Trade Organization (WTO) Appellate Body could identify an earlier decision of the ICJ on a particular rule, it automatically accepted the customary law character thereof.⁸⁴ The issue lies in the assumption that the ICJ’s decision is dispositive on the question whether or not a given rule is part of CIL.

Second, this shortcut quite often leads to improper generalisations of the scope of earlier judicial pronouncements on CIL. The Territorial and Maritime Dispute serves as a perfect example of this phenomenon. In that case, the ICJ recalled its previous jurisprudence, namely Qatar v. Bahrain, in which it had recognised that the principles of maritime delimitation in Articles 74 and 83 of the UN Convention on the Law of the Sea reflect CIL, and so too does Article 121, paragraphs 1 and 2 thereof.⁸⁵ However, in Qatar v. Bahrain, the Court ‘did not specifically address paragraph 3 of Article 121’, which qualifies maritime entitlements of a rock as opposed to those of an island. Despite that, the Court merely observed in the Territorial and Maritime Dispute that ‘the legal régime of islands set out in UNCLOS Article 121 forms an indivisible régime, all of which (as Colombia and Nicaragua recognise) has the status of customary international law’.⁸⁶ This approach is methodologically questionable. The Court makes no attempt at demonstrating the State practice or opinio juris in respect of the rule expressed in that treaty provision. Instead, it merely cross-references its earlier judgment, while recognising that that judgment contained no demonstration whatsoever as to the customary law character of the above-mentioned provision.

Perhaps a more worrying example can be found in the Certain Activities case, where the Court directly transposed the taxonomy of substantive and procedural obligations from a specific treaty regime as applied in its earlier case law, namely the 1975 Statute on the River Uruguay in the Pulp Mills case, to its analysis of the state of CIL in the context of transboundary environmental harm.⁸⁷ The Court went on to consider that substantive and procedural obligations with similar content apply as a matter of CIL to any non-industrial activities.⁸⁸ It did so without examining in any detail State practice or opinio juris.⁸⁹ The trend of generalising the scope of previous decisions on CIL is, of course, not limited to the ICJ’s practice. For instance, in its 2011 Advisory Opinion, the Seabed Disputes Chamber noted that both the International Tribunal for the Law of the Sea (ITLOS) and the ICJ had considered some of the specific provisions of the Articles on the Responsibility of States for Internationally Wrongful Acts (ARSIWA) as reflecting CIL and appeared to apply that conclusion to the ARSIWA more widely.⁹⁰

Third, the opposite side of this shortcut consists in simply not pronouncing on whether a given rule is part of CIL, in the absence of a pre-existing decision by international courts or tribunals on the subject. In principle, there is nothing improper in refraining from pronouncing on whether a given rule is or is not customary or the scope thereof, particularly when such determination is unnecessary for the court or tribunal to dispose of the issues before it.⁹¹ Indeed, it is in line with the common judicial avoidance techniques and may be justified in the interests of economy of means.⁹² However, the absence of an authoritative decision of the ICJ or any other court or tribunal as a reason not to pronounce on the customary law character of a given rule shows a complete misunderstanding insofar as the authority of any judicial or arbitral decision is concerned. A perfect illustration can be found in the EC – Hormones Report of the Appellate Body⁹³ or the subsequent EC – Biotech WTO Panel Report,⁹⁴ both avoiding pronouncing on the customary law status of the precautionary principle in the absence of an ‘authoritative decision by an international court or tribunal which recognises the precautionary principle as a principle of general or customary international law’.⁹⁵ Other convincing reasons may certainly explain the reluctance of the WTO Appellate Body and Panel to pronounce on the question whether the precautionary principle is part of CIL. However, it is certainly striking that the basis relied on, first and foremost, is the absence of an authoritative decision by an international court or tribunal recognising the principle as such.

3.3 ‘It Is CIL Because the ILC Has Said So’

Finally, one of the most common shortcuts is to refer to the ILC work as direct evidence of the state of CIL. As noted by the Commission itself, the ‘weight to be given to [its] determinations depends, however, on various factors, including the sources relied upon by the Commission, the stage reached in its work, and above all upon States’ reception of its output’.⁹⁶ Functionally, this shortcut is understandable to the extent that the ILC is tasked with codification and progressive development of international law. In the eyes of international courts and tribunals, the ILC’s work is rightly ‘most valuable, primarily due to the thoroughness of the procedures utilized by [it]’.⁹⁷ This is particularly the case where a given set of guidelines or articles produced by the ILC is rooted in a thorough survey of state practice and has garnered widespread support among states at the UN Sixth Legal Committee. In such circumstances, it is understandable that there are numerous examples where the ICJ, for instance, has ‘referred to provisions of the ILC’s codification work as customary with no or little further comment’.⁹⁸

However, at times, the reasoning of international courts and tribunals jumps too quickly to the conclusion that a given ILC end-product reflects the state of CIL. More fundamentally, the reasoning in respect of one particular provision tends to be almost automatically generalised to one or the other provision of the same end-product or, even worse, to the end-product of the ILC as a whole, without any inquiry as to whether that finds support in state practice and opinio juris. On many occasions, such generalising techniques may be harmless, but at times they may also forestall the development of the law, which might have purposefully been left in an open-ended texture to be refined by subsequent state practice and opinio juris.

For instance, the ICJ has often rubber-stamped statements of the ILC as representative of CIL. It has done so, even when such conclusions were only temporary or provisional, without the final product having been yet adopted. For example, the Court famously did so in the Gabčíkovo-Nagymaros Project with respect to the state of necessity, without much of a probing inquiry,⁹⁹ even if earlier tribunals had failed to recognise it as a circumstance precluding wrongfulness under CIL.¹⁰⁰ Subsequently, courts and tribunals have simply embraced with approval the ICJ’s finding as to the customary law character of what were to become Article 25 ARSIWA.¹⁰¹ In a similar vein of almost blindly approving the work-product of the ILC, the Court in the Application of the Convention of Genocide case had recourse by analogy to Article 16 ARSIWA, which it considered to reflect CIL and which informed the Court’s analysis whether Serbia breached Article III(e) of the Genocide Convention.¹⁰² In doing so, the Court applied and interpreted restrictively the requirements set out in Article 16 ARSIWA, even if there are serious doubts as to whether some of those requirements adequately reflect the existing state practice and opinio juris.¹⁰³

The use of the ILC’s work as a reflection of CIL has been part of judicial reasoning for years. It has indeed become a sort of ‘ping-pong’¹⁰⁴ or a ‘normative Ponzi scheme’,¹⁰⁵ whereby the Court heavily draws on and consolidates the work of the Commission, which in turn looks up to the ICJ’s case law for authoritative recognition of the rules or principles it seeks to codify. That is understandable on many levels, considering ‘a special vantage and authority’ that the ILC enjoys as a result of its close relationship with states.¹⁰⁶ Although some have argued that that special position makes ‘its pronouncements less tendentious, and more conservative, in character’,¹⁰⁷ the reality is that in many areas of international law the ILC is the only mechanism through which the views of states may be directly ascertained and made known in a systematic way. In this sense, the ILC seeks to ‘adopt a real-world approach and provide drafts that will hopefully prove useful and acceptable to the international community’.¹⁰⁸ However, more often than not, its drafts are a combination of elements of codification proper and progressive development of international law. This means that a thorough analysis may be required when an international court or tribunal later has to determine whether a given article, guideline, or rule, as presented by the ILC, reflects CIL.

As the examples above (and many more could be cited) demonstrate,¹⁰⁹ the relationship between international courts and tribunals and the ILC has grown increasingly symbiotic over time. One is thus left under the impression that judges do not scratch beneath the surface when making relatively bold pronouncements on rules that until their adoption by the ILC had an uncertain status in international law and were so regarded by states, including in their views as expressed at the UN Sixth Legal Committee. More fundamentally, the tendency to generalise conclusions as to the customary character in respect of one specific provision to other provisions, or to the entirety of projects under consideration, shows the over-reliance by courts and tribunals on the work of the ILC, without always adequately probing into the underlying evidence of state practice and opinio juris.

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’.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’.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’.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.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’.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.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.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.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’.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.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.

12 Eureka! On Courts’ Discretion in ‘Ascertaining’ Rules of Customary International Law

Letizia Lo Giacco

1 Introduction

A number of scholarly contributions on the theme have tackled the determination¹ of rules of customary international law (CIL) under the umbrella of the methodological dualism between induction and deduction.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.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.⁴ As such, ‘lawyers move behind the law and cannot pretend to lead it’.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 states⁶ or the principle of good faith.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.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.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 interpretation¹⁰ 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 cases¹¹ 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 CIL¹² intervenes in the debate about the determination of CIL rules by tackling the long-standing question of the ‘methodology’ that interpreters ‘must’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:

To decide whether these two factors [state practice and opinio juris] in the formative process of a customary law exist or not, is a delicate and difficult matter. The repetition, the number of examples of State practice, the duration of time required for the generation of customary law cannot be mathematically and uniformly decided. Each fact requires to be evaluated relatively according to the different occasions and circumstances.¹⁴

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’.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.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.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 practice¹⁸ and opinio juris; and the myth of a hypothetical ‘out there’ where to identify already existing rules of CIL.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.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.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.²² However, this position has largely displayed its limits,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.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.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.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 ILC²⁷ – 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 elements²⁸ in determining the existence and the content of rules of CIL.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’.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’.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’.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.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.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.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.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.

[F]or a rule to pass into customary international law, the International Court of Justice has authoritatively restated in the North Sea Continental Shelf cases that there must exist extensive and uniform state practice underpinned by opinio juris sive necessitatis. To the extent that the domestic decisions and national laws of States relating to the issue of duress as a defence to murder may be regarded as state practice, it is quite plain that this practice is not at all consistent.³⁷

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

the criminal codes and legislation of 14 civil law jurisdictions in which necessity or duress is prescribed as a general exculpatory principle applying to all crimes. … Indeed, the rejection of duress as a defence to the killing of innocent human beings in the Stalag Luft III and the Feurstein cases, both before British military tribunals, and in the Hölzer case before a Canadian military tribunal, reflects in essence the common law approach.³⁸

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’,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’.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,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’.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 case⁴³ in which the ICTY Trial Chamber acknowledged that opinio juris may play a primary evidentiary role at the expense of state practice.⁴⁴

The question nevertheless arises as to whether these provisions [Article 51(6) and Article 52(1) of the First Additional Protocol of 1977], assuming that they were not declaratory of customary international law, have subsequently been transformed into general rules of international law. … This is however an area where opinio iuris sive necessitatis may play a much greater role than usus, as a result of the aforementioned Martens Clause. In the light of the way States and courts have implemented it, this Clause clearly shows that principles of international humanitarian law may emerge through a customary process under the pressure of the demands of humanity or the dictates of public conscience, even where State practice is scant or inconsistent. The other element, in the form of opinio necessitatis, crystallising as a result of the imperatives of humanity or public conscience, may turn out to be the decisive element heralding the emergence of a general rule or principle of humanitarian law.⁴⁵

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.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’,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’.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’.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.⁵⁰ This finding was finally sealed by reference to relevant ICJ judicial decisions.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.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,⁵³ the ICJ was to determine ‘when the right to self-determination crystallised as a customary rule binding on all States’.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’.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,⁵⁶ but also to determine ‘the content and scope of such a right’,⁵⁷ namely to interpret such a right.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,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 1992.⁶⁰ 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.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]’.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’.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”)]’,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’.⁶⁵ Similarly, in the Tadić Appeal Judgment, the ICTY Appeals Chamber found case law to be reflective of CIL.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’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.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.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.⁷⁰ Importantly, such verbalisation stems from an evaluative process – entrenched in an exercise of discretion – channelled through the judges’ normative ideology⁷¹ 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.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’,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:

(2) A general practice and acceptance of that practice as law (opinio juris) are the two constituent elements of customary international law: together they are the essential conditions for the existence of a rule of customary international law. The identification of such a rule thus involves a careful examination of available evidence to establish their presence in any given case. This has been confirmed, inter alia, in the case law of the International Court of Justice, which refers to ‘two conditions [that] must be fulfilled [North Sea Continental Shelf, Judgment, I.C.J. Reports 1969, p. 3, at p. 44, para. 77] and has repeatedly laid down that ‘the existence of a rule of customary international law requires that there be “a settled practice” together with opinio juris’.[See, for example, Jurisdictional Immunities of the State (Germany v Italy: Greece intervening), Judgment, I.C.J. Reports 2012, p. 99, at pp. 122–123, para. 55; Continental Shelf (Libyan Arab Jamahiriya/Malta), Judgment, I.C.J. Reports 1985, p. 13, at pp. 29–30, para. 27; and North Sea Continental Shelf (see footnote above), at p. 44, para. 77]. To establish that a claim concerning the existence or the content of a rule of customary international law is well-founded thus entails a search for a practice that has gained such acceptance among States that it may be considered to be the expression of a legal right or obligation (namely, that it is required, permitted or prohibited as a matter of law). The test must always be: is there a general practice that is accepted as law?⁷⁴

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 interpretation⁷⁵ and may not be regarded as performing a merely declaratory function. Although this outlook bears the marks of legal realism,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’.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,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’.

13 Identification of and Resort to Customary International Law by the WTO Appellate Body

Mariana Clara de Andrade ^*

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.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)’.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,³ the practice of international tribunals does not always follow such methodology.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,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).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.

2.1 Customary Rules on Treaty Interpretation

Article 3.2 of the WTO Dispute Settlement Understanding (DSU) determines that the DSM ‘serves to preserve the rights and obligations of Members under the covered agreements, and to clarify the existing provisions of those agreements in accordance with customary rules of interpretation of public international law’. Rules on treaty interpretation, and even more specifically Articles 31, 32 and 33 of the Vienna Convention on the Law of Treaties (VCLT) are the only norms of general international law on the law of treaties that the AB has qualified as customary rules. This is so even though other canons of treaty interpretation can be found in the reports.⁷

A possible explanation for this approach is that, in fact, the DSU Article 3.2 reference to ‘customary rules of interpretation’ was originally intended to refer to the VCLT provisions codifying these customary interpretative guidelines. However, because not all members of the GATT/WTO were parties to the VCLT, the drafters chose to refer to ‘customary rules of interpretation of public international law’ instead.⁸ This shows that from the outset the intention was to resort to the VCLT rules in the interpretation of the agreements. Therefore, the early references by the AB to these rules as those reflected under Article 3.2 of the DSU were but a formality.

Indeed, the rules on treaty interpretation of the VCLT were invoked on the first WTO controversy to reach the appeals stage, the US – Gasoline dispute.⁹ In this report, the AB held that Article 31 of the VCLT had ‘attained the status of customary or general international law’.¹⁰ To ground that statement, the AB inserted a footnote with reference to decisions of the ICJ, European Court of Human Rights and the Inter-American Court of Human Rights, in addition to a few handbooks of international law.¹¹ It should be recalled that the practice of international courts and tribunals is not the main method for the determination of state practice and opinio juris. In fact, as stated by ILC Conclusion 12 on the identification of CIL, ‘Decisions of international courts and tribunals, in particular of the International Court of Justice, concerning the existence and content of rules of customary international law are a subsidiary means for the determination of such rule.’¹²

While a similar method was employed for Article 32,¹³ the AB did not follow the same approach in relation to Article 33.¹⁴ Instead, the adjudicators merely stated the latter was customary law, perhaps because it was already a tautological statement after granting this recognition to Articles 31 and 32.¹⁵

In subsequent reports, the AB does not seem to have considered it necessary to re-examine the customary value of Articles 31–33 of the VCLT. The adjudicators, including panellists, simply refer authoritatively to these rules, sometimes referencing also DSU Article 3.2. For example, in US – Carbon Steel, the AB noted that ‘It is well settled in WTO case law that the principles codified in Articles 31 and 32 of the Vienna Convention on the Law of Treaties (the “Vienna Convention”) are such customary rules [mentioned in Article 3.2].’¹⁶ In US – Softwood Lumber IV, the adjudicators noted that ‘As we have observed previously, in accordance with the customary rule of treaty interpretation reflected in Article 33(3) [of the VCLT], the terms of a treaty authenticated in more than one language – like the WTO Agreement – are presumed to have the same meaning in each authentic text.’¹⁷ Therefore, once the customary status has been determined within WTO case law, adjudicators consider it sufficient to refer back to previous adopted reports to make the same claim with respect to the status of a VCLT provision.

The AB’s reference to VCLT Article 26 in Brazil – Desiccated Coconut corroborates the conclusion that the organ has refrained from declaring other rules on the law of treaties as reflecting CIL. In that dispute, the panel had invoked the principle of non-retroactivity as reflected in Article 28 VCLT as ‘an accepted principle of customary international law’.¹⁸ This terminology was not followed by the AB.

The AB case law also features references to non-VCLT rules and principles on treaty interpretation which are derived from the provisions in that convention (in particular Article 31). The references to the interpretative principles of effectiveness, systemic integration and in dubio mitius are all, in one way or another, connected to VCLT Articles 31 and 32. In other words, this is the legal basis employed by the AB to invoke these interpretative canons. Nonetheless, the AB did not explicitly consider these principles to reflect CIL. For example, in Japan – Alcoholic Beverages II, the AB explicitly invoked the ‘principle of effectiveness or ut res magis valeat quam pereat’, and stated that it was a ‘fundamental tenet of treaty interpretation flowing from the general rule of interpretation set out in Article 31’.¹⁹ In EC and Certain Member States – Large Civil Aircraft, the AB invoked VCLT Article 31(3)(c) as ‘an expression of the “principle of systemic integration”’.²⁰ Finally, the AB has also referred to the principle of in dubio mitius, on a footnote in the EC – Hormones²¹ and in the China – Publications and Audiovisual Products reports.²² In EC – Hormones, the AB cited the ‘interpretative principle of in dubio mitius’, widely recognised in international law as a ‘supplementary means of interpretation under VCLT Article 32’, and added a quote from an international law handbook with a definition of the principle, in addition to reference to relevant case law (including ICJ, PCIJ and arbitral decisions) as well as other doctrinal works.²³ As opposed to the principles of effectiveness and systemic integration, in dubio mitius is not expressly codified in the VCLT. Still, the legal basis indicated by the AB in the EC – Hormones report was not entirely detached from the convention.

What is important to remark is that none of these references is considered to reflect CIL, even though they are based on VCLT Articles 31 and 32.

2.2 Customary Rules on State Responsibility

References to general rules on state responsibility appear more frequently in WTO case law since the adoption of the ILC’s Articles on Responsibility of States for Internationally Wrongful Acts (ARSIWA) in 2001. Before 2001, the AB mainly resorted to scholarship on the topic.²⁴ After 2001, the codification of rules on state responsibility by the ILC left little need for the AB to engage in other methods of identification. Since then, the methodology of reference to these sources of international law in AB reports consists mainly in citing the works of the ILC on state responsibility.

The AB adopted a cautious approach in determining that a rule of state responsibility reflects CIL. The only concept of state responsibility the AB has considered to have attained customary status is the principle of proportionality in the context of countermeasures for wrongful acts.²⁵ In the US – Line Pipe report, adopted in 2002, the AB stated that the ARSIWA was ‘not a binding instrument as such’, but its Article 51 nevertheless ‘sets out a recognized principle of customary international law’.²⁶

To support this statement, the AB added a footnote referencing the Nicaragua and the Gabčíkovo-Nagymaros decisions of the ICJ.²⁷ Additionally, the AB stated that ‘also the United States has acknowledged this principle elsewhere’,²⁸ referencing remarks the United States had made in the commentaries to the works of the ILC in 1997 and its position in proceedings before an arbitral tribunal. The AB, however, did not indicate that this reference to the US’s position was reflective of opinio juris. Moreover, the AB did not refer to the recognition of the customary status of the principle of proportionality by other states, and the conclusion that Article 51 ‘sets out a recognized principle of customary international law’ is not further explained in the report. Indeed, the ILC commentaries to Article 51 state that ‘[p]roportionality is a well-established requirement for taking countermeasures, being widely recognized in State practice, doctrine and jurisprudence’.²⁹ The AB could have referred to these commentaries as an authoritative source for advancing the customary status of the proportionality principle.

In sum, the method employed by the AB in this case, both in quoting the ICJ case law and the US’s position, was to seek purposive legitimation to the conclusion that proportionality reflects CIL. The indication of the US’s (the interested party in that dispute in the quality of defendant) position can thus be viewed as a way of ascertaining opinio juris to reinforce the reference to proportionality as a customary rule.

References to general rules on attribution and to Article 14 of the ARSIWA are two instances of AB practice that corroborate the organ’s reluctance to declare the customary status of general rules on state responsibility. In the case of attribution, in US – Definitive Anti-Dumping and Countervailing Duties on Certain Products from China (US – AD and CVD (China)), the AB deliberately avoided taking a position on the status of Article 5 of the ARSIWA as a customary rule. The ARSIWA provision had been invoked by China, the complainant, as a tool for interpreting the term ‘public body’ in Article 1.1(a)(1)(iv) of the Agreement on Subsidies and Countervailing Measures (ASCM) as a ‘relevant rule’ under VCLT Article 31(3)(c).³⁰

In that dispute, the AB submitted that ARSIWA Article 5 was not binding per se but, insofar as it reflected CIL or general principles, it could be taken into consideration as ‘applicable in the relations between the parties’ in the terms of Article 31(3)(c).³¹ Thus, to assess whether the provisions were ‘rules of international law’, the AB would have to consider whether they constituted customary law or general principles. Instead, the adjudicators circumvented the question by concluding that, in fact, their interpretation (based on the general rule in VCLT Article 31(1)) of ‘public body’ in Article 1.1(a)(1)(iv) of the ASCM ‘coincide[d] with the essence of Article 5 [of the ILC ARSIWA]’.³² In other words, because the content of the general rule coincided with their interpretation of the WTO provision under scrutiny, it was not necessary to ascertain the customary status of Draft Article 5.³³

Article 14 of the ARSIWA, on the ‘extension in time of the breach of an obligation’, was invoked by the European Communities in EC and Certain Member States – Large Civil Aircraft. Similarly to the US – AD and CVD (China) dispute, there was disagreement between the disputants regarding the status of the rule as customary.³⁴ In its reasoning and findings, the AB bypassed the discussions on whether Article 14 of the ARSIWA reflected CIL, if it could be considered a customary rule for purposes of interpretation, and whether there is a legal basis for the invocation of this rule in the WTO legal system. Instead, the adjudicators went on to analyse whether ARSIWA Article 14(1) and Article 5 of the ASCM had the same scope, similar to its approach in US – AD and CVD (China).³⁵ The AB did dismiss the EC’s argument, but not based on the allegations that it did not reflect CIL, but because its substance was not relevant to provide support to the interpretation advanced by the European Union.

2.3 The AB’s Trends in Ascertaining CIL

Three remarks can be made regarding the methodology employed by the AB when ascertaining the existence of CIL. First, as advanced earlier, there are only two fields of customary law whose existence the organ has explicitly exploited: rules on treaty interpretation and rules on state responsibility. This may be explained because they are not norms of substantive, primary (understood as ‘rules that place obligations on States, the violation of which may generate responsibility’)³⁶ nature: they are structural rules which arguably are necessary for the functioning of any legal system.³⁷ Moreover, another factor that could explain this choice is that engaging in the two-element methodology for identifying customary rules of primary nature and applying those rules in the WTO DSM would fall outside the jurisdictional scope of WTO adjudication.

The second remark is that the AB seems hesitant to determine the customary status of a rule. It has actively refrained from doing so in at least two cases (general rules on attribution and the relationship between the duration of a conduct and its effects). As described above, in the US – AD and CVD (China) and EC and Certain Member States – Large Civil Aircraft disputes, there was express disagreement between the parties with respect to the customary status of certain ARSIWA canons, and yet the AB avoided making a finding.

Additionally, the AB followed the same dismissive approach with the precautionary principle in the EC – Hormones report. In that dispute, the status of the precautionary principle was challenged by the parties: the EC argued that the precautionary concept reflected a ‘general customary rule of international law’,³⁸ while the United States claimed it represented neither a customary rule nor principle, but merely an ‘approach’.³⁹ The AB considered that the question was controversial under international law, and that ‘it [was] unnecessary, and probably imprudent, for the Appellate Body in this appeal to take a position on this important, but abstract, question’.⁴⁰ The AB concluded only that the precautionary principle ‘finds reflection’ in Article 5.7 of the Agreement on the Application of Sanitary and Phytosanitary Measures (SPS).⁴¹ In any case, it is relevant to remark that AB’s hesitancy to declare the customary status of the precautionary principle is not unique in international adjudication.⁴²

Relatedly, the AB’s practice also demonstrates that this organ was hesitant in declaring the customary status of a given principle in particular when there was a controversy between the parties. The customary status of the treaty interpretation rules in the VCLT is virtually undisputed, and the concept of proportionality is also enshrined in legal logic. Moreover, in these cases, there was no explicit disagreement regarding the customary value of these sources of law. In the other cases, there was explicit disagreement, and the AB refrained from taking a position.

The third remark is that the AB does not make a full-blown assessment of opinio juris and state practice to ascertain the customary status of a rule. Put differently, the AB does not properly ‘identify’ the existence of a customary rule, it ‘asserts’ such existence.⁴³ In doing so, it relies on codified instruments of international law. In the case of rules of treaty interpretation, it relied on scholarship and decisions from other international law tribunals to state the customary status of Articles 31 and 32 of the VCLT. In the case of rules on state responsibility, it simply stated Article 51 of the ARSIWA reflected a ‘principle of customary international law’, quoted previous ICJ decisions and referred to the acknowledgement by the United States of this customary status. This methodology hints that the organ was mostly preoccupied with ensuring the acceptance of the legal reasoning by the affected party, rather than determining the customary status of a fundamental principle as a matter of law with legal implications for the WTO legal system. In fact, proceeding with a full-blown query of state practice, opinio juris or comparative study of national legal systems for the determination of a customary rule or a general principle seems not only unnecessary but also a potential source of controversy in the context of WTO dispute settlement.⁴⁴

The limits of resorting to CIL in WTO dispute settlement are not evident. In US – Cotton Yarn, the AB held that the concept of proportionality was a ‘customary principle’ of state responsibility that had not been derogated by WTO law. The adjudicators considered that an ‘exorbitant derogation from the principle of proportionality … could be justified only if the drafters of the [WTO Agreement] had expressly provided for it, which is not the case’.⁴⁵ Perhaps the determination of the customary status of a rule entails the possibility of fall-back to general international law in the case of a gap in the tool-box of secondary norms of the WTO legal system. This is relevant for the purposes of the distinction between using a non-WTO source of law for interpretative purposes and applying such norm. The next section will address the practical implications of this distinction in light of WTO case law.

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.⁴⁶ It seems well-settled that WTO dispute settlement only has jurisdiction over the so-called WTO covered agreements,⁴⁷ while its applicable law can range further than that.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.

3.1 The Jurisdiction of WTO Dispute Settlement and the Notions of Interpretation and Application of International Law

Disputes which call into question the use of sources of law outside the jurisdictional limits of a court are controversial because ‘[t]hey do not fall plainly within the scope of the jurisdictional clause, nor clearly outside it; they straddle the dividing line’.⁴⁹ In the context of the WTO, there is no equivalent to Article 38 of the ICJ, enlisting the sources of law that can be invoked. However, Article 3.2 of the DSU determines that the DSM: ‘serves to preserve the rights and obligations of Members under the covered agreements, and to clarify the existing provisions of those agreements in accordance with customary rules of interpretation of public international law. Recommendations and rulings of the DSB cannot add to or diminish the rights and obligations provided in the covered agreements.’ Article 3.2 of the DSU sets forth an express limitation to WTO adjudication. Applying non-WTO sources of law in the multilateral trading system arguably amounts to ‘adding to or diminishing the rights and obligations’ of the covered agreements, while using extraneous norms for purposes of interpretation of WTO provisions can serve for the ‘clarification of the existing provisions’.⁵⁰ Therefore, the distinction between interpretation and application of norms can serve as a valuable tool to understand the limits of WTO jurisdiction with respect to CIL.⁵¹

Distinguishing ‘interpretation’ and ‘application’ of norms entails the debate of whether these are two different processes, two processes which are interconnected, or if in fact one cannot draw such a distinction.⁵² Gardiner sustains that they reflect a ‘natural sequence that is inherent to the process of reading a treaty: first ascribing meaning to its terms and then applying the outcome to a particular situation’.⁵³ The distinction was discussed during the works of the ILC on the codification of the law of the treaties, in particular in connection to the question of intertemporal law,⁵⁴ but the topic was revealed to be highly controversial⁵⁵ and any attempt to make a clear-cut distinction was dismissed.⁵⁶ The matter was addressed in the document preceding the works of the commission – the Harvard Draft Convention on the Law of Treaties.⁵⁷ The following distinction is found in the commentaries to the Harvard Draft Convention:

Interpretation is closely connected with the carrying out of treaties, for before a treaty can be applied in a given set of circumstances it must be determined whether or not it was meant to apply in those circumstances. … There is, however, a recognized distinction between the two processes. Interpretation is the process of determining the meaning of a text; application is the process of determining the consequences which, according to the text, should follow in a given situation.⁵⁸

The question addressed here is of a slightly different nature. The use of this distinction is intended to clarify the limits of the use of an international source of law for the interpretation of another source of law.⁵⁹ This being so, the inquiry seeks to clarify the line between using CIL as an interpretative tool for the application of another norm, and when this interpretative recourse transfigures into the actual application of that customary rule, which in principle should have only a subsidiary character.⁶⁰ In particular, the question here addressed is not related to the interpretation of CIL as such.⁶¹ The above definition can be thus adapted in the following manner: ‘The use of a norm for interpretative purposes is the process of resorting to an auxiliary source with the aim of determining the meaning of an original norm; application of a norm is the process of determining the consequences which, according to its content, should follow in a given situation.’ Interpretation is a cognitive process, while application is a practical one. This does not mean the two phases cannot overlap. Overlap may happen when different sources of law are used to interpret an obligation under dispute by an international adjudicator. The end conclusion will thus reflect an intersection between use of sources different than the one originally being ‘interpreted’ (i.e., a WTO covered agreement term, provision or obligation) – and its final application may be an indirect application of these other sources. Thus, it is of interest whether this final application of the WTO provision entails the incidental application of a non-WTO rule. The practice of the AB may shed some light in the position taken by WTO adjudicators in this sense.

3.2 The AB’s Resort to CIL: Interpretation or Application?

As described in Section 2, the AB declared the customary status of rules deriving from outside of the WTO system only in a limited number of instances, and it has refrained from taking a position regarding this status in other instances. The rules the AB considered to reflect customary law are those related to state responsibility and treaty interpretation. Two sets of ‘boundaries’ can be inferred from the AB’s practice described in Section 2. These boundaries seem to ensure that CIL is used solely for interpretative purposes, in detriment of their ‘application’.

The first boundary relates to the content of the norm. The AB only declares as customary rules those that are ‘structural’: meta-norms,⁶² such as those related to treaty interpretation, and rules of state responsibility. In particular, rules on treaty interpretation are operational: they lack substantive implications and they relate to the cognitive process of interpreting a norm. Perhaps more crucial is the fact that customary rules on treaty interpretation have been expressly incorporated by the WTO legal system. Although they can be applied within the scope of their operational function, they cannot be applied as to add to or diminish the substantive obligations provided for in the WTO Agreements.⁶³ Because these rules are ‘structural’ (their role is more procedural or instrumental), they lack substantive content (i.e., they are not ‘primary’ rules as their content does not prescribe obligations per se)⁶⁴ and thus are less likely to ‘add to or diminish rights and obligations of WTO members’.⁶⁵

One reason for this approach is that adjudicators may feel that recognising the customary status of a rule with substantive content may give the impression that they are creating substantive obligations or even overriding WTO law. The general reluctance of the AB to refer to non-WTO rules as customary, even when there is ground for doing so, can be regarded as a cautious approach in not overemphasising the role of these sources in the WTO legal system. This possibly explains why the AB granted this status to rules on treaty interpretation and the proportionality principle – as they are operative concepts, and not concepts entailing autonomous substantive obligations. One can infer that this gives more leeway for the AB not to be accused of overstepping its jurisdictional mandate.

The dispute on whether the precautionary principle reflected a customary rule further illustrates this possibility. In EC – Hormones, the AB refrained from answering this question, and held that it was ‘unnecessary, and probably imprudent, for the Appellate Body in this appeal to take a position on this important, but abstract, question’.⁶⁶ It concluded that the precautionary principle found reflection in WTO law, and could not, by itself, override the provisions of the SPS Agreement. Put differently, the AB stated that the principle is incorporated by the SPS, and plays an ‘internal’ role in the WTO system. Conversely, the adjudicators indicated that an ‘extraneous’ (i.e., not codified in WTO Agreements) reflection of the precautionary principle has very limited, if any, role in WTO law.

The emphasis that the AB put in the statement that the precautionary principle is part of WTO law can be read as a sign that the adjudicators acknowledged the importance of the concept, but were cautious so as not to overstate – or give the impression that they overstate – its authoritativeness in the WTO legal system. It may be useful to consider that the particularity of the precautionary principle with respect to treaty rules and secondary rules of international law is that the first denotes, at least to some extent, a substantive dimension: even if not consisting of a clear rule of conduct, it nevertheless can be a source of obligations to guiding the conduct of states.⁶⁷

The second, related, ‘boundary’ that limits the role of customary law in WTO adjudication concerns regarding the way in which the AB employed concepts which it considered having attained this status. The AB’s practice has ensured that its reliance on CIL remained subordinated to the prevalence of WTO legal texts.⁶⁸ As delineated in Section 3.1, ‘application’ can be understood as ‘the process of determining the consequences which … should follow in a given situation’.⁶⁹ Accordingly, one can consider that a WTO adjudicator is applying a non-WTO rule to the extent that the findings of violation or non-violation contained in the report ensues from non-WTO language.

In US – Cotton Yarn, the AB had to ascertain the meaning of ‘serious damage’ under Article 6.4 of the Agreement on Textiles in order to determine whether the United States could attribute damage caused by the importation of a certain category of products to one member only and imposing safeguards measures only against that particular country, disregarding proportionality.⁷⁰ In US – Line Pipe, the AB had to ascertain the meaning of ‘serious injury’ that justifies the application of a safeguard measure under Article 5.1 of the Agreement on Safeguards. In both cases, the focus of the interpretation was on the wording of the WTO provisions under scrutiny, and the AB used the principle of proportionality to shed light on and give meaning to specific treaty provisions.

Moreover, neither the principle of proportionality nor the rules on treaty interpretation have an autonomous content or entail legal consequences per se. Rather, by definition, they are operational to interpret other rules of international law. In this sense, both the principle of proportionality and the rules on treaty interpretation must be employed in conjunction with other rules of primary content. It can be concluded that the AB has resorted to (what it declared to reflect) customary law insofar as these concepts were subordinate to the interpretation of WTO provisions.

The principle of good faith could arguably also be considered as having attained the status of a customary rule. However, the AB has declared only VCLT Article 31(1) as a whole to reflect customary law, rather than the concept of good faith as an ‘independent’ principle. It can be speculated that the AB has only referred to VCLT Articles 31–33 as customary because these provisions are implied in the text of DSU Article 3.2 as the ‘customary rules of interpretation of public international law’. Moreover, if good faith were to be declared a customary rule, and not ‘just’ a principle of treaty interpretation and treaty performance, it could be understood that there are textual grounds to bring claims based on violations of good faith.⁷¹ This could raise criticisms from the membership and amount to accusations that the AB is ‘adding to or diminishing rights and obligations of Members’.

Extraneous principles and rules that do not create rights and obligations for WTO members provide safer grounds for the AB not to overstep its jurisdictional mandate. By determining as customary rules norms related only to state responsibility and treaty interpretation (and in the case of the former, even a limited set thereof), the AB ensures that they will remain subordinate to WTO obligations. This approach allows adjudicators to resort to these sources of law for the interpretation of WTO provisions, rather than to create doubts as to whether they are being applied, thereby adding to or diminishing rights and obligations contained in the covered agreements.

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.

14 The Practice of Non-state Armed Groups and the Formation of Customary International Humanitarian Law Towards Direct Relevance?

Zhuo Liang

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.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’.² The use of words ‘general practice’ has led some to argue that international custom is not restricted to the practice of states only,³ as practice may emanate also from non-state actors.⁴ While such an interpretation arguably runs against the drafters’ intention,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.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]’,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’.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’.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.¹⁰ Be that as it may, there is little other support for such a stance.¹¹ It is also noteworthy that post-Tadić ICTY jurisprudence tended to revert to the elements of state practice and opinio juris.¹²

It may be concluded that under lex lata, only states and international organisations have actually been entrusted with a law-making power.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.

3.1 The Legitimacy Problem of Customary Rules Binding NSAGs

There is little dispute today that NSAGs are bound by IHL.¹⁴ It is also generally accepted that NSAGs are bound by customary IHL. As the Special Court for Sierra Leone asseverated, ‘[insurgents] are bound as a matter of international customary law to observe the obligations declared by Common Article 3.’¹⁵ Alleging that NSAGs are bound by CIL, in addition to treaties and their unilateral commitments, substantially expands the ambit of their obligations under IHL¹⁶ and is thus vital for the protection of war victims.

In the eyes of some, the legitimacy of customary rules binding NSAGs is flawed, inasmuch as their formation does not take into account the practice of those actors. As argued, one of the legitimising premises of CIL is that it originates in the actions and beliefs of those whom it later comes to bind.¹⁷ Since CIL binds not only states, clinging to a state-centric notion merely evinces a perceived loss of its democratic legitimacy.¹⁸ In the process of CIL’s formation, the unfair lack of the participation of the entities that the law intends to regulate would severely compromise its legitimacy.¹⁹ The need to cure this legitimacy deficit hence justifies incorporating the practice of non-state actors, such as NSAGs, as an element of CIL.²⁰

Indeed, depicting CIL as such a behaviourally self-generated norm-forming process represents the dominant approach, according to which no conceptual obstacle, at first blush, seems to arise for non-state actors to create customary rules for themselves.²¹ A closer inspection, however, unveils a major deficiency of this logic chain that is followed by a dramatic corollary: the legitimacy of CIL is fully restored only if all participants of the international legal order, ranging from states to international organisations to individuals, are conferred the capacity to create customary rules for themselves.²² This theoretical prospect, explicitly noted by a few authors,²³ confuses states which play a cardinal role in international law with others which do not, based on a taken-for-granted equivalence between them. By doing so, it improperly overlooks the built-in but defensible inequality between states and non-state actors before international law.

Such a presupposed equivalence is a fiction. According to Hugh Thirlway, a unique characteristic of states, which bars any non-state actor being promoted to full state rank, is that they best represent the interests and needs of human beings in international society.²⁴ It is possible that the idea of an entity creating self-governing law fits only the club of sovereign states. Traditional voluntarist approach to international law has confirmed that by virtue of sovereign equality of states, no legal obligations can be imposed on any one of them without its consent.²⁵ A legitimacy criticism is hence devised for addressing the lack of involvement of some states in the formation of CIL.²⁶ Non-state armed groups and other non-state actors, however, are not sovereign entities, and the principle of voluntarism never precludes the imposition of obligations on them in the absence of their consent.²⁷ Thus, there is no inherent reason why they cannot be subjected to the will of states in international law.²⁸ If the creation of CIL should be monopolised by states, even recognising the self-generating nature of this body of law cannot automatically bring out a legitimate appeal for non-state participation. In effect, the theory of legitimacy accommodates the possibility ‘for A to have legitimate authority over B even if A’s rule is neither consented to nor democratic’.²⁹ This modality of legitimacy especially caters to international law which ‘does not now enjoy, and is unlikely to achieve in the foreseeable future, a significant grounding either in the consent of its subjects or in democratic law-making processes’.³⁰

The ILC’s approaches to the role of international organisations in the formation of CIL reveal the intransigence of the state-centred CIL system. In its Draft Conclusions on Identification of CIL, the ILC proffered that ‘[i]n certain cases, the practice of international organizations also contributes to the formation of [CIL].’³¹ The ILC foresaw two clear circumstances where such practice arises as an element of CIL: states have transferred exclusive competences to international organisations, or have conferred upon them competences that are functionally equivalent to powers exercised by states.³² Rossana Deplano argued that the ILC’s seemingly ground-breaking conclusion is stealthily anchored in an unspoken premise: international organisations are empowered by states, and their practice can contribute to the formation of CIL only insofar as it is conceived as a surrogate of state practice.³³ To say the least, ILC’s work brings to light that contemporary international law offers little ground for elaborating on the contribution to CIL by international organisations as independent actors.³⁴ In other words, states seem never to have lost, or will lose, monopoly over the creation of CIL by countenancing limited participation of international organisations in this process.

Even if it may be contended that international organisations’ capacity to contribute to CIL accords with CIL’s character as a set of rules arising from the practice and usage of a distinctive community,³⁵ their case is not comparable to that of NSAGs. Kristina Daugirdas articulated three reasons justifying why international organisations can directly contribute to CIL: first, the states establishing an international organisation may subjectively intend for that organisation to be able to do so; second, this capacity may be an implied power of the organisation; and third, this capacity may be a byproduct of other features of the organisation, such as international legal personality, the capacity to enter into treaties, incurring responsibility for violations and making claims.³⁶ None of these is neatly applicable, by analogy, to NSAGs. First and foremost, NSAGs are not created or empowered by states. On the contrary, their presence in the territory is essentially illegal under national laws. Second, NSAGs do not have ‘implied powers’ – a principle that has no concern with them. Third, NSAGs cannot participate in treaties, and there is no agreement as to whether they can assume direct responsibility under international law.³⁷ Among all these considerations, the lack of states’ empowerment is most fatal. The absence of such a process manifesting states’ privilege helps explain why non-state actors other than international organisations are not considered creators of practice that contributes to the formation of CIL.³⁸

In sum, although it may be said that the legitimacy of CIL comes from its self-generating character, it is doubtful whether this formula is applicable beyond states and state-empowered entities. At present, ‘states always retain the final word’ to decide whether to bestow upon non-state actors a law-making power.³⁹ This is a status quo, and is not bound to suffer from legitimacy flaw if the legal asymmetry between states and non-state actors is not convincingly repudiated.

3.2 The Sense of Ownership and the Compliance Dynamics

Moving one step further from the legitimacy criticism, scholars arguing in favour of incorporating the practice of NSAGs put forward an argument concerning compliance. Marco Sassòli questioned how NSAGs could be expected to abide by IHL if they are not involved in the law-making process.⁴⁰ Indeed, sometimes NSAGs denied the binding force of IHL norms on them by arguing that they did (and could) not participate in the creation of those norms which is monopolised by states.⁴¹ This was the case with the Fuerzas Armadas Revolucionarias de Colombia, the Frente Farabundo Martí para la Liberación Nacional of El Salvador, and the National Liberation Front of Vietnam.⁴² Therefore scholars argued that to overcome this quandary, it is necessary to adopt an ownership approach, whose core message is that engaging NSAGs in the creation of norms vests them with a ‘sense of ownership’ which would strengthen their incentive for compliance with obligations.⁴³ On the part of CIL, such engagement efforts mean incorporating their practice into the formative process.⁴⁴

This line of reasoning postulates that an actor’s compliance can be improved if it feels a sense of ownership of the rules. As Hyeran Jo expounded, ‘[a] sense of ownership increases the likelihood that rules will take root within a rebel movement and be perceived as meaningful and worthwhile. When rules are internalized in this way, compliant behavior may eventually become a matter of habit … When this happens, self-implementation and self-policing replace outside supervision.’⁴⁵ Notwithstanding the plausibility of such an explanation as it appears, it is doubtful to what extent this ownership approach would work, in the context of custom-making, for NSAGs whose compliance poses a perennial threat to IHL.

The first uncertainty is whether creating a sense of ownership is serviceable enough for fostering compliance. Although the same question can be asked vis-à-vis states, that the compliance mechanisms of NSAGs are more fragile may render the ownership approach particularly feeble. In fact, the ownership consideration is far from the whole picture of the compliance dynamics. Besides the lack of ownership of norms, NSAGs may refuse to observe or implement IHL on various grounds, including strategic military concerns, the likelihood of prosecution, the lack of knowledge of applicable norms and political or religious ideology.⁴⁶ More importantly, the sense of ownership arguably stands among the least influential factors affecting NSAGs’ behaviour. While it is logically sound for Western scholars that the reluctance to accept IHL norms is due to non-participation in their creation, such a stance is not really expressed by most of NSAGs today.⁴⁷ Instead, the ownership argumentation appears only to be retained by some Colombian NSAGs.⁴⁸ It is also reported that Geneva Call, in its experience of engaging NSAGs, has not confronted any NSAG citing its exclusion from norm formation as a ground for rejecting the application of humanitarian standards; even many NSAGs which have agreed to adhere to IHL do not raise objections concerning non-participation.⁴⁹ Given such a nebulous nexus between the ownership approach and NSAGs’ compliance, it might be too exaggerated to suggest that the compliance record of NSAGs would be palpably altered once a sense of ownership is in place, in light of the sundry temptations to ignore the law.

Even if it is acceptable that more engagement and a sense of ownership are better than nothing for cultivating NSAGs’ willingness for compliance, another issue that requires careful unpacking is whether such a sentiment can really be aroused in the context of custom-making. Unlike treaty negotiations or issuance of unilateral commitments in which NSAGs may have their voice and (un)acceptance of norms plainly heard, the codification process of customs would foreseeably remain in the hands of expert groups conducting research from a third-party’s standpoint. Being remote from that pursuit, NSAGs would stand little chance to express their concerns or mount effective challenges when they have issues. It is thus questionable whether a mere promise to take into account their practice in the codification of CIL, without providing occasions for them to speak their mind, would equip NSAGs with a sense of ownership and induces their adherence to the resultant norms. In fact, criticisms have already been enunciated against the ICRC’s study of customary IHL for not being sufficiently reflective of what states truly think and do pursuant to the law.⁵⁰ If these criticisms are apposite, to what extent can a sense of ownership be conveyed from the codification of CIL?

It may be argued that to ensure NSAGs’ opinions and concerns being genuinely heard and addressed, unprecedented participatory mechanisms are needed. However, there are numerous NSAGs operating in widely scattered areas of the world. Is it physically feasible to come into contact with all, or at least a substantial part of them? If not, should certain NSAGs be selected as ‘representatives’ getting involved in the codification process? How to then assure that the competence of those who are invited to participate would be recognised by other NSAGs? Moreover, even if participatory mechanisms for NSAGs are adequately devised, they cannot fix the participation and representativeness plights in the long run. Unlike states, NSAGs are often ephemeral. After the close of hostilities, an NSAG may be disbanded or form the legitimate government of a (new) state. Hence, ‘[t]here remains a risk that even if the armed groups of today contribute to the formation of international law, the armed groups of tomorrow will still not feel any ownership of these norms and will use that as an excuse not to comply with them.’⁵¹

Hitherto, even though an ownership approach should be adopted for the sake of compliance, constructing the direct relevance between the practice of NSAGs and customary IHL would sit among the most specious propositions for approaching this goal.⁵² In fact, there are no statements on the part of NSAGs suggesting that they are aware of the possibility that their acts could contribute to CIL, or that they consent to it.⁵³ It might eventually become one’s own wishful thinking to purport that the compliance predicament could be practically remedied through admitting NSAGs into the formative process of CIL.

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.

4.1 Scope Ratione Personae

Non-state armed groups are extremely diverse. They differ in the extent of territorial control, internal structure, capacity to train members, and the disciplinary or punitive measures that are taken against members.⁵⁴ It seems intricate to establish detailed criteria for including some of them while excluding the rest. Pondering over the appropriateness of engaging selected NSAGs only, Marco Sassòli submitted that ‘the international community should try to apply all the legal mechanisms suggested to all armed groups.’⁵⁵ This view holds up inasmuch as it seems to be the only likely solution in harmony with the legitimacy concern as displayed in Section 3.1, which implies that all non-state actors shall be treated as law-creators because excluding any of them would defeat the argumentation itself.⁵⁶ Now that it is insisted that CIL can be legitimately binding only for those whose practice and beliefs are constitutive to its formation, and that it shall bind all NSAGs which are parties to armed conflicts, it would be groundless to disqualify any of them for the purpose of custom-making.

Therefore, since customary IHL applies only in the situations of armed conflict, an NSAG shall be taken into account for the formation of this body of law so long as it can engage in a NIAC.⁵⁷ Criteria of a NIAC include ‘the intensity of the conflict and the organization of the parties to the conflict’.⁵⁸ Accordingly, to be a party to a NIAC and then a prospective candidate for custom-making, an NSAG is required to have ‘a sufficient degree of organization’ and ‘be able to and does conduct, or is otherwise involved, in an armed campaign which reaches the required degree of intensity’.⁵⁹

This approach seems counter-intuitive, as even Al-Qaeda and the Islamic State of Iraq and Syria would attain the law-creator status. It has to be emphasised that an all-inclusive approach for NSAGs per se is not tantamount to indiscriminate incorporation of any sort of their practice. As will be illustrated in Section 4.2 below, the practice should be subject to scrutiny before being utilised as the basis of customary IHL.

4.2 Scope Ratione Materiae

The practice of NSAGs consists of various types and forms. They may include, but are not limited to: first, verbal acts, such as codes of conduct, internal legislations, unilateral commitments, instructions to armed members, special agreements, peace treaties and statements in international fora; and second, physical acts, such as battlefield behaviour, the use of certain weapons and the treatment afforded to different categories of persons.⁶⁰

As with state practice, the practice of NSAGs is innately Janus-faced. It contains both the positive practice of committing to and complying with IHL and the negative acts of disregarding, rejecting and wilfully breaking the rules. As often reported in NIACs, some NSAGs have blatantly denied, in whole or in part, the application of IHL. In the context of custom formation, those persistently defying existing norms and continuing waging their rebellions without any restraints could arguably be labelled as ‘persistent objectors’.⁶¹ This reality entails a pressing question: should the anticipated NSAG-contributed customary IHL incorporate both the positive and negative components of their practice? With a view to curbing the implications of the contrary practice of NSAGs, Anthea Roberts and Sandesh Sivakumaran suggested that ‘armed groups, acting alone, would not have the power to create a new custom or undermine or change an existing custom.’⁶² It is also noteworthy that the ICTY, which affirmed the direct relevance of the practice of NSAGs, has only cited their practice consistent with the objectives of protecting war victims.⁶³ Nonetheless, the lack of articulated justifications and criteria for doing away with what looks repugnant risks opening the door to arbitrariness and subjectivity.

The imbroglio caused by the negative practice of NSAGs might be defused through the mutatis mutandis application of the norms governing custom-making to the case of NSAGs. On one hand, deviation from a rule should generally be treated as ‘breaches of that rule, not as indications of the recognition of a new rule’.⁶⁴ Therefore, the misdeeds of NSAGs can seldom result in the emergence of new norms. On the other hand, the formation of CIL demands state practice be ‘extensive and virtually uniform’⁶⁵ rather than universal, and be generally consistent with, rather than ‘in absolutely rigorous conformity with’, the rules.⁶⁶ While NSAGs have often been demonised as nothing but criminal gangs, some scholars testified that contrary to the public’s stereotype, NSAGs have frequently committed to and/or implemented humanitarian norms,⁶⁷ and that only a few NSAGs entirely reject the pertinence of IHL.⁶⁸ As Raphaël van Steenberghe analysed, ‘practice evidences that most of the armed groups being party to an armed conflict are ready to respect IHL … [T]he rare official oppositions from armed groups to IHL application do not seem to be sufficiently important [to the formation of CIL rules].’⁶⁹ If the empirical evidence as offered is solid, the negative practice of NSAGs may not carry as much weight as it seems to have in determining the contents of CIL norms.

Another mechanism that would function as a powerful filer of practice is jus cogens (peremptory norms). One of the legal consequences of jus cogens in relation to CIL is that even if constituent elements of CIL are present, a putative customary rule does not come into existence if it conflicts with jus cogens.⁷⁰ In this sense, there would be no customary rule that may conflict with jus cogens. As a logical result, the practice of NSAGs potentially contributing to such a ‘rule’ would simply play no role in its formation which is forestalled.

This inhibitory effect of jus cogens works for those that have persistently objected to an emerging customary rule and maintain their objection after the rule has crystallised. The persistent objector rule does not prevent the emergence of a customary norm of a jus cogens character to which one or more states have persistently objected.⁷¹ Likewise, even if an NSAG has persistently maintained its objection to a customary rule of a jus cogens character since that rule was in the process of formation, it could not thwart the crystallisation of that rule. In other words, its practice signalling persistent objection would be irrelevant to the formation of such a rule.

Customary IHL contains abundant rules of a jus cogens character. In its advisory opinion on the Legality of the Threat or Use of Nuclear Weapons, the ICJ described the fundamental rules of IHL as constituting ‘intransgressible principles’ of CIL.⁷² According to the ILC, the intention of the world court was to treat these principles as peremptory.⁷³ Thus far, it has been widely agreed among scholars that basic rules of IHL have achieved the jus cogens status,⁷⁴ although there is no consensus as to which basic norms exactly fall into this genre. Also, it has been determined in both international and domestic jurisprudence that IHL rules prohibiting war crimes form part of jus cogens.⁷⁵ Given the coverage of jus cogens in the realm of IHL, it is promising that not much room is left for NSAGs to create customary rules that contravene existing ones, or to maintain objections to them.

With these principles in mind, the implications of the frustrating negative practice of NSAGs should not be overstated. The positive and negative facets of their practice may be weighed together as a whole for determining the contents of customs.

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.

5.1 Implications on the Structure of CIL

The conceptual distinction between state practice and the practice of NSAGs gives rise to a question: if the practice of NSAGs were to count as an element of CIL, what would the hierarchy between such practice and state practice? Put another way, in determining the contents of norms, would the practice of NSAGs be superior, equal or inferior to state practice be? In fact, this question posits the existence of a unitary body of CIL norms applicable to states and NSAGs alike. A legal pluralist outlook would suggest another way of perceiving the structure of CIL. To address this hierarchy issue, the two scenarios should be examined respectively.

Traditionally, CIL is understood as a unitary legal system. As Daragh Murray maintained, CIL shall ‘be regarded as a unitary body of law, binding all entities possessing international legal personality. While elements of the law may be limited ratione personae, this does not imply the existence of distinct bodies of customary law relevant to particular categories of entity.’⁷⁶ However, the alleged self-generating nature of CIL as noted in Section 3.1 would imply that each and every type of entities which are bound by CIL should be vested with a power to create customary rules for itself. Such a theoretical perspective was envisaged by Anthony Clark Arend: ‘[i]f, however, the [S]tate were to lose its monopoly in a neomedieval system, … [t]here could, in fact, be multiple levels of CIL … [A] scholar … would need to examine the practice of this entire panoply of actors … [I]t is also possible that there could be rules of CIL that are binding on some, but not all, international actors.’⁷⁷

Indeed, if the creators of CIL are no longer to be confined to states, the traditional structure of CIL would be radically stretched. It is conceivable that this body of law is likely, if not bound, to become multi-layered to accommodate multifarious sets of norms created by and applied to different actors.⁷⁸ As Jean d’Aspremont wrote, ‘if one accepts that non-state actors can contribute to the formation of [CIL] … the practice of non-state actors can only be germane to the emergence of customary rules whose object is to regulate non-state actors’ behaviour.’⁷⁹ As to IHL, the expected diversification of CIL, resulted from non-state participation in custom-making, would engender two main layers of customary norms: (1) norms created by and applicable to states;⁸⁰ and (2) norms created by and applicable to NSAGs. In such a double-layered system, the hierarchy problem between state practice and the practice of NSAGs would not arise, because wherever a discrepancy exists between their practices, it can be handily deposited onto two entities’ respective exclusive domains of norms whose scopes of application ratione personae would not overlap.

In contrast, the hierarchy issue would emerge vis-à-vis the traditional concept of CIL if the practice of NSAGs were to become directly relevant. This is because as long as CIL is considered a unitary system, customary IHL norms binding states and NSAGs alike should be conceptually co-authored by both. In determining the contents of this body of law, the practices of the two entities should be weighed along with one another in a holistic manner. To bridge the potential discrepancies, Anthea Roberts and Sandesh Sivakumaran proposed that states and NSAGs’ practices ‘need not be treated equally. The centrality of the role of States in international law means that their practice should still be given more weight.’⁸¹ Admittedly, in affirming the direct relevance of the practice of NSAGs, neither the ICTY nor the Darfur Commission pointed to any of such practice that contradicted IHL norms created by states.⁸² However, there appears to be a tension between conferring upon NSAGs a custom-making role in the name of legitimacy on one hand and subordinating them to states on the other hand: if CIL’s legitimacy – in the proposed meaning of the term – is to be maximised, it is puzzling why states should retain the authority to deplete at will the value of the conduct of other actors. Probably a more sensible approach is to treat, in theory, the practice of state and NSAGs on an equal footing and to determine the superiority between them on a case-by-case basis, in consideration of the specific problem that a norm seeks to address.⁸³

5.2 Implications on the Application and Protective Standards of Customary IHL

Besides challenging the unitary character of CIL, incorporating the practice of NSAGs would lead to reappraising some crucial aspects of IHL. The most conspicuous is the risk of nullifying the equal application of IHL in NIAC, a principle denoting that IHL ‘applies equally to both sides of a conflict’.⁸⁴ Apparently, this principle perches at the opposite end of a conceived double-layered system of CIL, in which states and NSAGs are bound by different norms. Two options for assuaging this tension would arise. One is to incorporate the practice of NSAGs into the traditional unitary body of CIL, on account of the importance of equal application. The other option is to close the door for the equal application of IHL in NIAC, with the possibility of the stratification of CIL preserved.

At first glance, the price of taking the latter path seems exorbitant. In fact, however, there is an ongoing debate in academia over whether the principle of equal application should be established in NIAC. Those who contest the applicability of this principle in NIAC first looked for support from the historical origin of this principle. It was argued that the principle of equal application resulted from the separation between jus ad bellum (the law governing resorting to force) and jus in bello (the law regulating the conduct of war) – a dichotomy that exists only in inter-state conflict.⁸⁵ The principle of equal application is a necessity to ensure that the same IHL norms apply to both belligerent states regardless their respective causes for resorting to war. International law, however, does not traditionally regulate the legality of the use of force in internal strife.⁸⁶ In other words, there is no ad bellum/in bello separation and the resultant theoretical demand for equal application in NIAC. In addition to this argument, it was also noted that the transplant of the equal application from inter-state conflict into NIAC is practically problematic and undesirable, for requiring NSAGs to implement the same obligations as states do is unrealistic, and may result in a high incidence of non-compliance.⁸⁷

Nevertheless, it is the majority view that the principle of equal application is doubtlessly applicable to NIAC.⁸⁸ While they did not gainsay the inter-state origin of the equal application, the defenders of this principle warned that abandoning it in NIAC would reduce the incentive of government forces involved to comply with their IHL undertakings, as they would hardly benefit from an asymmetric IHL.⁸⁹

It is not the intention of this chapter to work out a definitive answer for this debate. The suggestion is that there is some room to dispute, in theory, the equal application of IHL in NIAC, at the service of the envisaged customary norms based exclusively on the acts of NSAGs. Be that as it may, this would turn out to be a thankless, if not only dangerous, exercise as it denotes radical deviation from the classical perception of how IHL is applied.⁹⁰ It may then be safer to stick to a unitary body of CIL under the co-authorship of states and NSAGs, with the same customs regulating both sides. As can be seen, concerns related to equal application of IHL would restrain the choice of theoretical pathways for those in favour of giving NSAGs a direct role in custom-making.

Another implication that incorporating the practice of NSAGs might have is the risk of regression of IHL, that is, the downgrading of its existing protective standards.⁹¹ While, as analysed in Section 4.2, the negative practice of NSAGs may not have a scathing impact on the formation of CIL, the crux here is that even the acts of NSAGs aimed at regulating their conduct or protecting war victims appear to be more primitive and less humane than states’. For instance, some NSAGs in countries like Sierra Leone, Uganda, Philippines, Nepal and India adopted their codes of conduct based on an instruction of the Chinese People’s Liberation Army, with which they share a similar ideology.⁹² This instruction encompasses a number of rules of humanitarianism such as ‘do not hit or swear at people’, ‘do not take liberties with women’ and ‘do not maltreat captives’.⁹³ Notwithstanding such laudable efforts, it is evident that those norms, being oversimplified and crude, would not afford war victims protection as comprehensive as that offered by state-crafted ones. Further examples can be found in the codes of conduct of the Ejército de Liberación Nacional in Colombia, Ejército Zapatista de Liberación Nacional in Mexico, the Sudan People’s Liberation Army and the Taliban in Afghanistan. All of those codes contain merely rudimentary or concise rules, and fail to cover many critical aspects of IHL.⁹⁴ Should such instruments, reflective of the practice of NSAGs, be taken as a basis for customary IHL, it would be hard to formulate norms reaching the protective standards that have already been achieved in IHL. In effect, this dilemma arises also in relation to the law-making of other non-state actors. As Hilary Charlesworth observed, engaging non-state actors in the creation of CIL ‘often has the effect of generating weak norms on a wide variety of topics’.⁹⁵

However, such a pessimistic prospect foretelling the inevitable regression of law is not infallible. It should be equally noted that NSAGs have sometimes embraced more protective rules with obligations whose scopes are wider than those agreed among states. An illustrative example is Geneva Call’s Deed of Commitment on landmine ban which has been signed by many NSAGs already.⁹⁶ While the Ottawa Convention on anti-personnel landmines, signed among states, prohibits mines that are ‘designed to be exploded by the presence, proximity or contact of a person’,⁹⁷ Geneva Call’s deed bans mines that have such an effect, whether they are designed for that purpose or not.⁹⁸ Furthermore, there are also cases in which unilateral commitments of NSAGs may be more humane than IHL standards in certain respects. For instance, the Sudan People’s Liberation Movement committed to apply the Convention on Certain Conventional Weapons at a time when the convention was applicable to international armed conflicts alone, and the National Transitional Council of Libya issued a communiqué that prohibited the use of anti-vehicle mines and anti-personnel mines during the 2011 civil war in Libya.⁹⁹

These examples illustrate that the alleged regression of IHL may not necessarily be the case, or at least, may not take place in every corner of the law. In certain areas where NSAGs assent to more protection, more progressive norms are in sight. Nevertheless, more empirical studies are needed for judging the exact implications of the practice of NSAGs on the protective standards of customary IHL.

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.¹⁰⁰ More efforts to engage NSAGs are certainly desirable,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.¹⁰² In essence, states may have a keen interest in maintaining their exclusive or dominant role in law-making.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.¹⁰⁴ For the sake of theoretical completeness and practical utility, proponents of this project are invited to make further elaborations on it.

15 Identifying Custom in Universal Periodic Review Recommendations

Frederick Cowell

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.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.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.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.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.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.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.⁷ Rosalyn Higgins concluded in a 1965 paper that the repeated practice of UN political organs was of ‘probative value as customary law’.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.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.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.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’.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.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’.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.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.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.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.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.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.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.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.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”?’.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.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.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.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.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’.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.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.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.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.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.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.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.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.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.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.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.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.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.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’.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.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.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.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.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.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.⁴⁸ Sometimes this relates to a course of action already decided upon by a state party and recommendations help reinforce this course of action.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.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.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.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.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.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.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.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.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.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.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.

4.1 Acceptance of Recommendations and Practice

Although state practice was historically conceived as the physical acts of states, for example by controlling which ships were allowed into a particular area, there is now a general recognition that verbal acts can in certain circumstances constitute state practice.⁶⁰ UPR recommendations are issued through an official process, created by a GA resolution that has a broad-based international acceptance and is treated by human rights treaty bodies as being authoritative evidence of state practice in relation to a particular norm. Accepted recommendations go beyond mere statements of practice.⁶¹ International decisions – such as a GA resolutions – are exceptional as their institutional provenance means that they are considered indicative of either how states are acting, ought to act or ought not be acting.⁶² Conclusion 4 of the ILC’s Draft Conclusions on the Identification of Customary International Law makes specific reference to international organisations in the context of state practice, and Draft Conclusion 7 notes that practice includes a ‘wide range of forms’ including ‘resolutions adopted by an international organization’.⁶³ Strictly speaking recommendations are not resolutions but they are advanced as part of an organisational process and therefore would be analogous to the processes outlined in both the ILC’s Draft Proposals and the International Law Association’s (ILA) final report.⁶⁴ In fact as the ILA’s report goes on to note the practice of international tribunals ‘is replete with examples of verbal acts being treated as examples of practice’ so the concept of practice is viewed in relatively expansive terms.⁶⁵

Yet, this raises the issue of what precise moment in the UPR process – acceptance of a recommendation or implementation of the substance of the recommendation – constitutes state practice. Implementation of a recommendation would demonstrate the existence of a concrete human rights protection within a state and therefore be the physical manifestation of a principle. But, as the ILA notes, ‘statements in international organizations and the resolutions these bodies adopt’ are more common than ‘physical acts, such as arresting people or seizing property’ leading to the conclusion that if a claim is publicly communicated it would constitute an act for the purpose of custom.⁶⁶ Following UNHRC Resolution 16/21, a state is required to ‘clearly communicate to the Council … its positions on all received recommendations’ entailing that there is a requirement on states to take a public position in relation to the recommendations they have been offered.⁶⁷ Acceptance of a recommendation is therefore made in public, recorded in an official UN document, with an expectation that its terms will be put into practice by a state; thus when a recommendation is accepted, it is state practice. As Malcom Shaw puts it, if practice is considered as simply ‘what States actually do’, then publicly making a commitment to implement a specified human rights reform and consenting to be examined on progress towards that reform in four years’ time, is what states ‘do’.⁶⁸ Yet, an individual recommendation and acceptance of it by a state under review would not really be sufficient to establish that there was state practice as a recommendation applies to a particular state.⁶⁹ Practice, according to the ICJ, needs to be widespread as well as ‘sufficiently extensive and convincing’ in order for it to be considered the basis of custom.⁷⁰ Therefore, multiple accepted recommendations of the category four or five type, which by their nature require a specific course of action on a human rights norm by states, would need to be shown in order to demonstrate a practice.

Even if a chain of accepted recommendations on the same subject can be identified, there are likely to be some rejected recommendations on the same subject. The rejection rate of recommendations in action category five supports the idea that states act with the belief that because of their framing, such commitments are in some way consequential.⁷¹ Yet Elvira Domínguez Redondo notes that this pattern of behaviour can be interpreted narrowly, as simply the state under review ‘asserting its reluctance to be monitored by the UPR on the implementation of such a recommendation during its next review’.⁷² Therefore, a state may not actually object to the substance of the recommendation but wish to avoid, for a variety of reasons, accepting a UPR recommendation on the subject. The multiplicity of motivations behind rejected recommendations means that they are difficult to read as a conclusive manifestation of the persistent objector doctrine.⁷³ As Joel Trachtman notes, the doctrine only applies ‘when the customary rule is in the process of emerging’ but that this is a somewhat problematic part of the principle as custom ‘is always in a zen-like process of becoming and un-becoming’.⁷⁴ As UPR recommendations (as shown below) on the same subject might be accepted in some cases but rejected in others, this means that it is difficult to pinpoint the moment of becoming for a norm. Furthermore, in relation to customary human rights law, the persistent objector rule could mean that a right that ought to be universal is essentially opted out of by a state, although human rights tribunals have rejected this argument where the right is considered jus cogens.⁷⁵

Rejected recommendations also do not give much of an insight into the substantive objection to a recommendation. As Lynn Loschin identifies in her four-part model for analysing the persistent objector doctrine in international human rights law, ‘the quality and quantity of the State’s objection’ would be important for validating whether the objection reflects a genuine preference of a state.⁷⁶ It is, for example, entirely possible that a recommendation is rejected based on part of its text and not as a reflection of the whole recommendation. Even if a state’s rejection of a recommendation is relatively consistent over review cycles, that also may not be grounds for saying that a customary norm should not be universal. As Lepard argues, customary human rights law should be about what rights ought to be protected as a matter of international law, not an assessment of the often-inconsistent nature of state practice.⁷⁷ Accepted and rejected recommendations therefore need to be considered in tandem in order to ascertain the nature of a norm which emerges from recommendations, but there would need to be a high number of accepted recommendations which leads onto the next issue – quantification.

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’.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.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’.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.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.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.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.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.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.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.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.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.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.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’.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’.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’.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.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.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.

4.3 UPR Recommendations: The Sense of Obligation

The participatory nature of the UPR process arguably makes any consensus identified in accepted recommendations more justifiable on democratic grounds as the basis of a shared belief that a particular principle ought to be binding, in accordance with the interpretation of opinio juris set out in the first section of this chapter. All states have participated in at least one UPR review and all states are treated equally before it, in that they all get to be reviewed and can contribute to other states’ reviews. Unlike human rights treaty bodies, which subject states to review by panel of experts, the UPR process is genuinely participatory. Nicole Rouhgan’s work on the democratic formation of custom attempts to reconcile the way that custom’s formation ‘falls short of contemporary ideals of democracy’ and is characterised by an absence of a mechanism ‘to protect formal equality in the development of customary rules’.⁹⁶ Emmanuel Voyiakis echoes this criticism, noting how international systems are riven with inequalities, reflecting the interests of powerful states in the formation of custom to the extent that CIL as a concept lacked a firm ‘justification for generating rules with normative force’ over other states.⁹⁷ Most customs, as Anthea Roberts notes, are based on the practice of fewer than a dozen states, meaning that formation of custom skews towards states with power and knowledge of legal formation creating a situation which by default privileges powerful states.⁹⁸ In an attempt to re-found an understanding of CIL’s formation that is more democratic Roughan argues that it should be ‘understood at its core to be a matter of social participation’.⁹⁹ A series of accepted recommendations would be representative both of a broad commitment on a particular human rights norm from individual states accepting recommendations on that norm, but also would represent a positive statements from the states offering those recommendations on what they believed the law ought to be. But, whilst this would affect the validity of the consensus behind a particular norm, it would not give an insight into the subjective belief that the norm is or ought to be binding.

The process of taking part in the review and being scrutinised on the implementation of accepted recommendations is a form of ongoing interaction, which can build a sense of obligation. Jutta Brunnée and Stephen Toope argue that processes of institutional interaction on the part of a state can build a sense of fidelity to the institution encouraging them to reshape their behaviour so as to create a sense of legality.¹⁰⁰ Their thesis has received criticism from different directions, including claims that it is too reductive about the nature of obligations and fails to really interrogate the nature of international society within which states’ values are supposedly shaped.¹⁰¹ Yet, interaction has instrumental value in showing how the understanding of a norm as obligatory can emerge. As Brunnée and Toope note, within all systems of law (national or international) ‘law is constructed through rhetorical activity producing increasingly influential mutual expectations or shared understandings of actors’.¹⁰² Research on the politicised nature of UPR recommendations actually underscores the conclusion that states view the acceptance of recommendations as a process that involves accepting responsibilities.¹⁰³ This is because recommendations, when offered in a partisan manner, still appear consequential to the state under review and the overlap between the UPR and other legal processes means that the UPR process itself is seen as important. A reaction to recommendations would need to be actually observed in order to establish that changes were arising in part as a result of interaction with the UPR process, in order to meet the commonly accepted requirements of the subjective element of custom.

One example of such a reaction is child marriage; there have been a few hundred recommendations issued to states in relation to the issue of early forced marriage, 68 per cent of which have been accepted by states across all three cycles. Early forced marriage is prohibited in CEDAW and in the CRC but there is a tension about both the scope of the prohibition and the age of marriage – CEDAW specifies no minimum age of marriage but the CRC implies eighteen.¹⁰⁴ The HRC has interpreted the provisions in the ICCPR on the right to a family in a way which allows for individual states to reach their own conclusion about marriage laws.¹⁰⁵ Recommendations made to Indonesia in their second review cycle to eliminate early marriage prompted the government to investigate the enforcement of marriage laws and to draft a new law raising the age of marriage to eighteen across the country.¹⁰⁶ At their third cycle review they accepted recommendations on the outright prohibition of forced early marriage.¹⁰⁷ Benin accepted recommendations during its second cycle to abolish early marriage and in its third cycle national report detailed measures it had taken to implement new legislation protecting children’s rights.¹⁰⁸ In its third cycle review during the interactive dialogue states expressed concern about the persistent prevalence of early forced marriage in spite of changes to the law and in response Benin committed to prosecutions of the perpetrators of forced marriage.¹⁰⁹ In both cases the state undertook actions indicative of a belief they were under an obligation to fulfil the substance of the recommendation. These are just two states and many other states have accepted recommendations on this subject, but these examples serve to illustrate how opinio juris can be inferred by looking at a state’s subsequent conduct in the UPR process in relation to the recommendation.

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.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.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.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.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).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.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.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.