1. Introduction
In its 2018 Draft Conclusions on Identification of Customary International LawFootnote 1 (hereinafter Conclusions), the International Law Commission (ILC) sets out the following caveat for its efforts to provide organized doctrinal guidance on custom identification:
[T]he draft conclusions do not address, directly, the processes by which customary international law develops over time. Yet in practice identification cannot always be considered in isolation from formation; the identification of the existence and content of a rule of customary international law may well involve consideration of the processes by which it has developed. The draft conclusions thus inevitably refer in places to the formation of rules of customary international law. They do not, however, deal systematically with how such rules emerge, change, or terminate.Footnote 2
This position is intriguing in the way it distinguishes two types of questions. To begin with, custom’s ‘formation’ and other development-related issues are considered essentially a matter of fact. They hinge on how members of the international society interact with each other and how custom gradually gains its ground in this process.Footnote 3 In contrast, the ‘identification’ of custom is deemed a matter of law and of how facts are to be assessed, regarding which doctrinal efforts may provide useful guidance.Footnote 4 This presents an analytical rather than descriptive approachFootnote 5 of approaching custom’s nature. However, other than the somewhat evasive language of ‘[cannot] always be…in isolation’, the ILC provides no further elaboration on when and how ‘formation’ and ‘identification’ may overlap or diverge or on the extent to which they inform the ILC’s work.
This article seeks to clarify the relationship between the formation and identification of custom. It is structured in five sections. Section 2 introduces the analogy between custom identification and historiography and explains the rationale behind this comparison. Drawing on this analogy, Section 3 proceeds to defend the position that custom’s formation and identification are an inseparable undertaking of narrative construction.Footnote 6 It also highlights that the common notion of ‘custom identification’, as understood by the ILC, does not carry an independent meaning. Equally importantly, Section 3 evaluates the extent to which the different criteria for good practices in law and history overlap. The central argument is that while a direct, vertical ‘transplant’ of such criteria across disciplines is untenable, it would be an overstatement to claim that the standards of good historical scholarship cannot inform those of legal argumentation.
Based on the idea that custom’s formation and identification are conceptually inseparable, Section 4 shows the extent to which some doctrinal controversies arise from the view that custom emerges spontaneously. Three issues will be highlighted there: the logically flawed definition of custom and opinio juris, the controversy over the sufficient duration for custom’s formation, and the possibility of interpreting ‘already existing’ customary norms to resolve legal disputes. As will be further explained, these issues are problematized precisely due to their close link to the idea of custom’s chronological development. Section 5 concludes the main arguments made throughout the article.
2. Historiography: Its meaning and value as an analytical lens for custom identification
This section introduces certain basic concepts, definitional and methodological issues relevant to the arguments that follow. The starting point is a triad of key terms constantly employed in the literature to refer to historical inquiries: ‘the past’, ‘history’, and ‘historiography’. The first point to note is the well-established distinction that historians draw between ‘the past’ and ‘history’. ‘The past’ is a collective, umbrella concept consisting of whatever events, incidents, and human activities took place in any previous temporal moment, whereas ‘history’ denotes ‘the ‘meaning’ [produced] from the past’.Footnote 7 In other words, while any historical writing is based on archives or documents recording past events,Footnote 8 it is the meanings constructed by us today – rather than those archives and documents per se – that constitute ‘history’ in the proper sense of this term. This definitional distinction reflects a commonsense view that the study of history neither seeks to, nor is capable of, exploring everything that happened in the past.Footnote 9 Rather, historians’ ability to construct meaning is inherently limited by various factors, such as the availability of documents.Footnote 10
A second point lies in the separation between ‘history’ and ‘historiography’. According to Oxford English Dictionary, the term ‘historiography’ can be broadly conceived to mean ‘[t]he writing of history; written history’ and ‘[t]he study of history-writing, [especially] as an academic discipline’.Footnote 11 While ‘historiography’ and ‘history’ may be used interchangeably for this reason,Footnote 12 throughout this article ‘historiography’ will be understood in the second sense, referring to scholarly discussions on the nature and assumptions of historical investigations.Footnote 13 This terminological choice is made to avoid confusion. Since, as has just been noted, history is a meaning one constructs from the past, historiography becomes an appropriate alternative that describes the study of meaning production (i.e., historical writing) itself. In this sense, historiography, sometimes termed ‘philosophy of history’Footnote 14 or ‘methodology of history’,Footnote 15 concerns not only such technical issues as verifying the authenticity of archival documents, but also fundamental issues of how research questions are formulated; how evidence is gathered, selected, and presented; and conceptual problems of historical objectivity, and so on.Footnote 16
As will be further elaborated in the following sections, an analogy between custom identification and historiography here is far more than another rhetorical device which serves only aesthetic purposes. Rather, since both undertakings aim at making sense of the past from today’s perspective,Footnote 17 insights from historiography can help unravel some deep-rooted puzzles surrounding customary international law that, albeit seemingly unrelated, are conceptually linked.
Indeed, it might be argued that two fundamental differences remain between legal and historical practices, rendering the analogy made in this article inappropriate. To begin with, history is generally descriptive, concerned with the ‘is’ rather than the ‘ought’,Footnote 18 while the study of law focuses on the opposite.Footnote 19 However, from a technical and formalist perspective, this criticism is somewhat misplaced. As clarified above, ‘history’ is defined here as the ‘meaning’ established from the past, and ‘historiography’ as the study of meaning-making. In this article, the analogy is drawn between custom identification and the latter, not the former.
Next, even setting aside this potential misunderstanding for the sake of argument, the distinction between ‘is’ and ‘ought’ is not a sufficient reason to preclude analogies between law and history. Although a detailed engagement with the relevant literature goes beyond the scope of this article, it is important to note that the distinction should not be taken as absolute; and accordingly, a reference to what ‘is’ can still inform our understanding of what ‘ought’ to be.Footnote 20 For example, as MacIntyre remarks in a famous essay:
To say that we ought to do something is to affirm that there is a commonly accepted rule; and the existence of such a rule presupposes a consensus of opinion as to where our common interests lie. An obligation is constituted in part by such a consensus and the concept of ‘ought’ is logically dependent on the concept of a common interest and can only be explained in terms of it.Footnote 21
For him, it is the ‘common interest’ or ‘consensus’ – as a fact – that forms part of the basis for what ‘ought’ to be. Importantly, this seems precisely what international custom, defined as ‘a general practice accepted as law’ in Article 38(1)(b) of the Statute of the International Court of Justice (ICJ), is meant to achieve. Following MacIntyre’s argument, the provision essentially seeks to transform an ‘is’ – namely, the presence of a patterned practice – into an ‘ought’ (i.e., a binding custom), by way of another fact: the presence of a sense of a legal obligation.Footnote 22 Hence, as is put by Tomuschat, ‘to recognize custom … is an imaginative device to establish congruence between the is and the ought’.Footnote 23 It is also in this vein that Orford asserts that ‘any description of past practice [is] a legal argument’Footnote 24 in itself.
Lastly, the use of history has long been a common component of the (international) legal argument. For example, historical narratives are at the core of the so-called ‘originalist’ interpretive approach to US constitutional law, and any claim to ‘discover’ the original intent of framers inevitably entails an element of meaning-making from the past.Footnote 25 The same is true for international law, where past practice of states and travaux préparatoires of treaties, for example, are institutionalized sources of information that can be drawn to produce doctrinal guidance on today’s legal matters.Footnote 26 In other words, there is no inherent reason why historiography – the study of how history is written – should not serve as a fruitful theoretical resource for legal scholarship.Footnote 27
The use of history in legal argument, however, leads us to a second potential challenge against the present article’s analogy between custom identification and historiography. For some, history drawn in legal argumentation concerns not so much about pursuing historical objectivity; rather, it is employed selectively to advance a normative claim and win a case.Footnote 28 As such, Tushnet writes that ‘Law-office history is a legal practice, not a historical one. The criteria for evaluating it, for determining what is a successful performance, must be drawn from legal practice rather than from historical practice’.Footnote 29 To ensure a smooth flow of argument, this point will be dissected below in Section 3.2.2. For the time being, it is sufficient to note that the distinction between law and history does not render historiographical insights irrelevant to the present context.
3. The inseparability of custom’s formation and identification
3.1 Formation and identification as narrative-construction
By distinguishing custom’s formation and identification, the ILC implies that it is one thing to say that a customary norm exists and quite another to say that the norm is, or can be, identified. In line with this understanding, identifying a customary norm is synonymous with finding a pre-existing custom as it is. Analogizing custom identification to a physicist’s identification of gravity laws through the patterned movement of planets,Footnote 30 D’Amato asserted in his speech during the 1988 annual meeting of the American Society of International Law that ‘We don’t dictate what [customary] international law is; we find it out there’.Footnote 31 Parisi, in the same vein, wrote that ‘Judicial recognition of spontaneous norms amounts to a declaratory (rather than constitutive) function that treats custom as a legal fact. The legal system ‘finds’ the law by recognizing social norms but does not ‘create’ the law’.Footnote 32 More recently, Johnston has contended that as long as ‘a general practice accepted as law’ is present,Footnote 33 custom becomes present automatically according to Article 38(1)(b) of the Statute of the ICJ, independent of any process of law-ascertainment/identification:Footnote 34
Whether or not someone actually conducts an evaluation of state practice and opinio juris and concludes that sufficient evidence exists for a given customary international law rule to be identified has no impact on the existence of customary international law. The mere existence of that practice and opinio juris is necessary, but also sufficient, for custom to exist. It is in this way that the formation of customary international law is an informal process.Footnote 35
One may nonetheless raise a conceptual question regarding this understanding: how can we even know the existence of ‘a general practice and opinio juris’ before someone brings it up and provides evidence to argue that case? An intuitive response may be that this question is largely epistemological, and Johnston’s point is ontological, which, according to Westerman, should be clearly distinguished.Footnote 36 Nevertheless, this separation would create more problems than it would solve. For example, it seems to assume that some customary law, albeit present (i.e., ontologically existing), cannot be identified (i.e., epistemologically known to us) for some reasons. However, there is no way to demonstrate a norm’s existence independently of our knowledge – and even if there is, the practical benefits of doing so remain unclear. After all, the law would serve no function in shaping social behavior if it is unknown.
Indeed, attempts to ‘find’ already-existing custom are intuitively comprehensible due to custom’s essential connection to actual past events – signature of treaties, promulgation of unilateral declarations, protest, and so on. In historiographical terms, ‘[w]hether or not a historian has actually carried out the act of verification is irrelevant to [a fact’s] factuality; it really is there entirely independently of the historian’.Footnote 37 ‘A general practice’, therefore, is traditionally seen as ‘[the] less problematic’ and ‘objective element’Footnote 38 of custom. And relatedly, that custom gradually evolves through engagement among international actors can be presented and understood as a valid sociological description. Nevertheless, contra Parisi, it seems that any ‘judicial recognition’ of custom involves an inexpungable element of law-creation.Footnote 39 That element pervades almost every aspect of our ‘finding’ of facts/practice. First, on a micro-level, even the description of a single past event is not an automatic or mechanical undertaking. It is lawyers themselves that decide the way to describe ‘what is going on in the world’.Footnote 40 For example, one needs to describe the fact that ‘a man opens the door for a woman’Footnote 41 to formulate a proposed norm about it. Various options highlighting the different aspects of this act may be available:
‘[O]ne should always assist others who may benefit from such assistance’; ‘a man should always assist women who may benefit from such assistance’; ‘one should open the door for others’; ‘a man should open the door for others/women’ and so on and so forth.Footnote 42
Of course, there could be more to highlight: the man and woman’s ethnicity, social status, the colour and size of that door, etc. The limitless number of ways to present a fact further attests to the previous point that ‘a general practice’ can never be something ‘out there’ and automatically accessible.Footnote 43 Furthermore, even if the factuality of a door-opening act can be affirmed, this does not dictate or guarantee a singularly truthful manner in which that act can be described.Footnote 44
Next, articulating a customary norm’s existence requires more than ascertaining the factuality of isolated events. According to the ICJ’s jurisprudence, acts contributing to customary norms must ‘amount to a settled practice’Footnote 45 and ‘be consistent with [the alleged customary] rules’.Footnote 46 But an ocean of scattered and disjointed facts does not create such patterns and consistency automatically. Rather, we need to ‘make a pattern’Footnote 47 ourselves by choosing, synthesising, connecting, and interpreting various historical facts and, at the same time, assigning meaning to that set of carefully curated facts. For example, as Aral points out, three particular historical events are repetitively singled out in various scholarly works in their justification for a customary right to democratic governance – ‘the 1991 military takeover in Haiti [,] 1997 military coup in Sierra Leone [, and the] intervention of the Security Council in the post-election crisis in Côte d’Ivoire’.Footnote 48 Meanwhile, a similar meaning is inferred from these events: they are turning points that repeatedly reinforce the international community’s commitment to democratic governance. For some academic lawyers, these events create a picture of ‘a cumulative practice’Footnote 49 leading to the crystallization of customary international law. On the other hand, ‘any possible setbacks from the linear advancement of democratic rule’ would be ‘disregarded’Footnote 50 in such scholarly accounts. To borrow the words of Carr, it is indeed up to historians (lawyers) themselves to ‘[decide] to which facts to give the floor, and in what order or context’.Footnote 51
Noteworthily, in the process of ‘judicial recognition’ of custom, not only are the substantive, primary rules of customary international law established, but the interpretation and scope of Article 38(1)(b) of the Statute of the ICJ are also shaped and reaffirmed. For instance, Article 38(1)(b) does not dictate what behaviour/acts constitute ‘a general practice’ within its meaning. Thus, whether states’ verbal statements and physical acts are to be counted in equal measure presents a classic difficulty in this context.Footnote 52 It is also unclear whether the behaviour/acts of international actors other than states should count, leading to intense scholarly debates over international organizations’ role in building custom.Footnote 53 A ‘theoretically neutral’ observation of reality is non-existent.Footnote 54
In brief, while it is tempting to be in the shoes of a third-person observer and ‘describe things as they are’, any articulation of the two elements’ presence per se is already a subjective construction and intervention in reality.Footnote 55 Custom formation and identification are therefore inseparable and synonymous with narrative construction.
3.2 Narrative construction, massivity of evidence, and ‘objective’ law-making
Acknowledging custom’s character as narrative construction is more than a matter of purely academic interest. In the present section, this article describes how narrative construction works and how this notion helps clarify some theoretical uncertainties surrounding custom identification.
3.2.1 Purpose, evidence, and the massivity of facts
First, every undertaking of narrative construction is based on a guiding normative claim.Footnote 56 A claim contains a number of elements/‘qualifiers’,Footnote 57 such as the identity of the right holder and duty bearer, the content of rights and obligations (e.g., a prohibitive or permissive norm), the conditions in which rights and obligations are triggered, and so on. In other words, international lawyers have to determine what they want to argue and possess more or less a general idea of what the customary norm in their mind looks like. Thus, in the previous door-opening example, lawyers may choose to highlight the information they consider relevant and ignore everything else according to that claim.Footnote 58 Also, since the determination of that normative claim is essentially inspired by practical purposes of solving problems and achieving a desired outcome,Footnote 59 this qualifier selection is something to which the ILC’s Conclusions have little to say. Hence, there is no general legal requirement for or prohibition of the inclusion or exclusion of any element.Footnote 60
To determine the normative claim is also to determine the boundary of relevant evidence of that claim. For historians, facts or ‘raw material’Footnote 61 of the past do not become historical evidence unless used to respond to a question and for a purpose.Footnote 62 Similarly, as the ICJ notes in the case of the Right of Passage over Indian Territory:
In the course of the written and oral proceedings, [a right of passage claimed by Portugal] was discussed with reference to the different categories making up the right, namely private persons, civil officials, goods in general, armed forces, armed police, and arms and ammunition. The Court will proceed to examine whether such a right as is claimed by Portugal is established on the basis of the practice that prevailed between the Parties during the British and post-British periods in respect of each of these categories.Footnote 63
The starting point of the ICJ’s examination, therefore, was the normative claim Portugal put forward, and this claim simultaneously determined the scope of relevant materials which need to be consulted by the Court. By focusing narrowly on the bilateral interactions and the aforementioned categories contained in the customary-right claim, the Court justified its finding of a general practice and a customary norm governing the passage of Portuguese personnel in question. Likewise, in the case of Colombian–Peruvian Asylum,Footnote 64 a qualifier of locality was introduced initially by Colombia, leading to the Court’s limited inquiry into the diplomatic-protection practice of Latin America only.Footnote 65 Similarly, in SS Lotus, where the Permanent Court of International Justice (PCIJ) required that ‘a close analogy’Footnote 66 between past and present cases be carried out to identify binding norms, the Court substantially limited the scope of relevant state practice. Given the nature of the French–Turkish jurisdictional dispute in question, ‘[the PCIJ] omitted from its consideration purported precedents of practice involving crimes committed against one state’s nationals in another state’s actual territory [and] precedents relating to the exercise of jurisdiction concerning single-vessel incidents’.Footnote 67 Whilst the Court found no customary norm supporting the claim of France’s ‘exclusiv[e] flag state jurisdiction’Footnote 68 over its nationals involved in (a) maritime collision in this case, it managed to evaluate practice within a narrowly defined ambit and established the existence of a usage on this basis.
Hence, although the vast body of historical events and state practiceFootnote 69 has long presented a challenge to customary law-making, acknowledging the link between claims and their evidence clarifies the scope and extent of this problem. It figures most prominently when one, as D’Amato does, looks at a random mass of state acts, tries to reveal the underlying norms ‘out there’, and yet possesses no normative claim or guiding question in mind. This method offers no guidance whatsoever as to which direction one should go when collecting evidence. The same problem also becomes acute when a question or a claim is ill-defined and vaguely worded. For example, a claim to a ‘customary human right to life’ tells us nothing about the permissibility of capital punishment, euthanasia, or the government’s suppression of abortion. Conversely, states’ acceptance or prohibition of these acts can neither substantiate nor disprove the existence of such a claim to customary right. Similarly, it is one thing to argue that ‘prohibition against torture’ is a customary norm and another to say that there exists a duty to ‘prosecute’ acts of torture.Footnote 70 The former is a negative right by nature and the latter positive, and they necessitate different types of evidence showing that states follow them in negative and positive ways, respectively. The bottom line is that context-specific claims are required to address context-specific issues and achieve context-specific outcomes.Footnote 71
3.2.2 Objectivity of customary law-making
Linking customary law-making to narrative construction raises concerns about the minimal (degree of) objectivity of the former.Footnote 72 If, as stated above, customary law-making inherently hinges on lawyers’ discretion and practical considerations of how to present a fact, then one could potentially construct whatever kinds of ‘general practice accepted as law’ by choosing different qualifiers, giving weight to different events, excluding some of them from consideration, and inferring different meanings therefrom. As we always seem able to construct different narratives based on the same set of facts,Footnote 73 customary law-making cannot provide any definite or ‘irreproachable’Footnote 74 solution to a dispute on the existence and scope of a customary norm.Footnote 75
But such challenges are indeed common to all exercises of narrative construction. Deconstructivist literary critics, for example, have attacked the objectivity of history as an intellectual discipline, dismissing historians’ work as identical to that of novelists.Footnote 76 For most historians, this attack has generally been unsuccessful since the objectivity question is essentially a definitional issue, and contemporary historiansFootnote 77 do not proceed on the assumption that historical writings hinge on finding a singularly truthful understanding of the past. As is well-established in historiography today,Footnote 78 objectivity is generally understood as the extent to which a historical account conforms to standards of good scholarly investigation as adopted by the professional community of historians. These include, but are not limited to,Footnote 79 the meticulous examination of authentic documents and the requirement for a narrative’s coverage and coherence.Footnote 80 Coverage denotes that the wider the scope of past events a narrative incorporates, the better the narrative is.Footnote 81 Coherence requires the absence of logical contradictions in a narrative. Because the satisfaction of these requirements is inherently a matter of degree, objectivity can never be absolute and static. Moreover, in order to make a choice between competing narratives based on an identical body of events, we need to compare the extent to which these narratives conform to such criteria.Footnote 82 As such, one can only speak of a higher or lower objectivity of a narrative vis-à-vis that of other competing narratives. Objectivity is relative in this sense.
It is certainly true, as briefly mentioned in Section 2, that legal practice and historical research pursue distinct aims. And for this reason, the criteria for good historical writing and lawyering may differ. Nevertheless, it would be an overstatement to assert, according to this view, that historiographical debates on historical objectivity and ‘good history’ offer no insight into the process of customary law-making. To start with, it is precisely the question of objectivity that has been raised by some academic lawyers mentioned above, including Joyner, Hakimi, and Staubach. Indeed, these authors have refrained from stating positively what they mean by objectivity, and whether their understanding is akin to that of most historians. However, it is sensible to see the classic doctrinal requirement for a general practice in customary law-making as presuming certain substantive criteria for historical objectivity. For a claim to custom to be persuasive, the ILC insists, the practice ‘must be sufficiently widespread and representative, as well as consistent’.Footnote 83 In other words, the more instances which can be explained by the presence of an alleged customary norm, the better a claim to custom is. Ideally, a persuasive claim should also not include potential inconsistencies in the general practice. Whenever inconsistencies arise, international lawyers have to explain why it does not undermine the overall picture of consistent practice – just as writers advocating a customary right to democratic governance do.
The relative nature of historical objectivity also explains why setting a fixed and precise threshold for the generality, consistency,Footnote 84 or exhaustivenessFootnote 85 of custom-forming practice always fails. They are criteria for claims to custom to satisfy to the greatest extent possible, rather than black-and-white requirements which are either met or unmet. This leads us to an even broader conceptual point about (customary) international law. Higgins, for example, characterizes international law as a ‘process’ of making authoritative decisions by citing, among other things, Lauterpacht’s view that the international law judges are choosing ‘between claims which have varying degrees of legal merit’.Footnote 86 Collins has also observed that in international legal decision-making, ‘[a]ll participants share the same burden of presenting the best evidence in favour of one rule or another, whilst the system itself forecloses the possibility that any of those determinations can have absolute finality’.Footnote 87 The puzzle of custom’s alleged lack of objectivity is thus resolved: first, under this view, the ‘legal merit’ (in Lauterpacht’s words) of each argument – in the present context, the extent to which claims to custom conform to the criteria for ‘objectivity’ – becomes a matter of gradation. Second, there is nothing worthy of lament about the parallel existence of different, competing narratives about custom. In contrast, this is the premise of any international legal decision-making, as there needs to be different narratives as options to be chosen.
4. Some practical implications of narrative construction
The separation between the formation/development of custom and its identification not only assumes the former to be ‘natural’ and immune to lawyers’ subjective will/control, but it also presumes that the element of time plays a central role in custom’s establishment. This is at the core of much confusion about custom identification. In the following sections, this point will be illustrated via practical examples.
4.1 The alleged circularities of custom’s definition
The ILC defined opinio juris in its 1950 handbook as a ‘conception that the practice is required by, or consistent with, prevailing international law’.Footnote 88 In highly similar wording, the ICJ declares in North Sea Continental Shelf that this concept denotes ‘a belief that this practice is rendered obligatory by the existence of a rule of law requiring it’.Footnote 89 Whilst the mainstream literature today has not had difficulties embracing this construction, what is simultaneously (and paradoxically) acknowledged is that this understanding is riddled with conceptual problems. As is famously pointed out by KelsenFootnote 90 and widely recognized by later generations of international lawyers, in order for states to believe they are acting according to a law, the relevant law must preexist.Footnote 91 Even more fundamentally, as Koskenniemi argues, the relationship between custom’s two elements is ‘hopelessly circular’Footnote 92 because the two-element definition relies on opinio juris to distinguish custom-forming practice from non-custom-forming ‘usage’, and, at the same time, it presumes that (state) practice is the basis from which opinio juris can be inferred.
Once again, these interconnected doubts can be demystified by discarding the idea of custom as spontaneous law. Two points need to be reiterated. First, observation and identification of custom are never purposeless – we need to decide what to argue in order to achieve the desired outcome; it is such a decision that guides our observation. Second, to ‘identify’ what states have considered legally obligatory is a process of assigning meaning to facts. It involves building a narrative of ‘this/that is what states have willed to be law’. Consequently, one does not ‘find’ opinio juris which is already there; rather, the content of opinio juris is in fact a manifestation of the desired conclusion one aims to reach, or the guiding question one possesses. For example, in the North Sea Continental Shelf case, when contending that Nordic states’ delimitation of continental shelf boundaries reflected an opinio juris accepting the ‘equidistance principle’ as law, Denmark was arguing, simultaneously, for the ‘equidistance principle’ itself as law. Similarly, In Right of Passage, by arguing that India’s non-protest against Portugal’s passage for a long time reflected an opinio juris acknowledging Portugal’s legal right of passage, Portugal was also defending its right of passage per se as part of custom. The chicken-and-egg relationship between custom’s two elementsFootnote 93 is thus clarified: we indeed should ‘know’ opinio juris in the first place, i.e., determine what to argue. One might still insist, for example, that this idea does little to address the most fundamental element of circularity. That is, we rely on a pre-determined hypothesis to determine the scope of relevant evidence on the one hand but revise that same hypothesis when evidence does not support it on the other.Footnote 94 Indeed, all narrative construction and theory building (including those in natural sciences)Footnote 95 concern a ‘continuous interaction between hypothesis and evidence’,Footnote 96 thereby showing a degree of ‘circularity’ in this sense. But it is unclear why this should be considered a devastating criticism. It has not prevented us from making sense of the reality we observe or creating some guidance on evidence collection and analysis.
A related issue can also be illuminated in this connection. For the classic discussion on the ascertainment of opinio juris, the issue of dispositive importance resides in the extent to which an act – such as acts of international organizations and the voting behavior of states at the UN General Assembly – can be said as opinio juris-reflective.Footnote 97 An underlying assumption here is that there exists a one-to-one correspondenceFootnote 98 between ‘a general practice’ and opinio juris. However, the assignment of meaning to facts does not operate on that assumption. Take, for example, a hypothetical criminal case where a male killed a female. During the trial, the court examined the following facts and convicted him of premeditated murder:Footnote 99 he went to a shop and purchased a knife, stalked the victim for a while and knew her daily routine, and eventually committed the crime with that knife one day on the victim’s way home. Except for the killing itself, these facts, if observed and described in isolation, tell the lawyers nothing about the criminal’s mens rea of the ultimate crime. Buying a knife is not prohibited by criminal law; it may even be for totally legitimate purposes like cooking.Footnote 100 Being a stalker, albeit likely illegal, does not automatically reveal an intent to kill, either. Nevertheless, these facts, when taken together, are capable of constructing a coherent narrative that the defendant in question engaged in a premeditated killing. Two points are clear in this example. First, for the purpose of dispute settlement, a legal meaning (mens rea) is retrospectively assigned to facts by lawyers.Footnote 101 It could well be that the defendant actually committed that killing on the spur of the moment or that he actually had no intention to do so when purchasing that knife. Nevertheless, these possibilities, in the end, become completely immaterial as long as no narratives are formulated to defend them.Footnote 102 The truth about the defendant’s psychological reality, if any, remains inherently agnostic to lawyers and judges.Footnote 103
Next, a legal meaning is not only retrospectively assigned, but it is assigned to a narrative as a whole.Footnote 104 Namely, in a narrative, ‘[t]he meaning of the whole is always a product of the parts yet, simultaneously, the parts derive their meaning from the whole’.Footnote 105 It is thus unnecessary – and practically difficult – to speak of the opinio juris-reflecting character of an individual act, an instance of omission or abstention, and so on. Even scholarly views which attempt to, or claim to have succeeded in, pinpointing the opinio juris-reflecting nature of some practice are conditioned by factual contexts. For instance, if taken in isolation, inaction and silence of a state may simply reflect its indifference or unawareness of the proposed norm in question, and therefore, no ‘acceptance’/opinio juris can be straightforwardly inferred therefrom. However, when observed together with facts that the state has been ‘in a position to react and the circumstances called for some reaction’,Footnote 106 an inference of opinio juris becomes justifiable. In a somewhat different context, in the case of Nuclear Tests,Footnote 107 the ICJ has also emphasized that the French statement conceding its own wrongs must be ‘considered within the general framework of the security of international intercourse’Footnote 108 to establish France’s subjective intent to be bound.Footnote 109 In any case, attempts to conceptualize an unmistakable expression of opinio juris are essentially beside the point.
4.2 The temporal requirement for custom
As Oppenheim pointed out, when we understand custom as emerging from a natural chronology,Footnote 110 a question that naturally arises would be how long the duration must be for a customary norm to gain ground. In the classic literature, a wide variety of views have been put forward on this matter. These have ranged, for example, from the requirement that custom needs to emerge from ‘a continuous practice from time immemorialFootnote 111 to the idea that ‘instant custom’ is conceptually possible. In the North Sea Continental Shelf case, the ICJ has declared that ‘the passage of only a short period of time is not necessarily, or of itself, a bar to the formation of a new rule of customary international law’.Footnote 112 For this reason, contemporary lawyers may be less troubled by this ‘duration’ problem than their intellectual predecessors. But, in any case, the notion of custom’s slow emergence remains central to the traditional understanding of custom in the literature. This is part of the reason why theoretical devices which promise to ‘accelerate’ the customary law-making processFootnote 113 have not ceased to attract academic attention.
Sender and Wood have recently examined these questions in further detail. Based on a survey of the existing doctrinal approaches to custom’s temporality, they present two basic observations on ‘how far the passage of time is necessary for the creation of rules of customary international law’.Footnote 114 First, there is no requirement for any ‘particular duration’ to be present in customary law-making.Footnote 115 Importantly, however,
[S]ufficient time must elapse so as to allow for the emergence of a general practice and an assurance of its acceptance as law (where the matter is otherwise equivocal). Thus the ‘time element’ has a role in the formation as well as the evaluation of the two constituent elements of customary international law.Footnote 116
Second, it indeed would be difficult to pinpoint the exact moment of custom’s formation, primarily due to the fact that custom emerges from engagements between international actorsFootnote 117 But for practical reasons, they write, ‘it seems there must be a “custom-making” moment: at one point there is no rule, and then there is’.Footnote 118 In this connection, they cite the ICJ’s Chagos Advisory OpinionFootnote 119 for example: in that case, the Court had to consider ‘whether the right to self-determination was a rule of customary international law in the period from 1965 to 1968 (and if so what its content then was)’.Footnote 120
These heavily qualified observations reveal the doctrinal puzzle scholars encounter when attempting to make sense of custom’s formation, identification, and the relations between the two. For example, asserting that ‘sufficient time’ is necessary for custom identification adds little to the existing debate. The ‘how far’ question they set for themselves remains largely unanswered. Moreover, if one concedes that the exact moment of custom’s emergence is not always discernable, to what extent is it meaningful to insist on the continued relevance of the temporal element? After all, ‘how much time is sufficient’ is an unanswerable question from a logical perspective unless we can discern the point at which a norm emerges. The two authors offer no further elaboration on such issues.
In truth, these issues cease to present challenges once we approach custom’s formation and identification through the lens of narrative construction. Here, the well-established conceptual distinction between ‘the past’ and ‘history’ provides explanatory aid. As briefly introduced in Section 3.1, past events are what we can single out and arrange in a framework of natural time/chronology.Footnote 121 But just because past events belong to a bygone age does not entail that history itself – as a meaning constructed on the basis of such events – belongs to a bygone age as well. Meanings and narratives are constructed in response to today’s challenges and from today’s perspectives.Footnote 122
Along the same line, one arrives at the conclusion that the only ‘moment’ meaningful to customary law-making is the moment when international lawyers put forward a narrative about ‘a general practice accepted as law’.Footnote 123 The fact that practical considerations call on determining a norm’s customary status on a given point, e.g., the Chagos case, does not transform custom identification from an ‘ex post facto construct’Footnote 124 into the excavation of already existing yet hidden norms. This point is partially, but clearly, illustrated by the presence of much scholarly disagreement on the customary status of self-determination before the ICJ ‘declared’ it as such in 2019.Footnote 125
Equally importantly, even the Chagos case offers little doctrinal guidance on custom’s allegedly temporal requirement in the sense of Sender and Wood. In that case, as the two authors correctly observed, the Court was answering whether a customary norm of self-determination could be said as present between 1965 and 1968. Whilst the Court speculated that such a norm might have started to evolve ‘since the adoption of the Charter of the United Nations and of resolution 1514 (XV) of 14 December 1960’, it never asserted either (i) that the day 14 December 1960 or that (ii) a given day between 1965 and 1968 could be seen as the exact moment of the norm’s formation. In this sense, the Chagos Advisory Opinion is not special or different as a reference point from numerous other case-law concerning custom identification, regardless of the Court’s eventual finding. For instance, in all the previously mentioned precedents, e.g. SS Lotus, Right of Passage, North Sea Continental Shelf, etc., the only question the international tribunals had to resolve was the alleged presence of a customary norm regulating the disputing parties’ interactions when the disputes in question arose. No clear determination of the precise, custom-forming moment was ever necessary/possible. And to recall, since the question of ‘exact moment’ and ‘sufficient time’ are logically connected, the inability to answer the former renders the latter equally unresolvable.
4.3 The ‘post-identification’ interpretability of custom
According to a dictionary definition, to ‘interpret’ something means ‘[t]o expound the meaning of (something abstruse or mysterious); to render (words, writings, an author, etc.) clear or explicit; to elucidate; to explain’.Footnote 126 Transposed in the legal context, it can be broadly understood as an act of ‘determining the meaning of a rule’.Footnote 127 If this definition is adopted, then essentially every step of custom’s identification may be said to count as interpretation. For instance, selecting qualifiers and drawing an analogy between cases are, by their nature, steps in clarifying the content of custom at the same time.Footnote 128
Many contemporary writers advocating custom’s interpretability, however, have taken a very different view on what is meant by interpreting a customary norm. For them, the foregoing process of choosing qualifiers is to be labelled as custom ‘identification’, and on the other hand, there is a separate stage of post-identification ‘interpretation’, through which the meaning of a ‘previously identified’Footnote 129 custom can be further pinpointed in the same way as interpreting a written treaty provision,Footnote 130 without regard to custom’s two constitutive elements. Writing in a recent symposium article, Merkouris and Mileva have made this stance explicitly:
It is important to acknowledge that at the stage of [custom] identification, a variation of interpretive reasoning also takes place, insofar as the evaluation of evidence of state practice and opinio juris is concerned. However, this is not legal interpretation stricto sensu. [For example, regarding treaty interpretation,] [i]n applying Art. 2(1)(a) [of the Vienna Convention on the Law of Treaties (VCLT)], courts also examine the text and the language employed in order to determine whether the states involved intended to create binding obligations. This process has some interpretative features and undeniably leads to some rudimentary content determination, but no court has ever argued that this is legal interpretation in the strict sense. When they seek to interpret, they apply Arts. 31–33 VCLT, not Art. 2(1)(a).Footnote 131
And also:
The reasoning employed at the stage of identification is concerned with questions about the relevance and weight to be given to evidence of state practice and opinio juris, and the outcome of this reasoning is a binary one – a [customary] rule is determined to either exist or not. The reasoning employed at the stage of interpretation is concerned with the determination of the content of the [customary international law] rule and how this rule applies to the case at hand.Footnote 132
As further support for this construction, Merkouris insists that we cannot afford not to recognize this post-identification interpretability for practical reasons. He writes: ‘[A] rule of customary international law, once identified by an international court or tribunal, does not cease to exist’.Footnote 133 Without recognizing this, customary norms ‘would have to be identified each and every time it is applied’.Footnote 134 As Milanovic has similarly insisted:
[T]here is simply no doubt that the content of some rules – whether based in treaty or custom – is of a general character and cannot be reduced to an enumerated list of specific applications. For example, if the human rights to privacy or freedom of expression exist in custom (as they do), applying them to novel contexts, such as cyberspace, does not require a new inquiry into state practice and opinio juris. Similarly, we can apply the customary principles of distinction or proportionality of the law of armed conflict to novel weapons, and do not need a weapon-specific inquiry into custom whenever a new weapon is developed.Footnote 135
A few scholars have pointed out the potential problems underlying custom’s post-identification interpretation. Lando, for example, emphasizes the infeasibility of applying the VCLT rules of treaty interpretation (i.e., Articles 31 and 32) in custom’s post-identification interpretation. To elaborate, Article 31 expressly limits its own applicability to treaty interpretation only.Footnote 136 And even if the provision can be said as constituting custom, this fact does not expand Article 31’s applicability to the post-identification interpretation of other customary norms.Footnote 137 Another issue which can be raised here concerns the legitimacy of interpreting a ‘previously identified’ custom – without repetitively affirming the existence of ‘a general practice accepted as law’, customary law-ascertainment can risk losing an adequate basis of state consent.Footnote 138
These challenges are generally valid: it would seem extraordinarily difficult to expect the interpretation of an unwritten legal norm to resolve doctrinal disputes and produce tenable guidance on state behaviour. But it is Merkouris and Mileva’s conceptual starting point – the separation between pre- and post-identification of custom – that presents the most obvious difficulties. As explained in the preceding framework of ‘formation as construction’, custom’s content is not to be intelligible and ‘identifiable’ until one puts forward a claim to custom and presents relevant evidence to substantiate it. Hence, any gap-filling through ‘interpretation’ can also be characterized as identifying a new customary legal normFootnote 139 with higher precision.Footnote 140 By the same token, no custom is ‘previously identified’ and ‘out there’ for international lawyers to access when a legal dispute arises. Evidence of state practice and opinio juris is constantly revisited, with different meanings being constructed therefrom to solve different problemsFootnote 141 – just like history always can, and needs to be, rewritten.Footnote 142
The next point to be addressed is the extent to which custom ‘continues to exist’ after initial identification. As the foregoing discussion reveals, a definitive answer to this question is absent. It all depends on whether, at a given point where the identification of custom is called for, we will be able to construct a claim to it and provide evidence to corroborate our position.Footnote 143 By acknowledging the ‘ever-changing’,Footnote 144 ‘ever-developing’,Footnote 145 and ‘dynamic’Footnote 146 character of custom, we embrace the possibility that the existence and content of a customary norm may vary. Meanwhile, we accept that custom’s development arises from acts and opinio juris which are prima facie incompatible with what we previously believed to be customary norms.Footnote 147
Of course, it seems intuitively appealing to think of legal norms as always existing – like treaty provisions whose binding force does not diminish or disappear simply due to the passage of time.Footnote 148 For instance, by claiming that provisions of widely ratified human rights treaties can be considered part of custom,Footnote 149 one seems able to ‘entrench’ certain fundamental values in international law and prevent it from being influenced by subsequent, non-conforming state acts.Footnote 150 Nonetheless, this ‘stable existence’ is achieved only by foreclosing the development of custom towards ‘undesirable’ directions and by denying the flexible character of custom in response to evolving global affairs. This poses obvious doctrinal challenges. For example, even concerning treaty interpretation, the VCLT explicitly requires that ‘any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation’Footnote 151 be taken into account. As such, it would be bizarre to contend that the content of a treaty provision may be influenced by states’ subsequent practice while claiming that the customary norm arising from that provision would always remain intact.
In truth, even if a customary norm is not regarded as ‘always existing’ after initial identification, its practical usefulness would not be adversely affected. As illustrated in the previous ICJ jurisprudence, for problem-solving, we may indeed have to construct a narrative about the status of custom at a bygone moment. The narrative is juxtaposed and compared against other competing narratives, and their objectivity may be assessed according to the criteria for objectivity. This has two implications. First, the crux of the matter lies in whether we can establish the (non-)existence of a customary norm at a past moment relevant to problem-solving. We do not need to assume that a customary norm will always ‘stay with us’ to achieve this purpose. Second, as Beham has commented, ‘a full assessment [of a norm’s status as custom] is only necessary once a specific argument is put forward’.Footnote 152 Before that full assessment, the current status of customary norms is unknowable.Footnote 153
Finally, a few words may be added to the relationship between ‘identifying’ a norm, ‘interpreting’ a norm, and ‘applying’ a norm to a given situation, which Milanovic does not seem to clarify. Indeed, Milanovic is right to point out the practical impossibility of exhaustively examining all the information that can potentially be extracted from state practice and opinio juris. The conclusion that privacy and free speech rights may be extended to cyberspace also seems supportable. However, what is missing is an account of why this must be the case. Except for the implied regulatory necessity, Milanovic offers no explanation as to why cyberspace, despite a new ‘context’, has no impact whatsoever on the automatic application of privacy and free speech rights. Hence, one does not know when and how far an ‘abstract frame’ can be expanded to new situations. For example, although the (customary) right to life seems broad and general enough and may be extended to various contexts and situations, as practice shows, its applicability has been contested where norms of lex specialis character exist, such as jus in bello.Footnote 154 Regardless of the reasoning and outcome of this debate, its very presence shows that the extension of an ‘abstract frame’ is far from an automatic or intuitive undertaking. To take another example, ‘[t]he rule that local remedies must be exhausted before international proceedings may be instituted’, recognized as ‘a well-established rule of customary international law’Footnote 155 by the ICJ in the case of Interhandel,Footnote 156 is perhaps ‘of a general character’ as well. In human rights treaties, it forms a basic prerequisite for anyone seeking remedies before international courts and treaty bodies.Footnote 157 By contrast, in the context of investment treaties, this rule garners scant international support,Footnote 158 and in most situations, ‘an investor may directly commence arbitral proceedings against a state for breaches of their rights by a public authority of that state’.Footnote 159 To what extent a context is (ir)relevant to an ‘already-existing’ norm can only be answered on a case-by-case basis, with a reference to practice. Hence, a better approach to justify Milanovic’s conclusions that privacy and free speech rights are applicable in cyberspace seems to be the fact that states, in practice, have not treated these rights differently in real life and in cyberspace.
Relatedly and more fundamentally, ‘applying’ customary norms requires us to draw an analogy between events which previously gave rise to a customary norm and the new event in question. And by acknowledging the ‘applicability’ of a norm to an event, we also acknowledge the relevance of factual information contained in that event as custom-forming when we are to ‘apply’ that norm again in the future. In other words, what hides behind the seemingly automatic law application remains an act of law-creation.Footnote 160
5. Conclusion
The ILC presents an ambivalent position on the formation and identification of customary international law in its Draft Conclusions. On the one hand, by indicating the separability of custom’s formation and identification, it attempts to adopt the viewpoint of an objective and third-person observer. However, while the Draft Conclusions provide useful guidance for international lawyers on how to pinpoint the two elements of custom, they also implicitly concede that custom identification lies in narrative construction. As a result, even the requirement for customary law’s temporality – the most typical and distinguishing feature of a norm that emerges over a long time – is conveniently obscured by the ILC in Draft Conclusions through a reference to ICJ’s equivocal claim that ‘the passage of only a short period of time is … a bar to the formation of a new rule of customary international law’.Footnote 161
The relationship between the formation and identification of custom is that they are the same. Although it could be argued that custom’s formation and existence – independent of identification – may be ‘presumed’,Footnote 162 there is no reason why we must do so. As explained above, this presumption provides little, if any, clarification of how customary law-making can be understood and carried out. The crucial matter in terms of customary law-making lies in how international lawyers exercise their discretion and tell their own stories about custom.Footnote 163
Open access
