I. An Overlooked Aspect in the Debate on Custom Identification
Since the 1986 Nicaragua judgment of the International Court of Justice (ICJ or “the Court”),Footnote 1 the identification of customary international law has been a constant focus of scholarly debate.Footnote 2 After Nicaragua, writers criticised the Court for only paying lip service to the two-element theory of custom, set out in North Sea Continental Shelf, under which rules of customary international law are identified by an inductive inquiry into the existence of “extensive and virtually uniform” state practice and “a general recognition that a rule of law or legal obligation is involved”, or opinio juris.Footnote 3 In Nicaragua, the Court had not been methodologically rigorous in establishing the existence of customary rules concerning the use of force, because it over-relied on opinio juris to make up for inconsistent state practice.Footnote 4 This criticism eventually resulted in academic commentary about “modern” approaches to custom identification. Under these “modern” approaches, the ICJ can identify customary rules by deduction, rather than induction, from basic principles of international law. If inductive processes are applied, the moral content of putative customary rules can influence the relative weight of state practice and opinio juris.Footnote 5
Diverging ideas among scholars and the hesitation of domestic courts on custom identification led to the International Law Commission (ILC) starting a study on the identification of customary international law in 2012.Footnote 6 The ILC's Special Rapporteur, Sir Michael Wood, released five reportsFootnote 7 culminating in the adoption of a set of Draft Conclusions in 2018,Footnote 8 of which the General Assembly has taken note.Footnote 9 The ILC aimed to “clarify the current rules on formation and evidence of customary international law”,Footnote 10 thus offering “guidance, not only to international lawyers, but also to domestic lawyers … who were often called upon to apply rules of customary international law”.Footnote 11 According to the ILC, the realities of international practice indicated that “an exhaustive review of State practice and opinio juris was exceptional, as more often than not evidence of a rule is first sought in the decisions of the [ICJ], the work of the [ILC], or in resolutions of the General Assembly and treaties”.Footnote 12 Sir Michael Wood also wrote that judicial organs other than the Court, such as the International Tribunal for the Law of the Sea (ITLOS) and the Appellate Body of the World Trade Organization (WTO), do not use any peculiar custom identification methodology.Footnote 13
The ILC's work on customary international law generated further academic debate on judicial custom identification methodologies.Footnote 14 In 2015, Talmon published an influential article arguing that the “the main method employed by the Court is neither induction nor deduction but, rather, assertion”.Footnote 15 This academic debate built on the view that it is beneficial for international tribunals to record, in their judicial decisions, the inductive reasoning they may employ to identify customary international law. Detailing evidence of state practice and opinio juris can also invite criticism.Footnote 16 However, justifying custom identification by reference to evidence of state practice and opinio juris helps make the exercise of the judicial function appear more legitimate.Footnote 17 Criticising an international tribunal's reliance on such evidence can contribute also to improving the quality of judicial reasoning in future cases and to developing international law, including by the “provision of information about and syntheses of State practice and opinio juris in order to establish customary international law”.Footnote 18
Academic writings and the ILC's study do not explicitly discuss the impact of the Court's internal working mechanisms on custom identification. Judicial decisions are the final, and the only public, stage of a complex deliberation process. All other stages are secret. Scholars of custom thus can see and analyse only the words on the pages of those decisions and the parties’ pleadings. Judge Tomka wrote extrajudicially that “the Court does not in every case undertake an inductive study of the relevant elements, at least in the text of its judgments”.Footnote 19 Although scholars may only see assertion, it may well be that induction, deduction or other processes leading to it have simply remained secret and never made it into the Court's decisions. By not considering this aspect in criticising the ICJ's approach to custom identification, scholars have held the Court to an ideal, unrealistic standard.Footnote 20
Different from the existing literature, this article discusses the connection between the ICJ's deliberation process and the Court's choice among custom identification methodologies. The ICJ's deliberation process strikes a balance between the two opposites of individuality and collegiality.Footnote 21 While the initial stages of deliberations are individual-driven, individual input gives way to collegial work as the subsequent stages of deliberations unfold.Footnote 22 This article argues that there is a significant correlation between the choice of methodology by which the custom identification exercise is recorded in judicial decisions and whether that choice is made by individuals or collegially. Although this article focuses on customary international law, its conclusions may inform one's analysis of how the ICJ identifies other rules of international law, such as general principles, or approaches choosing among different methods of treaty interpretation.
Section II frames the argument's focus more fully and outlines the methodologies adopted to develop that argument. Section III outlines the ICJ's little-known deliberation process. Section IV elaborates on how the custom identification exercise is recorded at the various stages of ICJ deliberations, also by reference to evidence gathered by empirical methods. The focus of Section V is twofold: first, to verify doctrinally the correlation between the choice of methodology by which custom identification is recorded in ICJ decisions and the degree of collegiality of deliberations; second, to discuss legitimacy implications. Section VI concludes.
II. Focus and Methodology
This article focuses on the ICJ for three reasons: first, as the only international judicial organ with general jurisdiction, it may have to identify custom across all areas of international law;Footnote 23 second, distinct from international tribunals created by treaty régimes, the main point of ICJ cases often concerns custom identification; third, the ICJ has been the centre of a considerable part of the debate on custom identification. This article borrows Talmon's analytical framework which distinguishes three methodologies by which the ICJ identifies customary international law: induction, deduction and assertion, the last one's being either simple assertion or assertion by reference to the Court's earlier decision or external sources, such as the ILC's work.Footnote 24 The findings of this article are complementary to the ILC's work on customary international law, because they can explain why the ICJ, despite subscribing to the two-element theory, does not always appear to identify customary rules based on evidence of state practice and opinio juris. Understanding how deliberations affect the choice among custom identification methodologies can also help assess the legitimacy of exercising the judicial function and thus the ICJ's accountability to its audience.Footnote 25
This article argues that the individuality or collegiality of different stages of the ICJ's deliberation process impacts the choice among methodologies to record the Court's custom identification exercises. Individual-driven stages of deliberations favour recording the outcome of custom identification as an inductive exercise.Footnote 26 Collegial action in deliberations promotes recording that outcome as a non-inductive exercise. This correlation does not mean that, as the final stages of deliberations are collegial, one should expect the ICJ always to identify custom by non-inductive processes. There are other factors which, despite collegiality, may promote recording the result of custom identification as an inductive exercise, such as the background of the decisions’ drafters and the subject matter of a case requiring the identification of customary rules.Footnote 27
Developing this article's argument would be challenging on a purely doctrinal approach. One may verify the argument's soundness by reference to how judges have recorded, in their individual opinions, the results of their custom identification exercise and by how such results have been recorded by other international judicial organs that adopt a “less collegial” deliberation model relying on drafting by judge rapporteurs. This doctrinal verification, however, only takes one so far. This article develops its argument also based on empirical analysis conducted by an anonymous survey.
A survey was found to be better suited to obtaining the desired empirical data than interviews, primarily because it could ensure a greater degree of anonymity: while the respondents’ identity is known to the author, one cannot link particular answers to particular respondents. Respondents were requested to answer 21 multiple-choice questions. To ensure that the survey would capture more nuance, each multiple-choice question allowed respondents to elaborate on each answer with a short comment. Respondents were requested to answer the questions exclusively based on their direct experience as ICJ officials, discounting perceptions or knowledge acquired in other pursuits. The survey was distributed among 13 individuals who have worked in the ICJ's Department of Legal Matters, all of whom have had direct experience of the deliberation process. Some judges and former judges were asked to take the survey, but declined based on their duty of confidentiality. The sample size is admittedly not as large as one could hope, but it is significant given that approximately 25 individuals work in the ICJ's Department of Legal Matters at any one time, and that the ICJ started recruiting one Associate Legal Officer per judge only in the mid/late-noughties. The sample of 13 respondents covers a time period from 2010 to date, which should ensure responses based on experience obtained from work done on a variety of ICJ cases.
III. Deliberation Process at the International Court of Justice
If the ICJ's deliberation process is not well known,Footnote 28 the literature on it at least indicates little engagement with it.Footnote 29 ICJ deliberations are governed by the Resolution concerning the Internal Judicial Practice of the Court (Resolution), last reviewed in 1976, which applies to deliberations in both contentious and advisory proceedings.Footnote 30
Under Article 1 of the Resolution, deliberations may begin after the written proceedings close, but before the hearings. Judges may convene to “exchange views concerning the case, and bring to the notice of the Court any point in regard to which they consider it may be necessary to call for explanations during the course of the oral proceedings”.Footnote 31 Article 1 deliberations have played a modest role in the Court's practice.Footnote 32
A key step between the opening and closure of the oral proceedings is the preparation by the President of a list of issues for decision.Footnote 33 This list takes the form of neutral questions, arising from the parties’ arguments,Footnote 34 and is distributed to all judges before the next steps in the deliberations. The list of issues is central to the phase following the closure of the oral hearings. Article 3(i) of the Resolution states that, after the expiration of “an appropriate period … allowed to the judges in order that they may study the arguments presented to the Court”,Footnote 35 the Court holds a meeting where “the President outlines the issues which in [their] opinion will require discussion and decision by the Court”. This meeting takes place shortly after the oral proceedings. At this meeting, judges mainly discuss amendments to the list of issues, as Article 3(i) suggests.
Over a few weeks following the Article 3 deliberation, each judge individually writes a note on how they would decide the case.Footnote 36 This exercise provides judges with a first chance to set out their views in full. Judge Jennings wrote that “notes tend now to be considerable studies of the issues and the law involved in the entire case”.Footnote 37 Nonetheless, much depends on each judge's writing style. While some may write crisply and only address the issues for decision, others may prepare lengthier notes repeating the parties’ submissions. Whichever the approach, judges use written notes also to persuade their colleagues of their views. Judges thus may put considerable effort into preparing convincing pieces of legal writing, but their views are not final until the final vote on the text of the decision.
The written notes are deposited with the Registrar, translated into the other official language and distributed to all judges.Footnote 38 Under Article 5 of the Resolution, “[a]fter the judges have had an opportunity to examine the written notes, a further deliberation is held, in the course of which all the judges … must declare their views”.Footnote 39 Article 5 indicates that all judges must take the floor, which emphasises collegiality. The Article 5 deliberation finishes with judges electing two from among those in the majority to a committee tasked with writing the Court's decision, although some informal discussions in this regard may have taken place in advance.Footnote 40 The President sits on the drafting committee ex officio, unless they are not in the majority or are not presiding.Footnote 41 Electing the drafting committee is crucial, as judgments differ depending on the drafters’ disposition and expertise.Footnote 42 Although, in a drafting committee, the President should be primarily a tie-breaker,Footnote 43 they may greatly influence the drafting process.Footnote 44
Working on the basis of a provisional table of contents prepared by the Registry alongside the President's Office, the judges on the drafting committee jointly write a preliminary draft judgment or advisory opinion in one of the Court's official languages, assisted by their law clerks,Footnote 45 circulated to all judges after translation into the other official language. These judges make written suggestions to amend the preliminary draft, such as to expand the Court's reasoning or reduce it better to reflect the majority's position. The drafting committee considers the written amendments in preparing the draft for the first reading, which all judges examine, paragraph by paragraph, at a plenary meeting often lasting a few days.Footnote 46 While that draft is live-edited during the first reading, structural and far-reaching amendments are for the drafting committee to consider and eventually implement in preparing the draft for the second and last reading.Footnote 47 Judges wishing to append “separate or dissenting opinions” make the text of those opinions available after the first reading.Footnote 48 The second reading follows the same format as the first and closes with all judges voting on the operative paragraphs.Footnote 49 Judges finalise their individual opinions after the second reading.Footnote 50
Individuality is a key feature of the initial stages of the deliberation process, emerging especially from the agenda-setting role of the President in preparing the list of issues, the preparation of the judges’ written notes and the presentation of a provisional table of contents by the Registry and President. With the drafting committee's election, deliberations acquire a collegial dimension, which becomes more marked as the entire Court gathers at the first and second readings.
IV. Custom Identification in the Context of Judicial Deliberations
ICJ deliberations unfold from an individual-driven process to one where the Court reaches its decision collegially. Each stage of the deliberations affects the choice among methodologies by which to record the custom identification process: although the initial, individual-driven stages favour showing custom identification as an inductive exercise, the increasingly collegial character of the later stages promotes showing that process as non-inductive.
A. Action by President, Judges and Registry: Individual-driven Custom Identification
The initial stages of the ICJ's deliberations are characterised by individual action by the President, other judges and Registry officials. At least formally, the Court adheres to custom identification as an inquiry into evidence of state practice and opinio juris. The President, individual judges and Registry officials heed this formal adherence to custom identification by induction not only because they might see induction as the proper methodology to establish the existence of customary rules, but also because of the persuasive weight of earlier judicial decisions implementing that method.Footnote 51 In disputes where decisions depend on the existence of customary rules, their approach to custom identification would likely reflect an inductive exercise.
This view finds support in one's consideration of the potential impact of the parties’ pleadings on the initial stages of the ICJ's deliberations. In their arguments, states aim to satisfy the Court's expectation of how a case should be pleaded,Footnote 52 which can include framing their submissions on custom identification around the Court's established position that the existence of customary rules is established by induction from evidence of state practice and opinio juris. The President's list of issues largely follows the structure and content of the parties’ submissions, considering that it will contain references to relevant parts of the parties’ pleadings.Footnote 53 The Article 5 deliberation notes, based on the President's list of issues, also conceivably mirror how a case was argued. Similarly, the Registry and President prepare the provisional table of contents, which guides the drafting committee's work, based on the parties’ arguments and the President's list of issues. Two cases indicate that the parties’ submissions can determine the Court's choice of induction as the framework to record its custom identification exercise. In North Sea Continental Shelf, the rule of the 1958 Geneva Convention on the Continental ShelfFootnote 54 requiring delimitation by equidistance was binding on Denmark and the Netherlands but not on Germany. Addressing whether continental shelf delimitation by equidistance was required under customary international law, the Court considered the evidence of state practice and opinio juris submitted by the parties and found that there was no such a requirement.Footnote 55 In Jurisdictional Immunities of the State, the Court had to establish whether there were customary exceptions to state immunity justifying the Italian courts’ decisions to deny Germany's jurisdictional immunity. The ICJ analysed state practice and opinio juris, by which it established by inductive reasoning that the territorial tort principle was not a customary exception to Germany's immunity.Footnote 56 In both cases, the parties presented evidence of state practice and opinio juris to show, on an inductive basis, the existence or inexistence of customary international law.Footnote 57 The Court, likely influenced by the individual action by the President, other judges and Registry officials, recorded its custom identification exercise as an inductive approach.
Other decisions turning on the existence of customary rules suggest that early individual action in deliberations is not always a guarantee that decisions reflect custom identification as an inductive exercise. The ICJ's advisory opinion in Chagos depended on whether the right to self-determination was part of customary international law, but the Court, instead of ascertaining the existence of that customary rule by induction, relied on the wording and voting record of certain General Assembly resolutions.Footnote 58 The ICJ adopted a comparable approach in Nicaragua.
One could explain the difference between North Sea Continental Shelf and Jurisdictional Immunities of the State on one hand, and Chagos and Nicaragua on the other hand, by reference to how the parties have framed their submissions on custom identification.Footnote 59 In North Sea Continental Shelf and Jurisdictional Immunities of the State, all parties had used induction as the analytical framework of their submissions on custom identification. Conversely, in Chagos the participant states had taken a variety of approaches to identifying self-determination as a customary right, including induction, deduction and reliance on General Assembly resolutions.Footnote 60 In Nicaragua, because the US refused to appear at the merits, the Court could only rely on Nicaragua's submission, which either assumed the existence of the relevant customary rules, or argued for their existence by reference to judicial decisions or scholarly writings.Footnote 61 The parties’ framing of custom identification in turn influences individual action in the deliberation process,Footnote 62 but the extent to which the pleadings influence that individual action depends also on the approach of the President, judges and Registry.Footnote 63 Nothing prevented the Court, in Chagos or Nicaragua, from establishing the existence of the relevant customary rules by induction. Even if states had not provided sufficient evidence of state practice or opinio juris, the Court could have sought out that evidence for itself.Footnote 64 In Jurisdictional Immunities of the State, the Court relied on domestic judicial decisions not cited by the parties as evidence of state practice.Footnote 65 While pleadings influence actions by individuals in the deliberation process, they do not necessarily determine its course.
The Court's different approaches in Chagos and Nicaragua on one hand, and in North Sea Continental Shelf and Jurisdictional Immunities of the State on the other hand, could also be attributed to the analytical approach to custom identification of the individuals operating in the initial stages of the deliberation process. Factors which could affect these choices may be difficult precisely to identify, and one should not generalise. Possibly, the more familiar these actors are with the content and evidence of given customary rules the identification of which the parties dispute, the more likely it seems that the Court would identify such rules by induction.Footnote 66 Furthermore, whichever methodology to record custom identification these individual actors were to suggest, other judges might be unlikely to disagree with the chosen approach on a methodological level, provided that the chosen approach results in the desired outcome.Footnote 67
B. Drafting Committee Stage: Between Individuality and Collegiality
The approach to custom identification of the drafting committee's members can depend on their personal and professional backgrounds. Although one should not infer too much from the mere requirements under Article 2 of the Court's Statute, in practice ICJ judges are selected from distinct backgrounds. Judges typically come from legal academia, diplomacy, national judiciaries, and governments and international organisations.Footnote 68 Currently, these four categories appear equally represented on the Court's bench.Footnote 69 The extent of a judge's prior experience in international law varies; their professional background before joining the bench plausibly influences their familiarity with custom, including identification methodologies.Footnote 70 This influence is exacerbated by the processes for selecting international judges, to which political negotiation is not foreign,Footnote 71 sometimes leading to the election of persons lacking appreciable acquaintance with custom identification.
National judges are rarely well-versed in international law, let alone the intricacies of custom identification.Footnote 72 Diplomats and government officials may have a better grasp while being sensitive to how state behaviour manifests itself, depending both on the breadth and depth of their legal training, and on their career focus before joining the Court. Although judges chosen from among legal academics should have a robust understanding of custom identification methodologies, merely having worked in international legal academia does not guarantee that understanding. Moreover, while judges trained in civil law systems might favour terser reasoning, judges trained in the common law tradition are more familiar with discursive judgments engaging with legal arguments in scrupulous detail. One might also expect judges who have been practising advocates to pay special attention to the parties’ arguments.
The individual inclinations and style of the drafting committee's judges will determine what the preliminary draft says and does not say on custom identification. It is more likely than not that non-inductive methodologies provide judges who are not familiar with custom identification with a sense of security, because identifying custom by induction may appear more cumbersome and require more extensive research and greater understanding of both the custom-formation process and the subject matter of a putative customary rule.Footnote 73 These perceived difficulties could explain why judges who have limited familiarity with custom identification may find it less attractive to identify customary rules by induction in the judgments they are called upon to draft.
The presence of three judges on the drafting committee could compensate for the less extensive acquaintance of certain members with custom identification methodologies. However, working in a three-member committee is a collegial endeavour to prepare an agreed text. The President sits on drafting committees without actually writing, as the two judges elected by their peers split the exercise between them. Although each judge is responsible only for part of the writing, the drafting committee produces a draft decision circulated as the work of the entire committee. In cases raising issues of custom identification, the drafting committee's collegial working methods entail discussions on the Court's choice of how to record the results of custom identification exercises. Unless its three members agree on recording custom identification as an inductive process, the drafting committee is likely to prepare a draft decision in which the custom identification is recorded as a compromise between the members’ positions, which might be custom identification by induction.Footnote 74 Further compromises on the methodology to identify customary rules made in the final stages of the deliberation process likely result in recording custom identification as a non-inductive exercise, as stated below.
C. Amendments, First Reading and Second Reading: Collegial Custom Identification
In the final stages of the deliberations, starting with the circulation of the preliminary draft, the Court acts collegially. Because judges must agree on a text for the Court's decision,Footnote 75 collegiality promotes minimalistic reasoning and the use of argumentative shortcuts, which results in showing custom identification as a non-inductive process.Footnote 76
1. Minimalistic reasoning
The Court is known for saying nothing more, and sometimes even less, than necessary to decide cases.Footnote 77 Commentators have perceived certain decisions to be “oracular” and “cryptic”.Footnote 78 This criticism extends to the Court's decisions on custom identification and is usually framed as a reproach for not showing the reasoning to establish the existence of customary rules as induction based on evidence of state practice and opinio juris. For example, in Nicaragua v Colombia a question concerned whether the definition of “continental shelf” under Article 76(1) of the 1982 United Nations Convention on the Law of the Sea (UNCLOS)Footnote 79 was part of customary international law. Because Colombia was not a party to UNCLOS, that definition could have been part of the applicable law only as custom. In its 2012 judgment, the Court noted that the parties had agreed that Article 76(1) reflected a customary norm, based on which it found that the definition of the continental shelf was part of customary international law.Footnote 80 This finding was not a foregone conclusion.Footnote 81
Critics of the ICJ's terse reasoning rarely enquire into the reasons underlying its perceived failures to examine evidence of state practice and opinio juris, the assumption being that the Court has not undertaken that examination unless it appears in its decisions. This assumption is misleading, because the “concern for agreement inherent in a collective drafting process”Footnote 82 pushes judges, in the later stages of the deliberations, to identify a common position that maximises the number of supportive votes.Footnote 83 Identifying this common position can start even before the deliberations enter their final, collegial stage. Already in their notes for the Article 5 deliberation, judges address how the Court should identify relevant customary rules. Preparing the Article 5 deliberation note allows each judge the freedom to frame custom identification as each sees fit. These written notes are the basis to start identifying the potential approaches to custom identification capable of being approved by the largest possible majority. The variety of approaches emerging from the Article 5 deliberation notes could also affect the writing of the preliminary draft by the drafting committee. A preliminary draft changes considerably in the deliberations.Footnote 84 However, how much it actually changes varies depending on, for example, how receptive judges on a drafting committee are to their colleagues’ suggestions and the President's more or less laissez-faire approach. Even assuming limited changes to the preliminary and subsequent drafts, the necessity for the Court collegially to approve a single text requires judges to compromise on how certain points are formulated in the relevant decision. Such points are likely those on which a case turns, which may include how to establish whether certain rules of customary international law exist.
One of the central questions in the 2019 Chagos advisory opinion was whether, at the time of Mauritius's independence, the Mauritian people had a right to self-determination under customary international law.Footnote 85 Although the ICJ was nearly unanimous in its affirmative reply,Footnote 86 the participant states adopted different lines of argument to show that the right to self-determination was part of customary international law at the relevant time. The US sought to establish that there was no customary right to self-determination based on lack of evidence of state practice and opinio juris,Footnote 87 while the UK emphasised practice inconsistent with the status of self-determination as a customary right.Footnote 88 Some states relied on the passing of General Assembly resolution 1514(XV), without additional evidence of state practice or opinio juris.Footnote 89 Other states pointed to numerous territories achieving independence as examples of practice and used resolution 1514(XV) as evidence of opinio juris,Footnote 90 or criticised the relevance of the practice on which the UK relied.Footnote 91 Brazil simply asserted that self-determination existed as a customary right.Footnote 92 Similar to states, judges are also likely to have favoured different approaches to identifying self-determination as a customary right in their Article 5 deliberation notes.Footnote 93 Although the drafting committee would have endeavoured to find a compromise between the different approaches to custom identification suggested in the advisory proceedings, other judges would have scrutinised the approach chosen by the drafting committee during the first and second readings of the draft advisory opinion, also with a view to finding a text agreeable to the largest possible majority. The result was not to record custom identification as an analysis of state practice and opinio juris, but by reference to General Assembly resolutions.
Distinct from Nicaragua v Colombia, Chagos is not even a case where the Court identified customary international law by simple assertion. If the Court asserts that a rule is part of customary international law or simply does not show that it ascertained its existence by inductive reasoning, it is more likely than not because, in the first and second reading of the draft judgment or advisory opinion, judges could not agree on induction being the most suitable approach to identify that rule.Footnote 94 While one may, and perhaps should, criticise the ICJ's terseness in recording its custom identification exercises, this criticism should recognise how collegiality in the final stages of the Court's deliberations can impact the text of the Court's judgments and advisory opinions.
2. Argumentative shortcuts
Argumentative shortcuts are drafting devices allowing the ICJ to reduce its reasoning to what is strictly necessary to decide cases. Although the drafting committee's members might use these shortcuts to pre-empt disagreements in the first and second readings, use of these shortcuts also results from the collegial character of the final stages of the Court's deliberation process.Footnote 95 During deliberations, the ICJ often deals with issues which could become sticking points in the drafting process. To avoid becoming entrenched in their positions, which may itself reflect on the quality of their common endeavour, judges use argumentative shortcuts to circumvent the cause of their disagreement. Argumentative shortcuts for custom identification include: reference to the ILC's work;Footnote 96 citing earlier judicial decisions;Footnote 97 relying on the resolutions of international organisations;Footnote 98 listing established treaty rules;Footnote 99 and accepting the parties’ agreement that rules are customary in character.Footnote 100
Argumentative shortcuts are drafting strategies which could affect judicial drafting both on a macro-level, by influencing the Court's overall approach to deciding a case where the outcome depends on custom identification, and on a micro-level, where using them impacts only a narrower aspect of the Court's decision. An example of the former is the Chagos advisory opinion, which turned on the existence of a customary right to self-determination identified based on a number of General Assembly resolutions.Footnote 101 An example of the latter is Nicaragua v Colombia, in which the parties’ agreement was the basis for the Court to establish that Article 76(1) of UNCLOS was part of customary international law, which was anyway unnecessary owing to Nicaragua's failure to prove that its continental margin overlapped with Colombia's 200-nautical-mile continental shelf entitlement.Footnote 102
Although deliberations may explain the Court's economy of expression, they do not necessarily excuse it.Footnote 103 Argumentative shortcuts can result from collegiality in the final stages of the deliberation process, but could determine the perception that the Court's approach to recording the custom identification process lacks transparency and cogency. This perception could adversely affect the legitimacy of the Court's decisions.Footnote 104 Although scholars have perceived argumentative shortcuts mainly to have a negative impact on the ICJ's role as a custom identification agency, this perception is not entirely justified. Whether argumentative shortcuts are sound means of custom identification depends on the circumstances in which the Court uses them.
Relying on the ILC's work could be problematic because its function is not only to codify, but also progressively to develop, international law.Footnote 105 The ILC often states whether it regards certain proposals as progressive development, based on which the ICJ can assess whether referring to the ILC's work could be a sound shortcut to identify custom.Footnote 106 Citing earlier judicial decisions in which the Court or other international tribunals have established the existence of a customary rule could also be a reliable means of custom identification, provided that the earlier identification exercise is grounded in convincing, preferably inductive, reasoning. Using these argumentative shortcuts does not inevitably warrant the perception that custom identification by the ICJ lacks cogency, as the Court could be relying on an earlier, methodologically sound, custom identification exercise.
Relying on the resolutions of international organisations and accepting the parties’ agreement that a certain rule has customary status is problematic. Resolutions of international organisations could be evidence of opinio juris; they could also be evidence of state practice, which could take the form of conduct in connection with the adoption of such resolutions. The ILC's Commentary to Draft Conclusion 3(2) states that “acts forming the relevant practice are not as such evidence of acceptance as law”Footnote 107 because, in principle, “the existence of one element may not be deduced merely from the existence of the other”.Footnote 108 However, Draft Conclusion 3(2) “does not exclude that the same material may be used to ascertain practice and acceptance as law”,Footnote 109 although it “must be examined as part of two distinct inquiries”.Footnote 110 Where the Court relies on resolutions of international organisations as evidence of both state practice and opinio juris, it should use them as part of separate inquiries into evidence of state practice and evidence of opinio juris. In Chagos, the Court could have shown that it had established the existence of self-determination as a customary right by inductive reasoning by adding an assessment of relevant state practice to its consideration of opinio juris emerging from General Assembly resolutions. Difficulties would remain in Nicaragua-like cases, in which state practice is not sufficiently consistent to justify finding that a customary rule exists. In such cases, the ICJ's choice could be between declaring that a putative customary rule does not exist and declaring that it does, but, in the latter case, be seen to exercise a lawmaking function.Footnote 111 An additional problem is the one of “cheap talk”, as states may make statements in multilateral forums without necessarily intending to align their practice with those statements.Footnote 112
By relying on the parties’ agreement that a rule has customary status, the Court may seem to concede the generalisability of its legal reasoning and refrain from actively developing international law.Footnote 113 The parties’ agreement used as an argumentative shortcut for custom identification seems to achieve the exact opposite. Even though the ICJ's judgments only bind the states partiesFootnote 114 and advisory opinions are not binding,Footnote 115 the Court's decisions are regarded as authoritative statements of what international law is. By relying on the parties’ agreement in establishing the customary status of a rule, the Court pronounces on the existence of a rule which, once identified, is at least perceived to be binding on all states. The parties’ agreement is an unsound argumentative shortcut for custom identification, best avoided especially where, such as Nicaragua v Colombia, there is no need to pronounce on the existence of a customary rule.
V. Assessment and Implications
There are two bases helpful to verify doctrinally the correlation between the choice of methodology to record custom identification and the degree to which the stages of deliberations are individual-driven or collegial: first, how judges have shown to identify customary rules in their individual opinions; second, how international tribunals that rely on drafting by judge rapporteurs record the results of their custom identification exercises. This correlation has implications for the legitimacy of judicial custom identification.
A. Induction as Dissent
As individual works, the judges’ opinions can help one verify the existence of a correlation between, on one hand, individual-driven stages of the deliberation process and, on the other hand, recording custom identification as an inductive exercise. In their individual opinions, judges have identified customary international law using various non-inductive methodologies. Nevertheless, individual opinions show that whether judges writing separately record custom identification as an inductive exercise depends on whether they agree with the Court's decision.
Individual opinions show that judges identify customary international law by non-inductive methodologies, mostly assertion, when they agree with the result of the ICJ's custom identification exercise, irrespective of the Court's chosen methodology. In Nicaragua, the Court found that the principle of non-intervention was part of customary international law based on evidence of opinio juris emerging from certain General Assembly resolutions.Footnote 116 Writing separately, Judges Nagendra Singh and Sette-Camara simply asserted that non-intervention was a customary obligation.Footnote 117 In Djibouti v France, the Court relied on the existing customary rule on immunity of diplomatic agents to ascertain, by deductive reasoning, the existence of a putative customary rule on Head of State immunity.Footnote 118 Judge ad hoc Yusuf only asserted that “[t]he customary validity of the immunity of a Head of State abroad has been recognized by several conventions”.Footnote 119 The Court asserted the customary status of the definition of “continental shelf” as codified in Article 76(1) of UNCLOS.Footnote 120 Judge Abraham and Judge ad hoc Mensah agreed with the Court's finding simply by making the same assertion as the Court.Footnote 121 Judges also assert custom when agreeing with the result of the Court's inductive inquiry into state practice and opinio juris, as Judge Koroma did for the territorial tort exception in his separate opinion in Jurisdictional Immunities of the State.Footnote 122
Judges writing separately seldom identify customary international law by induction while also agreeing with the outcome of the Court's custom identification exercise. The only two examples of this use of induction in individual opinions are Right of Passage over Indian Territory and Chagos. In Right of Passage over Indian Territory, Portugal had requested the Court to find that it had a right of passage over the territory of India to transport goods, persons and troops between Portuguese enclaves in India.Footnote 123 Showing its reasoning as inductive, the Court ascertained the existence of a local customary rule conferring the right claimed by Portugal, except for the passage of armed personnel.Footnote 124 In his separate opinion, Judge Wellington Koo similarly examined evidence of state practice and opinio juris since the time of the British Raj, reaching the same conclusion as the Court.Footnote 125 In Chagos, Judges Sebutinde and Robinson agreed with the Court's finding that the people of Mauritius had, at the relevant time, a customary right to self-determination which included the UK's obligation to preserve Mauritius's territorial integrity.Footnote 126 Nonetheless, in their individual opinions they examined evidence of state practice and opinio juris to establish the existence of that customary right by inductive reasoning.Footnote 127 Judges who use induction to frame their approach to custom identification while also agreeing with the Court's findings on the existence of certain customary rules may intend to provide the Court's reasoning with further support.Footnote 128 Judges likely have this purpose in Chagos-like cases, in which the Court's reasoning does not show custom identification as an inductive exercise. These judges may appear to achieve the exact opposite, as their inductive analysis can suggest that the Court failed to identify the relevant customary rules based on evidence of state practice and opinio juris.
Judges identify customary international law by induction in two scenarios. First, judges always use induction when dissenting from the result of the ICJ's identification exercise. Judges Moreno Quintana and Sir Percy Spender dissented from the Court's judgment in Right of Passage over Indian Territory; in their individual opinions, they respectively sought to show, based on evidence of state practice and opinio juris, that there was no right of passage as claimed by Portugal and that its right of passage extended to armed personnel.Footnote 129 In Arrest Warrant, Judge ad hoc Van den Wyngaert dissented from the Court's finding that Foreign Ministers enjoyed immunity under customary international law, listing instances of state practice and opinio juris inconsistent with the Court's finding.Footnote 130 Judges have also dissented from the result of the Court's identification exercise without setting out specific instances of state practice and opinio juris, but by reproaching the Court for not having identified the relevant customary rules by induction. This was the case for Judge Sepúlveda-Amor in Navigational and Related Rights, in relation to Costa Rica's right to subsistence fishing along the San Juan River,Footnote 131 and Judge Donoghue in Certain Activities, in relation to the obligation to conduct an environmental impact assessment where domestic activities risk causing significant transboundary harm.Footnote 132
Second, judges identify customary international law by induction when analysing the customary status of rules which the Court found it unnecessary to discuss in its decisions. In Arrest Warrant, the Court did not discuss whether Belgium had based its assertion of universal jurisdiction on customary international law, because it could dispose of the case on grounds of immunity.Footnote 133 Judges Guillaume, Higgins, Kooijmans and Buergenthal discussed this aspect in their individual opinions by reference to evidence of state practice and opinio juris.Footnote 134 In Belgium v Senegal, the Court had found that there was no dispute on the alleged breach of Senegal's obligation to exercise universal jurisdiction under customary international law.Footnote 135 Judge Abraham wrote that there would have been insufficient evidence of state practice and opinio juris to find that the exercise of universal jurisdiction was a customary obligation, although without listing evidence of either element.Footnote 136
It should come as no surprise that individual judges agreeing with the result of the Court's custom identification exercises, however conducted, tend not to record their custom identification exercise by induction. These judges do not have strong practical reasons to list evidence of state practice and opinio juris only to reach the same conclusion as that contained in the judgment or advisory opinion, apart from providing the Court's reasoning with further support. Significantly, induction is the methodology invariably used by individual judges disagreeing with the result of the Court's custom identification exercise. As their disagreement on custom identification should be grounded in convincing legal reasoning, dissenting judges have an incentive to identify customary international law by induction, as the most rigorous methodology available. The consistent use of induction in the opinions of judges dissenting on custom identification suggests that there is a correlation between the choice of induction as a custom identification methodology and drafting by judges acting individually. This correlation supports the view that, in the Court's deliberation process, custom identification is recorded as an inductive exercise in the initial, individual-driven stages of that process, and that, conversely, collegiality in the final stages of the deliberation process can explain the Court's failure to show custom identification by reference to evidence of state practice and opinio juris.
B. Induction on a Sliding Scale
There is a correlation between individual action in the ICJ's deliberations and recording custom identification as an inductive process. As individual influence gives way to collegiality, the likelihood that the Court identifies custom by induction decreases. There appears to be a sliding scale of likelihood of induction, linked to the degree of collegiality in different stages of the ICJ's deliberations.
The deliberation processes of other international tribunals that also may be called upon to identify customary international law show the same tension between their individual-driven and collegial stages, but the balance between individuality and collegiality does not always mirror that of the ICJ's deliberations. ITLOS and the WTO Appellate Body have deliberation processes following the ICJ's model, which also replicate the balance between individual-driven and collegial stages of the Court's deliberations.Footnote 137 Decisions of the European and Inter-American Courts of Human Rights (ECtHR and IACtHR) are drafted by judge rapporteurs and receive more limited collegial scrutiny than at the ICJ,Footnote 138 also owing to the heavier docket of these two courts.Footnote 139 The Court of Justice of the European Union (CJEU), which also has a heavy docket,Footnote 140 similarly relies on drafting by judge rapporteurs, both at the Court of Justice and at the General Court.Footnote 141 Moreover, at the Court of Justice each case raising novel points of law is assigned to an Advocate General, who delivers an opinion after the oral pleadings.Footnote 142 The deliberation process of the ECtHR, IACtHR and CJEU is mainly driven by individuals, such as judge rapporteurs and Advocates General, which limits the impact of collegiality on judicial decisions.Footnote 143
Because deliberations at ITLOS and the Appellate Body promote collegiality in decision-making similar to the ICJ, the likelihood that these tribunals show custom identification as an inductive exercise should also be on a sliding scale comparable to that described for the Court. Showing custom identification as induction at ITLOS and the Appellate Body thus should be unlikely, as confirmed by their jurisprudence. ITLOS has only been shown to assert the existence of custom, especially for rules found in UNCLOS.Footnote 144 Where it has previously asserted the existence of a customary rule, ITLOS has relied on that assertion in subsequent decisions,Footnote 145 but, the first time that ITLOS had to identify customary rules not codified by UNCLOS, it did so by reference to their prior identification by the ICJ.Footnote 146 The Appellate Body rarely shows to identify customary rules. Since 1996, the Appellate Body has identified only few such rules: the general rule of treaty interpretation,Footnote 147 the principle of good faith,Footnote 148 proportionality as a condition for lawful countermeasuresFootnote 149 and the attribution to states of their organs’ conduct.Footnote 150 The Appellate Body has never recorded custom identification by induction, but has always shown to assert the customary status of rules either by reference to external sources, such as decisions of other international tribunals and scholarly writings, or by justifying identifying a rule which it had already found to have customary status by reference to its own earlier decisions.
The stronger individual element, and correspondingly more limited role of collegiality, in the deliberations of the ECtHR, IACtHR and CJEU should increase the likelihood of recording custom identification by induction. For these judicial organs, the likelihood of showing custom identification as an inductive exercise thus could not be represented by the same sliding scale described in relation to the ICJ. That showing custom identification by induction is more likely in individual-driven deliberations emerges from the jurisprudence of the ECtHR, IACtHR and CJEU. The ECtHR's Chambers have identified few rules of customary international law and have mostly asserted their existence.Footnote 151 However, in Naït-Liman v Switzerland a Chamber showed that it had used inductive reasoning to establish the inexistence of a customary rule permitting the exercise of universal civil jurisdiction for liability arising from acts of torture.Footnote 152 Although mere assertion is its most frequently used manner of recording custom identification,Footnote 153 alongside referring to external sourcesFootnote 154 and its own earlier jurisprudence,Footnote 155 the Grand Chamber has identified custom also by induction. In Naït-Liman v Switzerland, the Grand Chamber used the same inductive approach as the Chamber's judgment in the same case and upheld the Chamber's finding.Footnote 156 In Kononov v Latvia, the Grand Chamber analysed conduct and declarations by states to conclude that individual criminal responsibility for the acts of one's subordinates has customary status.Footnote 157 Custom identification by the IACtHR follows a similar pattern. Beyond recognising the customary duty to make reparation for internationally wrongful acts, codified by Article 63(1) of the 1969 American Convention on Human Rights,Footnote 158 the IACtHR tends to show that it asserts the existence of the relevant customary rules, either without reference to supporting material,Footnote 159 or by referring to external sourcesFootnote 160 or its own jurisprudence.Footnote 161 That jurisprudence includes only one case of custom identification shown as induction. In its Asylum advisory opinion, the IACtHR examined evidence of state practice and opinio juris to find that it was insufficient to ground the identification of regional custom on diplomatic asylum.Footnote 162
At the CJEU, the General Court shows that it identifies customary international law by merely asserting that certain rules have customary status,Footnote 163 often by reference to internal or external jurisprudence;Footnote 164 rarely has it shown to ascertain the existence of customary international law by deductive reasoning.Footnote 165 The Court of Justice also shows merely to assert the existence of customFootnote 166 or to do so by relying on internal or external sources.Footnote 167 The General Court has not recorded custom identification by induction. The Court of Justice has shown to have identified custom mainly by agreeing with the Advocates General's conclusions on custom identification, which has sometimes taken place by induction.Footnote 168 Advocates General have shown that they ascertained the existence of customary international law both by assertionFootnote 169 and by deductive reasoning,Footnote 170 but have also done so by induction, although not always with the requisite degree of rigour.Footnote 171
The jurisprudence of international tribunals other than the ICJ confirms that the less collegial the deliberations, the higher the likelihood of custom identification being recorded as an inductive exercise. ITLOS and the Appellate Body, which use collegial working methods, have never identified custom by induction. The ECtHR, IACtHR and CJEU, where custom identification is the province of judge rapporteurs or Advocates General, have occasionally identified custom by induction. The sheer number of cases of custom identification recorded by induction in the latter group of tribunals might not seem to be sufficiently significant to support the existence of a correlation between individual-driven stages of judicial deliberations and the use of induction in identifying customary international law. Distinct from that of the ICJ, the jurisdiction of these tribunals is limited to ascertaining breaches of rules under their constitutive treaties, which may make it unnecessary to identify customary rules of international law, thus explaining the significance of the existing cases of custom identification recorded by induction.
C. Induction and Legitimacy
The impact of collegial decision-making on how custom identification is recorded in judicial decisions raises legitimacy questions concerning the use of non-inductive methods of custom identification and the transparency of deliberations.Footnote 172
Custom identification by induction is consistent with the ICJ's view of customary international law as the union of state practice and opinio juris. Because induction requires the Court to review evidence of state practice and opinio juris, it allows states, as the relevant constituency,Footnote 173 to assess the strength of the Court's reasoning and, consequently, the acceptability of its decisions against the background of its own definition of customary international law.Footnote 174 Induction also promotes the perception that custom is a democratic source of international law, identified based on the conduct of the majority of states the practice of which is relevant in the custom identification exercise.Footnote 175
Showing custom identification by non-inductive methodologies reduces the breadth and depth of the Court's reasoning, lowering its perceived quality and adversely affecting the output legitimacy of the Court's decisions.Footnote 176 In her 2021 statement to the ILC, President Donoghue compared the reception of the Court's texts, including its decisions, to the situation of a student being given “only partial credit for a correct answer on a mathematics test because of a failure to indicate the steps taken to reach that answer”.Footnote 177 President Donoghue specifically linked her example to the identification of customary international law, stating that “an unsubstantiated conclusion on the existence or content of a rule could easily be criticized and might detract from the overall credibility of the Court”.Footnote 178 This problem applies especially to custom identification shown by simple assertion, which, because it entails little to no engagement with evidence of state practice and opinio juris, can justify the perception that the ICJ exercises its judicial function arbitrarily. Asserting the existence of customary rules by certain argumentative shortcuts may limit this perception, and, in limited cases, might even benefit legitimacy. For instance, the ECtHR has relied on European consensus, presumably as a proxy of evidence of opinio juris, to decide whether a customary rule existed allowing abortion on grounds wider than under Irish law.Footnote 179 Although European consensus could not substitute evidence of opinio juris, the ECtHR's reliance on it could boost the legitimacy of its decisions.Footnote 180
When the ICJ uses argumentative shortcuts, states can assess whether the chosen shortcuts ground the Court's decisions in cogent reasoning. This assessment is case-specific because it depends on the circumstances in which the Court uses specific shortcuts.Footnote 181 Shortcuts thus are not necessarily problematic for the legitimacy of custom identification by the Court. For example, recording custom identification by reference to the ILC's review of evidence of state practice and opinio juris would not adversely affect legitimacy. For the ICJ, the only exception is its occasional reliance on the parties’ agreement. That reliance results in generalising the positions of specific states into customary rules that one may perceive as applicable to all states. This generalisation entails not evaluating the practice of the majority of states, which negatively impacts the perceived democratic legitimacy of the relevant customary rules. For other international tribunals, relying on the ICJ's jurisprudence when first identifying a new customary rule may be problematic. This reliance appears to be an informal preliminary reference to the ICJ in matters of custom identification.Footnote 182 This mechanism could promote uniformity in custom identification,Footnote 183 but would entail that international tribunals other than the ICJ would outsource the exercise of a limited, but potentially significant, part of their judicial function to the Court, especially if their constitutive instruments include custom as part of the applicable law.
Although non-inductive methodologies are not necessarily problematic for legitimacy, the processes resulting in their adoption lack transparency. Confidentiality of deliberations is integral to judicial ethics,Footnote 184 which benefits legitimacy, but also prevents the relevant constituency from assessing why the Court has identified custom by non-inductive processes, which can negatively impact legitimacy. Despite this tension, the protection of the judges’ independence and impartially are reasons to maintain the secrecy of deliberations, also considering that ICJ judges are entitled to append individual opinions in which they elaborate on the reasons that have led the Court to choose a particular approach to record its custom identification exercise. Where the Court's decisions do not show custom identification by induction, not disclosing whether inductive processes have taken place in deliberations may be detrimental to the legitimacy of the Court's exercise of its judicial function. Yet, legitimacy assessments should recognise the possibility that inductive processes have simply remained secret. Although this consideration may not offset entirely the legitimacy deficit determined by not expressing inductive processes publicly in judicial decisions, it may give insights into the reasons behind the Court's perceived failure to apply a rigorous custom identification methodology. Furthermore, distinct from deliberations, evidence of state practice and opinio juris is public, which affords a basis for examining whether custom identification recorded as a non-inductive exercise resulted in sound decisions, although the public justification of those decisions remains methodologically unpersuasive.
VI. Wider Significance
How the ICJ records the results of its custom identification exercises is affected by the shape of the Court's deliberation process and the emphasis on collegiality in its later stages. There is a higher likelihood of showing custom identification as induction in individual-driven stages of judicial deliberations, which finds support in the use of induction in the individual opinions of ICJ judges and in the decisions of international tribunals which adopt a less collegial deliberation model. In its 2018 Draft Conclusions, the ILC only clarified the secondary rules governing custom identification, without touching on the impact that deliberation processes may have on the choice among methodologies to record custom identification in judicial decisions. The conclusions of this work thus complement the ILC's Draft Conclusions, by providing insight into why, despite adopting the two-element theory of customary international law, the ICJ does not always identify customary rules by assessing evidence of state practice and opinio juris.
For their part, commentators have written on custom identification by international tribunals assuming that such a process is carried out by decision-makers operating free from any procedural constraint. However, the reality of international tribunals is far from this assumption underlying the literature on custom identification. Although the secrecy of deliberations may seem effectively to exempt the ICJ's approach to custom identification from scrutiny, evidence of state practice and opinio juris remains publicly available. While custom identification exercises are secret, evidence of custom's constitutive elements is not. The public availability of that evidence could benefit the legitimacy of custom identification by the Court, partly offsetting the legitimacy problems connected with the secrecy of deliberations.
Further study could examine the content of the customary rules which international tribunals other than the ICJ identify, how the relationship between specialised treaty régimes and general international law affects the choice of those tribunals among custom identification methodologies, and how those tribunals explain why and through which means they identify custom despite their function being that of interpreting and applying treaty provisions. Moreover, questions arise as to the impact of deliberation processes on other sources of international law, such as the identification of general principles and treaty interpretation. One may also ask the evaluative question as to whether, when showing that they identify custom by non-inductive methodologies, the ICJ and other international tribunals “get it right”. It remains that methodology is an important means for judicial decisions to be seen as grounded in reasoned, public justification. Whether international tribunals record their custom identification exercises as inductive or not, the cogency of their reasoning will depend not only on “getting it right”, but also on the soundness of the chosen methodology.
