2 The Sources of International Law
International law provides a normative framework for the conduct of interstate relations. In this sense, international society is no exception to the maxim of ibi societas, ibi ius: where there is social structure, there is law. The sources of international law define the rules of the system: if a candidate rule is attested by one or more of the recognized ‘sources’ of international law, then it may be accepted as part of international law. Simultaneously, the diffuse character of the sources highlights the decentralization of international law-making.
The formally recognized sources of international law are reflected in Article 38 of the Statute of the International Court of Justice.1 These sources are oft en presented—as in Article 38—as separate, but they influence each other in practice.
It is common for writers to differentiate between formal and material sources of law. Formal sources are those methods for the creation of rules of general application which are legally binding on their addressees. The material sources provide evidence of the existence of rules which, when established, are binding and of general application. In the context of international relations, however, the use of the term ‘formal source’ is misleading since it conjures up notions associated with the constitutional machinery of law-making within states. No such machinery exists for the creation of international law. Decisions of the International Court, unanimously supported resolutions of the General Assembly concerning matters of law, and important multilateral treaties seeking to codify or develop rules of international law are all significant to varying degrees. Nonetheless they are not binding on states generally. In this sense ‘formal sources’ hardly exist in international law. As a substitute, and perhaps as a ‘constitutional’ equivalent to formal sources, international law works on the basis that the general consent or acceptance of states can create rules of general application. The definition of custom in international law is essentially a statement of this principle, and not a reference to ancient custom as in English law.
(p. 21) In international law the distinction between formal and material sources is consequently difficult to maintain. The former reduces to a quasi-constitutional principle of inevitable but unhelpful generality. What matters more is the variety of material sources. These are the all-important evidence of a normative consensus among states and other relevant actors concerning particular rules or practices. Decisions of the International Court, resolutions of the General Assembly, and ‘law-making’ multilateral treaties are evidence of the attitude of these actors toward particular rules and of the presence or absence of consensus. Moreover, there is a process of interaction which gives these a status somewhat higher than other ‘material sources’. Neither an unratified treaty nor a report of the International Law Commission (ILC) to the General Assembly has any binding force as a matter of treaty law or otherwise. However, such documents stand as candidates for public reaction, approving or not as the case may be. They may approach a threshold of consensus and confront states which wish to oppose their being given normative force in a significant way.
The law of treaties concerns the content of specific obligations accepted by the parties (states and other persons with treaty-making power), that is, it concerns the incidence of obligations resulting from express agreement. Treaties may be bilateral or multilateral,2 but even if multilateral the obligations they create may run primarily between the two parties concerned—for example, the sending state and the receiving state in the case of diplomatic relations. But even if genuinely multilateral, the constraints of the treaty form still apply: in principle, treaties neither oblige nor benefit third parties without their consent. Thus the incidence of particular conventional obligations is a matter distinct from the sources of general international law, which is made by more diffuse processes. Treaties as such are a source of obligation and not a source of rules of general application. Treaties may however form an important material source in that they may be reflective of, or come to embody, customary international law.3
2. The Statute of the International Court of Justice
Historically the most important attempt to specify the sources of international law was Article 38 of the Statute of the Permanent Court of International Justice,4References(p. 22) taken over nearly verbatim5 as Article 38 of the Statute of the International Court of Justice:
(d) subject to the provisions of Article 59, judicial decisions and the teachings of the most highly qualified publicists of the various nations, as subsidiary means for the determination of rules of law.
2. This provision shall not prejudice the power of the Court to decide a case ex aequo et bono, if the parties agree thereto.
Article 59 provides that decisions ‘have no binding force except between the parties and in respect of that particular case’.
These provisions are expressed in terms of the function of the Court. However they reflect the previous practice of arbitral tribunals, and Article 38 is often put forward as a complete statement of the sources of international law.6 Yet the article makes no reference to ‘sources’ and, on close inspection, cannot be regarded as a straightforward enumeration.
The first question is whether paragraph 1 creates a hierarchy of sources. There is no express hierarchy, but the draftsmen stipulated an order, and in one draft the word ‘successively’ appeared.7 In practice sub-paragraphs (a) and (b) are the most important: we can explain the priority of (a) by the fact that it refers to a source of obligation which will ordinarily prevail as being more specific.8 But it is unwise to think in terms of hierarchy as dictated by the order (a) to (d) in all cases. Source (a) relates to obligations; in some circumstances a treaty does not give rise to a correspondingReferences(p. 23) obligation of a state party, notably when it is contrary to a peremptory norm of international law;9 and in all cases the content of a treaty obligation depends on the interpretation of the treaty, a process governed by international law.10 A treaty may even be displaced by a subsequent rule of customary international law, at least where its effects are recognized in the subsequent conduct of the parties.11
Dating back to 1920, Article 38 has been described, inter alia, as out of date, narrow and ill-adapted to modern international relations.12 But in practice it is malleable enough, and its emphasis on general acceptance is right: customary law is not to be confused with the last emanation of will of the General Assembly.13
3. International Custom14
(A) The Concept of Custom
Article 38 refers to ‘international custom, as evidence of a general practice accepted as law’. The wording is prima facie defective: the existence of a custom is not to be confused with the evidence adduced in its favour; it is the conclusion drawn by someone (a legal adviser, a court, a government, a commentator) as to two related questions: (a) is there a general practice; (b) is it accepted as international law? Judge Read has described customary international law as ‘the generalization of the practice of States’,15 and so it is; but the reasons for making the generalization involve an evaluation of whether the practice is fit to be accepted, and is in truth generally accepted, as law.
Although the terms are sometimes used interchangeably, ‘custom’ and ‘usage’ are terms of art with different meanings. A usage is a general practice which does not reflect a legal obligation: examples include ceremonial salutes at sea and the practice of granting certain parking privileges to diplomatic vehicles.16 Such practices are carried on out of courtesy (or ‘comity’) and are neither articulated nor claimed as legalReferences(p. 24) requirements. International comity is a species of accommodation: it involves neighbourliness, mutual respect, and the friendly waiver of technicalities.17 However, particular rules of comity, maintained consistently without reservation, may develop into rules of customary law.18
The material sources of custom are manifold and include: diplomatic correspondence, policy statements, press releases, the opinions of government legal advisers, official manuals on legal questions (e.g. manuals of military law), executive decisions and practices, orders to military forces (e.g. rules of engagement), comments by governments on ILC drafts and accompanying commentary, legislation, international and national judicial decisions, recitals in treaties and other international instruments (especially when in ‘all states’ form),19 an extensive pattern of treaties in the same terms, the practice of international organs, and resolutions relating to legal questions in UN organs, notably the General Assembly. The value of these sources varies and will depend on the circumstances.
(B) The Elements of Custom
(i) Duration and consistency of practice
The question of uniformity and consistency of practice is very much a matter of appreciation. Complete uniformity of practice is not required, but substantial uniformity is, and for this reason in Anglo-Norwegian Fisheries the Court refused to accept the existence of a 10-mile rule for the closing line of bays.20
Although 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 on the basis of what was originally a purely conventional rule, an indispensable requirement would be that within the period in question, short though it might be, State practice, including that of the States whose interests are specially affected, should have been both extensive and virtually uniform in the sense of the provision invoked;—and should moreover have occurred in such a way as to show a general recognition that a rule of law or legal obligation is involved.22
References(p. 25) This sets a high standard: it was met in the case of some of the rules concerning the continental shelf articulated in the Truman Proclamation, but not the delimitation rule which the ILC had proposed as a matter of convenience and which was not contained in that Proclamation.23
(ii) Generality of practice
Complete consistency is not required; often the real problem is to distinguish mere abstention from protest by a number of states in face of a practice followed by others. Silence may denote either tacit agreement or a simple lack of interest in the issue. It may be that the Permanent Court in the Lotus case misjudged the consequences of absence of protest and the significance of fairly general abstention from prosecutions by states other than the flag state.24 In the event the Geneva Convention on the High Seas adopted the rule which the Court had rejected—a rare example of the overruling by treaty of a decision of the Court on a point of custom.25
In Fisheries Jurisdiction(UK v Iceland) the International Court referred to the extension of a fishery zone up to a 12nm limit ‘which appears now to be generally accepted’ and to ‘an increasing and widespread acceptance of the concept of preferential rights for coastal states’ in a situation of special dependence on coastal fisheries.26But while refusing to ‘render judgment sub specie legis ferendae, or [to] anticipate the law before the legislator has laid it down’,27 the Court did in fact articulate a rule of preferential coastal state rights, a transitional step towards the Exclusive Economic Zone regime which would be included in the United Nations Convention on the Law of the Sea28 (UNCLOS).
(iii) ‘Accepted as law’: opinio iuris sive necessitatis
The Statute of the International Court refers to ‘a general practice accepted as law’. Some writers do not consider this psychological element to be required for custom,29 but something like it must be necessary.30 It is ordinarily expressed in terms of the Latin neologism opinio iuris sive necessitatis, a phrase which has, perhaps regrettably,References(p. 26) become established.31 But the idea of normativity—the articulation of a practice as binding—is not new: as a necessary requirement of a customary rule it goes back to Isidore of Seville (c540–636CE) and beyond.32
Even if the rarity of the judicial decisions to be found among the reported cases were [established]…it would merely show that States had often, in practice, abstained from instituting criminal proceedings, and not that they recognized themselves as being obliged to do so; for only if such abstention were based on their being conscious of having a duty to abstain would it be possible to speak of an international custom. The alleged fact does not allow one to infer that States have been conscious of having such a duty; on the other hand…there are other circumstances calculated to show that the contrary is true.34
Presumably the same principles should apply to both positive conduct and abstention, yet in the Lotus the Court was not ready to accept continuous conduct as evidence of a legal duty and required a high standard of proof of opinio iuris.35
Again in North Sea Continental Shelf Denmark and the Netherlands argued that the equidistance–special circumstances method of delimiting the continental shelf had become accepted as law by the date of the Convention on the Continental Shelf.36 The Court declined to presume the existence of opinio iuris based on the practice as at that date. Nor did it accept that the subsequent practice of states based upon theReferences(p. 27) Convention had produced a customary rule. However, the decision is not incompatible with the view that existing general practice raises a presumption of opinio iuris. Before 1958, there was little practice concerning the equidistance principle apart from the records of the ILC, which revealed the experimental aspect of the principle at that time.37 As to post-1958 practice, the Court’s rejection of the argument rested primarily on two factors: (a) Article 6 was directed at agreement and was not of a norm-creating character;38 (b) the convention having been in force for less than three years, the state practice was inadequate ‘to show a general recognition that a rule of law or legal obligation is involved’.39 But the tenor of the judgment is hostile to the presumption of opinio iuris.40
In considering the instances of the conduct…the Court has to emphasize that, as was observed in the North Sea Continental Shelf cases, for a new customary rule to be formed, not only must the acts concerned ‘amount to a settled practice’, but they must be accompanied by the opinio juris sive necessitatis. Either the States taking such action or other States in a position to react to it, must have behaved so that their conduct is ‘evidence of a belief that this practice is rendered obligatory by the existence of a rule of law requiring it. The need for such a belief, i.e., the existence of a subjective element, is implicit in the very notion of the opinio juris sive necessitatis’.42
The fact…that various international agreements, such as agreements for the promotion and protection of foreign investments and the Washington Convention, have established special legal régimes governing investment protection, or that provisions in this regard are commonly included in contracts entered into directly between States and foreign investors, is not sufficient to show that there has been a change in the customary rules of diplomatic protection; it could equally show the contrary.43
The choice of approach appears to depend on the character of the issues—that is, the state of the law may be a primary point in contention—and on the discretion of the Court. The approach may depend on whether practice is largely treaty-based (in which case opinio iuris is sufficient to expand application of the treaty norms as custom), or whether the law on the question is still developing.
References(p. 28) (C) The Relativity of Custom
The term ‘general international law’ should not be taken to require universal acceptance of a rule by all subjects of international law. True, there are rules of international law which are universally accepted, and the system of international law is daily reaffirmed by states in making and responding to claims of right. But the principles of the system—consent, the requirements for custom, the persistent objector—mean that particular rules may have less than universal acceptance, yet still form part of international law. Similarly a rule of international law to which a state has not expressly or by implication accepted may not be opposable to that state.
(i) The persistent objector
The reduction of custom to a question of special relations is illustrated by the rule that a state may exempt itself from the application of a new customary rule by persistent objection during the norm’s formation.44 Evidence of objection must be clear, and there is a rebuttable presumption of acceptance. Whatever the theoretical underpinnings of the persistent objector principle, it is recognized by international tribunals,45 and in the practice of states. Indeed given the majoritarian tendency of international relations the principle is likely to have increased prominence.46 However, with the increasing emergence of communitarian norms, reflecting the interests of the international community as a whole, the incidence of the persistent objector rule may be limited.47 More common may be disagreement as to the meaning or scope of an accepted rule, as to which the views of particular disputing states will not be decisive.48 Nonetheless the persistent objector rule reinforces the principle of state consent in the creation of custom.
References(p. 29) (ii) The subsequent objector
In Anglo-Norwegian Fisheries part of the Norwegian argument was that even if the 10nm closing line for bays and certain rules were part of general international law, they did not bind Norway which had ‘consistently and unequivocally manifested a refusal to accept them’.49 The UK admitted the general principle, while denying that Norway had manifested its supposed refusal to accept the rules. Thus it regarded the question as one of persistent objection. The Court did not deal with the issue in this way, however. Its ratio was that Norway had departed from the alleged rules, if they existed, and that other states had acquiesced in this practice. But the Court was not explicit with respect to the role of acquiescence in validating a subsequent contracting-out.50Here one must face the problem of change in a customary rule.51 If a substantial group of states asserts a new rule, the momentum of increased defection, complemented by acquiescence, may result in a new rule,52 as was the case concerning the continental shelf. If the process is slow and neither the new nor the old rule has an overwhelming majority of adherents, the consequence is a network of special relations based on opposability, acquiescence and even perhaps historic title. This situation will normally be transitional in character—though in affairs of state, transitions can take some time.
It is difficult to see why the number of States between which a local custom may be established on the basis of long practice must necessarily be larger than two. The Court sees no reason why long continued practice between two States accepted by them as regulating their relations should not form the basis of mutual rights and obligations between two States.54
References(p. 30) When considering the formation of bilateral custom, general formulae concerning custom will not supplant the need for case-by-case analysis. Where a party seeks to vary the general law on a bilateral basis, the proponent of the special right has to give proof of a sense of obligation on the part of the territorial sovereign. In such circumstances the notion of opinio iuris merges into the principle of acquiescence.55 In Right of Passage, the transit arrangement dated back to the Mughal period, and went unquestioned by the British and later independent Indian governments.56
The Party which relies on a custom of this kind must prove that this custom is established in such a manner that it has become binding on the other Party. The Colombian Government must prove that the rule invoked by it is in accordance with a constant and uniform usage practised by the States in question, and that this usage is the expression of a right appertaining to the State granting asylum and a duty incumbent on the territorial State.59
The Court went on to remark that ‘even if such a custom existed between certain Latin-American States only, it could not be invoked against Peru which, far from having its attitude adhered to it, has on the contrary repudiated it’.60 Other attempts to establish a norm of local custom before an international court or tribunal have likewise failed.61
Treaties are the most important source of obligation in international law.62 ‘Law-making’ treaties moreover have a direct influence on the content of general international law, an influence not conveyed adequately by their designation as material sources.
References(p. 31) Bilateral treaties may provide evidence of customary rules,63 and indeed there is no dogmatic distinction between ‘law-making’ treaties and other treaties. If bilateral treaties, for example those on extradition, are habitually framed in the same way, a court may regard the standard form as law even in the absence of a treaty obligation in that case.64 However, caution is necessary in evaluating treaties for this purpose.
So-called ‘law-making’ treaties create legal obligations, the one-time observance of which does not discharge the obligation. Thus a treaty for the joint carrying-out of a single enterprise is not law-making, and fulfilment of the treaty’s objects will discharge the obligation. Law-making treaties create general norms, framed as legal propositions, to govern the conduct of the parties, not necessarily limited to their conduct inter se— indeed the expression of an obligation in universal or ‘all states’ form is an indication of an intent to create such a general rule. The Declaration of Paris of 1856 (on neutrality in maritime warfare), the Hague Conventions of 1899 and of 1907 (on the law of war and neutrality), the Geneva Protocol of 1925 (on prohibited weapons), the General Treaty for the Renunciation of War of 1928, the Genocide Convention of 1948, and the four Geneva Conventions of 1949 (on the protection of civilians and other groups in time of war) are examples of this type. Moreover, those parts of the UN Charter that do not spell out the constitutional competence of the organization’s organs, and other organizational questions, have the same character—notably the principles set out in Article 2 and further articulated in the Friendly Relations Declaration of 1970.65 UNCLOS is a more recent example.66 Although treaties are as such binding only on the parties, the number of parties, the explicit acceptance of these rules by states generally and, in some cases, the declaratory character of the provisions combine to produce a powerful law-creating effect.67 Non-parties may by their conduct accept the provisions of a conventionReferences(p. 32) as representing customary international law.68 This has been the case with Hague Convention IV of 190769 and the annexed rules on land warfare. In special circumstances even an unratified treaty may be regarded as evidence of generally accepted rules.70
In North Sea Continental Shelf71 the principal issue was the extent to which, if at all, Germany was bound by the provisions of the Geneva Convention on the Continental Shelf (GCCS) which it had signed but not ratified. The Court concluded that only the first three articles represented emergent or pre-existing customary law.72 The basis on which the Court distinguished between articles included reference to the faculty of making unilateral reservations, a faculty which applied to some articles but not to those which, by inference, had a more fundamental status. That was a case where the treaty itself made the distinction; by contrast the mere existence of reservations where no provision for reservations is made in the treaty will not by itself annul the probative value of its provisions.73 The Court concluded, further, that the provision on delimitation of shelf areas in Article 6 of the Convention had not become a rule of customary law by virtue of the subsequent practice of states and, in particular, of non-parties.74 The six dissenting judges regarded the Convention as having greater potency, particularly with respect to the generation of rules after the conclusion of the Convention.75 In both Gulf of Maine76 and Continental Shelf (Libya/Malta),77 considerable weight was accorded to aspects of UNCLOS, although it was not yet in force.
According to Baxter, aft er North Sea Continental Shelf it became clear that ‘the treaty-making process may also have unwelcome side-effects’: this is the so-called ‘Baxter paradox’.78 In particular, he notes that treaties declaratory or constitutive of custom may ‘arrest’ its further development and that until ‘the treaty is revised or amended, the customary international law will remain the image of the treaty as it was before it was revised.’79
References(p. 33) (B) Relation of Treaties to Custom
When norms of treaty origin crystallize into new principles or rules of customary law, the customary norms retain a separate identity even where the two norms may be identical in content. Thus a state which fails to become a party to a law-making treaty may find itself indirectly affected by the norms contained in the treaty—unless its opposition rises to the level of persistent objection. Even then its position may be awkward: it will be unable to invoke the new rule itself but unable also to secure from other states continued adherence to the old. This was the experience of the US and Japan in continuing to assert a maximum 3nm territorial sea once it became clear that most states rejected that standard in favour of 12nm.80 More generally the US has sought to rely on provisions of UNCLOS—for example in the field of maritime transit—despite its repeated failure to ratify.
In the long run, one significant effect of non-participation in a law-making treaty is inability to invoke its dispute settlement provisions: a dispute can only arise under a treaty as between parties to the treaty. This may not matter if there is a separate basis for jurisdiction, for example under the Optional Clause or a free-standing dispute settlement treaty,81 and if the customary law rule is arguably the same as that contained in the treaty. In Nicaragua, the position was unusual: the US relied on an Optional Clause reservation that excluded the Court from applying the Organization of American States (OAS) Charter, under which the dispute arose, in the absence of other affected states. The Court avoided the effect of the jurisdictional reservation by holding that it was free to apply customary international law (the content of which was, it held, the same as the OAS Charter).82 But this was to confuse jurisdiction and applicable law: states do not cease to have disputes under a treaty merely because the Court has, in consequence, no jurisdiction over those disputes. The views of the dissenting judges on this point are to be preferred.83
As a general rule, the requirements of duration, consistency, and generality of practice, as well as opinioiuris, means that customary law is often outpaced by specific treaties. But this is not always the case; in the longer term, customary law may be called on to mould and even modify treaty texts which cannot realistically be amended, however desirable amendment might be. A case in point is the law of self-defence as expressed in Article 51 of the UN Charter.84 This parallels the right of self-defence that existed in customary international law prior to the Charter, but makes no mentionReferences(p. 34) of necessity and proportionality. Despite the absence of these words in Article 51, the International Court has read them in.85 The principle does not, however, cut both ways, and the requirement in Article 51 that any exercise of the right be reported to the Security Council has not been imported into custom.86
5. General Principles of Law87
Article 38(1)(c) of the Statute of the International Court refers to ‘the general principles of law recognized by civilized nations’.88 This source is listed aft er treaty and custom, both of which depend more immediately on state consent. Nonetheless, these general principles are not considered ‘subsidiary means’, a term confined to Article 38(1)(d). The formulation appeared in the compromis of arbitral tribunals in the nineteenth century, and similar formulae appear in draft instruments on the functioning of tribunals.89 In the Committee of Jurists that prepared the Statute there was no consensus on the significance of the phrase. Descamps (Belgium) had natural law concepts in mind; his draft referred to ‘the rules of international law recognized by the legal conscience of civilized peoples’. Root (US) considered that governments would mistrust a court that relied on subjective concepts associated with principles of justice. However, the Committee realized that the Court must have a certain power to develop and refine such principles. In the end a joint proposal by Root and Phillimore (UK) was accepted, and this became the text we now have.90
Root and Phillimore regarded these principles as rules accepted in the domestic law of all civilized states, and Guggenheim held the firm view that paragraph (c) must be applied in this light.91 However, Oppenheim’s view is preferable: ‘[t]he intention is toReferences(p. 35) authorize the Court to apply the general principles of municipal jurisprudence, in particular of private law, insofar as they are applicable to relations of States’.92 The latter part of this statement is significant. Tribunals have not adopted a mechanical system of borrowing from domestic law. Rather they have employed or adapted modes of general legal reasoning as well as comparative law analogies in order to make a coherent body of rules for application by international judicial process. It is difficult for state practice to generate the evolution of the rules of procedure and evidence as well as the substantive law that a court must employ. An international tribunal chooses, edits, and adapts elements from other developed systems. The result is a body of international law the content of which has been influenced by domestic law but which is still its own creation.93
(A) General Principles of Law in the Practice of Tribunals
Arbitral tribunals have frequently resorted to analogies from municipal law. In the Fabiani94 case between France and Venezuela the arbitrator had recourse to municipal public law on the question of state responsibility for the state’s agents, including judicial officers, for acts carried out in an official capacity. The arbitrator also relied on general principles of law in assessing damages. The Permanent Court of Arbitration applied the principle of moratory interest on debts in Russian Indemnity.95 Since the Statute of the Permanent Court was concluded in 1920, tribunals not otherwise bound by it have generally treated Article 38(1)(c) as declaratory.96
In practice tribunals show considerable discretion in matters involving general principles. Decisions on the acquisition of territory tend not to reflect domestic derivatives of real property, and municipal analogies may have done more harm than good here. The evolution of the rules on the effect of duress on treaties has not depended on changes in domestic law.97 In North Atlantic Fisheries the tribunal considered the concept of servitude but refused to apply it.98 Moreover, in some cases, for example thoseReferences(p. 36) involving the expropriation of private rights, reference to domestic law might yield uncertain results, and the choice of model reveal ideological predilections.