I’d like to divide
myself in order to see,
among these mountains,
each and every flower
of every cherry tree.1

1. Introduction

At the turn of the century, States began to cooperate more in creating international courts and tribunals meant to oversee novel legal regimes.2 One thinks of the World Trade Organization (WTO) and its dispute settlement bodies; human rights courts implementing their constitutive treaties; temporary international criminal tribunals or the permanent International Criminal Court; countless more investment and arbitral tribunals; and other expert fora, such as the International Tribunal for the Law of the Sea. At the same time, Presidents of the International Court of Justice (ICJ) began to query how cohesive the international legal system could remain amidst this ‘institutional fragmentation’.3 Thus Judge Guillaume described international law as ‘very deficient’ without the sort of appeal or review procedure that national legal systems use to resolve conflicts between judicial institutions.4 The international legal system was in turn threatened by how ‘the same rule of law might be given different interpretations in different cases’.5 He took as ‘example’ the Nicaragua and Tadic decisions of, respectively, the ICJ and the International Criminal Tribunal for the Former Yugoslavia (ICTY).6 He considered the rulings to conflict since the former posits a State is responsible for the wrongdoings of private actors once it has ‘effective control’ over these by directing or enforcing their actions; the latter ostensibly relaxes this standard to ‘overall control’ so that a State is sooner responsible by ‘equipping and financing’ a private group along with ‘helping in the general planning of its military activity’.7 The President underlined such divergent interpretations posed a ‘particularly acute risk’ because specialised fora were ‘inclined to favour their own disciplines’.8 This situation also presaged conflicts between the rules of special regimes, as fora would not read these as concordant.9

In contrast, the International Law Commission (ILC) pioneered a ‘keep calm and carry on’ attitude in 2006.10 To anaesthetise ‘anxiety’ about regime conflicts,11 its fragmentation report assured certain ‘legal techniques’ could help avoid or resolve serious incidences of disunity.12 To this end, Article 31(3)(c) of the Vienna Convention on the Law of Treaties (VCLT) was to play a central role. For while fora accompanying special regimes may have rather exclusive ratione materiae and ratione personae jurisdictions—eg trade law, the law of the sea, regional human rights law—the rule of interpretation within the VCLT reminds any treaty interpreter, including a judicial forum, that these areas of international law are not so afield as to be perfectly isolated from one another. The rule requires that, when interpreting a treaty, ‘There shall be taken into account, together with the context: (…) Any relevant rules of international law applicable in the relations between the parties’.13 Conversely, it asks of, say, a human rights court to not adjudicate without proper notice of how its constitutive treaty relates to the rest of international law in a given case.

It appears then vital that judicial fora just as well interpret and apply the rule of interpretation coherently.14 Yet the literature has hitherto paid little attention to this issue. Support, criticism and clarification towards the harmonising role of the provision have been mounted on its interpretations via Articles 31 and 32 of the VCLT, whereby jurisprudence is only a ‘supplementary means’15 to finding its ‘correct understanding’.16 Then where research does emphasise case law, it focuses on a single ad hoc tribunal,17 trade body,18 or regional human rights court,19 whilst comparisons between fora have not been comprehensive enough to, nor meant to per se, investigate whether they interpret and apply Article 31(3)(c) harmoniously.20

This article thus inquires to what extent judicial fora have conflicted in interpreting and applying Article 31(3)(c). Section 2 reviews what meanings fora have given to clauses of the provision. While the scope is restricted to decisions directly referring to and applying the provision, the article still throws out a large net; the case law belongs to the world court, regional human rights courts, ad hoc investment or arbitral tribunals, WTO bodies and the Iran-US Claims Tribunal. Notably, when judicial fora give different meanings to the terms of Article 31(3)(c), they diverge in ascertaining the contours of the rule of interpretation, ie its normative content. This does not automatically mean that a jurisprudential conflict exists. Whether one does must be ascertained according to a definition of conflict, which can be strict or flexible. Therefore Section 3 analyses the decisions from Section 2 to see if these conflict in stricto or in lato sensu. The article concludes international courts and tribunals have, in either case, not used the rule of interpretation in such diverse ways as to indubitably contradict one another in their determinations of its normative content; but the conclusion comes with some important qualifications.

2. How have judicial fora interpreted and applied Article 31(3)(c)?

This section presents the meanings judicial fora have given to clauses of Article 31(3)(c) without reflecting on how any detail could figure into a narrative of its incoherence. This job is left to Section 3. The clauses relate to what rules of international law are under the provision; how these rules can be relevant; who the parties are; how the rules are applicable between the parties; and when and how these rules must be taken into account.

2.1. What are rules of international law?

Judicial fora readily accept that, under Article 31(3)(c), rules are those belonging to the primary sources of international law, as encapsulated by Article 38 of the ICJ Statute.21 These are then international customary law,22 treaties,23 and general principles of law.24

On the fringe of this scope, matters can become tricky. Most fora forgo taking into account so-called soft law, ie intrinsically nonbinding instruments, unless the relevant norms form customary law. Thus the tribunal in Access to Information under OSPAR refused to consider the Rio Declaration as it was ‘not a treaty’ and still only ‘evolving’ into custom.25 Two years later, the Iron Rhine tribunal found ‘emerging principles’ within the Rio Declaration to signify a ‘trend’ of requiring the prevention of transboundary harm caused by development.26 It found this duty had become part of general international law and could thus be used for interpretation.27 Conversely, whereas one party in Rhine Chlorides argued the polluter pays principle, present in several treaties, was of value to Article 31(3)(c), the tribunal disagreed, finding it remained absent from customary law.28 Some fora, evidently, are quicker than others to deem a norm in statu nascendi to have utility for the provision; each attempts though to square such a norm with the aforementioned primary sources.

There are fora that have not shied away from using Article 31(3)(c) to interpret treaties in light of rules that are neither general principles of law nor part of treaties or custom. In Iron Rhine, it were EU decisions and regulations; their status is arguably sui generis.29 The European Court of Human Rights (ECtHR) has taken into account decisions from international organisations.30 And in Mamatkulov it used numerous human rights bodies’ rules of procedure, along with the ICJ’s Statute, to determine it could prescribe interim measures.31 Rules of procedure, of course, are not treaties. Furthermore, in Case No A/18, the Iran–US Claims Tribunal interpreted a jurisdictional clause in light of the effective nationality rule without mentioning its legal status; it is telling though that the dissenting opinion remarked that no mere ‘trends in jurisprudence or doctrine’ could create a ‘rule of international law’, at least not one for the use of Article 31(3)(c).32

2.2. How must the rules be relevant?

The WTO Appellate Body has tightly conditioned how a rule may be relevant under Article 31(3)(c).33 In PeruAgricultural Products, for example, clauses in a bilateral Free Trade Agreement were not relevant because these did not ‘concern the same subject matter as the treaty terms being interpreted’.34 The Agreement allowed Peru to keep a price range system, setting import duties on some agricultural products, while the interpreted provisions prompted the detachable question of whether to categorise the produce as ‘ordinary customs duties’.35

The Ambiente Ufficio tribunal expressed a looser notion of relevance.36 An investment treaty’s jurisdictional clause mandated claimants to submit disputes to Argentina’s national courts 18 months prior to seeking international settlement, a requirement found to be ‘sufficiently comparable’ to the exhaustion of local remedies rule in diplomatic protection.37 In particular, the tribunal considered the external rule to have the same ‘purpose of honoring the host State’s sovereignty by providing [it] the opportunity to settle a dispute in its own fora’.38 The exhaustion of local remedies rule had then also a ‘futility exception’ that could be read into the jurisdictional clause, meaning if Argentinian proceedings were futile the claimants could seek international settlement without further ado.39

Such drawing on ‘strong structural parallels’ to indicate relevance happens repeatedly.40 The Inter-American Court of Human Rights (IACHR) noted for instance that the American Convention on Human Rights (ACHR) and the Convention on the Rights of the Child established ‘a very comprehensive international corpus juris for the protection of the child’.41 The former’s Article 19, obligating States to maintain ‘measures of protection’ for children, thus had to be interpreted in view of certain provisions in the latter that related to the case at hand, showing street children could not be discriminated against, required special assistance, etc.42 By the same token, the ECtHR interpreted the right to family life under Article 8 of the European Convention on Human Rights (ECHR) in light of, inter alia, the Hague Convention on the Civil Aspects of International Child Abduction.43 Article 8 does not even make mention of children, a fortiori the problem of child abduction, at issue in Neulinger.44 Yet the forum declared the legal material shared the ‘same philosophy’ of appreciating the child’s best interests.45

Then the ICJ has understood relevance more flexibly than the WTO Appellate Body. In Mutual Assistance, a treaty of friendship was relevant for interpreting the Convention on Mutual Assistance in Criminal Matters.46 This was so notwithstanding that the latter did not involve the subject matter of rules in the former, which were also ‘formulated in a broad and general manner’.47 That the parties had once agreed to fostering ‘co-operation and friendship’ did not contribute any ‘specific operational guidance as to the practical application’ of the Convention, but still had some relevance.48

Having said that, in Somalia v Kenya, the ICJ found Article 83 of the United Nations Convention on the Law of the Sea to be relevant for interpreting a Memorandum of Understanding because the latter ‘expressly mentioned’ the former one, while their particular provisions had a ‘similarity in language’ that made it ‘reasonable’ to read one in light of the other.49 Here the ICJ, like the WTO Appellate Body, constructed relevance in a thoroughly textual way.

2.3. Who are the parties?

In EC–Biotech, a WTO Panel appeared to suggest that, for Article 31(3)(c) to operate, the relevant rule must bind all States party to the interpreted treaty.50 The term parties hereby implies ‘parallel membership’.51 However, the forum was rejecting the EC’s argument that an external treaty to which it was party, but not all other parties to the dispute were, could be used under the provision.52 Accepting the argument would, without proper justification, presume a

State would agree to a mandatory rule of treaty interpretation which could have as a consequence that the interpretation of a treaty to which that State is a party is affected by other rules of international law which that State has decided not to accept.53

Panel members then explicitly said they ‘need not, and do not’ have regard for when ‘the relevant rules of international law are applicable in the relations between all parties to the dispute, but not between all WTO Members’.54 Rather, the Appellate Body elaborated hereon in EC–Large Civil Aircraft.55 Seeing as interpretation reveals the parties’ ‘common intention’ behind treaties, it found none must be haphazardly read in light of another without parallel membership.56 Even so, and quoting the fragmentation report of the ILC, it acknowledged how Article 31(3)(c) was key to the ‘systemic integration’ of States’ commitments with the broader legal landscape, such that

a delicate balance must be struck between, on the one hand, taking due account of an individual WTO Member’s international obligations and, on the other hand, ensuring a consistent and harmonious approach to the interpretation of WTO law among all WTO Members.57

Other fora have less consistently shown that relevant rules bind the parties in dispute. Technically, any time a human rights or investment forum settles a dispute between a State and a private actor, ie a legal person not bound by the majority of treaties or custom, any reliance on Article 31(3)(c) would be greatly restricted if parties signified States alone.58 But it goes without saying the ECtHR and IACHR have applied the provision in such contexts, suggesting relevant rules need not bind both parties. The matter will be returned to later, particularly when discussing how rules are applicable under Article 31(3)(c).

It has anyhow become exceptional for human rights fora to take heed of whether relevant rules bind the respondent State. The trend is not too subtle at the ECtHR. When Turkey, in Demir, argued the court could activate Article 31(3)(c) only in connection to rules binding it, the ECtHR refuted the claim, saying that ‘in searching for common ground among the norms of international law it has never distinguished between sources of law according to whether or not they have been signed or ratified by the respondent State’.59 It supported this statement with a wealth of case law. In one dispute, the relevant material only became binding upon EU organs; in another, only eight ECHR parties had ratified the material; in still another, the respondent had not even signed it.60 Meanwhile, the IACHR has often remarked how treaties relevant for interpretation bind States party, but without saying if Article 31(3)(c) requires it to.61 Its conduct may rather be attributed to Article 29(b) of the ACHR, prohibiting any interpretation isolated from parties’ other human rights obligations.62 The IACHR has on one occasion searched for ‘regulatory trends in international law’ by reference to, inter alia, an EU agreement, which obviously binds no ACHR parties.63

Of course, controversy about who the parties are under Article 31(3)(c) is only really pertinent to interpretations in light of conventional rules. Customary law usually binds States unless these are persistent objectors, have contracted out of the custom, or, if this is a regional custom, fall outside the territory of application.64

2.4. How are the rules applicable between the parties?

Noted earlier was how some fora appear to take into account rules that bind not all the disputing parties because these are not all States. Yet these rules can be ‘applicable’ between individuals and States when the former are under the jurisdiction of the latter.65 The IACHR hinted at this construction in a case engaging Article 31(3)(c), asserting the relevant body of human rights law had contributed to international law’s ‘faculty for regulating relations between States and human beings within their respective jurisdictions’.66 That relevant rules are applicable means these must somehow bind parties legally.67

On a related note, the EC–Biotech Panel found the ‘not accepting’ of a conventional rule included a signatory State’s nonratification.68 The rule is not applicable to that State and cannot impact the interpretations of treaties it has otherwise ratified.69 For as the OSPAR tribunal put it, ‘A treaty is a solemn undertaking and States Parties are entitled to have applied to them and to their peoples that to which they have agreed and not things to which they have not agreed.’70

The issue of applicability also involves the ageless dilemma of whether rules coming about after a treaty’s conclusion may still influence its interpretation.71 Fora have repeatedly employed Article 31(3)(c) for dynamic interpretations—though note constructing an ‘evolving meaning’ need not involve Article 31(3)(c).72 The IACHR did do so in one case concerning a medical procedure unknown when the ACHR was drafted.73 The ECtHR has also highlighted that the provision complements its understanding of the ECHR as a ‘living’ instrument, to be read alongside ‘evolving norms’.74 Then the Iron Rhine tribunal interpreted treaties from 1839 and 1873 in accordance with more recent customary environmental principles.75 It had to demonstrate at length though that, beyond textual elements and the travaux préparatoires, the treaties’ object and purpose warranted their application thus.76 If the newer rules were applicable thus hinged on diligently discerning the parties’ intentions.77

2.5. When are relevant rules taken into account and how?

This subsection’s question concerns where Article 31(3)(c) has its place in the general process of interpretation and with what effect. If an interpreter finds, for example, that a treaty’s ordinary meaning and object and purpose already satisfy a certain way of reading it, must she still take into account an external rule? If she does, how may this rule influence the meaning a treaty is eventually given?

The EC–Biotech Panel wrote that if exhausting the other means of interpretation within Article 31 ‘results in more than one permissible interpretation’, paragraph (3)(c) prioritises whichever ‘is more in accord with other applicable rules of international law’.78 Also the Iron Rhine tribunal appears above to have recognised it as a rule of last resort. And the Iran–US Claims Tribunal observed a jurisdictional clause’s ordinary meaning was ambiguous enough for recourse to the provision.79 In contrast, some judges have disapproved of the ECtHR using Article 31(3)(c) to ‘align’ the ECHR with external rules ‘without duly taking into account the wording of the provisions under interpretation’.80

Some of the ICJ’s separate opinions wrote similarly in Oil Platforms, a case worth detailing. The US argued its military campaign against Iran fell under Article XX(1)(d) of the 1955 Treaty of Amity, which permitted measures ‘necessary to protect its essential security interests’.81 Using Article 31(3)(c), the world court interpreted the exception in light of jus ad bellum rules, mindful of the treaty’s object and purpose of ensuring ‘firm and enduring peace’.82 It could never have ‘intended to operate wholly independently’ from these outside rules, so ‘measures’ could not include illegal uses of force.83 The ICJ then lengthily determined the US’ actions were illegal as such and, consequently, not captured by the provision concerned.84 Judge Higgins commented that, had the court analysed the exception more textually, it would have applied it as it had an alike text in Nicaragua, simply finding the military activity unnecessary for protecting essential security interests.85 But Judge Kooijmans noted the word ‘necessary’ had to inevitably correspond with use of force rules; he noted their ‘peremptory’ character meant the US’ actions had to be tested against ‘stricter legal norms’ than others; the treaty itself did not offer these; general international law had therefore to step in and fill the lacunae.86 He likened the way the ICJ used Article 31(3)(c) to ‘putting the shoe on the wrong foot’ though, for the majority very nearly applied the relevant rules to the US’ actions instead of applying the Treaty of Amity upon interpreting it with those same rules.87

This last comment bears more though on what ‘normative weight’ to give outside rules when using them to interpret a treaty—and to a further extent, decide a case.88 In this connection, the ICJ in Oil Platforms gave the use of force prohibition such appreciable normative value—fully testing if the US’ actions breached the rule—that Judge Higgins said the forum did ‘displace the applicable law’ instead of only interpreting it.89 The ECtHR has likewise been accused of ‘extending the scope of treaty obligations’.90Neulinger exemplifies this again, concerning Article 8 of the ECHR and the Hague Convention on the Civil Aspects of International Child Abduction.91 The forum imported from the latter the principle of the child’s best interest and its various requirements into Article 8 and subsequently tested for the provision’s breach by seeing if the respondent’s domestic courts had properly implemented the Hague Convention—it found these to have not, and so the State breached the ECHR.92 Thus it took the relevant rules into account by practically applying them together with Article 8 instead of merely using them to inform the meaning of that provision; had it done so, it would not have needed to see if the State breached the external treaty but just the ECHR.

Mutual Assistance showcases how taking outside rules into account can pervade a treaty’s meaning much less.93 Here, as aforesaid, the relevant rules were worded broadly, aspirational, and not even addressing the subject matter concerned. Their relevance for interpreting the Convention on Mutual Assistance in Criminal Matters was minimal as a result, and could therefore not ‘stand in the way of a party to that Convention relying on a clause contained in it which allows for nonperformance of a conventional obligation under certain circumstances’.94 How much a rule had to be taken into account via Article 31(3)(c) depended then on its degree of relevance. This is complemented by how, in Somalia v Kenya, one rule with a ‘similarity in wording’ was ‘particularly relevant’ for interpreting another, and with stronger implications.95 The outside rule was decisive in determining that the parties had not intended to create a dispute settlement procedure through the interpreted treaty, which would otherwise have barred the ICJ’s jurisdiction—but it should be said the particular interpretation was already suggested by the terms of the treaty as well as its object and purpose.96

2.6. Preliminary conclusions

Section 2 has now reviewed how various judicial fora have used Article 31(3)(c), the interpretations and applications of which clearly appear divided. At times rules must bind all parties to a treaty or—alternatively and where appropriate—just the dispute, or are relevant only if overlapping with a certain provision’s subject matter; other times rules are relevant by how the outside instruments serve a comparable purpose, and this may be true even if those rules do not bind any parties to the dispute. Given the above, Section 3 now uses strict and lenient definitions of jurisprudential conflict to see if fora have substantially contradicted one another in finding the normative content of Article 31(3)(c).

3. To what extent do the interpretations and applications of Article 31(3)(c) conflict?

As aforementioned, the possibility of judicial decisions conflicting was partly what stoked so much worry about the number of international courts and tribunals rising. Yet the literature on fragmentation is wanting in clarity as to what a conflict exactly is.97 When some authors do commit to an explanation, they form their definitions in line with ‘preferences regarding the nature of law and its function in the international community’.98 Since these preferences do not always align, the proclaimed ‘sine qua non condition’ for jurisprudential conflict can vary from one thinker to the next.99 Thus both the strict and broad definitions of conflict are used to analyse the case law detailed above.

3.1. Does the case law reveal a conflict in stricto sensu?

Section 2 crystallised how judicial fora have given dissimilar meanings to Article 31(3)(c). For a conflict in stricto sensu to arise though, fora must have requested certain parties to disputes—certain States—to use the rule of interpretation in a genuinely contradictory way. This means their decisions must ‘create conflicting obligations for the parties involved, which inevitably forces states to breach one or the other law’.100 It follows that a decision requesting compliance with an obligation and another allowing enjoyment of a permission can never genuinely conflict. A State is never technically forced to exercise the latter and, by not doing so, can avoid a violation if need be. The same is true if one decision sets out an obligation that is exhausted by compliance with another (stricter) obligation; that is, if one decision says workers have the right to x days of leisure and another says x + 10.101 Complying with the second fulfils the first.102

Those opting for the strict meaning intend hereby to strongly presume against conflict, for conflicts must be anomalous if the telos of ‘some coherent legal order’ is to instruct behaviour therein.103 If this is the minimal purpose of international law qua law, the legal system must not do so irrationally.104 When permissions and obligations seem to create conflicts the strictest understanding perseveres in saying these actually do not exist and international law remains untainted by their ‘negative connotation’.105

Moreover, the parties and subject matter between cases must correspond for a genuine conflict to appear. Although Judge Guillaume claimed, as noted above, that Nicaragua and Tadic clashed on how a State can be responsible for private actions, the narrow view of conflict rebuts this. It asks whether the ICJ and ICTY did not apply different rules altogether if the former’s decision preceded the latter’s by enough time—over a decade—for the law of responsibility to have changed accordingly.106 And the parties to the cases were different, as well as the facts, which further suggest the effective control test to have been more appropriate for establishing responsibility over the relatively disorganised contra rebels in Nicaragua than the more hierarchical military force that Tadic focused on.107 And if these fora indeed applied different rules to different facts and concerned different parties, their decisions ipso facto cannot contradict, for it becomes inconceivable that any State A has to breach decision x by complying with a decision y that neither binds it nor addresses the same matter.108

So, under the strict reading of Nicaragua and Tadic, one holds these decisions developed the law of responsibility in a not incompatible way, touching upon different areas therein.109 Put differently, the strict definition of conflict, just as the broad one dealt with below, remains sensitive to the notion that norms, judicial decisions, can accumulate instead of conflict.110

Now, as to what the decisions in Section 2 command of States, one must bear in mind that Article 31(3)(c) creates an obligation for an interpreter by being phrased in a mandatory way, stating relevant rules ‘shall’ be taken into account.111 The interpreters of rules in international law are, in principle, States by default, although these can delegate that competence of authoritative interpretation to a judicial forum as part of its jurisdiction to decide a case.112 It stands to reason that when a forum interprets and applies Article 31(3)(c), it binds parties to a dispute with an obligation as to how they must interpret a treaty in light of external rules.113

Also worth knowing is that Article 31(3)(c) of the VCLT codified a customary rule of treaty interpretation that binds all States and has the same normative content.114 Being not solely part of a treaty with limited membership, the rule can more easily generate the overlap of subject matter and parties preconditioning a strict conflict.

In spite of the decisions in Section 2 concerning the same rule though, these never address the same parties. On top of that, even if the parties converged in two cases involving Article 31(3)(c), the subject matter might not. Suppose States A and B submit a dispute about treaty x’s rules to forum α, which applies Article 31(3)(c); these also bring a dispute regarding treaty y’s rules to forum β, and this applies the rule of interpretation, too. If these fora appear to attach to the provision different meanings though, their divergence may be explained by how the interpreted treaties are different, much like how Nicaragua and Tadic were noted to perhaps not strictly conflict if the ICJ and ICTY applied the law of responsibility to sufficiently different facts.

Article 31(3)(c) jurisprudence does seem to attest to the conclusion that divergent uses of the provision might not conflict as much as accumulate. The EC–Biotech Panel, for example, explained it took a particularly strict understanding of parties only insofar as to answer whether an external treaty could inform the meaning of WTO law if not all the disputing parties were party to that treaty, as was the case. It did not preclude that broader meanings of parties might be warranted in other cases. The Panel also chose its particular interpretation because it found no good reason why—verbatim, ‘it is not apparent why’—a treaty should be interpreted in relation to rules a party has not consented to be bound by.115 One can take this statement as pointing out how the EC failed in that case to persuade the forum of an alternative reading of Article 31(3)(c); the statement does not disqualify that there can be a good enough reason for the relevant rules to be taken into account without binding every, or any, party to the dispute.

This is significant to how especially human rights courts have interpreted their constitutive treaties in light of rules either binding on the respondent or not any party to the dispute. As shown above, the IACHR has reasoned the relevant rules are applicable to States and individuals by how these regulate their relations. That said, it has also, without a clear reason, applied Article 31(3)(c) when the relevant rules did not bind the respondent State. The ECtHR did elaborate on a similar application, saying it had to probe for consensus on a human rights norm without needing to rely on whether or not this was binding on parties either to the dispute or the ECHR.

One could argue that the WTO Panel’s interpretation of parties diverges from those of human rights fora because the latter use Article 31(3)(c) to interpret different types of treaties. The ECtHR does constantly underline it must consider outside rules while keeping in mind the nature of the ECHR as an evolving human rights treaty.116 The ECHR, as much as the ACHR, is also not a traditional treaty; it creates no reciprocal obligations inter se States party but ‘objective obligations’.117 The normative content of such ‘erga omnes partes’ obligations must then not become ‘bilateralized’.118 By extension, interpreting a human rights provision in light of an outside rule must never depend upon the latter binding the certain ECHR party that happens to participate in a dispute.119 Simultaneously, it would be disagreeable for external rules to seldom be of use under Article 31(3)(c) if these must strictly bind all 47 parties to the human rights treaty. If this were the reality, the ECHR and other multilateral treaties with objective obligations would become ‘islands’ with little connection to the rest of international law.120 The human rights treaty could also not have the evolving meaning it has today—through Article 31(3)(c) at least—since any evolution in line with newer rules would have to wait for all States party to consent to these first.121 The ILC has in fact maintained that, under Article 31(3)(c), other rules may inform the meaning of a treaty with objective obligations even if not all the treaty parties have ratified these but have ‘“implicitly” accepted or at least tolerated’ them.122 This approach appears to be the ECtHR’s, too. In Demir, for instance, the European Social Charter was relevant for reading the freedom of association—traditionally a political right—as also protecting the economic right of collective bargaining, even though the respondent

invoked the absence of political support on the part of member States (…) for the creation of an additional protocol to extend the Convention system to certain economic and social rights. The Court observe[d] (…) that this attitude of member States was accompanied (…) by a wish to strengthen the mechanism of the European Social Charter. The Court regard[ed] this as an argument in support of the existence of a consensus among Contracting States to promote economic and social rights.123

The approach also explains how the ECtHR has presented little restraint in considering soft law, such as in Mamatkulov. To conclude that other rules of international law informed it could prescribe interim measures, the court did not need to show that other human rights bodies’ rules of procedures, as relevant material, were somehow binding on States party since these bodies were exercising an alike power thereunder without hinderance; this already showed the international community accepted they had that competence.

Furthermore, the WTO Panel may very well have rejected such a wide construction of Article 31(3)(c) because WTO rules are essentially bilateral,124 with the qualification that the trade regime means to ‘help sustain a very delicate equilibrium between the states parties and other stakeholders’ so as to ‘ensure that they obtain the benefits associated with membership of the regime’.125 Thus it is understandable that the WTO Appellate Body clarified an interpreter of a WTO agreement must not be too quick to take into account other rules when these bind only the parties in dispute and not all WTO members, lest the agreements inter se the latter group become too divergent.126

However, the implication here is that the contours of Article 31(3)(c) are almost limitlessly adaptable to the interpreted treaty’s context. Any specific prescription for how the rule of interpretation must be used in a case can then become anchored in, or even emphasise, how special the treaty regime is that needs to be interpreted. This crucial feature of Article 31(3)(c) will be revisited shortly.

It may for now be posited that, if the jurisprudence on Article 31(3)(c) does accumulate and not conflict, judicial decisions may oblige States to interpret their treaties in light of outside rules in a not very streamlined way. There is no one straightforward answer to how the rule of interpretation must be employed in all cases. Sometimes the parties might have to be all those to the dispute; not necessarily all of these; all treaty parties; or none of the above in some instances. Meanwhile, one provision’s interpretation may only allow other rules with overlapping subject matter to be relevant; this does not prevent interpreting another provision in light of rules that are more relevant teleologically, serving the same purpose of protecting a child’s rights or regulating jurisdictional issues in disputes between States and individuals. Then how and when rules must be taken into account may vary by, say, how clear the interpreted provision is on its own merits; Article 31(3)(c) may add little to what is already unambiguous. And the outside material might provide hardly any practical guidance if not too relevant, as it was in Mutual Assistance, but not Somalia v Kenya. But in a case such as Oil Platforms or Neulinger, the rules may be so relevant towards deciding a case due to its facts that entire norm complexes containing principles, prohibitions, requirements and exceptions are imported into the interpreted treaty without there even being a strong textual link. It may seem as though the judicial forum actually applies these then.

Although Article 31(3)(c) has been likened to a ‘master key’ for the entire building of international law, in view of how its contours may mould to the context of a case, the hefty key ring of a janitor may be a more apt metaphor.127 For it is not true that one key fits every lock. A formula for how the rule of interpretation must be used to connect one treaty’s rules to other rules of international law will not do for those of another.128 It is sensible to remark here that while Article 31(3)(c) clearly provides an obligation it is above all a tool of interpretation, and tools can be used to accomplish more than a single task; the obligation is a flexible one. It may precisely be impossible to attain a one-size-fits-all method of interpretation in international law; the ‘actual divergences as to the results of the interpretative process may be considered inherent in the nature of this legal system’.129 One detects then no strict conflict in the use of Article 31(3)(c) because, besides the parties never overlapping, the interpretations and applications of the provision can be argued to accumulate one way or the other. The following subsection tests if there is still a broader kind of conflict at hand.

3.2. Does the case law reveal a conflict in lato sensu?

It remains to be seen whether fora, in determining the normative content of Article 31(3)(c), have conflicted in a wider sense. This contrasts from the strict sense by being built on an entirely different logic, which the ILC preferred in its fragmentation report. Regarding legal rules, it said these ‘may possess different background justifications or emerge from different legislative policies or aim at divergent ends’ that can ‘frustrate’ one another even when there is no strict contradiction in how a certain State must act.130 In other words, that certain rules may bind different parties or concern different subject matter takes away nothing from the fact that their objectives can ‘get in each other’s way’.131

The same goes for rulings, as illustrated with Nicaragua and Tadic once again. The point is not that the parties were different and the fora had different mandates and applied the law of responsibility differently to different facts; what matters is that the ICJ and ICTY decisions stemmed from ideologies that so glaringly contrast.132 The former follows ‘Westphalian logic’ to its conclusion that a State cannot be held responsible for acts not conducted by its own organs unless these outside actors are clearly under its effective control; the reasoning is unpalatable for international criminal law and its mantra of ending impunity for the gravest crimes that States commit, since these—or rather, the leaders of these—may attempt to thereby hide behind private actors and the letter of the law.133 By adopting the overall control test, the ICTY conflicted with the ICJ, but in lato sensu. Such ‘policy conflicts’,134 are thus found when judicial fora substantially ‘use different doctrines, tests and justifications to handle similar cases’.135

The fragmentation of international law may have amplified the problem of disparate policy ‘forces’ not concerting well.136 Fora might be so isolated from one another that norms within one regime are seen as having an underlying philosophy completely separate from, and possibly antagonistic to, another’s. Thus, as Martti Koskenniemi and Päivi Leino put it,

If a human rights treaty body or a WTO panel interprets the [VCLT] so as to reinforce that body’s jurisdiction or the special nature of the relevant treaty, and in so doing deviates from the standard interpretation, then this is bound to weaken the authority of that standard interpretation and to buttress the interests or objectives represented by the human rights body or the WTO panel.137

One would assume then that a conflict in lato sensu exists if there are telltale signs that judicial fora have interpreted and applied Article 31(3)(c) from the perspective of their own regime’s interests exclusively. Such an assumption of what a broad conflict would look like in the case of Article 31(3)(c) may be further supported, and clarified, by formulating the provision’s purpose. To be sure, this article empirically analyses how fora have used Article 31(3)(c) in terms of jurisprudential conflict, but to apply the broad definition thereof, it must delve into a normative enquiry as to what the provision is normally meant to do if its use is not thoroughly hijacked by regimes’ special interests.

The telos of Article 31(3)(c) is not straightforward. Richard Gardiner lists a few functions, eg resolving ‘time issues’ or ‘conflicting obligations’.138 One writer provides that the provision works towards ‘preserving coherence within the subset of international laws applicable between all parties to an interpreted treaty’.139 Another says more conservatively that its ‘primary purpose’ is ‘not to bring coherence to the world of international legal rules, but to offer interpretive guidance’.140 Still another finds it aspires to ‘an appropriate accommodation between conflicting values and interests in international society’.141 And perhaps the ILC articulated the most overwhelming role. Admittedly, it prescribed Article 31(3)(c) as a case-by-case tool to keep international law coherent as a ‘system’, meaning as a minimum that connections between rules are made reasonably enough that they ‘appear not randomly related to each other’.142 But it finally said, too, that Article 31(3)(c) made it possible ‘to give expression to and to keep alive (…) any sense of the common good of humankind, not reducible to the good of any particular institution or “regime”’.143 So the provision should not only connect parts of international law—a mostly technical exercise—but such harmonisation must also accord with thinking of the legal system as securing not regime interests but ‘above all rights and obligations that have backing in something like a general, public interest’.144

This article temporarily entertains the ILC’s more imposing vision of what Article 31(3)(c) should accomplish. It revisits this choice later on though, because crucially, that Article 31(3)(c) should be a vehicle for cohesion that is unspoiled by regimes’ special interests makes the findings in the previous subsection very curious. Under a strict understanding of conflict one must conclude that, besides the requisite overlap of parties missing, fora have interpreted and applied Article 31(3)(c) harmoniously exactly because the rule does not neutrally serve only a single purpose; that its use in many different cases must allow for its normative content to have a construction adaptable to the context of a particular treaty being interpreted here, another very different one there.

When it comes to divergence and isolation from other fora’s jurisprudence, one may as well start with human rights courts, the usual suspects.145 Such nomenclature may carry some pretence. In Al-Adsani, for instance, the ECtHR clarified it had to use Article 31(3)(c) to interpret the ECHR ‘so far as possible’ not inside a ‘vacuum’ but ‘in harmony with other rules of international law of which it forms part’.146 In turn, the rule of State immunity could not ‘in principle’ disproportionately restrict the right of access to domestic courts, even if that access was sought for reason of a violation of the jus cogens prohibition against torture.147 This decision appears to then lend credence to what the ILC described as Article 31(3)(c)’s purpose, since the rights and obligations in all directions—the State’s under the law of immunity, the individual’s under human rights law—were found to be in harmony.148

Then again, the Strasbourg court admitted it took external rules into account only insomuch as the special attributes of its constitutive treaty allowed it to. It finds rules relevant ‘in particular’ if these pursue a similar, humanistic philosophy.149 The IACHR has, too, shown reluctance to give the ACHR an interpretation involving external rules not intent on securing human dignity,150 for it said:

according to the systematic argument, norms should be interpreted as part of a whole, the meaning and scope of which must be defined based on the legal system to which they belong (…) in other words, international human rights law.151

Undoubtedly international human rights law is but a part of international law, so the IACHR implies not to admit any and all relevant rules from the latter into an interpretation. But this policy might be sensible for human rights fora; taking into account certain outside rules could effectively not expand or evolve the meaning of human rights but, as Al-Adsani appears to testify to, restrict their contents instead.152

It must again be stressed that from the strict perspective of conflict such tailored use of Article 31(3)(c) by a human rights court is partly key to arguing there is no incoherence. The special context of human rights treaties allows the forum to use the rule of interpretation in a way that its normative content, revealed and specified thereby, accumulates with how other fora have interpreted it. From the in lato perspective though, such specialisation becomes suspicious if it suggests that Article 31(3)(c) is being used for the special interests of a regime—exactly as the ILC posits it must not be.

Yet how other fora have used the provision also appears to suit their particular judicial context best. As noted above, the WTO Appellate Body has, due to the structure of its regime, urged any interpretation of WTO agreements via Article 31(3)(c) to balance the systemic integration of some parties’ external rules with the need to keep those agreements consistent for WTO members as a whole. Here again the purpose that the ILC envisions for the provision is not realised, but finds compromise with the cause of consistent and global trade liberalisation and the ensuing motive of ‘strengthening the credibility of the parties’ original undertakings and reducing the incentive for defections from the relevant regime’.153 Meanwhile, the need to check whether the bilateral relations between two parties to a dispute do not inappropriately alter the interpretation of a treaty is not as immediately present in, say, ad hoc arbitrations or before the ICJ, where the problem that was posed in EC–Biotech has also not arisen so far.154

Moreover, fora exclusively settling interstate disputes generally equate treaty interpretation with discovering the parties’ intentions behind a text; an approach human rights courts do not so much adhere to.155 Such philosophical division in how to construe a text’s meaning also affects the operation of Article 31(3)(c). The Iron Rhine tribunal observed new external rules to interpret an older treaty only once its plain text and object and purpose suggested the parties intended for its meaning to evolve. The ECtHR, as seen in Section 2, has accurately been criticised for not undertaking a judicial exercise as meticulous. And, from the perspective of the OSPAR tribunal and the EC–Biotech Panel, the ECtHR clearly disavows reverence to State sovereignty when it takes into account rules not binding parties to a dispute. A stronger alignment to sovereignty dictates that only once a State has ratified a treaty can it affect the interpretation of other treaties.156 If the ICJ interpreted the law of responsibility in a State-centric way, and the ICTY in a more humanitarian way, and that thereby the two fora conflicted in lato sensu, it seems Article 31(3)(c) has also received a similar treatment; but this article does not posit such a conclusion just yet.

It is nevertheless clear that while fora use Article 31(3)(c) for systemic integration, they do so not without considering the interests and needs their special regimes militate, be it evident in the very ethos of their treaty interpretation. The provision is, as Nele Matz-Luck describes,

a viable principle for several tasks: filling gaps left by a treaty, clarifying unclear terms and generally safeguarding that terms used in a similar context are understood in a similar way. However, the viability of this specific kind of interpretation as a conflict-solution tool is less obvious. Systemic integration may fail if it is the system that is in many ways incoherent.157

And what has been shown is that the contours of Article 31(3)(c) yield to the context of the interpreted treaty, and there are over a thousand treaties in the international legal system.

In order to see no lato conflict in the provision’s use, one must do either one of two things. One can downgrade from what the ILC purports the rule of interpretation to be, namely an ideal means of systemic integration that remains untouched by the institutional biases of regimes and serves the good of humankind. Even in the fragmentation report a less demanding purpose is discernible, since for international law to be a system the ILC basically requires its rules to have some sort of relationship that is not random.158 If Article 31(3)(c) humbly pursues finding meaningful relationships between rules in spite of—or by virtue of—its paying attention to the special contexts of these rules and their regimes, even the broad perspective of conflict sees no contradiction in how its normative content is determined; the purpose of the rule of interpretation is then fulfilled by simply using Article 31(3)(c) in one case, then another, and where it is appropriate to the context of a treaty. The in lato understanding of conflict is then even reconciled with its in stricto sensu counterpart.

Otherwise, to attain an elevated purpose of Article 31(3)(c), the structure of the international legal system as it stands today would need to be reformed. For, as Judge Crawford writes, ‘fragmentation is the product of a system of laws that, by and large, lacks a sense of vertical integration, of hierarchy’.159 And restructuring international law into a more hierarchical, less decentralized system is a project not for any single State or international subject to manage anyhow. In that case, where one desires Article 31(3)(c) to be an absolutely uncorrupted tool of harmonisation, one will not only be forever disappointed but ask of it what it cannot and never will deliver on. Attaching to the provision such a purpose is inappropriate, and this article cannot entertain it so as to find a conflict in lato sensu. In fine, the case law on the rule of interpretation, reviewed in Section 2, serves not as conclusive evidence of any jurisprudential conflict.

4. Conclusion

Little has been written on how Article 31(3)(c) might not fall victim to the narrow interests of the regimes it is supposed to integrate with the rest of international law; whether its interpretations and applications in various rulings do not actually conflict, strictly or broadly speaking. This is what the article pursued to find out. Ultimately, no in stricto sensu conflict can be detected though since the subject matter and parties of cases never directly overlap. Specifically, the diverse interpretations and applications of Article 31(3)(c) accumulate, for how the provision should be used adjusts to the context of the interpreted treaty. This characteristic of the rule of interpretation appears difficult to reconcile with the relaxed sense of conflict at first sight, which tests whether the policies and philosophies behind decisions frustrate one another. If one construes the rule’s true purpose of systemic integration as harmonising international law perfectly, without any influence from the interests of special regimes and only a vague general, public interest, one may be especially quick to say in lato conflict is observable. But it would be inappropriate, too. In a legal system as decentralized as international law, judicial fora must be given some flexibility in determining the normative content of Article 31(3)(c), lest in making connections between disparate rules their decisions be repeatedly accused of causing conflict. The purpose of the provision as a systemic integration tool must after all be qualified by how the interests of special regimes are always to some extent incorporated in its use. For the provision to tour the full range of treaties in international law, it must be divisible, malleable, and yet still be seen as harmoniously interpreted and applied, at least for now.