Is international law a sphere of authority?

Given this definition, is international law a sphere of authority? At first glance, one might be inclined to think so. The statement ‘international law possesses authority’ is uncontroversial as long as one does not hold a purely realist view claiming that international law is nothing but hubris, a collection of empty phrases exclusively masking power plays. It is not so clear, however, whether the statement ‘international law is an authority’ is equally uncontroversial. An international authority, in the understanding of this Special Issue, exists where an international institution’s decisions, judgments and interpretations are recognised as binding or at least relevant by a critical number of actors for their behaviour (see also Zürn Reference Zürn2018: Ch 3). International law as such does not provide us with interpretations or decisions – actors who refer to and apply international law do.

Of course, there are multiple actors in the international sphere – states, non-governmental organisations, international bureaucracts, and perhaps most importantly courts and court-like actors – who do exactly that: they provide interpretations of norms and rules of international law and issue decisions on its basis. But this means then that international law in and of itself is not the site of authority, at least not in any comparable way to other spheres of authority examined in this Special Issue, such as the sphere of free trade, or the sphere of human rights, or of international security. There is no goal pursued by international law as such; it is not geared towards a clearly identifiable social purpose outside of the law. This is not a trait unique to international law, but rather applies to law more broadly. The very function of law might best be described as providing an institutional setting through which attaining a given social purpose is subjected to a formalism that requires actors to pause, to argue not with the moral valence of the social purpose (or at least not exclusively), but through a specific form that enables self-reflection (Möllers Reference Möllers2019). Thus, law does not have a ‘common goal’ outside of the law. If at all, law is an end in itself.

Norms of international law can thus wield authority,Footnote ² but international law as such does not constitute a sphere of authority in and of itself, at least not if we understand a sphere of authority as a governance space that is delimited by actors’ perception of a common good or goal. But what is it then, and how does this thing called ‘international law’ relate to spheres of authority? We need to clarify how international law is positioned vis-à-vis spheres of authority before we can examine in more detail how it might function as a conflict management tool. This requires some reflection on whether we can usefully speak of ‘international law’ as a set of legal rules, or, put more traditionally, whether or not international law constitutes a system.Footnote ³

International law as a system

Whether or not international law can be considered as a system has been a perennial question of the discipline, and is often tacitly assumed. Throughout all times, general international lawyers have not got tired of asserting that international law could constitute a unified legal order (cf. Simma and Pulkowski Reference Simma and Pulkowski2006), that it is a system (ILC 2006a: paragraph 251 (1)).Footnote ⁴ To be sure, this was for the longest time an aspirational project pursued by idealistic international lawyers (von Bernstorff Reference von Bernstorff, Krieger, Nolte and Zimmermann2019). Hans Kelsen acknowledged that international law in the 1930s had been but a ‘primitive law’, even if he considered this to be a transitory stage and was convinced that it might be assumed that ‘the technical development of international law is progressing on the same path as that already taken by the development of the legal orders of the States’ – he saw international courts as the first step towards centralising international legal institutions (Kelsen Reference Kelsen1941: 97).

Famously, and in contrast to Kelsen’s optimistic outlook, HLA Hart argued 20 years later that international law was precisely not a system (Hart Reference Hart1961: Ch 10). Rather, Hart found, international law was a set of primary rules, but lacked any secondary rules (of recognition, of change and of adjudication). While individual rules of international law were binding – rules of international law were clearly rules of law – Hart did not consider the ensemble of existing rules to form a legal system. Rather, his vision of international law was one of individual rules regulating single instances, but with many gaps in between and many situations in which no applicable legal rule existed. Hart’s view, articulated over half a century ago, has been criticised vehemently since, both on grounds of changed empirical circumstances one finds today (Payandeh Reference Payandeh2010) and on grounds of Hart’s own inconsistencies (Murphy Reference Murphy2017). But engaging with his assertion that international law is not a system, and recent criticism of this position, helps sharpen the contours of what is understood here as ‘international law’ and to position it in relation to one of this Special Issue’s main analytical entities, namely spheres of authority.

Recently, Liam Murphy (Reference Murphy2017: 207–13) has questioned Hart’s finding that international law does not know a rule of recognition. As mentioned, Hart insists that single rules of international law can be individually legally binding, i.e. valid, and that this does not require a rule of recognition. What changes with a basic rule of recognition is that law becomes predictable: ‘in a system with a basic rule of recognition we can say before a rule is actually made, that it will be valid if it conforms to the requirements of the rule of recognition’. Such a requirement, Hart posits, would be fulfilled if there was general acceptance that e.g. multilateral treaties could bind states that are not parties to the treaty (Hart Reference Hart1961: 235). Against this assessment, Murphy (Reference Murphy2017: 210–11) argues that substantive criteria exist for assessing whether a rule is a rule of international law.Footnote ⁵ These criteria are provided by what is traditionally known as ‘sources theory’. Sources theory, concerned not only with how international law is made, but also, more importantly for our purposes here, with how it is identified, how to ‘speak the language of the sources of international law’ (Besson and D’Aspremont Reference Besson and D’Aspremont2017: 7), provides a framework within which any claim to legality must be made. While sources theory is far from being unequivocal or undisputed, there is a general consensus that international treaties, international custom and general principles can be discerned as accepted sources of international law, and that thus, some form of rule of recognition exists in the form of sources theory (Besson Reference Besson, Besson and Tasioulas2010: 163–85).

That states may sometimes not accept the substantive content of individual rules of international law – i.e. by not ratifying a specific human rights treaty or by persistently objecting to the binding force of customary international law – means that we cannot know ex ante and in concreto upon which individual subjects of international law a rule of international law will exert binding force (Hurd Reference Hurd2015: 376–81). But it is in principle possible to know, ex ante, under which conditions a rule will be binding. There is a categorical difference between abstractly accepting the existence of international law as such, and accepting the substantive content of an individual rule of international law.

International law’s institutional challenge

Murphy (Reference Murphy2017: 212) further notes that the ILC and other defenders of the systematicity of international law might not necessarily use the term ‘system’ in a Hartian sense. While Hart requires the union of primary and secondary rules for a collection of legal rules to be called a ‘legal system’, Murphy suggests that the term ‘system’ with regard to international law might perhaps be better understood in terms of how the rules of international law relate to each other, as a system of ‘interlocking norms’.Footnote ⁶ International law and domestic law differ in their institutional structures in obvious ways (Murphy Reference Murphy2017: 213). The most obvious such difference, investigated at length by Collins (Reference Collins2016), is international law’s decentralised institutional nature, i.e. the absence of a central or unified authority similar to the state in domestic law that is exclusively in charge of interpreting and administering the law.Footnote ⁷

One terminological note of caution is in order. When international lawyers talk about the ‘institutional’ dimension of (international) law, they are usually referring to the kinds of organisational entities which administer the law and how these are arranged in relation to one another. The institutionalisation of international law is typically associated with the emergence of a range of organisations that are in charge of administering, interpreting and further developing rules of international law (see e.g. Collins Reference Collins2016: Ch 8). International relations scholarship in turn standardly defines the term ‘institution’ as ‘persistent and connected sets of rules (formal and informal) that prescribe behavioral roles, constrain activity, and shape expectations’ (Keohane Reference Keohane1989: 3) and differentiates between international organisations, understood as entities, and international institutions, understood as rules (see e.g. Martin and Simmons Reference Martin, Simmons, Carlsnaes, Risse and Simmons2013: 326). Thus, legalisation is often characterised as a particular form of institutionalisation (Abbott et al. Reference Abbott, Keohane, Moravcsik, Slaughter and Snidal2000: 401). These two viewpoints are not incompatible. But there is a danger of confusion. Very often, when international relations scholarship focuses on institutions that are not law, these institutions risk to be easily conflated with international organisations that embody them. And because international law scholarship’s main focal point is a particular rule-set, namely law as opposed to other, non-legal rules, it sees no need to explicate its narrower understanding of the term ‘institutional’.

This terminological confusion contains a first hint towards understanding better how international law is positioned vis-à-vis other rule-sets in the international sphere. While it is true that a legal rule is a particular type of rule, and that thus, we can usefully speak of ‘legalisation’ as an instance of institutionalisation in certain contexts, this obscures at the same time the particularities law possesses vis-à-vis other forms of institutionalisation. Law might be an institution in the sense of international relations, but the way in which legal rules are related to one another is different from the way in which rules within other institutions connect to each other: legal rules are mainly held together by their systematicity, i.e. individual rules can in principle be traced back to some form of acknowledged pedigree, to a recognised source. Thus, formal law cuts across institutions that are each held together by a common substantive purpose, or function. Put in terms of this Special Issue: international law is not embodied by one authority, or several authorities that are connected by a common substantive purpose. Rather, international law cuts across spheres of authority. Simultaneously, most spheres of authority studied in this Special Issue have a plethora of different types of rules and norms, not all of which are claimed to be rules of international law. The rules of international law within a sphere of authority will standardly be less than the total number of rules simpliciter within that sphere.

The institutional (or rather: organisational) challenge of international law consists in the fact that the entities administering international law are organised in a decentralised fashion. Because these entities are often and inadvertently compared with the state and its centralised organisation, expectations that people have in international law are perpetually frustrated, because they rest on assumptions of organisational administration of international law that are simply not accurate as a matter of fact (Collins Reference Collins2016). For example, Hart’s requirement that international treaties should bind non-parties sits fundamentally at odds with the decentralised nature of international law, which specifically and inherently provides that many rules of international law will not be applicable to all subjects of international law equally (Hurd Reference Hurd2015: 389–90). Not each individual rule of international law is universally applicable in concreto, but its claim to be law is of a universal nature. Because of this inherent feature of international law, its status as a legal order has perhaps always been more volatile than that of domestic law. Even if domestic law also knows rules that are not applicable to every single one of its subjects (e.g. rules that apply to homeowners, but not to people who rent; rules that apply to adults, but not to children; rules that apply only to members of a certain profession), this is different from the case of international law, in which the subjects themselves are in many cases capable of determining whether or not a rule applies to them.Footnote ⁸ Domestic law makes its distinctions based on differentiations external to the subjects in question. In international law, in turn, the excluded and included subjects belong to exactly the same category, even if not all rules apply to all subjects.

If we look at fragmentation from this perspective, it can be viewed ultimately as a form of partly uncoordinated differentiation which also exists in domestic legal systems (Möllers Reference Möllers2017) – but whereas in the domestic legal system, such uncoordinated differentiation is the exception, at the international level, it is the rule. This is not necessarily a bad thing: in fact, it might be an advantage, as it permits more actors to contend with one another, rather than hierarchically imposing a solution.

Legal argument as regulated interface conflict management

The framework proposed in this Special Issue distinguishes three types of cooperative conflict management (Introduction to this Special Issue): it is constitutionalised where norms of meta-governance are applied through institutionalised procedures to authoritatively solve interface conflicts; it is norm-based where the conflict is solved with reference to third norms, i.e. norms different from the ones that are in conflict with one another; and it is decentralised (yet cooperative) where the parties show a willingness for mutual accomodation and political compromise, but do not resort to third norms or delegate solution of the conflict to a third party (for case studies of such decentralised conflict management see Krisch et al. in this Special Issue).

Conflict management through international legal mechanisms belongs to the category of norm-based conflict management, i.e. conflict management with reference to third norms.Footnote ⁹ Norm conflicts in international law are often solved through resorting to norms of interpretation, which are different from the norms in conflict with one another and thus third norms. Consider for example the conflict between the obligation to prosecute serious human rights violations on the one hand and the rules of immunity from jurisdiction for both public officials as well as states, which pre-empt prosecutions of such cases, on the other hand (ECtHR 2001; ICJ 2002; ICJ 2012; ECtHR 2014).Footnote ¹⁰ International courts that were faced with this conflict applied both procedural norms as well as norms of interpretation in such a way that one commentator opined that ‘any apparent conflict’ had been ‘removed’ through the application of international law (Greenwood Reference Greenwood, Andenas and Bjorge2015: 46).

It should not be surprising that Greenwood considers international courts to be the main actors in ‘removing’ – i.e. managing, in terms of this Special Issue – a conflict between norms. With an increasing number of international courts and tribunals as well as quasi-judicial bodies,Footnote ¹¹ actors who claim for norms to be in conflict often delegate their application and interpretation to such a third party. While third parties can come from within an authority, they are characterised by a certain amount of independence, and thus constitute authorities in their own right.

The above example also shows that more than one court or court-like institution might be competent to address a given interface conflict. This is common for many interface conflicts. Consider for example the conflict between the sphere of international security and the sphere of international human rights, manifested as a conflict between the obligation to freeze the bank accounts of individuals on the Security Council’s sanctions lists without providing information about the listing and the obligation to ensure that such individuals be properly informed of the grounds for their inclusion on such lists, so as to be able to seek judicial review. While this interface conflict is probably best-known as the ‘Kadi’ saga at the European Court of Justice (ECJ 2008, 2013), it was also managed at the European Court of Human Rights, where another listed individual, Youssef Moustafa Nada, challenged his listing first in Swiss domestic courts and subsequently at the European Court of Human Rights (ECtHR 2012; for an overview of both the Kadi and Nada cases see De Wet Reference De Wet2013).

Lastly, the second example also shows that the actors who maintain positional differences are different from the authorities embodying the respective norm(s). In Kadi, the norms in question were embodied by the UN Security Council on the one hand and the EU on the other hand, whereas the actors with opposing views in front of the European Court of Justice were the claimant, Yassin Abdullah Kadi, and the EU organs responsible for imposing the freeze, i.e. the Commission and the Council. In addition, while it was these actors who opposed each other in the court proceedings, they were each part of a broader actor coalition (cf. Introduction to this Special Issue: 15).

The limits of focusing on courts

The Kadi saga illustrates that court proceedings will often capture only a fraction of a larger interface conflict. The focus on courts thus warrants explanation. Courts are neither unique sites in which interface conflicts are successfully handled, nor are they necessarily better positioned than other actors in handling them. Framing interface conflicts through international law is only one possible way of interface conflict management, and it happens not only within courtrooms. In fact, I agree that courts play only a limited role in resolving interface conflicts (Dunoff Reference Dunoff and Young2012: 156). Despite these caveats, I believe that scrutinising the way in which interface conflicts are handled when submitted to courts is instructive if we want to understand how international law functions.

There is a simple, but powerful reason for this: courts and court-like institutions are required to use a particular type of rules when addressing interface conflicts – legal rules. This is also true when the adjudicating institution is part of one sphere of authority, e.g. the European Court of Human Rights as an authority in the sphere of international human rights or the WTO Appellate Body within the sphere of free trade. The actors involved will use rules of international law to frame the interface conflict before courts and court-like institutions, and these third-party actors will use rules of international law to bring about a legal solution to the dispute at hand.

This does not mean that courts, especially those that are rooted exclusively within a given sphere of authority, do not discharge functions other than dispute settlement, law-interpretation and law-application, e.g. the creation of unity within an organisational entity (Shany Reference Shany, Fauchald and Nollkaemper2012: 19–20), or that a court’s hermeneutic does not entail essentially political considerations (Howse Reference Howse, Broude and Shany2011). But as Rosalyn Higgins (Reference Higgins2006: 7) has emphasised, international law continues to be an important point of reference for all international courts. If we want to understand better what role international law can possibly play in managing interface conflicts, then we need to first understand those cases that are squarely recognised as being legal. Studying the legal forms deployed in international courts might not teach us much about how whole regimes impact on and interact with each other (Dunoff Reference Dunoff and Young2012: 156–73). But it can illuminate the specificity of legal conflict management vis-à-vis other forms of conflict management.

Three dimensions of common legal form

With this caveat in mind, I propose a systematisation of how legal argument alters the way in which interface conflicts are framed. Three dimensions of commonality in legal form can be usefully distinguished: the procedural dimension, the argumentative dimension and the substantive, or ‘solution’, dimension. This is an attempt at ordering the different types of legal form that are standardly invoked in front of international tribunals.

Procedurally, actors will be constrained through legal rules as to which conflicts they can submit to third-party actors. Not all actors are subject to the jurisdiction of all courts or court-like institutions – the court needs to establish its jurisdiction first (ratione personae, materiae, temporis, loci). In many cases, particular actors will be altogether unable to seize third-party actors (e.g. individuals only have limited access to international jurisdictions, and non-governmental organisations are generally excluded from activating proceedings). Sometimes, other conflict management steps need to have been undertaken (e.g. exhaustion of local remedies, or mediation or conciliation procedures); often, there is only a particular set of substantive rules that can be invoked. Where a situation is under consideration in one judicial forum, it might be banned from submission in another forum (lis pendens, for examples see Boisson de Chazournes Reference Chazournes and Laurence2017: 46–9). The rules regulating under which conditions a court can adjudicate a case or not – the rules of jurisdiction and admissibility – are thus of primordial importance and can enable courts to avoid pronouncing on the substance of conflicting legal regimes. One such example is the Southern Bluefin Tuna arbitration, in which an arbitral tribunal established under the UN Convention on the Law of the Sea avoided ruling on whether the Convention substantively prevailed over a trilateral fisheries agreement between the litigating parties by pronouncing that it did not have jurisdiction over the case (Arbitral Tribunal 2000). In addition, the procedural dimension also, and crucially, includes rules on how the third-party actor will determine the underlying facts and how they will be evaluated. This is often framed through rules of burden of proof and evidentiary standards as well as the standard of review applied by the respective third-party actor. All of these procedural rules – or third norms, in the language of this Special Issue – will play a role in how an interface conflict is managed.

At the argumentative level, only legal arguments will be accepted in front of courts. This is one of the specificities of courts vis-à-vis executive agencies and the legislature, and, of course, this is inextricably linked with the procedural aspect (Möllers Reference Möllers2013: 92). This means that in order to justify one’s behaviour or claim, the actor cannot exclusively refer to a norm, but must at least implicitly (and better yet: explicitly) suggest that the invoked norm fulfils specific requirements of pedigree. Put differently, an actor invoking a norm as a norm of international law must either rely on such attributes that are generally accepted as sources of international law (e.g. that it figures in an international treaty), or must make an effort to justify why a norm should be recognised as a legally binding norm. This has consequences for the kinds of interface conflicts that are likely to be channelled through court proceedings: it only makes sense to seize a court or court-like institution if the actor who wants to prevail can make at least a plausible claim that there exists a legal norm. In addition, this claim must be such that ideally, the opponent cannot make a plausible claim to legality as well. It is a common strategy to argue that the norm the opponent relies upon is not, or not yet, a norm of international law. For example, the United States argued in the EC–Hormones case, examined more fully below, that the precautionary principle was an ‘approach’ and not a norm of customary international law, as claimed by the European Communities (WTO 1998: paragraph 43). A similarly effective strategy, which leads back to the procedural dimension, is to argue that the seized court does not have jurisdiction over the norm that the opponent relies upon, even if the norm is a legal norm.

Lastly, the conflict management devices that courts and court-like institutions choose to make explicit are again exclusively legal rules. Legal rules come into play at two levels in resolving interface conflicts. The first level might be called ‘legal norm conflict resolution’ proper: Where an interface conflict has been framed as a conflict between two substantive norms of international law, norm-conflict resolution rules come into play. These include the rules of lex specialis, lex posterior and lex superior (ILC 2006b), but also conflict of laws approaches (Michaels and Pauwelyn Reference Michaels and Pauwelyn2012) as well as rules of justification, excuses and the principle of proportionality (Jeutner Reference Jeutner2017: Part II). Take for example the Kadi litigation in the Court of First Instance: here, the Court had resolved the conflict between imposing the sanctions without informing the listed individual and the duty to inform the individual and ensure a right of judicial review of the listing decision through the application of Article 103 of the UN Charter, an explicit precedence rule. The Kadi litigation shows, however, that a conflict can also be solved without being explicitly framed as a conflict between two legal rules. The European Court of Justice ruled exclusively on the basis of European law, thus avoiding explicitly pronouncing on the conflicting rules. This is not a singular instance. Often, international norms are vague and cannot unequivocally be constructed as conflicting. Instead, they are in need of interpretation (Jeutner Reference Jeutner2017: 22–7). This is where rules of interpretation come in – rules that are in turn again rules of international law.Footnote ¹²

Pulkowski (Reference Pulkowski2014: Ch 6) has suggested that some basic concepts of international law (sovereignty, the right to have human rights, the Vienna Convention on the Law of Treaties and the UN Charter) ought to be viewed as ‘constitutive rules’ that enable discourse between different legal bodies, and thus lead to ‘communicative compatibility’. These ‘discourse rules of international law’ (Pulkowski Reference Pulkowski2014: 238) correspond largely to the rules mentioned in the third, ‘solution’, dimension above. While I do consider these substantive rules to be important when courts are called upon to provide a solution, I believe that a broader notion of legal conflict resolution that also encompasses the procedural and formal legality dimensions, as outlined above, captures more fully the elements of the communicative compatibility that is characteristic for the thing that we call ‘international law’.

The EC–Hormones litigation as an interface conflict

In EC–Hormones, as will be examined in more depth below, the WTO AB issued its decision on the basis of WTO law alone and did not make any finding as to the alleged norm conflict between WTO law and general international law. But this does not mean that the underlying dispute could not have been reasonably conceived of as a norm conflict, or that no interface conflict existed. The US had brought the case to the WTO Dispute Settlement Mechanism in 1996, claiming that the European Communities’ import ban on meat was violating several provisions of the GATT and other WTO agreements.Footnote ¹⁴ The European Communities countered that the import ban was justified under the precautionary principle as a principle of international environmental law. In other words, the European Communities argued the existence of a norm conflict: they suggested that the prohibition to impose import restrictions, stemming from the obligation to ensure free trade, was in conflict with the obligation to prohibit the import of potentially harmful food products, stemming from the obligation to protect the environment and health.

The EC–Hormones litigation is thus a prime example of what, in this Special Issue, is conceptualised as an interface conflict across spheres of authority: we have actors who explicitly disagree about how different norms and rules relate to one another, with a set of the norms in question associated with the World Trade Organization as an international authority. As such, the EC–Hormones case has also been important for the fragmentation debate in international law. The International Law Commission’s Study Group on fragmentation considered EC–Hormones to be representative of fragmentation as the result of ‘different special laws’, namely international trade law on the one hand and international environmental law on the other hand, and cautioned that the WTO AB’s approach ‘suggests that “environmental law” and “trade law” might be governed by different principles’ (ILC 2006b: paragraph 55, emphasis added). These ‘principles’ are roughly equivalent to what this Special Issue conceptualises as differing social purposes of different spheres of authority.

Against this backdrop, it is perhaps not surprising that the 1998 WTO AB Report did not settle the conflict definitively. Rather, it is only one singular instance of a protracted interface conflict between the United States and Canada on the one hand and the European Communities on the other hand, lasting over two decades. Following the 1998 Appellate Body report, which found that the European Communities’ beef import ban was violating WTO law, the European Communities amended its internal EC law to allow for the administration of certain growth hormones. The United States and Canada judged this insufficient and suspended concessions granted to the European Communities on the grounds of non-compliance. The European Communities, however, considered that they had in fact removed the measures found to be inconsistent with WTO law and thus complied with the Appellate Body Report. Now, the EC requested consultations, complaining that the continued suspensions of concessions were in turn violating WTO law (WTO 2004a; WTO 2004b). Both disputes were ultimately managed through mutually agreed solutions, in the case of the conflict between the EC and Canada through the negotiation of the Comprehensive Economic and Trade Agreement (CETA; WTO 2017) and in the case of the conflict between the EC and the US through a Memorandum of Understanding (WTO 2014). From this perspective, the interface conflict is much more recent than the date of the Appellate Body report might suggest.

Lastly, while the EC–Hormones litigation concerned exclusively the import of meat where animals had been treated with certain growth hormones, additional cases at the WTO have addressed questions of genetically modified food and agricultural technology, a question that ultimately remained unresolved politically, despite several attempts at finding a cooperative solution (WTO 2006; Pollack and Shaffer Reference Pollack and Shaffer2009: Ch 5). The EC–Hormones case is thus part of a larger complex of related transatlantic confrontations over genetically modified products and other forms of biotechnical engineering. At the bottom lies a fundamental disagreement as to whether food production ought to be subject to various forms of bioengineering, so as to improve the products at hand and enhance trade in them, or whether such treatment ought to be subject to a precautionary approach in order to guarantee maximum environmental and health protection – or, put in terms of this Special Issue, an interface conflict between spheres of authority. In the remainder of this section, I shall address the three dimensions of common legal form as they appear in the EC–Hormones Appellate Body report, all the while keeping in mind that this case only represents one particular instance of conflict management in a larger and much more protracted interface conflict (see Introduction to this Special Issue: Figure 2, for an analytical model of the interface conflict framework).

Proceduralising interface conflicts

As mentioned before, the procedural dimension is concerned with framing the conflict in legal terms. This includes, crucially, rules about how the court or court-like institution will establish the facts upon which it will base its judgment. In the EC–Hormones appeal, several points on the evidentiary standards and the underlying facts were raised. The European Communities argued inter alia that the Panel had erroneously put the onus of burden of proof on the EC, despite its status as responding party,Footnote ¹⁵ that the standard of review for the Panel ought to have been whether the measure was reasonable, and not whether it was conforming to an assessment of the scientific evidence as carried out by the Panel, and lastly, that the way in which the Panel had assessed the facts and evidence presented to it had not been adequate, and that it ought to have established an expert review group (WTO 1998: paragraphs 9–18, 37). Here, we can observe that the underlying interface conflict is not really visible – in fact, the main consequence of proceduralisation is that the interface conflict is precisely not being framed as an interface conflict. Thus, in response to the EC’s claim that the Panel ought to have established an expert group, the US not only argued that no such obligation existed under the Dispute Settlement Understanding, but additionally, and primarily, claimed that the EC would have needed to make that claim earlier in the proceedings. Because it had not raised this procedural grief during the Panel proceedings, the US argued that it was now precluded from doing so (WTO 1998: paragraph 54). This is a good example of how proceduralising an interface conflict radically alters the way in which the conflict is framed: no matter of substance is raised here at all.

Legal argument as a form: The importance of formal sources

The EC–Hormones case is perhaps best-known amongst international lawyers for its (non-)treatment of whether the precautionary principle is a rule of customary international law and if so, how it would relate to WTO norms. Famously, the Appellate Body did not decide this question, but rather argued that it was ‘unnecessary, and probably imprudent, for the Appellate Body in this appeal to take a position on this important, but abstract, question’ (WTO 1998: paragraph 123). However, the very fact that the EC had argued that the precautionary principle formed part of customary international law, or constituted a general principle of law, evidences that the EC needed to claim that the precautionary principle was indeed part of formal international law. It would not have been sufficient to claim that the precautionary approach was, as a matter of fact, superior to the approach taken on growth hormones by the US and Canadian authorities. Appealing to the substantive function of the rule is not enough. Rather, for the precautionary principle to be applicable to the situation at hand, the EC had to argue that it could be traced to a specific source, or sources, of international law (WTO 1998: paragraph 16). In turn, and accordingly, both the US and Canada tried to fend off that claim by denying that the precautionary principle was a rule of customary international law or a general principle in the sense of Article 38(1)(c) ICJ Statute. Instead, they suggested that something called a ‘precautionary approach’ might exist, but that such an approach was legally irrelevant (WTO 1998: paragraphs 43, 60). Both parties reiterated these stances a few years later, in the Biotech litigation addressing the same underlying interface conflict between the obligation to import bioengineered products and the precautionary principle prohibiting import of those products (WTO 2006; for comment see Pollack and Shaffer Reference Pollack and Shaffer2009: Ch 5), with the panel concluding again that it did not have to pronounce on whether the precautionary principle was a legal norm (WTO 2006: 341, paragraph 7.89).

The ‘solutions’ level: interpretation rather than conflict rules

Since the Appellate Body made no finding on the precautionary principle, but considered it to be reflected in several provisions of the SPS Agreement, the appeal mainly turned around the Panel’s interpretation of these provisions.Footnote ¹⁶ What is important for the purposes of this article is that the Appellate Body did not apply any ‘proper’ conflict resolution norms: it did not frame the underlying interface conflict between the social purposes of free trade and environmental and health protection as a conflict between two legal norms. Rather, all substantive issues were issues of interpretation of various SPS Agreement provisions.Footnote ¹⁷ Again, the fact that the Appellate Body in EC–Hormones, and, subsequently, the panel in Biotech, did not make a finding as to the legal status of the precautionary principle and opted to resolve the case exclusively through the SPS Agreement does not mean that there was no norm conflict to begin with. Indeed, it has been argued that from a legal-doctrinal point of view, it would have been more appropriate for the WTO AB to pronounce on the status of the precautionary principle, and, if it had found the precautionary principle to constitute a rule of customary international law, to resolve the conflict between the two norms through applying the lex posterior and lex specialis rules, respectively (Pauwelyn Reference Pauwelyn2003: 482).

This underlines that interpretation of legal rules by courts is an important way in which interface conflicts are managed through international legal rules – and that such conflict management can also entail avoidance of the conflict. In the EC–Hormones case, the Appellate Body rephrased the norm conflict that it had been presented with in terms of the WTO’s Covered Agreements and could thus address it through exclusively applying WTO law. A characteristic of courts and court-like institutions is that they are very likely to suggest some kind of solution, even if the underlying conflict is not squarely addressed. This was the case here: the WTO Appellate Body ultimately upheld the Panel’s main finding, namely that the EC’s import ban did indeed constitute a violation of WTO law (WTO 1998: paragraph 253). In fact, situations in which international courts decline to decide a dispute are rare and almost never occur in contentious proceedings.Footnote ¹⁸ Once the jurisdictional and admissibility thresholds are passed, courts and court-like institutions will suggest a solution, and they will do so by using legal interpretative and argumentative techniques.

This does not mean that international law provides a single answer for all or even most cases of interface conflict. The use of legal interpretative and argumentative techniques does not unequivocally yield one single solution, as the above example shows. Rather, international law provides a specific type of conflict management, a common language that is characterised by its formalism. Such a solution will always be of an interim nature, as it will never solve the underlying substantive political interface conflicts (ILC 2006b: paragraphs 487–8; Krisch Reference Krisch2010: 23). A legal solution provided by a court cannot provide a solution on exclusively substantive considerations: even if the solution is a substantive one, the court, in justifying it, will need to do so within the confines of its procedural and argumentative rules.Footnote ¹⁹