Debating interpretation: On the road to Ithaca

1. Introduction

In the articles published in the Leiden Journal of International Law (LJIL) and irrespective of the topic analysed, a theme that is touched upon time and time again is that of interpretation. Similarly, in recent years both the International Law Commission (ILC) and the Institut de Droit International (IDI) have devoted part of their work and the expertise of their members to addressing topics directly tied to interpretation. The ILC, for instance, in 2018 produced a set of Draft Conclusions with commentaries on ‘Subsequent agreements and subsequent practice in relation to interpretation of treaties’. ¹ Along the same vein, the IDI in the summer of 2021 published a report on the topic of limits to the dynamic interpretation of the constitutive instruments of international organizations by their internal organs. ² In addition to these, there have been several topics that in some shape or form share a connection with interpretation, as is, for instance, the case with the ILC’s work on ‘Unilateral Acts of States’, ³ ‘Identification of Customary International Law’, ⁴ ‘Jus Cogens’, ⁵ and the ‘Immunity of State Officials from Foreign Criminal Jurisdiction’. ⁶

Despite the ubiquity of interpretation in all fields of public international law, the nature, function and content of its rules is still debated. So much so, in fact, that even whether in the case of interpretation we can talk about ‘rules’ in a strict sense, ⁷ is also contested. It has been argued that it is more appropriate to refer to these ‘rules’ as ‘principles’, ‘canons’, ‘maxims’, or simply ‘methods’, with all the accompanying baggage that such terms come with regarding the normativity of the thing they attempt to describe. ⁸

Does this variety of theoretical approaches to interpretation and its ‘rules’ have any effect on the value of the discussions surrounding their content, such as the recent efforts of the ILC and the IDI? The aim of the thoughts put down in the following pages is to demonstrate that irrespective of where one stands on the spectrum of the debate on the ‘rules’ of interpretation, ⁹ we can only benefit from the continuation of this discussion. Section 2 will discuss that the Vienna Convention on the Law of Treaties (VCLT), ¹⁰ and its Articles 31–33, have contributed to the adoption of a common vocabulary on which interpretative arguments are constructed. This common vocabulary has had as a consequence the refinement of interpretative argumentation and an increased emphasis of international courts and tribunals on providing more detailed explanations of the steps of their interpretative reasoning (often based on the VCLT structure and vocabulary). Even when such reasoning is flawed (Section 3) the ability to identify such flaws, which stems from both the common vocabulary and the more detailed reasoning, is in and of itself an important benefit, as it can further the discussion and refinement of the rules of interpretation. As Section 4 will illustrate there is still a plethora of issues surrounding the content of the rules of interpretation that are still to be resolved, and many more yet to uncover. It is through this continuous process of debate and refinement that we have the most to gain, in understanding not only the rules of interpretation but international law as a system as well.

2. Setting out for Ithaca

When one talks of interpretation in international law, the almost Pavlovian response is to think of Articles 31–33 VCLT. Although, we now consider them as a given, that was not always the case. When debating the law of treaties there was a lengthy discussion among the ILC members as to whether such articles could and should be included in the draft articles on the law of treaties. ¹¹

Despite this, the ILC decided to include articles on interpretation. Although this may seem prima facie to be somewhat irreconcilable with the extensive debate within the ILC on the existence of rules of interpretation, the ILC clarified the reasons behind this choice in its commentary:

[T]he Commission confined itself to trying to isolate and codify the comparatively few general principles which appear to constitute general rules for the interpretation of treaties. Admittedly, the task of formulating even these rules is not easy, but the Commission considered that there were cogent reasons why it should be attempted … In addition the establishment of some measure of agreement in regard to the basic rules of interpretation is important not only for the application but also for the drafting of treaties. ¹²

It is quite striking how heavily the interpretative dialogue has been influenced by the choices made by the drafters of the VCLT. This has been also accentuated by the further linkage of customary rules of interpretation with their corresponding VCLT rules, to the degree that they are now considered to be one and the same. ¹³ This is further evinced if one compares the language and terms used in international jurisprudence pre-VCLT and post-VCLT, with the former being oftentimes mere assertions of the interpretative outcome or, in the best of scenarios, an invocation of a variety of interpretative tools but without any unified frame of reference. Although such instances continue to occur, the VCLT has led to the emergence of a common vocabulary that has streamlined, although far from perfected, the interpretative process. All ‘users’ of international law ¹⁴ argue on the basis of the language, and structure of the interpretative rules enshrined in the VCLT (both as conventional rules and as reflective of CIL).

The adoption of a common lexicon for interpretation is not, however, the only effect of the VCLT rules. An immediate corollary is that the interpretative reasoning can become both simpler, more streamlined, but also vastly more nuanced and complex. Interpretative arguments are increasingly structured in such a way as to demonstrate that certain materials, or methods fall within or outside the elements mentioned in the VCLT rules, and follow very closely the structure and choices of those articles to such a precise degree sometimes even when one is applying CIL rules of interpretation. This increased streamlining and complexity in interpretative argumentation on why certain interpretative outcomes are preferable through an application of the VCLT rules, and of judicial reasoning in general, can also provide more data and opportunities for the ‘users’ of international law to more readily and clearly evaluate whether the reasoning (though not necessarily the outcome) of the court or tribunal in question may be flawed, from the perspective of the application of the rules of interpretation. It is for this that we shall set sail in the next section.

3. Losing the path to Ithaca

In this part of our journey, I will examine two cases that best highlight how the evolution of interpretative reasoning in the decisions of international courts and tribunals is still a good thing even if they commit methodological errors. ¹⁵ Perhaps this holds true even more so in such cases of flaws precisely for that reason, namely that we, the ‘users’ of international law, can identify such errors because of the effort these courts and tribunals undertook to explain their interpretative choice on the basis of the rules of interpretation. The two cases that will be examined are Vattenfall AB and Others v. Germany and EC – Biotech.

The Vattenfall AB and Others v. Germany case arose out of Germany’s decision to phase out the use of nuclear energy. The applicants claimed that this decision was in violation of several obligations under the Energy Charter Treaty (ECT). ¹⁶ Germany, among other things, raised a jurisdictional objection, namely that following the Achmea judgment ¹⁷ the ICSID Tribunal did not have jurisdiction. Although the EU was not a disputing party, it intervened as a non-disputing party, because the dispute concerned issues that affected EU law and policy, and also because, as it had argued, the Achmea judgment (which concerned a BIT) could be applied mutatis mutandis to multilateral treaties. ¹⁸

The main point here was whether EU law could be considered a ‘relevant rule’ for the purpose of interpreting the ECT. I will leave aside the issue of whether the Tribunal was interpreting and applying customary international law (CIL) or the VCLT, ¹⁹ which depends on one’s interpretation of Article 4 VCLT as applying on a bilateral/izable basis or as a si omnes clause, ²⁰ and which the Tribunal evaded by the mercurial statement that the ECT must be interpreted in accordance with ‘the principles of international law relating to treaty interpretation … [which] are reflected in the VCLT’. ²¹ As far as the ‘relevance’ of EU law the Tribunal had this to say:

[t]he [European Commission’s] approach [ie that the Tibunal should take into account EU law as “relevant rules” under Article 31(3)(c) despite the fact that several State-parties to the ECT are not EU member-States] is unacceptable as it would potentially allow for different interpretations of the same ECT treaty provision. The Tribunal considers that this would be an incoherent and anomalous result … The need for coherence, and for a single unified interpretation of each treaty provision, is reflected in the priority given to the text of the treaty itself over other contextual elements under Article 31 VCLT … While the [European Commission] refers to the use of Article 31(3)(c) VCLT as a “systemic or harmonious interpretation”, the Tribunal finds that the effects of such an interpretation in the manner proposed in the [European Commission’s] submissions would not ensure ‘systemic coherence’, but rather its exact opposite. ²²

The Tribunal should be commended for attempting to justify its interpretation of Article 31(3)(c) VCLT and for analysing in the course of several paragraphs the reasons why in its view Article 31(3)(c) VCLT should be interpreted as referring to ‘relevant rules of international law applicable in the relations between the parties to the treaty’ rather than ‘relevant rules of international law applicable in the relations between the parties to the dispute’. ²³ It is exactly this detailed elaboration of its reasoning, and the various steps it had to take in order to arrive to its conclusion, that allows us to scrutinize its interpretative reasoning and identify any potential flaws. This discussion, in turn, may gradually lead to an overall further refinement of the rules of interpretation, and a clarification of their content.

However, although the Tribunal’s attention to a detailed reasoning should be commended, regrettably the same cannot be said about the reasoning itself. It suffers from certain presumptions that do not hold up to scrutiny as will be shown in the following paragraphs. The Tribunal’s interpretation is fuelled by its view that interpretation must lead to coherence and to a single unified interpretation of each treaty provision. This can, in the Tribunal’s view, be secured only if we interpret Article 31(3)(c) VCLT as referring to ‘parties to the treaty’, and that such an interpretation would ensure the stability of the treaty. ²⁴ The underlying presumption here is that a ‘parties to the dispute’ interpretation of Article 31(3)(c) VCLT would lead to incoherence, as the same provision would have a different meaning ‘depending on the independent legal obligations entered into by one State or another, and depending on the parties to a particular dispute’. ²⁵

But is this really true? At closer look, this seems to be an oversimplification. Firstly, a ‘parties to the treaty’ reading does not eo ipso mean that all rules binding between the parties to the treaty would be useful for the purposes of interpretation, since they would have to be proven ‘relevant’. Second, the ‘parties to the treaty’ approach far from ensures the stability and uniformity of the interpretation. The reason is quite simple. The participation to treaties is far from static. So, if one state leaves the ECT in our case, then potentially this would affect the set of ‘relevant rules’. The same could also happen if a new state joins the ECT. A whole set of ‘relevant rules’ would potentially come into or out of play. But that is not all. The same fragility and instability would occur the other way around. If a state party to the ECT, joins or withdraws from other treaties. Then potentially these treaties depending on the participation could become ‘relevant rules’ whereas previously they were not, or become irrelevant, whereas previously they fell under Article 31(3)(c). In fact, if one breaks these scenarios down, it becomes apparent that the stability and uniformity requirement is better served by the ‘parties to the dispute’ approach rather than the ‘parties to the treaty’. ²⁶ That is not to say that the interpretative outcome of the Tribunal is erroneous. The same result, of irrelevance of the EU law, most likely could have been arrived at by focusing on relevance rather than on ‘parties’. However, just because a broken clock is right twice a day does not heal the flaw in the Tribunal’s reasoning.

A further methodological criticism can be also raised because only after this analysis, does the Tribunal as a supplementary argument raise the issue that no ‘rule’ had actually been invoked, as the EU had simply referred to EU law as a whole rather than specifying a particular rule. ²⁷ Since no ‘rule’ existed, and from a judicial economy perspective, the analysis should have started from there as this would be an automatic bar for the application of Article 31(3)(c) VCLT.

A similar error in methodology can also be seen in EC-Biotech. This case concerned the claim that measures taken with respect to biotech products by EC member states were in violation of WTO law. The EC argued that the Convention on Biological Diversity (CBD) ²⁸ and the Biosafety Protocol ²⁹ were ‘relevant rules’ for the interpretation of the SPS Agreement. ³⁰ The USA, Canada, and Argentina objected on the ground that they were not bound by the Biosafety Protocol, ³¹ while the USA was also not a party to the CBD. ³²

The Panel, as the Tribunal in Vattenfall, engaged in an interpretation of the applicable rule of interpretation, and came to the conclusion that Article 31(3)(c) VCLT refers to ‘parties to the dispute’. However, this seems to be quite problematic since the Panel’s conclusion seems to be in stark contrast with the method it used. One point that needs to be clarified from the start, is that the rule applicable was not Article 31(3)(c) VCLT but customary law. Article 3.2 DSU clearly specifies that the WTO Dispute Settlement Bodies (DSBs) apply the customary rules on interpretation. ³³ Consequently, what the Panel was doing in EC-Biotech was interpreting the customary rule on interpretation. In fact, the Panel’s analysis is peppered with mentions of the rule of interpretation being ‘interpreted’. ³⁴ The Panel does this by reference to the text of the VCLT, and not only the text of Article 31(3)(c) but also of Articles 18, 31(2)(a) and (b), 31(3)(a), 66, and even the ILC commentary to the draft VCLT. ³⁵ All these references to the structure and terms of a variety of provisions within the VCLT are rather problematic, because it treats the VCLT as being in its entirety reflective of and identical with customary international law. This becomes even more difficult to account for, if one considers, first that the Panel concluded that for that case in order for a rule to be ‘relevant’ it had to be applicable in the relations between ‘the parties to the treaty’, and second that neither the USA nor the EU (then EC) were parties to the VCLT. However, when interpreting the customary rule on interpretation the Panel used the VCLT as a ‘relevant rule’. The problem with this is that this seems to fly in the face of its eventual finding. Since the VCLT has not been ratified either by the USA or the EU, it could not have possible qualified as a ‘relevant rule’ through which the customary rule on interpretation, and more particularly the ‘principle of systemic integration’ itself could have been interpreted. ³⁶

Consequently, the Panel applied a more expansive version of the principle of systemic integration only to conclude that the more restrictive version is the correct one. It interpreted the customary ‘principle of systemic integration’ by reference to the VCLT, which was nothing more than a ‘relevant rule applicable to any one of the parties to the dispute’ (not even all the parties to the dispute). It then, without missing a beat, concluded that ‘relevant rules’ are only those that are ‘applicable between all the parties to the treaty’. ³⁷ In essence, the Panel applied a version of the rule of interpretation that it eventually held should be rejected!

4. The multiple stops on the road to Ithaca

The above examples illustrated why the discussion surrounding and elaboration of the elements of the process of interpretation is so crucial. It not only provides a common lexicon and common ground for all ‘users’ of international law, ³⁸ but gently forces courts and tribunals (both international and domestic) to give more substantiated and clearly argued judgments. This, in turn, helps to hold these courts and their judgments accountable to a higher standard of reasoning and to a methodological coherence that is necessary for our discipline, thus contributing to the further refinement of the language and tools to be employed in the application of international rules, whatever their source.

Despite this, the road to Ithaca is far from over. There are numerous topics that still haven’t been resolved, and many more that we are still not aware. To mention a few:

• the relevance of interpretative maxims not explicitly mentioned in the VCLT for the interpretative process and whether they fall implicitly under Articles 31–32 VCLT or have an existence praeter-VCLT; ³⁹

• whether Article 32 VCLT can be used to correct the meaning of the text, in favour of which Judge Schwebel argued passionately in his Dissenting Opinion in Qatar v. Bahrain; ⁴⁰

• this also ties to the more overarching question of the (in)existence of a hierarchy between Articles 31 and 32 VCLT, with Judge Torres-Bernárdez very eloquently opined that ‘[a]s to the relationship between the “general rule of interpretation” (Art. 31) and the “supplementary means of interpretation” (Art. 32), it is also clear that the fact that they are presented as two different Articles does not at all mean that there are two interpretative processes. The interpretative process is a single one and, the interpreter is free at any moment to tum his attention to the supplementary means of interpretation concerned without waiting for completion of the application of the general rule of Article 31’; ⁴¹

• the role of in pari materia interpretation in the interpretative process, a somewhat missed opportunity in the recent Qatar v. United Arab Emirates case; ⁴²

• whether if subsequent practice not meeting the criteria of Article 31(3)(b) VCLT can still be relevant for the purposes of interpretation (with a similar logic to be equally applicable in other parts of Article 31, such as 31(3)(c)), which was touched upon in the Whaling case, by then Judge ad hoc Charlesworth and more clearly identified by the ILC in its work on subsequent practice; ⁴³ and also

• whether non-conventional rules abide to the same ‘rules’ of interpretation as treaty rules. ⁴⁴

But it is in this gradually etched out map that the discussions of the ILC and IDI, and all debates surrounding interpretation, need to be situated.

5. Hoping that the road to Ithaca is a long one

Such discussions would not be possible without the existence of the rules of the VCLT, on the basis of which or against which the arguments are being made. That is not to say, that post-VCLT, we live in a utopian interpretative paradise, where courts and tribunals always get it right, ⁴⁵ far from it. After all, the etymology of the word utopian, which is a compound word from the ancient Greek terms ‘oὐ’ and ‘τόπος’ ⁴⁶ means ‘no place’, or in context a place that does not exist. ⁴⁷ In fact, it is for this reason that the examples provided above were chosen. As instances where courts and tribunals despite using the accepted common vocabulary erred methodologically. But it is exactly this point that highlights the importance of debating and clarifying the content of the ‘rule/s’ of interpretation, as it pushes interpreters to justify their interpretations rather than merely assert them, and such justification happens within a frame provided by the ‘rule/s’ of interpretation. Through this iterative and incremental process of refinement, if not clarification, not only courts and tribunals but any interpreter can be and is held to a higher standard of argumentation, reasoning and justification.

To add to that, and again irrespective of where one stands on the existence or not of a correct interpretation and of ‘rules’ of interpretation in the strict sense, such refinement is not without its consequences, as it helps delineate with increasing clarity what Kelsen called the ‘frame’ of the rule being interpreted. According to his ‘frame theorem’ the result of interpretation ‘can only be the discovery of the frame that the norm to be interpreted represents and, within this frame, the cognition of several possibilities for implementation’. ⁴⁸ However, such a frame changes not only with the passage of time, but also with the gradual refinement of the ‘rule/s’ of interpretation. In this manner the contours of the rule being interpreted come gradually into greater focus.

In this process of gradual refinement of the ‘rules’ or of the arguments surrounding interpretation, it is not the destination but the journey that matters. Or, and let me close on this note, as has been infinitely more eloquently put by Kafavis in a poem inspired by the theme of Odysseus’ journey: ⁴⁹

• As you set out for Ithaka hope your road is a long one, full of adventure, full of discovery …

• Keep Ithaka always in your mind. Arriving there is what you’re destined for. But don’t hurry the journey at all. Better if it lasts for years, so you’re old by the time you reach the island, wealthy with all you’ve gained on the way, not expecting Ithaka to make you rich.

• Ithaka gave you the marvelous journey. Without her you wouldn’t have set out. She has nothing left to give you now.

• And if you find her poor, Ithaka won’t have fooled you. Wise as you will have become, so full of experience, you’ll have understood by then what these Ithakas mean. ⁵⁰