Book Symposium: Accounting for Difference in Treaty Interpretation over Time

[Julian Arato is an Associate-in-Law at Columbia Law School.]

Interpretation in International Law is something of an iconoclastic volume, from its critical ethos to its provocative structure around the metaphor of the game. The object of its revisionism, above all, is an apparently stagnant formalism that seems too prevalent in the theory and practice of interpretation in international law today. Symbolic of this antiquated formalism – for the editors and for many of the contributors – are the rules of interpretation embedded in the Vienna Convention on the Law of Treaties (VCLT). There is something funny about defending the rules in such a collection – like arguing for reform in a volume on revolution.

My chapter, ‘Accounting for Difference in Treaty Interpretation over Time’, embraces the Vienna rules in principle. I reject neither rules nor formality, and I happily hew toward the VCLT’s project of bringing order to the practice of interpretation. In my view, there is real value in the regulatory ideal that a “correct” interpretation of any treaty exists. The Vienna rules reflect a crucial language for approaching that ideal. The rules don’t simply provide access to the correct interpretation – they help constitute it. And, crucially, they do so according to a scheme accepted by states, the plenary subjects of international law.

Still, in an important sense I agree with the volume’s critical spirit. In my view the real problem lies not with the Vienna rules themselves, but with a reverence for them that sometimes borders on the fetishistic. It’s not the rules, in other words, but the way they tend to be received. The misstep lies in the notion that the Vienna rules comprise a fully self-contained approach to treaty interpretation – universally applicable to all treaties, and always in the same way. This article of faith is neither borne out in practice nor in theory, and nowhere is this more evident than in the interpretation of treaties over time.

All students of public international law are at some point taught that as a matter of doctrine all treaties are subject to the same unified rules of interpretation. Yet at the same time, everyone knows that some treaties are special. Time and again we hear that some kinds of treaties are different. Courts, tribunals, and scholars often intone that certain treaties are entitled to special treatment when it comes to interpretation, especially as regards changes of circumstances and intentions attending the passage of time.

Often the argument is that human rights treaties are somehow special – capable of progressive evolution over time, with or without the continued consent of the parties. By the same token, it is sometimes suggested that subsequent party practice, usually an authentic criterion for the interpretation, reinterpretation, or even the modification of treaties, is somehow of less value in the context of interpreting treaties that confer rights on natural persons. The argument is that in such cases the mastery of the parties over time is somehow reduced. We hear similar statements about environmental treaties, territorial treaties, and of course the constituent instruments of international organizations. The problem is that the explanations offered for differentiating among types of treaties are rarely satisfactory.

The argument that a particular treaty provision is entitled to special treatment tends to be explained in one of two ways, both of which prove ultimately unsatisfying. One approach simply invokes the general subject matter of certain treaties in singling them out for special treatment. In light of their special subject matter, the argument goes, treaties on subjects like human rights or the environment should be understood as insulated from the changing will of the parties, and sometimes capable of autonomous evolution. But such statements cannot withstand serious criticism. In the words of the ILC’s Fragmentation Report (at [21]): “characterizations (‘trade law’, ‘environmental law’) have no normative value per se … The characteristics have less to do with the ‘nature’ of the treaty than the interests from which it is described.”

A second school of thought tries to work within the Vienna rules – focusing especially on one criterion of interpretation, object and purpose, in accounting for ascribing differential weight to the other codified rules. On this view, the touchstone must always be the intention of the states parties, as reflected in the goals that the treaty seeks to achieve. Adherents of the object and purpose approach have the advantage of formality. Its proponents can argue that the answers lie within the rules after all, in the invocation of object and purpose at VCLT Article 31(1). And yet something important still seems to be missing. Even where a treaty’s goals are sufficiently determinate, the interpreter must still ask how far the parties were willing to go to achieve their goals. Though a treaty may enshrine certain values, it remains critical to ask to what extent the states parties intended to entrench those values – to what extent, in other words, they agreed to tie themselves to the mast.

My suggestion is that international lawyers’ tendency to focus doggedly on the canons of interpretation codified in the VCLT draws attention away from a crucial consideration in the interpretive puzzle: the nature of the treaty obligations under interpretation. What gets left out is any inquiry into how far the states parties intended to commit themselves in acceding to a treaty obligation.

Put more schematically, the critical issue elided by the Vienna rules is whether a treaty provision entails a merely reciprocal exchange of rights and duties, or rather incorporates a more absolute commitment by the parties to take on an obligation insulated from their changing intentions, and over which their subsequent mastery might prove relatively limited. Some treaty norms represent mere exchanges of rights and duties, wholly dependent on mutual performance. If one party breaches its obligations, the other is well within its rights to do the same. Other norms represent a shared commitment to abstain from, or engage in, a certain behavior – whatever the other parties do. It is well understood that differences in the level of party commitment are relevant to determining the consequences of treaty breach, or to resolving conflicts with subsequent treaties. And indeed, as Pauwelyn has noted, the VCLT itself recognizes the importance of drawing distinctions between different types of treaty norms for these non-interpretive purposes. I argue that this distinction is just as central to the resolution of problems of interpretation over time.

Starting from the perspective that the Vienna rules are essential, I suggest that the problem of differential interpretation over time reveals a deficiency in how we think about VCLT Articles 31–32. Specifically we ought to avoid treating it as a total interpretive mechanism. The problem lies not in our interpretive rules as such, but in the assumption that any set of rules can do all the work. Without arguing for anything like revising or abandoning the VCLT, I suggest that the process of interpretation may sometimes require taking into account considerations left unmentioned by Articles 31–32. Distinguishing between types of obligations based on the level of the parties’ commitment is a case in point. Doing so helps account for quite a bit of interpretive practice that would otherwise appear anomalous under the Vienna rules; hopefully it can provide a more principled justification for differential interpretation going forward.