AJIL Symposium: Is the Vienna Convention Hostile to Drafting History? A Response to Julian Davis Mortenson, Part 1

[Dr. Ulf Linderfalk is a Professor of International Law at the Faculty of Law at Lund University, Sweden.]

Julian’s article focuses on a single proposition (p. 780)

“[W]hen an interpreter thinks a text [of a treaty] is fairly clear and produces results that are not manifestly unreasonable or absurd, she ought to give that prima facie reading preclusive effect over anything the travaux [préparatoires] might suggest to the contrary.”

Specifically, Julian argues (p. 781), that this proposition – while today shared by an overwhelming majority of international judiciaries and legal scholars – “cannot be reconciled with the agreement actually reached in 1969” and embodied by Articles 31 and 32 of the 1969 Vienna Convention on the Law of Treaties (VCLT).

In critically assessing Mortenson’s article, I find that it builds on three assumptions:

• In the final analysis, the legally correct meaning of a treaty is determined by the intention of its parties. Thus, when interpreting a treaty, the ultimate purpose is to find out how the original parties to the treaty actually intended it to be understood.
• Articles 31 and 32 of the VCLT guide interpreters to discovering the common intention of treaty parties. Thus, ordinary meaning, context, preparatory work, and other means of interpretation help interpreters understand the legally correct meaning of a treaty.
• A detailed analysis of the preparatory work of the Vienna Convention is an appropriate method for a scholarly analysis of the legally correct meaning of Articles 31 and 32 of the VCLT.

As I will explain in my two posts for this Symposium, I think all three of Julian’s assumptions are either fundamentally flawed or seriously debatable. Readers with a particular interest in issues of treaty
interpretation might want to consult the slightly more elaborate working paper that I have recently posted on the SSRN.

What is the concept of party intention assumed by the VCLT?

The ultimate aim of a process of interpretation, as described by Articles 31-32 of the VCLT, is to establish the legally correct meaning of the interpreted treaty. By the legally correct meaning of a treaty, international lawyers generally understand it to mean the communicative intention of the treaty parties, that is to say, the meaning which the parties intended the treaty to express. The relevant provisions of the VCLT certainly confirm this approach to treaty interpretation. However, contrary to the assumption of Julian Mortenson, as clearly indicated by the wording of Articles 31-32 of the VCLT, it is not necessarily the communicative intention of the original parties to a treaty that the interpretation process aims to determine. The aim, rather, is to determine the communicative intention of those states, which are parties to a treaty at the time of interpretation. Any other understanding of the VCLT would render Article 31, paragraph 3 inexplicable, considering the definition in Article 2, paragraph 1(g) of a party as “a State which has consented to be bound by the treaty and for which the treaty is in force”. If it had been the ultimate purpose of the interpretation of a treaty to establish the communicative intention of only those states that took part in its negotiation and final conclusion, then why is it a requirement that a subsequent practice, for instance, establishes the agreement of all treaty parties, which in the case of a multilateral treaty open for general accession typically include also states that have entered the treaty at a later stage? It does not make sense.

Now, about the assumption adopted by Julian that is the ultimate purpose of the interpretation of a treaty to determine the actual communicative intention of its parties, I can only refer to pragmatics. According to this branch of linguistics, an utterance can be understood only on the assumption that whoever produced it acted rationally. That is to say, the addressee of the utterance has to assume that the utterer (whether a writer or a speaker) acted in conformity with some particular standard or standards of communication. Like pragmatics, I will refer to any such assumption made by an agent in the interpretation of an utterance as a communicative assumption.

The interpretation of a treaty is no different than the understanding of just any verbal utterance produced by a person or group of persons, whether orally or in writing. To think otherwise would seem somewhat naïve. Naturally, different kinds of communicative assumptions may be relevant for the interpretation of utterances depending on such things as, for instance, the functional or social reason causing them. To give just a few examples, in the specific context of treaty interpretation, as can be seen from the practice of international courts and tribunals, law-applying agents operate on assumptions such as the following:

• that treaty parties arranged things so that the treaty conforms to the lexicon, grammar and pragmatic rules of the language used for every authenticated version of it.
• that treaty parties arranged things so that a consistent meaning can be conferred on all words and lexicalized phrases used in the interpreted treaty.
• that treaty parties arranged things so that no parts of the treaty comes out as redundant.
• that treaty parties arranged things so that the application of the treaty results in the realization of its object and purpose.
• that treaty parties arranged things so that the treaty does not derogate from any other international legal norm applicable in the relationship between them.
• that treaty parties arranged things so that the treaty corresponds to whatever can be inferred from the subsequent practice developed in its application rather than to whatever can be inferred from its preparatory work.

Obviously, the role of the various means of interpretation listed in Articles 31-32 of the VCLT is to permit assumptions such as those just stated.

The point of this methodological detour is that it tells us something important about the ultimate purpose of the treaty interpretation process, as described in Articles 31-32 of the VCLT. Obviously, if Articles 31-32 presuppose the existence of a series of communicative assumptions, then for their application it is immaterial whether in the case of the interpretation of a particular treaty provision relative to a particular set of facts, the treaty parties had any actual communicative intention or not. A proposition about the meaning of treaty provision can be legally correct although in the final analysis it remains an assumption, the truth of which cannot possibly be verified. In the conceptual universe of the VCLT, consequently, the communicative intention of treaty parties is a rational construct.

Some commentators of the VCLT, and among them Julian, express themselves as if ultimately, Article 31 was applied to determine the meaning of the text of a treaty, and if because of this, an argument based on studies of preparatory work was in some sense truer of the communicative intention of the treaty parties. To illustrate, Mortenson writes on page 787:

“[A] failure [of drafting history] to confirm might show that the provisional interpretation would produce absurd results in the sense of Article 32(b) – not necessarily absurd in the abstract (if the VCLT rejected anything, it was the idea of interpretation in the abstract) but absurd measured by the thing that the interpreter now knows that the parties were trying to achieve”.

This idea builds on a misunderstanding of the organizational structure of the VCLT and the concept of party intention. First, conventional language, context, the object and purpose of a treaty, as well as preparatory work, are all means to determine the communicative intention of the parties. The determination is done differently, of course. Using conventional language, for example, the communicative intention of the parties is determined based on the assumption that the parties arranged things so that the treaty conforms to the lexicon, grammar and pragmatic rules of the language used for every authenticated version of it. Using preparatory work, the communicative intention is determined based on the assumption that the parties arranged things so that the treaty corresponds to whatever can be inferred from its preparatory work. Second, no argument based on any means of interpretation is truer of the communicative intention of the parties to a treaty than any other. If the concept of the communicative intention of the parties is a rational construct, the truthfulness of an argument favoring a particular understanding of a treaty is not an issue. Certainly, an argument based on Article 31 often forms a stronger reason for the adoption of a meaning than an argument based on preparatory work, but that is an entirely different thing.