AJIL Symposium: The Travaux of Travaux

[Julian Davis Mortenson is Assistant Professor of Law at Michigan Law]

It is often asserted that the Vienna Convention on the Law of Treaties relegates drafting history to a rigidly subsidiary role in treaty interpretation. Many commentators go so far as to suggest that the VCLT entrenches a categorical prejudice against travaux préparatoires—the preparatory work of negotiation, discussions, and drafting that produces a final treaty text. Because of this alleged hostility to history as a source of meaning, the conventional wisdom is that when an interpreter thinks a text 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 might suggest to the contrary.

As The Travaux of Travaux: Is the Vienna Convention Hostile to Drafting History? demonstrates, however, this conventional wisdom cannot be reconciled with the agreement that was actually concluded in 1969. Careful analysis of the multi-decade process that led to the VCLT shows that, far from adopting a doctrinally restrictive view of drafting history, the Vienna Conference sought to secure the place of travaux as a regular, central, and indeed crucial component of treaty interpretation. In reaching this conclusion, the article draws on a range of published and unpublished sources, including minutes from meetings of the Institut de droit international, the International Law Commission, the UN General Assembly in both its plenary and Sixth Committee sessions, and the Vienna Conference itself; internal memoranda and other documents circulated at each of those institutions; and proposed drafts and amendments that were submitted throughout the process.

It is true (and likely a source of modern confusion) that Vienna Conference delegates rejected a U.S. proposal to formulate the rules of treaty interpretation as a totality-of-the-circumstances balancing test. But that had nothing to do with hostility to travaux as such, much less with any desire to impose strict threshold requirements on their use. Rather, the delegates were rejecting Myres McDougal’s view of treaty interpretation as an ab initio reconstruction of whatever wise interpreters might view as good public policy. They objected to the purpose for which New Haven School interpreters wanted to use travaux—not to drafting history as a source of meaning per se.

To the contrary, the drafters repeatedly reiterated that any serious effort to understand a treaty should rely on the careful and textually grounded resort to travaux, without embarrassment or apology. They themselves leaned heavily on travaux when debating any legal question that turned on the meaning of an existing treaty. And each time a handful of genuinely anti-travaux delegates attempted to restrict the use of drafting history to cases where the text was ambiguous or absurd, those efforts were roundly rejected.

The understanding that emerged was of interpretation as a recursive and inelegant process that would spiral in toward the meaning of a treaty, rather than as a rigidly linear deductive algorithm tied to a particular hierarchical sequence. In any seriously contested case, interpreters were expected automatically to assess the historical evidence about the course of discussions, negotiations, and compromises that resulted in the treaty text—in short, the travaux. The modern view that Article 32 relegated travaux to an inferior position is simply wrong. The VCLT drafters were not hostile to travaux. They meant for treaty interpreters to assess drafting history for what it is worth in each case: no more, but certainly no less.