Syria, Chemical Weapons, and the Incoherence of the VCLT

My friend Dapo Akande has a superb post at EJIL: Talk! discussing whether the ICC could prosecute the use of chemical weapons by the government in Syria. I agree almost entirely with Dapo’s analysis, but I do want to offer a couple of thoughts about his discussion of the Vienna Convention on the Law of Treaties:

The argument that chemical weapons are not covered by Art. 8 is thus based on the removal of the explicit prohibition and the fact that it was thought that it would be the annex to the Statute that would set out those weapons like chemical weapons that are deemed are indiscriminate and cause unnecessary suffering.

However, it is erroneous to interpret a treaty primarily by reference to drafting history. Under Art. 32 of the Vienna Convention on the Law of Treaties, the drafting history is only to be used as a supplementary or secondary tool of interpretation to resolve ambiguity. As Art. 31 of the VCLT indicates and ICJ has stated, the interpreter must start with the text of the treaty. Thus one must start by looking to see what the words of the treaty as agreed actually means.  It seems to me that the words “poison or poisoned weapons” and more clearly “asphyxiating, poisonous or other gases” would cover a variety of chemical weapons. The latter wording is taken from the Geneva Gas Protocol which was intended to cover chemical weapons. Although the Chemical Weapons Convention of 1993 does not use this wording (referring instead to “Toxic chemicals”), this does not mean that the wording of the ICC Statute does not extend to chemical weapons. Although weapons not in the form of a gas are not covered by para xvii, it is also arguable that ‘poison’ and ‘toxic chemicals’ are largely synonymous.

I have no doubt that Dapo is correctly applying the VCLT. But that means — as I have long thought — that the VCLT’s approach to treaty interpretation is incoherent. Treaties are an expression of sovereign will; states are free to draft treaty provisions however they want, subject only to peremptory norms of international law, and no state is required to ratify a treaty that does not reflect its values, even one it played an important role in drafting. States should thus be able to expect that the VCLT will not lead judges to interpret critical treaty provisions in a manner that runs directly counter to their intended meaning.

Yet that is precisely what the VCLT seems to allow with regard to chemical weapons. No one contests the idea that numerous states in the Global South would not have ratified the Rome Statute if Art. 8 had directly criminalized the use of chemical and biological weapons (the poor man’s WMD), but not nuclear weapons (the rich man’s WMD). That is why the explicit prohibition on their use was removed during the drafting of the Rome Statute. So what possible justification could there be for the VCLT analysis Dapo provides above, one that ignores the actual intent of the Rome Statute’s drafters in favor of a mechanical application of dictionary definitions? That analysis quite literally says the following to all of the states that opposed including chemical weapons in Art. 8: “sorry, we know you would have refused to ratify the Rome Statute if it had directly criminalized chemical weapons, but Art. 8 criminalizes chemical weapons anyway because you allowed a word to remain in Art. 8 (“poison”) whose dictionary definition can be construed to include them. Better luck next time the international community creates a permanent international criminal court.”

I’m not exaggerating. That is precisely what the VCLT says to states in the Global South if Dapo’s analysis is correct — which, again, I’m sure it is, because he knows more about the VCLT than I ever will. That said, I can imagine at least one analysis that is both arguably VCLT-consistent and would not directly undermine the sovereign will of the numerous states that ratified the Rome Statute thinking it did not directly criminalize the use of chemical weapons. That analysis would emphasize Art. 31(4), which provides that “[a] special meaning shall be given to a term if it is established that the parties so intended” — a provision that indicates the drafters of the VCLT recognized, however imperfectly, that it is impossible to separate the “ordinary meaning” of a word from its intended meaning. Given that Art. 8 was specifically drafted to exclude a direct prohibition on chemical weapons, it seems clear that the drafters intended “poison” to have a special meaning, one that did not include chemical weapons.

I agree with Dapo that, under the VCLT, “it is erroneous to interpret a treaty primarily by reference to drafting history.” But it shouldn’t be. When the intent of the drafters is clear, there is no conceivable justification for ignoring it in favor of “ordinary meaning.”