This is a non-substance post just to say thanks from all of us to Richard Gardiner and all the commenters for offering, and to Duncan for organizing, the treaty interpretation symposium this past week. It was marvelously intellectual and subtle and, even having read the book, I am still reading the posts carefully. I think they will be read and cited for a long time in the academic literature, and I am certainly pleased to say that they are up on Opinio Juris.
As we come to the end of our discussion of Richard Gardiner’s book and the subject of Treaty Interpretation more generally, I wanted to thank our guest bloggers and, of course, Richard, for what has proven a sophisticated (and hopefully useful) discussion. I have thoroughly enjoyed the examination Isabelle, Malgosia, Jan and Richard brought to bear on issues about the New Haven School’s continued salience, dynamic treaty interpretation, prepatory work, the VCLT as rules or standards, etc. My only regret is that there were many more topics we could have addressed but did not — e.g., the role of unilateral statements under Article 32; questions on whether our method of interpretation should shift depending on the function it serves (whether articulating national or international law or perhaps enabling conduct versus restricting it), and the problems multiple languages pose. In any event, I hope that we’ll have future opportunities for such dialogue, and that our readers will feel free to weigh in with further views and comments of their own.
Treaties constitute one of the largest effective components of modern international law. Lawyers routinely have to give meaning to their terms. Mostly this is straightforward, but often enough there are interesting difficulties in deciding what the terms mean and how they apply in novel or unexpected situations.
I have found all the posts and comments extremely interesting, and I am grateful to those who have participated in this discussion. I think that the discussion has reinforced my opening remark that too much should not be expected of the Vienna rules. However, the rules do need to be considered as a starting point for treaty interpretation, giving at least an infrastructure.
As provisions in a treaty, the rules are not themselves all easy to interpret. Professor McDougal pointed this out at the 1969 Vienna conference noting the presence there of the ILC Special Rapporteur (Waldock) as “the best testimony, not always mute” of the impossibility of applying the textuality approach, and that reference to Professor Waldock had often been necessary in tribute “not to his skill in flipping the pages of a dictionary or as a logician, but rather to his very special knowledge of all the circumstances attending the framing of our draft Convention.”
I do, however, agree with Isabelle, that practice in the application of the Vienna rules has shown them not to shut the door on much of what McDougal sought. I think it emerges from the discussion what significance attaches to the further skills required once the relevant material has been identified in accordance with the Vienna rules. These are the skills deployed in evaluating the material and shaping interpretative arguments. However, they are skills which cannot readily be listed, still less reduced to further rules. Nevertheless, at the very least, it may prove to be helpful if treaty interpretation is based on a common foundation in the Vienna rules.
It seems that our discussion these past few days confirms that 40 years after the Vienna Conference the meta-questions relating to treaty interpretation remain unaltered. But perhaps it also appears that the ILC ultimately made the right decision to codify the relatively few basic principles on which agreement could be found. The Commission always made it clear that the effect of its codification was not to take away the freedom of interpreters to assess a range of considerations in reaching a conclusion on how to interpret treaty language. With varying degrees of success, international courts and tribunals have responded well to these principles, using them as guidance and justification, as tools to build credibility and exercise and assert their judicial function (sometimes in an institutional context), as instruments to achieve accountability, as techniques to order and structure their reasoning process, as means to make their decisions acceptable and comprehensible.
Practice informs that the ILC’s codification exercise might have served more useful purposes then the Commission initially expected, and in that sense the choices made by the ILC in codifying these principles and the position of the New Haven School might not have been as irreconcilable as originally perceived.
I much enjoyed reading the views of Duncan Hollis on the art element in treaty interpretation and on auto-interpretation. I agree that who it is who is making an interpretation may play a key part in the outcome. That two tribunals arrived at different interpretations of similar “umbrella” provisions in bilateral investment treaties shows this all too vividly (SGS v Pakistan and SGS v Philippines).
The discussion of Leo Gross’s theory of auto-interpretation adds what may be a dimension of realism, reflecting the most common circumstances of treaty interpretation. It may help explain the tolerance of the divergent interpretations in national courts of the “uniform” provisions on carriage by air described in my previous comments. But, absent specific indications favouring auto-interpretation (such as the fall-back to national definitions for terms undefined in double taxation agreements following the OECD model), auto-interpretation seems an uneasy fit with the notion of agreement (unless one accepts the generality of Philip Allott’s description of a treaty as “a disagreement reduced to writing”).
The question put to the ILC on the art of treaty interpretation was “whether there were any rules for practising that art”. The end result was the Vienna rules. Certainly these only amount to a rather loose framework. Switching to the analogy of musical composition as an art form, there are understandings as to what a fugue or sonata is, and their composition can be seen as being to some extent governed by rules; but within either form very different results can be produced. To the extent that there are rules for treaty interpretation, they constitute a springboard rather than a straightjacket.
The ILC seems to have adopted the term “rules” rather casually, their Special Rapporteur (Waldock) musing that “[i]n a sense all ‘rules’ of interpretation have the character of ‘guidelines’…”. The Commission took more care in explaining why the content of Article 31 was described as the singular “general rule”; but elsewhere in the Vienna Convention, selection of the terms “rule” and “rules” does not suggest profound consideration of terminology. Use of the description “the Vienna rules” seems generally a convenient shorthand for the three articles forming the set of relevant provisions. Their content is more of an indication of what is to be taken into account when interpreting a treaty, with only some hints of how to achieve an interpretation – hints left to be gleaned from quite sparse indications.
I have to plead guilty to hedging my bets on whether there is a “correct” interpretation to be given to a treaty provision. Stating that “no claim is made that the Vienna rules … lead directly to a necessarily correct result in every case”, and suggesting that further guidance is required “to set the ground for a ‘correct’ result, or at least one which has been correctly ascertained” (pp 6-7), I had in mind that the Vienna rules provide an agreed starting point for interpretation and something of a framework for the interpretative process.
My initial aim was, in fact, even more basic – to persuade interpreters to consider all relevant elements in the Vienna rules. Some otherwise quite respectable courts and tribunals used to give the impression that they saw the opening reference in the rules to “ordinary meaning” as a green light for a literal interpretation, perhaps dipping their toes timidly in the waters of context and purpose, giving a nod to preparatory work, and ignoring the rest. In assessing and applying the Vienna rules it is at least necessary to check out the whole package.
I am actually not sure whether it makes much sense, as Isabelle fleetingly suggests, to think of articles 31-32 in terms of either rules or principles. Following e.g. Rosenne, I tend to think that they lack any ‘norm-creative character’ (to use the phrase from the ICJ’s 1969 North Sea Continental Shelf cases), and are best seen as methodological devices: as instructions to whoever gets to be in a position to interpret a text. And methodological devices tend to be, so to speak, non-imperial: there are various different ways to engage in many activities, without it being possible to specify which one would be the best – this may well be why so many cookbooks are being produced…. Consequently, if it is not plausible to think of articles 31-32 as rules, it is not plausible to think of them as customary rules either (despite anything the ICJ may have said on the topic).
This does not mean that anything goes: interpretation is not an open-ended ‘free-for-all’. Typically, within the professional group of international lawyers, some arguments will be deemed more acceptable than others. Indeed, the same holds true within the sub-regimes, which may help explain why some in the human rights community were shocked (or mildly upset) by decisions such as the EC’s Court of First Instance decisions in Kadi and Yusuf a few years ago. The upset responses were not the result of the CFI somehow misapplying established rules of interpretation or applying the wrong rules of interpretation, but rather stemmed from the CFI’s unfamiliarity with human rights law. What may have made sense as a matter of EC law (one may have one’s doubts on this though), turned out to be less sensible from the viewpoint of the human rights community. The CFI ended up applying a body of rules without having acquired the sentivities of those who would normally be working with these rules, and as a result came up with readings of the right to property or access to justice that will not impress your average human rights lawyer.
By the same token, many in Europe at least have expressed some dismay at the US Supreme Court’s approach to consular cases, as highlighted by Duncan. These stem not so much (or not solely) from different methods of interpretation, but rather from the Supreme Court’s reluctance to adopt (or at least familiarize itself with) the sensitivities of international lawyers. This may be politically justified, perhaps; and it may be legally justifiable, perhaps, but those are different matters. As an explanation, the idea of there being different ‘invisible colleges’, in Schachter’s famous phrase, or different ‘interpretive communities’, as Stanley Fish puts it, seems to make some sense.
That also suggests that perhaps the most proper methodological device is that interpreters, judicial or otherwise, should show some humility and at least try to understand the mindset prevailing amongst the communities engaged in the body of law they are about to apply. An interesting exploration hereof is a recent study by H. Jefferson Powell with respect to the US Supreme Court (I forgot the title, and am not in a position to check right now). If one insists, one may see articles 31-32 as appeals to do just this, given their reliance on context, and the injunction to read in good faith.
As Jan suggested at the end of his post, it seems that the broader question underlying our debate is how principles of treaty interpretation are used to empower international courts and tribunals and what the limits are of that function.
On interpretation as a rule-governed activity, it seems that we all approach Articles 31 to 33 VCLT as principles rather than rules, despite the fact that they are formally treaty law reflecting customary international law on the matter. But it seems that this is not necessarily (anymore) the common perception of many commentators on treaty interpretation. This impression is based on many discussions and writings about the interpretation Article 31(3)(c) VCLT, characterized by a high-level of technicality and very much the belief that this is ‘a rule’. I’ve written elsewhere that Article 31(3)(c) VCLT is perhaps the only mistake in Articles 31 to 33 VCLT, in the sense that the value and function of the principle of harmonious interpretation might have been better served had the ILC decided not to codify the principle, for the same reasons why they decided not to codify the principle of effectiveness (which is widely applied, especially by courts and tribunals operating within an institutional context and interpreting continuously the same treaty text). The ILC made it clear that it did not intend to ‘exhaustively’ codify the matter of interpretation; it merely intended to state the basic principles based on the available practice at that time.
On the function of principles of treaty interpretation, I still think that they, on the one hand, help guide and structure the reasoning process, but, on the other hand, also help to justify the conclusion a court or tribunal reaches on the meaning of the treaty (to some extent, for whatever reason). They help make this conclusion acceptable for its audience (which introduces another contextual element). But their role cannot entirely be reduced to merely one of post hoc rationalization.
As international courts and tribunals multiply and judicial review of States’ measures and laws in the light of their international obligations increases, Articles 31 to 33 VCLT become more attractive. They seem to offer some level of objectivity, making the applicable standard of review more acceptable – especially at the earlier stage of a court’s existence when it might still be developing that standard of review (depending on the applicable procedural rules). Perhaps this is the start of a complex answer to Jan’s question.
I should also clarify that my earlier comment on the mild trend from formalism to informalism in justifying treaty interpretation in the WTO applies to the Appellate Body. It appears that panels have yet to fully respond to this development.
First off, I should say how much I admire Richard Gardiner’s book; it has already earned a prominent spot on my bookshelf and I expect it will become a regular reference work for me in any future interpretative exercises. At the same time, I have lots of questions and comments about the book and the growing density of the VCLT system it purports to describe. Let me start here with the premise that Richard himself uses to introduce the book–the idea that treaty interpretation is an “art” and that the VCLT now provides us with “rules for practising that art.” (p. 5). In doing so, Richard suggests that we can employ the VCLT rules to achieve a “correct” result (p. 6) and, more specifically, that the result should be a “single autonomous interpretation.” (p. 30). Accepting the proposition that treaty interpretation is an art, however, leads me to ask the normative question of whether as an art it should strive to some, single common approach for each case, not to mention the descriptive problem of whether states, international organizations, international tribunals, and other subjects of international law have actually done so.
Simply put, I wonder what the artistic axiom — that beauty lies in the art of the beholder — does for our art of treaty interpretation. What constitutes a “good” interpretation of a treaty may be as difficult to agree upon as what constitutes good “art.” Indeed, I see the question of the continuing salience of Prof. McDougal’s work (or the concept of textuality raised by Professor van Damme) as essentially a debate over which treaty interpretation techniques we should celebrate and which we should disapprove. We might analogize it to debates among various schools of art. Do we consider photorealism (i.e., textualism) to be better than abstract art (i.e., the New Haven School)? Or, is the answer somewhere in between a la impressionism (i.e., the VCLT rule)?
Perhaps Prof. Gardiner’s point is simply that treaty interpretation has evolved to the point where the community of interpretators has agreed upon a single, acceptable technique for our art, namely the VCLT. It’s become, to use Jan Klabber’s phrase, our battlefield for discourse. I’d agree with such descriptions to a point–would any international lawyer contest the necessity of consulting the VCLT rule as the start of an interpretative exercise? But, even assuming that’s true, I wonder about the assertion in at least other three respects.
First, is it true that the VCLT sets the whole battleground for debate, i.e., is it the only technique or simply the dominant one. It certainly could be an exclusive method if read broadly. For example, I for one have never found Article 32 too troublesome on the question of using prepatory work. I find plenty of space to “confirm” or “determine” whatever meaning a good faith interpretation of the text’s ordinary meaning in context produces. Now, some might say I go too far, but here I agree with Jan that most lawyers will do this work and marshal the VCLT to confirm it whenever that research favors them (and invoke the “subsidiary” nature of the exercise when it does not).
Second, I wonder how, if at all, we should conceive of the VCLT and petitions for “special” interpretative rules. Are they simply variations on the VCLT, or competitors with it? In particular, can we say the VCLT drives both the ECtHR’s rulings that hint at dynamic treaty interpretation with the WTO’s that don’t, because of special contexts–human rights vs. trade–that lead us to emphasize certain components of the VCLT rule over others. Might similar reasoning justify broader readings of constituent instruments of international organizations even as it would generate strong objections if applied in a bilateral context?
Third, even if the VCLT is, so to speak, the governing school of our art, I’m not sure it follows that it requires a single, common, autonomous intepretation in every instance. On the contrary, the “crucible” approach adopted by the ILC (and Prof. Gardiner) leaves lots of different moving parts in play to produce competing “reasonable” interpretations of a treaty’s text. Here, I’d cite another idea percolating around at the time the VCLT and the New Haven school gained prominence–Leo Gross’s theory of auto-interpretation. Gross argued against too quickly disassociating the subject of interpretation (i.e., the treaty) from the interpretator (i.e., states, IO, or international tribunal). In other words, worry less about the interpretative rule than finding a way to get someone to authoritatively apply it. For Gross, the horizontal nature of the international legal order frequently hinders the prospects of doing this. Parties can always agree on one ”authentic” treaty interpretation, just as they can delegate to some international organization or tribunal authority to pronounce it. And certainly, examples like the ECtHR or the ICJ reflect the product of such agreements. But, in most cases states have not agreed to common or third-party interpretations. More often, a state retains the right to self-interpret what the treaty means, i.e., an “auto-interpretation.” And other states and subjects of international law can do the same. The international legal order is thus more than likely to produce competing, unresolvable, auto-interpretations of a treaty’s text than any single, common interpretation. The VCLT may have improved things by forcing states to adopt a common framework (and language) for their disagreements, but it alone can’t move subjects of the law beyond that framework to some, actual ”correct” interpretation, however, dense the various components of the VCLT rule have become in recent years.
Indeed, I wonder if Gross didn’t overestimate the value of authoritative interpretations by third parties to produce a single, autonomous meaning. Witness the U.S. Supreme Court’s disagreements with the ICJ over the Vienna Convention on Consular Relations’s requirements of a remedy for a failure to provide consular notification under Article 36. Some argue that the ICJ’s interpretation is authoritative, and the United States apparently agreed at least with respect to certain named Mexican nationals even as it found domestic law lacking a method to comply. But, beyond those named parties, the Supreme Court believed the ICJ got it wrong, and offered its own competing vision of the article’s “correct” interpretation. Who’s right? Isn’t a decision as to which side gave the ”right” interpretation essentially a function of who we ask? (in fact, won’t those we ask answer based mostly on pre-conceived notions of whether international and domestic tribunals should have a vertical, horizontal, or diagonal relationship than the actual interpretations each employed?). In the end, is there any way to achieve some finality of interpretation, absent an agreed reconciliation negotiated by the United States with Mexico, or the ICJ itself? Otherwise, as much as I enjoy treaty interpretation, it remains a pretty subjective exercise, with each of us having our tastes and views of what’s good and bad work, even if we use only the VCLT to judge it.
First, my thanks for Malgosia, Isabelle and Jan for all their comments.
The work of the ECtHR has an interesting aspect in the present context. One of the first to take up the Vienna rules systematically, the Court (as Malgosia shows) has apparently not found them to constrict its development of a distinctive line of case law appropriate to the Convention which the Court applies.
Isabelle’s first post also addresses case law of a single institution, indicating that the WTO Appellate Body has in effect reflected one of the central ideas of the ILC in formulating the first part of the general rule – that the ordinary meaning of terms has to be drawn as much from the context as from the dictionary (ie not simply seeking what Judge Higgins critically described as ‘a mythical “ordinary meaning”’, Kasikili/Sedudu Island (Botswana/Namibia)).
Isabelle also takes us deep into the difficulties over preparatory work. It will not always be possible to treat silence on a particular point as “constructive ambiguity”; nor will it always be clear that a matter was intentionally excluded from a treaty. A contemporary example is brewing up in different circuits of the US Court of Appeals.
The Warsaw Convention provides that claims relating to accidents and incidents in carriage by air must be brought in a court in one of four specified places (destination, domicile of carrier etc) “at the option of the plaintiff”. The next provision states that “questions of procedure” are for the law of the place of the court to which the case is submitted. Can that court reject the choice of the plaintiff, finding itself inappropriate (forum non conveniens)? Is the right offered by the treaty to the claimant to opt for a jurisdiction a matter of procedure, or an exercise of that option subject to being overruled as a matter of procedure?
Courts have differed on this, even before the 1999 Montreal Convention added a fifth jurisdictional possibility. Some negotiators argued in 1999 that the safeguard against improper use of the fifth jurisdiction would be the doctrine of forum non conveniens. In 1929 a proposal to allow courts to refuse to hear cases had not been taken up. In 1999 the matter was still not resolved in the Convention. The text remained substantially as in the 1929 Convention, retaining the “option” of the claimant and “procedure” being governed by the law of the forum.
The Court of Appeals, Ninth Circuit, held in 2002 that the Warsaw Convention did not permit a finding of forum non conveniens (Hosaka v. United Airlines Inc, cert. denied 537 U.S. 1227 (2003)). In contrast, the doctrine has been allowed application in relation to the Montreal Convention in a case now on appeal to the Court of Appeals, Eleventh Circuit (In re: West Caribbean Airways, S.A., Sept 27, 2007 (S.D. Fla. No. 06-22748)).
What is the extent of admissible preparatory work? Although the Montreal Convention is in form a new treaty, much of it has its origins in the 1929 Convention. What is the significance (if any) of the mixed practice under that Convention? Does the discussion at the 1999 Montreal Conference leave interpretative options open (other than for the plaintiff!). It looks as if this case could give the rules of treaty interpretation a good workout. Perhaps the failure of the Montreal Conference to resolve the issue (along with some other key issues in carriage by air) allows scope for what Jan describes as a “battlefield” and for continuation of unresolved political issues.
I would agree with Richard (and Isabelle) that not too much should be expected from any rules on interpretation. Interpretation is, so to speak, not entirely a rule-governed activity, in much the same way as playing the violin or the piano is not entirely rule-governed. Or building do-it-yourself bookshelves by following the manual, for that matter. I might be particularly clumsy in these matters (my wife would nod affirmatively if she would read this), but have the experience that trips to Ikea always end up in great frustration….
If this is plausible (and many seem to accept the proposition that interpretation is at least in part a matter of art and therewith of talent), then it would seem to follow that rules on interpretation can do little more than set the outward parameters: certain things are not acceptable to begin with. The strange thing now about articles 31 and 32 is that the one serious limit these rules set, is constantly ignored: I cannot think of a serious lawyer who would not at least have a look at some of the preparatory work to bolster her conclusion or, if necessary, reconsider her conclusion, regardless of whether an interpretation without the preparatory works would lead to ambiguous or absurd results. It is understandable that the drafters of the Vienna Convention felt the need to limit recourse to the preparatory works, if only for practical reasons (never mind that there are also good political reasons to limit recourse to the travaux): one would have a hard time figuring out relevant from not so relevant materials (plus debates, of course, on what exactly would constitute such relevant materials) and, more importantly still perhaps: it would lead to all sorts of difficult negotiations on what exactly the record should reflect. Imagine the sort of record created if the UNCLOS travaux would have been thought to be decisive for future interpretations; the process would in all likelihood have taken twice as long. Still, any lawyer worth her salt will consult whatever record is available.
The main function served by the Vienna Convention’s rules, then, is as something of a battlefield: the continuation of politics by other means. If you don’t get your way when drafting the treaty, you can still try to get your way when a particular provision or term comes to be applied, by trying to control the meaning assigned to it. And in order to be able to argue that something ought to be read in some particular manner rather than any competing manner, it proves convenient to have a rule to fall back, if only to suggest that you’re not just making a political or partisan point. It is this circumstance, I would guess, which gives articles 31 and 32 much of their salience, and which makes it attractive to invoke them. Isabelle hints at much the same when suggesting that the WTO panels initially felt the need to solidify their judicial function in the rough and tumble world of trade politics, and I’ll take her word for it that later panels (if not litigating parties, perhaps) have been able to relax their attitudes somewhat.
Whether that can be generalized or not is a different matter, perhaps. One factor militating against this is that even parties before the European Court of Justice, for all its self-containment, have seen fit to invoke articles 31 and 32 in recent years, and I don’t think the Court will have done much to tell parties otherwise. So perhaps the trend of the last 15 years, as Richard identifies it, is a bit more general indeed than merely the WTO trying to solidify its existence. But that raises another question: where does such a general tend stem from? What problem is it supposed to be a response to? Something to reflect upon…
In the introduction to our discussion, Richard also raised the question of: ‘Should preparatory work be investigated only to identify agreement among the negotiators or, particularly in the case of treaty provisions with a codifying aspect, to assess the state of international law leading up to the treaty and how this affected the negotiators?’
With respect to the first aspect of the question of whether preparatory work be investigated only to identify agreement among the negotiators:
The proposition of McNair that preparatory work reflects the common intentions of the contracting parties is troubling, especially when applied to multilateral treaties. Legal fictions can be helpful and necessary to move from theoretical impasses to practical solutions. It is probably the case that there ‘is [no] single legislative will behind international law’ (ILC Fragmentation Study Report 2006), even treaty law. Treaties such as the WTO covered agreements are negotiated by a heterogeneous group with particular and diverse interests. The larger this group the more unlikely it becomes that negotiators converged to a single, specific meaning of the treaty language. Preparatory work cannot, almost by definition, concern the common intention of all contracting parties. The common intention is only formed once the treaty language has been drafted and will develop over time and with the accession of new parties. Everything before the conclusion of the treaty expresses interests, values, objectives, and concerns of a single signatory or a group of contracting parties. Negotiations are as much about common intentions as about self-interests and strategy. Using preparatory work is a matter of relevance and availability but also of weighing and balancing.
A recent Appellate Body report illustrates this well. In US – Stainless Steel (Mexico), the Appellate Body found that there was no need to consider negotiating history. But it did, nevertheless, as one of the disputants extensively relied on the negotiating history of the Anti-Dumping Agreement and its predecessor Tokyo Round Anti-Dumping Code. The US contended that historical materials demonstrated that the permissibility of zeroing was left undecided by WTO Members. The silence in the agreement on this matter was not constructive ambiguity but was intentionally excluded from its scope. As a result, the Appellate Body should read the text of the agreement to include a prohibition on zeroing. The Appellate Body rejected this argument on several grounds. Ratione personae, the negotiating history only reflected the positions of a few negotiating parties. Rationae materiae, it was also not substantively relevant, as it ‘did not resolve the issue of whether the negotiators … intended to prohibit zeroing’ [para 131]. Ratione temporis, much of the invoked materials pertained to the interpretation and drafting of the plurilateral Tokyo Round Anti-Dumping Code. This was a separate treaty from the GATT 1947 and was terminated with the adoption of the 1995 Anti-Dumping Agreement [para 132]. They are two separate treaties, with different texts, membership, binding force, and mainly different temporal scopes. Their respective negotiating history cannot easily be interchanged. As a result, the Appellate Body found that the Anti-Dumping Code ‘is of little relevance for the interpretation of differently phrased or new provisions of the Anti-Dumping Agreement’ [para 132]. In response, the US strongly disagreed with this reasoning because ‘the Division’s conclusions regarding the negotiating history simply cannot be reconciled with that history’. Ratione temporis, the US argued that relevant negotiating history of treaties no longer in force, such as the Tokyo Round Anti-Dumping Code, could serve to interpret another treaty, such as the Anti-Dumping Agreement.
Any action and inaction of negotiators preceding the conclusion of the treaty is part of the history of text, but history is not applicable law. From the practice of judicial treaty interpretation, it seems that there exists a presumption that materials evidencing such conduct, statements, or lack thereof cannot inform the meaning of the treaty text. It can be rebutted but the burden of proof is considerable. The larger the group of negotiators and signatories of a treaty, the higher that burden. It involves proving that there was a common intention of that group that the treaty text or term has a particular meaning, with written evidence.
But this raises another question, namely, if this burden of proof is met how does the negotiating history proven to reflect a common intention among (not necessarily all) negotiating partners still qualify as a ‘supplementary means of interpretation’? Perhaps, it may be possible to qualify such type of negotiating history as context under Article 31(2) VCLT, or even as evidence of a special meaning under Article 31(4) VCLT? These questions illustrate the earlier point made about understanding the relationship between different principles of interpretation, and their relative value and meaning.
With respect to the second aspect of the question of whether preparatory work be investigated particularly in the case of treaty provisions with a codifying aspect, to assess the state of international law leading up to the treaty and how this affected the negotiators:
These are just some preliminary remarks: I believe this is a different matter and that here negotiating history serves a different purpose depending on the substance of, for example, statements of negotiators. Statements of negotiating partners may be relevant to establishing what the state of customary international law is at the time of negotiating the treaty and how to codify such customary norm. Such statements may inform whether the constitutive elements of opinio iuris and state practice are available, to the extent that they inform whether States belief they are bound by a norm as a customary international law.
But when it comes to subsequent interpretation of the treaty that codifies the customary norm, the same statements can only be relevant to inform the meaning of the treaty if they clarify the extent to which negotiating States intended to be bound by the same norm as treaty law – on the assumption that pure codification is possible. Each statement should thus be examined to determine their scope. To the extent that such statements only relate to the norm that will be binding as treaty law, these statements cannot be taken into account to examine the state of customary international law (though admittedly, this may lead to a paradox, especially in the event of a large multilateral treaty being negotiated). A single statement may thus simultaneously inform whether a state believes to be bound by a norm of customary international and what that norm is, and explain how a state perceives the relationship between the treaty norm to be negotiated and the pre-existing customary norm (codification, crystallization, contracting out, etc). A single statement may thus serve two functions and an examination of the substance must determine what part is relevant to informing the state of customary international law and what part serves the purpose of treaty interpretation.
Such statements might also become relevant in the event that a court or tribunal decides to interpret the treaty against the background of other rules of international law. To the extent that such statements inform that States intended to contract out of pre-existing customary international law, it signals that the interpretation of the treaty norm acknowledges this decision. In that sense, the interpretation of the treaty against the background of other rules of international law is not about ensuring ‘harmony’ but, instead, useful to respect the choice to contract out of customary international law, in whatever degree or extent.
As I mentioned, these are just a few preliminary remarks on a most intriguing question.
The initial post introducing our discussion of Treaty Interpretation raises various issues of interest to the study and practice of treaty interpretation. In this first reply, I focus on the first question relating to the tragedy of ‘insistent emphasis upon an impossible, conformity-imposing textuality’. My responses are informed by my study of treaty interpretation by the WTO Appellate Body.
I believe that the practice of international courts and tribunals has demonstrated that the general principle in Article 31(1) VCLT does not emphasize ‘impossible, conformity-imposing textuality’. The text of Article 31(1) itself reads: ‘A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose’. The text itself thus emphasizes that the starting point is the ‘contextualized ordinary meaning’, a notion that was actually already introduced before the ILC started its work on the law of treaties (eg, writings by Stone and Huber, Judge Anzilotti in Interpretation of the Convention of 1919 Concerning Employment of Women During the Night, the ICJ in Constitution of the Maritime Safety Committee of the Inter-Governmental Maritime Consultative Organization).
In the context of the WTO, the claim has been made that the Appellate Body has used Article 31 VCLT to the extent of, indeed, something alike ‘impossible, conformity-imposing textuality’. The Appellate Body’s apparently excessive attention to the words of the WTO treaty language might seem overdone, but it is correct nevertheless. However, such excessive attention to the words of the treaty does not justify the claim of strict textualism. Much has been said about its use of dictionaries, but in fact it has hardly ever solely relied on dictionary meanings, and already a substantive amount of ‘contextualization’ takes place when choosing the relevant dictionary meanings. The Appellate Body itself has explicitly recognized the limits of dictionary definitions and the need to contextualize them, through various methods. Of course, the fact that it consults dictionaries is not problematic, it is a common technique. But the Appellate Body was unique in the way that it openly acknowledged its use of dictionaries, in great detail. Its initial apparently excessive use of dictionaries and references to Article 31 VCLT was probably instigated by the need to assert its judicial function against the backdrop of a not fully developed institutional model and under-developed procedural rules in the DSU. To a large extent, this has been a successful strategy. In response, the explanation of its interpretative process has become less formalistic in recent cases.
If the general thread in an interpretative practice is to use the contextualized ordinary meaning as the keystone upon which the interpretation is built, this raises the question of how to define context? Article 31 VCLT gives the impression that it defines context in abstracto in an exhaustive manner. But when studying the interpretative practices and techniques used by one or more courts and tribunals, it becomes easily visible that a broader notion of contextualism applies. The ILC did not explain what ‘ordinary meaning’ entails. The Appellate Body’s jurisprudence suggests that the ‘ordinary’ meaning already implies a considerable amount of contextualization of the treaty language as part of what is commonly perceived as literal interpretation. Even if the VCLT drafters realized that context has a broader meaning and impact in treaty interpretation, the language in Article 31(2) mirrors the ultimate consensus that could be reached. The consensus among States on the final text of the VCLT was grounded on the shared understanding that consensualism was the basis for entering into treaty obligations. This helps to understand and appreciate how the VCLT reflects to what States ultimately could (only) agree. Also, at the time of the VCLT negotiations there was little practice of courts and tribunals like WTO panels and the Appellate Body, who continuously revisit the same treaties and within a particular institutional context. This might point to a certain disjuncture between the treaty language in Articles 31 to 33 VCLT and the interpretive practice of international courts and tribunals, without calling the text of Article 31 to 33 VCLT into question.
In essence, I believe that the usefulness and value of any principle of treaty interpretation, including the use negotiating history, is relative and depends on its relationship to other principles of treaty interpretation, whether codified or not.