1. Introduction
Among the typical lectures of an introductory course on public international law devoted to the law of treaties, few have as many practical applications as the lecture where students learn about the rules on treaty interpretation. In that lesson, students learn that all international agreements, irrespective of their state parties or subject matter, are elucidated through the same set of rules which, for the most part, are set forth in Articles 31 and 32 of the Vienna Convention on the Law of Treaties (VCLT).Footnote 1 Aspiring lawyers learn the core of Article 31: that they should interpret a treaty in good faith, in accordance with the ordinary meaning of its terms in their context, and in the light of the treaty’s object and purpose. In addition, they learn that there are limited circumstances in which recourse may be had to the preparatory work of the relevant treaty, as contemplated by Article 32.
There is some irony in the fact that these VCLT provisions governing treaty interpretation, which reflect customary international law,Footnote 2 are recurrently read in different and incompatible ways. The differences in academic and judicial opinions on questions of treaty interpretation should not, however, be exaggerated. A careful study of state practice, international case law, and scholarship can reveal normative patterns of modest consistency, the systematization of which allows us to tell good and bad interpretations apart.
This article is devoted to one of the most troubling areas of treaty interpretation: the use of preparatory works (also known as travaux préparatoires). For the most part, the longstanding debates in legal practice and scholarship over these materials have focused on two main problems: firstly, with defining the sort of documents that fall within the notion of travaux (or, more broadly, within the category of ‘supplementary means of interpretation’)Footnote 3 and, secondly, with delineating the conditions under which recourse to travaux is ‘appropriate’ or, as others would rather say, ‘permissible’.Footnote 4 In light of the very elaborate arguments advanced in these discussions, it is surprising that international lawyers have mostly glossed over the problem that would invariably follow once a set of documents is correctly identified as part of the preparatory works of a treaty (assuming, of course, that the conditions enabling an interpreter to consult them are fulfilled). This is the question of how the travaux themselves should be interpreted. Indeed, for all the ink that has been spilled discussing the application of the VCLT rules to the treaties themselves, remarkably little has been written about the factors that should be considered in order to draw conclusions from the preparatory works. Although abstract, this is a subject with very practical consequences. As much can be illustrated by the study of a judicial discussion about the travaux of the Refugee Convention of 1951.Footnote 5 That treaty’s cornerstone is found in its Article 33(1), which establishes the prohibition of refoulement (the expulsion or return of refugees) in the following terms:
No Contracting State shall expel or return (‘refouler’) a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened on account of his race, religion, nationality, membership of a particular social group or political opinion.
The centrality of the non-refoulement principle to this Convention is hard to overestimate. Unless refugees are protected from forcible returns to the dangerous places from which they have fled, the normative scaffolding provided by the rest of that treaty could be rendered meaningless. Needless to say, a flawed reading of Article 33(1) by the executive and judicial organs of a state can have tragic consequences if it narrows the scope of refugee safeguards.
In this regard, a vital query, over which the terms of Article 33(1) are arguably silent, is whether a state’s policy that prevents the arrival of asylum seekers at its territory is outlawed by this provision. In other words, would a state that frustrates an asylum seeker’s attempt to enter its territory with the purpose of requesting international protection be said to comply with the principle of non-refoulement? The answer given to this question in a very influential judicial decision hinged, in no small part, on the interpretation of the Refugee Convention’s preparatory work. Indeed, in Sale v. Haitian Centers Council, the Supreme Court of the United States assessed the compatibility of a policy of interdiction of asylum seekers in the high seas with the Refugee Convention, concluding that non-refoulement could only apply to individuals that are already within the territory of the United States. To some extent, the Court’s reasoning relied on the academic opinion of Nehemiah Robinson and Atle Grahl-Madsen,Footnote 6 two influential academics who argued that the Convention’s travaux showed that a policy of absolute border closure would be compatible with the non-refoulement obligation.Footnote 7
The problem with the position of both Robinson and Grahl-Madsen is that their views were based on a narrow reading of the Refugee Convention’s negotiating records, which highlighted the position of a select group of state representatives concerning a specific scenario of migratory control (a case of ‘mass influx’ of asylum seekers) while ignoring other elements within the travaux that may have strongly supported the extraterritorial applicability of non-refoulement.Footnote 8 Indeed, a comprehensive reading of the preparatory works of the Refugee Convention might have rendered a much more nuanced understanding of its Article 33.Footnote 9
Aiming to assist interpreters in hard cases like this, the present article describes a sound interpretive approach for understanding a treaty’s travaux, inferred from a representative range of case law from the International Court of Justice (ICJ) and other dispute settlement bodies, as well as the opinions of legal scholars. The following discussion is divided into five parts. Section 2 will set the overall doctrinal framework by explaining what travaux are. Section 3 will then summarize the main viewpoints about the conditions under which recourse to travaux is permitted by the VCLT. Afterwards, Section 4 will sketch seven factors that should be taken into account when elucidating preparatory works. Critical remarks are offered in Section 5, followed by a conclusion in Section 6.
2. The boundaries of the notion of preparatory work
The notion of ‘preparatory work’ in Article 32 of the VCLT should be interpreted in accordance with its ordinary meaning. The plain meaning of the adjective ‘preparatory’ clearly indicates that the ‘work’ in question not only precedes the treaty itself (chronologically speaking), but that it has also been developed during the process through which the treaty comes into being – and with this very purpose in mind. There is a broad range of materials that may fit this description, including drafts prepared by states, the works of the International Law Commission (ILC) for the codification of international law,Footnote 10 correspondence exchanged by the parties,Footnote 11 negotiation minutes,Footnote 12 and perhaps even the resolutions from the United Nations General Assembly requesting the elaboration of a convention.Footnote 13
One can be easily misled by the apparent simplicity of this definition. Assuming that the object and purpose of the rules on treaty interpretation is to elucidate ‘the common intention of the parties’,Footnote 14 and thus a meaning the parties have agreed upon, it might be reasonable to suggest that not all ‘works’ created in ‘preparation’ of the treaty fall within this notion. Arguably, a document prepared by one of the parties in the course of treaty negotiations (or perhaps even after the agreement’s conclusion), which does not reflect the understanding of other treaty parties, might not constitute ‘preparatory works’ under a teleological interpretation of Article 32 of the VCLT. Moreover, other scholars have posited that accepting unilateral documents as travaux would run counter to the principle of good faith in treaty interpretation.Footnote 15
This issue can be complicated by the practice of states in the application of the rule codified in Article 32 of the VCLT.Footnote 16 Indeed, state practice on this point is mixed,Footnote 17 as states have voiced different views on whether documents prepared by one set of negotiators that were never shared with others could constitute ‘preparatory works’.Footnote 18 Similarly, there is no consensus among states on whether documents drafted and opinions voiced after the conclusion of a treaty could fall within the notion of travaux.Footnote 19 Some controversies on this point have focused on whether domestic explanatory documents and parliamentary debates on the appropriateness of a treaty’s ratification can be relied upon as preparatory works. For illustration, Detlev F. Vagts once recalled a curious episode from the history of diplomatic relations between the United States and Germany, as follows:
In a diplomatic interchange concerning the question whether the Treaty of Berlin automatically gave the United States rights accorded by the Treaty of Versailles, Germany referred to statements made by Senator Lodge in debate in the Senate upon the Treaty of Berlin. Concerning this, Secretary [of State] Hughes wrote the Ambassador to Germany:
‘Should occasion arise, you may orally explain to the German Foreign Office that expressions of opinion as to the meaning of the treaty of August 25, 1921, such as those to which the Foreign Office refers, occurring in general debate, cannot be regarded as affecting the interpretation of that treaty’.Footnote 20
The preparatory works of the VCLT itself fail to resolve the aforementioned questions surrounding the notion of travaux. As Esmé Shirlow and Michael Waibel have noted, during the drafting process of the VCLT – a project spearheaded by the ILC – the concept of ‘preparatory work’ was purposedly left undefined to assuage concerns that a categorical definition might lead to the exclusion of relevant interpretative evidence.Footnote 21
To round up the study about the meaning of travaux under the VCLT, judicial decisions (including those rendered by international and domestic courts and arbitral tribunals)Footnote 22 and qualified academic commentary should also be considered, as these materials constitute subsidiary means for the determination of international law, under Article 38(1)(d) of the ICJ’s Statute. As elaborated below, judicial and academic discussions on the matter have centred upon the temporal dimension of preparatory works (up until what point are materials produced by the parties considered to be ‘preparatory’) and their authorship and accessibility (whose work counts as travaux). No consensus has emerged from such discussions.
It seems clear that a treaty’s ‘conclusion’ represents the point in time that separates travaux (under Article 32 of the VCLT), from subsequent agreements or practices of the state parties that may also be relevant for interpretive purposes under Article 31. Ironically, it can be difficult to pinpoint the moment at which a treaty is ‘concluded’. As Richard Gardiner has argued, ‘[u]nfortunately, this is not a term having a clear meaning’,Footnote 23 as different provisions within the VCLT point in slightly different temporal directions – some related to a treaty’s signature, but others to its ratification and to a party’s accession.Footnote 24 In practical terms, this ambiguity might favour an expansive definition of travaux, which could encompass materials produced by a contracting party to facilitate the domestic approval of the treaty (typically by a legislative body) before its ratification. Importantly, however, the commentaries to the ILC’s Draft Articles on the Law of Treaties suggest that the authentication of the treaty’s text is the decisive moment for determining which materials are preparatory in nature.Footnote 25
International case law on the use of internal communications between state organs during ratification processes suggests that materials produced after a treaty’s signature are not travaux, but that they may nonetheless be relied upon under the more general heading of ‘supplementary means of interpretation’. For illustration, the arbitral tribunal in the 1985 Guinea/Guinea-Bissau maritime boundary case affirmed that the executive’s domestic explanatory memoranda to their legislature did not constitute, strictly speaking, ‘preparatory works’. The tribunal nonetheless took these documents into account as other supplementary means of interpretation.Footnote 26 Similarly, in the Jan Mayen case, the ICJ took into account an official communication sent by the Norwegian Government to the Storting (the Norwegian Parliament) in order to confirm the absence of a relationship between two international agreements.Footnote 27 More recently, in Arbitral Award of 3 October 1899 (Guyana v. Venezuela), the Court described a statement from the Venezuelan Foreign Ministry before its Congress on occasion of the ratification of a treaty as part of the circumstances surrounding the conclusion of such agreement.Footnote 28
Relatedly, judicial opinions are divided on the question of whether documents prepared by one party to the negotiations that were not shared with its counterparts may be part of a treaty’s preparatory works. Perhaps the most famous decision on this matter is that of the arbitral tribunal in Young, according to which
the term [travaux] must normally be restricted to material set down in writing – and thereby actually available at a later date. This means that oral statements and opinions not recorded in minutes or conference papers can apparently be regarded as a component of travaux préparatoires only in exceptional cases. They can in any event be considered only if made in an official capacity and during the negotiations themselves… A further prerequisite… is that [the material] was actually accessible and known to all parties. Drafts of particular articles, preparatory documents and proceedings of meetings from which one member or some members of the contracting parties were excluded cannot serve as an indication of common intentions and agreed definitions unless all parties had become familiar with the documents or material by the time the treaty was signed.Footnote 29
A more comprehensive record of judicial decisions on the matter would show that this question is far from settled. Acting as an arbitrator in the 1910’s North Atlantic Coast Fisheries Case, the renowned Argentinian jurist Luis María Drago was willing to rely on the journal of one negotiator to clarify a treaty term.Footnote 30 Somewhat similarly, in Border and Transborder Armed Actions (Nicaragua v. Honduras), the ICJ treated as travaux a report published by the United States’ Department of State of that country’s delegation to the Conference of Bogotá.Footnote 31
In light of the above, it may be concluded that a definitive, all-encompassing definition of travaux continues to elude us. Nevertheless, it may still be argued that the scope of preparatory works has a core, settled meaning, and that this notion only remains uncertain at its margins. In an effort to escape this jurisprudential impasse, Esmé Shirlow and Michael Waibel have recently advanced a ‘sliding-scale approach’ for the identification of and reliance upon preparatory works, arguing that ‘the focus should not be on the binary categorization of materials as “travaux” or not, but instead upon the utility of any given material to the interpretive exercise by reference to its precise qualitative features and the context of interpretation’.Footnote 32 As explained by these two scholars, ‘an interpreter might, for example, give more weight to materials that are capable of manifesting the joint intent of the parties’.Footnote 33
This ‘sliding scale approach’ for the identification of travaux is certainly a step in the right direction in the development of a coherent, reasonable, and systematic method for the use of preparatory works during the interpretation of treaties. However, a plausible drawback of this approach is that the lack of a lower-end point in its sliding scale could perhaps over-expand the range of materials to be consulted, drawing an interpreter’s attention farther and farther away from the actual treaty. In comparison, a broad but ultimately binary categorization of materials as preparatory works could perhaps be preferable as a matter of legal predictability and practicality.
Moreover, a binary categorization of travaux could be inherent to the process of treaty interpretation, understood as the ascertainment of the common intention of the parties to the agreement. For illustration, the tribunal in the Iron Rhine arbitration decided that the extracts from the prolonged diplomatic negotiations leading to the conclusion of the relevant treaty did not ‘have the character of travaux préparatoires on which it may safely rely as supplementary means of interpretation’, because
[t]hese extracts may show the desire or understanding of one or other of the Parties at particular moments in the extended negotiations, but do not serve the purpose of illuminating a common understanding as to the meaning of the various [treaty] provisions…Footnote 34
More importantly, a sliding scale approach needs to be supplemented with some parameters for the interpretation of travaux. Indeed, while this scale would allow us to compare the normative ‘weight’ of different materials that might fall somewhere along its spectrum of preparatory works, it nonetheless obviates the need to identify the meaning and implications of whatever is contained in such documents. To put it differently, one can only assess the relative persuasiveness of different travaux after determining what is revealed by them.
3. The relationship between Articles 31 and 32 of the VCLT
Any doctrinal study about the role of preparatory works in international law would be incomplete without addressing the stage of the interpretive process in which these documents may be used and the purpose that they serve. The questions about ‘when’ and ‘what for’ concerning travaux boil down to the relationship between Articles 31 and 32 of the VCLT, with the first provision establishing the general rule of treaty interpretation and the second one setting out the scenarios where recourse to supplementary means may be had.
The many theories purporting to explain the relationship between Articles 31 and 32 may be found along a spectrum, with diametrically opposed approaches at its poles. On one end of the spectrum, one can find restrictive and hierarchical approaches, pursuant to which recourse to preparatory works may only be had in order to confirm the meaning resulting from the application of Article 31, or to determine the meaning when the application of Article 31 renders an ambiguous or absurd reading. A restrictive approach would have two main practical implications. Firstly, the components of the rule set forth in Article 31 (i.e., the ordinary or special meaning of terms, the treaty’s object and purpose, etc.) would have to be ascertained without reference to the preparatory works. All such components should be identified within the treaty itself, its context (as this notion is defined in Article 31), the subsequent agreements and practice of the parties, and any relevant and applicable rules of international law. Secondly, a clear and reasonable reading of the treaty, resulting from the application of Article 31, would stand even in the odd scenario where this interpretation was contradicted by the travaux.
At the other end of the spectrum, one can find flexible interpretive approaches that would authorize and even encourage the use of supplementary means more generally. In practical terms, a flexible approach would have the opposite effects of a restrictive one. Therefore, travaux could be used to determine the meaning of a treaty term in situations other than those that are expressly set forth in Article 32, perhaps even to the point of correcting a ‘clear’ meaning ascertained through the general rule of Article 31.Footnote 35 Moreover, some of the components of Article 31, like the object and purpose of the treatyFootnote 36 or the special meaning of some of its termsFootnote 37 could be ascertained by reference to the preparatory works. More generally, it may be posited, as the ICJ recently suggested, that ‘a good-faith interpretation… must take into account the concern of State parties when drafting [a treaty]’, which entails a review of the travaux.Footnote 38 The implication might be that Articles 31 and 32 should be applied simultaneously, without any hierarchical distinction between its components.
There is no shortage of arguments in international legal scholarship in favour of either perspective. Scholars that favour a restrictive approach to the use of travaux tend to emphasize the presumptive ordinary meaning of the terms employed in Articles 31 and 32, which arguably reveal a hierarchy between these two provisions.Footnote 39 By contrast, their opponents argue that the preparatory works of the VCLT suggest that the drafters of this treaty shared the understanding ‘that Articles 31 and 32 were intertwined halves of a single, unitary whole’.Footnote 40
International case law provides a more nuanced picture. Early ICJ case law embraced a very restrictive approach to the use of these materials. For illustration, in an advisory opinion rendered in 1950, the Court declared that ‘[w]hen [it] can give effect to a provision of a treaty by giving the words used in it their natural and ordinary meaning… it is not permissible… to resort to travaux préparatoires’.Footnote 41 By contrast, more recent ICJ decisions describe the recourse to travaux in such situations as ‘unnecessary’, rather than ‘impermissible’.Footnote 42
As these last few paragraphs reveal, the questions about when preparatory works should be consulted and the purpose of this endeavour have provoked significant academic and judicial controversy. The theoretical challenges involved in justifying the recourse to travaux, coupled with the imprecise boundaries of this notion described in Section 2 above, might be part of the reason for which scholars and adjudicators have said little about the factors that should be considered to elucidate these documents. Indeed, the uncertainty about ‘when’ and ‘why’ travaux should be consulted could discourage the attempts to determine ‘how’ they should be used, as outlined in the following section.
4. Seven factors to consider when interpreting travaux préparatoires
Only so often, the preparatory works will contain ‘interpretative notes’, that is, statements of unanimity of intention amongst all the negotiating parties concerning the meaning of a treaty provision that are recorded for the avoidance of doubts. Such interpretive notes would, obviously, be deserving of deference.Footnote 43 For illustration, while interpreting the Palermo ConventionFootnote 44 in the Immunities and Criminal Proceedings case, the ICJ took note of an ‘interpretative note’ that was included in the travaux upon the agreement of the negotiating parties with the specific purpose of outlining ‘the intention of the Convention’.Footnote 45
When interpreters lack the benefit of such guidance, that is, when the travaux lack the express indications of the unanimous agreements of the negotiating parties, interpretation becomes more complicated. With the aim of assisting interpreters in such situations, this section outlines seven factors that may be considered in order to elucidate travaux. The first five factors are related to the treaty’s drafts. The remaining two factors have less to do with the treaty’s drafts than with the attitudes and comments expressed by state representatives during the negotiations.
Under the list of factors suggested below, interpreters should first assess the completeness of the preparatory works (Section 4.1). Secondly, they should determine what impact, if any, the authorship of a treaty’s original drafts can have (Section 4.2). Thirdly, interpreters should consider the possible implications of the translations of such original drafts (Section 4.3). Fourthly, they should assess the impact, if any, of the presence of multiple drafting groups during the treaty’s negotiations (Section 4.4). Fifthly, interpreters should examine the evolution of the drafts through the diplomatic negotiations (Section 4.5). This requires the interpreters to consider, inter alia, whether a drafted provision remained untouched throughout the negotiations, whether a treaty clause was moved from one section of the agreement to another, and whether the rejection of an amendment would allow the interpreters to draw adverse inferences from such a fact. Sixthly, interpreters should appraise the views expressed by an individual or by groups of delegates (Section 4.6). Finally, interpreters must assess the implications that a state’s own views during the negotiating process may have in future disputes concerning the interpretation of the treaty in question (Section 4.7).
It must be emphasized that the order in which these seven factors are presented is not intended to reflect a methodological chronology. Moreover, no claims are made about their normative status under international law. What follows are guidelines of interpretation that are mainly supported by judicial decisions and the views of publicists, that is, by subsidiary means for the determination of international law. Admittedly, the case law and scholarship on the matter is relatively limited, and has not been subject to intense scrutiny from states. As a consequence, the plausible status of these guidelines as reflections of customary international law or expressions of general legal principles may be contested.
4.1. The ‘completeness’ of travaux
On at least two occasions, the ICJ has cautioned against the overreliance on incomplete preparatory works. In Border and Transborder Armed Actions (Nicaragua v. Honduras), the Court looked into the travaux of the Pact of Bogotá, clarifying that ‘[i]n this case these [preparatory works] must of course be resorted to only with caution, as not all the stages of the drafting of the texts at the Bogotá Conference were the subject of detailed records’.Footnote 46 A similar observation was made in Maritime Delimitation and Territorial Questions between Qatar and Bahrain, where the Court noted that ‘[t]he travaux préparatoires of the Doha Minutes must be used with caution in the present case, on account of their fragmentary nature’.Footnote 47
As these cases illustrate, the completeness of preparatory works is a significant factor when determining their interpretative weight. Naturally, the precise implications of the absence of specific records within the travaux can only be assessed on a case-by-case basis.
4.2. The original draft’s authorship
The starting point of many treaty negotiations is a draft prepared by one of the parties. It may thus be open to question if an ambiguity resulting from such a draft could be interpreted to the detriment of its author, as the contra proferentem principle might require.Footnote 48 This maxim has a long history in international law, finding a place within the works of Grotius and Vattel.Footnote 49 The principle’s lineage can then be followed through a set of early arbitral decisions, where it was sometimes ‘understood in a rather wide sense of in favorem debitoris’.Footnote 50 In a handful of decisions the principle was more appropriately portrayed as requiring ‘that in case of doubt, the text of a treaty is to be interpreted against the party which drafted it… when, as in the case of the Treaty of Versailles, one of the parties handed a prepared text to the other party for signature’.Footnote 51
Similarly, Pierre d’Argent has argued that the contra proferentem principle could only be relevant in the ‘rare situations where ambiguous terms are found in a treaty that was both drafted but also somehow imposed by one party upon the other, leaving the other party no opportunity to review its content carefully’.Footnote 52 Nevertheless, this maxim has also been invoked to prevent a state party from taking advantage of an ambiguity to which it has ‘contributed’ and is, consequently, ‘to a certain extent responsible’.Footnote 53 For illustration, in an arbitration between the United States and Italy concerning the interpretation of a treaty about air transport services, Italy argued that the contra proferentem rule should be applied against the United States, given that the agreement under analysis had been proposed by the Americans.Footnote 54 The tribunal decided not to discuss the applicability of the contra proferentem rule, based on the fact that the relevant treaty provision was ‘clear and not ambiguous’,Footnote 55 arguably implying that this rule might had otherwise played a role in its legal assessment.
Be that as it may, in recent decades the contra proferentem principle has barely made appearances in international case law. In fact, the majority of cases that are commonly used to illustrate this maxim belong to the pre-VCLT era. Consequently, little has been said about the validity of contra proferentem in present days. At most, in its 1998 judgment in Fisheries Jurisdiction (Spain v. Canada) the ICJ observed that this rule ‘may have a role to play in the interpretation of contractual provisions’.Footnote 56 The Court made such pronouncement in response to a claim advanced by Spain, which described contra proferentem as a ‘rule, under which, when a text is ambiguous, it must be construed against the party who drafted it’.Footnote 57 Arguably, the ICJ’s cautious stance suggests that, in addition to appraising the completeness of the travaux, an interpreter should also take notice of the origin of the treaty’s draft.
Relevantly, the cautious approach of the ICJ with respect to the contra proferentem principle has not deterred states from arguing in favour of its applicability in other dispute settlement contexts. As Michael Waibel has noted, several states have invoked contra proferentem in investment arbitrations, arguing that standardized treaty clauses about ‘fair and equitable treatment’, based on ‘boilerplate provisions’ of model investment treaties, should be interpreted narrowly. While such arguments have not been endorsed by arbitral tribunals, Waibel has defended the applicability of this maxim where contracting parties had unequal bargaining power during the treaty’s negotiations.Footnote 58
Finally, it must be noted that the authorship of an original draft may have legal consequences beyond the plausible applicability of the contra proferentem principle. For illustration, in the Lighthouses case between France and Greece, the Permanent Court of International Justice (PCIJ) took notice of the history of the negotiation of the Special Agreement conferring jurisdiction to the Court. The dispute was about the validity of a concession contract between the Ottoman Empire and a French firm concerning the management and maintenance of some lighthouses along the Ottoman coastline, on territory that was ceded to Greece. According to France, the validity of this contract was exclusively determined by Ottoman law; by contrast, Greece asserted that international law on state succession would also be applicable. The Court noted that the first draft of the Special Agreement was framed by the Greek Government, and deduced from this fact the assumption that the Greek contention (the applicability of international law for the determination of the validity of the contract) was expressed in the Agreement.Footnote 59 For this reason, the Court decided ‘not to omit from considerations the objections of an international character opposed by the Greek Government to the arguments of the French Government’.Footnote 60
4.3. The translations of original drafts
There is some limited judicial authority to the effect that, if there is a disparity of meaning between different versions of a plurilingual treaty, the original version of which was drafted in one language, and subsequently translated into another, then preference should be given to the original version.Footnote 61 If, however, the original and subsequent drafts were translated ‘word-for-word’, the terms in both versions may reasonably be taken as synonymous,Footnote 62 and there would arguably be no a priori reason to prefer either translation. Moreover, when a term in the language of the original draft has an excellent equivalent in the second language, but has nonetheless been deliberately translated in a different way, the nuances introduced by the translation may be preferred.Footnote 63
Importantly, the negotiating procedure of some multilateral agreements could provide strong reasons to prefer a translation over an original draft. Indeed, there are cases in which a proposal is originally made in one language, and then translated into another language that is more widely in use (like English) in order to facilitate the participation of a wider group of states in the negotiations. In such scenarios, the final version of the agreement in the first language can differ significantly from its original proposal,Footnote 64 and so it might be more appropriate to focus one’s attention on the drafts in English.
It must be emphasized that all these propositions are controversial. Importantly, when discussing the groundwork for the VCLT, the ILC dismissed the possibility that divergencies between the authenticated language versions of a treaty should be resolved in favour of the language in which the agreement had been drawn up first.Footnote 65 However, some influential scholars on the law of treaties have supported the proposition that the language of the original draft may, in appropriate circumstances, be preferred.Footnote 66 For illustration, in a study published before the VCLT was concluded, Jean Hardy argued that
It is only normal that the presumption should be in favour of the original version, because that was the basis on which the negotiators in fact first reached agreement and the authoritative value of the other texts is subordinated to their equivalence to the original. The strength of the presumption in favour of the original version depends on the circumstances in which the other versions were drawn up. It will be weak if the negotiators all participated directly in the elaboration of those texts; stronger if they only exercised partial control over it, as, for example, by entrusting the task to a small drafting committee; and decisive if they left the entire job of drawing up those texts to one of the parties or to some specified body.Footnote 67
More recently, the obiter dicta observations of the ICJ in the LaGrand case, through which the Court explained the rationale behind the choice of language in its Statute, concerning the legal weight of provisional measures by primary reference to a French version of an amended draft of this document (subsequently translated into English), could perhaps be cited in support of favouring the language of an original draft.Footnote 68 Admittedly, the analysis of this matter in the Court’s judgment was very sparse, so it is perhaps advisable not to read much into it.
While inconclusive, the case law and scholarship referenced in the paragraphs above suggest that once the completeness of the travaux has been evaluated and the authorship of the first draft of the treaty (or the relevant provision) has been determined, interpreters should take into account how the original texts were translated to confirm the equivalence of terms or to attribute greater weight to the version that most likely expressed the shared view of the parties.
4.4. The possible impact of multiple drafting groups
There seems to be a presumption in treaty interpretation according to which one and the same term, which is employed at different parts of a single document, is expected to have a consistent meaning throughout the text. A contrario, this implies that different terms ordinarily have different meanings.Footnote 69 Some scholars have argued that the strength of this presumption may be determined by the individual or collective identity of those who drafted different parts of a treaty. For instance, James F. Hogg once remarked with respect to multilateral conventions that:
Such is the scope of the problems involved that often one committee is unable to cover all the topics in the time available. For this reason separate drafting committees are frequently constituted to deal with particular aspects of the agreement. It follows that words are more likely to be used in the same sense when they have been inserted by the same committee.Footnote 70
For those who might adhere to this view, the presumption of consistency in the usage of terms could be rebutted if the parts of the treaty where the terms at hand are located were drafted by different negotiators. Naturally, other factors should be taken into account before reaching this conclusion, such as the degree of coordination between the working groups in charge of drafting different parts of the treaty and the extent to which the agreement as a whole was considered by a sole set of delegates after its ‘committee’ stage.
The history of the negotiations of the Rome Statute of the International Criminal CourtFootnote 71 illustrates the impact that fragmented treaty drafting can have. During the final negotiations of the Statute, there were up to 11 sub-working groups simultaneously drafting different treaty provisions.Footnote 72 All these groups transmitted their preliminary texts to a ‘committee of the whole’ (assisted by a ‘drafting committee’) for final approval.Footnote 73 This piecemeal drafting procedure ‘resembled the assembly of a large jigsaw puzzle’,Footnote 74 hindering the internal consistency of the final treaty.Footnote 75 A notable inconsistency is the ‘asymmetric liability’ that seemingly results from the disparity of the terms used in Articles 12(3) and Article 13(a) and (b).Footnote 76 The former provision regulates the ad hoc acceptance of the Court’s jurisdiction by non-state parties, which may apparently be limited to specific ‘crimes’. By contrast, the latter provision regulates the exercise of the Court’s jurisdiction after the referral of a ‘situation’ by a state party or the Security Council. Since the term ‘crime’ is narrower than the term ‘situation’, a literal reading of the Statute would entail that the Court’s jurisdiction would vary in scenarios involving Articles 12 and 13. This disparity was not deliberately included in the Statute, and it has been redressed through the Court’s Rules of Procedure and Evidence.Footnote 77
4.5. The evolution of drafts through negotiations
A fifth factor in the interpretation of preparatory works is concerned with the consistency in a treaty’s draft text throughout the negotiation and the level of attention that the participants in such deliberations devoted to the provisions that must be elucidated. Such matters are perhaps best illustrated by direct reference to international case law.
The ICJ’s advisory opinion on the Constitution of the Maritime Safety Committee of the Inter-Governmental Maritime Consultative Organization (IMCO) stands out for its careful consideration of preparatory works. The question before the Court was concerned with the interpretation of Article 28(a) of IMCO’s Constitution, which on its face guarantees a place within one of its organs – the Maritime Safety Committee – to the eight largest ship-owning nations. Liberia and Panama, who were among such nations by reference to gross registered tonnage, were excluded from this Committee by a decision from IMCO’s Assembly in 1959, which had among its attributions the ‘election’ of the Committee’s members. Liberia and Panama argued that this election ‘was not an election in the usual sense of the word, for once those eight nations had been determined, the Assembly was bound to elect them’.Footnote 78
The Court agreed with Liberia’s and Panama’s argument, based on a straightforward application of the rules of treaty interpretation to IMCO’s Constitution. The ICJ then went on to study the drafting history of that Constitution’s Article 28(a) in order to confirm its interpretation. The Court’s analysis of the three main drafts of the treaty revealed that the choice of words in this provision was careful and deliberate. Indeed, the Court stressed the fact that certain words and phrases were ‘inserted with evident deliberation’.Footnote 79 Another important factor supporting the Court’s interpretation was that the negotiators’ intention over the meaning of Article 28(a) ‘remained constant’ throughout its three drafts.Footnote 80 By contrast, the negotiators’ choice to describe the Assembly’s attribution to designate members to the Committee as an ‘election’ was portrayed by the Court as ‘a mere drafting change’,Footnote 81 that was not discussed nor explained in the travaux.
Another prime example of the Court’s assessment of the course of diplomatic negotiations is found in Application of the Convention on the Prevention and Punishment of the Crime of Genocide (Bosnia and Herzegovina v. Serbia and Montenegro). In its judgment of 2007, the ICJ had to determine if Article I of the Genocide Convention creates self-standing legal obligations different from those which appear in the treaty’s subsequent articles or if this provision serves only as a contextual introduction to those later provisions. The Court endorsed the first alternative, confirming its reasoning by a reference to the preparatory work of this Convention. It observed that, in the course of the diplomatic negotiations, the state parties’ undertaking in Article I to prevent and to punish genocide was moved from the Preamble to this first Article.Footnote 82 In the Court’s view, this change ‘emphasizes the operative and non-preambular character of Article I’.Footnote 83
Deletions of clauses in earlier drafts are often taken to imply that whatever was erased cannot be read into the parties’ agreement.Footnote 84 For illustration, some scholars have suggested that the deletion of a draft article to the VCLT that would have established a broad principle of effet utile in treaty interpretation implies that the principle of effectiveness may only be relied upon to prevent a treaty (or parts of it) from being rendered redundant or meaningless.Footnote 85 Post-VCLT case law suggests, however, that a broader understanding of effet utile may be applicable in some scenarios.Footnote 86 On its own, the abandonment of an expression within the treaty’s draft may thus not be sufficient to draw adverse inferences.Footnote 87
Similar observations can be made with respect to defeated amendments, which have occasionally been recalled by the Court to draw inferences.Footnote 88 Nevertheless, in the absence of an explanation for the amendment’s rejection, this sole fact cannot signify a collective agreement to the opposite effect, for there can be many reasons behind the rejection.Footnote 89 For example, in LaGrand the Court made reference to the preparatory works of the PCIJ’s Statute while discussing the binding legal status of its own orders indicating provisional measures. The final draft of this Statute, in French, granted the Court ‘the power to suggest, if it considers that circumstances so require, the provisional measures that should be taken…’.Footnote 90
In later stages of the negotiating process over the French version of the text, an amendment was proposed to strengthen the language of this provision, by changing the word ‘indicate’ for ‘order’. Such proposal was not accepted. Nonetheless, the Court did not consider that the amendment’s rejection implied that its provisional measures were not binding since, in the Court’s view, such amendment’s dismissal ‘was motivated by the consideration that the Court did not have the means to assure the execution of its decisions’ and not by the lack of binding force of its orders.Footnote 91
Likewise, in Application of the International Convention on the Elimination of All Forms of Racial Discrimination (Qatar v. United Arab Emirates), the ICJ discussed whether this treaty’s prohibition of discrimination on the basis of ‘national origin’ applies to distinctions based on current nationality. The Court discussed the relevant travaux in some length, addressing the implications of an amendment jointly proposed by the negotiators from France and the United States to specify that ‘the expression “national origin” does not mean “nationality” or “citizenship”’, which was ultimately withdrawn.Footnote 92 In the Court’s opinion, no adverse inference could be drawn from the amendment’s withdrawal, since the concern that motivated France and the United States to table it in the first place had been addressed in another provision that made its way to the final text of the treaty.Footnote 93
A final example is provided by a 1992 decision from the Inter-American Court of Human Rights in the case Neira Alegría and Others v. Peru, where this Court discussed the impact of stylistic changes to a treaty during its negotiations. In that decision, the Court resolved an inconsistency between the authentic languages of the American ConventionFootnote 94 by having recourse to the travaux. It noted that during the final stages of the diplomatic negotiations, the final text of the treaty was modified by a stylistic commission, without there being any further discussion by state representatives about the final text. The Court argued that the final, stylized text should bear the same meaning as the previous, non-stylized one, which arguably reflected the views of the parties more faithfully.Footnote 95
4.6. The views of drafters and negotiators
Interpreters must not only consider the completeness of the preparatory works, but also the origin of a treaty’s drafts, their respective translations, and the changes made to such documents along the different stages of negotiations. The travaux often include minutes or records of discussions featuring the views of individual drafters and negotiators, which may be of value in the interpretive process.
Statements from such individuals may be appraised with caution, especially when they do not reflect a consensus of all the parties to the negotiations. For this reason, some domestic legal systems severely limit the scenarios under which travaux can have a discernible impact on the interpretation of a treaty. A famous example is the ‘bull’s eye’ rule in English law,Footnote 96 under which the travaux may only serve as ‘a legitimate guide to interpretation … if [they] yield a clear and authoritative answer’.Footnote 97
There is no equivalent to the bull’s eye rule under public international law,Footnote 98 where preparatory works may be relied upon even if they do not reveal the unanimous view of the parties. In other words, isolated comments from those attending the negotiations may provide sufficient evidence in support of one interpretation, when there are no indications to the contrary in the rest of the travaux. For illustration, in Jadhav (India v. Pakistan) the Court relied upon the comments of an ILC Special Rapporteur who was invited to one of the diplomatic conferences that resulted in the Vienna Convention on Consular RelationsFootnote 99 which, in the Court’s opinion, confirmed its reading of this treaty’s Article 36.Footnote 100 Similarly, in Application of the International Convention on the Elimination of All Forms of Racial Discrimination (Georgia v. Russian Federation), the ICJ made the following observation:
the usefulness of the travaux préparatoires in shedding light on the meaning of Article 22 is limited by the fact that there was little discussion of the expression ‘a dispute which is not settled’. A notable exception and one to which some significance must be attached is the statement by the Ghanian delegate… Footnote 101
A different but related question is concerned with the relative importance of viewpoints voiced in early stages of the negotiations as compared to those articulated later on. In this regard, the Appellate Body of the World Trade Organization has supported a parameter of temporal proximity, measured against the treaty’s conclusion, in order to assess the relevance of preparatory works. For illustration, the Appellate Body has asserted:
there is ‘some correlation between the timing of an event, act or other instrument… and their relevance to the treaty in question’, in the sense that ‘the further back in time of the event, act or other instrument took place, was enacted or was adopted relative to the conclusion of a treaty’, the less relevant it will be for interpreting the treaty in question. What should be considered ‘temporally proximate will vary from treaty provision to treaty provision’ and may depend on the structure of the negotiating process.Footnote 102
However, one should not rush to assume that whatever is said last should always prevail over what preceded it. In this article’s introduction, reference was made to the judgment of the Supreme Court of the United States in Sale v. Haitian Centers Council, which validated a policy of interdiction of asylum seekers on the high seas as in conformity with Article 33(1) of the Refugee Convention. To a significant extent, that Court reached its decision by emphasizing the views expressed by the Dutch delegate at a late stage in that treaty’s negotiating process – views which were ‘placed on record’. However, as the United Nations High Commissioner for Refugees (UNHCR) argued in its amicus curiae for this case, during such negotiations
formal amendments to the Convention to which delegates consented were ‘agreed’ or ‘adopted’. The Dutch delegate’s comment, by contrast, was merely placed on the record, a tacit acknowledgement that the views he expressed did not enjoy sufficient support to alter the actual language of the treaty.Footnote 103
The distinction made by the UNHCR, endorsed by Justice Blackmun in his dissent in Sale,Footnote 104 is all the more relevant since, in earlier stages of the drafting procedure, other delegates voiced positions supporting the applicability of non-refoulement to rejections at a country’s frontiers, implicitly acknowledging that this principle had at least some extraterritorial effects.Footnote 105 Irrespective of how the conflicting positions recorded on the travaux in this case should be resolved, the UNHCR’s amicus curiae highlights the importance of the specific rules, even tacit ones, of each diplomatic procedure, which caution against an unreflective preference for whatever the delegates happened to say last.
Similarly, in a detailed study of the travaux of the ICSID Convention, Julian Davis Mortenson argued that inter-State negotiations supporting a broad interpretation of the term ‘investment’ under Article 25 of this agreement were unduly downplayed by a later report from the World Bank’s Executive Directors, which has been relied upon by a string of investment tribunals favouring a narrow definition of ‘investment’.Footnote 106 The whole point of Mortenson’s article is that the earlier preparatory work should be afforded more weight than the report prepared after it.
4.7. Opposability of a state’s own views during the drafting process
The views expressed by a state’s representative during the diplomatic negotiations could perhaps be relevant in future disputes involving that state as a consequence of the principle of good faith, which ‘may require international actors to be consistent through time with respect to their approaches to treaty interpretation’,Footnote 107 thereby ‘preclud[ing] the simultaneous affirmation of incompatible legal positions’.Footnote 108 As a matter of fact, a few international decisions suggest that the interpretations advanced by a state’s representatives during a treaty’s negotiation and ratification process may be opposable to that state in future disputes.Footnote 109
For illustration, in HICEE B.V. v. The Slovak Republic the respondent supported its interpretation of a bilateral investment treaty between the Kingdom of the Netherlands and the Czech and Slovak Federal Republic by reference to the ‘Explanatory Notes’ submitted by the Minister of Foreign Affairs of the Netherlands to the States-General with the request of approval for the ratification of this treaty.Footnote 110 In the tribunal’s opinion, these ‘Dutch Explanatory Notes do not fit within any of the categories of extraneous material specified in Article 31 or Article 32’ of the VCLT; nevertheless, the tribunal asserted that such notes could be considered to be supplementary means of interpretation.Footnote 111 The tribunal felt compelled to grant significant weight to the content of the Dutch Explanatory Notes, for the benefit of the Slovak Republic, as a consequence of logic, good sense, and the principle of good faith in treaty interpretation.Footnote 112 It went on to state the following, obiter dicta:
[l]et us suppose, as a hypothesis, that the same question of interpretation of the Agreement was in the dispute between the Contracting Parties themselves, and that that dispute was in litigation … Had the matter gone into inter-State arbitration (or indeed to the International Court of Justice) it is surely inconceivable that the Netherlands would have wished to adopt a position in the litigation materially different from its formal public position at the time of ratification; or (in the unlikely contrary event) that the court or tribunal would have permitted it to do so. As the Latin maxim has it: nemo audietur venire contra factum suum.Footnote 113
Similar perspectives can be found in case law from other international legal regimes. A trade dispute that illustrates the plausible opposability of the views expressed by one state during the negotiations of a treaty is United States – Certain Measures on Steel and Aluminium Products.Footnote 114 In this case, a WTO Panel resolved an interpretative disagreement between the United States and China on whether the security exception in Article XXI(b) of the GATT 1994 was self-judging (as argued by the United States) or not. In the course of its deliberations, the Panel consulted ‘internal documents of the US delegation at the negotiations for the [International Trade Organization] Charter and the GATT 1947’.Footnote 115 Even though the Panel acknowledged the ‘limitations’Footnote 116 to the interpretive relevance of these documents, resulting from the fact that they ‘were not in the public domain, nor were … accessible to other delegations at the time the negotiations… took place’,Footnote 117 it did not consider that they were ‘in principle, entirely irrelevant as supplementary means for the interpretation of Article XXI(b)’.Footnote 118 In the Panel’s view, these records revealed that during the drafting process the delegation from the United States rejected the possibility that this security exception could be self-judging, a fact described by the Panel as ‘noteworthy’.Footnote 119
To some extent, the tribunal in the South China Sea Arbitration reasoned along similar lines, although for present purposes this case can only be examined with caution. After all, China did not appear in the arbitral proceedings, and so the tribunal endeavoured to appraise the legal position of such state through indirect means, like public statements from the Chinese Ministry of Foreign Affairs.Footnote 120 Notwithstanding the controversies surrounding this case, it is interesting to note that in order to determine whether the United Nations Convention on the Law of the Sea (UNCLOS)Footnote 121 allows the preservation of rights to resources which are at variance with that Convention and established before it entered into force, the Court considered the views expressed by the Chinese representative during the negotiation of that treaty. In the tribunal’s opinion, some comments of that representative contradict China’s current position, as ascertained by the tribunal. The relevant passage of the award reads as follows:
[t]he Tribunal notes these comments not because the remarks of any particular State during the negotiation of a multilateral Convention are indicative of the content of the final treaty, but because China’s resolute opposition to any accommodation of historic fishing is largely representative of the position that prevailed in the final text of the Convention. The Tribunal also notes that China’s position, as asserted during the negotiation of the Convention, is incompatible with a claim that China would be entitled to historic rights to living and non-living resources in the South China Sea that would take precedence over the exclusive economic zone rights of other littoral States …Footnote 122
Admittedly, the interpretive approach followed in the three cases studied above can be controversial. The internal deliberations of a state during a treaty’s negotiation (i.e., United States – Certain Measures on Steel and Aluminium Products) or ratification (i.e., HICEE B.V. v. The Slovak Republic) might not be reliable indicators of the common understanding of all treaty parties. At most, these records will usually contain evidence of how the executive (and, in the case of domestic approval proceedings, also the legislature) of one state read the treaty at hand. If treaty interpretation is the art of finding a resolve that was shared by all the state parties, the elucidation of one of the parties’ intentions can only tell one side of the story. Other pieces are needed to complete the puzzle. Relatedly, the record set by diplomatic negotiations proceedings (i.e., South China Sea) can be of little relevance in the absence of evidence that one party’s intentions were shared by its counterparts. It was for this very reason that Quincy Wright posited almost a century ago that preparatory works could hardly be relevant when interpreting multilateral treaties.Footnote 123
A fourth and final case concerning the opposability of a state’s position during the negotiations of a treaty, where the abovementioned concerns would not be present is the Case concerning the Interpretation of the Air Transportation Services Agreement between the United States of America and France. In this case, the tribunal had to decide if the treaty between the parties granted a United States airline the right to provide international aviation services between the United States and Turkey (on the one hand) and Iran (on the other), travelling in either case via Paris. The relevant treaty provision authorised air transit from the United States ‘to Paris and beyond via intermediate points in Switzerland, Italy, Greece, Egypt, the Near East, India, Burma and Siam to Hanoi, and hence to China and beyond, in both directions’.Footnote 124
The parties disagreed on whether the term ‘Near East’ encompassed Turkey and Iran. The United States argued in favour of a broad meaning that equated it to the entire Middle East, while France claimed that such term only designated the Arab part of the former Ottoman Empire.Footnote 125 The tribunal favoured the arguments advanced by France, based purely on the text of the treaty itself and ‘a general principle indispensable to the orderly development of air traffic’, concerning the ‘reasonable direct route’ between the countries that are immediately before and after the term ‘Near East’ in the provision under examination (that is, Egypt and India).Footnote 126
Afterwards, the tribunal proceeded to verify whether the travaux préparatoires confirmed its interpretation of the treaty. The tribunal noted that at various points during the diplomatic negotiations, the United States provided documents to France (including a first draft of the treaty) that could reasonably imply that the term ‘Near East’ was intended to portray a corridor between Greece and India that, roughly speaking, would pass through Cairo and Dhahran. The tribunal also noted that the United States’ choice to employ the term ‘Near East’ in its communications with France and the draft of the agreement was plausibly explained ‘[by] the fact that the political status of the territories concerned did not, at the time, appear to be very definite or stable’.Footnote 127 More generally, in a passage that is worth quoting in full, the tribunal observed:
the American authorities contributed by similar communications delivered on three different occasions, to establishing in the mind of the French Government a well-formed idea of both the general and the specific path of the route that was being proposed, and on which it was being requested to grant rights. By so doing, they allowed the negotiators of the other Contracting Party to convince themselves that… the term ‘Near East’ was not being employed in the widest of its various possible meanings, but to designate a region that, in spite of its breadth was nevertheless limited to the cartographic indications furnished by the Americans on their own initiative. The American Government, in the opinion of the Tribunal, must thus accept the fact that conduct of this nature during the preliminary negotiations could be urged against it on an issue of interpretation of the Agreement which resulted.Footnote 128
The tribunal’s reasoning was therefore influenced by notions of estoppel and contra proferentem. Indeed, since officials from the United States had made overtures to their French counterparties that could reasonably be interpreted as narrowing the geographical scope of the ‘Near East’, it would be inequitable to allow the United States to take advantage of the vagueness of a treaty term it had originally drafted.
Relevantly, a similar rationale would be difficult to sustain with respect to multilateral treaties that are adopted for purely humanitarian purposes. As once explained by the ICJ, when concluding such conventions ‘the contracting States do not have any interests of their own’ and ‘one cannot speak of individual advantages or disadvantages to States, or of the maintenance of a perfect contractual balance between rights and duties’.Footnote 129
Under international law, estoppel may be invoked to prevent a state from backtracking or contradicting a representation it made to another state, when the latter was induced by the former ‘to act to its detriment, to suffer a prejudice, or to convey a benefit upon the representing State’.Footnote 130 It stands to reason that the principle of estoppel paradigmatically operates in contract-like arrangements, and not in multilateral treaties with purely humanitarian purposes. In these scenarios, considerations of estoppel cannot provide a justification for the opposability, against a state, of the views expressed by its representatives in the course of treaty negotiations because, in such context, the relative positions of the parties between themselves are not affected by the agreement they might reach.
5. Critical analysis – Is there any need (or wisdom) in systematizing the aforementioned factors for the interpretation of travaux?
When Sir Hersch Lauterpacht delivered his course at the Hague Academy in 1934, he observed that there are no ‘ready made rules’ for interpreting travaux. Importantly, he asserted that such rules should not be drawn up because ‘the process of examining preparatory work is a pragmatic one’.Footnote 131 In response to Lauterpacht’s concern, it can be observed that international courts and tribunals routinely study these materials following unclear and often undeclared methodologies. The lack of clear guidance on the matter does not make an interpreter’s task easier; to the contrary, it weakens the analytical rigour with which the treaties themselves are elucidated and opens avenues for cherry-picking and arbitrariness.
Notwithstanding the above, there are weighty policy-oriented reasons that could militate against the standardization of a set of factors to interpret travaux, because they could end up complicating the already delicate work of diplomats.Footnote 132 Indeed, state negotiators might become warier of expressing any views on the record about the meaning of a drafted treaty provision to prevent such positions from becoming opposable to them in future international disputes. Negotiators may no longer be inclined to table amendments unless they are confident that other parties will accept them, because either a rejection of their proposal or a recorded divergence of views could be interpreted by future adjudicators as indications to dismiss an otherwise sound reading of the final treaty text. In this regard, some states have expressed the view that ‘[c]rucial negotiations might be seriously impaired – if not made impossible – when everything that is said risks to become part of a record and thus part of the public domain’Footnote 133 and that detailed records might ‘inhibit supple and productive negotiation of differences’.Footnote 134
In short, a strict set of rules on the interpretation of preparatory works could give rise to unproductive incentives in the drafting and negotiating of treaties, as diplomats would need to second-guess the potential consequences of any words that are placed on record.Footnote 135 The resulting framework could be inimical to the atmosphere of trust and frankness in which state parties could find the most convenient terms to express their agreement, as measured by the clarity and specificity of the words they employ. One could thus imagine that a rigid set of rules for the elucidation of travaux could have the unintended effect of producing slimmer diplomatic records and imprecisely worded treaties.
More generally, the systematization of factors outlined above could in many cases distract the attention of interpreters from the ultimate goal of their endeavour, which under the VCLT rules consists in the identification of the will of the parties as reflected in the treaty itself. In the context of dispute settlement, a project like this also entails a risk of overburdening the task of litigants and adjudicators, to the prejudice of judicial economy. As it was once observed by Sir Percy Spender,
there is always a risk that, instead of interpreting the relevant treaty or convention, one will find oneself tending to interpret the preparatory work and then transferring that interpretation across to the treaty or convention which is the sole subject of interpretation.Footnote 136
This concern might be all the more pressing considering the fact that travaux rarely prove to be decisive in dispute resolution. There is therefore a need to balance any prospective benefits derived from the further refinement of a technique to elucidate these materials against the probability that this endeavour might not be worth the candle.
Notwithstanding the above, it is submitted that the potential benefits of a systematization of factors that should be considered in the interpretation of travaux outweigh their drawbacks. If recent ICJ jurisprudence can serve as an indication on the matter, in most cases it will be unnecessary to devote much attention to the preparatory works of a treaty. However, in the minority of scenarios where these materials might be relevant, it can be helpful to have some broad guidelines on how to read them. Interpreters should not have to reinvent the wheel every time they have to deal with travaux. As outlined in this article, there are reasonable guard rails, scattered throughout international case law and scholarship, to assist them in this process.
6. Conclusion
Most of the academic literature concerning the preparatory works of treaties has been devoted to the definition of this notion and the conditions under which an interpreter may rely on such materials. By contrast, little has been said about the way in which the travaux themselves should be elucidated.
This article has attempted to fill this gap, discussing several factors to identify the implications that may be legitimately drawn from preparatory works. As elaborated above, when interpreting preparatory works one should assess their completeness, determine if any consequences may derive from the drafts’ authorship, consider the variations in meaning conveyed by the drafts’ translations, study the evolution of the treaty text throughout its negotiations, appraise whatever relevance the views expressed by some delegations may have, and contemplate the consequences that may follow from a contradiction between the views currently held by a state and those that were embraced by its officials before the treaty entered into force.
Open access
