1. Introduction
In the universe of international law, the rules of treaty interpretation occupy a singular place: the claim that they actually direct behaviour of legal operators is met with scepticism not only by proponents of critical approaches or other variations of rule-scepticism, but also by international legal scholars who otherwise believe in international law’s normative power. It might not be a stretch to analogize the rules of treaty interpretation with the month-long weather forecasts that Kenneth Arrow and his team of statisticians were tasked to produce during the Second World War. Having realized that their forecasts were almost consistently inaccurate, the team asked to be relieved of its duty, but was told the following: ‘The Commanding General is well aware that the forecasts are no good. However, he needs them for planning purposes.’Footnote 1 Paraphrasing this story, one could say that international lawyers know full well that the rules of treaty interpretation do not really guide the process of identifying the meaning of treaty provisions, but they continue to use them for other purposes ranging from post-hoc rationalization of the outcomes generally reached through other means to the strategic occultation of law-making commonly thought to be involved in most interpretive decisions.
This general attitude about the rules of treaty interpretation is well captured by the often-repeated adage that treaty interpretation is more akin to art than to science. Exactly what is meant by this adage may not be fully clear, all the more so because the adage itself has several variations. In the most common formulation simply contrasting art and science, the message sought to be conveyed seems to be that, similarly to art, no general rules can be developed as to the best way in which interpretation can be practiced.Footnote 2 The version originally formulated by the International Law Commission, i.e., ‘the interpretation of documents is to some extent, an art not an exact science’,Footnote 3 seems, however, subtly different in that the qualifiers ‘to some extent’ and ‘exact’ can be taken to suggest that there is something meaningfully law-like and general that can be said about interpretation even though no full predictability or determinacy can reasonably be expected, unlike what happens with general laws formulated by exact sciences. Such nuances aside, the adage that interpretation is an art rather than a science is generally used to suggest that treaty interpretation cannot be subject to regulation in the same way as other issues of international law. This general sense that there is something special about the rules of treaty interpretation is made clear by the fact that the ‘art versus science’ comparison is rarely, if ever, used with respect to other rules of international law.
Scepticism about the rules of treaty interpretation is sometimes expressed through another dichotomy, namely the distinction between esprit de géométrie and esprit de finesse introduced by Blaise Pascal.Footnote 4 A closer look at Pascal’s discussion of the distinction makes clear that the latter is no different from the ‘art versus science’ dichotomy. According to Pascal, ‘accustomed to the exact and plain principles of mathematics, and not reasoning till they have well inspected and arranged their principles, [mathematicians] are lost in matters of intuition where the principles do not allow of such arrangement’.Footnote 5 This is so, argues Pascal, because matters of intuition require judging ‘at a single glance’, in other words, proceeding ‘tacitly, naturally, and without technical rules’.Footnote 6 When transposed to the domain of treaty interpretation, the Pascalian dichotomy is meant to suggest that interpretation is a matter of intuitive rather than mathematical mind and that, as such, it cannot be a proper subject of legal regulation, since one cannot design technical rules governing intuitions.
Despite what it implies with respect to rule-like constraints over treaty interpretation, the above-described scepticism rarely prevents international lawyers from formulating rules allegedly governing treaty interpretation and discussing them at length, which reminds one of the following thought experiment Wittgenstein proposes in Philosophical Investigations: ‘I have locked the man up in the room – there is only one door left open.’Footnote 7 As Wittgenstein points out, faced with a statement like this, one is tempted to say that the man was not locked up at all. Isn’t ‘[a]n enclosure with a hole in it … as good as none?’Footnote 8
The present article asks the latter question about the rules of treaty interpretation: if all those rules can manage to give us is ‘[a]n enclosure with a hole in it’, can we say that they are as good as not having those rules at all? Stated differently, the objective of this article is to inquire whether, despite their much-discussed shortcomings, the rules of treaty interpretation have any ‘cash value’ in the sense given to this expression by William James, one of the founding fathers of the philosophy of pragmatism. In James’ writings, the ‘cash value’ of a concept designates the practical difference, if any, that it makes in actual experience.Footnote 9 Asking what the ‘cash value’ of the rules of treaty interpretation is amounts to asking what practical impact these rules have and whether a world without these rules would be different in a practically meaningful sense from our world.
In order to respond to these questions, it is helpful to first consider what international lawyers expect from the rules of treaty interpretation (Section 2). It is the contention of this article that the rules of treaty interpretation cannot deliver what is traditionally expected from them, since there is a mismatch between what they are designed to do and their traditionally expected role (Section 3). The last section of this article will revisit the perennial question of whether treaty interpretation is a constrained activity in light of this article’s main contention that the rules of treaty interpretation are not designed to govern meaning-ascertainment as strictly understood (Section 4).
2. What do international lawyers expect from the rules of treaty interpretation?
The scepticism generated by the rules of treaty interpretation has much to do with what international lawyers commonly expect from those rules. A closer look at this scepticism seems, thus, appropriate in order to assess the ‘cash value’ of the rules of treaty interpretation. Admittedly, it would make little sense to assess the ‘experiential worth’Footnote 10 of the rules of treaty interpretation in light of the fact that they have disappointed a set of expectations if one has no reason to have such expectations in the first place.
A first, frequently-voiced, criticism is that, given their high level of generality, the rules of treaty interpretation are unlikely to be of much assistance in resolving concrete interpretive disputes.Footnote 11 This was the position of the European Commission of Human Rights in Golder v. United Kingdom. Faced with the issue of whether the rules of interpretation of the Vienna Convention on the Law of Treaties (VCLT or Vienna Convention) could be applied even though the Convention was not in force yet, the Commission pointed to ‘the rather general’ character of those provisions and stated that ‘For the limited guidance one may find in these provisions, it is … not material whether the Vienna Convention is in force for the contracting parties.’Footnote 12 More recently, Joe Verhoeven expressed the view that Article 31 of the Vienna Convention titled ‘General Rule of Interpretation’ was largely devoid of practical significance, since it was hard to imagine a rule prescribing that one should ignore the terms of the treaty, interpret them in bad faith or without due regard to what the parties sought to achieve.Footnote 13 Even the official discourse has abandoned the pretense that the rules of treaty interpretation lead to one single correct meaning, accepting that there may be ‘more than one permissible interpretation’Footnote 14 or that ‘interpretation … leaves room for discussion’.Footnote 15
A second source of discontent with the rules of treaty interpretation is that they are not of immediate application. This criticism in turn can take several forms.
First, it is often pointed out that the rules of treaty interpretation themselves may need to be interpreted. As was nicely summarized by the Greek delegation at the United Nations Conference on the Law of Treaties:
If a treaty contained one or more rules as to its interpretation, those rules themselves would need to be interpreted, but at that point no rules of interpretation would be available. Even if a treaty provided rules for the interpretation of clauses regarding interpretation, those provisions would require to be interpreted by means not contained in the treaty. There [is] a vicious circle and thus it would be vain to set down rules about interpretation.Footnote 16
This difficulty can easily be illustrated. Consider the interpretive directive that the interpreter shall take into account: ‘the relevant rules of international law applicable between the parties’. One cannot apply such a rule before first establishing what is meant by ‘relevant’, ‘rules’ or ‘the parties’.Footnote 17 ‘Subsequent agreement and subsequent practice’ is another example, as shown by the recent lengthy discussion of their possible definitions by the International Law Commission.Footnote 18
Second, because the rules of treaty interpretation make a plurality of interpretive approaches available for use without establishing a meaningful order of priority among those approaches, it is thought that they leave room for a large amount of discretion unregulated by rule-like constraints.Footnote 19 While the interpretive regime under the Vienna Convention is presumptively text-based, the latter also prescribes that the terms of the treaty should be considered in light of the treaty’s object and purpose, and that the ordinary meaning should be disregarded when it is established that such was the intention of the parties.Footnote 20 This was described as ‘a bolting-together of conflicting ideas’, arguably making Article 31 of the Vienna Convention ‘worthless as a general rule of interpretation’.Footnote 21
Lastly, it is argued that the ‘means’ of interpretation to which the rules of interpretation direct the interpreter can hardly resolve interpretive disputes when those means themselves call for interpretation. For instance, the ‘ordinary meaning’ rule set forth in Article 31 of the VCLT cannot end an interpretive inquiry when the ordinary meaning of the term being interpreted remains open to interpretation or there is more than one plausible ordinary meaning.Footnote 22 Similarly, ‘[i]t is not normal practice in treaty drafting to spell out the “object and purpose” as if one were defining technical terms’.Footnote 23 What this means is that the object and purpose of a treaty need to be interpretively identified just as the meaning of the treaty provision they are supposed to help determining. The same is true of travaux préparatoires: they cannot be used in support of an interpretation if what they mean is not interpretively determined in the first place.
Although the sceptical charges described above seem to highlight varied concerns, they ultimately stem from one unarticulated expectation, namely the expectation that the rules of treaty interpretation should directly establish the meaning of a treaty provision. This expectation in turn rests on the assumption that the point of the rules of treaty interpretation is to bridge the gap between the signifier (the treaty provision being interpreted) and the signified (the meaning of the treaty provision). In other words, it is assumed that when applied to a treaty provision, the rules of treaty interpretation should lead the legal operators to the correct meaning of that provision without the intermediation of anything else.
This assumption is nothing more than the myth of an unmediated access to the signified, and as such, is not specific to the rules of treaty interpretation, or to law for that matter. That no phenomenon is accessible to human mind in an unmediated manner is the primary lesson conveyed by the famous Derridean slogan that ‘there is nothing outside of the text’.Footnote 24 As Derrida clarified, ‘the text’ here is not limited to something ‘graphic’ or a set of signifiers inserted in a book, but refers to ‘all possible referents’ that are part of what is called ‘reality’, the point being that the latter cannot be accessed outside ‘an interpretive experience’.Footnote 25
But it would be simplistic to believe that the above misapprehensions about treaty interpretation are only due to international lawyers’ lack of philosophical sophistication, for they also reflect a structural anxiety about interpretation in general. Law’s official image that it reflects governance by rules, not by men who happen to be in charge of administering the latter, can only seem plausible if the domain of interpretation is limited to the minimum. Interpretation represents a danger to the integrity of law, implying as it does a certain level of discretion, which is antithetical to certainty that is seen as an essential part of any legal system worth the name. If the rules of interpretation can justify more than one interpretation without any guidance as to which interpretation is the right one, then the choice of a particular interpretation cannot be said to be imposed by law.Footnote 26
H.L.A. Hart’s famous discussion of the separation of law and morals clearly displays this concern. Hart concedes that application of law in ‘penumbral cases’ is not a mechanical exercise, since it involves consideration of ‘aims, purposes, and policies’.Footnote 27 But he insists that what is ‘centrally important’ to law is that it has ‘the hard core of settled meaning’, which supposedly imposes itself without any interpretive effort.Footnote 28 To say that ‘all legal questions are fundamentally like those of the penumbra’ is, Hart argues, tantamount to saying that ‘there is nothing in the nature of a legal rule inconsistent with all questions being open to reconsideration in the light of social policy’.Footnote 29 In other words, a legal system would not be a genuine legal system if all of its rules were open to interpretation.
One finds such reluctance vis-à-vis interpretation when international lawyers are confronted with the necessity of interpreting the means of interpretation. For instance, Sir Percy Spender expressed his concerns with the use of travaux préparatoires by pointing to ‘the danger 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 30 More recently, an arbitral tribunal confronted with Explanatory Notes provided by one of the parties to the treaty as part of its domestic ratification process stated:
[The tribunal’s] task is to interpret the Agreement, not to interpret the interpretative materials, although it must be obvious that no tribunal could bring into play any of the interpretative materials mentioned in Articles 31 and 32 of the Vienna Convention without forming a view as to what the material in question indicates.Footnote 31
What is remarkable in this passage is the tribunal’s attempt to limit its interpretive task to the agreement itself, with the tribunal avoiding the very term ‘interpretation’ in the description of what it sets out to do with respect to the interpretive material at issue. But the euphemistic substitute ‘forming a view as to what the material in question indicates’ is, of course, no different from interpreting ‘the material in question’, since the meaning can only be accessed interpretively. As will be shown in the next section, such attempts to evade interpretation evince a fundamental misunderstanding about what the rules of treaty interpretation are designed for.
3. The proper province of the rules of treaty interpretation
To understand where the ‘cash value’ of the rules of treaty interpretation lies, it would be helpful to look at another discipline in which interpretation plays a central role, namely, literary studies.Footnote 32 Theoretical accounts of meaning offered by literary theorists have traditionally differed over the proper role of the author, the text, and the reader in the production of meaning. For some, interpreting a text is a matter of finding out what its author meant.Footnote 33 For others, interpretation is only about identifying the meaning of a text.Footnote 34 Still others believe that the source of the meaning lies in the readers.Footnote 35 It is, however, widely accepted that the plausibility of a particular reading is a matter of interpretive modi operandi shared within the discipline of literature. Steven Mailloux labels those modes of reading ‘interpretive conventions’, which he defines as ‘shared ways of making sense of reality’ or ‘group-licensed strategies for constructing meaning’.Footnote 36 Jonathan Culler talks about ‘reading conventions’ in light of which literary works are read and recognized as pieces of literature.Footnote 37 Following Derrida, Robert Scholes refers to ‘protocols of reading’, which he describes as ‘canons’ of interpretation.Footnote 38 For Stanley Fish, ‘recognized interpretive strategies’ act as ‘canons of acceptability’ of particular interpretations.Footnote 39 However called, these constraints have a disciplining effect, marking as they do the boundaries of acceptable readings, and their mastery is an integral part of ‘literary competence’.Footnote 40 No reading of a literary work can lay claim to legitimacy unless it is produced by a recognized mode of reading.
The rules of treaty interpretation can be seen as performing the same function in international law. A reading of a treaty provision is legally competent only if it can plausibly be derived from means to which a treaty interpreter can admissibly have recourse under the rules of treaty interpretation.Footnote 41 In this regard, the ruling of the WTO Appellate Body that ‘permissible interpretation’ is ‘one which is found to be appropriate after application of the pertinent rules of the Vienna Convention’ can be said to reflect a systemic reality in modern international law.Footnote 42 When a reading can plausibly be said to be derivable from particular means of interpretation cannot be a matter of rule-like constraints, since the response depends on numerous variables such as the power of persuasion and the authority of the interpreter or the nature of the audience. What can be, and is, subject to constraints is how the game of interpretation can be played and what moves the players are allowed to make.Footnote 43 In this sense, the rules of interpretation are similar to the rules of evidence that govern the introduction and use of evidence in legal proceedings. While they do not dictate the final outcome by themselves, they frame and channel the proceeding that leads to it by acting as the gate-keeper of legitimate interpretive moves.Footnote 44 In other words, they determine what ‘sources of interpretive data’Footnote 45 are admissible, and as such, are part of the ‘legal grammar’ of international law defined by Koskenniemi as ‘the system of production of good legal arguments’.Footnote 46
What constraints such gate-keeping actually imposes on treaty interpreters can again be illustrated by examples from literature. James Miller famously offered a new reading of Eliot’s The Waste Land according to which, during his stay in Paris, Eliot had developed a love relationship with a Frenchman named Jean Verdenal who was later killed in the First World War, suggesting that The Waste Land and some of Eliot’s later poems expressed his grief over the loss of Verdenal.Footnote 47 Some scholars disputed this readingFootnote 48 but often times they did so not because they thought that personal episodes from Eliot’s life cannot illuminate his work, but because they challenged the factual accuracy of Miller’s account of Eliot’s relationship with Verdenal.Footnote 49
Similarly, the letters written by Gustave Flaubert to his mistress Louise Coulet recording his progress with the writing of Madame Bovary, and offering his views on the novel’s themes, art, literature, and literary criticism are relied upon to interpret Madame Bovary to such an extent that those letters are thought to ‘form an inseparable pendant to the Bovary project’.Footnote 50 Such sources of interpretation are potentially limitless in literature, ranging from personal life of the author to his broader worldviews, provided that they can plausibly illuminate the meaning of a literary work.
Imagine now a treaty interpreter offering an interpretation based on a private letter of a negotiator or an episode of the personal life of the chairperson of a drafting committee that produced the text of the treaty. Such interpretive moves would unlikely be recognized as a competent reading, precisely because they do not fit the repertory of admissible modes of treaty interpretation consecrated by the rules of treaty interpretation.
To appreciate this difference, we can try to imagine a world with a different regime of treaty interpretation. As recognized in practice, there is, in theory, no cogent reason why treaties cannot be interpreted in a manner departing from the rules set forth in the Vienna Convention on the Law of Treaties.Footnote 51 One need look no further than the pre-Vienna Convention debates on treaty interpretation in scholarly writings, at the Institut de droit international and in the International Law Commission to realize that such a scenario is not speculative. For instance, writing at the beginning of the twentieth century, John Westlake stated:
The important point is to get at the real intention of the parties, and that enquiry is not to be shackled by any rule of interpretation which may exist in a particular national jurisprudence but is not generally accepted in the civilised world.Footnote 52
Charles Cheney Hyde later expressed a similar view:
As no principle of law deters contracting States from employing the terms of their agreement in any sense they choose, or prevents those charged with the task of interpretation from giving heed to all circumstances probative of the choice actually made, it may be doubted whether the enunciation of technical rules of construction serves a useful purpose. It is the freedom of such States, and likewise of tribunals interpreting their acts, which rather deserves emphasis; and there must be intolerance of whatever serves to obscure perception of it.Footnote 53
In his report to the Institut de Droit International, Lauterpacht defended a similar conception of treaty interpretation, labelling the common intention of the parties ‘a predominant factor that must not be sacrificed to presumptions or technical rules of interpretation’.Footnote 54 This conception was shared by some members of the International Law Commission during the codification of the law of treaties.Footnote 55 The theory that there should be no restriction on the means of identification of the common intention of the parties was also supported by some governments. For instance, the Greek government pointed out that, ‘since a treaty is an expression of the common intention of the parties, the only basic rule of interpretation is to ascertain that intention by every possible means in every possible way’.Footnote 56
However, this interpretive philosophy based on the common intention of the parties unrestrained by any rigid rule-like framework did not prevail. The Vienna Convention introduced a double constraint drastically limiting the place of intention in treaty interpretation. First, interpretation under the Vienna Convention is not supposed to start with an autonomous search for the intention of the parties; the interpretive enterprise under the Vienna Convention rests instead on ‘the elucidation of the meaning of the text’.Footnote 57 This is so not because the common intention of the parties has no relevance under the Vienna Convention but because of the assumption that the text of the treaty is the primary vehicle through which the parties express their intention.Footnote 58 Second, under the Vienna Convention, the meaning of a treaty text cannot be elucidated ‘by every possible means in every possible way’ but by resorting to a limitative set of means, which imposes ‘certain common disciplines on treaty interpreters’.Footnote 59
The ‘cash value’ of these disciplines is clear to every practitioner who has argued a treaty interpretation issue before international adjudicatory bodies. No competent member of the international legal community could believe that an interpretive argument can be made ‘by every possible means in every possible way’. On the contrary, it is well understood that an interpretive argument has little chance to succeed if it is not based on a means recognized by the rules of treaty interpretation.
This function of the rules of treaty interpretation is crucial to a successful communication between treaty makers and treaty interpreters. Treaty making is, among other things, an instance of communication between treaty makers and the actors called upon to apply the treaty. As influentially argued by Donald Davidson, a speaker who wants to be understood knows that he must speak ‘in such a way that he will be interpreted in a certain way’.Footnote 60 But the speaker cannot appreciate how the interpreter will interpret his utterances without forming ‘a picture of the interpreter’s readiness to interpret along certain lines’.Footnote 61 This picture is based on what Davidson calls ‘the prior theory’ of the interpreter, which reflects how the interpreter is disposed ‘in advance’ to interpret the speaker’s utterances.Footnote 62 The success of the communication in this Davidsonian model is, however, a matter of convergence between the respective ‘passing theories’ of the speaker and the interpreter, namely, ‘the theory the speaker intends the interpreter to use’ and ‘the one the interpreter actually uses to interpret [the speaker’s] utterance’.Footnote 63 The prior theories and the passing theories may or may not coincide. If there is evidence that the speaker intended the interpreter to use a passing theory different from his prior theory, the interpreter is expected to adjust his prior theory based on such evidence. With more ‘speech transaction’ and more familiarity between the speaker and the interpreter, their respective prior and passing theories become more aligned with each other and, as a consequence, the success of their communication becomes more likely.Footnote 64
The rules of treaty interpretation form the basis of the theory of interpretation that the international legal operators use in the sense of the Davidsonian model.Footnote 65 They, indeed, enable the treaty makers to appreciate how their utterances will likely be interpreted, since they know what ‘interpretive data’ the interpreters will likely use. To put it in Davidson’s terms, the widespread use in practice of the rules of interpretation of the Vienna Convention has given rise to a convergence of the prior and passing theories of the treaty makers and treaty interpreters. This alone would be significant enough, but there is more, since by so doing, the rules of treaty interpretation have made what can be called ‘authorial control’ over treaty interpretation possible. Treaty drafting would be a meaningless exercise if the treaty drafters had no way to constrain the future treaty interpreters. How treaties are read have an influence on how they are drafted because ‘[t]o intend a meaning is to postulate reactions of an imagined reader who has assimilated the relevant conventions’.Footnote 66 This connection between the rules of interpretation and treaty drafting was noted by the International Law Commission.Footnote 67 As pointed out by Allott, ‘[t]he drafters, knowing that the text will be interpreted by others, are seeking to influence their interpretation in advance, to programme it so far as possible’.Footnote 68 It is well-known that by varying the level of precision of treaty provisions, the drafters can control the level of discretion enjoyed by treaty interpreters.Footnote 69 A treaty provision prescribing that a maritime space between two states with opposite coasts should be delimited by agreement ‘in order to achieve an equitable solution’ leaves more freedom to the interpreter than the rule stating that a monitoring body is composed of X number of members. But no such control can be exercised if the treaty drafters cannot anticipate what means treaty interpreters will use when interpreting the treaty. An important part of the ‘cash value’ of the rules of treaty interpretation is that they make such anticipation possible. For instance, the premium placed on textual precision as a way of controlling interpretation has much to do with the prominent status of text-based interpretive moves under the current regime of treaty interpretation in international law. The impact of textualism in this regard was highlighted during the Vienna Conference on the Law of Treaties by the representative of Sweden:
Whereas the textual approach [does] not entail [the dangers of the subjective approach], it [has] the drawback, or hardship, that it [requires] representatives of States drafting the text of a treaty to consider all the implications of a subsequent textual approach to interpretation in the event of a dispute; it [calls] for energetic efforts to achieve the utmost clarity and completeness in formulating the text of a treaty.Footnote 70
It is also reasonable to assume that, if a particular means of interpretation makes one interpretation look more plausible than another in a specific instance, the fact that it is available to be resorted to would make a difference. A recent empirical study showed, for instance, that whether and under what circumstances travaux préparatoires can be used can have a significant impact on interpretive outcomes.Footnote 71 A similar point was made by a WTO Panel with respect to the mandate of Article 31(3)(c) of the Vienna Convention that ‘any relevant rules of international law applicable in the relations between the parties’ be taken into account.Footnote 72 In other words, the rules of interpretation matter because ‘they mobilize some categories for interpretive duty and leave others uncalled’,Footnote 73 and in so doing, they impact the outcome of interpretive disputes.Footnote 74
4. Is treaty interpretation a constrained activity?
This article has so far attempted to establish that the rules of treaty interpretation do not, by themselves, determine the meaning of treaty provisions, since they only specify what means can be used in the course of that exercise.Footnote 75 If one accepts this contention, which basically means that the dedicated rules of treaty interpretation do not control the interpretive process entirely, the question arises whether treaty interpretation can still be characterized as a constrained exercise. Stated differently, what would be the point of knowing which means can be legitimately used in treaty interpretation if the actual use of those means and what can be made of them remains outside the reach of legal rules?
A possible response inspired by Wittgenstein’s philosophy of rules could be that the very assumption that treaty interpretation can be meaningfully subject to rule-like constraints only if it is a practice regulated by rules in every aspect is misguided. To use Wittgenstein’s game analogy, the game of interpretation is never a game ‘everywhere bounded by rules’, one ‘whose rules never let a doubt creep in, but stop up all the gaps where it might’.Footnote 76 But the fact that a game is not bound by rules in every aspect does not necessarily mean that it is not properly rule-bound. As Wittgenstein observes, ‘there are [no] rules for how high one may throw the ball in tennis, or how hard, yet tennis is a game for all that, and has rules too’.Footnote 77
This response may, however, not look fully satisfactory. After all, what is left unregulated in the game of tennis is not as consequential as what is left untouched by the rules of treaty interpretation in the overall schemes of these two enterprises. To respond to the question above, we need to revisit our very understanding of the phenomenology of rules. A fundamental point to bear in mind in this regard is that no rule can determine the conditions of its own applications. Even such a simple and arguably universal sign such as showing direction with a finger cannot determine what the proper way of interpreting it is. As Wittgenstein points out, there is nothing in the gesture itself telling us which way we have to go: ‘whether in the direction of [the] finger or (for example) in the opposite one’.Footnote 78
Considered in isolation, rules also cannot tell us anything about the identity, object, and purpose of the practice they regulate. As stressed by Hubert Schwyzer, if we were to meet a tribe whose members move the chess pieces around on a chessboard in perfect accordance with the rules of chess, but perform those moves not to win or lose like in the game of chess, but only as an integral part of a sacred rite in order to find out what fate the gods have reserved for the tribe, it would be meaningless to say that they play chess. This is so because ‘the rules of a practice do not tell us what the practice is of which they are the rules’.Footnote 79 The rule that an activity should be treated as a game or a sacred rite is not itself part of the rules governing that activity.Footnote 80 What this means is that the ‘grammar’ of a practiceFootnote 81 can never be exhausted by its rules, as it is also a function of the attitude of the relevant community towards that practice, the latter’s place in the life of the community or what the members of the community consider to be appropriate things to do when they engage in such practice. For instance, it is the grammar of the game of chess, not its rules, that tells us that it is appropriate to ask who won, praise the player because of a particular move or, more generally, see the activity of playing chess as involving a recreational contest and not a sacred rite.Footnote 82
The rules governing treaty interpretation are not an exception in this regardFootnote 83 but this does not mean that treaty interpretation is not a regulated activity. The lesson of the Wittgensteinian insights above simply is that the gap between a signifier and a signified is not bridged by a rule-dictated interpretation but by the practice into which the members of the relevant community have been socialized.Footnote 84 That the means of treaty interpretation listed in the Vienna Convention have themselves to be interpreted because one cannot determine what a sign means without referring to other signs is not a reason to think that interpretation actually involves an endless process of ‘deferral’ in the universe of signs.Footnote 85 As Roland Barthes observed, a dictionary is ‘a vertiginous object’, since each word in a dictionary is defined by reference to other words, which in turn, are defined by reference to still other words, with no possibility to put an end to this process in theory.Footnote 86 However, when we use a dictionary, we never go through an endless process, since the practice of consulting a dictionary is always informed and shaped by the specific context in which we are involved whose purposes and constraints put an end to a potentially endless regression. The parameters of every interpretive practice are ‘always already’ known by the competent members of the international legal community, who know that there are only a limited number of options that are plausibly available in the game of persuasion that law largely is about.
This does not mean that interpretive questions leave no room for doubts, but part of being a competent member of a professional community means that one does not experience every theoretically imaginable doubt as an actual doubt.Footnote 87 Wittgenstein invites us to imagine ‘someone always doubting before he opened his front door whether an abyss did not yawn behind it, and making sure about it before he went through the door’Footnote 88 but the fact that we can imagine such a doubt does not trigger actual doubts in our mind each time we attempt to open a door. The legal self does not operate with an empty mind open to any conceivable doubt because it is constituted by the very conventions governing the legal field that ‘define and delimit the universe of the thinkable and the unthinkable’.Footnote 89 For instance, it internalizes the practical knowledge that ‘a text cannot be “legal” if it is open to all meanings’Footnote 90 or that the rule of ‘ordinary meaning’ does not mean that one can ‘arbitrarily select dictionary meanings when construing treaty texts’.Footnote 91 It knows what is appropriate and what is not appropriate to argue in view of the history and structure of the discipline. What this means is that the fact that interpretive decisions remain ‘choices’Footnote 92 does not make them arbitrary because the legal operators do not see all the choices as ‘equally choiceful’ as a result of their socialization into such ‘para-legal’ conventions.Footnote 93
5. Conclusion
The main question raised in this article is whether the rules of treaty interpretation have any ‘cash value’ despite the virtually universal scepticism expressed about their constraining power over interpretation. This article has attempted to unpack this scepticism by showing that it is based on a misguided assumption about what we can expect from the rules of treaty interpretation. Because the rules of treaty interpretation are not designed to directly bridge the gap between the signifier and the signified, the standard scepticism with respect to those rules largely misses the point. The function of the rules of treaty interpretation is not to directly determine the meaning of treaty provisions but to impose a ‘common discipline’ as to the legitimate means that can be used in the process of determining that meaning. This explains why the rules of treaty interpretation of the Vienna Convention can be said to apply to ‘any treaty, in any field of public international law … irrespective of the content of the treaty provision being examined and irrespective of the field of international law concerned’.Footnote 94
Does this make enough of a difference to justify the existence of the rules of treaty interpretation, bearing in mind that those rules say nothing else about how the meaning of treaty provisions is determined? As this article has attempted to show, there are two reasons why the response to this question is a resounding ‘yes’. First, even if we were to see the rules of treaty interpretation as establishing a regulatory regime that gives us an enclosure with a hole in it, this does not mean that such an enclosure is as useful as none at all. As pointed out by Baker and Hacker, ‘[i]f there is only one hole in an enclosure, then there is only one way out, and we need to keep an eye on one point only’.Footnote 95 If all the rules of treaty interpretation could do is specify a limited set of means that can be used in the process of meaning-ascertainment then they at the very least exclude the means that have not received the imprimatur of those rules.
Second, the fact that the province of the rules of treaty interpretation is limited does not mean that ‘anything goes’ in treaty interpretation. A practice can never be determined or exhausted by the formalized rules designed to regulate it; it is also defined by a series of informal conventions governing the relevant actors’ attitudes towards it and determining what they consider to be permissible or appropriate things to do when they engage in such a practice. Those conventions limit the freedom of treaty interpreters, making sure that even though in theory, interpretive decisions involve choices, in practice, treaty interpretation is a constrained activity.
