Implied Terms In Treaties

Abstract

Can treaty terms be implied? And, if so, what does that mean? This Article draws on concepts from the branch of linguistics known as pragmatics to analyze how the rules on treaty interpretation allow, in exceptional cases, for the identification of implied terms in otherwise express treaty texts. Its key insight is that implied terms fit within the framework of Articles 31 and 32 of the Vienna Convention on the Law of Treaties and are derived from the associated interpretation of express terms. They cannot be derived from a separate process—and indeed such a separate process is not possible under the positive law.

Table Of Contents

1. I. Introduction 58

2. II. Defining Implied Terms 61

   1. A. Express Term 62

   2. B. TImplied Terms64

      1. 1. Narrowing the Concept 64

      2. 2. Difficulties 67

3. III. Implied Terms Prior to the VCLT 69

   1. A. Case Law 69

   2. B. Codification 75

      1. 1. The Harvard Research in International Law 76

      2. 2. The Institut de Droit International 77

      3. 3. The International Law Commission 79

      4. 4. The Vienna Conference 81

4. IV. Implied Terms in the Scheme of the VCLT 83

   1. A. The Interpretive Framework of the VCLT83

   2. B. Implied Terms Within VCLT Articles 31 and 32 86

      1. 1. Good Faith and Effectiveness 86

      2. 2. Ordinary Meaning, Context, Object, and Purpose 88

      3. 3. Subsequent Conduct and Authentic Interpretation 90

      4. 4. Systemic Integration and Relevant Rules of International Law 91

      5. 5. Supplementary Means 93

5. V. The Modern Practice of Implied Terms96

   1. A. Implied Terms Are Derived from Express Terms 96

   2. B. An Implied Term Cannot Contradict an Express Term 98

   3. C. An Implied Term Must Be Necessary 100

   4. D. An Implied Term Must Take Account of the Nature of the Treaty102

6. VI. Conclusion105

I. Introduction

A cardinal feature of international law, as reconstructed post-1945, is the predominance of treaties as a source of legal obligation for states.¹ This has diminished the role of customary international law,² which underpinned the medieval jus gentium and was one of the pillars of the later law of nations.³ The reasons for this are not difficult to divine. In the first place, the new fields of international cooperation that developed through the United Nations (UN) required an articulation that custom was unable to provide.⁴ In the second, states emerging from decolonization saw custom as deriving from a pre-existing and essentially imperialist system of law in which they had no involvement, save as victims of its unfair and reactionary outcomes.⁵ Treaties provided a partial solution to both these problems.

But that is not to say that the law of treaties was ready to take up this burden. Notwithstanding codification efforts in the early twentieth century,⁶ the field in 1945 was rudimentary by modern standards. In his 1949 survey of international law, setting the agenda for the International Law Commission (ILC), Hersch Lauterpacht noted that “there is hardly a branch of the law of treaties which is free from doubt and, in some cases, from confusion.”⁷ In the list of defects that followed, he observed that “[t]he field of interpretation of treaties continues to be overgrown with the weed of technical rules of construction which can be used—and are frequently used—in support of opposing contentions.”⁸

The picture in 2025 is different—at least as a matter of first impression. The ILC took up Lauterpacht's invitation as it was issued. In 1966, it concluded a set of Draft Articles on the Law of Treaties⁹ (ILC Draft Articles), which the UN General Assembly submitted to a diplomatic conference (Vienna Conference) for conversion into a convention.¹⁰ This led, in 1969, to the conclusion of the Vienna Convention on the Law of Treaties¹¹ (VCLT), which entered into force in 1980. The VCLT marked “the beginning of a new era in the law of treaties.”¹² It was immediately hailed (somewhat breathlessly) as “the cement that holds the world community together.”¹³ Today, it is considered a “bible” for practitioners of international law.¹⁴ One of its “most remarkable achievements” was the reduction of the thicket of rules identified by Lauterpacht to a series of concise statements in Articles 31 and 32,¹⁵ purportedly settling the disputes that bedeviled the pre-VCLT landscape.¹⁶ These are today invoked ritualistically in international and domestic litigation, and regularly confirmed as reflecting customary international law.¹⁷

But, if the VCLT “is a bible, it is a rather short one.”¹⁸ This includes its provisions on treaty interpretation. While admirable for their brevity, Articles 31 and 32 are additionally significant for what they leave out—namely any reference to maxims or canons of construction. In most systems of domestic law, the need for legal documents such as contracts¹⁹ or legislation²⁰ to be interpreted consistently gave rise to maxims designed to guide the interpreter to the “correct” result. Although such principles certainly existed in the law of treaties prior to the VCLT,²¹ Articles 31 and 32 consciously departed from their seriated approach for a generalized method that could be applied holistically to an individual case.²² Although some of the earlier maxims persist in practice, they no longer stand alone. Any attempt to use them must be squared with VCLT Articles 31 and 32.²³

In this Article, I am concerned with another (apparent) lacuna in the interpretive scheme of the VCLT: the identification of implied terms in treaties. Like interpretive canons, implied terms are well-known in domestic law. In the English law of contract, implied terms are identified within agreements to give effect to the unexpressed intention of the parties (terms implied in fact) or because the law otherwise requires their introduction (terms implied in law).²⁴ Similarly, the English law of statutory interpretation allows for implied terms in legislation—particularly when considering the powers of an administrative body to carry out a jurisdiction expressly granted by Parliament.²⁵ In both cases, direction is given to the interpreter as to when and how such terms are to be identified. While confusion may develop as to the content of that direction, it plainly exists and is seen to exist.

We look in vain, however, for similar direction in the law of treaties. Implied terms are mentioned nowhere in VCLT Articles 31 and 32. This omission continues in the major scholarly elaborations of those provisions, which make only brief (if any) reference to the concept,²⁶ occasionally dealing with it under the rubric of “necessary implication.”²⁷ And while more discrete contributions have alighted on the issue, these are limited either ratione materiae ²⁸ or ratione personae.²⁹ At the same time, however, implied terms are identified by those who interpret and apply treaties, though interpreters may not identify them as such. But exactly what interpreters are doing in identifying such terms, and why and how they are doing it, remains in many cases obscure.

Before going further, it is worth setting out some of the concerns that have motivated the writing of this Article. It is beyond doubt that treaty interpretation carries with it a subjective element, personal to the interpreter. This is unavoidable, but undesirable if treaties are to maintain their normative force. This is why the rules on treaty interpretation exist, and are rooted in a textualist, as opposed to a teleological, framework.³⁰ Simply put, an interpretive approach that cleaves close to the treaty's text as the perfection of the parties’ bargain does a better job of realizing that bargain than an alternative that focuses on a generalized concept of its drafters’ intent.³¹

This conclusion is not altered when dealing with treaty terms that are implied as opposed to express. Although it is sometimes necessary to identify implied terms in order to realize the parties’ bargain, any theory of such terms must impose strict limits on the ability of treaty interpreters to do so. The fact that a treaty is silent on a particular matter cannot be considered an invitation for the interpreter to fill it that silence by reference to his or her subjective understanding of what the treaty was intended to achieve.

With the above in mind, I attempt here an analysis of implied terms in treaties, with a view to (1) justifying their recognition as a distinct category of treaty terms, and (2) distilling workable principles, situated in the framework of the VCLT, for their identification. In Part II, I begin with a discussion of the process(es) of implication by reference to the branch of linguistics known as pragmatics, with a view to identifying the parameters of an implied term—as well as explaining and resolving difficulties with the concept as it applies to treaties. In Part III, I undertake a legal analysis of the pre-VCLT practice concerning implied terms in treaties in order to better understand their grounding in the positive law. In Part IV, I discuss implied terms within the framework of the VCLT, with a particular focus on the textually bounded principle of effectiveness that drives the process. In Part V, I identify further principles for the identification of implied terms under the current law. In Part VI, I offer some concluding remarks—as well as a distilled framework for the identification of implied terms in treaties.

II. Defining Implied Terms

Any analysis of implied terms in a legal instrument takes place in the shadow of that instrument's express terms. The law of treaties is no different. But the distinction between the two categories of term is occasionally slippery and rendered inaccessible by the fact that they exist on a spectrum. As a linguistic matter, the information communicated by a treaty term may be in varying degrees express or implied, and frequently in combination. But somewhere along this spectrum, as a matter of law, a term stops being express and starts being implied. A sensible discussion of the latter is only possible once the tipping point from the former is identified.³²

A. Express Terms

VCLT Article 31(1) draws no formal distinction between express and implied terms but refers to “terms” generally. Nevertheless, as a matter of practice, treaty interpreters distinguish between them.³³ Acknowledging the distinction, however, does not get us very far in drawing a workable line between the two. That process is the more difficult, and prima facie neat solutions collapse under scrutiny.

For example, it might be said that an express term is confined to that which the interpreter can derive from the words that actually appear in the treaty—its “semantic” content. That is useful enough as a starting point, but it is not a complete explanation. Article 31(3) of the Statute of the International Court of Justice³⁴ (ICJ Statute) provides that “[if] the Court includes upon the Bench no judge of the nationality of the parties, each of these parties may proceed to choose a judge as provided in paragraph 2 of this Article.” This is logically consistent with the idea that there may be other circumstances in which parties before the International Court of Justice (ICJ) could appoint a judge ad hoc, but no international lawyer would read Article 31(3) that way.³⁵ The term is clearly exclusive: if, and only if, the Court includes upon the Bench no judge of a party's nationality, that parties may proceed to choose a judge in accordance with Article 31(2).³⁶ But the and only if that gives the provision its accepted meaning is not contained within Article 31(3) itself. It must be inferred by the interpreter. If follows that an express term does not necessarily begin and end with its semantic content. The term must also contain information communicated by, but not literally expressed within, the term—its “pragmatic” content.³⁷

From a linguistic perspective, this means that an express term includes not only the term's language, but certain information directly implied by that language—which together constitute an “explicature.” But there are limits to this insight. Not every implication can be considered part of an express term, or the distinction between express and implied terms would collapse. The question therefore becomes what type of implication will properly be considered part of an express term, and what will be considered an implied term.³⁸

There is an intuitive boundary here. Consider a variation on the example already given. The Air Service Agreement arbitration³⁹ considered an aviation treaty between France and the United States. Before the tribunal was the question of whether the treaty granted a U.S. carrier the right to operate a service from the West Coast to Paris with transshipment to a smaller aircraft in London. Section IV(b) of the treaty's Annex provided that “[t]ransshipment when justified by economy of operation will be permitted at all points mentioned in the attached Schedules in territory of the two Contracting Parties.”⁴⁰

Keeping in mind what has already been said in relation to Article 31(3) of the ICJ Statute, we might have expected the tribunal to adopt a similar approach and read Section IV(b) as permitting transshipment in the territory of the United States or France and nowhere else. But it did not. By a majority, it held that the U.S. carrier had a right to unrestrained transshipment in any third state.⁴¹ It did this not by reference to the actual language of Section IV(b)—save to confirm that it was silent on the issue—but the wider text of the agreement, its negotiating context, and aviation practice. Leaving aside the question of whether its conclusion was legally correct,⁴² the majority was doing more than developing the semantic content of Section IV(b) by reference to its closely related pragmatic content. It identified what it thought was an implied term.

Pragmatics provides a ready explanation for this practice through the concept of an “implicature.”⁴³ An implicature is the content of a statement that is derived not from the direct subject of interpretation, but from the entire statement (including what is not said), situated within its context, and read against the reason why that statement was made.⁴⁴ Where the implicature begins, the express term (that is, the explicature) ends.⁴⁵

This much is easily said. But determining when this shift occurs requires close attention to the process by which implied terms are identified. I turn to that now.

B. Implied Terms

If we consider the legal concept of an implied treaty term to be synonymous with an implicature,⁴⁶ then identifying an implied term means acknowledging the work required to derive it from the express terms of the treaty in question. Interpretation to determine the meaning of an express term will focus on the language of the term and the very limited inferences that can be drawn therefrom immediately and almost unconsciously—which together form the explicature/express term. Interpretation to identify an implicature/implied term, however, will focus on a silence within the instrument and interpret it—by reference to the applicable rules on interpretation and such other material as may be admissible—in a considered way to determine how, if at all, that silence is to be filled by reading words into the text.

Two things follow from this. The first is that it is right to consider that interpretation and implication are distinct processes. The second is that this does not mean that the implication derives from something other than interpretation, merely that it is interpretation of a different type that may produce a different interpretive result.⁴⁷ In that sense, the Air Service Agreement majority was correct to treat what it was doing as treaty interpretation, and not something else.⁴⁸

1. Narrowing the Concept

With this concept in hand, we can improve our analysis by identifying those things that are not implied terms, at least in the sense that we define them here.

We can include first within this category the generally applicable elements of the customary law of treaties, as largely reflected in the VCLT. These include, inter alia, the default rules for resolving treaty conflict,⁴⁹ as well as the rules for their amendment,⁵⁰ termination, and suspension.⁵¹ And they also include—perhaps self-evidently—the rules of treaty interpretation themselves.⁵² Secondly, we can add aspects of the customary law of state responsibility, as largely reflected in the ILC Articles on the Responsibility of States for Internationally Wrongful Acts.⁵³ These include the concepts of circumstances precluding wrongfulness,⁵⁴ of remedies,⁵⁵ and of countermeasures.⁵⁶ Finally, we can add discrete principles such as the exception of non-performance—the idea that a party that fails unlawfully to perform a treaty obligation may be met with lawful non-performance in response.⁵⁷

While each of these can be dealt with by the treaty parties directly, as default rules they are not so much as terms implied in treaties by their drafters as terms imposed by international law itself.⁵⁸ The process for their introduction pays little attention to the treaty's express terms—save to confirm they have not been excluded or modified via lex specialis. As such, they proceed from a different theoretical premise, being introduced not to determine the meaning of the treaty text but meet some other priority. While it was fashionable in the past to repackage some of these concepts as implied terms,⁵⁹ that view is today not only unhelpful but heretical: they reflect for the most part freestanding customary rules or general principles of law.

Another province of international law that is often considered proximate to implied terms is the doctrine of inherent powers of international organizations, courts, and tribunals. Where the doctrine is engaged, the body in question is deemed to possess powers not set out in its constitutive treaty. This may include specific procedural functions,⁶⁰ elaborations on substantive jurisdiction⁶¹—and even a legal personality separate from the states parties.⁶² For some commentators, the existence of these powers for a particular body is determined through interpretation of its constitutive treaty in accordance with the principle of effectiveness, by which the interpreter determines the objectives of the body in question and infers the powers that will best promote those objectives within its governing instrument.⁶³

Such notions must also be excluded from our analysis. While identification of inherent powers as a legitimate act of treaty interpretation may have had currency previously,⁶⁴ it is difficult to see how it could be maintained fully today.⁶⁵ As we will see, the operating notion of effectiveness within the doctrine of inherent powers owes much to teleological conceptions of treaty interpretation that were mostly (but not entirely) done away with by the ILC in preference to the mostly textual solution of VCLT Article 31(1).⁶⁶ Under the current law, the open-ended principle of effectiveness (encapsulated in the maxim ut res magis valeat quam pereat) was subsumed within the more restrictive concept of effet utile, being the notion that treaties should be interpreted such that their terms are not rendered meaningless.⁶⁷ While some distinction persists, effectiveness in its modern guise does not “denote that agreements should always be given their maximum possible effect, merely that they be prevented from failing altogether.”⁶⁸ It is difficult to think of a power or attribute of an international organization sufficiently essential to meet this threshold.⁶⁹

The better view, therefore, is that inherent powers—whether arising in an international judicial, technical, or diplomatic body—have little to do with interpretation of the body's constitutive instrument in anything but the broadest sense, and therefore cannot constitute implied terms for the purposes of our analysis. They arise not from the treaty's text but from a different source—namely wider assumptions about what the states parties intended when they created the body in question.⁷⁰ Where considering a judicial body, for example, its creators undoubtedly expected it to behave as a court, and therefore to possess “an inherent jurisdiction, the power to exercise which is a necessary condition of . . . any court of law . . . being able to function at all.”⁷¹ In such a case, the inherent power may be identified not as an exercise in treaty interpretation, but as a corollary of the judicial function—with the question becoming whether anything in the body's constitutive instrument would seem to exclude the existence of the power.⁷² The process of their identification is therefore different to that for implied terms, reflecting a more open-textured approach to the body and the assumptions of its creators.

This same logic is applicable, mutatis mutandis, to international organizations generally, although the greater variation of such bodies may render identification of inherent powers more challenging. In this, however, international organizations are assisted by the presumption that their own statements as to their jurisdiction, when accompanied by an assertion of propriety, are legitimate.⁷³

2. Difficulties

Before moving on, it is necessary to address and resolve two apparent difficulties with the notion of implied terms in treaties: failure by treaty interpreters to employ the concept of implied terms, and deliberate ambiguity in treaty terms.

The first problem so identified arises from diffuse practice by treaty interpreters. In many cases, implied terms are not identified expressly: even if what the interpreter is doing as a matter of linguistics is identifying an implied term, the concept is subsumed within the general rubric of treaty interpretation and given no separate life of its own.

The Air Service Agreement case is emblematic of this. There, neither the award nor the dissenting opinion referred to the idea of an implied term, although that was functionally the issue in view. In contrast, however, one may point to the practice of the Iran–U.S. Claims Tribunal, which expressly identified in Case A15(IV) an implied term regarding the obligation of the United States to terminate proceedings concerning in Iran in its own courts following conclusion of the General Declaration.⁷⁴ And one can equally point to the decision of the tribunal in SAUR International v. Argentine Republic, which identified a “tacit condition” in the relevant bilateral investment treaty (BIT) that an investment would only be protected if made in accordance with the laws and regulations of the host state.⁷⁵

This division arises in large part from differences between the common and civil law traditions. Implied terms (and especially contractual implied terms in fact) are a common law creature, with no precise analogue in civil law systems. Indeed, a civil lawyer would be dismissive of the idea: under that tradition, implied terms are best understood as one expression of the common intention of the parties derived from evidence beyond the contract itself. This common intention, which exists outside the contract, is the true subject of the interpretive exercise,⁷⁶ the civil lawyer says, and so any distinction between express and implied terms misses the point.

Were the teleological methods of treaty interpretation that had validity prior to the VCLT still in effect today, this point would have merit—the civil law approach to interpretation being essentially teleological in character. But they are not, and the qualified triumph of textualism that the VCLT's interpretive regime represents means that this way of thinking does not excuse treaty interpreters from saying when they identify an implied term.⁷⁷

A second difficulty is the nature of a treaty itself, famously described as “a disagreement reduced to writing.”⁷⁸ While some points of disagreement are capable of resolution through negotiation, others are not. When this happens, drafters are often minded to mask the disagreement with opaque language and make it a problem for some unknown future interpreter.⁷⁹ Deliberate or “creative” ambiguity of this kind is known in the drafting of other legal texts, but rarely is it considered an acceptable, even useful, part of the drafter's repertoire.⁸⁰ Its use may pose an existential quandary for any notion of implied terms in treaties. On the one hand, ambiguity in a treaty's text, if unconscious, may widen the scope for the identification of implied terms by an interpreter. On the other, if the ambiguity was intended to communicate a lack of common intent, its resolution via an implied term may not only compromise the treaty parties’ project but disturb relations between them by solving what was meant to remain unsolved.⁸¹

We may brush such concerns to one side. Treaty drafting carries with it the understanding, if not the expectation, that the treaty will have to be interpreted sooner or later. As Philip Allott reminds us, that interpretation will invariably be “an act of violence,”⁸² constructed by reference to the objective intention of the drafters—being the interpreter's understanding of the treaty derived from the empirical reality of its text together with such other material as is considered admissible. The subjective intention of the drafters—i.e., what they actually meant when they committed the ambiguity to paper—is irrelevant. Put another way, deliberate ambiguity is only a temporary solution to disagreement, and a later interpreter may be duty-bound to resolve it⁸³—including through identification of an implied term if necessary.⁸⁴

III. Implied Terms Prior to the VCLT

With implied terms properly defined, the positive law concerning those terms may be considered. I start with setting out the landscape prior to the conclusion of the VCLT in two parts. First, I examine judicial and arbitral practice on implied terms prior to and during the work of the ILC. This examination focuses in large part on the work of international courts up to 1969, most notably the Permanent Court of International Justice (PCIJ), and the early decisions of its successor, the ICJ. Secondly, I look at the various codification efforts of the same period, up to and including the work of the Vienna Conference.

A. Case Law

The idea that a treaty's content might stretch beyond what the parties committed to paper is not new. Traces of the thesis appear in the foundational texts of modern international law. Grotius opined that “[t]he best Rule of Interpretation is to guess at the Will [of the parties] by the most probable Signs, which signs are of two Sorts, Words and Conjectures; which are sometimes considered separately, [and] sometimes together.”⁸⁵ Vattel similarly considered that focusing on the parties’ shared intent “not only serve[d] to explain the obscure or ambiguous expressions” in the treaty, but “also to extend or restrict its several provisions independently of the expressions, and in conformity to the intention and views of the legislature or the contracting parties, rather than to their words.”⁸⁶

The world in which these writers were operating, however, was one of law without courts—reflecting an international system in which the principal interpreters of treaties were the states that drafted them. Treaties themselves were nearly always bilateral and confined to technical subjects.⁸⁷ While much ink was spilled framing the rules of the game in terms of states’ rights and obligations, little attention was paid to how those rights and obligations were to be understood—or interpreted—by a third party.⁸⁸ Coherent practice on the law of treaties was rare, let alone practice on treaty interpretation, less still practice on implied terms.

Treaty interpretation in this period centered on the canons of construction, being common and civil law maxims that had developed in the context of contractual or statutory interpretation:⁸⁹ effet utile, contra proferentem, ejusdem generis, in dubio mitius, and so forth.⁹⁰ These were not applied in any systematic fashion but deployed ad hoc depending on the identity of the interpreter and the problem posed. This resulted in an interpretive regime in which participants could disagree on nearly all the major premises, including the validity and hierarchy of the canons themselves. The resulting chaos was summed up by James Yü, who declared that “text writers have accumulated a mass of fanciful rules which, on account of their sweeping, mechanical nature, have caused the task of interpretation to become dangerous.”⁹¹

This paradigm began to shift from 1920 onward—not so much because of a systematization of the rules on treaty interpretation, but through the creation of a foundation on which systematization could occur. The principal mover of this was the emergence of third-party adjudication as a recognized form of international dispute settlement, resulting in a nascent jurisprudence on treaty interpretation.⁹² The most significant development was the creation of the PCIJ. This was the first permanent international judicial body,⁹³ independent of the parties before it, that could develop a distinct case law on treaties and their interpretation,⁹⁴ albeit one that was occasionally inconsistent.⁹⁵

So far as implied terms were concerned, the case law of this period does not give us much assistance in distilling a workable set of principles, although the possibility of such terms was acknowledged.⁹⁶ But we can perhaps identify a few features to act as the foundations of a systematic enquiry. First, and as a general matter, interpreters tended to view their function—to a greater or lesser extent—as interpreting the text of the treaty itself. But while the text of the treaty was to prevail where clear, in cases where it was not, the intentions of its drafters, as determined by means other than the text, was determinative, as made clear by, inter alia, the France–Mexico Mixed Claims Commission in the Georges Pinson case.⁹⁷ Secondly, in searching for evidence of that intent beyond the express words of the treaty, the primary source was the travaux préparatoires. As the Turkey–Greece Mixed Arbitral Tribunal noted in Polyxène Plessa, this could be used “to interpret or supplement the text,” with a further distinction being drawn between supplementation and modification—the former permissible, the latter not.⁹⁸

The line between supplementation and modification was demonstrated by the PCIJ in the Acquisition of Polish Nationality advisory opinion. The case concerned the Polish Minority Treaty,⁹⁹ Article 4 of which provided that:

Poland admits and declares to be Polish nationals ipso facto and without the requirement of any formality persons of German, Austrian, Hungarian or Russian nationality who were born in the said territory of parents habitually resident there, even if at the date of the coming into force of the present treaty they are not themselves habitually resident there.¹⁰⁰

Poland argued that, notwithstanding this provision, it was entitled to treat as German any individual whose parents were not habitually resident in Polish territory both on the date of birth of the person concerned and on the date of the entry into force of the Treaty—arguing, in effect, for an implied requirement of continuous parental residence over the relevant period. The PCIJ was quick to disagree, finding that “[s]uch an assertion is in contradiction with the terms of the provision which it claims to interpret and is not supported by the precedents supplied by international practice.”¹⁰¹ In particular, the Court observed that Article 4 expressly linked the question of an individual's nationality to where their parents habitually lived when the individual was born, and that “to require furthermore the continuance or the reestablishment of this habitual residence at the time of the coming into force of the Treaty would amount to an addition to the text and would go beyond its terms.”¹⁰² In so finding, it took particular note of the fact that the proposed exclusion: was not included in Article 4, which already contained express exclusions; was not a feature of other treaties of annexation; did not account for the obvious situation in which someone's parents died before the Treaty entered into force; and would not eliminate the mischief identified by Poland in justification, i.e., dual nationality.¹⁰³ The Court concluded:

The Court's task is clearly defined. Having before it a clause which leaves little to be desired in the nature of clearness [i.e., clarity], it is bound to apply this clause as it stands, without considering whether other provisions might with advantage have been added to or substituted for it. . . . To impose an additional condition for the acquisition of Polish nationality, a condition not provided for in the [Polish Minority Treaty] would be equivalent; not to interpreting the Treaty, but to reconstructing it.¹⁰⁴

These tendencies also appear in the early jurisprudence of the ICJ, notably in the Asylum ¹⁰⁵ and Haya de la Torre ¹⁰⁶ cases concerning Article 2 of the Havana Convention on Asylum.¹⁰⁷ While these cases are useful, they give an impression of monolithic opposition to implied terms, an idea that the Court seemed to give theoretical credence only. But a better idea of where legitimate identification of an implied term ended and simply rewriting the treaty began is provided by the two Peace Treaties advisory opinions. These concerned the identically worded dispute settlement provisions of the 1919 treaties between the Allies and Bulgaria, Hungary, and Romania. The first Peace Treaties opinion considered the treaties’ provision for the creation of a Commission in the event a dispute could not be resolved by negotiation. In such an event, the treaties provided that a dispute could “be referred at the request of either party to the dispute to a Commission composed of one representative of each party and a third member selected by mutual agreement of the two parties from nationals of a third country.”¹⁰⁸ It further provided that “[s]hould the two parties fail to agree within a period of one month upon the appointment of the third member, the Secretary-General of the United Nations may be requested by either party to make the appointment.”¹⁰⁹

With relatively little intermediate reasoning, the ICJ held that the phrase “at the request of either party” entailed an obligation by the other party to “co-operate in constituting the Commission, in particular by appointing its representative.”¹¹⁰ The logic behind this position was the need to render the treaty effective—if there was no obligation of cooperation, “the method of settlement by Commissions provided for in the Treaties would completely fail in its purpose.”¹¹¹ Put another way, the additional language was necessary for the treaties to function as intended.¹¹²

In the second Peace Treaties opinion, the Court considered an elaboration on the question earlier posed: what was to happen if, in breach of the implied obligation of cooperation identified in the first Peace Treaties opinion, a party refused to appoint its representative? In such a case, could the secretary-general appoint the missing representative, as he was empowered to do with respect to the third (neutral) member of the Commission?

The Court answered this question in the negative, holding in effect that such an interpretation would stretch the clause beyond its breaking point. The provision was clear in its stipulation that the secretary-general had the power to appoint the third member of the Commission only, and the treaties did not consider this applicable in “the much more serious case of a complete refusal of co-operation by [the other party], taking the form of refusing to appoint its own Commissioner.”¹¹³ The Court further explained why its reasoning in the first Peace Treaties opinion could not be transposed to fill the gap:

The breach of a treaty obligation cannot be remedied by creating a Commission which is not the kind of Commission contemplated by the Treaties. It is the duty of the Court to interpret the Treaties, not to revise them. . . . The principle of interpretation expressed in the maxim: Ut res magis valeat quam pereat, often referred to as the rule of effectiveness, cannot justify the Court in attributing to the provisions for the settlement of disputes in the Peace Treaties a meaning which, as stated above, would be contrary to their letter and spirit.¹¹⁴

The line drawn by these two cases is a useful one. The implied term identified in the first Peace Treaties opinion filled a genuine silence in the treaties with a relatively discrete obligation that did not contradict the treaties’ express terms. The situation in the second Peace Treaties opinion was different. The treaties specified that the other party was responsible for appointing its representative, while at the same time stipulating that the secretary-general had responsibility for appointing the third member of the Commission. The Court also had the benefit of relevant treaty practice showing that parties were perfectly capable of introducing express procedures to deal with such situations.

From the point of view of pragmatics, the explicature of all this was clear: per the provision, the dispute would be referred to a Commission composed of one representative of each party to be appointed only by that party and a third member to be selected by mutual agreement or the secretary-general if agreement was not forthcoming. An implicature that gave the secretary-general the power to appoint the representative of a recalcitrant party would cut across this scheme—and, by extension, contradict party intent. But the Court's conclusion was not unanimous. A detailed dissent was appended by Judge Read. In his view, effectiveness compelled the contrary result, on the basis that the Court's preferred answer “would destroy the Disputes Article as an effective guarantee of the substantive provisions of the Treaty [and] would render largely nugatory the undertakings given to secure the enjoyment of human rights and fundamental freedoms.”¹¹⁵

The distinction between these two approaches gives an indication as to where the law of treaties stood on implied terms—and other interpretive issues—around this time. On the one hand, there were the proponents of the textual approach, who considered the text of the treaty as the perfection of its drafters’ vision, and thus to be taken as the ultimate expression of their intent. Under this approach, “[t]he intentions or presumed intentions of the framers cannot be invoked to fill in gaps, or import into the treaty something which is not there, or to correct or alter words or phrases the meaning of which is apparently plain, or to give them a sense different from that which they possess according to their normal and natural meaning.”¹¹⁶ This school was suspicious of, if not openly hostile toward, the concept of implied terms in treaties—and can be seen in, inter alia, the Polish Nationality case.

On the other, there were the proponents of the teleological approach, a school of thought that arose in response to the textualists and, more particularly, the canons of interpretation. A largely scholarly endeavor with some prominent judicial proponents, the teleologists reserved the right to expand or supplement a treaty's terms by reference to its object and purpose and a broad conception of effectiveness. Under this methodology, words could readily be added to otherwise express texts by an interpreter, provided they advanced the treaty's wider mission and the intentions of its drafters.¹¹⁷ The canons, meanwhile, were discarded as useless at best and misleading at worst.¹¹⁸

Whilst there were differences between these schools of thought, there were also similarities. The textualists did not preclude implied terms outright—as the first Peace Treaties opinion shows. And the teleologists still considered a treaty's text to be valid evidence of its drafters’ intent. The sticking point—as Judge Read's dissent in the second Peace Treaties opinion demonstrates vividly—was the role of effectiveness in the interpretive exercise. Used responsibly, and with an eye to the primacy of the text, it was a useful device to give a treaty its fullest value and effect consistent with its wording. Used irresponsibly, was judicial legislation,¹¹⁹ a fact the most extreme teleologists¹²⁰ cheerfully admitted, justifying their views by reference to the nature of the treaties they were charged with interpreting, and in general confining their more outlandish pronouncements to multilateral treaties serving a social or humanitarian function.¹²¹ In so doing, they made clear that what they were doing was not interpretation at all, but a post-interpretive exercise that used effectiveness to fill perceived gaps in the treaty.¹²²

A further sticking point was the role to be played by travaux préparatoires. As already seen, early teleologists encouraged its use as a means to determine the true sense of the text, or the intention of the parties,¹²³ a moderate view that persisted in certain influential quarters.¹²⁴ But in the age of the PCIJ and ICJ, the practice waned—being first confined by the textualists to situations in which the text interpreted was ex facie unclear,¹²⁵ and then excluded entirely by some extreme teleologists as shackling the treaty impermissibly to the will of its drafters.¹²⁶

What was absent from this interpretive contest, however, was any detailed consideration of what an interpreter was really doing when reading words into a treaty. While there was a line drawn by all sides of the argument as to when such an exercise was permissible (admittedly at different points) no court or tribunal undertook to explain the relationship between express and implied terms or how the latter related to or derived from the former. Everything took place under the umbrella of the textual/teleological debate, forestalling more subtle enquiries. To that end, the limited discussion of the issue in the textbooks and commentaries of the era usually (but not always¹²⁷) treated it within the wider discourse concerning the extent to which the “plain meaning” of a treaty text could be supplemented or usurped by external material.¹²⁸ But the seeds of a doctrine of implied terms were there and could be seen to sprout.

B. Codification

These varying debates rendered the question of treaty interpretation ripe for codification.¹²⁹ Movements dedicated to this enterprise were well-known in the first part of the twentieth century.¹³⁰ The Institut de Droit International, a learned society dedicated in part to the “gradual and progressive codification” of international law,¹³¹ was founded in 1873 against the backdrop of the great civil law codification movements of the eighteenth and nineteenth centuries and toiled to develop codifying resolutions on discrete topics. In the 1920s and 1930s, however, the utopian thinking that followed the end of World War I and the founding of the League of Nations, created enormous scholarly enthusiasm—particularly in the United States¹³²—for the idea that the idea that a broad-based codification of international law could yield a firm and enduring peace.¹³³

Between 1930 and 1970, three significant attempts at codifying the rules on treaty interpretation were undertaken, each with corresponding implications for the doctrine of implied terms. The first two—by the Harvard Research in International Law (Harvard Research) and the Institut—were, at best, preliminary to the third and ultimately successful effort by the ILC, which via the Vienna Conference became the VCLT.

1. The Harvard Research in International Law

The Harvard Research arose out of the codification efforts of the League of Nations. These culminated in the notorious failure that was the 1930 Hague Conference for the Codification of International Law, which attempted to conclude three substantial conventions—on nationality, territorial waters, and state responsibility—in under a month, and promptly imploded under pressure of time.¹³⁴ Before it could fail, however, a group of scholars at Harvard Law School, led by Manley O. Hudson, undertook to organize a private research project on the relevant topics in parallel.¹³⁵ This was done with a view to shaping the outcome of the Conference, but continued and expanded after the Conference collapsed. Between 1927 and 1939 it produced thirteen draft conventions with accompanying commentaries. Many of these contributed substantially to the development of international law in the middle part of the twentieth century.¹³⁶

One of the Harvard Research's more influential projects was its Draft Convention on the Law of Treaties (Harvard Draft Convention), which appeared with a commentary in 1935.¹³⁷ Its view on interpretation was set down in two paragraphs in Article 19: one on interpretation generally, and another on issues concerning plurilingual treaties. So far as the first was concerned, Article 19(a) provided that “[a] treaty is to be interpreted in light of the general purpose which it is intended to serve,”¹³⁸ further noting various extraneous elements to be considered in connection thereto, including: the historical background of the treaty; its travaux préparatoires; the circumstances at the time of its conclusion; the change in these circumstances sought to be effected; the subsequent conduct of the parties in applying its provisions; and the conditions prevailing at the time of the interpretation itself.

As the thirty-four pages of commentary attached to Article 19(a) made clear,¹³⁹ it reflected the contemporaneous chaos surrounding treaty interpretation. The solution of the Harvard Research was to adopt the teleological approach. In this sense, Article 19(a)'s most significant advances were to: first, junk the accumulated canons of interpretation as a useful tool of construction on the basis that due to the “fictitious ring of unassailable truth” that accompanied them, they could cause an interpreter to approach a treaty with a closed mind, preventing consideration of an individual case;¹⁴⁰ and, secondly, give general license to an interpreter to examine a treaty's travaux préparatoires “as they are an inherent part of the ‘whole picture’ of a treaty.”¹⁴¹ At the same time, the Harvard Research did not place an overt premium on the treaty's actual text, noting that the words used by the drafters “have significance only as they may be taken as expressions of the purpose or design of the parties which employed them.”¹⁴² The result was a loose “guide post” to the proper process of interpretation that broadly reflected the views of the moderate teleologists,¹⁴³ as summarized in John Westlake's injunction that “[t]he important point is to get at the real intention of the parties” and that, to that end, “a large and liberal spirit of interpretation will reasonably correspond.”¹⁴⁴

This conceptualization of the interpretive function meant that the Harvard Research did not give much attention to the concept of implied terms. While acknowledging that “[w]hen interpreting a treaty . . . the interpreter must not alter it or substitute a new text,”¹⁴⁵ it skirted the issue of when it was permissible to read words into a treaty—even as it examined the Polish Nationality advisory opinion that was the leading PCIJ elaboration on the topic.¹⁴⁶ This was consistent with the teleological understanding of treaty interpretation in which text was merely one of several mechanisms for communicating the drafters’ intent, said intent (as in civil law systems) being the true subject of interpretation. If the literal meaning of the words used contradicted that intent as determined from other sources, “the general practice of tribunals has been not to accept a construction, subversive of or tending to thwart, the manifest design of the contracting parties.”¹⁴⁷ The result was a doctrine of treaty interpretation that encouraged silences to be filled via effectiveness where necessary to advance the treaty's wider objectives as ascertained by the interpreter.

2. The Institut de Droit International

The next codification body to tackle treaty interpretation was the Institut, which commenced work on the topic at its 1948 Brussels session. Lauterpacht was appointed rapporteur.¹⁴⁸ The Institut's progress across subsequent sessions—notably the 1950 Bath session, the 1952 Siena session, and the 1956 Grenada session—reflects the general direction of travel at this time, as textualism became the dominant approach to treaty interpretation.

Lauterpacht's initial approach to the question was avowedly teleological.¹⁴⁹ It considered the treaty's text as merely the starting point of the inquiry,¹⁵⁰ placed great weight on the travaux préparatoires,¹⁵¹ and included a broad conception of effectiveness as a central pillar.¹⁵² Through the efforts of figures such as Eric Beckett¹⁵³ and Gerald Fitzmaurice,¹⁵⁴ it was gradually supplanted within the Institut by an approach confirming that “[s]ince the agreement of the parties has come about on the text of the treaty, the natural meaning of this text should be the basis of the process of interpretation.”¹⁵⁵ But this was not a decisive victory for the textualists, as the Institut remained divided on the relevance of travaux préparatoires and wider means of interpretation.¹⁵⁶

The solution proposed was for the Institut's work on the question to be confined to a general resolution that made no definitive reference to the internal sticking points.¹⁵⁷ This appeared at the Institut's 1956 Grenada session. Lauterpacht having been elected to the ICJ, Fitzmaurice took his place as rapporteur and led the Institut through a revision of his predecessor's work, resulting in a slimmed-down text.¹⁵⁸ The final product focused on the text of the treaty as the principal object of interpretation, eschewing any reference to the canons of interpretation. Article 1(1) provided that “[t]he agreement of the parties having expressed itself in the text of the treaty, the natural and ordinary meaning of the terms of this text should be taken as the basis of interpretation.”¹⁵⁹ It continued to note that “[t]he terms of the provisions of the treaty shall be interpreted in their entire context, in good faith and in light of the principles of international law.”¹⁶⁰ Article 2 maintained a potential role for, inter alia, the treaty's preparatory work and wider mission—but, perhaps strangely, only when the interpreter was an international court or tribunal.¹⁶¹ It was the first recognizable forerunner of VCLT Articles 31 and 32.¹⁶²

While again not a complete victory for textualism, the Institut's resolution was the beginning of the end for overtly teleological methods of treaty interpretation. Although considerations beyond the text of the treaty would play a role in the interpretive process moving forward, they no longer had a credible claim to being the focus of the exercise. This is significant for our attempt to carve out a role for implied terms within treaties—for it is only by prioritizing a treaty's express terms, as reflected in its text, that we can identify (and appreciate the distinct role of) implied terms. As already noted, such a distinction is unnecessary if the object of interpretation is a conception of the drafters’ intent that sits outside the treaty, an approach that at its most extreme considers the concept of treaty's terms to be mildly passé.

3. The International Law Commission

The ILC commenced work on the law of treaties in 1949. Four special rapporteurs, all British—James Brierly (1950–1952), Lauterpacht (1952–1955), Fitzmaurice (1955–1961), and Sir Humphrey Waldock (1961–1966)—carried the project through to the adoption of the ILC Draft Articles in 1966. But treaty interpretation was left unaddressed as a topic until the end of the project, appearing only in Waldock's Third Report¹⁶³ of 1964.¹⁶⁴ Waldock was inspired by the approach taken by the Institut in 1956; in particular, its prioritization of the treaty's text as the principal object of interpretation. He noted that “[t]he basic rule of treaty interpretation [is] the primacy of the text as evidence of the intention of parties,” such that “the text must be presumed to be the authentic expression of the intentions of the parties; and that, in consequence, the starting point and purpose of interpretation is to elucidate the meaning of the text.”¹⁶⁵

Waldock's primary rule of interpretation, contained in his Draft Article 70, was that “[t]he terms of the treaty shall be interpreted in good faith in accordance with the natural and ordinary meaning to be given to each term—(a) in its context in the treaty and in the context of the treaty as a whole; and (b) in the context of the rules of international law in force at the time of the conclusion of the treaty.”¹⁶⁶

That said, Waldock did not set his face entirely against teleological interpretation—provided it did not supplant the text as the perfection of the parties’ bargain. His interpretive scheme reserved, in Draft Article 71, a role for “other evidence or indications of the intentions of the parties” (including travaux préparatoires) when (1) confirming the natural and ordinary meaning of a treaty term, (2) determining it where that meaning was obscure, absurd or unreasonable, or (3) establishing that a special meaning was intended by the parties.¹⁶⁷ It also made express allowance, in Draft Article 72, for effectiveness—subject to the limitation that its use had to be consistent with the plain meaning of the text, and the treaty's object and purpose.¹⁶⁸ To an extent, this was somewhat redundant, as Waldock conceived of effectiveness as being a corollary of good faith interpretation, and hence part of his primary rule. But he justified its separate inclusion by reference to implied terms:

If the principle of effective interpretation may be said to be implicit in the requirement of good faith, there are, it is thought, two reasons which may make it desirable to formulate it in a separate article. The first is that the principle has special significance as the basis upon which it is justifiable to imply terms in a treaty for the purpose of giving efficacy to an intention necessarily to be inferred from the express provisions of the treaty. The second is that in this sphere—the sphere of implied terms—there is a particular need to indicate the proper limits of the application of the principle if too wide a door is not to be opened to purely teleological interpretations.¹⁶⁹

Draft Article 72 met with a negative reception in the Commission and was deleted. The reasons why help us understand the proper scope of implied terms within the VCLT regime. Two concerns are apparent in the minutes of the Commission. The first was, as Waldock himself averred, that references to effectiveness were superfluous in light of the fact that good faith was mentioned as a tenet of interpretation in Draft Article 70.¹⁷⁰ But, more importantly, the Commission's members expressed reservations to the expansive concept of effectiveness favored by the teleologists, arguing that the law would be better served by a notion of effectiveness that saved treaty provisions from redundancy, rather than maximizing their effect.¹⁷¹ For that reason, Draft Article 72 was not only dropped, but its implicit survival in Draft Article 70 was in a form analogous to the maxim of effet utile, and not the teleologists’ understanding of ut res magis valeat quam pereat. States subsequently accepted this development with equanimity.¹⁷²

The provisions of Waldock's scheme were then reorganized into a Draft Article 69 on the general rule of interpretation and a Draft Article 70 on further means of interpretation, which included travaux préparatoires.¹⁷³ Following further reworking, they appeared in the ILC Draft Articles as Articles 27 and 28,¹⁷⁴ which do not differ materially (for present purposes) from VCLT Articles 31 and 32.

The commentary to Articles 27 and 28 of the ILC Draft Articles affirms the general trend described. The ILC confirmed that the correct approach to treaty interpretation was principally textual—emphasizing “the primacy of the text as the basis for the interpretation of a treaty, while at the same time giving a certain place to extrinsic evidence [e.g., travaux préparatoires] of the intentions of the parties and to the objects and purposes of the treaty as means of interpretation.”¹⁷⁵ It sidelined the canons of construction, noting that any attempt to codify them “would clearly be inadvisable.”¹⁷⁶ And it further indicated that it considered the possibility of implied terms to be part of the general rule of interpretation, to be deployed only rarely and by reference to a narrow conception of effectiveness. It said:

Properly limited and applied, [effectiveness] does not call for an “extensive” or “liberal” interpretation in the sense of an interpretation going beyond what is expressed or necessarily to be implied in the terms of the treaty. Accordingly, it did not seem to the Commission that there was any need to include a separate provision on this point. . . . The Court, which has by no means adopted a narrow view of the extent to which it is proper to imply terms in treaties, has nevertheless insisted that there are definite limits to the use which may be made of the principle ut res magis valeat for this purpose.¹⁷⁷

We may cavil with the ILC's description. The statement that the ICJ had “by no means adopted a narrow view” of implied terms in treaties is correct in the sense that the Court did not exclude them entirely. But that is not the same as saying it had adopted a broad view of the topic. As we have seen, the “definite limits” set out in the jurisprudence of the Court and its predecessor were strict: the text of the treaty was the perfection of the parties’ bargain and the best evidence of their intent, to be elaborated on cautiously and where the addition was plainly necessary.

4. The Vienna Conference

Following the passage of the ILC Draft Articles, Articles 27 and 28 were submitted to the Vienna Conference, where they were debated during the First Session in 1968.¹⁷⁸ There was really only one topic that exercised the Conference—which came as a consequence of the first intervention by U.S. delegate, and founder of the New Haven School of international law,¹⁷⁹ Myres McDougal. A persistent critic of the ILC's work on treaty interpretation, McDougal considered the principal “defect and tragedy” of Articles 27 and 28 to be “their insistent emphasis upon an impossible, conformity-imposing textuality.”¹⁸⁰ His preferred approach was one in which “decision-makers undertake a disciplined, responsible effort to ascertain the genuine shared expectations of the particular parties to an agreement” supplemented by “the policies of the larger community which embraces both the parties and decision-maker.”¹⁸¹ Included with this, moreover, was a notion that the interpreter would also undertake a “policing” function—refusing to give effect to the parties’ expectations where they would conflict with “world public order.”¹⁸²

As a matter of substance, Fitzmaurice accurately described McDougal's approach as reflecting a “powerful plea for a completely ‘open-ended’ technique of interpretation” that was similar in some respects to that put forward by Lauterpacht in the Institut.¹⁸³ But whereas Lauterpacht proceeded by sweeping away the various interpretive canons, McDougal embraced them, giving oxygen to every technique of interpretation devised over the previous century and special weight to those which would allow the greatest violence to the treaty's text.¹⁸⁴ This was most obvious in the case of his policing function, which would allow the interpreter, via his or her subjective understanding of world public order (an amorphous concept considerably broader than jus cogens) to rewrite or neuter the parties’ bargain¹⁸⁵ in a manner inconsistent with the international judicial function¹⁸⁶ but in keeping with the wider rule-skepticism of the New Haven School.¹⁸⁷ Such an approach went beyond the most extreme examples of teleology, subordinating party autonomy to the caprice of the interpreter and “open[ing] the door to anarchy and abuse.”¹⁸⁸

At the Vienna Conference, McDougal continued his attack on the ILC Draft Articles, arguing in the Committee of the Whole that Articles 27 and 28 were based on an “obscurantist tautology” that the text could be interpreted without reference to extrinsic material.¹⁸⁹ To that end, on behalf of the U.S., he suggested that Articles 27 and 28 be merged to provide that “[a] treaty shall be interpreted in good faith in order to determine the meaning to be given to its terms,” but requiring that said interpretation was to be carried out in light of a non-exhaustive laundry list of “relevant factors” that did not expressly include the plain meaning of its text.¹⁹⁰ In short, it was an invitation to undo the previous fifteen years of work in the Institut and ILC, and to return to or possibly exceed Article 19(a) of the Harvard Draft Convention. It was met with the polite rejoinder that McDougal had misunderstood the ILC Draft Articles, which were not ex facie hostile to treaty interpretation by reference to extrinsic material, provided this took place within sensible limits that gave proper weight to the treaty's text.¹⁹¹ The matter was closed by a resounding rejection of the U.S. approach.¹⁹² VCLT Articles 31 and 32 were thereafter confirmed in their familiar terms.

This final episode confirms what came before, which is that a predominantly, but not exclusively, textual view of treaty interpretation is contained in the VCLT. More expansive conceptions of treaty interpretation—as reflected in, inter alia, the Harvard Draft Convention—were definitively rejected by the Institut, the ILC, and the Vienna Conference. So far as implied terms are concerned, the doctrine that this understanding produces cleaves closely to the text and that which can be drawn from it by inference with the assistance of admissible wider material and a narrow conception of effectiveness. All other approaches—and, in particularly, the teleological understanding that treaty texts could be modified on the basis of the interpreter's understanding of its drafters’ intent as derived from extra-treaty materials—are to be rejected. A four-decade battle had concluded. The textualists emerged the winners.

IV. Implied Terms in the Scheme of the VCLT

This brings us to the concept of implied terms within the present law of treaties. We can approach this in two parts. First, we can parse the customary international law rules on treaty interpretation, as set out in VCLT Articles 31 and 32, to determine how their various elements play a role in the identification of implied terms. Secondly, we can posit particular rules for the identification of implied terms that are drawn from, but not necessary neatly expressed within, these same elements. In this Part, we consider the first of these discussions. The second follows in Part V.

A. The Interpretive Framework of the VCLT

Absent a lex specialis, the rules contained in VCLT Articles 31 and 32 are applicable whenever a treaty fails to be interpreted. As noted, both provisions are generally accepted as reflecting customary international law,¹⁹³ together with the legal fiction that the principles they espouse are equally applicable to treaties entered into before the conclusion of the VCLT,¹⁹⁴ despite elements of Articles 31 and 32 doubtlessly reflecting the progressive development of international law.¹⁹⁵

VCLT Article 31, the “[g]eneral rule of interpretation,” provides as follows:

1. 1. A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.

2. 2. The context for the purpose of the interpretation of a treaty shall comprise, in addition to the text, including its preamble and annexes:

   1. (a) any agreement relating to the treaty which was made between all the parties in connection with the conclusion of the treaty;

   2. (b) any instrument which was made by one or more parties in connection with the conclusion of the treaty and accepted by the other parties as an instrument related to the treaty.

3. 3. There shall be taken into account, together with the context:

   1. (a) any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions;

   2. (b) any subsequent practice in the application of the treaty which establishes the agreement of the parties regarding its interpretation;

   3. (c) any relevant rules of international law applicable in the relations between the parties.

4. 4. A special meaning shall be given to a term if it is established that the parties so intended.¹⁹⁶

VCLT Article 32, dealing with “[s]upplementary means of interpretation,” provides:

Recourse may be had to supplementary means of interpretation, including the preparatory work of the treaty and the circumstances of its conclusion, in order to confirm the meaning resulting from the application of article 31, or to determine the meaning when the interpretation according to article 31:

1. (a) leaves the meaning ambiguous or obscure; or

2. (b) leads to a result which is manifestly absurd or unreasonable.¹⁹⁷

Before considering these provisions, and how they support the concept of implied terms, a few points about their operation are warranted. So far as VCLT Article 31 is concerned, the fact that the provision refers to a unitary “rule” of interpretation is deliberate—the whole of the rule, and not merely Article 31(1), must be consulted in each case as part of a “single combined operation,” with each element “thrown into the crucible” to furnish the proper interpretation.¹⁹⁸ That notwithstanding, there is little doubt that despite the lack of formal priority between the individual paragraphs of Article 31, the cornerstone of the provision is Article 31(1), as supplemented by Article 31(2). It is not difficult to see why. Article 31(1) is the only provision that refers expressly to the treaty's terms, on which the interpretive exercise rests as the objective manifestation of party intent.¹⁹⁹ And it is also the only element of Article 31 which, it can safely be said, will be relevant in all cases.

That being said, the structure of Article 31 suggests that while the terms of the treaty are intended as the starting point of the interpretive exercise under Article 31(1), Articles 31(3) and (4) furnish the interpreter with additional tools by which the ordinary meaning of those terms, read in good faith and in context, may be supported or modified by suite of wider and iteratively applied factors.²⁰⁰ While Article 31(1) may comprise the beginning and end of an interpretive endeavor in a given case, it is too much to suggest that it alone constitutes the general rule.²⁰¹

With respect to VCLT Article 32, while the provision is expressly subsidiary in character, the extent of that subsidiarity is often overstated.²⁰² By the word “determine,” Article 32 makes clear that in circumstances of ambiguity, obscurity, or absurdity, supplementary means may be definitive in determining a treaty's meaning.²⁰³ In all other cases, supplementary means may be used only to “confirm” an interpretation reached via application of the general rule in Article 31, but in these as well, the importance of such material must not be undersold. As Richard Gardiner has pointed out, its use to confirm an Article 31 meaning entails the possibility that the meaning will not be confirmed.²⁰⁴ Where this occurs, it will be incumbent on the interpreter to revisit the application of Article 31, potentially leading to the identification of a hitherto unnoticed ambiguity. In such a case, consideration of the supplementary means “is transformed from a potential confirming role of one of determining the meaning.”²⁰⁵ It is therefore an overstatement to say, as some courts or tribunals have, that where supplementary means conflict with an Article 31 interpretation, the latter takes priority by default.²⁰⁶

Further limitations on Article 32 arise through the character of supplementary material itself. The categories of supplementary means are famously not closed: any material that could reasonably be thought to assist the interpreter is admissible, from the travaux préparatoires to academic commentaries.²⁰⁷ But admissibility and weight are not the same thing, and it is always to be remembered that Article 32 has the same fundamental objective as Article 31: to determine the common intention, objectively constructed, of the treaty parties by reference to the text of the treaty itself. Material that reflects intention at or shortly before the treaty's conclusion (i.e., the travaux préparatoires) will therefore be of greater value as an aid to interpretation than partial or third-party materials produced after the fact.²⁰⁸

A regime such as this, which places emphasis on a treaty text as the starting point of the interpretive enquiry, does not necessarily map neatly onto a concept of implied terms. As already explained, an implied term is characterized by an absence of text—a silence emerging from and situated within the treaty's terms that may, if necessary, be filled by the interpreter.²⁰⁹ This initial awkwardness, however, does not justifying throwing up our collective hands: VCLT Article 31(1) does not specify whether the “terms” that are its subject are express or implied, and the ILC in considering ILC Draft Article 27 made clear that it considered it to apply in cases of implied terms.²¹⁰ What it does require, however, is an awareness of how each element of Articles 31 and 32 interacts with the concept of implied terms—and how their application may involve interpretation of a different type that may produce a different interpretive result.

B. Implied Terms Within VCLT Articles 31 and 32

As demonstrated in Part III, the VCLT regime of interpretation developed in fits and starts. Rooted in a disparate case law and diverging conceptions of the interpretive function, it was then subjected to multiple codification efforts and the ministrations of four ILC special rapporteurs before its finalization at the Vienna Conference. Like the documentary hypothesis,²¹¹ the result is the synthesis of multiple sources, each of which depicts a slightly different God, each of which has a slightly different agenda.

Stepping back from the text of Articles 31 and 32, the approach of the VCLT's interpretive engine to implied terms is relatively stable—or stable since the resolution of the textual/teleological conflict, at any rate. The basic approach developed in the case law of the early-to-mid twentieth century remains applicable; delineated, most usefully, in the two Peace Treaties advisory opinions. The default position is that the text of the treaty is complete and that if the parties had intended for it to include a particular term, they would have done so expressly. Any departure from this position will principally be justified by the reduced conception of effectiveness that survived the whetstone of the ILC: a minimalist tool to be resorted to rarely and where required to make the treaty work, whilst imposing additional obligations on the treaty parties only to the extent necessary to meet that objective. While effectiveness is teleological in character, it is teleology of a very particular type, bounded by textualism and subject to strict limits.

1. Good Faith and Effectiveness

As Waldock made clear, the requirement in VCLT Article 31(1) that treaties must be interpreted “in good faith” contains within it the residual conception of effectiveness that permits departure from the general position that treaties are complete on their own terms.²¹²

Effectiveness in this sense has two functions.²¹³ First, the interpreter may identify an implied term in order to prevent an express term from being rendered meaningless (an extension of the rule of effet utile). Secondly, when faced with two equally open constructions of the treaty, one of which contains an implied term, and one of which does not, the interpreter may deploy effectiveness in defense of the implied term, provided that the implied term permits better realization of the drafters’ intent. In deploying effectiveness in this sense, however, the interpreter must bear in mind the general hostility of the VCLT framework to extending the treaty's language past its express limits. Effectiveness “does not entitle a Tribunal to revise a treaty”;²¹⁴ fealty to language, too, is an aspect of good faith interpretation.²¹⁵ The case for the implied term must be correspondingly strong—and unanswerable under the interpretive regime of the VCLT.

A good example of the use of effectiveness in the identification of an implied term arose in the Bosnian Genocide case, concerning the potential responsibility of Serbia for acts said to be carried out by it in Bosnia and Herzegovina during the Yugoslav Wars. As its title suggests, the Convention on the Prevention and Punishment of the Crime of Genocide²¹⁶ (Genocide Convention) contains no positive obligation on states to refrain from committing genocide; rather, it requires only that states prevent and punish genocide and related acts committed by individuals.

These limits notwithstanding, the ICJ had little hesitation in identifying in the Genocide Convention an implied obligation on states not to commit genocidal acts, arising, inter alia, from the need to make effective the obligation to prevent genocide as set out in Article I, as well as the wider humanitarian and civilizing purpose of the treaty.²¹⁷ The Court concluded:

It would be paradoxical if States were thus under an obligation to prevent, so far as within their power, commission of genocide by persons over whom they have a certain influence, but were not forbidden to commit such acts through their own organs, or persons over whom they have such firm control that their conduct is attributable to the State concerned under international law. In short, the obligation to prevent genocide necessarily implies the prohibition of the commission of genocide.²¹⁸

This passage reflects the use of effectiveness in both guises identified. On the one hand, the implied duty identified by the Court was required to prevent an absurd outcome that would render Article I of the Genocide Convention functionally hollow; and on the other, it was needed in order to perfect the objective intention of the treaty's drafters, who had plainly intended to eliminate genocide as a general practice, and not to render it the exclusive preserve of states.²¹⁹

A similar process was undertaken more recently by the arbitral tribunal convened under Annex VII of the UN Convention on the Law of the Sea (UNCLOS) in the South China Sea case. The tribunal was called on to interpret UNCLOS Article 192, placing on the states parties a positive “obligation to protect and preserve the marine environment.”²²⁰ Recognizing the absurdities that could arise if the parties’ obligations were confined to the literal words on the page, the tribunal observed that Article 192 “thus entails the positive obligation to take active measures to protect and preserve the marine environment, and by logical implication, entails the negative obligation not to degrade the marine environment.”²²¹

2. Ordinary Meaning, Context, Object, and Purpose

At the core of VCLT Article 31(1) is the principal subject of the interpretive exercise, viz. “the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose.”²²² This language is better suited to the interpretation of express terms, and its application to implied terms may seem counterintuitive. If an implied term is characterized in the first instance by a silence in the treaty, then the ordinary meaning to be given to that silence is nothing.

Further consideration displaces this initial reaction. It is true that nearly all treaty silences are just that. But Article 31(1) requires that a silence be interpreted in the context of the treaty as a whole, which context includes, per Article 31(2), instruments made by all or some of the parties in connection with the conclusion of the treaty and acknowledged as such.²²³ The ordinary meaning of the express terms of the treaty and any associated instruments, read in light of the treaty's object and purpose, therefore form part of the context in which the silence arises—confirming not only that the silence is there, but also that, in certain cases, the interpreter is compelled to fill it by reference to the principle of effectiveness and other additional materials.

An example of this process arose in Case A15(IV) before the Iran-U.S. Claims Tribunal. The General Declaration set out the basic settlement between the United States and Iran that led to the Tribunal's creation following the Iranian Revolution and its associated Hostage Crisis. It recorded in General Principle B the United States’ commitment to “to terminate all legal proceedings in [its] courts involving claims of [U.S. nationals] against Iran and its state enterprises, to nullify all attachments and judgments obtained therein, to prohibit all further litigation based on such claims.”²²⁴ Iran's case was that the United States was obliged to bring about the termination of relevant legal proceedings in its courts immediately upon signature of the General Declaration. The Tribunal, however, held that the United States was only under an obligation to bring about such termination “within a reasonable period of time”—words that were not included within General Principle B, and which had to be inferred through a good faith reading of the provision (i.e., effectiveness).²²⁵

In particular, the Tribunal hinged its interpretation of General Principle B on the terms of the Claims Settlement Declaration that established the Tribunal itself.²²⁶ This was issued on the same day as the General Declaration, and qualified as a related instrument under VCLT Article 31(2).²²⁷ The Tribunal said:

The [General and Claims Settlement] Declarations set forth no express deadlines for carrying out the obligations they imposed on the United States with respect to terminating legal proceedings in United States courts. Article VII, paragraph 2, of the Claims Settlement Declaration requires the United States to consider claims excluded from the jurisdiction of United States courts “as of the date of filing of such claims with the Tribunal.” Thus, one might understand that the United States obligation to terminate legal proceedings came into existence only when the particular claim at issue in the proceeding was filed with the Tribunal. But a similar obligation is also implicated by Article I of the Claims Settlement Declaration, which requires the two Governments to encourage negotiated settlement and requires that claims not settled by negotiation be brought to the Tribunal. Permitting litigation to continue until the day a claim is filed with the Tribunal seems inconsistent with those obligations under Article I. Thus, in the absence of an express deadline, the Tribunal relies on the general treaty interpretation principle of good faith, which requires the conclusion that the United States was obliged to terminate, within a reasonable period of time, legal proceedings and litigation against Iran in United States courts that were arguably within the jurisdiction of the Tribunal and that consequently should be referred to the Tribunal.²²⁸

Effectiveness alone, however, did not furnish a precise length for the implied reasonable period in General Principle B. The Tribunal therefore engaged in a further inferential exercise based on time limits in the Claims Settlement Declaration. Taking note of the six- to nine-month period in Article I of the Claims Settlement Declaration, during which Iran and the United States would promote the settlement of outstanding claims falling within the Tribunal's jurisdiction pursuant to Article II of same, the Tribunal held that “a similar period logically would constitute a reasonable time frame for terminating the corresponding United States litigation of such outstanding claims.”²²⁹

A further example of the use of treaty context to identify an implied term arose before the International Tribunal for the Law of the Sea (ITLOS) in its Climate Change advisory opinion. There, ITLOS engaged in a comprehensive interpretation of UNCLOS with respect to the obligations of the states parties to combat climate change.²³⁰ One provision on which the Tribunal opined was UNCLOS Article 203, providing for certain preferential treatment for developing states “by international organizations in . . . the allocation of appropriate funds and technical assistance [and] the utilization of their specialized services.”²³¹ It was silent, however, on the role that the states parties were expected to play in this process as members of the relevant organizations. ITLOS filled the silence by reading UNCLOS Article 203 in context of UNCLOS Article 202, which levied similar obligations on states directly. It concluded, accordingly, that “[UNCLOS Article 203] implies the obligation of States to take, through the international organizations of which they are members, the measures necessary to put into effect preferential treatment for developing States as envisaged in this provision.”²³²

3. Subsequent Conduct and Authentic Interpretation

On occasion, the treaty parties may be more direct in telling the interpreter what their treaty means. VCLT Articles 31(3)(a) and (b)—and, on occasion, Article 31(2)(a)—all concern circumstances where the treaty parties have, by their agreement or practice, directed the interpreter as to what they consider to be the proper construction of the treaty. Such material is a powerful aid to construction, constituting a form of authentic interpretation that cannot ordinarily be controverted.²³³ That said, there are limits to the possibilities that such interpretation affords. As the ILC has noted, “[i]t is presumed that the parties to a treaty, by an agreement or practice in the application of the treaty, intend to interpret the treaty, not to amend or modify it.”²³⁴ Put another way, an interpretive agreement or practice cannot do overt violence to an interpretive result reached through application of Article 31(1). Whilst it may require the interpreter to reach a conclusion that would, all things being equal, be considered unlikely, it does not permit the plain words on the page to be abandoned in preference to something else.

When considering implied terms, however, the possibilities of authentic interpretation are widened. Given that silence is the starting point of the analysis, party agreements or practice provide a means by which that silence may be filled without the need to resort to more subtle interpretive tools. Provided the silence is not fulfilled in such a way as to compromise or modify the express terms of the treaty, the interpretation will be admissible, and the implied term directly identified by the parties themselves.

One example of how this might come about occurred in Clayton & Bilcon v. Canada, in which an investment treaty tribunal was charged with interpreting the North American Free Trade Agreement ( NAFTA). NAFTA Article 1116 permitted claims by an investor where the respondent state had breached a relevant provision of the treaty, and “the investor has incurred loss or damage by reason of, or arising out of, that breach.”²³⁵ The provision gave no indication as to whether it permitted claims with respect to the indirect losses of the investor—i.e., losses suffered by directly by a company owned by the investor and passed on to the investor in its capacity as shareholder. Under the logic of most arbitral tribunals, including those operating under NAFTA, the absence of any exclusion for indirect losses would mean that such losses could be claimed²³⁶—despite such claims not being available under customary principles of diplomatic protection.²³⁷ In Clayton & Bilcon, however, the tribunal noted that each of the state parties to NAFTA had, at one point or another, made submissions that Article 1116 should be interpreted to exclude indirect claims. Taken together, the tribunal held that “the NAFTA Parties’ subsequent practice [under VCLT Article 31(3)(b)] militates in favour of adopting the Respondent's position on this issue.”²³⁸ In other words, the tribunal identified in Article 1116 an implied exclusion of indirect claims.

4. Systemic Integration and Relevant Rules of International Law

VCLT Article 31(3)(c) requires, in interpreting a treaty, that account be taken of “any relevant rules of international law applicable in relations between the parties.”²³⁹ This principle, the rule of systemic integration, is necessary to ensure consistency between the treaty and other norms—customary and conventional—that regulate the treaty parties’ relationship, erecting a bulwark against international law's atomization. This explains the prominence given to VCLT Article 31(3)(c) by the ILC study group considering the “fragmentation” of international law²⁴⁰ and the wider literature.²⁴¹ That being said, it is important to be aware of the principle's limitations: it merely requires other rules of international law to be taken into account in interpreting the treaty and construed consistently with them to the extent possible.²⁴² As such, it is a principle allowing the selection of the correct interpretation from a number of candidates, and does not allow swathes of external rules to be shoehorned uncritically into the parties’ bargain. Where the terms of the treaty as construed under VCLT Article 31(1) are clear, Article 31(3)(c) cannot be deployed to add to or modify them.²⁴³

As with the sources of authentic interpretation, the silence that is the initial foundation of an implied term allows VCLT Article 31(3)(c) greater freedom of movement. While this freedom of movement stops short of the statement made in the Georges Pinson case, in which it was held that “[a]ny international convention must be deemed to refer tacitly to customary international law for all questions which it does not itself resolve in express terms and in a different manner,”²⁴⁴ it is nevertheless potentially significant—particularly when supported by other elements of VCLT Article 31.

A pertinent example of this arose in Golder v. United Kingdom, concerning the right to a fair trial under Article 6(1) of the European Convention on Human Rights (ECHR). This provided that “[i]n the determination of his civil rights and obligations or of any criminal charge against him, everyone is entitled to a fair and public hearing within a reasonable time by an independent and impartial tribunal established by law.”²⁴⁵ The applicant was a man convicted of armed robbery serving a lengthy custodial sentence in a British prison. Following a riot at the prison, the applicant was accused by a guard of having assaulted him. The prison then refused the applicant's request to consult a lawyer with a view to bringing defamation proceedings against the guard. The question that arose before the European Court of Human Rights as to whether ECHR Article 6(1) included a right of access to court—a matter on which it was silent.

The Court held with little hesitation that ECHR Article 6(1) did include an implied right of access to court. Its reasoning was based principally in the words of Article 6(1) itself, and notably the right of individuals to have their civil rights determined by an “independent and impartial tribunal.”²⁴⁶ But it also gave significant airtime to systemic integration in its analysis:

Among [the rules to be taken into account under VCLT Article 31(3)(c)] are general principles of law and especially “general principles of law recognized by civilised nations” . . . . The principle whereby a civil claim must be capable of being submitted to a judge ranks as one of the universally “recognised” fundamental principles of law; the same is true of the principle of international law which forbids the denial of justice. [ECHR Article 6(1)] must be read in the light of these principles.²⁴⁷

To that end, the Court concluded that “it follows that the right of access constitutes an element which is inherent in the right stated by Article 6(1).”²⁴⁸ The Court further emphasized, moreover, that in so finding, the Court had not added to the obligations already contained in the ECHR, but merely interpreted Article 6(1) in context, having regard to its object and purpose, and in light of wider international law. Put another way, it claimed merely to have made express what the treaty drafters had previously only implied.²⁴⁹

The ICJ made clear the limits of such possibilities in the Allegations of Genocide case. There, Ukraine sought to rely on the Court's statement in Bosnian Genocide that in carrying out the obligation to prevent genocide, states must “employ all means reasonably available to them . . . within the limits permitted by international law.”²⁵⁰ What this meant, said Ukraine, is that Articles I and IV of the Genocide Convention should be read as preventing states from abusively invoking the obligation to prevent genocide to justify violations of the jus ad bellum—claiming, in turn, that the Russian Federation had done just that in the context of its 2022 invasion of Ukraine.²⁵¹ Articles I and IV say no such thing—and so, although not framed as such, Ukraine was arguing that these provisions contained an implied term.

The Court was unmoved. It noted that its statement in Bosnian Genocide did not “interpret the Convention as incorporating rules of international law that are extrinsic to it,” but provided only that “a state is not required, under the Convention, to act in disregard of other rules of international law . . . [or authorized] to act beyond the limits permitted elsewhere by international law.”²⁵² However, it continued, “[t]hose limits are not defined by the Convention itself but by other rules of international law.”²⁵³ The Court therefore held Ukraine's allegations concerning Russia's alleged abuse of the Genocide Convention were not rooted in the treaty on which its jurisdiction was founded and so fell outside its jurisdiction ratione materiae, i.e., it denied the existence of the contended implied term.²⁵⁴

5. Supplementary Means

The final element of an implied terms analysis under the VCLT is Article 32. For the teleologists, supplementary means—especially the travaux préparatoires—were one of the principal sources of implied terms, permitting in extreme cases the wholesale rewriting of the treaty. The Vienna Conference, however, signaled the demise of that school of thought. The permissible role for supplementary means under the VCLT is a reduced one.

Multiple factors may play a role in determining the relative force of supplementary means in an implied terms analysis. The first is the nature of the interpretation produced by Article 31. If that interpretation is relatively clear, and the supplementary means confirm it, then the latter will reinforce the presence or absence of the implied term respectively. If the interpretation is obscure or absurd, however, the supplementary means will obtain a yet-greater role—guiding the interpreter to the “correct” interpretation, including as to the existence and content of an implied term.

The second is the nature of the supplementary means themselves. Not all supplementary means are created equal. Given the function of the interpretive exercise under the VCLT is to determine the objective intention of the parties, material generated at or immediately before the treaty's conclusion that sheds light on that intention is the most valuable—as Article 32 makes clear by its express reference to “the preparatory work of the treaty and the circumstances of its conclusion.”²⁵⁵ To that end, the travaux préparatoires as a record of the treaty negotiations are usually considered the best supplementary evidence of the parties’ intent, comprising as they do a written record of the negotiations. Where they are available, recourse to them is to be expected—even if there is no impact on the final interpretation.

At the same time, use of travaux préparatoires carries a significant health warning. The first concerns what comes within the ambit of the concept. In short, a document can only be considered part of the preparatory work where it forms part of the negotiating record and was available to all parties participating in the negotiating process.²⁵⁶ It must also demonstrate the common understanding of the parties, and not merely the unilateral views of one or some of them.²⁵⁷ The second concerns the specificity of the material itself. Many commentators have cautioned against placing too much weight on the negotiating record. It is often, at various turns, incomplete, contradictory, and difficult to deploy with full cognizance of its context.²⁵⁸ As Paul Reuter pointed out, “recourse to preparatory work means treading uncertain ground: its content is not precisely defined nor rigorously certified, and it reveals the shortcomings or possible blunders of the negotiators as well as their reluctance to confront the true difficulties.”²⁵⁹ For this reason, the guidance provided by Lord Wilberforce in Fothergill v. Monarch Airlines has much to recommend it: while a treaty's travaux préparatoires is “[a well-established] supplementary means of interpretation,” they can only assist if they “clearly and indisputably point to a definitive legislative intention,” and not merely if the general thrust of the negotiating record supports a particular view.²⁶⁰ Put another way, as Lord Steyn did in The Giannis NK, “[o]nly a bull's eye counts. Nothing less will do.”²⁶¹ However, as Lord Goff recognized in R v. Bow Street Stipendiary Magistrate; Ex parte Pinochet (No. 3), the absence of something from the preparatory material may also constitute the bullseye—especially if the implication in question is so significant that one would expect it to be the subject of considerable commentary during negotiations.²⁶²

The Bosnian Genocide case provides an example of how travaux préparatoires may be used to support identification of an implied term. As already explained, the Court's finding that the Genocide Convention contained an implied obligation on states not to commit genocidal acts was based principally on the general rule of interpretation in VCLT Article 31.²⁶³ But the Court in confirming this interpretation had recourse also to the preparatory work of the Genocide Convention. In particular, it noted that while the drafters of the Convention had made clear that it could not encompass the criminal liability of states, a series of amendments to what would become Article IX allowed for liability under the ordinary rules of state responsibility—and that the Chairman of the Sixth Committee understood the language of Article IX to provide for exactly that. The Court concluded:

[T]wo points may be drawn from the drafting history just reviewed. The first is that much of it was concerned with proposals supporting the criminal responsibility of States; but those proposals were not adopted. The second is that the amendment which was adopted—to Article IX—is about jurisdiction in respect of the responsibility of States simpliciter. Consequently, the drafting history may be seen as supporting the conclusion reached by the Court [under VCLT Article 31].²⁶⁴

The Court's methodology further confirms that any attempt by an interpreter to identify an implied term on the basis of supplementary means alone is likely ex facie suspect. In nearly all cases, the principal source for the inference will be the toolbox of the general rule. The only exception would be where supplementary means operate in determinative mode per VCLT Articles 32(a) and (b). But even in these cases, it would be remarkable for the interpreter to merely cast Article 31 to one side and proceed on the basis of Article 32 alone—reverting, in effect, to the teleological ideal. Numerous elements of Article 31—good faith, object and purpose, treaty context—would still serve as yardsticks against which to deploy supplementary means. Put another way, even if it produces an initially unsatisfactory interpretation, Article 31 cannot be excluded from the interpretive analysis.

Another supplementary means that may play a role in the identification of implied terms are the canons of construction (remember them?). As noted, these can remain relevant to the process of treaty interpretation, insofar as they can be squared with the VCLT regime. One way in which they can be so squared is by classifying them as supplementary means within the meaning of Article 32.²⁶⁵

One canon of potential relevance in the context of implied terms is that of expressio unius est exclusio alterius—the idea that the express reference to one thing excludes implied reference to something else.²⁶⁶ This can be useful in excluding the existence of an implied term in certain cases. To take one example, in the Admission advisory opinion, the ICJ was required to interpret Article 4 of the Charter of the United Nations, providing that UN membership was “open to all other peace-loving states which accept the obligations contained in the present Charter and, in the judgment of the Organization, are able and willing to carry out those obligations.”²⁶⁷ Faced with this express language, the Court refused to imply any additional conditions to membership. According to it, the factors listed in Article 4:

[C]onstitute an exhaustive enumeration and are not merely stated by way of guidance or example. The provision would lose its significance and weight, if other conditions, unconnected with those laid down, could be demanded. The conditions stated in paragraph 1 of Article 4 must therefore be regarded not merely as the necessary conditions, but also as the conditions which suffice.²⁶⁸

Use of the canons as supplementary means carries with the advantage of allowing these principles—which, admittedly, often seem sensible when stated in isolation—a role in the treaty interpretive process in a subsidiary sense without overcoming the carefully-formulated general rule of VCLT Article 31. Some alternatives—for example, seeing the canons as an aspect of good faith interpretation²⁶⁹—risk letting the difficulties of the pre-VCLT era return via the very means through which they were supposed to be removed.

V. The Modern Practice of Implied Terms

The analysis in Part IV is (hopefully) a useful illustration on how the various elements of VCLT Articles 31 and 32 can be deployed to assist in the search for an implied treaty term. Nevertheless, the illustration is not complete—for although the VCLT interpretive regime applies equally to express and implied terms both, it gives itself more easily to the interpretation of express terms.

For this reason, the application of Articles 31 and 32 in the context of an implied term requires an additional layer of analysis beyond the bare words of those provisions. The interpreter must root their frame of reference in (1) the distinction between express and implied terms and how that distinction interacts with (2) the concept of effectiveness inherent in good faith interpretation, and (3) the particular treaty under consideration.

A. Implied Terms Are Derived from Express Terms

To understand how this works, we must recall the observation from the beginning of this Article. The process by which an implied term is identified is a process of interpretation deriving in the first place from the express terms of the treaty, which represent the perfection of the parties’ bargain. An implied term arises from a prima facie silence within a treaty—but determining the scope and, in due course, the content of that silence requires an understanding of the express terms of the treaty, to be reached through the usual tools of VCLT Articles 31 and 32.

The process of identifying an implied term therefore commences with the interpretation of the treaty's express terms, so far as these are relevant to the case at hand. At the end of that process, the interpreter should be able to identify what the express terms of the treaty say (foreclosing, as we will see, an implied term to the extent of any contradiction), identify what they do not say (locating a silence that an implied term may fill), and have an understanding of its drafters’ intent sufficient to determine the content of any term to be implied. As Fitzmaurice observed:

As a matter of strict logic it is only . . . on the basis of a text whose meaning is prima facie or pro tanto clear, or made clear, that the process of inference can properly begin: for in logic no inferences at all can be drawn from something the meaning of which is itself unestablished or obscure. To use the inferential process in order to establish the meaning is therefore to put the cart before the horse and enter upon a vicious circle. Once the meaning is established, and clear so far as it goes, it becomes legitimate to draw those inferences which the text demands or necessarily gives rise to.²⁷⁰

An attempt to circumvent this principle arose in the recent decision of the High Court of Australia in Kingdom of Spain v. Infrastructure Services Sàrl and Another. The case concerned proceedings in Federal Court of Australia for the recognition of an arbitral award rendered for the benefit of the claimants against Spain under the terms of the Convention on the Settlement of Investment Disputes between States and Nationals of Other States (ICSID Convention).²⁷¹ As Spain claimed had immunity from the adjudicative jurisdiction of the Federal Court with respect to those proceedings, the question to be determined was whether the ICSID Convention contained a submission by Spain to the jurisdiction of the Australian courts for the purposes of s 10(1) of the Foreign States Immunities Act 1985 (Cth). This, in turn, fell to be read in light of the rule of customary international law that such a submission could only be express, and not implied.²⁷²

The difficulty for the claimants was that the ICSID Convention contained no such provision. Its high point was Article 54(1), providing that each state party would “recognize an award rendered pursuant to this Convention as binding and enforce the pecuniary obligations imposed by that award within its territories as if it were a final judgment of a court in that State.”²⁷³ Read on its face, this was a promise by Spain that it would recognize any ICSID award presented before its own courts. It did not provide in terms that Spain agreed to submit to the jurisdiction of the courts of any other state with respect to proceedings seeking the recognition of ICSID awards against it.²⁷⁴

Nevertheless, the High Court held that Spain had indeed submitted with respect to the proceedings at bar, on the basis, inter alia, that an express submission to the jurisdiction should be understood as “as requiring only that the expression of waiver be derived from the express words of the international agreement, whether as an express term or as a term implied for reasons including necessity.”²⁷⁵ Put another way, the High Court seemed to say that while a submission to the jurisdiction by prior agreement had to be express, and implied term could nevertheless meet this requirement—as all implied terms descend or are derived from express terms. It then concluded that Spain, by agreeing that other states were under an obligation, per Article 54(1), to recognize awards presented in their courts, had agreed of necessity (that is, impliedly) to submit to the jurisdiction of those courts with respect to recognition proceedings.²⁷⁶

The High Court's position, while an ingenious way to avoid a situation in which an ICSID award could only rarely be recognized in Australia,²⁷⁷ had the effect of collapsing or rendering illusory the distinction between express and implied terms, an eyebrow-raising conclusion. Moreover, as Lord Goff explained in Pinochet (No. 3) (on which the High Court purported to rely²⁷⁸) the rule that adjudicative immunity can only be waived expressly is intended to prevent “international chaos as the courts of different state parties to a treaty reach different conclusions on the question of whether a waiver of immunity was to be implied.”²⁷⁹ By effectively stating that all implied terms were merely variants of an express term, the High Court invited precisely that chaos.

B. An Implied Term Cannot Contradict an Express Term

The need for implied terms to derive from express terms means that the former may not contradict the latter, as this would involve not interpreting the parties’ bargain, but rewriting it to suit the interpreter. In positing this rule, we need to bear in mind a further observation made at the beginning of this Article, which is that an express term includes not only the words on the face of the treaty itself, but certain limited and obvious implications of those words—explicatures. An explicature is an implication so closely linked to the express text of the treaty that it may be seen as part of that text.²⁸⁰ It is to be contrasted with the concept of the implicature, which is commensurate with the concept of an implied term developed here, produces a more complex expansion of the express text of the treaty, and requires greater effort by the interpreter to identify and populate. As a blunt rule of thumb: if the interpreter is seeking to “add” a few additional words (e.g., and only if) to an otherwise complete provision to clarify its meaning, almost for the avoidance of doubt, they have identified an explicature; if they are seeking to add language that expands the substance of the treaty, or changes the focus of an existing obligation, they have identified an implicature.

The upshot of this is that, when we say that an implied term cannot contradict an express term, we must include in the latter category not only the words on the page, but also any explicature emerging from those words. This is important in cases where an explicature closes off the possibility of an implied exception to the express term from which it emerges.

An example of where this process failed is the Air Service Agreement arbitration discussed earlier.²⁸¹ In that case, the majority of the tribunal held that Section IV(b) of the Annex to the U.S.–France Air Services Agreement, providing that “[t]ransshipment when justified by economy of operation will be permitted at all points mentioned in the attached Schedules in territory of the two Contracting Parties,” contained an implied term that gave a U.S. carrier a right to unrestrained transshipment in any third state.²⁸² The majority reached this conclusion through an involved process that took into account the text of the treaty as a whole, its preparatory work, and wider contextual factors—deploying the full toolbox that the VCLT affords. However, it failed to take account of the obvious explicature that Section IV(b) contained. The treaty contained a defined regime for transshipment that was plainly the product of considerable negotiation. It contained no further elaboration on the topic, reflecting a failure by the parties to agree on anything beyond the words on the page. Reuter's opinion in dissent bears consideration:

When a treaty between two or several Parties is intended to establish general rules on a specific subject, it is tempting to infer general principles from that treaty and thereby to collect elements which will make it possible to “fill a gap,” i.e. to settle matters which were not specifically resolved. However, the use of such a constructive approach is only permissible if it truly corresponds to the intention of the Parties as is ascertainable from specific and consistent evidence. That is not the case where the treaty is silent, not because the Parties did not want to lay down detailed rules, nor because the negotiators had neglected to do so, nor because of a development which had not been foreseen at the time the Agreement was concluded, but because of the conscious acceptance of an unresolved disagreement which resulted in a gap being left in the Agreement. . . .

In the present case, the 1946 Agreement devotes a rather substantial provision to “transshipment” . . . . It is therefore apparent from the very structure of that text that the matter was not overlooked by the negotiators, that they realised its importance, and that it was intentional that the issue was left unresolved when the Agreement was concluded. . . . [T]he Parties probably did not regard that silence as a final situation, but they must have accepted the fact that pending a new agreement, general rules which had hitherto been applicable between them should remain in force, . . . . Under these rules, the French Government had to give its agreement to air services relating to French territory in all their aspects . . . .²⁸³

Put another way, given that the parties had plainly not neglected the question of transshipment in their agreement, but had regulated it closely, the obvious inference for the interpreter to draw was not of an implicature setting out a parallel and far more liberal regime of transshipment to Section IV(b); it was to assume that Section IV(b) was the beginning and end of the parties’ bargain, providing (by addition of an explicature) that transshipment when justified by economy of operation would be permitted at all points mentioned in the attached Schedules in territory of the two Contracting Parties and nowhere else. Given that this explicature was more closely tied to the treaty's actual text than the majority's identified implicature, the explicature took priority, and ruled out the possibility of the implicature.

A treaty may also preclude implied terms globally, by expressly stating that the parties’ bargain is contained within the four walls of its text and cannot be improved upon.²⁸⁴ Such terms are rare, but not unheard of. One example is contained in Article XI of the Indus Waters Treaty between Pakistan and India. Intended, inter alia, to regulate the use of rivers flowing through the disputed territory of Kashmir, the parties were careful to that nothing affecting their competing territorial claims could be inferred from the treaty's terms. To that end, Article XI(1)(b) provides that:

It is expressly understood that . . . nothing contained in this Treaty, and nothing arising out of the execution thereof, shall be construed as constituting a recognition or waiver (whether tacit, by implication, or otherwise) of any rights or claims whatsoever of either of the Parties other than those rights or claims which are expressly recognized or waived in this Treaty.²⁸⁵

This term constitutes a powerful argument against any attempt to identify an implied term in the terms of the Indus Waters Treaty—and certainly any implied term that could affect Pakistan or India's claims concerning Kashmir. To that end, the Court of Arbitration convened in the Kishenganga case affirmed that its decisions, and the interpretations of the Indus Waters Treaty set out therein, could have no implications for the status of Kashmir as a contested territory.²⁸⁶

C. An Implied Term Must Be Necessary

The two principles set out above are intended to guide the interpreter in identifying a silence that is available to be filled by an implied term. Once this is done, the question whether it is appropriate for the silence to be filled and, if so, with what.

It is here that the concept of effectiveness arises once more. As explained, effectiveness is inherent in the good faith interpretation mandated by VCLT Article 31(1) and was flagged by Waldock as the conceptual engine room of implied terms in treaties. Initially a wide-ranging notion that underpinned the teleological approach to treaty interpretation, the ILC and Vienna Conference left it much reduced, and applicable in two circumstances only: first, to allow the treaty to function; and second, to select between competing “correct” interpretations of the treaty.²⁸⁷ The context of implied terms gives itself far more readily to effectiveness in the former guise—although the latter cannot be excluded.

What is clear, however, is that effectiveness will only permit identification of an implied term to the extent necessary. Fitzmaurice provides a useful explanation of this notion:

[T]here tends to be confusion between possible and necessary inferences. Inferences are not to be drawn for the fun of it, so to speak: it must be necessary to draw them in the given circumstances, and in consequence they themselves must have a character of necessity and not merely of possibility. Given the element of the speculative which . . . must, in some degree at least, enter into the process of inference, the familiar concept of the “reasonable” inference is a legitimate one, provided it is borne in mind that the only truly reasonable inferences are those which impose themselves, and not those which merely suggest themselves as possible or desirable.²⁸⁸

So far, so good. But what is necessity in this context? Two forms suggest themselves. In the first place, identification of the implied term may be irresistible as the application of VCLT Articles 31 and 32 has rendered the inference so obvious that it cannot be denied.²⁸⁹ In the second, the implied term may be essential to preserve the integrity of the treaty's system and prevent frustration of its object and purpose, even if it cannot be said that the drafters ever turned their mind to the issue. From a certain point of view, however, these two forms collapse into one: it is undeniable that the drafters intended their creation to work, and so if an implied term is necessary to preserve its operation or to avoid absurd results, it follows that it must have been intended.²⁹⁰

This analysis only covers half the problem. It is one thing to identify that an implied term is necessary as a matter of fact. It is another to give it content. Necessity applies here as well, in the sense that the interpreter may give no greater life to the implication than necessary to fill the gap identified. Put another way, the interpreter must ensure that the implied term adds no more to the express terms of the treaty than is required—not just in the case at bar, but generally.

Again, the Peace Treaties advisory opinions show how this may work in practice. There, the ICJ was confronted by a dispute settlement clause that could readily be frustrated by one party failing to appoint its representative to a dispute settlement Commission. The ICJ agreed that an implied term was necessary to prevent the clause from being frustrated, and so identified, in the first Peace Treaties opinion, an implied obligation on each party to appoint its representative at another party's request.²⁹¹ When certain of the parties failed to comply with this implied duty, however, the Court refused, in the second Peace Treaties opinion, to identify further implied terms to remedy the position: the Court having done what was necessary to save the clause in the first Peace Treaties opinion, anything additional was logically unnecessary and amounted to rewriting the clause to suit the preferences of the interpreter.²⁹²

The Peace Treaties opinions also reveal a further relevant factor in the application of necessity, which is the proximity of the proposed implication to the actual text of the treaty. One of the reasons that explicatures are readily identified is because they do not stray overmuch from the express terms of the treaty. They are rendered necessary by the words on the page. But, in the case of an implicature, when the inference suggested strains the semantic boundaries of those words, or obviously seeks to move beyond them, or change their focus, the case for the inference must be that much the greater in light of the wider framework of VCLT Articles 31 and 32. In the first Peace Treaties opinion, it was relatively easy for the Court to infer, from a party's right to appoint a representative, a corresponding obligation of that party to appoint that representative when asked. In the second Peace Treaties opinion, the Court was reluctant to infer that the drafters of the clause intended for the UN secretary-general to act as an default appointing authority when the parties (1) had already given the secretary-general such a role in relation to the third (neutral) member of the Commission, and (2) could easily have extended this role to the case of a delinquent party failing to appoint its representative, but (apparently) chose not to.

D. An Implied Term Must Take Account of the Nature of the Treaty

A final consideration to be borne in mind when considering an implied term is the nature of the treaty itself. At a high level, this is a general argument against the implication of terms, for the reasons given by Lord Goff in Pinochet (No. 3):

I recognise that a term may be implied into a treaty . . . . It would, however, be wrong to assume that a term may be implied into a treaty on the same basis as . . . an ordinary commercial contract . . . . This is because treaties are different in origin, and serve a different purpose. Treaties are the fruit of long negotiation, the purpose being to produce a draft that which is acceptable to a number, often a substantial number, of state parties. The negotiation of a treaty may well take a long time, running into years. Draft after draft is produced of individual articles, which are considered in depth by national representatives, and are the subject of detailed comment and consideration. The agreed terms may well be the fruit of “horse-trading” in order to achieve general agreement, and proposed articles may be amended, or even omitted in whole or in part, to accommodate the wishes or anxieties of some of the negotiating parties. In circumstances such as these, it is the text of the treaty itself which provides the only safe guide to its terms, though reference may be made, where appropriate, to the travaux préparatoires. But implied terms cannot, except in the most obvious cases, be relied on as binding the state parties who ultimately sign the treaty, who will in all probability include those who were not involved in the preliminary negotiations.²⁹³

Pinochet (No. 3) concerned the Convention against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment²⁹⁴ (Torture Convention) a treaty to which these considerations undoubtedly apply. And while the leitmotif of Lord Goff's reasoning, that treaties are not contracts, is obviously correct, it is equally correct to point out that not all treaties are the same. Whilst some are the consequence of many years of painstaking multilateral negotiations, others are not. Early BITs, for example, were often concluded on the basis of model agreements produced by capital-exporting states, and concluded by capital-importing states with little additional negotiation or awareness of what their terms imported.²⁹⁵ While not a contract, one can see how the identification of implied terms may operate differently in the case of such a BIT than for the Torture Convention.²⁹⁶ The character of the treaty is an essential part of the context of its terms, which forms part of the interpretive calculus of VCLT Article 31(1).

This insight—that different legal texts compel different interpretive approaches—is not a new one. Ahron Barak famously proposed drawing interpretive distinctions based on the legal character of the text being interpreted, with wills on one end of the spectrum and constitutions on the other.²⁹⁷ But whereas domestic legal systems have devised a range of instruments to serve as vehicles of legal rights and obligations, international law in the post-1945 era tends (as we have seen²⁹⁸) to use the treaty as the sole means for the granular organization of international society.²⁹⁹

Nevertheless, distinctions may be drawn between different types of treaties—and these distinctions may reflect our willingness to infer the existence of implied terms within them as a general matter. This is not to say that the inference must not be necessary—merely that our perception of what is necessary changes with the treaty in view.

Various distinctions suggest themselves in this respect. The first concerns the membership of the treaty, which in turn impacts on the complexity of its negotiation. A short form bilateral treaty, hastily concluded, without significant discussion, may lend itself more readily to the identification of implied terms, as the drafters’ work may be more obviously incomplete. Not every BIT, for example, provides that an investment procured by the investor through corruption or other serious illegality will not be subject to protection—but the existence of an implied term to this effect is so obvious that it may be identified simply by saying it out loud.³⁰⁰ But, when dealing with a multilateral treaty, an additional level of caution is required. Such treaties are often the result of detailed negotiation in which the drafters were careful to write down only what they wanted committed to paper.³⁰¹ They also allow for the possibility that states may join the treaty after the finalization of its text. In such cases, the objective intention of the drafters may preclude the identification of implied terms—or ready identification, at any rate.³⁰²

A second distinction concerns the scope of the issues dealt with by the treaty, together with its aspirations for completeness. Where the subject matter of the treaty is narrow and technical, it is objectively less likely that its drafters intended to supplement it via implied terms. Where it is broad-ranging and comprehensive, it is more likely that the treaty parties intended any apparent hole to be plugged by the interpreter through identification of an implied term. The ICSID Convention, for example, is a technical treaty that sets up a process of investor-state dispute settlement together with the means for enforcing resulting awards. Its drafters are unlikely to have intended anything to be added to it beyond the words on the page. Conversely, UNCLOS declares in its preamble that it is intended to establish “a legal order for the seas and oceans,”³⁰³ a concept with constitutional implications.³⁰⁴ Its drafters intended it to be comprehensive, and as such, it is permissible to approach the question of whether an implied term is necessary to fill an apparent silence in its text with a broader appreciation of what necessity actually entails—as the Climate Change advisory opinion makes clear. At the same time, it is important to recall that UNCLOS is also a framework convention,³⁰⁵ which deliberately leaves some areas to be addressed by subsequent treaties.³⁰⁶ When dealing with these areas, a narrower appreciation of necessity could well be appropriate—lest the interpreter trespass on an area that the drafters deliberately left over to be addressed by the states parties in the future.

A third distinction concerns treaties intended to develop over time to meet the objectives of their drafters. This is reflected in the concept of evolutionary interpretation; a departure from the usual rule that treaty terms must be interpreted as at the time of the treaty's conclusion.³⁰⁷ Where a treaty is subject to evolutionary interpretation, it is objectively more likely that its drafters intended that evolution to occur not only through construction of the treaty's express terms, but also—where necessary—through the identification of implied terms. This is most obviously the case for treaties such as the ECHR, which is a “living instrument”³⁰⁸ intended to move with the times. It is also a “law-making treaty”³⁰⁹ that sets out rules of general and enduring significance, to be interpreted in the manner “most appropriate in order to realise the aim and achieve the objective of the treaty.”³¹⁰ Such leeway is less obvious where the treaty is intended to have permanent and absolute effect erga omnes—as in the case of a boundary treaty.³¹¹

These distinctions are not exhaustive. There may be others that suggest themselves.³¹² They are also not absolute criteria to be applied rigorously to each treaty. They reflect the fact that the essential character of a treaty is part of its context, and that, by virtue of that context, some treaties give themselves more readily to the identification of implied terms than others. In this sense, it is possible to imagine a spectrum. At one end stand agreements like the Indus Waters Treaty: technical instruments that expressly deny the possibility of implied terms. At the other stand treaties like the ECHR: instruments grounded in mutable principles designed to evolve over time and intended to be comprehensive within their fields. The conception of such a spectrum does not reflect the abandonment of textual interpretation, nor the general principle that an implied term can only be identified where necessary. Rather, it reminds us that the conception of what is necessary will vary from treaty to treaty.

VI. Conclusion

Implied terms remain a challenging area of inquiry within the law of treaties. Here, I have attempted to establish the principled framework for their identification within the textual rules of interpretation under the VCLT, together with an understanding of how the elements of those rules apply in the context of implied terms. It is the start, not the end, of a conversation.

The question of whether to identify an implied term in the otherwise complete text of a treaty is ultimately one that is personal to the individual interpreter. The VCLT regime does not purport to limit the entities or persons that may undertake the task of interpretation;³¹³ it is a toolbox intended for use by all. While some interpretations may carry greater force than others, the final judgment in each case is a subjective appreciation. While the interpretation may be given a veneer of respectability by claiming as its source the objective intention of the parties, the fact remains that the interpreter's understanding of that objective intention is itself subjective.³¹⁴

In a similar vein, Martti Koskenniemi famously claimed that international law is an argumentative system in which any outcome is capable of justification with impeccable legal reasoning.³¹⁵ This statement self-evidently cannot be correct in its entirety, otherwise international law would be little more than an exquisite set of policy justifications.³¹⁶ This Article attempts to limit the number of plausible interpretive outcomes in the context of implied terms to counter such assertions. The framework it has proposed is a simple one, rooted in the essentially textual character of modern treaty interpretation. Beginning with the express text of the treaty and its explicatures, it seeks to identify silences and then fill them (or not) using the tools provided by VCLT Articles 31 and 32. In summary, it requires the following in any given case:

• • First, interpretation of the treaty's express terms in accordance with the usual and predominantly textualist approach of VCLT Articles 31 and 32, bearing in mind that the express terms include not only the semantic content of the treaty (i.e., the words on the page), but also that limited pragmatic content that can be drawn immediately from those words—which together form an explicature.

• • Secondly, against the background of the express terms, identification and delimitation of a silence, situated within the treaty in view, that is not filled through interpretation of the express terms.

• • Thirdly, determination of whether the silence so identified is capable of being filled by identification of an implied term (or implicature³¹⁷), using the full toolbox of VCLT Articles 31 and 32, but bearing in mind:

  1. ○ That there is a heavy presumption that a silence within the treaty is just that, and that the principle of effectiveness will only allow it to be filled where objectively necessary.

  2. ○ That, if the determination is made that the silence in the treaty can be filled by identification of an implied term, the implied term so identified (1) cannot contradict an express term, (2) can be no broader or more onerous than necessary to fill the silence identified, and (3) cannot be defined or otherwise shaped by the contemporary policy preferences of the interpreter, save in cases where evolutionary interpretation is appropriate.

  3. ○ That in both cases, what is necessary will depend on the character of (1) the particular treaty, and (2) the particular silence to be filled within it.

The fundamental insight of this approach is that implied terms may only be inferred where necessary as part of a rule-governed interpretation of express terms—with the threshold of necessity differing from treaty to treaty, and even from interpretation to interpretation. This variability, however, is grounded in the basic observation that most of the time, treaty parties say what they mean, and implied terms remain correspondingly exceptional. As the Appellate Body of the World Trade Organization observed in Canada—Patent Term, “[s]ometimes the absence of something means simply that it is not there.”³¹⁸

But even more fundamentally, this Article is a plea for the concept of implied terms to be adopted more widely by treaty interpreters—even if the circumstances in which such implied terms are identified are rare. No small amount of confusion has bedeviled treaty interpretation, even in the post-VCLT era, because interpreters do not say when they are purporting to identify an implied term. Rather, they are content to simply wrap the written and unwritten parts of the treaty up in a generalized process called “interpretation” and call it a day. This does not help anyone and leads only to uncertainty as to what the interpreter has concluded a treaty means and how he or she has reached that conclusion. Solving this problem starts, as with most problems in international law, with a proper taxonomy—which means recognizing that implied terms are not only distinct from express terms, but produced via a different interpretive process that will produce a different interpretive result. It is only by properly distinguishing between the two classes of term that these different results can be appreciated—and the different process (or the version of that process set out here, at any rate) properly delineated and consistently applied.