I. INTRODUCTION
Article 18 is one of the most puzzling provisions of the Vienna Convention,Footnote 1 since it establishes a relatively vague ‘interim obligation’ for States ‘to refrain from acts which would defeat the object and purpose of a treaty’ between its signature and ratification. This obligation exists until the State in question has made its intention clear not to become a party to the treaty (Article 18(a)), or when it has expressed its consent to be bound by the treaty, pending its entry into force and provided that such entry into force is not unduly delayed (Article 18(b)). Although the existence of such an interim obligation has been recognized by States and in various international legal regimes, it remains problematic since Article 18 does not define its own contours nor when and under which conditions it is being breached.Footnote 2 What is clear, conversely, is the general idea behind it, namely that the value of a treaty-based undertaking between States shall not be weakened before the treaty effectively enters into force.Footnote 3 This is a principle with a long tradition, as even before the drafting of the Vienna Convention, it was argued that neither party shall (without repudiating the proposed treaty itself) ‘voluntarily place itself in a position where it cannot comply with the conditions as they existed at the time the treaty was signed’.Footnote 4 This provision protects the commitment of the treaty signatories not to defeat the object and purpose of the treaty in question, which also means that even a simple signature entails certain legal obligations. It represents a middle ground between there being no obligations at all and full commitment to the treaty. This allows States to review the treaty before it becomes fully binding on them,Footnote 5 while, internationally, the interim obligation of Article 18 furthers and facilitates cooperation between signatory States.Footnote 6
Yet, the question remains as to what exactly this interim obligation is. The 1935 Harvard Draft Convention on the Law of TreatiesFootnote 7 and the International Law Commission in 1965Footnote 8 provided altogether eight examples of violations of this interim obligation, fleshing out the problem how a treaty might be seriously hampered or even rendered nugatory.Footnote 9 But there are further questions which this article seeks to answer: first, as examined in section II, it remains uncertain how the interim obligation of Article 18 fits within the general international law of treaties; what is its legal nature and temporal scope; the role which the principle of good faith plays as a possible underlying principle; and how one should understand the object and purpose of a treaty and how it can be defeated. A further issue concerns the scope of application of Article 18. These examples suggest its apparent focus is on bilateral rather than multilateral treaties, but the application of this interim obligation exclusively to bilateral treaties would contravene both the expressed and implied intent of the drafters.Footnote 10 Therefore, section III discusses how Article 18 fits within the normative system of international law and law-making treaties, such as human rights treaties. The problem of the concept of interim obligations is that reprehensible State behaviour (such as human rights violations) does not so much defeat the object and purpose of a treaty aimed at its prohibition: on the contrary, and paradoxically, it rather emphasizes the need for such a treaty.Footnote 11
II. Article 18 VCLT in the General International Law of Treaties
A. The Normative Nature of Article 18 VCLT
1. From a moral duty…
The exact nature of the interim obligation now enshrined in Article 18 has been controversial from the very beginning. Its origins lie in Article 9 of the Harvard Draft Convention on the Law of Treaties which, for the first time, formulated this interim obligation arising from the principle of good faith in a codified version.Footnote 12 The comment to Article 9 then clearly states that it ‘does not envisage a legal duty […] under international law’.Footnote 13 The drafters continue by explaining that the reason for this being a moral rather than legal obligation is that, in contrast to the non-performance of a legal duty (which would involve important legal consequences such as the liability to make reparation for any losses or damages sustained by the other party or parties), the non-performance of such an interim obligation produces no such legal consequences. Such behaviour may certainly impair the non-performing State's reputation or expose it to the charge of bad faith, but it does not subject that State to the rules of State responsibility for the violation of a legal obligation, because no legal obligation yet exists which could have been violated.Footnote 14
In his comment on Article 9, JL Brierly, Special Rapporteur on the Law of Treaties, therefore emphasized that this interim obligation was a good faith duty rather than a legal obligation and thus without legal consequences.Footnote 15 Fernand Dehousse similarly argued that in the interval between the signature and ratification of a treaty there are only moral reasons for a signatory's good faith behaviour, and hence no legal obligations.Footnote 16
This view that a treaty cannot have legal effects before its entry into force stands to reason, and is supported by various other principles of the law of treaties. First, according to the principle of pacta sunt servanda, only treaties in force are binding upon the parties and must be performed by them in good faith. However, since the obligation of good faith performance is dependent on the treaty's being in force, the determination of when exactly a treaty enters into force with respect to a party is critical.Footnote 17 Second, this determination is based on the basic rules of Article 24(1) VCLT, which states that ‘a treaty enters into force in such manner and upon such date as it may provide or as the negotiating States may agree’, and Article 24(2) VCLT, providing that ‘failing any such provision or agreement, a treaty enters into force as soon as consent to be bound by the treaty has been established’. Both provisions reflect customary international lawFootnote 18 and consequently clarify that, in the context of the current discussion, the potential imposition of a legal obligation on a State before a treaty's entry into force is difficult to justify. Third, and finally, Article 28 VCLT, which is regarded as a codifying either an existing customary ruleFootnote 19 or a general principle of law,Footnote 20 denies retroactive effect to a treaty, ‘unless a different intention appears from the treaty or is otherwise established’. When read together, these provisions strongly appear to express the general principle that treaties do not have any legal effect before their entry into force.Footnote 21
2. …to a legal obligation
The more sceptical Brierly was succeeded as Special Rapporteur on the Law of Treaties by Sir Hersch Lauterpacht, who showed more enthusiasm for a legally binding interim obligation in the law of treaties. In his report for the Fifth Session of the International Law Commission in 1953, he regarded signature as having ‘the effect of obliging the signatories to abstain, prior to ratification, from a course of action inconsistent with the purpose of the treaty’ and thus argued that this ‘obligation constitute[d] a legal, and not merely a moral, duty’.Footnote 22 Furthermore, Lauterpacht suggested that this interim obligation only referred to intended and not merely calculated acts capable of impairing the value of the obligation as signed, and that consequently, its purpose was to prohibit action in bad faith which deliberately aims at depriving the other party of benefits legitimately expected from the treaty. Lauterpacht's approach appears to suggest that what matters is whether a State subjectively acts in bad faith, but he subsequently provides an example which supports the view that what actually matters is whether an action by a State can be regarded as a manifestation of bad faith.Footnote 23 The next Special Rapporteur, Sir Gerald Fitzmaurice, followed a more cautious approachFootnote 24 by dropping any reference to good or bad faith and describing the interim obligation as a duty ‘not to take any action calculated to impair or prejudice the objects of the treaty’.Footnote 25 More importantly, however, and like Lauterpacht before him, he stressed the legal nature of the interim obligation.Footnote 26 Nevertheless, it was not until the 1962 report of Sir Humphrey Waldock that the International Law Commission (ILC) began to study the problem of the interim obligation more seriously and in more detail.Footnote 27 It is important to note that, first, Waldock concurred with Lauterpacht and Fitzmaurice in regarding the interim obligation as being a legal duty, and second, he reintroduced good faith as an essential element in the obligation ‘to refrain from any action calculated to frustrate the objects of the treaty or to impair its eventual performance’.Footnote 28 This brief illustration of the development of Article 18 shows that it underwent considerable discussion and modification, and that there is a general agreement that as of the 1953 Lauterpacht draft, the interim obligation was considered to be a legal obligation.Footnote 29
B. The Temporal Scope of Article 18 VCLT
Although Article 18 is a legally binding obligation, it nonetheless suffers from certain deficiencies which weaken its normative force.Footnote 30 The first of these deficiencies concerns the exact temporal scope of the interim obligation which requires further interpretation, especially in the light Article 18(a) and (b) setting out the interim obligation both before and after the State's consent to be bound, respectively.
1. The interim obligation before the consent to be bound
Article 18(a) covers situations in which a State has formally accepted a negotiated treaty, but not yet expressed its consent to be bound in the manner agreed upon, eg through ratification or approval. For this reason, the State in question can at any time end the interim obligation by ‘making its intention clear not to become a party’.Footnote 31 This ‘provisional status’Footnote 32 of a treaty, and thus a State's interim obligation, is established by its signature or the exchange of instruments constituting the treaty. Either of these acts approves and authenticates the treaty.Footnote 33 It is important to note that ‘signature’ does not refer to the mere initialling or signature ad referendum within the meaning of Article 10(b) VCLT, nor to the definitive signature of Article 12 VCLT. The word ‘signature’ in Article 18(a) is to be understood as both a ‘final signature’ which approves and authenticates the text of the treaty as well as a ‘simple signature’ of the treaty, which is then subject to ratification. Similarly, the exchange of instruments, which is also subject to ratification, must be distinguished from the ‘exchange of instruments’ provided for in Article 13 VCLT. This means that in order to give rise to the interim obligation, the signature must be legally effective and binding, and must accordingly be carried out by a person with the required official capacity.Footnote 34
The interim obligation under Article 18(a) ceases once the State ratifies the treaty or makes clear its intention not to become a party to the treaty qua ratification. This reflects the sovereign right of any State not to conclude a signed treaty; and even after signature, eventual ratification remains a discretionary act and is, unless agreed otherwise, not obligatory and may be withheld for any reason.Footnote 35 In this vein, Article 8 of the Harvard Draft stated that ‘[t]he signature of a treaty on behalf of a State does not create for that State an obligation to ratify the treaty’,Footnote 36 and the ILC rejected a proposal to include such an obligation in the future Vienna Convention as early as 1951.Footnote 37 Unfortunately, Article 18(a) remains silent as to the means by which a State has to express its intention not to become a party in order to bring the interim obligation to an end. Silence is not sufficient to do so,Footnote 38 nor is an action which frustrates the object and purpose of the treaty, as this would render Article 18 meaningless.Footnote 39 Especially in the light of the Dissenting Opinion by Judge Lachs in the North Sea Continental Cases,Footnote 40 a State expressing its intention to invalidate the signature should employ a certain measure of formality, and the act of withdrawal should be as final and explicit as an act of signature.Footnote 41
The prime example of how a State can make clear its intention not to become a party to a treaty is the withdrawal of the United States signature from the Rome Statute in May 2002, which became known as ‘unsigning’. In a letter to the United Nations Secretary General, the Bush administration announced in December 2000 that the US did ‘not intend to become a party to the treaty’ and that ‘[a]ccordingly, the United States has no legal obligations arising from its signature’,Footnote 42 without, however, mentioning Article 18 itself. The underlying reason for this act of unsigning related to a series of controversial bilateral non-surrender agreements which the US had concluded with various States which were either signatories or already parties to the Rome Statute before it entered into force. Article 98(2) of the Rome Statute meant that these agreements effectively prevented the International Criminal Court from exercising its jurisdiction over US citizens. At that time, it was unclear how unsigning a treaty fitted within the law of treaties, as there was no established custom related to ‘unsigning’.Footnote 43 Interestingly, however, an academic legal opinion on the US practice of concluding such bilateral non-surrender agreements considered that they would indeed prevent a signatory (such as the US) from fulfilling its obligations to the International Criminal Court and to other State parties to the Rome Statute,Footnote 44 and that the act of unsigning would effectively relieve it from its obligations to abide by the interim obligation of Article 18.Footnote 45 It has been argued that it was exactly to avoid any obligations under the law of treaties (as opposed to the Rome Statute itself) flowing from this interim obligation that may have prompted the act of unsigning.Footnote 46 This ‘easy way out’ greatly limits the potential force of the interim obligation,Footnote 47 and in the future, States pressuring signatories into compliance must consider whether their actions could prompt these signatories to unsign rather than to ratify the treaty in question.Footnote 48 One could also argue, of course, that the US was well aware of the possible consequences of acting in violation of its interim obligations, and that its unsigning of the Rome Statute was perfectly in line with Article 18 and its customary law status (as the US is only a signatory to the Vienna Convention and therefore not legally bound by it).Footnote 49 In this vein, its actions were fully compatible with international law.
Unsigning could, however, also be perceived as an action which defeats the object and purpose of a treaty and thus in violation of the interim obligation,Footnote 50 particularly since there is no such thing as ‘physical unsigning’; the UN Treaty Collection still lists the US as a signatory to the Rome Statute, albeit with the caveat of a footnote referring to the letter from 2002.Footnote 51 Yet regarding signature as the first step in a one-way ratchet process towards ratification is mistaken on two counts: first, such an approach would be inconsistent with the wording of Article 18, which certainly allows States to make clear their intention not to become a party to a treaty. And second, we must assume that if it is possible for a State to take the even more drastic step of withdrawing from a ratified treaty (either under the terms of the treaty itself or Article 56 VCLT), then the less drastic step of withdrawal from the interim obligation must also be possible. A State cannot be expected to ratify a treaty which it does not intend to uphold in order to allow it to withdraw from it formally after ratification. This scenario would be extremely disruptive to the treaty regime in question and to the law of treaties in general.
2. The interim obligation after the consent to be bound
In contrast, Article 18(b) deals with situations in which a State has already expressed its consent to be bound by the treaty, ‘pending the entry into force of the treaty and provided that such entry into force is not unduly delayed’. This scenario usually occurs where multilateral treaties require that in order to enter into force all, or a specific number of, parties have expressed their consent to be bound (typically through ratification, but any other means under Article 11 VCLT equally suffices), and/or that a specific period of time has elapsed since the necessary number of ratifications has been attained. Therefore, Article 18(b) applies in the interval between the valid declaration of consent to be bound made by each contracting State and the eventual entry into force of the treaty for that State. Given that in such a situation the State in question has already done everything necessary on its part to support the entry into force of the treaty in question, one can presume that its behaviour is more oriented towards compliance with the substance of the treaty than that of a mere signatory States under Article 18(a):Footnote 52 This means that there is less need to protect sovereign freedom of action than in the earlier stages covered by Article 18(a).
Article 18(b) also introduces the element of ‘undue delay’ which is somewhat ambiguous and needs further interpretation. During the Vienna Conference, Argentina, Ecuador, and Uruguay proposed a fixed 12-month period between ratification and entry into force, in which States might freely withdraw from the treaty,Footnote 53 whilst Special Rapporteur Waldock had even suggested a limit of ten years.Footnote 54 However, these attempts to introduce a fixed time-limit ultimately failed, and ‘undue delay’ in the treaty's entry into force must be ascertained in the light of the circumstances of every individual case and several factors need to be taken into account, such as the number of contracting States, the complexity of the treaty subject-matter, the amount of political controversy, and the time taken to negotiate the treaty.Footnote 55 Considering the time taken for some multilateral treaties to enter into force, periods of less than five years do not appear to amount to an undue delay.Footnote 56 However, even beyond that period of time, the interim obligation does not necessarily cease to exist. For example, it has been argued that there is no such undue delay in the case of the Comprehensive Nuclear Test Ban Treaty, which was opened for signature in 1996 and still has not entered into force. Reasons for this include the treaty's built-in mechanism in Article XIV which address such a delay, requiring the UN Secretary-General to convene a Conference of the States that have deposited their instruments of ratification to consider the situation and decide on further measures.Footnote 57 Moreover, the General Assembly continues to adopt a resolution promoting the treaty's entry into force every year, which is supported by the positive votes of almost all UN member States.Footnote 58 Consequently, one cannot certainly say that this treaty has fallen into desuetude and that the interim obligation has ceased to exist.Footnote 59
Lastly, and contrary to paragraph (a), paragraph (b) does not provide a procedure to ‘unratify’ a treaty, as it does to ‘unsign’ it. The question of whether a State which has already consented to be bound may withdraw such consent before the treaty enters into force is not addressed by Article 18 (nor by any other provision of the Vienna Convention), and such a withdrawal is not, therefore, a breach of the interim obligation.Footnote 60 This means that since the withdrawal of the consent to be bound is not prohibited by the Vienna Convention, a contracting State is legally bound by a treaty only after it entered into force and prior to that date, the State's freedom of action still exists.Footnote 61 There is State practice where the withdrawal of instruments of accession or ratification before the entry into force of a treaty was not met with any objections by the other contracting States.Footnote 62 Therefore, as the depositary of multilateral treaties, the UN Secretary-General in principle allows the withdrawal of an instrument of ratification or instruments of a similar nature until the entry into force of the treaty in question, implying that until then, States are definitely not bound by the treaty.Footnote 63
C. Good Faith as an Underlying Principle
It is evident that the common enterprise of an international agreement between or among States creates a mutual bond of trust, requiring that none of those States should attempt to defeat the object and purpose of the treaty before it enters into force. If a State no longer wishes to participate in this common endeavour, it is free to withdraw its signature or ratification, depending on the stage of treaty formation. It follows from this, conversely, that a State contemplating withdrawal should not be free to ‘torpedo’ the treaty by undertaking disloyal acts. The issue is, therefore, one of protecting legitimate expectations and of ensuring minimally loyal behaviour, or of acting in good faith.Footnote 64 There is no doubt that good faith plays a role in ensuring the interim obligation of this provision, but it is not self-evident what exactly its role is and what it entails, particularly as Article 18 itself does not mention this principle.
Good faith ‘is not itself a source of obligation where none would otherwise exist’,Footnote 65 but it can certainly help refine an otherwise vague provision, such as Article 18. An example of such as approach (predating the Vienna Convention by almost 40 years) can be found in the 1928 Megalidis decision by the Greco-Turkish Mixed Arbitral Tribunal. Turkey had signed the Treaty of Lausanne in 1923 with, inter alia, Greece, yet prior to its entry into force, it had seized the property of a Greek national, claiming that this property was exempt from restitution under Article 67 of the Lausanne Treaty. The Tribunal, however, considered the seizure to be illegal, as ‘already with the signature of a treaty and before its entry into force, there is for the contracting parties an obligation not to do anything that could harm the treaty […]. This […] amounts to a manifestation of good faith […].’Footnote 66 This view is also supported by the conclusions of the ILC concerning the Vienna Convention, which considered here to be an obligation of good faith to refrain from defeating the object and purpose of a treaty to be a legal obligation inherent in Article 18.Footnote 67
Nevertheless, this does not answer the question whether there must be a subjective intention on part of the State in question to deprive the treaty of its efficacy, or whether it is sufficient that, objectively, it causes this to happen.Footnote 68 Given that good faith ‘must mean more than just “good form”’ and that the ‘signature of a treaty is […] not an empty gesture’,Footnote 69 a subjective approach to good faith seems plausible. This argument is further bolstered by the ICJ in the Nicaragua case when it referred to acts that are ‘calculated to deprive’,Footnote 70 ‘calculated to defeat’ or ‘directed to defeating’Footnote 71 a treaty's object and purpose. These variations on a theme, namely bad faith, not only seem to reflect the actual terms of Article 18, but they also suggest that the concern might be more than whether the actions in question actually and objectively defeat the object and purpose, thus also suggesting that the subjective intention behind the relevant acts is important.Footnote 72 Consequently, the interim obligation would only be violated by acts which are intended to frustrate the object and purpose of a treaty, and not by acts which frustrate it unintentionally.
Having said that, there are, however, reasonable grounds to assume that the principle of good faith inherent in Article 18 is to be understood in an objective, rather than subjective, fashion. This view is supported by the drafting history of this provision, which does not suggest subjective bad faith was an essential element.Footnote 73 It is also significant and ironic that, although the 1953 Lauterpacht draft provisionally introduced an element of subjectivity into the precursor of Article 18,Footnote 74 Lauterpacht himself warned, as a judge at the ICJ, that caution would be needed if international courts were to have competence to attribute bad faith to a sovereign State.Footnote 75 Ever since the Certain German Interests case, it has been clear that bad faith cannot be simply presumed, and the burden of proof rests with the party seeking to show it.Footnote 76 This means that if it were necessary to prove the intention of a State, this would render the interim obligation potentially ineffective and thus meaningless.Footnote 77 Accordingly, no proof of subjective bad faith is required to activate the interim obligation, and the standards of Article 18 are to be construed objectively, namely whether the acts in question defeat of the object and purpose of the treaty in an actual and objectively ascertainable fashion.
D. Defining the Object and Purpose of a Treaty
1. ‘Object and purpose’
The question remains what this substantive, yet at the same time vague wording means: what is, in fact, the object and purpose of a treaty, and how can a State defeat it in the context of the interim obligation of Article 18? The phrase ‘object and purpose’ is used eight times in the text of the Convention,Footnote 78 yet it remains a puzzle. Given the ILC's statement that ‘the expression “object and purpose of the treaty” has the same meaning in all of [the provisions of the Convention]’,Footnote 79 it may be helpful to refer to its draft guidelines on reservations.Footnote 80 Draft Guideline 3.1.6 provides that the object and purpose of a treaty should be determined in good faith and by considering the terms of the treaty in their context. Recourse should also be had to the title of the treaty, its travaux préparatoires, the circumstances of its conclusion, and any subsequent practice agreed upon by the parties.Footnote 81
Although a teleological approach geared towards the effectiveness of a treaty can yield tangible legal results,Footnote 82 caution is nevertheless required, for whilst the object and purpose of a treaty might be often regarded as obvious, in many cases it may be elusive.Footnote 83 In the Reservations to the Genocide Convention advisory opinion, the ICJ explained that the phrase meant ‘what is essential to the object of the Convention’ in the sense that, acting against this object, ‘the Convention would be impaired both in its principle and its application’.Footnote 84 The Court also distinguished between the ‘object’ and the ‘purpose’ as in French public law where there is a clear distinction between ‘l'objet’ of a legal instrument and ‘le but’, which denotes the reasons for establishing ‘l'objet’ in the first place.Footnote 85 It identified the objects of the Genocide Convention as being ‘to safeguard the very existence of certain human groups’ as well as ‘to confirm and endorse the most elementary principles of morality’,Footnote 86 whilst its purpose was the condemnation and punishment of genocide as an international crime which involves a denial of existence of entire human groups that shocks the conscience of humankind and results in great losses to humanity.Footnote 87 The Court nonetheless failed to explain why it understood the two words as having separate meanings, or how they are distinct from one another.Footnote 88
The problem remains that this provides neither notion with sufficient substance.Footnote 89 The general theory of the law of obligations suggests that the ‘purpose’ of a treaty refers to the ‘objective(s)’ which it attempts to achieve, and that two different types of criteria are to be considered when ascertaining the object and purpose: first, a functional criterion for determining the purpose, emphasizing the ends pursued by the treaty in question; and second, a material criterion for determining the objective, focusing on the body of norms that has to be established in order to realize that purpose. The object and purpose are therefore two complementary and interdependent elements.Footnote 90 To assess these functions and elements in practice, four tests have been proposed: (i) under the ‘essential elements test’, Article 18 enjoins States to comply with the most important or essential parts of a treaty, which constitute its object and purpose.Footnote 91 The disadvantage of this test is, however, that it offers no objective method for determining those parts of a treaty which are ‘essential’.Footnote 92 (ii) The ‘impossible performance test’ prohibits States from taking action that would render a treaty's subsequent performance impossible.Footnote 93 The problem with this test is that it does not distinguish between bilateral and multilateral treaties. It is impossible to apply this test to multilateral treaties in a meaningful and successful way, and thus it offers only little restraint on States in such a context.Footnote 94 (iii) The ‘bad faith and manifest intent test’ considers whether a State's action seems unwarranted and condemnable, regardless of actual proof of bad faith.Footnote 95 Although this test relies on objective evidence and tries to avoid delving into the subjective intent of a State, the difficulty of determining which actions manifest or demonstrate bad faith remains.Footnote 96 (iv) The ‘facilitation test’ preserves the status quo before signature, but in one direction only: States may move toward eventual compliance with the treaty's object, but not away from it. The problems with this test are that it could create double standards if one of the treaty signatories has already moved further towards achieving the treaty's objective; and it could create perverse incentives for States to step back from achieving the treaty objective before signature, in order to allow it, de facto, to maintain the status quo ante afterwards.Footnote 97
As an alternative, a two-step objective method has been proposed which relies on a combination of contextual, historical, and teleological interpretation. In the first step, one seeks to ascertain the object and purpose by having recourse to the title,Footnote 98 the preamble,Footnote 99 the programmatic provisions,Footnote 100 and the travaux préparatoires Footnote 101 of the treaty. In the second step, this is then tested against the text of the treaty and other available material and adjusted accordingly, if necessary. The result of that process can subsequently be used as a guide when interpreting other treaty provisions or for assessing compliance with them.Footnote 102 In practice, one can also ascertain the object and purpose of a treaty by referring to other agreements. In international environmental law, for instance, a State which has ratified a framework convention would be obliged under Article 18 not to defeat the object and purpose not only of that convention itself, but also of the protocols and other instruments associated with it, as the latter reinforce and guarantee the continuity of the object and purpose arising from that framework convention.Footnote 103
2. ‘Defeating’ the object and purpose
Another phrase which requires further interpretation is the interim obligation of States ‘to refrain from acts which would defeat’ the object and purpose of a treaty. Interestingly, the word ‘defeat’ is not used elsewhere in the Vienna Convention and it is therefore not clear what elements are to be taken into account in determining whether this threshold has been met or not.Footnote 104 Even though the language used is emphatically cast in negative terms, Article 18 sends mixed signals by employing the strong word ‘defeat’ in the same sentence as the much weaker phrase ‘would refrain’.Footnote 105 The consequence of this particular wording is that not every deviation from the provisions of a treaty, pending its ratification or entry into force, will automatically result in its object and purpose being defeated—if that were the case, the treaty would de facto already enter into force upon signature.Footnote 106 The threshold for violating the interim obligation is consequently much higher than that for violating the treaty itself. However, ‘defeating’ the object and purpose requires action of a much more severe nature than actions that are merely ‘incompatible’ with the object and purpose, as Article 19(c) VCLT requires.Footnote 107
The threshold of ‘defeat’ will be met if the performance of the treaty (or one of its provisions) is rendered meaningless and loses its object.Footnote 108 Alternatively, one could presume that the other party or parties would not have concluded the treaty under the same conditions, had they known that such acts would be undertaken.Footnote 109 Two examples might help elucidate further this phrase: First, the object and purpose of a treaty is defeated if a State signs a treaty concerning the return of works of art previously taken from the territory of another State, and then actively destroys these works of art or allows them to be destroyed.Footnote 110 Second, the object and purpose is also defeated if a State, having signed a disarmament treaty obliging the parties to reduce their armies by one-third, does not maintain the status quo and increases its army during the period before ratification so that even if it then does reduce its size by a third, there has been no real reduction at all.Footnote 111 What becomes evident is that there is a clear parallel between these situations covered by Article 18 and situations in which the unilateral conduct of a State party to a treaty permits the other parties to suspend or terminate the treaty. The first example finds a clear parallel in Article 61 VCLT (termination of a treaty for supervening impossibility of performance), whilst the second example evokes the situation of Article 62 VCLT (fundamental change of circumstances). Thus, it may be argued that the criteria laid down in these two provisions may also indicate whether a certain form of conduct amounts to an act defeating the object and purpose of a treaty for the purposes of Article 18.Footnote 112 Perhaps an even more instructive comparison can be drawn from EU law and the Member States’ obligation to transpose directives within a given time frame. In a very similar fashion to signed treaties, Member States are under no obligation to adopt the relevant measures before the transposition deadline.Footnote 113 However, as the Court of Justice of the EU has clarified, they must nonetheless ‘refrain from taking any measures liable seriously to compromise the result prescribed’.Footnote 114 Such a detrimental effect can be avoided if the Member States take those measures that are ‘necessary to ensure that the result prescribed by the directive is achieved at the end of that period’.Footnote 115 Although the Court does not consider this obligation to be a rigid ‘stand-still’ obligation,Footnote 116 as Member States retain the right to transpose directives in stages,Footnote 117 this effectively amounts to a ‘ratcheting-up’ and moving towards the desired object.
This finding is also crucial in terms of understanding the word ‘refrain’, which suggests that compliance with the interim obligation simply requires passive conduct on part of the States bound by it.Footnote 118 Yet, in certain cases it is possible that a treaty requires active conduct in order to prevent it from becoming meaningless, for example, if a State has promised to deliver products from a forest or mine. In such cases, the State would certainly be obliged to maintain this forest or mine in order not to jeopardize the production and delivery of the promised goods.Footnote 119 It therefore appears that the exact meaning of the obligation to ‘refrain’ can only be ascertained in each individual case and in the light of the commitments made under the treaty in question.Footnote 120
III. Article 18 and the Normative System of International Law
The question, which has not so far been analysed in relation to Article 18, is whether the legal consequences of its application are the same in relation to all norms of international law, ie whether the ‘soft’ consequences of Article 18 should be the same or whether they should perhaps be applied differently in the context of norms of a higher order, ie those of a jus cogens, erga omnes, and erga omnes partes nature. Article 18—as drafted—does not distinguish between these norms of international law. However, this is not an entirely academic issue but also might have practical consequences. In the realm of the law of treaties, this question is best illustrated by the ICJ's Reservations to the Genocide Convention advisory opinion. The Court stated that ‘[t]he first consequence arising from this conception is that the principles underlying the Convention are principles which are recognized by civilized nations as binding on States, even without any conventional obligation’.Footnote 121 In addition, the Court held:
In such a convention the contracting States do not have any interests of their own; they merely have, one and all, a common interest, namely, the accomplishment of those high purposes which are the raison d'être of the convention. Consequently, in a convention of this type one cannot speak of individual advantages or disadvantages to States, or of the maintenance of a perfect contractual balance between rights and duties. The high ideals which inspired the Convention provide, by virtue of the common will of the parties, the foundation and measure of all its provisions.Footnote 122
The Court also referred to the ‘special character’ of the Convention, which was adopted ‘for a purely humanitarian and civilising purpose’Footnote 123 and thereby introduced a distinction between so-called contractual (traités contrats) and normative conventions (traités lois). As already discussed above,Footnote 124 the main distinction between the Genocide Convention and other treaties between States concerned its object and purpose. This was not to seek a perfect balance of interest between States, but—as a humanitarian convention—to provide, ‘by virtue of the common will of the parties, the foundation and measure of all its provisions’.Footnote 125 Accepting or rejecting reservations rests upon the decision of individual States, and therefore in effect reducing the law-making character of the Convention to bilateral and reciprocal relations between States. There is no independent body to decide on the admissibility of reservations, which would reflect the normative nature of the Convention.Footnote 126
It is also noteworthy that certain provisions of the Vienna Convention, without expressly stating this, are applied to treaties commonly known as traités lois. The structure and the legal consequences of the application of these provisions to such treaties differ from their application to traités contrats. The best example of this distinction is Article 60 VCLT (‘material breach)’ and reflects, to a degree, the classification of the norms of the law of treaties presented by Gerald Fitzmaurice in his reports when serving as Special Rapporteur of the International Law Commission during its work on the VCLT (see below).
Does a different type of treaty norms call for a variable application of Article 18?Footnote 127 In general, the typology of treaties is a very complex matter. Without embarking on a detailed discussion of the correct typology of treaties, it should be noted that treaties can contain mixed obligations, and this does not sit easily with the practicalities of the law of treaties. The general character of a treaty is defined by its object and purpose (such as the Genocide Convention). Therefore, law-making treaties should be identified as such in accordance with their general object and purpose, and if a nuanced approach to the application of Article 18 is warranted in relation to these treaties, the ‘object and purpose’ should be a guiding test, and not result in a word-by-word analysis of the treaty's provisions. Admittedly, a test based on the object and purpose is certainly not perfect due to its inherently vague character. However, the process of interpretation of treaty provisions can reduce its vagueness and the ‘object and purpose’ test is a constructive way forward when applying Article 18 to different types of treaties. There is no doubt that human rights treaties are the foremost candidates for such law-making treaties, giving effect to obligations erga omnes partes. This was confirmed by the Inter-American Court of Human Rights in its Advisory Opinion Regarding Reservations to the American Convention on Human Rights, in which the Court emphasized the special legal order established by human rights treaties to which States submit themselves, ‘towards all individuals within their jurisdiction’, but not towards other States.Footnote 128 The UN Human Rights Committee (HRC) argued in its famous, but equally divisive and controversial, General Comment 24 that the VCLT provisions are not appropriate for human rights treaties, since they do not ‘form a web of inter-State exchanges of mutual obligations, but rather confer individual rights’.Footnote 129 Brölmann, however, notes that Article 60(5) VCLT extends the exception of non-reciprocity to humanitarian treaties.Footnote 130
The question of the Vienna Convention's compatibility with human rights treaties, which arguably have a special character, is not new. Many scholars express the view that, generally, the application of the Vienna Convention to human rights treaties is irreconcilable with their special character, as the Convention embodies ‘an empty, amoral world where States have reciprocal dealings only with other States, where there are no people hurt by States’ actions, and demanding reparations, no international institutions creating special mechanisms peopled by experts monitoring and reporting and no non-governmental organisations […] demanding accountability’.Footnote 131 The question of reservations to human rights treaties is the most compelling example of such a debate and was raised both by the ILC and in the Reports of Alain Pellet during their work on reservations to treaties.Footnote 132 According to Gerald Fitzmaurice, such treaties, being of a normative character, embody ‘integral obligations’ (that is, they have to be performed as such and in their entirety), and they establish a regime ‘towards all the world rather than towards particular parties', as contrasted with contractual treaties, which are based on reciprocity of the parties to the treaty.Footnote 133 The treaties characterized by interdependent obligations were of such a nature as to make the performance of one party dependent on that of all the other parties (such as a disarmament treaty).
However, the Vienna Convention, in its final form, has not followed the typology presented by Fitzmaurice, and hence the ILC decided to concentrate on the form of the treaty, and not on its substance, with the exception of Article 60(5) the VCLT.Footnote 134 Joost Pauwelyn also mentions Articles 41(1)(b)(i), 53, 58(1)(b)(i), 60(2), and 64 VCLT as being traces of the concept of interdependent treaties.Footnote 135 For example, Articles 41 and 58 do not go as far as Fitzmaurice did, and only prohibit, but do not invalidate, the inter se modification or suspension of multilateral treaty obligations that are, essentially, of a collective nature, ie binding erga omnes partes.Footnote 136
As explained above, the Vienna Convention was drafted to deal with reciprocal obligations of States and did not reflect the typology proposed by Fitzmaurice. The application of its system of reservations to human rights treaties has been subject to robust debate, it being argued that it is not compatible with the special nature of human rights treaties. Therefore, a similar question arises in concerning the usefulness of Article 18 in the context of normative treaties which are not based on a contractual quid quo pro as it too would be contrary to the erga omnes character of such treaties: as with reservations to normative treaties, the application of Article 18 to bilateral relations would be against the spirit and character of traités lois.
The view in the scholarly literature, in general, is that Article 18 is not suitable for normative treaties.Footnote 137 Jan Klabbers suggests that ‘[i]nstead of defeating the object and purpose of a law-making convention, any behaviour irreconcilable with it, prior to its entry into force, actually serves to emphasize the desirability of its entry into force. The behaviour, rather, strengthens the very point of the treaty.’Footnote 138 He also argues that in order to be applicable to human rights treaties, Article 18 should be reconceptualized. It should also be made subject to a test in order to determine whether the interim obligation is violated, which he calls a ‘manifest intent test’.Footnote 139 Such a test is based on the general law of treaties, which he considers in principle suitable to deal with neither community interests nor with contractual relations.
Klabbers explains that the ‘manifest intent test is based on the following premise:
[I]f behaviour seems unwarranted and condemnable, it may be assumed to have been inspired by less than lofty motivations and ought to be condemned, regardless of whether anyone's legitimate expectations are really frustrated or can reasonably be said to have been frustrated, regardless of actual proof of bad faith.Footnote 140
Such a test, according to Klabbers, is abstract and, due to its vagueness, it would be problematic, if applied in judicial proceedings.
A solution can be drawn from the practice of the reservations to treaties. Although, the ILC decided not to apply a different regime to reservations to treaties, these treaties were seen as ‘containing numerous interdependent rights and obligations’.Footnote 141 As Christine Chinkin observes, these Guidelines have implicitly highlighted that what distinguishes human rights treaties from others is their establishing expert bodies and the acceptance that such bodies may need to assess the permissibility of reservations in order to be able to function properly. For example, the monitoring body of the 1979 Convention on the Elimination of All Forms of Discrimination against Women has concluded that certain reservations were impermissible as being contrary to the object and purpose of the Convention. It has not, however, suggested that the offending reservation be severed,Footnote 142 as has been done by the Human Rights Committee and the European Court of Human Rights and which has resulted adverse reactions from some States.Footnote 143 However, the Pellet GuideFootnote 144 makes notable progress in the assessment of the permissibility of reservations which has attempted to free human rights treaties from the inherent bilateralism of treaties and to situate them in an objective, but also consensualist framework.Footnote 145 The human rights treaty monitoring bodies and, generally, the highest organs of other nominative treaties could be empowered to decide on the conduct of a State for the purposes of Article 18 and condemn an action which was considered to be contrary to the treaty's object and purpose. This may appear to be a modest suggestion. However, in general, States do not accept easily public naming and shaming. Such an approach could only be applicable to treaties that have already entered into force (as it presupposes that the monitoring body has been established), and it would therefore have only a quite limited scope of application. Practically speaking, it could only be applicable to acceding States.
Such a modest, but balanced, proposition would reflect the special character of human rights treaties. Admittedly, there are certainly difficulties concerning the definition of what actually are human right treaties, and that category of treaties is far from homogenous.Footnote 146 However, as Chinkin stated, relying on the analysis of the Inter-American Court of Human Rights, human rights treaties have two essential elements: first, all of these treaties have a horizontal effect regulating inter-State behaviour; and second, they constitute a framework enabling States to make binding unilateral commitments not to violate the human rights of individuals within their jurisdiction. Thus, ‘a human rights treaty represents a vertical relationship, a governmental pledge and limit to governmental powers’.Footnote 147 Even the most staunch supporter of the homogeneity of treaties has to admit that ‘traditionally international law had no stake in the substance of treaties, but this is changing and human rights law is at the forefront of this’.Footnote 148 Therefore, the traditional understanding of the law of treaties has to be adjusted and reformulated to fit these new developments, including Article 18. Otherwise, the Vienna Convention on the Law of Treaties will lose its importance and currency.
IV. Conclusion
This article has endeavoured to resolve certain difficulties in ascertaining and assessing the legal character of Article 18 VCLT, in particular its opaque nature which is evidenced by the hesitation of the ILC in determining whether it is a moral or legal obligation. The conclusion is that the interim obligation in Article 18 is not a moral, but is, in fact, a legally binding obligation. However, the normative force of that legal obligation is weakened as a result of its succinct wording. One of these problems is the short life span of the interim obligation codified in Article 18, unlike the more enduring obligation of pacta sunt servanda. This interim obligation can assume two forms: one before and the other after consent to be bound has been expressed. In the first scenario, the interim obligation ceases once a State ratified the treaty or made clear its intention not to become a party to the treaty at all. In certain circumstances, this may lead to the so-called ‘unsigning’ of a treaty and which may be justified under Article 18; at the same time, it may defeat the object and purpose of the treaty and thus violate the interim obligation. Such different interpretations of a seemingly simple act further exemplify the nebulous character of this provision.
The second possible application of Article 18 is after the consent to be bound has been expressed, but before the entry into force of the treaty. This part of the provision introduces the element of ‘undue delay which, whilst to some degree objectivises this procedure, is not without controversy concerning what the time-limits involved. For example, it has been argued that such ‘undue delay’ does not exist in relation to the 1993 Nuclear Test Ban Treaty. Furthermore, this does not permit the ‘unratification’ of a treaty, as contrasted with the phenomenon of ‘unsigning’ – although there is State practice concerning the withdrawal of instruments of ratification not encountering any subsequent legal or political difficulties.
Lastly, whilst one can argue that the reference to object and purpose in Article 18 constitutes its essential core, it without doubt remains one of the vaguest and most ill-defined tests in the Vienna Convention, both as regards its meaning and whether it is subjective or objective in character. Therefore, it does little to unravel the meaning of this provision. This problem becomes acute when seeking to apply Article 18 to normative treaties, as opposed to contractual treaties, with a special emphasis on human rights treaties. The special character of human rights treaties was already raised during the work of the ILC concerning reservations to treaties. The majority of the provisions of the Vienna Convention were drafted with a view to their application to contractual treaties. We agree with Chinkin that the application of Article 18 in the context of human rights treaties should be different and left to the discretion of monitoring bodies which they create.
At a minimum, however, the fact that the interim obligation in Article 18 is based on good faith can help make more precise this otherwise vague provision. This is particularly the case if good faith is understood as an objective criterion, as has been argued above that it should be.
