Jean Monnet Center at NYU School of Law



Previous|Title|Next

II. WTO obligations are continuing obligations

A. Interpreting the WTO treaty: "contemporaneous" or "evolutionary" interpretation?

1. The distinction explained

The question of contemporaneous versus evolutionary interpretation arises "whenever a rule refers to a concept the scope or significance of which has changed in the course of time".83 Such change may constitute an evolution in the law or an evolution in the way contracting parties to a treaty understand a concept (without such new understanding being reflected in new law). The words of Art. 31.3(c) VCLT leave this question open. Art. 31.3(c) requires that a treaty interpreter takes into account, together with the context, "any relevant rules of international law applicable in the relations between the parties". It does not limit this reference to other rules of international law to those that existed at the time of conclusion of the treaty, nor does it explicitly include other law that had developed at the time of application.84

The principle of contemporaneity requires that a juridical fact be appreciated in the light of understanding and law contemporary with it.85 For example, when the ICJ interpreted the provisions of Article 20 of the treaties between the United States and Morocco of 1787 and 1836 (substantially identical in terms) in the United States Nationals in Morocco case, it stated that "it is necessary to take into account the meaning of the word `dispute' at the times when the two treaties were concluded".86 This principle of contemporaneity is generally seen as the rule.87 In other words, the rule is that in interpreting terms in, for example, the WTO treaty, reference is to be had to their meaning in international law in April 1994, i.e., the time WTO agreements were concluded. On that basis, it would seem that the reference to the Berne, Paris and Rome Conventions in the TRIPS Agreement88 is a reference to these agreements as they stood in April 1994, i.e., at the time of conclusion of the WTO treaty.89 The incorporation of these WIPO conventions into the WTO treaty would only be a dynamic one -- automatically incorporating also modifications brought to these conventions in WIPO -- in case there were a clear indication on behalf of the drafters of the TRIPS agreement to that effect.90

There is, however, scope for exception to this principle of contemporaneity. Indeed, already in the preparation of the VCLT itself, it was realised that

"the content of a word, e.g. `bay' or `territorial waters', may change with the evolution of the law if the parties used it in the treaty as a general concept and not as a word of fixed content".91

This reflects the so-called "evolutionary approach" to treaty interpretation. This approach has been followed in three subsequent ICJ judgements, interpreting, for example, the term "territorial status".92 In those cases, the ICJ referred to the need to interpret the treaty "within the framework of the entire legal system prevailing at the time of interpretation".93

To answer the question as to which interpretative method must be applied (contemporary or evolutionary approach), the intention of the parties to the treaty must be decisive.94

2. The distinction applied to the WTO treaty

Where does this leave us when it comes to interpreting terms used in the WTO treaty? One Appellate Body stands out in this respect. In its report on US - Shrimp/Turtle the Appellate Body had to decide whether the terms "exhaustible natural resources" in GATT Art. XX(g) included only living resources or also non-living resources (in casu, certain sea turtles). It found as follows:

"The words of Article XX(g), `exhaustible natural resources', were actually crafted more than 50 years ago. They must be read by a treaty interpreter in the light of contemporary concerns of the community of nations about the protection and conservation of the environment".95

In support of this evolutionary interpretation, the Appellate Body referred to the preamble attached to the 1994 WTO Agreement which mentions "the objective of sustainable development". It continued as follows:

"From the perspective embodied in the preamble of the WTO Agreement, we note that the generic term `natural resources' in Article XX(g) is not `static' in its content or reference but is rather `by definition, evolutionary'. It is, therefore, pertinent to note that modern international conventions and declarations make frequent references to natural resources as embracing both living and non-living resources".96

Based on the preamble to the WTO Agreement and "the recent acknowledgement by the international community of the importance of concerted bilateral or multilateral action to protect living natural resources", the Appellate Body concluded that "it is too late in the day to suppose that Article XX(g) of the GATT 1994 may be read as referring only to the conservation of exhaustible mineral or other non-living natural resources".97

Are the terms in GATT Art. XX(g) special, or should this evolutionary approach be applied to the interpretation of most or all WTO provisions? As noted earlier, this will depend on the particular provision and the intentions expressed therein by the drafters. Hence, much like the bilateral/multilateral distinction discussed above, one can rarely say that an entire treaty must be interpreted in an evolutionary way. Each provision must be examined on its own merits. Nonetheless, three arguments can be made in support of evolutionary interpretation becoming the rule, rather than the exception, when it comes to interpreting the WTO treaty.

Firstly, WTO obligations are increasingly of a regulatory nature. As compared to the GATT (based largely on country-specific tariff concessions), the character and underlying foundation of most WTO rules has changed. WTO agreements such as the SPS or TRIPS agreement98 are unrelated to the idea of balancing trade concessions that are (like in GATT or GATS99) additional to the rules set out in the agreement itself. Today, most WTO rules set out general standards and conditions for state conduct, which are binding not as part of some ephemeral balance of concessions, but as a result of their inherent and indefinite juridical force. In that sense, the WTO treaty is increasingly a law-making treaty or traité-loi. One of the features of general regulatory provisions is that they use broad, unspecified terms, such as "exhaustible natural resources", "public morals" or "essential security interests" in GATT Articles XX and XXI. As the Appellate Body found in US - Shrimp/Turtle this is an indication that evolutionary interpretation was intended.100 It is, indeed, an indication that WTO Members wanted these terms to evolve with society and international law or, at least, should have realised that the vagueness of these terms would result in their meaning being open to discussion and variation depending on the context and times.101

In this context, interpreting most WTO provisions in an evolutionary manner seems both logical and necessary.102 The same attitude was taken by the European Court of Human Rights when interpreting the European Convention on Human Rights, another "regulatory treaty". In the Tyrer case, for example, it found as follows:

"The Court must also recall that the Convention is a living instrument which ... must be interpreted in the light of present-day conditions ... the Court cannot but be influenced by the developments and commonly accepted standards in the penal policy of the member States of the Council of Europe in this field".103

Secondly, pursuant to Art. 31.1 VCLT, the terms of the WTO treaty must be interpreted "in their context and in the light of its object and purpose".104 When it comes to regulatory treaties, this object and purpose soon leads a life that is largely independent and de-linked from the subjective and temporal intentions of the treaty's original drafters. It becomes an objective, dynamic regime where "the whole is greater than the sum of the parts".105 As it was noted by a 1986 arbitral tribunal in a dispute between Canada and France:

"Dans les traités normatifs ou institutionels conclus entre un nombre élevé d'Etats et pour une durée indéterminée, le but du traité se détache aisément des finalités poursuivies par chacun des contractants primitifs et acquiert une autonomie objective que l'interprète ne peut méconnaître ».106

Consequently, when interpreting WTO provisions account must be had of the WTO's object and purpose as it may evolve over time.

Thirdly, most WTO provisions can be said to be of a "continuing nature" in the sense that WTO members regularly renew their consent to them by means of implementing, applying, monitoring and reconfirming these provisions in the context of the WTO as an international organisation. As noted by Simon in respect of treaties establishing international organisations:

"l'accord de volontés qui a présidé à la conclusion de la convention ne s'est pas épuisé dans la rédaction d'un texte; l'application d'une telle convention suppose nécessairement le renouvellement permanent de l'adhésion des Etats membres au contenu de normes juridiques dont l'instrument signé ne constitue qu'une expression solonnelle, mais, par essence, éphémère. Or la fonction de l'interprète est ... de donner effet au consensus continu des parties plutôt qu'à des volontés cristallisées à la date de signature ... ou, si l'on préfère, de reconnaître dans les traités constitutifs la présence d' `effectivités en action ` qui imposent une interprétation `évolutive' des intentions des fondateurs de l'organisation".107

The fact that WTO members could thus be said to "continuously" re-confirm their consent to the WTO treaty pleads in favour of interpreting this treaty in an evolutionary manner, that is, in line with understandings and law at the time of its application.

B. Putting a time-label on the WTO treaty in case it conflicts with another treaty

The third and last reason above in support of interpreting the WTO treaty in an evolutionary manner (WTO provisions can be said to be of a "continuing nature") provides an appropriate bridge to the other question of timing we set out to examine. That is, the question of putting a time-label on the WTO treaty in case it conflicts with another treaty so as to apply the lex posterior rule (pursuant to which the treaty later in time prevails).

Art. 30.3 and 4 VCLT provide for the following conflict rules in respect of "Application of successive treaties relating to the same subject-matter":

3. When all the parties to the earlier treaty are parties also to the later treaty but the earlier treaty is not terminated or suspended in operation under article 59, the earlier treaty applies only to the extent that its provisions are compatible with those of the later treaty.

4. When the parties to the later treaty do not include all the parties to the earlier one:

(a) as between States Parties to both treaties the same rule applies as in paragraph 3;

(a) as between a State Party to both treaties and a State Party to only one of the treaties, the treaty to which both States are parties governs their mutual rights and obligations.

For present purposes, we shall examine only the question of timing, that is, the terms "successive treaties", "earlier treaty" and "later treaty", as referred to in Art. 30 VCLT.

1. Treaties as instruments with a time-label

The focus of the VCLT is treaties as an instrument, not obligations deriving from treaties.108 Art. 30 VCLT as well deals with the application of "successive treaties" and "an earlier treaty" as opposed to "a later treaty". This seems to imply that the timing of treaties under Art. 30 is a question of putting a date on the treaty as an abstract instrument, not a question of defining when the treaty imposes a particular obligation as between two given states.

This approach of putting a time-label on treaties as instruments makes sense in case one is faced with a treaty that is clearly concluded in order to amend an earlier one and where the parties to both treaties are exactly the same. One may think here of a 1990 bilateral investment treaty between A and B which is subsequently amended by a later 2000 treaty as between A and B; or a 1994 WTO agreement which is subsequently amended by a later 2000 treaty as between all WTO members, pursuant to Art. X of the Marrakesh Agreement.

In those instances, the domestic law analogy with "legislative intent" and "the legislator" being able to change earlier legislation by later legislation, may make sense. In those cases, one is, indeed, faced with one homogenous bloc of states, acting, arguably, as some kind of legislator in a particular field (albeit legislation applicable only as between the parties to the treaty).

2. The fiction of "legislative intent" unfolded

However, the fiction of later "legislative intent" overruling another, earlier one, looses its attraction as soon as the "same context - same parties" constellation changes. Indeed, even in respect of two AB treaties, it becomes hard to refer to one and the same "legislature" in case the earlier bilateral treaty was concluded in the context of a multilateral environmental agreement (MEA) and the later one in the context of, for example, the WTO. A state's consent must be seen as one and indivisible, irrespective of who negotiated the treaty, but the reality remains that in the context of an MEA a very different set of people and values are at work than those active in, for example, the WTO context. Hence, the position of one state in one context may, indeed, be diametrically opposed to that same state's position in another context.109

But it is not only the difference in context that may make the analogy with "legislative intent" unworkable. In many instances -- and especially in respect of the great regulatory treaties of modern times, such as certain MEAs, UNCLOS and the WTO -- the two treaties will have a different membership. The two treaties are then clearly the result of a different "legislature", i.e., a different composition of states. They will hence be the reflection of a different balance of interests and one state may well have been able to push through its interests more under one treaty than under another. This difference will be accentuated in case a norm in one treaty context is adopted by unanimity and a norm in another treaty context is adopted by majority voting, even as between states that are parties to both treaties. In that case, it may well be that one of the parties accepted the first norm but explicitly voted against adoption of the second norm. Such objection cannot mean that the state in question is not bound by the second norm since it agreed to the majority voting procedure in the first place. But it would make the comparison with one and the same "legislator" simply changing its mind over time more difficult.

The above considerations highlight the diversity as between different treaties and the difficulty of making a domestic law analogy based on changing "legislative intent". Nonetheless, these difficulties must, as a general rule, be accepted as a reality of international law. Adopting a lex posterior rule in this context may, indeed, provide a strong incentive for states to streamline their positions across international organisations and irrespective of the membership to particular treaties. States must realise that, in principle, whatever they consent to now prevails over what they agreed on earlier, irrespective of the context in which the obligations were entered into.

Hence, the lex posterior rule may be an important instrument that creates some order in the chaos of international law. However, given its shaky foundation of changing "legislative intent"", it cannot be seen as an absolute rule the way it is regarded, for example, in domestic law for inter-acting statutes. Exceptions to it must be allowed for in case the analogy with "latest legislative intent" looses touch with reality.

3. The difficulty of putting a time-label on a treaty as "instrument"

The lex posterior rule in Art. 30 may not only be put in doubt because of the shaky analogy it makes with changing "legislative intent". From a more practical point of view, it will in many cases also be difficult to "put a time-label" on a treaty.

To begin with, it is generally accepted that the timing of a treaty for purposes of Art. 30 is determined by the date of its conclusion or adoption. In case of the WTO treaty, for example, this is 15 April 1994. It is not the date of entry into force that determines the timing of a treaty under Art. 30. This was made explicit by the Expert Consultant at the Vienna Conference:

"for purposes of determining which of two treaties was the later one, the relevant date should be that of the adoption of the treaty and not that of its entry into force. His own understanding of the intentions of the [ILC] confirmed that assumption".110

The fact that the date of conclusion is decisive was explained with reference to the concept of "legislative intent" elaborated earlier: "when the second treaty was adopted, there was a new legislative intention; that intention, as expressed in the later instrument, should therefore be taken as intended to prevail over the intention expressed in the earlier instrument. That being so, it was inevitable that the date of the adoption should be the relevant one".111 That the date of conclusion counts to define a "later treaty" under Art. 30 finds support also in Art. 59 where reference is made to "conclusion of a later treaty" and "conclude a later treaty".112

Hence, it is not the difficulty of deciding between the date of conclusion, opening for signature, ratification or entry into force that makes it difficult to put a time-label on treaties. What counts is the date of conclusion, irrespective of the fact that the treaty may have been ratified by, or entered into force for, different parties at different times. Utilising the date of conclusion does, indeed, make logical sense when faced with a conflict between two treaties to which no parties acceded subsequently. For example, as between a 1990 bilateral investment treaty between A and B and a subsequent 2000 bilateral investment treaty between A and B.

The situation is more complicated in the event of treaties to which additional states have acceded. This is the case in respect of all regulatory treaties with a universal calling (such as the WTO treaty, UNCLOS, most MEAs and most human rights treaties) as well as many regional arrangements (such as the European Union, ECHR or NAFTA). The underlying objective of these treaties is that an increasing number of states accede to it (with or without regional restrictions). When faced with such "expanding" treaties, it becomes often untenable to stick to the date of original conclusion of the treaty. This problem was already pointed out by Sir Ian Sinclair at the Vienna Conference itself:

"supposing a multilateral convention was opened for signature in 1960, State A ratified it in 1961, and the convention entered into force in 1962. Then State A and State B concluded a bilateral treaty on the same subject in 1963 which entered into force in 1964, after which State B acceded to the multilateral convention in 1965. Which of the treaties was the earlier and which was the later? In State A's view, the multilateral convention was the earlier [1960] but in State B's view it was the later [1965]".113

The fact that the parties to a treaty did not "conclude" the treaty at the same point in time makes it impossible to put a single time-label on the treaty in question. It necessitates a shift away from the treaty as abstract instrument to an assessment of the treaty as source of rights and obligations resting on particular states.114 This fact is, moreover, a death blow to the fiction of each treaty being concluded by one and the same "legislative intent", expressed at one point in time. This fiction must, in turn, be brought back to a genuine principle of international law, namely the contractual freedom of states. We next examine Art. 30 from this perspective of treaty provisions binding on particular states and contractual freedom as between the two states in question.

4. Treaties as a source of rights and obligations as between particular parties

Although the crucial date under Art. 30 is the date of conclusion of the treaty, Art. 30 can be activated only as between two parties for which the treaty has entered into force. If not, there could not be conflict.115 However, once the treaty has entered into force for the two states in question, i.e., once a conflict may arise, in order to put a time-label on the treaty it is not the date of entry into force that counts -- that date that may well be different for the parties involved -- but the date of conclusion of the treaty.

But this leaves us with the situation referred to by Sinclair -- that is, the situation that prompted us to examine Art. 30 in terms of rights and obligations resting on particular states -- namely: treaty 1 to which state A is an original member and state B acceded at a later point in time, in conflict with a bilateral treaty 2 that was concluded by A and B in between these two points of time. One may think here of a conflict between a WTO rule and a provision in a bilateral treaty concluded in 2000 as between two parties, one of which is an original WTO member (1994), the other being a state that acceded to the WTO only in 2001. For the original (WTO) member (state A) treaty 1 (the WTO treaty) is "concluded" at the time of the treaty's original conclusion (15 April 1994). Hence, for A treaty 1 is the "earlier treaty". In contrast, for the acceding member (state B), treaty 1 (the WTO treaty) is "concluded" at the time its accession was adopted (in casu, 2001).116 Hence, for B the same treaty 1 is the "later treaty". The same situation would arise in case A and B are original WTO members, A concludes and becomes bound by a WTO amendment in 2000, there where B only does so in 2002 and in 2001 A and B conclude a treaty in conflict with the WTO amendment. For A, the 2001 treaty is the later in time; for B the WTO amendment is the later in time. In those cases, as Vierdag pointed out,

"[p]aradoxically, as a result of the lex posterior rule laid down in paragraph 3 of Article 30, a treaty [or amendment] to which a State was quick to become a party will be set aside by an incompatible treaty to which it became a party at a later date, perhaps reluctantly. For States that were slow in adhering to a treaty [or amendment] the effect of paragraph 3 will be that this treaty will supersede an incompatible treaty they were quick to enter into at an earlier date".117

At this juncture, two approaches are possible. Firstly, in application of Art. 30 one could search for the time of "convergence of state consent" in respect of the particular treaty provision. Secondly, one could dis-apply Art. 30 all together on the ground that the two treaties are not "successive". We next deal with these two alternatives in turn.

5. Look for the "time of convergence of state consent" in respect of the treaty provision concerned

Firstly, one could submit that in circumstances where for one state the treaty is "earlier" and for the other the treaty is "later", one ought to focus, not so much on the timing of the treaty as abstract instrument, but on the date at which the consent of the two states in question converged. Hence, instead of focusing on the date of "legislative intent" underlying the treaty as abstract instrument, one would then focus on the date when the expression of contractual freedom of the particular states in question met. Under this first approach, the hypothetical conflict outlined above between the WTO treaty and a bilateral treaty concluded subsequently in 2000 but before the second state acceded to the WTO in 2001, would then be resolved in favour of the WTO rule since the consent of both parties to the WTO treaty arose subsequently to that underlying the bilateral agreement (i.e., 2001, date of accession of the second state to the WTO, as opposed to 2000, date of conclusion of the bilateral treaty). In other words, when faced with two parties for which a treaty has a different date, the latest date should then be adopted as the date reflecting the time at which both parties' consent around the treaty emerged.

In the context of a multilateral treaty in conflict with a bilateral treaty this approach may work. But if one applies it also to a similar type of conflict between two multilateral treaties, the solution may be problematic for it becomes difficult to talk of an earlier versus a later convergence of state consent. Take the example of the WTO treaty (1994) in conflict with the Carthagena Biosafety Protocol (1999). States A and B are original WTO members which adopted the Protocol in 1999. State C acceded to the WTO in 2000 after it had adopted the Protocol in 1999. State D acceded to the WTO in 2001 and subsequently adopted the Protocol. In that situation our first approach offers the following outcome:

This differential approach would thus mean that for some states, WTO rules prevail, for others the Protocol. Depending on the circumstances -- i.e., is the Protocol obligation one of an integral nature? - the WTO rule may be "illegal" pursuant to Art. 41 as an inter se deviation from an integral obligation.118 But if no integral obligations are involved, can this "balkanisation" of multilateral treaties be tolerated?

As between states that both participated in the conclusion of the two treaties (in casu, states A and B), to say that the "later treaty" prevails as between them -- even if for other states this treaty may not be the later one - seems to make sense. However, as soon as one of the treaties is the "later" in time for one party and the "earlier" in time for another party, the solution offered becomes shaky. Indeed, in terms of state practice, would states A and B (original parties to both the WTO and the Protocol) realise that when they let C accede to the WTO in 2000 in their relationship with C, WTO rules prevail; but when they subsequently let D accede to the WTO in 2001 in their relationship with D, the Protocol prevails?

6. Dis-apply Art. 30 - the notion of "continuing" or "living" treaties

This brings us to a second approach. One could submit, indeed, that in situations where for one state a conflicting treaty is the earlier one, whereas for the other state it is the later one, it is impossible to define the treaty as either "earlier" or "later" in time as required in Art. 30 (for A it is "earlier" and for B it is "later"). Consequently, the conflict in question is not one of "successive treaties". Hence, Art. 30 does not apply and one must have resort to other conflict rules (in particular, the lex specialis principle).

A strong argument in support of this second approach applies when faced with a multilateral treaty of what I term a "continuing" or "living" nature. Such multilateral treaty norms are part of a regulatory framework or legal system that was created at one point in time but continues to exist and evolve over a mostly indefinite period. Most rules of modern multilateral conventions are of this nature, including EC treaties, WTO rules, UNCLOS and many environmental conventions and human rights treaties. They are rules part of a framework or system which is continuously confirmed, implemented, adapted and expanded, for example, by means of the accession of new state parties (for which not only the consent of the new party is required, but also the reciprocal acceptance of all, or a majority of, existing parties). Such treaty norms were not only consented to when they originally emerged, but continue to be confirmed, either directly or indirectly, throughout their existence, in particular when monitored and evolving within the context of an international organisation (such as the WTO).119 It would be inconsistent with the genuine will of states to "freeze" this type of rules into the mould of time at which they were originally created and to label them as an expression of state consent limited to, say, 15 April 1994.

This type of treaty norms derive from "continuing" or "living" treaties, not reflections of a "one-shot-end-all" expression of state consent.120 As a result, when such treaty norm conflicts with another treaty norm, in particular another continuing treaty norm, the "guillotine" approach of time of conclusion (the later in time prevailing) may not make sense and lead to arbitrary solutions.

Take the example of GATT 1947 and GATT 1994, on the one hand, in conflict with an MEA concluded in 1990, on the other. The GATT 1947 was concluded in 1947 and continued to exist up to 1995. In 1994, however, the GATT 1947 was incorporated into the WTO Agreement (as part of GATT 1994) under whose umbrella also a number of new treaties were put. As a result, when the WTO Agreement was concluded in 15 April 1994, also the provisions of GATT 1947 as incorporated into the WTO Agreement, were "re-concluded". Now, as against our 1990 MEA, did this re-conclusion of GATT result in a complete tabula rasa in the sense that whereas before 1994, the MEA prevailed, as of 1994, it was again the GATT that must prevail as the"later treaty"?121

If so, the same argument could be made in respect of the recent confirmation of WTO agreements at the Doha 2001 Ministerial Conference. Indeed, the Doha Ministerial Declaration "strongly reaffirm[s] the principles and objectives set out in the Marrakesh Agreement Establishing the World Trade Organization".122 Does this mean that before Doha, agreements concluded between 1994 and 2001 prevailed over the Marrakesh Agreement, whereas after Doha, it is again the "re-concluded" Marrakesh Agreement that prevails? In my view, this approach is erroneous and the WTO treaty should rather be seen as a "continuing treaty".

Taking the time of conclusion or re-conclusion as decisive in the context of "continuing treaties", would have an even more absurd effect when it comes to regular revisions of technical treaties. Such technical treaties may need to be revised on a regular basis to keep track of technological developments. But does this mean that after each revision, the revised treaty prevails again as against treaties concluded before the revision? If so, technical treaties subject to regular revision would then, through Art. 30, always prevail over, for example, regulatory treaties that are not often revised (but may be more important).123 It is highly questionable whether this result would still be in line with the underlying rationale of Art. 30, that is, giving preference to the "latest expression of state consent".

Finally, it must be pointed out that the notion of "continuing" or "living" treaties, suggested here, is a logical consequence also of the obligation to interpret certain treaties in an "evolutionary" manner (discussed above). Evolutionary interpretation and the impossibility of putting a singly time-label on the treaty in question go hand in hand. As Fastenrath put it:

"the adoption of a strictly static interpretation is consistent only with a view that identifies law with the will of a legislator as expressed in a specific law-creating process".124

Thus, the Appellate Body's evolutionary approach to treaty interpretation must be matched with an acknowledgement that the WTO treaty is a "continuing" or "living" treaty in respect of which the lex posterior principle in Art. 30 cannot always be applied.

7. Practical consequences for conflicts involving the WTO treaty as a "continuing treaty"

It should normally be sufficient that one of the two treaty norms in conflict is of a "continuing" or "living" nature for Art. 30 to be dis-applied (i.e., for the treaty norms not to be "successive"). This would be the case, for example, when two original WTO members conclude a bilateral treaty, addressing certain trade issues, in 2000. In case of conflict between the two treaties, it would then, indeed, be difficult to say that the two treaties are "successive" in time, since WTO rules existed before the bilateral treaty was concluded and continue to be confirmed after that. The conflict should then be better solved by, for example, the lex specialis principle, according to which the bilateral treaty would most likely prevail either because of its more specific subject-matter or its more specific membership, or both.

Nonetheless, even "continuing" or "living" treaties do have a starting point, even if this starting point is not the beginning and the end of state consent to the treaty. As a result, there may be treaties that were concluded before the starting point of a continuing treaty. Consequently, there may still be conflicts involving a continuing treaty where the two treaties are "successive" and the continuing treaty prevails as later in time under Art. 30. For example, the starting point of GATT is 1947. That of new WTO agreements, 1994. Thus, in case of a conflict between a GATT 1947/WTO norm and another norm (not of a continuing nature) which clearly pre-dates the GATT 1947/WTO norm (say, a bilateral agreement concluded in, respectively, 1930 or 1980), Art. 30 should still find application given that the two treaties are then, indeed, "successive" in time.125 Indeed, although the consent to the GATT 1947/WTO norm was given in 1947/1994 and continues to be confirmed thereafter, it remains the case that the (non-continuing) consent given to the other norm pre-dates the consent to the GATT 1947/WTO norm.

However, in the event of conflict between two norms of the "continuing" or "living" type, it seems difficult to apply Art. 30 at all, except in case one norm is an explicit amendment of the other in the sense of Art. 40 VCLT.126 Even if the starting point of one of the two treaties may then pre-date that of the other, given that the norms are in conflict, at the time of the conflict both exist and continue to be confirmed at the same time. Hence, the two treaties cannot be defined as "successive" in time. They are rather "parallel" in time. Thus, Art. 30 cannot be applied and, most probably, the lex specialis principle will resolve the conflict.


83 Preamble 4 to the Resolution of the Institute of International Law of 1 September 1995 on "Problems arising from a succession of codification conventions on a particular subject", 66 YEARBOOK OF THE INSTITUTE OF INTERNATIONAL LAW, PART I, 245-248 (1995), hereafter "IIL Resolution 1995".

84 Note, however, that Art. 31.3 (a) and (b) explicitly include « subsequent agreement » and « subsequent practice » as between the parties to a treaty regarding the application or interpretation of that treaty.

85 This was first expressed by Judge Huber in the Island of Palmas arbitration (1928, Netherlands v. United States), 2 R. Int'l Arb. Awards 831, 845.

86 ICJ Reports 1951, 189.

87 That such contemporaneous interpretation remains, in principle, the rule, even today see Guinea-Bissau/Senegal Maritime Boundary Arbitration (Award of 31 July 1989, 83 ILR 1, at para. 85 (1990):

"the 1960 Agreement must be interpreted in the light of the law in force at the date of its conclusion. It is a well established general principle that a legal event must be assessed in the light of the law in force at the time of its occurrence and the application of that aspect of intertemporal law to cases such as the present one is confirmed by case-law in the realm of the law of the sea".

See also the Individual Opinion of Judge Bedjaoui, in the Case Concerning the Gabcikovo-Nagymaros Project (Hungary v. Slovakia), ICJ Reports 1997 (in his view evolutionary interpretation is then the exception, if proof of intention to that effect can be pointed at).

88 Footnote 2 in the TRIPS Agreement.

89 See, in contrast, the final report of the Arbitration Panel Established Pursuant to Article 2008 of the North-American Free-Trade Agreement (NAFTA), in the matter of Tariffs Applied by Canada to Certain U.-S.-Origin Agricultural Products, 2 December 1996, where it was found that the use of the term "GATT" in the cross-reference provisions of the FTA and NAFTA had to be interpreted to mean GATT as it evolved into the WTO agreement".

90 Such intention to refer to a non-WTO agreement as it evolves over time can, for example, be found in item (k), second paragraph, in Annex I to the Subsidies Agreement. There, reference is made to « an international undertaking on official export credits to which at least twelve original [WTO] Members ... are parties as of 1 January 1979 (or a successor undertaking which has been adopted by those original Members) » (emphasis added). In a recent panel report, this reference was found to be one to « the most recent successor undertaking which has been adopted prior to the time that the second paragraph is considered », in casu, a 1998 OECD Arrangement (Panel report on Brazil - Export Financing Programme for Aircraft, WT/DS46/RW/2, adopted on 23 August 2001, para. 5.83).

91 Observations and Proposals of the Special Rapporteur, quoted in DIETRICH RAUSCHNING, THE VIENNA CONVENTION ON THE LAW OF TREATIES, TRAVAUX PRÉPARATOIRES 244 (1978) (hereafter "Rauschning, Travaux").

92 Namibia (Legal Consequences) Advisory Opinion, ICJ Reports 1971, 31:

"the concepts embodied in Article 22 of the Covenant ... were not static, but were by definition evolutionary ... The parties to the Covenant must consequently be deemed to have accepted them as such"

See also the Aegean Sea Continental Shelf Case, ICJ Reports 1978, 3; the more recent La Bretagne Arbitration Decision (90 RGDIP 716, at para. 49 (1986); the Guinea-Bissau/Senegal Maritime Boundary Arbitration (Award of 31 July 1989, 83 ILR 1, at para. 85 (1990); and the ICJ Case Concerning the Gabcikovo-Nagymaros Project (Hungary v. Slovakia), ICJ Reports 1997.

93 Namibia (Legal Consequences) Advisory Opinion, ICJ Reports 1971, 31.

94 See Rosalyn Higgins, Some Observations on the Inter-Temporal Rule in International Law, in JERZEY MAKARCYK (ED.), THEORY OF INTERNATIONAL LAW, ESSAYS IN HONOUR OF K. SKUBISZEWSKI 173 (1999). As the Commentary to Article 31.3(c) VCLT has put it: "the relevance of international law for the interpretation of treaties in any given case was dependent on the intentions of the parties" and "would normally be indicated by interpretation of the term in good faith" (Rauschning, Travaux, p. 254). In 1975 the Institute of International Law adopted a resolution in the same sense. 1975 Resolution of the Institute of International Law, paragraph 4, first sentence: "Wherever a provision of a treaty refers to a legal or other concept without defining it, it is appropriate to have recourse to the usual methods of interpretation in order to determine whether the concept concerned is to be interpreted as understood at the time when the provision was drawn up or as understood at the time of its application" (1975 YEARBOOK OF THE INSTITUTE OF INTERNATIONAL LAW, 537, at 539).

95 Appellate Body report on United States - Import Prohibition of Certain Shrimp and Shrimp Products, WT/DS58/AB/R, adopted on 6 November 1998, para. 129.

96 Ibid., para. 130.

97 Ibid., para. 131.

98 Respectively, the Agreement on the Application of Sanitary and Phytosanitary Measures and the Agreement on Trade-related Intellectual Property Rights.

99 General Agreement on Trade in Services.

100 See Tetsuo Sato, Evolving Constitutions of International Organizations 259 (1996): "because constituent instruments are reasonably considered to contain many concepts and provisions of mobile reference, the provisions concerned will, unless the intentions of the parties are proved to be reference to a fixed content upon the examination, be regarded to be based upon a mobile reference and will be so interpreted. To this extent, the evolutionary nature of constituent instruments will be supported by the theory of inter-temporary law and their evolutionary and teleological interpretation will have a legitimacy".

101 As noted in JENNINGS, R. AND WATTS, A., OPPENHEIM'S INTERNATIONAL LAW, VOL. 1, 1268: "There is however room for the view that a treaty of a `constitutional' character should be subject to somewhat different rules of interpretation so as to allow for the intrinsically evolutionary nature of a constitution". Whatever is meant by "constitutional" here, it seems that the same reasoning applies to what I termed regulatory treaties, such as the WTO treaty.

102 For a possible exception, see supra text at note 89. From a purely linguistic (or "ordinary meaning" of the words, Art. 31.1 VCLT) perspective, practice has, indeed, shown that the starting point for many WTO provisions is the time of application of the WTO treaty, not the time of its conclusion (April 1994): when construing the "ordinary meaning" of WTO terms, both panels and the Appellate Body consistently resort to the most recent version of, for example, the Oxford English Dictionary, not the version as it existed in 1994. As Fastenrath remarked: "If one takes modern philosophy of language into account, a dynamic understanding of legal rules becomes inevitable" (ULRICH FASTENRATH, LÜCKEN IM VÖLKERRECHT: ZU RECHTSCHARAKTER, QUELLEN, SYSTEMZUSAMMENHANG, METHODENLEHRE UND FUNKTIONEN DES VÖLKERRECHTS 295 (1991), hereafter ,,Fastenrath, Lucken").

103 Case A/26, 15. See also the Marckx case, Case A/31, 19.

104 See also supra note 84.

105 See Joseph Weiler, Introduction : The Reformation of European Constitutionalism, in JOSEPH WEILER, THE CONSTITUTION OF EUROPE 221, at 225 (1999).

106 Decision of 17 July 1986, 1986 RGDIP 713, at 734.

107 DENYS SIMON, L'INTERPRETATION JUDICIAIRE DES TRAITES D'ORGANISATIONS INTERNATIONALES 372-373 (1981). See, more generally, MCDOUGAL, M., LASSWEL, H., and MILLER, J., THE INTERPRETATION OF INTERNATIONAL AGREEMENTS AND WORLD PUBLIC ORDER 99 (1967): "the principal aim of an interpreter should be to give effect to the continuing consensus of the parties - that is, their contemporary shared expectations concerning problems of the type being disputed".

108 See SHABTAI ROSENNE, BREACH OF TREATY, 3-4 (1985). On that basis, Rosenne distinguishes "the law of the instrument" from "the law of the obligation".

109 The example of the position of some developing countries in respect of genitically modified organisms (GMOs) is telling. In the WTO, they are very much opposed to trade restrictions in respect of GMOs (thus safeguarding trade rights in a WTO trade context). In contrast, during the negotiations of the Carthagena Biosafety Protocol they pleaded very much in favour of granting as much leeway as possible to states wanting to protect themselves against GMO imports (thus safeguarding environmental protection rights in the biosafety, environmental context).

110 Official Records of the Vienna Conference, Vol. 2, 253, para. 39. The Expert Consultant was confirming an earlier statement to that effect by Sir Ian Sinclair, the UK representative who had expressed a similar view: "the decisive date should be that of the adoption of the treaty; it is based on paragraph 1 of Article 56 [now Art. 59], which referred to the conclusion of a later treaty" (ibid., Vol. 2, 222). Before the vote on what is now Art. 30 was taken, the Ceylonese delegation once again confirmed this approach as follows: "the crucial date ... should be the date when the text of the new treaty had been finally and formally established" (ibid., vol. 2, 56, para. 50). In support: Vierdag, The Time of the `Conclusion' of a Multilateral Treaty, 60 BYIL 75 (1989) (hereafter "Vierdag, Conclusion ...").

111 Ibid., Vol. 2, 253, para. 39.

112 See note 110. Consequently, one of the problems put forward by Sir Ian Sinclair at the Vienna Conference can be easily resolved. The problem was the following: "Supposing that Convention A was signed in 1964 and came into force in 1966, whereas Convention B was signed and entered into force in 1965, which of them would be earlier?" (Official Records of the Vienna Conference, Vol. 1, 165). The answer must be: Convention A since A was concluded in 1964 and B only later in 1965.

113 Ibid., Vol. 1, 165.

114 In this respect, Vierdag refers to « the distinction between abstract norms and concrete rights and obligations ... `Abstract norms' refers here to treaty rules as such, irrespective of the legal position of signatories or States bound by the rules. `Concrete rights and obligations' refers to the specific position of a particular State with respect to one or more treaties" (Vierdag, Conclusion ..., 94).

115 But recall Art. 18 VCLT on « Obligation not to defeat the object and purpose of a treaty prior to its entry into force ».

116 In the WTO, for example, this would be the date at which the Ministerial Conference approves "the agreement on the terms of accession" pursuant to Art. XII:2 of the Marrakesh Agreement. Note that such approval can be adopted by two-thirds majority of WTO members. Hence, here as well, the situation may arise that a WTO member does not agree to another state's accession but nonetheless that state can accede. As a result, earlier treaties as between these two parties would, without the consent of the original WTO member, be overruled by the later WTO agreement. Art. XIII of the Marrakesh Agreement allows, however, for existing WTO members to decide not to apply the WTO treaty in its relationship with an acceding state. Moreover, the WTO practice is to approve accession only by consensus.

117 Vierdag, Conclusion ..., 101. He adds that in those cases « [t]he dates of adoption (or opening for signature) or entry into force of a treaty are of limited relevance in this connection, and the relevance of Article 30 is limited accordingly".

118 See above Section I.

119 See supra text at note 107.

120 Such AB continuing treaties BB are, indeed, AB continuing acts BB, as referred to in Art. 14.2 of the ILC Draft 2000. On this notion, see Joost Pauwelyn, The Concept of a `Continuing Violation' of an International Obligation: Selected Problems, 1995 BYIL 415. In support, see also the references in note 107.

121 In contrast, if one were to take the singular act of conclusion of a multilateral treaty seriously, it could well be argued that GATT 1947 (as incorporated without any change in GATT 1994) remains, pursuant to Article II:4 of the WTO Agreement, a 93legally distinct94 instrument concluded not in 1994, but in 1947. Hence, in terms of timing, GATT remains the earlier treaty vis-E0-vis, for example, pre-1994 environmental treaties. At the same time, new WTO agreements (such as, the 1994 SPS agreement) would then, however, be later in time. Thus having to make a distinction between GATT and SPS rules shows the absurdity of the 93guillotine94 rule of time of conclusion in respect of 93continuing treaties94.

122 WTO doc. WT/MIN(01)/DEC/1, 20 November 2001, para. 1.

123 For an example, see Vierdag: 93the Radio Regulations [of the International Telecommunications Union] are subject to regular revisions, and they are adopted and become binding again on member States in the revised form. The regular conclusion of revised Regulations means that in the end the Regulations will always become the later treaty vis-E0-vis every other treaty that is not likewise regularly revised, such as the Covenant on Civil and Political Rights94 (Vierdag, Conclusion 85, 101).

124 Fastenrath, Lucken 85, 295.

125 This should be the case also in the hypothetical case outlined above, i.e., of an original WTO member A which concludes a bilateral agreement with B, before B's accession to the WTO. In that event the parties' consent around the WTO treaty emerged only at the date of adoption of B's accession so that, even if the parties' consent to the WTO treaty continues to be confirmed thereafter, the bilateral treaty (of a non-continuing nature) clearly predates the WTO treaty. As a result, Art. 30 must be applied and the WTO treaty prevails as lex posterior.

126 Art. 30 should continue to apply in case of amendments to a continuing treaty norm. There as well, there can be no doubt that the two norms are AB successive BB : the amendment comes later in time than the provision of the continuing treaty which is then no longer confirmed as of the date of the amendment. Hence the amendment must prevail as the later in time under Art. 30.

Top|Previous|Title|Next

 

 


This site is part of the Academy of European Law online, a joint partnership of the Jean Monnet Center at NYU School of Law and of the Academy of European Law at the European University Institute.
Questions or comments about this site?
Email Enfellows@exchange.law.nyu.edu