ESIL Reflection: Professor Baxter’s Legacy: Still Paradoxical?

PDF Version                                                 Vol 6, Issue 3
Editorial board: Samantha Besson, Jean d'Aspremont (Editor-in-Chief), Jan Klabbers and Christian Tams

Hugh Thirlway
Former Professor of International Law, Graduate Institute of International Studies, Geneva

Professor Baxter’s Legacy:  Still Paradoxical?

How quaint the ways of paradox!
At common sense she gaily mocks!

W. S. Gilbert: The Pirates of Penzance, Act II, No. 19.

A well-known problem in international law is the paradox named after the late Professor R.R. Baxter, who first pointed it out. It arises when a codifying convention in a particular field of law has been drawn up and has been accepted by a number of parties, amounting in fact to the majority of States (or of States to which the subject matter is relevant). The convention incorporates a rule (Rule X) which, at the time of the conclusion of the convention, had attracted some degree of State practice and evidence of opinio juris such that the rule might have been expected to develop into a rule of customary law. A majority of States (or of the States potentially affected by the rule) become parties to the convention, so their practice in the relevant domain is subsequently treaty-based. What of those States that, for whatever reason, choose not to become parties to the convention? The question will be whether their practice, in the domain to which Rule X applies, is consistent with that rule, so as to be capable of developing on this basis a rule of customary international law, in parallel with the convention.  Professor Baxter pointed out that

the proof of a consistent pattern of conduct by non-parties becomes more difficult as the number of parties to the instrument increases. The number of participants in the process of creating customary law may become so small that the evidence of their practice may be minimal or altogether lacking. Hence the paradox that as the number of parties to a treaty increases it becomes more difficult to demonstrate what is the state of customary international law dehors the treaty.[1]

Thus the more popular the rule is as a treaty provision, the less its chances (it would seem) of becoming a demonstrable customary rule also.

The matter does not appear ever to have arisen as a practical problem, in the sense that a specific inter-State dispute on the existence or applicability of a rule of law was found to involve the paradox; nor is it often the subject of academic discussion nowadays. (A valuable exception is the treatment of it by Crawford in his 2013 Hague Academy lectures, which will be looked at more closely below.[2]) It was considered by the International Law Commission (the ‘ILC’) in the course of its work on the Identification of Customary International Law; the Special Rapporteur discussed in particular the views of Crawford, just mentioned, and agreed with him that ‘the Baxter paradox is not a genuine paradox’.[3] However that may be, it appears to be still recognized as having some relevance or validity, even if it is ‘more apparent than real’,[4] i.e. principally of theoretical interest; but on that level, there is perhaps more to be said.

The paradox centres on the possible significance of a treaty (particularly a multilateral treaty) for the establishment or emergence of a rule of customary international law; and this has of course been considered by the ILC in recent years in its work on the Identification of Customary International Law. In its draft Conclusions the Commission indicated three situations in which ‘a rule set forth in a treaty may reflect a rule of customary law’.[5]

The first of these is not here pertinent: that in which the treaty was itself truly custom-codifying, in the sense that when it was adopted there was already a customary rule established.[6] As the ICJ found in in the Armed Activities case in 1986, the adoption of a multilateral convention re-enacting an existing rule of customary law does not cause the disappearance of the customary rule, nor are the treaty parties exempt from its application.[7]  In this situation, the paradox does not arise: the States, however few, that refrain from becoming parties to it, but align themselves with the customary-law practice, are not lone pioneers, who must muster sufficient numbers for a custom to develop; they are recruits to an existing custom, revealing themselves as such by the same practice and opinio as though the codifying convention had never been concluded.

The second scenario, in which the Baxter paradox may well become relevant, is that in which a convention ‘has given rise to a general practice that is accepted as law (opinio juris), thus generating a new rule of customary international law’.[8] Less clear-cut is the third scenario, where the convention ‘has led to the crystallization of a rule of customary international law that had started to emerge prior to the conclusion of the treaty’, because an element to be taken into consideration in deciding how far it has done so will be the practice of non-party States. The distinction between this and the first scenario may not be easy to make in practice; and it is arguable that, in case of doubt, the number of non-treaty States who adopt the relevant practice, and the degree of promptness (after the entry into force of the convention) with which they do so, may be invoked in favour of the view that there was a pre-existing custom.

Let us however consider more closely this ‘head-counting’ process, required to establish the generality of practice among non-convention States, out of which the paradox arises. The exclusion of the convention States from the calculation is based on what was said by the ICJ in the North Sea Continental Shelf cases. The Court posed the question: had the equidistance delimitation rule stated in the 1958 Geneva Convention on the Continental Shelf become a rule of customary law? Examining State practice,[9] the Court said:

To begin with, over half the States concerned [i.e., those whose delimitation practice had been invoked to support a customary rule], whether acting unilaterally or conjointly, were or shortly became parties to the Geneva Convention, and were therefore presumably, so far as they were concerned, acting actually or potentially in the application of the Convention. From their action no inference could legitimately be drawn as to the existence of a rule of customary law in favour of the equidistance principle.[10]

However, as regards States that ‘shortly [afterwards] became’ parties to the Convention, surely this is being wise after the event? At the time of each of these delimitations, the State concerned was perfectly free to apply the equidistance method or not to apply it, so its choice to apply it could have been based simply on the fact that the method seemed to give satisfactory results in the circumstances (and the existence of the Convention might have been seen as evidence in favour of this view). It could also have been based on the belief that there was already a customary rule requiring it, which would make the delimitation an act of State practice.[11]  What it could not have been based on was any perceived treaty-obligation, for none yet existed for those States.[12] If one had been assessing the state of the equidistance rule at the time of, and in the light of, each of these delimitations, all one would have been able to say is that the rule was regarded by those States as a good rule, and one appropriate to the specific delimitation.  One could not conclude that the State concerned regarded itself as bound by the Convention, to which it was not yet a party, to use equidistance: one could only say that either it did not feel itself bound at all, or that it felt itself bound by a non-treaty rule, i.e. a custom.

Secondly, practice between States both of which were, at the time, parties to the convention may certainly be taken to have been based on the obligations of the convention; but what of practice between a convention State and a non-convention State? Unless the convention specifically provides otherwise, States parties to it must be taken to be free to act without reference to the convention in their relationships with non-convention States.[13] Thus the practice that is to be excluded from any assessment of custom is solely that between convention States.

Having excluded, at least, the practice of States that were, at the time of relevant practice, parties to the relevant convention, in relation to each other, one turns to the practice of other States. The emphasis, in discussions of situations where the paradox may operate, tends to focus on the limited number of non-convention States, whose activities could therefore be custom-creating.[14] But is the question one of absolute number, or of proportion? Normally, one could not convincingly argue that a general custom had arisen from the practice of half-a-dozen States; but if the custom under discussion was, it was suggested, a ‘particular’ custom, involving seven or eight States in total, the practice of the half-dozen might well be held to establish it, provided it was regarded as indicating the existence of ‘a constant and uniform usage practised by the States in question’;[15] or as ILC Draft Conclusion 16 [15] puts it, ‘a general practice among the States concerned’.[16] To put the point more generally, it would seem that the relevant figure is the total number of States who could participate in a law-creating practice.  If the alleged customary rule is, for example, one of maritime delimitation, the number of States endorsing it by practice would be compared with the total number of States with a coastline;[17] the land-locked States would not be taken into account because they cannot contribute by practice to such custom-formation.

But the States parties to the convention also cannot contribute by practice to custom-formation; they are excluded from consideration by the North Sea Continental Shelf ruling. Does it not follow that for the purposes of calculation in the situation envisaged by Baxter, it is only the States that have not ratified the convention that form the overall group within which the enquiry into the ‘generality’ of customary practice should be directed? The number of such States remains limited, but that number appears in a different part of the equation.        If, say, 85% of the non-convention States endorse the alleged rule by practice, is that not sufficient to conclude in favour of a custom, even if they are fairly few in numerical terms, and greatly outnumbered by the convention-States?[18] That practice can also be backed up by any practice between convention and non-convention States, as noted above.

The objection that in such circumstances ‘only a few States would . . . be creating customary law’[19] takes on a different significance when these States are creating customary law which is at one and the same time general custom, and yet binding as such only on themselves – with the possible inclusion of an even smaller number of other non-convention States who have not put themselves in the position of ‘persistent objectors’.[20] 

Of recent writers, Fausto Pocar discounts the Baxter paradox by arguing for the inclusion of the practice of the convention States to support the customary rule:  ‘the treaty itself is an important piece of State practice for the determination of customary law, although its role in this regard must be carefully assessed, and the impact that any subsequent practice of the contracting States in the application of the treaty which establishes their agreement or disagreement regarding its interpretation may bear on the development of a customary norm’.[21]

The most recent study of the paradox is that of Crawford in his 2013 Hague Academy Lectures, in which he goes so far as to entitle a section ‘Solving the Baxter Paradox’.[22] His first suggestion is ‘to generate a presumption of opinio juris from wide participation in a treaty, at least in normative terms’.[23] He suggests that this was done by the Ethiopia-Eritrea Claims Commission with reference to the 1949 Geneva Conventions, which is correct, but it is material that the convention in question was a humanitarian one; in the case of such a convention non-participation by a State is likely to be attributable less to rejection of the principles embodied in it than to (for example) inertia, political prejudice, etc.  Furthermore in the Ethiopia-Eritrea case, the parties agreed that the relevant rule had become customary.[24]  Could such a presumption be imposed on a dissenting State? Crawford’s answer is that such a situation should be met by application of the ‘persistent objector’ principle.

A second point advanced by Crawford is, in effect, a denial of the North Sea dictum as to the non-relevance of the practice of States that are already parties to the convention. In his view, the sources that constitute the usual indicators of a State’s attitude to a possible customary rule (diplomatic correspondence, policy statements, etc.) ‘will continue to be created and shed light on a State’s position long after it has ratified a treaty’ on the same subject.[25] Certainly a State party to a convention may hold a view on whether the rule expressed in that convention has also become a customary one, and may urge that view upon (in particular) non-party States. But can that constitute ‘State practice’ as traditionally understood?  What a State in this position cannot say is that it is itself acting in a particular way because it is satisfied that there is a customary rule obliging it to do so – the elements of practice and opinio juris in the classic sense.


So long as there exist multilateral conventions that have not been ratified by every State in the world, and that deal with matters that are also regulated, to a greater or less extent, by customary law of more recent origin (in many cases having arisen partly as a result of the convention), the situation to which Professor Baxter drew attention will continue to be observable. It may even be regarded as coming to present a more serious issue, insofar as the subject matter of such conventions is frequently such that universal recognition of the principles they embody is to be seen as, at the least, highly desirable. If some of the rules for the delimitation of the respective areas of continental shelf of coastal States are rejected by a few States, this is acceptable in the context of inter-State relations generally, and lends emphasis to the continuing importance of State sovereignty; but in relation to such matters as, for example, the prohibition of genocide, and the protection of at least the most basic of the recognized human rights, claims to hold dissenting views (which imply potentially divergent action) cannot be viewed with the same equanimity. Yet, as noted above, the consequent collision remains merely potential; no inter-State dispute has been recorded that turned on the validity, or otherwise, of such a claim.

Such a dispute could of course yet appear,[26] and transform the paradox from an interesting intellectual problem to one of actual international relations. There have thus been a number of attempts to ‘solve’ the Baxter paradox, to get round it, or to demonstrate its lack of practical impact. Can the paradox be ‘solved’; or is it merely to be ‘avoided’? These are the two approaches to it suggested by Crawford.[27] A prior question is however, is it a paradox at all, and if so, of what kind ? It contains no self-contradiction, like the well-known ‘Cretan liar’ paradox, or otherwise logically unacceptable conclusion, nor an elusive definition, like the equally well-known sorites or ‘heap’ paradox.[28] It has merely a counter-intuitive element: one would expect that the more States show allegiance to a developing rule of law, by ratifying a treaty embodying it, the more easily it could be shown to have become a general customary rule. It states, or represents, in dramatic form a fact which is inconvenient for the development of international law, and its consistent application. There is no need to seek a ‘solution’ to the paradox, but rather a way of palliating that inconvenience.

One may first of all scrutinise the factual basis to ensure that it is not over-stated. With this in view, we have noted above (1) that it is not necessary to exclude from consideration the practice of States because they later became convention-parties, and (2) that for purposes of proportionate assessment, the total ‘heads to be counted’ are merely the non-convention States.

Secondly – de lege ferenda – one may introduce some adjustments into the classic analysis of custom-making: thus Crawford proposes, as we have seen, the adoption of a presumption of opinio juris from the simple fact of widespread participation in a law-making convention, and that account be taken of the attitude towards the relevant rule adopted by States who are committed to it in its convention form.  However, Professor Baxter’s name will, it seems, continue to appear in this context in the index of works on customary law, at least until these adjustments are accepted by the international community. Perhaps they should be made the subject of an international convention ...


Cite as: Hugh Thirlway, ‘Professor Baxter’s Legacy: Still Paradoxical?’ 6:3 ESIL Reflection (2017).

[1] R.R. Baxter, ‘Treaties and Custom’, (1970) 129 Recueil des Cours de l’Académie de Droit International, 64.

[2] James Crawford, ‘Chance, Order, Change: The Course of International Law’, General Course of Public International Law, 365 Recueil des Cours (2013) in Hague Academy Pocketbooks, paras128-174; see also Fausto Pocar, ‘Protocol I Additional to the Geneva Conventions and Customary International Law’ in Yoram Dinstein et al. (eds) The Progression of International Law (Brill 2011) 199.

[3] ILC, ‘Third Report on the Identification of Customary Law by Special Rapporteur Michael Wood’ UN Doc A/CN.4/682, 29, para 41, citing 365 Recueil des Cours (2013) 107, 110.

[4] cf Crawford (n 2) 107, 110; Mathias Forteau, ‘A New “Baxter Paradox”?  Does the Work of the ILC on Matters Already Governed by Multilateral Treaties Necessarily Constitute a Dead End? Some Observations on the ILC Draft Articles on the Expulsion of Aliens’, Harvard Human Rights Law Journal 2016.

[5] ILC, Draft Conclusion 11[12] on the Identification of Customary International Law, para 1.

[6] ibid, para 1(a).

[7] ‘The fact that the above-mentioned principles, recognized as such, have been codified or embodied in multilateral conventions does not mean that they cease to exist and to apply as principles of customary law, even as regards countries that are parties to such conventions’: ICJ Reports 1986, 93, para 174 (emphasis added).

[8] This is the situation contemplated by paragraph 1(b) of ILC Draft Conclusion 11[12] on the Identification of Customary International Law.

[9] The Judgment mentions that ‘fifteen cases [had] been cited . . . in which continental shelf boundaries have been delimited according to the equidistance principle’ (43, para 75); it specifies four of these (UK/Norway, UK/Denmark, UK/Netherlands and Norway/Denmark), but does not say which were the others.  From the pleadings, they would seem to be Bahrain/Saudi Arabia, FRG/Denmark, Netherlands/Denmark, USSR/Finland (2 agreements), and Belgian legislation concerning Belgium’s continental shelf (annexed to the Counter-Memorials) and Norway/Sweden, Italy/Yugoslavia, Australia/West Irian.

[10] [1969] ICJ Rep 3, 43, para 76.

[11] This might appear more likely if the convention in question was one containing provisions with an ethical dimension, rather than concerning the nature and extent of seaward claims.

[12] The Court refers to these States as ‘acting ... potentially in the application of the Convention’ (North Sea Continental Shelf, n10). However, this is an evasion; the question is what motivated the States concerned at the time of their delimitation.

[13] cf ILC: ‘[W]here States act in conformity with a treaty by which they are not (yet) bound, or towards States not parties to the treaty, the existence of “acceptance as law” may indeed be established’: ILC (n 3), 48-49, para 62, citing ‘the reference to Venezuelan practice in Colombian-Peruvian asylum case, Judgment of November 20th, 1950: ICJ Reports 1950 266, 370 (Dissenting Opinion of M. Caicedo Castilla)’; Dapo Akande in Opinio Juris, agreeing with the approach that ‘one can use the practice of States parties to treaties in their relations with non-parties . . . since in such a case the practice of the party cannot be said to be based on the treaty’. Dapo Akande, ‘Response’ to Duncan Hollis, ‘The Empire Strikes Back – Debating the Origins of the Customary Laws of War’ (2007) Opinio Juris  <> accessed 15 March 2017.

[14] Thus Sir Robert Jennings in the Military and Paramilitary Activities case, dismissing the idea that a customary law on force and self-defence existed alongside the Charter, referred to ‘those few States which are not parties to the Charter’: [1986] ICJ Rep 531; and Professor Akande too has emphasized that ‘The problem is that only a few states would then be creating customary international law’. ibid. But note that provisions of the convention that do not reflect custom may still have some relevance in relations between a party and a non-party: see Territorial and Maritime Dispute (Nicaragua v. Colombia) [2012] ICJ Reports, 624, 669, para126 in fine.

[15] Asylum case [1950] ICJ Rep 276; the text does not say ‘all the States in question’, though this has been ‘written in’ by some commentators: see e.g., Forteau in Max Planck Encyclopedia, s.v. ‘Regional International Law’, para 20. The same would apply if the custom related to a field in which only a limited number of States, not geographically related, were concerned, so that the possible rule would or could concern them.

[16] ILC, ‘Third Report on the Identification of Customary Law’ (n 3).

[17] This is presumably what is meant when the Court in North Sea refers to the cases ‘potentially calling for delimitation in the world as a whole’: ICJ Reports 1969, 43, para 75.

[18] Sometime around 1972/3, as a newcomer to international law, I put this point to Professor Baxter in a letter and received a very friendly reply, in which, as I recall, he conceded that my argument had validity; but unfortunately the letters have long since been lost!

[19] Akande (n 13).

[20] This however implies that the new custom will always remain limited to the small group; what will be the position if one or more States parties to the relevant convention withdraw from it? Unless any customary rule established by the small group of non-treaty States is a general custom, the withdrawing State would apparently be left in a uniquely lawless position!

[21] Pocar (n 2); but it is difficult to see any opinio juris in this context.

[22] Crawford (n 2), paras167-174.

[23] ibid, para 167.

[24] Prisoners of War: Ethiopia’s Claim 4, Partial Award (2003) 135 ILR 263, para 31.

[25] Crawford (n 2), para168.

[26] And not only would any dispute-settlement provisions in the relevant convention of course be inapplicable, but the ‘heretic’ State would probably be one of those holding aloof from compulsory settlement in general.

[27] ibid, Chap. IV, Sections C and D.

[28] Mentioned by Crawford (n 2), 143-4, para 169, who assimilates the Baxter paradox to it.