Determining Customary International Law: The ICJ’s Methodology and the Idyllic World of the ILC

Written by

Editor’s Note: This is the author’s concluding post in the EJIL:Debate! regarding an article in the current issue of EJIL Vol. 26 (2015) No 2, by Stefan Talmon. The original post is here. See also the posts discussing the article by Omri Sender and Michael Wood, Harlan G. Cohen and Fernando Lusa Bordin.

I am very grateful to Sir Michael Wood and Omri Sender, as well as Harlan G. Cohen and Fernando Lusa Bordin, for their thoughtful comments on my EJIL article. Both Harlan and Fernando seem to agree with my main propositions and, in particular, with the proposition that the ICJ, in order to determine rules of customary international law, uses induction and deduction as well as assertion. They raise interesting questions that I did not address in my article, such as why the ICJ was not more interested in developing a clear methodology and why States might actually prefer ‘methodological mayhem’, or the flexibility of methodological uncertainty, over the strict application of the inductive method or a relaxation of the demands of that method. Their contribution takes the debate further and may be read as a complement to my article.

In the following, I will focus on the comments of Sir Michael Wood and Omri Sender, who are more critical of my propositions. I will only deal with their substantive comments and leave readers to decide for themselves how many eyebrows they would like to raise while considering what the authors identified as ‘some bold statements’ in my article without, however, specifying their concerns. Let me respond to their counter propositions one by one before offering some final remarks on the work of the ILC, and thus Sir Michael’s work as its Rapporteur, on the identification of customary international law.

The ICJ has a ‘coherent methodology’

Sir Michael and Omri take issue with my proposition that the Court has hardly ever stated its methodology for determining the rules of customary international law. Quite on the contrary, they argue, a ‘coherent methodology does come into sight’ in the decisions of the ICJ. In these decisions the ICJ ‘made clear’ that ‘the existence of a rule of customary international law requires that there be a “settled practice” together with opinio juris’ and that the Court ‘must apply the criteria which it has repeatedly laid down for identifying a rule of customary international law’.

The authors have not taken issue with my definition of ‘methodology’ as meaning ‘the methods employed to determine the existence of rules of international law’, nor is it clear what their definition of ‘methodology’ is, if they have a different one. However, I have difficulties seeing how the ‘application’ of the two constituent elements of international custom can be called a ‘methodology’ for determining rules of customary international law. It seems that the authors are more concerned with the abstract criteria for the ‘existence’ of a rule of customary international law than with the method used for determining this existence in practice. The fact that the ICJ is paying lip-service to the two constituent elements of international custom does not say anything about the method it actually uses to determine the existence of rules of customary international law. As its case law shows, the ICJ does not live up to its self-proclaimed duty to prove the existence of customary international law by producing evidence of general and consistent State practice and opinio juris.

The authors also take issue with my finding that the literature has had little to say on the ICJ’s methodology for the determination of customary international law. While the term ‘little interest’ is a relative one, the five additional pieces of literature cited by the authors (Geiger, Ferrer Lloret, Hagemann, Benvenisti and Tams) do not make the ICJ’s methodology in determining customary international law a well-studied subject in international law; especially if one compares, for example, the number of papers on the question of the ICJ’s methodology with the wealth of material on customary international law in general. However, their selection of additional literature is revealing as it shows the authors’ misconception of the question of ‘methodology’. Again, they are not so much concerned with ‘methodology’ but with the criteria for the existence of customary international law and the proof of these criteria.

For example, Hagemann’s article deals exclusively with the questions of what, according to the ICJ’s case law, constitutes State practice (61-67), what quality that practice must have (i.e. frequency, uniformity, generality, specificity and legal import of practice) (68-73), what counts as opinio juris, how it is proven and what weight should be given to opinio juris compared to State practice (73-88). Ferrer Lloret’s piece does not deal with the question of the ICJ’s methodology as the partial title provided by the authors may suggest. Instead, as the full title of the article indicates, it is a Spanish language case-note (with a one paragraph abstract in English) on the Jurisdictional Immunities of the State case and deals with the question of the proof of State practice and opinio juris in that particular case and criticizes the Court for its limited use of State practice and, in particular, for not taking into account the decisions of US courts. It remains to be added that the excellent article by Christian Tams was published subsequently to my article and was thus difficult to take into account.

The ICJ does not employ the ‘deductive’ method

The authors’ statement that I provide my ‘own definitions for “inductive” and “deductive” methods’ seems to imply that I have made up definitions that do not conform to the general usage of these terms and thus that I should not complain of the Court’s inconsistent and non-technical use of induction and deduction. As set out in the article, the terms have their origin in Aristotle’s logic and the definitions provided are consistent with the meaning generally attributed to them (see e.g. Dominicé, Methodology of International Law, Encyclopedia of Public International Law, vol. III (1997), p. 356; Salmon, Dictionnaire de droit international public (2001), pp. 924-925). Sir Michael himself in his First Report on formation and evidence of customary international law contrasted ‘empirical research into State practice’ with ‘deductive reasoning’ (UN Doc. A/CN.4/663, 17 May 2013, 9, para. 21).

According to the authors, the word ‘deduce’, as used by the ICJ, simply means to ‘reach a conclusion’ rather than inferring, by way of legal reasoning, a specific rule from an existing and generally accepted rule or principle. With all due respect for authors’ reliance on ‘plain English’, the passages quoted from the Court’s jurisprudence in my article seem to employ the term in the latter rather than in the former meaning. In the Barcelona Traction case the Court speaks of ‘deducing from considerations of equity any broader right of protection’ (ICJ Reports (1970) 3 at 48, para. 94). President Bustamante y Rivero stated in his separate opinion in the North Sea Continental Shelf cases that ‘it is possible to deduce [other essential principles] from the accepted concept of the continental shelf’ (ICJ Reports (1969) 3 at 58, para. 3) and Judge Fouad Ammoun said that the Court ‘must enquire what rule should be deduced from the principle of equity’ (ICJ Reports (1969) 3, at 147, para. 48) and that it ‘is from the principle […] that the rules flow. It is therefore necessary, after having gone back to the principle, namely equity, to state what rules applicable to the matter can be deduced from it.’ (ibid, 147, para. 49). Further examples are given in the article.

The authors question my proposition that the ICJ also uses the deductive method to determine rules of customary international law because ‘the Court itself does not use abstract terms, such as “induction” and “deduction”, to describe what it does’. But, the Court does not always do what it says (see Geiger, in: From Bilateralism to Community Interest: Essays in honour of Judge Bruno Simma (2011), pp. 674, 692) and it does not necessarily always say what it does. While it is true that the ICJ ‘does not engage in theoretical speculation’, that does not mean that academics are not to engage in ‘theoretical analysis’ of what the Court does.

The authors’ conclusion ‘induction “yes”, deduction “no”’ is also rather surprising considering that the ILC chairman, Mr Niehaus, said in his report to the 6th Committee of the UN General Assembly in 2013 that the ‘Commission was aware that the Arrest Warrant case had been the subject of critical commentary in relation to the immunity ratione personae of the minister for foreign affairs, as it was predicated on deductive reasoning rather than on an analysis of State practice’ (UN Doc. A/C.6/68/SR.17, 8 November 2013, at 6, para. 20).

The ICJ does not use a method of assertion but a ‘“pragmatic” approach’

The authors are, of course, correct that assertion is ‘not a methodology’ (something never claimed) but it can be a ‘method’ – in plain English ‘a way or manner’ – of determining the existence of rules of international law. What the authors call a ‘“pragmatic” approach’, i.e. the reliance on ‘the considered views expressed by States and bodies like the International Law Commission’ or the use of ‘rules that are clearly formulated in a written expression’ is, in my opinion, nothing else but judicial window-dressing. That this ‘pragmatic approach’ has apparently been endorsed by the then President of the ICJ, Judge Tomka, does not make the matter any better. Reference to the ILC is exactly one of the techniques of assertion identified in my article (p. 437). I have, of course, not been alone in drawing attention to the ICJ’s ‘marked tendency to assert the existence of a customary rule more than to prove it’ (Pellet, in: Looking to the Future, Essays on International Law in Honour of W. Michael Reisman (2011), 1076).

The authors state that the ‘“pragmatic” approach’ is not precluded ‘in appropriate cases, for example where the written text relied upon by the Court is the result of a rigorous and systematic review of practice and acceptance as law, and as such has met with the general acceptance of States’. But who examines and guarantees that a written text – be it ILC draft articles or a treaty – has really been met with ‘the general acceptance of States’? It is exactly this unproven reliance on other actors and texts that opens up the possibility of assertion for the ICJ. Eyal Benvenisti (in his stimulating article mentioned by the authors, at pp. 85-86) has pointed to the 1997 Convention on the Law of Non-navigational Uses of International Watercourses as an example for the ICJ’s ‘unsupported assertion’ in the Gabčikovo-Nagymaros Project case (ICJ Reports (1997) 7 at 56, para. 85) that the Watercourses Convention reflected contemporary customary law although its text had been adopted less than four months earlier, had no signatories at the time and in fact was opposed by numerous States. Adopting the authors’ telegram style, I would thus conclude: pragmatic approach ‘no’, assertion ‘yes’.

On the basis of the authors’ comments I see no reasons to revise my main proposition that the main method employed by the ICJ is not induction or deduction but assertion. By declaring rules to constitute customary international law (with or without reference to the views of others) without proving either the existence of a general and consistent practice of States or its acceptance as law, the ICJ has become the Oracle of The Hague to which States make their occasional pilgrimage to hear the words of custom from the mouth of the judges.

The ICJ’s methodology and the idyllic world of the ILC

It is perhaps not surprising that the authors, and in particular Sir Michael as the ILC’s Special Rapporteur on the topic of ‘Identification of customary international law’, are more critical of my article than others as its findings are unsettling the idyllic world of the ILC. The ILC’s 16 draft conclusions on the identification of customary international law provisionally adopted by the Drafting Committee in July 2015 draw ‘heavily on case-law of the International Court’. The draft conclusions give the impression that the determination of the existence and content of rules of customary international law is an exact science. It seems as if the ILC has developed a little machine into which the two elements of ‘general practice’ and ‘opinio juris’ are put according to the recipe found in the ILC’s cookery book entitled ‘Conclusions on the Identification of Customary International Law’ and out come rules of customary international law.

The recipe, of course, is neither new nor particularly nutritious and most of its ingredients can be found in good textbooks on international law (see e.g. James Crawford, Brownlie’s Principles of Public International Law, 8th edn, 2012, pp. 23-30). While the draft conclusions are a helpful restatement of how courts and others should go about trying to identify rules of customary international law, the ILC’s recipe is only sufficient for the first course. I was rather surprised to learn from Sir Michael’s Third Report (UN Doc. A/CN.4/682, 27 March 2015, pp. 67-68) that this is it. There are no more draft conclusions to be expected. But what about situations in which State practice is non-existent or precarious, State practice is conflicting or too disparate and thus inconclusive, the opinio juris of States cannot be established, or there is a discrepancy between State practice and opinio juris (see my article, pp. 432-434)? In all these situations the ICJ has not been prevented from finding rules of customary international law. The ILC does not attempt to explain this.

There is also nothing in the conclusions addressing the subjectivity and selectivity of any assessment of State practice and opinio juris. These problems inherent in the determination of rules of customary international law do not disappear simply by not addressing them. What the ILC puts forward in its conclusions is a pure theory on the identification of customary international law, but the practice of the ICJ (and other courts) is more complex. If Christian Tams and Eyal Benvenisti are correct (and I think they largely are) that currently ‘anything goes in the ascertainment of custom’ (Tams, Meta-Custom and the Court: A Study in Judicial Law-Making, LAPE 14 (2015), p. 79) and that ‘the ICJ or other tribunals “cheat” by inventing what they refer to as custom’ (Benvenisti, in: The Impact of International Law on International Cooperation (2004), p. 87) one would have expected more of the ILC and its Special Rapporteur than to perpetuate the myth that international custom is established inductively, and only inductively, by ascertaining that there is a general practice that is accepted as law (opinio juris).


Print Friendly, PDF & Email


No tags available

Leave a Comment

Comments for this post are closed


Will Worster says

December 3, 2015

Thanks for this great discussion among so many scholars of customary international law. Initially, we need to be clear whether we are talking about the existence of CIL or the proof of CIL. Sir Michael's reports did a great job of highlighting this and this difference explains some of the divergence of views here. I understand assertion as reaching a conclusion without rigorous methodological substantiation, but the assertion/conclusion could still be correct after all, and potentially still justifiable with rigorous method. This is a failure of evidence but not necessarily of existence.

Also I think we need to be clear what step of the CIL methodology we are talking about when we argue over the application of deduction vs induction. In my view, the ICJ applies deduction and induction differently to, e.g., the rules on the rules of CIL, the elements of CIL, the method itself, the (stratified) sampling method, the admissibility of types of evidence, the weight and burden of proof of law, and the transition from empirical description to normative prescription, among other logical steps. So when we say the ICJ applies a particular method, we are likely focusing on only one or two of the small steps in the overall logical analysis - again another source of the divergence of views, making some of the eyebrow-raising bold statements less controversial.

And lastly, let us not omit that the inductive and deductive methods interact. The application of induction at all is based on a deductive conclusion that induction is the correct scientific method. We could correctly reach a rule through deduction from a rule reached by induction - the deduced rule might be implicit and necessary for the induced rule. And examples could go on.

In sum, I do not think the ICJ uses a singular method easily captured in an induction/deduction argument. It uses an eclectic method, quite normal and natural for lawyers!

Panos Merkouris says

December 4, 2015

First of all, I would also like to thank all the contributors to this discussion for some very interesting points regarding customary international law. If I may offer my two pennies’ worth on the matter, the mixed application of the inductive and deductive approaches by the ICJ with respect to customary international law, may also be reflective of the fact that these two approaches correspond to two entirely different stages in the ‘life-cycle’ of a rule of customary international law.

Whereas induction is apposite to determining the existence of a customary rule, i.e. through weighing state practice and opinio juris to induce the existence of a customary rule, deduction, and especially the manifestations of deductive reasoning that Professor Talmon has identified, reflect a process of interpretation of an existing customary rule.

The inductive approach can only go so far. Clearly no State practice or opinio juris can or will ever be so specific as to provide concrete solutions as to the application of a customary rule in any imaginable situation. Not even written rules included in treaties have that degree of specificity (I won’t go into a discussion of in claris non fit interpretatio, since not only does it not cover all treaty rules, but I also find myself agreeing with McDougal who stated during the Vienna Conference on the Law of Treaties that this proposition is nothing more than an ‘obscurantist tautology’ (UN Doc. A/CONF.39/C.1/SR.31, para. 41)).

Consequently, once a customary rule has inductively been identified, the clarification of its specific content (if the inductive process fails to do so, which in most instances it will) will be a matter of interpretation. In this sense, the deductive approach to customary international law is the direct result of an existing customary rule being interpreted. To make matters even more interesting this process of interpretation has strong similarities with that of interpretation of treaties. This is suggested in Geiger’s chapter (mentioned in Sir Michael Wood and Omri Sender’s post), and has earlier been analysed in various writings of Bleckmann, Orakhelasvili and myself (I apologise for this instance of self-promotion).

But if you don’t want to take my word for it, i.e. that the deductive process of interpretation of customary international law resembles somewhat the process and rules that govern the interpretation of treaties, perhaps the words of one of the ICJ judges in North Sea Continental Shelf might be more persuasive:

“Customary law, being vague and containing gaps compared with written law, requires precision and completion about its content. This task, in its nature being interpretative, would be incumbent upon the Court. The method of logical and teleological interpretation can be applied in the case of customary law as in the case of written law.”
(Dissenting Opinion of Judge Tanaka, p. 181).