An earlier version of this paper was presented at the 2006 meeting of the American Philosophical Association's Pacific Division. The author wishes to thank the participants in that session for their comments, as well as Andy Altman and this journal's Editor for helpful discussion of the ideas contained herein.

1. See, for example, the International Court of Justice’s opinion in the North Sea Continental Shelf Cases; Restatement (Third) of the Foreign Relations Law of the U.S. (St. Paul, MN: American Law Institute Publishers, 1987)Google Scholar: §102; Thirlway, H.W.A., International Customary Law and Codification (Leiden: A. W. Sijthoff, 1972) at 47.Google Scholar

2. Geny, Francois, Methode d’interpretation et sources en droit prive positif 2nd ed. (Paris: F. Pichon et Durand-Auzias, 1919) at 367–71 Google Scholar. See also Benson, Peter, “Francois Geny’s Doctrine of Customary Law” (1982) 20 Can. Y. B. Int’l Law 267 Google Scholar.

3. Byers, Michael, Custom, Power and the Power of Rules (New York: Cambridge University Press, 1999) at 130–33 CrossRefGoogle Scholar.

4. Throughout this paper I follow the common practice in discussions about international law by describing states as agents; for instance, I speak of states having beliefs and acting as legal officials in the international legal order. But as Antonio Cassese points out, “although the protagonists of international life are States as legal entities or corporate structures, of course they can only operate through individuals, who do not act on their own account but as State officials, as the tools of the structures to which they belong” ( Cassese, Antonio, International Law, 2nd ed. (New York: Oxford University Press, 2005) at 4 Google Scholar). On the justifiability of attributing not only acts but also mental states to states, see Lefkowitz, “The Principle of Fairness and States’ Duty to Obey International Law” [unpublished, draft on file with author]Google Scholar.

5. Byers, supra note 3 at 131.

6. By “conventional rule” (or “social rule”) I mean a rule that exists as a matter of convention. This use of the term “convention” should be distinguished from a common understanding of that term in international law as synonymous with “treaty”.

7. Hart, H.L.A., The Concept of Law (New York: Oxford University Press, 1961)Google Scholar. Hart’s discussion of this distinction is somewhat problematic. For one attempt to clarify it, see MacCormick, Neil, H.L.A. Hart (Stanford CA: Stanford University Press, 1981) at 103–06 Google Scholar.

8. See Hart, ibid. at 104-05, 108.

9. I assume that A, B, C, D, E, etc., all believe themselves to be members of the same group; that is, each believes she is a member in the group, and believes that the others are members of the group.

10. Kontou, Nancy defends this view in The Termination and Revision of Treaties in the Light of New Customary International Law (New York: Oxford University Press, 1994)Google Scholar. This procedure for the creation of new customary international law fits the image of claim and response some international lawyers use to describe certain examples of customary law creation, such as the so-called Truman Proclamation. See Walden, Raphael M., “Customary International Law: A Jurisprudential Analysis” (1978) 13 Israel L. Rev. 86 Google Scholar.

11. See Byers, supra note 3 at 130.

12. See Hart, supra note 7 at 56-57, 101-03. For discussion, see MacCormick, supra note 7; Coleman, Jules, The Practice of Principle (New York: Oxford University Press, 2001).Google Scholar

13. One can easily imagine the other bus riders, or neighbors, complaining that the deviator has done something she ought not to do. But we must be careful to distinguish between (a) the clothing of individual preferences in moral language, which likely underlies the claim of a bus rider who complains of being displaced from “my seat”; (b) non-rule based moral reasons, such as those grounded in the principle of fair-play, which might provide a basis for criticizing the neighbor who refuses to plant flowers; and (c) criticism of an agent by appeal to a social rule, one that exists (and so provides a reason) if and only if most members of a group adopt the internal point of view with respect to the conduct at issue.

14. Brownlie, Ian, Principles of Public International Law, 5th ed. (New York: Clarendon Press, 1998) at 5.Google Scholar Brownlie writes that the term “usage” in international law refers to non-legal customary (or social) rules, such as those mentioned in the text, while the term “custom” serves as a term of art that refers only to social rules that impose legal obligations.

15. The reader might reasonably wonder how there comes to be specific legal officials in the first place. After all, the existence of such officials are presupposed in the account set out in the text of how a customary rule becomes legally valid and the concept of an office seems best understood in terms of various normative rules, which at least initially might well be customary rules. Thus the initial creation of a legal office, and so of a legal official, would seem to require the temporally prior existence of some customary legal rules, and so the chronological paradox arises once again. I set this question aside here in order to focus on the question of how new customary norms of international law come to be part of an already existing international legal order (which includes, among other things, international legal officials). For one account of this issue regarding the origins of legal officials (within a Hartian legal positivist account of law and a legal system), see Coleman, supra note 12 at 100-01.

16. Note the difference between this belief and the belief by ordinary members of this society that since their legal officials adhere to the aforementioned rule of recognition, and the customary rule exists, that rule is law. This belief while true, does not make the rule in question legally valid.

17. Since the rule of recognition directs officials to pronounce customary rules as legally valid there can be no new rule of customary law until a customary rule already exists. Given such a rule of recognition, it will often be true that once such a rule exists, it exists as a legally valid rule. That is, it exists as law prior to its application by legal officials to settle a dispute. But again, the beliefs of those subject to the rule (qua subjects) as to the legal status of the rule have no bearing on what that status actually is.

18. Judges sitting on the International Court of Justice (and the International Criminal Court) may be analogues to the legal officials described in the earlier hypothetical example; that is, agents whose beliefs are relevant to what counts as law, but not to the formation of customary rules that are eligible for incorporation into law. Article 38(1) of the Statute of the International Court of Justice is usually described as denying judicial decisions the status of a source of law, but what matters is the actual practice of officials in the international legal order, and here the issue may not be so clear.

19. See Hart, supra note 7 at 226-31.

20. Hart aims his critique against the existence of a rule of recognition for the international legal order as a whole, and not simply customary international law. But it seems quite possible for a legal order to contain various rules of recognition for sub-classes of legal rules in that order without their also being a single (even very complex) rule of recognition at the base of the legal order as a whole. Such a legal order would give rise to less uncertainty than a primitive legal order, insofar as some disputes could be settled by an authoritative process, even if it also produced more uncertainty than would a legal order with a single, all-encompassing, rule of recognition. The international legal order may constitute just such a legal system. Thus, Hart correctly endorses the rejection of pacta sunt servanda as the all-encompassing rule of recognition for the international legal order, yet as Franck (and others) correctly observe, this conventional rule does serve as a rule of recognition for one part of international law, namely treaty law, ibid. at 228; Franck, Thomas, The Power of Legitimacy Among Nations (New York: Oxford University Press, 1990) at 187 Google Scholar). Pacta sunt servanda is clearly akin to a rule directing actors to “an authoritative list or text of the rules,” which Hart offers as an illustration of a relatively simple rule of recognition; see Hart, supra note 7 at 92. Even if no authoritative procedure exists in the international legal order for adjudicating conflicts between obligations imposed by different treaties, or between treaty obligations and customary international law obligations, points on which I take no position, this entails only that the international legal order contains less systematicity than do many domestic (or municipal) legal orders. It does not follow that the international legal order contains no rules of recognition. Note too, that the argument of this paper depends only on the existence of a rule of recognition for customary international law, and not a single all-encompassing rule of recognition at the base of the entire international legal order.

21. See Hart, supra note 7 at 230.

22. Ibid. at 231.

23. D’Amato, Anthony A. “The Neo-Positivist Concept of International Law” (1965) 59 Am. J. Int’l L. 321 at 321-24CrossRefGoogle Scholar.

24. Ibid. at 323-24; see Hart, supra note 7 at 231, and discussion of Hart’s remark below.

25. Ibid. at 323.

26. Ibid. at 322-23.

27. I place the term “abuse” in quotation marks to indicate that the concern lies not with the failure to adhere to the standards for creating legal rules, but rather with the purposes for which officials use their powers to do so.

28. D’Amato also reads Hart’s denial of a necessary connection between law and morality as implying that international law cannot incorporate moral principles, a conclusion he finds problematic in the case of international humanitarian law. As Hart later made clear in his postscript to the second edition of Concept of Law (which, it should be acknowledged, was published several decades after D’Amato’s article), he thinks it is possible, but not necessary for a legal system to treat morality as a source of law. See Hart, H.L.A., The Concept of Law, 2nd ed. (New York: Oxford University Press, 1994)Google Scholar. Thus, Hart’s theory of law does not commit him to excluding as sources of international law the moral principles constitutive of jus ad bellum and jus in bello. Whether they are in fact sources of international law depends on the rule of recognition to which officials in the international legal system adhere, or so I argue elsewhere. See “Inclusive Legal Positivism, International Law, and Human Rights” [unpublished, draft on file with author].

29. D’Amato, supra note 23 at 322. No such implication follows from Hart’s theory, and D’Amato’s cautious word-choice may indicate an awareness of this.

30. Interestingly, D’Amato, adopts Hart’s analysis of law and a legal system in his book The Concept of Custom in International Law (Ithaca, NY: Cornell University Press, 1971)Google Scholar. However, while he refers to secondary rules a number of times, D’Amato does not use the term ‘rule of recognition’, and refers several times to the article discussed in the text.

31. See Franck, supra note 20 at 183-94.

32. In doing so, he transforms Hart’s rule of recognition into something akin to Kelsen’s Grundnorm. Yet, as was indicated above, Hart attributes law’s normativity (as a descriptive matter) to the adoption of the internal point of view with respect to (particular) legal norms, rather than to the existence of a basic norm of the legal system. In fairness to Franck, it should be noted that on a number of occasions in the chapter on international law, Hart speaks as if the concepts of a Grundnorm and a rule of recognition are equivalent.

33. Ibid. at 184.

34. Ibid. at 187.

35. Compare to Franck’s remark that the international community is defined by an ultimate rule of recognition that specifies the sources of members’ obligations. Ibid. at 190.

36. van Hoof, G.J.H., Rethinking the Sources of International Law (Deventer: Kluwer, 1983) at 44–56 Google Scholar.

37. Ibid. at 53; see also Walden, supra note 10 at 90.

38. Van Hoof’s complaint is somewhat odd, since Hart conducts no search for the rule of recognition in either the English or American legal systems, but rather offers possibly oversimplified candidates for such rules, which at least initially appear to function as required by Hart’s theory.

39. See Hart, supra note 7 at 89-90.

40. See van Hoof, supra note 36 at 55.

41. Ibid. at 54. The quotation from Hart is in Hart, supra note 7 at 222.

42. See Hart, supra note 7 at 230.

43. Indeed, John Tasioulas has suggested to me that Hart may have written the chapter on international law only because Austin and Kelsen discussed it.

44. Ibid. at 231.

45. Ibid. at 231.

46. D’Amato, gives the same interpretation of Hart’s remark in The Concept of Custom Google Scholar; see D’Amato, supra note 30 at 28-29.

47. Though as was noted earlier, D’Amato correctly points out that this change would not necessarily entail an all-encompassing rule of recognition; that is, one that validated and systematized all international legal norms.

48. See Hart, supra note 7 at 229.

49. The parallel here would be to argue that an official of a domestic legal system ought not to adhere to a rule of recognition that directs him to treat legislative enactments as legally valid if the laws enacted are unjust, e.g., enslave a portion of the population.

50. Besides those mentioned in this paper, other discussants of the chronological paradox include Kelsen, Hans, Principles of International Law, 2nd ed. (New York: Holt, Rinehart and Winston, 1966)Google Scholar; Finnis, John, Natural Law and Natural Rights (Oxford: Oxford University Press, 1980)Google Scholar; and Tasioulas, John, “Customary International Law and the Quest for Global Justice” in Perreau-Saussine, A. & Murphy, J.B., eds., The Nature of Customary Law: Philosophical, Historical and Legal Perspectives (New York: Cambridge University Press, 2007) at 307–35 Google Scholar. Unfortunately, considerations of space preclude engagement with their views here.