1. H. L. A. Hart, The Concept of Law (2d ed. 1994), at 168.

2. This is Simon Blackburn echoing Jerry's Fodor's misgivings about the label “normativity.” Blackburn, Simon, Majesty of Reason , 85 Philosophy 5 (2010)CrossRefGoogle Scholar.

3. The terms “bindingness” and “normativity” are used interchangeably here, as they typically are in the metaethical literature. See, e.g., Moore, A.W., Categorical Imperative? , 63 Ethics 235, 237 (1953)CrossRefGoogle Scholar.

4. Luke Russell, Two Kinds of Normativity: Korsgaard v. Hume, in Hume on Motivation and Virtue (Charles Pigden ed., 2009), at 208–225, 208–209.

5. Swaminathan, Shivprasad, Projectivism and the Metaethical Foundations of the Normativity of Law , 7 Jurisprudence 231 (2016)CrossRefGoogle Scholar. In this wide semantic sense of “normative,” a multitude of standards are normative, among which are moral standards.

6. David Copp, Morality, Normativity and Society (1995), at 10.

7. This is discussed in greater detail in Swaminathan, supra note 5. A putative account of normativity of law cannot just stack up any semantic model on top of a model of moral bindingness without regard to their compatibility (see Section II infra). Here, our focus will primarily be on the second limb of the project of accounting for the normativity of law, i.e., that of moral bindingness.

8. For a discussion on the role of metaethics in understanding the “normativity of law” and in legal theory in general, see S. Delacroix, Legal Norms and Normativity: An Essay in Genealogy (2006), at xxiv.

9. Simon Blackburn, Ruling Passions: A Theory of Practical Reasoning (1998), at 84–91.

10. Simon Blackburn, Spreading the Word: Groundings in the Philosophy of Language (1984), at 181–182. The expressions “impinge” and “project” are borrowed from Blackburn. See Section II for an elaboration on these pictures of morality.

11. “I have the dim. . .outline of this book in my mind. . .One side saying: you are blind. . . The other: you are seeing ghosts.” This is from Hart's notebook, which was to become The Concept of Law. Nicola Lacey, A Life of H. L. A. Hart: The Nightmare and the Noble Dream (2004), at 222.

12. See infra note 18 for a precis of Hart's semantic project.

13. H. L. A. Hart, Essays on Bentham: Studies in Jurisprudence and Political Theory (1982), at 267.

14. The tangle Hart finds himself in is more in the nature of a conundrum than a paradox in the strict sense, but we will continue to use Hart's terminology here. The metaethical problem persists regardless of what we may choose to call it.

15. Hart, supra note 13, at 159–160.

16. Does this mean that the version of the rule of recognition that was not congruous with the impinging model was congruous with the projectivist model? No such assumption can be made and indeed it is not being made here. See discussion in note 59 infra and text accompanying it.

17. The choice of candidates is by no means arbitrary. As well as being instructive representatives of the two models of normativity, the two also weave into the narrative of this paper organically. The choice of the Scandinavian Legal Realists requires no justification, as they were, for a long time, the only adherents of the projectivist model of normativity. Perhaps a little more needs to be said in defense of Kelsen's invocation as an exemplar of the impinging model before we proceed any further. Roughly, Kelsen's Grundnorm is read here (see Section II for discussion) as taking the place of the objective moral requirement in the standard impinging model and acting as the fount of normativity and moral bindingness. This reading of Kelsen is, of course, far from universally accepted. For an instructive overview of the debate about Kelsen's normativity, see Julie Dickson, Interpreting Normativity, in Properties of Law: Essays in Honour of Jim Harris (Timothy Endicott, Joshua Getzler & Edwin Peel eds., 2006), at 22–42. The reading of Kelsen that presents his model of normativity along impinging lines was suggested by the Scandinavian Legal Realists (Karl Olivecrona and Alf Ross, in particular) and later endorsed by Joseph Raz. The impinging elements in Hart's model will be seen to bear resemblance to those found in Kelsen's account, and the rule of recognition in its impinging avatar seems to operate just as the Kelsenian Grundnorm did—or at any rate, as the Scandinavian Legal Realists and Raz thought the Grundnorm did. Indeed, a form of the paradox that Hart finds himself with was identified by Euginio Bulygin (see discussion in Section II, particularly note 58 infra) as the very antinomy inherent in Kelsen's work.

18. A “semantic” project is not only concerned with language but also has something to say about how language relates to “reality.” As discussed in note 19 infra, Hart's semantic model in The Concept of Law has both these elements and explains how legal statements convey what is required by extant social practices.

19. Swaminathan, Shivprasad, Toh's Hartian Project: Expressivist Turn, Relativist Returns , 39 Austl. J. Legal Phil. 87 (2014)Google Scholar.

Id. at 104–105. Norm-relativist statements are “normative” in the wide sense of “normative” but are distinct from moral statements. This semantic model allowed Hart to steer a middle path between reductionism on the one hand and the idea that legal requirements are normative in the same sense as moral requirements on the other. In The Concept of Law Hart never goes on to pose the further question of how legal requirements conveyed by statements of legal obligation may come to be morally binding.

20. Toh, Kevin, Hart's Expressivism and His Benthamite Project , 11 Legal Theory 75 (2005)CrossRefGoogle Scholar; Toh, Kevin, Raz on Detachment, Acceptance and Describability , 27 Oxford J. Legal Stud. 403 (2007)CrossRefGoogle Scholar; Toh, Kevin, Legal Judgments as Plural Acceptance of Norms , 1 Oxford Stud. Legal Phil. 107 (2011)CrossRefGoogle Scholar; Raz, Joseph, H. L. A. Hart (1907–1992) , 5 Utilitas 148 (1993)CrossRefGoogle Scholar.

21. This project does not, therefore, occupy the same field or tackle the same set of issues surrounding Hart's normativity as explored by Toh. For a discussion of Toh's semantic project and an alternative “norm-relativist” reading of Hart's semantic project, see Swaminathan, supra note 19.

22. For a more elaborate fleshing out of the metaethical positions discussed briefly here, see Swaminathan, supra note 5.

23. Aiken, Henry David, Authority of Moral Judgments , 12 Phil. & Phenomenological Research 513 (1952)CrossRefGoogle Scholar, and Evaluation and Obligation: Two Functions of Judgments in the Language of Conduct, 47 J. Phil. 5 (1950).

24. Blackburn, supra note 10, at 181.

25. Derek Parfit, Normativity, in Oxford Studies in Metaethics: Volume 1 (Russ Shafer-Landau ed., 2006), at 325–380, 352. As Parfit, who endorses a cognitivist account, points out, it is the moral fact that gives moral reasons for action:

Id. (emphasis added).

26. Moral realism postulates the existence of ontologically given moral facts whose existence is independent of our beliefs about what is right and wrong. David Brink, Moral Realism and the Foundations of Morality (1989), at 7, 14.

27. Constructivists differ from realists only on ontology insofar as they hold that the moral reality is not ontologically given but something that is “constructed.” Russ Shafer-Landau; Moral Realism: A Defence (2003), at 14.

28. Voluntarism is a cognitivist position as it claims that moral judgments purport to represent truth. However, the truth is fixed by the say-so of some entity. Christine Korsgaard, Sources of Normativity (1996), at 21.

29. Two assumptions make up noncognitivism: semantic nondescriptivism, the idea that moral judgments do not purport to describe moral facts, and psychological noncognitivism, the idea that moral judgments do not involve a cognitive state of mind such as belief.

30. See discussion in Toh, Kevin, Hart's Expressivism and His Benthamite Project , 11 Legal Theory 75, 79 (2005)CrossRefGoogle Scholar. This follows from the so-called “standard picture of psychology,” endorsed by cognitivists and noncognitivists alike, which dichotomizes mental states into motivationally inert cognitive states (beliefs) and motivationally charged conative or noncognitive states (attitudes and sentiments). Michael Smith, Moral Problem (1988), at 7–9.

31. The image of “spreading” and “projecting” normativity is Simon Blackburn's. Blackburn, supra note 10, at 181.

32. Simon Blackburn, Practical Tortoise Raising and Other Philosophical Essays (2010), at 32.

33. Swaminathan, supra note 5; Nigel Simmonds, Central Issues in Jurisprudence (3d ed. 2008), at 149.

34. See Sections II & III.E infra.

35. Toh sketches a semantic model that could be congruous with a projectivist model of normativity of law. See notes 20 & 21 supra and accompanying text; for a “norm-relativist” alternative, which could also be congruous with the model, see Swaminathan, supra notes 19 & 21.

36. In Kelsen's account, it was the Grundnorm that endowed all other norms in the legal system with normativity, validity, and bindingness. Kelsen argued, “to say that the behaviour of an individual is commanded by an objectively valid norm amounts to the same as saying that the individual is obligated to behave in this way.” Hans Kelsen, General Theory of Law and State (1961), at 30.

37. Alf Ross, Validity and the Conflict between Positivism and Natural Law, in Normativity and Norms: Critical Perspectives on Kelsenian Themes (S.L. Paulson & B.L. Paulson eds., 1998), at 147–164, 150. As I have already noted earlier in note 17 supra, this reading of Kelsen is contested; for an alternative reading, see S.L. Paulson, A “Justified Normativity” Thesis in Kelsen's Pure Theory of Law? Rejoinders to Robert Alexy and Joseph Raz, in Institutionalized Reason: The Jurisprudence of Robert Alexy (Matthias Klatt ed., 2012), at 61–111. But whether or not Kelsen can exegetically be read along impinging lines need not worry us, for the only reason it interests us here is its potential to illustrate the working of the impinging model of normativity. Throughout this article, therefore, when we refer to Kelsen's model of normativity it is this Ross's Kelsen—also endorsed by Olivecrona and Raz—that we will have in mind.

38. Id. at 153–154 (emphasis added).

39. Id. at 154.

40. Olivecrona, Karl, Realism and Idealism: Some Reflections on the Cardinal Point in Legal Philosophy , 26 N.Y.U. L. Rev. 125 (1951)Google Scholar. Raz's reading of Kelsen's normativity as “justified normativity” is of a piece with Ross's and Olivecrona's. Joseph Raz, Authority of Law (1979), at 134; Joseph Raz, The Purity of the Pure Theory, in Paulson & Paulson eds., supra note 37, at 237–252, 245. For a discussion on justified normativity, see Swaminathan, supra note 5.

41. Max Lyles, A Call for Scientific Purity: Axel Hagerstrom's Critique of Legal Science (2006), at 312.

42. A note of caution may not be out of place here as a great many positions in philosophy go under the label voluntarism. See Simon Blackburn, Oxford Dictionary of Philosophy (1996), at 397. In this paper, voluntarism will refer to the position that states that obligations are created by the will or say-so of a god or godlike sovereign.

43. Korsgaard, supra note 28, at 21.

44. The post-enlightenment versions of voluntarism such as Hobbes's tended to be secular, where the truth-making entity was an earthly entity such as the sovereign.

45. Voluntarism has a much older history. Earlier versions of voluntarism were theological in tone with the truth-making entity being God.

46. Korsgaard, supra note 28, at 21.

47. Id.; J.B. Schneewind, Essays on the History of Moral Philosophy (2010), at 202–205; J.B. Schneewind, The Invention of Autonomy (1998), at 95–100.

48. Korsgaard, supra note 28, at 23. There could be one further question posed here: Where does this sovereign get its authority from? This question might be seen to be an invitation to infinite regress for the voluntarist. All forms of voluntarism face this challenge, called Euthyphro's dilemma. See note 51 infra and text accompanying it for discussion of Euthyphro's dilemma. Hobbes too faced this problem. His solution was to base the sovereign's power on a social compact. Samuel Clarke, however, felt that this was not enough to deflect the Euthyphro's dilemma. Korsgaard, supra note 28.

49. Olivecrona, supra note 40, at 125.

50. Olivecrona uses the “chameleon-hued” word “right” somewhat loosely. What he means by “right” is, strictly speaking, “authority” in moral terms or “power” in Hohfeldian terms. Hohfeld, Wesley N., Some Fundamental Legal Conceptions as Applied in Judicial Reasoning , 23 Yale L.J. 16 (1913)CrossRefGoogle Scholar. Olivecrona's emphasis here is on the power to change the normative situation of those subject to the lawgiver's commands presupposed by the voluntarist picture.

51. These are versions of the so-called Euthyphro's dilemma. See John Gardner, Law as a Leap of Faith (2012), at 1.

52. Korsgaard, supra note 28, at 20–28. Korsgaard describes what we have outlined here as the problem of infinite regress of justification facing voluntarist accounts. Korsgaard does not consider Kelsen's account of “normativity of law,” but her line of criticism against the voluntarists could be easily extended to Kelsen as well.

53. Olivecrona, supra note 40, at 128.

54. Lyles, supra note 41, at 313.

55. Swaminathan, supra note 5. For an assessment of the project of “naturalizing” normativity, see Brian Leiter, Naturalizing Jurisprudence (2007) passim.

56. Lyles, supra note 41, at 320; Patricia Mindus, A Real Mind: The Life and Work of Axel Hagerstrom (2010) passim.

57. Richard Mervyn Hare, Sorting Out Ethics (1997), at 96. Patricia Mindus too notes that Hagerstrom was a precursor to twentieth-century noncognitivists such as A.J. Ayer and Charles Stevenson. Mindus, supra note 56, at 79.

58. Euginio Bulygin identifies an antinomy in Kelsen's model. Kelsen's model of normativity, Bulygin assumes, operates along the lines of the natural lawyers’ model. As such, Bulygin's reading of Kelsen is of a piece with those of Olivecrona, Ross, and Raz. Bulygin then points out how this conception of normativity sat oddly with Kelsen's semantics of statements of legal obligation. The statement of legal obligations congruous with Kelsen's model of normativity was the cognitivist one, namely, that statements of legal obligation purport to state moral reasons for action. The legal statements would be truth-apt in relation to such moral reasons. However, Bulygin observed that for Kelsen, statements of legal obligation were prescriptions that were not truth-apt. One consequence to follow from this was that they did not purport to denote any objective moral properties. This, he argued, created an antinomy in his account. See Eugenio Bulygin, An Antinomy in Kelsen's Pure Theory of Law, in Paulson & Paulson, eds., supra note 37, at 297–316, 298–300.

59. This is not, however, to suggest that, without anything more, this version of the rule of the recognition is, in and of itself, sufficient to yield a projectivist account of normativity of law. The claim here is the modest one that the version of the rule of recognition that holds the field in The Concept of Law is not incongruous with the projectivist model, unlike the versions to be found in Legal and Moral Obligation and Essays on Bentham.

60. Hart, supra note 1, at 59–60.

61. Id. at 58.

62. Id. at 60.

63. Id. at 58 (original emphasis).

64. Id. (original emphasis).

65. It will be argued in Section III.E infra that such a “rule of recognition” having a “justificatory” role would be congruous with an impinging model of normativity of law alone, as the obligation to obey imposed by the rule of recognition cannot but be understood as a moral obligation, despite Hart never claiming, in so many words, that it is a moral obligation.

66. Hart, supra note 1, at 96.

67. Had Hart persisted with this impinging model of the rule of recognition in the remainder of The Concept of Law, it would have been the case that legal rules deriving their validity from the rule of recognition provide authoritative moral reasons for action; this would have sat oddly with his norm-relativist semantics of statements of legal obligation discussed in note 19 supra where he viewed the legal “ought” as distinct from the moral “ought,” and this would have flown in the face of his long held view that statements of legal obligation have nothing whatsoever to do with moral obligations. See De Paramo, Juan Ramon, Entrevista a HLA Hart , 5 Doxa 340 (1988)Google Scholar, discussed in Timothy Endicott, Law and Language, in Stanford Encyclopedia of Philosophy (Edward N. Zalta ed., 2010), http://plato.stanford.edu/archives/fall2010/entries/law-language/. Hart would then, perhaps, have had to face the paradox he did in Essays on Bentham, twenty years earlier, in The Concept of Law itself. As we shall see, it is this long held commitment he finds his model of normativity of law in Essays on Bentham to go against.

68. H. L. A. Hart, Legal and Moral Obligation, in Essays in Moral Philosophy (I.E. Melden ed., 1958), at 82–107, 91–93.

69. Id. at 91.

70. Id.

71. It will be argued in Section III.E infra that this version of the basic norm purports to impose a moral obligation of obedience on the citizen and is congruous with the impinging model alone.

72. Hughes's intervention is termed “fortuitous” because even though it had nothing to do with the impinging nature of the rule of recognition, in response to Hughes's intervention, Hart unwittingly ended up significantly altering the nature of the rule of recognition by dropping its impinging elements.

73. Hughes, G., The Existence of a Legal System , 35 N.Y.U. L. Rev. 1001 (1960)Google Scholar.

74. Hart, supra note 1, at 60. Roger Shiner argues that Hart accepts Hughes's criticism as sound, albeit implicitly, and his move of breaking the acceptance requirement between citizens and officials was in response to Hughes's objection. Roger Shiner, Norm and Nature (1992), at 162.

75. Hart, supra note 1, at 60.

76. Hart perhaps placed too much emphasis on “knowledge” or cognizance of the rule of recognition being the crucial factor that set simple legal systems apart from complex ones. While it is true that it is easier for a citizen in a simple legal system of Rex to know the rule of recognition than it is in a complex legal system, most citizens in a complex legal system do nevertheless know what the law is, and by implication would have at least a working knowledge of the rule of recognition, though would not know of its specifics. Knowledge of something is not a binary all-or-nothing idea that admits of only two extremes. There is a sliding scale between knowledge and ignorance that Hart does not appear sufficiently sensitive to. Later in The Concept of Law Hart does admit that citizens in even in a complex legal system do know what the law is and utter internal legal statements—which would be impossible without having a working knowledge of the rule of recognition. Hart, supra note 1, at 114.

77. Hart, supra note 1, at 96.

78. On the contrast between the rule of recognition and the Grundnorm on this point see Grant Lamond, The Rule of Recognition and the Foundations of a Legal System, in Reading HLA Hart's ‘The Concept of Law’ (Luis Duarte D'Almeida, James Edwards & Andrea Dolcetti eds., 2013), at 97–112, 114.

79. Hart thought that the notion of content independent peremptory reasons was central to his account of “normativity of law”: “I certainly think. . .that the notion of a content independent peremptory reason for action is needed. . .to explain the ‘normativity of law’.” Hart, supra note 13, at 262.

80. Id. at 244 (parentheses and emphasis added).

81. Id. at 254. Hart also reminds us here that he had first introduced the idea of “content independent” reasons in Legal and Moral Obligation.

82. Id. at 253.

83. Id. at 258 (emphasis added).

84. Here, in this new version, the “recognition of the commander's words as constituting content independent reasons,” does the very job that “acceptance of the rule of recognition” by the citizen does in the Rex model. Indeed, Hart uses the terms “recognition,” “acceptance,” and “normative attitude” interchangeably.

85. In fact, given his metaethical ambivalence, Hart would have been loath to expressly refer to such obligation of obedience as a moral obligation.

86. Hart, supra note 13, at 253.

87. Id. at 267.

88. Hart argues in Chapter VI of Essays on Bentham, id. at 159:

In this passage, the terms “cognitive” account and “noncognitive” are used by Hart somewhat idiosyncratically to refer to the semantic model of statements of legal obligation. A cognitive account, such as Raz's, is one in which statements of legal obligation state moral reasons, and a noncognitive account is one that, like his, denies that connection. Hart here is merely confirming his long held view that statements of legal obligation do not convey or purport to convey any moral obligation. See Hart, supra note 13, at 144–145.

89. Id.

90. Hart remained steadfast in this commitment throughout his life. See supra note 67.

91. Hart, supra note 13, at 267.

92. Bulygin, supra note 58.

93. Hart, supra note 13, at 265.

94. Hart maintains that the argument—that his model requires an ulterior moral standard—ultimately “fails.” Id. It is fairly obvious from the discussion, id., that Hart seeks to tackle the objection that his account needs an ulterior moral principle as truth-maker elliptically by dismissing the argument that judges need to believe in the moral legitimacy of the law while acting in their capacity as judges. But this hardly amounts to an engagement with the crucial metaethical question he is purporting to respond to. All arguments that go into establishing the proposition that judges need not accept the commander's words as peremptory reasons at best only provide phenomenological support to the semantic thesis that statements of legal obligation on the lips of judges do not purport to state any objective moral reasons for action. However, it was this semantic thesis that was in direct contrast to the central theme of the chapter that legal directives were content independent peremptory reasons for action—and hence, the manifestation of the paradox. As we shall see in Section V, when he (mis)diagnoses the paradox in his account, he can be seen grappling with the same issue, albeit with the logic of judges morally accepting the laws.

95. Hart, supra note 13, at 256 (emphasis added). Note, here Hart is talking about the case of the simple legal system. The case of the complex legal system is discussed in text accompanying notes 103 and 104 infra.

96. In one place, Hart says “. . . to have such authority is to have one's expressions of intentions as to the actions of others accepted as peremptory content independent reasons for action.” Hart, supra note 13, at 256.

97. Having reasons to do something, in the impinging model, is tantamount to there being an objective moral standard that requires action. It is the objective moral standard that gives reasons. If the idea of an objective moral standard as the source of normativity is dropped, the idea of reason-givingness would also have to be dropped along with it.

98. Furthermore, even the idea of “acceptance” by citizens being the ground or source of normativity is at odds with an impinging model. The impinging model understands authority in terms of objective moral standards that give reasons for action. It is the objective moral standard that is the source of moral bindingness that gives reasons for action, regardless of the moral acceptance of the subject. It would be incongruous for a purported impinging model to rest the source of normativity on the moral “acceptance” of the subject (where moral acceptance is understood in terms of cognitive belief). The moral belief aims to track true moral standards and it is this true moral standard, not the belief, that is the source of normativity. Thus, a moral belief could never be the source of normativity in the impinging model. In any event, Hart rules out a cognitive understanding of moral acceptance as he categorically refers to the moral acceptance in his model as an “attitude.” An attitude is not even truth-apt and does not purport to track any moral standards.

99. This eclecticism was probably caused by Hart's having misunderstood the metaethical configuration and method of operation of the projectivist account.

100. The projectivist model, it will be recollected, does not understand moral bindingness in “justificatory” terms as the impinging model does. Rather, it understands moral bindingness in motivational, attitudinal terms. Consequently, the projectivist model explains moral bindingness without invoking the idea of reason-givingness or a lawgiver's moral right to issue directives: ideas that make sense only on a justificatory understanding of normativity and necessarily presuppose a truth-maker.

101. As Grant Lamond points out, influenced by the emphasis of his successors, the reference to “officials” in general in The Concept of Law becomes reference to “judges” in Essays on Bentham. Lamond, supra note 78, at 111.

102. Hart, supra note 13, at 267.

103. Id. at 257, 258.

104. What we find here is in many respects similar to the situation in which Hart had found himself in The Concept of Law in the transition from the simple to the complex system. In response to Hughes's objection, Hart conceded that while the citizens in the simple legal system of Rex may “accept” that what “Rex says ought to be done” and thus accept the authority of the lawgiver, in a complex legal system, they are in no position to do so. Consequently, he also ended up significantly altering the rule of recognition in the remainder of The Concept of Law from an impinging one to a nonimpinging one. As a consequence, when a judge made a statement of legal obligation, he did not thereby state any peremptory reasons that a citizen has for action. Here, in Essays on Bentham, a problem similar to the one that Hart faced in The Concept of Law in the transition from the simple to the complex system seems to recur. However, in Essays on Bentham, in response, Hart does not deprive the rule of recognition of its impinging nature (as he had done in The Concept of Law).

105. Hart, supra note 68, at 92.

106. Lacey, supra note 11, at 228.

107. This Rex model was in many respects identical to the Rex model in Legal and Moral Obligation.

108. For an illuminating discussion of Hart's metaethical commitments, see John Finnis, On Hart's Ways: Law as Reason and Fact, in The Legacy of HLA Hart: Legal, Political and Moral Philosophy (Matthew Kramer, Claire Grant, Ben Colburn & Antony Hatzistavrou eds., 2008), at 3–27.