1. This seems to be how most participants in the debate understand theoretical disagreement: see, e.g., Brian Leiter, “Explaining Theoretical Disagreement” (2009) 76 U Chicago L Rev 1215 at 1216; Shapiro, Scott J, “The ‘Hart-Dworkin’ Debate: A Short Guide for the Perplexed” in Ripstein, Arthur, ed, Ronald Dworkin (New York: Cambridge University Press, 2007) 22 at 40.Google Scholar I have argued elsewhere that theoretical disagreement includes, but is not limited to, disagreement about the criteria of legal validity: Dale Smith, “Theoretical Disagreement and the Semantic Sting” (2010) 30 Oxford J Legal Stud 635 at 644. However, for present purposes, I shall leave this point to one side.

2. For a more detailed presentation of the argument from theoretical disagreement, see Smith, ibid.

3. I shall, for short, refer to Dworkin’s “argument from theoretical disagreement” as the ATD. The fullest presentation of the ATD is in Chapter 1 of Law’s Empire: Dworkin, Ronald, Law’s Empire (Oxford: Hart, 1998) ch 1.Google Scholar

4. Shapiro, supra note 1 at 50.

5. Leiter, supra note 1 at 1249.

6. Ibid at 1227. Matthew Kramer has made a similar point: Kramer, Matthew, In Defense of Legal Positivism: Law without Trimmings (Oxford: Oxford University Press, 1999) at 136, 141–44.Google Scholar However, Kramer appears to be interested in agreement in legal judgments only in so far as it provides a basis for responding to the ATD, whereas Leiter is also interested in it as a distinct legal phenomenon that theories of law should seek to explain.

7. Leiter, supra note 1 at 1247 [emphasis in original].

8. Ibid. Leiter may have retreated from this claim in a more recent, unpublished article: Brian Leiter, “Why Legal Positivism (Again)?” University of Chicago Public Law Working Paper No. 442 (September 9, 2013) online: Social Science Research Network http://ssrn.com/abstract=2323013. In that article, the ability to explain agreement in legal judgments is presented as merely part of one of three reasons why positivism is so popular among legal philosophers. However, at 11, he describes agreement in legal judgments as “the most important fact about modern legal systems” (since, if there was not massive and pervasive agreement in legal judgments, modern legal systems would collapse under the weight of legal disputes).

9. Leiter, supra note 1 at 1249 (describing theoretical disagreements as “relatively marginal phenomena within the scope of a general theory of law”).

10. Ibid at 1220-22.

11. Hart, HLA, The Concept of Law, 3rd ed (Oxford: Clarendon, 2012) at 108–10.Google Scholar Hart adds that officials must also share a certain attitude towards this pattern of behaviour, but my focus will be on the pattern of behaviour itself.

12. This is not to imply that other versions of positivism cannot account for agreement in legal judgments (or that they are not vulnerable to the ATD). However, their explanation of agreement in legal judgments will differ from Hart’s (as will the basis, if any, on which they are vulnerable to the ATD).

13. For example, partly in response to the ATD, Shapiro has developed a form of positivism which denies that the criteria of legal validity are fixed by convergence among officials: Shapiro, supra note 1 at 43-49; Shapiro, Scott J, Legality (Cambridge: Belknap Press, 2011).Google Scholar

14. It is controversial whether the discussion of law in Dworkin’s penultimate book, Justice for Hedgehogs, is compatible with the theory of law presented in Law’s Empire. See Dworkin, Ronald, Justice for Hedgehogs (Cambridge: Belknap Press, 2011) at 402Google Scholar, 485 note 1; Jeremy Waldron, “Jurisprudence for Hedgehogs” NYU School of Law Public Law Research Paper No. 13-45 (July 5, 2013), online: Social Science Research Network http://ssrn.com/abstract=2290309 at 6-8. Be that as it may, my concern is with the theory of law presented in Law’s Empire.

15. I am grateful to Arie Rosen for a very helpful discussion of the issues canvassed in this paragraph.

16. Leiter, supra note 1 at 1227 [emphasis in original, footnote omitted].

17. Ibid at 1226-27 [emphasis in original, footnote omitted]. As parts of this passage hint, Leiter points to the existence of massive and pervasive agreement in legal judgments as indicating that theoretical disagreement is relatively uncommon, being confined primarily to appellate courts. I say more about this in Section 4.

18. Ibid at 1227.

19. Ibid at 1228 [emphasis in original].

20. Ibid at 1248 [footnote omitted].

21. Ibid at 1247 [emphasis in original].

22. He provides some empirical evidence from the United States to support this claim: ibid at 1226 note 54.

23. Though I suggest in Section 3 that this explanation is incomplete.

24. Cf the discussion of Riggs v Palmer 115 NY 506 (1889) in Section 5. For illustrative purposes, I assume that intentions can be ascribed to the legislature as a whole.

25. Even if judges generally consider the plain meaning of the words contained in statutes, this is not enough to show that there is the convergence on the relevant criterion of legal validity required by positivism. For one thing, there are many different roles that the plain meaning of a statute could play in determining what contribution the statute makes to the content of the law. (Even law as integrity can allow that it plays some role: see infra note 31.) On the version of positivism that Leiter seeks to defend, there must be convergence among officials on the particular role that the plain meaning of the statute plays.

One could imagine a different version of positivism, which claims that convergence among officials determines only what the sources of law are, and that there is a non-convergence-based way of getting from the sources of law to the content of the law. For example, convergence among officials might determine that statutes are a source of law, but not how statutes are to be interpreted—theories in philosophy of language and linguistics might settle the question of interpretation. I hope to show in future work that, while this view avoids some of the problems with the version of positivism that Leiter defends, it encounters other significant problems.

26. Leiter offers a different objection to law as integrity that is based on the possibility of moral error: see infra note 31.

27. It seems plausible to suggest that, when seeking to explain the fact that relatively few potential legal disputes result in formal litigation (and so on), one should be particularly concerned with whether there is agreement among officials such as judges and lawyers. In most cases, the parties will consult lawyers when deciding whether to litigate, and the existence of agreement among judges may deter litigation even in circumstances where there is disagreement within the broader community.

28. If different officials reach the same conclusions for different reasons, would we not expect them to mention this? Not necessarily. Provided their fellow officials agree with them about the outcome, there may be little incentive to engage in explicit disagreement about the basis on which that outcome should be reached. (As Lael Weis pointed out to me, this is especially so in jurisdictions in which concurring judgments are discouraged.) There may also be little opportunity for them to engage in such disagreement. On the view Leiter presents, it is less likely that a matter will come before the courts if there is agreement about the appropriate legal outcome.

29. Indeed, this is one of the moves Leiter makes when responding to the ATD. He acknowledges that Dworkin might argue that legal officials hold different theories about how to resolve legal disputes and yet agree about the outcome of most disputes, but queries whether we have any reason to suppose that this is in fact the case: Leiter, supra note 1 at 1228-29.

30. Dworkin, supra note 3 at 255.

31. Admittedly, there may be some disagreement about what counts as a “data point”. For example, do the data points include the legislative history of a statute? (For Dworkin’s preferred answer, see ibid at 343-45.) However, any plausible interpretation of our legal practice must accept that, at least, orders made by courts and the text of statutes enacted by the legislature count among the data points, and so it remains the case that there is a shared set of data points, even if some legal interpreters recognize further data points that others do not.

Incidentally, the point made in the text may help to allay a further concern of Leiter’s—namely, that law as integrity renders the law “esoteric”, because it entails that we might fail to identify most (perhaps all) of the law in our community, since we might fail to identify the moral principles that form part of the law and by reference to which concrete legal decisions are to be reached (Brian Leiter, “Book Review” (2006) 56 J Legal Educ 675 at 675). Even if there are widespread misunderstandings of the moral principles that form part of the law on Dworkin’s account, we may nevertheless know quite a bit of the law of our community, since there may be some data points that any plausible interpretation will fit (think, for example, of much of the traffic laws). In any case, unless Leiter’s concern is premised on his rejection of moral cognitivism, one might expect him to allow that people are sometimes capable of moral knowledge, including about some of the moral principles relevant under law as integrity.

32. Dworkin, supra note 3 at 256.

33. Ibid at 255-56.

34. This appears to be true of the jurisdiction with which I am most familiar. With regard to statutory interpretation, the High Court of Australia typically attaches a great deal of weight to the words of the statute, so as to respect the constitutional separation of judicial and legislative power: see, e.g., Project Blue Sky Inc v Australian Broadcasting Authority (1998) 194 CLR 355 at [78]; Thiess v Collector of Customs [2014] HCA 12 at [22]; Zheng v Cai (2009) 239 CLR 446 at [28]. Similarly, while there are notable exceptions, it has also generally taken a conservative view towards judicial reform of the common law: see Momcilovic v R (2011) 245 CLR 1 at [392]-[398].

35. Thanks to John Tasioulas for helpful comments on this point.

36. Blake, Susan, Browne, Julie & Sime, Stuart, A Practical Approach to Alternative Dispute Resolution, 2nd ed (Oxford: Oxford University Press, 2012) at 13–15.Google Scholar While Leiter acknowledges that there are other reasons (apart from agreement in legal judgments) why particular cases are or are not litigated or appealed, he does not mention this consideration.

37. Leiter, supra note 1 at 1228-29. He does not, however, consider the features of law as integrity (especially the operation of the dimension of fit) that support such an account in the way suggested in Section 2.

38. Ibid at 1229 [emphasis in original]. Leiter prefaces this statement by indicating that he is assuming, for the purposes of this response, that Dworkin is right to take legal discourse at “face value”.

39. There is another possible reading of this passage—namely, that Leiter is simply querying whether Dworkin has provided any affirmative evidence in support of the suggestion that there is widespread theoretical disagreement underlying the massive and pervasive agreement in legal judgments. Indeed, Leiter starts the relevant paragraph with this query. However, the passage quoted in the text seems to go further, by asserting that “agreement is agreement” and that only someone already committed to law as integrity would accept the Dworkinian explanation of agreement in legal judgments.

40. In presenting this objection, I draw on the account of the theoretical virtue of simplicity sketched in Dale Smith, “Must the Law Be Capable of Possessing Authority?” (2012) 18 Legal Theory 69 at 91. Note that Kramer also argues that positivism offers a simpler explanation of agreement and disagreement about law than does Dworkin, though for a somewhat different reason than the one suggested in the text: Kramer, supra note 6 at 139-40.

41. Leiter, supra note 1 at 1224.

42. Ibid at 1239.

43. Ibid [emphasis in original].

44. Ibid at 1247.

45. Ibid at 1226 [emphasis in original].

46. See, e.g., ibid at 1226 (claiming that the significance that Dworkin attaches to the phenomenon of theoretical disagreement is puzzling, given the relative rarity of such disagreement). This is not the only basis on which Leiter questions the emphasis that Dworkin places on the ability of a theory of law to account for theoretical disagreement. For example, he also contends that the onus is on Dworkin to show that theoretical disagreement is a central feature of law, and that Dworkin has not discharged this argumentative burden: ibid at 1220.

47. Ibid at 1225.

48. Ibid at 1232.

49. Ibid at 1226 [emphasis in original].

50. As mentioned in supra note 46, this is not the only basis for the contention.

51. Leiter’s view seems to be that we appeal to these virtues only where there is more than one explanation that “makes sense” of the relevant phenomenon: Leiter, supra note 1 at 1239-40. Of course, the idea of an explanation “making sense” of a phenomenon could do with elaboration, but Leiter claims that an intuitive understanding of the idea will suffice in the present context. I say a bit more about this below.

52. At least, this is his official position. In Section 3, I suggested that he sometimes appears to allow that law as integrity can explain agreement in legal judgments, while claiming that its explanation is less straightforward than positivism’s. I also suggested that one way of understanding this claim is as an appeal to the virtue of simplicity.

53. 115 NY 506 (1889).

54. Leiter, supra note 1 at 1218.

55. Ibid at 1242. Leiter bases this conclusion partly on the fact that, previously, Judge Earl had unsuccessfully opposed the view that convicts retain their property rights (including the right to inherit) (id), and partly on the fact that, in other probate cases (unlike in Riggs), Judge Earl sought to uphold the testator’s intentions and was generally unconcerned with the correct theory of statutory interpretation (ibid at 1244).

56. Ibid at 1243.

57. Ibid at 1246. More generally, Leiter appeals to the work of Karl Llewellyn and Philip Bobbitt to support his contention that, in difficult cases of statutory and constitutional interpretation, judges tend to act opportunistically: ibid at 1229-30.

58. This is the Error Theory account of theoretical disagreement.

59. This is the Disingenuity account of theoretical disagreement.

60. Dworkin, supra note 3 at 42-43 [emphasis added].

61. Leiter, supra note 1 at 1247. As we have seen, Leiter does not base this conclusion solely on his contention that positivism can, and Dworkin cannot, account for the judicial opportunism in cases like Riggs. He also bases it, inter alia, on his contention that positivism can, and Dworkin cannot, account for agreement in legal judgments.

62. Dworkin, supra note 3 at 37-39.

63. Cf Shapiro’s suggestion that the reason the public “has yet to pick up on the judicial ruse” is that “the law is a professional practice and lay persons are either ignorant of its ground rules or too intimidated by legal officials to challenge them.” (Shapiro, supra note 1 at 42; referred to in Leiter, supra note 1 at 1238.) For the reason given in the text, ignorance on the part of laypeople may not be an insurmountable obstacle to exposing judicial opportunism. Moreover, while we can easily imagine legal systems in which laypeople are too intimidated by legal officials to challenge them, we might query whether this is a plausible portrayal of those legal systems with which we are most familiar.

64. Again, we can imagine legal systems where officials get away with this through a campaign of terror and intimidation. But, again, it is far from clear that this is true of the legal systems with which we are most familiar.

65. Leiter, supra note 1 at 1239.

66. Dworkin, supra note 3 at 239.

67. See ibid at 184 (acknowledging that our legal structure often violates integrity in various ways, but insisting that we regard such violations as wrongful).

68. Ibid at 265.

69. Ibid at 184, 217.

70. Leiter, supra note 1 at 1248 [emphasis in original].

71. For Dworkin’s response to this objection, see Dworkin, supra note 3 at 101-04. The objector, of course, denies that this response is satisfactory: see, e.g., Hart, supra note 11 at 270-71.

72. See, e.g., Shapiro, supra note 13 at 46-49.

73. For an interesting attempt to dissolve the normativity problem, see Enoch, David, “Reason-Giving and the Law” in Green, Leslie & Leiter, Brian, eds, Oxford Studies in Philosophy of Law: Volume 1 (Oxford: Oxford University Press, 2011).Google Scholar

74. When, on another occasion, Leiter recommends the use of these criteria to assess a certain argument for moral realism, he describes them as “intuitively plausible”: Leiter, Brian, Naturalizing Jurisprudence: Essays on American Legal Realism and Naturalism in Legal Philosophy (Oxford: Oxford University Press, 2007) at 205.Google Scholar In that context, though, Leiter can (and does) claim, with some plausibility, that his opponents are also committed to the use of those criteria. We shall see that the same is not true in the present context.

75. Raz, Joseph, Between Authority and Interpretation (Oxford: Oxford University Press, 2009) at 31.Google Scholar

76. Dworkin, supra note 3 at 53.

77. It would be a mistake to equate the dimension of fit with the virtue of consilience. Dworkin emphasizes that our convictions about fit (where the threshold of fit should be set, how the dimensions of fit and justification should be balanced against each other, etc) reflect our convictions of political morality: ibid at 257. By contrast, judgments of consilience are not meant to be responsive to our convictions of political morality in this way.

78. That Dworkin would regard this as inappropriate is even clearer from his discussion in Justice for Hedgehogs. There, he argues for what he calls the “independence of value” thesis—namely, that morality and ethics represent a separate domain of inquiry from science and metaphysics, and that any moral or ethical argument must ultimately stand or fall on moral or ethical (not scientific or metaphysical) grounds: Dworkin, supra note 14 at 9-11. Moreover, the methodology that is appropriate in each domain differs radically. Dworkin seeks to defend accounts of moral and ethical values (which, on the view presented in Justice for Hedgehogs, include law) by showing that those accounts fit with each other in such a way as to reveal our moral and ethical values to be integrated and mutually supporting. Further, those accounts must be ones we can authentically embrace, in the sense that they must reflect genuine moral or ethical convictions on our part: ibid at 108. On this view, theory choice in the moral/ethical domain requires moral/ethical judgment, rather than appeal to the sorts of considerations that would be appropriate when choosing between scientific theories. Admittedly, though, it is controversial whether the discussion of law in Justice for Hedgehogs is compatible with the view Dworkin defends in Law’s Empire: see supra note 14. It is also debatable whether the discussion of methodology in the two books is consistent.

79. Concerning the normative character of the dimension of fit, see supra note 77. Further, on Dworkin’s view, we do not seek discrete explanations of different legal phenomena, but rather a unified interpretation of the entirety of our legal practices. Note, though, that there is also some pressure within Leiter’s methodology to adopt unified accounts of multiple phenomena. At least where there are competing explanations that make sense of a particular phenomenon, we adjudicate between those explanations in part based on whether they make sense of other phenomena as well.

80. On my reading of Leiter, he intends his critique of law as integrity in Leiter, supra note 1 to stand independently of the claim he makes elsewhere that we should adopt a naturalistic approach to legal philosophy: see Leiter, supra note 74 at Part II. I base this reading partly on the fact that Leiter does not explicitly appeal to his naturalism when presenting his critique of law as integrity, and partly on the approach he takes elsewhere, which is first to argue that Hart is the victor in the debate with Dworkin and then to consider whether there are further (methodological) problems with Hart’s approach: ibid at ch 6. However, given the point we have reached in the text, Leiter might appeal to his naturalism to support the claim that we should employ the theoretical virtues when choosing between competing theories of law. I cannot properly evaluate this possibility here, but I should point out that, at least in some moods, Leiter would regard the implications of naturalism for legal philosophy as being much more radical than this (though he has wavered as to what those implications are). It is also worth noting that Leiter’s claim that the theoretical virtues represent the appropriate criteria for choosing between rival scientific theories is disputed by some philosophers of science. See, e.g., Helen E Longino, “Gender, Politics, and the Theoretical Virtues” (1995) 104 Synthese 383 (arguing that there are alternative bases for choosing between rival scientific theories and that, in some contexts, the choice between employing the traditional theoretical virtues or alternative virtues must be made on socio-political grounds).

81. Indeed, it may be that each methodology must be assessed in conjunction with the theory of law that fits most naturally with that methodology, rather than us first choosing between the competing methodologies and then using the chosen methodology to adjudicate between rival theories of law. See Dickson, Julie, Evaluation and Legal Theory (Oxford: Hart, 2001) 13–14.Google Scholar

82. Dworkin, supra note 3 at 190-216.