1. Constructive interpretation in law is, roughly speaking, Rawlsian reflective equilibrium. John Rawls, A Theory of Justice(Cambtidge. MA: Belknap Press of Harvard University, 1971) at 46–53. Instead of Rawls’s considered convictions to be put in equilibrium with a formal theory of justice, one starts with preinterpretive institutional practices (constitutions, statutes, and precedents) and seeks a justification of that practice in equilibrium with the preinterpretive data. The method achieves coherence because—according to its proponents—reflective equilibrium requires that the justification of the practice entail and justify the bulk of the institutional practice. Likely, there will also be inferential connections among the principles that constitute part of the justification of the practice. Such internal architecture among the principles and data are a form of coherence. See Ken Kress, “Coherence and Formalism” (1993) 16 Harv. J. of L. & Publ. Pol. 639 at 657–59, 665–66; Ken Kress, “Coherence” in A Companion to The Philosophy of Law and Legal Theory. D. Patterson, ed., (Oxford: Blackwell, 1996) 533 at 545–46.

2. Joseph Raz, “The Relevance of Coherence” (1992) 72 Boston U.L. Rev. 273 at 286 and reprinted in Raz, Ethics in the Public Domain: Essays in the Morality of Law and Politics (Oxford: Clarendon Press, 1994) (subsequent citations are to the law review pagination); Ronald Dworkin, Law’s Empire (Cambridge, MA: Harvard University Press, 1986) at 65–68, 139, 230, 255 [hereinafter “LE”]; Ronald Dworkin, Taking Rights Seriously. rev. ed. (Cambridge, MA: Harvard University Press, 1978) at 121, 340, 360 [hereinafter “TRS”].

3. Joseph Raz, “Dworkin: A New Link in the Chain” (1986) 74 Cal. L. Rev. 1103 at 1111–12 (reviewing Ronald Dworkin, A Matter of Principle (Cambridge, MA: Harvard University Press, 1985)); Gerald J. Postema, “Bentham and Dworkin on Positivism and Adjudication” (1980) 5 Soc. Theory & Practice 347 at 364–66; David Lyons, “Justification and Judicial Responsibility” (1984) 72 Cal. L. Rev. 178 at 196–99; David Lyons, Ethics and The Rule of Law (Cambridge and New York: Cambridge University Press, 1984) at 92–104.

4. Interestingly, Rawls does not think that coherentist methods in ethics and justice need to meet a high threshold of fit with considered convictions. Rawls, supra note 1 at 49; Raz, supra note 2 at 285–86.

5. See infra Part H.

6. TRS, supra note 2 at 121.

7. I owe this distinction—but not the use I make of it—to Stephen Perry. For Perry’s objection to that (ab)use, see Stephen R. Perry, “Two Models of Legal Principles” (1997) 82 Iowa L. Rev. 787. For my co-authored reply, see Larry Alexander & Ken Kress, “Replies to Our Critics” (1997) 82 Iowa L. Rev. 923 at 932–36.

8. See infra text accompanying notes 63–65, 70–72.

9. TRS, supra note 2, at 340, 360 (emphasis added).

10. LE, supra note 2 at 65–68, 139, 230, 255. See infra text accompanying notes 34–53, 65–69.

11. Raz, supra note 2 at 286.

12. Raz, supra note 2 at 283 and passim.

13. LE, supra note 2 at 65–68, 139, 230, 255; TRS, supra note 2 at 121, 340, 360.

14. Raz, supra note 2 at 285–86. Of course, rejecting deviation from Razian sources of law is consistent with Raz’s overall views. Joseph Raz, Concept of a Legal System: An Introduction to the Theory of Legal Systems. 2d ed. (Oxford: Clarendon Press; New York: Oxford University Press, 1980); Joseph Raz, The Authority of Law: Essays on Law and Morality (Oxford: Clarendon Press; New York: Oxford University Press, 1979); Joseph Raz, The Morality of Freedom (Oxford: Clarendon Press; New York: Oxford University Press, 1986).

15. Joseph Raz, “Dworkin: A New Link in the Chain”, supra note 3 at 1111–12 (reviewing Ronald Dworkin, A Matter of Principle); Gerald J. Postema, “Bentham and Dworkin on Positivism and Adjudication”, supra note 3 at 364–66; David Lyons, “Justification and Judicial Responsibility”, supra note 3 at 196–99; David Lyons, Ethics and the Rule of Law, supra note 3 at 92–104. Dworkin argues that despite the high threshold of fit, law as integrity is—at least in some circumstances—less conservative than pragmatism. But his claim is unpersuasive because he argues against utilitarian rather than Kantian pragmatism, and it is also in tension with his discussion of remedies. LE, supra note 2 at 220–21, 389–92.

16. Otto Neurath, “Protokollsatze” (1932/33) 3 Erkenntis 204, reprinted as “Protocol Sentences” in A.J. Ayer, ed., Logical Positivism (Glencoe, IL: Free Press, 1959) 199 at 201. (“We are like sailors who must rebuild their ship on the open sea, never able to dismantle it in dry-dock and to reconstruct it there out of the best materials.”).

17. Laurence BonJour, The Structure of Empirical Knowledge (Cambridge, MA: Harvard University Press, 1985) at 107; Ralph C.S. Walker, The Coherence Theory of Truth: Realism, Anti-Realism, Idealism (London: Routledge, 1989) at 26.

18. In the case of a finite number of beliefs, this result is easily proven in combinatory theory.

19. Impartial observation of all evidence may be insufficient to guarantee that a coherence theory will reach a unique set of justified beliefs or truths at the end of the inquiry, if, as I have argued elsewhere respecting coherence accounts of legal reasoning, they are path dependent. Ken Kress, Legal Theory and Political Legitimacy: Positivism, Dworkin’s Rights Thesis, and Critical Legal Studies (Ann Arbor: University Microfilms, 1989) at 5–53, esp. at 50–53, 20–26; Ken Kress, “Legal Reasoning and Coherence Theories: Dworkin’s Rights Thesis, Retroactivity, and the Linear Order of Decisions” (1984) 72 Cal. L. Rev. 369 at 382–83, 400–02.

20. More precisely, in the infinite case there will still be the cardinally same infinite number of equally coherent sets, but they will be a subset of those sets which are coherent but are not constrained to fit the legal sources.

Among finite sets, no two sets can have the same cardinality (be put in one-to-one correspondence with each other), yet one be a subset of the other. By contrast, an infinite set is by definition a set that can be put in one-to-one correspondence with a subset of itself. For example the infinite set of natural numbers and its subset the even numbers can be put in one-to-one correspondence as follows:

0, 1, 2, 3, 4 …

0, 2, 4, 6, 8 …

21. Willard Van Orman Quine, Word and Object (Cambridge, MA: Technology Press of the Massachusetts Institute of Technology, 1960) at 26–79; Willard Van Orman Quine, “Two Dogmas of Empiricism” in William Van Orman Quine, ed., From a Logical Point of View. 2d rev. ed. (Cambridge, MA: Harvard University Press, 1961) at 20.

22. Melvin A. Eisenberg, The Nature of the Common Law (Cambridge, MA: Harvard University Press, 1988).

23. LE, supra note 2; TRS, supra note 2.

24. Steven J. Burton, An Introduction to Law and Legal Reasoning. 2d ed. (Boston: Little, Brown, 1995).

25. Ernest Weinrib may hold such a position, although I do not know of any other modern theorist who does. Moreover, Weinrib’s coherentism, as I have argued elsewhere, is more akin to modern monistic foundationalism grounded in corrective justice than to modern pluralist coherentism. Ken Kress, “Coherence and Formalism” (1993) 16 Harv. J. of L. & Publ. Poi. 639 at 665.

26. Some, for example, Dworkin, might urge that moral considerations are internal to law.

27. Joseph Raz, “Dworkin: A New Link in the Chain”, supra note 3 at 1111–12 (reviewing Ronald Dworkin, A Matter of Principle); Gerald J. Postema, “Bentham and Dworkin on Positivism and Adjudication”, supra note 3 at 364–66; David Lyons, “Justification and Judicial Responsibility”, supra note 3 at 196–99; David Lyons, Ethics and the Rule of Law, supra note 3 at 92–104.

28. Raz, supra note 2 at 285–86.

29. Dworkin employs the first two idioms, the third theory is the perspective I prefer.

30. The same is true of most noncoherence theories: conventional or critical morality and institutional history are either the dominant or exclusive elements.

31. LE, supra note 2 at 65–68.

32. Compare TRS, supra note 2 at 121, 340–41, 360 (absolute minimum) with Ronald Dworkin, “Natural Law Revisited” (1982) 34 U. Ha. L. Rev. 165 at 171–72 (rule of thumb) and LE, supra note 2 at vii, 225, 228–32, 238–50, 257–58 (rule of thumb), 410–11 (best combination of fit and moral appeal).

33. Ken Kress, “The Interpretive Turn” (1987) 97 Ethics 834 at 845.

34. TRS, supra note 2 at 340–41, 360.

35. LE, supra note 2 at 171–72.

36. But see the qualification in the text infra accompanying notes 65–71.

37. Gerald J. Postema, “Protestant Interpretation and Social Practices” (1987) 6 L. & Phil. 283 at 293–97.

38. LE, supra note 2 at 231, 242, 255.

39. Jules L. Coleman, Risks and Wrongs (Cambridge and New York: Cambridge University Press, 1992) at 6–13, 211, 362; Jules L. Coleman, “Second Thoughts and Other First Impressions” in Brian Bix, ed., Analyzing Law: New Essays in Legal Theory (New York: Oxford University Press, 1998) (forthcoming); Jules Coleman, “Incorporationism, Conventionality and the Practical Difference Thesis” (1998) 4 Legal Theory 381.1 specifically reserve, as a question for further research, whether Coleman is correct that the purpose of jurisprudence is to most fruitfully illuminate the practice of law and that this means that stage one fit is epistemic, or whether some nonepistemic or partially nonepistemic constraint better captures stage one fit.

40. Michael S. Moore, “The Interpretive Turn in Modern Theory: A Turn for the Worse?” (1989) 41 Stan. L. Rev. 871 (arguing that interpretation is not a sui generis category distinct from description or normativity, but rather that interpretation collapses into one or the other; in particular, Dworkinian interpretation collapses into normativity).

41. There are, of course, other logical possibilities. To give just two: the practice of constructive interpretation in legal practice could be justified by constructive interpretation in jurisprudence, which justificatory practice could itself be justified by constructive interpretation, which practice could itself be justified either (1) normatively or, alternatively, (2) epistemically.

42. LE, supra note 2 at 231, 242–45, 255–56.

43. Ibid. at 231, 246–47, 257.

44. Ibid. at 165–67, 172–75, 184, 257.

45. Indeed, I shall note later that fit actually comes into Dworkin’s theory at four different places; therefore, there are probably four different conceptions of fit in the best interpretation of his view. Because jurisprudence is interpretive, there are two stages of fit at the level of jurisprudence. Because law as integrity is interpretive, there are an additional two stages of fit at the level of legal practice. See infra text accompanying notes 69–70.

46. LE, supra note 2 at 131, 242–45, 255–56.

47. Ibid. at 67. Dworkin’s use of the term “reform” is misleading. Different perspectives on the purpose of the practice, and their resulting post-interpretive views of the requirements of law, are best understood as conflicting theories of what the law really is. Postema, supra note 37 at 293.

48. LE, supra note 2 at 66.

49. Ibid. at 67. See also ibid. at 139, 230, 255. Although Rawls does not believe that constructive interpretative methods in ethics and justice need meet a high threshold of fit with considered convictions, he agrees with Dworkin that these methods in law must fit institutional facts to a high degree. His reasons, however, focus on the requirement that judges be exemplars of public reason.

Echoing Dworkin, Rawls claims that “the best [constitutional] interpretation is the one that best fits the ‘constitution and constitutional precedents,’ and justifies that interpretation ‘in terms of the public conception of justice or a reasonable variant thereof.’” John Rawls, Political Liberalism (New York: Columbia University Press, 1993) at 236. See generally ibid. at 231–40. Yet Rawls’s reasons for requiring a high threshold of fit in law do not depend upon Dworkin’s argument that interpretation must meet a high level of fit to be interpretation not invention of a new subject matter:

This account of what the justices are to do seems to be the same as Ronald Dworkin’s view as stated say in “Hard Cases” in Taking Rights Seriously … or in Law’s Empire …. chap. 7, except for possibly one proviso. I have said that the justices in interpreting the constitution are to appeal to the political values covered by the public political conception of justice, or at least by some recognizable variant thereof. The values the justices can invoke are restricted to what is reasonably believed to be covered by that conception or its variants, and not by a conception of morality as such, not even of political morality. The latter I think too broad. Thus, though an appeal to a social minimum specified by basic needs is appropriate …, an appeal cannot be made to the difference principle unless it appears as a guideline in a statute…. I believe Dworkin thinks that his requirement of fit alone leads to roughly the same conclusion, as he takes the requirement of fit to distinguish interpretation from invention and that a reasonable interpretation suffices to show what is already implicit in the law as articulated within the political conceptions, or one of its recognizable variants. He may be correct about this but I am unsure. I incline to require, in addition to fit, that in order for the court’s decisions to be properly judicial decisions of law, that the interpretation fall within the public political conception of justice or a recognizable variant thereof. I doubt that this view differs in substance from Dworkin’s. (Rawls, ibid. at 236–37, n. 23.)

Rather, Rawls’s reasons for a high threshold of fit appeal to the requirement of employing public reason as a condition of legitimacy. Judicial decisions must be reasonably derived from a public political conception of justice. Moreover, legitimacy requires that judicial authority be embedded in a constitutional framework that all reasonable and rational citizens can be expected to accept. Ibid. at 217. In consequence of its focus on public reasons, Rawls’s conception of adjudication appears potentially more conventionalist than Dworkin’s.

50. Michael Green, “Negligence = Economic Efficiency: Doubts >” (1997) 75 Tex. L. Rev. 1605 at 1609, 1616 (reciting Bohlen’s anticipation of the Learned Hand test for negligence).

51. For example, Star Trek: The Next Generation: The Neutral Zone (Rebroadcast KCRG, July 27, 1998) (“This is the 24th century: material needs no longer exist.”)

52. LE, supra note 2 at 165–67, 172–75, 184, 257.

53. Larry Alexander & Ken Kress, “Against Legal Principles” in Andrei Marmor, ed., Law and Interpretation: Essays in Legal Philosophy (Oxford: Clarendon Press; New York: Oxford University Press, 1995) 279 at 308–27, reprinted in (1997) 82 Iowa L. Rev. 739 at 768–86. For discussions of the argument, see “Colloquy” (1997) 82 Iowa L. Rev. at 738.

54. LE, supra note 2 at 134, 88, 116–17, 166, 184, 435–36 n.7; Alexander & Kress, supra note 53 at 311, reprinted in (1997) 82 Iowa L. Rev. at 771.

55. Raz, supra note 2 at 293; LE, supra note 2 at 189.

56. I have argued elsewhere that Dworkin exaggerates the extent to which our legal practices exemplify integrity. Ken Kress, “Why No Judge Should be a Dworkinian Coherentist” (1999) 77 Tex. L. Rev., forthcoming.

57. LE, supra note 2 at 188.

58. Ibid. at 192–95.

59. Ibid. at 199–201.

60. TRS, supra note 2 at 14–45.

61. See, e.g., Jeremy Waldron, “Special Ties and Natural Duties” (1993) 22 Phil. & Publ. Affairs 3. But see Mark C. Murphy, “Acceptance of Authority and the Duty to Comply with Just Institutions: A Comment on Waldron” (1994) 23 Phil. & Publ. Affairs 270.

62. LE, supra note 2 at 93; Alexander & Kress, supra note 53 at 312–14, reprinted in (1997) 82 Iowa L. Rev. 739 at 771–73.

63. Adarand Constructors, Inc. v. Pena. 515 U.S. 200 (U.S. 1995), 115 S.Ct. 2097, 132 L.Ed. 2d 158.

64. For a contrary view, see Michael S. Moore, “Legal Principles Revisited” (1997) 82 Iowa L. Rev. 867; Stephen R. Perry, “Two Models of Legal Principles” (1997) 82 Iowa L. Rev. 787. For our reply, see Larry Alexander & Ken Kress, “Replies to our Critics” (1997) 82 Iowa L. Rev. 923 at 928–32.

65. See supra text accompanying notes 32–46.

66. Compare TRS, supra note 2 at 340–41, 360, (absolute floor) with Ronald Dworkin, “Natural Law Revisited” (1982) 34 U. Fla. L. Rev. 165 at 171–72 (rule of thumb). See also supra notes 31–46 and accompanying text.

67. TRS, supra note 2 at 360. Ironically, Dworkin’s more elaborate two-stage interpretation of fit in Law’s Empire is at odds with his reply to Mackie, especially if Dworkin believes that the level of stage one fit equals the level of stage two fit, because at stage two one looks for the best overall combination of stage two fit and moral appeal. See supra note 43 and accompanying text; see generally text accompanying notes 34–46.

68. LE, supra note 2 at 231 (discussing a chain novelist as an analogy to a judge).

69. LE, supra note 2 at 90, 226–27, 380.

70. Jeremy Waldron, “The Circumstances of Integrity” (1997) 3 Legal Theory 1.

71. Alexander & Kress, supra note 53 at 297–98, reprinted in (1997) 82 Iowa L. Rev. at 757–58; Kress, supra note 56 (developing the argument further).

72. Ibid.; Alexander & Kress, supra note 53 at 297–98, reprinted in (1997) 82 Iowa L. Rev. at 757–58.

73. Joseph Raz, “Authority, Law and Morality” (1985) 68 The Monist 295.

74. As a bright line test in First Amendment doctrine chills less exercise of free speech rights, so predictability of law generally increases liberty because citizens are better able to approach the boundaries of the law without crossing them.

75. Moreover, some would urge that the latter is necessary to adequately explain the authority of law. Joseph Raz, The Morality of Freedom supra note 14 at 23–105; Raz, supra note 73.

76. See Rolf E. Sartorius, Individual Conduct and Social Norms: A Utilitarian Account of Social Union and the Rule of Law (Encino, CA: Dickenson Publishing Co., 1975) at 192; Ken Kress, “Legal indeterminacy” (1989) 77 Cal. L. Rev. 283 at 304–05 and n.86 (criticizing and reconstructing Sartorius’s position).

77. This theory resembles, but may not be identical to, that in Eisenberg, supra note 22. While Eisenberg employs conventional morality, he does not discuss whether his foundational principles and generative principles are critical or conventional.

78. See, e.g., TRS, supra note 2 at 37–38.

79. See sources cited supra note 76.

80. For a more detailed argument considering, in addition, factors other than the proper threshold of fit, including the existence of moral disagreement among judges, see Ken Kress, supra note 56.