Two Models of Legal Principles

Iowa Law Review

Volumn 82, Issue 3, 1997, Pages 787-818

  Perry, Stephen R ^a  

^a UNIVERSITY OF PENNSYLVANIA   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0346441501     PISSN: 00210552     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (65)

1. See Ronald Dworkin, Is Law a System of Rules?, 35 U. Chi. L. Rev. 14, 22-29 (1967), reprinted as Ronald Dworkin, The Model of Rules I, in Taking Rights Seriously 14, 22-28 (rev. ed. 1977) [hereinafter Dworkin, Taking Rights Seriously].
2. See Ronald Dworkin, Is Law a System of Rules?, 35 U. Chi. L. Rev. 14, 22-29 (1967), reprinted as Ronald Dworkin, The Model of Rules I, in Taking Rights Seriously 14, 22-28 (rev. ed. 1977) [hereinafter Dworkin, Taking Rights Seriously].
3. I borrow this term from H.L.A. Hart, The Concept of Law 261 (2d ed. 1994).
4. See Joseph Raz, Practical Reason and Norms 36-37 (2d ed. 1990).
5. In addition to the rule/principle distinction, Dworkin also drew a distinction in his early work between principles and policies. The latter distinction cuts across the former, since it is concerned with different kinds of content that principles, understood in the sense of the rule/principle distinction, might have: principles were said by Dworkin to be concerned with individual rights, whereas policies were concerned with social goals. See Dworkin, Taking Rights Seriously, supra note 1, at 22, 82-84. In this essay I shall be discussing only the rule/principle distinction.
6. Principles are, as noted, first-order reasons that are moral in character; as such, they will necessarily possess some minimum degree of generality. Not all first-order reasons need be general; a reason of self-interest might be relevant only to a specific person on a specific occasion, for example. The generality of principles means that, given their value-oriented content, they will ordinarily be relevant to more than a single type of action; thus a single principle could figure in the justification of more than one moral rule. On the other hand a moral rule could, given its action-oriented content, conceivably be justified by more than one set of principles.
7. See Raz, supra note 3, at 39, 58-59, 73; see also Joseph Raz, The Authority of Law 16-19 (1979) [hereinafter Raz, Authority]; Joseph Raz, The Morality of Freedom 41-42, 57-59 (1986) [hereinafter Raz, Freedom].
8. See Raz, supra note 3, at 39, 58-59, 73; see also Joseph Raz, The Authority of Law 16-19 (1979) [hereinafter Raz, Authority]; Joseph Raz, The Morality of Freedom 41-42, 57-59 (1986) [hereinafter Raz, Freedom].
9. See Raz, supra note 3, at 39, 58-59, 73; see also Joseph Raz, The Authority of Law 16-19 (1979) [hereinafter Raz, Authority]; Joseph Raz, The Morality of Freedom 41-42, 57-59 (1986) [hereinafter Raz, Freedom].
10. See Joseph Raz, Ethics in the Public Domain 194-204 (1994); Raz, Freedom, supra note 6, at 38-69.
11. See Raz, Freedom, supra note 6, at 60.
12. There are certain cases where something like an autonomous rule would be necessary even in a world in which morality was epistemically transparent. These involve situations where the no-difference thesis, i.e., the thesis that an exercise of authority should make no difference to what its subjects ought to do, does not hold. Raz gives three examples of such situations: first, where the precise action demanded by morality is underdetermined; second, where a convention is required to solve a coordination problem; and third, where an authoritative directive is required to solve a prisoners' dilemma. See Raz, Freedom, supra note 6, at 48-51; see also Andrei Marmor, Interpretation and Legal Theory 116-17, 176-81 (1992).
13. There are certain cases where something like an autonomous rule would be necessary even in a world in which morality was epistemically transparent. These involve situations where the no-difference thesis, i.e., the thesis that an exercise of authority should make no difference to what its subjects ought to do, does not hold. Raz gives three examples of such situations: first, where the precise action demanded by morality is underdetermined; second, where a convention is required to solve a coordination problem; and third, where an authoritative directive is required to solve a prisoners' dilemma. See Raz, Freedom, supra note 6, at 48-51; see also Andrei Marmor, Interpretation and Legal Theory 116-17, 176-81 (1992).
14. For a discussion of the methodological concerns in legal theory that make a discussion of point or purpose relevant, see Stephen R. Perry, Interpretation and Methodology in Legal Theory, in Law and Interpretation 97 (Andrei Marmor ed., 1995).
15. See Raz, Authority, supra note 6, at 50-51.
16. See Raz, supra note 7, at 199-204.
17. See id; see also Raz, Authority, supra note 6, at 45-52.
18. Cf. Dworkin, Taking Rights Seriously, supra note 1, at 338, 346-48.
19. This might involve, among other things, a change in view about what the "type" of the case is; sometimes this is what is at issue when a prior case is distinguished, for example.
20. See Stephen R. Perry, Judicial Obligation, Precedent and the Common Law, 7 Oxford J. Legal Stud. 215, 234-55 (1987) [hereinafter Perry, Judicial Obligation]; Stephen R. Perry, Second-Order Reasons, Uncertainty and Legal Theory, 62 S. Cal. L. Rev. 913, 963-93 (1989) [hereinafter Perry, Second-Order Reasons].
21. See Stephen R. Perry, Judicial Obligation, Precedent and the Common Law, 7 Oxford J. Legal Stud. 215, 234-55 (1987) [hereinafter Perry, Judicial Obligation]; Stephen R. Perry, Second-Order Reasons, Uncertainty and Legal Theory, 62 S. Cal. L. Rev. 913, 963-93 (1989) [hereinafter Perry, Second-Order Reasons].
22. There is an interesting debate as to whether consistency, understood not in loose pragmatic terms but rather as calling for equality of treatment among persons, is in fact a value to which courts should give weight. See, e.g., Larry Alexander & Ken Kress, Against Legal Principles, in Law and Interpretation, supra note 10, at 279, 294-95, 305, reprinted in 82 Iowa L. Rev. 739, 754-55, 765 (1997). For reasons which I shall discuss in section VI, I believe that consistency is such a value. But even if I am wrong in this, the argument in the text remains unaffected. It is enough for present purposes that there be some set of conservative values (referred to later in the text as the rule-of-law values), that incline towards doing what has been done before. The most important and least controversial such value is the need for some degree of predictability or stability in the legal decision-making process.
23. See Dworkin, Taking Rights Seriously, supra note 1, at 43-44, 64-65.
24. See id.
25. This is not an uncontroversial assumption; see supra note 17.
26. There is, of course, a problem of scope that arises here, which can be described in the following way: What exactly is comprehended by the phrase, "the total set of relevant and binding autonomous legal rules"? Is it the entire common law? Is it the entire common law together with all other autonomous legal rules, including in particular those created by legislation? Or is it perhaps some subset of the common law, such as, in an appropriate case, the law of torts? Or is it some still smaller set of rules? These are not questions that need to be addressed for present purposes. See further Ronald Dworkin, Law's Empire 250-54 (1986).
27. Cf. Perry, Judicial Obligation, supra note 16, at 239-55.
28. Cf. Perry, Second-Order Reasons, supra note 16, at 932-36, 966-68. One way to understand the notion of a second-order principle is as a generalization of Raz's notion of a second-order reason. For Raz, a second-order reason is a reason to act on, or not to act on, a first-order reason. On the generalized understanding, a second-order reason is a reason to treat a first-order reason as having a degree of weight which differs from the weight one would otherwise attribute to it. The idea of "weight" can, however, be understood in either a substantive or an epistemic sense. The substantive weight of a principle is the weight that is attributed to it in practical reasoning, in the process of assessing the overall balance of reasons. Epistemic weight is the degree of confidence that one must have in one's belief that the principle has heretofor been incorrectly formulated before a presumption in favor of the principle's (moral) correctness can be rebutted. ("Moral correctness" concerns, in this context, both the content of a principle and its substantive weight.) On a "substantive" interpretation of the notion of a second-order principle, it is substantive weight that is assumed to be variable: thus an exclusionary reason is just the special case of a reason to treat a first-order reason as having zero weight. But an "epistemic" interpretation is also possible, since, as we shall see shortly, the required degree of confidence in one's own present moral beliefs can vary, so that it makes sense to speak of attributing a greater or lesser epistemic weight to the relevant principle. At one time it seemed to me that both the substantive and epistemic interpretations of the generalized notion of a second-order reason were possible: see id. However, I now have some doubts about the coherence of the substantive interpretation, and have therefore confined discussion in the text to the epistemic understanding.
29. See, for example, O'Brien v. Robinson, 1973 App. Cas. 912 (H.L.): While it would be open to your Lordships to do so, this is not, I think, a suitable case in which to exercise the recently asserted power of this House to refuse to follow one of its own previous decisions. An examination of the reasoning in the judgments in the cases on this subject during the last hundred years suggests that the law might easily have developed on different lines from those which it in fact followed. But, for my part, I am not persuaded that this development was clearly wrong or leads to results which are clearly unjust . . . . Id. at 930 (Lord Diplock); see also Fitzleet Estates Ltd. v. Cherry, [1977] 3 All E.R. 996, 1000 (H.L.) ("If the decision in the Chancery Lane case was wrong, it certainly was not so clearly wrong and productive of injustice as to make it right for the House to depart from it") (Viscount Dilhorne).
30. Cf. Perry, Judicial Obligation, supra note 16, at 24142. In speaking of "the nature of the case," I have in mind two things. First, stability and predictability might be more important in some areas of the law than in others. Thus we should expect a higher epistemic threshold in contract cases, where legal rules are intended in part to facilitate the structuring of consensual relationships, than in tort cases (or at least in tort cases involving involuntary harm between strangers). Second, we should expect the epistemic threshold to vary inversely with the degree of perceived injustice of the present legal rule (or, alternatively, with the extent of the bad consequences for which it is thought to be responsible). Thus the threshold should, all things being equal, be lower for a rule that is thought to involve great injustice than for one that is thought to involve only some lesser degree of wrong.
31. An earlier formulation that I gave of this idea in Perry, Judicial Obligation, supra note 16, was criticized by Heidi Hurd on the grounds that it was just a variant on Raz's conception of justified authority, and hence vulnerable to a critique similar to one she had offered of Raz. See Heidi M. Hurd, Challenging Authority, 100 Yale L.J. 1611, 1639-40 (1991). But it was never my intention to offer an account of authority, either justified or de facto. The basic thesis has always been, rather, that certain values which ought to be taken into account in the balance of reasons, namely, the rule-of-law values, cannot plausibly be regarded as ordinary first-order reasons. These values are instead best understood as second-order reasons; they are, in effect, functions that take as arguments the first-order principles that have figured in judicial reasoning in the past.
32. Cf. Alexander & Kress, supra note 17, at 296, 309, reprinted in 82 Iowa L. Rev. 739, 756, 769 (1997).
33. It is important not to confuse expectations with the rule-of-law values of predictability and stability; the former are psychological states of persons, whereas the latter are general desiderata in a legal system.
34. On the notion of substantive weight, see supra note 23.
35. For purposes of simplicity, I am assuming that the stated rule is the only autonomous rule that is both binding on Julia and relevant to the type of case in question, and hence the only member of the set of rules for which Julia must, within the rationalization model, find the best justification. On the notion of relevant and binding autonomous rules, see supra note 21.
36. It also bears mention that a sufficiently drastic reformulation of the underlying principles could lead to a redefinition of the relevant type of case. As was explained in Section II, a case's type is determined by the general summary rule - and hence, ultimately, by the principles underlying the rule - that we take the case to be governed by. The general rule has this effect because it treats only certain (types of) facts as material to the resolution of the case. If the characterization of relevant facts changes, so does the case's type. This is, however, a complication that for present purposes we may ignore.
37. See, e.g., Alexander & Kress, supra note 17, at 283-85, reprinted in 82 Iowa L. Rev. 739, 743-45 (1997).
38. 88 Harv. L. Rev. 1057 (1975), reprinted in Dworkin, Taking Rights Seriously, supra note 1, at 81.
39. 88 Harv. L. Rev. 1057 (1975), reprinted in Dworkin, Taking Rights Seriously, supra note 1, at 81.
40. See particularly Dworkin, supra note 21, at 225-58.
41. Dworkin, Taking Rights Seriously, supra note 1, at 37.
42. Id.
43. Id. at 38.
44. Id. at 44.
45. Dworkin's version of the primacy model is not, however, the same as the one put forward in section III above: on Dworkin's account, the conservatively-inclined principles underlying the doctrines of precedent and legislative supremacy are not characterized as second-order in nature, nor are they said to take effect through the epistemic entrenchment of first-order, substantive principles.
46. Dworkin, Taking Rights Seriously, supra note 1, at 40.
47. See, e.g., Joseph Raz, Legal Principles and the Limits of Law, 81 Yale L.J. 823 (1972).
48. See Dworkin, Taking Rights Seriously, supra note 1, at 343.
49. I take this to be the gist of Larry Alexander's complaint that Dworkin's theory of law cannot overcome the problem of "bad beginnings," i.e., the problem of moral errors in past judicial decisions. Since the theory must supposedly take all past decisions as constituting the settled law, whether those decisions contain moral errors or not, Dworkin's account of legal reasoning "requires that we justify the unjustified." Larry Alexander, Bad Beginnings, 145 U. Pa. L. Rev. 57, 84 (1996).
50. See id.
51. See Dworkin, supra note 21, at 164-67.
52. 111 N.E. 1050 (N.Y. 1916).
53. 1932 App. Cas. 562 (H.L.).
54. See Dworkin, Taking Rights Seriously, supra note 1, at 111.
55. In Dworkin's terminology, positivism requires validity to be a matter of pedigree. See Dworkin, Taking Rights Seriously, supra note 1, at 17, 346-48.
56. See Dworkin, supra note 21, at 164-67, 176-86.
57. Alexander & Kress, supra note 17, reprinted in 82 Iowa L. Rev. 739 (1997).
58. Alexander and Kress clearly conceive of legal rules in terms of, to use my terminology, the autonomous conception of rules.
59. See Alexander & Kress, supra note 17, at 293-94, reprinted in 82 Iowa L. Rev. 739, 753-54 (1997).
60. Id. at 295, reprinted in 82 Iowa L. Rev. 739, 755 (1997).
61. Id. at 309, reprinted in 82 Iowa L. Rev. 739, 769 (1997).
62. See id. at 305, reprinted in 82 Iowa L. Rev. 739, 765 (1997).
63. See id. at 301-04, reprinted in 82 Iowa L. Rev. 739, 761-64 (1997).
64. See Alexander & Kress, supra note 17, at 296, 309, reprinted in 82 Iowa L. Rev. 739, 756, 769 (1997).
65. Id. at 309, reprinted in 82 Iowa L. Rev. 739, 769 (1997).