1 “The Model of Rules,” 35 Univ. of Chicago L.R. 14 (1967), reprinted as Chap. 1 of Hughes, G. (ed.), Law, Reason and Justice (New York: N.Y.U. Press, 1969Google Scholar), henceforth cited as “Model” with page references to the original (and to Hughes in square brackets); “Philosophy and the Critique of Law,” in Grimm, R. H. & Mackay, A. F. (eds.), Society: Revolution and Reform (Cleveland: Case-Western Reserve Univ. Press, 1971) 59–81Google Scholar, henceforth cited as “Critique”; “Social Rules and Legal Theory,” 81 Yale L.J. 855 (1972), henceforth cited as “Social Rules.”

2 The Concept of Law (Oxford: Oxford Univ. Press, 1961Google Scholar), particularly Chaps. V and VI.

3 “Model,” pp. 44–45 [39–41].

4 Op. cit., n. 1.

5 See references cited in “Social Rules,” p. 855, n. 2.

6 Sartorius, Rolf, “Social Policy and Judicial Legislation,” 8 Am.Phil.Q. 151 (1971).

7 Coval, S. C. & Smith, J. C., “Some Structural Properties of Legal Decisions” (1973) 32 C.L.J. 81.Google Scholar

8 See ns. 6 and 7.

9 Hart, op. cit., n. 2.

10 A complete statement of a rule, as Dworkin points out (“Model,” p. 25 [16]) incorporates all exceptions to the rule so that some of A, B, C,… are descriptions of the absence of certain factual circumstances.

11 Hart, op. cit., n. 2, pp. 92–94, deals only with the operation of secondary rules upon primary rules. But a moment's reflection indicates that secondary rules operate upon secondary rules as well. Thus, for example, in considering the jurisdiction of a particular court to hear a particular case we look for the appropriate rule of adjudication (e.g., a statute constituting the court and defining its jurisdiction) which we must then test against the appropriate rule of ascertainment to ensure that it is, indeed, a rule of the system, and so the chain continues. A similar convolution involving secondary rules of ascertainment would be encountered in assessing the legal status of a rule issued by an officer under the authority of an administrative agency empowered so to delegate rule-making functions by the terms of regulations which an executive agency is authorised to make under a statute enacted by the Queen in Parliament.

12 Hart, op. cit., n. 2, p. 113.

13 Ibid., pp. 55–56, 113–114.

14 This is not to suggest that our curious Martian will locate a single sovereign á la Austin. In a parliamentary democracy he will find himself climbing different hierarchies or branches of hierarchies leading sometimes to Parliament, sometimes to the highest appellate court of the land. When a federal system is superimposed, the parliamentary branch is split into branches for each of the local, as well as the national, Parliaments, converging via the constitutional document which defines the division of powers, in the body authorised to amend the constitution. Let me be quite clear that I am not suggesting that the authority of the institution at the pinnacle is entailed by its identity; rather, its authority is accepted by the relevant people, a purely factual statement. The vital point of Hart's system is that the statement, “you are breaking the law by doing X” is a descriptive, not a normative statement. It is analogous to the statement, “if you allow the batter more than three strikes, you are not playing baseball.” This statement entails not the normative conclusion, “you must allow the batter only three strikes,” but the hypothetical, “if you wish to play baseball, you must allow the batter only three strikes.”

15 Hart, op. cit., n. 2, pp. 119–120, 123–125.

16 Ibid., p. 124; emphasis added. I believe that it is fair to suggest that this passage describes what Dworkin terms discretion in the first weak sense (see “Model,” p. 32 [25]).

17 Ibid., p. 131; emphasis added. This passage appears to describe discretion in Dworkin's strong sense (see “Model,” p. 33 [26]), the existence of which he denies.

18 Ibid., p. 132. The obscurantism complained of in the last sentence is no doubt one of the great unifying features of the common law as practised around the planet.

19 There appears to be a certain element of circularity in this statement as the identification of the officials of the system would seem logically to depend upon identification of the rules of the system. Ignoring the problem of how the system came to exist “in the beginning.” however, we may say that a legal system exists at a given point of time if there exists a body of individuals (whom we shall call officials) all of whom concur in the acceptance of a fundamental rule (rule of recognition) which indicates to them the rules by which they must govern themselves as to tenure and conduct of office, as well as the rules which they are to enforce vis-a-vis the general populace. In addition, the populace must generally obey the latter rules. So formulated, the identification of the rules of the system and the officials of the system are logically simultaneous not sequential.

20 “Model,” pp. 30–31 [23].

21 Ibid., pp. 22–29 [13–21].

22 (1889) 115 N.Y. 506, 22 N.E. 188.

23 (1960) 32 N.J. 358, 161 A. 2d 69.

24 The analysis of the Sartorius and Coval/Smith papers, infra, s. V, achieves, as a by-product, a more rigorous description of the required premise.

25 “Social Rules,” p. 857.

26 Ibid., pp. 869–871; emphasis in original. I must point out that some of the “…” represent large portions of text, but I believe what I have given is an entirely faithful representation of the argument.

27 Ibid., p. 879; emphasis added. What are here termed the first and second senses of “discretion” were in “Model” referred to as the first and second weak senses, and what is here termed the third sense was in “Model” called the strong sense. See “Model,” pp. 32–33 [25–26].

28 Supra, pp. 140–141; “Social Rules,” pp. 857–868.

29 “Model,” p. 23 [14].

30 “Model,” p. 38 [32]. I will argue, infra, pp. 146–147, that it is indeed a consequence of Dworkin's formulation that no rule is, in its own right, binding.

31 Supra, p. 140.

32 To repeat, the term “open” relates not to the admitted incompleteness of Dworkin's model, but to the fact that its criterion for separating legal from extralegal material is independent of general acceptance or other objective manifestation. This feature will survive in the completed model. I use the terms “open” and “closed” only as a shorthand description of the structural relationships between Hart's model, Dworkin's model, and the Coval/Smith and Sartorius. models which are closed models of principle (infra, s. V).

33 Supra, p. 138.

34 Supra, p. 141; see also n. 27.

35 Ibid.

36 “Model,” p. 27 [19].

37 E.g., “The specific overrides the general,” or “The later takes precedence over the earlier.”

38 “Model,” p. 25 [16].

39 “ Social Rules,” pp. 871–872.

40 Ibid.

41 Infra, pp. 147–150.

42 Supra, p. 138.

43 “Model,” pp. 22–29 [13–21].

44 (1889) 115 N.Y. 506, 22 N.E. 188.

45 (1960) 32 N.J. 358, 161 A. 2d 69.

46 “Social Rules,” p. 888.

47 Supra, s. IV, second paragraph.

48 “Model,” pp. 37–38 [31–32].

49 Supra, p. 141.

50 “Critique,” pp. 65–66; emphasis in original.

51 The principles man might even think correct a decision which the rules man labels a mistake of law. In such a case the court would be pronouncing a new rule.

52 Supra, p. 146.

53 Op. cit., n. 6.

54 Supra, p. 141. On the significance of discretion in Dworkin's first (weak) sense for the law as a system of entitlements, see supra, pt. (i) of s. IV.

55 Sartorius, op. cit., n. 6, p. 155; emphasis in original.

56 Ibid., p. 156.

57 Ibid., p. 158.

58 Op. cit., n. 7.

59 Ibid., p. 83.

60 Ibid., p. 95.

61 Ibid., p. 102.

62 (1889) 115 N.Y. 506, 22 N.E. 188; see supra, p. 146.

63 Coval and Smith, op. cit., n. 7, p. 85. The authors conclude (p. 86) that the rule against a culpable killer taking from the estate of his victim applies only in the case of intentional killing. That does not appear to be the law in England or Canada. Thus in In the Estate of Hall, Hall v. Knight and Baxter [1914] P. 1, Baxter was found guilty of manslaughter in the death of Hall, but it was proved that the killing was unintentional. On the question of whether Baxter could take under Hall's will, the Court of Appeal held that it was bound by Cleaver v. Mutual Reserve Fund Life Association [1892] 1 Q.B. 147, which denied the proceeds of the deceased's life insurance policy to his convicted murderer on the ground that “no man shall profit from his own wrong.” All three judges in the Hall case expressed shock at the notion that any other result was possible. On the distinction between murder and manslaughter, Hamilton L.J. said (pp. 7–8): “The distinction seems to me either to rely unduly upon legal classification, or else to encourage what, I am sure, would be very noxious—a sentimental speculation as to the motives and degree of moral guilt of a person who has justly been convicted and sent to prison.” To the same effect is Re Giles, Giles v. Giles [1971] 3 All E.R. 1141 (Ch.D.). The refusal to distinguish between murder and manslaughter is also law in Canada: Lundy v. Lundy (1895) 24 S.C.R. 650; Re Charlton [1968] 2 O.K. 96Google Scholar (Ont.H.Ct.).

64 See, for example, Samuelson, Paul, A., Economics: An Introductory Analysis, 6th ed. (New York: McGraw-Hill, 1964)Google Scholar, Chap. 21. For a rigorous mathematical treatment, see Henderson, , James, M. and Quandt, , Richard, E., MicroeconomicTheory: A Mathematical Approach, 1st ed. (New York: McGraw-Hill, 1958)Google Scholar, Chap. 2.

65 It is not necessary that the utility function give us a cardinal ordering, i.e., enable us to say that situation one is three times as good as situation two. In mathematical terms, it is only necessary that the utility function be defined up to a monotonic transformation.

66 When a statute or precedent brings a whole new field of activity under legal regulation where there was none before, we may not be able to picture that action as a choice between two points. We may learn nothing more, with respect to the utility function, than that the principles embodied in the statute or precedent are henceforth to count as legal principles and thus to enter into the utility function. In such a situation we gain no further insight into the ordering of sets of principle attainments, i.e., into the shape of the utility function.

67 A point X is said to “dominate” a point Y when each of the co-ordinates of X exceeds the corresponding co-ordinate of Y so that X is preferable in every respect. Of course, if a case involves a choice between two points, one of which is dominated by the other, the choice is easy. Thus the mere knowledge that a certain principle is a legal principle does provide some ordering information. However, it seems most unlikely that a case worth appealing will have this character. Once we get into constellations of three, four, or more principles in a single case, it seems improbable that either outcome will be dominant as here defined.

68 In Dworkin's open model of principles, the ordering is the normative premise to be specified as the proper guide to a judge's deliberations.