This paper was prepared for presentation at a conference on the “Law and Economics of Environmental Policy” sponsored jointly by the European Association of Law and Economics and the Geneva Association, Paris, April 4–5, 1991. I am grateful to the conference participants, particularly Lewis Kornhauser, as well as to Ian Ayres, Daniel Farber, Brian Langille, Eric Rasmusen, Michael Trebilcock and an anonymous referee of the journal for helpful comments.

1. Sagoff, M. The Economy of the Earth: Philosophy, Law and the Environment (Cambridge: Cambridge University Press, 1988).Google Scholar

2. Ibid,at 3 and 57–59.

3. Ibid, at 22: “To set these priorities, we need to distinguish the pure from the polluted, the naturalfrom the artificial, the noble from the mundane, good from bad, and right from wrong. These are scientific, cultural, aesthetic, historical, and ethical—not primarily economic—distinctions.”

4. Ibid,at 7–14.

5. Ibid, at 12, quoting Richard Rorty,“Science as Solidarity”, manuscript, November, 1984 at 4.

6. Ibid, at 13, quoting Rorty, supra, note 5 at 2.

7. See, for example, the comments provided by Amartya Sen in reply to a paper written by Jon Elster which had made some points comparable to Sagoff’s. Both Sen’s paper, “Foundations of Social Choice Theory: An Epilogue”, and Elster’s paper, “The Market and the Forum: Three Varieties of Political Theory” are in Elster & Hylland, eds, Foundations of Social Choice Theory (Cambridge: Cambridge University Press, 1986). For some criticism of Sen’s reply to Elster, see Pildes, R.H. & Anderson, E. “Slinging Arrows at Democracy: Social Choice Theory, Value Pluralism, and Democratic Politics” (1990)0 Columbia L. Rev. 2121 at 2200.CrossRefGoogle Scholar

8. See for example, Sugden, R. The Political Economy of Public Choice (Oxford: Martin Robertson, 1981) at 10.Google Scholar

9. Leading examples of the literature would include: Ackerman, B. “The Storrs Lectures: Discovering the Constitution,” (1984) 93 Yale L. J. 1013;Google Scholar Maclntyre, A. (Notre Dame: University of NotreDame Press, 1981); Google Scholar Michelman, F. “Foreword: Traces of Self-Government” (1986)100 Harv. L. Rev. 4;Google Scholar Michelman, F. “Law’s Republic” (1988)97 Yale L. J. 1493;CrossRefGoogle Scholar Sandel, M. “The Procedural Republic and the Unencumbered Self” (1984) 2 Political Theory 81;CrossRefGoogle Scholar Sunstein, C. “Interest Groups in American Public Law” (1985) 38 Stanford L. Rev. 29;CrossRefGoogle Scholar Sunstein, C. “Interest Groups in American Public Law” (1985) 38 Stanford L. Rev. 29;CrossRefGoogle Scholar Sunstein, C. “Beyond the Republican Revival” (1988) 97 Yale L. J. 1539; andCrossRefGoogle Scholar Taylo, C. “Kant’s Theory of Freedom” in Philosophy and the Human Sciences: Philosophical Papers (Vol. 2) (Cambridge: Cambridge University Press, (1985)CrossRefGoogle Scholar

10. For argument to this effect, see Posner, R. The Problems of Jurisprudence (Cambridge: Harvard University Press, (1990) at 418.Google Scholar Republicans admit the point themselves; see Sunstein, “Beyond the Republican Revival”, supra, note 9 at 1576.

11. Geoffrey Brennan has provided a very interesting economic argument to support the idea that individuals might reveal different and, arguably, less self-interested, preferences in political choices from those they would express over the same alternatives in private or market choices; see his paper “Irrational Action, Individual Sovereignty and Political Process: Why There Is a Coherent ‘Merit Goods’ Argument” in Brennan, G. & Walsh, C. eds, Rationality, Individualism and Public Policy (Canberra: Research on Federal Financial Relations,The Australian National University, 1990).Crudely, Brennan’s argument is that because oneșs vote is so unlikely to be decisive for the choice of any particular outcome in political processes, the cost of voting one’s conscience, or of the expressing of some higher order concern, is less expensive than it is in a market, where such a choice is, typically, absolutely decisive for what the individual chooser receives. Political processes, all thingsconsidered, result in quite different aggregate outcomes.Google Scholar

12. See, for example, Nagel, T. “Ruthlessness in Public Life” in Mortal Questions (Cambridge: Cambridge University Press, 1979) at 85–86: “ …the balance of outcome-oriented and action-oriented morality will justify the design of public institutions whose officials can do what would be unsuitable in private life. Some of the deviations will be conspicuously consequentialist; others will express the impersonality of public morality in other ways.” For some discussion of howthis role- or institution- dependent morality might affect the collective rationality conditions of social choice theory in a way that is analogous to some of the arguments in this paper,Google Scholar See, Chapman, B. “Individual Rights, Good Consequences, and the Theory of Social Choice” (1982)12 J. for the Theory of Soc. Behaviour 317.CrossRefGoogle Scholar

13. Farber, D. & Frickey, P. & Law and Public Choice (Chicago: University of Chicago Press, 1991) at 42–47for some suggestions that the results of social choice theory may have republican implications.CrossRefGoogle Scholar

14. Any review of social choice theory would exemplify this method of setting up the problem of social choice; see, for example, Sen, A. Collective Choice and Social Welfare (Edinburgh: Oliver and Boyd, 1970).Google Scholar

15. Both this notation and the definition of a chosen alternative are standard form in social choice theory. See, for example, Arrow, K. Social Choice and Individual Values (New Haven: Yale UniversityPress, 1963).at 15and Sen, supra, note 14 at 10Google Scholar

16. For a discussion of the necessity of satisfying acyclicity if social choice is to satisfy the usual definition of a chosen alternative, see Sen, supra, note 14 at 16.

17. Arrow introduces his problem of social choice by way of the paradox of majority voting as an example; see Arrow, supra, note 15 at 2.

18. 1 have used the following example in my “Law, Morality and the Logic of Choice: An Economist’s View” (1979) 29 U. of Tor. L. J. 114 at 124–30.

19. The defining conditions for the method of majority rule were first axiomatized in social choice theoretic terms by May, K.O. “A Set of Independent, Necessary and Sufficient Conditions for Simple Majority Decision” (1952) 20 Econometrica 680.CrossRefGoogle Scholar

20. See, for example, Sen, supra, note 14 at 166–86.

21. For good discussion, see Mueller, D. Public Choice II (Cambridge: Cambridge University Press, 1989).at 58–95.Google Scholar

22. Sagoff, supra, note 1 at 197–98.

23. See the discussion in Mueller, supra, note 21 at 58–65. A simple example should illustrate how poorly a pure distribution problem is handledby majority rule. Imagine that for a group of three individualsA, B, and C, a simple majority must decide how one hundred dollars is to be divided amongst them. A and B might decide to form a majority coalition and vote for dividing the money evenly between them, leaving C with nothing. Faced with this, C would likely approach A and propose giving him more, say $75, keeping $25 for himself, and freezing B out altogether. This coalition of A and C, which can beat the previous coalition of A and B, is in turn defeatable by the coalition of B and C wherein C gets $75 and B gets $25. But then this last coalition can be defeated by the first and the voting cycle begins anew.

24. Gregory, R. The Price of Amenity (London: Macmillan, 1971) at 111–12.Google Scholar

25. See the discussion in Chapman, supra, note 18 at 125–30.

26. For a proof, see Sugden, supra, note 8 at 157–58. In a Latin square, each of the alternatives in a set of three appears as most preferred (“best”) as least preferred (“worst”), and as “between”, forat least one individual in the population of voters.

27. See Sen, supra, note 14 at 168 and 173–86.

28. For a discussion that argues for choice hierarchies on slightly different grounds, see , “Individual Rights and Collective Rationality: Some Implications for Economic Analysis of Law” (1982) 10 Hofstra L. Rev. 455.Google Scholar

29. Of course, it is possible that a majority voting cycle could arise just at the lower level locationaldecision in a more complicated example involving more locational alternatives. Thus, the two level choice hierarchy so far described, which simply puts the activity level issue ahead of the location issue, would not suffice to avoid all possible voting cycles. But this only suggests that in a more complicated example we need a more complicated hierarchy of choice. We might, for example, have to break the locational alternatives into different types and vote on the different locational types in some sort of sequence. This would have the effect of imposing not-between value restriction on individual preferences at the locational level and, therefore, would avoid the cycle. For some discussion of a more complicated choice hierarchy, developed in a quite different context, see Chapman, B. & Trebilcock, M. “” (1992) 44 Rutgers Law Review 797, at 849–67.Google Scholar

30. James Buchanan is the name that most easily comes to mind; see, for example, his Freedom in Constitutional Contract (College Station: Texas A & M press, 1977). Also see Buchanan, J. & Tullock, G. The Calculus of Consent (Ann Arbor: University of Michigan Press, 1962).CrossRefGoogle Scholar

31. Rawls, J. A Theory of Justice Oxford: Oxford University Press, 1971).Google Scholar

32. For a list of examples of choice hierarchic thinking from legal theory, see Chapman, supra, note 28 at 470.

33. See Sagoff, supra,note 1 at 101–04. Also see Sen, A. “Rational Fools: A Critique of the Behavioral Foundation of Economic Theory” (1977) 6 Phil. & Pub. Affairs 317 at 337ff,Google Scholar and Taylor, C. “Self- Interpreting Animals” in his Human Agency and Language (Cambridge: Cambridge University Press, 1985) at 66ff.CrossRefGoogle Scholar

34. Schwartz, T. “"Rationality and the Myth of the Maximum” (1972) 6 Nous 97.CrossRefGoogle Scholar

35. This assumes that the said preference relation is connected, i.e., that all the alternatives can be compared as either better than, worse than, or at least as good as each other according to the choice criteria. Otherwise, definition (1) of a chosen element, which says only that x cannot be chosen if there is another alternative y which is better, is not equivalent to saying that x is chosen if.tr is at least as good as all other alternatives y. For some discussion of the possible importance of connectedness and this distinction, see Mclennen, E. Rationality and Dynamic Choice Cambridge: Cambridge University Press, (1990) at 35–40.CrossRefGoogle Scholar

36. Strictly speaking, Schwartz also requires that no non-empty propersubset of C(A) should satisfy (2). That is, if we can discard some inferior alternatives from C(A) without reducing C(A) to the null set,we should do so.

37. Blair, D. Bordes, G. Kelly, J. & Suzumura, K. “” (1976) 13 J. of Eco. Theory 361 at Theorem 2.Google Scholar

38. For the original characterizations, see Marquis de Condorcet, Essai sur VApplication de V Analyse ä la Probabilité des Decisions Rendues ä la Pluralité des Voix (Paris, 1785); and Chernoff, H. “Rational Selection of Decision Functions” (1954)22 Econometrica 422. CrossRefGoogle Scholar Condition GC and condition C are equivalent to what Sen refers to respectively as property γ and property a in Sen, A. “Social Choice Theory: A Re-examination” (1977) 45 Econometrica 53.CrossRefGoogle Scholar

39. For good discussion of choice consistency conditions in these contraction and expansion consistency terms, see Sen, supra, note 38.

40. This is an important linkage to make since many theorists, for example, Rawls supra, note 31 at 24–25, link maximization with the sorts of theories that do not take process values seriously. Process values loom large in importance later in this paper when we come to discuss legal decision-making. See infra. Section V. for a full argument on this, see Chapman, supra, note 28.

41. Arrow, infra, note 15 at 118–20.

42. See, e.g., Buchanan, J. “Social Choice, Democracy, and Free Markets” (1954) 62 J. of Pol. Eco. 114).Google Scholar For good discussion of Buchanan’s criticisms of Arrow, see Levi, I. Hard Choices (Cambridge: Cambridge University Press, 1986) at 150–57.CrossRefGoogle Scholar

43. Arrow, infra, note 15 at 120.

44. Plott, C. “” (1973) 41 Econometrica 1075.Google Scholar

45. Ibid, at 1079–80.

46. See Blair et al, supra, note 37 at Theorem 1.

47. Ibid.

48. This is because there is a way to relax acyclicity, namely by relaxing condition GC without relaxing either condition C or the superset condition S which combine together to make up PI. (Condition S requires that a choice from a superset should not be a proper subset of a choice from one of its own subsets; for a formal definition and a choice theoretic axiomatization of PI, see Blair etal, supra,note 37 at 366.) However, as the text tofollow indicates, this route is not available to us. For thetheorist who takes choice hierarchies seriously, PI and acyclicity stand or fall together because thechoice hierarchic form violates condition C.

49. Ibid.

50. For an argument that the path independence condition also makes a nonsense of taking “individual rights partitions” seriously in social choice theory, see Chapman, supra, note 28 at 468, and Chapman, B. “Rights as Constraints: Nozick versus Sen” (1983) 15 Theory and Decision 1.Google Scholar

51. Kramer, G. “Some Procedural Aspects of Majority Rule” in Chapman, J. & Pennock, J. eds, Due Process (New York: New York University Press, 1977).Google Scholar

52. Ibid,at 285.

53. Ibid,at 267–68.

54. Ibid,at 292.

55. Ibid, at 276–78. Unfortunately, Kramer refers to the condition as the “independence of irrelevant alternatives”, which is the name usually given to a very different condition used in the proof of Arrow’s impossibility theorem. Arrow is himself responsible for this sort of confusion because of an example he provides in his book, supra, note 15 at 27. For clarification, see Ray, P. “” (1973) 41 Econometrica 987 Google Scholar and Sen, A. “On Weights and Measures: Informational Constraints in Social Welfare Analysis” (1977) 45 Econometrica 1539 at 1563.CrossRefGoogle Scholar

56. Sagoff, supra, note 1 at 100: “…social regulation should be based on communitarian or public values, not on the personal preferences of individuals.... Public values are goals or intentions that people ascribe to the group or community of which they are members; such values are theirs because they believe and argue they should be theirs; people pursue these values not as individuals but as members of the group. They then share with other members of their community intersubjective intentions or, to speak roughly, common goals and aspirations, and it is by virtue of these that a group or community is a group or community.”(emphasis in original)

Compare the following remarks from Susan Hurley in Natural Reasons (Oxford: Oxford University Press, 1989) at 330: “My position remains a version of the classical view according to which individual preferences and interests cannot ultimately be understood independently of public values and shared reason-giving practices....Reeognizing and providing decision procedures for such collective alternativesis one of the basic tasks of domain and agenda division and institutional design in a democracy.”(emphasis added)

57. This may seem to have the contrast between parliamentary and judicial decision-making exactly the wrong way around. It might be thought that legislatures or parliaments decide the general descriptions of acceptable outcomes, leaving the final details to be worked out by the courts in more specifically focused adjudicative processes. However, this objection only goes to how legislatures and courts work together. In deciding on its own what legislative policy is to be chosen, however, parliament does tend to follow what Kramer calls “parliamentary procedure”, viz. that amendments to motions are voted on first, and then the motion (amended or not) is voted on against the status quo. This allows everyone to bring their given preferences, unlaundered by any kind of not-between value restriction, to bear on the final choice of these policy outcomes.

58. For excellent discussion of the difference between “judgement” under a legal rule or category and the use of a conclusive formula that can uniquely determine a conclusion, see , “Legal Formalism: On the Immanent Rationality of Law” (1988) 97 Yale L. J. 949 at 1006–07.Google Scholar

59. This was essentially Wittgenstein’s point about rule following in his Philosophical Investigations (Oxford: BlackweU, 1953) at paras. 117,198,199,201,211 and 219. For good discussion of the importance of Wittgenstein’s account of rule following for an understanding of law as a rational enterprise, see , “Revolution Without Foundation: The Grammar of Scepticism and Law” (1988) 33 Mcgill L.J. 451.Google Scholar

60. The arguments of the critics are effectively summarized (and critiqued) in Langille, supra, note 59.

61. Weinrib, supra, note 58 at 1009.

62. Arrow, supra, note 15.