(1) Aanund Hylland points out to me that Napoleon's dictum might be interpreted as asserting Pareto-optimality (‘tout pour le peuple’) while denying democracy (‘rien par le peuple’). My refusal of this conception does not extend to the first part.

(2) For a more elaborate presentation of the conception of rational choice underlying this exposition, I refer to some of my other writings on the topic: Ulysses and the Sirens (rev. ed.: Cambridge University Press, 1984); Sour Grapes (Cambridge University Press, 1983); The nature and scope of rational-choice explanation, in Le Pore, E. and McLaughlin, B.P. (eds.), Actions and Events: perspectives on the philosophy of Donald Davidson (Oxford, Blackwell, 1985), pp. 60–72Google Scholar; Introduction to Elster, J. (ed.), Rational Choice (Oxford, Blackwell, 1986), pp. 1–33Google Scholar, and ‘When rationality fails’, forthcoming in K. Cook and M. Levi (eds.), Limits to Rationality.

(3) We may note, for later reference, that this implies that if the agent is indifferent between x and y and between y and z, he must be indifferent between x and z, since indifference between two options can be defined by saying that each of them is at least as good as the other.

(4) See Bloch, S. and Reddaway, P., Russia's Political Hospitals (London, Futura Books, 1978), p. 255Google Scholar.

(5) If one could demonstrate that myopia, risk-seeking and other self-destructive phenomena are necessarily non-autonomous attitudes, the two definitions of preference-rationality would coincide. I cannot see, however, how this demonstration might be carried out.

(6) Feinberg, J., Harm to Self (Oxford University Press, 1986), p. 32Google Scholar.

(7) On this see also my Sour Grapes, Ch. 1, 3 and Nagel, T., The View from Nowhere (Oxford University Press, 1986), p. 113 ffGoogle Scholar.

(8) See my Ulysses and the Sirens, Ch. 11.

(9) Schelling, T.C., Choice and Consequence (Cambridge, Mass., Harvard University Press, 1984), Chs 3 and 4Google Scholar.

(10) Ainslie, G., Beyond microeconomics, in Elster, J. (ed.), The Multiple Self (Cambridge, Cambridge University Press, 1989), pp. 133–175Google Scholar.

(11) See my Sour Grapes, Ch. 11.

(12) Farber, L., Lying, Despair, Jealousy, Envy, Sex, Suicide, Drugs and the Good Life (New York, Basic Books, 1976)Google Scholar.

(13) It is often argued that in such cases one ought to assume, at least initially, that all states are equally likely. The problem, however, is that it may not be possible to make a non-arbitrary individuation of states. Consider a man who is known to have faced three roads, two going up the mountain and one going down into the valley. We know he has taken one of the roads, but we do not know which. Should we assume equal likelihood of the states ‘up’ and ‘down’, or equal likelihood of each road?

(14) See, however, Chs 1–5 in J. Elster (ed.), The Multiple Self, op. cit.

(15) Gibbard, A., Manipulation of voting schemes: a general result, Econometrica, XLI (1973), 587–601CrossRefGoogle Scholar; Manipulation of voting schemes that mix voting with chance, Econometrica, LXLV (1977), 665–81.

(16) For a rare defense of random voting see Note, Choosing representatives by lottery voting, Yale Law Journal, LXXXXIII (1984), 1283–1308.

(17) For details, see my forthcoming Tanner Lectures (Oxford University) on ‘Taming chance: randomization in individual and social decisions’.

(18) See for example Ordeshook, P.C., Game Theory and Political Theory (Cambridge University Press, 1986), Chs. 5–6CrossRefGoogle Scholar.

(19) Of course, people may not have any incentive at all to express their preferences, whether sincerely or insincerely. One ‘paradox of voting’ is that, given the extremely small probability that any one vote will be pivotal, the expected utility of voting to any individual can be expected to be smaller than the direct and indirect costs of voting. This is likely to be the case even if the utility of voting takes account of the possible benefits to other people, and certainly true if people make the decision to vote on purely selfish grounds. Since people do vote in large numbers, their behaviour cannot be purely selfish. This also holds out the hope that opportunism, or the expression of insincere preferences, might be limited by similar non-selfish motivations.

(20) A seminal article on the principal-agent problem is Jensen, M.C. and Meckling, W.H., Theory of the firm: managerial behavior, agency costs and ownership structure, Journal of Financial Economics, III (1976), 305–360CrossRefGoogle Scholar.

(21) For a survey, see Mueller, D.C., Public Choice (Cambridge University Press, 1979), Ch. 8Google Scholar.

(22) For two examples of societies in which the level of opportunistic behavior approaches that which is assumed by the public-choice school to prevail universally, see Banfield, E. O., The Moral Basis of a Backward Society (New York, The Free Press, 1958Google Scholar), and Turnbull, C., The Mountain People (New York, Simon and Schuster, 1972)Google Scholar. Fortunately, not all societies are like the village of Montegrano in Southern Italy of the Ik of Uganda.

(23) This is a main theme in Rawis, John, A Theory of Justice (Cambridge, Mass., Harvard University Press, 1971)Google Scholar.

(24) For the most recent statement of this approach, see Brennan, G. and Buchanan, J.M., The Reason of Rules (Cambridge University Press, 1986)CrossRefGoogle Scholar.

(25) Imagine a three-person society and three options, x, y and z. Writing P for strict individual preference, we may stipulate that their preferences are as follows xP₁yP₁z, yP₂xP₂z, xP₃zP₃y. On the unanimity rule, society would be indifferent between x and y and between y and z, since in neither case is one option unanimously preferred, yet x is unanimously and hence socially preferred to z.

(26) For a summary of some recent results, see Riker, W., Liberalism Against Populism (San Francisco, Freeman, 1982), Ch. 7Google Scholar.

(27) For this argument, see Goodin, R., Laundering preferences, in Elster, J. (ed.), Foundations of Social Choice Theory (Cambridge University Press, 1986), pp. 75–102Google Scholar, and J. Elster, The market and the forum, ibid. pp. 103–32.

(28) This suggestion has been made by Adam Przeworski in unpublished work.

(29) Polanyi, K., Personal Knowledge (New York, Harper, 1962)Google Scholar.

(30) Nelson, R. and Winter, S., An Evolutionary Theory of Economic Change (Cambridge, Mass., Harvard University Press, 1982), Ch. 4 and passimGoogle Scholar.

(31) See in particular the writings of Hayek, F.A., beginning with ‘Economics and knowledge’, Economica, n.s. XIII (1937), 33–54Google Scholar.

(32) Subtle form of dictatorships or oligarchy may, however, arise at the level of agenda manipulation. If a small elite can control the agenda or the voting procedure so as to obtain the result they want in each particular case, the fact that the process is formally democratic is irrelevant. Democracy (i.e. non-dictatorship) could be ensured by using the same procedure in all cases or (preferably) by choosing procedures randomly in each case.

(33) Tractatus Politicus, VII.1. I am grateful to E. Balibar for indicating this passage to me.

(34) For a discussion of the changing function of the Rule of Law, from a protection against absolute monarchy to a protection against absolute democracy, see the contributions of Sejersted, F.in Elster, J. and Slagstad, R. (eds.), Constitutionalism and Democracy (forthcoming from Cambridge University Press)Google Scholar.

(35) MacDowell, D.M., The Law in Classical Athens (London, Thames and Hudson, 1978), p. 48 ffGoogle Scholar.

(36) Ibid. p. 50 ff.

(37) Najemy, J., Corporatism and Consensus in Florentine Electoral Politics 1280–1400 (Chapel Hill, University of North Carolina Press, 1982)Google Scholar.

(38) See the essays in Elster and Slagstad (eds.), Constitutionalism and Democracy.

(39) Nordhaus, W., The political business cycle, Review of Economic Studies, XLII (1975), 169–190CrossRefGoogle Scholar, at p. 188.

(40) A useful but ultimately somewhat disappointing study of this problem is Edgerton, R.B., Rules, Exceptions, and Social Order (Berkeley, University of California Press, 1985)Google Scholar. Coming from an anthropological rather than a decision-theoretic tradition, he neglects the incentive effects of rules and the conflict between long-term stability and short-term expediency.

(42) G. Ainslie, Beyond microeconomics.

(42) Cass Sunstein had pointed out to me two interesting American cases, which well bring out the ambiguity of exceptions to constitutional rules. In Korematsu v. United States (323 U.S. 214(1944)) confining measures imposed on American citizens of Japanese ancestry were deemed constitutional. In New York Times Co. v. United States; United States v. Washington Post Co. (403 U.S. 713 (1971)) the attempt to stop publication of the Pentagon Papers was deemed unconstitutional. In both cases the defenders of the restrictive measures argued that the Constitution must not become a straightjacket on governmental action when the military security of the nation is at stake. The ‘Constitution is not a suicide pact’. It is unclear, however, whether these views represent (1) the kind of temptation which the Constitution was set up to prevent in the first place, (2) a legitimate concern that the Constitution might impose excessively tight bounds on the government or (3) an argument that the First Amendment can sometimes be overridden by other parts of the Constitution.

(43) Farber, op. cit.

(44) Quattrone, G. and Tversky, A., Self-deception and the voter's illusion, in Elster, J. (ed.), The Multiple Self, pp. 35–58, at p. 48Google Scholar.

(45) Democracy in America (New York, Anchor Books, 1969), p. 243Google Scholar.

(46) Ibid. p. 244.

(47) For further discussion of these two examples, see my Sour Grapes Ch. 11. 9, ‘Is there (or should there be) a right to work?’, in Guttman, A. (ed.), Democracy and the Welfare State (forthcoming from Princeton University Press)Google Scholar.

(48) For the notion of pragmatic contradiction see my Logic and Society (Chichester, Wiley, 1978)Google Scholar, Ch. 4 and Sour Grapes, Ch. 11.

(49) In Ch. 1.5 of Sour Grapes; see also ‘The market and the forum’.

(50) Hume, , Essays : Moral, Political and Literary (Oxford University Press, 1963), p. 40Google Scholar. Actually, self-interested action is not the worst-case assumption. Even worse results may follow if people act out of spite or, more pertinently, out of fanaticism. On this theme see Hirschman, A., The Passions and the Interests (Princeton University Press, 1977)Google Scholar.

(51) See, for instance, Tocqueville on Turgot, in L'Ancien Régime et la Révolution, vol. II (Paris, Gallimard, 1953), p. 425–426Google Scholar. ‘Ce qui confond, c'est de lui voir décider si vite un plan si nouveau dont les conséquences peuvent être immenses, en digérer si peu les idées, en proposer l'adoption immédiate sinon pour l'année courante, au moins pour l'année d'après, comme s'il ne s'agissait que d'une simple réforme administrative, qu'on peut faire sans solennité. Enfin, ce qui confond, c'est la confiance et la tranquillité d'espoir avec lesquelles il termine son mémoire’.

(52) There are, then, three objections to the unitary-actor model of political decision-making: (i) From the general principle of methodological individualism, it follows that the community cannot literally be seen as an actor with desires, beliefs and the capacity to act. (ii) From the more specific arguments set out in (III) it follows that one cannot even in general speak as if the community were a unitary actor, (iii) Even assuming that one could, the scope for rational action would be severely limited because of the problem of uncertainty.

(53) For a discussion of gradient-climbing, see Ch. 1 of my Ulysses and the Sirens.

(54) Lipset, R.G. and Lancaster, K., The general theory of the second-best, Review of Economic Studies, XXIV (1956–1957), 11–32, at p. 12CrossRefGoogle Scholar. See also Margalit, A., Ideals and second bests, in Fox, S. (ed.), Philosophy for Education (Jerusalem, Van Leer Foundation, 1983), pp. 77–90Google Scholar.

(55) See T. Bauer, The unclearing market, in J. Elster and K.O. Moene (eds.), Alternatives to Capitalism, forthcoming.

(56) Putterman, L., Some behavioral perspectives on the dominance of hierarchical over democratic forms of enterprise, Journal of Economic Behavior and Organization, III (1982), 139–160, at p. 152CrossRefGoogle Scholar.

(57) Putterman, Critique of the Gotha Programme, op. cit. p. 149.

(58) See my ‘Consequences of constitutional choice: reflections on Tocqueville’, forthcoming in Elster and Slagstad(eds.), Constitutionalism and Democracy.

(59) The following draws heavily upon Williamson, C., American Suffrage: from property to democracy 1760–1860 (Princeton University Press, 1960)Google Scholar; McGovney, D.O., The American Suffrage Medley (University of Chicago Press, 1949)Google Scholar; Seymour, C. and Frary, D.P., How the World Votes, vols I-II (Springfield, Mass., G.A. Nichols, 1918)Google Scholar; Metthews, H.C.H., McKibbin, R.I. and Kay, J.A., The franchise factor in the rise of the Labour Party, English Historical Review, XCI (1986), 723–752Google Scholar.

(60) Actually, the argument only shows that bribing rich voters is more costly, which may be offset by the fact that with suffrage restricted to the rich there are fewer voters to be bribed.

(61) Stephen Holmes has pointed out to me that the Romans imposed economic conditions on the right to vote in order to elicit information from the citizens about their taxable property. In theory, this rule could also serve the purpose of sorting out the citizens who were sufficiently concerned about voting to risk bringing their property to the attention of the authorities. The richer the citizens, the greater their public-spiritedness would have to be for them to overcome this obstacle—or the greater their expected gains from participating in politics.

(62) See Ely, J., Democracy and Distrust (Cambridge, Mass., Harvard University Press, 1980), pp. 31, 120 ff., 146 ffGoogle Scholar.

(63) Athenian citizens were disfranchised for cowardice in war and for unpaid debts to the state (Mac Dowell, op. cit. pp. 160, 165).

(64) Ely, op. cit. p. 120.

(65) Thus Aiskhines in the speech Against Timarkhos: ‘The legislator considered it impossible for the same man to be bad privately and good publicly’ (cited after Mac Dowell, op. cit. p. 174).

(66) This is, of course, how rich countries today decide on aid to poor countries.

(67) Ely, op. cit. p. 120 ff.

(68) See for instance Freeman, J.R. and Snidal, D., Diffusion, development and democratization: enfranchisement in Western Europe, Canadian Journal of Political Science, XV (1982), 299–329CrossRefGoogle Scholar.

(69) One might consider a purely instrumental-utilitarian argument for extension of the suffrage, by arguing that the removal of this degrading discrimination ipso facto represented a gain in welfare. Yet again this instrumental consideration would be parasitic on a non-instrumental one, namely the perceived inherent injustice in the unequal treatment.

(70) Strictly speaking, there are no instances of compulsory saving. Yet we may distinguish between the welfare services, such as old age pensions, in which the element of saving is large and the element of risk-pooling is small, and those in which the latter predominates.

(71) Formally, this is often presented as contributions of the employer. Economists agree, however, that these are de facto payroll deductions, in the sense that without compulsory employer contribution the salaries of employes would have been higher by the same amount (Page, B., Who Gets What from Government? (Berkeley, University of California Press, 1983), p. 28)Google Scholar.

(72) Abraham, K., Distributing Risk (New Haven, Yale University Press, 1986), p. 84CrossRefGoogle Scholar.

(73) Ibid. p. 92.

(74) Ibid. p. 99.

(75) Page, op. cit., pp. 67, 75.

(76) Dworkin, R., What is equality? Part 2; Equality of resources, Philosophy and Public Affairs, X (1981), 283–345Google Scholar. For an illustration, consider three individuals, A, B and C. A and B have the same skills. B and C produce the same output. A and C work the same number of hours. In other words, the unskilled B is able to produce as much as the skilled C because he is willing to work more hours than his unskilled colleague A. Dworkin would endorse higher income to B, and relegate A and C to the same, lower income level. The principle ‘To each according to his unweighed labour time’ was also used by the early British socialists (see Pagano, U., Work and Welfare in Economic Theory (Oxford, Blackwell, 1985), Ch. 2.3Google Scholar.

(77) See notably Roemer, J., Equality of talents, Economics and Philosophy, I (1985), 151–187CrossRefGoogle Scholar.

(78) There are two aspects to ambition. On the one hand, one may need to encourage ambition for everybody's sake. Any theory of justice must take account of the need to pay people more when necessary to elicit socially useful work. On the other hand, one may (with Dworkin) or may not (with Rawls and utilitarianism) think that ambition is a morally relevant ground for higher rewards. Consider three equally qualified workers, A, B and C who choose to work, respectively, 4, 8 and 12 hours a day at a given wage. One may believe (with Rawls and utilitarianism) that it is morally justified to tax C (and perhaps B), and use the proceeds to subsidize A, and yet stop short of the tax rate that would make the three equally well off if this makes A worse off (Rawls) or reduces total welfare (utilitarianism) compared to a lower tax rate. Dworkin, however, would not accept any taxation on earning differentials due to ambition rather than skills.

(79) For the development of this proposal up to 1982, see S.I. Albrecht and Deutsch, S., The challenge of economic democracy: the case of Sweden, Economic and Industrial Democracy, IV (1983), 287–320Google Scholar, and Myrdal, H.G., Collective wage earner funds in Sweden, International Labour Review, CXX (1981), 319–334Google Scholar. The proposal finally adopted differs in some aspects from the versions discussed in these articles.

(80) Meade, , Efficiency, Equality, and the Ownership of Property (London, Allen and Unwin, 1964)Google Scholar; Krouse, R. and Mac Pherson, M., A ‘mixed’-property regime: equality and liberty in a market economy, Ethics, LXXXXVII (1986), 119–128CrossRefGoogle Scholar. See also my ‘Comments on Krouse and McPherson’. ibid. 146–53.

(81) See van Parijs, Philippe and van der Veen, R., The transition from capitalism to communism, Theory and Society, XVI (1987)Google Scholar, as well as my ‘Comments’ on their proposal, ibid.

(82) Weitzman, M., The Share Economy (Cambridge, Mass., Harvard University Press, 1984)Google Scholar.

(83) For discussion and further references, see the introduction to Elster and Moene (eds.), Alternatives to Capitalism.

(84) Meade, op. cit. p. 56–57.

(85) See my ‘Comments on Krouse and McPherson’.

(86) Here Meade is more realistic, when he argues that ‘investments [would have to be] chosen by specialists on behalf of the man in the street’ (op. cit. p. 40).

(87) For a related argument, see Frank, R.H., Choosing the Right Pond (Oxford University Press, 1985), p. 256 ffGoogle Scholar.

(88) See P. van Parijs and R. van der Veen, ‘Reply to six critics’, forthcoming in Theory and Society.

(89) For this argument see Shapiro, C. and Stiglitz, , Equilibrium unemployment as a worker discipline device, American Economic Review, LXXIV (1984), 433–444Google Scholar.

(90) Nuti, D.M., The share economy: plausibility and viability of Weitzman's model (European University, Florence : Working Paper No. 85[194 from the Department of Economics, 1985)Google Scholar.

(91) I am indebted to Karl O. Moene for pressing this point on me.

(92) See my ‘Is there (of should there be) a right to work?’ and ‘Self-realization in work and politics’, Social Philosophy and Policy, III (1986), 97–126Google Scholar.

(93) Dahl, , A Preface to Economic Democracy (Berkeley, University of California Press, 1985)Google Scholar.

* I thank Aanund Hylland, Karl O. Moene and Cass Sunstein for probing and helpful comments on an earlier draft of this article, and King K. Tsao for invaluable research assistance.

(94) Grua, G. (ed.), Leibniz : Textes inédits (Paris, Presses Universitaires de France, 1948), p. 94Google Scholar.