"Mill's Principle of Government as a Basis of Democracy," The Monist, Vol. 55, No. 1 (1971), pp. 51-60.

by Gordon A. Welty
Wright State University
Dayton, OH 45435 USA

[/51] It has now become historically possible for rational men to reject the basic
assumptions of the social order which made Hobbes' account of political behavior plausible.
- G. Lichtheim

The founders and 19th-Century leaders of the American democratic experiment sought to base their institution on a rationalist and individualist model of political reality. As characterized by D. M. Levitan, they distrusted representative government and powerful executives, subscribing to a laissez-faire philosophy: he governs best who governs least./1/ Levitan goes on to accentuate the intimate relationship of the political system to its philosophical foundation. While he notes that liberal ideas were well adapted to the needs of the plutocracy, he does not play down the importance of Liberalism to the particular form taken by political institutions in democracy./2/ A paradigm of this Liberal doctrine is John Stuart Mill's famous Principle: "the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will is to prevent harm to others."/3/ This Principle, if not the source, can certainly qualify as a rationalization for a great part of what passes for "policy" in a democracy. [51/52]

Indeed, if Mill's Principle is both necessary and sufficient for the greatest happiness for the greatest number, as Mill thought it to be, it is of prime importance today to the study of democracy, quite apart from the social factors of its genesis. Its individualist thesis, its simplicity, and its widespread recognition all recommend it as a basis for democratic theory. These virtues, perhaps, account for its recurrence in contemporary ethical theorizing./4/

Since J.F. Stephen's attack,/5/ to the Devlin-Hart exchange/6/ and Wolff's superb analysis/7/ of our day, criticism and defense of Mill's Principle continues. Our discussion will focus on the logical relationship between the Principle and the social welfare optimum (the greatest happiness for the greatest number).

Two points should be emphasized at the outset. First, we are concerned with the relevance of the Principle for contemporary social science and social engineering. We are not treating a man or his age. Second, as we are not particularly interested in textual exegesis, we will develop the discussion along lines of morphological rather than philological research.

The "Very Simple Principle"
Mill proposed a Principle which, he supposed, would "govern absolutely the dealings of society with the individual in the way of compulsion and control."/8/ As we have seen, this "very simple principle" permitted each individual to pursue his good in his own way; permitted each actor was autonomous./9/ The Principle sought to establish [52/53] "theoretical limits" to the state and its legitimate interference with the action of the individual.

In proposing these theoretical limits, Mill conceived two domains of action of the individual. One was wholly personal, void interpersonal interaction; hence, totally exempt from society's interference. The other domain was properly social, alone where the Principle was to hold sway./10/ The criterion for this distinction was the potential for harm, but as we shall see, Mill was unable to maintain the distinction.

He acknowledges that even legitimate freedom of action can be represented as harmful. He admits that plausible arguments can be constructed to malign even innocuous and private behavior./11/ This makes a shambles of the distinction between the wholly private and the social domains of action since, on Mill's argument, there is no criterion for private and social./12/

Even though the social versus private distinction cannot be maintained, the relationship between the Principle and the greater good is still to be established. Society's acts of constraints are not for Mill, "merely a subcategory of actions in general," as Wolff suggests./13/ Thus they are not necessarily within the domain of the good. A consideration of the social structure envisioned by Mill makes this clear.

At the basis of the Principle is the logical truth that when no actor can modify the norms or rules defining social interaction, then no actor can change the norms to his benefit. Thus, a mechanical symmetry is established: the social system is indifferent in its payoffs, to location (or "label") of the recipient.

An examination of this symmetry reveals several conditions. Actors accept the norms as given (for a finite number of actors); these norms define legitimate behavior in all cases; and the norms are applied equally in all cases./14/ If these conditions are sustained, the Principle guarantees the greatest good. Let us make a remark on the first of these conditions. [53/54]

The first condition asserts that the actors are ruletakers. In neoclassical economic theory, it was assumed that the number of individuals was so large that the possibility of any actor effecting a change of price (the norm) was nil. Strictly speaking, "so large" is infinitely large./15/

Since there is always a finite number of actors, Mill patches up the argument by introducing an omniscient/16/ impartial administrator. Omniscience is theoretically required by the need for making interpersonal cardinal utility judgements. (The practical problems of this requirement will be noted later.) This administrator will hold the norms of social interaction fixed and rectify any violations of their fixity (violations which occur due to the finite number of actors, and which result in lesser good). The economist would call this administrator a "regulatory agency,"/17/ while Mill labels it "society."

At this point it is clear that Mill has ignored the double-edged sword of Benthamite "interests."/18/ To guarantee what Bentham called the "dexter side" of the individual's action, Mill introduced an administrator subject to no such guarantee. Indeed, on the basis of independence of interests (an assumption promulgated by Mill in his Principle), Bentham had concluded that the public servant would have "sinister interests."/19/ To accept that social and individual interests coincide would debilitate Mill's entire argument; to assume independence leaves the individual at the mercy of society; either is contrary to Wolff's supposition.

The Principle as a Necessary Condition
It would be highly desirable to the Liberal to show that Mill's Principle was necessary for the greater good. Mill did consider it to be necessary, arguing by contraposition. This can be illustrated by his comparison of Europe and China. He finds the benefits of progress and the attendant good all declining in Europe, precisely [54/55] because it is moving away from individualism towards "the Chinese ideal of making all people alike."/20/ A decline of individualism, i.e. the negation of the Principle, suggests to Mill a decline of the good.

To reject the necessity of the Principle it suffices to show that, in outcome, some other set of assumptions is its equivalent. One other such set is that of an n-person game.

An essential n-person game has a number of coalitions. But unlike Mill's "combinations," these coalitions can and will have an effect on the interests of others. Further, the members of the coalitions can recombine at will, distinct from Mill's reservations./21/ This property affects one of the ends of the n-person game, which is to achieve social stability rather than merely to maximize utility.

For relatively small n's, game theory does not require the introduction of the impartial administrator as did Mill's. This is quite an important point, as Mill's system degenerates in the absence of the administrator and the concept of an administrator to fill Mill's requirements is highly implausible, even for Mill./22/

Suppose that a given coalition can assure itself of a set of payoffs of value v. This is a subset of all payoffs. Obviously, this requires that the set of payoffs to a coalition must be at least as great as the sum of payoffs to its members operating individually, for the coalition to continue to exist. Call the payoff to the i-th agent, functioning individually f(i). Then we require

f(1) + f(2) + f(3) + ... + f(n) < v.

This is the property of superadditivity, which makes individualism, in Mill's sense, a disvalue./23/

Lloyd Shapley has shown that if a coalition is balanced then the core of the game is not empty./24/ From this it follows that there is a set of outcomes to the game such that no coalition can be formed to obtain for its members more than is provided by membership in a coalition in the core. While Shapley's proof involved an assumption [55/56] of side payments (i.e. a medium of exchange such as money was required), Herbert Scarf has proved that the argument can be extended to the case of games without side payments./25/ It has further been shown that as n grows large, the core shrinks to the value given by the realization of the Principle./26/

Thus, the existence of the general good can imply an n-person game which could well include violations, in its "play," of any or all of the norms Mill stipulates./27/ As such, the realization of the Principle is clearly not necessary for the general good. We turn then to the question of the Principle's sufficiency.

The Principle as a Sufficient Condition
Mill considered the Principle to be a sufficient condition for a social optimum. He argues from the Principle to the general good. Arguing directly, Mill asserts: "Mankind are greater gainers by suffering each other to live as seems good to themselves than by compelling each to live as seems good to the rest."/28/ Mankind's gain is postulated to be conditional on the realization of the Principle.

As Mill continues, it is clearly crucial to distinguish actions which are "prejudicial to the interests of others."/29/ It is prima facie evidence that an action is prejudicial to another if it causes the other person physical pain. (Pathological response to pain only further complicates the problem noted here.)/30/

As part of the difficulty in anticipating or identifying those [56/57] actions which are harmful, however, consider that perceptions of pain, hence the behavioral correlates of pain, vary considerably with the person. As W.K. Livingston puts it: "Pain is a perception; it is subjective and individual..."/31/ If the sum of its variation was the magnitude of the error (a reliability problem), it would be more difficult (by its very nature) to predict accurately the effects of an action in producing pain or to explain adequately the experience of pain as the consequent of a given action. But the variation is not simply to be attributed to individual differences. There is a further problem of validity.

A taxonomic problem arises as the perception of pain varies with psychological state. H.K. Beecher's comparative study of wounded soldiers vs. civilians who had sustained comparable tissue damage in surgery, indicated that leaving the battlefield honorably made pain more tolerable than was the pain of operation to the civilian./32/

Moreover, the perception of pain varies with psychosocial and ethnic factors. Beginning with Josey and Miller's study,/33/ a survey of the literature indicates that behavioral correlates of pain have been found to vary with ethnic group,/34/ with age,/35/ and perhaps even with sex./36/ For instance, the threshold of pain for Jewish subjects is lower than that of Anglo-Saxons; the threshold of the aged is lower than that of the young, etc. The list of factors is far from complete. It would appear that there are a large number of factors, presently not all identified, which would need to be [57/58] accounted for in the prediction or explanation of harmful actions.

Our argument to this point can be summarized: In the absence of a viable taxonomy of "harmful" actions, the Principle will not be sufficient for differential treatment, either punitive or preventive by society. Mill nowhere provides this taxonomy. Certainly an appeal to common sense is precluded by his rejection of mass opinion. Today we might suggest that Mill avoids the naturalistic fallacy by not appealing to common sense or the common man.

But by providing only a generic and not a specific scheme for determining the good (or the harmful) he renders his Principle inoperable. Indeed, in terms of practical application, more than just a taxonomy, perhaps a complete theory of human action is required.

On analytic grounds, the insufficiency of the Principle can be demonstrated. Thus, more than impractical, the Principle can be shown impossible to achieve a social optimum.

Make a list of all the humans who have ever lived, from the beginning of time. Assume these men are autonomous, with perfect knowledge. This assumption means that we can eliminate social or governmental intervention from the discussion, by virtue of the Principle. Does the social optimum follow?

Since all men are autonomous, all social interaction is voluntary and mutually advantageous. If agent A gives a good to B, while B does nothing in return, then B must be giving a goodto agent C who does good (possibly via a longer chain) to A.

Let us assume there is no institutional or physical provision for intertemporal transfers of goods (which excludes the possibility of an agent hoarding for his future retirement.) This is to say an agent cannot trade with Mother Nature. Now let us assume there are two sorts of people. Type A are wage earners and type B are retired or disabled. Since type A is the totality of wage earners, they must give all type B persons sustenance. Mr. A expects some good in return, and must get it later in life, when he retires. Who will provide this good to A? Not B, who will probably be dead by then and isn't a wage earner anyhow.

Let us assume when type A becomes retired, they will get sustenance from a new group of wage earners, C. To complete the mutually advantageous set of transactions, we must ask if B can do [58/59] good for C. Certainly not, for B has ceased to be a wage earner long before the time any type C persons appear.

Hence, either type B (retired) people will die at the beginning of retirement, or conversely will need to be wage earners to the moment of death. This is because no type A (or type C) person will sustain a retired person in the absence of the guarantee of future remuneration. Under the Principle, this guarantee is of course impossible./37/ On the basis of this highly simplified and schematized argument, we see how Samuelson/38/ and others/39/ have proved the following impossibility theorem: The social optimum can never be reached by a simple laissez-faire system, such as that exemplified by Mill's Principle. This, despite the fact that each agent is maximizing his own utility. Further, no agent is threatening the interests of any other.

Two alternatives/40/ appear: (a) deny the plausibility of assuming the existence of retired or disabled persons, and maintain the Principle, or (b) reject the Principle as insufficient for the social optimum and propose that society can interfere in the activities of the individual, even though these activities are harmful to him alone.

The Principle as a Tautology
In conclusion, we will summarize our discussion, and make a conjecture. The Principle which Mill forwarded to "absolutely govern" the society-individual interface has been found neither necessary nor sufficient for the social optimum. Not necessary [59/60] because the realization of the general good may as well imply the outcome of an n-person game. This game makes substantially different assumptions than the social atomism of the Principle. It does not appear to be sufficient because there is no way to operationalize the Principle. Further, under plausible assumptions about the working life of man, the Principle does not realize the social optimum at all.

We then make a conjecture. The Principle is a tautology. Mill has included far more than an allegedly minimal set of social activities in the collection of legitimate interferences. He has included not only redistributive functions of society, but equity and productive functions as well.

Mill overtly includes only redistribution; he acknowledges society's right to reequilibrate the good following an unjust act of an individual. But as Gert Muller has persuasively argued, equity and productive functions are also necessary in the minimal set of social activities,/41/ and Mill has surreptitiously included them in the set legitimate under the Principle. Consider, for instance, the discussion of taxation/42/ and that of education and minority age./43/ This inclusion by Mill requires an assumption of the `compressibility' of social functions, which is a variant of psychologism. It appears more plausible to reject this radical individualism.

Our conjecture has an even more far-reaching base: the prejudice of 19th-Century thought was to a mathematics of analytic equations. Among other things, this prejudice included deterministic and mechanical world views. The Principle certainly illustrates this world view. The development of a mathematics of nonanalytic equations makes the shortcomings of compressibility of social functions more stark, as well as providing new possibilities for formal models of the foundation of democracy. [60/]

Footnotes

* I want to thank Professor Oskar Morgenstern for helpful comments and criticism, and L. Taylor-Dancy and G.E. Lundin for editorial assistance. Of course, they are not responsible for any errors remaining or opinions expressed in this essay.

1. David M. Levitan "The Responsibility of Administrative Officials in a Democratic Society," Political Science Quarterly, Vol. 46 (1946), pp. 562-63. Hereinafter cited as "Resp. Ad. Of."

2. Levitan, "Resp. Ad. Of.," p. 563.

3. John Stuart Mill, On Liberty, ed. C.V. Shields (Indianapolis: Bobbs-Merrill, 1956), p. 13. Hereinafter cited as OL.

4. Essentially the same Principle is found in Kurt Baier Moral Point of View (Ithaca: Cornell University Press, 1958), p. 230 [hereinafter cited as MPV]; and in Georg Henrik von Wright Varieties of Goodness (New York: Humanities Press, 1963), p. 202. Hereinafter cited as VG. We will note further similarities throughout this essay.

5. James FitzJames Stephen Liberty, Equality, Fraternity (London: Smith, Elder & Co., 1893).

6. Patrick Devlin The Enforcement of Morals (New York: Oxford University Press, 1968) [hereinafter cited as EM]; also H.L.A. Hart Law, Liberty and Morality (Stanford: Stanford University Press, 1963).

7. Robert P. Wolff The Poverty of Liberalism (Boston: Beacon Press, 1968). Hereinafter cited as PL.

8. Mill, OL, p. 13.

9. Devlin, EM, p. 106.

10. Mill, OL, pp. 13-14.

11. Mill, OL, pp. 116-117.

12. Cf. also Wolff, PL, pp. 23-25.

13. Wolff, PL, p. 7.

14. Cf. Baier, MPV, pp. 207-208.

15. Robert J. Aumann, "Markets with a Continuum of Traders," Econometrica, Vol. 32 (1964), pp. 39-41, 47-49.

16. Baier, MPV, p. 202 refers to the necessity of a "God's-eye point of view."

17. Baier, MPV, p. 206 calls it "a disinterested official."

18. Cf. J. Bentham Works, ed. John Bowring (Edinburgh: Tait, 1843), Vol. 6, p. 258.

19. J. Bentham Plan of Parliamentary Reform (London: Hunter, 1817), p. 223.

20. Mill, OL, p. 88.

21. For Mill's reservations on combinations, cf. OL, p. 16, pp. 124 ff.

22. Cf. Mill, OL, p. 102.

23. G.H. von Wright, VG, p. 195 recognizes this superadditivity.

24. Robert J. Aumann, "Survey of Cooperative Games Without Side Payments," in M. Shubik (ed.), Essays in Mathematical Economics (Princeton: Princeton University Press, 1967), p. 11.

25. Aumann, ibid.

26. Gerard Debreu and Herbert Scarf, "A Limit Theorem on the Core of an Economy," International Economic Review, Vol. 4

(1963), pp. 235-46. For further discussion of n-person games, see Martin Shubik, "Welfare, Economic Structure, and Game Theoretic Solutions," in Fritz Zwicky and A.G. Wilson (eds.) New Methods of Thought and Procedure (New York: Springer Verlag, 1967), pp. 228-45.

27. Cf. his list in OL, p. 95.

28. Mill, OL, p. 17.

29. Mill, OL, p. 114.

30. We might note that the specification of a taxonomy of the good and the harmful is empirical for a Utilitarian. If x is good for some actor y, then x is relevant to the taxonomy. If x has no such predicate for any actor, it is irrelevant to the discussion. Thus, questions of justice, etc., never arise as conceptual, but only practical problems, in the Utilitarian analysis.

31. W.K. Livingston, Pain Mechanisms (New York: Macmillan and Co., 1943), p. 62.

32. Henry Beecher, "Relationship of Significance of Wound to the Pain Experience," Journal of the American Medical Association, Vol. 161 (1956), pp. 1609-13.

33. C.C. Josey and C.H. Miller, "Race, Sex, and Class Differences in Ability to Endure Pain," Journal of Social Psychology, Vol. 3 (1932), pp. 374-76.

34. Mark Zborowski, "Cultural Components in Response to Pain,"Journal of Social Issues, Vol. 8 (1952), pp. 16-30.

35. E.G. Sherman and E. Robillard, "Sensitivity to Pain in the Aged," Canadian Medical Association Journal, Vol. 83 (1960), pp. 944-47.

36. Victor A. Christopherson, et al., "Social-Cultural Correlates of Pain Response," Final Report of Project No. 1390 (Tucson: University of Arizona Press, 1966).

37. For instance, K. Baier, MPV, p. 230, would reject the possibility of a societal guarantee by rejecting "meddling" as nothing but "paternalism." Of course, interpersonal guarantees are excluded ex hypothesi.

38. Paul A. Samuelson, "An Exact Consumption-Loan Model of Interest With or Without the Social Contrivance of Money," Journal of Political Economy, Vol. 66 (1958) pp. 467-82, esp. 476-78. Samuelson's discussion is more elaborate and interesting than that presented here. Our simplification, however, suffices our purposes.

39. Cf. Jerome L. Stein, "A Minimal Role of Government in Achieving Optimal Growth," Economica, Vol. 36 (1969), pp. 139-50 for further discussion and references.

40. A third alternative proposed by Frank Knight in Risk, Uncertainty and Profit (Boston: Houghton Mifflin, 1921; 1937), has great merit: abandon the individual as the unit of analysis, in favor of the family or household. But this abandons the whole of the social atomism of the Principle.

41. Cf. Gert Muller "Umriss einer analytischen Soziologie," Sociologia Internationalis, Vol. 5 (1967), pp. 141-66; and "History as a Rigorous Discipline," History and Theory, Vol. VI (1967), pp. 310-11.

42. Mill, OL, p. 123.

43. Mill, OL, p. 13; pp. 128-30.

-o-