Front Page Titles (by Subject) Inventing the State: The Social Contract - The State
The Online Library of Liberty
A project of Liberty Fund, Inc.
Search this Title:
Inventing the State: The Social Contract - Anthony de Jasay, The State 
The State (Indianapolis: Liberty Fund, 1998).
About Liberty Fund:
Liberty Fund, Inc. is a private, educational foundation established to encourage the study of the ideal of a society of free and responsible individuals.
The copyright to this edition, in both print and electronic forms, is held by Liberty Fund, Inc.
Fair use statement:
This material is put online to further the educational goals of Liberty Fund, Inc. Unless otherwise stated in the Copyright Information section above, this material may be used freely for educational and academic purposes. It may not be used in any way for profit.
Inventing the State: The Social Contract
Political hedonism requires a benign state or a conformist subject. Failing both, it is a foolhardy attitude.
Hobbes, who could be mischievous, saw that every man has reason to fear his fellow man if they are alike.
All men, needing self-approval, seek eminence over others. If I let my fellow man seek eminence, he will invade my property, therefore I must attack his first. Self-preservation must drive both of us to fight each other, and there will be “savage war for glory.” Both our lives will become “nasty, brutish and short.”
While self-preservation is said to be the spring of all Hobbesian conduct, it is clear that I would not have to worry about preserving myself if my neighbour, whether to become eminent or to forestall me, did not invade my property. Is there a way of persuading the neighbour to desist? Perhaps by letting him know that I am not seeking eminence over him and he has nothing to fear? If self-preservation no longer obliged him to keep up his guard, and he lowered it, I could pounce and gain eminence over him; and so could he if I agreed to let him be and lowered my guard. As he is like me, I have to fear him, and cannot prudently make the first step which would break the vicious circle if he were unlike me.
In modern decision theory, such situations are called “prisoners’ dilemmas.”15 As set up, they have no spontaneous cooperative solution. Left to themselves, both “prisoners” must, if they are rational, seek to get the better of each other by “confessing” first, and both end up with a longer sentence than if they had both played “thief’s honour” and refused to confess. In Hobbes, they both end up with a shorter and nastier life. Their sole escape is to abandon the state of nature and conclude a “covenant of mutual trust” whereby a designated sovereign is invested with whatever power it takes to enforce peace (or natural right). Thus nobody need fear that, by behaving trustingly, he will be taken advantage of by the others; therefore all can behave trustingly. The sovereign will, for some reason, use his absolute power only for obtaining this result. His subjects have no right to rebel but nor do they have any reason for doing so. It is not clear whether, if they did have cause, they would have a right to rebel.
The prisoners’ dilemma implicit in Hobbes requires, for its proper study, the state of nature where no sovereign authority stops the participants from making themselves miserable if they are so inclined.16 States are in a state of nature in that they retain the faculty of recourse to force against each other and do not transfer their arms and their sovereignty to a super-state.17
I will consider, in this context, two Hobbesian dilemmas, those of war and trade. While I am at it, I will go on briefly to look at Rousseau’s problem of general social cooperation also, though the latter is quite different in nature (it is not a “prisoners’ dilemma” and requires a special psychological assumption in order not to result in voluntary cooperation).
Let there be two sovereign countries (to borrow the language of army manoeuvres, “Blue” and “Red”). Both want “eminence” in Hobbes’s sense. The order of their preferences is: (1) victory in war, (2) disarmament, (3) armed peace and (4) defeat in war. They must choose between two “strategies”—arming and disarming—without knowing what the other country chooses. The “pay-off matrix” resulting from this situation will then be as in figure 1.
Though Blue does not know whether Red will arm or disarm, he will choose to arm because by doing so he avoids defeat, gets peace at a cost as the worst-case pay-off and may get victory if Red is a sucker. Red is like Blue, and reasons similarly. He, too, chooses to arm. They end up in the southeast corner of the figure, in armed peace which is the “maximin” (the best worst-case) solution proper to hostile players. The northwest corner of costless peace is denied them, though they would both prefer it, because of their even greater preference for victory over each other. Once in the northwest corner, Blue would try to go into the southwest and Red into the northeast quadrant, i.e. the “cooperative solution” of costless peace would be unstable in the absence of a super-state enforcing disarmament.
This is, broadly, the result we actually find in the real world. States are most of the time in the southeast quadrant of the figure, i.e. they live in costly armed peace. From time to time they slip into the southwest or northeast quadrant and make war. Whether this is by virtue of unequal arms, a freak cause, or for another of the innumerable historical causes of war, is beyond our present concern. Despite their preference for northwest over southeast, however, they do not surrender sovereignty. We must carefully note this fact and consider it presently.
The dilemma of trade is formally identical to the dilemma of war. Let there be the same two countries, Red and Blue. Each wants the other’s goods. Both have the same order of preferences: (1) get foreign goods for free, (2) trade home goods for foreign goods, (3) retain the home goods (no trade), and (4) forgo the home goods and get no foreign goods (total loss, confiscation, expropriation, write-off). The two countries contract to deliver goods to each other (or to lend for later repayment, or invest for a return). As there is no enforcing super-state, they can either perform the contract or default, as in figure 2.
Game theory would once again predict that neither trader will give the other the chance to play him for a sucker, so that “maximin” is the dominant strategy for both and they end up not trading. The structure of their preferences and the structure of the pay-offs jointly deny them the benefit of trading in the absence of a contract-enforcer. This prediction, of course, is belied by the widespread fact of trade, investment and lending across national jurisdictions, which those who engage in them find on the whole worthwhile in the face of a certain frequency of bad debts and defaults of one kind or another. States are in certain circumstances even prepared to give redress to foreign nationals and enforce performance by their own defaulting nationals; an altogether quixotic act by the standard conceptions of basic social contract theory. Equally quixotic is the voluntary submission, by medieval traders and bankers, of cases of default or disputed contract performance to the judgments of their peers appointed for the purpose but possessing no arms and commanding no police, especially when you consider the danger that the decision might have gone against them!
If history demonstrates that two ostensibly identical dilemmas regularly give rise to contrasting outcomes, the war dilemma resulting in armed peace (with occasional war) and the trade dilemma resulting in trade, the ostensible identity must hide some significant difference. Intuitively, war is more easily seen as a single isolated act than is trade. A war can even be fought “to end all wars,” to have hegemony in peace forever after. Trade is typically an indefinite series of recurrent acts, which the participants fully intend to prolong. Everything that mathematics and psychology finds conducive to cooperative solutions in “iterated” prisoners’ dilemmas applies to trade, much less of it to war. Neither dilemma and its real-world resolution, however, lends convincing support to the Hobbesian reason for inventing a state and escaping from the brutish misery of the state of nature, into its encircling arms.
Is there more force in Rousseau’s thesis, that people in the state of nature are unable to organize the social cooperation necessary for the realization of their common good (the general will)? His basic statement of the problem is in the Second Discourse, and is known as the parable of the Hunting Party.18 If (two) hunters stalk a deer, they are sure to catch it if only each one will stand faithfully at his post. They can in this way unconsciously acquire the idea of mutual obligation (which, for Rousseau, forms the passage from the state of nature to civil society), but only if their present and palpable interest demands it. However, they lack foresight and “hardly think of the morrow.” Therefore, if one sees a hare passing, he will quit the deer stalk and run off to catch it, depriving the other hunter of the deer and, indeed, of bagging anything at all. The pay-off matrix of their interaction will have the form of figure 3.19
As both hunters prefer the deer, or even half of one, to a hare, neither has an incentive to “sucker” the other, leaving him standing while he runs off after the hare. Neither would, therefore, rationally opt for a “maximin” strategy (go for the hare in the southeast corner). The deer hunt, then, is critically different from the genuine, Hobbesian prisoners’ dilemma. Social cooperation is not a dilemma and does not for that reason require coercion. A problem (but not a dilemma) is only created for the hunting party by the myopia of one of the hunters who cannot see that a sure deer at the end of the hunt is better than a sure hare. (If both hunters suffered from such complete lack of forethought, they might “objectively” have a prisoners’ dilemma without feeling it. Neither would worry about the end-result of the party; they would not perceive the missed deer, let alone invent an arrangement, such as the social contract creating a state, enabling them to catch the deer rather than the hare, which is the only reason they would have for not letting the hunt take its course, with both hunters running off after the game, if any, they happen to see.)
Supposing, then, that at least the second hunter is alive to the advantage of getting the first hunter to keep his place, what solutions are available for overcoming the latter’s myopia or fecklessness? The contractarian solution is to get him to become a party to the social contract, voluntarily submitting to coercion when needed. But it is difficult to see why he would see the advantage of the social contract if he does not see that of standing fast.20 He is either shortsighted and sees neither, or he is not and the hunters don’t need the social contract.
A more promising line of thought is to suppose that the hunters have hunted before and, as by happy chance no hare crossed their paths, they did catch the deer. The second hunter (the far-sighted one) has saved up a quarter. Next time out he dangled it before the myopic eyes of the first hunter, keeping him at his task and letting him have it at the end of the day while he kept the whole new deer they successfully caught together. (He has, of course, not forgotten once more to set aside a quarter, to maintain the “wage fund.”) This, in a slightly abridged version, is the story of abstinence, capital accumulation, natural selection, the differential contributions and rewards of entrepreneurial initiative and wage labour, and in fact the organization of social cooperation and the determination of the terms on which the participants are willing to carry it on. (In “How Justice Overrides Contracts” [pp. 160-173], we will meet the claim that willing social cooperation is not a matter of the terms the participants agree, but of the terms being reasonable. If the terms that have proved capable of bringing about social cooperation need not, for that reason alone, be considered reasonable, difficulties arise about the very meaning of social cooperation. What, then, is cooperation on unreasonable terms?)
The story, however, does not naturally lend itself to the sort of happy ending which we have learnt to associate with the exit from the state of nature. It does not explain why rational persons, living in a state of nature, should have a preference for the state and seek to invent one (and it is silent on the civic preferences of persons who have been educated in and by the state and have never had occasion to try the state of nature).
Persons are in states, have been there for many generations, and would have no practical means of getting out if they wanted to. States are in the state of nature; many of them have known something approaching the security offered by the super-state when they were part of the Roman Empire, or a British colony; and if they wanted to surrender their sovereignty to a super-state, there are at least some practical steps they could take to try and organize one. They do nothing of the sort. They are quite content to listen to their own voice at the United Nations, leaving it the fatuous irrelevance that it is. Is it, then, beyond reasonable doubt that persons would rush and negotiate a social contract if, like states, they had the option not to do so?
States have known both peace and war throughout history. Some states have died as such because of war, though more states have been born. Most, however, have survived more than one war and continue to muddle through, without finding existence so “nasty and brutish” as to make life within a world state look enticing. Even the very particular prisoners’ dilemma in which two nuclear superpowers are exposed to the threat of destruction and to the expense of maintaining a counter-threat, has not so far induced them to seek shelter and assured self-preservation in a Soviet-American contract.
On a less apocalyptic level, “beggar-thy-neighbour” policies in international trade seem to be a perfectly good practical illustration of the prisoners’ dilemma as applied to states. Generally speaking, all states could be better off if, by cooperative conduct, they allowed the potential gains from trade to be fully realized, just as all prisoners would be better off if none betrayed the other by confessing. The “dominant strategy” of each state (as the “optimum tariff” argument demonstrates), however, is to engage in discriminatory trade practices, high tariffs, competitive devaluations and so forth. This strategy is “dominant” on the argument that if other states behave nicely and adopt free-trader conduct, the first state will reap advantages from its misbehaviour, while if other states misbehave, it would suffer by not also misbehaving. The supposed outcome of every state adopting its dominant strategy is an escalating trade war with everybody rapidly getting poorer and being unable to do anything about it in the absence of a super-state with powers of coercion. In actual fact, many states much of the time behave reasonably well in international trade. They either do not have a dominant strategy or, if they do, it is not to misbehave. Most states most of the time adhere to GATT rules (which stand for the “cooperative solution” in game-theory parlance). Trade wars are generally minor skirmishes, limited to a few products of a few states and instead of escalating as they should, they usually subside. Such “partial free trade” is achieved, just like “partial peace,” without benefit of a state above states and the transfer of power to it. Complete free trade, like total peace, may from most points of view be more satisfactory, but the cost of the added satisfaction must appear prohibitive to the participants; states do not willingly submit to domination even if the dominant entity is to be called the Democratic Federation of Independent Peoples.
People, in the sense of natural persons, however, are supposed by contractarian theory to submit willingly. Unlike states in international relations, people as persons have no opportunity to contradict this supposition. For centuries, since Hobbes if not before, political theory has been assuming that people did not, in fact, very much mind the potential threat of being coerced, being too frightened of the hurt they might suffer in un-coerced “chaos” (this is the Hobbesian version of the social contract), or too interested in the beneficial results of coercion (which is the broader basis of the social contract, laid by Rousseau).21 I believe this is how one should read the cryptic and profound observation of Leo Strauss (few others have thought more powerfully and deeply about these matters), that Hobbes “created” political hedonism, which transformed life “on a scale never yet approached by any other teaching.”22 It is a not very important detail that instead of pleasure (as hedonists are supposed to seek), Hobbes spoke of self-preservation as the end which explains action.23 Since Hobbes, it is tacitly treated as a self-evident truth that people need, or want to have, the state because their hedonistic pain-and-pleasure calculus is ipso facto favourable to it.
Recent research into the prisoners’ dilemma, both deductive into its logical structure and experimental into actual behaviour in such situations, has established that acceptance of coercion by the participants is not a necessary condition for their finding a “cooperative solution.”24 Some of the crucial steps toward getting this result are: (a) to admit that the dilemma can be confronted more than once (it can be an iterative or sequential “game”), so that reliance on single-stage rationality may not correctly predict the moves of rational players; (b) to make a player’s move depend in part on the other player’s move in a previous stage of the sequential game, or in some other game altogether (i.e. to make it depend on experience), either player taking account of the reputation established by the other for toughness or softness; (c) to make him play as he ought to if the other player were playing tit-for-tat; (d) to introduce some regard for the relative value of present and future; and (e) to let the increased pay-off of a cooperatively solved game teach people to go for the cooperative solution in subsequent games. It is intuitively plausible that in a state of nature where people do not instantly club each other to death in a single-stage non-cooperative performance of the dilemma game, but where they survive for some time and have both occasion and incentive to assess and heed each other’s capacity for retaliation, vengefulness, mutual protection, gratitude, fair play, etc. the prisoners’ dilemma becomes both very much more complicated and loses much of its inexorability.
Nor need one limit the application of this result to the sole bellum omnium contra omnes. Hobbes makes people choose Leviathan to produce order out of purported chaos. But people need not have chosen Leviathan, since some kind of cooperative solution, some kind of order emerges in the state of nature, too, though it may not be the same kind of order as that produced by the state. Both qualitative and quantitative differences are possible, indeed extremely likely, though it is very hard to form sensible hypotheses about what the voluntary solution would exactly be like. Whether the voluntary product, in turn, is inferior or superior to the state product, will have to remain a matter of taste. The important thing is not to confuse the question of how we like either product, with the far more vital question of how we like the entire society in which order is state-produced, compared to the entire society (the state of nature) in which it is a voluntary arrangement.
What goes for order goes, by simple extension of the argument, for other public goods as well, whose production was supposed to have been altogether prevented by a rigidly interpreted prisoners’ dilemma and the related, rather looser free-rider problem.25 Once a public good, say clean air, paved streets or national defence, gets produced, people cannot be excluded from enjoying it regardless of whether they have paid their share of the cost of producing it. (We shall have occasion in chapter 4, pp. 234-6, to question what “their share,” in the sense of the part of the cost that a particular person ought to bear, can possibly mean.) Therefore, many will choose not to bear “their share” and the public good will not get produced or maintained, unless the state steps in to coerce all would-be free riders to pay, at the one hit both overcoming “isolation” by making each individual act as he would if all had one common will, and providing “assurance” to each individual who pays that he is not a lone sucker, for all others pay too.26 If the general dilemma is conceived of as a sequential game, a society’s perpetual learning process, it seems obvious that it can have a solution for each intermediate stage, and arbitrary to rule out the likelihood of at least some of the solutions being cooperative, so that as a general proposition, at least some quantities of some public goods will get produced on a voluntary basis.
“Some quantities” of “some public goods” as a result of non-coercive spontaneous solutions, sounds insufficiently affirmative. The reflex reaction of capitalism’s adversary may well be that, because of external economies and diseconomies, only an all-embracing compulsory arrangement, i.e. a state, can ensure that the right amount of public goods gets made. In this view, the prisoners’ dilemma would represent one limiting case, that of total failure to “internalize,” and the state would be the other limiting case in which the entire benefit of an external economy gets internalized from the state’s aggregative point of view. The in-between case of the voluntary association, the spontaneously formed interest group, would stop short of complete internalization and as a consequence would typically tend to fall between the two stools of the unresolved prisoners’ dilemma and state provision of the right amount of the good. Nor is it, of course, always true of any and all levels of output that if the state has in fact chosen that level, it considered it (given all constraints, scarcities and competing claims) the “right” one. If the claim that the output of a public good chosen by the state is the right output, is to be more than a tautological statement of the state’s “revealed preference,” it must somehow be related to some independently derived standard of the optimum.
In the case of individually consumed goods, this standard is, by and large, the Pareto-optimum reached by equating the marginal rates of substitution and transformation between any two goods. But as it is nonsense to speak of a marginal rate of substitution between a public and a private good (a person cannot decide to give up a dollar’s worth of chocolate to get a dollar’s worth of clean air, law and order or paved street), this standard does not work. When the post-1981 Polish state imports one more water-cannon and reduces chocolate imports by the corresponding sum, the decision can hardly be related to the marginal Polish chocolate-eater’s relative liking for law and order and chocolate. If it expresses anything, the decision must express the state’s balancing of the real interests of society that it considers important, in proportion to the importance it attaches to each. The individual chocolate-eater is obviously unable to attach the proper weights to the interest of the vanguard of the working class, of the Organs, of proletarian internationalism, etc. How much tax to surrender to the state so it may buy law and order or clean air for the use of the individual taxpayer in question is not a matter of any taxpayer’s choice. The state cannot buy a collective good for him.
A standard which will do for “collective choice” (if we must resort, for the sake of argument, to this question-begging concept) what Pareto-efficiency does for individual ones, can always be contrived by supposing either (a) that society has but one will (e.g. a will manifested by unanimity, or possibly the general will), or (b) that the several more-or-less divergent wills (including, arguably, the will of the state itself) which are present in society can, by a system of weights attached to each, be expressed as one will (what Robert Paul Wolff disdainfully calls “vector-sum democracy”).27
Whoever fixes the relative weights to be attached (i.e. makes the interpersonal comparisons, or reads the general will, or whichever way the reader prefers to phrase it), fixes the “right” output of public goods with respect to the standard he has thus set up for himself. Whatever he decides, he will, therefore, always be in the proud position of having fixed the right output; for there can never be independent proof to the contrary. It is a redundant apology for the state to say “by reading the general will,” “by balancing the merits of conflicting claims,” “by duly considering public need against the background of its disinflationary policy,” etc. it has determined the right output of public goods. For, whatever the output it chose on whichever considerations, it would not have been, according to its own lights, the wrong one, and no one can ever say that somebody else’s lights would have led it to a more nearly “correct” determination.
It remains to add that the political hedonist who is content to sign the social contract must somehow or other have convinced himself that he is getting a good deal. The incremental pleasure he expects to derive from having the state arrange the production of the correct amount of order and other public goods, instead of relying on a possibly quite inadequate patchwork of spontaneous arrangements, must outweigh the pain of coercion he thinks he will suffer at the state’s hands.
The obvious case where this must hold true is when he does not expect to suffer at all. He will, as a matter of fact, never be coerced if he wills what the state wills, or vice versa, if he can rely on the state to will only what he wills. He must either be the perfect conformist, or he must believe in a benign state which has the power of coercion but lets itself be controlled by those who have none.
[15. ]I submit that “prisoners’ ”; is preferable to the more usual “prisoner’s,” for the dilemma is always that of two or more persons and its essence is the fatality of mutual betrayal. It cannot ever be a game of solitaire.
[16. ]Hobbes’s dilemma is more natural and less rigorous than the one set up under the conventions of formal game theory and it should, much of the time, have a cooperative solution. In a formal game, the player must make his move all the way, he is not allowed pauses, feints or tentative half-moves whose second half depends on the equally tentative reactions, tâtonnements of the other player. In the state of nature a player, before even making a half-move, may make speeches, brandish his weapon, cajole, etc. Depending on the other player’s reaction or rather on his reading of it, he may walk away (if the other stands his ground), or strike a blow (either because the other looks about to strike first, or because he is looking the other way), or perhaps hear and consider an offer of Danegeld.
[17. ]In his remarkable Anarchy and Cooperation, Taylor rightly raises an eyebrow at Hobbes not applying to a state of nature composed of states, the analysis he applies to a state of nature composed of persons. This reproach looks particularly grave from an empiricist’s point of view: for a state of nature composed of states is available in the real world, while a state of nature composed of persons is a theoretical construct, or at least it was that for Hobbes and his readers who had no inkling of what modern anthropologists were going to find in remote corners of the world.
[18. ]J-J. Rousseau, Discours sur l’origine de l’inégalité parmi les hommes, 1755.
[19. ]I am borrowing this formulation from Raymond Boudon and Franç ois Bourricaud, Dictionnaire critique de la sociologie, 1982, p. 477. In attributing the crucial role in the creation of the problem to myopia, I have been preceded by Kenneth M. Waltz, Man, the State and War, 1965, esp. p. 168. Myopia can make the deer worth less than the hare because it is further away; the second hunter’s awareness of the first hunter’s myopia can induce the former to run off after a hare even though it is the latter who is too shortsighted to see the deer!
[20. ]To be led, by a scrutiny of the core structure of mutually advantageous cooperation, to the conclusion that Rousseau’s social contract has an insufficient basis in rational self-interest, is certainly unexpected. The theory of the social contract has always served as the rational foundation for the state, making mystical-historical foundations both of the pre-Reformation and the romantic-Hegelian type redundant. Part II of Ernst Cassirer’s The Myth of the State (1946) is entitled “The Struggle against Myth in the History of Political Theory,” and deals with the Stoic heritage in political philosophy, culminating in contractarian theory. In it, he writes: “if we reduce the legal and social order to free individual acts, to a voluntary contractual submission of the governed, all mystery is gone. There is nothing less mysterious than a contract.”
[21. ]“If, in a group of people, some people act so as to harm my interest, I may readily submit to coercion if this is the pre-condition of subjecting them to coercion” (W. J. Baumol, Welfare Economics and the Theory of the State, 2nd edn, 1965, p. 182). This statement is presented as enabling the state’s generally recognized functions to be logically derived from what its subjects want. It is not explained why the fact that some people act to harm my interest is sufficient to persuade me to submit to coercion (in order to submit them to it, too), regardless of the sort of harm they are doing to my interest, its gravity, eventual possibilities of a non-coercive defence, and regardless also of the gravity of the coercion I am submitting to, and all its consequences. Yet it is not hard to interpret history in a way which should make me prefer the harm people do to my interest, to the harm people organized into a state and capable of coercing me, can do to my interest.
[22. ]Leo Strauss, Natural Right and History, 1953, p. 169.
[23. ]Ibid., p. 169, note 5.
[24. ]Taylor, Anarchy and Cooperation ch. 3; David M. Kreps, Paul Milgrom, John Roberts and Robert Wilson, “Rational Cooperation in the Finitely Repeated Prisoners’ Dilemma,” Journal of Economic Theory, 27, 1982; J. Smale, “The Prisoner’s Dilemma and Dynamical Systems Associated to Non-Cooperative Games,” Econometrica, 48, 1980. For a broader review of the problem, cf. Anatol Rapoport, “Prisoners’ Dilemma—Recollections and Observations,” in Anatol Rapoport (ed.), Game Theory as a Theory of Conflict Resolution, 1974, pp. 17-34. The important point seems to be that the players must not be stupid and totally without foresight. Fairly alert, worldly-wise players will generally cooperate in iterated prisoners’ dilemmas. Cf. also Russell Hardin, Collective Action, 1982, p. 146.
[25. ]Prisoners’ dilemma and free riding are not just different words describing the same structure of interaction. The former imposes on each rational prisoner one dominant strategy, i.e. to confess before the other can betray him. This alone secures the least bad of two alternative worst-case outcomes (maximin). The free-rider problem imposes no dominant strategy, maximin or other. It is not inherently inconsistent with a cooperative solution. Where would the free rider ride free if there were no cooperative transport service?
[26. ]The allusions are to A. K. Sen’s “Isolation, Assurance and the Social Rate of Discount,” Quarterly Journal of Economics, 81, 1967.
[27. ]Robert Paul Wolff, Barrington Moore Jr and Herbert Marcuse, A Critique of Pure Tolerance, 1965.