Anarchy, State and Utopia Part 3

If each individual acts rationally, unlimited by moral constraints, (B, B' B') will emerge. How will things differ, if at all, with the addition of moral constraints? It might be thought that moral considerations require allowing another to do whatever you do; since the situation is symmetrical some symmetrical solution must be found. To this the fishy reply might be made that (B, B' B') is symmetrical, and hence someone performing a B-ish action recognizes that the other will do likewise. But recognizing that another will do likewise is not the same as allowing allowing him to do this. A person performing a him to do this. A person performing a B B-ish action is trying to impose a (B, C' C') solution. What moral right does he have to impose impose this asymmetry, to this asymmetry, to force force others not to behave as he does? But before accepting this strong counterreply as conclusive, we should ask whether each person faces or views himself as facing a symmetrical situation? Each person knows more about himself than he does about the other; each can be surer of his own intentions not to aggress against the other if he finds himself in the dominant power position, than he can be of the others' similar intentions. (Following Acton, we might wonder whether any of us can be sure, or even reasonably confident.) Given this asymmetry of each knowing more about his own intentions than about those of the other party, others not to behave as he does? But before accepting this strong counterreply as conclusive, we should ask whether each person faces or views himself as facing a symmetrical situation? Each person knows more about himself than he does about the other; each can be surer of his own intentions not to aggress against the other if he finds himself in the dominant power position, than he can be of the others' similar intentions. (Following Acton, we might wonder whether any of us can be sure, or even reasonably confident.) Given this asymmetry of each knowing more about his own intentions than about those of the other party,4 isn't it reasonable for each to pursue the isn't it reasonable for each to pursue the B B-ish action? Rather, since it's individually rational, does this asymmetry serve to rebut the argument from symmetry for the (A, A' A') solution and against the (B, B' B') solution? Clearly, things become very messy.

Rather than focusing on the total situation, it would be more promising to ask whether something special about the B B-ish actions excludes them as morally permissible. Does some moral prohibition rule out B? B? If so, we must distinguish the If so, we must distinguish the B B actions from those other prohibitings of actions on the grounds of the risk they present, which we have already held to be legitimate. What distinguishes prohibiting others from joining another protective agency, or forcibly acting to prevent another agency from getting more powerful than your own or yourself from an agency's forbidding others to punish its clients except by a reliable procedure (and punishing those who disobey this prohibition even should it turn out that the clients did wrong these others and were not innocent)? Let us first consider cases which commonly actions from those other prohibitings of actions on the grounds of the risk they present, which we have already held to be legitimate. What distinguishes prohibiting others from joining another protective agency, or forcibly acting to prevent another agency from getting more powerful than your own or yourself from an agency's forbidding others to punish its clients except by a reliable procedure (and punishing those who disobey this prohibition even should it turn out that the clients did wrong these others and were not innocent)? Let us first consider cases which commonly are are distinguished. distinguished.

PREEMPTIVE ATTACK.

According to usual doctrine, under some circ.u.mstances a country X X may launch a preemptive attack, or a preventive war, upon another country may launch a preemptive attack, or a preventive war, upon another country Y; Y; for example, if for example, if Y Y is itself about to launch an immediate attack upon is itself about to launch an immediate attack upon X X, or if Y Y has announced that it will do so upon reaching a certain level of military readiness, which it expects to do some time soon. Yet it is not accepted doctrine that one nation has announced that it will do so upon reaching a certain level of military readiness, which it expects to do some time soon. Yet it is not accepted doctrine that one nation X X may launch a war against another nation may launch a war against another nation Y Y because because Y Y is getting stronger, and (such is the behavior of nations) might well attack is getting stronger, and (such is the behavior of nations) might well attack X X when it gets stronger still. Self-defense plausibly covers the first sort of situation but not the second. Why? when it gets stronger still. Self-defense plausibly covers the first sort of situation but not the second. Why?

It might be thought that the difference is merely a matter of greater or lesser probability. When a nation is about to launch an attack, or has announced that it will when and if it reaches a certain level of readiness, the probability is very high that it will attack. Whereas the probability is not as great that any nation getting stronger will attack when it attains greater strength. But the distinction between the cases does not depend upon such probability considerations. For however low the probability, estimated by the "experts" of neutral countries, of Y Y's launching an attack on X X (in the second case) within the next ten years (0.5, 0.2, 0.05), we can imagine alternatively that Y now is about to wield a super-device fresh out of its scientific laboratories that, with (in the second case) within the next ten years (0.5, 0.2, 0.05), we can imagine alternatively that Y now is about to wield a super-device fresh out of its scientific laboratories that, with that that probability, will conquer probability, will conquer X; X; while with one minus that probability, it will do nothing. (Perhaps this probability is the probability of the device's working, or perhaps the device itself is probabilistic.) The device is set to be wielded within one week; while with one minus that probability, it will do nothing. (Perhaps this probability is the probability of the device's working, or perhaps the device itself is probabilistic.) The device is set to be wielded within one week; Y Y is committed to use it, the timetable is being followed and a countdown has begun. Here is committed to use it, the timetable is being followed and a countdown has begun. Here X X, in self-defense, may attack, or issue an ultimatum that if the device is not dismantled within two days it will attack, and so on. (And what if, though the timetable doesn't call for it, the device can can be used the next day or immediately?) If be used the next day or immediately?) If Y Y were spinning a roulette wheel and with probability 0.025 the damage of war would be inflicted on were spinning a roulette wheel and with probability 0.025 the damage of war would be inflicted on X X, X X could act in self-defense. But, in the second case even when the probability is equal, could act in self-defense. But, in the second case even when the probability is equal, X X may not so act against may not so act against Y Y's arming. Therefore, the issue is not merely a matter of how high the probability is. Upon what, then, if not the magnitude of the probability, does the distinction between the first type of case and the second type rest?

The distinction depends on how the harm, if it eventuates, is related to what Y Y already has done. For some actions that yield various outcomes with various probabilities, nothing more need be done by the agent (after the action is performed) to produce an outcome which, when it eventuates, is something he did or brought about or caused to happen, and so on. (In some cases, further actions of already has done. For some actions that yield various outcomes with various probabilities, nothing more need be done by the agent (after the action is performed) to produce an outcome which, when it eventuates, is something he did or brought about or caused to happen, and so on. (In some cases, further actions of others others might be needed, for example, soldiers obeying a commander's orders.) If such an action yields a high enough probability of a dangerous "border crossing," another may prohibit it. On the other hand, some processes might lead to certain possible consequences, but only if further decisions are made by the people engaging in them. Processes might, as in the cases we are considering, place people in a better position to do something, and so make it more likely that they will decide to do it. These processes involve further significant decisions by the persons and the border crossings depend upon might be needed, for example, soldiers obeying a commander's orders.) If such an action yields a high enough probability of a dangerous "border crossing," another may prohibit it. On the other hand, some processes might lead to certain possible consequences, but only if further decisions are made by the people engaging in them. Processes might, as in the cases we are considering, place people in a better position to do something, and so make it more likely that they will decide to do it. These processes involve further significant decisions by the persons and the border crossings depend upon these these decisions (made more likely by the process). It is permissible to prohibit the former actions where the person need do nothing more, but not to prohibit the latter processes. decisions (made more likely by the process). It is permissible to prohibit the former actions where the person need do nothing more, but not to prohibit the latter processes.af Why? Why?

Perhaps the principle is something like this: an act is not wrong and so cannot be prohibited if it is harmless without a further major decision to commit wrong (that is, if it would not be wrong if the agent was fixed unalterably against the further wrong decision); it can only be prohibited when it is a planned prelude to the further wrong action. So stated, the principle would protect actions that merely facilitate others' wrongdoing if the acts are harmless in themselves-for example, publishing the plans of the alarm systems of banks. The act would be tolerated were it known that others would not decide to do wrong. Among such actions, the clearest candidates for prohibition are those which, it is thought, could be done for no reason other than to facilitate wrongdoing. (Even here, can't one always imagine an eccentric with legitimate though odd reasons?) We may avoid this question of whether such actions so clearly intended only to aid the wrongdoing of others may be prohibited. All the actions we are concerned with could be done for perfectly legitimate and respectable reasons (for example, self-defense), and they require further decision to commit wrong by the agent himself, if wrong is to occur.

A stringent principle would hold that one may prohibit only the last wrong decision necessary to produce the wrong. (Or, the last act necessary to an alternative in a set, any one of which is necessary.) More stringent yet would be a principle holding that one may prohibit only the pa.s.sing of the last clear point at which the last wrong decision necessary to the wrong can be reversed. More lat.i.tude is given to prohibition by the following principle (hence it is a weaker principle against prohibition): Prohibit only only wrong decisions and actions on them (or dangerous actions requiring no further wrong decisions). One may wrong decisions and actions on them (or dangerous actions requiring no further wrong decisions). One may not not prohibit actions which are not based on decisions that are wrong, merely on the grounds that they facilitate or make more likely the agent himself later making wrong decisions and doing the wrong actions which follow from them. Since even this weaker principle is sufficient to prohibit actions which are not based on decisions that are wrong, merely on the grounds that they facilitate or make more likely the agent himself later making wrong decisions and doing the wrong actions which follow from them. Since even this weaker principle is sufficient to exclude exclude prohibiting others from strengthening their protective agency or joining another one, we need not decide here which principle is most appropriate. (The two stronger principles, of course, also would exclude such prohibitions.) prohibiting others from strengthening their protective agency or joining another one, we need not decide here which principle is most appropriate. (The two stronger principles, of course, also would exclude such prohibitions.) It might be objected that the principles adumbrated should not be applied to hold impermissible some group A' A's forcibly intervening in the process of B B's strengthening their protective agency. For that process is a special one; if it is successful, A A will be in a far weaker position, if not unable, to enforce the prohibition on wrong when finally will be in a far weaker position, if not unable, to enforce the prohibition on wrong when finally A A is ent.i.tled to do so. How can is ent.i.tled to do so. How can A A be asked to refrain from prohibiting the earlier stages when it knows that any wrongs will be done later when it is unable to oppose them as effectively? But if the early stages of be asked to refrain from prohibiting the earlier stages when it knows that any wrongs will be done later when it is unable to oppose them as effectively? But if the early stages of B B's process involve no commitment to any later wrong, and if B B has good (nonaggressive) reasons for its actions, then it is not absurd to hold that others may not interfere with the earlier and in themselves (supposing certain continuations) harmless stages, even though this abstention will put them in a less strong position later. has good (nonaggressive) reasons for its actions, then it is not absurd to hold that others may not interfere with the earlier and in themselves (supposing certain continuations) harmless stages, even though this abstention will put them in a less strong position later.5 We have found a distinction, which appears theoretically significant, that distinguishes a protective agency's forbidding others from using unreliable or unfair procedures to exact justice on its clients from other prohibitions-such as forbidding others to form another protective agency-which might be thought to be allowable if the first is. For our purposes in this essay we need not provide the theory which underlies this distinction and explains its significance, even though investigating these issues promises to lead very quickly to fundamental questions. It is enough to have reb.u.t.ted the charge we imagined earlier that our argument fails because it "proves" too much, in that it provides a rationale not only for the permissible rise of a dominant protective a.s.sociation, but also for this a.s.sociation's forcing someone not to take his patronage elsewhere or for some person's forcing others not to join any a.s.sociation. Our argument provides no rationale for the latter actions and cannot be used to defend them.

We have put forth a principle which excludes prohibiting actions not wrong in themselves, actions that merely facilitate or make more likely the commission of other wrongs dependent upon other wrong decisions the agent has not made (yet). (This statement is intentionally ambiguous so as to encompa.s.s the strong and the weak principles.) This principle does not not claim that no one may be held responsible or be punished for attempting to get others to do wrong because to succeed the attempt requires the decision of claim that no one may be held responsible or be punished for attempting to get others to do wrong because to succeed the attempt requires the decision of others others to do wrong. For the principle focuses on whether the thrust toward wrong already has been made and is now out of to do wrong. For the principle focuses on whether the thrust toward wrong already has been made and is now out of that person's that person's hands. It is hands. It is a further a further question whether and to what extent any decisions of others can eliminate his responsibility for the result of his original attempt. Prime candidates for responsibility continuing are attempts to get others to do some wrong, which attempt succeeds (not by accident and in the manner intended, and so forth) in getting them to decide and act wrongly. (In this case, isn't the original act wrong itself, and so question whether and to what extent any decisions of others can eliminate his responsibility for the result of his original attempt. Prime candidates for responsibility continuing are attempts to get others to do some wrong, which attempt succeeds (not by accident and in the manner intended, and so forth) in getting them to decide and act wrongly. (In this case, isn't the original act wrong itself, and so not not protected from prohibition under the conditions of the principle?) protected from prohibition under the conditions of the principle?) The contrasting view holds that the further decisions of others eliminate the responsibility of someone who succeeds in his attempt to get them to act in a certain way; though he persuades them or convinces them or whips them up to do it, they could have chosen to refrain. The following model might underlie this view. For each act, so the model runs, there is a fixed amount of responsibility; this might be measured by how much punishment there is to be for the act. Someone persuaded by another to do something may be punished fully for his action; he may be punished as much as someone who decides all by himself to do the same action. Since all of the punishment for that action is used up, so is all of the responsibility for it; there is no more responsibility or punishment for that action left over to place on another person. So, the argument concludes, a person who persuades another to decide to do something cannot be held responsible for or at all punished for the consequences of the other's action. But this model of a fixed amount of responsibility for an act is mistaken. If two persons each cooperate in murdering or a.s.saulting a third, then each a.s.saulter or murderer may be punished fully. Each may receive the same punishment as someone acting alone, n n years say. They need not each be given years say. They need not each be given n n/2. Responsibility is not a bucket in which less remains when some is apportioned out; there is not a fixed amount of punishment or responsibility which one uses up so that none is left over for the other. Since this model or picture of how responsibility operates is mistaken, a major prop is removed from the view that no one may be punished for persuading another responsible individual to do something.6

BEHAVIOR IN THE PROCESS.

We have argued that even someone who foresees that a protective a.s.sociation will become dominant may not forbid others to join up. But though no one may be forbidden to join up, might not everyone choose choose to stay out, in order to avoid the state at the end of the process? Might not a population of anarchists realize how individual efforts at hiring protection will lead, by an invisible-hand process, to a state, and because they have historical evidence and theoretical grounds for the worry that the state is a Frankenstein monster that will run amuck and will not stay limited to minimal functions, might not they each prudentially choose not to begin along that path? to stay out, in order to avoid the state at the end of the process? Might not a population of anarchists realize how individual efforts at hiring protection will lead, by an invisible-hand process, to a state, and because they have historical evidence and theoretical grounds for the worry that the state is a Frankenstein monster that will run amuck and will not stay limited to minimal functions, might not they each prudentially choose not to begin along that path? 7 7 If told to anarchists, is the invisible-hand account of how the state arises a self-defeating prophecy? If told to anarchists, is the invisible-hand account of how the state arises a self-defeating prophecy?

It will be difficult for such concerted effort to succeed in blocking the formation of the state, since each individual will realize that it is in his own individual interests to join a protective a.s.sociation (the more so as some others join), and his joining or not will not make the difference as to whether or not the state develops. (The B B actions of the earlier matrices are dominant.) However, it must be admitted that other individuals with special motivations would not behave as we have described: for example, people whose religion prohibits purchasing protection or joining with others in protective ventures; or misanthropes who refuse to cooperate with or hire any other persons; or personal pacifists who refuse to support or partic.i.p.ate in any inst.i.tution that uses force, even for their own self-defense. We must restrict our claim that a state would arise from a state of nature, so as to exclude these special psychologies which thwart the operation of the invisible-hand process we have described. For each special psychology, we may insert a specific clause in the claim to exclude it. Thus: in a territory containing rational individuals who also are willing to use force in self-defense and are willing to cooperate with others and to hire them, ... actions of the earlier matrices are dominant.) However, it must be admitted that other individuals with special motivations would not behave as we have described: for example, people whose religion prohibits purchasing protection or joining with others in protective ventures; or misanthropes who refuse to cooperate with or hire any other persons; or personal pacifists who refuse to support or partic.i.p.ate in any inst.i.tution that uses force, even for their own self-defense. We must restrict our claim that a state would arise from a state of nature, so as to exclude these special psychologies which thwart the operation of the invisible-hand process we have described. For each special psychology, we may insert a specific clause in the claim to exclude it. Thus: in a territory containing rational individuals who also are willing to use force in self-defense and are willing to cooperate with others and to hire them, ...

At the close of Chapter 5, we argued that a territory with a dominant protective agency contains a state. Would Locke agree that in such a territory there was a state or civil society? If so, would he say it had been created by a social compact? Clients of the same protective agency are in a state of civil society with respect to each other; clients and independents have exactly the same rights vis-a-vis each other as any two persons in a state of nature, and hence are in a state of nature with respect to each other (Two Treatises of Government, II, sect. 87). But does the fact that the independents yield before the superior power of the dominant protective agency and II, sect. 87). But does the fact that the independents yield before the superior power of the dominant protective agency and don't don't act as executioners of the law of nature against its clients (despite having a right to) mean that they are not in a Lockean state of nature with respect to the clients? Should one say they are in a act as executioners of the law of nature against its clients (despite having a right to) mean that they are not in a Lockean state of nature with respect to the clients? Should one say they are in a de jure de jure state of nature but not a state of nature but not a de facto de facto one? Would Locke use some notion of political or civil society under which there could be a civil society in an area even if not one? Would Locke use some notion of political or civil society under which there could be a civil society in an area even if not every every two people in that area stood in a civil-society relationship with respect to each other? One also would want this notion to be of political interest; if merely two of the many individuals in an area stand in a civil-society relationship with respect to each other, this should be insufficient for there to be civil society in that area. two people in that area stood in a civil-society relationship with respect to each other? One also would want this notion to be of political interest; if merely two of the many individuals in an area stand in a civil-society relationship with respect to each other, this should be insufficient for there to be civil society in that area.8 We have described a process whereby individuals in an area separately sign up for personal protection with different business enterprises which provide protective services, all but one of the agencies being extinguished or all coming to some modus vivendi, modus vivendi, and so on. To what degree, if any, does this process fit what Locke envisioned as individuals "agreeing with other men to join and unite into a community," consenting "to make one community or government" (sect. 95), compacting to make up a commonwealth (sect. 99)? The process looks nothing like unanimous joint agreement to create a government or state. No one, as they buy protective services from their local protective agency, has in mind anything so grand. But perhaps joint agreement where each has in mind that the others will agree and each intends to bring about the end result of this is not necessary for a Lockean compact. and so on. To what degree, if any, does this process fit what Locke envisioned as individuals "agreeing with other men to join and unite into a community," consenting "to make one community or government" (sect. 95), compacting to make up a commonwealth (sect. 99)? The process looks nothing like unanimous joint agreement to create a government or state. No one, as they buy protective services from their local protective agency, has in mind anything so grand. But perhaps joint agreement where each has in mind that the others will agree and each intends to bring about the end result of this is not necessary for a Lockean compact.9 I myself see little point to stretching the notion of "compact" so that each pattern or state of affairs that arises from the disparate voluntary actions of separately acting individuals is viewed as arising from a I myself see little point to stretching the notion of "compact" so that each pattern or state of affairs that arises from the disparate voluntary actions of separately acting individuals is viewed as arising from a social compact, social compact, even though no one had the pattern in mind or was acting to achieve it. Or, if the notion is so stretched, this should be made clear so that others are not misled as to its import. It should be made clear that the notion is such that each of the following arises from a social compact: the total state of affairs const.i.tuted by who is married to, or living with, whom; the distribution on a given evening in a given city of who is in what movie theater, sitting where; the particular traffic pattern on a state's highways on a given day; the set of customers of a given grocery store on a given day and the particular pattern of purchases they make, and so on. Far be it from me to claim that this wider notion is of no interest; that a state can arise by a process that fits this wider notion (without fitting the narrower one) is of very great interest indeed! even though no one had the pattern in mind or was acting to achieve it. Or, if the notion is so stretched, this should be made clear so that others are not misled as to its import. It should be made clear that the notion is such that each of the following arises from a social compact: the total state of affairs const.i.tuted by who is married to, or living with, whom; the distribution on a given evening in a given city of who is in what movie theater, sitting where; the particular traffic pattern on a state's highways on a given day; the set of customers of a given grocery store on a given day and the particular pattern of purchases they make, and so on. Far be it from me to claim that this wider notion is of no interest; that a state can arise by a process that fits this wider notion (without fitting the narrower one) is of very great interest indeed!

The view we present here should not be confused with other views. It differs from social compact views in its invisible-hand structure. It differs from views that "de facto "de facto might makes state might makes state (legal) right" in holding that enforcement rights and rights to oversee this enforcement exist independently and are held by all rather than confined to one or a small group, and that the process of acc.u.mulating sole effective enforcement and overseeing power may take place without anyone's rights being violated; that a state may arise by a process in which no one's rights are violated. Shall we say that a state which has arisen from a state of nature by the process described has replaced the state of nature which therefore no longer exists, or shall we say that it exists within a state of nature and hence is compatible with one? No doubt, the first would better fit the Lockean tradition; but the state arises so gradually and imperceptibly out of Locke's state of nature, without any great or fundamental breach of continuity, that one is tempted tempted to take the second option, disregarding Locke's incredulousness: "... unless any one will say the state of nature and civil society are one and the same thing, which I have never yet found any one so great a patron of anarchy as to affirm" (sect. 94). to take the second option, disregarding Locke's incredulousness: "... unless any one will say the state of nature and civil society are one and the same thing, which I have never yet found any one so great a patron of anarchy as to affirm" (sect. 94).

LEGITIMACY.

Some might deny, perhaps properly, that any normative notion is to be built into an account of the state, even the right to enforce rights and to prohibit dangerous private enforcement of justice provided compensation is made to those prohibited. But since this does not grant to the state or any of its agents any rights not possessed by each and every person, it seems a harmless inclusion. It gives the state no special special rights and certainly does not entail that all acts of rule by the state are presumptively right. Nor does it entail that persons acting as agents of the state possess any special immunity from punishment, if they violate another's rights. The public whose agents they are may provide them with liability insurance, or guarantee to cover their liability. But it may not rights and certainly does not entail that all acts of rule by the state are presumptively right. Nor does it entail that persons acting as agents of the state possess any special immunity from punishment, if they violate another's rights. The public whose agents they are may provide them with liability insurance, or guarantee to cover their liability. But it may not diminish diminish their liability as compared to that of other persons. Also, protective agencies will not have limited liability, nor will any other corporations. Those voluntarily dealing with a corporation (customers, creditors, workers, and others) will do so by contracts explicitly limiting the corporation's liability, if that is the way the corporation chooses to do business. A corporation's liability to those involuntarily intertwined with it will be unlimited, and it presumably will choose to cover this liability with insurance policies. their liability as compared to that of other persons. Also, protective agencies will not have limited liability, nor will any other corporations. Those voluntarily dealing with a corporation (customers, creditors, workers, and others) will do so by contracts explicitly limiting the corporation's liability, if that is the way the corporation chooses to do business. A corporation's liability to those involuntarily intertwined with it will be unlimited, and it presumably will choose to cover this liability with insurance policies.

Does the state we have described have legitimacy, does it legitimately rule? The dominant protective agency has de facto de facto power; it acquired this power and reached its position of dominance without violating anyone's rights; it wields this power as well as anyone would expect. Do these facts add up to its being the legitimate wielder of the power? As "legitimacy" is used in political theory, those legitimately wielding power are ent.i.tled, are power; it acquired this power and reached its position of dominance without violating anyone's rights; it wields this power as well as anyone would expect. Do these facts add up to its being the legitimate wielder of the power? As "legitimacy" is used in political theory, those legitimately wielding power are ent.i.tled, are specially specially ent.i.tled, to wield it. ent.i.tled, to wield it.ag Does the dominant protective agency have any special ent.i.tlement? A dominant agency and another tiny one, or a dominant agency and an unaffiliated individual person, are on a par in the nature of their rights to enforce other rights. How might they have differential ent.i.tlements? Does the dominant protective agency have any special ent.i.tlement? A dominant agency and another tiny one, or a dominant agency and an unaffiliated individual person, are on a par in the nature of their rights to enforce other rights. How might they have differential ent.i.tlements?

Consider whether the dominant protective agency is ent.i.tled to be the one which is dominant. Is a restaurant you choose to go to on a given evening ent.i.tled to your patronage? Perhaps one is tempted to say, in some circ.u.mstances, they merit it or deserve it; they serve better food, less expensively, and in nicer surroundings, and they work long and hard to do so; still, they are not ent.i.tled to your patronage.10 You do not violate any ent.i.tlement of theirs if you choose to go elsewhere. By choosing to go there, though, you do authorize them to serve and bill you. They have no ent.i.tlement You do not violate any ent.i.tlement of theirs if you choose to go elsewhere. By choosing to go there, though, you do authorize them to serve and bill you. They have no ent.i.tlement to be the one to be the one which serves you, but they are ent.i.tled to serve you. Similarly, we must distinguish between an agency's being ent.i.tled to be the one wielding certain power from its being ent.i.tled to wield that power. which serves you, but they are ent.i.tled to serve you. Similarly, we must distinguish between an agency's being ent.i.tled to be the one wielding certain power from its being ent.i.tled to wield that power.11 Is the dominant agency's only ent.i.tlement, then, its being ent.i.tled to wield the power? We can reach questions of ent.i.tlement by another route that illuminates further the situation of persons in a state of nature. Is the dominant agency's only ent.i.tlement, then, its being ent.i.tled to wield the power? We can reach questions of ent.i.tlement by another route that illuminates further the situation of persons in a state of nature.

A protective agency may act against or for a particular person. It acts against him if it enforces someone's rights against him, punishes him, exacts compensation from him, and so forth. It acts for him if it defends him against others, punishes others for violating his rights, forces other to compensate him, and so forth. Theorists of the state of nature hold that there are certain rights residing in the victim of wrong that others may exercise only only if authorized by him; and there are other rights that others may exercise, whether or not the victim authorized them to do so. The right to exact compensation is of the first sort; the right to punish of the second. If the victim chooses not to be compensated, no one else may exact compensation for him or for themselves in his place. But if the victim does wish to be compensated, why may only those whom he has authorized to act for him exact compensation? Clearly, if several different persons each exact full compensation from the offender, this would do him an injustice. How then is it to be determined which person acts? Is the one who may act the one who acts first to exact sufficient compensation for the victim? But allowing many to compete to be the first successfully to exact compensation will embroil prudent wrongdoers and victims alike in many independent time- and energy-consuming hearing processes, only one of which actually will result in a compensation payment. Alternatively, perhaps the person who first begins the attempt to exact compensation preempts the field; no others may also engage in the process. But this would allow the wrongdoer himself to have a confederate be the first to start compensation proceedings (which would be long, complicated, and perhaps inconclusive) in order to stop others from exacting compensation from him. if authorized by him; and there are other rights that others may exercise, whether or not the victim authorized them to do so. The right to exact compensation is of the first sort; the right to punish of the second. If the victim chooses not to be compensated, no one else may exact compensation for him or for themselves in his place. But if the victim does wish to be compensated, why may only those whom he has authorized to act for him exact compensation? Clearly, if several different persons each exact full compensation from the offender, this would do him an injustice. How then is it to be determined which person acts? Is the one who may act the one who acts first to exact sufficient compensation for the victim? But allowing many to compete to be the first successfully to exact compensation will embroil prudent wrongdoers and victims alike in many independent time- and energy-consuming hearing processes, only one of which actually will result in a compensation payment. Alternatively, perhaps the person who first begins the attempt to exact compensation preempts the field; no others may also engage in the process. But this would allow the wrongdoer himself to have a confederate be the first to start compensation proceedings (which would be long, complicated, and perhaps inconclusive) in order to stop others from exacting compensation from him.

In theory, an arbitrary rule could be used to select anyone as the one to exact (or to authorize another to exact) compensation-for example, "the exacter of compensation is to be that person whose name comes immediately after that of the victim in an alphabetical listing of the names of everyone in the territory." (Would this lead to people victimizing their immediate alphabetical predecessors?) That it be the victim who selects the exacter of compensation ensures, at least, that he will be committed to rest content with the upshot of the process and will not continue to attempt to get further compensation. The victim will not believe he selected a process by nature unfair to himself; or if he comes to believe this, he will have only himself to blame. It is to the advantage of the wrongdoer that the victim be involved in, and committed to, the process, for otherwise the victim will initiate a second process to obtain the remainder of what he believes he deserves. The victim can be expected to accede to a restriction against double jeopardy only if the initial process is one he is committed to and has some confidence in, as would not be the case if a confederate of the wrongdoer made the initial judgment. But what is wrong with double jeopardy, given that if its its upshot is unjust the person punished can act himself? And, why cannot a victim place his wrongdoer under double jeopardy, even though the first process was one that he himself had authorized? Cannot the victim say that he had authorized another to exact his just compensation, and that since the agent failed to do this fully, he himself is within his rights to authorize yet another to act? If the first person he sends against a wrongdoer fails to reach him, he may send another; if he reaches him but is bought off, the victim may send another; why may he not send another if his first agent fails to perform his task adequately? To be sure, if he does send another to exact something above and beyond what his first agent attempted to take, he runs the risk that others will think his added exaction unjust and so will oppose him. But are there other than prudential grounds for his not doing so? There is reason against double jeopardy in a civic legal system as it is usually imagined. Since all it takes is one conviction, it is unfair to allow the prosecution to keep trying and trying until it succeeds. This would not apply in the state of nature, where the matter is not settled absolutely and is not binding upon all when the victim's agent or agency reaches a judgment. It is unfair to give the prosecutor in a civic system many chances at a final and binding judgment, for if he is lucky one time there will be little recourse for the person found guilty. However, in a state of nature there is recourse for someone who holds the decision against himself unjust. upshot is unjust the person punished can act himself? And, why cannot a victim place his wrongdoer under double jeopardy, even though the first process was one that he himself had authorized? Cannot the victim say that he had authorized another to exact his just compensation, and that since the agent failed to do this fully, he himself is within his rights to authorize yet another to act? If the first person he sends against a wrongdoer fails to reach him, he may send another; if he reaches him but is bought off, the victim may send another; why may he not send another if his first agent fails to perform his task adequately? To be sure, if he does send another to exact something above and beyond what his first agent attempted to take, he runs the risk that others will think his added exaction unjust and so will oppose him. But are there other than prudential grounds for his not doing so? There is reason against double jeopardy in a civic legal system as it is usually imagined. Since all it takes is one conviction, it is unfair to allow the prosecution to keep trying and trying until it succeeds. This would not apply in the state of nature, where the matter is not settled absolutely and is not binding upon all when the victim's agent or agency reaches a judgment. It is unfair to give the prosecutor in a civic system many chances at a final and binding judgment, for if he is lucky one time there will be little recourse for the person found guilty. However, in a state of nature there is recourse for someone who holds the decision against himself unjust.12 But even though there is no guarantee that a victim will regard his agent's decision as acceptable, it is more likely than his so regarding that of some unknown third party; and so his selecting the exacter of compensation is a step toward ending the affair. (His antagonist also might agree to accept the result.) There is yet another reason, perhaps the major one, for the victim's being the appropriate locus of action to exact compensation. The victim is the one to whom compensation is owed, not only in the sense that the money goes to him, but also in that the other is under an obligation But even though there is no guarantee that a victim will regard his agent's decision as acceptable, it is more likely than his so regarding that of some unknown third party; and so his selecting the exacter of compensation is a step toward ending the affair. (His antagonist also might agree to accept the result.) There is yet another reason, perhaps the major one, for the victim's being the appropriate locus of action to exact compensation. The victim is the one to whom compensation is owed, not only in the sense that the money goes to him, but also in that the other is under an obligation to him to him to pay it. (These are distinct: I may be under an obligation to you to pay another person money, having promised to you that I would pay him.) As the person to whom this enforceable obligation is owed, the victim seems the appropriate party to determine precisely how it is to be enforced. to pay it. (These are distinct: I may be under an obligation to you to pay another person money, having promised to you that I would pay him.) As the person to whom this enforceable obligation is owed, the victim seems the appropriate party to determine precisely how it is to be enforced.

THE RIGHT OF ALL TO PUNISH.

In contrast to exaction of compensation, which it views as something done appropriately only by the victim or his authorized agent, state-of-nature theory usually views punishment as a function that anyone may perform. Locke realizes that this "will seem a very strange doctrine to some men" (sect. 9). He defends it by saying that the law of nature would be in vain if no one in a state of nature had a power to execute it, and since all in the state of nature have equal rights, if any one person may execute it then everyone has that right (sect. 7); he says also that an offender becomes dangerous to mankind in general, and so everyone may punish him (sect. 8), and he challenges the reader to find some other ground for a country's punishing aliens for crimes they commit within it. Is the general right to punish so counterintuitive? If some great wrong were committed in another country which refuses to punish it (perhaps the government is in league with, or is itself, the wrongdoer), wouldn't it be all right for you to punish the wrongdoer, to inflict some harm on him for his act? Furthermore, one might try to derive the right to punish from other moral considerations: from the right to protect, combined with the view that a wrongdoer's moral boundaries change. One might take a contract-like view of moral prohibitions and hold that those who themselves violate another's boundaries forfeit the right to have certain of their own boundaries respected. On this view, one is not morally prohibited from doing certain sorts of things to others who have already violated certain moral prohibitions (and gone unpunished for this). Certain wrongdoing gives others a liberty liberty to cross certain boundaries (an absence of a duty not to do it); the details might be those of some retributive view. to cross certain boundaries (an absence of a duty not to do it); the details might be those of some retributive view.13 Talk of a right to punish may seem strange if we interpret it strongly as a right which others must not happen to interfere with or themselves exercise, rather than as a liberty to do it, which liberty others also may have. The stronger interpretation of right is unnecessary ; the liberty to punish would give Locke much of what he needs, perhaps all if we add the duty of the wrongdoer not to resist his punishment. We may add to these reasons which make more plausible the claim that there is a general right to punish the consideration that, unlike compensation, punishment is not owed to the victim (though he may be the person most greatly interested in its being carried out), and so it is not something he has special authority over. Talk of a right to punish may seem strange if we interpret it strongly as a right which others must not happen to interfere with or themselves exercise, rather than as a liberty to do it, which liberty others also may have. The stronger interpretation of right is unnecessary ; the liberty to punish would give Locke much of what he needs, perhaps all if we add the duty of the wrongdoer not to resist his punishment. We may add to these reasons which make more plausible the claim that there is a general right to punish the consideration that, unlike compensation, punishment is not owed to the victim (though he may be the person most greatly interested in its being carried out), and so it is not something he has special authority over.

How would a system of open punishment operate? All of our previous difficulties in imagining how open exaction of compensation would work apply as well to a system of open punishing. And there are other difficulties. Is it to be a system of the first actor's preempting the field? Will s.a.d.i.s.ts compete to be first to get their licks in? This would greatly magnify the problem of keeping the punishers from exceeding the bounds of the deserved punishment and would be undesirable, the opportunities it offers for cheerful and unalienated labor notwithstanding. In a system of open punishment would anyone be in a position to decide upon mercy; and would another be permitted to negate this decision by punishing additionally so long as the sum did not exceed the amount deserved? Could the offender have a confederate punish him only lightly? Would there be any likelihood that the victim would feel that justice had been done? And so on.

If a system that leaves punishment to whomever happens to do it is defective, how is it to be decided who, among all those willing and perhaps eager, punishes? It might be thought that, as before, it should be the victim or his authorized agent. Yet though the victim occupies the unhappy special position of victim and is owed compensation, he is not owed punishment. (That is "owed" to the person who deserves to be punished.) The offender is not under an obligation to the victim to be punished; he doesn't deserve to be punished "to the victim." So why should the victim have a special right to punish or to be the punisher? If he has no special right to punish, does he have any special right to choose that the punishment not not be carried out at all, or that mercy be granted? May someone punish an offender even against the wishes of the offended party who morally objects to the mode of punishment? If a Gandhian is attacked, may others defend him by means he morally rejects? Others too are affected; they are made fearful and less secure if such crimes go unpunished. Should the fact that the victim was the one most affected by the crime give him a special status with regard to punishing the offender? (Are the others affected by the crime, or only by its going unpunished?) If the victim was killed does the special status devolve upon the closest kin? If there are two victims of a murderer, do each of the next of kin have a right to punish him with death, with a compet.i.tion for who will be the first to act? Perhaps then, rather than its being the case that anyone may punish or that the victim alone has authority to punish, the solution is that all concerned (namely, everyone) jointly act to punish or to empower someone to punish. But this would require some inst.i.tutional apparatus or mode of decision within the state of nature itself. And, if we specify this as everyone's having a right to a say in the ultimate determination of punishment, this would be the only right of this sort which people possessed in a state of nature; it would add up to a right (the right to determine the punishment) possessed by people jointly rather than individually. There seems to be no neat way to understand how the right to punish would operate within a state of nature. From this discussion of who may exact compensation and who may punish emerges another avenue to the question of a dominant protective a.s.sociation's ent.i.tlement. be carried out at all, or that mercy be granted? May someone punish an offender even against the wishes of the offended party who morally objects to the mode of punishment? If a Gandhian is attacked, may others defend him by means he morally rejects? Others too are affected; they are made fearful and less secure if such crimes go unpunished. Should the fact that the victim was the one most affected by the crime give him a special status with regard to punishing the offender? (Are the others affected by the crime, or only by its going unpunished?) If the victim was killed does the special status devolve upon the closest kin? If there are two victims of a murderer, do each of the next of kin have a right to punish him with death, with a compet.i.tion for who will be the first to act? Perhaps then, rather than its being the case that anyone may punish or that the victim alone has authority to punish, the solution is that all concerned (namely, everyone) jointly act to punish or to empower someone to punish. But this would require some inst.i.tutional apparatus or mode of decision within the state of nature itself. And, if we specify this as everyone's having a right to a say in the ultimate determination of punishment, this would be the only right of this sort which people possessed in a state of nature; it would add up to a right (the right to determine the punishment) possessed by people jointly rather than individually. There seems to be no neat way to understand how the right to punish would operate within a state of nature. From this discussion of who may exact compensation and who may punish emerges another avenue to the question of a dominant protective a.s.sociation's ent.i.tlement.

The dominant protective a.s.sociation is authorized by many persons to act as their agent in exacting compensation for them. It is ent.i.tled to act for them, whereas a small agency is ent.i.tled to act for fewer persons, and an individual is ent.i.tled to act only for himself. In this sense of having a greater number of individual ent.i.tlements, but a kind that others have as well, the dominant protective agency has a greater ent.i.tlement. Something more can be said, given the unclarity about how rights to punish operate in a state of nature. To the extent To the extent that it is plausible that all who have some claim to a right to punish have to act jointly, then the dominant agency will be viewed as having the greatest ent.i.tlement to exact punishment, since almost all authorize it to act in their place. In exacting punishment it displaces and preempts the actions to punish of the fewest others. Any private individual who acts will exclude the actions and ent.i.tlements of all the others; whereas very many people will feel their ent.i.tlement is being exercised when their agent, the dominant protective agency, acts. This would account for thinking that the dominant protective agency or a state has some special legitimacy. Having more ent.i.tlements to act, it is more ent.i.tled to act. But it is not ent.i.tled to be the dominant agency, nor is anyone else. that it is plausible that all who have some claim to a right to punish have to act jointly, then the dominant agency will be viewed as having the greatest ent.i.tlement to exact punishment, since almost all authorize it to act in their place. In exacting punishment it displaces and preempts the actions to punish of the fewest others. Any private individual who acts will exclude the actions and ent.i.tlements of all the others; whereas very many people will feel their ent.i.tlement is being exercised when their agent, the dominant protective agency, acts. This would account for thinking that the dominant protective agency or a state has some special legitimacy. Having more ent.i.tlements to act, it is more ent.i.tled to act. But it is not ent.i.tled to be the dominant agency, nor is anyone else.

We should note one further possible source of viewing something as the legitimate locus of the exercise of enforcing power. To the extent that individuals view choosing a protective agency as a coordination game, with advantages to their quickly converging upon the same one, though it doesn't matter very much which one, they may think the one that happened to be settled upon is the appropriate or proper one now to look to for protection. Consider a neighborhood meeting place for teenagers. It may not matter very much where the place is, so long as everyone knows the place where others will congregate, depending upon others to go there if anywhere. That place becomes "the place to go" to meet others. It is not only that you will be more likely to be unsuccessful if you look elsewhere; it is that others benefit from, and count upon, your converging upon that place, and similarly you benefit from, and count upon, their congregating there. It is not ent.i.tled to be the meeting place; if it is a store its owner is not ent.i.tled to have his store be the one at which people congregate. It is not that individuals must meet there. It's just the place to meet. Similarly, one might imagine a given protective agency's becoming the one to be protected by. To the extent that people attempt to coordinate their actions and converge upon a protective agency which will have all as clients, the process is, to that extent, not fully an invisible-hand one. And there will be intermediate cases, where some view it as a coordination game, and others, oblivious of this, merely react to local signals.14 When only one agency actually exercises the right to prohibit others from using their unreliable procedures for enforcing justice, that makes it the de facto de facto state. Our rationale for this prohibition rests on the ignorance, uncertainty, and lack of knowledge of people. In some situations, it is not known whether a particular person performed a certain action, and procedures for finding this out differ in reliability or fairness. We may ask whether, in a world of perfect factual knowledge and information, anyone could legitimately claim the right (without claiming to be its sole possessor) to prohibit another from punishing a guilty party. Even given factual agreement, there might be disagreement about what amount of punishment a particular act deserved, and about which acts deserved punishment. I have proceeded in this essay (as much as possible) without questioning or focusing upon the a.s.sumption common to much utopian and anarchist theorizing, that there is some set of principles obvious enough to be accepted by all men of good will, precise enough to give unambiguous guidance in particular situations, clear enough so that all will realize its dictates, and complete enough to cover all problems that actually will arise. To have rested the case for the state on the denial of such an a.s.sumption would have left the hope that the future progress of humanity (and moral philosophy) might yield such agreement, and so might undercut the rationale for the state. Not only does the day seem distant when all men of good will shall agree to libertarian principles; these principles have not been completely stated, nor is there now one unique set of principles agreed to by all libertarians. Consider for example, the issue of whether full-blooded copyright is legitimate. Some libertarians argue it isn't legitimate, but claim that its effect can be obtained if authors and publishers include in the contract when they sell books a provision prohibiting its unauthorized printing, and then sue any book pirate for breach of contract; apparently they forget that some people sometimes lose books and others find them. Other libertarians disagree. state. Our rationale for this prohibition rests on the ignorance, uncertainty, and lack of knowledge of people. In some situations, it is not known whether a particular person performed a certain action, and procedures for finding this out differ in reliability or fairness. We may ask whether, in a world of perfect factual knowledge and information, anyone could legitimately claim the right (without claiming to be its sole possessor) to prohibit another from punishing a guilty party. Even given factual agreement, there might be disagreement about what amount of punishment a particular act deserved, and about which acts deserved punishment. I have proceeded in this essay (as much as possible) without questioning or focusing upon the a.s.sumption common to much utopian and anarchist theorizing, that there is some set of principles obvious enough to be accepted by all men of good will, precise enough to give unambiguous guidance in particular situations, clear enough so that all will realize its dictates, and complete enough to cover all problems that actually will arise. To have rested the case for the state on the denial of such an a.s.sumption would have left the hope that the future progress of humanity (and moral philosophy) might yield such agreement, and so might undercut the rationale for the state. Not only does the day seem distant when all men of good will shall agree to libertarian principles; these principles have not been completely stated, nor is there now one unique set of principles agreed to by all libertarians. Consider for example, the issue of whether full-blooded copyright is legitimate. Some libertarians argue it isn't legitimate, but claim that its effect can be obtained if authors and publishers include in the contract when they sell books a provision prohibiting its unauthorized printing, and then sue any book pirate for breach of contract; apparently they forget that some people sometimes lose books and others find them. Other libertarians disagree.15 Similarly for patents. If persons so close in general theory can disagree over a point so fundamental, two libertarian protective agencies might manage to do battle over it. One agency might attempt to enforce a prohibition upon a person's publishing a particular book (because this violates the author's property right) or reproducing a certain invention he has not invented independently, while the other agency fights this prohibition as a violation of individual rights. Disagreements about what is to be enforced, argue the unreluctant archists, provide yet another reason (in addition to lack of factual knowledge) for the apparatus of the state; as also does the need for sometimes changing the content of what is to be enforced. People who prefer peace to the enforcement of their view of right will unite together in Similarly for patents. If persons so close in general theory can disagree over a point so fundamental, two libertarian protective agencies might manage to do battle over it. One agency might attempt to enforce a prohibition upon a person's publishing a particular book (because this violates the author's property right) or reproducing a certain invention he has not invented independently, while the other agency fights this prohibition as a violation of individual rights. Disagreements about what is to be enforced, argue the unreluctant archists, provide yet another reason (in addition to lack of factual knowledge) for the apparatus of the state; as also does the need for sometimes changing the content of what is to be enforced. People who prefer peace to the enforcement of their view of right will unite together in one one state. But of course, if people genuinely state. But of course, if people genuinely do do hold this preference, their protective agencies will not do battle either. hold this preference, their protective agencies will not do battle either.

PREVENTIVE RESTRAINT.

Finally, let us notice how the issue of "preventive detention" or "preventive restraint" is related to the principle of compensation (Chapter 4) and to our discussion in Chapter 5 of the extensive protection it requires the ultraminimal state to provide, even for those who do not pay. The notion should be widened to include all restrictions on individuals in order to lessen the risk that they they will violate others' rights; call this widened notion "preventive restraint." Included under this would be requiring some individuals to report to an official once a week (as if they were on parole), forbidding some individuals from being in certain places at certain hours, gun control laws, and so on (but not laws forbidding the publication of the plans of bank alarm systems). Preventive detention would encompa.s.s imprisoning someone, not for any crime he has committed, but because it is predicted of him that the probability is significantly higher than normal that he will commit a crime. (His previous crimes may be part of the data on the basis of which the predictions are made.) will violate others' rights; call this widened notion "preventive restraint." Included under this would be requiring some individuals to report to an official once a week (as if they were on parole), forbidding some individuals from being in certain places at certain hours, gun control laws, and so on (but not laws forbidding the publication of the plans of bank alarm systems). Preventive detention would encompa.s.s imprisoning someone, not for any crime he has committed, but because it is predicted of him that the probability is significantly higher than normal that he will commit a crime. (His previous crimes may be part of the data on