PROHIBITING PRIVATE ENFORCEMENT OF JUSTICE
AN independent might be prohibited from privately exacting justice because his procedure is known to be too risky and dangerous—that is, it involves a higher risk (than another procedure) of punishing an innocent person or overpunishing a guilty one—or because his procedure isn’t known not to be risky. (His procedure would exhibit another mode of unreliability if its chances were much greater of not punishing a guilty person, but this would not be a reason for prohibiting his private enforcement.)
Let us consider these in turn. If the independent’s procedure is very unreliable and imposes high risk on others (perhaps he consults tea leaves), then if he does it frequently, he may make all fearful, even those not his victims. Anyone, acting in self-defense, may stop him from engaging in his high-risk activity. But surely the independent may be stopped from using a very unreliable procedure, even if he is not a constant menace. If it is known that the independent will enforce his own rights by his very unreliable procedure only once every ten years, this will not create general fear and apprehension in the society. The ground for prohibiting his widely intermittent use of his procedure is not, therefore, to avoid any widespread uncompensated apprehension and fear which otherwise would exist.
If there were many independents who were all liable to punish wrongly, the probabilities would add up to create a dangerous situation for all. Then, others would be entitled to group together and prohibit the totality of such activities. But how would this prohibition work? Would they prohibit each of the individually non-fear-creating activities? Within a state of nature by what procedure can they pick and choose which of the totality is to continue, and what would give them the right to do this? No protective association, however dominant, would have this right. For the legitimate powers of a protective association are merely the sum of the individual rights that its members or clients transfer to the association. No new rights and powers arise; each right of the association is decomposable without residue into those individual rights held by distinct individuals acting alone in a state of nature. A combination of individuals may have the right to do some action C, which no individual alone had the right to do, if C is identical to D and E, and persons who individually have the right to do D and the right to do E combine. If some rights of individuals were of the form “You have the right to do A provided 51 percent or 85 percent or whatever of the others agree you may,” then a combination of individuals would have the right to do A, even though none separately had this right. But no individual’s rights are of this form. No person or group is entitled to pick who in the totality will be allowed to continue. All the independents might group together and decide this. They might, for example, use some random procedure to allocate a number of (sellable?) rights to continue private enforcement so as to reduce the total danger to a point below the threshold. The difficulty is that, if a large number of independents do this, it will be in the interests of an individual to abstain from this arrangement. It will be in his interests to continue his risky activities as he chooses, while the others mutually limit theirs so as to bring the totality of acts including his to below the danger level. For the others probably would limit themselves some distance away from the danger boundary, leaving him room to squeeze in. Even were the others to rest adjacent to the line of danger so that his activities would bring the totality across it, on which grounds could his activities be picked out as the ones to prohibit? Similarly, it will be in the interests of any individual to refrain from otherwise unanimous agreements in the state of nature: for example, the agreement to set up a state. Anything an individual can gain by such a unanimous agreement he can gain through separate bilateral agreements. Any contract which really needs almost unanimity, any contract which is essentially joint, will serve its purpose whether or not a given individual participates; so it will be in his interests not to bind himself to participate.
A principle suggested by Herbert Hart, which (following John Rawls) we shall call the principle of fairness, would be of service here if it were adequate. This principle holds that when a number of persons engage in a just, mutually advantageous, cooperative venture according to rules and thus restrain their liberty in ways necessary to yield advantages for all, those who have submitted to these restrictions have a right to similar acquiescence on the part of those who have benefited from their submission.1 Acceptance of benefits (even when this is not a giving of express or tacit undertaking to cooperate) is enough, according to this principle, to bind one. If one adds to the principle of fairness the claim that the others to whom the obligations are owed or their agents may enforce the obligations arising under this principle (including the obligation to limit one’s actions), then groups of people in a state of nature who agree to a procedure to pick those to engage in certain acts will have legitimate rights to prohibit “free riders.” Such a right may be crucial to the viability of such agreements. We should scrutinize such a powerful right very carefully, especially as it seems to make unanimous consent to coercive government in a state of nature unnecessary! Yet a further reason to examine it is its plausibility as a counterexample to my claim that no new rights “emerge” at the group level, that individuals in combination cannot create new rights which are not the sum of preexisting ones. A right to enforce others’ obligation to limit their conduct in specified ways might stem from some special feature of the obligation or might be thought to follow from some general principle that all obligations owed to others may be enforced. In the absence of argument for the special enforcement-justifying nature of the obligation supposedly arising under the principle of fairness, I shall consider first the principle of the enforceability of all obligations and then turn to the adequacy of the principle of fairness itself. If either of these principles is rejected, the right to enforce the cooperation of others in these situations totters. I shall argue that both of these principles must be rejected.
Herbert Hart’s argument for the existence of a natural right2 depends upon particularizing the principle of the enforceability of all obligations: someone’s being under a special obligation to you to do A (which might have arisen, for example, by their promising to you that they would do A) gives you, not only the right that they do A, but also the right to force them to do A. Only against a background in which people may not force you to do A or other actions you may promise to do can we understand, says Hart, the point and purpose of special obligations. Since special obligations do have a point and purpose, Hart continues, there is a natural right not to be forced to do something unless certain specified conditions pertain; this natural right is built into the background against which special obligations exist.
This well-known argument of Hart’s is puzzling. I may release someone from an obligation not to force me to do A. (“I now release you from the obligation not to force me to do A. You now are free to force me to do A.”) Yet so releasing them does not create in me an obligation to them to do A. Since Hart supposes that my being under an obligation to someone to do A gives him (entails that he has) the right to force me to do A, and since we have seen the converse does not hold, we may consider that component of being under an obligation to someone to do something over and above his having the right to force you to do it. (May we suppose there is this distinguishable component without facing the charge of “logical atomism”?) An alternative view which rejects Hart’s inclusion of the right to force in the notion of being owed an obligation might hold that this additional component is the whole of the content of being obligated to someone to do something. If I don’t do it, then (all things being equal) I’m doing something wrong; control over the situation is in his hands; he has the power to release me from the obligation unless he’s promised to someone else that he won’t, and so on. Perhaps all this looks too ephemeral without the additional presence of rights of enforcement. Yet rights of enforcement are themselves merely rights; that is, permissions to do something and obligations on others not to interfere. True, one has the right to enforce these further obligations, but it is not clear that including rights of enforcing really shores up the whole structure if one assumes it to be insubstantial to begin with. Perhaps one must merely take the moral realm seriously and think one component amounts to something even without a connection to enforcement. (Of course, this is not to say that this component never is connected with enforcement!) On this view, we can explain the point of obligations without bringing in rights of enforcement and hence without supposing a general background of obligation not to force from which this stands out. (Of course, even though Hart’s argument does not demonstrate the existence of such an obligation not to force, it may exist nevertheless.)
Apart from these general considerations against the principle of the enforceability of all special obligations, puzzle cases can be produced. For example, if I promise to you that I will not murder someone, this does not give you the right to force me not to, for you already have this right, though it does create a particular obligation to you. Or, if I cautiously insist that you first promise to me that you won’t force me to do A before I will make my promise to you to do A, and I do receive this promise from you first, it would be implausible to say that in promising I give you the right to force me to do A. (Though consider the situation which results if I am so foolish as to release you unilaterally from your promise to me.)
If there were cogency to Hart’s claim that only against a background of required nonforcing can we understand the point of special rights, then there would seem to be equal cogency to the claim that only against a background of permitted forcing can we understand the point of general rights. For according to Hart, a person has a general right to do A if and only if for all persons P and Q, Q may not interfere with P’s doing A or force him not to do A, unless P has acted to give Q a special right to do this. But not every act can be substituted for “A”; people have general rights to do only particular types of action. So, one might argue, if there is to be a point to having general rights, to having rights to do a particular type of act A, to other’s being under an obligation not to force you not to do A, then it must be against a contrasting background, in which there is no obligation on people to refrain from forcing you to do, or not to do, things, that is, against a background in which, for actions generally, people do not have a general right to do them. If Hart can argue to a presumption against forcing from there being a point to particular rights, then it seems he can equally well argue to the absence of such a presumption from there being a point to general rights.3
An argument for an enforceable obligation has two stages: the first leads to the existence of the obligation, and the second, to its enforceability. Having disposed of the second stage (at least insofar as it is supposed generally to follow from the first), let us turn to the supposed obligation to cooperate in the joint decisions of others to limit their activities. The principle of fairness, as we stated it following Hart and Rawls, is objectionable and unacceptable. Suppose some of the people in your neighborhood (there are 364 other adults) have found a public address system and decide to institute a system of public entertainment. They post a list of names, one for each day, yours among them. On his assigned day (one can easily switch days) a person is to run the public address system, play records over it, give news bulletins, tell amusing stories he has heard, and so on. After 138 days on which each person has done his part, your day arrives. Are you obligated to take your turn? You have benefited from it, occasionally opening your window to listen, enjoying some music or chuckling at someone’s funny story. The other people have put themselves out. But must you answer the call when it is your turn to do so? As it stands, surely not. Though you benefit from the arrangement, you may know all along that 364 days of entertainment supplied by others will not be worth your giving up one day. You would rather not have any of it and not give up a day than have it all and spend one of your days at it. Given these preferences, how can it be that you are required to participate when your scheduled time comes? It would be nice to have philosophy readings on the radio to which one could tune in at any time, perhaps late at night when tired. But it may not be nice enough for you to want to give up one whole day of your own as a reader on the program. Whatever you want, can others create an obligation for you to do so by going ahead and starting the program themselves? In this case you can choose to forgo the benefit by not turning on the radio; in other cases the benefits may be unavoidable. If each day a different person on your street sweeps the entire street, must you do so when your time comes? Even if you don’t care that much about a clean street? Must you imagine dirt as you traverse the street, so as not to benefit as a free rider? Must you refrain from turning on the radio to hear the philosophy readings? Must you mow your front lawn as often as your neighbors mow theirs?
At the very least one wants to build into the principle of fairness the condition that the benefits to a person from the actions of the others are greater than the costs to him of doing his share. How are we to imagine this? Is the condition satisfied if you do enjoy the daily broadcasts over the PA system in your neighborhood but would prefer a day off hiking, rather than hearing these broadcasts all year? For you to be obligated to give up your day to broadcast mustn’t it be true, at least, that there is nothing you could do with a day (with that day, with the increment in any other day by shifting some activities to that day) which you would prefer to hearing broadcasts for the year? If the only way to get the broadcasts was to spend the day participating in the arrangement, in order for the condition that the benefits outweigh the costs to be satisfied, you would have to be willing to spend it on the broadcasts rather than to gain any other available thing.
If the principle of fairness were modified so as to contain this very strong condition, it still would be objectionable. The benefits might only barely be worth the costs to you of doing your share, yet others might benefit from this institution much more than you do; they all treasure listening to the public broadcasts. As the person least benefited by the practice, are you obligated to do an equal amount for it? Or perhaps you would prefer that all cooperated in another venture, limiting their conduct and making sacrifices for it. It is true, given that they are not following your plan (and thus limiting what other options are available to you), that the benefits of their venture are worth to you the costs of your cooperation. However, you do not wish to cooperate, as part of your plan to focus their attention on your alternative proposal which they have ignored or not given, in your view at least, its proper due. (You want them, for example, to read the Talmud on the radio instead of the philosophy they are reading.) By lending the institution (their institution) the support of your cooperating in it, you will only make it harder to change or alter.4
On the face of it, enforcing the principle of fairness is objectionable. You may not decide to give me something, for example a book, and then grab money from me to pay for it, even if I have nothing better to spend the money on. You have, if anything, even less reason to demand payment if your activity that gives me the book also benefits you; suppose that your best way of getting exercise is by throwing books into people’s houses, or that some other activity of yours thrusts books into people’s houses as an unavoidable side effect. Nor are things changed if your inability to collect money or payments for the books which unavoidably spill over into others’ houses makes it inadvisable or too expensive for you to carry on the activity with this side effect. One cannot, whatever one’s purposes, just act so as to give people benefits and then demand (or seize) payment. Nor can a group of persons do this. If you may not charge and collect for benefits you bestow without prior agreement, you certainly may not do so for benefits whose bestowal costs you nothing, and most certainly people need not repay you for costless-to-provide benefits which yet others provided them. So the fact that we partially are “social products” in that we benefit from current patterns and forms created by the multitudinous actions of a long string of long-forgotten people, forms which include institutions, ways of doing things, and language (whose social nature may involve our current use depending upon Wittgensteinian matching of the speech of others), does not create in us a general floating debt which the current society can collect and use as it will.
Perhaps a modified principle of fairness can be stated which would be free from these and similar difficulties. What seems certain is that any such principle, if possible, would be so complex and involuted that one could not combine it with a special principle legitimating enforcement within a state of nature of the obligations that have arisen under it. Hence, even if the principle could be formulated so that it was no longer open to objection, it would not serve to obviate the need for other persons’ consenting to cooperate and limit their own activities.
Let us return to our independent. Apart from other nonindependents’ fear (perhaps they will not be so worried), may not the person about to be punished defend himself? Must he allow the punishment to take place, collecting compensation afterwards if he can show that it was unjust? But show to whom? If he knows he’s innocent, may he demand compensation immediately and enforce his rights to collect it? And so on. The notions of procedural rights, public demonstration of guilt, and the like, have a very unclear status within state-of-nature theory.
It might be said that each person has a right to have his guilt determined by the least dangerous of the known procedures for ascertaining guilt, that is, by the one having the lowest probability of finding an innocent person guilty. There are well-known maxims of the following form: better m guilty persons go free than n innocent persons be punished. For each n, each maxim will countenance an upper limit to the ratio m/n. It will say: better m, but not better m + 1. (A system may pick differing upper limits for different crimes.) On the greatly implausible assumption that we know each system of procedures’ precise probability of finding an innocent person guilty,5 and a guilty person innocent, we will opt for those procedures whose long-run ratio of the two kinds of errors comes closest, from below, to the highest ratio we find acceptable. It is far from obvious where to set the ratio. To say it is better that any number of guilty go free rather than that one innocent person be punished presumably would require not having any system of punishment at all. For any system we can devise which sometimes does actually punish someone will involve some appreciable risk of punishing an innocent person, and it almost certainly will do so as it operates on large numbers of people. And any system S can be transformed into one having a lower probability of punishing an innocent person, for example, by conjoining to it a roulette procedure whereby the probability is only. 1 that anyone found guilty by S actually gets punished. (This procedure is iterative.)
If a person objects that the independent’s procedure yields too high a probability of an innocent person’s being punished, how can it be determined what probabilities are too high? We can imagine that each individual goes through the following reasoning: The greater the procedural safeguards, the less my chances of getting unjustly convicted, and also the greater the chances that a guilty person goes free; hence the less effectively the system deters crime and so the greater my chances of being a victim of a crime. That system is most effective which minimizes the expected value of unearned harm to me, either through my being unjustly punished or through my being a victim of a crime. If we simplify greatly by assuming that penalties and victimization costs balance out, one would want the safeguards at that most stringent point where any lowering of them would increase one’s probability of being unjustly punished more than it would lower (through added deterrence) one’s vulnerability to being victimized by a crime; and where any increasing of the safeguards would increase one’s probability of being victimized by a crime (through lessened deterrence) more than it would lessen one’s probability of being punished though innocent. Since utilities differ among persons, there is no reason to expect individuals who make such an expected value calculation to converge upon the identical set of procedures. Furthermore, some persons may think it important in itself that guilty people be punished and may be willing to run some increased risks of being punished themselves in order to accomplish this. These people will consider it more of a drawback, the greater the probability a procedure gives guilty people of going unpunished, and they will incorporate this in their calculations, apart from its effects on deterrence. It is, to say the least, very doubtful that any provision of the law of nature will (and will be known to) settle the question of how much weight is to be given to such considerations, or will reconcile people’s different assessments of the seriousness of being punished when innocent as compared to being victimized by a crime (even if both involve the same physical thing happening to them). With the best will in the world, individuals will favor differing procedures yielding differing probabilities of an innocent person’s being punished.
One could not, it seems, permissibly prohibit someone from using a procedure solely because it yields a marginally higher probability of punishing an innocent person than does the procedure you deem optimal. After all, your favorite procedure also will stand in this relation to that of someone else. Nor are matters changed by the fact that many other persons use your procedure. It seems that persons in a state of nature must tolerate (that is, not forbid) the use of procedures in the “neighborhood” of their own; but it seems they may forbid the use of far more risky procedures. An acute problem is presented if two groups each believe their own procedures to be reliable while believing that of the other group to be very dangerous. No procedure to resolve their disagreement seems likely to work; and presenting the nonprocedural principle that the group which is right should triumph (and the other should give in to it) seems unlikely to produce peace when each group, firmly believing itself to be the one that is right, acts on the principle.
When sincere and good persons differ, we are prone to think they must accept some procedure to decide their differences, some procedure they both agree to be reliable or fair. Here we see the possibility that this disagreement may extend all the way up the ladder of procedures. Also, one sometimes will refuse to let issues stay settled by the adverse decision of such a procedure, specifically when the wrong decision is worse even than the disruption and costs (including fighting) of refusing to accept it, when the wrong decision is worse than conflict with those on the other side. It is dismaying to contemplate situations where both of the opposed parties feel that conflict is preferable to an adverse decision by any procedure. Each views the situation as one in which he who is right must act, and the other should give in. It will be of little avail for a neutral party to say to both, “Look, you both think you’re right, so on that principle, as you will apply it, you’ll fight. Therefore you must agree to some procedure to decide the matter.” For they each believe that conflict is better than losing the issue.* And one of them may be right in this. Shouldn’t he engage in the conflict? Shouldn’t he engage in the conflict? (True, both of them will think the one is themselves.) One might try to avoid these painful issues by a commitment to procedures, come what may. (May one possible result of applying the procedures be that they themselves are rejected?) Some view the state as such a device for shifting the ultimate burden of moral decision, so that there never comes to be that sort of conflict among individuals. But what sort of individual could so abdicate? Who could turn every decision over to an external procedure, accepting whatever results come? The possibility of such conflict is part of the human condition. Though this problem in the state of nature is an unavoidable one, given suitable institutional elaboration it need be no more pressing in the state of nature than under a state, where it also exists.6
The issue of which decisions can be left to an external binding procedure connects with the interesting question of what moral obligations someone is under who is being punished for a crime of which he knows himself to be innocent. The judicial system (containing no procedural unfairness, let us suppose) has sentenced him to life imprisonment, or death. May he escape? May he harm another in order to escape? These questions differ from the one of whether someone wrongfully attacking (or participating in the attack of) another may claim self-defense as justifying his killing the other when the other, in self-defense, acts so as to endanger his own attacker’s life. Here the answer is, “No.” The attacker should not be attacking in the first place, nor does someone else’s threatening him with death unless he does attack make it permissible for him to do so. His job is to get out of that situation; if he fails to do so he is at a moral disadvantage. Soldiers who know their country is waging an aggressive war and who are manning antiaircraft guns in defense of a military emplacement may not in self-defense fire upon the planes of the attacked nation which is acting in self-defense, even though the planes are over their heads and are about to bomb them. It is a soldier’s responsibility to determine if his side’s cause is just; if he finds the issue tangled, unclear, or confusing, he may not shift the responsibility to his leaders, who will certainly tell him their cause is just. The selective conscientious objector may be right in his claim that he has a moral duty not to fight; and if he is, may not another acquiescent soldier be punished for doing what it was his moral duty not to do? Thus we return to the point that some bucks stop with each of us; and we reject the morally elitist view that some soldiers cannot be expected to think for themselves. (They are certainly not encouraged to think for themselves by the practice of absolving them of all responsibility for their actions within the rules of war.) Nor do we see why the political realm is special. Why, precisely, is one specially absolved of responsibility for actions when these are performed jointly with others from political motives under the direction or orders of political leaders?7
We thus far have supposed that you know that another’s procedure of justice differs from your own for the worse. Suppose now that you have no reliable knowledge about another’s procedure of justice. May you stop him in self-defense and may your protective agency act for you, solely because you or it does not know whether his procedure is reliable? Do you have the right to have your guilt or innocence, and punishment, determined by a system known to be reliable and fair? Known to whom? Those wielding it may know it to be reliable and fair. Do you have a right to have your guilt or innocence, and punishment, determined by a system you know to be reliable and fair? Are someone’s rights violated if he thinks that only the use of tea leaves is reliable or if he is incapable of concentrating on the description of the system others use so that he doesn’t know whether it’s reliable, and so on? One may think of the state as the authoritative settler of doubts about reliability and fairness. But of course there is no guarantee that it will settle them (the president of Yale didn’t think Black Panthers could get a fair trial), and there is no reason to suppose it will manage to do so more effectively than another scheme. The natural-rights tradition offers little guidance on precisely what one’s procedural rights are in a state of nature, on how principles specifying how one is to act have knowledge built into their various clauses, and so on. Yet persons within this tradition do not hold that there are no procedural rights; that is, that one may not defend oneself against being handled by unreliable or unfair procedures.
HOW MAY THE DOMINANT AGENCY ACT?
What then may a dominant protective association forbid other individuals to do? The dominant protective association may reserve for itself the right to judge any procedure of justice to be applied to its clients. It may announce, and act on the announcement, that it will punish anyone who uses on one of its clients a procedure that it finds to be unreliable or unfair. It will punish anyone who uses on one of its clients a procedure that it already knows to be unreliable or unfair, and it will defend its clients against the application of such a procedure. May it announce that it will punish anyone who uses on one of its clients a procedure that it has not, at the time of punishment, already approved as reliable and fair? May it set itself up as having to pass, in advance, on any procedure to be used on one of its clients, so that anyone using on one of its clients any procedure that has not already received the protective association’s seal of approval will be punished? Clearly, individuals themselves do not have this right. To say that an individual may punish anyone who applies to him a procedure of justice that has not met his approval would be to say that a criminal who refuses to approve anyone’s procedure of justice could legitimately punish anyone who attempted to punish him. It might be thought that a protective association legitimately can do this, for it would not be partial to its clients in this manner. But there is no guarantee of this impartiality. Nor have we seen any way that such a new right might arise from the combining of individuals’ preexisting rights. We must conclude that protective associations do not have this right, including the sole dominant one.
Every individual does have the right that information sufficient to show that a procedure of justice about to be applied to him is reliable and fair (or no less so than other procedures in use) be made publicly available or made available to him. He has the right to be shown that he is being handled by some reliable and fair system. In the absence of such a showing he may defend himself and resist the imposition of the relatively unknown system. When the information is made publicly available or made available to him, he is in a position to know about the reliability and fairness of the procedure.8 He examines this information, and if he finds the system within the bounds of reliability and fairness he must submit to it; finding it unreliable and unfair he may resist. His submission means that he refrains from punishing another for using this system. He may resist the imposition of its particular decision though, on the grounds that he is innocent. If he chooses not to, he need not participate in the process whereby the system determines his guilt or innocence. Since it has not yet been established that he is guilty, he may not be aggressed against and forced to participate. However, prudence might suggest to him that his chances of being found innocent are increased if he cooperates in the offering of some defense.
The principle is that a person may resist, in self-defense, if others try to apply to him an unreliable or unfair procedure of justice. In applying this principle, an individual will resist those systems which after all conscientious consideration he finds to be unfair or unreliable. An individual may empower his protective agency to exercise for him his rights to resist the imposition of any procedure which has not made its reliability and fairness known, and to resist any procedure that is unfair or unreliable. In Chapter 2 we described briefly the processes that would lead to the dominance of one protective association in a given area, or to a dominant federation of protective associations using rules to peacefully adjudicate disputes among themselves. This dominant protective association will prohibit anyone from applying to its members any procedure about which insufficient information is available as to its reliability and fairness. It also will prohibit anyone from applying to its members an unreliable or unfair procedure; which means, since they are applying the principle and have the muscle to do so, that others are prohibited from applying to the protective association’s members any procedure the protective association deems unfair or unreliable. Leaving aside the chances of evading the system’s operation, anyone violating this prohibition will be punished. The protective association will publish a list of those procedures it deems fair and reliable (and perhaps of those it deems otherwise); and it would take a brave soul indeed to proceed to apply a known procedure not yet on its approved list. Since an association’s clients will expect it to do all it can to discourage unreliable procedures, the protective association will keep its list up-to-date, covering all publicly known procedures.
It might be claimed that our assumption that procedural rights exist makes our argument too easy. Does a person who did violate another’s rights himself have a right that this fact be determined by a fair and reliable procedure? It is true that an unreliable procedure will too often find an innocent person guilty. But does applying such an unreliable procedure to a guilty person violate any right of his? May he, in self-defense, resist the imposition of such a procedure upon himself? But what would he be defending himself against? Too high a probability of a punishment he deserves? These questions are important ones for our argument. If a guilty person may not defend himself against such procedures and also may not punish someone else for using them upon him, then may his protective agency defend him against the procedures or punish someone afterwards for having used them upon him, independently of whether or not (and therefore even if) he turns out to be guilty? One would have thought the agency’s only rights of action are those its clients transfer to it. But if a guilty client has no such right, he cannot transfer it to the agency.
The agency does not, of course, know that its client is guilty, whereas the client himself does know (let us suppose) of his own guilt. But does this difference in knowledge make the requisite difference? Isn’t the ignorant agency required to investigate the question of its client’s guilt, instead of proceeding on the assumption of his innocence? The difference in epistemic situation between agency and client can make the following difference. The agency may under some circumstances defend its client against the imposition of a penalty while promptly proceeding to investigate the question of his guilt. If the agency knows that the punishing party has used a reliable procedure, it accepts its verdict of guilty, and it cannot intervene on the assumption that its client is, or well might be, innocent. If the agency deems the procedure unreliable or doesn’t know how reliable it is, it need not presume its client guilty, and it may investigate the matter itself. If upon investigation it determines that its client is guilty, it allows him to be punished. This protection of its client against the actual imposition of the penalty is relatively straightforward, except for the question of whether the agency must compensate the prospective punishers for any costs imposed upon them by having to delay while the protective agency determines to its satisfaction its own client’s guilt. It would seem that the protective agency does have to pay compensation to users of relatively unreliable procedures for any disadvantages caused by the enforced delay; and to the users of procedures of unknown reliability it must pay full compensation if the procedures are reliable, otherwise compensation for disadvantages. (Who bears the burden of proof in the question of the reliability of the procedures?) Since the agency may recover this amount (forcibly) from its client who asserted his innocence, this will be something of a deterrent to false pleas of innocence.*
The agency’s temporary protection and defense against the infliction of the penalty is relatively straightforward. Less straightforward is the protective agency’s appropriate action after a penalty has been inflicted. If the punisher’s procedure was a reliable one, the agency does not act against the punisher. But may the agency punish someone who punishes its client, acting on the basis of an unreliable procedure? May it punish that person independently of whether or not its client is guilty? Or must it investigate, using its own reliable procedure, to determine his guilt or innocence, punishing his punishers only if it determines its client innocent? (Or is it: if it fails to find him guilty?) By what right could the protective agency announce that it will punish anyone using an unreliable procedure who punishes its clients, independently of the guilt or innocence of the clients?
The person who uses an unreliable procedure, acting upon its result, imposes risks upon others, whether or not his procedure misfires in a particular case. Someone playing Russian roulette upon another does the same thing if when he pulls the trigger the gun does not fire. The protective agency may treat the unreliable enforcer of justice as it treats any performer of a risky action. We distinguished in Chapter 4 a range of possible responses to a risky action, which were appropriate in different sorts of circumstances: prohibition, compensation to those whose boundaries are crossed, and compensation to all those who undergo a risk of a boundary crossing. The unreliable enforcer of justice might either perform actions others are fearful of, or not; and either might be done to obtain compensation for some previous wrong, or to exact retribution.9 A person who uses an unreliable procedure of enforcing justice and is led to perform some unfeared action will not be punished afterwards. If it turns out that the person on whom he acted was guilty and that the compensation taken was appropriate, the situation will be left as is. If the person on whom he acted turns out to be innocent, the unreliable enforcer of justice may be forced fully to compensate him for the action.
On the other hand, the unreliable enforcer of justice may be forbidden to impose those consequences that would be feared if expected. Why? If done frequently enough so as to create general fear, such unreliable enforcement may be forbidden in order to avoid the general uncompensated-for fear. Even if done rarely, the unreliable enforcer may be punished for imposing this feared consequence upon an innocent person. But if the unreliable enforcer acts rarely and creates no general fear, why may he be punished for imposing a feared consequence upon a person who is guilty? A system of punishing unreliable punishers for their punishment of guilty persons would help deter them from using their unreliable system upon anyone and therefore from using it upon innocent people. But not everything that would aid in such deterrence may be inflicted. The question is whether it would be legitimate in this case to punish after the fact the unreliable punisher of someone who turned out to be guilty.
No one has a right to use a relatively unreliable procedure in order to decide whether to punish another. Using such a system, he is in no position to know that the other deserves punishment; hence he has no right to punish him. But how can we say this? If the other has committed a crime, doesn’t everyone in a state of nature have a right to punish him? And therefore doesn’t someone who doesn’t know that this other person has committed the crime? Here, it seems to me, we face a terminological issue about how to merge epistemic considerations with rights. Shall we say that someone doesn’t have a right to do certain things unless he knows certain facts, or shall we say that he does have a right but he does wrong in exercising it unless he knows certain facts? It may be neater to decide it one way, but we can still say all we wish in the other mode; there is a simple translation between the two modes of discourse.10 We shall pick the latter mode of speech; if anything, this makes our argument look less compelling. If we assume that anyone has a right to take something that a thief has stolen, then under this latter terminology someone who takes a stolen object from a thief, without knowing it had been stolen, had a right to take the object; but since he didn’t know he had this right, his taking the object was wrong and impermissible. Even though no right of the first thief is violated, the second didn’t know this and so acted wrongly and impermissibly.
Having taken this terminological fork, we might propose an epistemic principle of border crossing: If doing act A would violate Q’s rights unless condition C obtained, then someone who does not know that C obtains may not do A. Since we may assume that all know that inflicting a punishment upon someone violates his rights unless he is guilty of an offense, we may make do with the weaker principle: If someone knows that doing act A would violate Q’s rights unless condition C obtained, he may not do A if he does not know that C obtains. Weaker still, but sufficient for our purposes, is: If someone knows that doing act A would violate Q’s rights unless condition C obtained, he may not do A if he has not ascertained that C obtains through being in the best feasible position for ascertaining this. (This weakening of the consequent also avoids various problems connected with epistemological skepticism.) Anyone may punish a violator of this prohibition. More precisely, anyone has the right so to punish a violator; people may do so only if they themselves don’t run afoul of the prohibition, that is, only if they themselves have ascertained that another violated the prohibition, being in the best position to have ascertained this.
On this view, what a person may do is not limited only by the rights of others. An unreliable punisher violates no right of the guilty person; but still he may not punish him. This extra space is created by epistemic considerations. (It would be a fertile area for investigation, if one could avoid drowning in the morass of considerations about “subjective-ought” and “objective-ought.”) Note that on this construal, a person does not have a right that he be punished only by use of a relatively reliable procedure. (Even though he may, if he so chooses, give another permission to use a less reliable procedure on him.) On this view, many procedural rights stem not from rights of the person acted upon, but rather from moral considerations about the person or persons doing the acting.
It is not clear to me that this is the proper focus. Perhaps the person acted upon does have such procedural rights against the user of an unreliable procedure. (But what is a guilty person’s complaint against an unreliable procedure. That it is too likely to mispunish him? Would we have the user of an unreliable procedure compensate the guilty person he punished, for violating his right?) We have seen that our argument for a protective agency’s punishing the wielder of the unreliable procedure for inflicting a penalty upon its client would go much more smoothly were this so. The client merely would authorize his agency to act to enforce his procedural right. For the purposes of our subargument here, we have shown that our conclusion stands, even without the facilitating assumption of procedural rights. (We do not mean to imply that there aren’t such rights.) In either case, a protective agency may punish a wielder of an unreliable or unfair procedure who (against the client’s will) has punished one of its clients, independently of whether or not its client actually is guilty and therefore even if its client is guilty.
The tradition of theorizing about the state we discussed briefly in Chapter 2 has a state claiming a monopoly on the use of force. Has any monopoly element yet entered our account of the dominant protective agency? Everyone may defend himself against unknown or unreliable procedures and may punish those who use or attempt to use such procedures against him. As its client’s agent, the protective association has the right to do this for its clients. It grants that every individual, including those not affiliated with the association, has this right. So far, no monopoly is claimed. To be sure, there is a universal element in the content of the claim: the right to pass on anyone’s procedure. But it does not claim to be the sole possessor of this right; everyone has it. Since no claim is made that there is some right which it and only it has, no monopoly is claimed. With regard to its own clients, however, it applies and enforces these rights which it grants that everyone has. It deems its own procedures reliable and fair. There will be a strong tendency for it to deem all other procedures, or even the “same” procedures run by others, either unreliable or unfair. But we need not suppose it excludes every other procedure. Everyone has the right to defend against procedures that are in fact not, or not known to be, both reliable and fair. Since the dominant protective association judges its own procedures to be both reliable and fair, and believes this to be generally known, it will not allow anyone to defend against them; that is, it will punish anyone who does so. The dominant protective association will act freely on its own understanding of the situation, whereas no one else will be able to do so with impunity. Although no monopoly is claimed, the dominant agency does occupy a unique position by virtue of its power. It, and it alone, enforces prohibitions on others’ procedures of justice, as it sees fit. It does not claim the right to prohibit others arbitrarily; it claims only the right to prohibit anyone’s using actually defective procedures on its clients. But when it sees itself as acting against actually defective procedures, others may see it as acting against what it thinks are defective procedures. It alone will act freely against what it thinks are defective procedures, whatever anyone else thinks. As the most powerful applier of principles which it grants everyone the right to apply correctly, it enforces its will, which, from the inside, it thinks is correct. From its strength stems its actual position as the ultimate enforcer and the ultimate judge with regard to its own clients. Claiming only the universal right to act correctly, it acts correctly by its own lights. It alone is in a position to act solely by its own lights.
Does this unique position constitute a monopoly? There is no right the dominant protective association claims uniquely to possess. But its strength leads it to be the unique agent acting across the board to enforce a particular right. It is not merely that it happens to be the only exerciser of a right it grants that all possess; the nature of the right is such that once a dominant power emerges, it alone will actually exercise that right. For the right includes the right to stop others from wrongfully exercising the right, and only the dominant power will be able to exercise this right against all others. Here, if anywhere, is the place for applying some notion of a de facto monopoly: a monopoly that is not de jure because it is not the result of some unique grant of exclusive right while others are excluded from exercising a similar privilege. Other protective agencies, to be sure, can enter the market and attempt to wean customers away from the dominant protective agency. They can attempt to replace it as the dominant one. But being the already dominant protective agency gives an agency a significant market advantage in the competition for clients. The dominant agency can offer its customers a guarantee that no other agencies can match: “Only those procedures we deem appropriate will be used on our customers.”
The dominant protective agency’s domain does not extend to quarrels of nonclients among themselves. If one independent is about to use his procedure of justice upon another independent, then presumably the protective association would have no right to intervene. It would have the right we all do to intervene to aid an unwilling victim whose rights are threatened. But since it may not intervene on paternalistic grounds, the protective association would have no proper business interfering if both independents were satisfied with their procedure of justice. This does not show that the dominant protective association is not a state. A state, too, could abstain from disputes where all concerned parties chose to opt out of the state’s apparatus. (Though it is more difficult for people to opt out of the state in a limited way, by choosing some other procedure for settling a particular quarrel of theirs. For that procedure’s settlement, and their reactions to it, might involve areas that not all parties concerned have removed voluntarily from the state’s concern.) And shouldn’t (and mustn’t) each state allow that option to its citizens?
If the protective agency deems the independents’ procedures for enforcing their own rights insufficiently reliable or fair when applied to its clients, it will prohibit the independents from such self-help enforcement. The grounds for this prohibition are that the self-help enforcement imposes risks of danger on its clients. Since the prohibition makes it impossible for the independents credibly to threaten to punish clients who violate their rights, it makes them unable to protect themselves from harm and seriously disadvantages the independents in their daily activities and life. Yet it is perfectly possible that the independents’ activities including self-help enforcement could proceed without anyone’s rights being violated (leaving aside the question of procedural rights). According to our principle of compensation given in Chapter 4, in these circumstances those persons promulgating and benefiting from the prohibition must compensate those disadvantaged by it. The clients of the protective agency, then, must compensate the independents for the disadvantages imposed upon them by being prohibited self-help enforcement of their own rights against the agency’s clients. Undoubtedly, the least expensive way to compensate the independents would be to supply them with protective services to cover those situations of conflict with the paying customers of the protective agency. This will be less expensive than leaving them unprotected against violations of their rights (by not punishing any client who does so) and then attempting to pay them afterwards to cover their losses through having (and being in a position in which they were exposed to having) their rights violated. If it were not less expensive, then instead of buying protective services, people would save their money and use it to cover their losses, perhaps by jointly pooling their money in an insurance scheme.
Must the members of the protective agency pay for protective services (vis-à-vis its clients) for the independents? Can they insist that the independents purchase the services themselves? After all, using self-help procedures would not have been without costs for the independent. The principle of compensation does not require those who prohibit an epileptic from driving to pay his full cost of taxis, chauffeurs, and so on. If the epileptic were allowed to run his own automobile, this too would have its costs: money for the car, insurance, gasoline, repair bills, and aggravation. In compensating for disadvantages imposed, the prohibitors need pay only an amount sufficient to compensate for the disadvantages of the prohibition minus an amount representing the costs the prohibited party would have borne were it not for the prohibition. The prohibitors needn’t pay the complete costs of taxis; they must pay only the amount which when combined with the costs to the prohibited party of running his own private automobile is sufficient for taxis. They may find it less expensive to compensate in kind for the disadvantages they impose than to supply monetary compensation; they may engage in some activity that removes or partially lessens the disadvantages, compensating in money only for the net disadvantages remaining.
If the prohibitor pays to the person prohibited monetary compensation equal to an amount that covers the disadvantages imposed minus the costs of the activity where it permitted, this amount may be insufficient to enable the prohibited party to overcome the disadvantages. If his costs in performing the prohibited action would have been monetary, he can combine the compensation payment with this money unspent and purchase the equivalent service. But if his costs would not have been directly monetary but involve energy, time, and the like, as in the case of the independent’s self-help enforcement of rights, then this monetary payment of the difference will not by itself enable the prohibited party to overcome the disadvantage by purchasing the equivalent of what he is prohibited. If the independent has other financial resources he can use without disadvantaging himself, then this payment of the difference will suffice to leave the prohibited party undisadvantaged. But if the independent has no such other financial resources, a protective agency may not pay him an amount less than the cost of its least expensive protective policy, and so leave him only the alternatives of being defenseless against the wrongs of its clients or having to work in the cash market to earn sufficient funds to total the premium on a policy. For this financially pressed prohibited individual, the agency must make up the difference between the monetary costs to him of the unprohibited activity and the amount necessary to purchase an overcoming or counterbalancing of the disadvantage imposed. The prohibitor must completely supply enough, in money or in kind, to overcome the disadvantages. No compensation need be provided to someone who would not be disadvantaged by buying protection for himself. For those of scanter resources, to whom the unprohibited activity had no monetary costs, the agency must provide the difference between the resources they can spare without disadvantage and the cost of protection. For someone for whom it had some monetary costs, the prohibitor must supply the additional monetary amount (over and above what they can spare without disadvantage) necessary to overcome the disadvantages. If the prohibitors compensate in kind, they may charge the financially pressed prohibited party for this, up to the monetary costs to him of his unprohibited activity provided this amount is not greater than the price of the good.11 As the only effective supplier, the dominant protective agency must offer in compensation the difference between its own fee and monetary costs to this prohibited party of self-help enforcement. It almost always will receive this amount back in partial payment for the purchase of a protection policy. It goes without saying that these dealings and prohibitions apply only to those using unreliable or unfair enforcement procedures.
Thus the dominant protective agency must supply the independents—that is, everyone it prohibits from self-help enforcement against its clients on the grounds that their procedures of enforcement are unreliable or unfair—with protective services against its clients; it may have to provide some persons services for a fee that is less than the price of these services. These persons may, of course, choose to refuse to pay the fee and so do without these compensatory services. If the dominant protective agency provides protective services in this way for independents, won’t this lead people to leave the agency in order to receive its services without paying? Not to any great extent, since compensation is paid only to those who would be disadvantaged by purchasing protection for themselves, and only in the amount that will equal the cost of an unfancy policy when added to the sum of the monetary costs of self-help protection plus whatever amount the person comfortably could pay. Furthermore, the agency protects these independents it compensates only against its own paying clients on whom the independents are forbidden to use self-help enforcement. The more free riders there are, the more desirable it is to be a client always protected by the agency. This factor, along with the others, acts to reduce the number of free riders and to move the equilibrium toward almost universal participation.
We set ourselves the task, in Chapter 3, of showing that the dominant protective association within a territory satisfied two crucial necessary conditions for being a state: that it had the requisite sort of monopoly over the use of force in the territory, and that it protected the rights of everyone in the territory, even if this universal protection could be provided only in a “redistributive” fashion. These very crucial facets of the state constituted the subject of the individualist anarchists’ condemnation of the state as immoral. We also set ourselves the task of showing that these monopoly and redistributive elements were themselves morally legitimate, of showing that the transition from a state of nature to an ultraminimal state (the monopoly element) was morally legitimate and violated no one’s rights and that the transition from an ultraminimal to a minimal state (the “redistributive” element) also was morally legitimate and violated no one’s rights.
A protective agency dominant in a territory does satisfy the two crucial necessary conditions for being a state. It is the only generally effective enforcer of a prohibition on others’ using unreliable enforcement procedures (calling them as it sees them), and it oversees these procedures. And the agency protects those nonclients in its territory whom it prohibits from using self-help enforcement procedures on its clients, in their dealings with its clients, even if such protection must be financed (in apparent redistributive fashion) by its clients. It is morally required to do this by the principle of compensation, which requires those who act in self-protection in order to increase their own security to compensate those they prohibit from doing risky acts which might actually have turned out to be harmless12 for the disadvantages imposed upon them.
We noted in beginning Chapter 3 that whether the provision of protective services for some by others was “redistributive” would depend upon the reasons for it. We now see that such provision need not be redistributive since it can be justified on other than redistributive grounds, namely, those provided in the principle of compensation. (Recall that “redistributive” applies to reasons for a practice or institution, and only elliptically and derivatively to the institution itself.) To sharpen this point, we can imagine that protective agencies offer two types of protection policies: those protecting clients against risky private enforcement of justice and those not doing so but protecting only against theft, assault, and so forth (provided these are not done in the course of private enforcement of justice). Since it is only with regard to those with the first type of policy that others are prohibited from privately enforcing justice, only they will be required to compensate the persons prohibited private enforcement for the disadvantages imposed upon them. The holders of only the second type of policy will not have to pay for the protection of others, there being nothing they have to compensate these others for. Since the reasons for wanting to be protected against private enforcement of justice are compelling, almost all who purchase protection will purchase this type of protection, despite its extra costs, and therefore will be involved in providing protection for the independents.
We have discharged our task of explaining how a state would arise from a state of nature without anyone’s rights being violated. The moral objections of the individualist anarchist to the minimal state are overcome. It is not an unjust imposition of a monopoly; the de facto monopoly grows by an invisible-hand process and by morally permissible means, without anyone’s rights being violated and without any claims being made to a special right that others do not possess. And requiring the clients of the de facto monopoly to pay for the protection of those they prohibit from self-help enforcement against them, far from being immoral, is morally required by the principle of compensation adumbrated in Chapter 4.
We canvassed, in Chapter 4, the possibility of forbidding people to perform acts if they lack the means to compensate others for possible harmful consequences of these acts or if they lack liability insurance to cover these consequences. Were such prohibition legitimate, according to the principle of compensation the persons prohibited would have to be compensated for the disadvantages imposed upon them, and they could use the compensatory payments to purchase the liability insurance! Only those disadvantaged by the prohibition would be compensated: namely, those who lack other resources they can shift (without disadvantaging sacrifice) to purchase the liability insurance. When these people spend their compensatory payments for liability insurance, we have what amounts to public provision of special liability insurance. It is provided to those unable to afford it and covers only those risky actions which fall under the principle of compensation—those actions which are legitimately prohibited when uncovered (provided disadvantages are compensated for), actions whose prohibition would seriously disadvantage persons. Providing such insurance almost certainly would be the least expensive way to compensate people who present only normal danger to others for the disadvantages of the prohibition. Since they then would be insured against the eventuation of certain of their risks to others, these actions then would not be prohibited to them. Thus we see how, if it were legitimate to prohibit some actions to those uncovered by liability insurance, and were this done, another apparent redistributive aspect of the state would enter by solid libertarian moral principles! (The exclamation point stands for my surprise.)
Does the dominant protective agency in a given geographical territory constitute the state of that territory? We have seen in Chapter 2 how the notion of a monopoly on the use of force is difficult to state precisely so that it does not fall before obvious counterexamples. This notion, as usually explained, cannot be used with any confidence to answer our question. We should accept a decision yielded by the precise wording of a definition in some text only if that definition had been devised for application to cases as complicated as ours and had stood up to tests against a range of such cases. No classification, in passing, by accident can answer our question in any useful manner.
Consider the following discursive description by an anthropologist:
The concentration of all physical force in the hands of the central authority is the primary function of the state and is its decisive characteristic. In order to make this clear, consider what may not be done under the state form of rule: no one in the society governed by the state may take another’s life, do him physical harm, touch his property, or damage his reputation save by permission of the state. The officers of the state have powers to take life, inflict corporal punishment, seize property as fine or by expropriation, and affect the standing and reputation of a member of the society.
This is not to say that in societies without the state one may take life with impunity. But in such societies (e.g., among Bushmen, Eskimo, and the tribes of central Australia) the central authority that protects the household against wrongdoers is nonexistent, weak, or sporadic, and it was applied among the Crow and other Indians of the western Plains only as situations arose. The household or the individual is protected in societies without the state by nonexplicit means, by total group participation in suppression of the wrongdoer, by temporarily or sporadically applied force that is no longer needed (and so no longer used) when the cause for its application is past. The state has means for the suppression of what the society considers to be wrongs or crimes: police, courts of law, prisons, institutions which explicitly and specifically function in this area of activity. Moreover, these institutions are stable within the frame of reference of the society, and permanent.
When the state was formed in ancient Russia, the ruling prince asserted the power to impose fines and to wreak physical pain and death, but allowed no one else to act thus. He asserted once again the monopolistic nature of the state power by withholding its power from any other person or body. If harm was done by one subject to another without the prince’s express permission, this was a wrong, and the wrongdoer was punished. Moreover, the prince’s power could only be explicitly delegated. The class of subject thus protected was thereby carefully defined, of course; by no means were all those within his realm so protected.
No one person or group can stand in place of the state; the state’s acts can only be performed directly or by express delegation. The state in delegating its power makes its delegate an agent (organ) of the state. Policemen, judges, jail guards derive their power to coerce, according to the rules of the society, directly from the central authority; so do the tax-collectors, the military, frontier guards, and the like. The authoritative function of the state rests on its command of these forces as its agents.13
The writer does not claim that the features he lists all are necessary features of the state; divergence in one feature would not serve to show that the dominant protective agency of a territory was not a state. Clearly the dominant agency has almost all of the features specified; and its enduring administrative structures, with full-time specialized personnel, make it diverge greatly—in the direction of a state—from what anthropologists call a stateless society. On the basis of the many writings like that quoted, one would call it a state.
It is plausible to conclude that the dominant protective association in a territory is its state, only for a territory of some size containing more than a few people. We do not claim that each person who, under anarchy, retains a monopoly on the use of force on his quarter acre of property is its state; nor are the only three inhabitants of an island one square block in size. It would be futile, and would serve no useful purpose, to attempt to specify conditions on the size of population and territory necessary for a state to exist. Also, we speak of cases where almost all of the people in the territory are clients of the dominant agency and where independents are in a subordinate power position in conflicts with the agency and its clients. (We have argued that this will occur.) Precisely what percentage must be clients and how subordinate the power position of the independents must be are more interesting questions, but concerning these I have nothing especially interesting to say.
One additional necessary condition for a state was extracted from the Weberian tradition by our discussion in Chapter 2: namely, that it claim to be the sole authorizer of violence. The dominant protective association makes no such claim. Having described the position of the dominant protective association, and having seen how closely it fits anthropologists’ notions, should we weaken the Weberian necessary condition so that it includes a de facto monopoly which is the territory’s sole effective judge over the permissibility of violence, having a right (to be sure, one had by all) to make judgments on the matter and to act on correct ones? The case is very strong for doing so, and it is wholly desirable and appropriate. We therefore conclude that the protective association dominant in a territory, as described, is a state. However, to remind the reader of our slight weakening of the Weberian condition, we occasionally shall refer to the dominant protective agency as “a statelike entity,” instead of simply as “a state.”
THE INVISIBLE-HAND EXPLANATION OF THE STATE
Have we provided an invisible-hand explanation (see Chapter 2) of the state’s arising within a state of nature; have we given an invisible-hand explanation of the state? The rights possessed by the state are already possessed by each individual in a state of nature. These rights, since they are already contained whole in the explanatory parts, are not provided an invisible-hand explanation. Nor have we provided an invisible-hand explanation of how the state acquires rights unique to it. This is fortunate; for since the state has no special rights, there is nothing of that sort to be explained.
We have explained how, without anyone having this in mind, the self-interested and rational actions of persons in a Lockean state of nature will lead to single protective agencies dominant over geographical territories; each territory will have either one dominant agency or a number of agencies federally affiliated so as to constitute, in essence, one. And we have explained how, without claiming to possess any rights uniquely, a protective agency dominant in a territory will occupy a unique position. Though each person has a right to act correctly to prohibit others from violating rights (including the right not to be punished unless shown to deserve it), only the dominant protective association will be able, without sanction, to enforce correctness as it sees it. Its power makes it the arbiter of correctness; it determines what, for purposes of punishment, counts as a breach of correctness. Our explanation does not assume or claim that might makes right. But might does make enforced prohibitions, even if no one thinks the mighty have a special entitlement to have realized in the world their own view of which prohibitions are correctly enforced.
Our explanation of this de facto monopoly is an invisible-hand explanation. If the state is an institution (1) that has the right to enforce rights, prohibit dangerous private enforcement of justice, pass upon such private procedures, and so forth, and (2) that effectively is the sole wielder within a geographical territory of the right in (1), then by offering an invisible-hand explanation of (2), though not of (1), we have partially explained in invisible-hand fashion the existence of the state. More precisely, we have partially explained in invisible-hand fashion the existence of the ultraminimal state. What is the explanation of how a minimal state arises? The dominant protective association with the monopoly element is morally required to compensate for the disadvantages it imposes upon those it prohibits from self-help activities against its clients. However, it actually might fail to provide this compensation. Those operating an ultraminimal state are morally required to transform it into a minimal state, but they might choose not to do so. We have assumed that generally people will do what they are morally required to do. Explaining how a state could arise from a state of nature without violating anyone’s rights refutes the principled objections of the anarchist. But one would feel more confidence if an explanation of how a state would arise from a state of nature also specified reasons why an ultraminimal state would be transformed into a minimal one, in addition to moral reasons, if it specified incentives for providing the compensation or the causes of its being provided in addition to people’s desire to do what they ought. We should note that even in the event that no nonmoral incentives or causes are found to be sufficient for the transition from an ultraminimal to a minimal state, and the explanation continues to lean heavily upon people’s moral motivations, it does not specify people’s objective as that of establishing a state. Instead, persons view themselves as providing particular other persons with compensation for particular prohibitions they have imposed upon them. The explanation remains an invisible-hand one.
* Must their calculation about which is better include their chances of success? There is some temptation to define this area of conflict as one where such chances of wrong are for certain purposes thought to be as bad as the wrong for sure. A theory of how probability interacts with the moral weight of wrongs is sorely needed.
In treating the question as one of whether the benefits of conflict outweigh its costs, the text seriously oversimplifies the issue. Instead of a simple cost-benefit principle, the correct principle requires for an act to be morally permissible, not merely that its moral benefits outweigh its moral costs, but that there is no other alternative action available with less moral cost, such that the additional moral cost of the contemplated action over the alternative outweighs its additional moral benefit. (For a detailed discussion of these issues see my “Moral Complications and Moral Structures,” Natural Law Forum, 1968, pp. 1–50, especially the discussion of Principle VII.) One would be in a position to advance the discussion of many issues if one combined such a principle with a theory of the moral weight of harms or wrongs with certain specified probabilities, to get an explicitly probabilified version of this principle. I mention only one application here that might not spring to mind. It is often assumed that the only pacifist position which is a moral position absolutely forbids violent action. Any pacifist position that considers the effectiveness of pacifist techniques is labeled tactical rather than moral. But if a pacifist holds that because certain techniques of significant effectiveness are available (civilian resistance, non-violent defense, satyagraha, and so on) it is morally wrong to wage or prepare for war, he is putting forth a comprehensible position that is a moral one, and which does require appeal to facts about the effectiveness of pacifist techniques. Given the lack of certainty about the effects of various actions (wars, pacifist techniques) the principle to govern the moral discussion of whether nonpacifist actions are morally permitted is a probabilified version of the principle (Principle VII) described briefly above.
* Clients no doubt would empower their agency to proceed as described in the text, if the client himself is unable to say whether he is guilty or innocent, perhaps because he is unconscious, agreeing to replace any compensating amount the agency must pay to the prospective punisher.
This deterrent to false pleas of innocence might act also to deter some innocent people against whom the evidence is overwhelming from protesting their innocence. There will be few such cases, but it may be to avoid this undesirable deterrence that a person who is found guilty beyond a reasonable doubt after having pleaded innocent is not also penalized for perjury.