Ch. 7, Rules and Justice

… it is in the laws of the Commonwealth as in the laws of gaming: whatsoever the gamesters all agree on is injustice to none of them.
*48

From that law of nature … there followeth a third; which is this that men perform their covenants made: without which, covenants are in vain and but empty words, and … we are still in the condition of war. And in this law of nature consisteth the fountain and origin of
justice. For where no covenant hath preceded … every man has right to everything, and consequently no action can be unjust. But when a covenant is made then to break it is unjust; and the definition of injustice is no other than the not performance of covenant. And whatsoever is not unjust is just.
*49

This formulation raises three questions, however. First, what is it that one’s agreement commits one to do? Second, what are we to understand by “voluntary” agreement, and how does the requirement that the agreement be voluntary affect obligation? Third, how broadly can we interpret “agreement” or “consent” for the purposes of this argument? Let us deal with these questions in turn.

In response to the first, we offer only a minor clarification. When one freely agrees to the rules, one promises to do rather more than pursue what is best for oneself under the terms of the agreement. Specifically, although the rules will typically include instructions as to how violations are to be handled and what punishments are to attach to such violations, and although these instructions are therefore contained within the inclusive agreement, it seems wrongheaded to say that agreement implies only a willingness to accept the defined punishment for violations. Consent is to the
rules, and the moral force of promise keeping is such that one is obligated to other players to play by those rules. To violate the rules may sometimes be personally profitable, but it will not be “just,” and it will not become “just” simply by virtue of one’s acceptance of punishment. “Just conduct” will consist in keeping one’s promises to other players, that is, in abiding by agreed-on rules. A player, for example, who punches another with his fist in American football concedes a fifteen-yard penalty. But he also endures the moral opprobrium of having committed an “unjust” act, and it is expected that this purely moral dimension—the player’s sense of justice—will carry weight in moderating his behavior.

It is important to make this point because, in some economists’ discussions of the law, one obtains the impression that choosing whether to abide by the rules is like selecting a drink at a soft-drink machine; that is, one either abides by the rules and pays no penalty or fails to abide by the rules and simply pays the price of so doing, as reflected in the rules. But the legislated punishment is not to be construed simply as the “price” of an alternative course of action; it also symbolizes the fact that a “wrong” has been committed. Making a choice among alternative drinks is a morally neutral act; choosing between legitimate and illegitimate modes of behavior is not morally neutral (at least not if the legitimacy springs from the prior consent of the chooser).

The second question is related to the circumstances under which an agreement may be said to be genuinely voluntary. Clearly, a promise to abide by the rules that is obtained under duress is not binding in the way that a freely offered promise is binding. The victim of a holdup who promises to pay a million dollars from his bank account in exchange for his life does not fail to meet a totally legitimate expectation when he subsequently refuses to pay. In the same way, a starving man who agrees to run inordinate risks for a crust of bread cannot necessarily be held to be bound to the agreement, although he is presumably more bound than he would be had he refused, despite his adversity, to agree. In other words, the moral force of an obligation to keep a promise is blunted (but not necessarily eliminated) if the circumstances surrounding the giving of the promise involve nonvoluntary elements.

We must be careful here not to extend the argument too far. To say, for example, that the agreement is nonvoluntary if the bargaining strengths of the parties to it are not precisely equal seems absurdly restrictive. It seems that only in cases of extreme duress or outright coercion does agreement to the rules
not morally bind the players. When A and B agree to marry, for example, we do not normally demand that their bargaining strengths or premarriage “threat strategies” or income positions or beauty or physical strength (or whatever each brings to the marriage) be equal (or even not unduly unequal) before the exchange of vows can be regarded as binding. We do not typically inquire whether A and B had other offers of marriage or could have had such offers. All that is required to give the “covenant” moral force is the absence of extreme duress. Indeed, even the presence of the paternal shotgun is not normally construed to remove the moral obligations entailed in the marriage entirely (and sometimes not to blunt those obligations at all).

It is worth making this point clearly so as to guard against the prospect of admitting alien concepts of “justice” through the back door under the cover of the “voluntariness” constraint. In some literature, the “justice” of abiding by an agreement is made entirely contingent on the justice of the status quo, the latter notion of “justice” usually making appeal to the relative income positions of the parties. We emphasize that the voluntariness of agreement is not to be so construed in our conception. It is only in those circumstances in which a promise would be held to be nonbinding (extracted by force or under conditions of total duplicity by one of the parties) that the requirements of justice, as we have defined it, would be waived.

The third general question raised by this conception of justice involves the issue of how broadly agreement or consent to the rules can be construed. Clearly, in most social contexts, players do not explicitly agree to the rules that apply to their interactions. Drivers do not, for example, construct the rules of the road by explicit consensus. How, then, can considerations of promise keeping be construed to apply in such cases? This question is central and long standing in any “social-contract” theory. Some scholars have sought to establish a presumptive obligation on the part of citizens to abide by the “rules” of a prevailing social order
*50 on the basis of what “would have been agreed to.”
*51 To the extent that such obligations can be established, considerations of justice apply. For such purely
hypothetical consent, however, the argument seems less than totally persuasive. It is not clear exactly how a person can be bound by promises he has not made, or how a person can be construed to have agreed to rules simply on the grounds that those rules can be presumed to make him better off.

Tacit, or
implicit, consent, however, is another story. Tacit consent can be construed to be given to rules of a game by participants when they voluntarily participate. The mere fact of participation obligates each participant, as if by an explicit promise, to abide by the rules, provided that the participants have a genuine option not to participate if they so choose. Failure to abide by the rules would then be to treat other participants “unjustly.” Participants’ expectations that others will play by the rules become “legitimate” by virtue of the voluntariness of participation by all players. If, for example, a person asked to join an existing poker game and was permitted entry, he would seem to be no less obligated by considerations of justice to obey the rules than the other players, who had given explicit consent.

A similar point can be made with respect to another important group of “participants” in the social order—those who
administer the rules. Again, we do not require the explicit agreement of judges, umpires, and enforcers on the structure of the rules in order to argue that these agents would be acting “unjustly” were they to violate or modify the rules in accordance with personal preferences. Those who administer a race presumably accept the rule that the first person across the finish line will win. Those who administer the legal system purport to be acting within the rules of that system. Those who enter the “game” in either case do so on the basis of an understanding of the rules, and in some sense the existence of rules amounts to an implicit promise made by administrators to all participants (actual and potential) that the rules will be administered faithfully.

If participation in a sporting event or parlor game—either as a player or as an administrator, judge, or enforcer—is sufficient to obligate all parties to abide by the rules, by virtue of considerations of justice can we not extrapolate such obligation to all social settings? Whether or not we can depends on whether individual participants in a given social order can be construed to be
voluntary participants. For administrators of the rules and for those who voluntarily join (free immigrants, for example), the obligation seems clearly applicable. For general citizens, voluntariness of participation is not so clear. The problem is, of course, that participation amounts to playing the “only game in town.” There may be no effective alternatives. Even in this case, however, there is some sense in which violation of the rules is “unjust” to other participants. Just what this sense is deserves some attention. Suppose that A, B, …,
N are participants in a game the rules for which have been decided unilaterally by Z. Suppose, furthermore, that these rules have traditionally been observed—that A, B, …,
N have played by them for some time. If A were now to inflict harm on B by breaking the rules, would we not say that B did not
deserve to be harmed? Would not A’s harming B be unjust in this sense? Would not B’s
reasonable expectation that A would continue to hold by the rules be a legitimate basis for B’s conduct? Does not the mere fact that such rules have prevailed for a long time contribute to the legitimacy of B’s expectations? If so, rules may be considered to be given tacit consent simply by virtue of their history or regular observance—even if there is no effective option to not playing and participation is involuntary in that sense. Of course, such tacit consent as a source of obligation to abide by existing rules does not imply quiescence toward efforts to change the rules. But this is a matter we shall discuss below.

Where does all this leave us with respect to the relation between rules and justice? It indicates that prevailing rules, simply by virtue of their existence, project an aura of justice. Behavior that contravenes prevailing rules amounts to unjust treatment of other participants in the social arena, because the others have legitimate expectations that all persons will behave in accordance with the rules. The legitimacy, not the reasonableness, of the expectations is crucial, and this legitimacy arises because, and to the extent that, participation in the activities governed by the rules is, or can be construed to be, voluntary. Voluntary participation amounts to agreement to the rules. It constitutes a tacit promise to abide by prevailing rules, and the breaking of such a promise is equivalent to unjust conduct because it involves treating others in ways in which they do not deserve to be treated.

IV. Justice among Rules

Although the discussion of just conduct set out in the preceding two sections has some affinities with notions in contemporary social philosophy, it is not the most common approach to “justice.” Typically, justice is introduced as a criterion for
evaluating alternative rules or institutions. This criterial usage of justice characterizes many strands of normative political and social theory—from natural law to modern theses advancing the norms for “distributive” and “social” justice.
*52 It is, in this sense, much more common to think of justice constraining rules than to think of rules constraining justice.

In this section, we advance the proposition that the business of deciding what rules are just—that is, of deriving a meaning for justice as a criterion of choice among rules—is a particular instance of decision making within prevailing rules. Such an argument requires the recognition that the decision as to what the rules shall be is itself made in the context of more abstract rules that apply to the choice among rules. We shall call rules at this more abstract level
meta-rules. Then it follows from our earlier discussion that rules are just if agreed-on meta-rules dictating their selection have been observed.

Although this is an entirely natural extension of the Hobbesian position, we here part company with Hobbes, at least in the strict sense of Hobbes’s exposition. In Hobbes’s view, the notion of an “unjust law” is meaningless. “No law,” he claims, “can be unjust. The law is made by the sovereign power, and all that is done by such power is warranted and owned by every one of the people; and that which every man will have so, no man can say is unjust.”
*53 But as Hobbes’s discussion makes clear, all this is contingent on individuals’ agreement on the principle of sovereign authority. It is the more abstract agreement on the structure of civil order that makes the law derived under that structure “just.” And so it is for us—for the essence of the constitutionalist approach is that political action (including the making of laws) be conducted according to certain rules (or meta-rules). A “just law” is then a meaningful concept; it is one that is derived under agreed-on institutions or, equivalently, one that does not violate agreed-on rules under which those institutions operate.

It is important to recognize that we are here weakening the requirements for just
conduct. Hitherto, conduct was described as just if it did not violate agreed-on rules related to that conduct. Now, conduct is also considered to be just if it does not violate just rules (defined as rules emerging under agreed-on meta-rules). Since the meta-rules in question may not
require agreement on rules directly, an “agreed-on rule” and a “just rule” are distinct concepts. We may not require a rule to be agreed on for it to be binding, provided that the rule in question emerged under agreed-on meta-rules.

The distinction between agreed-on rules and just rules—the recognition that agreement can be applied at different levels of abstraction—somewhat complicates the discussion of the requirements of justice. This is an issue we shall attempt to clarify in Section V. At this point, we shall address a different issue, which is related to the connection between abstraction and voluntariness. That is, since justice within rules is contingent on the rules having been
voluntarily agreed to, we must make some sense of the notion of voluntary agreement at the more abstract, meta-rule level. What does it mean to refer to a meta-rule as the subject of voluntary consent? What, for example, would involuntary consent involve at that meta-rule level?

In fact, the shift to more and more abstract levels of discourse serves to modify concern about the voluntariness of agreement—at least in one major sense. One can, of course, in principle imagine direct physical coercion being applied by one agent to others at the meta-rule level. But one has to ask whether the application of force is at all likely. At the extreme level, where the thick Rawlsian veil of ignorance is drawn and individuals are totally uncertain as to their future positions, individuals have no interests to defend. Any reason that any one of them has for preferring one set of rules over another will be a reason for all others to prefer that set of rules as well. In what sense would, or could, any individual find reason to coerce others?

Moreover, we must bear in mind that the agents at this abstract level
wish to secure agreements voluntarily, and for good reason: It is the capacity of such voluntary agreements to establish moral obligations that drives the whole “constitutional exercise.” Agents desire a stable institutional order so that they will have an appropriate context in which to pursue their (imagined) future life plans. If a
forced agreement will not serve to legitimate the rules, it will not establish a moral obligation to abide by the rules; so forceful extraction of agreement has no point.

Furthermore, reference to extreme need, and more general concerns about relative positions in the constitutional status quo that might be taken to offend the “voluntariness” requirement, also seems irrelevant at this most abstract level. If individuals are totally ignorant as to their future positions, they have no separately identifiable interests; there is a fundamental equality of position. It seems impossible that agreements reached in such a context could reflect unacceptable differences in status quo positions. Accordingly, however exactly one might wish to express the requirements for voluntariness in agreements reached, all such requirements appear to be met to an increasing extent as one moves to higher and higher levels of abstraction in the rule formation exercise. This fact represents a possible major justification for the whole “constitutionalist” approach.

V. Just Rules, Agreed-on Rules, and Just Conduct

The central argument in the previous section was that justice among rules is simply a case of justice within rules once removed. By implication, justice among meta-rules is a matter of justice within meta-meta-rules, and so on. In this sense, justice within rules becomes the paradigmatic case. Any appeal to considerations of justice in an interaction is an appeal to agreements made with respect to that interaction or to agreements made with respect to the rules under which that interaction takes place.

This line of reasoning requires us to clarify the distinction between rules that are “just” in the sense defined (that is, rules that emerge from voluntarily agreed-on procedures for determining rules) and rules that are agreed on. We must also point out that our paradigmatic notion of just conduct, or justice within the rules, entails an extension of our definition of just conduct to include conduct that is obedient both to agreed-on rules and to just rules, as defined.

To amplify what is at stake here, we shall offer a simple example. A tennis match of some significance is in its fifth and final set. Individual A is leading 5-4, and B is to serve. Suddenly, the umpire announces a change in the rules: The net is to be raised one foot. Could not B legitimately complain that this rule change is “unjust,” that he does not deserve to be treated in this way? or that A does not deserve an unanticipated advantage? Following the reasoning set out in the previous two sections, it would indeed seem that A and B are being treated unjustly. The rules to which they tacitly agreed when they decided to participate have been changed. They have not implicitly promised to obey the new rule, so they cannot be accused of unjust behavior if they violate it. On the contrary, the rule to which A implicitly agreed has been violated in A’s favor: B has been treated unjustly.

Yet such a conclusion may well be premature. We must ask whether the rule change in question has been made in accordance with the recognized meta-rules. Suppose, for example, that a properly constituted body, say, the World Tennis Federation, has decided to raise the net in the game of tennis by appropriate adherence to its own rules of procedure. To take the simplest case, suppose that the decision is made some time in advance and is to take effect at midnight on New Year’s Eve, 1985, and that the match between A and B is the final of the 1985 Australian Tennis Championship, held under World Tennis Federation rules and occurring on New Year’s Eve. The match has taken rather longer than expected, and at 4-5 in the fifth set with B to serve, midnight strikes. In this case, since A and B were aware of the prospective rule change and the possibility that the match would last until midnight, both A and B can reasonably be held to be bound by the new rules.

If, however, the World Tennis Federation did not announce the rule change in advance and the umpire was simply notified by cable of the rule change at midnight, then neither A nor B could be construed to be bound by the new rule by virtue of direct implicit agreement. But if the federation’s actions were procedurally “just”—that is, obedient to agreed-on meta-rules governing its behavior in determining the rules—then the rule change would not be a violation of justice. The prevailing meta-rules would have been faithfully applied. The point here is that A and B are to be seen as tacitly accepting not just the rules of the game but also the meta-rules—indeed, a whole sequence of rules of various degrees of abstraction. The mere existence of such meta-rules testifies to the prospect that rules may be changed, and any change obedient to these meta-rules is presumptively just. The prospect of rule change is something that players agree to face, much like changes in the weather. Unlike changes in the weather, however, rule changes are the result of human activity and hence come under the scrutiny of “justice.” The question at issue is whether rule makers, in changing the rules, violate
legitimate expectations, and the answer is that they do not, provided that they follow the relevant meta-rules. The requirement that no rule change ever occur is one possible meta-rule, but if it is not the prevailing meta-rule a blind following of the meta-rule forbidding rule change would be
unjust. If, for example, the umpire refused to alter the height of the net, in spite of a clear specification in the meta-rules that he do so, B would receive an undeserved advantage and A would lose an advantage he did not deserve to lose.

This is a matter of some account in political-economic contexts. Consider the case in which government action creates change in the income of a specified group by means of some regulation restricting entry into a particular industry. Now suppose that after a time government removes such protective regulation. Individuals will lose income, and the individuals who lose will not necessarily be those who gained in the first place. Those who purchase New York taxi medallions and stand to lose if the number of medallions is increased are not necessarily those who benefited from the introduction of medallions. The case is sometimes made in this connection that the government action is “unjust,” either in removing the regulation or perhaps in introducing the regulation in the first place. Similarly, the charge is occasionally made that when governments do pay compensation to the victims of policy changes, such action is unjust to taxpayers who must foot the bill.

In the justice within rules perspective, however, none of these claims is necessarily valid. Provided that government action in each case is itself obedient to the prevailing and accepted rules governing public-sector conduct, there is no violation of precepts of justice. As a matter of fact, modern political arrangements seem to be such that those who allocate public revenues can allocate them more or less as they choose or as political considerations suggest, in which case, the government simply cannot in such contexts violate principles of justice. To argue that government action is unjust is to claim that it violates agreed-on rules of public conduct. That is simply the way arguments about justice must properly proceed.

Now, it may be that the rules are unclear, as they often are in social contexts. There is some question, for example, as to whether regulation may in some situations amount to a “taking” and hence violate existing constitutional rules. There is also some question as to whether judicial interpretation may, in some cases, amount to a change in the rules and, in this sense, raise the issue of what role the courts properly exercise in the whole institutional order. Furthermore, certain rules may overlap with others at the same level of abstraction and indicate rather different behavioral restrictions. In all such cases, justice requires clarification of the rules and reconciliation of ambiguities. One major clarification along these lines involves the recognition of the level of abstraction at which agreement is to be applied. But the
grammar of arguments about justice at least seems clear.

One other possible ambiguity should be addressed. What are we to make of the status of agreements made outside the structure of rules? Clearly, when the set of individuals party to the agreement includes all the individuals who compose the political order, such agreements cannot conceivably violate precepts of justice. But when these sets of people are different, such agreements may be unjust. That is to say, the agreed-on rule structure may preclude certain sorts of agreements.

It is important to make this point in order to clarify the status of simple two-person market exchange in our conception. The parties to that exchange agree voluntarily to its terms. If one party subsequently fails to live up to those terms, he is presumptively acting unjustly toward the other party. But this presumption of unjust conduct can only be that—for it may be that the more inclusive set of constitutional rules explicitly precludes small-number voluntary transactions of various sorts or otherwise establishes obligations that conflict with those imposed in small-number transactions. Consider, for example, the case in which the firms in an industry form a cartel by means of an agreement, voluntarily entered into, the object of which is to restrict entry into the industry, reduce output, and raise prices. Suppose, now, that one of the parties to that agreement reneges on the terms, cuts prices, and makes substantial profits at the other firms’ expense. Is that cartel breaker acting “unjustly” with respect to other firms in the industry? Surely yes. Yet in doing so, the firm is also ceasing to act unjustly with respect to consumers of the firm’s product, if the more inclusive set of constitutional rules constrains restrictions on cartel deals. Considerations of justice here conflict, precisely as they would if A made separate agreements with B and C that for unexpected reasons turned out to be in conflict. But whereas in the case of ordinary agreements that conflict, considerations of justice do not indicate which agreement should be decisive, there is a distinct presumption in the cartel case that the constitutional rules should be decisive. In this sense, some amplification of Hobbes’s characterization is in order. Although it remains true that “whatsoever the gamesters all agree on is injustice to none of
them” (emphasis added), this does not rule out the possibility that the gamesters are acting “unjustly” in a more global sense. Whether or not that is so will depend, however, on the same basic principle—namely, whether the gamesters have prior, more inclusive agreements with other agents that are, or ought to be, overriding.

VI. Conclusions

In this chapter, our aim has been to offer an account of the nature of justice that spills naturally out of our constitutionalist-contractarian perspective. We have specifically defined justice in terms of conduct that does not violate agreed-on rules. One major implication of this approach is that justice is not a “primary” concept. Rather, it is derived from two logically prior notions: first, that agreements carry moral obligations to abide by the terms of those agreements; and second, that appeals to justice take place within an institutional context that serves to assign justice its meaning.

The approach also suggests a grammar of “justice” arguments. In particular, appeals to justice are appeals to matters of fact, as well as value. What are the rules that govern particular areas of conduct? Were those rules agreed to, or can they be construed to have been agreed to? These are questions of fact. There remains, of course, the value question of the extent to which agreements should bind the agreers. But this issue cannot, as a general matter of principle, be decided by considerations of justice. Justice becomes relevant only when the force of voluntary agreement is taken as given.

We should finally emphasize the restrictive purpose of this chapter. Our central aim has been to relate the notion of justice to the rules of social order. The nature of the argument, however, has made it necessary to skirt perilously close to areas of deep philosophical inquiry for which our own claims to competence are, at best, marginal. We should therefore make clear the absence of any putative claim to have made a positive contribution to the analysis of justice itself.

The central point is important, however, to the overall thrust of our argument and, in our view, to a proper understanding of the concept of justice. Any discussion of justice, whether by learned philosophers or common people, takes for granted that the interaction of persons in society occurs within a structure of rules. Appeals to the considerations of justice are appeals to those rules. In this sense, once justice is deemed a desirable attribute of conduct or of social order, there is established an indirect “reason of rules”—and an independent demand for a constitutionalist or rule-focused perspective on social life.