A. The priority of liberty

Any examination of Rawls's central political views must commence with A Theory of Justice (1971),¹

It is followed by the second principle, in which economic inequalities are countenanced just so long as they work to the advantage of all.

Noteworthy is the maximalism of the initial statement of the first principle along the three dimensions of person, quantity, and kind. To the question, For which citizens is liberty to be secured? the answer is: All of them. To the question, How much liberty are they to enjoy? the answer is: As much as can be achieved. And to the question, Which liberties matter? the suggested answer is: They all do. These maximalisms display Rawls as the legitimate heir to the classical liberal tradition running from John Locke and, especially, Immanuel Kant,²

It cannot be denied on Twin Earth, where, as will be exhibited below, the first principle serves as the centerpiece of Twin Rawls's libertarianism.³

Conspicuously absent from this catalog are economic liberties, including freedom of contract to buy and sell, to employ and be employed, or to accumulate and invest.⁴

When Rawls turns (in Section 42 of A Theory of Justice) to comparison of economic systems, he observes that market-based economies possess several advantages over command systems. Markets are less administratively cumbersome for distribution of consumption goods, and they more efficiently respond to the preferences of households (TJ, 270). Markets afford scope for individuals to move between jobs in response to changes in relative wage levels; they also decentralize economic power (TJ, 272). But if this should seem to provide a strong rationale for endorsing private property regimes, Rawls demurs. “It is evident … that there is no essential tie between the use of free markets and private ownership of the instruments of production” (TJ, 271), he affirms, adding that “[w]hich of these systems and the many intermediate forms most fully answers to the requirements of justice cannot, I think, be determined in advance” (TJ, 274). That is because the appropriateness for a given society of a set of economic institutions will depend on particular facts about its history and circumstances. The theory of justice is neutral between socialist and individual ownership of productive resources.

When some two decades later Rawls rethinks these matters for the revised edition of A Theory of Justice (1999), the conclusion is unchanged:

In another respect, though, Rawls's ideas about the priority of liberty have shifted significantly. Under the influence of H. L. A. Hart's critique,⁵

These liberties are specified by a list that is then defended against the charge of being makeshift via an argument attempting to show that these are the sort of things that contractors in the original position might opt for so as to give full effect to their two moral powers of cleaving to a conception of the good and exercising a sense of justice.⁷

Throughout the evolution of the theory of justice as fairness, Rawls maintains the priority of a liberty principle, yet, as the preceding sketch indicates, the content of protected liberties becomes increasingly modest in successive iterations. Rawls deems this a theoretical improvement, but Twin Rawls draws the opposite conclusion. To understand the crux of their disagreement it will be useful to turn to the rationales given for the priority of liberty (either in a maximalist sense or through a list of specified liberties) over other goods that can be advanced via political means.

One rationale is the importance to individuals of standing one to another in a relationship of equal citizenship. In a well-ordered society, self-respect endures disparities in wealth because these are justified by reference to the position of the least well-off member. To be subordinate to others as a participating member of a social union is, however, demeaning. It is to possess less than a full share of the social bases of self-respect. That is why inequality of basic liberties, as opposed to inequalities of wealth, would be rejected by contractors in the original position.¹⁰

There is, however, a second rationale for the priority of liberty, one directly responsive to persons' capacity to formulate and pursue particular conceptions of the good. What citizens can be presumed to possess in common is a commitment to acting justly toward each other, where the duties of justice bear alike on everyone. But no such commonality characterizes the ends to which they direct themselves, the advancement of which confers meaning on their lives and which contribute no less importantly than does equal citizenship to their enjoyment of self-respect. A just order is one in which individuals, motivated by goods that they take to be of compelling, even transcendent value, nonetheless respect the rights of others to show themselves indifferent to these goods while instead pursuing other ends that may seem to be of negligible worth.¹¹

Rawls observes that under favorable conditions in which necessities of survival are reliably satisfied, “the obstacles to the exercise of the equal liberties decline and a growing insistence upon the right to pursue our spiritual and cultural interests asserts itself. Increasingly it becomes more important to secure the free internal life of the various communities of interests in which persons and groups seek to achieve, in modes of social union consistent with equal liberty, the ends and excellences to which they are drawn” (TJ, 543). With regard to these ends and excellences, the specifically political liberties enjoy no distinctive pride of place—or, rather, they do not for those individuals whose pursuit of the good life is primarily conducted in venues outside the public arena. What those individuals need from others is generalized noninterference, and because in the original position they do not know which particular ends will command their allegiance, they have reason to value a wider rather than narrower scope of liberty. Thus, they will regard the latter versions of the first principle as insufficiently responsive to whatever interests they will find themselves to have that lie beyond the practice of citizenship as such. Instead, they will resist shrinkage of the original guarantee of maximal equal liberty. They will understand this liberty in a wide sense, even if they do not arrive at one definitive account of where it is at its most expansive. That is, Hart's critique may lead them to agree that no maximizing function for liberties is derivable a priori. They will, however, nonetheless insist that we often have good reason between competing alternatives to judge that one is more favorable than the other to the exercise of unimpeded self-determination. In particular, they are very likely to hold that both theory and historical practice amply demonstrate that liberty is enhanced by an order of private ownership of productive assets rather than collective control by state bureaucracies. If so, they will be in accord with Twin Rawls, who shares with our Rawls a commitment to the priority of liberty, but who understands that principle to commend an order of generalized noninterference rather than one in which political freedoms are privileged.

B. Strains of commitment

Deliberators in the original position decline to adopt the principle of (maximizing expected) average utility, preferring instead the maximin strategy (of securing the best possible minimum) incorporated in the difference principle. The choice is overdetermined. It is stimulated in part by difficulties of ascertaining behind the veil of ignorance any probabilities concerning which social position they will occupy, and in part by an aversion to gambles that put them at risk of falling below a satisfactory minimum. Both of these prongs of the Rawlsian argument are controversial and have generated numerous responses. They will not be examined further here.

There is, however, a third prong of the argument that is substantially independent of the other two and that carries considerable conviction in its own right. Unless individuals will be regularly and reliably motivated to act in accord with chosen principles, at least under favorable conditions, those principles are unsatisfactory as the framework for a society's basic structure. The problem with utilitarianism is, simply put, that it demands too much of individuals. It requires them to acquiesce in the sacrifice of much, perhaps all, of that which they value for the sake of maximizing a putative social good. One who assents to a utilitarian strategy because it holds forth the prospect of a greater expected return for self than does any of its competitors is apt to find himself unable or unwilling to accede to its demands should he turn out to be unfortunate in the lottery of life's events. And even if one should prove to be among the fortunate beneficiaries of the utility principle, this good luck could be undone by the defections of less fortunate others. Thus, a rational individual alive to the risks of instability will be dissuaded from endorsing in the original position the principle of average utility. The integrity of society is unnecessarily imperiled by erecting its institutions on foundations from which individuals will predictably tend to defect. By way of contrast, justice as fairness rejects extreme sacrifice. It guarantees to each a decent minimum, indeed the most favorable minimum that can be secured by any social design. For that reason, invoking the Rawlsian idiom, it is burdened by lesser “strains of commitment” than is utilitarianism.¹²

Although Rawls does not explicitly draw the connection, concern for keeping the strains of commitment low for all parties supports the priority of the liberty principle. To be left alone to serve those ends that hold out surpassing value for oneself, rather than being required to drop them for the sake of what is proclaimed to be a superior social product, is mandatory for those whose personal projects hold out great significance. Such individuals could not rationally accede to a system in which it is likely that they will be forcibly separated from that which they hold dear. The strains of commitment counsel against endorsement of such potentially onerous demands.

One must ask, though, which liberty principle will find favor with contractors intent to specify a basic structure that they will subsequently find congenial to their moral powers? The Rawlsian “fully adequate scheme” of civil freedoms answers in some measure to the concern that one will not be involuntarily divorced from one's ends. To be in possession of a full range of political freedoms renders one well-provisioned in a democratic environment to campaign actively on their behalf. However, the assurance thereby provided is slim. One's potential opponents are equally graced with democratic freedoms, and if at the end of the contest they should command a majority, then it will not necessarily be much consolation to have had one's day at the polls. If the subjectively measured costs of the required sacrifice are high, then even those who possess a firm sense of justice may find themselves reluctant to go along. So the same considerations that lead Rawls to endorse the two principles over average utilitarianism support construing the liberty principle broadly, as broadly as it presents itself in its initial maximizing formulation.

There are further reasons to believe that strains of commitment cut against the official Rawlsian statement of justice as fairness. Rawls has argued that the difference principle meliorates those strains (against the utilitarian alternative) by rendering the losers in life's lottery as well-off as they can be. That may be true, but it leaves open questions about the impact of those strains on other strata of the population. Because individuals are represented as not only rational but also reasonable, they are motivated by a concern for reciprocity (see PL, 17–18). They are prepared to forgo some measure of advantage in order to ensure gains for their fellows. They are not, then, purely rational egoists. Neither, though, are they perfect impartialists. Between their own preferred ends and those of others, they are not neutral. The more they are required to forgo the former so as to advance the latter, the greater the strains of commitment under which they labor. It would seem, then, that the same sort of concern for social stability that spoke against utilitarianism will also pronounce against unchecked requirements of redistribution. A theorist has done only half the job by ascertaining the view from the bottom; it is requisite also to see how things present themselves from a top-down perspective. From that vantage, the nature of the impositions potentially placed on the more well-off are arguably more than they will be prepared to bear.

The strict logic of the difference principle entails that for the sake of even very small enhancements to the least well-off, their better-advantaged fellows must be prepared to bear enormous costs, limited only by the crossover point at which they would themselves assume the position of society's least well-off. “[I]t seems extraordinary that the justice of increasing the expectations of the better placed by a billion dollars, say, should turn on whether the prospects of the least favored increase or decrease by a penny” (TJ, 57). Can the demands of justice impose so stringently on the more advantaged, and if they do, will not the rigor of the system generate strains of commitment that only moral saints and heroes will find themselves able to abide?

Rawls's immediate answer is that “the difference principle is not intended to apply to such abstract possibilities” (TJ, 157). Rather, it is incorporated in a basic structure designed for the most part to create a tide that raises and lowers boats together.¹³

Rawls's most revealing and candid response to this worry is supplied in Section 17, “The Tendency to Equality.” In response to the fact that people come into society with vastly different abilities and vastly different prospects flowing therefrom, he observes:

To this the indicated response is: Exactly! Bonds uniting citizens in the Rawlsian conception reveal themselves to be extraordinarily tight. It is no small thing to agree to share one's fate with another. That is the sort of undertaking embarked on within a family, by friends or lovers determined to pursue the good life together, partners in a far-reaching enterprise, a platoon's soldiers guarding each other's backs, congregants joined in common worship, or devotees of a mutually adored good. In such settings, the success of one constitutes in no small measure the success of all, and an individual who is not prepared to sacrifice some of her own quota of the good, indeed sacrifice liberally, is an anomalous partner in the pursuit. And of course it is not uncommon for the strains of commitment inherent in such intimate relations to rupture the bonds that formerly held the parties close.¹⁴

It is quite otherwise when the tie among persons is nothing more than a citizenship held in common. Here intimacy is very much the exception rather than the rule. The comprehensive theories of one's compatriots will often incorporate commitments and ideas of the good that leave one unmoved—or worse. In such cases one is obliged to respect their liberty to follow their own fervid musings or strange gods, but it is to demand too much to add to the standing order of noninterference the further requirement that positive assistance be afforded—that, in a word, one join one's fate to theirs. The two principles of justice are seen in such instances to be in direct tension with each other. The liberty principle allows individuals to devote themselves to their own preferred conceptions of the good and thus to distance themselves from others who acknowledge ends that are indifferent or antithetical to their own. (Or rather, such permission is afforded by a liberty principle that extends beyond a basketful of privileges confined to the political arena.) The difference principle binds individuals together. Something has to give.

And so it does. Rawls has second thoughts concerning the intensity of social bonds under a justice as fairness regime. In the revised edition of A Theory of Justice, the striking passage about individuals sharing each other's fate disappears.¹⁵

It is not clear that such revision by omission is satisfactory. The reader may suspect that it camouflages rather than eliminates a fundamental problem with the project. The Rawls of 1971 must have believed that the bonds of social solidarity to be imputed to citizens in a well-ordered regime based on the two principles of justice are substantial enough to make it sensible, if perhaps edging toward the far end of the metaphorically appropriate, to use the language of sharing one another's fate. Either he had misjudged then concerning how much can legitimately be spun out from contractors' posited sense of justice, or by the time he prepared the revised edition he had devised a different strategy to explain why the strains of commitment embedded in the difference principle are not too onerous for individuals reliably to bear. Unfortunately, no explanation of the transition is offered in the 1999 edition. An author may, of course, admit to prior missteps. Such frankness is commendable. Indeed, in this spirit of generosity Rawls has often acknowledged the force of his critics' remarks and revised accordingly. The problem here, however, is that the withdrawn passage is, arguably, no misstatement at all. Rather, it serves to explain an otherwise puzzling feature of the Rawlsian account. There is no mystery why the least well-off individuals find justice as fairness agreeable; under no alternative basic structure for society will they do better. The more recalcitrant problem is to explain why the more well-off accede to a system in which they demonstrably are less well-served than they would be under other social arrangements. Will they not be inclined to regard themselves as unjustifiably exploited? The answer that immediately suggests itself is this: Not if they take themselves in some very strong sense to be implicated in the fate of their less favored brethren. Absent this understanding, it is doubtful that justice as fairness can survive the strains of commitment.

Exactly this line of reasoning can be seen in the celebrated works of Twin Rawls. He, too, determines that the difference principle expresses a profound depth of social solidarity. But for Twin Rawls, this amounts to an utterly convincing reductio ad absurdum of the so-called second principle of justice. In large and diverse liberal societies where persons are spatially spread out from each other, espouse markedly different religious and philosophical comprehensive theories, and differ no less substantially with regard to their personal affections than in their various conceptions of the good, it is outlandish to suppose that they can regard their own success or failure as tightly bound up with the achievements of distant others. Because they have a sense of justice, they will forswear aggressive interference against others for whom friendly regard may be limited or altogether lacking. And because they are prudent, they will insist on reciprocal noninterference from others. That is not quite all that Twin Rawls includes in his theory of justice (see Section III below), but it is the dominant strand. It is what makes him Twin Harvard's most illustrious libertarian.

That Twin Rawls zigs where Rawls zags does not constitute disproof of the latter's theory. Perhaps it is our Rawls who has gotten things (more nearly) right. It is impossible decisively to defend one Rawls over the other without performing a more thorough excavation of the theory of justice than is possible here. Instead, I offer an ad hominem argument against our Rawls's understanding. Although arguments ad hominem normally carry a somewhat unsavory flavor, in this case that may be mitigated by the fact that the one arguing against Rawls's understanding is Rawls himself. That argument is initiated in Political Liberalism (1993), where he does not abandon his support of the difference principle but concedes that it is one among several ways of reasonably distributing the benefits and burdens of social cooperation. But in The Law of Peoples (1999), Rawls distances himself further from the difference principle by denying, against the urging of his own disciples,¹⁶

C. Choice behind the veil of ignorance

Perhaps the most distinctive aspect of Rawls's methodology is his casting the problem of eliciting basic principles of justice as a choice-theoretic determination behind a suitably specified veil of ignorance. People are stripped of individuating knowledge that would allow them to tailor principles to suit their own particular situations. Deprived of such data, they rationally direct themselves toward principles under which they will do comparatively well, regardless of who they turn out to be or how favorable the circumstances they confront. Thus, they legislate to safeguard their basic liberties—the first principle—and to ensure an economic status that is at least minimally satisfactory for enabling them to live as ends-pursuers—the second principle. I argued in Section IIA that Rawls backs away from a rendering of the liberty principle that adequately serves the former aim. To that result can now be added equally cogent reasons to maintain that the second principle's invitation to engineer a massively redistributivist welfare apparatus will be declined in the original position. Rather, strong property rights, including robust freedom of contract, will be endorsed, modified only by guaranteed provision of a minimally decent floor beneath which individuals will not be allowed to fall: a “social safety net.”¹⁸

Although contractors in the original position are screened from individuating knowledge, they have access to general facts, including those of social theory (TJ, 142). That theory is not, of course, a completed science: oases of relative clarity are surrounded by expanses of obscurity, and despite a generally progressive course, the theory dead-ends into occasional cul-de-sacs and regressions. Nonetheless, intelligent design of social institutions, not least a society's basic structure, entails judicious use of the best approximations to knowledge that we have while simultaneously guarding against being impaled on unacknowledged spikes of ignorance. So, for example, any tolerably fair and efficient design will build in a recognition that “the point of the institution of property is that, unless a definite agent is given responsibility for maintaining an asset and bears the loss for not doing so, that asset tends to deteriorate.” This was well known already by the contemporaries of Adam Smith, although this statement in fact comes from Rawls (LP, 39). I believe that by the time of the original drafting of A Theory of Justice, enough was known about the equity and efficiency properties of collective ownership of the means of production to decisively rule out socialism as a feature of a society's basic structure. Three decades on and counting, this is beyond serious dispute. Any contemporary invocation of the original position will, therefore, make quick work of dispatching Marxist fancies to the dustbin of defunct social theories.

Are there other propositions about the working of economic/political systems that were not brought behind Rawls's original veil of ignorance but that now are well-attested presumptions without which the practice of intelligent social design cannot be conducted? One plausible candidate is this: There is no amount of self-seeking battening on the public fisc, of blithely robbing Peter to pay Paul, that cannot be given a semi-plausible justification suitable for thirty-second sound bites. Or, to phrase it more soberly, “[W]hether the aims of the principles covering social and economic inequalities are realized is … difficult to ascertain. These matters are nearly always open to wide differences of reasonable opinion; they rest on complicated inferences and intuitive judgments that require us to assess complex social and economic information about topics poorly understood” (PL, 229). If the rules of the political game allow for taking and then redistributing property just so long as a majority can be assembled to testify that they are acting for the sake of some lofty ideal of social justice—perhaps one that goes by the name of “justice as fairness”—then it is predictable that those rules will be frequently and extravagantly bent in the service of interests that are neither impartial nor likely to advance the positions of the least well-off. Knowing this, contractors in the original position will be loath to afford carte blanche to the proliferation of allegedly welfarist measures. They will realize that the point of the institution of property as noted above will be imperiled by such a loosening of rights to the continued possession of that which has been acquired through voluntary transactions. It is eminently knowable behind the veil of ignorance that rent-seeking, the investment of resources in the attempt to secure windfall returns through political subventions, is a negative-sum game. Moreover, it is a game that society's already well-advantaged are likely to be especially good at. The upshot is that even if the veiled deliberators are motivated by a strong concern for the least well-off (make that especially if they are motivated by a strong concern for the least well-off), the deliberators will erect constraints on the ability of political actors to redefine property rights in the service of some ostensible ideal.

Suppose that we judge it a good thing, all else being equal, that resources flow to their most economically valued use. One way this desideratum could be pursued is via a process of hearings and speeches by all parties who take an interest in the subject, followed by votes in the legislature on competing allocative proposals. Another way in which the desideratum can be pursued is to allow trades of rights assignments until an outcome is reached in which no parties are willing to exchange their holdings for some other set on offer (i.e., they have landed on some point on the Pareto frontier). It is a simplification but not an oversimplification to say that the most important conclusion of the first century of post–Adam Smith economics is that the second strategy pays off better than the first. Indirection triumphs over direct pursuit of the desideratum. The most important finding of its second century may be that the aims of social justice also are better pursued by an indirect strategy. This remains true even if one believes that how the upper strata of society fares is irrelevant to concerns of justice and that exclusive focus is to be directed to the least well-off. It is not paradoxical hyperbole to maintain that the prospects of the poor are enhanced by disallowing all special pleading in legislatures and regulatory bodies on behalf of the poor. Impartial enforcement of a regime of strong property rights and binding contracts almost certainly will better serve them.¹⁹

Admittedly, this is a first approximation. Before pulling back the veil and entering into the full light of political day, allowance will be made for collective provision of genuinely public goods, those for which features of nonexcludability and nonrivalry of consumption render market provision awkward. Additionally, it is appropriate to legislate a safety net for those who find themselves in exigent straits and who are unable to extricate themselves therefrom, either through their own actions or through the voluntary subventions of others. This latter condition is responsive to strains of commitment that otherwise might render the precepts of justice too onerous for the poor to embrace, while the positive-sum character of the remainder of economic transacting will lessen those strains on all. Twin Rawls, who enjoyed the benefits of a thorough grounding in the central insights of twentieth-century social theory, drew precisely this conclusion. It was not beyond the reach of the other Rawls. Therefore, it is worth considering why he, too, did not gravitate toward libertarianism.

A. Moral arbitrariness

Rawls uses the term “system of natural liberty” to characterize a social order in which individuals enjoy full and equal liberty, and in which they are free to capitalize on their talents and fortune so far as they are able (TJ, 59–60). This interpretation of the second principle of justice's requirement of mutual advantage treats all persons impartially by ensuring their rights to transact with willing others in their pursuit of advancement. It does not, however, take notice of their various starting points. But because those who have been well-favored by a lucky turn of the natural lottery's wheel (or in subsequent social engagements) will tend to capture more of society's goods than do those who have been less well-favored, consequent shares of the good things of life will not accurately reflect any antecedent merit. “Intuitively, the most obvious injustice of the system of natural liberty is that it permits distributive shares to be improperly influenced by these factors so arbitrary from a moral point of view” (TJ, 72).

This essentially completes the critique of libertarianism in A Theory of Justice. Little more could have been expected. The term “libertarian” was then barely recognizable in a political context, and the classical liberalism from which contemporary libertarianism springs had been sharply out of favor for more than a century. In an already very large book, it may have seemed excessive to devote more ink to what seemed a mere historical footnote. Accordingly, Rawls turns forthwith to other, more viable interpretations of the second principle of justice. Had he been conversant with the writings of Twin Rawls, however, he would have realized that this is to give too short shrift to the libertarian alternative. Even if it is granted that one's place in the natural lottery of talents and initial social position is morally arbitrary, it does not follow that a structure mostly dedicated to safeguarding the rights of individuals to act from those starting places is also morally arbitrary. The point from which one begins is neither here nor there, morally speaking, but that one proceeds therefrom to freely exercise one's two moral powers so as to advance favored ends while simultaneously affording due recognition and respect to other agents is laden with moral significance. Recall Rawls's remark: “The natural distribution is neither just nor unjust…. What is just and unjust is the way that institutions deal with these facts” (TJ, 102). There are two broadly contrasting ways in which the political institutions of a society can deal with these facts. One is through the adoption of policies and procedures contrived so as directly to bring about some favored distributive configuration. The other is through a system of law that reinforces the capacity of self-directing individuals to generate valued outcomes.²⁰

B. Libertarianism is not political

In lecture 7 of Political Liberalism, “The Basic Structure as Subject,” Rawls contrasts justice as fairness with Robert Nozick's libertarianism. There is, Rawls claims, no special role for the basic structure within libertarian theory. The state is on a par with private associations insofar as it emerges from a historical series of voluntary transactions with willing clients who are at liberty either to purchase the package of services the state offers at its stipulated selling price or to decline to deal. “By viewing the state as a private association the libertarian doctrine rejects the fundamental ideas of the contract theory, and so quite naturally it has no place for a special theory of justice for the basic structure” (PL, 265).

This critique is multiply problematic. First, it's not clear that it is a critique, rather than simply a characterization. What's wrong with applying principles of justice that hold for private transactions to the evolution of a minimal state? Isn't that a gain from the point of view of theoretical parsimony? Does it not make the birth of political structures seem a bit less ad hoc? Even if Nozickian transactors are removed from social contract understood as deliberation behind a suitably defined veil of ignorance, why should we take that to be a theoretical defect? It is not obvious that the actually agreeable carries less weight than the hypothetically agreeable.

Second, even if Rawls does nail Nozick dead to rights, that is not equivalent to having dispatched libertarianism. Perhaps the flaws exhibited in Nozick's theory—assuming that they are in fact flaws—are idiosyncratic. If something more than victory in a one-on-one skirmish is to be achieved, then Rawls must train his sights more broadly on the gamut of exponents of the “system of natural liberty.”

Third, Rawls grossly mischaracterizes Nozick's argument. Contra Rawls, it is not Nozick's view of the transition to the minimal state that “[n]o one can be compelled to enter into such an agreement and everyone always has the option of becoming an independent” (PL, 265). This is a remarkable inversion of the extended Nozickian argument for the permissibility of disallowing independent status and instead compelling individuals to enter into a civil order.²¹

C. Libertarianism is illiberal

Rawls maintains that reasonably just constitutional democratic societies assure “sufficient all-purpose means to enable all citizens to make intelligent and effective use of their freedoms.” Otherwise impeccable systems of formal liberties that fail to incorporate such a guarantee, he adds, “are an impoverished form of liberalism—indeed not liberalism at all but libertarianism. The latter does not combine liberty and equality in the way liberalism does; it lacks the criterion of reciprocity and allows excessive social and economic inequalities” (LP, 49).

It is not immediately evident why Rawls says that libertarianism fails the criterion of reciprocity. A libertarian order more than any other rejects forced transactions and the exploitation consequent thereon. Instead, it privileges voluntary exchange, the paradigm instance of mutual benefit and reciprocity. To make sense of Rawls's denial of libertarian reciprocity, he must be understood as maintaining that the freedom of transactors separated by large disparities of wealth and power is not genuine. Someone perched on the edge of exigency is obliged to take whatever bargains she can get, no matter how hard, no matter how divorced from her conception of the good life.

So interpreted, this becomes a variation on the strains-of-commitment theme. The least well-off individuals in a libertarian order will find the meager share they receive from the cooperative surplus scant return for their adherence to the rules of that order. They will regard its formal equality as imposing on them a far greater burdens-to-benefits ratio than is borne by the more well-off. Unlike the high-flyers of laissez faire, the least well-off have scant rational stake in the order's sustenance. It will be unreasonable to expect them to agree to adhere to a regime of rights in which their own rights are of little value.²³

If strains-of-commitment arguments are to be taken seriously—as Section IIB above argues they must be—then this is Rawls's most formidable anti-libertarian thrust. A libertarian order that allows individuals to fall through the cracks is not one to which all can reasonably commit their allegiance. But rather than constituting a decisive objection to libertarianism, this argument instead amounts to a criterion for choosing among libertarian programs. The decided majority of prominent libertarian proponents have maintained that the central principle of maximum equal liberty is to be accompanied by a contingent and secondary principle of publicly provided subventions for those who are unable through their own voluntary undertakings or the assistance of willing others to elevate themselves above the floor of exigency.²⁴