II. Consent vs. Minimalism

To begin, then, why this peculiar neglect by libertarians of the classical problem of political obligation? One possibility, of course, is that most libertarians simply never took seriously the claims of the state in the first place; another is that they have been too busy explaining and criticizing the excesses of bloated and tyrannical actual states to worry much about the more ideal conditions under which genuine political obligation and political legitimacy are possible. A third possibility is that many libertarians are simply presuming the truth of a consent theory of political obligation, without bothering to state this presumption (perhaps because they take it as too obvious to state). David Boaz of the Cato Institute, for instance, in his book Libertarianism: A Primer, without ever explicitly mentioning the problem of the source of our special obligations to governments or fellow citizens, at one point simply defines government as “the consensual organization by which we adjudicate disputes, defend our rights, and provide for certain common needs.”⁵

Here it is plain that Boaz really intends either to define only legitimate government or else by fiat to deny the title of “government” to apparent governments that are nonconsensual organizations. Either way, his remarks appear to simply presume consent theory standards for governmental legitimacy and political obligation.

What is striking about Boaz's pronouncement—aside from how much is left unsaid—is that it might seem to be in reasonably obvious tension with Boaz's earlier contention that the sole proper “role of government is to protect people's rights.”⁶

If citizens unanimously consent to governmental institutions, of course, those citizens will have consensual rights against one another that the institutions not be undermined and that their rules be uniformly enforced (and they will also have obligations to comply with governmental rules). So there is nothing odd on its face about a libertarian asserting both that legitimate governments are confined to protecting rights and that their legitimate powers are consensual in origin. But the tension between these two theses is more apparent if one attends to which rights Boaz thinks it proper for governments to be protecting. Like any right-libertarian, Boaz has in mind a relatively limited role for government: the rights a legitimate government enforces are simply those implied by the “one fundamental human right”—that is, “the right to live your life as you choose so long as you don't infringe on the equal rights of others.”⁷ This fundamental right, he thinks, implies (or perhaps is identical to) a natural right of self-ownership, which in turn implies rights to be free of interference by others in one's performance of innocent actions, including one's making, keeping, and using property and one's making and having honored contracts with others.⁸ Thus, there seem to be many rights that legitimate governments must not curtail, apparently even if such curtailments are consented to.

The tension to which I am pointing, then, is this: On the one hand, libertarians tend to believe that we enjoy virtually absolute freedom of contract, so that we can legitimate by our unanimous consent (and undertake obligations to uphold and comply with) virtually any kind of arrangement, organization, or institutional structure. This consensualist strain of libertarian thought thus appears to imply that virtually any kind of government or state can, in principle, be legitimate (and owed political obligations), including, say, socialist governments. On the other hand, the minimalist strain of libertarianism implies that only reasonably limited governmental arrangements can be legitimate—those that preserve and defend extensive rights of individual liberty and self-ownership, those that refrain from forced redistribution of property, those that do not limit freedom of contract by making morally alienable rights legally inalienable (and so on). But, of course, if all rights are alienable, then the right to enjoy minimal political arrangements is alienable as well. It may be imprudent for us to agree to surrender our right to be free of extensive government; but it must be morally possible for us to do so, since a person's imprudence does not, for libertarians, limit her right to alienate rights (provided only that this imprudence does not result from her having been, for example, reduced by illness or temporary incapacitation to a nonagent, coerced by others, or denied by them information to which she is entitled).

The observation that this tension exists in libertarian thought is not new or exciting.⁹

But it does allow us to see, I think, why libertarians have tended to ignore the problem of political obligation. In their consensualist moments, libertarians can see consent theory standards of legitimacy and political obligation as no more than painfully obvious implications of the strong natural rights to freedom of contract and freedom from coercion that they embrace. As such, the implied theory of legitimacy and political obligation barely requires stating. And in their minimalist moments, a theory of political obligation may simply seem to libertarians to be altogether unnecessary. After all, if governments are in fact just performing their minimalist role of enforcing our basic natural rights, then it may seem that they are doing no more than requiring each of us to discharge obligations toward others (i.e., obligations not to violate their natural rights) that we had all along—obligations that are natural and that we would possess and could be legitimately required to discharge even if no political societies existed. No special theory of political obligation (owed to our government or to other persons qua fellow citizens) seems to be required to explain how we come to have enforceable obligations toward others when those obligations are logically pre-political and enforceable in nature. Government is simply taking on the enforcement task that private individuals would otherwise have to bear themselves.

III. Problems for Minimalists

This tension between the consensualist and the minimalist strains of libertarian thought—along with the attendant obscuring of questions about political obligation—appears centrally in the most widely known and discussed of contemporary libertarian works in political philosophy: Robert Nozick's Anarchy, State, and Utopia (1974). The minimalist strain in Nozick's defense of libertarian political philosophy is obvious, especially in Part 2 of the book: “The minimal state is the most extensive state that can be justified. Any state more extensive violates people's rights.” This minimal state is “the only morally legitimate state.”¹⁰

The consensualist strain may not be quite so prominently displayed by Nozick, but it is equally present. Part 1 of his book is an attempt to show how a state could and (within limits) would arise from a state of nature without anyone's rights being violated in the process. One might think that an easy solution to Nozick's problem would be this: all of the people within a territory could simply come together and unanimously consent to the authority of a state, removing with their consent any concerns that the state might be violating their rights. Nozick could not, I think, deny that this could happen, and that the result would be a legitimate state. But in Anarchy, State, and Utopia, he is plainly concerned to explain how a state could arise under slightly less fanciful or idealized conditions. So the story Nozick actually tells involves the creation by individuals in the state of nature of numerous, small voluntary mutual-protective associations, aimed at addressing their lack of security and lack of power to enforce their own rights.¹¹

What makes the demands of such associations on their clients morally legitimate is simply that each client has freely consented to participate in (or to pay a share of the cost of) the defense of the rights of all of the association's clients. Similarly, when competition for clients among such associations eventually produces a dominant protective association within a territory,¹² the legitimacy of its demands will also be grounded in its clients' consent.

Only in the final stage of Nozick's story in Part 1—in which a genuine minimal state at last arises when nonclient “independents” in the territory are forced to accept membership and are compensated for their lost rights¹³

—do legitimacy and obligation appear to turn on anything other than free personal consent. But, of course, the “principle of compensation”¹⁴ by which Nozick attempts to justify this final move is probably the least libertarian-looking component of the entire book (as well as one of the least independently plausible basic principles defended in Part 1).¹⁵ So, while Nozick is surely correct in denying that he is (like Locke) explaining the state in terms of a unanimous social compact,¹⁶ the obligations of clients/citizens and the rights of associations/states are still in his story grounded for the most part in personal consent, even if this is consent given in stages and over a period of time. Nozick does not, of course, refer directly to the problem of political obligation, nor is his language the language of consent theory. But during that time (in Part 1) in which he keeps his argument within recognizably libertarian bounds, he reasons like a consent theorist.¹⁷

The minimal state of Part 2 of Anarchy, State, and Utopia, then, seems to be defended as the sole possible legitimate state in minimalist terms (i.e., as doing no more than enforcing our basic moral rights), while the minimal state of Part 1 gets its blessing from Nozick in consensualist terms. Yet we know that Nozick insists that we possess an unlimited moral power to alienate our rights and to undertake new obligations by free consent, such that even free contracts into slavery are morally binding.¹⁸

How, then, can he avoid the tension noted above, in which unanimous free consent could in principle legitimate an extensive state that was declared illegitimate by Nozick's own minimalist standards? It is presumably insufficient to mumble something about this being a very unlikely development or about the imprudence of consensually empowering a modern redistributive state.

We might try the following. Perhaps Nozick means (in Part 2) that the minimal state is the only state that can be legitimate (and can be owed obligations of support and compliance) without unanimous consent from its citizens/subjects.¹⁹

Since all such a minimal state does is enforce citizens' preexisting natural rights and obligations, there can be no moral objections to its operations or demands.²⁰ The minimalist standard of legitimacy defended in Part 2 could then be portrayed not as the sole such standard, but as the only standard of legitimacy applicable to actual, nonideal political societies, in which it is impossible to get unanimous consent to anything, no matter how reasonable it might seem. The consensualist standard of legitimacy and obligation, apparently defended in Part 1 of Nozick's book, could then be read as the correct (but in practice largely irrelevant) standard of legitimacy to apply to political arrangements more extensive than those of the minimal state.²¹ Alternatively, consent theory's standards could be dismissed altogether as no part of Nozick's real strategy. Nozick is no consent theorist, we might argue. Indeed, the whole point of the way Nozick deals with independents in Part 1, we could claim, is precisely to show that so long as the state extends to persons its service of enforcing basic rights (and no more than this), their consent to its authority is irrelevant to its legitimacy. And consent can legitimate no state more extensive than the minimal state.

The problems with this latter reading of Nozick strike me as pointing to quite general problems for libertarian political philosophy. The first, most obvious problem is that this solution simply leaves hanging the libertarian's well-known commitment to the sanctity of contract. If free contract can legitimate any kind of nonpolitical arrangement, why should it be thought incapable of legitimating non-minimal political arrangements? The second, slightly less obvious problem faced by a libertarian trying to minimize the role of libertarianism's consensualist strain is this: there is no good reason, on libertarianism's own terms, to suppose that the minimal state is legitimate without the unanimous consent of its citizens/subjects.²²

We have seen the likely libertarian argument on this point: since all the minimal state does is enforce preexisting, legitimately enforceable rights and obligations, no consent is required to legitimate its demands. But this position seems confused in a couple of ways, even ignoring all the ways in which it might seem confused if we helped ourselves to the resources of a nonlibertarian theory of rights or justice. For a minimal state, in fact, must do more than simply enforce preexisting rights and obligations; and the ways in which it must do more—by limiting the exercise of preexisting rights—would seem (from the libertarian perspective, at least) to require the consent of the rights-holders to legitimate them.

First, a minimal state must not only enforce rights and obligations—by threatening and (if necessary) using punishment to compel performance. If it is to function as a state, it must also deny to others the right of competitive enforcement of those rights and obligations. Nozick, of course, worries a great deal about this problem, but he addresses it only by the very nonlibertarian move mentioned earlier: by allowing others' rights of competitive enforcement to be legitimately infringed provided that compensation is made. Indeed, Nozick is really only able to argue for the minimal state's justifiably excluding “dangerous private enforcement” by competitor individuals or groups,²³

leaving it entirely unclear how anything but consent could legitimate the state's prohibiting private enforcement of rights by persons or groups that used procedures that were not dangerous, or procedures that were at least as reliable as the state's own. When push comes to shove, Nozick actually allows that “[a] dominant agency and another tiny one, or a dominant agency and an unaffiliated individual person, are on a par in the nature of their rights to enforce other rights.”²⁴ Nozick's only real defense of the state's special claim to be the one that does the enforcing of rights flows from his very hesitant and enormously ad hoc speculation that perhaps the right to punish is “the only [natural] right” that is possessed not individually, but jointly—so that the dominant agency that becomes the state turns out (by definition) to hold (through the consent of its clients) a larger share of the collectively held right to punish than does any competitor agency or individual.²⁵ Since Nozick himself can barely advance the argument with a straight face, we can, I think, safely disregard it. But the result of doing so is that it is very hard to see how even a minimal state could legitimately perform its required functions without receiving the consent of those in its territory who possessed rival rights to perform those functions.

There is a second, similar problem for the libertarian, perhaps equally obvious, but certainly less widely discussed. Even the minimal state must do more than simply enforce preexisting rights and obligations (and exclude others' competitive efforts to do the same). As Nozick sees, in order to be a state, it must enforce these rights within a geographical area. But, as Nozick apparently fails to see, enforcing rights within a territory requires a state to place all sorts of limits on the property rights over land held by individuals within that territory. According to most libertarians, of course, natural property rights of the sort at issue here (preexisting, nonlegal [i.e., moral] property rights) are pretty full-bodied rights, with property holders being entitled to use their property more or less as they please, provided they do no direct harm to others in the process (or do not have obligations to others—for example, obligations of reparation of injury—that would otherwise limit their options). But in order to enforce rights within a territory, states (including, of course, minimal ones) must (among other things) control or prohibit movement across the borders of that territory. States also must prohibit property owners from giving, selling, or trading their land to alien nations or citizens (i.e., to those not bound by the state's laws), and must prohibit individual or group secession by landowners within the borders of the territory. But each of these restrictions, necessary as it may be to statehood and to the task of enforcing rights in the state's territories, constitutes a limitation on the rights of private landowners, a limitation that is not motivated by any direct harm to others that the landowner has done or threatened to do. States can acquire the rights to impose such restrictions on private landowners, according to the libertarian's own position, only by obtaining the consent of those landowners.²⁶

If this is correct, then the libertarian would be ill served by any attempt to jettison the consensualist aspect of libertarian political thought. Indeed, given that (as we have just seen) the libertarian's minimalist standard of legitimacy cannot be sustained by the libertarian independently of a reliance on consent-based rights and obligations, it is this minimalist standard that ought to be disposed of by libertarians. Libertarians, in short, should be consent theorists on the subjects of political legitimacy and political obligation. To those who observe that on the libertarian's own terms unanimous consent can justify almost anything—certainly including more-than-minimal political arrangements—the libertarian's response should be that this is correct, but that it is highly unlikely that anything like unanimous consent in a territory will ever be secured in order to legitimate such extensive, liberty-limiting arrangements. The more minimal the limitations of individual rights required by a state's institutions, the libertarian can argue, the more likely that state will be to enjoy something approaching unanimous consent (since all persons benefit, even if not equally, from enforcement of basic rights, but only some persons benefit from more extensive government). And since the best we can hope for in the real world is approximate political legitimacy, the argument continues, the minimal state is more likely than others to earn that consent which legitimates states. I do not myself find this line of argument persuasive, but it is at least one way in which a consent-based libertarian theory of political legitimacy can be reconciled with libertarianism's minimalist ambitions.

My arguments here suggest that libertarianism should be not only less minimalist in principle (if perhaps not in practice), but also slightly more anarchist in tone. Libertarianism has, of course, always been prepared—indeed, eager—to acknowledge with the anarchist the illegitimacy of most incarnations of the modern state. But while my arguments do not require libertarians to side with those “libertarian anarchists” who maintain that no state can possibly be morally legitimate, my arguments do, I believe, force the libertarian to acknowledge that only the actual consensual history of a state, not the nature of the functions it performs, can determine its legitimacy or illegitimacy. And in light of the apparent general scarcity of popular political consent in the real world, this means that merely downsizing or streamlining the state cannot have the legitimizing effects that libertarians often aver (though such downsizing might, of course, play a role in encouraging popular political consent).

None of this should sit too badly with libertarians. Given their preference for historical or process-oriented accounts of justice and property, a historical account of state legitimacy should seem reasonably natural. In fact, abandoning minimalism in favor of consensualism seems to me to make a political philosophy look more thoroughly libertarian, given libertarianism's traditional emphasis on the idea that we are free to choose anything, provided only that the choice does not wrong anyone else. If liberty is indeed the central concern of libertarianism, we might say, then libertarianism should be more concerned to defend freely chosen arrangements than to defend minimal ones; for however minimal arrangements may be, they may still be imposed contrary to the wills of those subjected to them, and so be apparent limitations of their basic right to liberty. Abandoning libertarianism's minimalism also seems to me to make libertarianism look more defensible; for it has always been difficult (for me, at least) to see how libertarians could hope to mount a compelling minimalist argument that appeals to strong basic rights of individual liberty and property without justifying these rights in a way that provides redistributionist or egalitarian liberals with material sufficient to defend at least modest rights to charity or welfare for those in genuine need. After all, poverty (and the superior economic power of others) often appears to be just as liberty-limiting as does state coercion. And if the moral possibility of private property is defended in terms of its necessity to true liberty (or self-government), so, it seems, must be a minimally decent standard of living. Thus, far from painting libertarianism into a corner, I am trying here to suggest for libertarianism a form that seems to me far less vulnerable to familiar counterarguments.

IV. Consent Theory

Embracing a consent theory of legitimacy and political obligation will, of course, still leave the libertarian exposed to all of the numerous contemporary critiques of consent theory. And it is to those critiques that I will momentarily turn. One quick caveat first, however: I am aware that in claiming that libertarians should be consent theorists I have claimed something that may seem to be plainly false of some varieties of libertarianism. Utilitarian libertarians, for instance, might seem to be defending libertarian conclusions from foundations that are fundamentally opposed to consent theory. After all, the best-known of the classical utilitarians—Hume (whom for these purposes we can count as a utilitarian), Bentham, and Mill—are also among the best-known critics of consent theory. Hume argued against it so convincingly that the whole problem of political obligation in effect dropped off the philosophical radar screen for more than a century. Nonetheless, this much, I think, can be said in favor of the reasonably general truth of my claim that libertarians should be consent theorists: Any theorist, utilitarian or not, who thinks that his arguments entail the justice or moral legitimacy of free-market capitalism must be committed as well to a conception of personal property rights and personal rights to freedom of contract that makes these rights extremely strong and extensive. And anyone committed to such strong and extensive rights of property and contract should be moved by the arguments advanced above against minimalist libertarianism, arguments supporting the need for the libertarian to accept a consent theory of political legitimacy and political obligation.

Before I begin examining the arguments against such a consent theory, I should quickly make clear exactly what kind of consent theory I am arguing that libertarians (and the rest of us) should defend. According to any theory properly called a consent theory, person A's consent to government B's rule over A (or to membership in the political society governed by B) is a necessary condition for person A to have general political obligations (of obedience and support) to B (or to B's society) and for B to legitimately rule A. Typically, consent theorists hold that personal consent is not sufficient to bring about these results if consent is in important ways defective—for example, if it is nonvoluntary, unintentional, uninformed, or given to arrangements that violate others' rights. But consent that is free, informed, and non-rights-violating (etc.) is both necessary and sufficient for political obligation. Further, since no libertarian will subscribe to a theory that grounds obligations for one (competent) person in consent given by others (such as the consent of one's ancestors, the consent of the wise, or the consent of a majority of others in one's society or territory), I consider here only theories that ground obligation and legitimacy in personal consent.

The idea of binding personal consent can, of course, be understood in broader or narrower fashions. I do not have in mind here the familiar distinction between express and tacit consent. I take both to consist in the deliberate and effective performance of actions (or omissions) whose conventional or contextual point is to communicate to others the agent's intention to undertake new obligations toward and/or convey new rights to those others.²⁷

Thus understood, libertarians should have no concerns about consent being only tacit. The distinction I have in mind here is, rather, the distinction between conceptions of personal consent that identify it only with certain actual performances (or series of actual performances) and those that expand the account of consent to identify it as well with certain dispositional, attitudinal, or related counterfactual features of persons—such as their attitudes of approval of or dedication toward something, their dispositions to consent, or the fact that they would have consented had they been able to do so, or been asked to do so, or had the occasion to do so arisen.

Libertarians seem unlikely to be sympathetic to such expansions of the concept of consent, and in that I agree with them. The facts that I approved of buying IBM stock at $48 per share, was disposed to do so, and would have done so had someone asked me to, unhappily do not establish that I did so. (As of this writing, the shares would now be worth more than twice that amount.) We do, of course, sometimes talk as if we think we can justify our treatment of others in terms of what they would have agreed to: if you are smashed by a bus in front of the hospital, and I drag you into the emergency room for treatment, I may say that it was permissible for me to manipulate your body without your consent because you clearly would have consented to my doing so. When, however, further investigation reveals that you are a suicidal Christian Scientist who deliberately walked in front of the bus—with the result that the medical treatment I procured for you not only frustrated your plans but violated one of your most cherished principles (your opposition to receiving any form of medical treatment)—I am likely to quickly abandon any consent-based justification for my action. But I am unlikely, I think, to believe that my action was therefore morally impermissible. My action was justified by the fact that I did what I had reasonable grounds to believe would be best for you, at a time when a critical choice about your well-being needed to be made and you were unable to make it for yourself. No appeal to your hypothetical consent is necessary to justify my actions, suggesting that even in cases where it is true that you would have consented to what I did (had you been able), the appeal to consent is probably really best understood as confused or as shorthand for some form of non-consent-based justification.²⁸

Of course, we need not settle here what counts as personal consent or which sorts of consent can actually ground personal obligations. For the purpose of the libertarian, any kind of consent that can be sufficient to undertake obligations and to alienate or renounce strong personal rights to liberty and property (including the right to enforce rights) will be an acceptable form of consent within libertarian consent theory. But libertarians do seem mostly to advocate actual personal consent as the only clearly binding form of consent. Actual personal consent theory is the strongest, most libertarian-looking form of consent theory, and it is the form of the theory advocated by many of the classical liberals to whom libertarians look for inspiration.²⁹

So I want to consider now the question of just how defensible such an actual personal consent theory of political legitimacy and political obligation really is.

V. Critiques of Consent Theory

Consent theory has, of course, a reasonably respectable pedigree, having been defended in the works of major figures in the history of political philosophy and having played a prominent role in the political theory appealed to in the foundational political documents of a number of modern states.³⁰

But there has also long been a tradition of skepticism about consent theory that dates at least to Hume's famous essay “Of the Original Contract” (1748),³¹ and this skeptical tradition is continued in the works of some of the most important contemporary social and political philosophers, such as Ronald Dworkin and John Rawls. Indeed, Dworkin's criticisms of consent theory amount to little more than a contemporary restatement of Hume's most famous objection: that nothing in the behavior of ordinary citizens of ordinary states (including their failure to emigrate) can be taken to count as binding consent to their governments' rule.³² Even political behavior that might look consensual is in fact neither sufficiently deliberate (or intentional) nor sufficiently free (given the severely limited and extremely costly alternatives) for it to be properly understood as binding consent.

Notice, however, that such arguments, even if convincing (as I believe they are),³³

are fatal only to a consent theory that is conjoined with perfectly inessential conservative ambitions regarding the legitimacy of contemporary states. That contemporary citizens have not authorized their states by free consent argues against consent theory only on the assumption that at least some contemporary states must turn out to be legitimate, according to any acceptable standard of legitimacy. Libertarians, of course, reject these conservative ambitions with respect to the modern state, so libertarian consent theorists are unlikely to find objections of the Hume/Dworkin sort very interesting. Nor should consent theorists, libertarian or not, be unnerved by such objections. The absence of legitimate contemporary states according to consent theory's standards argues no more strongly against those standards than the presence of extensive inequality in the world argues against liberal egalitarian standards of justice.

It would, perhaps, be a fair complaint against consent theory if its critics could show that it was practically impossible for a political community to satisfy its requirements for legitimacy. But it is, on the contrary, quite easy to imagine how legitimate societies (according to these requirements) could arise,³⁴

and even how they could sustain themselves in the face of newcomers and successive generations. Provided only that the society founded by its original members was able to acquire (via the consent of those members or others) appropriate rights of control over the land the society claimed as its territory, that society could legitimately exclude would-be immigrants, admit them only on the condition of their consent to existing arrangements, or charitably accept them as members of a different sort (not bound by the terms of the original charter). The same is true of the offspring of original members. Complications will abound, of course, if the opportunities of those excluded from the consensual society are in various ways severely limited by the conditions outside that society. But such complications are certainly not (for practical purposes) a necessary feature of the world.³⁵

It might seem that offspring present special problems here, since they will only be capable of achieving the autonomy necessary for them to make meaningful choices if they are raised and socialized in certain ways. Accordingly, offspring might seem to have some claims against the consensual illiberal society, in particular, rights that it be structured at least liberally enough to facilitate their achievement of autonomy. (Perhaps the desired conclusion would then actually be disjunctive: either unanimous consent is insufficient to justify all kinds of illiberality in an enduring polity, or illiberal consensual polities may not extend beyond one generation.) In fact, such an argument ignores several obvious lines of defense available to the consent theorist. One is that the rights of offspring to conditions for the achievement of autonomy are held not against societies, but (principally) against parents, who may provide such conditions within the family (or in subsidiary groups) even in tremendously illiberal societies (just as these conditions have, of course, been regularly provided for children in actual illiberal societies, both historical and contemporary). Another possible line of defense against special worries about offspring is to argue that autonomy is not a good to be maximized but only a condition to be respected when it is present in a person. As such, there is no general moral responsibility (on society or on persons generally) to provide the conditions within which autonomy may (or may most easily) develop. Autonomous choices must be respected when made; but their frequency need not be maximized, nor need their possibility be guaranteed to all with the required capacities.

The mere fact that there are not now, and are unlikely to be in the foreseeable future, political societies in the world that are legitimate according to consent theory is no objection to that theory. It is simply a lamentable fact about the world, one to add to the other such facts lamented by libertarians. We need a deeper or more substantial attack than this on the consent theorist's account of legitimacy if we are to regard the libertarian's adoption of consent theory as a mistake. It might seem, perhaps, that we could mount just such an attack by relying on contemporary observations about various “deep” features of social life that were uniformly (and perhaps deliberately, self-servingly) ignored by the classical consent theorists. Feminist political theorists, for instance, have frequently charged that consent theorists ignore the fact that women are routinely denied opportunities for genuine political consent, and that in the case of women some form of submission is typically identified with binding consent.³⁶

Similarly, and more generally, it is often claimed that the conditions of actual political life are normally such that even apparently free consent given by many individuals will be “tainted” (and rendered nonbinding) by the unfair background conditions and unequal bargaining power in the context of which any agreements must be made.³⁷ Thus, Rawls writes:

If the apparent consent given in actual political societies is not sufficiently intentional or free to be binding, consent theorists' standards of legitimacy may appear unrealistic.

But here again, these observations constitute objections to consent theory only if we begin by assuming that at least some actual political societies must satisfy defensible standards of legitimacy. And this is not something that libertarians (or we) ought simply to concede. Feminists surely do not want to argue that it would be impossible or undesirable for societies to include women as fully equal participants in the political community, or for societies to identify as women's binding political consent only those free, informed performances (or omissions) that should be so understood. If not, then the objection posed by feminists is not an objection to consent theory's standard of legitimacy, but only an objection to society's failure to take that standard seriously. As to Rawls's related but more general concerns, the same kind of response seems possible (though I will consider below a more potent aspect of the Rawlsian argument). Libertarian consent theorists will, of course, worry far less than Rawls does about the concentrations of wealth and power that may be produced by a history of free exchanges. But even those of us who do worry about the possibility that those concentrations will create unfair background conditions for political consent should be worrying principally about how to correct those conditions, not about the consent theory standard of legitimacy that condemns societies exhibiting the conditions. Only if it were impossible or wrong to correct unfair background conditions for consent would this seem to be a worry for consent theory. But Rawlsians cannot, I suspect, intend to claim that such changes are impossible or wrong, given that their aim (in their nonideal theory, at least) is in part precisely to identify the practical steps that we ought to take in order to accomplish this correction.

A serious attack on consent theory must argue more directly that genuinely free, informed (etc.) personal consent is either not necessary or not sufficient for political obligation and political legitimacy. To show that consent is not necessary for political obligations, one should defend an alternative account of possible grounds for such obligations that would force the consent theorist to accept at least a pluralistic theory of political obligation and legitimacy.³⁹

Or one could, like Hume, try to undermine consent theory more completely by arguing that the alternative account defended in fact shows consent to be utterly superfluous to the generation of political obligations—so that even in those cases where consent might appear to bind us, there is another more basic ground of obligation already instantiated.⁴⁰

I have tried over quite a few years to defend the view that none of the alternative accounts of political obligation and legitimacy yet proposed is successful.⁴¹

Indeed, it seems to me that one of the strongest arguments in favor of consent theory is precisely the failure of all of its rivals to satisfy the reasonable ambitions of a theory of political obligation.⁴² For those who find my arguments against these rivals unpersuasive, I am afraid I have nothing new to offer here. Instead, I would like to consider the other natural strategy for attacking consent theory, which is to argue that, regardless of how rival accounts might fare, consent theory itself cannot successfully defend the kind of political consent on which it concentrates as sufficient to ground political obligations.

One can, of course, pursue this strategy by defending a general, wide-ranging skepticism about the obligating force of free, informed (etc.) promises or consent, maintaining that there is simply no good reason to regard them as morally binding, as always morally binding, or as binding within anything but a just society. We might argue, in short, that consent is simply not a privileged source of special obligation, contrary to the apparent assumptions of consent theorists. We value the choice involved in promising or consenting, say, only because respecting our choices normally leads to our getting what we want. But our wants are often ill considered or immoral, and even where our wants are not thus criticizable, we often choose unwisely or wrongly. It will not do to reply (with Mill) that our individual choices should be respected because, after all, we know ourselves best. For our choices often adversely affect others, whose true wants we cannot consistently claim to know best. Why should even free, informed consent that issues from unwise choices or choices that adversely affect others be taken seriously as a ground of moral obligation or legitimacy?

The answer to such general skepticism about consent must, I think, be to reemphasize more persuasively the reasonably obvious advantages of the practice of promising/consent, both in consequentialist/mutual-advantage terms and in terms of the promotion of the value of autonomy. Emphasizing the latter, for instance, we could argue that we respect consent (and, through this, the choices on which consent is based) wherever doing so is necessary to respecting every person's right to be self-directing or self-governing, a right that is itself perhaps grounded (as many libertarians would have it) in special claims that we have to our own bodies and labor. Where consent is so defective that we are disinclined even to impute it to a self-governing agent—as in cases of coerced or seriously misinformed consent—or where consent fails to respect the equal entitlements of others to self-governance, we take the obligating and legitimating powers of consent to run out. But consent that is merely unwise or that merely adversely affects others (without violating their rights) is nonetheless binding and legitimating, being just an unfortunate aspect of a morally acceptable—indeed, a morally mandated—regime of universal self-government.

Whether or not such arguments are found persuasive, I think that a more promising approach to defeating consent theory is to defend a skeptical view not about consent generally, but more specifically about political consent. Even if promises and consent are perfectly respectable sources of obligation and perfectly respectable as legitimating acts in our everyday lives, the argument would go, consent to a state, a government, or a political society is another matter. Because of the nature of states and governments, even free, informed consent to their authority may neither legitimate them nor ground political obligations for consenters. I will consider two variants of this approach. On the first line of argument, political consent is alleged to be insufficient to bind us because what this consent must be consent to is too vague and open-ended to constitute the content of binding consent. On the second line of argument, which we might take to be the real point of the earlier Rawlsian objection, political consent cannot bind us to or legitimate that to which we consent, because even free, informed (etc.) consent can (and regularly does) issue in plainly illegitimate political outcomes.

Consider first, then, what political consent would have to be consent to. Since in order to count as undertaking political obligations (and as legitimating government) we would have to consent to the permanent authority over us of some state, government, or society, the substance of our consent would appear to have to be both objectionably vague and objectionably long-termed.⁴³

We must consent to a government's making and enforcing binding law. But while we might be perfectly content with the government and laws as they stand when we give our consent, we simply cannot reliably predict the changes that will occur over time. Similarly, we cannot reliably predict other significant changes that might occur in the composition or character of our political society, changes that might also make us deeply regret our consent to membership in that society.⁴⁴ Unwise choices to consent are one thing. But consent that purports to undertake utterly unspecific future obligations is another. It simply cannot bind us, since it can be no part of a self-governing life.⁴⁵ Similarly, consenting to a government's authority over us seems to be completely indefinite as to term, much like a contract in perpetuity; and such open-ended consent might also be regarded as objectionable on the grounds that it frustrates, rather than promotes, self-government (except, apparently, in the case of marriage).

It seems plausible to respond to such arguments, however, by maintaining that they underestimate the extent to which a consent theory of political obligation and legitimacy can just absorb the criticisms without any apparent weakening of the theory. If it is true that consent's binding force is lost or diminished as the content of the consent is more open-ended, the consent theorist can simply accept this fact as one more case in which consent is insufficient for obligation. Since consent theory already accepts other such limits to consent, accepting another will not seriously affect the theory's force, unless this limit cannot be motivated by the theory justifying the relevant obligations (which, in this case, it plainly can be), or unless accepting this limit will show that the ideal of government by consent thereby becomes unacceptable, unattractive, or impossible to achieve in practice. But nothing of this sort follows from worries about vague or open-ended consent. Perhaps the consent theorist will insist in response that genuine, binding political consent must be (or must be understood to be) relatively specific as to its content, or that in order to continue to bind and legitimate, political consent must be given regularly, rather than once and for all. The theorist could argue that, in fact, consent's binding (and legitimating) force diminishes over time; or she could accept in her moral theory analogues of the contract law doctrines of “frustration of purposes and impossibility,” which “allow relief from contractual obligations when circumstances turn out very differently from what the parties expected.”⁴⁶

However the consent theorist opts to respond to this class of objections, though, it seems clear that her response need not involve anything that interestingly weakens the appeal of consent theory or that argues against its claim that political consent is sufficient for political obligation and legitimacy.

VI. The Illiberal Consensual Society

I now arrive at the second line of objection to consent theory, the one that strikes me as the most formidable: that if consent is accepted as sufficient for legitimacy, this commits us to accepting the possible legitimacy of political societies that are uncharitable, intolerant, bigoted, and, more generally, illiberal. Since illiberal regimes are palpably illegitimate, the argument goes, there must be fundamental defects in consent theory's standard of legitimacy.

Suppose, to begin, that a group of persons (without obligations that would preclude this) comes together on a piece of land to which they have managed to establish a valid collective claim (say, an unwanted territory purchased from some sovereign state). By unanimous consent, they create for themselves a set of political institutions that provide for lawmaking, law enforcement, and more generally for both internal and external sovereignty over their territory. All are excluded from membership or residence in the territory of the new polity except on condition of consent to the authority of its constitution, laws, and government. Since, according to most libertarians, genuinely free and informed, unanimous consent can legitimate any kind of arrangement that does not violate the rights of outsiders, let us assume that the consenters are perfectly free (subject to the one constraint set by others' rights) to frame their new polity's institutions and laws as they wish.⁴⁷

The laws could provide for an absolute monarch or dictator (who is free to prey upon the masses), for political participation for only a privileged few, for draconian censorship and invasion of privacy, or for a police state. The consenters could provide that a small minority control the vast majority of the society's wealth and resources, that all members be required to practice one state religion, or that no one be permitted to practice any religion. They could require regular gladiatorial games, provide that members of disliked racial or ethnic groups be denied citizenship or even entry into the society, ban contraception and “unnatural” sex, and guarantee that no societal resources whatsoever be used to improve circumstances outside of the society (except where such neglect would violate outsiders' rights, as opposed to their needs or wants). Indeed, if a strong form of libertarianism is correct and all rights are alienable, then the laws could provide for human slavery, allowing slaves and masters to be selected by any means one can imagine. In all of these ways and more, the society arising from free, informed, unanimous consent could be despicable almost beyond description, yet plainly legitimate nonetheless according to consent theory's standard. Virtually the only categories of vice such a society could not in principle exhibit, one might say, would be those that involve violations of the rights of nonconsenting outsiders. Doesn't such a catalogue of possibilities constitute a virtual reductio of consent theory?

Surprisingly, perhaps, the answer seems to me to be no. Consent theorists can grant, of course, that imposing a liberal egalitarian regime on persons without their consent may typically be less wrong than would be imposing an intolerant, inegalitarian regime (because the former is likely to violate fewer rights and do less wrong than the latter). Thus, if we are considering only the question of the least-bad illegitimate polity to impose on a captive audience, liberal egalitarians might be right to favor what they favor. Libertarian consent theorists, of course, will favor more minimal political arrangements than those favored by liberal egalitarians, even where the question is only that of the least-bad illegitimate state to impose on the unwilling. But consent theorists with more robust moral theories than those generally embraced by libertarians could concede this much to the liberal egalitarian without in any way weakening their defense of consent standards for legitimacy and political obligation. What must be addressed more directly by the consent theorist, however, is the claim that the unattractive features of illiberal political societies obviously render them illegitimate.

What, then, is it about the ugly, illiberal, consensual society that is supposed to render it plainly illegitimate? In my view, a state or political society, like any other institutional arrangement, is legitimate if and only if it actually possesses the rights to do what it does (or, perhaps, the rights to do what it most centrally does). Territorial states, on this analysis, are legitimate if and only if they actually possess the rights to exclusively make and enforce law, to be obeyed by their subjects, to be free of various kinds of external interference, and to exercise jurisdiction and control over a geographical territory.⁴⁸

But there is no reason to suppose that simply by virtue of being ugly and illiberal the consensual state described above must lack the rights to do these things. After all, ugly and illiberal people have all sorts of rights to govern their own lives as they please. If the members of the illiberal society genuinely own the territory on which they conduct their business (and so are entitled to control over it), and if each member has freely alienated the rights that might give him valid grounds for complaint about being subjected to illiberal arrangements, then it is initially hard to see from what quarter complaints of illegitimacy might arise. However unattractive their society may be to us, it seems to have the rights it needs to be legitimate with respect to its subjects, and they seem to owe it political obligations grounded in their free consent; and that, of course, is precisely the kind of legitimacy with which we are here concerned.

VII. Objections and Replies

I can imagine three kinds of responses to the argument as I have presented it thus far. The first is that some rights simply cannot be alienated, so there must be some “internal” limit to the illiberal arrangements that could be legitimated even by free consent. The second is that the illiberal consensual society cannot be legitimate because it is unjust. The third, closely related to the second, is that while the members of the illiberal consensual society may have no reasonable grounds for complaint against each other (or against their state), other groups of persons (e.g., the members' children or outsiders) plainly do, and the grounds of their complaints are what delegitimate such societies. In responding to these objections, I will have to try in places to distinguish how a right-libertarian consent theorist would respond from how a more moderate consent theorist (like myself) should reply, so that we can appreciate both the force of these objections against a libertarian consent theory and what I take to be their true force.

I will begin with the charge that the creation of my imaginary illiberal consensual polity involved the alienation of rights that are in fact inalienable. I begin here, but perhaps contrary to first appearances, it is not really necessary for the consent theorist to remain here long or to say much in response to this line of argument. The libertarian consent theorist, of course, will strongly object to the very idea of inalienable rights, and for her the battle can be waged on that point. I am not convinced that the libertarian will lose that battle, but I will not try here to defend that view.⁴⁹

Suppose, instead, that my illiberal consensual polity cannot in fact be legitimated by consent, because some of the rights that would have to be alienated in order to legitimate it are actually inalienable. That result might affect the libertarian element of libertarian consent theory, but it is hard to see how it would affect the consent theory. Consent theory maintains that genuine, nondefective consent is both necessary and sufficient for political obligation and legitimacy. If only defective consent—for instance, “consent” that involves an attempt to alienate the inalienable—could (apparently) legitimate a particular polity (or kind of polity), then that polity is simply in principle illegitimate by consent theory's standards.⁵⁰ Indeed, if the liberal egalitarian (e.g., Rawlsian) critic of consent theory were to press this point harder, arguing that all illiberal societies could acquire the rights in which their legitimacy consists only by citizens' alienating the inalienable, then it might seem that he had done the consent theorist a favor—by bringing the consent theorist's standards of legitimacy into conformity with his own and depriving himself of the benefit of exploitable differences in the implications of the two views.

It seems unlikely, however, that the liberal egalitarian critic of consent theory will want to carry the argument to such extremes. For surely not all of the rights alienated in my imaginary consensual polity (that is, not all of the rights whose alienation contributed to the illiberality of the resulting polity) could be plausibly argued to be in principle inalienable. The rights to associate with whomever we please or to engage in consensual sexual relations of whatever sort and with whomever we please, for instance, seem to be rights that we normally take ourselves to be perfectly empowered to alienate, even for very long terms, as when we join a monastic order, enlist in the military, commit ourselves to a monogamous intimate relationship or marriage, and so on. If rights like these are in fact alienable, then there is no apparent reason why they could not be alienated en masse in the creation of an illiberal polity, a polity that will use the acquired rights to enact and enforce bigoted or puritanical policies. In short, there is no real possibility that liberal egalitarians will turn out to approve of everything that consent theory regards as legitimate.

And that, perhaps, may finally get us near the real point of the liberal egalitarian critique (the second possible response imagined above). It is not that people cannot make binding devil's bargains in their private lives, but rather that their doing so en masse cannot legitimate the resulting polity. Of course people are entitled to make private choices that flow from very unsavory motives, such as dislike of certain races, sexual preferences, or appearances. But polities, unlike our private relationships, are governed by principles of justice, and illiberal polities are illegitimate precisely because they are unjust. Political authorities are bound as a matter of justice to treat their subjects fairly and well, as equals and as equally inviolable. As Dworkin puts it, “No government is legitimate that does not show equal concern for the fate of all those citizens over whom it claims dominion and from whom it claims allegiance.”⁵¹

And the authorities of the consensual illiberal society cannot discharge the duty to treat all as equals if those subjects have agreed to arrangements that are degrading or that make them unequal.

Once again, of course, the libertarian consent theorist will simply reject the liberal egalitarian theory of justice that is alleged to have these implications, arguing that if one cannot point to clear wrongdoing in the process that creates the polity, one has no grounds for claiming that the resulting polity is unjust. But suppose that one finds it difficult to follow the libertarian this far, committed as one might be to a more moderate, left-libertarian or liberal egalitarian conception of social justice. Even in that event, I think, there is nothing to prevent one from nonetheless accepting consent theory's standards for legitimacy and political obligation.

A liberal egalitarian theory of justice and a consent theory of political legitimacy can be consistently conjoined if one accepts the distinction between what I have called “justification” and “legitimacy”—which I take to be two kinds of favorable evaluations of political institutions within quite different dimensions of evaluation.⁵²

General good qualities and virtues of a political society—such as benevolence or justice—are what is appealed to in justifying that society's existence, in showing why such a society is a good thing. But the society's legitimacy is a function of its specific interactions with particular subjects and consists (in part) in its having a particular set of rights over a subject. A state's legitimacy (or illegitimacy) with respect to me and its possession of other good (or bad) general characteristics are, in my view, simply independent variables, in the same way that a business's general good (or bad) qualities are independent of its right to have me as a client. States can be evaluated independently in either dimension, sometimes producing mixed evaluations (as when we refer to “benevolent dictatorships”).

If this is correct, then even if the liberal egalitarian critic of consent theory is right—contrary to the claims of the libertarian—that consensual political societies can be unjust (simply because they are illiberal), it does not follow from this that such societies are also illegitimate, or that consent theory's standard of legitimacy is flawed. These conclusions about legitimacy seem to follow only to those who conflate two importantly independent dimensions of institutional evaluation: justification and legitimacy. All that in fact follows from the liberal egalitarian critique is that illiberal societies are (in a variety of ways) bad, that the world would be morally better without them (or, at least, that it would be morally better if they lacked those bad traits). Institutional legitimacy is a function of the rights arising from specific relations between those institutions and particular persons, considered one by one. And consent theory maintains, persuasively I contend, that a person's genuine, nondefective consent to arrangements simply removes for that person reasonable grounds for complaint about those arrangements, at least so far and so long as his consent specifies.

The reason many insist, as a condition of legitimacy, that political authorities must treat all persons as equals is that we normally lack the opportunity to decide for ourselves how we wish to be treated in the political realm (and being treated as equals at least preserves most of our private-life options). The members of modern political societies are a captive audience. And the liberal egalitarian (e.g., Rawlsian) theory of justice is precisely about the kinds of institutional coercion to which such a captive audience may reasonably be subjected when its members have no choice in the matter. But my imagined consensual illiberal society is not a captive audience; it is a body of persons that has freely chosen the coercion to which it is subject. There is no need to appeal to “second-best” standards of legitimacy for captive audiences. The second-best theory is simply about the best and the worst that illegitimate societies can be. Such a theory is morally interesting, but it is not a theory of legitimacy. And finally, of course, there is also a clear sense in which the government of a consensual illiberal society does treat its subjects as equals and as equally inviolable—precisely by taking seriously their free, nondefective choices about how they wish to live.

How might the liberal egalitarian critic of consent theory respond? Perhaps he would answer with the third response imagined above. What you say may seem well and good, he might allow, so long as we focus solely on the happy consenters who live out their lives in the despicable illiberal society of their dreams. But what about some other classes of persons who may not be made so happy by the structure of such a political society? These classes might include: (a) those who give their binding consent to illiberal government, but who later change their minds or find themselves with new personal preferences or commitments of a sort that will make continued citizenship in such a society oppressive; (b) those of subsequent generations who are born within the illiberal society's territories; (c) those who are excluded from illiberal societies, or those who are excluded except on condition of consent to illiberal arrangements and whose opportunities for membership in other societies are few, costly, or unattractive; and (d) those who need what the illiberal society possesses.

How we understand the moral situation of the persons in these various groups will differ depending on how strictly we are prepared to defend narrowly (right-) libertarian aspects of libertarian consent theory. For instance, if we accept the familiar libertarian position with respect to positive rights—namely, that there are no positive rights that do not arise from consent or from obligatory restitution—then it will be relatively easy to dismiss the complaints of most of these groups. Provided only that the illiberal consensual society is not responsible for the plights of these persons, they can be treated according to whatever policies the society has elected to adopt.

But again, even those who find such libertarian arguments unpersuasive need not be led by consideration of these unfortunate groups of persons to conclude that consent theory is defective. It is perfectly consistent to accept the consent theorist's standards of political obligation and legitimacy while also maintaining