Exorcising Hobbes’s Ghost: A Future for Constitutional and International Law

Monica Hakimi* April, 2025

Law for Leviathan: Constitutional Law, International Law, and the State. By Daryl Levinson. New York: Oxford University Press. 2024. Pp. viii, 301. $34.95.

Introduction

Daryl Levinson¹ wants to liberate constitutional law from the ghost of Thomas Hobbes. More than 350 years ago, Hobbes advanced a conception of law that, for all the efforts to move past it, continues to resonate. In the strictest iteration of the Hobbesian conception, law cannot exist apart from Leviathan, a sovereign state that is the “sole source of legal authority” in a defined territory (p. 24). As the only possible source of law, Leviathan cannot itself be subject to law. “[I]t exercise[s] legally limitless power over its citizens and interact[s] with other Leviathans in a lawless international arena” (p. 24). It is also the source of “supreme coercive force” in its territory (p. 23). This unique combination—the absolute authority to issue binding commands for its subjects to follow, plus the capacity to use overriding coercion against those who might stray—is, in the Hobbesian conception, necessary for law to work.²

Levinson argues in Law for Leviathan: Constitutional Law, International Law, and the State that that conception is erroneous. In contemporary U.S. and international law, the state is not absolutely sovereign. It does not have “legally limitless power” to do whatever it pleases (p. 24). Although the state can still be an important source of law, the state is also a critical subject of law—expected to work within, rather than outside or above, the legal structures established to govern its interactions. “[T]he development of international and constitutional law has proven,” in Levinson’s view, that Hobbes was “wrong about the impossibility of imposing positive law on the sovereign state” (p. 7). “[S]tates can, in fact, be subject to law” (p. 7). But, Levinson claims, “Hobbes was right to see that states could not be subject to the familiar form of law created and implemented by and through states for the purpose of governing the lives of ordinary people” (p. 7). Levinson’s opening insight is that “[l]aw for states” works differently from “law by states, for ordinary people,” as Hobbes envisioned (p. 6).

This insight raises two large questions that drive the book. First, how does law for states work, if not as Hobbes imagined that law must? Second, how might it be made to work better than it now does? Levinson does not purport to offer definitive answers to these questions so much as he seeks to focus attention on, and present an alternative analytic apparatus for thinking through, them. For guidance, he looks to the field of international relations, which in his telling largely accepts that international law is “different in kind from the paradigmatic legal system run by and through the state” (p. 1). By comparison, he says, “constitutionalists have lagged, and still have much to learn from, their internationalist counterparts in . . . making the state the subject rather than the source of law” (p. 1). Thus, the self-described project of the book is “to bring together international and constitutional law to develop a unified theory of law for states” that is free from Hobbes’s errors (p. 7).

I am fully on board with Levinson’s project and write this Review in the spirit I think it invites. I share Levinson’s claim that the Hobbesian conception still, at deep levels, informs a lot of contemporary thinking about law—including, I would say, international law.³ I also agree with Levinson that, because this conception is specious, it corrupts the analyses of those who rely on it (p. 4). But, I will argue, Levinson himself has not managed to escape it. Throughout the book, he employs a conception of law that is still fundamentally Hobbesian in nature and, for that reason, unable to provide the analytic apparatus we need to think through the important questions that he raises. The core problem with this conception is that it conflates authority with coercion; it fails to account for what authority does in the legal enterprise independently of, and differently from, coercion.

Disentangling authority from coercion, as I propose that we do, allows us to break completely from the Hobbesian conception, with three clarifying effects.⁴ First, it illuminates how law works unlike Hobbes imagined. Second, it presents a metric for identifying when law is working well or poorly (and for whom). Third, it brings into sharper focus the urgency of finally exorcising Hobbes’s ghost from our legal and political imagination. Today, both U.S. constitutional and international law are being radically transformed, creating a once-in-a-generation opportunity to reconstitute their basic foundations. As long as Hobbes’s ghost retains its stranglehold on us, we risk not only squandering the opportunity to improve our lot but also stumbling confusedly into a world that is considerably more oppressive than the one that we are leaving behind.

I. Hobbes’s Ghost

Levinson and I agree on the defining features of the Hobbesian conception of law. “For Hobbes,” Levinson explains, “the idea of law without the state was a logical absurdity” (p. 24). Law, according to Hobbes, is the command of a sovereign.⁵ It is backed by the threat or use of coercion against those who might stray.⁶ And it is justified by the need to control those who might trample on others, destroying their peaceful coexistence. “Law was brought into the world for nothing else but to limit the natural liberty of particular men, in such manner, as they might not hurt but assist one another, and join together against a common enemy.”⁷ “Following in Hobbes’s footsteps,” Levinson explains, “a long tradition of positivist jurisprudence, defining law as the command of a sovereign backed by the threat of sanctions, has denied the reality of any form of law not stemming from the state” (p. 24). If law must come as a command from the sovereign, it cannot itself be commanded to the sovereign. Law “is essentially a coercive apparatus” that the sovereign uses to control its subjects for their own (supposed) common good.⁸

Levinson breaks from the Hobbesian conception on the form that law must take—that is, on the criteria for defining what law is. He follows a more contemporary line of jurisprudence that contends that law can, but need not, take the form of a sanction-backed, sovereign command. The break is skin deep, however, because he preserves the Hobbesian conception on the function that law must serve—that is, on the criteria for defining what law does. In this conception, law “is, in the end, an instrument of social control.”⁹ It works insofar as it in effect controls what its addressees do. Put differently, it works by getting them to do what it tells them to do, rather than what they would themselves choose to do, if they were exercising their own judgment. Levinson’s quandary is that form and function are entwined. Law that departs too dramatically from the Hobbesian form also struggles to serve the Hobbesian function. Thus, as long as we assume that law must have the practical effect of controlling people to work, we invariably will overlook or discount what law for states does. Let me unpack what I mean.

A. From Form to Function

The reason for insisting that law must take the form of a sanction-backed, sovereign command is that that is how law does what law, in the Hobbesian conception, exists to do: control those who would otherwise disrupt the social order. When a single sovereign has absolute authority in law, it can issue clear commands that settle, “finally, and authoritatively” what its subjects must do.¹⁰ When it also has the capacity to use overwhelming coercion against its subjects, it can ensure that they comply with its commands. Thus, lodging both supreme authority and overriding coercion in a single source—the sovereign state—is the Hobbesian recipe for making law controlling. Form follows function.

The significant break from the Hobbesian conception in contemporary jurisprudence has been to posit that law need not adhere to that form to serve the same function.¹¹ Control is still commonly taken to be the raison d’être of law, the thing that law really must do to be worthwhile.¹² But law is said to serve this function, even when it neither emanates from a single sovereign nor “link[s] back to coercion.”¹³ The jurisprudential task is then to explain how law—or more precisely, how the authority of law, as distinct from the coercion with which it might or might not be conjoined—can function to control people. How, exactly, does a source with authority get people to conform their behavior to law and, in this sense, control them if it is not an all-powerful sovereign that unilaterally issues and uses coercion to enforce its commands?¹⁴

Levinson takes on this jurisprudential task for law for states. He contends that law for states faces two serious challenges—what he calls “law without settlement” and “law without the sword”—by virtue of its departures from the Hobbesian form (p. 27). First, for law to be “a means of social control,” legal settlement is essential.¹⁵ If law does not settle what should be done, “every person is left to serve as ‘his own judge,’ ” and law cannot possibly through its directives control what they do (p. 27). Thus, Levinson contends, “[t]he first task of any legal order is to establish what the law is” (p. 27).

Legal settlement is relatively easy to achieve when authority is fully vested in a single sovereign that can unilaterally use that authority to dictate what everyone else must do. But in law for states, that vector is reversed; authority is stripped from and addressed back to the state, which becomes the subject. For law for states to function as an instrument of control, authority must first be allocated to a source that is not the state. That source must then settle what should be done for the state. In constitutional and international law, the source with authority is not a single entity, like the sovereign king. The sources are multiple—separate organs or agents that have to decide together what should be done if they are to settle their terms. For example, the ultimate source with authority in U.S. constitutional law is the collective We the People, reflected in the construct of popular sovereignty (p. 69). Other sources of U.S. constitutional authority include officials who staff the three branches of the federal government and the fifty state governments. In international law, the classic sources with authority are groups of states that collectively settle their terms, usually by ratifying a treaty or converging on common practices that become customary law.¹⁶

This departure from the Hobbesian form—transferring authority from a single sovereign to a broader collective—makes law for states possible. It vests authority in sources that are not identical to the sovereign. But Levinson explains that it also presents the challenge of law without settlement (pp. 27–28). Legal settlement is necessarily more difficult to achieve when multiple agents must decide together what should be done than when one can unilaterally dictate the terms for everyone else. The larger the collective, and the more complex the societal problem that it needs to address, the harder time its members are likely to have settling what should be done at every level of social organization. For example, even if the agents who comprise a given collective all agree on the general principles to govern their relations, they would still need to decide how these principles cash out in concrete cases if they are to settle what should be done. Even when they agree on what should be done in concrete cases, they will invariably face future cases with different facts, in which decisions again need to be made. Unless they keep regrouping and deciding together what should be done, or delegate to a single entity authority to decide what should be done for them, legal settlement will be lacking. And any authority that they have will not effectively work to control their subjects. Thus, law that departs too dramatically from the Hobbesian form is limited in its capacity to overcome the “defect” of uncertainty and serve the Hobbesian function.¹⁷

The second challenge that Levinson identifies goes to the issue of enforcement (p. 39). After a source with authority settles what should be done in law, the subjects must conform their behavior to it for it to control them. When law takes the Hobbesian form, the source with authority is also the source with overriding coercion. It uses both authority and coercion to control its subjects. By contrast, in law for states, the source with authority has to work without and very often against the source with overriding coercion. There are longstanding questions about its capacity to do that. Authority lacks the materiality to exert physical control, like a jail cell can. For it to have a comparable effect independent of or without coercion, its addressees must themselves decide to conform their behavior to it.

The question then becomes to what extent it is doing the work.¹⁸ Do people conform their behavior to what is settled in law because its authority is working on them; because they would have behaved that way in any event; or because they are however subtly being coerced by another source of power that lies behind it? The question is salient, because when authority is truly independent of and used against the sources of coercion, it evidently struggles to be controlling. One need not look hard to find cases in which those who wield disproportionate coercive power do not do exactly as others say they should. Thus, here again, law that strays too far from the Hobbesian form—this time, by separating the source with authority from the source with overriding coercion—has a lesser capacity to serve the Hobbesian function. Just as form follows function, function follows form.

Levinson is adamant, however, that law must have the practical effect of controlling people—it must get them to conform their behavior to what it prescribes—for it to work:

An effective system of law requires the validity and content of legal norms to be authoritatively specified and broadly complied with. If subjects are free to decide for themselves what the law is, or ignore it, then law will fail in its essential function of making its subjects do things they would not otherwise want to do.¹⁹

Many international relations and legal scholars share Levinson’s view. These scholars of course recognize that international law breaks from the Hobbesian form. They also take different positions on the extent to which international law can overcome the above two challenges to satisfy the Hobbesian function. But they widely concede that that is what international law must do if it is to operate as “real” law does. Insofar as international law’s content is unsettled, it does not clearly enough define what its subjects must do, so it cannot control them, and we are constantly told, its practical effect must be diminished.²⁰ Insofar as international law’s content is settled, but its subjects do not conform to it, it still does not control them or, we are again told, work.²¹ In either event, it fails to do what Hobbes and his acolytes insist that law must do to justify its very existence: control people.

B. The Analytic Trap

The quandary is that form and function are related. When we discard the Hobbesian form by locating authority beyond the confines of a single, all-powerful sovereign but preserve the Hobbesian function by insisting that law be controlling, we inevitably end up with what seems like lesser law—law that cannot as effectively serve its essential function. And Hobbes is proven to have been mostly correct.

This, however, is an analytic trap. The Hobbesian conception offers us only two options. One is to accept law on the terms that it presents. In this conception, the social value of law comes when it controls people.²² It works by overriding their independent judgment and depriving them of the liberties that they might otherwise enjoy, such that they do as it directs them to do, rather than as they would independently decide to do.²³ Neither constitutional nor international law has any role. The second option is to reject the Hobbesian conception and try to create a more emancipatory law—law that does not take as its central mission the subjugation of its subjects. But the Hobbesian conception suggests that this effort will fail. Although we might call the alternative that we create “law,” it will not be able to do the bare minimum that it must do if it is to serve its basic function as law. Thus, the choice, once one assumes this conception, is between law that is controlling and law that is, for all practical purposes, inert.²⁴

To Levinson’s credit, he refuses to accept this choice. Both of the above options are repulsive. Experience also tells us that the choice between them is false. In fact, we can create law that works differently from what Hobbes imagined, not just to subjugate but also to emancipate people. Because Levinson clings to the Hobbesian conception on function, however, he cannot guide us out of this analytic trap—or toward a more emancipatory law. Again, he is not alone. Many analysts share the assumption, derived from the Hobbesian conception, that law must have the practical effect of controlling people to work. Law for Leviathan illustrates how this assumption leaves us trapped between thirsting for law to be controlling and conceding that it will not really work to get us where we need to go.

“The fundamental dilemma of state power,” Levinson argues, “is that a state powerful enough to deliver valuable goods is also powerful enough to inflict grave harms” (p. 87). On the one hand, an uber-powerful state might serve humanity well. Most people do not flourish when militants roam from territory to territory, using violence to conquer their enemies and all the innocents who live there. Better to have a state with enough power to control these people for us, as Hobbes suggested. On the other hand, a state with all that power would necessarily also have what it needs to subjugate the rest of us—or at least to disregard our priorities and perspectives. Given these “risks and rewards of state-building,” Levinson explains, the governance project is to “manag[e]” state power so that the state has enough power to provide for us but not so much that it can constantly subjugate us (pp. 86, 92). For Levinson, this means using law to control those who are themselves in positions to control state power.²⁵ And we are trapped.

Consider three concrete examples of how the trap manifests in Levinson’s analysis. First, Levinson recognizes that the separation of power—and by “power” here, he means “authority”—is a desirable “feature of the constitutional design” insofar as it helps to prevent a slide toward “authoritarian tyranny” (pp. 84, 85). Dividing authority among different branches of government limits the extent to which any one can use law for tyrannical control. But, Levinson suggests, the division of authority also limits the government’s capacity to use law for large-scale social programs for the common good. “[T]here is a fine line,” he says, “between limiting the power of the state and crippling it” (p. 85). Our collective “embrace of limited government and dread of tyranny” might have helped to control the government but has also, he argues, left the government unable to “solve major problems or advance the public good,” turning it into a dysfunctional, or even “ ‘failed[,] state’ ” (pp. 83, 85). Levinson offers as an example the U.S. response to the COVID-19 pandemic. “While other rich countries implemented comprehensive public health strategies to combat the COVID pandemic, the United States ‘reacted . . . like a country with shoddy infrastructure and a dysfunctional government whose leaders were too corrupt or stupid to head off mass suffering’ ” (p. 83).

Here, Levinson assumes, without showing, that “[t]he failures of the country” are attributable to a constitutional design that divides authority.²⁶ The assumption is, in the Hobbesian conception, incontrovertible. When authority is divided among different organs of government, rather than consolidated in only one, it cannot as easily be used to settle what should be done or therefore have a practical effect. But as a matter of historical record, the assumption is false. Or at least, it reflects such a simplistic diagnosis of the cause of our failures as to be deeply misleading. Although the division of authority dates back to our founding, the United States has not always been so “crippl[ed]” a state (p. 85). In the post-World War II period, the United States has been among the most powerful, the most technologically savvy, the most economically dynamic, and the most culturally dominant of states. It has also displayed massively disproportionate power in shaping governance structures worldwide.²⁷

If by dividing authority we “intentionally design[ed] for ourselves a state built, if not to fail, at least not to fully succeed”—in other words, a “weaker” rather than a “stronger” state—what accounts for our state’s enormous strength along each of these dimensions (p. 85)? And if the division of authority is to blame for our dysfunction, why did our problems become so acute and seemingly irresolvable as authority, according to many experts, was increasingly centralized in the federal executive branch?²⁸ These questions warrant serious consideration before we summarily conclude, with Levinson and Hobbes, that “[f]or those who hold out hope that national state power might serve the public interest (or their own interests), . . . separation of powers is the enemy of effective governance” (p. 102). In fact, the separation of authority might once have contributed to our governance successes.

Second, given the frequency with which state power is abused, Levinson looks for law to control those in government who might use this power against us. Rights and votes are, according to him, “the two basic methods of controlling state power available to those who are subject to it” (p. 111):

If the goal is to protect subjects, or particularly vulnerable groups of them, against abuses of state power, that can be accomplished in two ways. One is by legally disallowing state power to be used in ways that threaten their fundamental interests, using rights. Another is by politically empowering groups of subjects to protect their own interests by influencing what the state will or will not do, using votes. (p. 111)

Levinson then analyzes how rights and votes work, “as comparable tools for performing the same basic task of controlling state power” (p. 113).

In analyzing rights and votes as two “methods of state control,” Levinson both breaks from the Hobbesian conception on form and retains the Hobbesian conception on function (p. 111). Contrary to the Hobbesian form, rights and votes place authority outside the state, permitting ordinary people to decide what should be done, rather than always submit to decisions that the state makes for them. But consistent with the Hobbesian function, Levinson posits that these forms of authority work insofar as they control their subject, which in this context is the state.

The analysis is flawed. Giving people authority to make their own decisions through rights or votes does not also give them the capacity to control the government. For example, if the government abuses its power by violating one of my rights, the right might (or might not) entitle me to advance a legal claim against it. However, for my right to affect what it does, it must itself recognize and take steps to realize the right. The right gives me a legal claim, but the government retains control—and with that control, the capacity to decide how to use its massive power in relation to me. Likewise, my exercising the right to vote might (but probably would not) affect who is elected into office. And our elected officials might (or might not) make decisions that, on some level, reflect my preferences or the preferences of the voting public at large. But the idea that I or that the American people can by voting control how the government uses its power strikes me as, at best, hyperbolic. Even putting aside the many ways that votes are negated in practice, their effect on the vast majority of governmental decisions is extremely attenuated.

Levinson on some level recognizes but lacks the tools to overcome the limits of his analysis. Even as he theorizes that rights and votes are “the two basic methods for controlling state power,” he harnesses facts that “raise some difficult questions about how rights against the state, regardless of substantive content, can be . . . made to work” (pp. 111, 185). For example, if rights and votes work to control state power, why is “nearly all influence over government decisionmaking . . . possessed by economic elites, whereas ‘mass-based interest groups’ and ‘average citizens’ have ‘little or no’ actual influence over policy outcomes” (p. 159)? Why does U.S. law betray such “categorical indifference to what many see and experience as the deepest injustices of American society” (p. 165)? And why do large majorities of Americans consistently report that they do not much trust the federal government to make the right decisions?²⁹ Without answers to these “difficult questions,” we have reasons to doubt that “rights work alongside votes to empower citizens to control the state” (pp. 185, 188). And insofar as they do not, we remain trapped, unable to identify how to use law to improve our lot.

Third, Levinson offers a proposal for making America more egalitarian. “Starting from the premise that ‘power can only be controlled by power,’ ” he advocates for using rights and votes to “distribute power more broadly and equally among social groups” (pp. 139, 157). For example, he argues that the wealthy should no longer be permitted “to leverage their economic power into disproportionate political power” and that Blacks and other members of minority groups should “be empowered to influence actual policy outcomes” in ways that they now do not (pp. 158–59). He recognizes the inherent difficulties in implementing this proposal. To start, we would need more accurate metrics for distinguishing the groups that are now “politically powerless” from the ones that “possess political power much greater than their numbers,” so that we can better balance power “among conflicting groups in society” and create the conditions under which each might more effectively control the others (pp. 150, 160).

I, too, would like the United States to be more egalitarian. But Levinson’s proposal for getting us there is ill-considered and potentially destructive. He nowhere grapples with the risk that pitting social groups against one another on the basis of their group identities might, in a polity as diverse as ours, aggravate already deep racial, economic, and cultural divisions. It might blind us to the range of experiences within groups or exacerbate certain forms of intragroup subjugation. Or it might augment the “impediments to effective governance” that he attributes to the division of authority—the “fragmented, disorganized, and conflicting governmental authority; politicized, denuded, and crippled bureaucracy . . . political polarization, [and] gridlock” (p. 83).

Moreover, even assuming that these risks could be mitigated, the devil would be in the details. As Levinson concedes, deciding “what ‘equality’ of political power should be understood to mean and how it should be operationalized” would raise “notoriously difficult and contested issues” (p. 158). Americans invariably would disagree about how to carve up the pie, presenting again the challenges of law without settlement and law without the sword. Who ultimately would decide the terms for redistributing power among us? How would that decision be enforced against those who are deemed to have too much power along whatever metric is deemed to be most relevant? And as long as we lack a clear vision of “what ‘equality’ of political power should be understood to mean and how it should be operationalized,” can we really be confident that inviting “conflicting groups in society” to try to control one another would revitalize, rather than further decimate, our democracy (pp. 150, 158)? After all, populist movements that insist on controlling those who are deemed to have exceeded their rightful stations in life sometimes do turn tyrannical.³⁰ Without more rigorous thinking on how we can use law to avoid that fate, we remain stuck in the trap.

II. The Conceptual Corrective

If we want to free ourselves from the Hobbesian trap—and exorcise his ghost, once and for all—we need to lose the false conceit that authority functions like coercion to control people. This conceit is baked into the Hobbesian conception.³¹ In that conception, the sovereign uses its authority and its coercion symbiotically together to control its people. The extent to which the former works independently of the latter is impossible to assess because the two are conjoined in the same source. Neither is there any practical reason to try to distinguish between them. All that matters, in this conception, is for that source to be able to use its law to control people in service of its vision of their common good.

Law for states explicitly severs authority from coercion by placing authority in sources who do not also have the levers of overriding coercion at their disposal. In practice, the two are severed even in what Levinson calls “law by states.”³² Judges who specialize in using the state’s authority do not themselves step down from the bench and whip the parties into submission. The state has other officials—other sources—who dole out its coercion. But as long as law by states is treated as the paradigmatic form of law, and control, the necessary function of authority will bleed into coercion, and its independent utility will be suspect. This is the conceptual mistake that Levinson and his cohort make. They assume that law has to control people to work, so they overlook how law works through its authority independently of coercion and without exerting control. The mistake is significant, because authority never really works to control people. It works first and foremost by doing what might reasonably be described as the opposite of control: empowering people.

A. Discarding the Hobbesian Form

Form and function remain connected. So, to identify what authority does, if not control people, we need a clearer sense of what authority is, if not an instrument for control. A starting definition, which should be uncontroversial, is that authority is: (1) a socially constituted, (2) entitlement, (3) to make governance decisions. This definition overlaps with one common in political and legal theory: Authority is the right to rule.³³ A right is a normative construct. Like other normative constructs, it has to be socially constituted to exist. When it is, it entitles the person who holds it to rule—that is, to make governance decisions.

Many theorists take the definition a step further and add that authority consists in the right to rule “over” other people. In this account, authority entitles one side in a relationship to make “binding” decisions that the other is “obliged to obey.”³⁴ However, this account again invites the conflation of authority with control. It suggests that the one who has authority uses it to get the other to subordinate her own judgment to his—and do as he tells her to do.³⁵

That dynamic of control is not, in my account, indicative of authority at work.³⁶ Because authority must be socially constituted to exist, no one agent can control what happens with it. It must be constituted through a social process in which multiple agents interact and decide what should be done (and by whom) in the situations before them. The exchanges that constitute it can be highly formal and deliberate, as when parties to a treaty give an arbitral body authority to interpret the text. Or these exchanges can be more passive and implicit in a set of interactions. For example, when I attend a faculty meeting and wait patiently for it to start, I by default give my dean authority to define the agenda. The constitutive exchanges can also occur in a wide range of settings, whether among officials of a single state, among the state’s officials and its people, or among people who do not define their relations by reference to a state. The key is that, because authority is socially constituted, it depends for its existence, its salience, and its content on what the disparate agents who put it at issue in their interactions decide to do with it.³⁷

Significantly, each side in the exchange both makes and permits the other to make a decision about what will transpire between them. One first claims authority for the decision that he wants to make, in effect asking the others to give it to him. He permits them to decide whether and on what terms he has it. He then decides either what to do with this authority or in light of the fact that he did not receive it. His authority extends only as far as they (actively or passively) grant it. In other words, to the extent that they do not permit him to make the decision at issue, he does not have authority from them for that decision. He might still make the decision, without regard for their views, but in that event, he acts without authority from them. He imposes his decision on them in a way that is reminiscent of control.

This account of authority is purposefully bare bones. It captures the essence of what authority is, without insisting that authority take any particular form. It breaks from the Hobbesian conception, because it highlights that authority and control represent fundamentally different ways to order social relations. Whereas an instrument of control enables one side to rule “over” another, overriding her independent judgment, authority is an instrument for ruling “with” her, inviting her to exercise that judgment. Moreover, my account exposes that authority cannot exist in any form that satisfies the Hobbesian conception. Authority cannot reside entirely with one agent, because it depends on what happens in the exchanges between disparate agents. There can be and often are difficult empirical questions about who has authority from whom to make which decisions. But as a conceptual matter, no one can unilaterally “settle” its terms for anyone else.

A simple hypothetical helps make the point concrete. A judge claims authority to decide that a defendant should be jailed. In the strictest iteration of the Hobbesian conception, authority can manifest only in the form of a sovereign command—represented here by the judge’s decision to jail the defendant. Many legal theorists who reject that strict iteration nevertheless assume, consistent with it, that authority resides with and manifests in forms that emanate from the state.³⁸ For example, Levinson assumes that states have authority in what he calls “law by states.”³⁹ In standard accounts of popular sovereignty, the populace channels its sovereign decisions through the state (pp. 67–74). Likewise, although rights and votes plainly place some decisionmaking outside the state, many assume that they are still constituted by state institutions and exist on whatever terms the state establishes.⁴⁰ And in international law, formal legal theory instructs that almost all authority comes from states, whether for each to make its own governance decisions or for groups of them to make decisions collectively.⁴¹ These accounts all assume, with the Hobbesian conception, that states have authority. If the assumption is sound, the judge in my hypothetical would have authority for her decision, so long as she acts within the terms that the state establishes for her, no matter the views of anyone else.

But the assumption is false. Regardless of what authority the state claims for the judge, her authority also depends on what others give. For example, if observers in her courtroom deny her the authority that she claims—if they decide that she is running a show trial or otherwise abusing state power—she does not have authority for the decision from them. And no amount of gavel-pounding can settle her authority for them or obligate them to give it to her. For the state, as represented by the judge, to have authority from its people, the state has to constitute its authority with its people. It cannot just impose its authority on them by passing laws that they have no choice but to obey; that way of interacting with them might create a dynamic of control but would be lacking in authority.

The same point can be illustrated at a larger scale of social organization. The U.S. president has what authority others give him. He routinely acquires authority from other officials within the U.S. government, as when Congress asks him to adopt particular regulations, when judges defer to his positions, or when agency personnel look to him for guidance in making decisions. He also acquires authority from sources outside the U.S. government, although formal legal theory struggles to recognize this authority as legal in kind. For example, the president acquires authority when the UN Security Council permits militarily powerful states to use force in other states; when media sources persuade viewers that he should exceed the bounds of what has formally been enacted in law; and when the masses show up at his rallies or march in the streets to support him. Thus, as a matter of fact, if not in formal legal theory, the president’s authority is dependent on and constituted with the people who give it to him. He has only what they allocate to him, on the terms that they give it to him. He cannot settle it for or impose it on them, as he would have to do if he were to use it to control them.

B. Departing from the Hobbesian Function

The principal function of authority in the legal enterprise is to empower, not to control, the people to whom it is addressed. It again consists in a grant of permission to decide what should be done in the situations before them. And as a general rule, people can more easily do things when others permit them to make the associated decisions, or when others participate with or support them in the endeavor, than when they have to do everything alone, without any social support or (worse) in the face of social resistance.

The Hobbesian conception also, on some level, recognizes that authority empowers people.⁴² According to it, authority empowers the state’s officials to make decisions for the state’s people. But in the Hobbesian conception, authority is a kind of power that always resides, in some form, with the state, and it functions by working alongside the state’s levers of coercion to control people. As Leslie Green explains: “Plan A is that [the] subjects should conform to it without further supervision.”⁴³ “Sanctions are the law’s Plan B.”⁴⁴

By contrast, in my account, authority does not function like coercion to control people. It is a kind of power that functions differently from, and in three critical respects antithetically to, the levers of coercion that actually can control people. First, coercion disempowers the people on its receiving end. It strips them of liberties that they might otherwise enjoy. Authority empowers them. It permits them to do what they otherwise could not, or could not as easily, do. Take my right to vote. The right is a form of authority that I have because (and insofar as) other Americans give it to me. It empowers me to do what I could not otherwise do—and what those who are not afforded it cannot do: walk freely into a polling station and cast my ballot for the candidates of my choice, without social incident. As such, it empowers without in any way controlling me. And as I explained earlier, neither does it empower me to control others. It is a kind of power that works without indicia of control.

Second, coercion works when one agent usurps from another the decision of what should be done. Authority works when each permits the other to make that decision. Recall the mechanics of how authority is constituted; each side in the exchange permits the other to make a decision about its allocation. Thus, when one unilaterally asserts the prerogative to make a decision, without permitting the other a say, he creates the dynamics of control and can no longer be said to have authority for the decision from her. Indeed, his assertion of control reveals an absence, not the presence, of authority. If he actually had authority for the decision from her, he would not need to disregard her views and override her. He would instead be able to make the decision with support from her.

Consider again the courtroom hypothetical. Without more information, we cannot know to what extent the defendant went to jail because he gave the judge authority to decide that he should or because he did not really have a choice, with the bailiff standing behind him, ready to cuff him. But the bailiff’s coercion sure seems to be doing the lion’s share to control him. We can better see the judge’s authority in the behavior of those who are not controlled—those who themselves engage in a constitutive exchange, involving both give and take, with her. For example, the bailiff empowers the judge to decide whom he should cuff; in exchange, she empowers him to cuff the defendant in plain view, rather than by sneaking around in the shadows like a kidnapper. Victorious plaintiffs empower the judge to decide what relief they are owed; in exchange, she empowers them to secure relief. And peer judges empower her to guide their thinking in subsequent cases; in exchange, she empowers them more effectively to justify their decisions to their audiences. In all of these examples, authority is divided among disparate agents who each decide what should be done with it in the circumstances before them. Authority is not consolidated in or used to control any one of them.

Third, coercion works through a zero-sum exchange between the one who uses it and the one on the receiving end of it. It empowers the former at the latter’s expense. Authority is additive. It empowers one while also empowering the other—although it does not empower each in exactly the same way. The judge’s authority empowers her, while also empowering the bailiff, the victorious plaintiffs, and the peer judges who give it to her. Each can do more because of it than they could have done without it. My right to vote is similar. My exercise of the right does not prevent anyone else from doing the same. To the contrary, the more Americans create the social conditions that empower me to vote, the more others benefit from these same conditions and can also vote. Moreover, as we each exercise this authority, we in turn give some authority to others. For our votes to be effective, we permit others to decide what counts as a vote, how we should register to vote, where we should go to vote, on what date we should vote, and so forth for the many micro-level decisions that need to be made to establish a functional voting system. As long as we are willing to give others in our ecosystem authority to make such decisions, we can create large-scale social structures, like the voting system, that depend on it. Alternatively, we can deny one another this authority and make the associated structures—with whatever benefits they carry—harder to sustain.

In short, the Hobbesian conception is wrong both on form and on function. Authority does not work by settling what should be done for, and in effect controlling, the people on the receiving end. And the fact that authority does not work like that does not mean, as the Hobbesian conception would lead you to believe, that authority does not work. Authority works by empowering people. It empowers me to vote, bailiffs to coerce defendants, plaintiffs to secure relief, and U.S. presidents to exceed the bounds of what any Congress prescribed. The more people use and permit others to use authority in their ecosystems—that is, the more they divide and share it amongst themselves—the more of them it empowers. By its very terms, it empowers each who acquires a piece of it to make the decisions that they think are best in the circumstances before them, even as they also empower others to do the same. Thus, unlike the levers of coercion, which work to control and subjugate people, authority, as I have described it, creates the potential to emancipate them.⁴⁵ It is the secret ingredient for exorcising Hobbes’s ghost and escaping the false choice between law that controls and law that is inert.

Conclusion

The Hobbesian conception of law is erroneous. It is not just erroneous at the margins or for what Levinson calls “[l]aw for states” (p. 6). It is completely specious. It insists that the only possible kind of law is a kind in which authority—which most would say is an essential ingredient of law—does not work. Thus, those who adhere to this conception consistently overlook or downplay how law, through its authority, does work. Worse, they lack the analytic tools to diagnose when (and for whom) it might be working, when (and for whom) it might be perishing, and why we should care about the distinction.

The acute problem that we Americans now confront is not that our government is too powerful and insufficiently controlled. It might be both, but neither would be new—or necessarily a problem. The problem is that our government has increasingly become, along a critical dimension, weak. It faces a crisis of authority.⁴⁶ Put simply, too many Americans have lost trust in its capacity to address their priorities or are angry at it for what they perceive to be its failures. The sentiment has become sufficiently strong and widespread that we are, across many issue areas, miserly about giving it the authority that it would need to make the very decisions that could, at scale, empower us. For example, we have not given it enough authority to help us mitigate the effects of climate change; to build and maintain public infrastructure; to achieve adequate levels of education, housing, or health care; to address the high levels of violence in public and private spaces; to exploit entrepreneurial opportunities; or to work in decent conditions for meaningful wages. To the extent that we have empowered the government to advance each of these projects, so that it could in turn empower us, we have done so tepidly, without the support that broad-based authority would confer and with significant social resistance that suggests an absence.⁴⁷

The crisis of authority is replicated beyond U.S. borders. In the decades since World War II, the United States worked with other states to constitute unprecedented levels of authority in international arenas. This authority empowered both it and those who supported and were supported by it—most consistently, but not exclusively, other states in the global north—to do things that they otherwise could not as easily have done, on projects ranging from food safety⁴⁸ to trade⁴⁹ to fishery regulation⁵⁰ and national⁵¹ and human security.⁵² Much of this authority has now atrophied and is actively being resisted.⁵³ Enough people around the world are sufficiently disenchanted by how the U.S. government has asserted its power that it can no longer harness the same levels of support for the decisions that it wants made––whether to satisfy its own agendas or to help others who rely on it to satisfy theirs.⁵⁴ Although the U.S. government still can use its military and economic power to coerce people into submission, it more often struggles to harness authority for its global governance decisions, including the decisions that involve some use of coercion.⁵⁵

This dynamic—in which the U.S. government is bleeding authority, despite moves that might reconstitute it with one or another constituency at a time—is not in my view a problem in the abstract. I think Levinson would agree. He ends the book with a thought experiment about “a world in which states have ceded their position of centrality and power to other actors” (p. 228). There is, we agree, appeal in trying to change how authority has historically been allocated, so that it can better be used for human emancipation. Moreover, neither of us sees reason in the abstract for states always to be the principal loci of authority. States can effectively use authority to empower people, but they often also use coercion to control people, and they even more regularly neglect the people whom they purport to serve. Depending on the issue area, we might do better by allocating authority to local or global institutions, to nonstate entities, or to individuals to decide for themselves what to do. The U.S. government’s declining authority thus presents opportunities to create new governance structures with different allocations of authority that better realize law’s emancipatory potential.

We cannot effectively seize these opportunities, so long as we are trapped by the Hobbesian conception, because the only options in that conception are for law to be controlling (as authority cannot be) or inert. These are also Levinson’s options. Even as he strips law for states from the formal trappings of sovereign states, he insists that law must function like the law of sovereign states. As such, his proposals for reform are thirsty for control, deficient in authority, and if the goal is to make law more emancipatory, unlikely to work. We could eliminate the division of authority and consolidate decisionmaking in one place to create the conditions for control. But even if we could identify the right decisionmaker for every domain of human activity, why should we think that she would make decisions that are good for the rest of us? We could try to control her with rights or votes. But rights and votes do not effectively control people. We could also pit social groups against one another, such that each tries to control the others. But if we do that, we might descend into lawless violence. “In such condition,” Hobbes warned—and here, he could at least rely on historical memory—

[T]here is no place for Industry; because the fruit thereof is uncertain: and consequently no Culture of the Earth; no Navigation, nor use of the commodities that may be imported by Sea; no commodious Building; . . . no Arts; no Letters; no Society; and which is worst of all, continuall feare, and danger of violent death . . . .⁵⁶

No doubt, some want to live in a fantasy in which a Leviathan swoops in to save us. Others want to be Leviathan so that they can impose their views on us. However, those who are serious about creating a more emancipatory future do not have the luxury of fantasy or control. Our task is to find ways to constitute authority for new governance structures that have the broad-based support they need to be both effective and justifiable. This will be exceedingly difficult to do. It cannot be done in the top-down manner of control. It has to be done with the people whom one wants or needs involved in the enterprise, because it depends on their participation to create something different.⁵⁷ They need to be willing to engage in the constitutive process, with both give and take, for the benefits of empowering and being empowered by others. The more they distrust one another to make socially relevant decisions, vilify those who hold opposing views, or insist on claiming full control for themselves, the harder the task will be. But if we do not succeed at it, we are very unlikely to improve the human condition. A world with decrepit authority is a more brutal world, in which competing groups more often vie for control and coerce into submission those who stand in their way.⁵⁸ Thus, the question that America and the world now confront is how far down this path of violent coercion and control we are willing to go before enough of us say mercy and again look for opportunities to empower one another, despite our innumerable differences, so that we can build something better.

* William S. Beinecke Professor of Law, Columbia Law School. I am grateful to Kate Andrias, Jacob Katz Cogan, Harlan Grant Cohen, Kostiantyn Gorobets, Don Herzog, Julian Davis Mortenson, and Samuel Moyn for comments on earlier drafts.