Anyone puzzled by what sort of entity a state could possibly be, or how responsibilities could possibly be attributed to such an entity, will find much of interest in Sean Fleming's Leviathan on a Leash: A Theory of State Responsibility. In lively prose, with contemporary illustrations, Fleming develops Runciman's approach to understanding states on a Hobbesian model. Even those unpersuaded by Fleming's theory should appreciate the way he taps literatures across the disciplines of philosophy, political theory, international relations, and international law, as either supports or foils for his arguments, while describing a metalandscape of theorizing that should prove useful for organizing one's own inquiries into state responsibility.

Fleming's aim is to develop a theory of state responsibility that vindicates existing practices under international law that attribute treaty, debt, and reparative obligations to states rather than to their governments and that require subjects to bear the costs of fulfilling those obligations even when subjects have no say in, control over, or culpability for the actions that generated them. Fleming describes his intended apologia for burdening subjects with the costs of actions done by previous governments in the name of the state as an exercise in realpolitik and “nonideal” theory, largely because it seeks to accommodate what he acknowledges are unjust, indeed “tragic” practices.

His central diagnosis of the inability of the main theories of state responsibility—the “agential” and the “functional”—to account for what he takes as facts is that they lack the triadic structure that he, following Runciman, deems needed to explain the relation among the state, its government, and its subjects. The main theories of state responsibility address only the relation of government to state, omitting subjects. The “agential” theory understands the state as a corporate moral agent capable of deliberating and of forming intentions to act that are not reducible to the intentions of the individuals who comprise the government. Fleming objects that it is difficult to infer corporate intentions from the laws and policies of nondemocratic states, because these may instead reflect only the intentions of government leaders (108). More generally, he is skeptical that metaphysically unproblematic accounts of corporate intention— which he concedes hold for small participatory groups—“scale up” to large organizations in which not everyone is or has authorized an active participant. The “functional” theory understands the relation between a state and its government as a principal-agent relation. Actions of governmental organs are attributed to the state. Fleming objects that this theory does not distinguish authorized organs of the state from entities that do fulfill state functions but lack authorization, say, successful insurrectionists or usurpers. It would insulate corrupt officials and predatory governments from the stigma and costs of their misuse of power, something any acceptable theory of state “should” disallow. Here Fleming accidentally introduces a normative constraint from his personal morality in tension with his realpolitik project.

Fleming's proposed theory is that the state is a “fictional character” played by a “fictional actor,” the government. States exist only by virtue of being represented by governments “authorized” by subjects to do so. The government in turn authorizes officers and other functionaries to represent itself. The identity of a state is determined by its representative, and the continuity of a state over time depends on the mutual acknowledgment of predecessor and successor governments as representing the state. The costs of actions attributed to the state are properly distributed to subjects because subjects, including subjects not yet born, have authorized the government so long as their interests are somehow “present” in governmental deliberations. Actions by authorized government officials that are not “plausibly” done as part of their “proper governmental role” are solely the responsibility of those officials rather than of the state, and those officials rather than subjects should bear their costs. Some governments simply cannot be authorized, because they exercise too much coercion or indoctrination or limitation of access to information for people “freely” to consent to them. Nonauthorized governments cannot represent states, and states exist only if represented.

Sounds like a whole lot of ipsedixitism? Fleming's cloak of Hobbesian provenance for his theory borrows and centers (a radically revised notion of) Hobbes's concept of authorized representation, a concept Hobbes did without in the first three versions of his own theory of state. Fleming elaborately presents, as if he thought it his original discovery (but see, e.g., Paul Weithman, “Hobbes on Persons and Representation, in Interpreting Hobbes's Political Philosophy, ed. S. A. Lloyd [Cambridge University Press, 2019]), Hobbes's view in Leviathan and later works that we can count something as a receptacle for the attribution of actions in virtue of its being represented by another who acts in its name, even if it cannot itself be an agent. A temple or a bridge can be regarded as a “person” to which actions are attributed if its owner authorizes someone to act in its name. This appeals to Fleming because it makes possible attribution of actions to something that cannot act. A basic difficulty for Fleming in using Hobbes's idea is that no individual “owns” the state such that she can authorize a government to represent it. She could authorize a government to represent herself, but she has no moral authority to confer power on a government to coerce dissenters and others incapable of consenting, including future generations, or to act for any entity comprised of those others, whose rights are not hers to give away. If she cannot authorize a government to represent some fictional character to which the actions of every subject are attributed, it will be hard to see any justification for saddling her with liability for that fictional character's actions.

Fleming's effort to create a theory of state responsibility by mixing together (a) realpolitik accommodations of existing international legal practice with (b) the cachet of some Hobbesian concepts, though detached from the theoretical framework that gives sense to them, along with (c) appeal to Fleming's own personal moral intuitions, for example, about what ought to count as consent, results in a wobbly structure, taped together at the weak points with apparently ad hoc arguments that appeal variously to each of these competing sources as seems convenient. I will not speak to the plausibility of Fleming's personal moral intuitions, nor to whether he is correct in his understanding of what existing international practice requires. I will address just his misuse of Hobbes in a way that undermines Hobbes's central mission of resolving social disagreement. Fleming's positive theory introduces irresoluble disagreement at every point of application. Are the “background conditions” it requires for authorization satisfied? Have “enough” subjects authorized the government? Is the government “plausibly” acting within its “proper role”? Is the chain of representation sufficiently “continuous” and intergovernmental acknowledgments sufficiently “mutual” and “plausibly grounded” that the current state is the same state that incurred responsibilities? Are incapable subjects and future generations sufficiently “present” in the government's deliberations that they should be counted as having authorized it to undertake state responsibilities? How great a burden is “too great,” and how temporally far removed is “too far,” for current and future subjects to be held liable for paying the costs of the state's actions?

A Hobbesian diagnosis of the pervasive indeterminacy of Fleming's theory is that it lacks a coherent method, or principled approach to theory construction. Detaching Hobbes's concept of authorized representation from the theory that gives it sense leaves a vacuous, circular notion of statehood (the state is that which is represented by an entity authorized to represent the state) according to which, implausibly, anything could count as a state, and so necessitates Fleming's scramble to constrain its content by appeals to contingencies of existing international legal practice—however indefensible—and to his own, equally contingent, personal moral intuitions. Whereas Hobbes's leviathan is fundamentally a unique mechanism for the legitimate resolution of any possible dispute within a society at a time, Fleming's state is a formless fiction itself the object of irresoluble disputes. Make what you will of Fleming's theory of the state, but let us not implicate Hobbes in it. That theory yields, not Leviathan, nor Leviathan on a Leash, but at most, Behemoth in a Bog.