Online Colloquium (5): Reply to Critics by Sean Fleming
This online colloquium has been established to discuss Sean Fleming’s recent book, Leviathan on a Leash: A Theory of State Responsibility. We began with an introduction to the text by Dr Fleming, followed by responses from Silviya Lechner, Jerónimo Rilla and Philippe Crignon. We conclude this week with a reply by Sean Fleming. Many thanks to Princeton University Press for supporting this colloquium.
I sincerely thank Silviya Lechner, Jerónimo Rilla, and Philippe Crignon for their thoughtful responses to Leviathan on a Leash, and Robin Douglass for organizing this Colloquium. For the reader’s sake, let me begin with a roadmap. The respondents focus on three different aspects of the book. Lechner focuses on the kind of theory I develop: is it primarily normative or metaphysical? Rilla focuses on my interpretation of Hobbes’s theory of the state: does it stand up to scrutiny? Crignon focuses on the implicit premises of my Hobbesian theory of state responsibility: what does it assume or presuppose, especially about international law, and how Hobbesian is it really?
Lechner: Authorization and Representation
The crux of Lechner’s critique is that I hollow out the normative content of Hobbes’s theory of the state. I put too much weight on Hobbes’s account of representation and give short shrift to his account of authorization. ‘Throughout the book’, she says, ‘the emphasis is on the metaphysics of the state (questions of identity and continuity) rather than on normative analysis’.
I do not think the book is as tilted toward metaphysics, or away from normative analysis, as Lechner suggests. Metaphysics, very broadly defined, is the focus in three chapters: Chapter 1, which compares and critiques the two dominant theories of state responsibility; Chapter 2, which reinterprets Hobbes’s idea of state personality; and Chapter 4, which develops an account of state identity. The other two main chapters are primarily normative. Chapter 3 addresses the question of when actions should be attributed to the state, while Chapter 5 addresses the question of when the costs and burdens of the state’s responsibilities should be distributed to its subjects. Authorization, not representation, is the focus in both of these chapters. The bulk of Chapter 3 is about what it means for a government to be authorized, and the central concept in Chapter 5 is ‘authorization by fiction’. Even in Chapter 4, which is the most metaphysical of them all, the metaphysics is derivative of the normative analysis. In that chapter, I use Hobbes’s idea of succession to explain how a state can persist over time despite changes in its population, territory, and government. For Hobbes, a multitude of individuals becomes one person when the members of the multitude authorize a common representative; it remains one person as long as it has a continuous series of representatives, or an unbroken chain of succession. However, as I emphasize, not just any representative can unite a multitude; only an authorized representative can. The reader first has to accept the normative account of authorization that I develop in Chapter 3 in order to accept the ‘metaphysical’ account of state identity that I develop in Chapter 4.
Here, by the way, lies the answer to one of Lechner’s questions: ‘what is the relation between authorisation and representation?’ On my view, as on Hobbes’s, authorization is the normative relation that makes representation valid or legitimate. To authorize an actor is to grant that actor the right to represent: ‘done by authority’ means ‘done by Commission, or Licence from him whose right it is’. Authority thus marks the difference between representatives and misrepresentatives—between those who ‘personate’ and those who impersonate.
Although Leviathan on a Leash is more normative than Lechner suggests, there is something to her critique. She is right that the normative part of the bookis not nearly as Hobbesian as the metaphysical part. I take up Hobbes’s accounts of representation and personhood, but I jettison much of his account of authorization. As I said in my Introduction to this Colloquium, and as I argue at length in Chapter 3 of the book, Hobbes’s account of authorization has to be extensively modified to fit contemporary politics. Political authorization need not be unanimous, unlimited, or irrevocable, and the state need not have a single locus of authority—a supreme, sovereign representative. Through a critique of Hobbes, Chapter 3 develops a new account of political authorization, which begins with the thought that authorization requires an authentic expression of an agent’s will. The normative part of the book owes as much to Jean-Jacques Rousseau and Bernard Williams as it does to Hobbes.
So far, I have followed Lechner in using the word ‘metaphysics’ to refer to the non-normative aspects of the theory that I develop. But I should emphasize here, as I do in the book, that one of the advantages of my theory of state responsibility over the alternatives is that it is metaphysically thin. My Hobbesian theory of state responsibility provides a way of making sense of how states can be held responsible without positing corporate wills, agents, or intentions. The familiar concepts of authorization and representation do most of the theoretical work.
Lechner also asks where, if anywhere, I diverge from David Runciman. Let me first list the three points on which I follow him: (1) the state is ‘represented by Fiction’, like a bridge or an idol; (2) Hobbes’s triangular or triadic model of representation is essential for understanding the modern state; and (3) Hobbes’s account of representation can be separated from his account of authorization. The third point is the one on which Lechner diverges from Runciman and I. Whereas we think the structure of Hobbes’s theory of the state can be separated from its normative content, she argues that Hobbes’s thought should be understood ‘as a system of propositions’ (emphasis in original). I agree with Lechner that Hobbes’s thought should be interpreted holistically, which is precisely why I read his political and theological accounts of representation and personhood together. But it does not follow that his concepts are inextricable from his philosophical system, or that they all stand or fall together. Ideas from one philosophical system can often be redeployed within another. The history of political thought is, in large part, the history of how political thinkers have selectively borrowed ideas from others. We can separate Hobbes’s account of representation from his account of authorization, just as he separated the idea of representation from parliamentarianism.
My interpretation of Hobbes diverges from Runciman’s in only one major way. He makes the mistake (as many others do) of trying to compress Hobbes’s usage of ‘person’ into the definition of ‘person’ as ‘representative’ at the beginning of Chapter 16 of Leviathan. As I explain in Chapter 2, this renders Runciman’s reading of Hobbes inconsistent: ‘On one hand, he maintains that all persons are representatives. On the other, he maintains that states (and other incapable entities), which are not representatives, are nevertheless persons’ (51). He reaches the right conclusion—that Hobbes’s state is a fictional character—through a faulty line of reasoning. By showing that Hobbes’s concept of personhood is actually double-sided, and that persons can be ‘representees’ as well as representatives, I resolve this crucial inconsistency in Runciman’s interpretation.
Rilla: Personhood and Agency
Rilla accepts my argument that Hobbes’s state is like a fictional character. The state has to be represented by an actor because it cannot speak or act on its own. But Rilla maintains that Hobbes’s state is nevertheless an actor or ‘agent’ in its own right. Since Rilla is not persuaded by the textual evidence to the contrary that I present in Chapter 2, I will focus here on the conceptual problems with his interpretation.
Before I respond to Rilla’s arguments, let me identify the main point on which we disagree. I argue that Hobbes’s state is not an actor because it does not have a distinct will. The will of the state is nothing more than the will of the sovereign: ‘a Common-wealth hath no Will, nor makes no Lawes, but those that are made by the Will of him, or them that have the Soveraign Power’. The actor, or the source of will, is the sovereign; the state is a character, merely a passive vehicle for the sovereign’s will. Rilla replies that the state is an actor, because its will is not reducible to the sovereign’s will. What he and I disagree about is whether Hobbes’s state has a will that is distinct from that of the sovereign. The italicized part of the last sentence is crucial. I do not deny that Hobbes ascribes a will to the state (as in the passage I quote above); what I deny is that the state’s will can differ or diverge from the sovereign’s will. I say this only for the benefit of the reader, since Rilla understands my argument well. Throughout his response, he is careful not to make a straw man out of me. He develops the most thoughtful critique of my position that I have yet encountered.
Rilla’s first argument is that the state must have a distinct will, because otherwise it would be difficult to distinguish appropriate from inappropriate ways of representing the state. If the state’s will is nothing more than the will of the sovereign, then how would it be possible for the sovereign to play the role of the state well or badly? For the sovereign to represent the state appropriately is, Rilla says, for the sovereign to ‘conform to the will and action proper to the person of the state’, which is ‘different from his or her natural attitudes’. Thus, he argues, ‘a set of intentions are attached by default to the person of the state: if a sovereign grants liberties that undermine his or her authority to a subject, “it is to be understood it was not his will” (L, XXI.20, 342)’.
The state need not have a distinct will in order for there to be ‘appropriate and inappropriate ways of personating it’. By Rilla’s logic, Hobbes must also ascribe distinct wills to bridges and idols, because there are appropriate and inappropriate ways of representing them, too. The point of Hobbes’s examples of representation of incapable entities—’Children, Fooles, and Mad-men’, ‘Inanimate things, as a Church, an Hospital, a Bridge’, and ‘An Idol, or meer Figment of the brain’—is that the represented entity need not have any intrinsic capacity for will, action, speech, or rationality. The same point applies to the state, which is implied by the parallel Hobbes constructs between representation of incapable entities and representation of a multitude. It is Rilla who ‘fails to notice an important trait of this dramatis persona’, and of dramatis personae in general. Representation is constrained not only by the will of the represented person (since it may not have one of its own), but also by the actor’s need to play the role of the person in a plausible way. The plausibility of representation is determined externally by an audience as well as internally by the will of the represented person. Thus, the fact that the state’s role can be played well or badly does not imply that the state has a will distinct from that of the sovereign, any more than the fact that Robin Hood’s role can be played well or badly implies that Robin Hood has a will distinct from the actor who plays his role.
At this point, Rilla might fall back on the word ‘fiction’: ‘To be sure, this intentionality concerns the state as a person by fiction, and not in metaphysical terms, as a mental event of an emergent mind’. But to distinguish representation from misrepresentation of the state, it is no more necessary to posit a distinct fictional will of the state than it is to posit a real one. The argument in the previous paragraph holds even if every instance of ‘will’ is qualified by ‘fictional’.
Rilla’s next objection is interesting and formidable. As I argue in Chapter 2, Hobbes did not want the state to have a distinct will, because this would raise a seditious possibility: the subjects might object that the sovereign has misrepresented the will of the state. However, as Rilla points out, Hobbes also saw a danger on the other side: ‘If the will of the state boils down to the natural will of the sovereign, subjects may feel alienated or disaffected from this person and disavow its actions’. This is certainly true, but Hobbes’s solution to this problem was not to posit a free-standing will of the state. Instead, he uses the concept of authorization to close the gap between subjects and the sovereign. Since subjects have authorized the sovereign, they should see his actions as their own: ‘by this Institution of a Common-wealth, every particular man is Author of all the Soveraigne doth; and consequently he that complaineth of injury from his Soveraigne, complaineth of that whereof he himselfe is Author; and therefore ought not to accuse any man but himselfe’.
Rilla later raises a related point. My metaphysically thin interpretation of Hobbes’s state, he says, ‘leave[s] it defenceless against other personifications perceived as speaking and acting through rebellious representatives (e.g. Liberty, the Holy Ghost, idols or the People mobilised by the Parliamentarians)’. It is true that Hobbes could have countered all claims against the state with the argument that only the state is ‘real’ and that all of these other entities and ideas are fictitious or illusory. But that is not what Hobbes does—for one thing, because it would have been suicidal to dismiss God and the Holy Ghost as fictitious. Instead, he relies on authorization to establish a hierarchy of fictions. The state is the arch-fiction: the only one whose representative is authorized ‘without stint’ by all subjects. The sovereign is thus authorized to keep all of the other fictions in their places. Corporations, the Church, and God can be represented only as the sovereign permits. ‘Liberty’, like all other terms, is defined by the sovereign. And since ‘the people’ and ‘the state’ are the same fictional entity, represented by the sovereign, it is conceptually impossible for the people to act in opposition to the state (see below). Hobbes did not need a metaphysically thicker theory of the state.
Rilla also objects to the distinction that I draw between true representatives (monarchs) and fictional representatives (assemblies). He points out that assemblies are natural persons according to Hobbes’s definition in Chapter 16 of Leviathan, since they are actors capable of representing themselves. Rilla is undoubtedly right about this. But here, as elsewhere, I think it is a mistake to rely solely on Hobbes’s account of personhood from Chapter 16 of Leviathan. In Chapter 22, he refers to assemblies as ‘artificiall, and fictitious’: ‘if it [i.e. an act] be a crime, the Assembly may be punished, as farre-forth as it is capable, as by dissolution, or forfeiture of their Letters, (which is to such artificiall, and fictitious Bodies, capitall,)’. In The Elements of Law, again using ‘body politic’ to refer to an assembly, Hobbes says that ‘a body politic, as it is a fictitious body, so are the faculties of will thereof fictitious also’. On my view, Hobbes’s state is a fictional character, whereas a Hobbesian assembly is a fictional actor.
Rilla also takes issue with my claim that what makes Hobbes’s idea of state personality novel and valuable is ‘that it decouples personhood from agency’ (65). Historically, he points out, the idea of the state as fictional person predates Hobbes. Rilla is surely right. But my claim in the offending passage is not that Hobbes’s idea of state personality is historically novel, but that it is novel in relation to contemporary conceptions of the state. As I say in the very next sentence, Hobbes’s decoupling ‘allows us to sidestep the protracted debates about the metaphysics of corporate agency and intentionality’—debates in contemporary political theory and philosophy (65). On the question of how historically novel Hobbes’s state is, I follow Quentin Skinner, as Rilla does: ‘More clearly than any previous writer on public power, Hobbes enunciates the doctrine that the legal person lying at the heart of politics is neither the persona of the people nor the official person of the sovereign, but rather the artificial person of the state’. What Hobbes offers is not an entirely new theory of the state, but an especially clear and powerful articulation of a theory that contemporary political theory has misunderstood or forgotten.
Rilla concludes that Hobbes’s state is, contrary to my claim, a ‘fictional agent’. But this leaves him, and anyone else who thinks Hobbes’s state is an actor or agent, unable to make sense of Hobbes’s theatrical metaphor. If the state is the actor, then who or what is the character that the state represents?
One possibility is that the state is both the actor and the character; the state represents itself. Although Rilla maintains that Hobbes’s state is an actor, he says at the outset that he accepts my claim that Hobbes’s state is a fictional character. So he must think it is both. This is conceptually possible: the state (assuming that it is an actor) could play its own role, much as celebrities ‘play themselves’ when they make cameos in movies (e.g., Bob Barker played Bob Barker in Happy Gilmour). But what is the textual evidence that Hobbes understood the state as both actor and character? And what purpose could this self-representing fiction possibly serve in Hobbes’s political thought?
Another possibility is that the state (as actor) represents the people (as character). This could be what Rilla is suggesting here: ‘Hobbes attributes personhood and agency to the state because as long as it (and not merely the sovereign) acts, we-the-people act. Either channelled by a monarch or an assembly, it is “the action of the people” (L, XI.20, 158) that is at stake’. Rilla’s reference to ‘personhood and agency’ is odd, because it seems to affirm precisely the distinction that I insist on. More importantly, the relation between the state and the people is not merely representation, but identity. The state is the people. As Hobbes says in De cive, ‘Ordinary people and others who do not notice this point, always speak of a large number of men as the people, i.e. as the commonwealth’. Hobbes makes the same point in the passage of Leviathan from which Rilla quotes above. Subjects ‘are disposed to take for the action of the people [i.e. the state], that which is a multitude of actions done by a multitude of men’.
It may be that Rilla has taken the identity between the state and the people into account. Maybe this is his thought: ‘as long as [the state] acts, we-the-people act’, precisely because the state is the people. But if the state represents the people, and the people is the state, then the state represents itself. Once again, Hobbes’s state is back to being a self-representing fiction—both actor and character.
In sum, those who interpret Hobbes’s state as an actor have yet to provide a compelling answer to this crucial question: who or what is the corresponding character? Unless they can reconcile the state-as-actor interpretation with Hobbes’s theatrical metaphor—not to mention the textual evidence that I present in Chapter 2—they do not have a compelling rebuttal to my state-as-character interpretation.
Finally, Rilla argues that recognizing the state as an agent (susceptible to culpability, and subject to norms of rational consistency) would help to make sense of how Argentina has taken responsibility for the atrocities of its last military regime. I cannot adequately address this case here, so I will only offer one point of clarification. Rilla points out that, ‘[i]nstead of alleging “misattribution” due to the unauthorised character of the dictatorship, the democratic government that took office afterwards acknowledged the crimes as acts of “state terrorism”’. His assumption is that, on my account, the atrocities of the military regime are not attributable to Argentina. On the contrary, I think these atrocities are attributable to Argentina. Although the military regime was repressive and murderous, it probably met the minimal threshold for authorization—namely, acceptance as legitimate by a substantial number of subjects. As I argue in Chapter 3, ‘We must not fall into the trap of thinking that an authorized government is necessarily a “good” one’ (89). If only democratic, rights-respecting, or ‘decent’ governments count as authorized, then authoritarian states will never be responsible for anything, because the actions of repressive governments will never be attributable to the state. The state thus becomes an ‘artificial angel’. The claim that only good governments count as authorized backfires badly when it comes to state responsibility.
Crignon: Hidden Assumptions and the Law of Nations
Crignon begins by noting what is not Hobbesian about Leviathan on a Leash. He argues that my book depends on three premises or assumptions that Hobbes did not accept: ‘(1) the existence of an international forum where states acknowledge each other as persons, (2) a system of international law which introduces rules to which states are subjected and (3) human rights limitations on state sovereignty’. Crignon is right that my theory of state responsibility presupposes a system of international law—one in which the ‘law of nations’ is more strictly binding than Hobbes’s ‘law of nature’. Because Hobbes thought states were ‘Absolute, and Independent, subject to none but their own Representative’, he did not think they could be bound—not to each other, nor to their own citizens. I make room for state responsibility by modifying Hobbes’s account of authorization and casting off his absolutist theory of sovereignty. However, I think Crignon is wrong about the first and third premises. My theory of state responsibility does not assume or require an international forum of recognition or human rights limitations on sovereignty.
In relation to the first premise, Crignon argues that ‘[t]he idea that there is a common arena where states can acknowledge each other as peers is implicit’ in the book. On his reading, I am committed to the claim that ‘it is possible [for a state] to be a person regardless of the authorization/representation process’. The idea that he attributes to me is called the ‘constitutive theory of recognition’ among international lawyers: external recognition constitutes the state. But in Chapter 4, I actually defend the ‘declaratory theory of recognition’, which holds that the existence of the state is independent of external recognition: ‘The corporate identity of the state is not externally constituted by the recognition of other states; it is internally constituted by the state’s representatives and by the subjects who authorize them’ (143, emphasis in original; see also 130, note 10). Earlier in that chapter, I use the cases of Poland and Ethiopia to demonstrate the ‘irrelevance of external recognition to the continuity of the state’ (130). I argue that the Polish Republic and Ethiopia continued to exist for years despite an almost-total absence of external recognition (and total annexations of their territories and populations), because these states had representatives who continued to speak and act in their names. My account of state identity is, if anything, too dismissive of the role of external recognition in constituting the state.
Crignon next argues that my understanding of state personality is more Leibnizian than Hobbesian: ‘It is Leibniz, not Hobbes, who construed the idea that a state is a person in an international society, a “persona juris gentium”’. My view is certainly Leibnizian in that I see states as persons in relation to each other, whereas Hobbes saw states as persons only in relation to their own subjects. But if Leibniz argues that the ‘personality of the state is unrelated to the authorization and representation process’, then, in this respect, my understanding of state personality is decidedly Hobbesian. There is no contradiction between my understanding of states as ‘international’ persons (which is Leibnizian) and my claim that the personality of the state is constituted ‘internally’ by authorization and representation (which is Hobbesian). The fact that states are persons in relation to each other does not imply that their personhood is constituted by external recognition. Similarly, the fact that human beings are persons in relation to each other does not imply that their personhood is constituted by external recognition (rather than by some pre-social feature of the individual). How persons relate to each other is one question; how persons are constituted is another.
My theory of state responsibility also does not assume ‘human rights limitations on state sovereignty’. It does require limitations on sovereignty, but these limitations derive from authorization rather than from human rights. Whereas Hobbes thought political authorization had to be unanimous, irrevocable, and limitless, I argue that political authorization is always partial, temporary, and conditional (69, 85–88, 90–92). Authorization is temporary for two reasons: first because authority has a short shelf-life (which is why we hold elections every few years), and second because the authors themselves (subjects) have limited lifespans. Authorization is partial because there are always dissenters who refuse to authorize the government, as well as some subjects (such as young children) who do not have the capacity to authorize anyone. Authorization is conditional because it can be withdrawn; authorizing subjects can later become dissenters. It is for these reasons that sovereignty is limited. On my account, governments that do not respect human rights often do count as authorized. We might consider these governments to be ‘unjust’, but we should be careful about condemning them as ‘illegitimate’ or ‘unauthorized’: ‘The claim that only rights-respecting governments can be authorized implies that states cannot be responsible for violations of human rights’ (89). As I said at the end of my response to Rilla, the claim that only good governments count as authorized recoils on us when it comes to state responsibility.
Crignon argues that some of my examples depend on a conception of the state is that is un-Hobbesian: ‘A Hobbesian state can wage war, but cannot be bound by a covenant’. For Hobbes, ‘there can be no genuine treaty or compact between states’. I completely agree. As I argue in Chapter 2, Hobbes’s ‘theory of sovereignty rules out the possibility that states could be held responsible’ (68). Hobbes does not think states can be obligated (in a legal sense, at least) because he does not think they are subject to any higher authority; states cannot be bound (to each other, or to their own subjects) because there is no one to bind them. This is why I jettison Hobbes’s theory of sovereignty. My conception of the state is structurally Hobbesian, since I ‘retain the structure of Hobbes’s theory of representation’ (69), but Crignon is correct it is no longer Hobbes’s conception of the state after I am done with it.
Crignon says he ‘cannot see why the state cannot be accountable for its actions or even culpable for them, in addition to owning them’. Let me restate, and hopefully clarify, the arguments that I made in Chapter 3.
The conclusion that accountability resides with the state’s representatives, rather than with the state itself, follows from my earlier argument (discussed in my response to Rilla) that the state is a fictional character. Accountability, in common parlance and by my definition, is ‘an obligation to explain or justify’ an action—an obligation to answer for it (99). Since the state cannot speak or act on its own, it is simply incapable of real accountability. The task of answering for actions that are attributed to the state falls to its representatives. Public apologies may be issued in the name of the state, but they are always issued by the representatives of the state. Accountability thus resides with them.
The conclusion that culpability resides with the state’s representatives follows from an asymmetry between intention and action. Culpability requires both a wrongful act and a corresponding intention—in the terms of criminal law, an actus reus and a mens rea. But while actions can be attributed from a representative to a representee, intentions cannot be. To take a simple example, which I use in Chapter 3, suppose that I authorize an estate agent to buy a house for me. I am indifferent about the colour of the house. But the estate agent happens to like red houses, so she buys a red house for me. Since I authorized the estate agent, her action is attributable to me, but it would be a mistake to attribute her intention to me: ‘I bought a red house’ is true even though ‘I intended to buy a red house’ is false. Here lies the asymmetry between action and intention. The actions of an authorized representative are attributable to the person whom she represents; her intentions are not.
Now consider a case of wrongful action. Suppose that the prospective buyer of my old house requests a copy of the deed and the property survey, and I send these documents to him via my estate agent. Unbeknownst to me, these documents are forgeries. My estate agent recognizes this, but she is desperate to make the sale, so she keeps quiet and sends the forged documents to the prospective buyer. My estate agent is clearly culpable: she passed forged documents, and she intended to do so. Am I culpable, too? I think not. It is true that I passed forged documents; this act of my estate agent is attributable to me, since I authorized it. But it is not true that I intended to pass forged documents. The fact that my estate agent intended to do so does not imply that I intended to do so (unless I was aware of the forgery). As in the previous example, her action is attributable to me, but her intention is not. And since intent is necessary for culpability, I am not culpable.
Now consider the state. Since the state is a fictional character, it has no intentions of its own. If the state is to be culpable, the intentions of its representatives have to be attributed to it. Suppose that a soldier intentionally kills a civilian. On my account, this act is attributable to the state, because ultra vires acts of state officials still count as acts of state. However, the intention of the soldier is not attributable to the state, because intentions are not attributable at all. Here, again, the asymmetry between action and intention presents a barrier to vicarious culpability. Those who think the intentions of state officials should be attributed to the state have to explain why the intentions of representatives are not attributable to representees in other cases, such as the ones I describe in the previous two paragraphs. If the estate agent’s intention to pass forged documents or to buy a red house is not attributable to her client, then why should the intention of the soldier be attributed to the state? Why is vicarious liability so common but vicarious culpability so rare? My answer is simple: because it is possible to act vicariously but not to intend vicariously.
Crignon argues that ‘[t]he very notion of “doing wrong” or “being unjust” entails … culpability’. I do not think so, for the reasons that I describe above. A wrongful act can be attributed to a state, but a wrongful intention, which is necessary for culpability, cannot be. ‘Ownership’ and culpability can come apart. It is worth noting here that, under international law, states can be held responsible for wrongful acts but not criminally responsible for them. The UN International Law Commission’s Articles on Responsibility of States for Internationally Wrongful Acts say nothing about culpability; guilt, punishment, and criminality are conspicuously absent. States have reparative obligations under international law even though they cannot be culpable under international law.
Finally, Crignon asks precisely what role authorization and representation play in my Hobbesian theory of state responsibility. The answer that he suggests is mostly right: authorization and representation determine whose actions are attributed to the state and to whom the costs and burdens of the state are distributed. But I do not agree with his suggestion that authorization and representation play only ‘a minor role in the constitution of the state’. Once again, my account of state personality is more Hobbesian (and less Leibnizian) than Crignon suggests. For me, as for Hobbes, the personality of the state is a product not of external recognition, but of political representation.
Dr Sean Fleming
(Christ’s College, University of Cambridge)
 Leviathan, XVI. 244, emphasis in original. I cite Hobbes’s Leviathan according to the chapter numbers and the page numbers from the 2012 Clarendon edition, edited by Noel Malcolm.
 Leviathan, XVI. 248.
 See especially David Runciman, ‘Hobbes’s Theory of Representation: Anti-Democratic or Proto-Democratic?’, in Political Representation, eds. Ian Shapiro et al. (Cambridge: Cambridge University Press, 2009), pp. 15–34.
 Leviathan, XXXI. 570. See also De cive, VI.19 and VI.1a. I cite De cive by the chapter and paragraph numbers.
 Leviathan, XVI. 246–48.
 On this point, see Arash Abizadeh, ‘Hobbes’s Conventionalist Theology, the Trinity, and God as an Artificial Person by Fiction’, The Historical Journal 60, no. 4 (2017), pp. 915–41, at 926.
 Leviathan, XVIII. 270.
 Leviathan, XVI. 250; see also XXII. 348.
 Leviathan, XXII. 352.
 The Elements of Law, Natural and Politic, XXI.4. I cite The Elements by the chapter and paragraph numbers. On this passage, and on the role of fiction in Hobbes’s thought more generally, see Robin Douglass, ‘The Body Politic “is a fictitious body”: Hobbes on Imagination and Fiction’, Hobbes Studies 27, no. 2 (2014), pp. 126–47.
 Quentin Skinner, Visions of Politics Volume 2: Renaissance Virtues (Cambridge: Cambridge University Press, 2002), p. 404, quoted on p. 9 of Leviathan on a Leash.
 De cive, XII.8, emphasis in original.
 Leviathan, XI. 158.
 Leviathan, XXII. 348.
 As I explain elsewhere, the UN International Law Commission rejected a proposal to recognize international crimes of states. Sean Fleming, ‘Moral Agents and Legal Persons: The Ethics and the Law of State Responsibility’, International Theory 9, no. 3 (2017), pp. 466–89. I also discuss state criminality in ‘Leviathan on Trial: Should States Be Held Criminally Responsible?’, International Theory (forthcoming).