States as moral agents: the agential theory

The core idea of the agential theory is that states can be held responsible for the same reasons that we hold human beings responsible. Goodin argues that ‘the state is a moral agent, in all the respects that morally matter’ (Reference Goodin1995, 35). The state, ‘like the natural individual, is capable of embodying values, goals and ends; it, too, is capable (through its legislative and executive organs) of deliberative action in pursuit of them’ (Reference Goodin1995, 35). Erskine (Reference Erskine2001, 69–70) argues that the disanalogy between states and human beings ‘is often over-stated’ and that states are ‘capable of acting and knowing in a way that is analogous—but not identical—to that of most individual human beings’. Because states are capable of deliberating and of acting intentionally, they are ‘moral agents in the same way that we understand most individual human beings to be moral agents’ (Erskine Reference Erskine2008, 2).Footnote ²

The agential theory was developed by philosophers, most notably French (Reference French1979, Reference French1984, Reference French1995, Reference French1998), and later adopted by political theorists and IR scholars.Footnote ³ French argues that certain groups, called ‘conglomerate collectivities’, are moral agents over and above their members and can therefore be held responsible separately from their members.Footnote ⁴ Conglomerate collectivities have two defining features: (1) corporate identities that do not depend on determinate memberships and (2) corporate internal decision (CID) structures. First, unlike ‘aggregate collectivities’ such as a mobs and crowds, conglomerates retain their identities despite changes in their memberships (French Reference French1984, 29–30). Ireland is the same state as it was yesterday even though some of its members have died and others have been born, and Microsoft remains the same company over time even though its employees and shareholders change. Second, conglomerate collectivities have CID structures that allow them to deliberate and to combine the intentions of individuals into corporate intentions (French Reference French1984, 47–48). We might say that Ireland intends to raise taxes or that Microsoft intends to develop a new operating system. French concludes that conglomerate collectivities, such as states and corporations, are distinct agents that can be blamed, praised, punished, and obligated separately from their members. In an often-quoted phrase that sums up his argument, he declares that corporate agents are ‘full-fledged members of the moral community, of equal standing with the traditionally acknowledged residents: human beings’ (French Reference French1984, 32).

O’Neill (Reference Olleson1986) was the first to apply the agential theory to ethical issues in international affairs. She argues that many international ethical issues are intractable if individuals are assumed to be the only moral agents (O’Neill Reference Olleson1986, 51–53). ‘Individuals have remarkably few options to reduce nuclear dangers’ (O’Neill Reference Olleson1986, 55), so it is futile to say that they have duties to prevent nuclear war. Many responsibilities, such as duties to prevent war or climate change, must be assigned to states because states are the only agents with the capacities to act on them. The key premise of O’Neill’s argument is that ‘some institutions may be agents in the literal and unmetaphorical way in which individuals are agents’ (Reference Olleson1986, 58). She argues that ‘the two sorts of agents are similar enough to suggest that if ethical reasoning is accessible to individuals then it is not inaccessible to states’ (O’Neill Reference Olleson1986, 62, emphasis in original). The decision-making structures of states, like the minds of individuals, allow them to process ethical imperatives and to set goals (O’Neill Reference Olleson1986, 61–66). O’Neill thus uses the analogy between states and human beings to scale up Kantian moral agency.

Erskine (Reference Erskine2001) introduced the agential theory to the discipline of IR, where she remains its most prolific and influential proponent (Erskine Reference Erskine2003, Reference Erskine2004, Reference Erskine2008, Reference Erskine2010, Reference Erskine2014, Reference ErskineForthcoming). Drawing from both French and O’Neill, she argues that a group is a moral agent if and only if it has five features:

These five criteria determine which groups are sufficiently analogous to human beings to count as moral agents. For example, mobs and crowds are not moral agents because they do not have corporate identities or decision-making structures. Neither their personalities nor their intentions are distinct from the personalities and intentions of their members. Nor do puppet states or shell companies count as moral agents, since their identities are created and sustained by other agents. Examples of corporate moral agents include (most) states, business corporations, unions, intergovernmental organizations, rebel groups, and drug cartels.

For early proponents of the agential theory, corporate moral agency was a contentious proposition that had to be defended at every turn. Erskine once lamented that there is a ‘general reticence to accept that the class of moral agent might extend from the individual human being to encompass certain types of groups’ (Reference Erskine2003, 2). However, since the mid-2000s, the agential theory has ceased to require much justification, save for some obligatory citations of the works of these early proponents. The next generation of scholarship in the agential tradition focuses on applying the concept of corporate moral agency to particular issues, such as great-power responsibility (Brown Reference Brown2004), evil (Lu Reference Lu2004), Europe’s international citizenship (Dunne Reference Dunne2008), national defence (Eckert Reference Eckert2009), and state punishment (Lang Reference Lang2007, Reference Lang2008, Reference Lang2011; Erskine Reference Erskine2010). Many works of international political theory now take the idea of corporate moral agency as a basic premise (e.g. Pasternak Reference Pasternak2010, Reference Pasternak2013; Stilz Reference Stilz2011; Crawford Reference Crawford2013b; Collins Reference Collins2016). As Valentini declares, as if to state the obvious, ‘states, universities, churches, and hospitals clearly have a capacity for “collective will formation” sophisticated enough to warrant attribution of moral agency’ (Reference Valentini2011, 133, emphasis in original). The agential theory has a near-monopoly over the ethics of state responsibility. Its only significant rival is the lingering skepticism of corporate agency (e.g. Wight Reference Wight1999, Reference Wight2004, Reference Wight2006; Lomas Reference Lomas2005, Reference Lomas2014; Gould Reference Gould2009).

Just as the agential theory originated in philosophy, its dominance is largely due to developments in philosophy. The most important of these developments is List and Pettit’s (Reference List and Pettit2002, Reference List and Pettit2005, Reference List and Pettit2006, Reference List and Pettit2011) work on judgment aggregation.Footnote ⁶ Two of their insights have bolstered the argument for corporate agency. The first is that groups with certain kinds of decision-making procedures can have ‘intentions’ that none of their members have. For example, Tollefsen (Reference Tollefsen2015, 60–62) describes a scenario in which a three-member admissions committee evaluates PhD applicants using four criteria: test scores, grades, letters of recommendation, and writing samples. Only applicants who meet every criterion will be admitted, and whether a given applicant meets a given criterion will be determined by a majority vote. The committee votes as follows on Michael’s application (Table 1).

Table 1 Tollefsen’s (Reference Tollefsen2015, 61) admissions committee

Although none of the committee members believe that Michael meets all of the criteria for admission, the majority of them believe that he meets each criterion. As a result, when the votes are tallied, the committee decides to admit Michael to the PhD program even though none of its members think he is a suitable candidate. ‘We intend to admit Michael to the program’ is true even though ‘I intend to admit Michael’ is false for each committee member. The decision-making structure of the committee thus produces a corporate intention that cannot be ascribed to any individual.

The second important insight is that corporate intentions are ‘multiply realizable’, which means that the same corporate intention can result from different combinations of individual intentions (List and Pettit Reference List and Pettit2011, 65–66; Tollefsen Reference Tollefsen2015, 87–88). The committee might still have decided to admit Michael if each of its members had voted differently, and even if the committee had entirely different members. List and Pettit (Reference List and Pettit2011, 78) conclude that, since the intentions of corporate entities are irreducible to the intentions of individuals, ‘we must think of group agents as relatively autonomous entities—agents in their own right, as it is often said’. Wendt (Reference Wendt2004, 299–300, Reference Wendt2005), Stilz (Reference Stilz2011, 191), Erskine (Reference Erskine2014, 119), and Collins (Reference Collins2016, 344) have drawn on these arguments from judgment aggregation theory to strengthen the agential theory of state responsibility.

The broad picture that we get from the agential theory is of many nested and layered moral agents (List and Pettit Reference List and Pettit2011, 40). Human agents can be members of multiple corporate agents simultaneously, and corporate agents can be members of other corporate agents. I am a member of a state, a university, and a union; my state, in turn, is a member of many international organizations. Although the memberships of moral agents overlap, their responsibilities are mutually independent (Erskine Reference Erskine2003, 23; Lang Reference Lang2007, 245). A state can be responsible for breaching a contract even if none of its members – corporate or human – are responsible for the breach. Conversely, a state, a university, and several individuals can each be responsible for the same breach. As Erskine argues, ‘moral agency exists simultaneously at different levels, and moral agents at all levels can be responsible for concurrent, complementary, or even coordinated acts and omissions’ (Reference Erskine2010, 266). The responsibilities of agents at one level cannot be deduced from the responsibilities of agents at other levels. Each moral agent has its own personality, intentionality, and responsibilities.

States as legal persons: the functional theory

The international law of state responsibility has entirely different origins and influences. Whereas the agential theory grew out of Anglo-American philosophy, the law of state responsibility developed largely from post-World War One reparations law. As Crawford describes, the Treaty of Versailles ‘placed issues of responsibility for the major events of international war and peace irrevocably within the domain of the “legal”’ (Reference Crawford2013a, 27–28). Efforts to codify the law of state responsibility began in 1924, when the Assembly of the League of Nations convened a committee to identify and codify the most important areas of customary international law (Matsui Reference Matsui1993, 25). The United Nations (UN) ILC (2001) continued this work after the Second World War, and, after several decades, finished its Articles on Responsibility of States for Internationally Wrongful Acts.Footnote ⁷ Although the Articles have not yet been turned into a treaty or convention, they are widely considered to be an authoritative codification of the customary law of state responsibility (Crawford Reference Crawford2013a, 42–44, Olleson Reference O’NeillForthcoming). The Articles are of theoretical interest because they contain a functional theory of state responsibility that has little in common with the agential theory.

The idea of corporate agency is notably absent from the international law of state responsibility. States are held legally responsible not because they are agents, but because they act vicariously through human agents. The ILC (2001, 35) makes this point clearly in the Commentaries that accompany its Articles:

There are many similar remarks in the secondary legal literature and in the decisions of international courts. For instance, Cassese writes that ‘for a State to be responsible it is necessary first of all to establish whether the conduct of an individual may be imputed to it’ (Reference Cassese2001, 187), and Nollkaemper writes that ‘in factual terms states act through individuals’ (Reference Nollkaemper2003, 616). Crawford and Watkins emphasize that ‘states, lacking bodies of their own, can only act through the agency of others—in the end, of natural persons’ (Reference Crawford and Watkins2010, 287). In other words, states can only act vicariously. The relevant ‘agents’ for international lawyers are state officials.

The idea of corporate intentionality, which is foundational for the agential theory, is treated as both irrelevant and mysterious by international lawyers. The ILC’s Articles include ‘no requirement of mens rea on the part of a delinquent state: an act incurring state responsibility could occur even where a state did not undertake the act intentionally’ (Crawford Reference Crawford2013a, 37). Crawford adds that ‘the “intention” underlying state conduct is a notoriously difficult idea, quite apart from questions of proof’ (Reference Crawford2013a, 62). Although proof of intent is sometimes necessary for state responsibility, it is the intent of state officials that counts. Brownlie (Reference Brownlie1983, 38) similarly argues that ‘metaphors based on intention (dolus) or negligence (culpa) of natural persons tend to be unhelpful’ for understanding state responsibility, and he suggests that a more appropriate analogue is a principal–agent relation: ‘the issues in inter-state relations are often analogous to those arising from the activities of employees and enterprises in English law, where the legal person held liable is incapable of close control over its agents’ (Brownlie Reference Brownlie1983, 38, emphasis in original). In accordance with this analogy, international lawyers often describe state officials as ‘state agents’ (e.g. ILC 2001, 78–80; Cassese Reference Cassese2007, 656, 661; Momtaz Reference Momtaz2010, 237–38). States are like principals on behalf of which their officials act rather than like agents in their own right.Footnote ⁹

In place of the concepts of corporate agency and intentionality, international lawyers employ the concepts of attribution and function. ‘Attribution’ is the process of attaching or imputing an action of an individual to a state (Condorelli and Kress Reference Condorelli and Kress2010). For example, in United States v. Iran, the International Court of Justice (1980, 35) ruled that the actions of Iranian protestors who occupied the American embassy in 1979 were attributable to the state of Iran and that Iran was therefore responsible for them. Although all actions are performed and intended by individuals, some actions are attributed to states, meaning that they count as acts of state.Footnote ¹⁰

The most basic rule of attribution is that ‘[t]he conduct of any State organ shall be considered an act of that State under international law, whether the organ exercises legislative, executive, judicial or any other functions’ (ILC 2001, Art. 5). The use of ‘organ’ rather than ‘official’ in the Articles is indicative of the theory of corporate responsibility that underpins international law. Whereas ‘official’ implies an office, ‘organ’ implies a function (Kelsen Reference Kelsen1970, 150–58). For example, soldiers are state organs because they perform the function of defence. Other state organs include legislators, judges, diplomats, federal units, mayors, and police officers (ILC 2001, 41; Crawford Reference Crawford2013a, 118–24). Whether an entity is a state organ depends on its function: ‘The key element is the role of the entity or official as part of the administration of the state’ (Brownlie Reference Brownlie1983, 141).

The actions of private entities that ‘exercise elements of the governmental authority’ are also attributable to the state (ILC 2001, Art. 5). The idea of governmental authority, although not functional on the face of it, ultimately collapses into the idea of function. Brownlie uses ‘authority’ and ‘function’ interchangeably: ‘state authority has been delegated to local traditional and religious authorities … state functions have at other times been farmed out to private individuals’ (Reference Brownlie1983, 136). Crawford (Reference Crawford2013a, 129) notes that ‘[t]here is no consensus as to precisely what constitutes “governmental authority”’, and he defines it primarily in terms of three functions: detention and discipline, immigration control and quarantine, and seizure of property. For example, employees of private prisons exercise governmental authority because they perform the state function of detention and discipline. In other words, they are ‘de facto organs’ (Cassese Reference Cassese2007, 656; Momtaz Reference Momtaz2010, 243). The rule is that the actions of entities that ‘exercise functions of a public character’ are attributable to the state, even if these entities do not have formal state authority (Brownlie Reference Brownlie1983, 162; ILC 2001, 43).

The actions of a de jure or de facto state organ are attributable to the state ‘even if it exceeds its authority or contravenes instructions’ (ILC 2001, Art. 7). For example, in Caire v. Mexico, two Mexican military personnel murdered a French citizen after trying to extort money from him. Although the personnel acted outside of their authority, the French-Mexican Claims Commission ruled that their actions were nevertheless attributable to Mexico because they acted under the guise of their status as state organs (UN 2006, 517). What ultimately determines whether the actions of an individual are attributable to the state is not whether the individual acted under the authority of the state, but whether the individual ‘acted using the means and powers pertaining to his public function’ (Cassese Reference Cassese2001, 188). Function, not authority or intentionality, is thus the foundational concept in the international law of state responsibility.

An important implication of the functional theory is that there is a direct line of responsibility that runs to the state from each of its organs. Although the chain of command may matter for determining individual criminal responsibility, ‘the position of an official in the internal hierarchy has no relevance to the question of state responsibility’ (Brownlie Reference Brownlie1983, 134; ILC 2001, 40). The actions of a corporal are attributable to the state no less than the actions of a general, since both perform the function of defence. In the functional theory, unlike in the agential theory, there is a clear separation between agency and responsibility. Only individual organs can act, but the state is responsible for their actions.

Whereas the agential theory in almost uncontested in IR and political theory, the functional theory has a few rivals in international law. Some lawyers and legal theorists have adopted the agential theory (e.g. Tanguay-Renaud Reference Tanguay-Renaud2013), and others have proposed still other ways of conceptualizing state responsibility (e.g. Luban Reference Luban2011, 78–85). However, the functional theory is by far the dominant theory of state responsibility in international law and, crucially, the one that underpins the ILC’s Articles.

The gulf between the ethics and the law of state responsibility

To summarize, the agential and functional theories employ different concepts and rely on different models of collective action (Table 2).

Table 2 The agential and functional theories

The key concepts for the agential theory are agency and intentionality, and the model for state responsibility is a simple case of individual responsibility. States, like human agents, are responsible for actions that they perform intentionally or willfully. The key concepts for the functional theory are attribution and function, and the model for state responsibility is a principal–agent relation. States are responsible for actions that their organs perform on their behalf. Behind the superficial similarities, there are deep theoretical differences between the dominant ethical and legal approaches to state responsibility.

The agential and functional approaches might also be thought of as intellectual traditions or ‘subcultures’ (Bell Reference Bell2009, 19–22; Ashworth Reference Ashworth2012). In addition to using different models of collective action, the two approaches use different vocabularies, have different histories, and belong to different communities of scholars. This is not to say that the agential and functional theories are incompatible or incommensurable; they are not ‘paradigms’. They may even be complementary, although no one has yet tried to combine them.

Proponents of the agential theory do not recognize that international law contains a fundamentally different theory of state responsibility. They often appeal to international law for support, with the assumption that international law considers states to be corporate agents. Lang writes that, ‘[s]ince the heyday of positivism in the nineteenth century, states have been considered the primary agents of international law’ (Reference Lang2007, 244) and that ‘[t]he passage of the Articles on State Responsibility by the International Law Commission suggests that states can be considered responsible agents’ (Reference Lang2008, 23). Wendt declares that ‘[i]nternational politics as we know it today would be impossible without attributions of corporate agency, a fact recognized by international law’ (Reference Wendt1999, 10).Footnote ¹¹ Tollefsen argues, more generally, that ‘[o]ur practice of attributing responsibility to organizations … seems to presuppose that organizations literally have intentional states. For we could not hold them legally and morally responsible for an act unless they intended to commit the act’ (Reference Tollefsen2002, 396, emphasis in original). Many proponents of the agential theory have overlooked the functional theory because they have projected their own understanding of state responsibility onto international law.

The source of this misinterpretation is the idea that ‘only agents can be held responsible in the relevant way’ (Copp Reference Copp2006, 216). Although non-agents can be causally responsible, such as when a hailstorm is responsible for damage, they cannot be normatively responsible. Moral and legal responsibility are typically understood to require a will. As Goodin puts it, ‘[r]esponsibility implies agency, and agency implies some capacity for intentional action’ (Reference Goodin1995, 35). If only agents can be held responsible, then states cannot be held responsible unless they are agents: ‘the thesis that no collective is a moral agent … implies that no state has ever wrongly bombed a city and that no state has ever deserved blame for imprisoning an innocent person’ (Copp Reference Copp2007, 283).Footnote ¹² Proponents of the agential theory tend to reason as follows: international law holds states responsible; only agents can be held responsible; therefore, international law assumes that states are agents.

The flaw in this line of reasoning is far from obvious. The conclusion follows logically from the premises, so, since it is false that international law assumes that states are agents, one of the premises must be false. The first premise is clearly true: international law does hold states responsible. The fault lies with the second premise. Although it is true that an entity must, in some sense, be able to act in order to be held responsible, it is not true that only agents – construed as intentional actors – can be held responsible. It is possible for an entity to be vicariously responsible for actions that it neither performs nor intends. To use Brownlie’s (Reference Brownlie1983, 38) example, an employer is usually liable for damage that his employee causes, even if the employer did not specifically order the employee’s actions. Even non-agents can be vicariously responsible. Someone who is in a coma is not, at that time, an agent, but her guardian or representative might still sign contracts and incur debts in her name. The international law of state responsibility operates on the assumption that the capacity for vicarious action is sufficient for responsibility. States are not agents in their own right, but they can nevertheless be held responsible because the actions of their organs are attributed to them.

The idea of vicarious action can be used to explain how states can have moral responsibilities as well as legal responsibilities. The vicarious responsibilities that employers, officers, and ministers bear for the actions of their subordinates are both moral and legal. Just as employers are morally responsible for the actions of their employees – for instance, they have moral obligations to repair damage that their employees cause – states can be morally responsible for the actions of their organs. France has a moral obligation to compensate people who have been abused by French soldiers, but we do not need to assume that France is a corporate agent or to prove France’s intentions in order to make this judgment. States can be ‘moral principals’ instead of moral agents.

One advantage of conceptualizing state responsibility in terms of vicarious action is that it allows us to avoid the most common objection to holding states responsible – that collective entities cannot truly be agents, let alone moral agents (Ludwig Reference Ludwig2007, Reference Ludwig2014; Miller Reference Miller2007). Despite the dominance of the agential theory, skepticism of corporate agency remains common in IR. As Wight argues, the idea that the state is an agent ‘rests on the classic error of methodological structuralism: the attribution of the agential powers and attributes of human agents to a collective social form’ (Reference Wight2006, 188).Footnote ¹³ The functional theory provides a justification for holding states responsible that skeptics of corporate agency can accept.

Conceptualizing states as moral principals rather than moral agents also allows us to sidestep the thorny issue of whether states have intrinsic rights. Moral agency is often taken to imply rights as well as responsibilities (cf. Erskine Reference Erskine2014, 141). If ‘the state is a moral agent, in all the respects that morally matter’ (Goodin Reference Goodin1995, 35), then states seem to be entitled to rights that are analogous to human rights. There is a tension between corporate moral agency and ‘normative individualism’, or the principle that human beings are the ultimate units of concern (Hindriks Reference Hindriks2014). The idea that individuals are the primary bearers of rights, and that the rights of states are merely conditional or derivative, is difficult to sustain if we take states to be moral agents. Wendt is one of the few proponents of the agential theory who recognizes this: ‘if states really are people too, then we need some other, non-metaphysical way to justify liberalism, which may force us to confront possibly uncomfortable truths’ (Reference Wendt2004, 293).Footnote ¹⁴ But if states are moral principals rather than moral agents – that is, if they are only vicarious actors with no agency of their own – then there is no reason why they must have intrinsic rights. The functional theory allows us to take the asymmetrical position that states have non-derivative responsibilities but that their rights, such as sovereignty, are derived entirely from the rights of individuals.

Proponents of the agential theory may have shunned the ideas of vicarious action and vicarious responsibility in part because they have associated them with juridical fictions. But vicarious actions and responsibilities need not be fictional. For instance, if Margaret buys a house through an estate agent, we say ‘Margaret bought a house’. The action of buying a house really counts as Margaret’s, and she is really responsible for it, even though her estate agent physically performed it (see Fleming Reference FlemingForthcoming). Contemporary international lawyers treat the actions and responsibilities of states in the same way – as vicarious but real. The ILC (2001, 35–36) emphasizes that ‘[t]he State is a real organized entity’, not a juridical fiction, and it uses ‘attribution’ rather than ‘imputation’ in its Articles because the former term ‘avoids any suggestion that the legal process of connecting conduct to the State is a fiction, or that the conduct in question is “really” that of someone else’ (see also Crawford Reference Crawford2013a, 113n). The functional theory is not a fictional theory.

Conceptualizing states as moral principals instead of moral agents gives us a much thinner conception of state responsibility. If the responsibilities of states are purely vicarious, then they can be restitutive but not retributive. It does not seem possible to be vicariously guilty or culpable. Guilt, culpability, and punishment require intent, and intent is not something that can be transmitted from agents to principals. For example, suppose that an employee assaults a customer while on the job. We might, in some circumstances, hold the employer liable for compensating the victim, but we would not charge the employer with assault. Assault requires intent, and the intent of the employee cannot be attributed to his employer. Similarly, international law holds states liable for repairing harms that their organs cause, but it does not attempt to punish states for these harms. As I explain in the next section, the functional theory forecloses the very possibility of ‘state crimes’.

Can states commit crimes?

Even if we ultimately favor the agential theory, we ought to understand the functional theory, since we can neither understand nor effectively criticize the international law of state responsibility without it. The debate about state crimes illustrates this point.

Although states are held responsible under international law, they are not held criminally responsible. The Nuremberg International Military Tribunal’s (1947 [1946], 221) often-quoted declaration that ‘[c]rimes against international law are committed by men, not by abstract entities’ remains the rule. As Crawford and Watkins describe, ‘there has been no development of corporate criminal responsibility to parallel the introduction of individual criminal responsibility on the international plane, nor has there been any trend among arbitral tribunals to impose punitive damages on states’ (Reference Crawford and Watkins2010, 285). It is significant that the International Criminal Court has jurisdiction over natural persons only (UN 1998, Art. 25.1). However, earlier drafts of the ILC’s Articles did include a controversial article about international crimes of states.

Article 19

International Crimes and Delicts

…

The ILC deleted Article 19, along with all other references to international crimes, before the UN General Assembly adopted the final version of the Articles in 2001. In order to avoid the penal connotations of ‘crime’, the ILC (2001, Ch. III) chose to designate severe violations of international law as ‘serious breaches’ instead.

Some proponents of the agential theory have criticized international law for omitting the concept of state crime. As Lang argues, ‘[t]he passage of the Articles on State Responsibility by the International Law Commission suggests that states can be considered responsible agents, although their refusal to include punitive sanctions as part of those articles vitiates their impact’ (Reference Lang2008, 23). His argument has two prongs. The first is that ‘the failure of the ILC to include Article 19 in its final articles is not a conceptual one’ (Lang Reference Lang2011, 110), since states are literally intentional agents. The elements of a crime – actus reus and mens rea – apply to states as well as to individuals (Lang Reference Lang2007, 244–45). Tanguay-Renaud similarly argues that a ‘corporate agent, which in an important sense has a mind of its own, may … exhibit the types of mens rea attitudes that are deemed so central to modern criminal culpability—namely, intention, recklessness, negligence, and the like’ (Reference Tanguay-Renaud2013, 262). The second prong of Lang’s argument is that holding states responsible but denying that they can be criminally responsible is pedantic if not incoherent. He is particularly critical of the law of genocide, which holds that states can be responsible for genocide but that only individuals can be criminally responsible for genocide: ‘To distinguish between committing a crime and being responsible for committing something that is called a crime seems to be stretching the meaning of language beyond comprehension’ (Lang Reference Lang2011, 104). If genocide is a crime, and if states can be responsible for genocide, then how could genocidal states not be criminals? Lang concludes that the law of state responsibility is impoverished without the concept of state crime.

The ILC had two reasons for omitting Article 19. One reason is that the concept of state crime was ‘[t]he single most controversial element in the draft articles on State responsibility’ (Crawford Reference Crawford1998, 9). Replacing ‘international crime’ with ‘serious breach’ was, in part, a pragmatic compromise that allowed the ILC to move forward. The deeper reason for omitting Article 19 is that the concept of state crime is at odds with the functional theory. International lawyers do not employ the concepts of corporate agency and intentionality, so, from their perspective, it is not clear how states could form the requisite mens rea to commit crimes.Footnote ¹⁶ As Marek (Reference Marek1979–80, 464) argues, the application of ‘crime’ to the state ‘founders on the truism that penal law deals necessarily with individuals … No amount of anthropomorphism can make the State an object of psychological analysis’. Caron makes the same point: ‘Criminal jurisprudence ties moral culpability to intent; crimes require both an act and a mens rea. But what then does it mean to speak of an organization being criminally responsible? Corporations and states do not have intent’ (Reference Caron1998, 311).Footnote ¹⁷ Part of the reason that there was so much opposition to the concept of state crime is that ascriptions of mens rea to states are difficult to square with the functional theory. States cannot have ‘guilty minds’ because they do not have minds in the first place.

Even defenders of Article 19 in IL reject analogies between state crime and individual crime. Their justification for the concept of state crime was never that states have analogous capacities for actus reus and mens rea, as Lang (Reference Lang2007, Reference Lang2011) and Tanguay-Renaud (Reference Tanguay-Renaud2013) argue, but that the most serious breaches of international law (such as aggression) must be distinguished from ordinary breaches (such as trade violations). State crimes were to be distinguished according to their seriousness, not by analogy with individual crimes (Spinedi Reference Spinedi1989, 50–57; Abi-Saab Reference Abi-Saab1999, 345–46; Pellet Reference Pellet1999, 426–27). As Crawford (Reference Crawford1998, 13) noted in his ‘First Report on State Responsibility’, ‘even those Governments which support the retention of Article 19 in some form do not support a developed regime of criminal responsibility of States, that is to say, a genuine “penalizing” of the most serious wrongful acts’. Although most governments agreed that a special category was necessary for the most serious breaches of international law, it was ‘quite widely felt that the terminology of “crimes” of State is potentially misleading’ (Reference Crawford1998, 13). It is notable that Jørgensen (Reference Jørgensen2000), who provides the fullest defence of the concept of state crime, draws an analogy with corporate crime in domestic law rather than with individual crime in international law (see also Caron Reference Caron1998). International lawyers on both sides of the state crime debate reject the state–human analogy that underpins the agential theory. We cannot understand the arguments of either side until we recognize that the international law of state responsibility employs the idea of vicarious action rather than corporate agency.

As for the second prong of Lang’s argument, there is nothing incoherent about holding individuals criminally responsible and the state non-criminally responsible for the very same action. Once again, the idea of vicarious action is the key to the puzzle. One person can be criminally responsible for an action while another is vicariously responsible it. For example, an employee could be guilty of criminal negligence for causing an industrial accident while her employer, who was not personally negligent, has a duty to compensate the victims. The industrial accident was indeed a crime, but this does not mean that everyone who bears some responsibility for it is a criminal. Questions about who is guilty are distinct from questions about who has an obligation to repair the harm (Crawford Reference Crawford1998, 23). Just as we assign criminal responsibility to the employee but only vicarious responsibility to her employer, we can assign criminal responsibility to individuals but only vicarious responsibility to their state. A state can have obligations to compensate the survivors of genocide, to promote reconciliation, and to reform the institutions that made the genocide possible even though only individuals are criminally responsible for genocide. Understanding the functional theory dissolves the apparent contradiction.

What initially appears to be a first-order normative debate about whether states ought to be held criminally responsible is a symptom of a deeper disagreement about how to conceptualize state responsibility. IR scholars and international lawyers often disagree about how states can be held responsible because they disagree about why states can be held responsible in the first place. This is not to say that the debate about state crimes cleaves perfectly along theoretical lines. For instance, a proponent of the agential theory could argue against criminalizing states on the grounds that, in the absence of a world state, it creates a pretext for vengeance. However, the deeper, theoretical dimension of the state crime debate has escaped many commentators (though see Gould Reference Gould2009), in large part because they do not recognize that IR and IL rely on different theories of state responsibility.

Which theory we accept affects how we respond to the international wrongs of states, such as genocides and cyberattacks. If we subscribe to the agential theory, then we are likely to see the offending states as criminals and to demand that they be punished. What we need is an International Criminal Court that is capable of trying states in addition to individuals. If we subscribe to the functional theory, then we are likely to see the offending states as pools of resources that can be used to compensate victims. Although the state is liable for repairing the harm, it is particular individuals, if anyone, who should be put on trial. What we need is an International Court of Justice with compulsory jurisdiction and an international sheriff’s office that is capable of enforcing judgments against states. The agential theory and the functional theory suggest different responses to the international wrongs of states, as well as different proposals for international legal reform.