What's the problem with responsibility?

World political events are increasingly framed in moral terms, and perhaps no moral concept is as readily deployed as that of responsibility. From the vague appeal to the obligation of citizens of wealthy countries to ‘make poverty history’, to the International Criminal Court's (ICC) ongoing efforts to bring figures like Joseph Kony and Saif Gaddafi to trial, and the resurgence of calls for humanitarian intervention from the international community, responsibility seems to be the order of the day. The shift to hold individuals and regimes accountable to universal obligations is a remarkable change in world politics, but for all the progress – epitomized by institutions like the ICC and doctrines such as the Responsibility to Protect (R2P) – there are serious problems with our contemporary practices of responsibility and dominant understandings of moral agency that inform them. In particular, emerging understandings of responsibility in world politics have been too focused on the actions of individuals, leading to the neglect of structural causes of mass violence and more indirect lines of responsibility revealed by attending to the wider social context. Related to this limited focus on individual actors, academic accounts of moral agency have tended to downplay the political aspects of this focus on the individual's failure to uphold their obligations, which obscures the power inequalities and particular interests that are served by focusing on individual actors over enabling conditions and social structures.

The problematic consequence of this can be seen in the difficulties faced by the ICC in trying to serve the ends of both individual accountability and the peaceful resolution of conflicts, as well as in the troubling violence that results from using ‘humanitarian’ violence for moral ends. Given the increasingly central place of responsibility within world politics, attending to the limits and blind spots of our contemporary practices is a pressing intellectual and political project. The United Nations (UN) has begun to orient itself around the idea that sovereignty entails a ‘responsibility to protect’ (UN General Assembly 2005a, 2005b; UN Secretary General 2008; UN General Assembly 2009), leading to new obligations that hold both states and the international community accountable to ensure human security, while also making clear that negligent states, and individuals, should be held responsible for the harms they inflict. These claims are rooted in a reinterpretation of traditional state sovereignty and in the moral discourse and legal architecture of human rights that has developed over the past 50 years, implying that moral concern does not stop at state borders and that sovereignty does not provide immunity from judgement or, increasingly, punishment (International Commission on Intervention and State Sovereignty (ICISS 2001)). Further, the ICC is at the forefront of a movement to implement international criminal law, building on the tribunals set up throughout the 1990s and 2000s, which gives legal reality to the notion of individual criminal responsibility in world politics. Yet, at the heart of these developments there is a difficulty – How is the act of holding a person or a specific group responsible for mass violence and suffering justified? While there may be a practical simplicity to this question, it is philosophically and ethically fraught – which is highlighted in philosophical debates over whether moral responsibility is even possible (Clarke Reference Clarke2005; Lenman Reference Lenman2006) – and the ambiguities that make holding actors responsible a questionable one on the personal level are increased when we consider events and actors in world politics (Lang Reference Lang2007, 242–243).

As the idea of responsibility has become a central part of world politics, investigating the meaning of responsibility has emerged as a topic in international political theory (Erskine Reference Erskine2003; May Reference May2005; Ainley Reference Ainley2006; Jeffery Reference Jeffery2008; Lang Reference Lang2008). Research has focused on two aspects of moral responsibility; first, the development of a system of international criminal law that is able to hold individuals legally accountable for a limited set of serious international crimes (May Reference May2005, 117–178; Robertson Reference Robertson2005), and second, the attempt to hold states and other collective actors politically accountable for their harmful actions (Erskine Reference Erskine2001; Lang Reference Lang2007). The exploration of what forms of responsibility are possible in world politics has led to an analysis of responsibility across its different aspects: causal, moral, political, and legal. Determining how these aspects are, and should be, related is at the centre of current intellectual controversies, such as whether collective actors can have moral responsibilities that attach to their causal responsibility for events; how far international legal responsibilities are a reflection of moral responsibilities; and whether our primary political responsibilities are to humanity as such or to our exclusive political communities.Footnote 1 Further, establishing international/global practices of responsibility requires that we define the relevant actors whose behaviour is to be shaped by these practices. On one hand, they can refer to universal obligations that fall to all natural individuals to protect human rights, and to attempts to hold particular individuals accountable for egregious harms. On the other, it can refer to responsibilities assigned to collective actors, to states primarily, but also possibly to institutions, such as international organizations, NGOs or multi-national corporations, who can be bearers of obligations and subjects of accountability. The desire to assign responsibility internationally/globally generates controversy not only because international politics has historically been defined by the absence of authoritative institutions capable of holding states accountable (Onuf Reference Onuf1979, 247–252), but also because assigning causal responsibility, to an individual or a group, for the complex social activities that define world politics is difficult and counter to much established practice (Erskine Reference Erskine2001, 69–74).

As successful and important as this work has been, there is a central question that has not been sufficiently considered: What is the value of seeking to assign responsibility in world politics? The full importance of this question emerges when we consider the difficulties and limitations of assigning responsibility for mass violence through an individualist and rationalist understanding of moral agency. Practically speaking, this is an easy question to overlook because the answer is intuitive. Enforcing moral responsibilities through political and legal institutions is necessary to any just society, it would seem. Much of the work in international political theory has been devoted to locating moral agents that bear moral obligations (soldiers, presidents, democratic publics, states, the international community) and considering how and when they can be held accountable (though trials, truth commissions, amnesties, public inquiries, apologies, lustrations, boycotts). My central claim is that this question is neither simple nor unproblematic and that when we consider practices of responsibility at the international/global level we must return to the question of what purpose the act of holding individuals and collective actors responsible serves.

The objection I raise to contemporary efforts to hold actors in world politics responsible is that they are based on a problematic understanding of moral agency. In short, the rational and autonomous agent presumed in practices of responsibility in world politics is inadequate in its inattention to the socially embedded nature of responsibility and for that reason obscures that these practices affirm particular political ends and limit the critical reach of our attempts to pursue responsibility. I find support among scholars working in international political theory, but argue that a still more critical reading of responsibility is needed (Frost Reference Frost2003; Crawford Reference Crawford2007; Ainley Reference Ainley2008; Dauphinee Reference Dauphinee2008; Lang Reference Lang2008, 45–57). The responsible agent is a socially constructed agent and the act of holding responsible is a coercive and creative political act. Holding agents in world politics responsible for their actions does more than make clear their moral obligations or hold them accountable. Holding an agent responsible moulds her agency in a particular way, and when this process is not consciously acknowledged the ideals, institutions and social structures that form our understanding of moral agency are withheld from scrutiny and tacitly reinforced. Criticism of contemporary practices of responsibility has looked at the inadequacy of focusing on the autonomous individuals as the primary responsible agent and the possibility of assigning responsibility to collective actors; what I hope to add, drawing on the work of John Dewey, is an analysis of responsibility as a practice aimed towards achieving particular ends and a defence of more democratic practices of responsibility that attend to social conditions as much as individual actions, while seeking to enable forms of moral agency that focus on reconstructing our social context as much as controlling the behaviour of individuals.

I begin by tracing out a series of critiques of a broadly accepted account of moral responsibility that has informed scholarship in international political theory and continues to influence world politics. This account of responsibility is based on an individualist and rationalist understanding of moral agency, which understands the moral agent as a self-determined chooser able to follow objective moral principles. Yet, if one had good reason to revise this account of moral agency then the practice of holding individuals responsible becomes problematic (Tallgren Reference Tallgren2002; Crawford Reference Crawford2007; Ainley Reference Ainley2008). By analogy, a similar account of moral agency is applied to collective actors as well (Brown Reference Brown2001; Erskine Reference Erskine2001). After tracing out contemporary critiques of the conventional account of responsibility, I go on to suggest that insufficient attention has been paid to the social construction of responsible agents and turn to an alternative understanding of moral agency, which I describe as practice-based. Even where contemporary work in international political theory begins from the social construction of responsible agents, I argue it pays insufficient attention to the disciplinary function of contemporary practices of responsibility that reinforce political and legal inequalities in the contemporary international order without taking adequate measure of the consequences of these developments. I then suggest a more critical practice-based account of responsibility can be developed from a situated and democratic understanding of moral agency. The need for such an account is illustrated by analysing recent attempts to assign responsibility to political leaders in recent conflicts in Sierra Leone and Uganda, where the limitations of conventional practices are highlighted and the possibilities opened up by a practice-based account are explored.Footnote 2

The good, the bad, and the incapable: moral agency in world politics

In contemporary world politics there are two distinct developments of responsibility: first, the pursuit of individual responsibility, and second, a push to hold collective international actors responsible. In both cases, responsibility is assigned to autonomous agents who have an obligation to respect universal human rights, which prohibit grievous international crimes. As a consequence, academic work on the place of responsibility in world politics has required a search for capable moral agents who can be held accountable. I argue that contemporary international/global practices of responsibility depend upon an individualist and rationalist understanding of moral agency, while also suggesting that this practical understanding of moral agency is in part drawn from influential philosophical accounts. This understanding of responsibility is premised on a series of related claims about moral agency: there are objective moral principles, which the moral agent is able to recognize independently of convention, and the moral agent is capable of meaningful choice, which makes punishment effective and just, so long as her free will in such matters is sufficient to make her the cause of her actions. As a general account of moral agency it is broadly Kantian and distinctly modern (Smiley Reference Smiley1992, 72–101). To the extent that actors can determine their obligations and act freely we hold them accountable. Failures to act morally are either wilful failings or the result of incompetence. Holding actors responsible, then, allows us to blame harms on the actions of immoral individuals and collective actors, and to punish them accordingly.Footnote 3 The purpose of what follows is to substantiate my claim that this account of agency both underlies international/global practices of responsibility and is influential in understandings of responsibility in international political theory.

While there are international practices of responsibility that have longer histories, the major developments in world politics have come since the end of WWII and represent a profound shift in the modern international order. The traditional notion that sovereignty entailed non-interference in the domestic affairs of states, and that it was granted as a matter of mutual recognition between states, has been replaced by an understanding of sovereignty as conditional, such that a state's failure to protect its citizens justifies intervention. Further, the development of the contemporary idea of human rights, as both a legal regime and an ethico-political ideal, has altered the standards of legitimate statehood and furthered the development of transnational and global forms of political order. Our contemporary return to an ethical vision of world politics is seen in a variety of practices, all of which entail forms of moral responsibility. For the purpose of illustration I look at the R2P doctrine and the Rome Statute of the ICC.

R2P was originally developed in response to the controversies around the use of military force for humanitarian purposes in the 1990s and 2000s (ICISS 2001, 1–10), which was itself a reflection of the new importance given to protecting human rights. The responsibilities of sovereignty articulated in R2P are not simply drawn from the UN Charter or existing international law, as the strong emphasis on the existence of a clear responsibility to ensure human security and human rights required an appeal beyond established custom and written law: ‘Whether universally popular or not, there is growing recognition worldwide that the protection of human security, including human rights and human dignity, must be one of the fundamental objectives of modern international institutions’ (ICISS 2001, 6). This kind of moral appeal is characteristic of practices of responsibility in world politics. From Justice Robert Jackson's invocation at Nuremberg of ‘abnormal and inhuman conduct’ that went beyond ‘mere technical or incidental transgression of international conventions’ (International Military Tribunal (IMT) 1947, 98–102) to Tony Blair's justification of the NATO intervention in Kosovo in terms of a ‘moral purpose’ to defend the ‘values of liberty, the rule of law, human rights and an open society’ (Reference Blair1999), the development of international/global practices of responsibility depends upon an appeal to universal principles; this appeal requires that obligations are known by the actors in questions – individuals, states, and the international community (ICISS 2001, 13) – so far as they are appropriately reasonable. Further, when considering the actors responsible for protecting human security, R2P invokes a concept of the state as an autonomous actor able to choose whether or not to uphold its obligations, hence making it possible for the state to be accountable for its actions.Footnote 4 A similar agency is ascribed to the UN Security Council, General Assembly, and other international actors – though it is less clear how such actors are to be held accountable (Erskine Reference Erskine2008b). ‘The “responsibility to protect” implies above all else a responsibility to react to situations of compelling need for human protection. When preventive measures fail to resolve or contain the situation and when a state is unable or unwilling to redress the situation, then interventionary measures by other members of the broader community of states may be required’ (ICISS 2001, 29). It is presumed that actors freely make choices for which they are held accountable – and where they prove to be incapable, their responsibilities are passed on to more capable actors.

This understanding of moral agency is also seen in the Rome Statute of the ICC. In establishing a court to try international criminals the statute reinforces the legal status of these crimes, however, the violations in question are generally considered jus cogens norms and their prohibition is not simply based in the written law but more deeply in universal morality (Fisher Reference Fisher2009; Macleod Reference Macleod2010). According to Article 32.2 of the statute, individuals under the jurisdiction of the ICC cannot escape responsibility by pleading ignorance of the law (2011, 21), as it is assumed that the crimes under consideration are not only clearly articulated but also reflect basic moral norms – the only exemptions, in Articles 30 and 31, are for individuals lacking the mental capacity to know the content of the law or the consequences of their actions (Rome Statute 2011, 20–21). Even within the Rome Statute, which is the clearest legal expression of the requirements of individual responsibility, moral claims external to treaty law remain important. Regarding the issue of whether superior orders can exculpate an individual from responsibility (Article 33), the statute makes an exemption for individuals legally required to follow orders and who did not know their orders were unlawful, but only if the committed acts that were not manifestly unlawful (2011, 21–22). The issue of which crimes are manifestly unlawful is dealt with in part; crimes against humanity and genocide are specified, but the basis for that distinction and the potentially manifestly unlawful nature of other crimes is not clarified. The force of these claims rests on the notion that certain acts are clearly and universally wrong, which should be apparent to any reasonable individual (May Reference May2005, 196–198). In addition to knowledge of the impermissibility of international crimes, individuals subject to the ICC are also assumed to be capable of making free and rational choices, which ensures that they can be held accountable. This requirement can be seen in the exemptions in Article 31 based on mental incapacity or extreme necessity (Rome Statute 2011, 21), which establishes that ‘The “ideal type” perpetrator of international crime is reasonable, rational, intentional and knowledgeable, and his actions are entirely under his volitional control’ (Ainley Reference Ainley2006, 7). Looking to both R2P and the Rome Statute as exemplars of international/global practices of responsibility reveals a nascent understanding of moral agency based on strong claims regarding individual autonomy and the rationality of moral principles; understanding why this is so requires a turn towards the philosophical accounts of moral responsibility.

The practical understanding of responsibility seen in R2P and the Rome Statute reflects a distinctly modern conception of moral agency. Assigning moral responsibility is treated as the discovery of a fact about the world, of both the cause of an action and its moral worth.Footnote 5 Responsibility obtains where the actor is the cause of events and her intention is expressed through her actions. This understanding of responsibility echoes back to Aristotle's discussion of voluntary action in book three of the Nicomachean Ethics (Reference Aristotle1985, 53–59), though our contemporary understanding is rationalist and individualist in a way that Aristotle's was not (Smiley Reference Smiley1992, 33–37). As Marion Smiley makes clear, Aristotle's discussion of voluntary actions is intended as a guide to assigning blame as part of a social practice of ascribing responsibility to individuals, such that exemptions from responsibility based on the involuntary nature of an act or ignorance of its consequences did not depend on the notions of an absolutely free will or universal principle that guide modern accounts responsibility, which are based on an account of responsibility as a determination of truth rather than the social ascription of blame (1992, 39).Footnote 6 Modern conceptions of responsibility remove the act of assigning responsibility from its context, from its function within social and political life, and render it as an independent judgment of the moral worth of the agent – thus leading to understandings of moral responsibility as individual accountability for upholding one's rational obligations (Williams Reference Williams1985, 176–179).

This understanding of responsibility has led to a focus within the philosophical literature on the issue of free will, as the possibility that forces outside the agent's control determine human action undermines the possibility of holding individuals responsible (Strawson Reference Strawson2010). What is most important about these debates is that the primary concern is to preserve the modern account of responsibility, rather than to question the premises on which it is based (Smiley Reference Smiley1992, 95; Clarke Reference Clarke2005). Affirming some notion of compatibalism is the most common way to understand moral responsibility in the contemporary philosophical literature, which maintains that human beings remain free enough for their actions to be considered rational and voluntary.Footnote 7 What is striking, however, is that while this cedes some ground to the social construction of responsibility, dominant understandings of responsibility remain committed to the view that holding individuals responsible is an act of judgment that serves to determine the moral worth of the individual based on her choice to meet her obligations, and to hold her accountable when she does not. However, without a strong defence of the fundamental reality of free will and objective moral principle, compatibalism fails to fully acknowledge the social, and political, construction of the free and rational moral agent – the consequences of which, I will consider in the next section.

While I have only offered a brief account of the complex debates on moral responsibility, this should be adequate to make clear how the conventional understanding of responsibility, and the account of moral agency it draws on, is deployed in international political theory. Drawing on a distinction between different liberal traditions of international thought made by Kirsten Reference AinleyAinley, I look at one distinctly cosmopolitan line focused on individual responsibility and another internationalist one focused on states, in order to show how the conventional account of moral agency is fundamental to both (2008, 6). I focus on these particular liberal traditions because they not only draw heavily on Kantian notions of moral agency, but also because they have been extremely influential within international political theory.Footnote 8

Cosmopolitan liberalism is defined by its concern for the individual, her rights and responsibilities, not as a member of some particular polity but as an individual. First, the moral rights of the individual provide an authoritative basis for political legitimacy beyond de facto state sovereignty. Second, violating the moral rights of individuals is a serious wrong regardless of the victim's physical or social location, which undermines the traditional notion that moral relationships do not cross borders (Beitz Reference Beitz1999). This justifies the prosecution of individuals responsible for rights violations by international or global institutions, as well as various forms of intervention, military and otherwise, to protect individual rights (Luban Reference Luban1980, Reference Luban1987). The rights bearing individual replaces the state as the foundation of legitimate authority in world politics, as the moral rights of the individual provide the foundation for legitimate legal and political orders to be built (Held Reference Held2009, 540).

In order to hold these rights, the individual at the centre of cosmopolitan liberalism must be a particular type of moral agent. The rights and responsibilities of the individual are dependent upon his ability to recognize authoritative moral principles and freely direct his own action.Footnote 9 If the individual cannot know moral principles through the exercise of reason, or does not recognize the same moral principles as everyone else, the authority of universal rights is undermined. Further, if individuals cannot freely chose to follow the moral law then the assignation of responsibility is thrown into doubt. The wilful failure to uphold universal principles then becomes an expression of moral evil (Connolly Reference Connolly1991, 1). We can see this in the identification of criminal individuals in world politics who must be held responsible for their actions: dictators, warlords, and terrorists. These threatening figures are different from agents that fail to uphold moral principle because of a lack of capacity. This is seen in the contrasting identification of irrational genocidaires, child soldiers, and exploited victims of poverty, who have a role in violence but are not fully culpable. These designations, of evil and incompetence, are necessary to the logic of conventional moral agency as those denied responsible agency reveal the essential traits of moral agency through contrast.

Like liberal cosmopolitanism, liberal internationalism provides a vision of a moral international order in which responsibility is a key component. The difference is the value given to the state. The moral value of the state arises because the political community enables the fulfilment of individual rights, particularly rights to self-determination expressed through democratic representation.Footnote 10 The privilege accorded to the state is not given to the state as such, rather the quality of domestic arrangement becomes a matter of universal concern as political community has value so far as it protects and enables individual rights. Legitimacy, therefore, is based on a particular arrangement of the state, namely a liberal-democratic one. This privilege then justifies coercive practices and supra-state institutions to ensure a more liberal international order populated by responsible states.

The conventional account of moral agency remains at the core of liberal internationalism, even as the focus changes from individual to collective actors. This undermines the principle of sovereign equality and leads to a distinction between ‘good’ and ‘bad’ states, making hierarchy a key feature of a liberal internationalist order.Footnote 11 Liberal internationalism, and the discourse of good and bad states that it generates, depends upon assigning responsibility to the state as a collective actor that has similar powers of agency to the responsible individual. For the conventional account of moral responsibility to be traced onto the state, it must be treated as an agent who freely acts with reason and volition. If a state cannot meet the requirements of the conventional account of moral agency, including both having the right domestic arrangement and the required state capacity, then it is not fully legitimate.Footnote 12 Full sovereignty is only granted to the good liberal state, which is able to control its borders, prevent and punish crime, provide social services, maintain economic growth, and respect international law. States that fail to meet this standard are exposed to the prosecutions of their leaders and agents, humanitarian military interventions, and outside interference such as development and structural adjustment programs. As liberal internationalism depends upon state practice rather than cosmopolitan institutions, good states gain the right to violate the sovereignty of states acting wrongly or those that are incapable of acting morally.Footnote 13

The traditional presumption of state legitimacy is lost when a state fails to abide by universal standards, but the way in which states fail to abide by these standards matters. As with individuals, tracing the identities that are denied status as responsible agents reveals the essential traits of privileged agents. The prosecution of rogue states that intentionally violate the norms of international politics is relatively rare, and in those instances innocent civilians are quickly acknowledged as victims of an evil state.Footnote 14 If good states were granted privilege as collective actors, then it would seem that bad states should face punishment as collective actors. This happens to a limited degree with the use of sanctions and embargoes, but the suffering caused to civilians undermines the morality of such actions (Lang Reference Lang1999). The more common response is to seek out responsible individuals that can be held accountable. The concern for individuals’ rights that distinguishes liberal internationalism from nationalism also constrains discourses of responsibility to specific actors, whether individual or collective, rather than a wider community or society as such. This is also seen in the discourse on failed states, in which the state's failure to abide by universal standards is due to a lack of state capacity. In this case the citizens are still victims to be protected, but neither the state nor its authorized actors are necessarily seen as criminals, though there may be particular individuals that can be held responsible for their role in undermining the state. Within liberal internationalism, the assignment of responsibility is focused on criminal individuals or groups that are seen to be the cause of the negative consequences of bad states. This means that a strong individualist orientation is maintained, which gives emerging practices of international responsibility a degree of continuity despite the differences that exists between these perspectives in institutional terms. International criminal law, for example, is not only supported by a cosmopolitan orientation. Despite the cosmopolitan celebrations of the ICC (Mégret Reference Mégret2001; Roach Reference Roach2005), the institution remains embedded in a statist international system and consideration of the liberal internationalist line of thinking is necessary to fully understand the emerging practices of responsibility in international criminal law (Peskin Reference Peskin2009).

It is against this backdrop of emerging international/global practice and conventional theoretical understandings that recent inquiries into the place of responsibility in world politics have brought important insights to the fore. The key objections I want to develop are that the influential perspectives in international political theory are too focused on finding autonomous individuals to be held accountable, such that they are inattentive to the social construction of our obligations. These are important omissions because they obscure the power inequalities that make our contemporary practices of responsibility possible – for example, the social privilege and political power necessary for individuals and collective actors to exercise the autonomous agency assumed in dominant understands of responsibility cannot be presumed to exist naturally. Further, as we acknowledge that our understandings of moral agency and our practices of responsibility are social constructions, we are confronted with the question of whether they are the best constructions and on what basis such determinations are made.

We're talking about practice: critical account of responsibility in international political theory

In the previous section I examined how the rationalist and individualist account of moral agency informs the ICC and R2P, as well as how it functions in liberal visions of moral order in world politics.Footnote 15 There are, however, reasons to think that this account of moral agency is implausible and leads to problematic consequences, suggesting that another way of thinking about responsibility would be desirable and alter our evaluations of contemporary practices. The first line of development addresses collective forms of agency. In particular, two questions are raised: first, can institutions be moral agents, and if so, how do we hold them responsible; and second, what forms of collective responsibility can be assigned both to formal institutions and more informal groups. Liberal practices of responsibility have tended to address these issues obliquely. For example in international criminal law, the individual is the key moral agent, such that charges of conspiracy and joint criminal enterprise, while used to try individuals thought to be morally responsible, are controversial, as they potentially undermine the understanding of moral agency that underlies the practice itself (Danner and Martinez Reference Danner2005). While in liberal internationalist accounts the state is the primary collective agent, this depends upon the presumed fundamental nature of the social bonds of the state, and in particular on the forms of accountability established by representative government. For example, Michael Walzer's defence of the principle of non-intervention is based on the moral and basic quality of the state as a social institution, such that it must be treated as a collective actor by those external to it (Reference Walzer1980), rather than on an examination of its institutional capacity as a responsible agent.Footnote 16 These are important limitations, as the lack of a clear account of institutional moral agency and a fuller consideration of collective accountability impairs emerging practices of responsibility by failing to identify collective actors that may be most responsible for harms, and, by misidentifying individuals subject to the force of institutions as autonomous agents.

Toni Erskine addresses the first question, suggesting that institutions can be moral agents (Reference Erskine2001, Reference Erskine2004). Erskine argues that moral agency extends to institutional actors, stating that ‘one must first be able to understand and reflect upon moral requirements. One must also have the capacity to then act in such a way as to conform to these requirements. Furthermore, to exercise moral agency, one must not only have the capacity to act in response to moral requirements, but also the freedom to do so’ (2001, 69). The capacity of institutions to act as moral agents in this way is established if the institution has ‘an identity that is more than the sum of the identities of its constitutive parts and, therefore, does not rely on a determinate membership; a decision-making structure; an identity over time; and a conception of itself as a unit’ (Erskine Reference Erskine2001, 72). This extension of moral agency helps to specify which states can be held responsible, as Erskine distinguishes between morally competent states and quasi-states not simply on the basis of their form of government but rather on their capacity as institutional agents. This is an important critical insight but there are limitations to Erskine account of institutional agency, as it maintains some aspects of the conventional account of moral agency,Footnote 17 and in doing so Erskine continues to treat responsibility as a feature of moral agents, such that we are interested in the truth or falsity of the claim that some agent is or is not worthy of blame. Once the moral agency of institutions is accepted, identifying non-natural moral agents that are a product of the social practice of assigning responsibility, the plausibility of the conventional account of moral agency is undermined. So, while Erskine exposes the way moral agents are constructed, such that individuals and corporate agents can be said to have moral responsibilities, and accepts that social structures can enable and disable, moral agency, she pays insufficient attention to the social construction of responsible agency itself. This limitation can be seen in her attempt to bracket off the questions of where the obligations we hold moral agents accountable to come from, which means that the emphasis on autonomous agency in her account is not fully justified.Footnote 18 Erskine also fails to attend to the question of who sets the standard for responsible agency and how that standard is used to discipline and privilege some actors over others. As its constructed nature is more fully recognized, the modern conception of agency requires a different justification, namely one based in its value as a social practice (Frost Reference Frost2003, 85–86).

Neta Reference CrawfordCrawfordCrawford addresses the second question of collective responsibility in her article looking at how responsibility is attributed in cases of military atrocity (2007). The insight that she pursues is that the attribution of responsibility to individuals who commit military atrocities may be sufficient in isolated cases, but where the commission of atrocities is systematic further attributions are needed. Crawford argues responsibility for systemic atrocities should be attributed, in part, to those who create and maintain atrocity-enabling conditions (2007, 190–191). Individual soldiers, whether acting alone or in groups, are constrained by the social structures that define modern warfare. Further, the structures held in place by the operating guidelines of the military, the policy of the government and the actions of the public could actually be the cause of systematic atrocities. Crawford argues that responsibility should be attributed to these collective actors because they constrain the actions available to soldiers and define the ends they can pursue. This is an important development as she acknowledges that the actors we seek to hold responsible – soldiers, commanders, government officials – are not fully free or rational, but social subjects following orders and fulfilling roles. Crawford's argument, however, does not take the full measure of its insight – it is not simply something that is true in the case of military atrocity but social life in general.Footnote 19 Thus, the limitation in her account is that she continues to search for responsible agents first, namely those that determine the conditions under which wars are waged, rather than focusing on practices of responsibility that shape both individual and collective actors as subjects of responsibility.

Erskine and Crawford open up our understanding of moral agency, locating collective actors that can fail in their obligations and highlighting the enabling conditions that individuals should be held accountable for creating. Yet, as long as we think of moral agency as a feature possessed by actors, such that the assignment of responsibility is a judgment that an actor intentionally caused some harm that he knew was wrong and need not have done, the practice through which we hold actors responsible will receive secondary consideration. A key contribution of critical understandings of moral agency is that they enable, and push, us to examine the structures that shape responsible agents – as the responsible moral agent is no longer presumed to be natural or given. This leads to a further line of development, which is a turn towards thinking about responsibility as a social practice in order to better account for collective responsibility, as well as to open up critical interrogations of the construction of responsible agents.Footnote 20

Mervyn Frost offers an account of responsibility as a social practice that begins with a critique of the conventional account of moral agency, stating that a central feature of a practice-based account ‘is that it rejects any suggestion that we might start an ethical inquiry by considering the reasoning of a moral agent who could be envisaged as being in some sense free standing, socially unconnected, and outside of any particular historical period’ (2003, 90). Going further than either Erskine or Crawford, Frost focuses on the way in which moral agents are socially and historically constructed through practice. Rather than viewing moral agency as a capacity that agents possess, he argues that through our participation in social practices, which provide those involved with an account of their roles and the values they are pursuing, we take on responsibilities and are treated as responsible agents whose behaviour is susceptible to praise and blame. This avoids the persistent problem of whether or not human beings are truly free and focuses on the way in which assigning responsibility has a social function, namely as a part of maintaining and developing social norms. For Frost this makes the issue of assigning responsibility to collective agents less problematic, as the issue is not whether institutions or collective actors are moral agents, but rather whether our practices establish institutional agents that can be held culpable (2003, 92–94). In making this argument he distinguishes between integrated practices, in which responsible institutions and individuals are clearly identified, and dispersed practices, where such formal roles are more ambiguous (2003, 94–98). This account of moral agency as a feature of social practice avoids the difficulties of Erskine's and Crawford's understandings.

While Frost acknowledges that practices of responsibility will change and develop in response to broader social transformations and conflicts thrown up within practices, his analysis runs the risk of privileging convention because practices are defined by established norms and institutions that are not easily changed or challenged. He responds by arguing that the overall effects of a practice, rather than simply the actions of participants, should be subject to forms of accountability (2008). His argument appeals to the practice of critical theory, in which the non-volitional and unintended consequences of social structures are shown to result in objectionable practices and are thus opened up to reform. Yet this returns Frost to a persistent dilemma within critical theory: on what basis does one judge a practice from the outside (2008, 80–83) – which is the key riposte offered by an understanding of responsibility grounded by objective moral principles recognized by autonomous agents. My criticism is not that an external perspective is necessary but impossible, but that within Frost's own account, an external perspective on a practice would itself have to be part of a practice of general moral critique. Smiley makes this point when she suggests preserving elements of our modern conception of moral agency is vital because it provides the starting point for a general practice of moral critique (1992, 177–205). Given the way he sets out his constitutive theory, Frost creates a real difficulty by focusing on the historical construction of moral agency and the diversity of responsibilities that individuals take on through their engagement in social practices, which undermines the possibility of a critical moral viewpoint from which convention is challenged.

Alasdair MacIntyre addresses this problem directly in his discussion of responsibility (1999), arguing that the capacity for moral reflection and self-accountability that modern understandings of responsibility presume is difficult, if not impossible, to achieve in a modern bureaucratic society. MacIntyre rejects the notion that critical moral reflection is a possession of rational agents, rather it must be supported by social practices, but he does not think that a critical practice of moral reflection is likely given the type of agency that individuals living in contemporary societies develop (1999, 320–321). Contemporary agency, he argues, is fractured and compartmentalized, such that the integrity and constancy necessary for independent moral responsibility – as opposed to responsibilities taken on only as part of one's social roles – are available to a vanishingly small number of people (MacIntyre Reference MacIntyre1999, 325–327). While MacIntyre thinks that these are virtues that could be pursued to redeem a critical moral agency,Footnote 21 and that individuals should be held responsible for their shared role in creating societies in which these virtues are not developed, he more accurately identifies the barriers to this task than Frost does, and in doing so undermines the critical aspirations of Frost's theory.

MacIntyre focuses on the social conditions within modern societies that make moral agency difficult to sustain; this difficulty increases when we look to the international/global level. Ainley (Reference Ainley2008) draws out these difficulties when considering the attempt to attribute responsibility for large-scale atrocities. She explores how the structures of world politics legitimate forms of physical and structural violence that are as destructive as those forms identified as ‘atrocities’ (for example, wars of self-defence or for humanitarian causes, as well as the poverty and suffering caused, or simply allowed, as part of the global capitalist economy), while also questioning the idea that collective violence can be understood in terms of individualist and rationalist moral agency, such that we can identify criminal agents in the context of ethnic cleansing and genocide (Ainley Reference Ainley2008, 19–24). This suggests that international politics is not institutionalized in a way amenable to conventional accounts of moral agency. While Reference MacIntyreMacIntyre has a localized vision of the moral community, in which there is social space for the virtues required for responsibility (2007, 252–255 and 258–263), this option is not available if we are concerned with world politics, as the degree of diversity will not support a singular account of the virtuous individual, nor is their sufficient institutionalization to support the habituation of moral agency on his model. This failure, however, is not solely a feature of responsibility in world politics – the forms of identity privileged by the social ascription of moral agency are based on exclusions that MacIntyre does not sufficiently interrogate.

William Connolly does take up this interrogation, adding an examination of the construction of the moral agent as a privileged identity to our understanding of responsibility (Reference Connolly1991, Reference Connolly1995). Connolly investigates how the responsible agent is affirmed through the construction and devaluation of a constitutive other. When we affirm objective moral principles as part of the practice of holding individuals responsible, those who fail – or refuse – to respect their obligations are rendered as evil or lacking moral capacities. When we insist that responsible agents act autonomously we obscure the way in which privileged identities benefit from social support and deny the dependencies and weakness that afflict the marginalized. ‘We insist that we must be sovereign agents and our state must be a sovereign entity, and the desire to punish targets exactly those ambiguous constituencies whose conduct would otherwise call these modes of agency into question’ (Connolly Reference Connolly1995, 48). Therefore the responsible agent is also a political identity, held in place by force and authority, such that an unwillingness to acknowledge this risks naturalizing hierarchical relations of domination. It is this aspect of coercion that both Frost and MacIntyre underemphasize, as they appeal to the coherence and acceptability of practices of responsibility. Connolly does not suggest that there is a form of moral agency that can avoid this political element, the challenge is rather to acknowledge it and develop better practices of responsibility that remain open to contestation.

This insight into the importance of privileged identities has distinctive relevance in world politics. For example, as the privileged identities of the good individual/state are taken as givens, the identification of evil dictators or rogue states can be used to uncritically justify mass violence, while limiting our analysis to opposing and punishing evil, rather than considering the wider context that leads to collective violence. The role of the moral other is to render those who put privileged identities into doubt exceptional, so that the privileged are not found to be complicit. This can be seen, for example, in how the identity of ‘African warlords’ is used to understand conflicts on the content: the men responsible for such violence are irrational, violent, animal-like, and profoundly evil. ‘Even if you could coax these men out of their jungle lairs and get them to the negotiating table, there is very little to offer them… All they want is cash, guns, and a license to rampage. And they've already got all three. How do you negotiate with that’ (Gettleman Reference Gettleman2010).

Further, if we look to practices of war-making to discover forms of responsibility built into the practice itself – such as war crimes trials and collective security arrangements – without interrogating the privileged identities mobilized, we may too easily endorse the exclusion of certain violent acts from acceptable war making, which is problematic because assigning responsibility for exceptional crimes to abnormal individuals implies that those involved in the general practice of war are not guilty of immoral behaviour. ‘The war crimes trial marks a state of exception – a supposed deviation from ‘normal’ war – and employs a set of procedural logics that has as their main goal the conceptual and material excision of the war criminal from the landscape of legitimate war-related killing’ (Dauphinee Reference Dauphinee2008, 51). Criminal atrocities are the work of volitional individuals that have exceeded the moral limits of collective violence. They cannot be the consequence of soldiers predictably doing what men and women do in such situations – if civilian massacres, sexual abuse and rape, forced migration and other war crimes are normal outcomes of war, then the institution of war itself would be undermined. As an example, Mahmood Mamdani contrasts the violence in Darfur and Iraq, illustrating both the destructiveness of legitimate violence and the importance of assigning responsibility for illegitimate violence to culpable agents. Examining the number of deaths, we find that the human suffering from the wars in Sudan and Iraq are quite similar, difficult as it is to ascertain definitive numbers for either conflict (Mamdani Reference Mamdani2010, 58). Yet, the way in which the violence is understood is very different: in Darfur the violence is labelled as genocide, while in Iraq it is conceptualized as counter-insurgency (Mamdani Reference Mamdani2010, 59). Mamdani argues that Louis Moreno-Ocampo, lead prosecutor for the ICC, tries ‘to connect all consequences in Darfur to a single cause: Bashir’. Pointing out that ‘Moreno-Ocampo told journalists in The Hague, “What happened in Darfur is a consequence of Bashir's will” ’ (Reference Mamdani2008). This framing of violence in Darfur over-emphasizes one man's ability to control events and minimizes the wider social causes of the conflict. Yet, similar levels of violence in Iraq are seen as generally legitimate counter-insurgency, despite the controversy over the initial invasion, and there have been few attempts to assert singular individual responsibility for the violence and suffering, as the insurgent and sectarian violence has been understood as arising within a wider context.

Collective forms of punishment create a similar need for critical accounts of privileged identities, such as good states and decent peoples, which are contrasted with failed states and rogue regimes (Lang Reference Lang2008). In considering the place of punishment in world politics, Anthony Lang critiques the way powerful states impose accountability upon weaker states. Even where these practices of punishment may have justice as a motive they claim international authority by coercion, enable hierarchical political relationships and harm individuals who are largely innocent (Lang Reference Lang2008, 14–24). Attending to the construction of responsible agents as a political project highlights the power dynamics at work in practices of responsibility. The practical need to hold individuals responsible is interwoven with desires to punish and discipline, which helps to normalize contingent forms of agency, to not only regulate deviancy but also shield social practices from criticism. We can see this in the way dominant liberal states have sought to punish bad states, which are seen as security threats and sources of instability (Morton Reference Morton2005, 371–379). What this contrast obscures is the legitimized physical violence perpetuated by good states, the power hierarchies of the international system and the structural violence of the global economy, by assigning responsibility for harms and abuses to evil individuals or governments. This neutralizes the possibility that these negative consequences are the result of the normal operation of broader social structures.

The insights of MacIntyre and Connolly focus on the role of social institutions and identities in practices of responsibility, which provide necessary tools but also present challenges if we seek reconstructed practices of responsibility. The move to a practice-based account of moral agency suggests several requirements for developing critical practices of responsibility. First, these practices should take into account structures that affect moral agency, including damaging structures such as a war-prone international system and a global capitalist economy that sustains mass poverty, as well impediments to critical moral agency thrown up by the psychology of collective violence and the diverse and fragmented social space in which world politics are conducted. Second, it should acknowledge that moral agency and practices of responsibility are developed through forms of coercion and are inherently political, which reveals the hierarchical relationships that define current international/global practices of responsibility. This leads to a final requirement: we must ask what kind of practices of responsibility would be best without presuming that the answer is incontestable or final, which brings out the question of how practices are established and maintained. Critical interrogations of responsibility must ask who has power in, who benefits from, and who wields authority in our practices (Smiley Reference Smiley1992, 255–272), while also offering a vision of just social relations. In the final part of this section, I want to suggest that turning to Dewey's work provides a promising way forward, as taking up aspects of his account furthers a critical reconstruction of international/global practices of responsibility.

A Deweyan account of moral agency fits within the broader practice-based model discussed above, in which responsibility is understood not as a metaphysical attribute of individuals but as a social means of regulating human behaviour.Footnote 22 To hold an agent responsible is to insist that her behaviour takes social obligations into account, but being held to account also moulds individuals and collective actors into agents capable of changing their behaviour in specific ways (Dewey Reference Dewey2002, 315). Responsible agency is achieved through the formation of individual habits that enable self-control and reflective action as a feature of the character of the agent, as well as the establishment of customs and institutions that enable us to continually reconstruct our social roles and the ends pursued by the community in a critical and democratic manner.

Dewey focuses on the way individual impulses are cultivated into habits, which are established ways of acting in pursuit of given ends that are largely automatic, in order to understand both conventional practices of responsibility and to offer a guide to their reconstruction (2002, 316–318). Unlike MacIntyre, whose account depends upon the maintenance of tradition, Dewey's view of moral agency is adaptive, emphasizing the possibility for both individual habit and social custom to change in response to new conditions, while maintaining a critical ethical standpoint. In this, he splits the difference between MacIntyre's and Connolly's insights by offering an account of moral agency that disavows a single or permanent understanding of the good life in its critical dimension while also emphasizing that in any specific situation we can and do assign responsibilities. A Deweyan understanding of moral agency is best described as both situated and democratic because along with analysing the idea of responsibility in terms of its social function, he argues that our practices of responsibility should enable us to see the effects of our social conditions as well as our volitional actions, while also empowering us to remake our practices of responsibility democratically.

For Dewey responsibility is central to the ongoing development of moral agency, as it makes explicit the type of individuals we take ourselves to be, holds us accountable to our ethical standards and defines our social obligations (Reference Dewey2009, 127–138 and 152–158). Holding someone responsible is a judgment of the consequences of his actions, of whether he upheld his obligations and how he should be held accountable. Dewey, however, insists that there are further questions to be asked about the consequences of our conventional practices, namely whether they uphold the ideals they are intended to serve and whether those ideals are worthy of our effort. Where practices of responsibility become problematic by not upholding the ethical ideals they are intended to serve or leading to forms of individual personality and community life that are found wanting, they call out for critical reconstruction, which includes an analysis of the consequences of our customary and habitual forms of agency, a reconsideration of the ends that our established modes of action are intended to secure, and practical action to test our reconstructed understanding of responsibility.

Dewey places special emphasis on the disruption of habit and custom, on those moments when they fail as guides to action, in order to motivate and guide changes in social practices (Dewey Reference Dewey1991, 316). Whether practices are reaffirmed or altered they continue to be evaluated by their internal ideals, by the success with which they uphold the principles and relationships that they are intended to enable (Dewey Reference Dewey2002, 326–327). Moral agency is situated in that it grows out of particular contexts in which the capabilities and obligations of individuals are developed and enforced by social institutions, but it is not dependent upon convention, as changing conditions and the varieties of human experience give impetus to reconstruct our understandings of what we expect of individuals and institutions, to reconsider the ends we pursue and to alter our communities in profound ways. Returning to the broader philosophical understandings of moral agency discussed above, Dewey is able to maintain the critical edge that Kant's insistence on the separation of morality from convention offers, but without losing sight of the social function of practices of responsibility that comes from the Aristotelian tradition. Beyond splitting the difference between these influential approaches, Dewey also brings out the inherently contested and political nature of our practices, as they are always drawn from our social context and therefore always partial and open to future reconstruction.

The development of critical moral reflection is a key aspect of moral agency for Dewey, as it enables individuals to act as participants in the ongoing reconstruction of social customs, giving rise to forms of conduct that express the individual's will in, and accountability to, the moral community (1969). This critical reflection, however, is not dependent upon affirming the modernist conception of the free and rational agent. Dewey's understanding of moral agency is democratizing in that it offers an analysis of how agency develops and is rendered problematic, while also suggesting that moral agency should be reconstructed such that agents are socially and politically empowered to participate in the ongoing formation of social practices. In seeking improved forms of critical moral agency, we should acknowledge that the capabilities of actors depend upon the institutions that define society and the types of relationships that define the community, such that a better account of moral agency is one based on empowering actors to have an active role in defining social practices and ethical ends, and recognizing that they are also responsible for themselves and their place in the community. Given this, the question of responsibility changes. Practices of responsibility that make us into agents capable of both determining and following our ends lead to forms of community in which every member is asked the ethical question: What kinds of individuals and communities should we aspire to become? This is achieved by holding agents responsible not only to the discipline of just social institutions but also to empower them to challenge and reconstruct those institutions (Dewey Reference Dewey1991, 56–58). The democratic ethos that Dewey supports emerges from his understanding of responsibility and suggests a forward-looking programme of ethical reform (Reference Dewey1991, 88), which can inform our thinking about international/global practices of responsibility.

The current limitations of responsibility in world politics and the potential for reconstruction

In the introduction I suggested that emerging practices of responsibility are limited by an excessive focus on individual accountability and inattentiveness to the politics of assigning responsibility in world politics. The key practical problems that come out of this critique are that contemporary practices of responsibility place too much emphasis on individual accountability, and have tended to be hierarchical and exclusive. Thus far I have highlighted the way critical scholars of responsibility have addressed these issues. In what follows I offer some initial explorations of how a Deweyan approach informs these criticisms and suggests possible ways of reconstructing practices of responsibility in world politics.

Action taken to reconstruct social practices, Dewey argues, should be participatory, enabling each member of the community to have a say in the social relationships and institutions that affect her life (1927, 143). In turn, a reformed practice of responsibility should support a democratic moral agency. For Dewey, this means that we should encourage forms of agency (Reference Dewey1931) that are critical and active in evaluating ethical ends, enabled by social relationships and institutions that foster cooperative decision making. Democratic moral agency corrects key defects identified in contemporary international/global practices of responsibility, namely the naturalization of privileged forms of agency through exclusion and violence, and the occlusion of structural conditions in favour of a focus on the actions of autonomous individuals.

The limitations of our contemporary practices of responsibility can be seen in the ICC's indictment of Joseph Kony, leader of the Lord's Resistance Army in Uganda. This was the court's first indictment and in many ways is an obvious case, as Kony is an extremely unpleasant character that has fomented appalling violence in Uganda. Despite this, it is insufficient to reduce either the conflict in Uganda to Kony's intentional actions or his motivations to an irrational evil will. Louis Moreno-Ocampo, says of Kony, ‘He will never make peace, his goal is to abduct children, torture and kill and so he must be stopped’ (Ladu Reference Ladu2010). The danger of constructing criminal identities in terms of irrational evil or otherwise exceptional forms of agency is that the identities of ‘good’ and ‘bad’ are insulated from contestation and the social conditions and power relations that enable them are naturalized. In the Kony case, the point goes beyond claiming that the ICC's actions in Uganda may undermine the peace process (Glasius Reference Glasius2009, 506–508). The court's cooperation with the Ugandan government, despite atrocities committed by its soldiers, is not incidental, and the denunciation of Kony as a moral monster is sadly not based on the singular horror of his actions, as the ICC position on the conflict in Uganda reflects the presumption of legitimacy afforded to state violence. Opening up the question, what are the social conditions that enabled Kony to operate with impunity for over 20 years, Adam Branch points to the importance of the failures and violence of the Ugandan government in generating opposition among the Acholi people from whom the rebels drew support, and the complicity of international groups providing aid to support the camps in which the Acholi have been forced to live (Branch Reference Branch2007). Further, the Ugandan government's manipulation of the ICC investigation, in which the abuses of government forces are not being investigated, has arguably prolonged the conflict (Branch Reference Branch2008). What is revealed by refusing to be satisfied with explanations for violence that depend upon the irrational evil of Kony, is that in the case of international criminals it is all too easy to see these individuals as extraordinary people acting in ordinary times – as moral monsters – but a key feature of much international crime is that the criminal is in fact acting in extraordinary circumstances.

Within emerging practices of responsibility Kony's actions need not, even cannot, be examined in social context because the ICC's model of moral agency assigns culpability only to individuals. The violence of government forces is treated as legitimate; the rightful use of state authority is presumed, except where the actions of an individual soldier goes too far, and thereby the everyday violence of the state is excused. The Acholi people are confined to government camps as a matter of national security and anti-terrorism measures, the international aid which sustains these camps and the governments counter-insurgency actions in the north are exempt from criticism because they are supported and enabled by powerful states, such as the United States, leading to the criticism that the ICC investigation is only a sign of ‘the depoliticizing language of humanitarian intervention [that] serves a wider function; [as] “humanitarian intervention” is not an antidote to international power relations, but its latest product’ (Mamdani Reference Mamdani2010, 59). The state violence in Uganda is enabled by practices of international/global responsibility that focus excessively on the accountability of individuals and exclude an examination of the role that the sovereign state system plays in sustaining and encouraging conflict.

Where Connolly encouraged us to expose those identities that are demonized and disempowered through conventional understandings of moral agency, as well as the disastrous social conditions that it ignores (1995, 41–74), Dewey enables us to put those experiences to work in identifying problematic situations and motivating social reconstruction (Dewey Reference Dewey2004, 113–115). The everyday experience of marginalized individuals and communities is significant, as the experience of poverty, abuse, and neglect reveals pernicious customs and institutions that hold onerous social relations and ideals in place. So, rather than making Kony's violence exceptional, and therefore limiting responsibility to one man, a Deweyan approach would ask us to understand how such terrible actions came about, to know not only how a ‘Kony’ was made but what can be done to prevent such violence in the future.Footnote 23

These insights, into the way the experiences of those who are excluded and marginalized can facilitate critical reconstructions, can be applied at the international/global level. The examples used here come from post-colonial African states and point to common problems arising from these states’ placement in the international system and global economy, suggesting roots that go much deeper than the actions of singular evil individuals. The response, both politically and intellectually, that treats the violence in Uganda as attributable to individuals avoids troubling questions while also uncritically supporting the view that there is a privileged liberal community that is empowered to intervene, at times violently, to hold these individuals accountable to putatively universal standards. If our practices of responsibility are excessively focused on assigning blame to autonomous individuals, we risk failing to address wider social conditions that lead to the harms we want to prevent. In contrast, our conventional understanding of moral agency focuses on the commission of war crimes, human rights violations and other mass atrocities as singular events requiring individual accountability, both in the form of trials and in terms of enforcement and intervention (Orford Reference Orford2003, 35–36). If evil is caused by the free actions of autonomous agents and moral principles are rationally knowable by everyone, then ‘Morals withdraw from active concern with detailed economic and political conditions’ (Dewey Reference Dewey2004, 113), and need be concerned only with confronting evil. A Deweyan understanding of responsibility, then, suggests a dramatic reorientation in our practices, one that focuses on the context in which atrocities are made possible and carried out.

A further example of the limits of a reductive understanding of moral agency can be seen in the international community's response to the civil war in Sierra Leone. The general explanation that the conflict was about access to the country's diamonds, and the prosecution of Charles Taylor and other individuals, presents an easily understandable narrative of criminality, in which the violence was motivated by immoral greed and irrational hatreds (Keen Reference Keen2003). This basic narrative obscures both the complex social conditions that lead to the civil war and the part that was played by external actors, particularly structural adjustment policies of the International Monetary Fund (IMF) and World Bank (Keen Reference Keen1997, 67–76). David Keen's work suggests that understanding the conflict in Sierra Leone requires an appreciation of the poverty, inequality, social divisions and corruption in that country, which reveals that conflict over diamond resources was not the result of sinful greed, but a rational reaction to poverty and government policies taken in response to the structural adjustments required by international financial institutions (2005, 73–89). Further, the rebel's use of child soldiers and extreme forms of violence should be placed in a wider context, including the exclusion of young people from education and work, and the absence of alternatives to violent criminality as a means of social advancement (Keen Reference Keen2005, 78–80). While the causes of the conflict were diverse, Keen highlights the role of liberalization policies demanded by international institutions and supported by the international community. Whatever the theoretical virtues of these neoliberal policy prescriptions, the reality is that they did not take account of how government officials and private individuals would respond to the weakening of the state and privatization of services and resources (Keen Reference Keen2005, 74). The result was increased poverty, greater incentives for diamond smuggling, the reduction of social services – which importantly undermined support for the state and in the case of education encouraged radical student groups – and new opportunities for corruption, both for local chieftains and the military (Keen Reference Keen2005, 84–86). To insist on a wider account of the causes of violence is not to excuse the individual actions of Charles Taylor and other defendants being tried by the Special Court for Sierra Leone, but it does point to the limits of contemporary practices of responsibility. It is not just that the heads of the World Bank or IMF are not being brought before an international court (a provocative suggestion), but that wider social causes and more diffuse forms of responsibility remain unexamined. This is especially problematic for international/global responsibility, as collective violence and mass suffering never have simple causes or single lines of accountability. Further, the failure to address social causes and structural violence reinforces established power and authority.Footnote 24

A Deweyan perspective emphasizes that thinking about responsibility on a rationalist and individualist model of agency promises a hollow transformation. If our current practices of responsibility obscure systematic deprivation and mass violence, while naturalizing the construction of moral agency in individualistic terms, a critical practice-based account of moral responsibility should deepen and widen our analysis, while suggesting more radical paths for reconstruction. What has been said thus far is general, and concerned primarily with the needed conceptual reconstruction of practices of responsibility – but a detailed examination of specific practices of responsibility is not possible here. Therefore, in conclusion I offer suggestive remarks that point to how a Deweyan account of responsibility would support ongoing developments in practices of international/global responsibility and suggest further reforms.

With the creation of the ICC and the ascendency of responsibility as a constituent of legitimate sovereignty, international criminal law and humanitarian intervention have become influential practices, and while I have suggested that a Deweyan account of moral agency is preferable to the conventional model that is embodied in these developments, the move to consider responsibility as a practice is not hostile to the goal of expanding justice beyond national borders, undermining the privilege of state agents and where appropriate holding individuals accountable through trials. In fact, it is the experience of the inadequacy of domestic institutions to constrain violence, along with the unacceptability of an amoral understanding of world politics, which creates a need for international/global practices of responsibility. A Deweyan account, however, is more open to institutional transformations of world politics, not only by advocating for democratic forms of action that would insist on a central role for victims, affected communities and the public at large, but also insisting that the wider social structures that feed into violence must be reconstructed along with the behaviour of individuals. Initial practical starting points would include expanding on the ICC's inclusion of victim's representatives, reconsidering the use of amnesties or the revocation of indictments if a trial would no longer serve the interests of those most affected, and focusing more on supporting local justice mechanisms over holding international trials far from the societies seeking reconstruction. Further, it would also entail rethinking the purposes that the ICC or R2P serve, moving beyond a concern with protecting victims and oriented towards empowering individuals and communities. Much more so than contemporary practices of responsibility, the democratic and situated practices that a Deweyan account of responsibility suggest, place the presumed legitimacy of state-violence under question, undermine the assumption that world politics should be structured by the sovereignty of the nation-state, or that the organization of the global economy can be exempt from scrutiny. Further, the transformation that attempts to render world politics more ethical requires, from a Deweyan perspective, is not pre-given, as is the case with cosmopolitan or internationalist practices of responsibility. The value of a cosmopolitan global order, for instance, cannot be assumed, as the need to reconstruct the problematic politics of the domestic and international spheres of the sovereign state system need not necessarily lead to a singular global political order. A Deweyan-internationalism (Cochran Reference Cochran2010) would begin from the identification of particular problems and the creation of publics that transcend the territorial state (Dewey Reference Dewey1927), and from there ask what forms of institutional reconstruction are best. Identifying common problems and forming new transnational publics is part of the process of reconfiguring moral agency beyond the form it takes in a conventional national context, which requires the development of shared ideals, common practices and new forms of institutional order. The form that this might take is not pre-given and even where the understanding of responsibility we find in contemporary liberal practices are seen to be valuable, their value is neither guaranteed nor assumed (Bray Reference Bray2009; Hoover Reference Hoover2011, 166–198).

The upshot of this prescriptive modesty is that a Deweyan account of responsibility provides critical insights into how practices form and when they may be in need of reconstruction, while also providing a general democratizing ethos of just social relations that can guide our thinking on this question (Manicas Reference Manicas1981). A Deweyan practice of international/global responsibility might support attempts to hold individuals accountable for enabling conditions of violence as well as their own individual acts, for example. This accountability would not, however, be based simply on punishing or constraining evil individuals – instead the act of holding accountable would focus on controlling and improving the actions of individuals and groups, as well as the social conditions in which they find themselves, in particular by attending to the ability of excluded actors to influence social behaviour and to act with greater freedom as a result. For example, in the case of Uganda, the power of the army and state officials to control the context in which violence is ongoing gives us reason to hold them responsible in ways that rebels are less likely to be. Using practices of responsibility to interrogate the role of powerful social actors would not only potentially expand accountability based on command responsibility or joint criminal enterprize, but could also lead us to consider the role that communities play in enabling violence, opening up difficult questions of how they could be held accountable.

In international criminal law the increasing use of the joint criminal enterprize as a charge in international trials points towards less individualistic and voluntarist forms of responsibility (Powles Reference Powles2004, 606–619), though to be properly effective the idea of collective responsibility would need to be more fully developed in international criminal law. More broadly, the move to attribute human rights responsibilities to international organizations, NGOs and multi-national corporations also develops alternative notions of responsibility (Clapham Reference Clapham2006), and a Deweyan approach to responsibility would push us to engage with and expand upon these creative developments. These developments are controversial in part because they challenge conventional understandings of moral agency, therefore the effort to reconstruct practices of responsibility both informs and is supported by alternative accounts of moral agency. The scope for accountability could potentially be extended to consider the unique responsibility of international/global actors, such as hegemonic states and international organizations, which support and develop social institutions that sustain exploitative and violent practices. Not only could international organizations like the IMF potentially be held responsible, but reconstructed practices of responsibility could support efforts to assign obligations to powerful non-state actors that generally avoid consideration, particularly multi-national corporations (Ruggie Reference Ruggie2007), but which could include international NGOs, religious and civic groups with international influence and even powerful individuals. Beyond supporting controversial developments in international/global responsibility, the practice-based account of moral agency developed here also reorients practices of responsibility in world politics, both legal and political, towards wider programs of reconstruction and reform, and away from the interests of the powerful and privileged actors in world politics. Because the focus of our practices of responsibility would be on the experience of those people facing social problems and violent catastrophe, practices of international responsibility would not primarily be concerned with punishing individual violations, but rather with reforming and rebuilding damaged societies, and reconstructing international/global political relations and social structures. This opens up space for alternatives to international criminal proceedings and provides impetus to challenge international norms and hierarchies that limit solutions to intractable social conflicts. For example, it gives additional support to alternatives to prosecutions, such as those partly developed through the Special Court for Sierra Leone, which may have a better chance of ensuring peace and enabling reconstruction. The key features of a democratic account of international/global responsibility would include: a focus on the effects of individual and social action on people's everyday experience – such that the problem of responsibility is not only about holding actors accountable for specific wrongs, but also holding institutions and collectives accountable for practices that sustain objectionable conditions; a forward-looking concern for improving the conduct of actors as well as the institutions and conditions that develop moral agency; the direct participation of affected communities in practices of responsibility, rather than a reliance on international legal prosecutions, which are often held far from the people involved, and foreign interventions or programs of development by disengaged experts; and a recognition that practices of responsibility express an ethical and social judgment that affirms particular ideals and social institutions – and that this exercise of judgment is ongoing and open to contestation rather than fixed.

The argument presented here is only an opening, but the hope is that by criticizing the often unarticulated presumptions about moral agency that inform practices of responsibility in world politics the ideas presented here open up an important line of inquiry. Emerging practices of responsibility are to be encouraged, but a focus on individualist and rationalist accounts of responsibility risks missing influential social conditions and forms of political privilege that exacerbate and contribute to profound suffering. The appeal to Dewey's democratic understanding of responsibility is intended to respond to this blind spot by altering how we think about responsible agency and opening up new lines of inquiry that may lead to reforms in international practice.