Types of moral responsibility

What kinds of responsibility are we concerned with when we talk about moral responsibility? There are two main connotations of the term, which I will distinguish as attributive responsibility and assignment responsibility.Footnote 4 The attributive conception of responsibility helps us decide to whom we should attribute (retrospectively) praise or blame for a particular state of affairs. Frequently, attributive responsibility is interchangeable with the concept of causal responsibility – one acquires attributive responsibility for a state of affairs by causing (or permitting) it to exist (or continue). Alternately, we sometimes feel that those who have not, strictly speaking, caused a state of affairs should nevertheless be held responsible as if they had, such that we are willing to attribute moral responsibility to them on the basis of negligence, strict liability, or some similar criterion.Footnote 5 In any event, attributive responsibility is backward-looking or retrospective: it focuses mainly on such questions as who deserves credit or blame for what has occurred or developed.Footnote 6

The assignment conception of responsibility, by contrast, is forward-looking or prospective: it asks not who is to blame for a state of affairs but rather who should bear the burden of caring for it. Having assignment responsibility means that a state of affairs is in your charge, or up to you, that it is your duty or obligation or role to deal with it. In this sense, responsibilities are very like moral duties as that term is popularly used in contemporary philosophical ethics, but admit of somewhat broader application since they are compatible with consequentialist as well as deontological systems of thought (Goodin, Reference Goodin1995: Ch. 5).Footnote 7 In this paper, I will discuss both attributive responsibility and assignment responsibility, but I will principally be concerned with moral responsibility understood as assignment responsibility.

It is of course possible to have both attributive and assignment responsibility for the same state of affairs – indeed, we should expect that this will typically be the case. Nevertheless, practically all those who have written on responsibility agree that assignment responsibility cannot simply be derived from attributive responsibility in any straightforward sense. There are some cases in which, though individual x is to blame, it is instead up to individual y to clean up x’s mess: as for example where x lacks some relevant necessary capacity (where x is a child, it may be up to her parent y to make things right). Likewise, while questions of capability are clearly relevant to assignment responsibility, there will be plenty of cases where those questions are not sufficient to assign all responsibilities, as when x and y are both capable, but x is to blame for the situation, or where both are capable but cleaning up is y’s well-remunerated job. At least in many of these cases, therefore, we will want to introduce considerations about the past as well as about the future into our deliberations.Footnote 8

So typically, there will be at least some connection between our assessment of attributive responsibility and assignment responsibility in a given situation – and this connection is part of what enables us to distinguish both personal and collective responsibility from the more general duties associated with humanity and benevolence alone. In this regard, Robert Goodin (Reference Goodin1995: Ch. 16) invokes a traditional distinction between general moral duties or responsibilities, and special moral duties or responsibilities.Footnote 9 The former are those duties we have toward all humanity universally; the latter are those we hold toward specific persons due to specific features of our social connection to them. Assignment responsibility, in the sense I will be using here, refers only to special moral responsibilities. For, in contrast to a general moral imperative, an assignment responsibility will typically be a moral imperative derived specifically from one’s actions or from action-related features of one’s situation.

How is collective responsibility assigned?

How is responsibility attributed or assigned to collectives or groups?Footnote 10 Contemporary theoretical accounts of moral responsibility, when they take up the problem of collectives, have tended to offer one of three alternative answers to this question, each identifying a relevant form of connection between some collective and the state of affairs under consideration.Footnote 11 One version assigns responsibility on the basis of a claim about causation: a group acquires responsibility for an action or outcome to the extent that they themselves caused that action or outcome to happen (whether directly or indirectly). For example, if a group of well-trained warriors decide together, on their own authority, to go out and invade a neighboring country, they are morally responsible for that action and its consequences.

The causal mechanism can of course be more complex. To take an easy case as an example, if in a direct democracy authorization by the electorate is a necessary and sufficient condition to provide for the invasion of a neighboring country, and should the electorate choose of their own free and informed will to endorse a call for invasion, then we can say, at least of those citizens who voted for the war, that they collectively are morally responsible for that action and its consequences. However, though this account resonates with our core intuitions about moral responsibility, it is also a very hard account to square with the mechanisms of modern democratic life. In contemporary representative democracies, citizens rarely, if ever, have the opportunity to affect such policy choices directly. If direct causation is required to assign responsibility for actions and outcomes, then the complex mechanisms of modern representative government will leave a large share of moral responsibility for state action assigned to no one in particular.

Perhaps because of these difficulties, many contemporary moral theorists have focused instead on assigning responsibility by means of association. The crudest form of this associative conception of responsibility occurs when responsibility is assigned by means of identity: those who belong to a certain national, ethnic, or religious group are assigned responsibility for those actions and outcomes caused by their fellow nationals, ethnics, or co-religionists. The most extreme versions of this mode of assigning responsibility are notorious, and indeed it is these identity-based methods that for much of the twentieth century gave the whole idea of ‘collective responsibility’ a bad name.Footnote 12 This notion has not by any means disappeared: the attacks by Al Qaeda and other terrorist organizations have implicitly justified their targeting of American, European, Israeli, and other civilians by reference to those citizens’ national identity, in much the same way that Western forces have detained and tortured many civilians based not on hard evidence but rather on their status as Arabs or Muslims. What seems unfair about these assignments of responsibility by association is that they incorporate no element of voluntary connection. Identity is to a considerable extent involuntary; so too are the associations which an identity may carry. Yet we feel it is not right to attach much responsibility to someone – particularly, to attach any attributive responsibility to them – unless they have in some way done something to acquire it.

It is probably in response to this weakness of simple identity-based attributions of responsibility that several contemporary theorists have sought to give an account of responsibility mediated not through identity simply but rather through a more active and consensual form of association. Since few forces in our world exert more causal influence than the nation-state, it should not be surprising that the most interesting accounts of responsibility through association have been accounts of national responsibility.Footnote 13 Farid Abdel-Nour, in his article ‘National Responsibility’, argues that national responsibility is distributed principally by means of feelings of national pride. Abdel-Nour argues, ‘An individual’s pride in the achievements of her nation connects her imaginatively to the actions of those who brought them about, and ‘in this way (and only in this way) can she be meaningfully implicated in the cause of distant outcomes by virtue of her national identity alone’.Footnote 14 Thus, to the degree that I take pride in the actions of my government, I acquire responsibility for them; moreover, perhaps by implication, to the degree that I do not feel pride or even feel shame at those actions, I evade responsibility for them when they are blamable.

There is something to be said in favor of this view. In particular, I think it helps to expand (in ways that are ordinarily counter-intuitive to us) our sense of the many kinds of actions, attitudes, and relationships that can incur collective moral and political responsibility.Footnote 15 But Abdel-Nour’s account also specifies the mechanism by which collective responsibility is acquired in such a way that the liability of citizens for the actions of their states is sharply limited. For while identity simpliciter as a mechanism for acquiring responsibility affords too little room for the role of voluntary assent, Abdel-Nour’s account of national responsibility seems to assign too much importance to voluntariness.

On Abdel-Nour’s account, it is attitudes such as pride and emotional assent that help to identify those who hold them with the actions of their nation. Since responsibility is transmitted through identification and emotion, a voluntary rejection of that identification and emotion may be sufficient to detach oneself from the responsibilities associated with these attitudes.Footnote 16 ‘National responsibility is actively incurred by individuals with every proud thought and every proud statement they make about the achievements of their nation’, Abdel-Nour argues, but he goes on to assert that this ‘is also the limit of their national responsibility, which only extends to the actions that have historically brought about the objects of their national pride’.Footnote 17 Abdel-Nour later reiterates that a citizen acquires collective responsibility for the actions of her nation-state ‘by means of her national identity alone’ and that whether she bears present responsibility for such past atrocities as slavery or genocide ‘will depend entirely upon her objects of national pride and the relationship between the actions that brought them about, on the one hand, and the horrors in question on the other’.Footnote 18

I do not want to deny that the phenomenon Abdel-Nour describes as national responsibility exists or that it can, in many cases, be morally significant. But by tying national responsibility so closely to feelings of pride and emotional identification, Abdel-Nour significantly limits the potential reach of responsibility for the actions of nation-states. Ultimately, Abdel-Nour’s view, like that of several other contemporary theorists, is an affect-based model of collective responsibility.Footnote 19 But the particular problem with the affect-based view in this context is its suggestion that divesting oneself of pride and emotional identification with the achievement’s of one’s nation-state might be sufficient to relieve oneself of responsibility for the consequences of its actions.Footnote 20 Such a view might lead us to conclude prematurely that collective responsibility is assignable only to those whose consent and indeed commitment to national projects had been firmly secured. At the level of individual ethics, such a view would, in many cases, seem to encourage political quietism, as good people might seek to avoid associating too closely with the affairs of their national communities for fear of incurring unwanted moral responsibilities for the policies those communities pursue. At the larger level of normative theory, an account that derives responsibility for state action solely from its citizens’ affective attachments and voluntary endorsements risks leaving a great deal of moral responsibility assigned to no one in particular. For surely there are many actions taken by states which many of its citizens have not endorsed, but which nevertheless are taken – or at least claim to be taken – in their name, through means supplied by them, for ends that purport to be their own collective interests and purposes. If states can act on this basis without committing its citizens to responsibility for state actions, the modern nation-state can effectively become a very efficient responsibility-laundering machine: taking actions in the world with the gravest consequences for which there is simply no one available to take responsibility.

Among recent theories of collective responsibility, the general approach best suited to address these deficiencies of the causation- and affect-based approaches is a version of the social connection model defended by Iris Marion Young. On Young’s account, we should understand collective responsibility as being acquired and transmitted most often not as a result of direct causality or through affective identification but rather through our active – though not necessarily causal – participation in a chain of social connections that helps to create or sustain the negative outcome in question.Footnote 21 The social connection model, Young explains, ‘says that all agents who contribute by their actions to the structural processes that produce injustice have responsibilities to remedy these injustices’.Footnote 22 Young’s social connection model aims to identify the full range of people who may bear some degree of (both attributive and assignment) responsibility for a state of affairs, but is forward-looking in its focus on how to ensure that an essentially shared assignment responsibility for remedying the situation is discharged as fully as possible by means of various forms of collective action.

The social connection model constitutes a significant improvement over either the causation model or the affective identification model. Ascribing collective responsibility through social connection makes responsibility more widely available than it would be under a strict causation model. At the same time, the social connection model also makes responsibility farther reaching and less easily evadable than it would be under the affective identification model – for now, in order to evade being caught up in collective responsibility, individuals must make not only internal changes to their feelings but also substantial external changes to their behavior as well.

However, Young herself expresses skepticism about the validity of ascribing responsibility by means of ‘nation-state membership’, on the grounds that ‘political communities have evolved in contingent and arbitrary ways that are more connected to power than to moral right’.Footnote 23 This argument rightly asserts that there are many important ways to acquire responsibility apart from state membership (illustrated by Young’s particular case of concern, the responsibility of consumers for international trade practices they help to support). This however does not show that state membership is not a form of social connection in Young’s sense, only that we should be skeptical about attempts to make state membership the exclusive focus of our analysis of responsibility-laden social connections. Young’s own criteria, however, point to compelling reasons why the state must constitute a key mode of responsibility transmission:

This, I want to argue, aptly describes the way in which many citizens are involved with the activities of their states; consequently, the social connection model may apply equally well to state membership in addition to other forms of social connection.

The authorized state account asserts that there are different types of social connections we have with one another, and that it stands to reason that the parameters of responsibility will vary from type to type. In this paper, I will be adopting a modified version of the social connection model as an overall approach, but then going on to explore one particular type of connection – the kind of connection that citizens have through their states to other individuals affected by their states’ actions. Assuming that this is one widespread form of social connection, I will then go on to ask what kind of connection authorized state representation really is, and consequently what kind of collective responsibility we would expect to see in cases where that is the dominant mode of social connection.

Essential features

If the nation is not the only means by which we acquire collective responsibility, what alternatives are available? The idea I want to pursue here is that there is another morally significant form of association within the national community, apart from that provided by citizens’ cognitive assent or affective identification with the community as they imagine it. If we return to the powerful and extensive form of moral association discussed above, the nation-state, I suggest that with respect to certain problems, we should perhaps detach our attention from the nation side of the equation and refocus it on its companion concept, the state. In the next three sections, therefore, I want to sketch out an alternative account of collective responsibility grounded in the state’s role as a representative or agent of its citizens. This alternative account will propose that an authorized state model of collective responsibility, as opposed to the identity-centric or affect-centric models, best describes how collective responsibility is acquired in some (though explicitly not in all) circumstances, including particularly (but not exclusively limited to) the state’s actions in the international arena.

We can begin by supposing that in a case in which all these factors – the citizens’ general authorization of the state’s powers, their affective endorsement of its specific actions in a given instance, and the causal indispensability of their support in bringing those actions into effect – were in place, the grounds for assigning responsibility to citizens collectively for the state’s actions would be clear and indeed over-determined. In this ideal-type situation, at least, they would certainly be responsible, for all the potentially relevant conditions – causality, affective identification, and authorization – would obtain. The question then is: under what circumstances would a general authorization of the state alone be sufficient to assign responsibility to citizens, regardless of their causal role in or affective identification with the policy in question? We can certainly imagine cases where, causality and state authorization obtaining, affect might not be necessary to assign responsibility: for example, in the case of a very lazy citizen population, which could easily prevent a harmful policy but fails to, even though they do not approve the policy. Similarly, we can, with little difficulty, imagine cases where, affect and state authorization obtaining, causality might be dispensable in assigning responsibility: for example, in the case of a jingoistic citizenry who supported a government’s wrongful war warmly, even though the government would certainly have undertaken it without their approval. So neither causality nor affect is a necessary condition for assigning responsibility. The question is: when is state authorization alone a sufficient condition?

The answer is to be found in the implicit logic of the authorized state model, which points to the peculiar ends and means associated with the state as grounds for connecting its actions with its citizens in a chain of shared responsibility. I will say more about these peculiar ends and means below, but to anticipate: the state claims to pursue certain necessary social functions (such as protection and provision of public goods) through powerful and effective means (such as violence and coercion exercised under unitary control), and claims further that its citizens authorize it to pursue such functions by such means. So where these conditions obtain – where the state does pursue such functions through such means – it is there that the authorized state model of collective responsibility will best apply.Footnote 25

That is the basic intuition behind the authorized state account, which I will unpack in greater detail below. However, for expository purposes, I want to limit my initial sketch of this alternative authorized state model to the easiest and most plausible case, and therefore I propose that we begin by assuming three favorable conditions that apply to many but by no means to all political circumstances. First, I propose here to restrict my analysis in the first instance to the responsibilities of citizens of democratic states whose authorization of their state is essentially consensual. (If the theory seems plausible in that circumstance, we can then consider how far it may or may not generalize to subjects of non-democratic states.) Second, I want to consider initially those cases in which citizens witness and judge their state’s actions through a public process that is reasonably fair and open. (Again, if the theory seems agreeable in that circumstance, we may then ask to what extent deception and manipulation might tend to mitigate citizen responsibility for state action.) And third, I wish for the moment to keep entirely separate the question of punishment or indeed of any coercive means of enforcing the responsibility we might attribute or assign to citizens through state action. (Once again, if the argument framed this way holds up in broad form, we may then ask how far we wish to permit those to whom remedial action is due to employ force or coercion to punish or otherwise recover what is owed to them.)

Authorization and the social contract

What then does it mean to say that we can acquire collective responsibility for our state’s action because the state acts as our representative? To answer this question, our best resource will be the body of thought which most thoroughly tries to specify the normative dimensions of the relationship between citizens and their state: the social contract tradition. From Thomas Hobbes to John Rawls and beyond, the social contract tradition is a rich and varied source of thinking about the relationship between citizens and their state, and the notion of the social contract itself is powerfully intuitive in the modern world. The concept also has notorious analytical difficulties, especially with regard to the crucial issues of authorization and consent.Footnote 26 The most successful attempts to overcome these difficulties have therefore offered a more modest, bare-bones account of the contract itself; one that downplays the literal grounding of authorization in consent, and instead interprets citizen authorization in terms of the indispensability of state functions.

In what follows, I will rely on one such contemporary account, that of George Klosko, which grounds state authority not in voluntarism or literal consent, but rather in our shared need for the state as a background condition for living acceptably commodious lives.Footnote 27 In the modern world, the state has become the indispensable mechanism for providing its subjects with, at a minimum, the following: collective self-defense, law enforcement, economic coordination, and basic social services (including education, welfare, health, etc.). Since only the state is capable of providing such goods as protection from harm, rule enforcement, and provision of key public goods, anyone who has signed on to the project of living the sorts of lives that require those goods may be seen from a normative point of view as authorizing the existence of the state that alone makes that project feasible. The fact that citizens typically feel entitled, as a matter of right, to enjoy such goods, provides at least some evidence of the reality that the state acts as their agent to procure these goods for them.Footnote 28

Klosko does not extend his argument to incorporate the idea of authorization, but I think this extension can reasonably be justified as compatible with his basic argument.Footnote 29 In doing so, it is important to recognize explicitly that the question of state authorization is analytically distinct from the question of consent (though it is of course related to it). The problem of consent is concerned with the origin of the moral connection that binds individuals to their state, while the question of authorization is concerned instead with the consequences of that connection. It is about what citizens authorize their state to do – do they authorize it to conduct wars, for example? – not how they do the authorizing. Authorization of the state in this sense is compatible with a variety of different answers to the problem of consent, including models such as Klosko’s that purport to establish state legitimacy without reference to ‘consent’ in the ordinary sense at all.Footnote 30 My argument here is that authorizing the state to act as our agent (whether via consent or by means of a more subtle mechanism like Klosko’s argument from need) tends to establish a principal/agent relationship between citizens and their states, one that commits us to prima facie obligations to take responsibility for the consequences of its actions.Footnote 31

In addition to authorizing the state in general, it also seems logical to extend the argument to assert that citizens are responsible not only for the existence of their state generally but also, to an extent, for the consequences of the form of their state (to the extent that its having that form is a necessary condition of it serving the purposes its citizens pursue). For example, if the state’s effective functioning seems to require a unitary executive with military authority, certain morally pregnant consequences will follow from that fact (for example, that the state is capable of getting its citizens into a war which may be controversial among them). If so, I will eventually be arguing, it is certainly plausible that certain moral responsibilities may attach themselves to those consequences.

The state

What precisely do we mean by ‘the state’ in this context? I want to offer a two-part description, the first part based on the well-known functional definition of Max Weber, the second on a recent etymologically based insight unearthed by Quentin Skinner. Weber famously defined the state as that agency which exercised a monopoly on the legitimate use of violence or coercion within a given territory (Weber, Reference Weber1946: 77–78).Footnote 32 This part of the description helps to explain why the state might be an indispensable mode of social connection in terms of distributing moral responsibility for collective action. The state is a peculiar form of social connection through which we exercise our collective capacity for the always morally hazardous activity of violence and coercion; consequently, at least in those instances where a violent or coercive collective action is an issue, we will want an account specifically of the set of moral understandings that forms the background condition enabling this sort of action to proceed. On the authorized state model, this implies that we (that is, the citizens) authorize the state to exercise force on our behalf, in pursuit of our shared purposes; that this will include, at the extreme, even such acts as conducting wars; and that this sort of authorization is granted to the state as a general power, rather than licensed to it on a specific, case-by-case basis.

It is also important to clarify an ambiguity in Weber’s definition, namely, what precisely is meant by the restriction of the state’s claimed monopoly within a given territory. This could mean that the state claims a monopoly on violence and coercion exercised over the people within the territory; but this restriction would be odd, because it is clearly also central to the state’s functioning that it be able to exercise violence against others outside the territory when necessary for the security of those within. So we might reformulate Weber’s definition to assert that a state holds a monopoly on the legitimate use of force (a) over the people within a particular territory and also (b) on behalf of the people within a particular territory.

If this Weberian definition fixes the function of the state, however, it is important also to distinguish its location (figuratively speaking) as well. Skinner’s account identifies the emergence in the seventeenth century (and specifically in Hobbes’s political theory) of the idea of a state that is intermediate between, and distinct from, either the rulers on the one hand or the ruled on the other.Footnote 33 The state is not simply the people – it is an institution independent of the people, established by them (in some sense or other) to do things on their behalf.Footnote 34 But if the state is not simply the people, it is equally not simply the government – that is, the persons in government at any particular moment in time (what the British call a government and what Americans call an administration).Footnote 35 This is an important distinction for our purposes, because it points out that the claim of the authorized state model is not that the government at any particular moment in time represents all its citizens’ wishes – clearly that will not be true in any majoritarian politics in a pluralistic society – but rather that the state over time represents its citizens’ interests and purposes.Footnote 36 Consequently, there are a variety of actions we can meaningfully attribute to a state – for example, undertaking a war – where we can also point to the government as the cause of the state’s actions without diminishing the force of the claim.Footnote 37 It may be true that the Bush Administration caused the United States to go to war in Iraq, and that without that administration’s choices we would not have gone to war. But it is no more meaningful to say as a consequence that the United States did not go to war with Iraq than it is to say that the murderer of King Duncan was not Macbeth but rather Macbeth’s id. If true, it is trivial. It is not just that the Bush Administration works for the United States; it is that the Bush Administration is the United States at that moment, so far as work is concerned.

Before proceeding, let me dispense with two possible objections to the authorized state model of collective responsibility. First of all, some political observers tend to assume that it is meaningless to talk about states as moral actors at all. For example, as Toni Erskine points out, in the study of international relations, states are universally assumed to be the standard-issue ‘agents’ whose behavior must be explained; at the same time, however, their moral agency is just as standardly denied.Footnote 38 But as Erskine and Robert Goodin both argue, there are plentiful reasons to challenge this view. The state possesses most of the qualities we normally associate with moral agency, such as a capacity for deliberation and unitary action and a more-or-less stable identity over time. Goodin indeed argues persuasively that ‘the state is a moral agent, in all the respects that morally matter’.Footnote 39 I will not argue the point further here, but simply take the coherence of describing the state as in some sense a moral agent for granted as an assumption of the authorized state model.

Secondly, let me acknowledge that the relation between authorization and consent is not at all straightforward and obviously problematic for the authorized state model, and I will engage with some of these problems in more detail below. For reasons of scope, however, I will not in this essay try to argue against those who want to claim that there is no sense in which we as citizens authorize the state, and will concede that if that claim is true there are of course no circumstances in which the authorized state model, as I have described it, applies. Instead, I will adopt the hypothesis that the notion of citizen authorization of state action is at least meaningful, and instead try to determine more precisely in what circumstances it would be applicable, and to what degree.

Hobbes on representation and authorization

Despite its emphasis on the normative dimensions of the relationship between states and citizens, the social contract tradition as a whole gives scant attention to the problem of citizens’ moral responsibility for state action. Among the canonical texts of the tradition, the only extended treatment of the subject is to be found in one chapter of Hobbes’ Leviathan.Footnote 40 Despite this overall inattention to the problem, however, Hobbes is certainly not the only social contract theorist who adopts and operates from a citizen-authorized state model of the polity: that is essentially true of the account of the state found in all the classic social contract theorists.Footnote 41 Instead, what is unique about Hobbes’ account of representation is its willingness to address directly the issue which all the other theories seem anxious to ignore: namely, the extent to which citizens, in authorizing the state, also acquire ownership of the state’s actions as their agent.

Hobbes’ account of state representation explicitly claims not only that citizens authorize the state to act on their behalf but also that this makes them responsible for what the state does.Footnote 42 By ‘the state’ – Hobbes actually uses the term ‘Commonwealth’ most frequently, but tells us in the Preface to Leviathan that this is synonymous with the idea of the ‘State’ – Hobbes means an enduring institution of common life intermediate between the people on the one hand and the current administrators of the government on the other.Footnote 43 The people so represented are to be understood, Hobbes tells us, as ‘many Authors, of every thing their Representative saith, or doth in their name; Every man giving their common Representer, Authority from himself in particular; and owning all the actions the Representer doth’, provided the representative does not exceed the legitimate extent of its commission.Footnote 44

How does the state act as a representative? What does it represent? According to Hobbes’ argument, states rightly act for citizens – in their place and on their behalf – when they exercise on a representative basis the rights those citizens themselves possess. For Hobbes, the social contract is at its root best understood as a transfer of rights from those who naturally possess them to an artificial sovereign/state agency that exercises those rights on their behalf. The subjects therefore come to ‘own’ the actions of their representative because they are the owners of the rights by which the actions of the representative are performed.

So on Hobbes’ account, whatever responsibility for the sovereign’s actions the subjects possess results in some way from the fact that the sovereign acts on the basis of their rights, or, as Hobbes also defines them, their ‘blameless liberties’.Footnote 46 So states can only rightly take actions which its citizens collectively possess the right to take. For Hobbes, this is the ultimate rationale for responsibility mediated through state representation. Since citizens create the state to exercise on their behalf a set of rights that the citizens own, they necessarily come to own also the actions that the state takes on their behalf (that is, on the basis of their purported rights).

At first glance, there might appear to be an unwelcome tension here: for an increased emphasis on the importance of authorization for incurring moral responsibility might seemingly tend to weaken the grip of the kinds of involuntary aspects of collective responsibility we have so far been stressing. But this is misleading, because the form of representation Hobbes has in mind to deal with all the most important questions – that of the state – relies on a far more lenient standard of legitimacy in questions of authorization than we are used to thinking about, especially in political contexts. Central to Hobbes’ argument is the claim that either brute force and the immediate threat of death, on the one hand, or the strategic knowledge that such force and death are almost inevitable eventual consequences of one’s situation, on the other, are fully sufficient bases for motivating legitimate authorization, at least in the case of the state.

Up to a point, therefore, Hobbes’ justification for the authorized state model mirrors the rationale for the state offered by Klosko. It is not necessary for the state’s claim to represent us that it can show that we have authorized its every action individually and specifically; indeed, establishing ‘consent’ may not be necessary in any direct sense to ground responsibility for its acts. Rather, Hobbes (like Klosko) claims that the state can properly claim to be authorized by its citizens because those citizens need their state like they need no other form of social connection.Footnote 47 The peace and stability which the state alone can provide is the necessary precondition of all our other activities, and thus its authority to rule over us and act for us ought to take precedence (at least when it chooses to assert it) over almost any rival moral claim. This, Hobbes thinks, creates a powerful presumptive claim on the state’s part that when it acts in the name of its citizens, it does so legitimately: on their behalf, for their benefit, on the basis of their rights, and with their ultimate authorization.

This insight from Hobbes’ account of the social contract, then, is the core intuition that lies at the heart of the authorized state account. Hobbes’ notion that the state is the authorized representative of its citizens remains central to the subsequent development of the social contract tradition, down to our contemporary intuitions about the role and legitimacy of the state. We can follow Hobbes this far, therefore, without feeling obligated as a consequence to adopt the rest of Hobbes’ controversial account of moral life.Footnote 48 Indeed, it is important for the argument to follow that we explicitly reject much of that account, since taken on its own terms, the extravagant moral permissions Hobbes’ system affords to those in the state of nature (which Hobbes says includes the actions of sovereigns with respect to other states) would tend to undermine any strong account of moral responsibility for state action.

Fortunately there are ample normative grounds for rejecting much of the remainder of Hobbes’ theory. As most of the leading social contract theorists who followed Hobbes (e.g., Locke, Rousseau, etc.) noted, even if Hobbes has not exaggerated the perilous position of the individual in the state of nature, the position in which the social man finds himself after the establishment of the social contract is not truly analogous to the dangers of his pre-social existence. If so, then Hobbes is incorrect to advance his much broader claims about the dire necessities of the world and the broad moral permissions those necessities entail; consequently he is also incorrect to argue that the state (like the natural men it comes to represent) rightly acts without any meaningful moral limits.Footnote 49 Nevertheless, Hobbes’ account (like Klosko’s) rightly emphasizes that to the extent that the state is authorized by its citizens, that authorization can be grounded in certain objective features of those citizens’ social and strategic situation, rather than in any specific instances of their consent to particular actions or even their explicit consent to its agency generally. It is this claim that is crucial to the authorized state account.

Authority and the authorized state

In addressing the idea of ‘authority’ in social contract theory, contemporary philosophy has tended to focus almost exclusively on its implications for the problem of political obligation.Footnote 50 The problem thus framed tries to explain to a subject who might question her state’s authority why she should do what the sovereign commands rather than what she herself wishes.Footnote 51 But this focus on obedience is not at all the exclusive problem raised by the concept of authorization as we encounter it for example in Hobbes’ theory. As Hobbes recognized, from a different perspective the concept of authority itself just as significantly implies citizen ownership of state action.Footnote 52

Indeed, once one begins to think about it in this way, responsibility and obligation seem almost to imply one another. For how can a state have the moral right to command or to rule if it does not also possess the right to commit its citizens morally to certain responsibilities? If a state has any right to act externally on its citizens’ behalf, surely it also has a right to act externally in ways that commit its citizens as principals. Furthermore, even if consent is insufficient to ground full-blown obligation, it may still conceivably be sufficient to ground moral responsibility for citizens. Perhaps one cannot infer from my minimal acceptance of a state that I will obey whatever it commands, but one still might be able to infer that I think of it as my present agent – and that I will be willing to stand by what it does in my name. For all its problems, therefore, the argument from tacit consent becomes noticeably less weak when applied collectively to questions of responsibility rather than individually to questions of political obligation. To ground political obligation, we need to show that each of our various actions-in-concert imply our individual consent to obey the state’s every command, which is tenuous. But to ground collective responsibility for state action in something like tacit consent, we need to maintain only that the citizens collectively have acted in concert in such a way as to recognize the state as their agent in pursuit of public purposes – a more plausible proposition (though not without its difficulties).

Still, the issue of consent remains problematic for social contract theory, and I will not attempt to resolve at any great length the various problems it raises here.Footnote 53 Instead I acknowledge the problematic aspects of consent, and merely point out that they are no more (or less) a problem for the authorized state model of collective responsibility than they are for the concept of the state as a whole. If we can generate enough of consent, or something resembling it, to get the notion of an authorized state off the ground at all, why should it be any more difficult to draw the inference that a state that truly was authorized by its citizens would be capable of generating at least an occasional case of moral responsibility through the association? And if it is not able to generate moral responsibility in this fashion, the moral ties between citizens and their states must be very weak indeed.

The authorized state model, therefore, provides an alternative way of thinking about citizen responsibility for state action that makes the assignment of collective responsibility potentially applicable to a broader range of actors, and also less easily evadable by those who might wish to avoid it. We now need to turn our attention to the question of how on the authorized state account we are to understand the kind of responsibility that state action creates for citizens. It is difficult to accept that the authorized state account can provide us with good reasons for connecting attributive moral responsibility for state action to individual citizens. In a strong sense, the most serious forms of moral criticism – attributing praise or blame – can only be justified when individuals have played some role in bringing about the action in question, whether through voluntary action or negligence. Whether we have attributive responsibility for a state policy that has harmed others will therefore turn substantially on what our attitude was toward the policy, and what action we took to promote it (or did not take, or took ineffectively, to defeat it) – in much the same way as the causal and affective identification accounts of collective responsibility suggest. But assignment responsibility is another matter. We can easily acquire a significant degree of assignment responsibility through the actions the state takes as our agent on our behalf. And this assignment responsibility is a special, as opposed to general, instance of responsibility – it is not the same as our general obligations of benevolence and care as human beings, but rather constitutes a specific form of responsibility occasioned by our relationship as principals of the agency which is the state.

The authorized state account commits us to the view that, rather than our causal relationship to or voluntary endorsement of state policies, it is instead our complex but deeply rooted moral relationship with our nation-states that grounds many forms of collective responsibility; in particular responsibility for most state action in the international arena. This commitment makes the authorized state account especially well suited to analyze the operations of modern representative democracies, in which what is at most a very minimal layer of consent seems in practice to count as a sufficient basis for authorizing state action, even while citizens play no discernible causal role in the actual formation of policy decisions. As a matter of causality, there is little that any citizen can effectively do to, for example, stop a war.

This fact has led some commentators – even those operating within the same broad conceptual framework as Hobbes regarding representation – to conclude that the state cannot in fact transmit responsibility for its actions to the citizens it represents, because the burden of that responsibility is such that no individual would be able or willing to bear it. So David Runciman, for example, argues that the state’s moral agency is distinct from that of the individuals who compose it because ‘no individual or group could possibly take onto themselves the entire burden of the state’, in particular ‘the fiscal burden of the state’s debt, or the moral burden of the state’s acts of violence’.Footnote 54 (Anyone who did, he adds, ‘would go bankrupt or mad, or perhaps be bankrupt or mad already’.)Footnote 55 Instead, Runciman argues, in cases of large-scale state actions such as a decision to go to war, we attribute or assign ‘liability’ to the state rather than to the people it represents. ‘No collection of individuals can be liable for the actions of the state, no matter how those liabilities are distributed’.Footnote 56 This is because in the end, ‘it is states that go to war, not peoples, and it is the existence of the state that allows peoples to know when they are at war, when the war is over, and whether they have won’.Footnote 57

Viewed from the perspective of the authorized state theory of collective responsibility, Runciman’s argument correctly identifies the morally significant fact that the state, rather than the nation proper, is the mechanism by which (to continue the example) both the decision to go to war is taken and the waging of war itself is executed. It would only follow from this fact that states rather than peoples held exclusive responsibility for war if states were morally unconnected to the peoples they represent. And of course according to the authorized state theory they are not. Instead, states represent individuals, not only because they assert that they act on their behalf but also because they really do exercise certain generalized powers on those individuals’ behalf.

Non-democratic states

The authorized state account of collective responsibility maintains that citizens can acquire collective responsibility for the actions of their nation-states, not only by means of the causal role they play in choosing and implementing its policies, nor also by means of their affective responses to those policies, but simply by means of their relationship as the (ultimately) authorizing principals of the entity that selects and enacts those policies, namely the state. This kind of view seems to open the door to assigning collective responsibility to the subjects of non-democratic states more easily than perhaps our intuitions would suggest, since it is those subjects’ status of involuntary association and causal irrelevance to the policy process that seems to afford them a persuasive claim to be excused from collective responsibility for their states’ actions. After all, the authorized state account rests in part on the insight that with respect to their causal role in the selection of policies, the citizens of democratic states are not so very different after all from the subjects of non-democratic states. Both sets of individuals are marginalized with respect to the actions taken in their name; it is only a matter of how far to the margins individuals are pushed.

In addressing this point, I should reiterate that the authorized state account of collective responsibility does not claim that the causal role that individuals play in a policy process or the attitudes and preferences they have about those policies are irrelevant to questions about their collective responsibility. On the contrary, it asserts that such roles and attitudes will also prove to be significant factors in assigning collective responsibility for state action. Those who help to implement a policy, and those who approve and support it, are by those actions and attitudes further implicated in responsibility for the consequences that follow; likewise, those who do not know about, disapprove of, or resist a policy can by those actions partly mitigate their degree of responsibility for the ensuing consequences. The point is not that causality and affective response are insignificant parts of assigning responsibility, but rather that they do not constitute the complete scheme by which moral responsibility may be incurred.

For this same reason, we can say of the subjects of non-democratic states that their responsibility for their states’ actions is likely to be significantly mitigated by their lack of voice within their community.Footnote 58 Non-democratic citizens face at least two problems that democratic citizens generally do not: (1) closed or dysfunctional public spheres that effectively deny voice to citizens; and (2) greatly increased risks of retaliation and punishment for opposing government policies. In non-democratic states where these conditions exist (and to some extent also in ‘democratic’ states where they do), citizen responsibility for state action is likely to be reduced accordingly.Footnote 59 Indeed, we may be able to go even further and adopt as a general guiding maxim: the more voice, the more responsibility.Footnote 60

This does not mean, however, that subjects of non-democratic regimes have no moral responsibility whatsoever for the things their state does in their name. For, as we previously noted, there is an extent to which all individuals in the modern world need a state.Footnote 61 With regard to international action, for example, there is a real sense in which all people need armed defense in a dangerous world, and consequently also need a military to represent their security interests. But this fact carries with it certain moral risks, risks that are shared with and to some extent similar to those that apply to democratic citizens. Consequently, to the extent that the state is engaged in doing those things that all states do on behalf of their citizens, which any state would do on behalf of its citizens, then even citizens of non-democratic states remain responsible for the outcomes.Footnote 62 But as a state’s actions go beyond this threshold into other, more voluntary, spheres of policy choice – for example, adopting a pattern of military aggression or of internal oppression – the moral responsibilities of non-democratic subjects diminish accordingly.

Why then does the same excuse not avail itself to citizens of democratic regimes when they disagree with the policy choices of their government? I think the answer (on the authorized state account of collective responsibility) will be that, however they might disapprove of the outcomes, democratic citizens collectively do continue to have a proprietary role and voice in their national government. We might consider an analogy between the relationship an employee has to their company, on the one hand, and the relationship a shareholder has to their company, on the other. Citizens of non-democratic regimes are like employees: they have an association with their company, and it may even do things on their behalf for which they acquire some degree of responsibility. But they are not shareholders, and therefore they do not own the company and its actions in quite the same way. Shareholders, by contrast, can certainly fail to convince their company to make the choices they would prefer, but it remains their company, their representative, that subsequently acts. Of course it is easier for shareholders to divest from ownership in a company than it is for citizens to divest from their nation; but this is merely a measure of the greater complexity of the underlying ownership relationship in question. The kind of responsibility involved remains the same.

Manipulated citizenries

Next, how should we think about democratic citizenries when their public discourse – the process by which they relate to their government and its policy choices – is corrupted by some significant form of deception or manipulation?Footnote 63 We may want to begin addressing this question by asking ourselves why we should consider factors like deception and manipulation at all. Why suppose that a fair and open politics has any bearing on the question of collective responsibility for state action? After all, the authorized state account of collective responsibility maintains that citizens’ causal role in producing policies – in particular, their consent to those policies – is not the primary source of the moral responsibilities for the consequences of their states’ actions. If their responsibility is not grounded in consent, how can the fact that their awareness of policy choices that has been subject to deception and manipulation be relevant to the degree of responsibility they incur?

Once again, it is important to note that the authorized state account of collective responsibility does not claim that citizens’ attitudes toward policies, participation in policy processes, and consent to policies is irrelevant to questions of responsibility, but rather that citizen attitudes toward policy choices will usually not eliminate their responsibilities, only mitigate them. For this reason, a public that is maneuvered by means of deception and manipulation into undertaking an unjustified war, for example, may nevertheless retain a significant degree of responsibility for the consequences that ensue. Even if they were manipulated into acquiescing in state action, it is nevertheless their state that acts in the world, representing their rights as the basis for its action. This is especially the case when, as often occurs in representative democracies, it is a small group of elites who does the deceiving and manipulating in question. Since one important component of assignment responsibility includes possessing the available resources for taking the appropriate action, it will frequently be the case that the most blamable perpetrators of a collectively underwritten wrong are the least capable of rectifying the damage they have done – and then it may fall to other responsible parties to shoulder their fair share of the resulting burden (before it falls to humanity more generally).

Still, in such circumstances, deception and manipulation must serve as an important mitigating factor in assigning collective responsibility. As a general matter, we might even formulate the following principle: all things being equal, the more the manipulation, the less the responsibility.Footnote 64 However, it is important to qualify this last claim with the observation that frequently citizens bear considerable responsibility (in the attributive sense) for creating the context in which deception and manipulation are able to be effective. Daniel Boorstin, in his influential book The Image, eloquently argues that the increasingly manipulable world of modern democratic public discourse is largely a ‘world of our own making’ and an act of ‘national self-hypnosis’.Footnote 65 Through our ‘extravagant expectations’, Boorstin claims, we ‘have used our wealth, our literacy, our technology, and our progress, to create the thicket of unreality which stands between us and the facts of life’; and therefore, he concludes, ‘each of us individually provides the market and the demand for the illusions which flood our experience’.Footnote 66

This view has its own potential pitfalls, of course, not the least of which is the temptation to engage in ‘blaming the victim’, in this case, the manipulated public. Nevertheless, there is an important difference between this and the standard cases of ‘blaming the victim’, because in a national community the public collectively – and only the public – truly sets the conditions that determine how easy it is for its leaders to manipulate it. An inattentive, easily distracted public, which habitually acquires its information through superficial and sensationalist media, collectively bears much of the blame for any deception or manipulation that results from these receptive conditions that they help to facilitate and maintain. Just as a principal rightly may be blamed for not performing due diligence in supervising its agent, so too citizens must show due diligence in observing and discussing the actions of their state, or else assume responsibility for the consequences that result from their negligence.

Punishment and coercion

Finally, we should consider to what extent it is appropriate to enforce our assignment of collective responsibility to citizens of other states by means of collective punishment or coercion. If citizens can acquire responsibility collectively, can they not also be held responsible collectively by other states? The authorized state account might tend to suggest that they can, which is borne out by our approval of some practices, such as reasonable war reparations, that rely on the idea that citizens can be held collectively responsible for the actions their state takes on their behalf.

Nevertheless, it is important to qualify this conclusion heavily, as it is the primary aspect of the idea of collective responsibility that is liable to significant abuse. States themselves can of course be held liable directly through the imposition of reparations, for example, or even criminally responsible, as Anthony Lang persuasively argues; moreover, recent innovations in international law make it possible to assess and impose such responsibilities on states.Footnote 67 But ever since the precedent established by the Nuremberg trials, there has rightly been a significant hesitancy about extending punishment beyond specific identifiable actors in war crimes to the realm of ordinary citizens.Footnote 68 Part of what inspires this hesitancy is a sense that punishment per se is inextricably bound up with questions of attributive responsibility in a way that is not true of questions of assignment responsibility. The dominant rationales for punishment (deterrence, retribution, moral education) all require a specific attribution of responsibility (usually causal) to distinguish justice from mere compulsion.Footnote 69 Where such attributive responsibility cannot be specifically established – and the authorized state account says nothing about citizens’ attributive responsibility for state action – punishment and coercion will not be appropriate, though of course assignment responsibility may well be.Footnote 70

While it may possibly be legitimate to assess some forms of collective responsibility by force, there are good reasons for imposing significant limitations on this impulse. One reason concerns the problem of self-interest. So long as we are dealing with non-coercive means of enforcing responsibility, we are essentially concerned with instances of individual judgment – either our external judgments regarding the conduct and obligations of others, or our internal judgments about our own conduct and obligations. Once coercion and punishment enter the picture, however, we have moved onto a new playing field with a corresponding new set of incentives. To the extent that we can coerce others into fulfilling their putative responsibilities, we have an incentive to add to their responsibilities so as to diminish our own. We may naturally be motivated to construe others’ responsibilities in such a way as to promote our own benefit, and to skew our own internal and external judgments accordingly. For this reason, we will be safest imposing at least a presumption against punishment or coercion as a mode for addressing instances of collective responsibility.

In addition, there may be some categories of punishment or coercive action that it is desirable to rule out altogether, when applied to individuals, as inconsistent with the greatest utility, respect for persons, or some similar normative principle. The prohibition against killing non-combatants in traditional just war theory is an example of this kind of categorical restriction on collective punishment.Footnote 71 This prohibition does not rest on questions of responsibility at all, since violence administered in war is not conceived of as punishment of any form. Rather, in just war theory, violence is permitted solely because soldiers constitute (reasonably) imminent dangers of life and limb to those who oppose them.Footnote 72 On this logic, Osama bin Laden cannot use arguments about collective responsibility to justify attacks on American civilians because, whatever sort of complaint he might conceivably have against them, he cannot legitimately have a violent complaint against them: their status as non-combatants effectively rules this out.Footnote 73

This does not mean that it will never be appropriate to enforce assignment responsibilities, even those acquired through association or social connection rather than causation, by means of punishment or coercion. In such cases, we must also take into account the urgency of the claim of right or need being asserted, on the one hand, and the clarity of the mechanism that underlies the responsibility, on the other. But even in the clearest such cases, we should exercise caution in enforcing collective responsibility through punishment or coercion rather than through external judgment and internal acceptance of the rightful claims of others. In particular, we should as a general rule be much slower to permit coercive enforcement of responsibility when we are unable to make out a clear and compelling case of attributive responsibility, or where responsibility is acquired by means of association or social connection rather than direct causation.