Responsible nations: Miller on national responsibility

Kasper Lippert-Rasmussen
Responsible nations: Miller on national responsibility

Ethics & Global Politics

Responsible nations: Miller on national responsibility

Kasper Lippert-Rasmussen*

Department of Politics, University of Aarhus, Århus C, Denmark


In National Responsibility and Global Justice, David Miller defends the view that a member of a nation can be collectively responsible for an outcome despite the fact that: (i) she did not control it; (ii) she actively opposed those of her nation's policies that produced the outcome; and (iii) actively opposing the relevant policy was costly for her. I argue that Miller's arguments in favor of this strong externalist view about responsibility and control are insufficient. Specifically, I show that Miller's two models of synchronic collective responsibility—the like-minded group model and the cooperative practice model—ground neither synchronic nor diachronic national responsibility, nor apply in the case of nations generally speaking.

Keywords: collective responsibility; David Miller; nations; historical responsibility; national responsibility

*Correspondence to: Kasper Lippert-Rasmussen, Department of Politics, University of Aarhus, Nordre Ringgade 1, DK-8000 Århus C, Denmark, Email:

Published: 19 May 2009

Citation: Ethics & Global Politics. Vol. 2, No. 2, 2009, pp 109–130. DOI: 10.3402/egp.v2i2.1935
©2009 K Lippert-Rasumussen. This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial 3.0 Unported License (, permitting all non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.


In his recent book, National Responsibility and Global Justice, David Miller sets out how many of us have conflicting intuitions with regard to global poverty and misfortune.1 On the one hand, we have a firm intuition that the existing massive global inequalities are manifestly unjust. Trying to account for this intuition, we might appeal to a cosmopolitan theory of egalitarian justice, according to which it is bad if some people are worse off than others through no fault or choice of their own regardless of whether they are co-nationals.2 When addressing inequalities within a nation, many theorists hold inequalities not reflecting differential choice or fault to be unjust precisely because those who are worse off are not worse off through their own fault or choice. While some egalitarian theorists, e.g. Nagel, who consider such inequalities between co-nationals unjust, explicitly resist extrapolation to the case of international justice, it seems tempting to infer that since few people grow up belonging to a particular nation as a result of a choice or fault of their own, holding that intranational inequalities are unjust for the stated reason commits one to hold that so are international inequalities.3

While Miller thinks that existing global inequalities are to a significant degree, but not wholly, unjust though not qua inequalities (see the next sentence), he rejects cosmopolitan egalitarianism. He accepts the former proposition because, in his view, present global inequalities violate a sufficientarian principle of distributive justice and because the present scheme of international co-operation involves a distribution of benefits and burdens that is unfair to poorer nations.4 Miller's main reason for rejecting cosmopolitan egalitarianism is that, in various ways, it fails to accommodate the normative significance of nationality. First, implementing cosmopolitan egalitarian policies will prevent the realization of the values realized by people having a shared national identity, e.g. trust and solidarity among co-nationals and the sense of belonging that members of a nation enjoy.5 Second, to apply any principle of justice we need to rank the goods whose distribution it governs. However, any such ranking presupposes a shared understanding among those individuals among whom these goods should be distributed, according to the relevant principle, and shared understandings of such rankings obtain only among co-nationals (56–62).6 Finally, and most important for the purposes of this article, Miller thinks that global egalitarianism conflicts with the other core intuition pertaining to global poverty and misfortune, to which I referred in the previous paragraph: namely, that global inequality to some extent reflects different collective choices by different nations and that it would be unjust not to hold members of a nation responsible for their nation's bad choices and simply bring about equality through making members of other nations bear these burdens. For instance and with some cautious reservations, Miller points out that Ghana and Malaysia had similar per capita Gross Domestic Products (GDP) when they gained their independence from Britain in 1957 and that Malaysia's per capita GDP now is 10 times greater than Ghana's (241). In Miller's view, this inequality reflects domestic failure in Ghana and, according to Miller, and, provided certain other conditions are met, such cases speak strongly against any egalitarian cosmopolitan principle of justice that favors transferring resources from richer to poorer countries, e.g. from Malaysia to Ghana, to eliminate global inequality: such a principle would unjustly ignore the different choices made by different nations.7 Hence, considerations about national responsibility undermine the case for cosmopolitan egalitarianism, which represents at least one plausible bid at a theory of international justice. Having briefly introduced the dialectical context in which Miller addresses the issue of national responsibility, I shall set this context aside to focus on Miller's theory of national responsibility itself. Ultimately, of course, if Miller's views about national responsibility turn out to be flawed, his case against global egalitarianism is weakened.8

Often we think a nation is responsible for a bad outcome in such a way that it is morally justified to impose costs on each member to compensate those harmed.9 This is so, even if some members played no causal role with regard to their nation's bringing it about. For instance, post-Nazi Germany was responsible for Holocaust such that it was morally justified to make it pay damages even if some of the costs fell upon individual Germans who actively resisted the Nazi regime.10

As David Miller points out this idea about national responsibility ‘goes against’ the view that: ‘[I]t is only what a person does herself that can make her responsible for harmful outcomes’ (120).11 Call the view presupposed by the pertinent idea of national responsibility the externalist view about control and responsibility:

A person can be responsible for an outcome12, even if she did not control it.13

Positive views about national responsibility often involve the strong externalist view:

A person can be responsible for an outcome despite the fact that: (i) she did not control it; (ii) she actively, albeit unsuccessfully, opposed those of her nation's policies that produced the outcome; (iii) actively opposing the relevant policy was costly for her (and further opposition would have involved even greater costs).14

Some might even subscribe to the extreme externalist view:

A person can be responsible for an outcome despite the fact that: (i) she did not control it; (ii) she actively, albeit unsuccessfully, did whatever she could to oppose those of her nation's policies that produced the outcome; (iii) actively opposing the relevant policy was extremely costly for her (and further opposition would have involved even greater costs).

While Miller rejects the extreme view, he accepts the strong externalist view (120–2). The reason he does not endorse the former position seems to be that he thinks that what one can be responsible for depends on what normal people can reasonably be expected to do (121–2) and that the extreme view demands more of people than they can reasonably be expected to do.15

National responsibility is just one species of collective responsibility.16 However, Miller is particularly interested in national responsibility and submits that judgments about this are ‘more basic’ than judgments of state responsibility (111).17 While I shall follow Miller in focusing on national responsibility, I fail to see the strength of his arguments for thinking of national responsibility as a more basic form of collective responsibility than state responsibility.18 For instance, the ‘disadvantages’ that he mentions in limiting responsibility to states—‘we may want to hold nations responsible for actions performed by states that no longer exist’ (see 112; see also 140–1)—only motivate the view that besides state responsibility there is also national responsibility, not that the latter is more basic. Moreover, this ‘disadvantage’ is matched by a comparable (albeit hypothetical) ‘disadvantage’ in the case where a state persists, yet nationhood changes, e.g. due to extreme patterns of emigration.19

Although I would want to allow for cases of collective responsibility, I am more critical of externalist views of responsibility than Miller is.20Specifically, in view of the unfairness of holding people outcome responsible for bad outcomes which they did not bring about and which, at some cost to themselves, they even tried to prevent I deny the strong externalist thesis about control and responsibility. My main contentions in this paper are that Miller has not offered us sufficient reasons to accept his strong externalist view and, accordingly, that he has not shown that the scope of national responsibility is as broad as he thinks it is. For that purpose I need and shall not deny that collectivities, including nations, can be responsible or play a causal role that is irreducible to the sum of the individual causal contributions of all members; that agents who feel responsible for, e.g. experience shame about, outcomes produced by collectivities of which they are members have a certain depth that people who all too readily feel that the relevant outcome has nothing to do with them do not possess; that members of a collectivity may voluntarily agree to divide remedial responsibility amongst themselves in a way that does not reflect each individual member's causal contribution; or particular all things considered judgments about distributable, national remedial responsibility.

The structure of the paper is as follows: Section Two clarifies the nature of the relevant notion of responsibility. Section Three presents Miller's two models of synchronic national responsibility—the like-minded group model and the cooperative practice model—while Section Four addresses Miller's ‘ties of identification’-based case for diachronic, national responsibility. Assuming that nations constitute like-minded groups and cooperative practices, Sections Five and Six canvass a number of objections to Miller's two models of synchronic national responsibility. Section Seven argues that even if we set aside the criticisms in the two preceding sections, Miller's two models will at most ground national responsibility in a quite narrow range of cases, since nations rarely constitute like-minded groups or cooperative practices. Section Eight concludes.


Before presenting Miller's case for national responsibility, I need to clarify two preliminary issues. First, Miller shows helpfully that ‘responsibility’ may mean different things. Specifically, he distinguishes between moral responsibility and outcome responsibility21: ‘In order to be morally responsible for P's condition, A must have acted in a way that displays moral fault: he must have deprived P deliberately or recklessly, or he must have failed to provide for P despite having a pre-existing obligation to do so’ (100). Moral fault is not required for outcome responsibility, but outcome responsibility is a necessary condition of moral responsibility (89).22

When we ask about outcome responsibility ‘(w)e want to know whether a particular agent can be credited or debited with a particular outcome—a gain or loss, either to the agent or to other parties. There is a presumption that where A is outcome responsible for O, the gains and losses that fall upon A should stay where they are, whereas gains and losses falling upon P and Q may have to be shifted … this presumption can be set aside. There may be overriding reasons why the gains and losses should be distributed differently’ (87).

This normative presumption about where gains and losses should be located is what renders outcome responsibility different from mere causal responsibility (87). In the light of Miller's views of collective outcome responsibility, I take it that he does not consider causal responsibility a necessary condition for outcome responsibility. For instance, Miller thinks that if I am a member of a democracy I must bear the costs of a bad decision even if I voted against it and did everything I could to fight it. Presumably, this is a case where, in Miller's view, I might be outcome responsible for something for which I bear no causal responsibility.23

The externalist views mentioned in Section One all concern outcome responsibility and this is the notion of responsibility that is the exclusive focus in this paper. Miller does not want to commit himself to an externalist position regarding national, moral responsibility. On his view, a person cannot be blamed on account of what her nation did irrespective of what she herself did or supported or opposed, although she can be held outcome responsible for it (130n21).

Obviously, since Miller seems to hold that moral responsibility is internal to control whereas outcome responsibility is not, he must say something about the warrant for this asymmetry. It is not immediately obvious what he would want to say here. Many of the reasons why one would want to be an internalist about moral responsibility and control, e.g. that it is unfair to blame people for something they did not control,24 would appear to apply just as well to outcome responsibility and some of the reasons that Miller would offer in support of his externalist position about outcome responsibility would appear, if applicable there, to be applicable in the case of moral responsibility as well—e.g. if I can be outcome responsible for something out of the pure bad brute luck of belonging to a certain group that brings about a bad outcome (123), why can I not similarly be to blame or at fault for my group's bringing about this outcome? Since I shall offer some direct criticisms of Miller's externalist view about outcome responsibility below, I will set aside the issue of justifying the asymmetric treatment of moral and outcome responsibility.

There is a second preliminary issue that needs to be dealt with to bring Miller's view of national responsibility into sharp focus. Most moral theories imply that under some circumstances we can permissibly impose harm on some morally faultless people in order to save a greater number of people from comparable harm. And we can impose harm on some morally faultless people in order to save the same number of people from much graver harm. Utilitarianism implies this. But so do most other moral theories. However, in none of these cases do the relevant moral theories imply that those people upon whom costs are imposed for the greater good, are outcome responsible for the predicament of others. Or at least, if what we mean by ‘outcome responsibility’ is no more than that, then collective outcome responsibility is acceptable on all but the most extreme theories, e.g. ethical egoism. This raises the question of what then captures the core commitment of views that endorse collective outcome responsibility. I suggest that the best way to construe this core commitment is the following: the fact that a collective agent has brought about a certain outcome makes each individual member of the collective share in outcome responsibility for this outcome (to a significant degree) irrespective of this member's own attitude toward or causal role in respect of this outcome, only if this makes it pro tanto justified to hold this member remedially responsible for the outcome. So, on this view, if I am collectively outcome responsible for a bad outcome there is a pro tanto, but potentially overridden, reason for making me bear the cost of remedying the relevant bad outcome—assuming that it needs to be remedied—that does not apply to other individuals who are not members of the relevant collective (and played no other comparable role in bringing about the outcome).25


Miller's argument in favor of national, synchronic outcome responsibility has a pleasingly clear structure. Basically, he sets out two—in Miller's view: intuitively plausible—ideal-typical models of collective, synchronic outcome responsibility and then argues that nations are sufficiently alike these for nations to be responsible.26 The first model is the like-minded group model. Members of a group, say a mob, are collectively responsible for an outcome because they ‘share aims and outlooks in common, and [they] recognize their like-mindedness, so that when individual members act they do so in the light of the support they are receiving from other members of the group’ (117). This—shared aims and outlooks and the resulting fact about in the light of which members of the group act — is what makes it the case ‘that even those who play no direct role in producing the outcome that concerns us may nevertheless properly be brought within the scope of collective responsibility’ (p. 117).

In fact, Miller endorses the weaker requirement that shared outlooks—not shared aims and outlooks—suffice for collective outcome responsibility27: ‘where a community of people shares a set of cultural values, one of whose effects is to encourage behavior that results in outcome O, then everyone who belongs to the community shares in the responsibility for O, even if they disapprove of it’ (p. 118). It is important but unclear what ‘encourage’ means here. It may mean that the values imply, together with known or believed facts, that the relevant behavior is desirable. Or it may mean that people who hold these values often, or sometimes, are caused to engage in this sort of behavior even if it is in no way endorsed by, perhaps even contradicts, the cultural values. On the latter view, every member of a community will bear responsibility for acts of violence due to this group's cultural values which, say, despite including the value of non-violence also include a strong positive valuation of belonging to the group that sometimes leads its adherents to acts of violence against non-members although this goes against the relevant cultural values and, accordingly, frowned upon by most other members of the community. But even on the former interpretation people need not share aims despite their shared values.28

The second model is the cooperative practice model. Here members of a group can bear outcome responsibility for a certain outcome even if they do not share ‘aims and outlooks’ provided that ‘they are beneficiaries of a common practice in which participants are treated fairly … and so they must be prepared to carry their share of the costs, [e.g.] … the cost that stem from the external impact of the practice’ (119). Because Miller only says that cooperating in a fair cooperative practice ‘may be sufficient to create responsibility’ (119; my italics), he allows that responsibility is not present in all such cases. For some participants in a practice to be responsible for the actions of other participants they need not benefit from the participation of those particular participants—indeed, it may be that the former group of people would be better off if the latter did not participate in the practice. It suffices that they benefit from participating in the practice relative to not participating provided that they are ‘treated fairly’. Being ‘treated fairly’ requires that one has a fair chance to influence the shape of the relevant cooperative practice (119–20). Hence, people who are forced to participate in a certain way in a cooperative practice are not responsible despite benefiting from it.

While Miller thinks that these two models of responsibility are not the only models of collective responsibility, he thinks that they are ‘the models that are most relevant in thinking about national responsibility’ (114).


So far I have focused on Miller's argument for why present members of a nation can be outcome responsible for present outcomes that they opposed and acted to prevent. However, there is a similar question about how future (past) generations of members of a certain nation can be responsible for what past (future) generations of members of their nation did (do) given that present (past) members do not favor these acts and would have acted to prevent them had they had the chance.

Ideally, our answer to how present members of a nation are related to other present members of their nation in such a way that they can bear outcome responsibility for what these other members do also provides an answer to how present members of a nation can bear outcome responsibility for what past members did, i.e. because present members stand in the same relation to past members of the nation as the relation they stand in to other present members—say, the relation of being members of the same spatio-temporally extended nation.29

Neither of Miller's two models of synchronic national responsibility, however, seems to be able to provide such a unifying answer. First, in view of the indisputable fact that past generations do not act ‘in the light of the support they are receiving from’ future generations and in view of how unlike-minded different generations of the same nation can be, it is not clear how the first model of synchronic national responsibility—the like-minded group model—can be extended to the diachronic case.30 One of Miller's paradigm cases of diachronic national responsibility is the outcome responsibility of post-Nazi Germany. However, if that case were to be accounted for along the like-minded group model, the degree of like-mindedness required for collective responsibility would have to be set so unattractively low that almost anyone would bear collective responsibility for almost anything anyone ever did.31 Moreover, it is conceivable that in the future a nation will turn into a nation that is perfectly identically minded with Nazi Germany in terms of those ‘outlooks and aims’ relevant to the Holocaust, e.g. political authoritarianism, anti-Semitism, and virulent nationalism. But we would not for that reason hold this nation responsible for the Holocaust even if it was much more like-minded with Nazi Germany than, say, Germany was in the sixties.32 This imaginary example points to a more general problem concerning the like-minded group model of responsibility, namely that the underlying relation here is one of qualitative identity and that the obtaining of this relation does not secure numerical identity, which is what—initially, at least—most people think matters in relation to diachronic national responsibility.33

Second, the cooperative practice model as expounded above does not extend to the diachronic case, because while the benefit component of the cooperative model may be satisfied, clearly, the present members of a nation had no fair chance of influencing their nation's actions prior to their coming into existence and, arguably, past generations had precious little—not ‘no’—chance of influencing the actions of their nation in the future.34

It might be suggested that the fairness component applied to diachronic national responsibility should be understood in such a way that if each generation has a fair chance to influence their nation's actions during the time they exist that suffices for their having had a fair chance of influencing their nation's actions and, thus, suffices for their being responsible for the actions for past and future generations even though they had no fair chance of influencing these actions. This suggestion is not very attractive. How can the fact that a group of people controls one set of actions be a reason for their being responsible for another set of actions they do not control? In everyday life, retorting to someone who denies responsibility for a certain outcome on the ground that it was not up to her that she is nevertheless responsible for the outcome because there was another outcome which was up to her is unconvincing.

Admittedly, Miller claims that ‘nations can be regarded as cooperative practices in which costs and benefits … are fairly distributed among the members’ and that each member therefore ‘falls within the scope of collective responsibility’ (160), which may suggest that he disagrees with my scepticism about the applicability of the cooperative practice model to the case of diachronic responsibility. However, the brief passage in which he airs the quoted view does not explain how the cooperative model can account for collective responsibility for what a nation did. Rather, he uses the model to argue that presently marginalized minority groups have no national diachronic responsibility, because the benefit component is unsatisfied in their case.35

The sort of relationship that underpins diachronic collective outcome responsibility might of course differ from the two main kinds of relationship that underpin synchronic collective outcome responsibility. In fact, Miller seems to make three different suggestions as to what relations exist between a present-day individual and this person's nation in the past in virtue of which she can bear collective responsibility for what her nation did in the past, all of which differs from the relations that bear on diachronic responsibility. None of them is convincing.

First, in connection with his discussion of the communitarian idea about social identity, Miller seems to suggest that one is responsible for one's nation's actions in the past, because one identifies oneself as a member of a certain nation and that nation is an entity which exists as a numerically identical entity over time (140).36 Apart from some further problems with the notion of identification that I note below, this proposal suffers from the defect that, generally speaking, it is not the case that one is responsible for an outcome simply because it was brought about by some entity with which one identifies. One might be more willing to compensate others for harm caused by an entity that one identifies with, but that does not reflect the view that one has a (presumptive) duty to do so and that those who are harmed by the relevant outcome have a (presumptive) claim on one to compensate them.

Second, at one point Miller might be read as suggesting that the relation of ‘inheritance of benefits between the present generation and its national predecessors’ (158) is that relation which is the bearer of diachronic national responsibility. However, this relation does not seem to be of the right sort. That it obtains may show that, say, a nation has a remedial duty to compensate those to whom these benefits would have accrued in the absence of past injustice, but it is a relation that clearly is not the bearer of diachronic responsibility. A nation can have benefited in the past from injustices of other nations and the fact that it has inherited these benefits does nothing to show that it bears outcome responsibility for these injustices. If it has had ample opportunities to return the relevant benefits to their rightful owners in the past, it may be responsible for failing to do so, but outcome responsibility for this omission is responsibility for a different matter.

Third, especially in connection with his discussion of apologies for historic injustice Miller appears to hold that direct ‘ties of identification’ between present and past members of a nation can ground diachronic collective responsibility (158, 160), where presumably one can identify with people with whom one neither cooperates, nor shares much in terms of aims and outlooks.37 This account of diachronic responsibility raises the question of why ties-of-identification cannot underpin synchronic outcome responsibility, so that in order to uphold Miller's endorsement of diachronic collective outcome responsibility one will have to accept forms of synchronic collective outcome responsibility that corresponds to neither the like-minded group model, nor the cooperative practice model. As noted, Miller explicitly allows other models of synchronic responsibility than these two, so this possibility is one that he has allowed for. However, the ties-of-identification account is problematic in that while, arguably, identification may be a necessary condition of collective responsibility, clearly it is not a sufficient condition.38 This is especially clear in the case of synchronic responsibility where the scope of one's identifications may extend much beyond (or may not cover) the scope of one's responsibilities. People in Western Europe who for political reasons identified with China during the Cultural Revolution may have to account for their sympathies, but given that they were causally irrelevant to what happened during that tumultuous period of Chinese history identification cannot render them responsible for those events.39 The reasons one would offer for why a tie of identification does not suffice for synchronic responsibility would seem to apply to the diachronic case as well.40


In this section and the remaining ones I return to synchronic responsibility. One argument Miller makes in favor of an externalist view about collective outcome responsibility appeals to the difficulties of assigning responsibility for various sub-outcomes to individual members of a group. Suppose a mob vandalizes a neighborhood. For any item that is destroyed there will be particular individuals who destroyed that item and particular individuals who encouraged, inspired or otherwise influenced these individuals to do so. Typically, for each fact there is one or more individuals who know that fact, but there is not one individual, e.g. a judge, who knows all the facts. Hence, since somebody has to bear the costs of the mob's rampage and since it is fairer that members of the mob do so—whatever the exact nature of the causal role they played in the mob's destructive behavior—rather than people who did not participate in the mob, it is tempting to say that all members of the mob share outcome responsibility equally for the losses resulting from the mob rampage. Miller writes: ‘it may be impossible to assign specific shares of responsibility for what has happened to individual members of the mob. We may not know what causal contribution each made to the final outcome, and even if we did, it might still be controversial how responsibility should be divided (if there are recognized community leaders among the group conducting the rampage, should they be assigned a greater share of responsibility simply by virtue of that fact?) … So our starting point must be that the group is collectively responsible, that other things being equal they are remedially responsible for restoring the damage that they have caused, and that every participant bears an equal share of that responsibility. It may then be possible for the participants themselves to make finer-grained allocations, depending on what is known about the activities of each member, whether some can be identified as ringleaders, and so forth’ (116–7).

I interpret these considerations as ones that bear on identifying outcome responsibility. One might question this interpretation on the ground that, in an earlier chapter, Miller explicitly distinguishes between identifying responsibility and assigning responsibility: ‘Identifying responsibility is a matter of looking to see, if anybody, meets the relevant conditions for being responsible … Assigning responsibility, by contrast, involves a decision to attach certain costs and benefits to an agent, whether or not the relevant conditions [pertaining to knowledge about the facts about who is outcome or remedially responsible for the relevant outcome] are fulfilled’ (84).41 Although Miller uses the phrase ‘assign responsibility’ in the passage quoted in the previous paragraph, he wants the case of the mob to support the conclusion that the idea of nations being outcome responsible ‘makes (ethical) sense’ (113). But it can do this only if it involves identifying outcome responsibility and not just considerations about a certain assignment of (remedial) responsibility being correct. No one denies that it may be justified, all things considered, to assign outcome responsibility to nations. Also, Miller writes that the focus of the chapter in which the mob case appears is ‘outcome responsibility’ (134) and in a passage immediately after introducing the mob case, he raises the question of what justifies us in saying that the ‘individual members [of the mob] share in the collective responsibility [for the damage it has caused]?’ Before answering the question, he reminds the reader of what our interest in outcome responsibility derives from (115; see also 115n5). These considerations, I hope, suffice to justify my interpretation of Miller's discussion of the mob case as one that pertains to identifying outcome responsibility.

While I am inclined to share Miller's view as specified in the previous paragraph about how the costs of the damage caused should be distributed given the impossibility of identifying the contribution of each individual member of the mob, I do not think this view supports any ascription of collective responsibility to all members of a nation at a suitably fundamental level. Note, first, that the appeal to what is controversial (see the second sentence in the long quotation from Miller two paragraphs above) is beside the point, because it cannot support a particular view about the matter. Specifically, it is not as if the ‘equal shares of responsibility’-view is a fall-back position that wins in the face of disagreement. Presumably, the relevant controversy exists in part because some people reject the ‘equal shares of responsibility’-view and the mere fact that some disagree with them is a not more of a reason for them to accept the ‘equal shares of responsibility’-view than the fact that they disagree with friends of the ‘equal shares of responsibility’-view is a reason for the latter to reject their view.

Second, imagine a case where, unrealistically, it is really possible to identify what each member of the mob did. If in such a case, we would still say that ‘every participant bears an equal share of that responsibility’, we would ascribe collective outcome responsibility to members of the group. However, if, as I suspect, we would not do so and if we would welcome the availability of the relevant finer-grained information, this suggests that we do not really think that all members of the mob are equally collectively outcome responsible for the damages. Rather, we think that, due to our limited information and for the purpose of having a rule of regulation, we might act as if all members of the mob are equally outcome responsible.42 Such a rule has desirable incentive effects and due to the regrettable unavailability of finer-grained information this is the least unfair way to distribute costs. 43 Some members of the mob may be made to pay more than their share of responsibility warrants, but this injustice is smaller than making people pay who do not have any outcome responsibility at all for what happened. Unlike others, members of the mob are individually morally responsible for participating in the mob conduct. Still, in making each member pay we recognize that ideally each should bear costs corresponding to each person's specific share of the outcome responsibility.

In other words, there are good reasons why our thinking about responsibility in relation to mob behavior does not support any positive view about equal collective outcome responsibility.44 But even if we set these aside, there are features of the case that render it relevantly different from the case of national responsibility which it is supposed to cast light upon. In the mob case, each person who was part of the mob, and knew what was going on, does bear some outcome responsibility for what happens simply by virtue of contributing to what are the predictable results of mob behavior by virtue of having control over his own participation in the mob. But for that reason, the appeal to the case of the rampaging mob cannot really support the view that members of a nation can bear outcome responsibility for a certain outcome brought about by the nation regardless of what causal role they played in bringing about this outcome, which is after all the central issue of contention.45 To elaborate this point, suppose that a handful of die-hard Nazis got lost in the Amazon jungle in 1933, were presumed dead by the outside world, and suppose they did not influence political events in Europe prior to or after 1933. Until May 1945, they remain isolated in the jungle sharing the ‘aims and outlooks’ of Hitler's regime. Yet, we cannot hold them outcome responsible for World War II (WWII) events simply because they shared ‘aims and outlooks’ with non-isolated Nazi members of the German nation.46

In response, Miller might stress the causal component in the ‘like-minded group model’, i.e. that ‘individual members act as they do … in the light of the support they are receiving from other members of the group’ and add that people only share views in the relevant sense—as opposed to hold the same view separately—if people are aware of other people's holding that view and that this somehow is part of the explanation of why all members of the group hold that view (117).47 The exegetical relevance of this line of argument is supported by the fact that Miller writes that ‘the like-minded group model requires’ that its members play ‘a causal role in the genesis of’ the policies for which they are collectively responsible (132). The like-minded group model so construed is immune to the case of the Nazis in the jungle. But by insisting that each member of a group played a causal role in the genesis of that which they are collectively responsible for, it is a rather attenuated sense of collective responsibility that the like-minded group model offers. In particular, it becomes unclear that the model actually extends to the hard cases of national responsibility, where people who opposed certain policies are nevertheless held responsible for those policies because they reflected what their nation did. Finally, if we stress the causal element, then it raises the question whether sharing ‘aims and outlooks’ bears any weight at all. For consider a German WWII citizen who hates the Nazi ‘aims and outlooks’, but pretends otherwise for purely opportunistic reasons.48 Suppose, moreover, that this citizen is so good at faking that she plays the same causal role with regard to the ‘aims and outlooks’ of other German citizens and with regard to the support they receive in terms of acting in the light of their beliefs. This person seems to bear no less outcome responsibility than an otherwise identical German citizen who wholeheartedly endorses the ‘aims and outlooks’ that our first citizen simply fakes. This suggests that the agent's causal role, not her attitudes or outlooks, matters to outcome responsibility. But assuming that the agent's causal role was under her control, this suggestion does not support the externalist views.49


Having raised some questions in the previous section about Miller's first model of collective responsibility, I now turn to his second model. In this model participants in a practice have to benefit from it for collective responsibility to apply. Note initially that it is not clear if we are talking benefits ex ante or benefits ex post. Consider the case in which participants expect their practice to benefit them, but it turns out otherwise. For example, members of an employee-controlled firm think that they will benefit from using a manufacturing process that has very bad environmental effects, but a public scandal breaks and the owners loose all their money. Clearly, members could not plausibly reject outcome responsibility for the pollution on the ground that, as it turned out, the pollution was not economically advantageous to them. Thus, what matters are benefits ex ante.

There is then a further issue of whether ‘benefits ex ante’ means the benefits that participants in the practice expect to derive from participation in the practice or the benefits that an outside observer will predict as likely even though these benefits are not anticipated by the participants themselves. This difference is relevant specifically for the issue of national responsibility because many people may engage in acts that together with lots of other acts reproduce the national culture even if they have given no thoughts to benefits accruing to them as a result of the reproduction thereof. Hence, under the former construal of benefits most members of a nation will not be collectively outcome responsible for those acts that somehow flow from that culture.

If by ‘benefits ex ante’ we mean benefits predicted by an outside observer then it is unclear that such benefits could really be a requirement for responsibility. For compare two cases that are alike in terms of all subjective aspects, i.e. in terms of what the agents intend to do, in terms of their expectations about the results of their practice, etc. In one case a reasonable observer predicts that the practice will turn out to be beneficial all things considered—here the agents themselves have positive, objectively warranted beliefs about what they can expect—whereas this is not what an observer would predict in the other case—here the agents themselves have the same non-culpably optimistic beliefs as the first group of agents have, but in their case these beliefs are objectively unwarranted. It is hard to see how this difference which is after all wholly external to the perspective of the agents could make a difference to what they are responsible for.

Let us now set aside the issue of ex ante versus ex post and consider some further cases that cast doubt on the benefit requirement. Consider a situation in which I am a member of an organized group that is involved in fair cooperation for mutual benefit and that it engages in a certain policy that I oppose and actively resist. Is the fact that I benefit from the practices enough to make me outcome responsible for what that group does? Suppose I am a US national, receive unemployment benefits, and actively oppose USA's war in Iraq. Suppose also that if I was denied the unemployment benefit I would no longer benefit, all things considered, from being a US national. Am I outcome responsible for whatever US actions in Iraq cause? I do not see how the fact that I receive unemployment benefits makes a difference here. Compare my situation with the situation of someone who is identical to me in all respects except that, due to some bureaucratic error, he has unjustly been denied unemployment benefits. On Miller's view this should make a difference such that while I am outcome responsible for what happens in Iraq, he is not. This is implausible. First, the difference between the two persons in question is due to factors that are wholly external to them. Second, to my knowledge when people discuss who shares collective responsibility for what their nation did, they do not appeal to facts about who benefits and who does not benefit from being members of the relevant nation. Nor do people think that the more a group of people benefits from being members of a nation, the greater their share of collective responsibility. But that seems to square badly with the view that enjoying benefits from participating in a fair cooperative scheme is what makes one responsible for outcomes brought about through that scheme even if one has no control over such outcomes and provided that being member of a nation is seen as being a member of such a scheme. Of course, to return to my example in the unlikely case that the money paid out in unemployment benefit derived from plundering Iraq, the former person might share responsibility for returning the money to their rightful owners, but that still would not amount to this person's having a collective outcome responsibility for US actions in Iraq.

Turn next to a case where I benefit from a practice (which is fair as far as the distribution between participants goes) and suppose that, unlike other participants, I am uninformed about the true nature of the practice. I think the practice is just an ordinary business, while in fact it is a drug smuggling ring and the benefits I receive result from drug smuggling. Now, suppose that damages from this practice have to be covered. It seems plausible to say that since I have benefited from the unjust practice and my benefits derive from this activity, it is justified to make me pay for some of the harms involved in the practice. But is this enough to say that I am outcome responsible for those harms? If we say so, are we then saying that there is no difference between me and those participants who were aware of the true nature of the practice and the resulting harm in terms of outcome responsibility? It seems this case is more like the case where, say, a flood takes some property from someone, moves it to my place, and where I then ought to return the property to the original owner, but not because I am responsible for the flood-induced move of the property.50

Some may object that people cannot ‘become responsible for outcomes as a result of chance factors over which they have no control’ (123). However, Miller thinks that there are cases where it is intuitively clear that this is possible, e.g. a case ‘where three pedestrians happen to be on the scene of an accident’ (123) and all three of them must act in concert to save the victim. Here, Miller rightly submits, ‘we would not be impressed if one of the pedestrians asked “why me?” … as an attempt to dodge responsibility for the situation that now confronted him. In a similar way, I may see it as regrettable bad luck that I belong to a political community many of whose members are willing to support policies with terrible outcomes, making it incumbent on me to get my hands dirty and help create a majority for some less objectionable (but still objectionable) alternative’ (122). The problem with these considerations, however, is that they do not address the objection initially presented, i.e. that Miller's view implies that people may be ‘responsible for outcomes simply by virtue of their membership of certain groups, or their participation in certain practices, even though they may not have chosen to be in that position’ (122). For the case of the pedestrians to address this objection, each pedestrian would have to be collectively responsible for the plight of the victim simply on account of belonging to the same group, e.g. co-nationals present at the scene or the group of people able to save the victim, whether or not they tried to save the victim.51 If one of the three pedestrians is responsible for the harm suffered by the unfortunate person, then it would have to be the case that the responsible pedestrian must somehow have failed to do his part to save the unfortunate person. But that, of course, was something he had a choice about. I conclude that the benefit requirement in Miller's formulation of collective outcome responsibility should be omitted. But since it would be implausible to hold that mere participation in a practice renders one responsible for what other members do, this means giving up on Miller's second model of collective responsibility.

At this point it might be objected that the cooperative practice model for collective outcome responsibility includes a fairness condition, which I have ignored in drawing the negative conclusion in the previous sentence. More precisely those to whom the practice is unfair do not on account of this model bear any collective outcome responsibility, whereas those who benefit from the unfairness may well bear collective outcome responsibility. Note, however, that if anything, this condition seems to restrict, not expand, the scope of national responsibility. Second, the condition does not seem right. Suppose the leader of a gang decides in a way that is procedurally unfair to rob a bank and to divide the proceeds in a way that is substantively unfair. Now, if this group is collectively outcome responsible for something—say, using violence against customers and employees during the raid—the members of the group that were unfairly treated with regard to the division of the proceeds cannot say that they are less outcome responsible than other members of the gang because they were treated unfairly by them with regard to the division of the proceeds.52 Some might respond that since gang members are engaged in an uncontroversially unjust practice, unlike members of a nation, the issue of fairness among participants in the practice does not arise and, accordingly, the example does not speak to the issue of national responsibility at all. I am not persuaded that there cannot be an unfair distribution of benefits and burdens between participants in an uncontroversially unjust practice, e.g. if the gang leader is a sexist and favors male members of the gang, we would still say that this is unfair to female members of the gang even though the gang is engaged in an unjust practice. But suppose I am wrong about that. In that case, I could simply revise the example so that it does not involve an uncontroversially unjust practice. Members of a group of co-operating salespersons are not engaged in such a practice. Yet, if they engage in wrongfully manipulative sales techniques, I take it that individual salespersons cannot deny responsibility on the ground that she was being unfairly treated in the distribution of the proceeds from the group's wrongful sales techniques.


Even if we grant Miller his two models of collective outcome responsibility, it is a further question whether there are many outcomes for which nations are responsible. For that to be the case, nations will have to correspond either to groups of people sharing ‘aims and outlooks’ or to fair common practices for mutual benefit. Consider the extent to which nations can be said to constitute groups of like-minded people. It seems obvious that people within a nation can vary tremendously in ‘aims and outlooks’. In fact, sharing aims and outlooks seems inessential to sharing nationality if, say, Londoners 1789 belong to the same nation as Londoners 2008. However, Miller thinks otherwise and suggests that belonging to the same nation involves among other things a set of understandings about how the nation's collective life should be led, recognition of special obligations to other members of the nation, valuation of the continued existence of the nation, and an aspiration for the self-determination of the nation (124–6). Even if we grant Miller these claims, cases in which the ‘aims and outlooks’ that constitute the nation are what leads to its bringing about an outcome for which we might want to hold it outcome responsible are rather unusual.53 For instance, the aims and outlooks that underpinned the imperialism of Victoria's England do not seem to be the aims and outlooks that according to Miller's list were a defining feature of English nationality.

Consider next the extent to which nations can be seen as cooperative practices. It strikes me as doubtful that national identity is best understood as a practice where the members only reproduce the identity and benefit from it. National identity is (re)produced and shaped by non-nationals, e.g. members of a neighbor nation who have certain (often stereotypical) ideas of what their neighboring nations are like, which influence members of the neighboring nation's self-understanding, and may even benefit members of neighboring nations who derive various gains from knowing that, at least, they are unlike their neighbors in certain respects.54 This suggests that the practice of reproducing nations involves cross-national agency. Even in those cases where national identity is reproduced by members only, I do not see that Miller has offered us any reason to think that this benefits all members—relative to other alternative practices or non-participation—or that those aspects from which mutual benefits derive are often tied to the production of those outcomes for which we would want to hold nations outcome responsible.55 Note that in saying this, I am not denying that, generally speaking, people derive benefits from being members of a nation—to deny this would be to display a much more fundamental disagreement with Miller than the one I bring to the fore here—I am simply saying that Miller has not done enough to clarify the exact nature of the benefit requirements that must be satisfied for national responsibility along the lines of the cooperative model to apply and that he has not done enough to show that these requirements, generally speaking, are satisfied to such an extent that they support Miller's views about concrete cases of national responsibility.


Miller's models of synchronic, collective responsibility seem neither persuasive, nor clearly transferable to national responsibility. Also, they are different from the ties-of-identification based account of diachronic responsibility, which in turn suffers from the problem that, generally speaking, identification is insufficient for responsibility. Miller stresses that we must respond to human suffering bearing in mind that victims are not just victims but also agents (e.g. 263). In my view, his position on collective responsibility errs in the direction of treating people as agents in relation to certain events, when their role often is that of passive observers (or even active opponents being dragged along).


A previous version of this paper was presented at CSMN, University of Oslo, 7 November 2008. I thank Lene Bomann-Larsen, Thom Brooks, Morten Brænder, Jakob Elster, Søren Flinch Midtgaard, Christel Fricke, Nils Holtug, Sune Lægaard, Raino Malnes, David Miller, Rasmus Sommer Hansen, and two anonymous referees of this journal for helpful criticisms of a previous version of this paper.


1. David Miller, National Responsibility and Global Justice (Oxford: Oxford University Press, 2007). All otherwise unattributed parenthetical references are to this work.

2. Charles R. Beitz, Political Theory and International Relations (Princeton, NJ: Princeton University Press, 1979), 151; Thomas Pogge, Realizing Rawls (Ithaca, NY: Cornell, 1989), 247.

3. Thomas Nagel, ‘The Problem of Global Justice,’ Philosophy and Public Affairs 33 (2005): 113–47. See also Michael Blake, ‘Distributive Justice, State Coercion, and Autonomy,’ Philosophy and Public Affairs 30 (2001): 257–96.

4. Cp. David Miller, ‘National Responsibility and Global Justice,’ Critical Review of Social and Political Philosophy 11 (2008): 383–99, 395. On David Miller's view, some inequalities neither violate the sufficiency principle, nor reflect terms of cooperation unfairly favoring better off countries (see the example of Ghana and Malaysia below).

5. For an account these values tied to nationality, see David Miller, On Nationality (Oxford: Oxford University Press, 1995).

6. Note that it is unclear why a comparable problem does not undermine principles of justice that apply only to distributions within, and not between, nations—after all, there might be considerable disagreement between co-nationals about how to rank goods, sometimes more so than across two like-minded, small nations. Moreover, it is unclear why the metric problem does not undermine the sufficiency principle of global justice. My focus in this paper, however, is not on the metric problem, so having flagged these two issues I shall set them aside.

7. David Miller, ‘National Responsibility and Global Justice,’ Critical Review of Social and Political Philosophy 11 (2008): 387.

8. A further problem, which I shall not address here, is whether any way of incorporating a concern for collective responsibility is necessarily incompatible with global luck-egalitarianism rather than very much in its spirit. See Kasper Lippert-Rasmussen, ‘Luck-Egalitarianism and Group Responsibility,’ in Egalitarianism and Responsibility, eds. Zofia Stemplowska and Frank Knight (Oxford: Oxford University Press, forthcoming).

9. We rarely say that nations are responsible for good outcomes such that each member ought to benefit whatever their causal role. The rationale for this asymmetric stance on national responsibility for bad and good outcomes is elusive. In this article my focus lies squarely on responsibility for bad outcomes.

10. Ultimately, this view is consistent with scepticism about distributable, national collective responsibility, since it merely says that, all things considered, it is morally justified to impose costs on people who opposed their nation's policies, e.g. because there is a strong case for compensation and because it would be extremely costly to distinguish them from their responsible co-nationals, even if it is unjust to make them bear costs because of policies they opposed.

11. See also David Miller, ‘National Responsibility and Global Justice,’ Critical Review of Social and Political Philosophy 11 (2008): 81–2, 136.

12. People sometimes protest against the foreign policies of their nation under the slogan: ‘Not in our name’. Such protests may reflect an externalist view about national responsibility, but other interpretations are possible, e.g. that citizens have an obligation to resist objectionable policies being carried out in their name albeit they are not outcome responsible for the results of these policies.

13. In the post-Frankfurtian literature on responsibility theorists distinguish between dual outcome and causal path control, see John Martin Fischer, The Metaphysics of Free Will: An Essay on Control (Oxford: Blackwell, 1994), 131–4. My argument is neutral with regard to these two conceptions of control.

14. I take as read here and in the case of the extreme externalist view below that the agent does not assist in realizing the relevant policies at the same time as in other ways she acts to prevent them from being realized, or if she does so, it is very difficult to avoid assisting in realizing the relevant policies—for example it is hard to avoid paying VAT, and by paying VAT we assist whatever policies the state pursues through the use of taxes collected—and her overall causal contribution is negative with respect to the relevant policies. Interestingly because of Thomas Pogge's and David Miller's significant disagreements about global justice, it appears that Thomas Pogge's view on how citizens of wealthy states are responsible for their government's actions bearing on the situation of the global poor simply by virtue of working in the economy or paying taxes commits him to share David Miller's strong externalist view about responsibility, see Thomas Pogge, World Poverty and Human Rights (Cambridge: Cambridge University Press, 2002), 66–7, 139–45.

15. I fail to see why the imputations of responsibility to extraordinary persons should not be based on what extraordinary persons can reasonably be expected to achieve. Note also that how bad the collectively produced outcome is likely to be surely makes a difference as to what one can reasonably expect people, normal or not, to do to avoid it. Thus, the suggested explanation of why one should not endorse the extreme view implies that in cases where the collectively produced bad outcomes are not very bad, people can avoid responsibility for what their nation did even though they did far from everything they could to prevent it. For given that not so much was at stake they could not reasonably be expected to do very much. Note that this is consistent with the strong externalist view because it only states that there are possible cases where having opposed a policy is not a sufficient condition of absence of responsibility and that is consistent with there being other cases where opposing a policy is not even a necessary condition of absence of responsibility.

16. A denial of the externalist view is compatible with endorsement of collective responsibility, since ascriptions of collective responsibility need not translate into ascriptions of individual responsibility to each member of the collective.

17. Cp. Thomas Scanlon, Moral Dimensions. Permissibility, Meaning, Blame (Cambridge, MA: Harvard University Press, 2008), 165; J. Angelo Corlett, ‘Collective Moral Responsibility,’ Journal of Social Philosophy 32 (2001): 573–84.

18. Cp. Sune Lægaard, ‘David Miller on Immigration Policy and Nationality,’ Journal of Applied Philosophy 24 (2007): 283–98; Roland Pierik, ‘Collective Responsibility and National Responsibility,’ Critical Review of International Social and Political Philosophy 11, no. 4 (2008): 465–83, 480–1.

19. Finally, in some cases it may not be a problem at all if the bearer of diachronic responsibility is not identity over time but, say, a relevant kind of continuity between the predecessor and the successor state. For instance, it might be held that there are sufficient continuities between the Soviet state and the post-Soviet Russian state for the latter to bear some responsibility for what the former did even if these are non-identical states.

20. See my ‘Egalitarianism and collective responsibility,’ in Egalitarianism and Responsibility, eds. Frank Knight and Zofia Stemplowska (Oxford University Press, forthcoming). The cases I have in mind are, roughly, ones in which people can do together what none of them can do individually and they fail because no one is willing to make an effort.

21. David Miller distinguishes these two forms of responsibility from a third: remedial responsibility. Roughly, one is remedially responsible for a certain bad state of affairs if, morally speaking, one has ‘a duty or obligation to put [it] right’ (84). One can be remedially responsible for something that others are morally (and, thus, outcome) responsible for, e.g. because one is extremely well off and can at little cost to oneself do something about it whereas others are extremely badly off and can only do something about it at great cost to themselves.

22. Some might doubt that we have any clear concept of outcome responsibility as opposed to causal, moral, or remedial responsibility and suggest that we can say all we want to say with regard to responsibility employing the last three concepts only. Such doubts are congenial to my rejection of the strong externalist thesis, but irrelevant to my argument against it which proceeds on the basis of the integrity of the concept of outcome responsibility.

23. It is not clear to me that David Miller does not endorse the extreme externalist view in the case of democratically organized collective entities. Note also that if I am collectively responsible for an outcome brought about through what my fellow citizens—all of whom may or may not be my co-nationals—democratically decide, this cannot ground national responsibility as opposed to the collective responsibility of the relevant demos. See Lægaard, ‘David Miller on Immigration Policy and Nationality,’ Journal of Applied Philosophy 24 (2007): 290.

24. Cp. Steven Sverdlik, ‘Collective Responsibility,’ Philosophical Studies 51 (1987): 61–76, 68.

25. David Miller needs the ‘assuming that it needs to be remedied’-clause to address cases where a group of agents are outcome responsible for a certain bad outcome and yet, on David Miller's view, there is not even a pro tanto reason for them to remedy the relevant bad outcome. E.g. sellers that on a free market drive other sellers out of business by supplying the relevant good at a lower price clearly are outcome responsible (together with those customers whose decisions to buy are sensitive to prices) for the bad outcome of driving the latter sellers out of business. Yet, there is no pro tanto reason for them to remedy this bad outcome, because this is not as such a remediable outcome in the first place.

26. Presumably, David Miller would also make the more specific claim that nations are alike the two models in respect of those outcomes for which, pre-theoretically, we think nations are responsible and unalike in respect of those outcomes for we think nations are not responsible.

27. Cp. Nicholas Rescher, ‘Collective Responsibility,’ Journal of Social Philosophy, 29 (1998): 46–58, 49.

28. A further question is whether shared values require identical value sets or whether it suffices that one group has all the values another group has leaving it open that the former group may have other and additional values.

29. There might be other such relations for which nationality is irrelevant, e.g. being members of the same state that authorized the relevant policies, upon which collective responsibility might supervene. However, we can disregard such relations in the present context, which concerns national responsibility.

30. The present nation may be connected with its past through a continuous series of time slices of national cultures such that each slice is very much alike the one preceding it and the one succeeding it, but the initial member of the series has no similarities with the last member of the series. Nowhere does David Miller suggest that it is such a relation of continuity that grounds responsibility.

31. I see no principled reason why the degree of like-mindedness required for diachronic responsibility should be different from the degree required for synchronic responsibility.

32. David Miller contends that those aspects of like-mindedness that ground collective responsibility must be aspects that are causally relevant to the bad outcome.

33. I say ‘initially’ because cases of fission or fusion of nations suggest that identity is not what matters.

34. Recall that the benefit requirement is satisfied if participants in a practice benefit relative to non-participation. They need not benefit relative to participation in the practice albeit only with a subsection of the existing participants. Hence, if nations are seen as practices existing over time, one generation of a nation may benefit in the relevant sense from participation, even if it does not benefit from the participation of future, non-overlapping generations of co-nationals.

35. It is true that David Miller does make a positive point about diachronic responsibility of minorities for their nation's past misconduct—i.e. that immigrant groups that have prospered ‘should pay their fair share of the costs of rectification’ (161)—but the motivation for this claim can hardly be that present immigrant groups are in a cooperative practice with past non-immigrant members of the nation.

36. In connection with states, David Miller writes that because states ‘persist over time despite changes in the personnel’ there can be ‘no problem of inheritance’ of responsibilities (140). So here the relation of numerical identity over time seems to be that in virtue of which an entity is responsible for what happened in the past.

37. So, unlike in the first model, identification here does not run through one's identification with a particular nation.

38. For a discussion of necessary, subjective conditions of individual responsibility, see Galen Strawson, Freedom and Belief (Oxford: Clarendon Press, 1986), 173–306. David Miller himself uses the tie of identification condition as a necessary condition for responsibility (to apologize), see 158, n27, as does Alasdair MacIntyre in his discussion of how we identify ourselves in social terms, which David Miller appeals to, see After Virtue (London: Duckworth, 1981), 204–5.

39. Cp. Bernard Williams, Shame and Necessity (Berkeley, CA: University of California Press, 1993), 56–7, on the primacy of causation in relation to responsibility.

40. A further complication is that whom one identifies with depends a lot upon the setting – see Samuel Scheffler, ‘Immigration and the Significance of Culture,’ Philosophy & Public Affairs 35 (2007): 93–125, 100–1; reprinted in Sune Lægaard, Nils Holtug, and Kasper Lippert-Rasmussen, eds., Nationalism and Multiculturalism in A World of Immigration (Houndmills: Palgrave Macmillan, 2009), 119–150 —and yet one's responsibility does not seem to vary with the setting.

41. The distinction between identifying and assigning responsibility appears to cut across the distinction between outcome and remedial responsibility.

42. For the distinction between the specification of ‘just institutions and just practices’ which ‘vary with circumstances’ and must rely on proxies for what really matters and rules of justice that do not, see Richard Arneson, ‘Luck Egalitarianism and Prioritarianism’, Ethics 110 (2000): 339–49, 345.

43. Cp. David Miller on our unease with ‘strict liability laws’ (86) and on the motivation behind our interest in outcome responsibility—‘as far as possible we want people to be able to control what benefits and burdens they receive, but we also want to protect them against the side effects, intended or unintended, of other people's actions’ (89).

44. See also Richard Child, ‘Should We Hold Nations Responsible?’, Res Publica. DOI 10.1007/s11158-008-9058-y (online first).

45. In effect, David Miller agrees with this point when he writes that moving from the mob case to the case of racism in the post-bellum American South we are brought ‘closer to the idea of national responsibility’ (117–8).

46. Cp. the related, but somewhat different criticism, in Roland Pierik, ‘Collective Responsibility and National Responsibility,’ Critical Review of International Social and Political Philosophy 11, no. 4 (2008): 476–8.

47. David Miller does say that in the mob case ‘each made some causal contribution to the final outcome’ (115).

48. One could even imagine a case where everyone opportunistically fakes abominable views.

49. A somewhat different line of argument says that what is crucial is that the agent has the option of avoiding outcome responsibility by exiting her nation. On this view, one might be outcome responsible for an outcome which one did not control provided one controlled whether one was outcome responsible for the upshot—cp. Michael Otsuka, ‘Incompatibilism and the Avoidability of Blame,’ Ethics 108 (1998): 685–701. Note, however, that it is not clear what exiting a nation amounts to—presumably, it involves more, or something other, than a change of citizenship. Also, David Miller's two models of collective responsibility do not seem to imply that collective, national responsibility is conditional upon the existence of an exit option, although he seems to hold that the existence of exit options from the relevant collective is relevant to one's sharing collective responsibility (117–18, n8).

50. David Miller seems to agree, see the example of the company ruled by a small deceiving clique (119).

51. Note that David Miller's example of the pedestrians is intended only to show that people can ‘become responsible for outcomes as a result of chance factors over which they have no control’ (123). Hence, the example need not be one that embodies the features of the cooperative practice model of national responsibility, i.e. one in which the three pedestrians participate, on fair terms, in a cooperative scheme that caused the plight of the person struck by an accident even if that participation was not voluntary. Of course, David Miller would say that in such a case the three pedestrians are collectively responsible for the bad outcome even though it arose as a result of factors that none of the participants had individual control over. However, the dialectic here is that David Miller is responding to people who reject the co-operative practice model of national responsibility on the ground that people cannot “become responsible for outcomes as a result of chance factors over which they have no control”. Hence, it would be question-begging for David Miller simply to appeal to the possibility of responsibility in the controversial case, which, presumably, is why he appeals to the possibility of responsibility for outcomes resulting from chance factors in a much simpler case. I thank an anonymous referee for pointing out the need for clarification at this point.

52. It might be replied that it makes a difference here whether members of the gang had an exit option and that if they had not, unlike if they had, it would make a difference to collective outcome responsibility for violence during the raid whether the proceeds of the robbery were fairly distributed among members of the gang. This view strikes me as implausible. But even if it were not it is not clear that David Miller would be able to appeal to it, since for many people all exit options from one's nation are not closed.

53. Jacob T. Levy, ‘National and Statist Responsibility’, Critical Review of International Social and Political Philosophy 11, no. 4 (2008): 485–99, 489–91. Cp. David Miller, ‘National Responsibility and Global Justice’, Critical Review of International Social and Political Philosophy 11, no. 4 (2008): 383–99, 386–7.

54. Cp. Benedict Anderson, Imagined Communities: Reflections on the Origin and Spread of Nationalism. 2nd ed. (London: Verso Book, 1991). Anderson's cases are almost cases in which the national identity of a certain group of people is largely a result of how members of other nations see their national identity.

55. Of course, David Miller has dealt extensively with the benefits that people derive from being members of nations elsewhere, e.g. David Miller, On Nationality (Oxford: Clarendon Press, 1995), but I think it is true also of this work that it does not show that all members of a nation benefit from their membership relative to the relevant alternative scenario. In fact, the question of which alternative to membership of a particular nation is relevant to answering questions about distributive responsibility for that nation's acts is a further question which I think is under-addressed in David Miller's work. David Miller seems to think that it is a situation without nations. If people are better off being members of their nation than they would be if no nations existed (or, at least, if they were not members of any), they benefit from their nationality and the co-operative practice model applies. But it is not clear why this alternative—as opposed to one in which people are members of a different nation or one in which people are members of the same nation albeit with somewhat different characteristics—is the one relevant for ascriptions of responsibility. Note that to raise this question one need not assume, implausibly, that the mere logical possibility of an individual's having been better off by being member of a different, logically possible nation suffices to absolve this individual for collective responsibility for her nation's actions.

About The Author

Kasper Lippert-Rasmussen
University of Aarhus, Department of Politics

Kasper Lippert-Rasmussen is a professor in political theory at the University of Aarhus, Denmark. He works primarily in ethics, political philosophy and informal logic. His articles on these matters have appeared in journals such as Journal of Political Philosophy, Ethics, Philosophy and Public Affairs, The Journal of Ethics, Philosophical Books, Politics, Philosophy & Economics, Philosophical Studies, and Theoria.