Miller 2004

Holding Nations Responsible
Author(s): by David Miller
Source: Ethics, Vol. 114, No. 2 (January 2004), pp. 240-268
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/10.1086/379353 .
Accessed: 20/06/2014 21:01
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to Ethics.
http://www.jstor.org
This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

All use subject to JSTOR Terms and Conditions
Holding Nations Responsible*
David Miller
Contemporary liberal theories of justice lay heavy emphasis on the idea

of individual responsibility. Where people can be shown to be respon-
sible for the outcomes of their choices and their actions, two things
standardly follow. First, if the outcome in question involves the individual
concerned having a greater or lesser share of resources than she would
otherwise have, justice does not require us to intervene. If, having been
given my fair share of resources in the first place, I decide to squander
them or gamble them away, I must bear the costs of my decision myself.
Second, if the outcome in question involves the wrongful imposition of
costs on some other person, then I owe that person compensation for
the costs in question. If, for instance, I carelessly light a fire that burns
my neighbor’s haystacks, then justice requires me to pay for his loss.
These principles of what we might call ‘self-regarding’ and ‘other-
regarding’ individual responsibility are very familiar. I shall not try here
to investigate either their basis or their limits. I shall not in other words
try to explain exactly how an agent must be connected to the conse-
quences of her action in order for us to hold her responsible for those
consequences (I shall, though, attempt to distinguish the kind of re-
sponsibility involved from other kinds, especially moral responsibility as
that idea is usually understood). Nor shall I try to establish when re-
sponsibility runs out—when the costs to the agent become so great that
we are no longer prepared to let her bear them single-handed. Most of
us believe, for instance, that if I become completely destitute as a result
of behavior for which I am responsible, others can still have obligations
of justice to help me. I shall assume that we know approximately how
* This article has benefited greatly from seminar discussion at the Department of
Government, London School of Economics, at the Department of Politics, University of
Toronto, and at the Nuffield Political Theory Workshop. I thank all the participants in
these discussions for their comments and especially also Samuel Black, David Copp, Cécile
Fabre, Catherine Frost, Andrew Mason, Matt Matravers, Monica Mookherjee, Samuel
Scheffler, Jacob Schiff, and the referees and editors of Ethics for sending me extensive
written comments which have much improved successive drafts of the article.
Ethics 114 (January 2004): 240–268

䉷 2004 by The University of Chicago. All rights reserved. 0014-1704/2004/11402-
0002$10.00
240

Miller Holding Nations Responsible 241
principles of responsibility work in the case of individuals. My interest
is in whether there could be collective, and especially national, ana-
logues to these principles. Does it make sense to treat nations as re-
sponsible agents who can be made to bear the results of their conduct?
It seems that often we do treat nations in this way. For instance,
when considering historical injustices perpetrated by one people against
another, we may say that nation A is responsible for the harm caused
to nation B, and that as a result the present-day members of A have an
obligation to the members of B to repair that damage. We are also
sometimes inclined to argue that nations are responsible for their own
present circumstances—that, for instance, a country that is now relatively
poor because of choices made in the past is responsible for its own
poverty, implying that there is no injustice (or less injustice, anyway) in
its members having fewer resources than people elsewhere. We appear,
in other words, willing to make judgments both about other-regarding
responsibility and about self-regarding responsibility in the case of col-
lectives such as nations as well as individuals. But is this just a case of
sloppy thinking, or is there some defensible sense in which we can hold
nations responsible (and if so, under what conditions)? This is the ques-
tion that I aim to tackle here.1
I
Paradoxically, perhaps, it is those liberals most closely wedded to ideas
of individual responsibility, whether self-regarding or other-regarding,
who are most likely to resist the thought that collective entities such as
nations can be held responsible for the results of their actions. For such
1. Some philosophers prefer to think about issues of international or global justice

wholly in present- or future-oriented terms. They advocate some form of global equality,
for instance, or an extended version of John Rawls’s difference principle, and choose to
disregard the question how the existing distribution of resources and opportunities came
about. Whatever can be said in favor of this approach, it does not square with common
opinion on the question of international distribution. Nor, interestingly, is it the approach
favored by Rawls himself, who rejects the application of egalitarian principles of justice
at the global level on the grounds that “the causes of the wealth of a people and the forms
it takes lie in their political culture and in the religious, philosophical, and moral traditions
that support the basic structure of their political and social institutions, as well as in the
industriousness and cooperative talents of its members” (John Rawls, The Law of Peoples
[Cambridge, Mass.: Harvard University Press, 1999], p. 108). In my view a full account of
global justice must involve both principles of national responsibility and principles setting
out what is owed to individuals and to nations regardless of responsibility. Here I focus
on one part of the picture, but elsewhere I have tried to fill in other parts: see esp. David
Miller, “Justice and Global Inequality,” in Inequality, Globalization, and World Politics, ed.
Andrew Hurrell and Ngaire Woods (Oxford: Oxford University Press, 1999), pp. 187–210,
“National Self-Determination and Global Justice,” in my Citizenship and National Identity
(Cambridge: Polity, 2000), pp. 161–79, and “Distributing Responsibilities,” Journal of Po-
litical Philosophy 9 (2001): 453–71.

242 Ethics January 2004
liberals are likely to insist that responsibility, of the sort that brings
considerations of justice into play, attaches only to individuals. It may
well be appropriate, on this view, to pin responsibility onto figures such
as military or political leaders in the event that they can be shown to
have committed war crimes or inflicted suffering on their people. But
this simply cannot be done in an aggregate way in the case of large
masses of people. In order to assign responsibility, we must be able to
connect a particular outcome to the will and intention of a particular
agent—and only individual people can be agents in the relevant sense.
The liberals I have in mind are likely to be happy with the corollary
that nations should not be treated as responsible agents who as a matter
of justice should both reap the benefits and bear the losses they have
created through their own actions and decisions. In particular, the re-
distributive claims made by poor nations against rich nations cannot be
turned aside by claiming that the poor nations are themselves respon-
sible for their present predicament.2 In consequence, such liberals will
want to endorse principles of global justice, whether egalitarian prin-
ciples or principles of sufficiency—principles specifying a minimum
standard of living that every human should be guaranteed as a matter
of right—that make no reference to collective responsibility. They are
confident that there is no relevant sense in which a poor Bangladeshi
farmer, say, could be held responsible for his condition, as part of the
collective responsibility of the Bangladeshi people. These liberals may,
however, be a bit less happy with a second corollary, which is that nations
cannot be held responsible for the effects of their own past actions, in
cases where these prove to be damaging to other nations or individuals.
It is, after all, now common practice to seek to extract symbolic apologies
and material compensation for historical events, such as the colonial
expropriation of native lands, massacres, and genocides, from present-
day members and representatives of the nations who perpetrated them.3
Yet if the notion of collective responsibility makes no sense, or cannot
be applied to nations, then one cannot hold Britons collectively re-
sponsible for the treatment of American Indians, or Turks for the Ar-
menian genocide, or Germans for the Holocaust. Nor does it make
sense for representatives of these nations to make public statements of
apology or to offer compensation. Of course anyone may properly de-
plore these historical events, and anyone may seek to improve the po-
sition of groups who are now disadvantaged by them. But apology and
2. For two recent examples, see Charles Beitz, “Rawls’s Law of Peoples,” Ethics 110
(2000): 669–96, esp. pp. 691–92, and Allen Buchanan, “Rawls’s Law of Peoples: Rules for
a Vanished Westphalian World,” Ethics 110 (2000): 697–721, esp. p. 715.
3. See the evidence amassed in Elazar Barkan, The Guilt of Nations: Restitution and
Negotiating Historical Injustices (New York: Norton, 2000).

compensation only make sense if they are ways of recognizing respon-
sibility, and, except in unusual cases where particular individuals can
be linked to particular outcomes, as in war crimes trials, this responsi-
bility must be collective.
It seems, then, that the question whether it makes sense to treat
nations as potentially responsible actors has important implications for
the way we think both about global justice in the present and about the
rectification of past injustice. So there are good reasons for trying to
analyze the idea of national responsibility. Unfortunately both of its
components, “national” and “responsibility,” are likely to invite misun-
derstanding and disagreement, so let me try to clarify each of them in
turn. I shall say more later about what I understand a nation to be, but
by way of brief characterization it is a community of people who share
an identity and a public culture, who recognize special obligations to
one another and value their continued association, and who aspire to
be politically self-determining.4 As this characterization makes clear,
nations are not to be confused with states, even though in many cases
they act through states. The connection between nation and state is a
contingent one: a nation may exist without having its own state, and a
state may exercise authority over several different nations. Moreover,
national identity does not depend on state identity, so that we can speak
of the German nation, for instance, persisting through the Weimar Re-
public, the Nazi regime, the Federal Republic and the German Dem-
ocratic Republic, and so forth.
The distinction between nation and state is important, because it
is sometimes argued that states, by virtue of their formal organization,
can be treated as responsible actors, whereas nations cannot.5 It is cer-
tainly more straightforward to assign responsibility to states than to
nations, because in the former case we can point to specific institu-
tions—governments, legislatures, armies, and so forth—as the bearers
of responsibility, and we can also point to particular acts—passing leg-
islation, signing treaties, declaring war, for instance—for whose conse-
quences states can be held responsible. So why do we need to consider
national responsibility at all? One reason is that if we divorce state
responsibility from national responsibility, then it becomes difficult to
show how individual people can share in the responsibility to compen-
sate those whom the state they belong to has harmed, whereas if we
treat states as acting on behalf of nations, such collective responsibility
4. For a more complete account, see David Miller, On Nationality (Oxford: Clarendon,
1995), chap. 2, and “In Defence of Nationality,” Journal of Applied Philosophy 10 (1993):
3–16, reprinted in Miller, Citizenship and National Identity.
5. See, e.g., Toni Erskine, “Assigning Responsibilities to Institutional Moral Agents:
The Case of States and Quasi-States,” Ethics and International Affairs 15 (2001): 67–85.

will be easier to establish. A second reason is that we may want to hold
nations responsible for actions performed by states that no longer exist,
as in the case of the continuing responsibility of the German people
for acts carried out by the Nazi state that was destroyed and replaced
in 1945. A third reason is that, although nations may act through states,
in which case national responsibility and state responsibility may coin-
cide, in other cases this may not happen. Think, for example, of a
stateless nation whose quest for self-determination leads it to carry out
a terror campaign against the people holding it in subjection. Of course
I have yet to demonstrate that assigning collective responsibility in such
circumstances is legitimate—that is the overall aim of the article. Here
I am just indicating why we need to investigate the idea of national
responsibility over and above the responsibility of states.
For present purposes I shall leave aside the special issues that are
raised by national minorities or by the phenomenon of territories in
which national populations are intermingled. These are key questions
in any general theory of nationality, and I have tried to address them
elsewhere,6 but here I want to narrow the inquiry to the case where the
inhabitants of a territory form a single national community. This is a
gross simplification, admittedly, but one that I think can be justified,
given the complexity of the issue that we are addressing.7
As for ‘responsibility’, this is potentially one of the most confusing
terms in the lexicon of moral and political philosophy. It is important,
therefore, to say precisely which sense of responsibility is at issue in
asking whether nations can be held responsible for their own condition
or the damage they inflict outside their borders. The relevant sense is
“outcome responsibility,” to borrow a term from Tony Honoré.8 When
we say that an agent is outcome responsible for the consequences of
6. See esp. David Miller, “Secession and the Principle of Nationality,” in Rethinking
Nationalism, ed. Jocelyne Couture, Kai Nielsen, and Michel Seymour (Calgary: University
of Calgary Press, 1998), pp. 261–82, and in Margaret Moore, ed., National Self-Determination
and Secession (Oxford: Oxford University Press, 1998), pp. 62–78, and “Nationality in
Divided Societies,” in Multinational Democracies, ed. Alain Gagnon and James Tully (Cam-
bridge: Cambridge University Press, 2001); both included in Miller, Citizenship and National
Identity.
7. Note that this is not tantamount to saying that the territory contains only like-
minded people who share a thick set of cultural values. My understanding of nationality
makes room for wide cultural differences among people who nevertheless share a common
public culture—see esp. Miller, On Nationality, chap. 5. Later in the article I shall explicitly
address the issue of how cultural and political differences within nations affect collective
responsibility. Note also that I do not wish to deny that the citizens of multinational states
can have collective responsibilities—it is merely that this more complex case requires
separate treatment.
8. See Tony Honoré, “Responsibility and Luck,” in his Responsibility and Fault (Oxford:
Hart, 1999), pp. 14–40.

her action, we are attributing those consequences to her in such a way
that, other things being equal, the resulting benefits and burdens should
fall to her. If the consequences include harm to others, then outcome
responsibility may, depending on the case, entail liability to compensate
for that harm.9 Outcome responsibility is a narrower notion than bare
causal responsibility because it does not extend to outcomes that arise
in bizarre and unpredictable ways: if I toss an orange pip over the edge
of a cliff and in doing so precipitate a rock fall, I am causally responsible,
but not outcome responsible, for the damage that ensues. Because the
purpose of assigning outcome responsibility is normative—we want the
concept to guide us in thinking about where the good and bad con-
sequences of action should be allowed to fall—and because as agents
we want to determine as far as possible what happens to us, we are not
held responsible for the remote causal consequences of our actions,
over which we have no control. For the same reason, ascriptions of
outcome responsibility also rest on implicit norms concerning the ca-
pacities that human beings can be expected to possess. Thus we do not
hold people responsible for the consequences of their actions in cases
where those consequences could only have been avoided by a super-
human display of strength. Likewise we do not hold people outcome
responsible in cases where they are coerced into acting as they do, and
in making these judgments we rely on our intuitive sense of how much
pressure a normal person could be expected to resist. At the same time,
diminished capacity does not relieve a person of outcome responsibility.
People cannot escape it merely because through ignorance they failed
to anticipate the results of their actions. An unusually clumsy person
can be held responsible for the damage he causes as he blunders about
the world—Honoré refers to such people as “shortcomers.”10 Our in-
terest in outcome responsibility arises from our interest in the fair dis-
tribution of benefits and burdens between different agents: 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.11
9. It does not do so, for instance, in competitive situations where A’s success entails
B’s defeat. A is here outcome responsible for the setback to B’s interests but is justified
in imposing this harm because of the nature of the competition. For discussion, see Tony
Honoré, “The Morality of Tort Law—Questions and Answers,” in his Responsibility and
Fault, pp. 67–93, pp. 77–78.
10. Honoré, “Responsibility and Luck,” p. 15.
11. I shall not discuss how these two concerns are to be balanced against each other
as the limits of outcome responsibility are set. For illuminating critical discussions of
Honoré’s concept, see Stephen Perry, “Honoré on Responsibility for Outcomes,” in Relating
to Responsibility, ed. Peter Cane and John Gardner (Oxford: Hart, 2001), pp. 61–80; Peter
Cane, “Responsibility and Fault,” in ibid., pp. 81–110, and Responsibility in Law and Morality
(Oxford: Hart, 2002), esp. chap. 3.

If outcome responsibility is more stringent than bare causal re-
sponsibility, it is less stringent than moral responsibility as that term is
usually understood. To be morally responsible for the consequences of
an action, you must be outcome responsible, but the converse does not
hold. Attributions of moral responsibility involve a moral assessment of
the agent and thereby open the door to praise and blame, reward and
punishment. In making them we ask whether the agent intended the
consequences that flowed from his action, whether he acted negligently,
whether he acted in breach of obligations that he should have met, and
so forth. If someone could not have avoided bringing those conse-
quences about, no matter how hard he tried, he cannot be held morally
responsible. Negative moral responsibility, the kind that attracts blame,
requires that the agent’s conduct reveals some kind of moral fault,
whereas outcome responsibility does not. Hence the stupid or clumsy
person is not morally responsible for the damage he causes and is not
liable to be blamed or punished for it (he might be morally responsible
for getting into a position where damage is likely to ensue—walking
into the china shop in the first place—but that is a different matter).
Nevertheless he remains outcome responsible and can be required to
compensate those who suffer as a result of his actions, as well as bearing
the direct costs of his deficiencies himself.
It is important to keep this distinction between different senses of
responsibility in mind when we discuss national responsibility. For in
these discussions it is not moral responsibility of the blame-incurring
kind that is centrally at issue. The questions we are asking are whether
the members of nation A can be held responsible for their relatively
low standard of living or whether they can be held responsible for
repairing the damage inflicted on the members of B. In both cases,
what is at stake is the bearing of burdens—the burden of living a rel-
atively poor life or the burden of transferring resources to the members
of another nation, B. It might seem that national apologies are a dif-
ferent case: that when nation A makes a public apology for what its
members have done to B, it is precisely moral blame that is being ac-
knowledged. But here too we need to move carefully. When people
apologize, they often do so as a way of acknowledging their moral re-
sponsibility for what has happened—in other words they are admitting
that they are to blame. But this need not always be so. If someone
pushes against me on the Underground and I topple back and tread
heavily on someone else’s feet, I will certainly apologize for what has
happened, as the agent who has directly imposed the harm, but it
would surely be wrong to interpret this as admitting moral responsi-
bility. Perhaps, then, we should treat national apologies as a way of
acknowledging outcome responsibility for the damage that has been

inflicted on the victims, which sometimes, but not always, will involve
admitting moral responsibility too.
It is also important to distinguish outcome responsibility from what
I shall call ‘remedial responsibility’. The idea of remedial responsibility
comes into play when we encounter a person or a group of people who
are suffering harm of some kind, and we want to identify an agent whose
job it is to put that situation right: we say that the agent in question
has a remedial responsibility to end the suffering. Now one very obvious
way to find the relevant agent is to ask whether any person or set of
persons is outcome responsible for the harm in question. We have a
well-entrenched belief that if A is outcome responsible for harming P,
then A should also be held remedially responsible for putting the harm
right. Nevertheless, this is not always the appropriate direction in which
to look. As I have argued elsewhere, we have several independent prin-
ciples for assigning remedial responsibility, and the right one to appeal
to will depend on the case.12 Sometimes, for instance, the agent with
primary outcome responsibility for the harm will be incapable of rem-
edying it, or would have to bear excessive costs in order to do so. In
such a case we will bring another principle into play.
So when we ask whether nations can be held collectively responsible
for the self-regarding or other-regarding costs of their actions, we are
asking in the first place about outcome responsibility. We are asking
whether they can appropriately be held liable for the costs in question,
whether there is a prima facie reason for making them bear the costs
they have inflicted on themselves or compensate other peoples for the
damage they have inflicted. But this does not settle the issue of remedial
responsibility in cases where the costs are sufficiently great that they call
for remedy. It is important to underline this point. Suppose we find
that there is a legitimate sense in which Bangladeshi farmers can be
held collectively responsible for their present impoverished condition.
It does not follow that they should also be held remedially responsible
for getting rid of that poverty. Other agents may be much better placed
to do so, and we may decide that the responsibility should be theirs.
Nonetheless, outcome responsibility, if we can establish it, is surely a
relevant factor to consider. In other cases—for instance, where peoples
enjoy living standards that are relatively low compared with those of
others but are nonetheless perfectly adequate—it may be a decisive
factor.
To conclude this part of the discussion, I have argued that when
we ask whether the idea of national responsibility makes sense, we are
asking in the first instance whether nations can be held outcome re-
sponsible for the burdens they impose on themselves or others. We are
12. Miller, “Distributing Responsibilities.”

not primarily interested in the more stringent idea of moral responsi-
bility that is required to justify blaming and punishing—though ques-
tions of moral responsibility may sometimes arise when we consider how
nations or their representatives have acted. Nor are we concerned in
the first instance about remedial responsibility for harmful states of
affairs—though establishing outcome responsibility will often be rele-
vant when we are considering the question of how to assign remedial
responsibilities. Outcome responsibility is the focus, and the issue is
whether a form of responsibility that we routinely attribute to individuals
in everyday life, as well as in legal settings, for example, can properly
be extended to those communities of people we call nations.
II
We can break down the move from individual responsibility to national
responsibility into three separate steps. First, we need to explain and
justify the idea of collective responsibility in general. We must show that
it makes sense to hold collective groups—teams, crowds, corporations,
and so forth—responsible for the effects of what they do in such a way
that the individual members of those collectives can properly be held
liable for the ensuing costs.13 Second, we must extend the argument to
nations, understood at this point as contemporaneous groups of people;
we must show that nations have features such that the general analysis
of collective responsibility applies also to them. Third, we need to de-
velop a further argument that can apply to nations considered now as
extended in historical time—we need to show, in other words, that it
makes sense to hold present-day nations responsible for the actions of
their forebears in such a way that their current members can be held
liable to bear the (self-regarding or other-regarding) costs of those ac-
tions. Each step raises difficulties, and in the present article I shall
attempt only the first two, leaving for another occasion the problem of
inherited national responsibility. So I shall adopt the fiction that nations
are groups of people with a fixed membership and restrict myself to
13. Some authors consider it important to distinguish between groups that have a
formal structure—with a defined membership, a fixed procedure for making decisions,
etc.—and those whose members are more loosely associated. This affects assignments of
collective responsibility—e.g., Cane (Responsibility in Law and Morality, chap. 5) ascribes
“group responsibility” to corporations and other such rule-governed bodies, but “shared
responsibility” to individuals acting in concert. He argues that in the former case but not
the latter, responsibility and the ensuing liability to pay costs remain with the group and
do not descend to individual members. Others have challenged this view: see, e.g., Larry
May, The Morality of Groups: Collective Responsibility, Group-Based Harm and Corporate Rights
(Notre Dame, Ind.: University of Notre Dame Press, 1987). I prefer to use the idea of
collective responsibility to cover all these cases, leaving it an open question for the moment
what difference the presence or absence of a formal structure makes when responsibility
is assigned. In the case of nations, I address this question in Sec. III of the article.

investigating the conditions under which members can properly be held
outcome responsible for the costs of national decisions, policies, and
actions taken in their lifetimes.
To help our thinking about collective responsibility generally, I want
to elaborate two models, which I shall call the ‘like-minded group model’
and the ‘cooperative practice model’. These are to be thought of as
ideal types to which real groups may approximate to different degrees
and that may in practice overlap—a real group, that is to say, may have
some features that belong to the like-minded group model and some
that belong to the cooperative practice model. I don’t want to claim
that it is a necessary condition for ascribing collective responsibility to
a set of individuals that they should display like-minded group or co-
operative practice features. It has been argued with some plausibility
that there are circumstances in which we are justified in holding even
randomly assembled collections of individuals responsible for the out-
comes of their actions.14 But I think that these are the models that are
most relevant in thinking about national responsibility; I shall argue
later that we are justified in holding nations responsible insofar as they
display like-minded group and/or cooperative practice features. So let
me begin with the idea of a like-minded group.
It is easiest to introduce this by means of an example. Consider a
mob rampaging through a neighborhood, terrifying the residents, de-
stroying property, and looting shops. Different participants in the mob
act in different ways. Some actively attack persons or property; others
shout abuse or issue threats; yet others play a more passive role, running
alongside the activists, urging them on and contributing generally to
the atmosphere of excitement and fear. If after the event we had to
apportion individual moral or legal responsibility for what has hap-
pened, we should need to identify the precise causal role that each had
played in creating the damage. But it is also the case, so I want to argue,
that the whole mob bears collective responsibility for the effects of the
riot, and together they can be held liable for the cost of repairing the
damage to persons and property. The specific intentions of each par-
ticipant at the beginning of the riot may have been different: some may
have started out meaning to inflict physical damage, others may have
wanted to make a political point, and so forth. What matters is that each
person took part with the same general attitude—“teaching them a
lesson,” “showing them that we mean business,” and so forth—and each
made some causal contribution to the final outcome, whether this in-
volved engaging directly in destructive acts or merely in supporting and
encouraging those who did. Indeed, we may not be able to disentangle
14. See Virginia Held, “Can a Random Collection of Individuals Be Morally Respon-
sible?” Journal of Philosophy 68 (1970): 471–81.

individual contributions. Consider several members of the mob throw-
ing bricks at a plate-glass window at roughly the same moment: we
cannot say that any particular brick thrower was (causally) responsible
for smashing the window, but we can say that the group as a whole is
outcome responsible for the damage they brought about.15
What justifies us in saying not only that the mob as a collective is
responsible for the damage that it has caused but also that its individual
members share in the collective responsibility?16 Recall that our interest
in outcome responsibility derives at least in part from our interest in
remedial responsibility. We want to know whose responsibility it is to
clean up the vandalized neighborhood. Attributing responsibility to the
mob as a collective will not help unless responsibility also descends to
the individual members, because these are the people who will actually
have to bear the clean-up costs (we might imagine them contributing
their labor or their money to the clean-up operation). But we also have
to show that attributing responsibility to individual rioters is justified,
and this we can do by recognizing that they contributed to a collective
activity that was certain to inflict damage on other people, whether or
not they specifically intended the overall outcome that actually occurred.
Recall that outcome responsibility does not in general require intention:
we hold people responsible for the consequences of their actions that
a reasonable person would have foreseen, whether or not these con-
sequences were intended and whether or not they were actually foreseen
by the person in question. This condition was surely met in the case of
the riot; anyone participating should have foreseen what a hostile crowd
entering a vulnerable neighborhood was likely to do.
Given that the responsibility of the collective descends to its indi-
15. See Michael Zimmerman’s argument that where more people than were necessary
to bring about O all acted in a way designed to bring about O, no participant can escape
responsibility by claiming that his or her actions were inessential to O’s occurring (Michael
Zimmerman, “Sharing Responsibility,” American Philosophical Quarterly 22 [1985]: 115–22).
16. For a thoughtful discussion of the circumstances in which the collective respon-
sibility of groups either does, or does not, descend to their individual members, see Joel
Feinberg, “Collective Responsibility,” in his Doing and Deserving: Essays in the Theory of
Responsibility (Princeton, N.J.: Princeton University Press, 1970), pp. 222–51. Feinberg takes
as an example of group responsibility without individual responsibility the case of a car
full of railway passengers who fail to prevent an armed bandit from robbing the train.
Acting together the passengers could have overcome the robber, but this would have been
heroic, since one or two would probably have been shot in the course of doing so. It
seems to me, however, that the group of passengers is only responsible for not preventing
the robbery in a causal sense. There is no collective outcome responsibility, in the sense
used here, since as we saw imputations of outcome responsibility depend on judgments
about what it is reasonable to expect of normal people. It was not reasonable to expect
the passengers to tackle the bandit, and so they should not be asked to bear the loss of
property that their inaction entailed.

vidual members, why not dispense with the notion of collective respon-
sibility altogether and instead focus entirely on the outcome responsi-
bility of each individual? As I indicated earlier, it may be impossible to
assign specific shares of responsibility for what has happened to indi-
vidual 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 recog-
nized 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 respon-
sible for restoring the damage they have caused, and that every partic-
ipant bears an equal share of that responsibility. It may then be possible
for the participants themselves to make finer-grained allocations of re-
sponsibility, depending on what is known about the activities of each
member, whether some can be identified as ringleaders, and so forth.17
This, however, is irrelevant from the point of view of achieving a fair
distribution of costs and benefits between the rioters and their victims.
From this perspective, all that matters is that the rioters as a group can
be held collectively responsible for the damage they have caused and
are therefore liable to bear the costs of repairing that damage. Similarly,
when nations behave in ways that are harmful either to themselves or
to others, our primary concern will be to establish collective outcome
responsibility for what has happened, insofar as this bears on the allo-
cation of costs between nations. Within each nation, particular individ-
uals or particular subgroups may then be identified as bearing a special
responsibility, depending on the circumstances. I shall not in this article
investigate principles for distributing responsibility within nations.
Returning now to the like-minded group model, this applies to
groups who share aims and outlooks in common, and who 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.
This is particularly clear in the case of the mob. As students of crowd
behavior have long recognized, people in crowds behave differently pre-
cisely because of the contagion of those around them.18 Groups that
exemplify this model are not then just collections of individuals who
happen to have aims in common; they are groups whose members in-
teract in such a way that even those who play no direct role in producing
17. Compare here Larry May’s argument that when groups are responsible for harm,
the share of responsibility that descends to each member should depend on the causal
role played by that person in bringing about, or failing to prevent, the harm. See Larry
May, Sharing Responsibility (Chicago: University of Chicago Press, 1992), chaps. 2 and 6.
18. See the analysis in May, The Morality of Groups, chaps. 2 and 4.

the outcome that concerns us may nonetheless properly be brought
within the scope of collective responsibility. And this allows us to widen
the model to take in cases that are less obvious than that of the rioting
mob but that bring us closer to the idea of national responsibility.19 A
good example is provided by Feinberg’s discussion of racism in the
postbellum American South.20 Acts of violence against blacks, Feinberg
suggests, were carried out in a context in which Southern whites gen-
erally passively sympathized with such acts, even if they were not actively
involved in perpetrating them, as a result of a widely shared culture of
racial inequality. In these circumstances it makes sense to hold all South-
ern whites collectively responsible for keeping blacks in a state of sub-
jugation. Feinberg argues that this includes whites who did not approve
of the beatings and lynchings, on the grounds of their solidarity with
the majority who did. This distinguishes the example from the case of
the rioting mob, where I claimed that relatively passive rioters who
shared in the general aim of the riot but took no physical part in in-
flicting damage on persons and property nonetheless were collectively
responsible for that damage. The argument in the case of the postbellum
American South is that 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.21 By participating
in the community they help to sustain the climate of opinion in which
the actions in question take place, even if they voice their opposition
to the actions themselves.22
This of course raises the question of what individuals have to do
in order to escape from collective responsibility for the results of the
19. The rioting mob example illustrates how individuals can share in collective re-
sponsibility for outcomes that they did not specifically intend, which is one important
aspect of national responsibility, but in other ways it works less well as a model of national
responsibility. It involves a specific event, limited in time, and it also allows for a relatively
clear demarcation between those who share in the responsibility and those who don’t: to
escape responsibility, in normal circumstances, all one has to do is to stay at home. I
therefore extend the model to include cases that exhibit neither of these features. I thank
Catherine Frost for her comments on this point.
20. Feinberg, “Collective Responsibility,” pp. 247–48.
21. See also here the discussion in May, Sharing Responsibility, chap. 2, sec. 4.
22. For another example, consider the collective responsibility of the Roman Catholic
Church for the sexual abuse of minors by priests. There is little doubt that the over-
whelming majority of church members condemn this behavior. Yet the general mind-set
of church officials has been such that effective measures to prevent such abuse have not
been put in place. We could say that the church has tolerated the abuse even while not
condoning it. So while individual responsibility clearly rests primarily with the small num-
ber of priests who have taken advantage of their position to abuse minors, we can hold
the church collectively responsible for the general ethos that allowed this to happen. (My
thanks to Catherine Frost for suggesting this example.)

actions of groups to which they belong. I shall postpone discussion of
this important question in order to introduce my second model of group
responsibility, the cooperative practice model. Again, an example may
help to bring out its main features. Consider an employee-controlled
firm whose manufacturing process has unwanted environmental ef-
fects—it involves depositing chemical substances in a river, for instance.
Members are divided on whether this practice should continue or
whether a different, more expensive, technology should be used, but
when the matter is discussed, the majority favors staying with the existing
process. The employees, I want to claim, are collectively responsible for
the environmental damage they are causing, and if they are required
to pay the costs of cleaning up the river, these costs should be borne
collectively by all the members. Why does collective responsibility extend
to the dissenting minority? They are the beneficiaries of a common
practice in which participants are treated fairly—they get the income
and other benefits that go with the job, and they have a fair chance to
influence the firm’s decisions—and so they must also be prepared to
carry their share of the costs, in this case the costs that stem from the
external impact of the practice. Here again we see the difference be-
tween holding people morally responsible for the results of their actions
and holding them outcome responsible. It would not in general be right
to blame (or punish) members of the minority for what their firm has
done to the river—they could quite properly defend themselves by say-
ing that they spoke out against the manufacturing process that caused
the pollution. But it is right to hold them, along with others, liable for
the damage they have caused.
The cooperative practice model goes further than the like-minded
group model in one direction, because there is no requirement here
that the group in question should share a common identity or have
aims in common; participating in the practice and sharing in the ben-
efits may be sufficient to create responsibility. In another way, however,
it is more restrictive, because it imposes fairness requirements that the
like-minded group model need not impose. Change our example in
such a way that the decisions about which technology to use are taken
by a small clique who keep the rest of the workforce in the dark about
the whole issue, or skew the distribution of rewards in such a way that
one section of the workforce could reasonably claim to be working on
exploitative terms, and collective responsibility no longer extends to all
members but at most to the decision makers or the leading beneficiaries
of the practice.23 The like-minded group model does not depend in this
23. One can envisage intermediate cases here, and one might also explore further
the question whether both procedural and substantive fairness are necessary in order for
the cooperative practice model of responsibility to apply, but I shall restrict my analysis
to the simple cases.

way on substantive fairness. So long as the group in question is genuinely
like-minded, its collective responsibility does not depend on how it al-
locates power, status, or other benefits among its members. We can
therefore see these models as indicating two complementary sources of
collective responsibility which may, as I have indicated already, overlap
in particular cases. You can share in collective responsibility for an out-
come because you form part of a like-minded group that has brought
the outcome about, or because you are a participant in a cooperative
practice that produces the outcome, or for both reasons at once.
My analysis of the sources of collective responsibility may, however,
set liberal alarm bells ringing. For it implies that in certain circumstances
membership in a group may be sufficient to establish responsibility for
acts performed by other members of that community even when one
is opposed to those acts, and this goes against an intuition that it is only
what a person does herself that can make her responsible for harmful
outcomes.24 If membership is sufficient for responsibility, then it seems
that no one can escape responsibility except by physically removing
himself from the group in question, a course of action that it may be
very difficult if not impossible for the person to take. So we need to
take a closer look at the conditions under which a member can legiti-
mately claim that he has acted in such a way that he bears no personal
responsibility for the harmful consequences of the policies and practices
of his group.
As already suggested, mere inactivity will not, in general, relieve
members of their group-based responsibility. A member who stands by
and does nothing still provides passive support to other members of his
group (in the like-minded group case) or still receives his share of the
benefits (in the cooperative practice case). Nor is it sufficient simply to
voice your opposition to the activities that are imposing the costs. Speak-
ing up is better than doing nothing, of course, but as our discussion of
Southern white racism revealed, even someone who voices opposition
to certain of her community’s actions may still, by virtue of her mem-
bership, contribute to the climate of opinion in which those actions
take place, because she subscribes to the community’s values in general,
reinforces them in her daily activities, and perhaps supports the com-
munity in material terms. Nor, if the group has a formal procedure for
reaching decisions, will voting against the action or policy in question
24. This intuition is not universally shared. Indeed, paradoxically, it may occur more
often to liberal observers looking in from the outside than to those on the ground who
find themselves included in collective responsibility. For a robust statement of the opposite
view—that mere membership may be enough to implicate someone in collective respon-
sibility no matter what she does—see Hannah Arendt, “Collective Responsibility,” in Amor
Mundi: Explorations in the Faith and Thought of Hannah Arendt, ed. James William Bernauer
(Boston: Martinus Nijhoff, 1987), pp. 43–50.

necessarily exempt you from responsibility. Democratic procedures work
on the basis that people who find themselves on the losing side of a
vote must regard themselves as bound by the result, unless that result
is so morally offensive, or so far outside the competence of the decision-
making body, that some form of civil disobedience (or its equivalent)
is justified. Just as a member of the minority must, except in these special
circumstances, comply with the majority’s decision even though she
strongly dislikes it, so she must bear her share of the costs if the decision
turns out to have costly consequences. After all, were she in the majority,
she would expect the losing minority to pay its share.
So what must a dissenting member do to escape from collective
outcome responsibility? Unfortunately it is difficult to say anything more
precise than that he or she must take all reasonable steps to prevent
the outcome occurring. What is reasonable in a particular case will
depend on how seriously harmful the prospective outcome is and what
costs different courses of action will impose on the dissenter. Consider
the case of the postbellum American South. Anyone who joined the
National Association for the Advancement of Colored People or one of
its predecessor organizations, who took part in public demonstrations
against white racism, and so forth, and who by virtue of these activities
faced hostility from his neighbors in the white community would surely
have met the condition. Equally, someone who embarked on these ac-
tivities but was then deterred by serious threats from racist groups to
the safety of her home and family should also be exempted from re-
sponsibility—this is not a cost that we can reasonably expect an average
person to bear in the course of trying to stop racist attacks on blacks
(some people may turn out to be willing to bear the costs, but in doing
so they reveal themselves to be heroic: our imputations of responsibility
must be based on [admittedly imprecise] judgments about what can
reasonably be expected of people in general, not on what exceptional
individuals are able to achieve).
But what if the most effective way to combat the outcomes that you
oppose is to work inside the relevant group or practice, rather than
adopting a stance of outright opposition? In the racism case, for in-
stance, a person opposed to violence toward blacks might believe, with
justification, that he would have greater influence by staying within the
white community and gently shaming it into adopting more liberal at-
titudes, whereas by actively opposing the community he would simply
be written off as a “nigger lover” whose opinions could henceforth be
ignored. Or again, where a group governs itself through democratic
procedures, the most effective way to change its policies may be to stay
within democratic parameters—accepting rather than contesting deci-
sions when you find yourself in a minority—since this will increase your
influence in the long term. In cases like this, it may be ethically better

to accept a share of collective responsibility for a bad outcome than to
seek to avoid responsibility by distancing yourself from the group or
the practice that produces the outcome. This is not a paradox, provided
we keep it in mind that we are talking about outcome responsibility
rather than moral responsibility of the blame-incurring kind. The person
who with good reason decides that he should use his position as an
insider to try to change the way that the collective behaves does the
right thing and is not morally blamable for the ensuing harm (when
his efforts fail or only partly succeed), but he does render himself liable
to pay his share of the costs.25
Is it an objection to the view of collective responsibility advanced
here that it makes people 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?
A liberal Southern white may bitterly regret that he finds himself in a
community that supports violence against blacks, with the result that
he has to choose between radical opposition that relieves him of col-
lective responsibility and working within the community to change at-
titudes while continuing to share in responsibility for the violence. Bear
in mind, however, that in other, less controversial, cases people can
become responsible for outcomes as a result of chance factors over which
they have no control. Virginia Held gives the example of three pedes-
trians who happen upon the scene of an accident. In order to save the
victim they must act in concert. It is only chance that brings them
together in that place, and yet, confronted by the accident and being
the only people able to rescue the victim, they become collectively re-
sponsible for the harm he suffers if they fail to form a team and act.26
This may be unlucky for them: they may have to miss appointments or
dirty their clothes to get the victim to safety, but this is luck of an
unavoidable kind. We would not be impressed if one of the pedestrians
asked “why me?”—not just as an expression of frustration at having to
miss the concert he was hurrying to attend but 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
25. A more elaborately described example in support of the conclusion that individ-
uals can be held responsible for the results of practices that they oppose can be found
in Juha Räikkä, “On Disassociating Oneself from Collective Responsibility,” Social Theory
and Practice 23 (1997): 93–108. Räikkä, however, attempts to argue that the individuals
involved in such cases may be morally blameworthy even though they are acting rightly,
all things considered. I believe that blame is inappropriate here, and that a different sense
of responsibility is at stake, as argued in the text.
26. Held, p. 479.

and help to create a majority for some less objectionable (but still ob-
jectionable) alternative. My responsibilities are thrust upon me by my
circumstances, but they do not cease to be my responsibilities because
of that.
The claim that people who belong to like-minded groups or who
participate in cooperative practices are collectively responsible for the
results of their behavior does not, then, depend on the assumption that
entry into such groups or practices was voluntary or consented to. This
is going to be important when we turn our attention to nations in the
following section because, exceptional cases apart, people do not choose
to belong to national communities: they are simply born into them.
What I have tried to do in this part of the article is to sketch two models
of collective responsibility, two cases in which people can justifiably be
held liable for the costs incurred by groups of which they are members.
The next step is to see whether the conclusions we have reached can
be applied to nations. Granted that nations are communities of some
kind, do they display the features that would justify ascribing collective
responsibility to their members?
III
What, then, is a nation, to repeat Renan’s famous question?27 It is first
of all a group with a common identity: belonging to the nation is partially
constitutive of the identity of each member (partially constitutive be-
cause national membership does not exclude belonging to other com-
munities of identity, such as religious or ethnic groups). In other words,
nations are not merely collections of individuals who happen to be
juxtaposed in physical space, in the way that the three pedestrians were
in Held’s example referred to above. They are groups of people who
feel that they belong together because of what they have in common.
Second, among the things they have in common is a public culture, a
set of understandings about how their collective life should be led,
including principles that set the terms of their political association (a
principle of political equality, for instance) and guide, in broad terms,
the making of political decisions (a principle of individual rights, for
instance). This shared public culture does not exclude significant cul-
tural differences among subgroups within the nation, nor does it
mean—this is important to stress—that there is no political disagree-
ment among the members. On the contrary, people who share a public
culture can disagree quite radically about what the principles embedded
in that culture entail in relation to particular issues. Third, nations are
groups whose members recognize special obligations to one another,
27. Ernest Renan, “What Is a Nation?” in Modern Political Doctrines, ed. Alfred Zimmern
(London: Oxford University Press, 1939), pp. 186–205.

so that in that respect they are not like groups formed on a contractual
basis to realize the predetermined aims and objectives of the members,
where the reason for becoming and remaining a member is entirely
instrumental. Fourth, the continued existence of the nation is regarded
by the members as a valuable good, so that even if we could imagine
the instrumental benefits of membership, such as personal security, be-
ing provided in some other way, they would regard with horror and
dismay any suggestion that the nation should be disbanded and its
individual members assimilated to other national groups, or that the
whole nation should simply be absorbed into a larger unit without its
distinct identity being preserved.
If these four features are necessary for a group of people to con-
stitute a nation, one might wonder whether any nations do, in fact, exist.
Nationality should not be confused with common citizenship: the citi-
zens of a given state may bear two or more national identities. But even
when that possible source of confusion is removed, we still need to ask
whether every co-national does, in fact, share the set of beliefs and
attitudes I have listed in characterizing nationality. Must every French
person believe that he or she has special obligations to co-nationals, or
that the continued existence of France is intrinsically valuable, if there
is to be a French nation? If so, it seems very unlikely that this or any
other nation actually exists.28
But the condition just proposed is too stringent. What is necessary
to the existence of a nation is that the beliefs and attitudes in question
should be generally held (and believed by those who hold them to be
correct), not that they be held by every single member. This is true of
communities of all kinds. For a religious community to exist, for ex-
ample, its members must hold certain beliefs in common and behave
in certain ways toward each other, but it can survive the presence of a
few dissident members whose beliefs are heterodox or whose behavior
violates principles of reciprocity. One can’t say precisely how much dis-
sidence can occur before the community ceases to exist as such, and
similarly with nations there must come a point where indifference to-
ward the national identity, or unwillingness to acknowledge national
obligations, would mean that the nation in question had become some-
thing else—a group of people who just happened to share a language
or some other cultural traits, perhaps. So when speaking of nations I
28. Couldn’t one sidestep the problem by defining as French only those people who
held the beliefs and attitudes in question? The problem here is that the full-fledged French
will want to include the deviants as part of the French nation, partly on the grounds of
cultural commonalties and partly because they think that these others should recognize
special obligations, etc. They regard the deviants as reprobates rather than as outsiders,
in other words. So one can’t simply adopt a narrower criterion for being French.

am making a broad empirical assumption that there exists, among many
peoples, a sufficient degree of convergence in attitudes and beliefs that
the four conditions are met for the great majority of members.
Belonging to a nation also involves a fifth feature: the aspiration
to be politically self-determining. But in the real world this aspiration
is met to very different degrees. There is a spectrum of possibilities here,
of which three in particular are worth singling out. A nation may lack
self-determination entirely, as when it is subject to imperial rule from
outside. Next, it may possess its own state but have a despotic or au-
thoritarian form of government, where the ruler or ruling elite is drawn
from the people and claims to be acting in their name, but is not subject
to any mechanism of popular control. Finally, the nation may be gov-
erned democratically, with major decision makers answerable to the
citizen body as a whole at periodic elections. We need to distinguish
these cases in order to decide how far nations can be regarded as col-
lective agents who might be held responsible for the consequences of
what they do. Nations can be said to act collectively in two different
senses. First, because their members share an identity and a public
culture, both the practices that they follow and the behavior of individual
members can be seen as expressions of that common identity and cul-
ture. We say, for instance, that Germans are hard working, meaning that
the way individual German workers behave reflects a shared norm of
industriousness that forms part of the public culture of Germany. Or
we find that the pattern of family relations in a particular country, and
the number of children who are on average produced, corresponds to
the religious or other cultural values of the nation in question. Here
there is no deliberate decision to behave in a particular way or to adopt
a particular practice, but nevertheless what happens reflects the national
culture in a fairly direct sense. Second, where a suitable political struc-
ture is in place, the political decisions that are taken will embody to a
greater or lesser extent the articulated beliefs and attitudes of the nation
in question. The closer we come to the democratic end of the spectrum,
the truer this will be. So we have two forms of collective national action:
action that is deliberately concerted through political channels and ac-
tion that is undertaken by individuals, or groups of individuals, but that
reflects some element in the national culture.29
29. Is the second case a genuine case of collective action, sufficient to ground col-
lective responsibility? Clearly the fact that a number of individuals follow the same norm
in their private or economic lives does not by itself allow us to say that they are engaging
in collective action. But where the prevalence of the norm is common knowledge, and it
is regarded as a component feature of national identity, then the fact that the behavior
in question is not formally coordinated does not mean that it cannot be regarded as a
form of collective action. Compare here the cases of the rioting mob and white culture
in the postbellum American South discussed in the previous section.

How closely do nations conform to either of the two models of
collective responsibility outlined in Section II? Does it make sense, first
of all, to regard them as like-minded groups liable to bear the conse-
quences of their actions? Since members share both a common identity
and a public culture—the first two features noted above—there is prima
facie reason to regard them as meeting this condition. But it is clearly
crucial to establish that their collective actions are a genuine embodi-
ment of the shared beliefs and values that go to make up the national
culture. Here the distinction just drawn between different levels of self-
determination becomes significant. Where nations are subject to outside
rule, any ascription of national responsibility becomes problematic. The
nation is governed in a certain way, but it does not act politically at all.
And even where we witness forms of collective behavior that significantly
affect the well-being of members or impose burdens on outsiders, it will
be difficult to say which of these are authentic expressions of national
culture and which are merely the work of individuals who claim that
what they are doing reflects that culture. In the absence of a political
forum in which national aims and values can be articulated and de-
bated, it will be difficult to establish how far the population as a whole
is implicated in support for the activities in question.30
At the other end of the spectrum we have nations that are dem-
ocratically self-governing. Here the policies pursued by the state can
reasonably be seen as policies for whose effects the citizen body as a
whole is collectively responsible, given that they have authorized the
government to act on their behalf in a free election (I shall return later
to the question whether political dissenters can also be held responsi-
ble).31 And even where the consequences flow from patterns of behavior
30. This is not to say that we can never attribute responsibility to nations that lack
political self-determination. The clearest cases may be those that resemble the racism of
Southern whites discussed earlier. Suppose two peoples, A and B, locked together under
the same system of imperial rule, feel mutual hatred and contempt for each other, and
this results in genocidal acts perpetrated by certain As against the Bs. Under these cir-
cumstances, it would not be wrong to hold members of nation A collectively responsible
for these savage acts, on the basis that almost everyone belonging to A contributed to a
climate of opinion in which such acts were regarded not as morally reprehensible but as
permissible, if not justified. Of course, before making this judgment we would need evi-
dence that the attitudes in question were indeed very widely shared; we would also need
to be sure that large sections of the population were not being bullied into offering their
support for the killers. But such evidence could in principle be found.
31. I leave aside here the difficult question of how far (if at all) democratic elections
can be seen as authorizing the governing party to carry out the policies contained in its
manifesto. Clearly, if the government acts in ways that were not announced beforehand,
and that could not reasonably be foreseen by the voters, responsibility for these policies
does not automatically extend to the citizens generally, though my comments below about
responsibility under autocratic regimes apply here too.

that are not the direct result of political decision, these patterns of
behavior are open to democratic control. Suppose, for example, that
the dominant religion encourages large families, and that as a result
the population is increasing at a rate that causes social problems of
various kinds. It is open to the government to adopt a population policy
that gives incentives for parents to limit the size of their families. If after
democratic debate such a policy is rejected, then we can legitimately
say that the nation in question is collectively responsible for the con-
sequences of population growth: its culture is such that it prefers large
families to, let’s say, less crowded roads and cities.
What now of the case where the nation in question is governed
autocratically by an individual or a small elite drawn from within? It is
certainly harder in these circumstances to lay responsibility at the feet
of the ordinary subjects. Yet two considerations must be taken into ac-
count even here. One is that the rulers may hold beliefs and values that
correspond more or less closely to those of their subjects even though
they are not formally accountable to them. To the degree to which their
authority depends on that fact, we can say that they are supported by
the people, and that when they act, or fail to act, the consequences flow
from beliefs and values that are common national property. Suppose,
for instance, that the state is a theocracy and that its rulers issue a decree
that results in the death of some person deemed to be an apostate. If
the issuing of the decree stems from religious beliefs and practices that
are generally adhered to throughout the population, then some share
of responsibility falls upon the nation as a whole, even if we want to say
that it rests primarily with members of the ruling group.
The second consideration is that subjects of the autocracy may have
a duty to resist it in the event that it begins to act in ways that are
manifestly wrong, whether the wrongness takes the form of injustice to
outsiders or simply of policies that are seriously damaging to the com-
mon interests of the nation itself. Mere passivity is then not sufficient
to escape responsibility for the policies in question. Everything will turn
on whether resistance is feasible, what the costs of resistance are, and
whether sufficient numbers of people can act together to make their
resistance effective. Unfortunately, correct judgment on these matters
may be difficult to achieve, particularly for outsiders who have no ex-
perience of living under a repressive regime. How far, for instance,
should we hold the Serbian people as a whole responsible for ethnic
cleansing in Kosovo, given that they had no avenues of direct control
over Milosevic and the army that he directed? Should they have been
expected to make greater efforts to coordinate their opposition to his
regime (we know that it was divisions among the opposition parties that
helped him to stay in power for as long as he did)? Or were the costs
of effective opposition greater than the average Serbian could be ex-

pected to bear? The difficulty in answering these questions should make
us hesitate before we jump to the conclusion that responsibility spreads
beyond the ruling elite to the nation as a whole.32
Another reason for hesitation is the possibility that the ordinary
subjects of the autocracy were effectively brainwashed into holding views
that support the policies in question. Attributions of national respon-
sibility depend on the idea that the activities that nations engage in
express beliefs and values that are genuinely shared by their members.
This does not require that each member should have thought it all out
for herself, so to speak; it does not exclude normal processes of so-
cialization whereby individuals are exposed to certain values and prac-
tices as they grow up and come to adopt and identify with those values
and practices. But where current political attitudes can be directly traced
to sustained propaganda efforts by an autocratic regime that allows no
dissenting voices to be heard—attitudes of extreme hostility, say, toward
a neighboring community—it is much less plausible to hold ordinary
people responsible for the consequences that follow. Just as we cannot
expect people to make superhuman efforts to oppose a regime, so we
cannot expect them to stand firm against the propaganda barrage that
descends on them (a few individuals will, just as a few individuals may
be willing to bear extreme costs to fight the regime, but our judgments
about responsibility should be based on what we can reasonably expect
of the average person).
What this shows is that the more open and democratic a political
community is, the more justified we are in holding its members re-
sponsible for the decisions they make and the policies they follow. Na-
tional values will still, to a large extent, be inherited in practice, but
they will be discussed and debated, alternative views will be expressed,
and so forth. There seems little objection in these circumstances to
requiring the members to bear the costs of what they decide to do. But
what of those who find themselves in a dissenting minority? Here we
32. Writing in the immediate aftermath of World War II, Karl Jaspers took a harder
line in The Question of German Guilt (Westport, Conn.: Greenwood, 1978). Jaspers distin-
guishes between legal, political, moral, and metaphysical guilt, where political guilt implies
“having to bear the consequences of the deeds of the state whose power governs me and
under whose order I live” (p. 31). In Jaspers’s view all citizens share in this political guilt,
irrespective of the nature of the regime that governs them, and so it seemed clear to him
that all Germans should be expected to pay reparations after the war. “We are politically
responsible for our regime, for the acts of the regime, for the start of the war in this
world-historical situation, and for the kind of leaders we allowed to rise among us. For
that we answer to the victors, with our labor and with our working faculties, and must
make such amends as are exacted from the vanquished” (p. 78). This included those who
had opposed the regime and those who stood wholly aloof from politics. “The sense of
political liability lets no man dodge” (p. 62). Jaspers thought, rightly, that ascriptions of
legal and moral guilt must be more discriminating.

need to turn to the second of our two models of collective responsibility,
the cooperative practice model. For as we saw in the previous section,
those who are engaged in cooperative practices from which they benefit
can be held responsible for the outcomes of those practices despite
their opposition to the policies which produced those outcomes. So how
far can we justifiably represent nations as cooperative practices writ
large?
The case for so regarding them rests on two claims. First, as indi-
cated earlier, nations are communities whose members see themselves
as having obligations of mutual aid that are more extensive than the
aid they owe to human beings generally. (I do not address here the
question how far these circumscribed practices of mutual aid are justified;
I am simply indicating that this is how fellow nationals standardly un-
derstand their relationships to one another.) These obligations are typ-
ically discharged by creating and supporting institutions that provide
protective and welfare services on which each member can call as the
need arises. To the extent that there is fairness in the way that these
services are funded and provided, we can say that each member belongs
to and benefits from a cooperative practice.33
Second, nations provide their members with a number of public
goods, foremost among which is protection of the national culture itself.
I am assuming here that, as indicated in my sketch of nationality at the
beginning of the section, people value their national membership and
want it to continue. They must also value, therefore, those cultural
features that lend their nation its distinct character—the national lan-
guage, for instance; the physical appearance of cities or of landscape;
cultural traditions that mark them off from other nations; and so forth.34
These features are often subject to erosion by outside forces, and so
members have to invest resources and accept restrictions on their own
behavior to preserve their cultural heritage. Again we see that nations
exhibit the features of a large-scale cooperative practice: each member
makes certain sacrifices in order to support a national culture from
whose continued existence each is presumed to benefit.
33. In practice, protective and welfare services are normally provided to all citizens
of the state in question, regardless of national identity. But their justification—in particular
the justification of their redistributive elements—rests on the idea that they are a way of
discharging obligations that fellow nationals owe to one another. In multinational states
there is a marked tendency for welfare services especially to be devolved to national
subunits, insofar as it is feasible to do so.
34. These features need not be valued individually by everyone who belongs to the
nation—it is possible both to recognize some aspect of culture as a distinctive national
trait and to dislike it and wish to change it. (David Copp has suggested American gun
culture as a relevant example here.) What is necessary is that the ensemble of cultural
features should be valued positively by the nation’s members.

So now let us return to the question whether responsibility for the
outcomes of political decisions and policies can be extended to those
who dissent from, and oppose, the decisions and policies at stake. Even
if they cannot be said to play any causal role in the genesis of those
policies (as the like-minded group model requires), are they nonetheless
involved in a cooperative practice that implicates them in collective
responsibility? There are two issues to consider. The first is whether the
nation in question does indeed distribute the benefits and burdens of
membership fairly, including the opportunity to participate in political
decision making. Where a minority group is exploited, or is excluded
from a significant range of benefits that members of other groups stan-
dardly enjoy, it will be hard to justify the claim that their membership
alone makes them responsible for the consequences of national deci-
sions. Whereas a group that loses on a particular issue in a democratic
forum can be included in responsibility for the result on the ground
that it will win on other occasions, and therefore benefits from an on-
going practice that allows collective decisions to be taken, no such ar-
gument applies to a group that is excluded from decision making al-
together or that forms a permanent and oppressed minority. Thus the
position I am defending does not lead to absurd conclusions such as
that German Jews share in responsibility for the effects of the Nazi
regime or that Iraqi Kurds share in responsibility for the actions of
Saddam Hussein. State membership itself does not entail collective re-
sponsibility if the conditions for a cooperative practice are not met.
The second issue is the extent to which the dissident group shares
with the majority the beliefs and values that constitute the national
culture. Recall that I have excluded from consideration here cases in
which the state includes two or more distinct national groups: where
this obtains, and the national cultures in question are somewhat at odds,
protection of the culture of either group ceases to be a genuine public
good for both communities. But even if we leave aside multinational
states, it remains an open question how much cultural overlap there is
between majority and minority. One important issue here is whether
the national culture of the majority includes elements that collide with
the ethnic or religious cultures of particular groups, as, for instance,
German national identity during the Nazi period embodied notions of
racial superiority that made it repugnant to Jews and other ethnic mi-
norities. So we need to draw a distinction between dissenters who oppose
the majority view on a particular issue—say, a pacifist minority vehe-
mently opposed to a war that the majority supports—while continuing
to subscribe to other aspects of national culture, and dissidents who
reject that culture in an across-the-board way and therefore see no value
in policies designed to promote it. In the latter case, the idea of the
nation as a cooperative practice fails, at least so far as this involves

contributing to public goods from which every participant subsequently
benefits. It is difficult to judge how often this case occurs in reality: to
the extent that national identities are liberalized, in the sense that they
are purged of ethnically or religiously exclusive elements, and are con-
stituted instead by political and cultural values that are accessible to all,
outright alienation from national culture will be rare.35
I have been concerned in this section both to defend the idea of
national responsibility and to identify its limits. To take the limits first:
where nations are subject to external or to autocratic rule, it is usually
difficult to identify acts undertaken by individual members or by the
state as genuinely national acts, and so it becomes inappropriate to
spread responsibility for those acts throughout the population in ques-
tion. Furthermore, where cultural divisions run deep, we may decide
that talk of a single nation (in the sense outlined at the beginning of
the section) is out of place. These cases aside, I have argued that where
nations act in ways that impose burdens on themselves or on others,
responsibility for such burdens falls on every member, even on those
who opposed the decisions or policies in question. The argument turns
on the sharing of beliefs and attitudes that characterizes national com-
munities and on the benefits that membership brings with it. So I con-
clude that we are not wrong, in general, to hold contemporary fellow
nationals responsible for actions performed in their name. But what
weight should we give to national responsibility when we are thinking
about global justice? I conclude the article with some brief reflections
on this question.
IV
I have already noted one significant limitation of the analysis presented
here: I have not tried to show that it makes sense to hold present-day
nations responsible for the deeds of their forebears, even though this
is a claim that is often made in current political debate, when compen-
sation is demanded for historical injustices associated with slavery, co-
lonialism, genocides, and so forth. In fact, I believe that it is possible
35. Are there any cases in which a nation has fractured along purely political lines,
meaning not just that people disagree sharply over concrete issues, but they find that they
also lack any common principles in terms of which they can seek a resolution of their
differences? Even where the disagreement is deep and long lasting, as in the case of slavery
in the United States, the two sides can still subscribe to many common values. But if,
hypothetically, we can imagine such a fractured community, then we would not have a
nation in the sense I am presupposing. Instead we would have a variant of the multinational
case: a political association formed between two distinct peoples, in this instance two
peoples divided by their basic political principles. In such a case it would be wrong to
hold the whole association responsible for what one of its constituent communities had
decided to do.

to extend the argument of this essay so that it includes inherited national
responsibilities, but this is a difficult task which requires new conceptual
tools to accomplish; neither of the two models of collective responsibility
I have employed—the like-minded group and the cooperative practice—
can be used in any straightforward way to explain inherited responsi-
bilities. So the analysis here only applies directly to national responsi-
bility among and between contemporaries.
If we look at the international distribution of resources—using re-
sources here in its broadest sense to cover rights and freedoms, op-
portunities, wealth, income, access to public amenities, and so forth—
and ask what justice requires of it, what role should ideas of national
responsibility play? How far are national communities justified in claim-
ing compensation from other nations on the grounds that their being
worse off is the responsibility of these others? And how far can in-
equalities between nations be justified on the grounds that each nation
is responsible for its own level of resources? As I noted at the outset,
the latter idea is often resisted by liberals because it appears to let rich
nations off the hook when confronted with world poverty, allowing de-
fenders of the affluent to argue that poor nations are responsible for
their own impoverishment.
Let me then begin by reiterating an earlier point, that outcome
responsibility for harm needs to be distinguished from remedial respon-
sibility, understood as the responsibility to put harm right. Sometimes
the agent who is outcome responsible is also remedially responsible, but
not always—most obviously in the case where that agent cannot now
remedy the harm in question. So showing that there are instances in
which nations that lack sufficient resources to keep all of their members
above a poverty line are themselves responsible for that lack, by virtue
of policies pursued in the recent past, or features of the national culture
that have prevented sufficient resources from being generated, does not
entail that other nations should not now step in. It may, however, affect
the terms of the intervention: it is in general more justifiable to set
conditions on the receipt of resources when it can be shown that the
receiving agent is responsible for the resource deficiency being ad-
dressed. Donor nations can legitimately try to protect themselves against
being asked to make further transfers in the future by requiring changes
in economic and social policy that will prevent the deficiency from
recurring.
It is also relevant that national poverty is rarely the sole responsi-
bility of the nation that suffers it. It is, instead, the result of a national
culture faced with an external political and economic environment for
which other nations are diffusely responsible. One should not treat this
external environment as though it were simply a fact of nature. Cultures
that might be successful in more benign circumstances may prove in-

capable of responding to the external circumstances that they actually
face. For instance, a culture that encourages traditional, somewhat in-
flexible, patterns of work may be perfectly viable in an economic en-
vironment where commodity prices remain reasonably stable over time
but prove unable to cope with rapidly fluctuating prices. To say this is
not to condone any culture whatsoever: a culture which, say, values
looting one’s neighbor’s property more highly than acquiring resources
by honest toil is likely to have disastrous consequences whatever the
external environment is like. But such extreme cases aside, we should
recognize that national cultures will be partly rather than wholly respon-
sible for the life chances of their adherents.
Finally, it is worth underlining a point made earlier in relation to
other-directed harm, that where nations are autocratically governed, it
is wrong to hold their members responsible for the policies that are
followed if they have little chance to resist or are brainwashed into
unquestioning faith in their leader and his ideas. It seems that extreme
poverty and effective brainwashing do sometimes go hand in hand:
North Korea is a relevant contemporary example. People who are strug-
gling to stay alive are in a poor position to resist the mixture of coercion
and propaganda that keeps regimes like that of Kim Il Sung and his
son in power. In these circumstances, to hold them responsible for their
own deprivation is to add insult to injury.
These are reasons why we should not overemphasize national re-
sponsibility when thinking about the problems of global justice. But we
should not underemphasize it either. Rich nation-states should be made
to accept collective responsibility for the external economic effects of
their trade policies and the activities of corporations based in their
territory. It is easy to slip into a worldview that sees the international
economy as subject to market forces that no state can control. But, in
fact, states are quite able to consider the impact of, say, tariff barriers
on dependent nations when deciding whether to impose them, and
equally they have the power to insist that corporations should observe
decent labor and environmental practices in their overseas operations.
If they fail to do these things, and as a result vulnerable communities
or individuals are harmed, the states in question should be held re-
sponsible for the harm and made to compensate the victims. And, sub-
ject to the caveats entered in Section III of the article, responsibility
here will descend from the state to the national group whose state it is.
Equally, poor nations should not be allowed to claim that they have
no choices in how they respond to international economic conditions.
As we can readily observe, developing nations that initially had rather
similar resource bases have shown a huge variance in their economic
performance. In this context, claims of powerlessness may serve as a
smokescreen behind which a large proportion of gross national product

is being siphoned off into Swiss bank accounts or presidential jets. Such
nations need good governance and, subject to the caveats already en-
tered, the people as a whole should be held responsible if they under-
mine its conditions by, for example, colluding in corrupt practices or
accepting bribes at election time.
If we are prepared, under suitable conditions, to apply ideas of
national responsibility to both rich and poor nations, then this shows
at least that global justice cannot be understood as requiring simple
equality between peoples or indeed the global application of the Rawl-
sian difference principle.36 We need to look not just at how resources
are distributed at any moment but at what brought the current distri-
bution into being and, in particular, at how far collective agents are
responsible for being better off or worse off. More positively, we may
come to think that a just world order would be one that gave maximum
scope to national responsibility, subject to the proviso that no one should
have to suffer a standard of living that falls below a decent minimum.
Nations should have the opportunity to make real choices in areas such
as economic policy, population policy, and environmental policy, and
the effect of this will inevitably be that some become relatively richer
and others relatively poorer when wealth and poverty are measured in
conventional money terms. The responsibilities of outsiders here are to
help create an external environment that allows choices of these kinds
to be made, to encourage and support institutions of self-government,
and to transfer resources directly in cases where people are unavoidably
falling below an acceptable minimum standard of life.37 We should aim,
in other words, for a world order in which national responsibility be-
comes feasible and genuine for everyone, not merely a lame excuse for
inaction on the part of the better off.
36. For a fuller argument to this effect, see my “Justice and Global Inequality.” For
Rawls’s own rejection of the global difference principle, on grounds similar to those
canvassed here, see Rawls, p. 117.
37. I have tried to tackle the question of how such an acceptable minimum can be
identified in David Miller, “National Responsibility and International Justice,” in The Ethics
of Assistance: Morality and the Distant Needy, ed. Deen Chatterjee (Cambridge: Cambridge
University Press, 2004), pp. 123–43.


Miller 2004

Uploaded by

Copyright:

Available Formats

Miller 2004

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Miller 2004

Uploaded by

Copyright:

Available Formats

Holding Nations Responsible

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

Contemporary liberal theories of justice lay heavy emphasis on the idea

Ethics 114 (January 2004): 240–268

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

1. Some philosophers prefer to think about issues of international or global justice

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

12. Miller, “Distributing Responsibilities.”

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

This content downloaded from 195.34.79.253 on Fri, 20 Jun 2014 21:01:36 PM

You might also like