Rights, Liability, and the Moral Equality of Combatants
U. Steinhoff (&) Department of Politics and Public Administration, University of Hong Kong
, Pokfulam Road,
Hong Kong, People's Republic of China
According to the dominant position in the just war tradition from Augustine to Anscombe and beyond, there is no ''moral equality of combatants.'' That is, on the traditional view the combatants participating in a justified war may kill their enemy combatants participating in an unjustified warbut not vice versa (barring certain qualifications). I shall argue here, however, that in the large number of wars (and in practically all modern wars) where the combatants on the justified side violate the rights of innocent people (''collateral damage''), these combatants are in fact liable to attack by the combatants on the unjustified side. I will support this view with a rights-based account of liability to attack and then defend it against a number of objections raised in particular by Jeff McMahan. The result is that the thesis of the moral equality of combatants holds good for a large range of armed conflicts while the opposing thesis is of very limited practical relevance.
their enemies on the justified side, at least not if they know or find it highly likely
that their own war is unjustified (Augustines view is more lenient, but does not
amount to fully embracing the moral equality of combatants).
The philosopher who has probably done the most to attack the doctrine of the
moral equality of combatants in recent times is Jeff McMahan.1 According to him,
the combatants on the just and the unjust side are not both liable to attack
(McMahan 2004, 706), do not have both a liberty-right2 to kill each other. Rather,
the unjust combatants have no right to kill the just ones, and the just ones
have a claim-right to kill the unjust ones (McMahan 2009, 64). This is why:
People dont lose moral rights by justifiably defending themselves or other
innocent people against unjust attack; therefore, unless they lose rights for
some reason other than acquiring combatant status, just combatants are
innocent in the relevant sense. So, even when unjust combatants confine their
attack to military targets, they kill innocent people. Most of us believe that its
morally wrong to kill innocent people even as a means of achieving a goal
thats just. How, then, could it be permissible to kill innocent people as a
means of achieving goals that are unjust?
McMahan thinks that arguments of this kind conclusively demonstrate the moral
inequality of combatants at the level of basic morality (McMahan 2006, 379).
We should be clear about what this moral inequality means. It is not that the
just soldiers are nice, and the unjust soldiers are not. Rather, the claim is
that unjust soldiers, by deliberately targeting and killing just ones, are deliberately
targeting and killing innocent people. And normally, of course, we think that
intentionally killing innocent people (without their consent) who do not pose an
unjust threat amounts to murder or manslaughter and is at any rate one of the most
egregious moral crimes one can commit. If this is true, however, it would seem that
soldiers would be required to very seriously rethink their obligations; and civilians
who consider it their civic duty to support their troops would have to face the fact
that doing so might mean nothing less than supporting murderers or
manslaughterers. Thus, the issue at hand here is hardly trivial.
Is McMahan right? I think that he is right in principle (and that therefore Walzer
is wrong); that is, I think it is not true that in all wars the combatants on both sides
have the same liberty-right to kill enemy combatants, provided they abide by the
traditional jus in bello restrictions. However, McMahan greatly exaggerates the
scope of his argument (and so do others).3 For many, if not most modern wars it has
1 Others include C. A. J. Tony Coady, David Rodin, Lionel K. McPherson, Gerhard verland, and
2 McMahan himself does not talk about liberty-rights. For my account, however, the distinction between
claim-rights and liberty-rights is important. That a person A has a liberty-right towards another person B
to do x means that she is under no duty towards B not to do x. If she has a claim-right towards B to do x,
this means that B is under a duty towards her, A, not to interfere with her doing x. B is not under this duty
if the person in question has only the liberty-right to do x. On the distinction between liberty-rights and
permissions, see n. 15.
3 I have made this argument in Steinhoff (2007a, 9597) and (Steinhoff 2008, 220226). Jeff McMahan
has responded to my criticism in McMahan (2008b, 227244) and (McMahan 2009, 3851).
little relevance, since in many, if not most modern wars just soldiers do kill
innocent and non-threatening people or non-innocently contribute to their being
killed.4 The military euphemism for this is collateral damage; I prefer the term
concomitant slaughter. By participating in or non-innocently contributing to the
killing of innocent and non-threatening people one wrongs these people (for
innocent and non-threatening people haveand McMahan agreesa right not to be
killed), and someone who wrongs others and violates5 their rights cannot be just.
Thus, the soldiers are at best justified, but that does not make them innocent in the
relevant sense (namely in the sense of not wronging others). They remain liable to
In developing this argument I will proceed as follows. In Sect. 2 I will clarify the
historical background and show that, contrary to a widespread belief among
presentday just war theorists, the moral inequality thesis is the traditional and orthodox one.
In Sect. 3 I will briefly sketch an account of liability to self-defensive killing that is
inspired by Judith Jarvis Thomsons approach but revises it in ways that make it
immune to certain criticisms that can be leveled against the original theory (3.1).
I then apply this approach to war as a collective action (3.2).
In Sect. 4 I will deal with McMahans objections to the account proposed here.
He argues, first, that it has counter-intuitive implications; and he claims, second,
that justification defeats liability so that just combatants are not liable to be
attacked by unjust combatants. I will argue, in response, that the claim that
unjust combatants are liable to attack by just combatants but not vice versa is
true by definitional fiat and thus irrelevant to the substantive question at issue (4.1);
that it is not the account presented here but McMahans that comes with
counterintuitive implications (4.2 and 4.3); and, most importantly, that justification does not
defeat liability (4.4).6
In Sect. 5 I counter McMahans claim that the best way for the participant in an
unjustified war to protect his fellow citizens is to lay down his arms and that
therefore he should do precisely that. I argue, first, that this claim, even if it were
correct, would not undermine the point that both sides would have a liberty-right to
kill each other: what liberty-rights you have towards a person P, and what you are
permitted to do to P all things considered, are not the same. Second, McMahan is
not correct: he ignores collective action problems and certain realities of war.
4 An additional reason why McMahans argument has little relevance for practice is that there are very
few if any justified wars. On this point see Steinhoff (2007b) and (2012).
5 I distinguish, incidentally, between justified rights violations on the one hand and unjustified rights
violations on the other by using the terms justified rights violation and unjustified rights violation.
Thus, I refuse to follow those who call the former infringement and the latter violation. It does not
make any difference for the victim, after all, and the distinction between violation and infringement
might mistakenly suggest otherwise.
6 Incidentally, even if justification did defeat liability, this would not even matter. For an argument to that
effect, see Steinhoff (2008, 223). McMahans elaborate reply to this argument can be found in (2008b,
236244). For a no less elaborate response, see Steinhoff, McMahan, Symmetrical Defense and the
Moral Equality of Combatants, unpublished ms.
2 The Moral Inequality Thesis in History: On Imagined and Real Orthodoxies
McMahan presents the thesis of the moral equality of combatants as the traditional
one and his own account as revisionary. Yet he is mistaken about the just war
tradition, and so are many other present-day just war theorists. In order not to let this
mistake go uncorrected, let me clarify the historical facts.
First of all, when McMahan started to criticize the orthodox view, as he calls it, he
never actually provided any textual evidence for his claim that the thesis of the moral
equality of combatants is orthodox or traditional. To be sure, in those articles he
(somewhat selectively) mentions a couple of twentieth century writers (McMahan
1994a, 256, fn. 7, and b, 194, fn. 3), but the question is precisely whether those authors
really represent the traditionwhich reaches through the centuries.
In fact, his main target has always been Michael Walzers theory; after all,
Walzer coined the phrase the moral equality of combatants to begin with (Walzer
1977, 3441).7 The irony, however, is that Walzer himself explicitly stated (in a
passage that seems to have been widely overlooked by those who take him to
represent the tradition) that his far-reaching thesis of the moral equality of
combatants is not shared by traditional authors:
Catholic writers have long argued that [men] ought not to volunteer, ought not
to serve at all, if they know the war to be unjust. (Walzer 1977, 39)
Indeed, these authors have claimed that under these circumstances it is sin to join
the fraywhile Walzer with his moral equality thesis denies that it is morally
Recently McMahan has admitted that his previous reading of the tradition is
mistaken and that Francisco de Vitoria and Francisco Suarez have beaten [him]
into print by more than half a millennium (McMahan 2009, 237238, n. 26; see
also the text on p. 33 that McMahans note refers to). On these pages and elsewhere
in the book he thus provides textual evidence to the effect that two central writers in
the just war tradition reject the moral equality thesis; while he still does not offer
quotes from any traditional authors who would actually support the moral equality
thesis. Yet he simply continues to call the moral equality thesis the traditional
Maybe he thinks that Suarez and Vitoria are the only of the older authors that
have beaten him. But he is wrong. He has also clearly been beaten by, for
example, Domingo de Soto, Melchor Cano, Luis de Molina und Thomas Cajetan.8
And, of course, by still a few centuries more, by Aquinaswhom Suarez and
7 Hence the moral reality of war can be summed up in this way: when soldiers fight freely, choosing one
another as enemies and designing their own battles, their war is not a crime; when they fight without
freedom, their war is not their crime. (Walzer 1977, 37).
8 De Soto, Questio XL De Bello, Articulus Primus, Dubium Septimum (120); Cano, Questio XL
De Bello, An Christianis sit licitum bella gerere, Secunda Quaestio (152); Molina, Questio XL De
Bello, Articulus Primus, Disputatio Tercera, 13 (273); Cajetan, Summula, bellum, Bellum, quando
dicatur iustum vel injustum, licitum vel illicitum (420). All these texts are to be found (in Latin and
German) in Justenhoven and Stuben (2006); the page references refer to this edition. I have used the Latin
designations to make it easier for the reader to find the relevant passages in the English (or X-ish)/Latin
edition of her choice.
Vitoria, incidentally, interpret in precisely this way. To be sure, in STh IIII, q. 40
(On War), Aquinas does not address this issue clearly, but elsewhere he explicitly
states that people are bound not to obey the ungodly commands of tyrants
(Aquinas 2006, 195). That statement removes the excuse that Augustine once seems
to have made available for certain combatants on the unjustified side and is thus
hardly compatible with a general moral equality of combatants.
Finally, even Augustine himself, still a few centuries earlier, does not embrace a
general equality of combatants. True, as just mentioned, he states in Contra
Faustum, 22.75, that a soldier can innocently obey the command of an ungodly
king (Augustine 1887, 301). (Incidentally, this is not the same as stating an
ungodly command; besides, Augustine arguably had an ungodly Christian king in
mind). Yet one of the reasons why even this does not amount to an endorsement of a
general moral equality of combatants is that some wars might not be waged under
the command of kings (or of their equivalents) on both sides, ungodly or not.
Certain civil wars or slave revolts are an example. (It should also be noted that
actually none of the authors mentioned in the last paragraph, with the possible
exception of Aquinas, took Augustine to believe that combatants are permitted to
fight in a war they know to be unjust).
Thus, once people understand the actual tradition of just war theory, it becomes
very difficult to call the moral equality thesis orthodox without appearing to
engage in false labeling. Yet Helen Frowe, for example, echoing McMahan, insists
on this terminology. Again: why? She does seem to have realized that going back to
Suarez and still further beyond him will not work. So instead, she goes only back to
Grotius and claims that the idea that ordinary citizenscombatants or otherwise
might be morally responsible for their countrys aggression was roundly rejected by
Grotius, and that what she calls the orthodox view of just war theory has been
the dominant view of how we should understand the relationship between jus ad
bellum and jus in bello since Grotius (Frowe 2011, 161 and 118). I do not know
whence Frowe takes these ideas about Grotius.9 After all, Grotius says, completely
in line with Suarez, Vitoria and the other older just war theorists already mentioned:
But if [free people as well as slaves and subjects] have Order given them to
take up Arms, as is usual, then if it plainly appears that the War is unlawful, it
is their Duty not to meddle in it. (Grotius 2005, 1167 [II.26.III.1])
Consequently, he also states:
Neither can I admit that even those who have given just Cause to take up
Arms against them, may lawfully [he refers here to natural law, that is, to part
of morality] defend themselves; [for] no Man [has] a Right to oppose Force
to a just Attack, no more than a Criminal can plea a Right of defending himself
against the publick Officers of Justice, who would apprehend him, by Order of
the Magistrate (Grotius 2005, 417 [II.1.XVIII.1])
9 They are not from Grotiuss text, apparently. While she mentions him a few times, she never actually
quotes him, nor is his book included in her bibliography. For a succinct interpretation of Grotius that
supports mine, see Forde (1998). For in-depth analyses, see the contributions in Onuma (1993).
This is a clear argument for the moral inequality of combatants, no less clear than
McMahans argument, quoted above, that nearly 400 years later conclusively
demonstrate[s] the moral inequality of combatants at the level of basic morality
according to him (McMahan 2006, 379).
Grotius makes even the same qualification that McMahan does (Grotius 2005,
11811183 [II.26.VI]), and McMahan acknowledges this and quotes the pertinent
statements from Grotius (McMahan 2009, 1617). However, he does not provide
the quote I just adduced, nor does he acknowledge that Grotius is a fervent
partisan of the moral inequality of combatants. That is curious, for since
McMahan says that Grotius is arguably the greatest writer in the tradition of
just war theory (McMahan 2009, 16), one would think that McMahan would be
eager to point out that he has him on his side. Of course, the downside might be
that once you acknowledge that arguably the greatest writer in the tradition of
just war theory embraces your allegedly unorthodox and revisionary moral
inequality thesis, that thesis does not appear all that unorthodox and revisionary
The traditional inequality thesis survived into the twentieth century and was
embraced, long before McMahan, by Elizabeth Anscombe. McMahan, however,
claims that she is inconsistent (McMahan 2005a, all quotes in this paragraph
are taken from fn. 17 of that aricle), pointing out that in one place she argues that
the non-innocent people that can rightfully be attacked are those engaged in an
objectively unjust proceeding, while in another place she states that they are
those who are causing harm. Presumably, the first idea would imply the moral
inequality thesis while the second does not. In reply, first, let me point out that in
light of the argument presented in the present article, namely that combatants on
the justified side are also engaged in an unjust proceeding (namely, of harming
innocent bystanders), the first statement of Anscombes would not actually have
the implication McMahan ascribes to it. Maybe it is because he is aware of this
that he opts for charitably interpreting that first statement of Anscombes as
meaning engaged in objectively wrongful action (my emphasis). However, to
be consistent himself, McMahan should then be equally charitable towards her
second formulation, for example by reasonably interpreting harming as
meaning unjustifiably harming here. (This is made all the more reasonable by
the fact that Anscombe seems to be concerned here with the combatants on the
German side in the Second World War, thus already assuming that these
combatants are on an unjustified side. McMahan is, in my view, simply taking the
quote out of context).
Second, to take this isolated slip in terminology on Anscombes part as indicating
her inconsistency in the substantive matter is like taking the fact that McMahan
sometimes states that liability to defensive attack lies in moral responsibility for an
unjustified threat and sometimes that it lies in moral responsibility for an unjust
threat as indicating his inconsistency in the substantive matter. This would be unfair
and somewhat strange because McMahan explicitly embraces the moral inequality
thesis. But so does Anscombe, unequivocally stating, completely in line with the
tradition: Nor, if we know that a war is wrong, may we take part in it without sin,
however grievous it may seem to stand apart from our fellow countrymen
(Anscombe 1981, 73).10
Third, McMahan also fails to mention that Anscombe writes, just a few sentences
after the quote about harming on which he hangs so much:
Nor is there ground for trying them [the combatants on the German side, it is
safe to assume] on a criminal charge; not indeed, because a man has no
personal responsibility for fighting, but because they were not the subjects of
the state whose prisoners they are. (Anscombe 1981, 67)
The statement after not indeed is nothing less than an endorsement of the quite
traditional moral inequality thesis combined with Grotiuss legal impunity thesis.
Thus, the fact of the matter is that Anscombe is not more ambivalent about the
moral inequality of combatants than is McMahan.
McMahan also claims that the moral equality thesis informs the international
law of war (McMahan 2009, 3). Again he provides no evidence for this claim.11 So
how does he know? After all, McMahan himself acknowledges that the law of war
does not assert the moral equality of combatants but the legal equality of
combatants (McMahan 2009, 105). And in fact, he himself has argued that for
moral reasons international law must not simply turn moral inequality into legal
inequality (McMahan 2008a). So why is it not this idea that underlies the
international law of war?
This question particularly suggests itself since Grotius was of the decided
opinion, as McMahan indeed knows, that the law of nations has to differ from
natural law and is designed and accepted in order to serve certain purposes,
including, in particular, moral purposes (McMahan 2008a, 34). But since Grotius
is not only arguably the greatest writer in the tradition of just war theory but also
considered by many the founder of modern international law, it seems natural to
assume that what actually underlies the modern laws of armed conflict is the Grotian
combination of the moral inequality thesis with the legal impunity thesis. Thus, far
from revising anything, McMahan is in essence repeating a view that Grotius
already expressed 400 years ago and which seems to be embodied in modern
Finally, it should be noted that while McMahan differs from Walzer with regard
to one thesis, namely the moral equality thesis, his views about, for example, just
cause, proportionality, and double effect are, aside from quibbles about details,
comparatively conservative and not so much different from both Walzers and
traditional opinions. He also has little to nothing to say about the just war criteria of
10 With due embarrassment I have to admit that in Steinhoff (2007a, 62) I correctly argued that her ideas
would imply the moral inequality of combatants and incorrectly pondered that she might not have realized
this. As the Anscombe quote just provided shows (which I had also overlooked, and which, incidentally,
McMahan does not mention), she definitely did realize it.
11 Frowe (2011, 124) follows McMahan here, too: both in making the claim and in not providing
evidence for it.
prospects of success, last resort, and right intention. Thus, it would seem,
one has to look elsewhere for a really revisionary and unorthodox account of just
After thus having clarified the historical background, let us now turn to the
3 Liability to Attack: A Rights-Based Account
Why are even combatants in a justified war13 liable to attack, that is, why can they
be attacked without thereby being wronged, without thereby having their rights
violated? I have already answered this in the introduction, but let me make it clearer.
3.1 Liability to Attack in General
The account of liability to attack presented here is a revised version of Judith Jarvis
Thomsons (1991) account (see also Steinhoff 2007a, 7198, for present purposes
esp. 8194). I agree with Thomson that violating a persons rights makes you liable
to attack. However, I disagree with her on two important points. First, in my view,
rights violations are not the only thing that can make a person liable to attack.
Second, I also disagree with Thomson as to what can count as a rights violation. In
Thomsons view, even an innocent threat, for example an innocent man who is
thrown by a villain from a cliff and now threatens to crush the person beneath,
violates the rights of the person he threatens.
Many people have found this view of Thomsons counter-intuitive. It would seem
that you can violate a persons rights only by acts. I tend to agree. However, I still
think that the innocent falling man is liable to attack. Why?
On the account offered here people have a general right to defend themselves
against unjust threats of all sorts (an unjust threat being one that does not have a
right to pose the threat and has not been set in motion by someone who had a right to
do so). That is, this right refers not only to threats posed by persons, but to all
threats, including to those posed by inanimate objects. That there is such a right is
12 The reader will not be surprised and might forgive me if I humbly recommend the account presented in
On the Ethics of War and Terrorism and elsewhere: I provide a sustained argument against legitimate
authority, argue for the possibility and permissibility of individual war, subsume proportionality
completely under just cause, am very lenient with regard to right intention, reject last resort and prospects
of success as necessary just war criteria, reject the principle of double effect, do not think that the ius in
bello principles of discrimination and proportionality are independent, defend terrorism to a certain
degree, and radically circumscribe the century-old orthodoxy of the moral inequality of combatants.
Whatever else that account is, it is certainly not orthodox.
13 When I mention a justified war, I am referring to a war as a complex action, not as a historical
event. The Second World War is a historical event, Britains war against Germany an action, and
Germanys war against Britain another action. See Steinhoff (2009a, 135136). See also McMahan (2009,
shown, for instance, by the injustice of a law that prohibits destroying a rock that
threatens to crush me.14
While a rock does not have any rights, the falling man has. He has a right to life
and a right to self-defense. He has no right to crush innocent people, however, and
therefore he poses an unjust threat. Now if the rights of the two persons are not
compatible in this situationthat is, if the innocent person below can only defend
himself and save his life by vaporizing the unjust threat above with his ray gun
then, on the account presented here, the (claim-)rights in their collision can no
longer be upheld as (claim-)rights (for as claim-rights they are not compatible) but
instead become mere liberties (liberty-rights) (Steinhoff 2007a, 8586 and 888915).
While a right, understood as a claim right, implies that the person I hold the right
against cannot interfere in my exercise of it without wronging me, a liberty (-right)
held against a certain person only implies that I am not duty-bound towards that
person not to exercise this liberty and thus implies that by exercising it I would not
wrong her. It does not imply that the other person would wrong me if she tried to
keep me from exercising my liberty. Thus, on my account, in this case both parties
have lost their claim-rights to life and to self-defense but retained their liberty-rights
to life and self-defense.
This account is perfectly consistent, Seth Lazars doubts notwithstanding. To wit,
he states that if one can only lose the claim right not to be killed by posing an
unjust threati.e., a threat that violates another persons claim right not to be
killedthen it cannot be the case that two parties are each liable to be killed by the
other, in virtue of the threat each poses to the other (personal communication).
However, the i.e.-part of this statement ignores both my actual use of unjust
threat and the fact that I explicitly reject the view that one can lose ones claim
right to be killed only by violating another persons rights.
Thus, the situation of the innocent threat and the (potential) innocent victim is a
symmetrical defense case in which both parties are liable to attack. The fact that the
falling man is morally innocent is not sufficient for him to retain his specific right to
life in this situation: innocent people posing a threat to someone who is not
responsible for initiating a threat to another innocent and non-threatening person do
not have a right to life (at least not if killing them is the only way to stop them).16
And thus the innocent man below may try to vaporize the falling innocent man, but
the falling innocent man may also try to prevent this by killing the innocent man
below (the falling man might be armed).
14 McMahan (personal communication) notes against this example that the reason why it is permissible
to destroy the rock is not that the rock is liable but that there is simply no moral objection to destroying it.
However, I do not make any claim about the liability of rocks here. I claimand my example shows that
claim to be correctthat people have a general right (which in a situation of rights-clashes can be
transformed into a specific liberty) to destroy objects that unjustly threaten to destroy them, whether those
objects violate their rights or not.
15 There I sometimes used the term permission where I meant liberty-right. In fact, however, it is
necessary to distinguish permissions from liberty-rights. For example, a person A might have a
libertyright against person B to kill B without being permitted to kill B. In other words, just as claim rights can
sometimes be justifiably violated for the greater good, liberties, for the sake of the greater good, must
sometimes not be exercised.
16 This distinguishes this symmetrical defense case from the asymmetrical soldier/civilian cases below.
3.2 Rights-Violations in War
So how do people on the just side of a war violate the rights of innocent people
and thereby become liable to attack?
One obvious way of doing so is by personally killing or mutilating them, or by
imposing upon them a significant risk of their being killed or mutilated.17 Many
just soldiers clearly do exactly that by shooting and dropping bombs.
Many other just soldiers, however, do not themselves kill or mutilate innocent
people or impose significant risks on those people. They only participate in or
contribute to so doing. At least three types of participation or contribution need to
be distinguished: (a) actual participation in an unjust (collective) attack or an unjust
(collective) imposition of risk; (b) non-participatory guilty contribution; and
(c) non-participatory innocent contribution.
Let me first give one example outside of the war context for each type. The
Jackal wants to kill an innocent person. Alex has marked the target in some way,
full well knowing that this is supposed to enable the Jackal to murder the innocent
person. This is participation or assistance, and is also called such in most Western
jurisdictions. Incidentally, it doesnt matter whether this marking was actually
necessary for the Jackals plan to succeed or not; that is, it does not matter whether
Alex causally contributed.
Bob has sold the Jackal the rifle he needed for the task, full well knowing that the
Jackal will try to murder an innocent person with it. This is non-participatory culpable
contribution, in most Western jurisdictions also called aiding and abetting.18
Cecilia, a cook in a restaurant, has served the Jackal a meal, full well knowing
that the Jackal will soon kill an innocent person (she has no means to call the police
or to stop the Jackal in another way, but she would have been able to deny the Jackal
the meal), and Cecilias son has sold the Jackal ammunition, not knowing that his
customer will use it to kill an innocent person. These two cases are
nonparticipatory innocent contribution. By law Cecilia and her son are considered
Under (most Western) law, Alex and Bob can be held liable for the Jackals
attempt to murder the innocent person; Cecilia and her son, however, cannot.
Let me now give examples from the context of war. Just army A invades
country B and kills collaterally many innocent people. Many soldiers do not
themselves shoot or drop bombs on innocent people, but they help other soldiers in
doing so, for example by maintaining the weaponry, supplying ammunition and
marking targets during the war. They participate in collective actions (whether they
actually causally contribute to any killing and mutilation or not) that, justified or
not, are nevertheless unjust since they kill, mutilate or unjustly endanger innocent
people. The participating soldiers are thus liable to attack. (Note, however, that I am
not claiming here that all soldiers of army A participate in unjust collective acts. In
17 However, if this happens on the civilians own behalf, for example in a humanitarian intervention that
endangers them while simultaneously trying to save them from genocide, this might not make the
combatants liable to attack. For an argument to this effect, see McMahan (2010, 342379).
18 For a legal account of complicity, see Smith (1991). For predominantly philosophical accounts, see
Kutz (2000) and Miller (2001).
fact, I think that many soldiers, on both the justified and the unjustified side, might
not be liable to attack).19
Entrepreneur E has delivered weaponry to army A before the war started (but not
since then), full well knowing that it would be used for the unjust collective actions
of the war. He would be liable to attack (if attacking him would help to diminish the
danger, which is of course highly unlikely).
Farmer F delivers foodstuff to the army. He is a non-participatory innocent
contributor and not liable to attack.20 The same holds for the foreign entrepreneur
EF, who before the war had delivered screws to the country of army A which are
now used by that army in some of its weaponry.
This account of liability is, I submit, intuitively quite compelling, coherent and
well-suited to the realities of law in Western jurisdictions, not least because it makes
use of familiar distinctions used in those jurisdictions. Admittedly, those
distinctions are not always clear cut; there are unclear cases and grey areas. Yet in many
cases we are quite able to distinguish between actual participation in an act and
guilty or non-guilty contribution. Case law and common sense morality (which do
not necessarily make these distinctions in these terms but make them nonetheless)
fare reasonably well in making these distinctions when it comes to the attribution of
Thus, on the account presented here, combatants who participate in unjust
collective actions that kill and mutilate innocent people or impose significant risks
on them21 violate the rights of those innocent people and are therefore liable to
attack even if they justifiably participate in those collective actions that (partly)
constitute the war.
These combatants are therefore also not innocent in the relevant sense. McMahan
(personal communication) disagrees and claims that if something is justified it
cannot be culpable or non-innocent. However, in German the terms guilt and
culpability are translated with Schuld, which in turn also means debt. Since, as
we will see, justification does not defeat liability, a person who justifiably wrongs
another person normally owes her compensation and is thus indebted to her. In
contrast to McMahan (it would appear), I find nothing odd about the idea that one
can incur guilt not only before the law or morality but also before another human
19 See also Lazar (2009) and (2010). Lazar, however, seems to accept McMahans account of liability to
attack according to which one becomes liable to attack by being morally responsible for an objectively
unjustified threat of harm. See for example McMahan (2009, 35). As we will see, however, that account is
not correct. Lazar also seems to think that causal contribution is necessary for liability. That is also
incorrect, as I have just argued.
20 The idea that one has to distinguish people who aid the soldier qua human being from people who
are aiding the soldier qua soldier or fighting man, to use Jeffrie G. Murphys (1973, 534) words, has
always been quite popular, and rightly so. Cecile Fabre (2009) has recently argued that it is not correct
that civilians who supply soldiers only with food are not liable to attack. I agree, since, as she correctly
points out, some ways of supplying soldiers with food are helping them as soldiers. However, this only
reinforces the distinction I have just endorsed instead of undermining it. Obviously, more would need to
be said here (but not for present purposes).
21 Again, this needs to be somewhat qualified, as stated in n. 17. A further qualifier, as will become clear
below, is that combatants who participate in a justified war are not liable to be attacked by those enemy
soldiers (if any) who participated in the aggressive act that provoked the justified war in the first place.
being. I do find it odd, however, to think that violating another persons rights,
whether justifiably or not, does not alter ones moral status. How can that be? Are
rights so trivial? McMahan offers the possibility of perhaps feeling remorse or
agent-regret; yet, feeling remorse if one is not somehow culpable or guilty of
something is simply irrational; and regret does not become better by putting an
agent before it. I can also regret that another person violates somebodys right;
but if I do it myself more than regret seems appropriate, namely indeed a feeling of
remorse. If this is to be rational, it has to correspond to guilt. McMahan might be
inclined to claim that all such feelings of remorse are irrational, but such a claim
would seem implausible and certainly unsubstantiated. Be that as it may, in the end
this might be partly a quibble about words. For McMahan justification is by
definition sufficient for innocence. My concept of innocence is different. Leaving
aside excuses, justification is sufficient for innocence only in conjunction with
justice. Thus, again, combatants who participate in unjust collective actions that kill
and mutilate innocent people or impose significant risks on them are non-innocent
and hence liable to attack.
4 Reply to Objections
McMahan claims that justification defeats liability so that just combatants are not
liable to be attacked by unjust combatants. However, that claim is true by
definitional fiat and hence irrelevant for the substantive question at issue.
McMahan gives the following definitions:
As I understand it, a just cause is an aim that satisfies two conditions: (1) that it may
permissibly be pursued by means of war, and (2) that the reason why this is so is at
least in part that those against whom the war is fought have made themselves
morally liable to military attack. With this notion as background, we can now
distinguish between just combatants, who fight in a just war, and unjust
combatants, who fight in a war that lacks a just cause. (McMahan 2009, 5)
Given these stipulations, however, the thesis of the unequal liberty-rights of just and
unjust combatants to kill each other or of their unequal liability to attack is not what
many people take it to be, namely a synthetic moral truth (or a synthetic moral
falsehood) but rather an analytically true sentence.
Why? Well, for example, we can formulate the thesis as follows (of course, there
are various possible formulations, but they make no difference to my argument here,
as long as they keep true to McMahans claims and definitions):
The moral inequality of combatants: Unjust combatants are liable to military
attack by the just combatants they fight against, while just combatants are not
liable to military attack by the unjust combatants they fight against.
And then we can proceed as follows: First we substitute, as McMahans definitions
allow, combatants who fight in a war that lacks a just cause for unjust
combatants and combatants who fight in a just war for just combatants. In a
second step we substitute, as McMahans definitions also allow, combatants who
fight in a war in which those the war is fought against have made themselves
morally liable to military attack for combatants who fight in a just war and
combatants who fight in a war in which those the war is fought against have not
made themselves liable to military attack for combatants who fight in a war that
lacks a just cause. We will then have transformed the thesis of the moral inequality
of combatants into this logically equivalent thesis (I have added UC for unjust
combatants and JC for just combatants to make it slightly easier to understand):
Combatants (UC) who fight in a war in which those (JC) the war is fought
against have not made themselves liable to military attack are liable to military
attack by the combatants (JC) they fight against and who fight in a war in
which those (UC) the war is fought against have made themselves morally
liable to military attack by the combatants (JC) they fight against, while
combatants (JC) who fight in a war in which those (UC) the war is fought
against have made themselves morally liable to military attack are not liable to
military attack by the combatants (UC) they (JC) fight against and who (UC)
fight in war in which those (JC) the war is fought against have not made
themselves liable to military attack.
While this thesis, being analytical, is true, it is also irrelevant, for given how
McMahan defines unjust combatants and given that justification does not defeat
liability, there simply are no unjust combatants.
For how could there be a war that lacks a just cause in McMahans sense? After
all, for a war to have a just cause it need not be fought for that just causeto claim
the contrary would be to confuse the criterion of just cause with that of right
intention. To be sure, McMahan might already be doing that to a certain extent since
he states that a just cause is an aim, and arguably something can only be an aim if
somebody did, does or will actually aim at it. However, the fact remains that there
can be the aim or just cause x (somebody can have that aim) without the alleged
legitimate authority or without all those (or even any of those) fighting the war
having that aim.
Therefore, of course, in every war there is a just cause (provided justification
does not defeat liability; I argue below that it does not), for example the just cause to
defend certain innocent people, in particular ones own innocent civilians, against
being killed or maimed by the enemy combatants.
Frowe struggles to see this as a just cause when the need for defence arises from
my own impermissible action (personal communication). However, first,
McMahans definition is entirely compatible with considering this as a just cause, and it is
McMahans definition I am discussing. Second, while the first aggressive soldiers
have acted impermissibly, it is simply question-begging to claim that their comrades
who later join the fray act impermissibly, too. One must not tar all combatants on
the unjustified side with the same brush. There are different unjustified soldiers, not
an amorphous mass called the unjust combatants. Besides, one must not ignore
collective action problems. An individual combatant joining the fray later cannot
reasonably be said to have provoked the justified war, nor can he reasonably be said
to be able to stop it by simply surrendering. (See on this also Sect. 5 below).
In addition, there is also the just cause of fighting excessive violence; or the just
cause of hindering the enemy from reaching certain grossly unjust goals (you can
have that aim even if the enemy does not have unjust goals himself in the first
place); or the just cause of saving innocent people in another country from brutal
suppression, etc. Any such cause (and there are a lot more) is certainly one that
satisfies two conditions: (1) that it may permissibly be pursued by means of war, and
(2) that the reason why this is so is at least in part that those against who the war is
fought have made themselves morally liable to military attack. If those the war is
being fought against violate the rights of innocent people, they are liable to attack
(provided justification does not defeat liability). Incidentally, many combatants who
fight in an unjustified war, for that matter, actually will be fighting for a just cause.
Conversely, in every war, including the justified or just wars, there is also an
unjust cause, for example (again, there are many more) the unjust cause of killing or
maiming large numbers of innocent people on the other side. The objection22 that
this is not a cause but a side-effect is mistaken. A cause, in McMahans account, is
an aim, and hence something you can want to achieve with a war, and of course
people can participate in or support a war because they want certain innocent people
to get killed or mutilated. This happens, for example, in extermination wars, but of
course people can have this aim in other wars too. Maybe McMahan would like to
claim that the aims of the combatants have to be distinguished from the aims of the
war. But it is his task to do this in his definition of just cause. So far he has not.
Besides, wars cannot literally have aims, only persons can. And this fact makes it
very difficult to avoid the point of this paragraph.
And many combatants on the justified side will actually be fighting for unjust
causesand even if they are not, they still objectively contribute to an unjust cause.
Thus, we again have a moral equality of combatants: there are just and unjust causes
and contributions and intentions on both sides. An unjust combatant in McMahans
sense, howeverthat is, a combatant who fights in a war that lacks a just causeis
nowhere to be found.
4.2 On Making Too Many People Liable to be Killed
Let us leave now the just combatants and unjust combatants behind and return
to the liability of combatants in a justified war. On the account of liability presented
above, somebodys violating another persons right is sufficient (but not, due to
symmetrical defense cases, necessary) to generate the right to necessary and
proportionate (or not grossly disproportionate) self- or other-defense.23 Since
innocents who are not posing an unjust threat are not liable to attack or to being
threatened by attack, initiating or sustaining such threats to them is unjust. By
initiating or sustaining such threats, by participating in activities that initiate or
sustain such threats, and, of course, by actually killing and maiming innocents one
would wrong them (and McMahan agrees). Therefore, in all those wars in which
combatants on the justified side are responsible for violating the rights of innocents
(namely by using their weapons in the relevant manner or threatening to do so, for
instance, or by protecting the flanks and rear of their killing comrades, or, quite
generally, by participating in a collective action that poses unjust threats to
innocents), they are liable to attack. It follows that the soldiers on the unjustified
side do not wrong combatants on the justified side when they kill them and thereby
protect innocent bystanders against an unjust threat. Moreover, there is no reason
why such attacks against just combatants could not be proportionate, given what
proportionality means in defense cases (see Steinhoff 2008, 224225).
McMahan, however, tries to undermine this argument by refuting two
assumptions he attributes to me:
In summary, Steinhoffs argument presupposes (1) that just combatants can be
liable to defensive attack by virtue of posing a threat of wrongful harm to
innocent people, even if the action by which they do so is morally justified,
and (2) that just combatants pose a threat in the sense relevant to liability not
only when their action puts innocent people at immediate or imminent risk of
wrongful harm, but continuously throughout the period in which they are at
war. (McMahan 2008b, 229)
Let me first point out that my argument does not presuppose the second assumption
at all. As I already stated in Sect. 3.2, I think that many soldiers on both the justified
and the unjustified side are not liable to attack. In any case, in order to defend the
moral equality of combatants (for most wars) in the sense discussed here, namely in
the sense of an equal liberty to kill the enemy combatant, it suffices to be agnostic
on the question whether just combatants are always liable to attack in war. To
establish equality it would be enough to show that if they are not, the unjust
combatants are not either.24 If both (or roughly the same percentage on both sides)
are not liable to attack, they have again an equal liberty-right to kill each other
(equality might consist here, of course, in none of them having a liberty-right to kill
the other). (In fact, even if the percentages were very different, nonetheless any
combatant on the unjustified side who has a liberty-right to kill combatants on the
justified one limits the scope of the general thesis of the moral inequality of
McMahan further specifies my alleged second assumption as follows:
Precisely stated, this is the assumption that in a war that exposes innocent
people to some risk of wrongful harm, all just combatants count, for the
24 McMahan (2009, 41) now recognizes this: In principle unjust combatants might also be liable only
when the threat they pose passes a threshold of seriousness, so that they might not be liable to attack while
asleep, or when stationed at bases remote from fighting. If so, we would still have a version of the moral
equality of combatants: one that treats just and unjust combatants as liable to attack, though only when the
threat they pose to innocent people passes a certain threshold of seriousness. Seth Lazar (2009)
elaborates on this.
purpose of assigning liability, as posing a continuous threat to the innocent,
simply by virtue of the fact that their belligerent activities might, at some
point, actively endanger innocent people.
And he then tries to turn the alleged counterintuitive implications of this against me:
Assuming that it has application outside the context of war, it implies, for
example, that all drivers are liable to preventive killing on the ground that the
activity in which they are engaged poses a lethal threat to innocent people.
(McMahan 2008b, 230)
This maneuver by McMahan is unsuccessful, since I do not make the precisely
stated assumption. My assumption, rather, is that people who participate in an
unjust (collective) attack or an unjust (collective) imposition of risk, and people who
guiltily contribute to such things, are liable to attack. By just driving around with
your car, however, you do not participate in an unjust (collective) attack or an unjust
(collective) imposition of risk (yes, you impose some risk, but below a certain
threshold the imposition of risk is not unjust), and you do not guiltily contribute to
such things either.
Thus, the account of liability I rely on does not have the counter-intuitive
implications McMahan wants to ascribe to it. His own account, on the other hand,
does have very counter-intuitive implications. Let me explain why: While
McMahan thinks that the ordinary driver is not liable to attack if all goes well,
things change if the probability that she will otherwise kill the pedestrian gets high
enough (for example, if, as a result of mechanical failure, her car is veering
uncontrollably toward him). She will then become liable to defensive killing. Her
liability derives from her earlier choice to drive, made in the awareness of the risks
she would impose, together with bad luck (McMahan 2008b, 231).
Note, however, that contrary to what McMahan seems to assume (remember that the
example is given in the context of a discussion of threats) the driver does not pose a threat
at all. In the example, she has lost control over the car. In fact, she could be removed from
the scene (Scottie might beam her up), and this would not make any difference to the
pedestrian at all. The car is posing the threat now, not the driver.
I have come across the objection that this is like saying that it is the bullet, not the
soldier, who is the threat. Indeed, it is, but only with respect to circumstances where
the bullet has already left the barrel and the soldier will not attack any further. Thus,
with respect to such circumstances it is also saying the right thing. Under
circumstances where the soldier will continue fighting, however, he is the threat. So
is the driver if, after missing the hiker, he tries to run him over again. But this is not
the situation we are facing in McMahans example.
Thus, the driver is at best a culpable cause in McMahan terms (see McMahan
1994a, 258).25 Thus if she is a taxi driver, she has the same moral status as a certain
tired hiker standing next to the pedestrian, for this hiker called the taxi that is now
lethally veering towards the pedestrian; and McMahan, after all, wrote immediately
following the last quotation: So voluntary engagement in an activity known to
25 I write at best. McMahan would actually regard her an innocent cause.
impose a risk of wrongful harm, even if the ex ante risk is very low, is a basis of
liability to defensive action if the probability of wrongful harm becomes
unexpectedly high (McMahan 2008b, 231). Yet, if driving is an activity known
to impose a risk of wrongful harm, so is calling a taxi (and thereby making someone
engage in the activity of driving). Consequently, on McMahans account the tired
hiker (and what about the persons who built the car, the mechanic who maintained
it, the man at the gas station who filled the gas tank?) would be liable to attack and
the pedestrian could justly throw him in front of the taxi if this would save his, the
pedestrians, life. This is, I submit, utterly counter-intuitive.26 Many more
counterintuitive examples could easily be offered. (To mention just one in the context of
war: You worked last year in a Guatemalan screw factory, full well knowing that
these screws can be used in weapons. Gaddafi does use some of the screws you
produced for this purpose. It seems that on McMahans account you now are liable
to attack by the Libyan freedom fighters if blowing you up in your living room in
Guatemala brought them a military advantage of some sort. McMahans account of
liability is far too permissive).
Ironically, McMahan provides one himself:
Suppose that the only way you can prevent yourself from being killed by a culpable
attacker is to kill his mother. If you do kill her, can you then claim that she was
liable to be killed because, as a morally responsible agent, she voluntarily chose to
engage in an activity (having a child) that had a tiny probability of resulting in an
unjust threat and that this made her responsible for the threat you faced from her
son? Obviously not. But it is less obvious what the right explanation is of why the
mother is not liable. (McMahan 2005b, 396)
This example, like my previous one of the tired hiker, is flatly a refutation of
McMahans account of liability to defensive attack.
McMahans vague assurance that the causal connections are not of the right
sort for the transmission of moral responsibility (McMahan 2005b, 396) can hardly
count as an answer to the request for an explanation.
Contrary to what McMahan thinks, however, the explanation of why the mother
is not liable is obvious: the mother is not liable to your defensive attack because she
does not violate your rights (nor does she pose an unjust threat to you). She is not
participating in the attack on you and she did not culpably contribute to the attack
(she did not bear the child full well knowing that it would kill you, and even if she
did know that, what was she supposed to dokill it on spot?).27 This is the answer
of the rights-based account of liability, and it is a very plausible answerone also
shared by law and common sense, I submit.
I conclude that McMahans account of liability to defensive killing is mistaken. It
is not correct that voluntary engagement in an activity known to impose a risk of
wrongful harm, even if the ex ante risk is very low, is a basis of liability to defensive
action if the probability of wrongful harm becomes unexpectedly high.
I just discussed one of McMahans attempts to burden me with certain unpalatable
implications of a certain assumption he attributes to me. That attempt failed, as we
saw, because I simply do not make that assumption. McMahan makes another
If Steinhoff were right and just combatants who foreseeably but
unintentionally threaten innocent civilians thereby make themselves liable to attack not
only by the civilians but also by third parties, these cases would instead be
asymmetrical in a way that seems highly implausible. If just combatants are
liable to attack by unjust combatants in these cases, and if there is no right of
defense against an attack to which one is liable, then just combatants have no
right of self-defense against unjust combatants. Yet if they were to attempt to
defend themselves, the unjust combatants would have a right of defense
against them. (McMahan 2008b, 241)
Fortunately for me, I do not accept that there is no right of defense against an
attack to which one is liable (if right here means, as it obviously does,
libertyright). Incidentally, in defining liability McMahan said that if the person to be
killed has acted in such a way that to kill him would neither wrong him nor violate
his rights, even if he has not consented to be killed or to be subjected to the risk of
being killed , I will say that the person is liable to be killed (McMahan 2005b,
386). Nothing in this definition implies that people who are liable to attack may not
defend themselves. On the rights-based account, an innocent victim (remember that
my concept of innocence does not apply to responsible agents who actively pose
unjust threats) does not wrong an innocent unjust threat by attacking this threat if
doing so is proportionate and his only means to defend himself against the threat.
Nor does the innocent unjust threat wrong the innocent victim if she tries to defend
herself against the victims self-defense, provided that the measures she uses are
proportionate and necessary.28 Thus, on the rights-based account they are both
liable to attack and counter-attack by the other (Steinhoff 2007a, 8889).
Recently, however, McMahan has dramatically changed his definition of
liability. He now claims:
What it means for a person to be liable to attack is that theres a substantial
moral asymmetry between him and those who might attack him. He has no
right not to be attacked, and is therefore not wronged by being attacked, while
the attackers retain their right not to be attacked.29
McMahan is begging the important and controversial question here as to whether
liability really involves a moral asymmetry. In any case, one cannot decide a
philosophical debate by definitional fiat.
The following remark of McMahans comes closer to being an argument:
Some have claimed that the bomber forfeits only his right not to be attacked
but retains his right of self-defense. I doubt that thats coherent.30
I know that this is incoherentif, that is, at least one of the rights mentioned is
supposed to be a claim-right. In that case, however, it is also certainly not a claim
I make. My claim about the symmetrical defense cases at issue (as already explained
above with the example of the falling man and his potential victim) is that both
parties have lost their claim-rights to life and to self-defense but retained their
liberty-rights to life and self-defense. There is nothing incoherent about that.
Thus, the account of liability and symmetrical defense cases presented here is
perfectly coherent, while McMahans new definition of liability is ad hoc (he
does not really explain why he suddenly changes the definition and gives no
argument why we should accept the new one) and question-begging (it simply
stipulates that liability implies asymmetry).
Even if we accepted it, however, that would still be irrelevant for the question of
the equal liberty-rights of combatants to kill each other, because McMahans
stipulative definition only implies that there cannot be symmetric liability to attack
but not that there cannot be symmetrical liberty-rights to attack (unless McMahan
wants to redefine the concept of a liberty-right, too).
Finally, let me briefly discuss the following example McMahan recently
challenged me with (the choice of names is mine).31
Suppose Anthony unjustly attacks Barbara and her two innocent friends.
Barbara is about to defend herself by throwing a grenade, the only defensive
option she has under the circumstances. This defensive action, however,
would also kill an innocent bystander, thus wronging him.
Would not Barbara thus become liable to attack, on my account, so that now
Anthony can actually kill Barbara justly?
No, she does not become liable to attack by Anthony (which, however, does not
mean that Anthony cannot permissibly attack her; it only means that he would
wrong her if he attacked her). After all, Anthony, not Barbara, initiated the unjust
attack (as do the first attacking soldiers, whether they are participants in an
unjustified war or a justified one, so my concession here does not weaken my claim
as to the severe limitations of the moral inequality of combatants32), he forced
Barbara to choose between the lives of herself and her two friends on the one hand
and the life of an unjust attacker and an innocent bystander on the other. He had no
30 See note 29.
31 See his comments on my talk at the War and Self-Defence conference in Sheffield.
32 Incidentally, even if the first attacking soldiers on the unjustified side are not moral equals to the
defending soldiers, this does not change the fact that the other soldiers on the unjustified side are (if their
enemies endanger innocent people on the unjustified side), which sets extremely severe limits to the
generalizing thesis of the moral inequality of combatants.
liberty-right (and, of course, no claim-right) against her to attack her in the first
place, and therefore he cannot now have a liberty-right against her to violently
interfere with her self-defense against the unjust attack.33 If he does, he is violating
her claim-right to life and to self-defense against him. (The structure here is that of
self-defense against unjustly provoked self-defense).
Barbara, however, would become liable to attack by some other people, namely
by those specially related to the innocent bystander. And even Anthony would still
be permitted to kill (albeit unjustly) Barbara if this would be the lesser evil or even if
it would merely be justified under necessity,34 for example if the side-effects of
Barbaras self-defense would be too severe (perhaps she would kill several innocent
bystanders). In other words, that a person Y is liable to be killed by another person
X does not yet mean that X can justifiably kill Y; and that Y is not liable to be killed
by X does not yet mean that X cannot justifiably kill Y. Liability is not the
alldecisive factor for permissible killing (see also Sect. 5).35
4.4 Justification Does Not Defeat Liability
McMahan claims that justification defeats liability, and that therefore justified
soldiers cannot be liable to attack. As far as the moral (in)equality of combatants is
concerned, this is a question-begging argument, for it only establishes the
inequality of combatants on the assumption that combatants on the unjustified
(collective) side cannot, individually, justifiably participate in the war. However,
that is not only far from obviousindividual participation in an unjustified
collective act can clearly very often be justifiedbut it is also precisely the
controversial issue. Thus, even if justification defeated liability, this would not be
sufficient to undermine the thesis of the moral equality of combatants.
Yet I will not follow up on this issue here.36 Instead, I will now argue that
justification does not defeat liability. To set the scene, allow me to first point out that
McMahan is well aware of the fact that his claim that combatants in a justified war
are not liable to attack might intuitively be much less convincing than a defender of
his position would wish. He says:
The claim that justification defeats liability may seem more compelling in the
case of justified acts that are morally required. If so, and if acts of war by just
combatants that kill innocent people as a side effect are optional rather than
required, this would weaken the case for the claim that those who engage in
such acts do not thereby become liable. I believe that there are wars that
countries are morally required to fight And there may be others that, while
morally optional, are nevertheless justified in the strong sense that they are not
only permissible but also impartially better than any alternativethat is, there
This, it seems, is nothing less than the admission that in all normal wars, even in
normal just wars, the case for the claim that those who engage in such acts do
not thereby become liable is weakened. But with this admission the case for the
claim that McMahans thesis of the inequality of combatants has great practical
significance is also weakened, and very much so. How many wars are there that
countries are morally required to fight? Is there any such real war? Maybe (though
even that is not entirely clear, for a variety of reasons) Great Britain was required to
fight a war against Germany, but it was certainly not required to wage the
indiscriminate war it actually waged (Anscombe 1981, 7281, esp. 73). And that
actual war, quite obviously, was also not impartially better than any alternative.
Besides, the issue here is not the war as a whole but the justification of the
individual, allegedly just combatants. Has there ever been any act by just
combatants that has been impartially better than any alternative? How likely is that?
More likely than the same happening on the other side? And how does one find out?
If just soldiers are only then not liable to attack when all their acts of killing or
contributing to killing innocents are impartially better than any alternative, the
inequality of combatants has for all intents and purposes gone overboard.
Moreover, the claim that justification defeats liability to defensive killing is ad
hoc, anyway. McMahan denies this by pointing to the fact that criminal liability is
defeated by justification. He is right, of course, but that is irrelevant. The issue was
and is not liability to punishment, but liability to defensive killing, as he himself
acknowledges (McMahan 2008b, 234) (which makes it odd that he discusses
criminal liability at all).
He also mentions tort law and explains that strict liability is the only kind of
liability in either criminal or tort law that is not defeasible by a justification, and it
governs only a very limited domain of the law of torts (McMahan 2008b, 233).
However, this statement is a tautology. Strict liability is defined as liability that is
not defeasible by a justification or an excuse. The more interesting question to ask,
therefore, is whether in tort law justification always defeats liability. The existence
of strict liability shows that it does not which proves my point.
Moreover, whether the domain of strict liability is limited or not, the fact
remains that it is exactly the domain under discussion here. In fact, only a few lines
before he banished strict liability to a very limited domain, McMahan discussed
Joel Feinbergs case of the hiker in a snowstorm who, in order to avoid freezing to
death, breaks into an unoccupied cabin and makes a fire by burning some furniture.
He infringes the owners rights and owes the owner compensation, but the owner
would have no moral right of defense against the trespasser (McMahan 2008b,
233). Now, this example shows that justification (the hiker was justified according
to McMahan) does not defeat liability.
McMahan, however, offers this example to undermine another objection: in the
case where, for instance, just combatants bomb and kill innocent bystanders in
the course of a proportionate attack on a military target, the just combatants
would (from a moral point of view) ex post have to pay compensation to the victims
of their attack, while the potential victims, if they harm the attackers through
defensive action, would not have to pay compensation to them. This, I think, shows
that there is a crucial difference between the two parties there is certainly no
Yet, McMahan objects to this that one cannot infer that a person is liable to
defensive action ex ante from his being liable to pay compensation ex post
(McMahan 2008b, 233). That is correct; one cannot infer this without a further
premise (and one does not have to). However, in tort law the fact that A has to pay
compensation to B means that A has wronged B. Thus, the factif it is a fact, and I
think it isthat the innocent victims of the bombers can demand compensation for
the mutilations and losses they have suffered from the bombers while the bombers
cannot demand compensation for the losses and harms they have suffered due to the
defensive action of the innocents, shows that a moral asymmetry is at play here. The
bombers have wronged the innocents, and not vice versa. Seeing that McMahan
thinks that even differences of comparatively slight moral significance can make
all the difference in who is liable to be killed (McMahan 2005b, 394), it is unclear
why this should not be the case hereespecially since the difference is not slight at
In that context, let us revisit Feinbergs example. McMahan claims that the
owner would have no moral right of defense against the trespasser. Well, leaving
aside the moral question for a second: from a legal point of view it is actually not
that clear that the owner does not have a right of defense against the trespasser
(Christie 1999, 1008). Moreover, if the backpacker by his actions threatens the life
of the owner (maybe the owner cannot survive without the furniture), the owner
definitely has the legal and moral right to defend himself with lethal violence, if
need be. And the backpacker would under those circumstances not be allowed to use
lethal countermeasures in turn against the self-defense of the owner. That is the
legal situation at least in US law (which is the law to which McMahan refers). I
think it is also the correct stance from a moral point of view.
What we are dealing with here is a case of so-called private necessity. This is
important since McMahan now explicitly argues that the justification of the tactical
bomber is a necessity justification (and of course it is).37 However, in law the
necessity justification definitely does not defeat liability.
In law, the defense of private necessity allows a person to damage or destroy or to
trespass anothers property in the course of self-help efforts that are reasonably
necessary to save the actors clearly more valuable property and/or his or her own
life (Sugarman 2006, 56). However, most US statutes do not allow this legal
defense in cases where the self-helper threatens the life of another person. In such
cases, the self-helper becomes liable to attack. Nevertheless, following the Model
Penal Code, the statutes of some other US states seem to allow deliberately killing
an innocent and non-threatening person in order to save many others. Still, this same
Model Penal Code makes clear that this privilege does not abolish or impair any
remedy for such conduct that is available in any civil action,38 which means, as
37 McMahan, Self-Defense Against Justified Threats, p. 2.
38 Official Model Code, Sect. 3.1, as quoted in Christie (1999, 1026).
George C. Christie (1999, 1026) points out, that a person killing another innocent
person out of necessity would be liable in tort for substantial damages in a
wrongful death action brought by [the victims] next of kin. It seems, however,
that if the potential victim killed the self-helper, such a wrongful death action
could not, for good legal reasons, be brought against the potential victim (Christie
1999, 10341039). As Christie (1999, 1039) further notes: If any of the parties
would be free from tort liability, it would be the [innocent potential victim of a
necessary attack]. I cannot conceive of any American court holding an innocent
person liable in tort for shooting another person to prevent that other person from
Frowe, however, objects that if we are discussing unjust combatants the
proper analogy would be a different one, and states that she cannot conceive of a
court finding you liable for the death of one person by diverting the trolley away
from two hundred towards the one when I tied all those people to the track and set
the trolley in motion (personal communication). In response, let me note, first, that
when we discuss just combatants the proper analogy is not to diverting existing
threats (like trolleys already set in motion by someone else) but to initiating
completely new ones (like dropping bombs). And I can very well imagine a court
finding you liable for wrongful killing if you blow up one hundred innocent
people when this is the only way to keep me from killing one thousand. Second,
Frowe again tars all unjust combatants with the same brush (see my discussion
of this above). Some of them have not tied the innocent people to the track but
joined the fray later. And the question is whether they would be held liable if they
kill you in order to defend themselves or people to whom they have special
responsibilities from your attack.
Thus, US tort law takes it to be the case, and very reasonably so, that a person
who kills another innocent person out of necessity wrongs this innocent person,
while the innocent person killing the attacker does not wrong the attacker. But this
then means, both on McMahans previous definition of liability to be killed as well
as on his current one, that the first person must be legally liable to be killed, while
the second is not. After all, in defining liability McMahan once explained that if
the person to be killed has acted in such a way that to kill him would neither wrong
him nor violate his rights, even if he has not consented to be killed or to be subjected
to the risk of being killed I will say that the person is liable to be killed
(McMahan 2005b, 386), and now says, to repeat:
What it means for a person to be liable to attack is that theres a substantial
moral asymmetry between him and those who might attack him. He has no
right not to be attacked, and is therefore not wronged by being attacked, while
the attackers retain their right not to be attacked.39
This is the situation in US law as it concerns attacking an innocent person out of
necessity: the attacker is liable, the innocent person is not. Thus, when it comes to
the question whether US tort law and US self-defense and necessity statutes
39 McMahan, Self-Defense Against Justified Threats, p. 5.
consider justification to defeat liability against self-defensive killing in cases of
attempted homicideand these are the cases we are talking about when the issue is
warUS tort law and US self-defense and necessity statutes clearly support my
position. This is so not least of all because it is also a morally very plausible
Incidentally, McMahan not only has the law against him, but, it seems, his earlier
self as well:
Of course, a moral asymmetry remains between the Innocent Threat or
Attacker and the Innocent Victim: for an Innocent Threat or Attacker is, by
definition, morally responsible for the threat he poses, and may also be at fault
(though excused) for that threat. And these considerations are relevant to how
the inevitable misfortune ought to be distributed as a matter of justice.
(McMahan 2003, 412)
Indeed, they are. Therefore it is somewhat surprising that McMahan now suddenly
states that in a conflict between those who act at the behest of morality and those
who are wholly innocent and act in self-preservation [again, the correct expression
is self-defense], justice is silent (McMahan 2008b, 243).
One might be tempted to object here that there is nothing ad hoc or surprising
about this. After all, in the indented quote McMahan discusses innocent attackers or
threats while in the last quote he obviously refers to justified attackers. Yet, the
problem is that contrary to what McMahan suggests, only those who are morally
required to attack act at the behest of morality and not those who are merely
justified in attacking.40 A combatant can be justified in engaging in a certain attack
even if all things considered it would have been morally better to do something else
(maybe dedicate the same amount of energy and money and resources to Oxfam). If
even under those circumstances he attacks anyway, then however justified he might
be he is still morally responsible for the morally suboptimal choice of posing an
unjust threat. The innocent bystander, however, is completely free of such
responsibility. And these considerations, I think, are indeed relevant to how the
inevitable misfortune ought to be distributed as a matter of justice.
I conclude that McMahans claim that justification defeats liability in the legal
cases relevant for the present discussion is wrong. In addition, his claim that
justification morally defeats liability is and remains ad hoc and implausible.
5 Liability, Permissibility and Strategies of Other-Defense
I have argued elsewhere that while combatants in an unjustified war collectively
could perhaps best protect their civilian population by ceasing to fight, no individual
such combatant can stop the war on his own.41 I therefore suggested that in most
circumstances the best way for each combatant to defend innocent bystanders on his
side is to continue to fight (Steinhoff 2008, 222). McMahan responds that even if
this were true, it would not be obvious what an individual unjust combatant ought
to do since there seem to be instances in which people ought not to participate in
some activity even if it will be worse for some innocent people if they do not
(McMahan 2008b, 243). Yes, there might be such activities, but McMahan certainly
has not shown that all unjustified wars are such activities. To simply assume they
are is to beg the question. (And incidentally, could not just wars also be such
activities?pacifists say they are).
McMahan also says that my suggestion is not obviously true (McMahan
2008b, 243). That might be so, but his counter-claim is not obviously true either:
since it seems that unjust combatants ought not simply to fight on one
occasion after another without considering whether this has the effect of
prolonging the war in a way that is counterproductive, and since it seems that
they ought not to aim at victory, perhaps what they should aim for is the early
defeat of their own side. This may be the option that would best protect their
own civilians from injury and death caused by the military action of just
combatants. (McMahan 2008b, 244).
Let me note first that even if the strategy McMahan describes were
counterproductive, this would not change the fact that the combatants in a justified war who
threaten innocent bystanders are liable to attack. As I noted before: that someone is
liable to be killed does not necessarily mean that he can be killed permissibly; and
that someone is not liable to be killed does not mean that he cannot be killed
Consider the village sentinels who face a threat by a certain sect living in the
jungle who want to sacrifice (justifiably, let us assumemaybe otherwise huge
numbers of other children would be tortured to death by extraterrestrial villains) two
children of the village each year. It is well known among the village people that
resisting the intrusions of the sect will only make them come back again and again,
killing more than just two children in the course of the different ensuing battles. It
might be counter-productive not to hand over the two children at once and to allow
the sect-members to cut their throats, but that does not change the fact that the
throat-cutters would still be liable to attack. If the sentinels resisted the intrusion
they might thereby do a bad job (I am actually not entirely sure about this, but will
leave this point open here), they might perhaps even thereby wrong those additional
village members who will be killed in future raids (but this would obviously also
depend on which strategy the civilians prefer) or act otherwise impermissibly, but
they definitely do not wrong the intruding throat-cutters. Similarly, the soldiers on
the unjustified side do not wrong the soldiers on the other side by killing them on
one occasion after another in defense of their own civilians. The soldiers on the
justified side remain liable to attack, and thus the soldiers on the unjustified side
have a liberty-right against them to kill them.
42 Although McMahan realizes this fact, he does not pay sufficient attention to it. While his book is
called Killing in War, it actually deals for the most part only with Liability to Be Killed in War.
Second, whether or not combatants who participate in an unjustified war may aim
at victory depends on how they conceptualize victory. If they conceptualize it as the
achievement of an unjust cause, perhaps they must not aim at itbut the same
would be true for the combatants on the justified side. However, they certainly may
aim at it if they take victory to be the other sides ceasing to kill innocent civilians.
Third, by laying down their arms, soldiers on the unjustified side might actually
promote unjust goals. For example, if the just side confronts an enemy who easily
gives up, they might think: This war is really easy. If it had been more difficult, it
would have been prudent not to press for the achievement of certain unjust goals in
addition to our just ones because our tough enemies would have made that very
difficult; but now it seems there is no real problem here Indeed, in many wars
the combatants on the unjustified side are entirely justified in distrusting the avowed
just intentions of the just side and in acting accordingly. By fighting on, they
might contribute to the other sides justified war remaining justified.
Fourth, that a justified war is fought in a proportionate way by the combatants
waging it does not mean that the havoc they wreak cannot be further minimized.
McMahan, however, claims:
Acting to defend some innocent people on one occasion after another is
counterproductive if the larger effect is to continue to prolong the war
indefinitely, thereby repeatedly sustaining the threat to other members of the
civilian population who have so far survived. (McMahan 2008b, 243).
For many situations, this is simply wrong. A boxer might also end a fight more
quickly by standing still and lowering his guard, but ending a fight quickly and
ending it without brain damage are not the same thing. Similarly, the soldiers on the
unjustified side do not automatically have to trust that the soldiers on the justified
side will not exploit weaknesses in the cover to wreak even more destruction.43
The facts that (a) soldiers on the justified side who kill innocents are liable to
attack, (b) soldiers on the unjustified side will in many circumstances be entirely
justified in distrusting the just side and in acting accordingly, and (c) soldiers
participating in an unjustified war will in many circumstances be able to reduce the
harm done to their own population by fighting back, show that McMahans
generalization that unjust combatants must not attack soldiers fighting in an just war
(unless they commit war crimes) is mistaken.
6 Concluding Remarks
As I have said already in the introduction, the claim that all combatants who abide
by ius in bello restrictions have an equal liberty-right to kill each other is wrong.
There are some wars in which even those combatants who abide by ius in bello
restrictions do not have an equal liberty right to kill each other. Nevertheless, the
argument presented in this paper shows that the moral equality of combatants ends
up holding true of many more wars than allowed for by McMahans general thesis
about the moral inequality of combatants. In addition, and on a more theoretical
point, it also has to be emphasized that the obsession with liability obscures the
fact that whom you may kill in war and why is not reducible to the question of who
is liable to attack and who is not.44
In any case, whether or not soldiers are allowed to participate in an unjustified
war cannot be decided without a closer look at the details and circumstances of the
specific unjust war in question. While there is no general permission to fight in
unjustified wars, in some unjustified wars you can fight permissibly and, as long as
you abide by ius in bello restrictions, you can do so without wronging enemy
combatants. However, in modern wars you will always wrong innocent bystanders;
but so will the combatants on the justified sidewho therefore cannot be just.45
Open Access This article is distributed under the terms of the Creative Commons Attribution License
which permits any use, distribution, and reproduction in any medium, provided the original author(s) and
the source are credited.