Notes on Terrorism and Intention
Note: I apologize for the rough state of this material. What follows is only some slightly
tidied up lecture notes. I had hoped to have time to get this material into a more
polished state, but have been unable to do so. Some of the material (in gray font) is
mainly just notes to myself. Sorry. Jeff McMahan
What is terrorism?
Terrorism is the intentional use of violence against innocent people as a means of
intimidating and coercing other people associated with them, usually for political
If this is right, it’s intention that relevantly distinguishes legitimate acts of war that kill
innocent people from acts of terrorism.
Illustration: the tactical bomber and the terror bomber
Munitions factory. A civilian area right next to it. We have just cause and destroying the
factory will make a significant contribution to victory. Because of this the inevitable
killing of civilians is proportionate. But killing the civilians would be comparably
effective as a means of achieving the just cause, by terrorizing and intimidating the
Tactical bomber drops a bomb on the munitions factory foreseeing that this will kill the
civilians but not intending to kill them.
Terror bomber drops a bomb in exactly the same spot because creating a big explosion by
detonating the munitions in the factory is the most effective way to kill the civilians.
He destroys the munitions factory as a means of killing the civilians, as a means of
achieving the just cause.
The difference between the two acts is in the intention only. If an agent’s intention in
acting doesn’t affect the permissibility of his act, it’s hard to find a moral difference
between acts of war that kill innocent people as a side effect and acts of terrorism that
kill innocent people as an intended means.
The difference isn’t in the choice of ends. Terrorist tactics can be used to achieve just
ends. Terrorism is a crime of means, not ends. In the examples, the tactical bomber
and the terror bomber have the same just end.
Arguments against the relevance of intention to permissibility
Many distinguished moral theorists have argued in recent years that the intention with
which an agent acts is irrelevant to the permissibility of the action.
(1) Reasons given by facts about the world, not by facts about the agent’s subjective
Many people are reluctant to accept that what goes on in a person’s mind can make a
difference to whether his action is wrong. The dominant view now is that the reasons
one has for or against acting are given by facts in the world beyond one’s own mind.
Thomson’s argument. Suppose one is deliberating about whether to bomb the munitions
factory. And suppose that objectively one has sufficient reason to bomb the factory.
She thinks it would be absurd to think that, before deciding whether one would be
acting permissibly in bombing the factory, one would have to engage in introspection to
see what one’s intention would be in dropping the bomb. “What a queer performance
that would be!”
There’s certainly something to this. Suppose that I have a good intention in doing an act
that’s otherwise wrong. Can my intention alone make the action permissible? It’s hard
to believe it could.
But consider another subjective state: belief. Suppose that I reasonably believe that a
certain act is permissible. That may be sufficient to make it permissible, in one
important sense of permissible, when it otherwise wouldn’t be – for example, when the
other beliefs that I reasonably have that support my belief that the act is permissible are
in fact false.
(2) Thomson’s Alfred and Bertha case
Alfred plans to administer some stuff to Bertha that he believes is poison. But the stuff is
actually the only stuff in the world that can cure her otherwise fatal condition.
Thomson asks: “Is it permissible for Alfred to give it to her? Surely yes.” She claims
that it is irrelevant to the permissibility of his giving her the stuff that he acts with the
intention to kill her.
I claim: it’s permissible for him to give it to her provided he does so for an acceptable
I think it’s plausible to say (1) that it’s impermissible for him to give her the stuff with
the intention of killing her, (2) that it’s impermissible for him not to give her the stuff at
all, and (3) that it’s therefore permissible for him to give her the stuff with an
acceptable intention, preferably the intention of saving her life.
But Thomson stipulates that “he will give it to her only with the intention to kill her.” So
if it’s permissible for him to give it to her, it must be permissible for him to give it to
her with the intention of killing her.
But Alfred’s options aren’t just giving it to her with the intention of killing her and not
giving it to her at all. What it’s permissible or impermissible for him to do doesn’t
depend on what he’s in fact willing to do. We don’t have to grant that it’s permissible
for him to do the best of all the acts he’s willing to do if in fact there’s an act that he
ought to do instead that he won’t do. That he won’t do what it’s obligatory for him to
do can’t make it permissible for him to do the next best thing.
(3) Third party obligations
Still, there’s something in Thomson’s insistence that it must be permissible for him to
give her the stuff rather than allow her to die, no matter what his intention is. This may
be because we want him to give it to her, even if he does it intending to kill her.
Thomson says that it’s natural for us to want this but that if intention matters we ought to
be ashamed to want him to do it.
But when the reason why a person’s act would be wrong isn’t that the effects would be
bad but that his intentions or motivating reasons are wrong, it can be reasonable for
third parties to want and indeed permissible for them to encourage him to act wrongly,
if the effects of his doing so would be desirable.
Compare the commander who uses a terror bomber to achieve a tactical goal. Suppose a
military facility is a legitimate target in a just war but that it is right next to a children’s
hospital. And suppose that the best pilot, or perhaps the only available pilot for the
mission, is a person known to believe that the best way to win the war is to bomb
civilian targets. He’s also known to enjoy killing children. If he undertakes the
mission, his intention in destroying the military facility will be only to kill the children
win the war through terrorism.
It’s plausible to say in this case that if he drops the bomb with these intentions, he’ll be
doing wrong. But the commander doesn’t intend the deaths of the children or the
causation of terror. He does no wrong in exploiting the pilot’s wrongdoing. If the
wrong is only in the intention, then only the one who acts with that intention does
Reductive explanations of the relevance of intention
(1) Thomson: intention relevant to evaluation of the agent
(2) Kamm: not subjective states but objective causal relations
The traditional view is that an act with a bad effect is wrong when that effect is intended
as a means. Kamm says that what matters is not that the effect is intended but that it is
causally related to the good effect as a means.
Her view seems to permit both the tactical bomber’s act and the terror bomber’s act if the
killing of the civilians is a consequence of the destruction of the factory, but condemns
both if the civilians are between the explosion and the factory, so that they are killed on
the causal path to the destruction of the factory.
(3) Scanlon: we conflate evaluation of action with evaluation of deliberation
Tim Scanlon thinks that our fundamental mistake in believing intention to be relevant to
permissibility is that we confuse the assessment of the considerations that are relevant
to the permissibility of action with the assessment of the way in which an agent
deliberates about those considerations.
But nothing follows about whether facts about the way in which reasons figure in a
person’s moral deliberations can affect the permissibility of his action. It’s not absurd
to think that defective deliberation about reasons can contaminate action, or that action
done for a wrongful reason can be wrong precisely because the motivating reason
conflicts with the normative reason.
(4) Walen: implicit commitment to objectively wrongful action
Intention is relevant in that, although a bad intention may not make present action wrong
if there’s an objective justification for the action, it may be wrong to act with that
intention because it commits or disposes one to act in an objectively wrong way in
Trilemma: if intention is irrelevant to permissibility
If intention is irrelevant to permissibility, there are three broad possibilities:
(1) Accept terrorism: We can retain the view that the proportionate killing of innocent
civilians as a side effect of military action is permissible and accept that the intended
killing of a comparable number of civilians as a means of achieving a comparably
important goal is also permissible.
(2) Pacifism: We can reject terrorism in almost all instances and also reject acts of war
that would cause comparable harm to innocent civilians as a side effect.
(3) Find alternative difference: We can find an alternative way of distinguishing
morally between terrorism and legitimate acts of war that kill civilians as a side effect.
Different moral philosophers have accepted different options.
(1) Embrace terrorism
Jonathan Bennett: “I have been arguing that what the terror bomber does may be morally
all right, but I do not accept that it could be all right for him to behave in this manner
without compunction, without considering the cost to the civilians, without looking for
less lethal alternatives. All of this holds equally, of course, for the tactical bomber.”
This looks as if Bennett is sternly insisting on imposing some stringent constraints on
terrorism. But what are the constraints? Terrorists have to be sensitive and
conscientious, reflect about what they propose to do to civilians, and make sure that
there isn’t some way to achieve their goals that would be less harmful to their victims.
(2) Embrace pacifism
Some moral theorists take the rejection of terrorism as a fixed point and therefore end up
with a view that’s tantamount to contingent pacifism. One strategy is to try to get
proportionality to take over the work of intention. That is, instead of weakening the
constraint on intentional harming (as Bennett does), these theorists strengthen the
constraint against foreseen but unintended harming.
David Rodin, for example, argues that tolerance for harming the innocent as a side effect
in war should be no greater than it is for harming the innocent as a side effect in
domestic police work. But if the proportionality constraint is strong enough to support
a general prohibition of terrorism, it’ll also be so strong as to make most acts of war
disproportionate as well.
(3) Thomson: sort of embracing terrorism while sort of rejecting it
Judy Thomson isn’t a pacifist. She believes that it can be permissible to resort to war
even when this inevitably involves killing the innocent as a side effect. She accepts, for
example, that it can be permissible for a tactical bomber to kill the innocent as a side
effect of destroying a military target. Yet in general she thinks that action that kills an
innocent bystander as a side effect can’t be permissible.
A runaway trolley will kill you unless you blow it up, but blowing it up will kill an
innocent bystander next to it. Thomson says that self-preservation is impermissible in
this case. Yet she concedes that it has the same structure as the case of the tactical
bomber. How, then, can the tactical bomber’s action be permissible? It’s partly, she
says, because “there’s a war on.” How’s that relevant?
One consideration is that the “stakes are higher.” It’s not 1-for-1. In the case of the
tactical bomber, assume that the just cause is much more important than the lives of the
civilians who’ll be killed. War is therefore more like a case in which 5 will be killed by
the trolley unless it’s blown up, killing one bystander next to it. Presumably Thomson
accepts that as permissible, since she accepts the permissibility of diverting the trolley
in the standard trolley case.
Introduce intention: shove the fat man
So one difference between Trolley preemption and the tactical bomber is proportionality.
But this can’t solve the larger problem. For we can imagine a variant of the trolley case
in which the numbers are the same but in which the killing of the one would be
intentional – for example, 5 will be killed by the trolley unless one innocent person is
shoved into its path to stop it. Most people think this would be wrong (including
Thomson, I think, because she thinks that it’s wrong to kill one to save 5 in Transplant).
Thomson’s view supports terrorism
So Thomson’s view implies that however many the tactical bomber may permissibly kill,
the terror bomber may permissibly kill that many as well. So she ends up committed,
like Bennett, to accepting the permissibility of terrorism in a great many situations.
Terrorism permissible within war but not in peace?
Thomson’s view may have a curious implication about terrorism that she doesn’t notice.
Suppose it’s really something unique about conditions of war that makes it permissible
knowingly to kill the innocent as a side effect. (It’s the appeal to conditions of war that
is supposed to explain why the action of the tactical bomber is permissible when killing
an innocent bystander in Trolley Preemption isn’t. Suppose this isn’t just a matter of
numbers and hence of proportionality.) Since the conditions of war (whatever it is
that’s special about them) don’t obtain during peacetime, her view seems to imply that
terrorism can be permissible in war but not outside the context of war.
This view has in common with realism the idea that conditions of war alter what morality
requires. Realists say that war renders morality inapplicable. Thomson implies it
weakens the constraint on harming the innocent.
Yet the only reason Thomson gives for thinking that war is different is that the stakes are
higher. We can’t evaluate this suggestion until we have an explanation of why war is
In the final paragraph of her discussion of intention in the paper on self-defense, she says
that she “will have to bypass as too hard the question how the fact of war affects
questions of self-defense. … we here bypass the question what makes it permissible for
the pilot to drop his bombs in [Tactical] Bomber, and impermissible (if it is) for him to
do so in Terror Bomber.” Here she suggests that conditions of war explain both why
it’s permissible for the tactical bomber to kill civilians as a side effect even though
killing an innocent bystander in Trolley Preemption isn’t, and why it may be
permissible to kill innocent people as a side effect but not for terrorist purposes.
By inserting the parenthetical phrase “if it is,” Thomson manages to suggest that
conditions of war may even have the alchemical effect of making intention relevant to
permissibility, thus making terrorism impermissible even in war.
One can see how badly she wants to evade the clear implication of her view: that
intentionally killing innocent people in war is permissible whenever killing them as a
side effect of military action would be.
In the end, Thomson has nothing to say that could distinguish her position from Bennett’s
candid embrace of terrorism.
(4) Scanlon: what matters is presence or absence of a military target
Scanlon tries to preserve common sense intuitions by arguing that the moral difference
between the tactical bomber and terror bomber is found in an account of the acceptable
reasons that can justify exceptions to the prohibition of knowingly killing innocent
He takes as given a familiar account of the morality of permissible killing in war. “In a
war,” he says, “one is permitted to use against one’s opponents destructive and
potentially deadly force of a kind that would normally be prohibited. But this is
permitted only when its use is justified by a military objective, and only with the
provisions that one takes sufficient care to minimize harm to non-combatants and that
the harm can be expected to occur despite this care is ‘proportional’ to the importance
of the military objective.”
Objective justification: there must be a military objective. That can justify killing
innocent people. But winning the war isn’t an objective that’s sufficient to justify
killing innocent people.
The reason that terror bombing – pure terror bombing when there’s no military target at
all – isn’t permissible isn’t that anyone would act with a wrongful intention. It’s
instead that there’s no legitimate target. “It’s impermissible because it will kill people
and the circumstances are not such as to provide a justification for doing this – that is to
say, not such as to bring the case under any exception to the prohibition against doing
what can be reasonably foreseen to cause loss of life.”
But tactical bombing can be permissible because there’s a legitimate target. “In the case
where there is a military target … and the raid that would destroy it would also kill
enough people to demoralize the country, this raid is permissible just in case it is
permissible considered as a raid on the [military facility] alone.”
But what’s so important about the presence of a military target? Suppose we could win
the war by intentionally killing 1000 innocent civilians. If there’s no military
advantage, that’s impermissible on Tim’s view.
But now suppose that we can also win the war by killing 1000 enemy combatants, who
happen to be intermingled with 1000 different innocent civilians – so that attacking the
combatants would kill all 2000. Since killing these 1000 combatants would enable us
to win the war by military means, the killing of the 1000 civilians would be
So Tim’s view implies that it’s permissible to kill 1000 civilians along with 1000
combatants but not permissible to achieve the same goal by killing 1000 civilians only.
The view that intention is relevant to permissibility may have the same implication, but
this suggests that Tim’s view doesn’t have the advantage of greater plausibility. Why
not just stay with the traditional explanation?
Thoughts on Tim’s new draft
Tim’s position seems to require a special morality for war. Here the killing of innocent
people has to be justified by reference to military advantage. (This presupposes that
war is different because there’s no other area of life in which military advantage has
any significance.*) Is this a different morality or just one possible exception to the
prohibition of killing the innocent – one that arises only in conditions of war? Military
advantage in achieving a just cause can justify the minimal and proportionate killing of
innocents, but achieving the just cause in a way that kills the same number of innocents
but not in the course of achieving military advantage cannot – even, presumably, when
the alternative would kill fewer innocent people. Is attributing moral significance to
military advantage more intuitively plausible than attributing moral significance to
intention? This is the point about reflective equilibrium.
On Tim’s view, killing innocent people is permissible if it is done for military advantage,
or in the course of gaining military advantage, but not if it is done as a more direct
means of achieving victory in a just cause. It can be an exception to the prohibition of
killing innocent people that killing innocent people is necessary for military advantage
in achieving a just cause, but not that killing innocent people is necessary for
demoralizing a population in achieving a just cause.
Why is it plausible to suppose that the permissibility of killing innocent people can
depend on whether one is using military or nonmilitary means to achieve a just cause,
but not plausible to suppose that the permissibility could depend on whether one
intends to kill them as a means? It cannot matter to the people killed whether they are
killed by action that is a military means any more than it can matter to them whether
they are killed as a side effect rather than intentionally. (This seems true, in particular,
on Scanlon’s contractualism.**) If I am one of the innocent people, I will not be
impressed to be told that you are permitted to kill me because the act whereby you kill
me is a military means to victory rather than a nonmilitary means.
The question is what can justify an exception to the prohibition of killing innocent
people. Tim says that achieving a military advantage in a just war, thereby increasing
the probability of achieving the just cause, can. But increasing the probability of
achieving the just cause to the same degree, but directly, without doing so via gaining a
military advantage, cannot. If we are simply concerned with reasons given by the facts,
this doesn’t make much sense. If we have reason to increase the probability of
achieving the just cause and the same number of innocent people will be killed
whichever way we act, what difference can it make whether our causal path involves
demoralizing the government or weakening the government’s military?
Implications of Tim’s view for self-defense?
*There might be an analogue in individual self-defense. Namely: one may direct force
only against the one who poses the threat. Compare a case in which an innocent
bystander will be killed as a side effect of one’s defensive action against an aggressor.
Suppose that’s permissible. Secondary permissibility might make it permissible to kill
that same innocent bystander as a means. But suppose it’s a different innocent
bystander who would have to be killed as a means. That, presumably, would not be
justifiable on Scanlon’s view. But killing that other innocent bystander could become
permissible as a side effect of aiming a defensive blow at the aggressor.
**Just as the relevance of intention won’t emerge from Scanlon’s contractualism, so the
relevance of the distinction between military advantage and other means won’t emerge
Intention and justice
(1) To win a just war, kill 100 innocent people as a proportionate side effect of military
action, or kill exactly the same 100 people intentionally, as a means.
(2) To win a just war, kill 100 innocent people as a side effect of military action, or kill
99 different people intentionally, as a means.
(3) To win a just war, kill 100 innocent people as a side effect of military action, or kill
99 of the same people intentionally, as a means.
If intention is relevant to permissibility, killing the 100 intentionally in the first case must
be more seriously objectionable. Indeed, killing 100 as a side effect is permissible
while killing the same number intentionally is impermissible. And in the second case,
it seems that killing 100 as a side effect would be permissible, while again killing 99
intentionally would not be. Yet in the third case, killing fewer of the same people
intentionally rather than killing more of them as a side effect seems not only
permissible but required. So it seems that considerations of justice are involved here.
It’s not just intention, but the way in which intention interacts with justice.
(5) Find an alternative explanation of the impermissibility of terrorism – for
example, in legitimate authority
Lionel McPherson: shift the objection to terrorism from the idea that it involves
intentionally killing the innocent to the idea that it involves the use of violation for
political purposes without legitimate political authorization.
But approval by the collective is only instrumentally or procedurally important. It helps
to guard against the use of terrorist means but the absence of it isn’t what constitutes
terrorism. Some unauthorized political violence against noncombatants can be
legitimate and, more importantly, some fully authorized attacks on innocent people
have all the objectionable features of terrorism: Nazi bombings of British cities,
Hiroshima and Nagasaki, etc.
Best solution: accept, at least provisionally, that intention is relevant to
It’s not implausible to suppose that in determining whether one would be acting
permissibly, one needs to determine whether one would in fact be acting on the reasons
one has, which are given by external considerations, rather than on the basis of other
considerations that don’t in fact provide normative reasons, or that in fact count against
acting in the way one proposes to act.
Acting for an acceptable reason may be a necessary condition of objectively permissible
action, but not a sufficient condition.
In other words, it’s sufficient for an act to be wrong that it’s done with an unacceptable
Yet for an act to be permissible, it’s necessary both that the act be objectively permissible
and that the agent do the act for an acceptable reason. On this view, it’s harder to act
permissibly than to act impermissibly.
Reflective equilibrium justification
If I were God inventing morality, I’d give people rights but it might not occur to me to
make it the case that whether those rights are overridden may depend on the intentions
of the agent.
The case for the relevance of intention isn’t so much theoretical as it is intuitive. If we
deny that intention is relevant to permissibility, that may necessitate giving up more of
ordinary morality (and in criminal law) than most of us could accept.
The question is whether the arguments against the relevance of intention to permissibility
are more persuasive or rationally compelling than the intuitions we’d have to reject if
we reject the relevance of intention to permissibility.
Admittedly hard to understand how intention could affect permissibility. But the
intuitive costs of rejecting the significance of the distinction are high. And the
arguments against the relevance of intention aren’t decisive. We shouldn’t reject the
distinction without seeing what can be said in its defense.