Terrorism without Intention

Author(s): by DavidRodin
Source: Ethics, Vol. 114, No. 4, Symposium on Terrorism, War, and Justice (July 2004), pp.
752-771
Published by: The University of Chicago Press
Stable URL: http://www.jstor.org/stable/10.1086/383442 .
Accessed: 12/01/2015 20:48
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to Ethics.

http://www.jstor.org

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Terrorism without Intention*

David Rodin
If we are to be engaged in a war on terrorism, then we had better get
clear about what terrorism is. Until we achieve a clear and coherent
understanding of what the morally relevant features of terrorism are,
we cannot hope to develop an appropriate moral response to the war
on terrorism now being prosecuted by the United States and its allies
its proper aims, scope, limitations, and potential exceptional permissions.
In this article, I present a unified way of understanding the moral
significance of terrorism. I begin by briefly identifying four different
strategies for defining terrorism. I then introduce a definition of terrorism which locates its moral significance in the object of attack
terrorism is given its distinctive moral character by the fact that it uses
force against those who should not have force used against them. In
the idiom of the just war theory, it uses force against noncombatants.
It has sometimes been claimed that certain military actions of key
Western powers such as the Unites States, NATO, and Israel are properly
described as acts of terrorism because they cause the death of a large
number of noncombatants. This is a counterintuitive claim. But by examining the doctrine of double effect and contrasting it with the categories of reckless and negligent harming, I will argue that there are
good reasons for thinking that these claims may sometimes be correct.
Some harms inflicted unintentionally on noncombatantsso called collateral damagemay indeed be properly categorized as terrorist. I will
conclude by briefly indicating some of the implications which this analysis has for our thinking about the ethics of war.
I. DEFINING TERRORISM
The concept of terrorism has been so deformed by rhetorical usage that
it is probably not possible to provide a full analysis of its use in common
* I would like to thank seminar participants at Atlanta, St. Andrews, Melbourne, and
Oxford for their helpful comments. In particular I would like to thank Jeff McMahan,
Roger Crisp, and Tony Coady for their help.
Ethics 114 (July 2004): 752771
2004 by The University of Chicago. All rights reserved. 0014-1704/2004/114040003$10.00

752

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

753

language. At best what we can discern is a loose family resemblance. I

shall suggest, however, that it is possible to provide a moral definition
of the term. By this I mean an analysis of the features of acknowledged
core instances of terrorism which merit and explain the moral reaction
which most of us have toward them. These reactions are undeniably
negative; most of us regard acts of terrorism with abhorrence. This in
turn raises the question of whether the moral definition of a pejorative
concept makes the specified act wrong by definition. There is certainly
truth in the thought that wrongness is part of the meaning of terrorism
in this respect the concept is more like that of murder than it is of
killing. Yet it is not the case that establishing a moral definition trivializes
the task of morally assessing terrorism by turning it into a matter of
simple definition. This is because, first, it is part of the function of a
moral definition to explain why the class of actions is wrong. The strategy
I employ in this article is to explain an explicitly normative concept
through definitional elements that are at least less normative. Second,
once we have discovered the set of features that accounts for our negative moral reaction to a class of action, it is still an open question
whether there exist cases in which the act so defined may be justified
or excused (e.g., because of overwhelming consequentialist considerations). Understood in this way, a moral definition does not seek to
encompass all potential usages of a term but rather to unearth and
explain the distinctive features which enable the concept to play the
role it does in our moral thinking.
Many philosophers and theorists of various kinds have attempted
to define terrorism. These attempted definitions are too numerous to
review here in detail, but they may be usefully classified as making appeal
to four different sets of consideration. The taxonomy is not meant to
be exclusive; theorists can and often do appeal to more than one. Nevertheless, the fourfold classification usefully identifies and orders the
differing theoretical approaches to terrorism.
1. Tactical and operational definition. This may be phrased in terms
of the weapons deployed; for example, Carlos Marighela, the Brazilian
revolutionary, defined terrorism simply as the use of bomb attacks.1
Clearly this is too narrow. Any such definition would have to be enlarged
to include a range of further operational modalities, some of which are
distinctive to terrorism (such as plane hijacking), but others may include
more familiar military practices. Another form of operational definition
focuses not on the weapon systems themselves but on the mode of their
employment. Michael Walzer claims that randomness is the crucial
1. Quoted in C. A. J. Coady, The Morality of Terrorism, Philosophy 60 (1985): 4769,
p. 47.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

754

Ethics

July 2004

feature of terrorist activity.2 The point here is not that terrorist attacks
are unplanned or untargeted but rather that the violence appears random from the perspective of the victim. Violence comes unpredictably,
often without warning, and people are killed or escape death by merest
chance. Some authors have insisted that terrorism involves violence or
its threat against persons.3 Others allow violence against property to
count.4 Still others have denied a necessary connection with violence,
claiming there can be nonviolent terrorism.5
2. Teleological definition. This focuses on the ends or goals of violence. Many writers claim that terrorism must be a political act.6 The
definition of the U.K. Terrorism Act 2000 is somewhat broader, specifying that it be for the purpose of advancing a political, religious or
ideological cause.7 Beyond this, some authors have argued that a necessary feature of terrorism is that it is coercive, aiming to get people or
groups to do things they would not otherwise do.8 Finally, terrorism
is etymologically connected to terror, and some authors have argued
that the creation of terror among a given community is definitional of
terrorism.9
3. Agent-focused definition. In contrast to tactical and teleological
definitions which focus on the nature of the act, this approach focuses
on the nature of the actor. For example, Walter Lacqueur defines terrorism as the substate application of violence.10 Unsurprisingly, such
a definition which excludes state action from the definition of terrorism
has been enthusiastically endorsed by states themselves: for example,
the U.S. State Department definition of terrorism restricts it to violence perpetrated . . . by sub-national groups or clandestine agents.11

2. Michael Walzer, Just and Unjust Wars (New York: Basic, 1977), p. 197. Walter Lacquer,
on the other hand, explicitly denies that randomness is necessary to terrorism. See Walter
Lacqueur, The Age of Terrorism (Boston: Little, Brown, 1987), pp. 14344.
3. Igor Primoratz, What Is Terrorism? Journal of Applied Philosophy 7 (1990): 12938,
p. 135.
4. Coady, The Morality of Terrorism, p. 52.
5. Carl Wellman, On Terrorism Itself, Journal of Value Enquiry 13 (Winter 1979):
25058, p. 251.
6. See Coady, The Morality of Terrorism, p. 52; Jenny Teichman, How to Define
Terrorism, Philosophy 64 (1989): 51213.
7. U.K. Terrorism Act, 2000, pt. 1, sec. 1.
8. Wellman, On Terrorism Itself, p. 250.
9. Ibid.
10. Walter Lacqueur, Postmodern Terrorism, Foreign Affairs 75 (1996): 2436, p. 24.
11. Quoted in George Lopez and Neve Gordon, Terrorism in the Arab-Israeli
Conflict, in Ethics in International Affairs, ed. Andrew Valls (Lanham, Md.: Rowman &
Littlefield, 2000), pp. 99113, p. 103.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

755

4. Object-focused definition. Finally, terrorism has been defined as

attacks against a particular class of target variously described as innocent, neutral, civilian, or non-combatant.12 Igor Primoratz
goes further, arguing that terrorism has a certain basic structure. It
has not one, but two targets: the immediate, direct target, which is of
secondary importance, and the indirect target which is really important.13
What are we to make of this multiplicity of definitional elements?
The first thing to note is that it is plausible that each of these definitional
features may have utility in different contexts. For example, for the
purposes of a military-strategic, or sociological, or psychological investigation it may be highly relevant to distinguish between violence committed by state and nonstate actors, between different tactical modalities,
and between violence directed to different teleological aims. However,
it will be my argument that for the purposes of a moral definition of
terrorism, the crucial factor is the fourth elementthe object against
which force is used. The key to a moral understanding of terrorism is
that it consists in the use of force against those who should not have
force used against them. A teleological definitional element is necessary
to distinguish terrorism from common domestic forms of criminal activity, but it is the object-focused aspect that really directs and explains
our typical moral judgments.
With this in mind, I propose to introduce the following moral
definition: terrorism is the deliberate, negligent, or reckless use of force against
noncombatants, by state or nonstate actors for ideological ends and in the absence
of a substantively just legal process. Like all definitions, this one contains
a number of suppressed arguments, and it is important that we make
them explicit in order to see what is doing the moral and conceptual
work. The aspect of my definition which is likely to be most controversial
is the claim that terrorism may include negligent and reckless as well
as deliberate uses of force, and this will be defended in a separate section
below. Before doing this I propose to explain each clause of the definition in some detail.
Use of force.This is a deliberately broader notion than that of violence, with its connotations of wild and explosive physical harming.
Clearly the kinds of force used and the methods of its deployment must
be interpreted widely. Use of conventional weapons, weapons of mass

12. For innocent, see Primoratz, What Is Terrorism? pp. 131, 133; for neutral,
see Teichman, How to Define Terrorism, p. 513; and for non-combatant, see Coady,
The Morality of Terrorism, p. 52.
13. Primoratz, What Is Terrorism? p. 131.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

756

Ethics

July 2004

destruction, hostage taking, poisoning, systematic incidences of rape,

and destruction of property may all be terroristic.14
There will inevitably be a certain difficulty in determining what
level of force is required to reach the threshold of terrorism, and there
will certainly be many borderline cases. But equally clearly, for a use of
force to count morally as terrorism it must cause very substantial harm
to the lives and interests of its victims. The paradigm of terrorism is the
employment of extreme and shocking force, for example, killing, maiming, and hostage taking. It is doubtful, indeed, whether lesser uses of
force such as slapping or spanking could ever count morally as terrorism.
It has often been claimed that the threat of force may count morally
as terroristic.15 This is plausible, but the threats must be credible, and
usually this will be achieved by the prior use of actual force.
Ideological ends.The term ideological is used here in its broadest
sense to signify a systematic scheme of ideas, usually relating to politics
or society, or to the conduct of a class or group, and regarded as justifying actions.16 It is meant only to signify a commitment to some
systematic and socially directed end beyond the motives of fear, anger,
lust, and personal enrichment, which are the typical motives of common
violent crimes.
This teleological element is required in the definition to distinguish
terrorism from common crime. Though terrorism is a species of crime,
it is distinguished from common murder, rape, and destruction of property by the fact it is directed toward a broader agenda. Usually the
ideological nature of terrorism will mean that it is committed by organized groups with an established infrastructure, but this is not an
invariable feature, as cases of lone terrorist activity have been known to
14. There may be some doubt as to whether the disruption of information systems
(cyberterrorism) can be properly described as a use of force. Some incidences clearly
are. For example, one of the most frightening possibilities envisioned by Pentagon scenario
planners is the use of a high-altitude electromagnetic pulse (or HEMP) to attack the
information infrastructure of the United States. A single nuclear device detonated at an
altitude of 300 kilometers could devastate electrical systems across 90 percent of the
continental United States, though the blast itself would cause no harm to persons on the
ground. Such an attack is certainly a use of force and would count as an act of terrorism
in the morally relevant sense. In contrast, the use of computer hacking to disrupt Web
sites and information systems does not employ physical force, though it is often included
under the heading of cyberterrorism. Such acts are probably best seen as borderline cases
of terrorism. Though they do not involve physical force, they are still properly described
as attacks upon property (albeit of an intangible kind) and may cause significant harm
to the lives and interests of their victims.
15. For example, Walter Sinnott-Armstrong includes the threat of force in his definition of terrorism (Walter Sinnott-Armstrong, On Primoratzs Definition of Terrorism,
Journal of Value Enquiry 8 [1991]: 116), and it is explicitly included in the U.K. Terrorism
Act, 2000, definition (pt. 1, sec. 1).
16. Oxford English Dictionary, 2d ed.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

757

exist (e.g., the Unabomber, who conducted a prolonged series of bomb

attacks in America).
Reference to political ends, which is the preferred characterization
of many authors, is not fully adequate because it fails to capture the
numerous species of terrorist motivation beyond the purely political.
Terrorism may, for example, be national separatist (IRA, ETA, PLO,
Tamil Tigers), religious (Al-Qaida), racist (Ku Klux Klan), environmental (Animal Liberation Front, Earth Liberation Front), millennial fatalist
(Aum Shinrikyo), antitechnologist (the Unabomber), or antifederalist
(Timothy McVeigh, who bombed the Oklahoma federal office building).
Noncombatants.Clearly this clause of the definition contains the
crux of the moral argument. It is designed to capture the intuition that
terrorism has the significance it does in our moral thinking because the
targets it attacks are morally inappropriate. Put simply, terrorism involves
the use of force against those who should not have force used against
them, and to do so is a moral crime. Thus, terrorism is the political or
ideological species of common violent crime. It is the criminality (and
in its most serious form, the murderousness) of terrorism which explains
its moral status and the reaction we rightly have to it.
Now to characterize terrorism in this way is implicitly to invoke
some principled way of distinguishing those who are morally liable to
the use of force and those who are not. In the Western tradition the
most widely accepted way of doing this in the context of political conflict
is by reference to the principle of discrimination within the just war
theory. This stipulates that the only appropriate objects of force in a
conflict are combatantsthose who are engaged in fighting and are
therefore either individually or collectively offering harm to the putative
agent of force or to another.17 All others are to be regarded as noncombatants and excluded from attack.
The just war principle of discrimination is not without its critics. It
has been doubted, for example, whether the notion of a combatant
really identifies sufficient conditions for being an appropriate object of
force. For being a combatant or offering harm is a material fact about
the agent; it need say nothing about his moral responsibility for the fact
that he is offering harm (he might, e.g., be an unwilling and largely
blameless conscript).18 But for our present purposes we need make no
judgment about the soundness of the permissive side of the principle
of discrimination. As Michael Walzer says, The theoretical problem is
17. Thus the rather confusing equation within just war theory between noncombatants
and the innocent. Innocent is the negation of nocent, meaning harmful or injurious.
18. For a critique of the claim that it can be justifiable to use force against a morally
innocent material aggressor, see David Rodin, War and Self-Defense (Oxford: Oxford University Press, 2002), chap. 4, esp. pp. 83 ff.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

758

Ethics

July 2004

not to describe how immunity is gained, but how it is lost. We are all
immune to start with; our right not to be attacked is a feature of normal
human relationships.19 This is surely right, the immunity of noncombatants from attack is the foundational element in our moral thinking,
and whether or not the just war theory is ultimately able to sustain the
permissibility of killing combatants is irrelevant to this fact. It is this
basic moral judgment that ordinary people who are not engaged in any
threatening combat operations should not be subject to attack that
explains and underlies the moral repugnance we justifiably feel about
acts of terrorism.
The term noncombatant is preferable to the related term civilian
often employed in international law.20 The reason for this can be seen
from consideration of the bomb attack on the American destroyer USS
Cole in the port of Aden in October 2000 (probably by members of AlQaida). The crew members killed were uniformed servicemen and
therefore not civilians, but they were noncombatants in the morally
relevant sense. They were in port on a routine friendly visit and were
not involved in any combat operations at the time. Because of this, the
attack on the Cole is rightly seen by most observers as terroristic. Similarly,
soldiers who are wounded, sick, or have surrendered (hors de combat)
are noncombatant in the relevant sense, and attacks against them may
be morally terroristic.
State or nonstate actor.Strictly this element is superfluous to the
definitionit has been inserted only to make the universality of the
definition explicit.
Universality is a basic principle of the interpretation of moral rules.
There are, to be sure, exceptions to this principle, cases in which the
identity of the agent is relevant to the permissibility of a given act. For
example, though I am not permitted to cut people open and remove
their kidneys, doctors often are. Cases like this, however, are exceptional,
and all depend upon a rich moral context which explains the normative
particularity (the doctors special status and training, his duty of care,
the presence of the patients prior consent, and so on). Though there
is no a priori way to rule out such a contextual moral distinction between
state and nonstate use of extralegal force against noncombatants, it is
certainly far from obvious how such an account could proceed.
On the contrary, the proposition that acts of terrorism may be
committed by state as well as nonstate actors should be obvious once
19. Walzer, Just and Unjust Wars, p. 145n.
20. See, e.g., Protocol Additional to the Geneva Conventions of 12 August 1949, and Relating
to the Protection of Victims of International Armed Conflicts (Protocol 1) (1977), art. 51, sec. 2:
the civilian population as such, as well as individual civilians, shall not be the object of
attack.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

759

one has accepted that the key to a moral understanding of terrorism is

that it uses force against inappropriate objects. One may consider the
1985 bomb attack (with the loss of one life) on the Greenpeace vessel
Rainbow Warrior in Auckland harbor by French secret agents. This was
a paradigmatic terrorist attack; it involved the killing of noncombatants
in the absence of substantively just legal process for ideological purposes.
It is difficult to see how the fact that the perpetrator was a state government (and a liberal democratic one at that) could be relevant to
our moral assessment of the attack as terroristic.
Substantively just legal process.The purpose of this clause is to create
a qualified exception to the definition. The exception is required because without it the use of force by agents of the state to enforce the
law would be classed as terrorism. That cannot be right. Proper enforcement of the law by state agents is not an instance of terrorism;
indeed it is a paradigm of morally justified use of force.
However, the wording of the clause is designed to capture the limits
of this claim. For not all uses of force which are sanctioned by a legal
process are morally justified. If the legal process itself is substantively
unjust, then it cannot confer moral justification on the use of force.
There are both historical and moral reasons for thinking that when this
is the case the use of force may properly be regarded as terroristic.
When the term terrorism was first coined in the late eighteenth
century, it was employed to describe the use of terror as a tool of government by the Jacobin party after the French Revolution. Soon it came
to signify the internal use of terror by states more widely, what today
we would principally call totalitarian or repressive government. Thus,
in contrast to many modern definitions which explicitly exclude state
violence, in its original meaning terrorism was an act which could only
be committed by states. Today the word is seldom used in this way, but
it is apparent that from a moral perspective such use of force shares
relevant features with more familiar acts of terrorism such as bombings
and hijackings by nonstate actors: it too directs force against inappropriate targets, attacking those who should not be attacked for ideological
ends. The wording of this clause, therefore, is designed to allow such
uses of force to be included in the moral definition of terrorism. Thus
regimes that systematically inflict unjust force on their own populations
are rightly described as terrorist regimes.
Reference to a substantively just legal process is meant to signify
a contrast with a more procedural interpretation of justice. Thus totalitarian and repressive states often use violence in a way that is procedurally legalistic but which fails to conform to a substantively just legal
process. It is the procedurally legalistic but substantively unjust use of
force by states which qualifies for inclusion under the moral definition
of terrorism.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

760

Ethics

July 2004

Just what makes a legal process substantively just is a central question of political philosophy and cannot be properly explored here. Indeed, it may seem a defect of my definition that it includes reference
to such a contested and explicitly normative notion as substantive justice.
But I take it that any adequate political philosophy must contain some
way of distinguishing substantively just from unjust legal processes. The
proper way of understanding this element of the definition is to view
the term substantively just as a marker or placeholder for the particular
theory of justice one wishes to endorse. As such, the definition appropriately recognizes the way in which ones commitments in the theory
of justice have implications for the moral scope of terrorism (as indeed
they will have implications for many other normative concepts). The
positive claim I want to make is that when a state uses force against
noncombatants in accordance with a legal system which fails the substantive justice test (however one chooses to fully characterize that test),
there is good reason to count this morally as terrorism.21
It is worth noting that it is the legal processes themselves which
must be substantively unjust in order to satisfy the definition. Isolated
miscarriages of justice within an otherwise just legal system do not constitute terrorism. However, the systematic application of unjust legal
processes in ways that constitute a significant use of force against the
lives and interests of noncombatants certainly may constitute terrorism.
One further element of my definition is operative by way of omission. I have made no mention of the further teleological and agentfocused features which some authors have thought definitional of terrorism: the claim that it is coercive, the claim that there is a necessary
distinction between direct and indirect targets, and the link between
terrorism and terror. Jenny Teichman says we will look pretty silly if we
do not mention terror in our account of terrorism.22 Michael Walzer
has said that what is distinctive about terrorism is that it reaches beyond
all limits; it is infinitely threatening to whole peoples, whose individual
members are systematically exposed to violent death at any and every
moment in the course of their (largely innocuous) lives.23
Well, is it silly to omit mention of terror in a moral definition of
terrorism? The first thing to note is that Walzers characterization is
itself a considerable exaggeration. Anyone who has lived or worked in
London or Jerusalem or New York knows that this is not (necessarily)
21. It would seem that as a minimum we may say that a substantively just legal system
must involve the application of laws that do not violate basic human rights; impartial
enforcement and judicial bodies; fair rules of evidence; and ideally effective procedures
of appeal and review.
22. Teichman, How to Define Terrorism, p. 511.
23. Walzer, Just and Unjust Wars, p. 200.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

761

what it is like to live under threat of terrorist attack. Only in the most
extreme cases do terrorist attacks cause a genuine terror or panic in
the population as a whole. Shocking as the attacks may be, ordinary
people generally get on with their lives and think about the threat only
when personally affected or when the violence is reported in the news.
Such, apparently, is the grim resilience of the human spirit.
It does not seem to be the case that terrorism invariably has the
effect of causing terror in a population. On the contrary, there have
been instances in which communities have grown stronger and more
resilient as a result of terrorist attacks. London during the blitz is the
classic (though doubtless over-romanticized) example. If terrorism does
not of necessity create terror in the population against which it is directed, then terror cannot be part of the moral definition of the crime.
But it may be that to focus on actual effects is misleading. It may
be that the term terrorism is akin to teaching; one can be engaged
in the activity of teaching even if there isnt much learning going on
in the classroom. Perhaps the relevant fact is that the terrorist intends
to cause terror in the population. But this too seems inadequate. Perhaps
the terrorist is just confused about what the effects are likely to be, or
perhaps his moral compass is so distorted that he believes himself to
be benefiting his victims and their community, as is apparently the case
with members of the Aum Shinrikyo cult.24 Should these suppositions
prove correct, I can see no reason to alter our moral appraisal that such
acts constitute terrorism. A similar line of argument can be constructed
to exclude from the definition the other putative features, namely, coerciveness and the double targeting of terrorist attacks.25
The key point to understand here is the distinction between definitional and aggravating features of an offense. A definitional feature
of an offense identifies, as it were, a certain quantum of opprobrium
a step change in moral seriousness. We may be guided here by a domestic
analogy. If a man in domestic society commits an act of murder and it
is a feature of his crime that it causes (or was intended to cause) widespread terror, or if it was used to coerce a certain group of people or
to attack an indirect target distinct from his direct target, then these
are all aggravating features. They will cause us to view his crime with
greater seriousness than otherwise, but they do not change the fact that
what he is guilty of is an act of murder (albeit a particularly egregious
24. Lacqueur, Postmodern Terrorism, sec. 2.
25. None of this, of course, should obscure the significance of such features for the
moral assessment of terrorism. It is a vitally important feature of the crime of terrorism
that it will frequently cause severe and widespread fear and social dislocation. The claim
is simply that these features ought not be made part of the definition of the crime. They
are rather, as Tony Coady says, important empirical insights into the sociology and motivational psychology of terrorism (Coady, The Morality of Terrorism, p. 53).

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

762

Ethics

July 2004

one). In a similar way it would seem that the coercive and terrifying
features of terrorism (when they are present) are aggravating but not
definitional features of the offense.
II. THE DOCTRINE OF DOUBLE EFFECT AND THE
DISTINCTION BETWEEN TERROR BOMBING AND
TACTICAL BOMBING
The final element of my definition is intended to be somewhat more
revisionary than the preceding analysis. Because the definition includes
negligent and reckless uses of force, it implicitly entails that some
harms unintentionally inflicted on noncombatants in the course of war
are a form of terrorism. The issue arises particularly in the case of aerial
bombardment against targets within or adjacent to civilian populations
which is almost certain to generate noncombatant casualties. Although
new targeting technologies have significantly increased the bombing
accuracy of Western air forces, the number of noncombatant casualties
in modern air campaigns such as Kosovo, Afghanistan, and Iraq remains
high in absolute terms. For example, though it is difficult to verify the
reliability of figures, on some estimates the number of noncombatants
killed by Operation Enduring Freedom in Afghanistan exceeded the
number killed in the attacks on the Pentagon and World Trade Center.26
According to traditional moral theory, such collateral casualties
are considered the regrettable but generally permissible side effect of
legitimate military activity and are strongly distinguished from terrorism.
In what follows I aim to cast doubt on this assessment. Doing so will
entail examining a feature of moral theory which has long been thought
to support it: the doctrine or principle of double effect. A standard
formulation of the principle of double effect is as follows. One may
never intentionally bring about an evil, either as an end in itself, or as
a means to some greater good. Nonetheless, one may use neutral or
good means to achieve a good end which one foresees will have evil
consequences provided that (i) the evil consequences are not disproportionate to the intended good, (ii) the action is necessary in the sense
that there is no less costly way of achieving the good.27
The double-effect principle has traditionally been taken to ground
26. Carl Conetta puts the figure at 1,0001,300 civilian deaths directly from aerial
bombardment and a minimum of 3,000 civilian deaths attributable to the impact of the
bombing campaign and war on the nations refugee and famine crises (Strange Victory: A
Critical Appraisal of Operation Enduring Freedom and the Afghanistan War, Project on Defense
Alternatives, Research Monograph no. 6 [Cambridge, Mass.: Commonwealth Foundation,
January 2002]).
27. See Frances M. Kamm, The Doctrine of Triple Effect and Why a Rational Agent
Need Not Intend the Means to His End, The Aristotelian Society Supplementary Volume 74,
no. 1 (2000): 2139, p. 23.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

763

a moral distinction between the deliberate targeting of noncombatants,

often called terror bombing, and tactical bombing, which aims at military targets but may cause collateral damage among noncombatants.
In terror bombing, so the argument goes, the agent intends the death
of noncombatants as a means to his ends. This is always wrong, though
the goodness and viability of the ends may be very great indeed. In
tactical bombing, on the other hand, the agent intends only to strike
at legitimate military targets. The death of noncombatants, while it is
a foreseen consequence of the action, is not intended either as a means
or as an end (in Benthams terms it is an oblique intention but not a
direct intention). Provided the requirements of necessity and proportionality are met, therefore, the action, together with its foreseen consequences, is not impermissible.
Now many philosophers have been deeply skeptical of the view that
the intention of the agent in performing an action can be determinative
of the permissibility of the act in this way. As several authors have pointed
out, we have no satisfactory theoretical explanation of why intention in
the sense here invoked should make a difference to the permissibility
of actions.28 Moreover, J. J. Thomson has spoken about the peculiarity
of the principle when applied in action-guiding contexts. If a bomber
pilot came to ask for your advice as to the permissibility of a raid which
will destroy a military installation and also cause significant civilian casualties, it would seem absurd if you said to him that it all depends on
your direct intention: if your intention is to destroy the installation,
foreseeing but not intending the death of the civilians, this is permissible. But if your intention is to kill the civilians, merely foreseeing the
destruction of the installation, this would be impermissible.29 Even writers sympathetic to double effect such as Elizabeth Anscombe and John
Ford have stressed the care with which it must be applied and the
dangersboth intellectual and moralof its abuse.30
Important as these criticisms are, I would like to raise a different
set of questions about the principle of double effect by contrasting it
with two considerations that explain how legal and moral culpability
28. T. M. Scanlon, Intention and Permissibility, The Aristotelian Society Supplementary
Volume 74, no. 1 (2000): 30117, p. 303.
29. See Judith Jarvis Thomson, Self-Defense, Philosophy & Public Affairs 20 (1991):
283311, p. 293.
30. G. E. M. Anscombe, War and Murder, in Moral Problems, ed. James Rachels (New
York: HarperCollins, 1979); J. C. Ford, The Morality of Obliteration Bombing, in War
and Morality, ed. Richard Wasserstrom (Belmont, Calif.: Wadsworth, 1970), pp. 1541.
Partly as a response to worries such as these, T. M. Scanlon has, in a recent paper, suggested
a way of interpreting typical double-effect cases in a way that does not invoke the distinction
between direct and oblique intention: T. M. Scanlon, Intention and Permissibility, The
Aristotelian Society Supplementary Volume 74, no. 1 (2000): 30117.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

764

Ethics

July 2004

may exist in the absence of direct intention: these are the concepts of
recklessness and negligence. Recklessness may be characterized as the
culpable bringing about of unintentional evil consequences (or the risk
thereof) that are in fact unreasonable and unjustified in the circumstances. Though recklessness is distinguished in law from intention
proper, it is nonetheless a form of mens reaa mental state sufficient
for criminal liability. In law there are two forms of recklessness, subjective
recklessness, which requires that the agent consciously foresaw the risk
of evil consequences, and objective recklessness, in which the agent did
not foresee the risk but where a reasonable person would have done
so. Objective recklessness is very similar in content to negligence, which
is the failure to take reasonable precautions in the face of a foreseeable
risk. Negligence is primarily a civil law concept, but it may constitute a
ground for criminal liability where the negligence is gross.
Judgments involving negligence and recklessness sit uncomfortably
with the principle of double effect. Take the case of a motorist who
drives across a crowded school yard to deliver a sick person to a hospital.
The motorist certainly has no direct intention to harm the childrenhe
aims at their death neither as a means nor as an end. His conduct may
be necessary in the context, and if we imagine that the risk of death to
his passenger and to the children are roughly balanced, then it will also
be proportionate. Yet if he strikes and kills a child he will be held liable,
in law and in morality, for manslaughter because of the recklessness of
his actions. Similarly, a medical researcher who prematurely tests an
unsafe new vaccine on humans out of a desire to speed the development
of the product and thereby save lives is guilty of negligent action and
may be liable for manslaughter. This is true even though his action may
fulfill all the requirements of the double-effect principle.
One might object that there can be no tension between the concepts of recklessness and negligence on the one hand, and the doctrine
of double effect on the other, because reckless and negligent risks must
be such that they are in fact unreasonable in the circumstances. But we
have already said that actions that pass the double-effect test must be
necessary and proportionate, and surely action that is necessary and
proportionate is ipso facto reasonable.31 To construe the notion of reasonableness in this way, however, is to miss a crucial point. Persons have
rights against being harmed or used for the benefit of others, rights
which can only be alienated in very specific ways, usually having to do
with actions and decisions they have freely and responsibly taken. Because of this there is an additional element to the reasonableness test
which goes beyond the necessity and proportionality requirements,
namely: is it justifiable to inflict such a risk upon this particular person?
31. I am indebted to Jeff McMahan for raising this point.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

765

What motivates our intuition in the case of driving to the hospital

through the school yard and the vaccine case is that those who are being
forced to bear the risk have a right not to have such grave risks inflicted
upon them. The fact that the risks are necessary (from the perspective
of the beneficiaries of the risky activity) and proportionate (from the
impersonal perspective of the world at large) is not sufficient to defeat
the personal right not to be endangered or used in this way.
Now these observations about negligence and recklessness do not
strictly contradict the doctrine of double effect because that doctrine,
on its traditional interpretation, does not purport to identify sufficient
conditions for innocence, only sufficient conditions for guilt. Nonetheless, there is clearly a tension here. The principle is supposed to show
how the absence of a direct intention on the part of the agent can
materially affect the permissibility of acts with harmful consequences.
The concepts of negligence and recklessness, on the other hand, approach the idea of permissibility by drawing on the idea of a requirement
to observe a reasonable standard of care in ones actions and not to
undertake unreasonable risks. But one may fail to observe a reasonable
standard of care either by possessing a direct intention to cause harm,
by possessing an oblique intention to cause harm, or by possessing no
intentional attitude toward the harm at all, as when one fails to foresee
a harmful consequence of ones action that one could have been reasonably expected to foresee and avoid. Considerations of recklessness
and negligence threaten not to disprove the principle of double effect
but rather in certain contexts to render it irrelevant.
III. STANDARDS OF CARE IN MILITARY OPERATIONS
What does all this imply for our discussion of aerial bombardment and
the definition of terrorism? Clearly it shifts the focus of attention away
from the direct intention of those engaged in the activity and onto the
question of what constitutes an appropriate standard of care in military
operations. In general, standards of care are contextual and depend on
a large number of conditions specific to the agent, activity, and situation,
but it is possible to identify some common elements. For instance, the
more dangerous the activity one is engaged in, the higher the standard
of care required. Thus a higher standard of care is required in the
operation of a car than in riding a bicycle, and a higher standard still
is required in the operation of a jumbo jet. Standards of care are also
higher in activities that have a higher causal immediacy to harm. Thus,
those who perform roadwork are subject to a higher standard of care
than those who play football, even though footballs sometimes roll onto
the road and cause accidents. Standards of care are often linked to
professional relationships, and the standard of care expected of a pro-

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

766

Ethics

July 2004

fessional is generally higher than that expected of a nonprofessional

engaged in the same activity.
If we apply these observations to the activity of aerial bombardment,
we will observe that bombardment is a highly dangerous activity, that
there is a high causal immediacy between the dropping of bombs and
resulting fatalities, and that the activity is carried out by highly trained
professionals. All of these observations will naturally lead us to the
thought that the standard of care required of those engaged in military
activities such as aerial bombardment is very high indeed, and that
consequently when noncombatant fatalities are caused as the unintended but foreseen side effect of bombardment, this must raise serious
questions of culpable negligence or recklessness.
Balanced against this is the fact that in certain circumstances, often
in emergencies, it is justifiable to observe a lower standard of care and
to incur a higher degree of risk than would otherwise be appropriate.
So, for example, it may be justified for a surgeon to perform an operation in which there is an extremely high risk of death to the patient,
if the alternative to surgery is an even higher risk of death or of great
suffering. Moreover, police cars and ambulances are permitted to exceed
the speed limit and pass red lights in emergencies (though of course
strict requirements of care remain, e.g., the requirement to display warning lights, the requirement to slow down when entering intersections,
and the requirement to generally minimize the risk to others). Finally,
road engineers are permitted to make a reasonable trade-off between
cost and safety when designing a road, though they know that this will
lead to a certain number of otherwise avoidable fatalities.
Perhaps military operations which carry a high risk of noncombatant casualties are not reckless or negligent because the risks are
justifiable in a similar way. But the analogy is not a good one. For in
all the above examples a crucial part of what makes the risk justifiable
is that the party assuming the risk is also the beneficiary of the riskproducing activity. This is obvious in the case of the patient undergoing
risky surgery, but it is also true, in a statistical sense, of the risks of
speeding police cars and less-than-perfectly safe roads. Because the risk
of death to any particular person is spread more or less evenly throughout the community, one may rightly say that the community which enjoys
the benefits of improved apprehension of criminals and an affordable
road system also bears the risks incurred. Why should this idea be important to our sense that the risk in these cases is justifiable? The answer,
I think, is that it links to the idea of autonomous agency and free
acceptance of risk. A patient decides whether to assume the risk profile
of a given operation, and every community makes its own collective
decision about how to make the trade-offs involved in building roads
and apprehending criminals. The risks are justifiable in these cases, at

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

767

least in part, because they have been autonomously assumed either

individually or collectively by those who bear them.
Contrast this with the case of aerial bombardment which imposes
a substantial risk upon enemy noncombatants. In standard cases of military conflict there is no sense in which the party who bears the risk of
harm benefits from the risky activity.32 Neither have they autonomously
chosen, either individually or collectively, to bear the risks of the bombardment. It is extremely doubtful, therefore, whether the risks imposed
on enemy noncombatants can be justified in an analogous way to the
risks imposed in domestic emergency situations.
If one is persuaded by these interlinked lines of argument then
one will feel that the standard of care required of those engaged in
military operations is highhigher than is currently reflected in the
laws of war. One will be inclined to view many of the noncombatant
casualties caused in the course of military operations (including those
of Western nations) to be culpably reckless or negligent. One will feel
this despite the considerations adduced by the principle of double effect.
IV. RECKLESSNESS AND NEGLIGENCE AS MENS REA
FOR TERRORISM
Even if one is not persuaded that standards of care in the course of
military operations are as high as I have suggested, one will have to
accept that there is some standard of care incumbent on soldiers engaged
in military operations and that when harm is caused as a consequence
of their action falling below this standard, those harms will be culpable
by reason of negligence or recklessness. The question then arises
whether such reckless and negligent harming of noncombatants may
properly be regarded as terrorist. Under the moral definition which I
have been defending, there is strong reason to suppose that it may, for
both share an underlying moral structure with more familiar forms of
terrorism. Both constitute a culpable use of force for ideological ends,
by a state or nonstate actor, against noncombatants, and without substantively just legal process. Under the principle that like cases be treated
alike, there is good reason to class both under the single moral heading
of terrorism.
There is no denying, however, that such an analysis is at odds with
the current common usage of the term, which generally restricts terrorism to acts of intentional harming. This may lead one to the following
response: why should we meddle with the moral definition of terrorism,
when we already possess an adequate set of conceptual tools for dealing
with the issues raised in this article, in the form of the proportionality
32. A possible exception to this observation may be in the case of humanitarian
intervention.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

768

Ethics

July 2004

requirement within the laws of war?33 Military action which causes unacceptable unintended harm to noncombatants can be condemned as
disproportionate use of force. What is significant about the arguments
I have presented here is that they show that the hurdle for proportionate
use of force is much higher than has been commonly believed. But the
appropriate way to deal with this is to strengthen the proportionality
requirement, not to deform the ordinary meaning of terrorism.
There are two reasons why I think this response is not adequate.
The first has to do with a general observation about the proportionality
requirement in the laws of war. The notion of proportionality has its
most natural application in the domestic sphere. If I kill an assailant
who was about to kill me, it is easy to see that this is a proportionate
use of force, just as it is easy to see that killing someone to defend my
prize apple pie from premature and wrongful consumption is a disproportionate use of force. But in the context of war the proportionality
requirement is much more difficult to interpret. The jus in bello proportionality requirement states that the harm done in the course of a
military operation must not be disproportionate to the concrete military
advantage likely to be gained. But it is not at all obvious that the two
values of military advantage and harm against noncombatants are morally commensurable. Exactly what metric is one supposed to use to
determine the level of military advantage required to offset the foreseeable deaths of a given number of noncombatants? Tweaking the
proportionality requirement is unlikely to be an effective response to
the concerns I have raised until we have a clearer conception of what
proportionality in war is and how precisely it functions.
The second and deeper reason for the inadequacy of the proportionality response is that it rests on a misunderstanding of the reasons
why the unintended infliction of harm or risk upon noncombatants is
morally problematic. The proportionality requirement is most naturally
interpreted as a quasi-consequentialist principle which states that there
is a moral limit upon prima facie justified actions which is triggered
when such actions do more harm than good. The doubts about collateral
damage which I have been raising in this article have an entirely different
source. They derive rather from a conception of persons as beings with
rights against being harmed, or exposed to risk of harm, in the absence
of justifying conditions relating specifically to their autonomous actions
and decisions.
Consider a domestic example. If the only way I can flee an assailant
is to kill or impose a significant risk of death upon an innocent bystander,
33. For the legal definition of jus in bello proportionality see Protocol Additional to the
Geneva Conventions of 12 August 1949, and Relating to the Protection of Victims of International
Armed Conflicts (Protocol 1) (1977), art. 51 (5)(b) and art. 57 (2)(b).

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

769

such action would certainly be proportionate on any reasonable interpretation of proportionality. But most people would view such action
as wrong, and the reason for this is that there is no relevant moral fact
about the bystander which could justify inflicting this (admittedly proportionate) harm upon him or her. In the same way, the conclusion
that much collateral damage caused by military operations is negligent
or reckless is not motivated only by the thought that the force is disproportionate (though it may well be that also). It is rather motivated
by the thought that it is inflicted upon those who should not have force
inflicted upon themthose who because they are noncombatants have
no liability to have force (not even proportionate force) used against
them.34 Proportionality, therefore, is not the appropriate locus for the
kind of moral issue I have been raising.
One may accept all of this, however, and still have reservations about
extending the definition of the term terrorism to include unintentional, albeit reckless and negligent, use of force against noncombatants.
One might feel that to do so would be an act of disreputable persuasive
redefinition akin to arguments that seek to define inequality and poverty
as forms of violence. But in reality, very little turns on the purely terminological question of the usage of the word terrorism. The important issue is whether one accepts the substantive argument put forward
in this article: that the unintentional killing of some noncombatants in
the course of military operations is morally culpable to the same degree
and for the same reasons that typical acts of terrorism are culpable.
If one does accept this conclusion, then I think there are two entirely reasonable ways to settle the terminological question. The first is
to restrict the definition of terrorism to the intentional (in the narrow
sense of directly intended) use of force against noncombatants and to
categorize the reckless and negligent use of force against noncombatants
as a separate class of offense. This has the virtue of mirroring the distinction made in criminal law between murder, which requires the mens
rea of intention, and the lesser offense of manslaughter, for which recklessness or negligence suffices (in the United States the distinction is
between first- and second-degree murder). This is not an intrinsically
34. This analysis, of course, raises a difficult question about how it can ever be justifiable to impose risk upon those who have not through their actions made themselves
specifically morally vulnerable to it. This question is beyond the scope of this article, but
my sense is that the answer has to do with two kinds of consideration. The first is the
kind of free collective consent to beneficial but risky activities assumed by communities,
such as was discussed in Sec. III above. The second has to do with the much rarer set of
cases in which there are such overwhelming consequentialist considerations that we are
inclined to believe that it is justifiable to impose risks upon those who have done nothing
to deserve their imposition. Such cases, of course, necessarily take us far beyond the bounds
of any plausible conception of proportionality.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

770

Ethics

July 2004

objectionable way of settling the terminological question, however the

manslaughter/murder distinction has been criticized by legal scholars
as unhelpful and ad hoc. Indeed, murder is one of the few criminal
offenses for which recklessness is not a sufficient mens rea. What is
more, we have no term in our moral vocabulary for the reckless, as
opposed to intentional, use of force against noncombatants for ideological reasons. An artificial term would need to be introduced (perhaps
reckless terrorism or terrorist manslaughter) which may have difficulty in gaining general currency.
The second response, reflected in my moral definition, is to hold
that the crucial point about terrorism is not the direct intention of the
agent but the wrongfulness of the act. Many acts which would generally
be considered as sufficient for terrorism such as assault, battery, or the
threat of the use of force, are lesser offenses than reckless manslaughter.
It is difficult to see why such intentional acts should be included in the
definition of terrorism, while the more serious offense of reckless manslaughter is excluded. Following this line of logic, my moral definition
of terrorism makes reference to the deliberate, reckless, or negligent
use of force against noncombatants. I believe that this definition best
articulates what is morally most important about the phenomenon of
terrorism.
V. CONCLUSION
What conclusions for policy can be drawn from this moral analysis of
terrorism? There are, I think, two broad and tentative lines of thought
which emerge and which concern the jus ad bellum and the jus in bello
of any engagement with terrorist adversaries. The first is that we should
be extremely cautious about extending the rules of jus ad bellum so as
to include terrorism as a just cause for war. What has come to be called
the Bush doctrine asserts that countries have the right to use force
(and indeed to engage in preemptive war) if they believe they may be
subject to a tangible terrorist threat. But if my analysis is correct, it is
apparent that terrorism in the morally relevant sense is a relatively broad
military and political practice which is frequently utilized by states, including the United States and her allies. This has been obviously true
in cases such as the terror bombing of Germany and Japan during the
Second World War. But, as I have here suggested, terrorism in the sense
of the reckless use of force against noncombatants has also been an
issue in several recent conflicts including NATOs bombing campaign
against the Federal Republic of Yugoslavia in 1999, Operation Enduring
Freedom in Afghanistan, the war in Iraq, and, perhaps most clearly,
Israels military response to the Palestinian intifada. Viewing terrorism
as itself a just cause for war would therefore result in a dangerous and
unwelcome extension of the right to resort to force, one which would

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions

Rodin

Terrorism without Intention

771

expose a number of Western powers themselves to the use of military

force.
Second, in terms of the jus in bello, it seems intuitively correct that
any engagement which has as part of its goal the elimination of terrorist
threats and which is called a war on terror should not, if it is to be
legitimate, use terrorist tactics itself. At a minimum this will require
much greater care to eliminate noncombatant casualties than has been
demonstrated by Western powers in recent operations. So-called smart
targetable munitions have the potential to revolutionize the humanitarian aspects of war. But aerial bombardments in Iraq, Afghanistan,
and Kosovo have all resulted in very significant noncombatant casualties,
and Western powers continue to use substantially indiscriminate weapons such as cluster munitions and land mines. Moreover, the increased
ability to target accurately means that we must pay greater attention to
the nature of targets selected. The increasing tendency to target dual
use facilities such as power stations, transport, and media infrastructure
is particularly worrying in this context, for destroying such targets has
the potential to cause very significant noncombatant deaths for months
and even years after the conflict has ended.35 If the war against terrorism
is to avoid being a terrorist war, then such practices must be seriously
reconsidered.

35. See Henry Shue and David Wippman, Limiting Attacks on Dual-Use Facilities
Performing Indispensable Civilian Functions, Cornell International Law Journal 35 (2002):
55979.

This content downloaded from 210.48.219.5 on Mon, 12 Jan 2015 20:48:00 PM

All use subject to JSTOR Terms and Conditions