Author: Seumas Miller

1       Definition of Terrorism

2       Definition in terms of Civilians

3       Definition in Terms of Innocents

4       The Definition of Terrorism in Terms of Non-Combatants


1        Definition of Terrorism

Providing an adequate or, at least, an acceptable definition of terrorism has proved to be difficult. However, it is not clear that the difficulties are primarily intellectual rather than merely political. ‘One man’s terrorist is another man’s freedom fighter’ is a slogan that might reflect the political reality but this does not necessarily reflect the conceptual reality.

I discuss the definition of terrorism in the context of the following assumptions: (1) Terrorism paradigmatically consists of violent actions directed at persons; (2) Terrorism is a strategy that involves methods standardly regarded as both immoral and unlawful, such as assassination, deliberate killing of innocent civilians, torture, hostage taking, kidnapping, ethnic cleansing, and the use of chemical, biological or nuclear weapons; (3) Terrorism involves terrorising or instilling great fear in one group in order to cause some other group or, more likely, sub-group (e.g. leadership of the terrorised group) to do what they otherwise might not have done; (4) Terrorism is a means to achieve political or military ends; (5) Terrorism relies on a high degree of publicity.

In the light of the assumption that terrorism is a strategy, we can define a collective as a terrorist organisation if (a) it is an organisation (or organisation of organisations) and (b) it has terrorism as one of its principal strategies. An organisation can be defined as an ongoing joint enterprise with a structure of organisation roles, relatively persisting organisational purposes or ends, and the capacity to reproduce itself (e.g. by recruitment, training etc.). Clearly on this account, Islamic State (IS), Al Qaeda (AQ) and Hamas are terrorist organisations. They are also non-state actors. However, notice that this account leaves open the possibility that there could be terrorist state actors. Indeed, the Soviet Union under Stalin looks to be an example and, if IS had succeeded in establishing its caliphate, it would have been another example.

The first assumption mentioned above is essentially a matter of standard linguistic usage and, therefore, meaning: the term, “terrorism”, is standardly used to refer to violence directed at persons. However, it is important to distinguish between violent actions performed by combatants in the context of a war and violent actions in the context of a well-ordered jurisdiction outside a theatre of war. Shooting an enemy combatant dead in a theatre of war is not typically an act of terrorism whereas shooting dead an ordinary person in a well-ordered jurisdiction might well be. That said, some violent actions directed at combatants in a theatre of war might well constitute terrorism, e.g. torturing enemy combatants or using chemical or biological weapons against them. I will not concern myself with the issue of morally unacceptable weapons of war in this paper.

The violence perpetrated by terrorists is somewhat diverse but would include deliberate bombing of occupied civilian buildings, deliberate shooting of ordinary citizens outside war-zones, torturing of political prisoners and the like. There is a residual issue as to whether or not violent actions directed solely at property, e.g., blowing up an empty building or unmanned electricity station, could count as acts of terrorism. Since this is controversial and marginal to my concerns here I assume that such acts are not integral to the meaning of the term terrorism. Accordingly, in what follows I restrict terrorist acts to violent actions directed at persons.

The second assumption is in part an empirical claim based on a consideration of the violent methods actually employed by those groups generally agreed to be terrorists. It is also in part based on a conceptual connection between these methods and the activity of terrorising. The prospect of being killed or tortured, for example, is inherently fearful.

The third assumption is in large part a conceptual claim. For an activity to count as terrorism someone has to be trying to terrorise someone else, and for terrorism to be a strategic activity — as opposed to, for example, merely an expressive activity — it has to be in the service of some further end, i.e., changing the attitudes and/or behaviour of some group.

The fourth assumption is largely stipulative, albeit not unmotivated. Terrorism is an activity directed to political or military ends. This is not to say that criminals, for example, do not use the methods of terror to achieve their criminal ends; clearly some criminals kidnap to extract a ransom, torture to instil fear and thereby extort money, and so on. Moreover, torture and other methods of terror can be, and have been, used for religious ends, e.g. the methods of torture used by the Spanish Inquisition in the Middle Ages. However, my concern here is with political or military ends. On the other hand, it needs to be noted that sometimes terrorism has multiple ends. Thus the terrorist methods of many extremist Islamist groups serve political as well as religious ends. Indeed, in the case of IS and AQ their religious ends and their political ends are two sides of the same coin – the Caliphate being a theocracy.

The fifth assumption is in part stipulative and in part conceptual. If fear is to be instilled in some group as a consequence of the harm done to some other group, then the first group needs to know that the second group has in fact been harmed. Accordingly, the terrorist strategy relies on a degree of publicity. Indeed, it might well be that other things being equal the higher the level of publicity the more successful the terrorist strategy is. This certainly was the case with Al-Qaeda’s terrorist attack on the Twin Towers on September 11, 2001.

In the light of these five assumptions I offer the following preliminary definition of terrorism.

By definition, terrorism is a political or military strategy that:

  1. Consists in deliberately using violence against persons;
  2. Is a means of terrorising the members of some social or political group in order to achieve political or military purposes;
  3. Relies on the violence receiving a high degree of publicity, at least to the extent necessary to engender widespread fear in the target political or social group.

I note that on this account a failed attempt at performing a violent act would not count as an act of terrorism; a violent act actually has to be performed in order to count as terrorism. On the other hand, if a group deliberately performs a violent act that receives a high degree of publicity but, nevertheless, fails to actually engender widespread fear and/or bring about political change then it would still count as an act of terrorism. What of a violent act which failed to attract any publicity? Although this is less clear, my inclination is not to count a violent act that failed to attract any publicity as an actual act of terrorism.

The above definition is radically incomplete in that it fails to specify the class of persons who are the targets of terrorism. This issue will be the main topic of this paper. I will consider three possible specifications and reject all of them. (I consider a fourth more plausible specification in a second paper – Part 2 of this paper.) The first of these three specifications is civilian.

2        Definition in terms of Civilians

Civilian contrasts with military, and the implication here is that terrorism is a strategy deployed in the context of war. But when Timothy McVeigh bombed the Alfred P. Murrah Federal Building in Oklahoma City in 1995 killing 168 people he performed an act of terrorism, but he did not do so in the context of a war. McVeigh might have been under the illusion that there was a war and he was a combatant; but, if so, he was deluded. Similarly, the Aum Shinrikyo’s Sarin gas attack on the Tokyo Underground in 1995 that killed twelve people was an act of terrorism, but not one carried out in the context of war. More recently, there have been a spate of IS inspired attacks on ordinary citizens going about their business in well-ordered jurisdictions in liberal democracies, such as France, Belgium, the UK, the US among others. These include the attacks in Paris on a single night in 2015 by gunmen and suicide bombers who hit a concert hall, a major stadium, restaurants and bars leaving 130 people dead and hundreds wounded, a lorry attack in Nice in 2016 which killed dozens of people, and an attack in 2017 in London when three attackers killed eight people by driving a van into pedestrians on London Bridge and then launched a knife attack in Borough Market.

The definition of war is problematic, and no doubt especially when it comes to wars involving a non-state actor as one of the protagonists. Nevertheless, non-state actors can engage in war, e.g., historical wars of revolution, the current war being waged by IS in Syria and Iraq. However, a well-ordered liberal democratic state is not in a condition of war merely because a an individual or group performs an act of violence against the state by killing ordinary office-workers and other citizens going about their day to day lives. Rather such acts are crimes of murder and, specifically, crimes of murder that are also acts of terrorism (given the aim of terrorising people in the service of political purposes). This point is not undercut by the fact that in some cases, such as the Paris, Nice and London attacks mentioned above, these crimes of murder were inspired by an organisation that is engaged in a war elsewhere. After all, attacks such as these will probably continue to be carried out after the cessations of hostilities of a kind and on a scale that constitutes war – for instance if, as seems likely, IS is defeated on the battlefields of Iraq and Syria. Yet these future terrorist attacks are no different in kind from the ones currently being perpetrated.

The general point to be made here is that an act of terrorism is not necessarily performed in the context of war. Accordingly, terrorists are not necessarily (military) combatants ­— even combatants guilty of war crimes — and the combatant/non-combatant distinction (in its military sense) might not relevantly apply to the victims of terrorists. A person killed in the Oklahoma bombing or in the Tokyo Underground attack or in the Paris, Nice or London attacks, and who happened to be a member of the police service, or indeed of the armed services, has not been killed by an enemy combatant in a war; rather he or she has been murdered by fellow civilians in peacetime, just as his or her fellow civilians were murdered. McVeigh, the members of the Aum Shinrikyo, and the IS inspired murderers are terrorists but not combatants. Moreover, police officers who appear on the scene and seek to apprehend such terrorists – or shoot them in self-defence or to prevent further acts of killing — are not combatants (in the required military sense); rather they are police seeking to enforce the law. Hence the actions of the police in question are not governed by rules of engagement for military combat; rather their actions are governed by rules the point of which is to uphold the law and, as far as is possible, preserve life. This is not to say that the distinction might become increasingly fine between the police response and military combat in cases involving groups of heavily armed terrorists in otherwise well-ordered jurisdictions, such as was the case in the Paris attack mentioned above, in particular.[1]

Moreover, this is not to say that some so-called armed struggles conducted by terrorist groups are not wars and that, therefore, some terrorist acts ought to be understood as acts of war performed by combatants. As is well known, many terrorists groups, such as the IRA and IS, engaged in protracted campaigns of violence that generated a military response and significantly disrupted civil society for lengthy periods of time (unlike, say, the actions of Timothy McVeigh). This is so even if, as was the case in Northern Ireland, the military forces happen to be under police control, and therefore civilian authority. Moreover, many of these terrorist groups insisted that they were combatants fighting a war, or at least combatants fighting an armed struggle against enemy combatants. Whether or not the members of the IRA in particular were, or ought to have been, regarded as combatants fighting a war is controversial. At any rate, my point is that in principle there could be — and in fact there have been — armed struggles conducted by terrorist groups that are wars, and in such cases the terrorists are essentially combatants (irrespective of what their legal status might have been).  The wars of national liberation conducted by terrorist groups in Algeria against the French and in Kenya against the British are cases in point as is the armed conflict engaged in by IS in Iraq and Syria. Accordingly, we need to distinguish between terrorists who are essentially combatants engaged in a war and terrorists who are not combatants, but merely civilians performing the ordinary crime of murder in peacetime (albeit in order to instil fear to secure political purposes). Naturally, this distinction is somewhat vague and, in many contexts, difficult to make; nonetheless, the distinction exists, as the McVeigh, Aum Shinrikyo, Paris, Nice and London examples demonstrate.

So the above definition is misleading or, at least, incomplete. It fails to specify the meaning of “civilians” and, relatedly, to distinguish between terrorism as a crime in peacetime and terrorism in the context of a war.

The question that now arises is whether or not the term “civilian” in the above definition ought to be replaced by the term “innocents”. After all the notion of an innocent is, at least on the face of it, independent of whether or not the terrorist or their victims are combatants or non-combatants; so it promises to breach the divide between terrorism in war and terrorism in peacetime. A definition of terrorism in terms of innocents is recommended by a number of theorists, including Igor Primoratz[2].

3        Definition in Terms of Innocents

The definition in terms of innocents needs to offer an account of what counts as an innocent. Here we need to make a threefold distinction between violent attackers, perpetrators of an injustice (other than violence) and revolutionaries (those seeking to overthrow a government, whether by violent or non-violent means). Let us assume that anyone belonging to any of these three categories of persons is not innocent in the required sense. But note that someone could belong to one or more of these three categories and yet their associated actions be (all things considered) morally justifiable. For example, a person could be a combatant (violent attacker) fighting a morally justified war of national liberation. This combatant is not an innocent in the required sense, but their violent actions are (ex hypothesi) morally justified. Again, a person could be attempting, jointly with others, to overthrow an authoritarian government by non-violent but unlawful and somewhat unjust means, e.g. boycotts, disruption of transport and communication systems, work stoppages, occupancy of government buildings etc. Such a person is not an innocent in the required sense, since their actions are, let us assume, harmful to the economy and to law and order, but their actions may well be morally justified all things considered.

Let us refer to the first of these three categories as combatants. As we have seen, the notion of a combatant (in its military sense) is tied to the condition of war. Assume there is a political group perpetrating acts of terror against the citizens of a well-ordered, liberal democratic state in peacetime, e.g. the Aum Shinrikyo releasing nerve gas in the Tokyo underground. Obviously, such terrorists are not attacking enemy combatants; there is no war in progress. Rather they are attacking innocent civilians. Now imagine that there is a war. Assume that there is a military force perpetrating acts of terror against non-combatants, e.g., dropping an atomic bomb on Hiroshma in order to terrorise the Japanese leadership into surrendering. Such terrorists are attacking non-combatants who are also innocent civilians. So far so good for the definition of terrorism in terms of innocents.

Let us consider a putative kind of counter-example to the definition of terrorism in terms of innocents. In September 2001, Al-Qaeda operatives attacked the Twin Towers and also the Pentagon. Leaving aside the deaths of the innocent passengers in the planes used in these attacks, what does the definition under consideration tell us in relation to these two attacks? Contrary to bin Laden’s claims, the Twin Towers principally housed innocents in our above-described sense; accordingly, the attack on the Twin Towers was clearly terrorism by the lights of the definition. Now consider the attack on the Pentagon. By contrast with the Twin Towers, the Pentagon principally housed military and intelligence personnel, i.e., people responsible for assisting Israel in the Israel-Palestinian conflict, maintaining air-force bases in Saudi Arabia, implementing sanctions against Iraq and otherwise engaging in activities regarded by bin Laden and his followers as an attack on Islam. That is, unlike most of the people in the Twin Towers, the people in the Pentagon are not innocent in the required sense. (As noted above, it does not follow from this that these actions of the people in the Pentagon were not morally justified; non-innocence in this sense is not to be equated with responsibility for morally unjustified actions.) Therefore, on the definition of terrorism in terms of innocents, the attack on the Pentagon was not terrorism. This is implausible. Moreover, once again it illustrates an important and already noted defect in the definition under consideration. On this account politically motivated, lethal attacks on public officials, including security personnel, in well-ordered, liberal democratic states at peace will turn out not to be instances of terrorism just so long as they were performed in order to address some  injustice perpetrated by, or some violent attack initiated or supported by, those officials.

Let us now consider a rather different, kind of counter-example to the definition of terrorism in terms of innocents. Imagine a non-democratic, indeed highly authoritarian, government pursuing policies that are manifestly and seriously unjust, e.g. policies that impoverish and disempower some sections of the community while privileging other sections. Assume that well-intentioned democrats with a social conscience attempt to mobilize opposition to the government; opposition in the form of non-violent protests, strikes, boycotts, sit-ins, dissemination of anti-government material, passive non-compliance, and so on. These opposition elements are seeking to overthrow the government, indeed the system of government, albeit by non-violent means. The ANC in its initial non-violent phase prior to the 1960’s is a case in point. Accordingly, they are not innocents in the required sense. (Indeed, from the perspective of the authoritarian government and its substantial cohort of supporters among the citizenry, these opposition forces are engaged in attempting to overthrow the legitimate government of the country.) Moreover, they may well succeed in their revolution if harsh counter-measures are not introduced. Accordingly, the government embarks on a campaign of killings (‘disappearances’) and torture of opposition elements in order to instil fear in the opposition forces as a whole, and thus put an end to the ‘insurrection’. Surely this is state terrorism of the kind practised by the Argentinian Generals in the 1980’s, the Assad regime in Syria at the present time (who have even used chemical weapons against sections of the civilian population deemed to be supportive of the opposition) and (to a lesser extent) by the apartheid government in South Africa against the ANC. Nevertheless, it remains the case that the opposition forces are responsible for attempting to overthrow the government, albeit initially by non-violent means; and the government believes itself — and is believed by many, let us assume — to be legitimate. By the lights of the definition of terrorism in terms of innocents, the killings and torture perpetrated by the government is not terrorism, since the opposition forces are not innocent in the required sense.

What response is available to defenders of the definition? One such response to both of the above kinds of counter-example (i.e., examples involving killing public officials in liberal democratic states and ones involving killing insurrectionary democrats in authoritarian states) is to add the further condition that unless the violent response is, objectively speaking, morally justified, then the violent response is terrorism, notwithstanding the fact that the target group is not, strictly speaking, innocent. If this condition is made then: (a) the killers of public officials in liberal democracies turn out to be terrorists, since their murderous actions are not morally justified, and; (b) the insurrectionary democrats in our example turn out to be victims of terrorism – after all, for surely it is completely morally unacceptable to kill and torture those who are simply using non-violent means to establish a democracy and implement policies of social justice (even if these insurrectionary democrats ultimately use violent means by way of self-defence).

The upshot of this manoeuvre is that there are now two mutually exclusive categories of terrorism: (1) terrorism in which the victims are innocent in the initial sense of the definition, e.g. children killed in marketplace bombings conducted by Hamas; (2) terrorism in which the victims are not innocent in the initial sense, but in which the violent response of the terrorists is not morally justified, e.g. members of the ANC who are victims of the state terrorism of the apartheid government.

If terrorism is defined in terms of innocents in the initial sense it is open to the objection that authoritarian regimes, such as that of Assad, do not engage in terrorism when they murder, torture and/or use chemical weapons against those sections of the population who oppose them. Since those who seek to overthrow authoritarian regimes are typically engaged in violent or otherwise harmful action they are not innocent in the required sense, even if their actions are morally justified all things considered. On the other hand, if the notion of an innocent is widened so as to include all those whose actions are morally justified (all things considered) then a pandora’s box is opened up. What is now required is a specification of the apparently open-ended notion of an action that is morally justified (all things considered). Moreover, a string of counter-examples now come into view. For instance, combatants fighting an unjust war may well now be defined as terrorists merely by virtue of the fact that they are killing enemy combatants who are fighting a just war and are, therefore, now deemed to be innocents.

4        The Definition of Terrorism in Terms of Non-Combatants

In the light of the failure of the above definition of terrorism in terms of innocents, let us consider the following one in terms of non-combatants. The notion of a combatant and, therefore, of a non-combatant, might seem to provide the required degree of determinacy; unlike the notion of innocents or of morally legitimate/illegitimate targets. The notion of a combatant is both in itself relatively determinate, and it provides a clear sense in which a person is guilty and, therefore, a legitimate target of violence. So on this view, if one is a combatant, then one is using lethal force in the context of some conflict. On the other hand, on this view, if one is not a combatant then one is innocent; non-combatants do not use lethal force and, therefore, are innocents. Terrorism is, on this account, distinguished from other modes of organised political violence in that it targets non-combatants.

Let us, then, offer the following definition of terrorism. By definition, terrorism is a political or military strategy that:

  1. Consists in deliberately using violence against non-combatants;
  2. Is a means of terrorising the members of some social or political group in order to achieve political or military purposes;
  3. And relies on the violence receiving a high degree of publicity, at least to the extent necessary to engender widespread fear in the target political or social group.

The concept of a non-combatant is different from the concept of being innocent, or of not being morally responsible for attacking or infringing the rights of others and so on. As a consequence, non-combatants might, or might not, be innocent in the sense of being morally responsible for acts of violence, injustice or other forms of wrongdoing. Clearly some non-combatants are innocent, e.g., young children, and some instances of using deadly force against non-combatants, such as bombing restaurants (Hamas) or napalming villages (US forces in the Vietnam War), are acts of terrorism — indeed, morally unacceptable forms of terrorism. However, it does not follow from this that all instances of using deadly force against non-combatants are acts of terrorism and/or are morally unacceptable.

Prima facie, some non-combatants are guilty in the sense that they are morally responsible for the attacks, rights violations or injustices that some group of putative terrorists are responding to. The leaders of combatants, for example, and those who assist combatants qua combatants, such as munitions workers and spies, are guilty in this sense; specifically, they are morally responsible for attacking or waging war against such a group (or contributing to such an attack or war effort).

For the sake of argument let us assume that such leaders and contributors are a species of combatant, on the grounds that they are jointly morally responsible with the combatants for the killings (and perhaps other serious violent acts) carried out by the combatants themselves.  This assumption carries with it the implication that the key moral notion in play is not that of the bare notion of a combatant per se, but that of someone who is morally responsible for a lethal attack, whether or not he or she was the one doing the actual fighting. So the definition of terrorism in terms of non-combatants is now helping itself to the notion of a person who is morally responsible for a lethal attack; it is no longer simply using the narrower and less morally loaded notion of a combatant.

The problem that now confronts the definition of terrorism in terms of non-combatants is that it faces a further extension to the category of legitimate targets. On the revised definition of terrorism in terms of non-combatants, moral responsibility for lethal attacks is sufficient grounds for justifying a lethal response, i.e., for forfeiting one’s civilian immunity. If so, then the question that now must be asked is whether or not there are other additional categories of guilty persons who might likewise have forfeited their right to civilian immunity, namely, persons who are morally responsible for serous rights violations or injustices. Consider, for example, corrupt senior government officials and civil servants who fail to organise the distribution of aid in the form of medicine and food to their starving, disease-afflicted, fellow citizens, but rather sell it to line their own pockets. Suppose the foreseen consequence of this corruption and dereliction of their humanitarian duty is that tens of thousands of the needy die. These officials are not combatants in the required sense; they are not themselves soldiers engaged in an armed attack, nor are they the leaders of such combatants or assisting such combatants qua combatants. Accordingly, targeting these public officials would be, on the definition before us, terrorism. But these officials are guilty in the sense that they are morally responsible for ongoing, widespread and serious rights violations. Moreover, using lethal force against some such officials in order to instil fear in their fellow guilty officials, and thereby bring about a cessations to these ongoing, widespread and serious rights violations, may well be, under certain circumstances, morally justifiable.

The question that now arises is whether or not those who respond with lethal force to serious, non-violent rights violations are terrorists. I suggest that the advocates of the definition of terrorism in terms of non-combatants face a dilemma in attempting to address this question. There are two salient options for them. They could stand firm and refuse to vary their definition in the face of this kind of example; in doing so they would be insisting that those who respond with lethal force to ongoing, widespread and serious rights violations of this kind are, nevertheless, terrorists. Alternatively, they could concede (rightly or wrongly) that the latter are not terrorists, and seek to adjust their definition. Consider the latter option first.

The definition could be adjusted by narrowing the notion of a non-combatant so as to exclude those guilty of serious rights violations. However, in doing so the definition is no longer a definition that distinguishes terrorists from non-terrorists on the basis of the combatant/non-combatant distinction. For by the lights of the adjusted definition, although all terrorists target non-combatants, some non-terrorists also target non-combatants, namely non-combatants who are rights violators. Accordingly, to accept this adjustment to the definition amounts to an abandonment of the definition.

Alternatively, the definition could resist this adjustment and insist that those who target only rights violators are, nevertheless, terrorists. However, this refusal comes at a price. For we are owed an account of why the difference between bearing moral responsibility for lethal attacks and bearing moral responsibility for serious rights violations (including fatal ones) should make a difference when it comes to defining terrorism. On the view in question someone who responds with lethal force to lethal attacks is ipso facto not a terrorist. However, someone who responds with legal force to ongoing and widespread, serious rights violations (including fatal ones) is ipso facto a terrorist. Ex hypothesi there is not, or might not be, any significant moral difference between the two responses; that is, as with lethal attacks, it could be morally justifiable to respond with lethal force to ongoing, widespread and serious rights violations. So the challenge to proponents of the definition of terrorism – the definition in terms of non-combatants – to identify a relevant difference between groups who target only rights violators and those who target only combatants is unmet.

I conclude that the dilemma facing the definition of terrorism in terms of non-combatants is unresolved, if not unresolvable.

[1] Seumas Miller Shooting to Kill: The Ethics of Police and Military Use of Lethal Force (Oxford University Press, 2016) Chapter 5.

[2] Igor Primoratz “What is terrorism?” in  Igor Primoratz (ed.) Terrorism: The Philosophical Issues (New York: Palgrave Macmillan, 2004) p. 24.