Does terrorism create terror?

Abstract

The conceptual and etymological meaning of the terms ‘terror’ and ‘terrorism’ is so closely connected that it seems difficult to distinguish the one from the other. However, by comparing the idea that terrorism inevitably creates terror with the results of recent empirical studies of the psychological effects of terrorist attacks, four different points of view emerge corresponding to four distinct interpretations of the results. It is thus clearly necessary to redefine the terminology relating to terrorism, in order to conduct the discussion on a more rigorous basis.

Keywords:

• terror
• terminology
• post-traumatic stress disorder
• bomb attacks
• 11 September 2001

Introduction

The idea that terrorism creates terror is so well entrenched in people's minds that it constitutes the basic premise of the discipline devoted to studying the phenomenon. This apparent truism has remained largely unquestioned until very recently because it was impossible to quantify the psychological effects of terrorism. However, the growing number of published investigations into the widespread occurrence of post-traumatic stress disorder (PTSD) among groups exposed to the most violent attacks at the beginning of the twenty-first century now provides a solid basis for discussion. Analysis of the results of these publications has proved to be most useful in evaluating the pertinence of the postulate, while at the same time it has contributed to a re-appraisal of the concept of terrorism.

Terror and terrorism

Etymologically, ‘terrorism’ derives from ‘terror’. We know that the term ‘terrorism’ was coined in the course of the French Revolution to describe the actions and policy of the government during the period designated as the Reign of Terror. What is rarely mentioned are the semantic changes which led to the emergence of the neologism. The term derives from the Jacobins' positively connoted use of the word ‘terror’ as a means of justice. However, after the downfall of Robespierre, the two new political terms – ‘terrorist’ and ‘terrorism’ – appeared almost simultaneously in September 1794. The use of these terms in the thermidorian press was provocative, if not politically insulting, and aimed at stigmatising the perpetrators of the Terror, while at the same time defending the legitimacy of the Convention (Geffroy 1979, p. 131). From the outset, the term ‘terrorism’ had a derogatory connotation and was assimilated into the discursive strategies used to censure political opponents.

Two successive semantic shifts extended the meaning of the notion of terrorism, one to describe other historical uses of terror as a method of governing, ¹ and the other to designate a specific form of subversive violence. This latter semantic shift, which was decisive for future discussions of terrorism, occurred at the end of the 1870s in Russia when the Norodnaja Volja (The Will of the People) explicitly defined terrorism as a strategy for fighting against an autocratic system of government. In its 1879 programme, the executive committee of the organisation defined the aims of the ‘terrorist action’: to eliminate the most important people in power, to demonstrate that it is possible to fight against the government, to promote the revolutionary spirit among the people, and to protect the party from spies. ² According to Stepniak-Kravchinski (1883, p. 59), the Russian revolutionary movement's new orientation rapidly resulted in establishing terrorism as a ‘system’, not of government, but of the struggle against the ruling powers. Most of the movement's leaders saw revolutionary terrorism as a specifically Russian method developed to fight against the archaic political structures of the empire. However, after leaving Norodnaja Volja, Nikolaj Morozov, one of the main theoreticians of the terrorist movement, proposed a wider use of the revolutionary method. From his exile in Switzerland, he conceptualised what Venturi (1983, p. 673) later described as ‘pure terrorism’. Without going into the details of Morozov's thesis, it is important to underline the universal nature of his ideas, since he envisaged recourse to terrorist action against any form of oppression, from absolutism to the brute force of any constitutional form of government. ³

In the Russian context, the term ‘terrorism’ briefly lost its pejorative connotation. The word became derogatory again when the Marxist wing of the revolutionary movement criticised the failure of the campaigns of terrorist attacks. Few activists later used the term to describe their own use of violence. From the end of the nineteenth to the beginning of the twentieth century, the term was progressively used to designate subversive action against symbols of bourgeois power, whether in Europe, North America, or the colonies. This was the dominant connotation of ‘terrorism’ in the twentieth century, but it never supplanted its etymological meaning, which was constantly reactivated when condemnation of terrorist attacks against symbols of power led to recurrent reminders that terrorism is also one of the tools of power. The diversity of the term's meanings makes it admirably suitable for rhetorical use since it can be employed in discourse strategies aimed at challenging the legitimacy of an adversary who resorts to violence perceived to be politically incorrect.

In a polemical context, or when the term is used in everyday language, the tangible effects of terrorism do not really enter into consideration. The negative connotation of the word suffices to denounce the action, and its underlying sense – to frighten the adversary – apparently satisfies the semantic requirements of the term. The notion that, by definition, terrorism creates terror was deemed to be so obvious that in 1937, when lawyers at the League of Nations drew up the Convention for the Prevention and Repression of Terrorism, they defined the phenomenon, in the first Article, as: ‘criminal acts directed against a State and intended or calculated to create a state of terror in the minds of particular persons or a group of persons or the general public’ (Saul 2005, p. 63). Criticised from the outset for being too general and tautological, ⁴ this definition, nevertheless, formed the basis of contemporary discussion on terrorism.

Since the 1970s, various definitions have rephrased the statement contained in the 1937 Convention. Authors have sometimes tried to overcome the tautological nature of the enunciation by using different circumlocutions. According to Hacker (1976, p. 23), terrorism is a form of aggression which engenders ‘a mental state characterized by indefinable fear and desperation’, and which aims to ‘scare those in power’. Wilkinson (1973, p. 291) stated that terrorists ‘attempt to instil maximum fear and uncertainty among members of the community or group they seek to intimidate’. Wardlaw (1982, p. 16) considered that the terrorist act aimed ‘to create extreme anxiety and/or fear-inducing effects in a target group’. As for Howard (1992, p. 156), he stated that terrorism is ‘the systematic and premeditated use of violence to create a climate of extreme fear for political purposes’. More recently, Butler et al. (2003, p. 24) and Burke (2008, p. 39) made a similar assertion by associating the notions of terror and fear to describe the methods used by terrorists to achieve their political objectives.

Since the 1930s, there have been countless definitions of terrorism which do not need to be examined in detail here. The result of decades of recurrent discussion and controversy is generally held to be unsatisfactory (Smelser 2007, p. 239). As Silke (2004a, p. 207) has emphasised, continued failure to reach an agreed definition constitutes the clearest indication that research in this field lacks a common conceptual framework. In this respect, we should perhaps examine the idea that the root of the problem lies in the notion of terror itself, both from a definitional point of view and from the perspective of developing a conceptual framework. ⁵

The relevance of the concept of terror has only rarely been contested. The scope of the concept has been questioned, for example, by the American psychiatrist Ochberg commenting on a definition of terrorism submitted to him by Alex Schmid in the 1980s. He wrote: ‘The equation of terror with a state of chronic fear is permissible in lay language, but in psychiatry terror is an extreme form of anxiety, often accompanied by aggression, denial, constricted affect, and followed by frightening imagery and intrusive, repetitive recollection’ (Schmid and Jongman 1988, p. 19).

Schmid and Jongman responded to this criticism, considering it to be ‘significant and accurate’, especially since they had to admit that terrorist groups were rarely capable of keeping a whole population in a permanent state of fear. They nevertheless continued to support the pertinence of the concept of terror:

While violence is the key element with murder, it is the combination of the use of violence and the threat of more to come which initiates a terror process. Many extremists might often not be able to produce a prolonged terror effect by unexpected, dramatic acts of violence; however, the fact that the evocation of terror is their intent is sufficient to justify placing them in the same category as those who succeed. (Schmid and Jongman 1988, p. 19)

Even though their reasoning was obviously questionable, since it was based on the principle that the initiator's intention is sufficient to characterise a social phenomenon independently of its observable effects, it was necessary for the coherence of their argument. No justification was offered by Schmid and Jongman, but it was implicit in the subtitle of the section responding to Ochberg's criticism: ‘Terrorism without terror ?’. An affirmative answer to this question would have called into question the analytical utility of the concept of terrorism.

Horgan (2005, p. 14) restated the terms of the controversy but did not arrive at a definition. Like Schmid, he considered that the ‘immediate goal of most terrorist groups [is] to cause terror’ and admitted that ‘to maintain a constant atmosphere of terror is difficult even for the most belligerent terrorist groups’. On this basis, Horgan stated that the most daunting challenge confronting terrorists is ‘to terrorize effectively’. In other words, he looked at the problem of terror first and foremost from the point of view of the terrorists themselves. However, he previously made two interesting observations directly linking the question to a conceptual context. The first concerned the actual psychological effects of violence: ‘While we might experience arousal from hearing of a terrorist attack (for fear that there may be further attacks, for example) we will not experience the terror that the terrorists seek to achieve unless we are subject to an attack, and even then, it tends to be of limited duration’. The second referred to an inherent paradox: ‘On reflection, it may be surprising that despite our exposure to terrorist events via the printed or television media, our memory of terrorist events appears to recede swiftly’ (Horgan 2005, p. 14). Whether or not terrorist attacks effectively do create a climate of terror remains an open question, since although the presumed intention of terrorists is to ‘terrorise’ the population, their projects have never been entirely successful.

‘Substantial stress’ and PTSD

Ochberg's opinion did not gain unanimous support among psychiatrists. The authors of A Critical Dictionary of Psychiatric and Mental Health Terminology stated that ‘by extension’ the notion of terror could refer to the ‘state of fear that is imposed on a population in order to break down its resistance’ (Kipman et al. 2005, p. 399), a definition which was taken from a language dictionary (Rey and Rey-Debove 1985, p. 1949). At first sight, the definition seemed reasonable and compatible with the ideas prevailing among specialists studying terrorism. The situation changed after the devastating terrorist attacks carried out at the beginning of the twenty-first century, which gave rise to a number of empirical studies. The results of these studies now provide the basis for a wider discussion.

The first wave of research into the psychological impact of terrorism took place in Northern Ireland and Great Britain. In a review of the research literature on the topic, Curran (1988, p. 473) concluded that: ‘Judging from hospital referrals and admission data, suicide and attempted-suicide rates, the practice of psychoactive drug prescriptions, and community-based studies […] the campaign of terrorist violence does not seem to have resulted in any obvious increase in psychiatric morbidity’. This conclusion reiterated an observation made in a scientific study carried out by Cairns and Wilson (1984, p. 635) according to which ‘The majority of people in Northern Ireland manage to deal effectively with stress generated by the troubles, but […] a small minority are at risk’.

In actual fact, studies focusing specifically on subjects directly exposed to violence present a less clear-cut picture, with sometimes inconsistent results. Following the Old Bailey bomb explosion (8 March 1973), Caro and Irving (1973, p. 1433) found that 15% of the victims were emotionally shocked, whereas Hadden et al. (1978) estimated that emotional distress affected 50% of subjects who received treatment following bomb attacks. These early conflicting results were partly due to an imprecise methodology, a situation that improved noticeably during the years 1980–1990 and especially after the introduction of a scale for measuring the symptoms of PTSD in victims of terrorism in general (Loughrey et al. 1988), and bomb attacks in particular (Curran et al. 1990). Although research was subsequently conducted within a more rigorous framework, the analysis and interpretation of results still require a degree of caution.

When measuring the psychological effect of terrorist attacks on a given population, significantly different results are produced depending on the criteria taken into consideration. For this reason, it is always necessary to examine the results in the context of the methodology used in the investigation, and to interpret the information according to the type of psychological effects the researchers aimed to measure.

A first group of studies based on improved methodologies examined the symptoms of stress that appeared after a bomb attack. In these studies, a sample population was asked to reply to five questions: (1) Feeling very upset when something reminds you of what happened? (2) Repeated, disturbing memories, thoughts, or dreams about what happened? (3) Having difficulty concentrating? (4) Trouble falling or staying asleep? (5) Feeling irritable or having angry outbursts? Each participant evaluated their experience by choosing from a scale ranging from 1 (not at all) to 5 (extremely). If the participant chose 4 (quite a bit) or 5 for any question, he/she was deemed to be suffering from substantial stress. On this basis, Schuster et al. (2001) found that 44% of a nation-wide sample of the population of the USA were affected by substantial stress during the five days following the 11 September 2001 (9/11) terrorist attacks. Three months after the event, this dropped to 21% (Schuster et al. 2004). Vázquez et al. (2006) recorded substantial stress in 59.2% of the population of Madrid during the three to four weeks following the 11 March 2004 (3/11) attacks, and Rubin et al. (2005) found that 31% of a London sample experienced substantial stress 11 to 13 days after the bombings on 7 July 2005 (7/7). According to Vázquez et al. (2006), these results were based on a very general definition of stress and therefore presented an excessively alarming picture which could lead to confusion. They also pointed out that in all the studies, when all criteria were taken into account, the average response of the participants did not reach the level of 2 (a little bit), ‘which casts some doubts about catastrophic discourses on the implications of this disaster for the general population’ (Vázquez et al. 2006, p. 69). Rubin et al. (2005, p. 607) noted that for four of the most severe symptoms cited (irritability, sleeping problems, difficulty in concentration, recurrent thoughts, or dreams relating to the events) substantial stress was mentioned by only 10% of the sample, whereas the mildest symptom (feeling upset by the memory of the event) was noted by 25%. It should also be pointed out that the questionnaire used by Rubin et al. (2005, p. 607), contained the question ‘feeling upset’ and not ‘feeling very upset’. One could argue that what is really surprising is that two weeks after the 7/7 bombings, 75% of London's population did not feel particularly upset by the memory of the event.

A second group of studies measured the prevalence (or probable prevalence) of PTSD, based on the Diagnostic and Statistical Manual of Mental Disorders (4th edition 1994, or revised 4th edition 2000). Five criteria, and 17 symptoms divided among three of the criteria were taken into consideration:

• Criterion A1. Exposure to a traumatic event that involved physical threat.

• Criterion A2. Subjective reactions of fear, helplessness, or horror.

• Criterion B. Re-experiencing the event (1 out of 5 symptoms):

  1. Intrusive recollections.

  2. Recurrent, distressing dreams.

  3. Acting or feeling as if the event was recurring.

  4. Distress at exposure.

  5. Physiological reactivity on exposure.

• Criterion C. Persistent avoidance (3 out of 7) :

  1. Efforts to avoid thoughts, feelings, or conversation associated with the trauma.

  2. Efforts to avoid activities, places, or people that arouse recollections of the trauma.

  3. Inability to recall important aspects of the trauma.

  4. Diminished interest to participate in significant activities.

  5. Feelings of detachment from others.

  6. Restricted range of affect.

  7. Sense of foreshortened future.

• Criterion D. Hyper-arousal (1 out of 5):

  1. Insomnia.

  2. Irritability or outbursts of anger.

  3. Difficulty concentrating.

  4. Hyper-vigilance.

  5. Exaggerated startle response.

• Criterion E. Significant distress or social impairment.

Two different strategies were used to measure PTSD. The first took into account only the symptoms attested by the subjects – criteria B, C, and D. The second included all the diagnostic criteria listed above.

Using the first strategy, Schlenger et al. (2002) observed that during the two months following the 9/11 attacks, PTSD symptoms occurred in 11.2% of the population of New York, in 2.7% of the inhabitants of Washington, and 4% of the population nation-wide. Galéa et al. (2002) focused their attention on the inhabitants of Manhattan, where they recorded PTSD symptoms in 7.5% of the subjects during the five to eight weeks after the event. In two further diagnostic tests, they found that this figure dropped to 1.7% after four months and 0.6% after six months, but this was in a sample population in which residents of Manhattan were over-represented (Galéa et al. 2003). When interpreting these results, it is important to take into account the permanent presence of PTSD symptoms in the American population, independent of any events connected with terrorism. Galéa et al. (2002, p. 985) stated that in the years preceding the 9/11 attacks, PTSD symptoms affected an average of 3.6% of the whole population of the USA. However, the Galéa studies do not differentiate clearly between the symptoms which were apparent before the 9/11 events and those directly connected with the attacks. Might we not infer that in the first group of post-9/11 PTSD studies, the symptoms recorded by subjects directly affected by the attacks reflect an accumulation of new symptoms and others that pre-existed? If this is the case, should we then logically assume that in the two later series of studies, this same category of subjects differentiated between those symptoms directly related to 9/11 and others related to later events? If the subjects did effectively differentiate between them, it indicates that for a significant section of the population, the 9/11 attacks reinforced existing symptoms, but also that the psychological effects of the bombings diminished rapidly in the face of other sources of stress.

In their analysis of the reactions to the ‘3/11’ bombings, Vázquez et al. (2006) distinguished two levels. The first level comprised subjects in Madrid who scored 44 or more points out of a total of 85 allotted on a scale of one to five for the 17 symptoms described; 13.3% of the sample population attained this level. The second level required a score of 50 or more points. This was reached by 3.4% of the sample. In another study, Bleich et al. (2006) studied the impact of a prolonged period (four years) of terrorist attacks on a representative sample of the Israeli population. The sample included subjects who were directly exposed to the attacks (11% of the sample), others whose family members or friends were directly implicated (20%), and subjects who were not directly exposed themselves, nor were any of their family members or friends (69%). In this sample, 9% selected the criteria corresponding to PTSD symptoms, in other words, a percentage identical to that reported for a similar study carried out in 2002 (Bleich et al. 2003).

Using the second strategy for measuring PTSD, Miguel-Tobal (Vázquez et al. 2006, p. 71) estimated that 4% of the population of Madrid satisfied all the criteria of PTSD and that 2.3% developed symptoms directly related to the 3/11 attacks during the four to 15 weeks following the events. Referring to this same population sample, Vázquez et al. (2006) estimated that 1.9% of the subjects satisfied the clinical criteria corresponding to PTSD.

Finally, let us look briefly at the research results of a group of studies based exclusively on people directly involved in terrorist attacks. Shalev et al. (1992) reported that 33% of the survivors of a bomb attack against an Israeli bus on 6 July 1989 suffered from PTSD. In another study concerning the 1995–1996 terrorist attacks in France, Verger et al. (2004) observed that two years after the events, 31% of the survivors presented symptoms that tallied with PTSD criteria. A study by Njenga et al. (2004) concluded that during the three months following the events, these same criteria were valid for 35% of the survivors of the 7 August 1998 bombing of the American embassy in Nairobi. An identical percentage was reported by North et al. (1999) for victims of the 19 April 1995 Oklahoma City attack.

Discussion

This discussion will concentrate on the psychological effects of violence on subjects who were not directly exposed to it, since the problem is not to establish whether or not the immediate victims of terrorist attacks experience terror, but rather whether it is reasonable to generalise about the existence of this state of extreme fear among an entire population confronted with violence. The fundamental problem facing researchers into terrorism as a field of study is to establish a distinction between the passive target of the perpetrators (the direct victims) and two other categories: the indirect target (the general population) and the active target (the authorities).

The various results presented above lend themselves to two different kinds of interpretation. Friedman (2005), for example, notes that the psychological impact of terrorism is greater than that caused by natural disasters; that it affects a far greater number of people than those directly involved in the physical destruction of a bombed target; and that PTSD only partially accounts for the state of anxiety generated by such events. Similarly, Lamberg (2005, p. 544), relying mainly on the results of Galéa et al. (2003), considers that terrorism ‘assails a nation's psyche’, and notes that a substantial proportion of the population studied after 9/11 displayed clinical symptoms linked to the terrorist events. On the contrary, Delahanty (2007, p. 190) judges that these same results ‘may seem surprisingly low’ – even though they actually concerned a considerable number of people – and adds: ‘Terrorism significantly and persistently affects a small percentage of both directly and non-directly exposed individuals’. This perspective is similar to that of Vázquez et al. (2006, p. 71) who emphasised that immediately following the 9/11 and 3/11 attacks, both the American and Spanish authorities predicted an upsurge of mental problems in the public health sector. However, these predictions were not fulfilled. ⁶ In other words, while the theoretical discussion of the impact of terrorism led to the belief that the psychological consequences would be proportional to the intensity of the attacks, the attacks turned out to produce no objectively verifiable major public health crisis.

Another idea conveyed by theoretical discussions but not corroborated by empirical research, maintains that if terrorist attacks occur over an extended period of time, they will increase the level of fear in the general population. Bleich et al. (2006, p. 9) clearly show on the contrary that the percentage of PTSD sufferers in Israel did not increase from 2002 to 2004. They note: ‘These findings are interesting, as one would expect increased rates as a result of more people being exposed over time, and the development of delayed and reactivated PTSD as a result of the repeated terror attacks’. They also comment on the interesting tendency towards a lowering of anxiety levels among the general population (a significant reduction in the number of individuals fearing for their lives, from 60% in 2002 to 47% in 2004; fearing for the lives of their friends and family, from 67.9% to 54.1%; or feeling demoralised, from 58.6% to 29.5%), which Bleich et al. (2006, p. 9) attribute to a process of adaptation to violence. Thus they state: ‘These figures suggest that even though the ongoing terror continues to rob much of Israel's civilian population of their peace of mind and sense of safety, fewer people than previously feel significantly threatened by it’. Nevertheless, the same authors postulate a slow attrition of resilience among the Israeli population, expressed by a growing need for psychological treatment (9.7% in 2004 compared with 5.3% in 2002), and a lesser degree of optimism concerning the future of the State of Israel. Yet there is an underlying paradox in the Israeli situation, since only 56% of the subjects admitted to being optimistic about the future of the state of Israel in 2004 (compared with 66% in 2002), while at the same time 82% pronounced themselves optimistic about their own personal future (the same percentage as in 2002). Finally, we should mention the work of Shalev and Freedman (2005) which invalidates the hypothesis according to which a reduction of PTSD symptoms can be correlated with the frequency of terrorist attacks. In the Shalev and Freedman study, anxiety levels reported by subjects diminished in the same proportions during a period marked by an increase in terrorist attacks as during a period marked by a limited number of attacks.

A comparison of studies concerning the psychological impact of terrorism and the idea that terrorism creates terror leads to four different points of view. The first, expressed by Schmid and Jongman in their commentary on the work of Ochberg, tends to minimise the results obtained from various empirical studies. They maintain that what counts is not the physical impact of terrorist bombings, but rather the intention of the perpetrators (Schmid and Jongman 2005, p. 19). A counter-argument to this point of view has already been stated, namely, that the description of a social phenomenon does not depend solely on the intention of the instigator, but also on the practical consequences of the action. I would add a second criticism which challenges the basic premise on which Schmid and Jongman based their case: it is not at all certain that all members of the social group categorised as terrorists aim to create terror. Those who take hostages illustrate my argument. The aim of hostage-takers is neither to terrify their victims nor to intimidate the general population. Their objective is rather to put themselves in a strong position in order to negotiate their demands with the authorities, who are not ‘frightened’ by the situation, but are nonetheless subjected to pressure and forced to make a series of rational choices in order to extricate themselves and the hostages from an uncomfortable state of affairs.

A second point of view holds that the results of empirical investigation tend to show that terrorism indeed creates terror. The incidence of PTSD among subjects who were not directly exposed to violence indicates that the psychological effects of terrorist attacks extends beyond the direct victims and that a wider section of the population experiences symptoms identical to those reported by people present at the scene of the violence. The fact that these effects are limited to a fraction of subjects indirectly exposed to the violence does not in itself invalidate the hypothesis, since it is unrealistic to expect an entire population to be uniformly affected.

However, in many respects this point of view poses problems. First, the presence of identical symptoms among subjects who were either directly or indirectly exposed to violence does not necessarily imply that both categories underwent similar experiences, even less that their individual involvement was similar. Yehuda et al. (2005, p. 1793) have emphasised that research has so far only partially investigated this question. As they note: ‘One of the gaps in our knowledge concerns whether the effects on the mental health of persons indirectly exposed to terrorism are qualitatively or quantitatively different from those who have been directly exposed’. Second, there is the question of the threshold at which one can identify the phenomenon of ‘terror’ or ‘collective fear’, taking into account the fact that only a relatively small percentage (less than 10%) ⁷ of subjects in the population is affected by PTSD. Third, it should be noted that the results obtained so far have related to particularly spectacular and deadly attacks. It therefore seems difficult to assert that other events, although also classified in the category of terrorism, might be linked to the phenomenon of collective fear among significant sections of the population. Fourth, research carried out in Israel has produced contradictory results, indicating that a prolonged terrorist campaign has tended not to increase the levels of anxiety among the population, but has rather reinforced their capacity to adapt to higher levels of violence. In this respect, the claim made by Bleich et al. (2006, p. 8) that ‘fewer people than previously feel significantly threatened by [the ongoing terror]’ also illustrates the prevalence of imprecise terminology in research on terrorism. It would have been more accurate to assert that ‘fewer people than previously feel significantly threatened by deadly attacks’, rather than to link the idea of ongoing terror to a lower level of anxiety in the general population.

A third point of view, suggested by Friedman (2005), is that PTSD is not a suitable tool for measuring the impact of terrorism beyond the circle of direct survivors of violence. Indeed, if one takes other evaluation criteria into consideration the percentage of people affected by violence increases significantly. In a study carried out by Pat-Horenczyk (2005), from 30% to more than 50% of the subjects were found to be affected by ‘substantial stress’ and 63% of Israeli adolescents questioned in the context of the study stated that, given the frequency of terrorist actions in Israel, they experienced a high level of fear, horror, and helplessness. However, we should not consider that ‘stress’ is equivalent to ‘terror’ without raising once again the problem of accurately defining the terms appropriate for describing the psychological impact of violence. Neither should we highlight subjects' responses to a single question in an inquiry, while at the same time glossing over contradictions which emerge from an analysis of the global results. In the case of Pat-Horenczyk's study (2005, p. 346), the PTSD diagnostic test was administered to only 5.1% of the subjects, thus providing an accurate picture of the resilience of Israeli youth who lived with recurrent violence. Another contradiction is apparent in the work of Bleich et al. (2006) who state that while 47% of subjects questioned assert that they fear for their lives, only 29% experience feelings of tension and anxiety. Finally, it is worth pointing out that one specific section of the population, namely, government and political authorities, has never undergone systematic investigation, yet they are the main people to be concerned by politically motivated violence. An accurate assessment of their reactions would be of great interest in evaluating the effective impact of violence on one of the targeted categories.

A fourth point of view is that notions of terror and terrorism are merely terminological approximations which imperfectly express the anxiety-provoking nature of the phenomena to which they refer. We cannot presuppose that the phenomena are uniform, in the sense that their anxiety-provoking characteristics are a function of the type of attack which is perpetrated. Similarly, we cannot consider that their psychological impact would uniformly affect subjects who were not directly exposed to the violence. Ganor (2005) states that the aim of a terrorist attack is to influence public opinion, sap people's spirits, undermine their confidence and sense of security, and create a climate of panic and anxiety. This analysis seems reasonable to the extent that it enlarges the scale of reactions to violence without considering that the critical threshold of ‘terror’ is necessarily attained. However, it does raise the following question: If terrorism does not systematically and literally create terror, is the concept itself still analytically useful? It is difficult to conceive that a field of scientific investigation could be organised on the basis of imprecise terminology.

The question is crucial and it is particularly complex because the term ‘terror’ is simultaneously applied to different factors coexistent in violence, namely: (1) the intentions of the instigators; (2) acts of violence per se ⁸ ; and (3) the psychological impact of the events. The use of the term in a general sense presupposes a combination of these three factors and is associated with a specific type of violence. In addition, certain types of discourse concretise the notion of terror, as though it referred to a specific object rather than an abstract, imperfectly defined, concept. In reporting sudden outbreaks of violence, the need to consider the niceties of semantic meaning is easily overlooked. Serious terminological complications arise in advanced discussions when, for example, an attempt is made to objectively analyse the different factors and actors involved in violence, in which case ‘terror’ (in the general sense) does not necessarily result in ‘terror’ (in its psychological sense). Similarly, applying the word in its general sense to volatile contexts may be enough to generate fear, whether the presumed threat is real, imagined, or simply invented.

In effect, while the polysemy of ‘terror’ may lead to terminological confusion, it is this very confusion that makes it a word of choice in government communication strategies. This factor weighs heavily in the discussion about the pertinence of the concepts of terror and terrorism. Mueller (2006) has postulated that in the United States the exaggeration of a security threat is more likely to increase the level of fear than the potential ‘terror’ itself; in other words, the discussion of terrorism generates greater anxiety than the actual violence. Nevertheless, it would be wrong to over-emphasise the link between the semantic imprecision surrounding these concepts and their potential for political manipulation. Fundamentally, the difficulty resides in finding a vocabulary for adequately describing extreme violence and in understanding the underlying nature of the threat that is made explicit in the form of what are currently termed ‘terrorist attacks’. In this respect, Furedi (2007) makes a cogent argument that the language we use betrays confusion about the threat we face and therefore undermines our capacity to engage with it. Whether or not we agree with Mueller's and Furedi's observations and analyses, we are bound to admit that terrorism creates a state of anxiety not only as a result of the psychological effects that immediately follow acts of violence, but also because of the uncertainty as to what exactly ‘terror’ is.

Conclusion

Among the four different points of view considered in the discussion, the first, which focuses on perpetrators' intentions, should be ruled out because of its lack of rigour. As far as the other three are concerned, they all depend on the difficult task of adequately defining violence. The word ‘terrorism’ has been adopted into common parlance because it carries negative connotations and because its use produces two effects that are inherent in the phenomena it designates, namely, the desire to mete out blame and the creation of anxiety, although these vary with the type of terrorist action perpetrated. In the context of scientific terminology, on the other hand, it is difficult to ascertain its heuristic value. Everybody would agree that a world in which no direct or indirect victims suffered from politically motivated violence would be a better world than the one we live in now. The issue does not involve underplaying the impact of violence on its victims, but wondering whether the social sciences are equipped to examine the phenomena in a coherent and rigorous manner. Nor should it be denied that among the acts which are associated with terrorism, some aim to ‘provoke an immense psychological shock’, as noted by Qiao and Wang (2002, p. 182). Rather, there is a need to devise a conceptual framework capable of analysing this anxiety-provoking aspect of violence. Similarly, we should try to improve the accuracy with which actions are analysed and assigned to the category of terrorism, by identifying the perpetrators' aims more clearly and describing the impact of such acts on their chosen targets more precisely. In other words, three of the four interpretations of terrorism and its consequences briefly outlined in this article should continue to stimulate unbiased discussion, free of pre-conceived ideas as to the outcome, with the ultimate aim of determining whether the concept of terrorism is a useful heuristic tool that can contribute to the analysis of politically motivated violence, or whether it in fact constitutes an obstacle to effective analysis.

Finally, I would like to outline a few approaches to the problem that might prove fruitful in future discussion. First, as Horgan (2005, p. 22) has pertinently emphasised ‘emotional reactions’ to violence constitute only one of the several subjects dealt with in studies of terrorism. In this regard, the fact that the notion of ‘terror’ has become so prominent as to now have a research field named after it is precisely the main stumbling block towards discussing the topic more lucidly. The intention of the instigators of violence is certainly one aspect among others that deserves serious study. However, as stated earlier, it is not clear that all so-called terrorists seek to create a climate of terror among the population or that their aim in using the threat of extreme violence is to acquire control over a given society. Such an objective is usually the prerogative of state terrorism, which uses methods of intimidation to neutralise any movement opposing the ruling power. However, the same methods can just as well be employed by groups of insurgents who want to silence sections of the population who do not actively support their cause. In other words, this particular form of violence aims to produce inaction or a sense of resignation. This is not usually the aim of subversive groups who, on the contrary, attempt to goad their enemy towards an active confrontation. In this context, an act of violence is designed to provoke a reaction, for example, to obtain satisfaction with regard to wage claims, instigate an escalation of violence, exacerbate social tensions for political purposes, or create a climate of insecurity in order to demonstrate the state's inability to impose its authority. In all these examples, the logic of violence is that of interaction in the Clausewitzian sense (or ‘reciprocal action’ according to the translation of Wechselwirkung). This point is crucial for understanding the dynamics of a conflict.

The distinction between unilateral coercive violence and reciprocal action does not collapse the opposition between state terrorism and subversive terrorism since both can operate alternately or simultaneously in the same manner. What is more, the distinction tends to become blurred when, for example, the state employs death squads to eliminate subversive elements. Nevertheless, despite some grey areas, the distinction remains valid because of its theoretical implications. In practice, we should first of all consider whether each of the various forms of violence currently included in the concept of terrorism should not be analysed separately according to models adapted to their specific characteristics. This would mean that rather than attempting to develop a conceptual framework capable of encompassing all the phenomena, which, as most specialists agree, are heterogeneous, ⁹ we should instead examine components such as strategic analysis, public security policy, the theory of social movements, or criminology, and then possibly combine these approaches, as circumstances require. Such a model does not fully comply with the hopes expressed by Horgan (2005). He emphasises the need for finding a conceptual anchoring and a set of starting assumptions in order to comprehend the heterogeneity of terrorism. However, since it is currently impossible to find a common anchoring for the notion of terrorism, the model proposed here seems to be a more reasonable, albeit less ambitious project, than that of formulating a general theory of terrorism.

Second, strategic analysis in the context of terrorism may not at first sight seem to be a very original idea. Although this approach has been frequently adopted since the 1960s, it has always lacked a sound theoretical foundation. ¹⁰ In the present case, it is not just a question of invoking a few isolated concepts concerning strategic analysis, but rather of systematically analysing some of the phenomena associated with terrorism as though they were ‘a type of, and a tool of war’ (Horgan 2005, p. 17), by using the most sophisticated methods that the theory of war can provide. War theory is pertinent from two points of view. On the one hand, it should enable us to determine whether or not the phenomenon that we call ‘terror’ (in the psychological meaning of the term) is an element that is intrinsic to armed conflict but which has taken on a specific and shocking form under contemporary conditions. In other words, it is the psychological aspect of the pressure exerted on an adversary in order to bend them to the aggressor's will. On the other hand, using war theory as a starting point, it should be possible to decide whether ‘terror’ (in the general sense of the term) is or is not a new expression of the ‘trend to the extreme’ (Clausewitz 1976, p. 79), thus marking one more stage in a process whereby armed conflict transgresses the limits that states have attempted, more or less satisfactorily, to impose. Jenkins (1974, p. 2) already partially broached this second question over 30 years ago; it has recently been taken up by Girard (2007, p. 129) from a philosophical viewpoint, by Rasmussen (2006, p. 143) from a strategic standpoint, and by Sloan (2006, p. 30) from a more classical perspective. However, it can only be adequately discussed in conjunction with the question of clarifying terminology and separating out the components of terrorism.

Third, even if researchers cannot reach a consensus on a definition of terrorism in general, it should at least be possible to agree on the meaning of ‘terror’ as a psychological term. Such a definition would make it possible to assess the appropriateness of the terminology currently used in the cases we examine and to appraise the significance of the results of empirical studies on the psychological effects of violence. These results have not yet received the attention they deserve among specialists in terrorism, although Silke (2004b, p. 521) recognises the volume of work devoted to this subject when he states that our understanding of the psychological impact of terrorism has considerably increased since the beginning of the century. However, his own approach focuses on issues like the measures that should be taken in the case of future attacks, the type of help victims would require, and how to deal with the reactions to violence of the different ethnic communities involved. These are all important problems but they should not obscure the underlying theoretical implications that this article has attempted to bring to light, namely, that clarifying some of the terminology and definitions around the notion of terrorism might make our discussions more relevant and stimulate maximum use of the considerable body of knowledge we have already acquired through empirical studies of terrorism. Bleich's statement and my reformulation of it above is a case in point. The original and the variant both refer to the same type of event, but the reformulation is more appropriate, because replacing ‘terror’ with ‘deadly attacks’ avoids the ambiguity inherent in the use of the term ‘terror’ and describes more succinctly the observable empirical impact of the violence. This illustrates in a small way that the current widespread, indiscriminate use of the term is counter-productive, in the sense that it tends to obscure that which it is intended to clarify. Like the ancient Chinese, we have to be aware that the lack of clear language causes confusion: ‘If names are not rectified, speech will not accord with reality; when speech does not accord with reality, things will not be successfully accomplished’ (Confucius 2003, p. 139). Thus, it may be that ‘terror’ and ‘terrorism’ are not the adequate terms for describing the phenomena observed.

Notes

1. Blanc (1848, p. 100), in his history of the years 1830–1840 mentions an attempt at ‘royalist terrorism’. Michelet (1858, p. 367, 1865, p. 203) refers to the ‘terrorism of Condé’ during the Fronde rebellion, and the ‘growing terrorism in Russia’ when describing the repressive methods practiced in the empire under tsarist rule in the first part of the nineteenth century.

2. Programm des Exekutivkomitees der Narodnaja Volja (1879), cited by Walther (1997, p. 389).

3. Nicolaj Morozov, Terroristicheskaya Borba, 1880, cited by Laqueur and Alexander (1987, p. 72).

4. Among the critics of the definition, see Sottile (1938, p. 95).

5. This problem has been partially discussed by Gupta (2008, p. XVII), Derrida and Habermas (2003, p. 102), and Webel (2005, p. 90), but we must go further into the heart of the matter.

6. The authors also observe that between September 2001 and January 2002 the prescription of psychotropic medicine did not increase significantly, neither did cases of PTSD or mental problems in the New York Veterans Hospitals Network (Vázquez et al. 2006, p. 71).

7. DiMaggio and Galéa (2006) show that PTSD affected a range of 12–16% on the basis of an analysis of results obtained from 61 different studies. These results refer to variously involved groups in which subjects that were directly exposed were over-represented.

8. Thornton (2006, p. 43) states that ‘in an internal war situation, terror is a symbolic act designed to influence political behaviour by extranormal means, entailing the use or threat of violence’.

9. According to Victoroff (2006, p. 456): ‘Terrorism is highly heterogeneous. It is more fruitful to examine the multiple types and instances of terrorism than to regard terrorism as a unitary phenomenon’.

10. Studies of terrorism have a lot of theoretical foundations (probably too many). However, among these, the strategic component lacks a conceptual framework that incorporates the macro level of the theory of war.