Targeting Strategies of Terrorist Groups

Permanent Link:

Material Information

Title: Targeting Strategies of Terrorist Groups A Comparative Case Study of Northern Ireland and Spain
Physical Description: Book
Language: English
Creator: Quirke, Stephen
Publisher: New College of Florida
Place of Publication: Sarasota, Fla.
Creation Date: 2010
Publication Date: 2010


Subjects / Keywords: Terrorism
Political Violence
Genre: bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: This study addresses the question of why terrorist groups target civilians by analyzing the targeting strategies of nationalist terrorist groups in Northern Ireland and Spain. Nationalist terrorist groups are considered to be the most likely to apply violence in a way that preserves and maximizes their base of support. This study tested the hypothesis that greater levels of violence inflicted on the supporting communities of these groups cause supporters to become radicalized, thus allowing the groups to engage in increasingly indiscriminate violence without sacrificing public support. The central hypothesis was broken down into four parts so that different measures of indiscriminate violence could be examined: sectarian killings, collateral killings, and civilian deaths carried out under novel, idiosyncratic justifications. Sectarian killings were measured in two time frames to determine whether the motivation was deterrence or attrition. The study examined the nationalist groups� targeting strategies in the periods surrounding the six individual months in each conflict that saw the highest number of civilian deaths within the nationalist group�s supporting community caused by both government forces and right-wing terrorist groups. One hypothesis was confirmed, but only in two periods under study. Both periods occurred in Northern Ireland in the unusual context of a British ceasefire with the IRA. The other hypotheses were not confirmed. The findings suggest that the use of indiscriminate violence by nationalist terrorist groups is better explained by organizational dynamics, and particularly organizational weakness, than by the level of violence sustained by their supporting community.
Statement of Responsibility: by Stephen Quirke
Thesis: Thesis (B.A.) -- New College of Florida, 2010
Bibliography: Includes bibliographical references.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The New College of Florida, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Local: Faculty Sponsor: Hicks, Barbara

Record Information

Source Institution: New College of Florida
Holding Location: New College of Florida
Rights Management: Applicable rights reserved.
Classification: local - S.T. 2010 Q8
System ID: NCFE004311:00001

Permanent Link:

Material Information

Title: Targeting Strategies of Terrorist Groups A Comparative Case Study of Northern Ireland and Spain
Physical Description: Book
Language: English
Creator: Quirke, Stephen
Publisher: New College of Florida
Place of Publication: Sarasota, Fla.
Creation Date: 2010
Publication Date: 2010


Subjects / Keywords: Terrorism
Political Violence
Genre: bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation


Abstract: This study addresses the question of why terrorist groups target civilians by analyzing the targeting strategies of nationalist terrorist groups in Northern Ireland and Spain. Nationalist terrorist groups are considered to be the most likely to apply violence in a way that preserves and maximizes their base of support. This study tested the hypothesis that greater levels of violence inflicted on the supporting communities of these groups cause supporters to become radicalized, thus allowing the groups to engage in increasingly indiscriminate violence without sacrificing public support. The central hypothesis was broken down into four parts so that different measures of indiscriminate violence could be examined: sectarian killings, collateral killings, and civilian deaths carried out under novel, idiosyncratic justifications. Sectarian killings were measured in two time frames to determine whether the motivation was deterrence or attrition. The study examined the nationalist groups� targeting strategies in the periods surrounding the six individual months in each conflict that saw the highest number of civilian deaths within the nationalist group�s supporting community caused by both government forces and right-wing terrorist groups. One hypothesis was confirmed, but only in two periods under study. Both periods occurred in Northern Ireland in the unusual context of a British ceasefire with the IRA. The other hypotheses were not confirmed. The findings suggest that the use of indiscriminate violence by nationalist terrorist groups is better explained by organizational dynamics, and particularly organizational weakness, than by the level of violence sustained by their supporting community.
Statement of Responsibility: by Stephen Quirke
Thesis: Thesis (B.A.) -- New College of Florida, 2010
Bibliography: Includes bibliographical references.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The New College of Florida, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Local: Faculty Sponsor: Hicks, Barbara

Record Information

Source Institution: New College of Florida
Holding Location: New College of Florida
Rights Management: Applicable rights reserved.
Classification: local - S.T. 2010 Q8
System ID: NCFE004311:00001

This item is only available as the following downloads:

Full Text


Targeting Strategies of Terrorist Groups: A Comparative Case Study of Northern Ireland and Spain by Stephen Quirke A Thesis Submitted to the Division of Social Sciences New College of Florida in partial fulfillment of the requirements for the degree Bachelor of Arts Under the sponsorship of Barbara Hicks Sarasota, Florida July 2009


Table of Contents List of Tables and Figures................................................................................iii List of Acronyms................................................................................................iv Abstract.................................................................................................v Introduction and Structure of Study........................................................1 Chapter 1: ...............................................................................................................4 Knowing it When We See It: Locating a Bounded Definition of Terrorism Chapter 2:................................................................................................................20 Strategy or Syndrome? Perspectives on Terrori st Decision Making Chapter 3: ...............................................................................................................40 Why Target Civilians? Theories of Violence and Legitimization Chapter 4: ...............................................................................................................80 The Other Civil Rights Movement: The IRA and Northern Ireland Chapter 5: ...............................................................................................................130 Violent Myth Making: ETA and Spain Chapter 6: ...............................................................................................................175 Comparative Analysis Appendix..............................................................................................................188 Bibliography........................................................................................................189 ii


List of Tables and Figures Figure 4.1 Northern Ireland Timeline of Events........................................................88 Figure 4.2 Northern Irel and Fatality Trends..............................................................119 Figure 4.3 Northern Ireland Ci vilian fatality trends..................................................119 Table 4.1 Religious Composition of civilian deaths by actor....................................120 Table 4.2 High-Intensity Months fo r Catholic Civilian Deaths.................................120 Table 4.3 Analysis of High-Intensity Months............................................................121 Table 4.4 Results Summary ......................................................................................124 Table 4.5 Public Support for the Use of Violence in Northern Ireland.....................127 Figure 5.1 Spain Timeline of Events.........................................................................139 Table 5.1 Basque Country Deaths Inf licted by the Extreme Right and GAL............163 Table 5.2 Peak Years for ETAs Civilian Fatalities .................................................165 (Adjusted for Security-Related Deaths) Figure 5.2 Percentage of Civilian Ca sualties Caused by ETA Over Time................166 Table 5.3 Selectivity of ETAs Nonsecurity Civilian Killings 167..........................170 Table 5.4 Electoral Support for ETAs Political Wing..............................................172 In the Basque Legislature Table 6.1 Percentage of Civilian Deaths in Relation to All Deaths Caused by Republican Paramilitaries and ETA................................................178 iii


List of Acronyms EH Euskal Herritarrok (Communist Party of the Basque Homeland) ETA Euskadi Ta Askatasuna ( Basque Homeland and Liberty) ETA-m Euskadi Ta Askatasuna militar (Basque Homeland and Liberty, military branch) ETA-pm Euskadi TA Askatasuna politic o-militar (Basque Homeland and Liberty political-military branch) HB Herri Batasuna (Unity of the People) HAS (Popular Socialist Party) INLA Irish National Liberation Army IPLO Irish Peoples Liberation Organization IRA Irish Republican Army PIRA Provisional Irish Republican Army PP Partido Popular (Peoples Party) PNV Partido Nacional Vasco (Basque Nationalist Party) PSOE Partido Socialista Obrero Espanol (Spanish Socialist Workers Party) RIRA Real Irish Republican Army OIRA Official Irish Republican Army SF Sinn Fein (We Ourselves) SDLP Social Democratic and Labor Party UVF Ulster Volunteer Force UVA Ulster Volunt eer Association RUC Royal Ulster Constabulary TO Terrorist Organization iv


Targeting Strategies of Terrori st Groups: A Comparative Case Study of Northern Ireland and Spain Stephen Quirke New College of Florida, 2009 Abstract This study addresses the question of why terrorist groups target civilians by analyzing the targeting strategies of nationalist terrorist groups in Northern Ireland and Spain. Nationalist terro rist groups are considered to be the most likely to apply violence in a way that preserves and maximizes thei r base of support. This study tested the hypothesis that greater levels of violence in flicted on the supporting communities of these groups cause supporters to become radicalize d, thus allowing the groups to engage in increasingly indiscriminate violence without sacrifici ng public support. The central hypothesis was broken down into four parts so th at different measures of indiscriminate violence could be examined: sectarian killi ngs, collateral killings, and civilian deaths carried out under novel, idiosync ratic justifications. Sectarian killings were measured in two time frames to determine whether the motivation was deterren ce or attrition. The study examined the nationalist gr oups targeting strategies in the periods surrounding the six individual months in each conflict that saw the highest number of civilian deaths within the nationalist group s supporting community caused by both government forces and right-wing terrorist groups. One hypothesi s was confirmed, but only in two periods v


under study. Both periods occurred in Northern Ireland in the unusual context of a British ceasefire with the IRA. The other hypothese s were not confirmed. The findings suggest that the use of indiscriminate violence by nationa list terrorist groups is better explained by organizational dynamics, and particularly organizational weakness, than by the level of violence sustained by their supporting community. Barbara Hicks Division of Social Sciences vi


Introduction Popular discourse on terrorism provides le ss information than it does a rhetorical bludgeoning. Despite a conspicuous lack of c onsensus on a number of important issues including the meaning of the term terrorism has become the buzz word of choice to justify some highly questionable American fo reign policies, and ma ny other countries are catching on. This function alone makes te rrorism an important object of study. The September 11 th attacks represent a watershed in the history of terrorism because they violated all previously held expectations about terrorists use of discriminate force. Prior to these attacks, it was commonly held that terrorists want a lot of people watching not a lot of people dead. Generalizations about terrorism were reversed almost overnight. The attacks also highlighted a lingering gap in the literature on terrorism the persistent willingness of some terrorist groups to target civilians, despite its liabilities. This study focuse s on the target selection strate gies of nationa list terrorist organizations in Northern Ireland and Spain. Nationalist terrorist groups are selected because they are uniquely committed to maxi mizing popular support, and therefore tailor their use of violence to the preferences of local supporters. The severity of local grievances is an important determinant of th e form and intensity of the violence they produce. 1


The central puzzle this proj ect seeks to answer is why civilians are selected as targets of violence. It does so by attempting to explain cha nges in targetin g strategies over time and asking whether these changes re sult from rational calculation, emotional response, or structural constr aints. The central hypothesis is that as more violence is inflicted on the groups base of support, supporters grow more accepting of the terrorist group attacking civilians and noncombatants. Terrorist groups should pick up on these preferences and act on them to conserve resour ces. They may also believe that attacking civilians and noncombatants is the most eff ective way to wear down the will of their opponents. The main constraint preventing the targeting of civilians and noncombatants in this formulation is the need to mainta in popular support. Supporte rs outside the hard core may have little tolerance for the killing of civilians. As the community of supporters and potential supporters increa singly become the targets of violence, however, they may become morally disengaged from the larger society and support more extreme acts of violence. Investigating this question gives us insights into how the preferences of supporters are affected by experiences with violence, and how changes in the preferences of supporters interact with and constrain the decision making of nationalist terrorist organizations. These two pieces of knowle dge are invaluable for understanding and evaluating policy options against terrorist groups. 2


Structure of the Study Chapter one provides an overview of the i ssues that arise in defining terrorism, outlines the features which make for an e ffective definition in different arenas, and creates an operational definition for the purpos es of this study. Chapter two investigates the psychological, rational choice, and organizational pe rspectives on the decision making processes of terrorist organizations. It also briefly summarizes the empirical evidence on the success of terrorist groups, cla ssifies groups according to their strategic goals, and describes how goals affect a group s targeting strategy. Chapter three reviews some of the major theories that seek to e xplain why civilians are targeted by terrorist organizations, and lays out the studys research design and hypotheses. Chapters four and five will present case studies on Spain and the United Kingdom. The hypotheses are tested after a brief history of each conflict a nd the terrorist organizations sustaining them. Chapter six concludes the study by comparing the findings and pointing out directions for future research. 3


Chapter 1 Knowing It When We See It: Locating a Bounded Defi nition of Terrorism One of the biggest impediments to studying terrorism is the lack of a consensus definition among academics. Definitional disputes also create significant complications for intergovernmental cooperation. 1 Despite the relative commonality of bilateral and multilateral agreements concerning a variety of crimes, extradition for political offenses is often explicitly excluded. Because terrorism is always political, this loophole allows countries to shirk their oblig ation to extradite wanted te rrorists (Ganor 2002, 300). The persistence of the debate over definitions is well illustrated by the fact that, in 1998, the most comprehensive database on terrorist at tacks significantly altered its operational definition of terrorism. In that year, the database began to utilize what its authors described as a novel approach that of not choosing any particular definition of 1 Even the U.S. Government cannot agree on a single definition. The US Department of Defense defines terrorism as the unlawful use or threatened use of force or violence against individuals or property to coerce or intimidate governments or societies, of ten to achieve political, religious, or ideological objectives. The Department of State defines it as p remeditated, politically motivated violence perpetrated against noncombatant targets by subnational groups or clandestine agents, usually intended to influence an audience (Enders, 5). 4


terrorism, but rather allowing us ers of the database, within limits, to filter out the data that does not accord with their own definitions (START 2007, 2). Definitions Terrorism has been described as crime, warfare, and clandestine political violence. 2 Charles Tilly defines it as a type of political violence falling outside a regimes routine forms of struggle (2004, 5). Others have suggested that terrorism is unique because it attacks civilians and noncombatants. A fruitful way to structure an investigation of these pe rspectives is to ask exactly what we want out of a definition, and whether we need one at all. For any academic inquiry, a clear definition is needed to bound the phenomenon under investigation. Leaving the definition implicit is the road to obscurantism, (Gibbs 1989, 329). In the polit ical and legal realm, we need clear boundaries to demarcate those actions we wish to proscribe. Of cour se, there are already laws against murder, arson, kidnapping, extor tion, and any other act a terrorist might commit. However, it may be desirable to el evate the punishment for certain classes of crimes based on their motivation to create a mo re effective deterrent, as we currently do with categories like treason a nd sedition. We can also approach the need for a definition by investigating how the need for different definitions has evolved over time. The meaning of the term terrorism ha s changed four times over the past two hundred years (Hoffman 1998, 15). It was origina lly used as a positive term describing the government enforcement of order followi ng the French revolution. Thus, Robespierre 2 Martha Crenshaw defines terrorism as a conspiratorial style of violence calculated to alter the attitudes and behavior of multiple audiences (Crenshaw 1995, 4). 5


appealed to virtue, without which terror is evil; [and] terro r, without which virtue is helpless (16). The term later became asso ciated with revolutionary movements, particularly Russian anarchists who carried out a campaign of ty rannicide to foment revolution. By the 1930s, the word terrorism changed again. It was now used less to describe revolutionary movements and more to denote the mass repression of totalitarian states (e.g., Fascist Italy, N azi Germany and Stalinist Russia). The wave of nationalliberation struggles in Africa following th e end of World War II caused the terrorism discourse to shift back towards sub-national groups. Today, the word terrorism is most commonly associated with the contemporary and predominantly religious wave that was launched in 1979 with the Iranian revolu tion and Soviet inva sion of Afghanistan (Rapoport 2006, 18). This wave has produced both sub-national and transnational terrorist groups. For non-state actors, terrorism is unique among crimes because its perpetrators imagine themselves to be selflessly committe d to an imagined collective the oppressed (Hallett 2004, 52). Crime is not the only relative of terro rism, but for those who are pessimistic about terrorisms ability to achieve political objectives, it is often seen as the closest. Thus, according to Brian Hallett, te rrorists have a well developed theory for the Machiavellian or instrumental use of terror, bu t they lack even the rudiments of a theory for the effective achievement of political and social goals, (58). Such theatrical crimes provide no instrumental bene fit, but only symbolic, psyc hological satisfaction (62). But is success really so elusive for terro rist groups? Hezbollahs car bombing of the US Marine barracks in Lebanon on Octobe r 23, 1983 is often credited for the US withdrawal from Lebanon (Abrahms 2006, 45). The October 1994-August 1995 Hamas 6


suicide campaign against Israel is said to have led to the pa rtial Israeli withdrawal from the West Bank (Enders 2005, 5). During the wave of national liberation movements following World War II, terrorism also seems to have been effective at expelling the British from mandate-Palestine and the French from colonial Algeria. These anecdotes indicate that terrorism can be seen as an effective strategy. Additionally, Pape (2006) argues forcefully that the proliferation of su icide terrorism can be explained by its ability to achieve tactical goals. Terrorisms unique motivations, combined with its potential to achieve political goals, makes it seem less lik e normal crime and more like a form of warfare or low-intensity conflict. These competing perspectives help to illustrate why taking any one position on terrorism requires us to make assumptions about many others. The debate over strategic utility versus emotional benefits requires a bounded definition in order to create an unbiased body of evidence. If we accept the idea that terroris m only targets civilians, we cannot count an attack on Marines in Lebanon as a victory for terrori sts (regardless of its status as a suicide bombing). Similarly, if we insist on using the term strictly as a pejorative, our collective ability to study any class of phenomenon to which the name applies will be considerably reduced. Ethica l debates about what is good, bad and sometimes necessary will quickly run ove r any investigation built on such a weak foundation. Alex Schmid distinguishes four arenas of terrorism discourse; academic, state, public, and that of terrorists and their symp athizers (where the focus is on ends rather than means). Schmid suggests that each aren a seeks to define the term for its own purposes, and he advises us to accept such di fferences. He also suggests what an ideal 7


definition might look like for each category. Di stinguishing among these arenas is very useful for conceptualizing the different meanings evoked by different actors who employ the term. Schmid ultimately offers consensus definitions for both the political and academic arenas of terrorism discourse. To get at the academic discourse, Schmid asks dozens of terrorism experts to rank th e acceptability of his own proposed consensus definition, and he asks those who find it unacceptable, to define the term. He devotes more than a hundred pages to over a hundred different definitions in hopes of discovering one that is both comprehensive and widely acceptable. He does this by breaking down the frequency of distinct definitional elements. Differentiation of the victim and the target of violence was mentioned 37.5% of the time; terrorisms arbitrary or indiscriminate nature was mentioned 21% of the time; civilians, noncomba tants, and neutrals were mentioned only 17.5% of the time; and in an additional 15.5% the innocence of victims was emphasized (Hoffman 1998, 39). From this exercise he produced a lengthy academic definition consisting of many potential attributes, none of which were privileged. The full academic definition is as follows: Terrorism is an anxiety-inspiring method of repeated violent action, employed by (semi-) clandestine individual, group or state actors, for idiosyncratic, criminal or political reasons, whereby in contrast to assassinatio n the direct targets of violence are not the main targets. The immediate human victims of violence are generally chosen randomly (targets of opportunity) or selectively (repres entative or symbolic targets) from a target population, and serve as message generators. Threatand violence-based communication processes between terrorist (organization), (imp eriled) victims, and main targets are used to manipulate the main target (audience(s)), turning it into a target of terror, a target of demands, or a target of attention, depending on whether intimidation, coercion, or propaganda is primarily sought, (Schmid 1993, 8). This definition was greatly simplified for its application in stat e discourse. In this context, Schmid argues that a good definition is one to which many can agree. A broad definition is likely to be objected to by more people than a narrow one since it turns 8


more practitioners of violence into terrorist s (1993, 11). A legal definition, according to Schmid, should be based on the extant law covering war crimes. More specifically, he argues that terrorism should be considered the peacetime equivalent of war crimes, or federal crimes against humanity (11). Th is definition might exclude some forms of violence and coercion that some governments currently call terrorism, such as attacks on the military, hijackings for escape, and the destruction of property. However, Schmid argues that a narrow and precise definition of terrorism is likely to find broader support than one that includes various forms of viol ent dissent and protest short of terrorist atrocities (1993, 12). In 2003, th e Supreme Court of India a dopted Schmids definition (Madan Singh vs. State of Bihar). There is much to recommend this definiti on. At the most basic level it can help to facilitate international lawmaking and thus enhance interstate counterterrorism operations (for instance, preventing conf licts over extradition). Sc hmid also argues that his definition can help us strike a balance be tween two different response models those based on crime and war. Recognition of terrorisms political motivation and disrespect for human rights can expedite the creation of sp ecial legal instruments. At the same time, explicit acknowledgement that terrorism occurs outside of war helps to prevent an overreaction that undermines st ate legitimacy which could potentially fuel more terrorism (Schmid 1993, 336). 3 Ganor (2002) endorses this approach, but alters is slightly. For Ganor, an international definition of terrorism s hould be based on the Geneva and Hague conventions, which established that it was not acceptable to target civilians for political 3 By contrast, successful invocati on of the war metaphor makes citizens more willing to embrace the suspension of civil rights in pursuit of the enemy. It can also be used to justify collateral killings. (Kruglanski, Crenshaw, Post and Victoroff 2008, 101). 9


ends. Extending these sanctions on state viol ence to sub-national groups, Ganor argues that terrorism should be considered the sub-na tional equivalent of either war crimes or crimes against humanity, depending on the sp ecific context (2002, 304). He is thus agnostic as to whether there is an official state of peace something Schmid assumes explicitly. Ganor is also more explicit about the fact that terrorism consists of attacks against civilians only Many definitions of terrorism describe its targets as noncombatants rather than civilians. The U.S. Depart ment of State, for instance, which lifts its definition from Title 22 of the U.S. Code, Section 2656f(d), defines terrorism as the political targeting of noncombatant s by subnational groups (Enders 2005, 5). Noncombatants, however, can refer to anyone who is not currently ready for battle. It might include unarmed soldiers, pe acekeepers, or armed occupiers. Defining terrorism in this way is problematic because militant organizations and their sympathizers can make a legitimate objection: they cannot be expected to attack only those military personnel who are armed and ready for battle. By narrowing the definition of terrorism to include only deliberate attacks on civilia ns, we leave room for a fair fight between guerillas and state armies (G anor 2002, 289). Classifying al l noncombatants as potential victims of terrorism creates an unnecessary impediment to international agreement and cooperation. Action-Centered and Actor-Centered Approaches Sanchez-Cuenca and de la Calle (20 08) distinguish between two broad approaches to defining terrorism: the action-centered and actor-centered approaches. Under the action-centered approac h, terrorism can be used by very different actors, with several potential features dist inguishing it from other types of political violence. Most 10


definitions emphasize either the intent to sp read fear or the discrepancy between the target of violence and the audience of vi olence (Sanchez-Cuenca and de la Calle 2008, 33). These distinctions fail because they are features of all type s of violent coercion. Violence will not deter other challengers unless it inspires fear of loss, or, in more sanitized terminology, it creates a credible th reat which raises the perceived costs of violence beyond the expected benefits. The de terrent effect therefore requires that actors be coerced psychologically if not physica lly. Action-centered approaches therefore cannot distinguish terrorism from guerilla warfare, civil war, or other types of political violence without reference to the targeting of civilians or noncombatants. A good example of this approach co mes from Schmid, who, as we saw above, offers a political and legal definition of terrorism as the [ non-state] peacetime equivalent of war crimes (1993, 11). Sanchez-Cuenca and de la Calle disagr ee with the argument that targeting civilians or noncombatants makes terrorism unique, deriding it as one of the most widely shared false beliefs about terrorism (2008, 34). This view has proliferated, they argue, due to the bias of international data which is often used inappropriately in terrorism studies. For intern ational terrorism, attacks on ci vilians are common. However, based on a dataset these authors constructe d for European terrorism between 1965-2000, they claim that domestic terrorism also kills a large percentage of combatants. According to their data, combatants constitute fully 46% of all victims of European terrorism. According to these authors, this makes the civilian distinction sim ply useless (ibid). This argument is exceedingly circular one definition cannot be deemed bad simply because it excludes data captured by an altern ative definition. According to this logic, 11


their dataset may be equally wrong for failing to include acts of an even more expansive definition. The argument also ignores the possibility that their definition corrupts data with false positives. 4 To buttress their claim that civilians shoul d not be considered the sole victims of terrorism, they point out that civilians are making up an increasing proportion of victims in interstate wars, and are currently the largest group of victims in intrastate wars. However, when we consider terrorism to consis t of attacks against ci vilians regardless of the perpetrator, building a theory to e xplain it becomes extrem ely problematic. If insurgent groups, armies, and governments can produce terrorism, what exactly is the dependent variable [e.g. the set of attacks we seek to explain, or the extent of coercive success]? (Sanchez-Cuenca and de la Call e 2008, 34). In other words, it is highly unlikely that there will be a common set of causes or effects for the violence of all of these disparate actors. The action-centered de finition is thus too s lippery and fails to indicate a bounded type of violence which can be coherently explained. 5 Such a definition, they argue, precludes the construction of theory. Sanchez-Cuenca and de la Calle advocate an actor-centered definition based on the existence of extreme power asymmetries. The importance of terrorism, they argue, is in the unique conflict dynamics which emer ge when a non-state group does not have the 4 Consider the African National Congress (ANC), which fought against apartheid in South Africa from the 1960s until the 1994 election. The ANC rejected indiscriminate violence against whites despite a deeply segregated society and despite their failure to form an effective guerilla military force. The exiled ANC leadership was so committed to portraying itself as a principled challenger that its president signed a protocol of the Geneva Convention which legally bound the ANC to avoid attacks on civilians the first time a guerilla group had ever done so (Goodwin 2006, 2034). The decision to describe this group as terrorist has thus caused a great d eal of controversy. Nelson Mandel a, the Nobel Peace Prize winner and former president of South Africa, had coordinated sabotage campaigns against military and government targets as the head of the ANCs armed wing, Spear of the Nation In 2008, just before he turned 90, his name was removed from an immigration watch list for potential terrorists. 5 At the very least, coherent explanations will not be possible with such definitions unless we introduce a host of additional distinctions, and these can quickly become confusing. 12


capacity to win militarily, but is still able to challenge the states monopoly on violence by inflicting punishment on a consistent basis. To avoid analytical c onfusion, things like state terrorism are ignored in actor-centered definitions. This omission is justified by the fact that the techniques employed by states are sufficiently different from those of non-state groups, such that combining them together does not help us understand their behavior. State terrorism, according to th is argument, is better conceived as state repression (Sanchez-Cuenca a nd de la Calle 2008, 34). The authors provide two criteria to dete rmine the strength of the non-state group: the number of fatalities and the control of territory. 6 The first variable for determining power is the number of casualties a central ingredient for definitions of civil war. In order for a conflict to be considered a civi l war, scholars conventionally set a minimum threshold of 1,000 victims (there is disagreemen t as to whether this should be an annual or accumulated figure). Terrorists, Sanchez-Cuenca and de la Calle argue, do not achieve this destructive capacity. Furthermore, the f act that terrorists do not control territory makes them, by definition, clandestine or underground. For guerillas, however, The existence of territory establishes partic ular dynamics. The guerrillas have to act as a proto-state in the liberated territory, imposing order and extracting rents from the inhabitants. Guerrillas depend on the support (voluntary or not) of the population. Terrorist groups, instead, have very superficial contact with the population because they cannot act in the open (Sanchez-Cuenca and de la Calle 2008, 35). Terrorism is distinct from communal viol ence in that it is not sporadic but is instead a long-term organized challenge (Sanchez-Cuenca and de la Calle 2008, 13). For Sanchez-Cuenca and de la Calle, the main shortcoming of the actor-centered 6 They leave ambiguous whether both conditions or only the more stringent one control of territory is necessary before a group can be regarded as guerilla and hence no longer terrorist. They do not define control of territory. Asal and Rethemeyer define it as both the ability to coerce nonmember civilians to act or forebear and to exclude police and military units from some define geographic space for a period greater than six months (2008, 442). 13


approach is that it does not cover what ar e obviously terrorist acts when guerillas leave their territory and operate in urban se ttings. However, one problem they do not acknowledge is the myopic focu s on territorial acquisition as the main determinant of resource mobilization. Even without complete te rritorial control, sufficient coercion of or complicity with citizens can al low terrorists to acquire signi ficant resources, and thereby achieve high destructive capabilities. Fo r instance, Euskadi Ta Askatasuna (ETA) successfully enforced a revolutionary tax against businesses, while the Irish Republican Army (IRA) receives gets large sums of m oney from both legitimate business enterprises and donations from the Irish-Ame rican diaspora. It is unclear at what point successful resource extraction should become synonymous with occupying territory, or where a high enough level of casualties should allow us to describe a group as guerilla regardless of its territorial holdings. Problems with the Action-Centere d and Actor-Centered Approaches The main weakness of the action-centered approach is that it is hard to move consistently from acts to groups. Ganor suggests two principles that can accomplish this task: a qualitative principle, in which even a sing le attack against a civilian is sufficient to justify the terrorist label, and a quantitativ e principle, which would compare the number of attacks against civilians to the number of attacks against legitimate military targets before labeling a group as terrorist. The pr oblem with these suggestions is that the qualitative principle can never achieve the international consensu s Ganor promises, and the quantitative principle is grossly underspecified (Ganor 2002, 297). For instance, it is hard to imagine the emergence of an inte rnational consensus regarding the acceptable level of civilian casualties before a group comes under the terrorist label and thereby 14


additional sanctions. 7 For Goodwin, another advocate of the action-centered approach, the very label terrorist fa lsely essentializes a groups na ture. This is problematic because it can lead us into falsely labeling ev ery action of the group as terrorist, while precluding the possibility that its members be have rationally to match political means and ends. Another problem with act ion-centered definitions concerns the extent to which rebel groups coerce civilians to raise resources, often under the threat of violence. The use of child soldiers is an illustrative, if extreme, example. An action-centered definition could certainly accommodate categories like child soldiers. However, it is hard to imagine a test that can divide more difficult cases into terrorist ve rsus non-terrorist acts. This formulation could be turned against stat es as well, so that the extent to which coercion or the threat of it is integrated into the very fabric of society makes it impossible to define terrorism without arbitrarily discriminating against sub-state actors. One of the more difficult problems for act ion-centered definitions is determining whether political assassination should count as a terrorist attack. Anal ysts who answer in the affirmative have attributed the outbreak of World War I to the Serbian terrorist group Black Hand, which assassinated Archduke Franz Ferdinand. Such scholars are then in a position to treat terrorism as an extrem ely dangerous and sophisticated strategy. Assassination, however, is a very grey area. Like civilians, po litical leaders are noncombatants, and therefore not generally cons idered legitimate military targets. On the other hand, political leaders might be direc ting combat operations or holding entire regimes together, which makes attacking them potentially justifiable targets on 7 A more likely scenario might be to lay out qualitative criteria which establish in some rough way the steps rebel groups can follow to avoid international censure. This approach, however, is also extremely unlikely. States have no incentive to give their violent oppositions guidelines on how to win legitimacy and thereby challenge them more effectively. 15


instrumental grounds. Schmid distinguishes te rrorism from assassina tion by arguing that in terrorism, the direct targets of violence are not the main targets, thus making terrorist assassinations de-individuated (1983, 8,11). 8 In practice, this is a very problematic basis for telling the phenomenon apart, as the distinction between ta rget and audience is ubiquitous to all forms of coercion (it only separates coercive vi olence from purely destructive violence aimed at the enemys co mplete annihilation). Furthermore, whether any given politician or his colleagues (whom he might have lobbied) are responsible for any particular policy cannot easily be established, but may ve ry well be argued as an expost justification for what was perceived to be an effective coercive strategy for other reasons. The question we must ask is this: what is it about civilians that makes them illegitimate targets, and do these reasons ex tend to political leader s? Civilians are off limits simply on normative grounds; even when targeting them confers a strategic advantage, doing so is prohibited on the basis of fundamental human rights. This principle is the closest we can get to an an swer. Assassination can th erefore be more or less terrorist in nature, depending on the context. When the speci ficity of the direct target becomes less important relative to the indi rect target, we can say it becomes more terroristic. Thus, Ganor argues that an a ttack aimed against government personnel should therefore be defined as terrorism if the target was not in a decision making position of the states Counter-Terrorism po licy (2002, 288). Obviously some states will object to this limitation, desiring instead a complete prohibition on the targeting of political leaders. 8 Schmid also sent out a questionnai re to the editors of news agencies television, radios, and the press, asking what types of political violence their medium labeled terrorism. 75% of editors reported labeling assassination as terrorism (Schmid 1983, 9). 16


Actor-centered definitions avoid these issues, but cr eate a few of their own. Actor-centered definitions cannot distinguish terrorism from other forms of political violence without reference to cap abilities. The basis of this approach is that there exists a discernable level of capabilities in which ther e is no military power, but only the power to hurt (Sanchez-Cuenca and de la Calle 2008, 35). This is a somewhat problematic distinction, for if the power to hurt can accomp lish military goals, how can it be distinct from military power? The answer is that m ilitary power involves the power to destroy physical capabilities (conventi onally defined) and to poten tially annihilate the enemy entirely. The power to hurt, on the other hand, is tailored towards destroying the enemys will to fight in the absence of the power to destroy (Sanchez-Cuenca 2007, 5). This distinction seems to bring us back to those features of action-centered definitions which Sanchez-Cuenca and de la Calle were so quick to discard: the creation of fear and insecurity, and the discrepancy be tween the targets and audiences of violence. Thus, the real essence of this focus on capabi lities is that the perpetrators of terrorism focus primarily on the audiences rather than the targets of violence, seeking to enhance their capabilities (and thus e nhance their future bargaining position) by playing conflict dynamics in their favor. Still, the ability to inflict punishment seems to be such a fundamental feature of coercion that it provides littl e guidance in determining what is or is not terrorist the power to hurt and the power to dest roy can never be completely separated. The authors therefore make a some what arbitrary decision when they deploy their theory usi ng the control of any territory as a proxy for m ilitary power. For SanchezCuenca, In principle, any actor might carry out terrorist deeds. In practice, however, terrorism is mainly practiced by terrorist orga nizations (2007, 6). This is because given 17


their material constraints, terrorist organiza tions cannot but resort to the kind of tactics that we tend to associate with terrorism (1). This definition therefore rests on a somewhat stylized distinction, and also re lies on the claim that territorial control produces unique conflict dynamics. In the end, however, this definition is far more reliable than any of the alternatives. Conclusion Terrorism can be defined based on either the means employed or the identity of the actor. Defining terrorism based on its mean s is smart from a lega l perspective because it can create a disincentive ag ainst attacking civilians, leadi ng rational rebel groups to reevaluate their actions and potentially shift away from civilian targets. The definitional leap from acts to groups is a serious problem for action-centered definitions. If terrorism is truly defined by target selection, and we wish to encourage a shift away from civilian targets, we must define a reasonable leve l at which groups can free themselves from stigmatization. If, on th e other hand, we are merely using the word as a pejorative to shift perceptions of legitimacy, we should not care whether 1 or 1,000 civilians are killed by a rebel group. Issues such as these will continue to create problems for international agreement on an action-centered definition. Additionally, it is not completely clear that te rrorists would respond to legal incentives at the international level. Even if they behave rationally, the perceived benefits of targeting civilians may be too great to discourage through, for instance, international cooperation that cuts off foreign sources of f unding. Before we can say with certainty that 18


an internationally agreed-upon action-cente red definition will mitigate human rights violations, we must examine the strategy of te rrorism, the process of target selection, and the extent to which terrorist groups operat e rationally. We cannot tell whether increased costs will change behavior unless we al so know the extent to which armed groups respond to incentives and the benefits they perceive to continuing thei r present course of action. On the other hand, defining terrorism based on the nature of the actor is attractive from a theoretical perspectiv e because it is analytically simple, and avoids a host of difficulties relating to intermediate categories like assassination and the qualitative grey area of civilian coercion. For this reason, an actor-based definition will be adopted for the remainder of this study. Terrorism will be treated as the use of clandestine political violence by an actor that does not control terri tory, but that is nevertheless able to mount a serious challenge to the states mo nopoly on the legitimat e use of force. Following Asal and Rethemeyer, control of territory is defined as the ability to coerce nonmember civilians to act or forebear and to exclude police and military units from some defined geographic space for a period greater than six months (2008, 442). Violence will not be considered political simply because it challenges the state monopoly in order to qualify, a political goal must be espoused. 19


Chapter 2 Strategy or Syndrome? Perspectives on Terrorist Decision Making A substantial body of literature seeks to e xplain the strategic value of terrorism. However, most of this literatur e does not explicitly address th e issue of target selection. This gap draws attention away from the aspect of terrorism most frequently described as fundamentally irrational and is therefore a w eakness for current theories on terrorism. For those who subscribe to the action-centered de finition of terrorism, failing to explain the targeting of civilians makes te rrorism indistinct from other types of political violence, rendering claims about its causes and strate gy extremely dubious. What we must explain in order to explain terrorism is not simply why some people are aggrieved and how they are organized, nor even why they resort to violence, but why they employ violence against civilians or noncombatants in particular (Goodwin 2004, 260-1). This knowledge gap is problematic in the political realm as well, because it matters a great deal whether a terrorist group or any political actor target s civilians instrumentally or 20


irrationally. Actor-centered definitions therefore face the same imperative to deal with the question of why civilians are the targets of terrorist violence. We can distinguish broadly between two opposing perspectives on terrorism: one which sees it as a tool to ach ieve an end, and the other whic h sees it as a syndrome that precludes the rational pursuit of alternative means. These perspectives will be referred to as the rational choice and psychological perspectives, respective ly. Between these two perspectives lies the organiza tional approach, which stresse s how the group dynamics of underground organizations shape an individu als psychology and therefore his or her potential for rational choice tailored to the ach ievement of external political objectives. Explaining any terrorist organization adequate ly will require a combination of all three perspectives, and in varying degrees. There is no reason to assume that all terrorist organizations behave according to the same l ogic. Violence, even of a specific type, may be used in different strategic contexts and for different purposes (Sanchez-Cuenca and de la Calle 2008, 38.). Unfortunately, the pr oponents of each of these perspectives sometimes over-generalize in a rush toward externally valid hypotheses, and therefore present the three perspectives as antagonistic rather than complementary. The Psychologica l Perspective The psychological perspective on terrorism examines micro-level variables, and focuses on the role of cognitive and affective distortions in the shaping of terrorist behavior. Cognitive distortions in this sense are perceptual deviations stemming from simplifying models used to interpret the worl d, while affective distortions are emotional 21


or personality factors that undermine a persons ability to see things as they really are. Psychological studies of terrorism concern th emselves with the processes through which these distortions influence decision ma king (McCormick 2003, 490; Reich 1990; Hallet 2008, 54). Psychological research on terrorism has clustered in four areas. The first attempted to locate psychologica l traits that terrorists mi ght share in common, with the hope of defining a terrorist personality. 9 In the end, the huge effort s put into this line of inquiry produced only two conclusions: that the vast majority of terrorists cannot be classified as psychotics or as suffering from any other diagnosable maladies of the mind, and that there is no stable profile of te rrorists or potential terrorists (Gupta 2006, 40). Some researchers have reacted to this resear ch by suggesting that terrorists are abnormal in more subtle ways. Most, however, si mply accept the fact that terrorists are disturbingly normal. The second line of inquiry has focused on the frustration-a ggression hypothesis, which states that individual and collectiv e violence often results from a discrepancy between individual expectation and achievemen t. Gurr is widely credited with bringing this hypothesis into the real m of political violence with his theory of relative deprivation. Feelings of deprivation, in his view, are arrived at subjectively and need not have any real basis in fact Nevertheless, these feelings can cause people to act out violently (Brush 1996, 527). Much more remain s to be said about relative deprivation 9 Most prominent of these studies was the project launched by the West Germ an Ministry of the Interior in 1982, which entailed the most exhaustive series of terr orist interviews to date. The project used 227 leftwing terrorists and 23 right-wing extremists. It revealed a number of patterns in the personal histories of the subjects that seemed significant co mpared to the comparable populatio n of West Germans. The results, however, were contradicted by existing psychological profiles from other studies (Reich 1990, 269). 22


theory. For now, it will suffice to say that the line from grievances and frustration to aggression was vastly oversimplified. Gurr hims elf has since abandoned the theory in its original form. A third line of inquiry has looked for th e sources of terrori sm in narcissismaggression models of violen ce. Narcissistically motivated violence is a type which enables the individual to defend his self from damage and harm. According to this research, narcissistic personali ties are subject to splitting and externalization, whereby negative self-images are split away and projec ted onto an external enemy who is subject to attack. This research suggests that, although other factors pl ay an important role in the individuals turn to terrorism, preexisting narcissistic injury is also an important source of terrorist behavior (McCormick 2003, 492). Acts of terrorism may serve to stabilize an individuals self-esteem. A fourth body of work on terrorist psyc hology views terrorism not as the product of a single decision, but as the final result of a long dialectical process that pushes an individual toward a commitment to violen ce. The process occurs within a political environment in which the terrorist group, th e state, and the gr oups self-designated constituency interact. This approach has the benefit of acknowledging the importance of individual personalities and the impact of social psychology, without the liability of describing terrorism as a pathology. Banduras contribution to this body of litera ture examines how terrorists justify their actions through a process of moral di sengagement. According to Bandura (1990), self-sanction plays a critical role in the regula tion of all human conduct. Internalized control mechanisms, however, can be sele ctively disengaged and reconstructed. This 23


change can be accomplished by redefining one s actions in moral terms, shifting responsibility, misrepresenting the conseque nces of ones actions, or blaming and dehumanizing the victim. This social process of moral disengagement becomes particularly interesting in the context of violence which is designed to hurt innocent third parties a process which will later be discussed in much greater detail. Another important contribution of the psychological perspectiv e relates to the difficulty of giving up terrorism. Terrorists tend to hold bifurcated worldviews that separate people strictly into good or evil, friends or enemies. Once in place, this belief system is highly resistant to change. Such pers istence is due not only to the nature of the underground, but also to defen sive avoidance. Defensive a voidance refers to the fact that, once a commitment to terrorism has b een made, the psychological costs of calling ones actions into question become extremel y significant, causi ng individual belief systems to become increasingly rigid in orde r to justify past behavior (McCormick 2003, 493). The Rational Choice Perspective The rational choice perspec tive on terrorist decision ma king views terrorists as strategic, value-maximizing actors. The perspe ctive has strong and weak variants, with each employing different assumptions related to the groups intellectual abilities and knowledge of its operating environment. The st rong or substantive variant, drawn from the idealizations of neoclassical economics, does not distinguish the world as it is from the world as the terrorist group sees it and assumes that terro rist groups can anticipate the 24


consequences of their actions. The procedural theory, meanwhile, assumes that terrorist groups may act rationally according to their perception of the world, but that this perception is an incomplete reflection of real ity. Rational choices in this perspective may lead to suboptimal outcomes due to in formation costs, information limitation, computational constraints, and time pressures. Too much information may also lead to suboptimal outcomes if it causes gr oups to filter, frame, and t hus distort their perceptions of reality (McCormick 2003, 482). Both variants of the rational choice or t ool view investigate the rewards of using terrorism and treat terrorists as ratio nal maximizers unaffected by psychological maladies. The strategic interaction between states and clandestine oppositions is structured by their relative strengths and weaknesses. According to McCormick, states enjoy a force advantage while clandestine groups hold an information advantage (2003, 484). Terrorist groups lack the cap ability to strike fatal blows, but are able to locate and strike their enemy at any time. States, on th e other hand, can easily dispose of terrorist groups if they discover the identities and loca tions of members, but striking without this information is a serious liability. This si mple but powerful formulation underlies the defining features of any conflict between th e state and an underground competitor, and it lays the foundation for the most common argumen t about the strategy of terrorism: that it aims to either coerce concessions (in most accounts, by eliciting fear or terror), or provoke a backlash which mobilizes more support for terrorists. Rapoport calls the backlash approach the po litics of atrocity (McCormi ck 2003, 484). More detailed analysis has produced over tw enty (mostly overlapping) goals of terrorism, including terrorization, intimidation, de moralization, polarization, and radicalization of the public; 25


the building of group morale; enforcement of control; disruption of control; mobilization of forces and resources; elimination of opposi ng forces or their resources; punishment for cooperation with the enemy; provocation of countermeas ures and repression; and advertisement of the movement (Kalyva s 2004, 98). McCormick also observes that terrorist groups win over passive supporters wh en they manage to perform attacks that demonstrate operational competence. T oo much violence, however, may be counterproductive if it exposes the group to high levels of arrest, or if it causes popular support to be withdrawn. Taken together, the need to project an image of accomplishment (the influence constraint) and the contradictory need to replenish the rank and file as members are arrested or killed (the securi ty constraint) establish the strategic space within which all terrorist groups must operate (McCormick 2003, 497). Crenshaw has been one of the most visible defenders of the rational choice perspective. She defines terrorism simply as underground political violence and argues that it is likely to be an informed choice am ong alternatives, some of which have already been tried unsuccessfully. Terrorism may beco me desirable due to popular repudiation of a groups ideology, a failure to mobilize e ffectively, or an in ability to mobilize effectively due to authoritarian enviro nmental conditions (Crenshaw 1990, 11-13). Crenshaw identifies a number of benefits that can make terrorism attractive: it serves an agenda-setting function (making grievances more salient), may facilitate revolutionary conditions by fomenting inst ability, serves an excitational function (inspiring resistance by example), and may pr ovoke indiscriminate repression against the population, thus pushing previously unmobilized groups into the arms of terrorists. For instance, the German Red Army Faction sought to provoke repression which would 26


demonstrate that West Germany was still a fasc ist state in disguise. Marighella attempted a similar strategy in Brazil in the hopes that discontent will spread to all social groups and the military will be held exclusively responsible for failures (Crenshaw 1990, 19). Terrorism can also be used to demonstrat e strength, as the Pr ovisional IRA did in the 1970s during periods of negotiations with the British (1990, 16). This decision, Crenshaw argues, demonstrated a calculat ed choice to strengthen their bargaining position while alienating public opinion (ibid). It is easy to see these attacks as selfsabotaging, however, and as proof that terrorists are simply more inclined to inflict violence than to achieve political goals. Cr enshaw argues that strategic motivation is merely one factor in the decision-making pro cess leading to terrorism but one crucial to counteracting stereotypes of terrorists as irra tional fanatics (24). The Organizational Perspective The organizational perspective draws attention to the fact that terrorist organizations are subject to a range of influe nces that may be only tangentially related to strategic goals and perhaps even opposed to them. These infl uences derive primarily from the fact that the organizations are both violent and underground. Looking at terrorist groups as politic al organizations reveals a number of behavioral tendencies necessary for their survival. The two ce ntral tendencies are derived from the influence and security constraints. The security constraint refers to the fact that as more attacks are carried out the authorities will acquire greater information about the group, and if group members are arrested or kill ed more rapidly than they are replaced, 27


the organization will not survive. The influence constraint, on the ot her hand, refers to the fact that terrorist groups must carry out attacks both to stay relevant politically and to attract resources and recruits Steering between these two constraints creates serious hazards, but is essential to the survival of any violent underground group. Secrecy is a serious requirement that can take up a great deal of time and resources. The effectiveness and long-run surv ival of the organization depends not only on having a secret, but keeping the fact you have a secret a secret (McCormick 2003, 486). The need to remain clandestine has th e general effect of isolating terrorist organizations from their larger social and pol itical surroundings. All else equal, the need for insulation from societal roots varies directly with changes in the security environment, which in turn vary directly with the organizations level of attacks. The more successful a group becomes, the more pressure it comes under to separate itself from its social and political base (ibid). Another important constraint is the preference constraint. This term refers to the fact that, to the extent that a terrorist organization relies upon a community of supporters, the moderation of supporter preferences w ill constrain the groups behavior. This dynamic can be conceptualized in a few diffe rent ways. We can think of supporters as existing in three more or less distinct groups. The first group consists of the organizations official members and othe r hardcore supporters. The second group consists of those who do not personally carry out the groups missions, but who nevertheless support its goals and provide it with resources and protection. The third group consists of those who support the gr oups goals but do not support its means. 28


It is often noted that groups who alie nate their supporters by using too much indiscriminate violence do not survive for l ong. It is far from clear, however, how this preference constraint really works. There is likely a diffe rential impact among the three types of supporters just descri bed. Hardcore supporters should be the last to withdraw support. Similarly, the pool of those who merely agree ideolo gically will probably not be highly responsive to the means employed. At any rate, loss of these supporters will not really damage the terrorist group. The most lik ely and most dangerous possibility is that those who provide material resources a nd cover begin to withdraw support. The consequences of this withdrawal can manifest in either the short or the long term. In the short term, supporters might become informers or create organizat ional splits. In the longer term a smaller constituency can make recruiting more difficult. This would make it more difficult to growth, to satisfy the secu rity constraint, and to make a generational transmission. As groups go further underground, deci sion making can become increasingly rigid, closed, and inward-looking. This tendency can reinforce preexisting interpretations of the environment and limit the groups ab ility to adapt. All organizations use simplifying mechanisms to establish priorities, offset deficiencies in information, resolve informational inconsistencies, interpret their surroundings, and evaluate the consequences of their actions. Terrorist or ganizations take these mechanisms to the extreme. As terrorist groups go furthe r underground, social solidarity among group members also becomes increasingly important. The tendency fo r groups to disconnect socially from all outsiders becomes a liability for future recruitment. 29


Terrorist groups may tailor their actio ns towards auto-propaganda by making themselves out to be soldiers, or as the defenders of a larger community in danger. Government excesses are regularly used both to justify a terrorist groups initial use of violence and to rationalize any future decision to escalate (McCormick 2003, 488). Groups also exhibit a bias towards ac tion. In DeNardos model of strategic decision making, a group that fails to achieve its objectives through nonviolence can take one of three paths: build the base, moderate demands, or escalate ta ctics to stimulate or replace mobilization (DeNardo 1985, 244). Escal ation to violence takes the form of terrorism because, without power in numb ers, going underground is the only way to protect group members from state violen ce. The fundamental difference between terrorism and other types of political violen ce is that terrorism emanates from the underground while rioting, looti ng, trashing, and pillaging rage in the streets for all to see (including the police) (229). The choi ce among the three paths, for DeNardo, is a function of three attributes: location of an ideal point, intensity of preferences, and political impatience. Moderates (in his formul ation) are impatient for results but have only weak preferences; basebuilders are patient and have strong preferences; terrorists are impatient but have strong preferences. The decision to go underground is often ma de deliberately to leave the nonviolent path behind. Some groups, like the Irish Republican Army (IRA) and Euskadi Ta Askatasuna ( Basque Homeland and Liberty; ETA), have sought to have it both ways by trying to maintain open connections with their political movements. But in most cases the relationship is impossible to maintain. It is also very common for terrorist organizations to adopt a pre-existing theory of victory rather than design one fit for their operational or 30


strategic objectives. The probl em is that these scripts are adopted uncritically. For instance, Carlos Marighellas Mini-Manual of the Urban Guerilla had a strong influence on the German Red Army Faction, the Italian Red Brigades, and the Canadian Front de Libration du Qubec none of whom met with much success the latter two provoked the state into repressing them right out of existence (Parker 2007, 169; Drake 1998b, 18). Similarly, Arrigo Cavallina of the Italian Ar med Proletarians for Communism admits to having systematically read the works of Ma rx, Lenin, and Mao in an attempt to turn himself into a professional revolutionary (ibid). Terrorist behavior can also be influenced by inter-organizational competition. Examples include the historic rivalries between the Irish Republican Army (IRA) and Irish National Liberation Army (INLA), th e Palestinian Libera tion Organization and Palestinian Front for the Liberation of Palestin e, and the various elements of the Basque nationalist movement. In these cases and others, terrorism is both an instrument of competition with the state and a means of crowding out ones political rivals (McCormick 2003, 488). The bias towards action is compounde d by an internal environment that encourages risk taking. The literature on the social psychology of small groups corroborates this argument. Decision making within highly insulated organizations is vulnerable to groupthink, which manifests in illusions of invulnerability, unquestioned belief in group morality, and a stereotyped view of the enemy as too evil to warrant negotiation (McCormick 2004, 489). The problem of groupthink is reinforced by the tendency toward self-censorship and c onsensus building. The importance of group solidarity grows as an organization distances itself from mainstream society. Though it 31


may not be apparent to members or outsiders, the need to preserve the group can supplant the original political mission entirely (490). Up until now the discussion has focused on insights from the organizational perspective that may supplement the rational ch oice (or strategic) perspective. There is, however, a version of the or ganizational perspective which views group survival as antithetical to strategic success. Abrahms lays out seven tendencies of terrorist organizations that he claims are all tailore d to group survival, but which undercut the groups strategic success. He presents thes e features as fundamental puzzles for the strategic view. They are 1) th e persistent willingness of terrorists to target civilians despite its coercive ineffectiven ess; 2) the fact that terrorism is never a last resort, and terrorists seldom abandon violence to become nonviolent political parties; 3) the reflexively uncompromising attitude of terrorist groups 10 ; 4) terrorists protean political platforms 11 ; 5) anonymous attacks 12 (actually very much exp licable in light of the preference constraint); 6) the fact that terrorist organiza tions with identical political platforms routinely attack each other more than their allegedly shared enemy; and 7) the 10 Failure to compromise is inconsis tent with the rational choice persp ective for Abrahms because he sees terrorism as an extremism of means, not ends (2008, 82). Abrahms also points out that most bargaining theorists do not accept issue indi visibility between rational adversaries as a viable explanation for conflict because the complexity and multidimensionality of issues enables parties to find linkages and side payments to create mutually beneficial ba rgaining solutions (Abrahms 2008, 87). 11 Al Qaeda has found many different wars and types of enemies; ETA went from trying to overthrow Franco to targeting the emergent democratic government. The PKK has vacillated between advocating jihad, a Marxist revolution, and a Kurdish homeland gov erned without Islamist or Marxist principles. The Abu Nidal Organization relentlessly attacked Syria in the 1980s, then switched allegiance almost overnight and became a Syrian proxy. Laqueur further notes that many well-known groups (e.g. Argentine Montoneros, Colombian M-19, Popular Front for the Liberation of Palestine) that began on the extreme right ended up on the far left as far as their phraseology was concerned, (Abrahms 2008, 88-9). 12 According to Abrahms, 64% of worldwide terrorist attacks have been anonymous since the emergence of modern terrorism in 1968, and this figure has risen to 75% since September 11 th 2001 (2008, 90). Anonymous attacks are very much explicable when we consider the propensity to make mistakes either through collateral killings or a misreading of internal target audience preferences. The security constraint may also create an incentive to engage in less discrimi nate attacks to put greater pressure on the state, but the attacks may not be claimed after the fact to avoid censure from the groups in ternal target audience. 32


resistance of terrorist groups to disband when they consistently fail to achieve their political platform or when their stated grievances are resolved (Abrahms 2008, 82). Natural systems theory emphasizes the discrepancy between a groups stated goals and the goals of its members, and focu ses on the individuals desire to develop strong affective ties with othe r group members as the most important incentive to join. Terrorists, according to this view, are le ss policy maximizers than they are solidarity maximizers. If Abrahms is correct, there are significant implications for counterterrorism policy. Rather than focus on deterrence and the denial of concessions, governments should take advantage of the fact that terro rist groups are composed of networks of friends and family, with ideology largely entering through the back door (Abrahms 2008, 104). Counterterrorism policies should seek to drive wedges between the members of terrorist organizations. For Abrahms, this policy has created the only success in counterterrorism since the advent of modern terrorism in the 1960s. 13 But if terrorists are really just solidarity maximizers, what prevents them from joining the Boy Scouts, playing sports, attending parties, climbing mountains, or joining a non-violent secret society, such as the Ma sons, or Skull and Bones? According to Abrahms, terrorist groups are far more tigh tly knit than other voluntary associations because of physical danger, high costs of pa rticipation, and their violation of societal expectations, and the members of terrorist or ganizations seek this type of solidarity (2008, 100). Abrahms suggests that this may explain why committing acts of terrorism generates new recruits and boosts member mo rale, even while the groups fail to achieve 13 His evidence comes from the successful rounding up of the Red Brigades in Italy. Not surprisingly, he suggests that similar deals should be cut with Al Qaeda when detainees involvement and likelihood of rejoining are minimal. Attacking the social bonds of terrorist groups may also be accomplished by investing more in seeding double agents (Abrahms 2008, 105). 33


their political goals. But this explanation may miss the mark acts of terrorism could also boost recruitment among those with similar id eological commitments who are desirous of decisive action (i.e., it may serv e an excitational function). Abrahms seems to conflate the motiv ation to join a movement with the disincentives to leave once a commitment is ma de. The tendency of individuals to stay in a group is probably better expl ained by the lock-in effects resulting from close personal relationships, the significant psychological costs to admitting guilt for unwarranted violence, and the serious physical danger of backsliding against tr emendous pressure toward conformity. Terrorists are always on the hunt for backsliders, and when found, they treat them harshly. The Japanese Red Army killed roughly one third of its own membership (who were said to have died of defeatism) in the two months between December 1971 and February 1972. At leas t 12 members were killed, with some apparently buried alive. The Abu Nidal Organi zation has also been responsible for killing a significant percentage of its own membership over the course of its operational life (McCrormick 2003, 490). The need for fraternization is probably not a good explanation for why individuals join terrorist groups. If it matte rs at all, it is probably condi tioned by other factors such as the personal experience of repr ession. In a study on recruitment for the Provisional Irish Republican Army (PIRA), White found that be fore endorsing political violence, the individual must experience repr ession, view the repressing authority as illegitimate, view peaceful protest in the fact of repression as ineffective, and consider the reactions to repression of people with whom they have close ties (White 1989, 1227). He also found that responses to repression tend to be conditioned by class background. Specifically, 34


members of the working class a nd student activists are more likely than members of the middle class to experience repression, be available for costly violent protest, and experience the efficacy of political violence (ibid). Synthesizing Strategic and Organizational Perspectives At least at the cross-national aggregat e level, Abrahms launches an effective critique of the strategic pe rspective. Terrorists are cl early operating under a mix of motivations which are not always oriented to political outcomes. But we should not underestimate the possibility that they are making mistakes ra ther than making calculated decisions to enhance social solidarity. In ot her words, we should not be too quick to impute the effects of their behavior to clearly elucidated motiv ations. Terrorists may focus too much on tactical victories without knowing how to tie them into an effective strategy. Crenshaw suggests that organizational th eory may provide a way of completing the strategic theory by determining the valu es of oppositions, how their preferences are determined, and how intensely they are held (2001, 28). She also suggests that the two approaches could describe different types of organizationa l categories into which real groups can be fitted. Some groups might closely approximate the strategic model, while the decisions of others may be heavily influe nced by organizational politics. For instance, if there is extensiv e popular support for an organizati on, the state will have a more difficult time severing the links between the group and its base of support, which will 35


mitigate the effects of the security constraint, create greater room for maneuver in terms of the preference constraint, and lighten the burdens of secrecy. Like any other social movement organi zation, terrorist groups do not have the luxury of existing as abstractions in an ideal world. They must maintain themselves as an organization in order to pursue their objectives, and the nature of that organization creates both cognitive and behavioral constraints. The question is no t which perspective (psychological, organizational, or strategic) is best, but rather when and under what circumstances the different levels of analysis are most revealing. Even the most perfectly rational terrorist group operates in a re al setting, and must react to organizational imperatives. The clande stine requirement forces groups to sever or at least limit contacts with nonviolent allies and therefore prevents them from obtaining full information on their operating environment. The security constraint creates an incentive to use increasingly indiscriminate violence over time so that the group can cause greater pain to the state while limiting its visibility. The infl uence constraint also tends to push groups towards violence, because without a demonstrable operational capacity, they will lose publicity and fade into obscurity. Working from the groups internal logic, the only factor which would appe ar to limit a terrorist groups violence is the preference constraint of supporters. Comb ined with the preference constraint, the security constraint becomes much more im portant, as groups are no longer so free to engage in less disc riminate attacks. 36


Evidence on the Strategies and Successes of Terrorism The rational choice literature is complicated by the fact that it encompasses (but does not always acknowledge) multiple levels of rationality used at both the tactical and strategic level. Unfortunately, there are ma ny debates on whether terrorism is a rational endeavor (or whether it pays) in which scho lars refer to the achievement of different levels of goals using different standards of rationality. Acts of terrorism can achieve tactical vi ctories, such as receiving media coverage that helps to put issues on the national agenda, but terrorist groups often fail to link tactical concerns with strategic success. Terro rism nearly always fails at the strategic level, and terrorist groups that primarily target civilians always fail. 14 There are two major exceptions to this formulation. First, when terrorism is employed as a strategy of national liberation against a foreign occupier, whose claims to legitimate rule are, by definition, dubious, terrorism has been eff ective relatively often. Anecdotal evidence about terrorisms effectiveness is almost alwa ys taken from national liberation struggles. The second exception relates to the possibility of conflic t escalation. When terrorist groups manage to take control of territory and transform themselves into rebel groups, they are able to extract greater resources (f or instance, through taxation) and to recruit militants in the open. Greater resource mobilization translates into greater coercive power over the state, which enhances both the pow er to hurt and the power to destroy. If 14 A potentially important exception to this rule is conservative groups operating with tacit government support for instance the Ku Klux Klan in the early years of reconstruction. 37


terrorist groups manage to tr ansform themselves into an insurgency, they drastically improve their chances of coerci ng the state into concessions. 15 In 2006, Abrahms assessed the success rates of the twenty-eight groups placed on the Department of States list of foreign terro rist organizations. He found that the groups accomplished their forty-two policy objectives only 7% of the time, and that groups whose attacks on civilians outnumber attacks on the military never achieved their objectives. The strength of this conclusion is bolstered by his decision to use very generous assumptions that limited the number of failures he counts both total and partial success as policy su ccesses, treats only complete ly unsuccessful outcomes (no success) as failures, 16 counts objectives achieved before the groups were placed on the foreign terrorist organization list in 2001, lim its the objectives of Al Qaeda affiliates to their nationalist struggles, and perhaps most importantly, attr ibutes all policy success to terrorism, even when there are intervening va riables that may have contributed to the outcome (Abrahms 2006, 51). In their study of 648 terrorist groups, including 268 that effectively ended (excluding splintering), Seth Jones and Martin Libicki found that peak size was the only variable significantly relate d to a groups longevity (2008, 40-1). They also found that larger groups are more likely to either achieve their goals or ente r politics, although this correlation is conditioned by the breadth of th eir goals (2008, 39). According to Jones and Libicki, 10% of terrorist groups that have ended since 1968 did so because they 15 The recent terrorist campaign in Nepal followed this pattern successfu lly. The Communist Party of Nepal (Maoist) waged a 10-year civil war to remove the m onarch from power, and was pl aced on the U.S. list of international terrorist organizations. In 2008, the Maoi sts formed an alliance with the seven major political parties of Nepal to permanently remove the monarch from power. Since then, the Maoists have won the largest number of seats in the Constituent Assembly and their leader was named the Prime Minister. 16 Partial success was treated as neither a success nor failure. 38


achieved their goals. In most cases, terrorism had little or nothing to do with the outcome with the important exceptions of Israel Algeria, and South Africa (2008 19, 33). 17 Jones and Libicki adopt an action-centered defi nition of terrorism inclusive of groups that have acquired territory. Looking at insurgent groups only, they find a 25% rate of victory for all groups that ended, w ith the percentage of groups engaged in insurgencies clustering at and above the 1,000 member mar k. Nearly half the tim e, they found that insurgencies ended with a negotiated settlement (Jones and Libicki 2008, 100). The fact that terrorism can be successful particularly when it evolves into an insurgency suggests that it is inappropriate to view terrorism as a strategy that never succeeds, and therefore the result of either psychopathology or ir rationality. Empirical evidence suggests that nationalist terrorism is especially likely to achieve its goals. Part of the reason for this outcome is that nati onalist terrorism employs a strategy unique from other types of terrorism. This difference ha s an important impact on both organizational resources and target selection strategies. An investigation of vari ations among different types of groups and the importance of target selection strategies for group success is the subject of the next chapter. 17 The authors argue that terrorism was an important, but not sufficient, condition for both the creation of Israel and Algerian independence (2008, 33). 39


Chapter 3 Why Target Civilians? Theories of Strategy and Legitimization Up until now the discussion has focused on the political goals of different terrorist groups, the organizational constraints within wh ich terrorist groups act, and the way that these constraints structure decisions. This ove rview has allowed us to frame the decision making process of terrorist groups. This chap ter will begin to deploy these ideas toward an understanding of why groups target civili ans. It will do this by distinguishing among different types of terrorist groups and laying out the theories that conne ct their violence to a political strategy. The focus of the investigat ion will be on the strategic implications of insurgent violence against civilians, and on th e processes of social legitimization that allow such violence to be sustained over time. 40


Classifying Terrorist Organizations Many classification schemes have been appl ied to terrorist organizations (TOs). Some common labels for TOs include ethno-nationalist (or separatist), religious, leftist, anarchist, communist, fascist, conservative, single-issue, and criminal (Drake 1998b, 17). Which of these categories are useful beyond mere description? The rational choice perspective indicates that we can improve our understanding of te rrorist group behavior by distinguishing among their strategic goals. Looking at TOs from this perspective, we can distinguish among three types of goals: re gime change, territorial change, and policy change. Nationalist (separatist) TOs pursue territo rial change by repeatedly punishing the state in a war of attrition. Th ey may also engage in indisc riminate attacks against rival ethnic groups to make centrali zed governance less feasible (however, existing literature indicates that this tactic is not generally pursued). Regime change is sought by revolutionary, fascist, and some religious organizations. 18 Their primary tactic is to create disorder. Revolutionary groups create disorder by attacki ng symbolic targets and thus spreading propaganda of the deed. Their hope is that disorder will create the conditions conducive for a popular uprising. Fa scist groups create disorder to force the public into choosing authoritarianism ove r chaos (Sanchez-Cuenca and de la Calle 2008, 38). There is less to unite policy-oriented groups. Th ey may be largely nonlethal and focused on single issues (such as extreme environmenta list groups), or conservative and focused on maintaining the status quo by intimidati ng some section of the population into 18 Al Qaeda is an interesting case in point it is the only terrorist organization which has pursued regime change in multiple countries, and it has done so simultaneously. 41


compliance (e.g., the Ku Klux Klan). Evidence indicates that police forces are more likely to repress left-wing dissidents than right-wing dissidents (White 1995, 333). This finding is consistent with the argument that states are more likely to react with violence against those who pose a gr eater threat to their rule (ibid) White also finds evidence that British soldiers in Northern Ireland did not repress Loyalist and Republican paramilitaries equally. Instead, British sold iers appear to have followed the lead of Loyalist paramilitaries in attacking the Catholic community (White 1995, 347). One key difference between nationalist gr oups and revolutionary groups is in the breadth of their goals. 19 Abrahms presents evidence that terrorist groups with maximalist goals are less likely to have their demands met. 20 Even more important, according to Abrahms, is a groups target selection. Abra hms argues that terrorist groups that mainly attack civilian targets create the impression of holding maximalist goals, and therefore miscommunicate to their external target audience. Among the groups in Abrahms sample, even those groups that expressed lim ited, ambiguous, or idio syncratic objectives failed to win concessions when they primarily attacked civilians (2006, 76). This finding reinforces the puzzle of why terrorist groups choose their targets so poorly. Ideology is an important determinant of target selection. 21 Ideology transforms people into symbols. This transformation is not limited to terrorist violence 19 Empirical research on interstate bargaining demons trates that limited issues are more likely to be resolved than maximalist. More recently this insight has been applied to explain the unusual longevity of civil wars. Civil wars defy political resolution, according to this argument, because they are frequently fought over competing ideologies where the costs of retreating are high (Abrahms 2006, 53). Similarly, Jones and Libicki report that the possibility of a political solution is inversely linked to the breadth of terrorist goals. Most terrorist groups that end because of politics seek narrow policy goals (2008, xiii). 20 Abrahms examined twenty eight terrorist organizations He found that coercion succeeded in three out of eight cases when territory was a goal, but failed in a ll twenty cases when groups aimed to destroy a target states society or values (2006, 55). Maximalist goals undercut the states incentive to bargain. 21 Drake defines ideology as the beliefs, values, principles, and objectives however ill-defined or tenuous by which a group defines its distinctive political identity and aims What is important is that ideology provides a motive and framework for action (Drake 1998b, 2). 42


rationalization and dehumanizati on are both common themes of war. The basic aim of a nation at war in establishing an image of th e enemy is to distinguish as sharply as possible the act of killing from the act of murder by making of the former an act deserving all honour and praise (Drake 1998b, 7) The glorification of killing may also extend to civilians, as when the British just ified aerial bombings against German cities by describing them as revenge for earlier raids by the German Luftwaffe. These missions were also justified as attacks on German economic capacity, or as th e result of the moral depravity of the German people (ibid). Demo cratic states as well as terrorists may dehumanize civilians when it seem s to be in their interest. Ideology is important for a terrorist groups target selection because it provides a frame for its worldview and sets out the moral parameters within which they operate (Drake, 1998b, 1). There are currently no sy stematic quantitativ e analyses relating ideology to target selection. Nevertheless, it has been argued, somewhat anecdotally, that we can distinguish different t ypes of target selection based on generic cate gories of group ideology. According to Wragg, Left-wing terrorist organizations are believed to exercise more care in the selection of targets in order to avoid harming members of the proletariat they claim to represent (Karber and Mengel, 1983; Fine, 2008). The disinclination to indiscriminate attacks among many left-wing terrorist organizations might enhance their survival prospects, as indiscriminate attacks often result in a loss of support from sympathizers and moderate members, (Wragg 2008, 2). There is some limited evidence to indicate that left-wing terrorist gr oups last longer than other types, and that their attacks tend to be less lethal on average. 22 In Italy, during the 22 Crenshaw analyzed the lifespans of 77 non-randomly selected terrorist organizations, and found that revolutionary (leftist) and nationalist groups are about equally represented among those groups lasting over ten years (Crenshaw 1991, 79). This finding supports th e proposition that left-wing groups last longer than the average group we would not expect them to last longer than nationalist groups, which last for longer than usual for multiple reasons. 43


1970s and 1980s, fascist TOs utilized completely indiscriminate attacks against civilians, setting off bombs in packed st reets, trains, and railway sta tions. Leftist TOs were more selective and also preferred to wound th eir victims often by kneecapping (SanchezCuenca 2007, 25). During this period the number of fatalities per leth al attack was 1.2 among revolutionary TOs, but 3.5 for fascist TOs. By comparison, the average number of fatalities per lethal attack among Basque nationalist TOs was 1.4 (Sanchez-Cuenca 2008, 24). Hoffman argues that left-wing TOs are the most selective because their overriding tactical and ethical imperative is to tailor th eir use of violence to appeal to perceived constituencies (1998, 158). They generally focus on kidnapping and assassinating those they blame for economic exploitation or pol itical repression. At the opposite extreme, religiously motivated terrorists engage in mo re indiscriminate violence against a far wider category of targets, potentially encompassing anyone who does not share their faith. Ethno-nationalist or sepa ratist groups arguably fall somewhere in between these two models (ibid). In Northern Ireland and Spain, separatist te rrorists have tended to attack those who are either members of or cooperate with organizations they s ee as representing the foreign occupier. Extreme righ t-wing terrorists in Northe rn Ireland and Columbia, by contrast, used terrorism to protect the existi ng state of affairs. However, because it is much more difficult for them to identity th e terrorist groups they oppose, they attack political activists or random members of the ethnic community, whom they view as either terrorist sympathizers or at least culpable for not putting more pressure on the terrorist group to end their campaign (Drake 1998b, 12) Evidence indicates that many ordinary 44


Protestants in Northern Ireland suspect tacit Catholic complicity with Republican terrorist groups (Drake 1998, 31). Ideological differences, as well as operat ional and strategic f actors, have also been found to create variati on in target select ion patterns of communist TOs in West Germany and Italy. Briefly, the West German groups saw the destruct ion of international capitalism and imperialism as their ultimate goals, while the Italian groups focused much more on national and local issues (Drake 1998b, 21). Thus, the German groups tended to focus more on international businessmen, the U.S. military, and symbolic attacks that made statements about interna tional affairs. The Italian groups focused more on targeting the police and lower-level industrial managers Given the emphasis that some communist ideologues have placed on adaptation to loca l conditions, we might expect more strategic variation within this category of TOs. Drake corroborates this proposition, observing that in most communist terrorist campaigns, ideology and strategy have been adapted to local conditions. This is in line with the thinking of communist ideologists such as Mao and Debray, who have both emphasized that the military strategy to be adopted in any given conflict is specific to that conflict (1998b, 19). Sometimes the justifications for attacks are only supplied after the fact. For instance, Drake notes that higher-ranking members of the PIRA seem to have tried justifying attacks after they occurred, even though they would not have sanctioned them with prior notice. This pattern indicates that the leadership does not exercise full control over the cells operating under their organizations name, but feels the need to maintain an illusion of control. Terrorists have also widened the limits of what can be considered a legitimate target when doing so facilitates the execution of attacks (Drake 1998b, 9). Terrorist attacks are not al ways preceded by detailed ideological considerations. When targets are easily identified and thei r guilt has already been decided, target 45


selection is rather straightforward. Drake th erefore argues that th e PIRAs decision in 1970-71 to kill British soldiers was simply a matter of resuming war against a traditional enemy. Republican ideology had long consider ed British soldiers as legitimate targets, since they were seen as representatives of an occupying force. A PIRA spokesman in 1987 explained that local units did not need permission from higher-ups before carrying out a bomb attack against soldiers they were automatically legitimate targets (Drake 1998b, 8). It is commonly agreed that religi ous organizations are among the most indiscriminate, and evidence for this pe rspective is more readily available. 23 Religious groups are believed to be more indiscrimi nate because they are able to address themselves solely to God, which leads to an almost open-ended category of targets and few moral sanctions. Another factor which can make religious groups and others more indiscriminate relates to the so cial attitudes that result from rigid social partitioning. Asal and Rethemeyer (2008) call this othering. In its most basic form, the argument is that terrorists will not target an entire category of people randomly if doing so might alienate potential support within that category. However, clear di viding lines between members and others, like those based on religion or ethnicity, will make it easier to see all those on the other side as legitimate targets. Categorically att acking members of a separate group may also facilitate a boundary activation process, which could potentially set off an ethnic security dilemma (Asal and Rethemeyer 2008, 438; Wintrobe 2005, 131). 23 For instance, Asal and Rethemeyer report that organizations with religious and religious-ethnonationalist ideologies are more lethal than other types of gr oups (Asal and Rethemeyer 2008, 446). This finding confirmed a research hypothesis that was based on the idea that religious groups are more indiscriminate than others. Presumably, indiscriminate violence was in fact the reason that these groups were especially lethal. However, the authors do not explicitly state this in their results. 46


Abrahms argues that target selection is the most important variable affecting a terrorist groups success. His th eory is that attacking civilians is never effective for terrorist groups because it conveys an uninte nded message that the groups goal is to destroy society or its values. This explanation relies heavily on correspondent inference theory, which was developed by psychologist Edward Jones in the 1960s and 1970s to explain how observers infer an actors motivatio n. In Jones theory, the attribute-effect linkage shows how the motives of actors are pr esumed to be encoded in the effects of their behaviors, with levels of corresponde nce describing the extent to which actors objectives are believed to be encoded in the outcome of their behavior (Jones and Edwards 1979, 222). According to Abrahms, terrorism has an extremely high correspondence civilians imput e the intentions of terrori st violence rather than accepting what terrorists say. Thus, they will not believe that terrorists desire limited ends, even when this is what terrorists clai m. If this is true, attacking civilians is never strategically beneficial. Targ et countries view the negativ e consequences of terrorist attacks on their societies and political syst ems as evidence that terrorists want them destroyed, and are therefore skeptical that co ncessions will placate terrorists motivated by such maximalist goals (Abrahms 2006, 59). Abrahms uses three cases to support his argument: US perceptions of Al Qaeda following the attacks of September 11 th 2001, Russian perceptions of Chechnya after the apartment and hotel bombings in 1999, and Isr aeli perceptions of Pale stine after the first intifada. He also argues that his theory ma y be able to explain how the communities in which terrorists thrive impute state motives from the consequences of their counterterrorism policies (Abrahms 2006, 78). His argument is complicated by the fact 47


that government officials appear to be th e foremost proponents of equating maximalist means with maximalist ends. It is clearly a mistake to presume that governments do not promote these views strategically. Furthermor e, as Abrahms admits, civilians tend to become disillusioned of such high-corresponde nce inferences over time. The tendency of civilians to accept this framing readily indi cates that correspondence theory has some merit, but it is certainly not the only factor at work. Decades of research in social psychology sh ow that the most reliable source of ingroup cohesion is outgroup threat. Terrorist [and count erterrorist] attack is an outgroup threat. Along with high cohesion comes a constellation of changes in ingroup dynamics that are sometimes referred to as the authoritarian triad: idealization of ingroup values, increased respect for ingroup leaders, and increased sanctions for those [who] violate ingroup norms (McCauley 2008, 5). This alternative explanation helps to account fo r the decrease in correspondent inferences over time, and calls attention to the sensitivit y of knowledge creation to material context. Group ideology sets the parameters for targ eting strategies, but it does not explain changes in that strategy over time. As Abrahm s points out, terrorist groups can alter their ideology quite dramatically. Ideology is too loos e to determine a strict course of action in the face of changing incentive structures. C hoosing a conflict strategy may therefore have more to do with estimations about the extern al environment and the need to maintain popular support. Strategies of Violence Abrahms suggests four explanations for why terrorists attack civilians, in spite of the strategic liabilities of doing so: 1) terrorist groups ex aggerate their ability to coerce policy change, 2) terrori st groups attach equal importance to achieving intermediate goals 48


and end goals, 3) terrorism may be a superior st rategy to others, such as peaceful protest, even if it almost never achieves its end goals, and 4) only comp aratively weak groups target civilians because attacking the milita ry requires greater combat sophistication (2006, 77). These hypotheses largely miss the mark the first three explain the differences between political violence a nd nonviolence instead of violence against civilians versus the government. Organizati onal weakness is the only explanation that even addresses the phenomenon in questi on, and Abrahms himself describes it as empirically dubious, since nascent terrorist groups generally focus their attacks on military targets and then graduate to attacking civ ilian targets (Abrahms 2006, 77 [emphasis added]). Another answer to this question comes from the literature on violence in civil war. Hultman suggests that, because using violence against civilians is militarily cheap but politically costly, it sends a signal of resolve to incumben ts (Hultman 2008, 32). Hultman finds weak support for the hypothesis that a part y that loses battles will compensate for the perception of weakness by ta rgeting civilians and sending a costly signal. She finds stronger support for the hypothesis that rebe ls are more sensitive to losses than governments, and that the more a rebel group is losing, the more civilians it is likely to target (Hultman 2004, 21). It is also costly for the government to protect the population from attacks (27). Stathis Kalyvas investigates the states us e of indiscriminate violence. According to Kalyvas, the logic of indiscriminate violence is that when the guilty cannot be identified and arrested, violence should ta rget the innocent who are associated with 49


them. 24 The assumption is that the innocent will either force the guilty to alter their behavior, or that the guilty w ill change their course of action when they see its impact on the innocent (Kalyvas 2003, 112). Indiscriminate violence may be used initially because it is cheaper than its alternative, but over time, as evidence accumulates that it is counterproductive, it should be abandoned. The failure to adapt can be prolonged by institutional distortions but not indefinitely. Kalyvas observes that indiscriminate violence may fail because it overestimates the ties between political actors and civilians (2003, 115) In a civil war context, indiscriminate violence inflicted by government s assumes that civilians are able to lobby insurgents, that they have access to and influence over them, and that insurgents care about civilians. There are a few cases where insurgents reduced or ended their activities due to damage sustained by the population. Th ere are also a few documented cases in Greece and Lithuania in which ci vilians lobbied rebels into suspending their activities (122). However, in almost all cases examin ed by Kalyvas, insurgents disregard such lobbying. Civilians may often blam e insurgents for government re prisals, but they cannot act against them without government protection. The lack of local information leads incumbents to overestimate the strength of ties between incumbents and insurgents (Kalyva s 2003, 123). This reaction defeats the point of indiscriminate violence, and shapes id entities endogenously (ibid). In a study of a Catholic ghetto in Belfast, Jeffrey Sluka found that the stereotype that all people in Divis 24 Indiscriminate violence can have two distinct meanings. Either the targets of repression are selected without adequate care (scope), or the proportionality of attacks is poorly controlled (scale). The former can be described as indiscriminate targeting, the latter as indiscriminate force. Indiscriminate scope can also be referred to as categorical (Goodwin 2006). Kalyvas uses the term indiscrimina te violence to refer to indiscriminate targeting. 50


either belong to or strongly support the IRA or INLA caused ma ny who did not support them previously to become sympathizers, supporters, or even members (ibid). A more sophisticated varian t of the concept of other ing presented by Asal and Rethemeyer can be found in Goodwins theory of categorical terro rism. Goodwin argues that terrorists never attack indiscriminately, but instead socially construct their enemies, attacking categories of comp licitous civilians. Civilians are seen as complicitous insofar as they routinely be nefit from the government actions opposed, support the government, or have a substantial capacity to influence or to direct the government (Goodwin 2006, 2037). According to Goodwin, the goal of attacking complicitous civilians is to induce them to stop supporting or even to proactively demand changes in certain government polic ies or in the government itself. 25 This is precisely the argument presented by Kalyvas, but with one important difference: Goodwin also provides a theory for who is targeted. The precise categories of civilians viewed as complicitous depend on how revolutionaries c onstrue the existing pol itical order they want to change, with different categories generally associated with different regime types (ibid). At one extreme, complicitous civilia ns are limited to the cronies of the ruling autocrat. In a democratic context, the en tire body of citizens may be considered responsible for the governments actions. 26 Finally, in an ethnocracy, complicitous 25 The argument can easily accommodate ot her goals such as provocation, internal excitation, or direct government pressure unrelated to citizen lobbying. 26 Nemeth (2006) claims to find empirical support for the hypothesis that terrorist attacks in democratic states target the general population, while terrorists in autocratic states target institutions (like the military) that keep the leader in power. This finding seems to provide partial preliminary evidence for Goodwins theory of categorical terrorism. Ho wever, we should be very cautious in interpreting the results because Nemeth relies exclusively on the ITERATE dataset, which does not cover domestic terrorist events. The two most comprehensive databases on international and domestic terrorism are ITERATE and the Global Terrorism Database (GTD). Comparing them shows th at domestic terrorist events occur seven times as often as international terrorist events. In reality the size of this difference is certainly much greater the media attention created by international terrorism makes it much harder to miss. 51


civilians could include the entire dominan t ethnic group (ibid). How these political regimes are construed is complex and depe nds, among other things, on the organizations ideology, collective memory, and practical expe rience (as filtered through ideology). Note that this second formula tion is significantly di fferent from the first one presented by Goodwin, giving more weight to micro-level variables like perceptions and experience than to objective features of the group deemed complicitous. Goodwin concludes with three hypotheses: 1) Regimes with a history of state terrori sm are more likely than those without to experience categorical terrorism (making stat e terrorism a necessary but not sufficient condition for extensive categorical terrorism), 2) Regimes whose citizens are less bounded 27 from government decisions are more likely to experience categorical terrorism, 3) Countries with le ss cooperation or fewer cross-cutt ing political alliances between rebels and complicitous civilians are more likely to experience categorical terrorism. Whether militants adopt the widest range of categorical targeting will depend on a number of factors. Low orga nizational capacity may push groups towards less easier, less discriminate violence. However, many cost s might deter categoric al attacks against citizens deemed complicitous. For instance, categorical attacks may be avoided if complicitous civilians are potential allies or revolutionary members, if nonviolent appeals may influence complicitous civilians more eff ectively than the thre at of violence, if categorical terrorism may anger or repel members and supporters of the movement, if it may harm or prevent alliances with actual or potential sympathetic third parties, if it provokes state repression that constituents blame on the revolutiona ries, or if it is 27 For instance, Goodwin quotes a leader of Hamas wh o argued that There are no civilians in Israel because every citizen is required to serve in the army (2004, 261). Similarly, more representative democracies should be more likely to enable the belief that average citizens are responsible for the policies of their government. 52


expected to provoke repressi on which severely weakens or destroys the movement (Goodwin 2006, 2039). Clearly, ideology is not the whole story. Ta rget selection is also affected by the resources of the group, the s ecurity environment within which groups operate, and the reaction to violence amongst va rious target audiences (both internal and external). As Drake observes, no single cause adequately explains terrorist target selection (1998b, 2). There is no reason to assume that the choice to attack civilians will always be motivated by the same tactical considerations. Governments may do it because they think they can intimidate the public into compliance, and if right-wing groups have the implicit backing of the government, they may beha ve similarly. Terrorist groups may also calculate that they can provoke a backlash th at helps them mitigat e their own collective action problem. This dynamic presents a unique opportunity for ethni cally based groups in particular, as increasing violence may facilitate a boundary activation process. Terrorist groups may also attack civilians in retaliation for attacks committed by their government or by others in their community. From a tactical perspective this strategy may be advantageous giving vent to th e desires of potential supporters could win resources and recruits. Finally, civilian casual ties may simply be accidental if terrorist groups escalate their tactics a nd set off more bombs, it may be difficult to ensure that civilians stay out of harms way. 53


Target Selection and the Preference Constraint Terrorist groups operate with unstate d limits on their use of violence. Many groups meet their demise when they cross this limit, causing crucial resources to be withdrawn rapidly by repulsed constituents (McCauley 2008, 4; Wragg 2008, 2). As the former PIRA 28 member Eamon Collins explains, we carried the guns and planted the bombs, but the community fed us, hid us, opened their homes to us, [and] turned a blind eye to our operations. the IRA tried to act in a way that would avoid severe censure from within the nationalist community; they knew they were operating within a sophisticated set of informal restrictions on their behaviour, no less powerful for being largely unspoken (de la Calle and Sanchez-Cuenca 2006, 6). Terrorists depend on their base of sympat hizers and supporters beneath them for cover, information, money, and weapons. Most of all they need the ba se as a source of replacement recruits. If the capacity to replace losses is outpaced by the rate of attrition, terrorist groups will violate the security constr aint, and they will cease to exist unless an adjustment is made (McCormick 2003, 495). Indiscriminate violence may also cause irreconcilable internal schism s within the organization. On th e other hand, terrorists must also satisfy the influence constraint they must maintain a certain violent presence in order to stay in the headlines and command po litical attention. If they cannot meet this constraint, their ability to mobilize support wi ll be undermined and they will not be able to last (McCormick 2003, 496). Steering between these two constraints creates hazards for terrorist groups. 28 A split formed within the IRA in D ecember of 1969. The group that stayed with th e original leadership renamed itself the Official IRA (OIRA), and the disse nters labeled themselves the Provisional IRA. The Provisionals split off from the main group because they were worried that the demilitarization of the IRA, combined with an overemphasis on politics, would lead to an organization that spoke about a united Ireland but would in fact become mired in constitutional politics (White 1997, 26). The PIRA has been much more willing to use violence than has the continuity organization. The OIRA abandoned its military campaign in 1972, and in the late 1970s it evolved into the Workers party. After 1972 the term IRA is often used to describe the PIRA. 54


The organizations normative preferences are at most soft constraints on its actions there is ample room for instrument al rationalization after the fact (Kalyvas and Sanchez-Cuenca 2005, 215). Constituency costs are a more important constraining factor on terrorist organizations. Suppor ters can prevent actions that the organization would be otherwise willing to carry out. There are hard core members, material supporters, and a diffuse community of supporters who pr ovide tacit approval. Given that many organizations claim to act as representati ves, their need for support may extend beyond the hard core to include those of the wide r community they supposedly defend (219). The preference constraint refers to the organizations constant reliance on active and passive supporters. Maintaining the bare minimum constitu ency necessary for survival is an everpresent constraint on terror ist groups, who might otherwise prefer to exert maximum pressure on the state. Constituency costs stem from the prefer ences and moral sentiments of supporters. However, these preferences have a greater or lesser constraining effect depending on the nature of relations between supporters and th e organization. In some degree, a trade-off exists between the intensity of killing or the selectivity of viol ence on the one hand and popular support on the other (Kalyvas and Sanchez-Cu enca 2005, 219). For instance, supporters of the IRA and ETA are more moderate than the terrorist organizations themselves, and reject indiscriminate attacks that provoke civilian victims. Were ETA or the IRA to carry out systematic indiscriminate bombings, support among important sectors of the nationalist communities woul d wane and the pool of volunteers would probably shrink. (ibid). This dilemma is well-illustrated by the IR As brief use of forced suicide missions. On October 24 th of 1990, it kidnapped a man working in the canteen of an army base (to them, this made him a legitimate target), took his family hostage, and forced him to drive a car full of explosives into an army checkpoint (Kalyvas and Sanchez-Cuenca 2005, 55


211). The same tactic was used on two other me n that same day; one was warned that two of his sons would be shot if he did not co mply. Two of the bombs detonated, killing six soldiers, one of the drivers, and injuring a nother 35 (Drake 1998, 67). The vast majority of Catholics in Derry were disgusted by these attacks. As a result, forced suicide missions were quickly ruled out of the IRAs playbook (Kalyvas and Sanchez-Cuenca 2005, 220) Along similar lines, both the IRA and ETA have denied responsibility for some of their worst attacks (ibid). Kalyvas and Sanchez-Cuenca suggest that the relation between support and strategy may be U-shaped, so that organizations with very limited or very strong support are most able to engage in indiscriminate attacks and even suicide misisons (Kalyvas and Sanchez-Cuenca 2005, 222). Highly decentralized movements with small, isolated cells may engage in indiscriminate killing because they have no popular backing to lose. On the other hand, groups that rely on a radical soci al base may be rewarded for engaging in extreme acts. Between these poles are groups like the IRA and ETA, who are crucially reliant on popular support from a more mode rate base of support. The level of popular support among these organizations is generally weak, at least compared to organizations like Hamas and Hezbollah. 29 An important question in this framework is what determines the preferences of supporters in the first place. It is possibl e that preferen ces are largely determined by the behavior of opponents: harsh and indiscriminate reprisals by incumbents may produce support for more extreme violence such as suicide missions (Kalyvas and Sanchez-Cuenca 2005, 221). 29 In 2001 92% of Palestinians defended armed struggle against Israeli troops, and 58% supported attacks against civilians (Kalyvas and Sanchez-Cuenca 2005, 223). 56


Due to either miscalculation or operational incompetence, terrorists do not always provoke the reactions they desire. As an example of the former, the Red Brigades alienated the very industrial workers they sought to mobilize when they shot a popular communist union official for denouncing a fello w car factory worker as a Brigadist. The killing led to large demonstr ations against the Red Brigad es (Drake 1998b, 2). Terrorist may also attack the wrong target by mistake. For instance, the PIRA killed two Australian tourists in May 1990 after mistaking them for British soldiers (ibid). Terrorist groups are usually sensitive to the danger posed by public backlash. For instance, the IRA grew increasi ngly wary of indiscriminate attacks after Catholics turned out to march with Protestants at a funeral for Protestant children and a mother killed by one of their bombs (McCauley 2008, 4). Bu t not all groups have responded swiftly enough. The Armenian Secret Army for the Li beration of Armenia lost the sympathy of the Armenian diaspora when it killed non-Turks, failing to adapt its tactics to the preferences of constituents. In parallel, AS ALA [The Armenian Secret Army for the Liberation of Armenia] attacks declined from nearly fifty in 1981 to two in 1986 (ibid). Similar processes have led to the rapid downfalls of the Front de Libration du Qubec in Canada and Red Brigades in Italy, where the kidnapping and murder of respected political leaders caused both in ternal organizational splits and a rapid decline in public support. Conceptualizing social support for viol ence is inherently complex because preferences for violence are largely endogenous to the conflict. A te rrorist attack which initially causes revulsion may evoke very different feelings if the state reacts in ways that justify the groups ideolo gy. Potential reversals such as this help explain why terrorist 57


groups will often seek to provoke state repression in order to reveal the true nature of the state, as some revolutionary groups are f ond of saying. Ironically, right-wing terrorist groups may also seek to uncover the states true repressive nature, but with the very different strategic goal of freeing the state from the rule of law and popular participation in government, which they view as inimical to stability (Engene 2004, 28). Repression, Dissent, and Co llective Action Frames For at least the past 30 years, scholars of political conflict have grappled with the problem of why government repression sometim es discourages and sometimes motivates dissent. The literature was inspir ed by the idea that state repression is a key variable to account for political dissent, and perhaps ev en the most important variable (Lichbach 1987, 267). The mystery was and continues to be why such an obviously important variable produces contradictory effects. Theoretical predictions of the effect of repression on dissent have employed a huge variety of approaches, none of which ha s been entirely successful. Scholars have hypothesized positive relationships, where more repression mobilizes greater dissent, negative relationships, where more repres sion causes less dissent, and quadratic relationships, in which either initial repression leads to decreased dissent below a certain threshold, but beyond that point, catalyzes more dissent, or initial le vels of repression lead to greater dissent, but su fficiently high levels reduce it. Koopmans (2005) argues that the relations hip between dissent and repression is mediated by the mass media. Others, such as Gupta et al. (19 93), argue that the 58


relationship depends on regime type, either because different regimes engage in qualitatively or quantitatively different form s of repression, or because the role of the media and norms against state violence are different. The failure of a related literature, which sought to predict aggregate levels of government repression, has led to efforts to disaggregate middle regime types (beyond the standard authoritarian /democratic binary) by their institutional structures in order to account for their relative incentives to engage in repression (Prorok 2008). There have also been numerous atte mpts to disaggregate repression and dissent. The simplest way to disaggregate repression is to break it down into discriminate and indiscriminate violence. As previously not ed, this distinction can refer either to scale (absolute level or proportionality of force) or to scope (how carefully and on what basis targets are selected). Li chbach (1987) argues that the effect of repression on dissent depends on the relativ e efficacy of the tactic repressed, and on whether the government pursues a consiste nt accommodative or repressive strategy. 30 Most importantly, he argues that the repression of nonviolent activities may cause nonviolent activities to d ecline, but increase violent activit ies due to the fact that their relative costs decline when nonviolence is repressed (Lichbach 1987, 293). Most scholars of political mobilization have identifi ed sustained dissent as a function of three main factors: cultural frames, mobilizing structures, and political opportunity structures. These explanatory vari ables can also explai n state repression. The interaction between repression and dissent can occur in three ways; 1) directly through 30 Repression of a relatively more effective tactic is predicted to cause an increase in a dissident groups total conflict activities, becau se the rationally maximizing group will at least want to remain as effective as it was before. Similarly, if the government rewards and punishes the same tactic, the group will have a reduced incentive to stop using it, which should lead to an increase in total conflict activities. Thus, if states wish to minimize conflict, they should never punish and reward the same tactic, but instead be consistent and punish one while rewarding another (Lichbach 1987, 287). 59


the outputs; for instance, in confrontations between the police a nd protestors, 2) less directly through effects on mobilizing struct ures and cultural frames, or 3) through impacts on the other actors political opportuni ty structures (Dave nport et al 2005, xvi). Many researchers studying state repressi on refer to these concepts, although sometimes with different labels. For inst ance, some scholars refer to government ideology rather than cultural frames as they locate motivations for group identity and action. Others speak of the ethos of coerci ve institutions. Mobilizing structures are generally discussed in the cont ext of very specific organizations: political institutions, military organizations, police squads, and secret police or intelligence organizations. Political opportunity st ructure is also discussed in the repression literature, but under the label political threat, whic h refers to the perceived nece ssity of state repression to counter challengers who might alter the politic al or economic system. State opportunity is thus reframed as necessity (Davenport et al. 2005, xiii). It has also been argued that it is f undamentally mistaken to seek out a decontextualized relationship between such br oad classes of events divorced from the plethora of intervening variable s that we know to exist at le ast at the levels of episodes or classes of episodes (Davenport et al 2005, 211). According to this perspective, repression and dissent are highly aggregat ed classes of human actions whose effects are contingent on too many local factors to be precisely understood at the general level. A less cynical (but more circul ar) approach is advanced by Brockett (1993), who argues that the effect of regime violence on dissident mob ilization is best explained by examining its temporal location in a protest cycle. 31 Indiscriminate repressi on is likely to provoke 31 Protest cycles occur when the structure of political opportunity turns more favorable, encouraging groups to act on long-standing or newly created grievances. Early mobilizers encourage other groups and 60


further popular mobilization only during the ascendant phase of the protest cycle. In contrast, indiscriminate repression deters popu lar collective action befo re the initiation of a cycle, and it can (and does) bring protest cy cles to an abrupt end (Brockett 1993, 4712). The problem here is that we can never know the full shape of a protest cycle until after it has ended. Protest cycles merely describe the ways in which mobilized contention rises and falls over time. It is therefore taut ological to use them as an explanation for changing patterns of mobilization. This dilemma also exhibits features of the structure-action problem. Certain types of repression create moral outrage and make individuals more desirous of action against the state. But if repression is effective in destroying the organizations required to coordinate mobilization, there can be no coll ective outlet for grievances, and mobilization does not occur. Of course, lower levels of indiscriminate violence committed by the state may mitigate the collective problem of mobilizin g violence if this repression shrinks the cost differential between action and inaction. Suff iciently high levels of violence such as what we see in civil wars may eliminate the collective action problem entirely. In these situations, we must answer why more individuals do not participat e in violence, since the additional risk assumed is so minimal (Kalyvas and Kocher 2007). In the empirical literature it is frequen tly noted that the violent repression of nonviolent protest causes a tactical shift toward violence. This c ould be explained not only by Lichbachs (rational c hoice inspired) mechanism of re lative costs, but also by emotional reactions (Sambanis and Zinn 2005, 5) and belief amplification (a type of movements to activate, causing conflict to diffuse at a higher than normal frequency and intensity. In a detailed study of the Italian protest cycle of the late 60 s early 70s, Tarrow records an average of under 150 protest events for each half-year betw een and but an average of over 500 for the twelve half-year points starting in late 1968 (peaking at about 750 for the last half of 1971) (Brockett 1993, 471). 61


frame amplification see Snow, Roch ford, Worden, and Benford 1986, 469). The repression of nonviolent protest has often led to the creation of terrorist groups, and in instances where terrorist groups already exis ted, the repression of nonviolent protest has led to tremendous increases in terrorist violence. The precipitating event leading to terroris m in West Germany was the shooting of student Benno Ohnesborg by the Berlin polic e in June 1967. Ohnesborg was attending a protest against the visiting Shah of Iran, a vis it which happened to coin cide with the peak of a long-lasting student mobilization (D ella Porta 1995, xiv). In Italy, a police intervention to clear Romes university bui ldings, occupied by students, led several thousand students to converge on Piazza di Spagna on March 1, 1968 to protest the intervention. Fights broke out and escalate d, eventually resulting in 211 injured (158 among the police), 228 arrested, and 4 impris oned (Della Porta 1995, xv). This conflict triggered a bitter debate that polarized the political system, and set the stage for skirmishes between left and right terrorist groups for years to come. In Northern Ireland, thirteen civilian demonstrators were kill ed and twenty-nine injured on Bloody Sunday January 30 th 1972. This date marked a turning poin t in terrorist violence, with the number of terrorist bombings increasing from around 150 in 1970 to 1,382 in 1972 (Parker 2007, 160). Despite clear evidence that the repression of nonviolent pr otest leads to outbreaks and escalations of violence (at least, when it does not successfully destroy targeted organizations), there do not seem to be any st udies that focus on whether such repressive experiences affect the way terrorist groups socia lly construct their enem ies. It is likely the case that targeting one groups civilians lead s to retaliation in ki nd. As Hoffman noted 62


with regard to terrorist groups, The point is less their inherent differences than the fact that their tactical and targeting choices correspond to, and are determined by, their respective ideologies and attendant mechanisms of legitimization and justification; and perhaps most critically, by their relationship with the intended audience of their violent acts (1998, 158). It seems relatively obvious that retaliation will occur the only questions are whether target se lection patterns move towards or away from civilians, and how long the change lasts. According to Drake, terrorists seek to achieve their political objectives by coordinating the group's resources, pattern of attacks, and any other actions possibly including overt political activities into an effective strategy. The strategy adopted has a fundamental effect upon the selection of targets in that given a choice of targets terrorists acting rationally will choose to atta ck those which confer the greatest benefits upon their cause (1998, 2). Changes in the level of state repression may create the opportunity for groups to attack civilians, which they may perceive as having greater coercive potential. However, targeting civilians may be rejected for reason s unrelated to the preferences of internal target audience. If it is rejected, the s cale of violence would increase without any qualitative changes in its deployment, or in the targets selected. State repression should mitigate the burden of the preference constrai nt, as well as the clandestine requirement and security constraints. Anger at the state makes preferences more radical, more recruits make arrests less dangerous to group survival, and increased contact with the larger social movement gives groups better informati on about their operating environment. 63


Research Design We do not know very much about why civi lians are targeted by terrorist groups. Is it a rational strategy, an emotional response, an expression of pathology, or something in between? One way we can look for an explan ation for the targeting of civilians is to study changes in target selection over time in response to theoretically important variables. This study will focus on two variables: government repression and sectarian violence. The general argument is that te rrorists target civilians because they are relatively easier to attack a nd because doing so may yield grea ter tactical benefits, either by sending a costly signal to the state, or by setting off an ethnically oriented security dilemma. The production of violence against civilians, however, is constrained by the preferences of supporters. These constraint s should be relaxed when government or civilian oppositional violence hurts the ci vilian population, leading to calls for retribution. Government repression is often stressed as a key variable affecting social movements (Della Porta 1995, 56). It s theoretical importance is so large, in f act, that the relationship between repression and disside nt activity has developed an almost autonomous sub-literature within the larger study on the causes of political violence. Most of this literature uses Li chbachs (1987) theoretical stud y as its point of departure. Lichbach catalogued many of the early approa ches to repression and dissent and argued that each one was insufficient for explaini ng repressions mixed effect on dissent. He advocated a rational choice model of tactical choice to explain the variation. The most important ramifications of this model, for our purposes, are that 1) repressing one activity lowers the relative cost of its alternative, and 2) when the cost of a relatively more 64


effective tactic increases, maintaining the same level of output requires an increase in total conflict activities (since, by definition, a less effective tac tic must be used more to achieve the same result). Since Lichbachs review a great number of approaches have sought to explain the contradict ory affects of repression on disse nt. This literature has yet to resolve the big questions but it has produced a number of useful insights. Given that organized agitational violen ce seems to often result from state violence, and that preferences for violence ar e endogenous to the conflic t itself, it is very likely that terrorist targeting is altered as its members a nd supporters are exposed to varying levels of repression. This prem ise does not mean that terrorism is caused by state repression the growth of violence is genera lly more of a reciprocal dynamic than a unilateral escalation. The point is rather that the choice to escalate can be justified by greater levels of state repr ession, and in some cases stat e repression also provides the most important means for escalation. Engene (2004) conceives of terrorism as a means to alter the allegiances of various groups in society by challenging the states legitimacy. Stat es that are provoked to act outside of the rules er ode their own legitimacy and en hance the prestige of groups challenging the state (Engene 2004, 27). Kill ing nonviolent protestors in nominal democracies could accelerate this process through a number of mechanisms, including lowered relative costs of violence, greater opportunities to build up an injustice frame, and emotional responses. Understanding terrorism requires more than just an explanation for the level of violence. We must also understand why some ta rgets are selected rather than others. For some, this is what distinguishes terrorism from other forms of political violence, and it is 65


therefore required in a defi nition (Goodwin 2004). More specifically, we would like to know why terrorists attack civilia ns, and to what extent this tactic reflects strategic considerations. Goodwin argues that terrorists do not strike at citizens randomly, but generally attack only those categories which ar e seen as complicitous with the regime those who support, benefit from, or have a substantial capacity to influence the government (Goodwin 2006, 2037). As we saw in Chapter 3, Goodwin enumerates several complicitous civilian categories to show that they depend on the groups perception of the political regime. At one ex treme, complicitous civilians are limited to the cronies of the ruling autocrat. In democr atic contexts the entire body of citizens can be considered responsible for the gover nments actions. And in an ethnocracy, complicitous civilians may include the en tire dominant ethnic group (Goodwin 2006, 2037). These examples help to illustrate how an acceptable class of targets may be constructed in different politic al contexts, but they do not te ll us what terrorist groups will actually do. Targeting stra tegies are selected for both ideological and pragmatic reasons, and the mix varies considerably. As we have seen, the South African ANC rejected indiscriminate terrorism against whit es because it sought to portray itself as a principled challenger to the apartheid governme nt. Its president even signed a protocol of the Geneva Convention which legally bound the organization to avoid attacks on civilians (Goodwin 2006, 2034). On the other hand, the Liberation Tigers of Tamil Eelam which was once a very powerful revolutionary group, has occasionally engaged in indiscriminate attacks on ordinary Sinhale se civilians, despite having long since decimated rival Tamil nationa list groups (Goodwin 2006, 2033). 66


Goodwin proposes three contextual f actors affecting the decision to use categorical (i.e. indiscriminate) terrorism: complicitous civilian support of extensive state violence or state terrorism, the number of complicitous civilians and the ease of attacking them, and the level of political alliances or coope ration between revolutionaries and complicitous civilians (e.g., level of linguist ic, religious, or terr itorial segregation) (Goodwin 2006, 2040). These are plausible variables if we want to account for the different manifestations of violence across various countries They change too slowly, however, to explain short-term variance with in a single terrorist group or cluster of groups. The only possible exception is the support for state violence variable, but this is difficult to measure states may engage in greater repression without reliable information on the responses of various groups. It is clear that the experien ce of real grievances is rele vant to support for political violence; the repression of legitimate dissent is a recurring theme in the outbreak of political violence. The central question is what mix of vari ables determines a terrorist groups targeting strategy during different periods of time. Gi ven that the preferences of supporters is a key constraint on the type a nd level of violence produced, the level of violence they sustain is a valid point of departure. This study will use case studies of Ireland and Spain to examine the relationship between terrorist violence a nd violence against the communiti es from which they draw support. The two cases are highly comparable: both countries experienced secessionist, ethnically oriented terrorism, both of the main terrorist groups have had fairly high public support for at least certain periods, both groups have ties with politic al parties, and both operate in highly developed, consolidated demo cracies with strong st ate capacities (Spain 67


experienced the worst outbreak of terrorism late in its transition). The degree of public support combined with the goals of each group make them good candidates for approximating the rational choice perspective on terrorist groups. Ther e is also detailed information on attacks in each country, listing fata lities that include the time of the attack, the identity of the vi ctim, and the strategy behind the ki lling. Both countries also have sporadic survey evidence indicating levels of public support for the main terrorist groups, the IRA and ETA. Terrorist groups in Northern Ireland and the Basque region have also had the good fortune of sharing a bord er with less vigilant governments. 32 Because states cannot simply enter one anothers territory when ever they please, the ability of terrorist organizations to move personnel and materials across borders allows them to mobilize greater resources and avoid arrest. Parker argues that close cooperation between the Dublin and London governments was key to redu cing IRA violence in earlier years, and that the later failure to coordinate was precise ly what led to the vi olent backlash against British internment. 33 French cooperation with Spain against the ETA beginning in the mid 1980s has been at least equally important. To test whether intense state and ethnic violence lead to more indiscriminate forms of terrorism, we will examine all known attacks of terrorist groups operating in the same territory and with the same stated pur pose for the three-month periods before and after major civilian casualties are inflic ted on their supporting community by the state 32 The Basque region is split by Spains border with Fr ance, while Northern Ireland sits in a small corner northeast of Ireland. Northern Ireland constitutes about one-seventh of the islands total area. The rest of Ireland won independence in the Anglo-Irish war with the passage of the Government of Ireland Act in 1920 (White 1997, 25). 33 According to Parker, during the border campaign intern ment was adopted in both the Republic of Ireland and in Northern Ireland, thus depriving the IRA of safe haven. Later, when the Republic of Ireland did not cooperate, internment hardened identities and made the British look like hostile and indiscriminate outsiders. At the same time, Dublins more moderate policies created an easier operating environment for the IRA and helped outrage translate in to violence. This is actually a rather dramatic oversimplification, but an elaboration will have to wait until later. 68


security apparatus and rival terrorist organizations. Focusi ng on a single organization is theoretically unjustifiable because many groups operate in concert and divide tasks among themselves. For instance, the Anti-capitalist Autonomous Commandos ( Comandos Autonoous Anticapitalistas, or CAA) in Spain specialized in internal security so that the military faction of Basque Ho meland and Liberty (Euskadi Ta Askatasuna militar, or ETAm) could focus on government targets and maintain a stronger public image (de la Calle and Sanchez-Cuenca 2006, 18). Similarly, the Irish National Liberation Army (INLA) and Irish Peoples Liberation Organization (IPLO) both have relatively low percentages of attrition killings compared to the Provisional IRA (ibid). Given the PIRAs overwhelming dominance, they have chosen to specialize in the dirty work of sectarian and securi ty killings (ibid). Attacks carried out by ETA-m, ETA-pm, and the CAA are combined together. In Nort hern Ireland, the Official IRA, Provisional IRA, INLA, and IPLO are combined together Attacks from non-specific Republican and Basque nationalist groups are also included, as they presumably come from one of the above organizations. We will test three hypotheses. All rely on the idea that experiencing violence, either directly or indirectly, leads individua ls to develop more radical preferences and thus support of more extreme forms of po litical action, including varying types and degrees of violence. 34 The independent variable is th e measure of violence experienced by the community; the dependent variables are the measures of violence inflicted by the terrorist organizations. If the physical repression of civilians creates more radical preferences, what exactly does that mean for the decision ma king of a terrorist group? More radical 34 Indirect experience with violence refers to knowing others who are direct victims. 69


preferences give the leadership more room to maneuver so that they can execute a plan that yields the highest expect ed return. More radical prefer ences on their own do not lead to a new strategy. If an altered support base looks more favorably on some course of action, the decision to take that course will still be made by leaders who are facing their own incentives. The potential decisions availa ble depend on the precise way in which the preferences of supporters change. We consider four distinct possi bilities within three categories: 1) Supporters may become more willing to expand the list of legitimate targets to include certain categories of civilians. This possibility has two variants. a. The first possibility is that supporters become more willing to accept all civilians from a rival ethnic group as legitimate targets. For such support to be significant, there must be a high degree of mu tually escalating violence between the groups. We would not genera lly expect for repression from the security forces to result in violence aimed solely at a rival ethnic group. However, from a simple cost-benefit perspective such a counter-intuitive reaction may be desirable to a terrorist gr oup if it serves to excite the internal base, thus offsetting costs among external audiences. The internal base would probably not desire attacks against a ri val group without a history of violence between them, or unless the group dominated the government to such an extent that its repression was seen as e quivalent to theirs. Even in the presence of either, however, the road to such indiscriminate targeting requires a significant degree of radicalization an d entails a high amount of risk. The strategy forecloses the possibility of future cooperation between the groups. 70


And as a coercive policy against the other community, targeting the entire group makes little sense. 35 This tactic is therefor e an unlikely possibility. If supporters in the social community feel that some means are off the table, the instrumental use of ethni c violence to divide soci ety probably ranks highly among them. b. The second possibility is that supporters become more willing to accept some categories of civilians as legiti mate targets based on their individual behavior. In this scenario targets can a lter their behavior and thus avoid punishment. However, the warrant linking their actions to legitimate violence against them may be weak, creating a risk of constituency costs. Assuming that other targets are readily available, terrorist groups may prefer to conserve the support generated by state repression ra ther than spread it thin with an unnecessary expansion of targets. This tactic is therefore an intermediate possibility. 2) Supporters may become more willing to accept collateral damage. As supporters experience more repression they can grow more angry and more determined to remove the cause of their grievances, becoming more callous towards innocent bystanders, and perhaps even blaming vict ims on the government for allowing the conflict to continue. Because collateral dama ge is a moral grey area for supporters of violence, this is an intermediate possibility. 35 If any citizen can be attacked there is no effective sanction on any particular behavior. Terrorist groups would have to miscalculate to the point of wishful thinking if they expect their killing of civilians to cause the entire group to submit rather than mobilize. As Mc Cauley observes, the social psychological effects of such out-group threats are well-established, and they point in the latter direction (2008, 5). 71


3) Supporters may become more willing to support reciprocal retaliation. This is the most likely possibility. In this scenari o, attacks from the security forces would lead to increased support fo r retaliation against them (and potentially another ethnic group if the state and its secu rity forces are sectarian), and attacks from rival groups would lead to attacks against civilians within that comm unity. This reaction can be defended morally as at least no worse than the techniques being employed by the opposition, and can be justified on the basis of deterrence or coercion. To the extent that such retaliation focuses on civilians, re ciprocal retaliation may seem to blend with the expansion of legitimate targets. 36 The difference is that retaliation is confined to a short period of time. Th e period is limited because the strategy is abandoned when it fails as a deterrent the patience of supporters only lasts for so long. There is no direct way to observe such changes in the attitudes of supporters unless they are manifested in action. Even when the action occurs it will not be clear from the outset what it reflects about the organizations internal dynamics. However, given a definite outcome and a limited number of explanations, we can judge some scenarios to be far more likely than others. The four possible changes in supporter preferences under the specific conditions outlined above can be formalized into testable hypotheses: H1 : There is a significant difference in the num ber of indiscriminate killings undertaken by a nationalist terrorist group in the three months before an d after either the security forces or rival terrorist groups kill a large number of civ ilians within their community. (Three months is chosen for simplicity if th e new targets are considered to be legitimate 36 The differential impact of Protestant paramilitaries and British security forces on the production of sectarian violence by Republican paramilitaries will al so be addressed. It is, however, secondary to the main hypotheses. 72


they should be the object of violence for as long as the violence persists, or until the strategy of violence changes). H2: High levels of fatalities among the nationali st terrorist groups community of support will immediately precede the expansion of legitimate targets to new categories of civilians based on their behavior. These ne w categories should appear within three months of the high levels of fatalities plenty of time for the organization to reformulate its strategy and take advantage of the new political situation. H3: There is a significant difference in the number of collateral killings committed by a nationalist terrorist group in th e three months before and afte r a large number within their civilian community die as a result of state or sectarian violence. H4: There is a significant difference in the number of indiscriminate killings undertaken by a nationalist terrorist group within the same month that either the security forces or rival terrorist groups kill a large number of civilians w ithin their community. Such indiscriminate killing will be limited to one month because the motivation is retaliation and not escalation supporters will not tolerate the expansion of legitimate targets to include all members of the rival group. If the violence used by a terrorist group violates the norms of supporters, either because of either the identity of victims or the sheer magnitude of violence inflicted, the group can be expected to lose sympathy and support. If we assume that state brutality and ethnic violence (both measured by fatalities ) are accurate proxies of a communitys support for violence against the state or th e other ethnic group, we can predict that changes in target selection cause d by strategic considerations unrelated to such support (for example, a failure to meet specified de mands, or perceived opportunities such as war entrance or upcoming elections), will erode popular support. This re lationship is very difficult to measure directly, but we can a pproximate it with the electoral results of political parties linked to te rrorist groups Sinn Fein for the IRA, and Batasuna (formerly Euskal Herritarok and even mo re formerly Herri Batasuna) for ETA. If electoral victories for these groups do not su ffer even as terrorist groups commit violence that cannot be legitimated by temporally proxima te state or ethnic violence, then either 1) 73


grievances from the past remain effective long into the future, 2) votes for these parties are not complementary to support for the terr orist groups with whom they are commonly associated, or 3) grievances are not accurate ly captured by the number of deaths inflicted on the community. In the more likely event that electoral support does decrease, this drop may not necessarily indicate decreased social support for vi olence. If support for these parties and support for violence are not complementary, decreased support could instead represent waning support for the partys goals am ong voters who would not have supported terrorist violence under any circumstances. Th e literature has tended to assume that support for a terrorist group and its associ ated political party is more or less complementary. In fact, the available ev idence supports this assumption (SanchezCuenca 2008, McAllister 2004). For instance, Sinn Fein and Batasuna have both made relatively blatant endorsements of armed str uggle, and have also employed abstentionist policies that send clear anti-system message s both to the state and their supporters. Classification of Attacks and Coding Rules Northern Ireland In order to test the hypotheses, many attacks had to be coded from raw data. Attacks were coded using Suttons Index of Deaths in Northern Ireland (1994). This database lists the status of victims and perpetrators but not the selectivity or stra tegy of violence. The selectivity of attacks was categorized accord ing to the following scheme, borrowed from de la Calle and Sanchez-Cuenca (2006): 74


Selective killing is based on the behavior of the vi ctim (e.g., the victim cooperated with ecurity forces, or was involved in a Loyalist paramilitary organization); s Generic (state) killing is based on the occupation of the victim (e.g., the victim was in he police or in the army); t Collateral killings are those which are merely incide ntal to an attacks purpose. Common examples include civilians shot by stray bullets during a gun fight, or hurt by ombs that target the military. b Indiscriminate killing is that which targets an entir e class of people based on some ascriptive trait re lating to ideology, religi on, nationality, or language. In the context of ethnic conflict, indiscriminate attacks are those which targ et members of an ethnic group irrespective of their individua l behavior, and are equivalent to sectarian killings. Following de la Calle and Sanchez-Cuenca, we will consider attacks as sectarian only when they are both 1) purposive, or non-acciden tal, and 2) not rationalized on the basis of the victims occupation (2006, 29). Unlike these authors, we will not treat the killing of rival paramilitary members as sectarian, as these attacks are obviously related to the ictims behavior. v Mistakes occur when terrorists misjudge their victims dentity. This typically happens when a person has the same car or apartment previously owned by a member of the security forces. Coding rules Classifying attacks into categories required that a number of technical choices be written down and consistently applied. The following is the list of guidelines that were followed during the coding procedure. Each rule is applied to cases of Pr otestant civilian casualties inflicted by Republican (Catholic) groups. 1 Coding sectarian attacks If civilians of a different reli gion are killed in an attack wh ich did not target members of the security forces, the attack is consid ered sectarian. A common example is a bomb placed in a civilian establishment. Other comm on examples of sectarian attacks are those n which: i -The victims death was isolated and t hus could not have been related to the commissioning of an attack on security forces. -An inadequate warning was given, and the re sulting civilian deaths were primarily Protestant. 2. Distinguishing sectarian from individual 75


The rule will be to put more weight towa rd a sectarian coding unless information is presented which demonstrates that the killing was based on the indivi duals occupation or personal behavior. 3. Distinguishing sectarian from collateral The rule will be to code attacks as sectarian by default, and code att acks as collateral only when the deaths were clearly unintentional or when members of the security forces are learly the real target. In practice this mean s that attacks are coded as collateral when: c -an inadequate warning 37 is given and the resultant fallout kills many Catholic civilians and only one or two Protestants (presumably the organization would not have chosen to ill a majority of Catholics civilians with an indiscriminate attack), or when k -civilians are killed when members of the security forces are targeted in a military setting checkpoint, army base, po lice station, etc), or when ( -the majority of an attacks victims are members of the security forces (Protestant civilians the minority). This rule applies to attacks that occur in civilian settings. In short, within a military setting, targeting the armed forces is sufficient to label any others victims as collateral, even if no secu rity force members are killed. On the other hand, if an attack carried out in a civilian setting results in a majority of civilian deaths, these deaths will be treated as sectarian instead of collateral. This last rule only applies in the presence of an inadequate warning (i.e. the majority of the time), because such a warning may be responsible for the appearance of security forces on th e list of casualties. In a civilian setting, the security forces must constitute a majority of victims in order for civilians to count as collateral rather than sectarian. 4. Classifying premature explosions Some attacks apparently kill more than inte nded due to either premature explosions or local incompetence. This suggest s that apparently sectarian deaths should instead be considered collateral. The problem with accepting this claim at face value is that there is no way of knowing how many people were intended to die. Furthermore if targets are selected on a sectarian basis, organizational incompetence is no reason to erase that fact. When members of the terrorist organization are killed in a prem ature explosion, the 37 The IRA commonly calls the police before an attack to give the police time to defuse the bomb or clear civilians from the area. These warn ings are often inadequate, which would seem to suggest that the resultant deaths should be treated as unintentional. However, the inadequacy of these warnings may be intentional, and the resulting fall out tends to mainly kill civilians. 76


civilian victims are considered collateral. However, when no members of the organization are killed in the explosio n, the normal coding rules are followed. D ifficult cases encountered -Bombings that killed several Protestant a nd Catholic civilians appear to be fully indiscriminate. It is probable that the Catholic deaths in these attacks were accidental, but in making that assumption we cannot rule out th e possibility that all civilian deaths were nintended. Due to this ambiguity these attacks are placed in an intermediate category. u -Sniper attacks on the security forces which kill Protestant civili ans. On the one hand, such attacks do not choose civilians as their pr imary targets. However, sniper rifles might seem sufficiently accurate that we can rule out accidents. The most important consideration here is the ex tremely poor or nonexistent trai ning given to IRA recruits before they are put into battle This suggests that the attacks represent colla teral killings. Along these lines it is worth nothing that the fi rst lethal sniper attack against civilians killed both a Protestant and a Catholic. Such fully indiscriminate attacks are almost certainly accidental in some regard, making at least some of their fatalities collateral. Again, the default will be to treat these att acks as sectarian unless the number of Catholic civilian deaths reaches parity with Protesta nt civilian deaths. In especially unclear circumstances where no defensible rule can be established, the attack is simply coded as intermediate -passersby who witness an attack in progress (these wi ll be classified as indiscriminate or ntermediate in special circumstances) i -individuals who are killed trying to protect th eir establishment from a bomb (these will be considered sectarian since the attack its elf was made for sectarian reasons. A death which results from a sectarian attack is not qui te as incidental as those which result from ttacks on legitimate targets). a -security men who are killed, sometimes duri ng robberies (these will be considered collateral when part of a robbery, but sectar ian when no evidence of a robbery exists). Spain Data on fatalities in Spain are taken fr om de la Calle and Sanchez-Cuencas Victims of ETA Dataset This dataset was constructed us ing several pre-existing lists, including one developed by the Basque government, and through comprehensive searches through two promin ent Spanish newspapers El Pais and ABC These newspaper searches eliminated some false pos itives in the other datasets and brought new attacks to light. 77


De la Calle and Sanchez-Cuencas classi fication scheme for the selectivity of attacks is mostly the same as what wa s outlined above for Northern Ireland. The difference is that the ETA dataset include s a category for social group killing. The Victims of ETA codebook fails to define social group killing, but describes it as having a sectarian strategy. Given th at ethnic conflict has been small or nonexistent in the Spanish case, it would seem that the author s used sectarian to refer to intra-group killing. However, in the code book just released for the authors Domestic Terrorist Victims Dataset (which encompasses their ETA data) they define so cial group killing as those based on specific social traits such as religion, id eology, or ethnicity (DTV 2009, 7). Thus, the authors clearly mixed strictly ethnic terrorism with intra-group policing based on ideology. Coding Rules In order to distinguish socia l group killing from indiscriminate killing, the authors lean towards the indiscriminate by default unless they find a clear indication that the attack was aimed at a particul ar ideological or ethnic group. Individual killings were coded when there was ev idence that the target was chosen for some particular behavior (such as judicial ev idence, statements by the victims relatives, r accusations made against the victim by terrorists). o State killings were coded when individuals w ho worked for the state were killed, but Social group killings were coded if those who work ed for the state were consistently ttacked based on their so cial traits (ibid). a Collateral killings were coded when they f ound good information indicating that the ictim was not the target of the attack. v Mistakes were coded when the victim was the target, but was selected improperly. Although the Victims of ETA Dataset is based on more detailed information than the data on Northern Ireland, th e classification scheme is almo st exactly the same, and the 78


coding rules employed are highly comparable. Using these classifications we will test the hypotheses outlined above by looking for changes in collateral killings, individuals who are killed based on an expanding notion of proscribed behavior, and changes in indiscriminate killing. (Indiscriminate killing in Spain is for the most part not ethnically motivated. The opposite is true in Northern Ireland). The quantitative analyses for each case are preceded by qualitative case studies. These case studies provide a basic backgr ound of each conflict, the organizational histories of the hegemonic terrorist groups in each case (with a focus on the numerous splits each have gone through), and the evolut ion of their strategies. Data cannot be directly compared between the two cases because each case has unique factors that limited the capacity to use violence and affected the desirability of using it in different ways. After testing the hypotheses for each case we will compare the results. These results will allow us to draw conclusions about the strategy of target selection, how it is constrained by supporters, and how changes in target selection affect both supporters and the terrorist group. 79


Chapter 4 The Other Civil Rights Movement: The IRA and Northern Ireland Between the late 1960s and the mid-1990s, more than 3,700 people were killed by political violence in Northern Ireland a period known euphemistically as The Troubles. This figure represents an averag e of just over 2 deaths a week, with a proportional equivalent in the United St ates of over 600,000 deaths, and 150,000 in Britain (Moloney 2002, xiii). The American equiva lent is nine times more war dead than the U.S. experienced in Vietnam, and even more than the number of Americans that died during World War II (Hayes and McAllister 2000, 3). Nearly 1 in 50 Northern Irish, or 30,000, have been injured in the violence. The comparable figure for the US would be 5 million, and just over 1 million for Britain (Moloney 2002, xiv). Very few people in Northern Ireland have not personally known at least one person who was killed in The Troubles. For some, this is a compelling reason to consider the conflict a civil war (ibid). 80


The Troubles were the latest and most protracted phase in the Anglo-Irish conflict, which began 400 years earlier with the Tudor wars and plantations of the sixteenth century. Ireland was Br itains first colony, and is sti ll one of its last. There has always been some form of resistance to the English presence ever since the English began establishing colonial settlements when th e Normans invaded in the twelfth century (Moloney 2002, xiv). The contemporary Irish Republican movement has organizational roots dating back to the 1840s. Its ideology was inspired by the French revolution and a failed rebellion by the United Irishmen in the 1790s. Ireland during this time was dominated by landlords and English economic interests. Landlords belonged to the Established (Anglican) Church and discrimi nated against Catholics and Presbyterian Dissenters. The United Irishmen were founded by Presbyterian merchants and manufacturers in Belfast and D ublin. They sought alliances wi th Irish Catholics who also wanted political and social relief (White 1997, 24). The United Kingdom of Great Britai n and Ireland was formed in 1801. Throughout the 1800s (1803, 1848, and 1867) Irish re bels tried to remove Ireland from the UK by force. Despite their repeated failure, many in Ireland remained committed to the idea of an independent Irish Republic. Anti-Catholicism was built into the state ideology and was actively promoted by its leaders. Unionist government ministers would urge their supporters to hire only Protestant s, and one Prime Minister famously described the parliament at Stormont as a Protestant Parliament for a Protes tant state (Moloney 2002, 42). This message was occasionally rein forced by sectarian violence. Riots, burnings, shootings and bombings, mostly from Protestant mobs, we re a regular feature of political life since the Irish nationalists began to agitate for Home Rule in the mid81


nineteenth century. When a liberal government suggested granting Irish Home Rule in 1912, Protestants rebelled against the British. Their leaders organized a private army called the Ulster Volunteer Force, smuggled t housands of rifles into the country from Germany, and threatened to resist Home Rule by force (ibid). Protestants banded together in a semi-secr et society called the Orange Order to protect their economic domination. From the mid-nineteenth century onward, industrialization changed Northern Ireland a nd the entire island. The desire to retain access to British markets bolstered Protestant resistance to Irish independence, while the flood of rural Catholics into Belfast attract ed by work opportunities created competition with Protestants and led to sectarian te nsion (Moloney 2002,43). Orangeism became an instrument of sectarian divi sion and privilege. By the end of the 1940s, no Unionist political official could win el ected office unless he or she was a member. Huge parades of Orangemen were held annually on July 12 th with tens of thous ands of men marching through the streets wearing bowler hats and orange sashes. Although baffling to outsiders, these parades were designed to re mind Catholics of their subordinate place in the political, social and economic order (Moloney 2002, 44). The Anglo-Irish war of 1919-21 saw the outbreak of fresh communal violence, as did the 1921 Treaty settlement. The British government passed the Government of Ireland Act in 1920 to partition Ireland and br eak up the Anglo-Irish war. The treaty was initially rejected by Rep ublicans. But in 1921, negotiations between British representatives and Republicans le d to a treaty that reinforced the partition of Ireland into the six counties of Northern Ir eland and the twenty-six counti es of the Irish Free State. 82


The IRA was divided on whether to suppor t the treaty. Some supported the treaty and stayed loyal to Michael Collins, the IRAs former Director Intelligence and plenipotentiary to the treaty negotiations. The opposition rallied behind Eamon de Valera, who was President of the Republic during th e negotiations. The split was not caused by partition, however. Even those who opposed th e Treaty believed that the nationalist areas would be removed from the six partitioned counties, as the Brit ish implied during the Treaty negotiation. Republicans believed th at once these areas were removed, the truncated remnants would not be viable, and th e new state would collapse into their hands (Moloney 2002, 47). Republicans were split over the Treaty because it imposed a duty on members of the Free State government to swear allegian ce to the British crown. Collins argued that the oath did not matter, since Ireland had secured the freedom to achieve freedom (Moloney 2002, 47). The resulting Irish civ il war was a highly unequal battle the IRA was on the defensive almost from the beginning, and by 1923 the war was over. Valera resigned from Sinn Fein and set up a constitutional Republican party called Fianna Fail (Soldiers of Destiny), which later won a majo rity of votes in the Irish Free State and broke with Sinn Feins abstentionist policy. 38 In 1955 Sinn Fein itself ended its policy of abstention in the Re public of Ireland. The British reneged on their promise to remove nationalist areas, and a majority report recommended that all six partitioned counties be incorporated into the new Northern Ireland state. The fledgling govern ment in Dublin (the location of central government in the Irish Free State), now le d by de Valera, had little option but to 38 De Valera eventually agreed to sign the oath, but claimed that this was not the same as swearing to it. Abstentionism was the practice of participating in elections but not in government. It represented an act of defiance, and an active refusal to give legitimacy to the government. 83


acquiesce. In this manner, nearly half a million Catholics and nationalists, a third of the population, were forced into a state with whic h they did not identify and that was openly hostile to them (Moloney 2002, 48). Meanwhile, No rthern Irelands poli tical leaders were faced with the dilemma of constructing a political order opposed by up to a third of its citizens. They resorted to reliable methods. The early 1920s saw scores killed in riots, guns battles, and burnings, a nd in the early 1930s violence erupted again. Catholics made up a disproportionate number of fatalities (43). Those who remained in Sinn Fein and th e IRA still refused to recognize partition and the governments in Dublin and Belfast that it created. From 1939 to 1945 and from 1956 to 1962 the IRA engaged in unsuccessful military campaigns in England and Northern Ireland in an attempt to create a united Ireland. De Valera drafted a new constitution in 1937 that set a united Irela nd as a central goal. The IRA responded by ending its conflict with the Southern state, and began a pr ocess that would eventually lead to full recognition and participation in its institutions. From then on, the main goal of the IRA was to drive the British from the North (Moloney 2002, 48). The six counties of Ulster in Northern Ireland were given some measure of selfrule based on the British Westminster model. One commentator describes it as a pathological specimen of majoritarian democracy (Engene 2004, 114). The system provided for a virtual political, economic a nd social monopoly of power for the segment of society loyal to the Britis h crown (Protestants). From the introduction of provincial self-rule in 1920 to the intr oduction of direct rule from London in 1972, the Unionist parties won every election and consequently held on to executive power in the province (ibid). 84


By 1939, the IRA felt confident enough to declare war against Britain. It opened up contacts with Nazi leaders then at war with Britain, but little came of the relationship (Moloney 2002, 48). The Forties Campaign (as this phase of the conflict came to be called) was a very poorly organized endeavor A bombing in Conventry in the English Midlands near the start of the campaign we nt wrong, killing five civilians and wounding sixty. Around the same time the IRA in Dublin raided the Irish armys weapons reserves and promptly lost their loot to the police (49). Th e killings at Coventry mobilized the English police, who used harsh methods agai nst IRA suspects. The Dublin arms raid permitted de Valera to seek emergency powers and intern IRA leaders. De Valera was concerned that the attacks on Britain and ove rtures to Nazi Germ any would give the British an excuse to demand his support in the war effort. By 1945 the IRA was forced to accept a crushing defeat. The organization began to rebuild itself in 1947, and by 1948 it was large enough to hold a Convention (Moloney 2002, 49). It was at this time that the IRAs Army Council issued an order forbidding units from making any attacks on the Irish police, the Garda Siochana (GS), or any other military forc es of the Irish free stat e. It did so out of fear that the government in Dublin might retaliate again, undermining the campaign against Northern Ireland as it did during the Forties Campai gn. From this point onward, the South was to be the logistical base, and the North the war zone (Moloney 2002, 50). In 1949, the Irish Free State declared itself a Republic. The IRA launched another campaign in 1955. In the midst of agitation over the arrest of suspected IRA activists, two Repub licans, one of them an IRA man imprisoned for partaking in an unsuccessful arms raid, were elected from Northern Ireland to the 85


Westminster parliament. The IRA took this as a sign that the Northern nationalists might be receptive to an IRA campaign (Moloney 2002, 50). The 1956-62 campaign, which came to be called the Border Campaign, focu sed on the border counties of Fermanagh, Tyrone, and Armagh. Northern auth orities introduced internment within days of the first acts of violence, and in 1957 de Valera in troduced internment and later set up military tribunals that meted out very harsh sentence s. The public support th e IRA expected never materialized, but the campaign lingered on anot her five years before the IRA leadership reluctantly acknowledged defeat (ibid). Twelve people were killed and another thirtyeight wounded in the Border Campaign. Un ionists those who supported Northern Irelands union with the rest of the U.K., we re alarmed at this level of IRA activity. Compared to the Troubles, it was quite tame. The Troubles: Conflict Background In the 1960s, a civil rights campaign was launched in Northern Ireland. The effort, inspired by the African-American civil ri ghts movement, was designed to address economic and political discrimination dati ng back to the part itioning of Ireland. 39 The campaign was very successful at mobilizing Catholics, which led to resentment and counter-demonstrations by Protestant extremists. This led to widespread rioting in August of 1969, and the British Army was called in by the government of Northern Ireland to help restore order (for a comprehensiv e timeline of conflict history, the IRAs organizational history, and strategic a pproaches, see figure 4.1 on page 88). 39 When statistics were first collected in the 1970s and 1980s, Catholics were found to be at least three times more likely to be unemployed than Protestants, and disproportionately represented in the poorestpaid, least-skilled, and most insecure jobs (Moloney 2002, 43). 86


The IRA at this time was a relatively minor organization. It had suffered a humiliating defeat in the 1956-62 border camp aign. Its members had hoped in vain for popular demonstrations and protests when they were arrested. There were neither protests nor welcome-home celebrations. Thoroughly demoralized, many members in prison pledged voluntarily to never take up arms again (Moloney 2002, 53). By 1962 the IRA was very weak, socially isolated, and largely without direction. Europe in the late 1960s was experiencing a wave of liberalism that brought new power to the left. In Northern Ireland, m odernizing reform was on the agenda of the Stormont parliaments new Prime Minist er, Terrence ONeill, and the new Labour government in London was instituting major social reforms (Moloney 2002, 53). The British had voted out Winston Churchill and his Conservative Party by a wide margin, ushering in a number of drasti c social changes. College education was for the first time thrown open to working-class children on the ba sis of merit. As more Catholics obtained a college education, their economic, social, and political expectations soared. Unionisms inability and refusal to satisfy these expect ations caused discontent among the Catholic community to grow rapidly (2002, 45). Accord ing to Moloney, educational reform was the central cause of social instability lead ing up to the Troubles no other factor was more responsible (ibid). While Catholics grew bolder and bolder in demanding their political rights, Protestants reacted fiercely to their political leaders who tried to accommodate. Unionists saw Prime Minister Terence ONeills modern izing policies as a th reat to continued union with Britain, and a danger ous concession to Republicans that might end in a united Ireland. Some of these critics rallied around the Reverend Ian Paisley, a young Protestant 87




preacher who embodied an explosive comb ination of political ambition, religious fundamentalism, and a hatred for the Roman Catholic Church. Paisley represented a virulent strain within the Un ionists that stretched back hundreds of years, when Belfast street preaches regularly incited crowds to riot, burn, and kill their Catholic neighbors (Moloney 2002, 61). Convinced that the IRA was planning a new violent offensive in 1966, Paisleys oratory inspired the newly cr eated Ulster Volunteer Force (UVF) to launch a preemptive strike. The UVFs official goal was to kill IRA leaders, but their actions demonstrated that they were quite co ntent to attack any Catholic they could find. By the summer of 1966 they had killed 3 civi lians, including 2 Catholic teenagers totally unrelated to the IRA (ibid). In 1965, the Northern Nationa list Party entered the Stormont parliament for the first time since 1930 to become the offici al opposition, thus recognizing the state of Northern Ireland (Moloney 2002, 62). The Republican movement as a whole was increasingly turning towards institutionaliz ed politics. Catholics were demanding their rights and becoming more willi ng to use violence to defend themselves. They also began compiling evidence of discrimination th e Campaign for Social Justice conducted research that showed that Catholics were hugely underrepresented in places like the central and local government civil service. Soon the idea was floate d of setting up a civil rights body modeled on the NAACP (ibid). In 1967, the Northern Ireland Civil Rights Association drafted a list of proposed reforms. Among these were the scrapping of gerrymandered electoral wards that gave minor ity Unionists control over local councils, the removal of the Special Powers Act, 40 a points list for pub lic housing, and the 40 This act had been passed in 1922. It gave the unionist government unprecedented police powers, including the powers to arrest, detain without trial, and suppress political dissent Penalties were so severe 89


scrapping of a rule that restri cted the right to vote in local council elections to property owners. 41 The political system responded very poorly to the demands being placed on it. Popular mobilization to increase pressure on the government was met with violence from Protestant crowds. In October of 1964 a two-da y battle erupted in th e lower part of the Falls Road after the Royal Ulster Constabular y (or RUC the Northern Irish police force) invaded the area to remove the Irish tricol or flag from the offices of a Sinn Fein candidate. Fifty civilians and over twenty RUC members were injured. Catholics clashed with the police agai n after Reverend Paisley brought a mob through the nationalist Markets area on their way to a protes t against ecumenical Protestants. Most dramatic were the clashes on Oct ober 5, 1968, when a poorly attended civil rights march in Derry was met with harsh repression from RUC officers. Republicans chose Derry for the protest because they saw it as a symbol of the greatest injustices committed against Catholics (not unlike Martin Luther Kings march through Birmingham). Indeed, for Unionists the city ha d been a symbol of Protestant supremacy ever since its inhabitants withstood the siege in 1688, helping the Protestant King William of Orange defeat the Catholic us urper, King James (Moloney 2002, 64). Eightyeight demonstrators were injured, and thirty-s ix arrested. The demons trators were joined by two British Labour MPs, who were hoste d by West Belfasts Republican Labour MP Gerry Fitt, who was himself beat over the h ead by police (ibid). Th e police even used a that a South African Prime Minister during the apartheid era famously d eclared that he would trade all of his emergency laws for one clause of the Special Powers Act (Moloney 2002, 39). 41 This rule had given some mainly unionist businessmen up to six votes, while thousands of working-class Catholics were disenfranchised (Moloney 2002, 62). 90


water cannon to disperse protestors, creating a television sc ene highly reminiscent of the police brutality in Alabama just five years earlier (ibid). The whole episode was filmed and broad cast to the world, complete with an interview from the blood-splattered MP Gerry Fi tt. The pressure intensified when another march in Derry brought 15,000 onto the street s, and a severely outnumbered RUC was forced to let the protestors through. As the Northern Ireland Civil Rights Association held more marches, the British began to demand action from ONeills Unionist government (Moloney 2002, 64). At the end of October, ONeill announced a package of reforms and appeared on television to reque st an end to civil rights ag itation. The Northern Ireland Civil Rights Association met his request, but the radical student group Peoples Democracy was not satisfied with the refo rms or their pace, and announced a 75 mile march between Belfast and Derry on New Y ears Day in 1969 to voice their disapproval (ibid). As Unionist discontent grew, the members of ONeills cabinet began to resign one by one. In February he called an unexpect ed election to the Stormont parliament hoping to strengthen his position. If anythi ng, it strengthened his opponents (Moloney 2002, 65). In March, Loyalist extremists plante d bombs at an electricity substation and punctured a pipeline supplying most of the water in Belfast. Most people assumed the IRA was responsible. This perception fed re sistance to ONeill and his reform project, leading him to resign in April of 1968 (ibid). On August 12 th a riot broke out when the Apprentice Boys, an offshoot of the Orange Order, clashed with Republicans and the RUC on their annu al march. Ferocious rioting broke out between nati onalists and the RUC. The riot ers repulsed waves of baton91


wielding police-men with rocks and gasoline bo mbs until the RUC turned to CS tear gas that flooded the streets, choking all of the areas residents. The much hated B Specials (an armed paramilitary of pro-British Protesta nts) were mobilized. In response, the Irish Prime Minister warned that his government would not stand by as nationalists were attacked (Moloney 2002, 66). Two days later, an exhausted RUC was forced to call in military assistance from Britain (ibid; Engene 2004, 115). The British did not leave for another 30 years. Given the nature of armies and the motives of the I.R.A., it was predictable that the British Army would be drawn into conflict with the ci vilian population (Coogan 1994, 260). It was not difficult to draw them into political blunders, taking advantage of what Rapoport calls the politics of atroc ity. Soldiers, police wardens, and British personnel in general seemed indoctrinated against the Irish. When internment was introduced, it was seen by the British as a solely anti-Republican and anti-Catholic measure. Despite a litany of crimes committed by Loyalists, none were seized until two years after internment began (Coogan 1994, 260-1; Melaugh 2009, 2). By this time, Loyalist terrorist groups had ki lled at least 138 people, 126 of whom were civilians (102 of whom were Catholic). From the start of internment in August 1971 until its end in December 1975, only 107 Loyalists were de tained, compared to 1,874 Republicans (Melaugh 2009, 2). Poor and out of date intel ligence meant that most internees had little or nothing to do with the IRA. This situation was worsened by the fact that the IRA had gotten word that internment was coming and had arranged escapes for its members (Molone 2002, 101). According to the IRA Chief of Staff at the time, of the 342 suspects 92


initially rounded up, only 30 were actual me mbers of the IRA (Parker 2007, 160). There were also widespread reports that in ternees were mistreated by the police. 42 Some of these reports were confirmed by the British government, who nevertheless maintained that they did not constitute physical brutality as it unders tood the term. The governments failure to address the abuses da maged its credibility even further, and in December of 1971 the Republic of Ireland filed the first ever inter-state case brought before the European Court, claiming that the emergency procedures against suspected terrorists violated the European Conventi on on Human Rights (161). The court ruled in its favor six years later, but stopped short of describing the techniques as torture. Predictably, internment led to a considerab le increase in support for the IRA (Woodwell 2005, 173; Coogan 1994, 261; Parker 2007, 160). The last major event to symbolize Cat holic grievances occurred on January 30 th 1972, when the Northern Ireland Civil Rights Association organized a march of 10,000 people against internment. After being assaul ted with rocks, British soldiers killed 13 unarmed civilians and wounded 17, one of whom died weeks later. The event came to be known as Bloody Sunday. In the wave of indignation that followed, the British Embassy in Dublin was burned to the ground. The soldiers later claimed that they were shot at, but no one else presen t that day heard any shots unt il the army started firing including some well respected journalis ts (Coogan 1994, 261). A Br itish tribunal later exonerated the soldiers, giving the Repub lican community yet another point for 42 The Royal Ulster Constabulary, with supervisio n from the British Army, applied five techniques previously practiced in colonial emergencies: hooding, wall standing, subjection to noise, relative deprivation of food and water, and sl eep deprivation. Detainees were also forced to run an obstacle course over broken glass and rough ground while being beaten, and were deceived into believing that they being thrown from high-flying helicopters (Parker 2007, 161). 93

PAGE 100

propaganda. Within weeks, the British suspended the Stormont government and assumed direct control over Northern Ireland (111). Indiscriminate violence suffered by Irish ci vilians created sign ificant support for the IRA, fueling anger and amplifying the nati onalist injustice frame. According to the Sutton database, British security killed 162 Catholic civilians from 1969-2001, compared to only 23 Protestants civilians (Sutton 1994). British security forces did not treat criminals equally either. Over the same period, British security killed 140 Republican terrorists, but only 16 Protes tant terrorists (ibid). According to an analysis by White between August 1969 and June 1972, the British state and its soldiers responded to viol ent Catholic disturbances and killings by Republican paramilitaries. 43 However, of the eight coeffici ents associated with variables that measured violent Protestant disturbances and persons killed by Loyalists, only one was associated with a signif icant increase in repression; when Loyalists killed more people, the British Army responded by killing more people. This is a surprising fact given that most of those killed by both parties were Catholic civilians (White 1995, 345). This adds to qualitative evidence a nd claims of moderate Republicans that the security forces were not only biased against Ca tholics, but that they also collaborated with Loyalist paramilitaries. Harsh and indiscriminate repression from the security forces and Loyalists paramilitaries 44 was responsible for most Republican violence, transforming a weak and marginalized IRA into the most powerful paramilitary in Europe. 43 The full extent of this collusion is not yet fully known, but both official and unofficial investigations are underway. What is certain is that, starting in 1987 a covert military intelligence unit within the British Army provided Army Intelligence to the Ulster Defens e Association to create a list of possible targets (Melaugh 2009b). 44 The term paramilitary is extremely common among students of the Northern Ireland conflict. In this section the term will therefore be used interchangeably with the term terrorist organization. 94

PAGE 101

Clashes between rival crowds would draw in the security for ces, who would often side with Loyalists and immediately come into conflict with nationalis ts. As the conflicts escalated Catholics would build barricades and throw gasolin e bombs at the police, who would respond with baton charges. Behind th e police, mobs of Loyalists, often armed with guns and gasoline bombs, set fire to homes and public houses (Moloney 2002, 66). White provides another illust rative example. On June 27 th of 1970, Protestant crowds attacked a small Catholic enclave in East Belfast. The British Army was called in to defend the community, but they claimed to be stretched too thin by other riots. The Provisional IRA stepped in and s hot at Protestant attackers from a local Catholic church. Two Protestants were shot dead, and one Cat holic defender was killed. Within a week the British government and British Army responded with more repression. On July 3 rd the British Army raided Catholic homes in West Belfast searching for weapons. Large crowds confronted them, afraid that they w ould be left defensele ss against Protestant paramilitaries. 45 Rioting broke out, accompanied by a gun battle with the Official IRA. Two civilians were shot dead, including a Polish photographer, and a third was crushed under an armored car. The entire area was placed under curfew for 36 hours while the people were kept in their home s. During the curfew, the Britis h Army gave a tour of the area to two Protestant ministers, apparently to demonstrate that th e Catholics had been subdued (White 1995, 347). Episodes like these indicate that British soldiers viewed their objective in Northern Ireland as cooperating with the Pr otestant community to control and subdue the agitating Catholic community (White 1995, 346) But indiscriminate violence from both 45 Obviously, an unarmed and terrorized populace might be even more dangerous if they are forced into the waiting arms of militants. In Ballymurphy, defens e against Loyalist attack was reportedly the main motivation for those who joined the local unit (Moloney 2002, 84). 95

PAGE 102

groups, and particularly the Loyalists, system atically drove civili ans into the violent Republican movement. According to Coogan, It is hardly an over-simplification to say that the Catholics were forced off the stre ets into the arms of the I.R.A., who were subsequently maintained in their recruiti ng by the activities both of the Protestant paramilitary mirror organizations which grew up to the combat the I.R.A. and of the British Army (1994, 259). Prior to August of 1969 the IRA had been largely maintained through family tradition. But as a result of ineffective repression, the IRA by 1971 had developed into a large and well-organized force (Woodwell 2005, 173). The new recruits, known as sixty-niners (after the year), were motivated by fear of Loyalist violence and an overwhelming desire to fight back (Molone y 2002, 80). According to Bell, the IRA by 1970 had serious trouble finding se rvice positions for the huge influx of recruits, creating danger of a guerilla overload (2000, 225). By 1974 a majority of British voters wanted to withdraw troops from Northern Ireland (McAllister 2004, 127). By the end of 1974, however, war wearine ss was taking hold in the Catholic community, and support for the Republican movement was beginning to fall (Woodwell 2005, 173). Negative publicity, internment, and other security policies were also beginning to take their toll on the IRA. Hundr eds of members were killed or imprisoned, and at the same time, secret meetings be tween IRA leaders and the British government through its Secret Service gave the impression that the state was w illing to compromise, and perhaps even withdraw British soldiers from Northern Ireland. The IRA declared a cease-fire at the end of 1974 that la sted until the end of 1975 (ibid). The ceasefire was tenuous. Deaths inflicted against security forces were cut in half. Civilian deaths actually increased markedly this year as a result of sectarian violence 96

PAGE 103

against civilians committed by the IRA, Ulster Defense Association (UDA) and Ulster Volunteer Force (UVF) and internecin e violence amongst Republicans (Woodwell 2005, 185). The British government used the cease-fi re to enhance its surveillance of the Republican movement. By the time the cea se-fire ended in November 1975, British intelligence networks had laid the groundwork for a massive and successful crackdown on the IRA in the years that followed (173) In response to the growing success of security forces, a cell struct ure was adopted in 1977, featur ing smaller operational units and a looser hierarchical stru cture. This structure made th e organization more difficult to penetrate (ibid). The first steps toward a lasting peace occurred with the signing of the 1985 Anglo-Irish (or Hillsborough) Agreement, wh ich granted the Republic new consultative roles in the governance of Northern Ireland. The centerpiece of the agreement was the creation of the Intergovernmental Conference, at which Irish ministers could express their views on issues related to the North. The agr eement also included a joint statement that the status of Northern Ireland could only ch ange through the consent of the populations in both the North and the South. This marked the first public recognition of Northern Irelands sovereignty with respect to reunification (176). The agreement represented a new era of cooperation between the UK a nd Ireland; Northern Ireland was no longer viewed as a point of contention, but rather as a common problem for the two governments. Many have marked the end of the conflict with the 1994 ceasefire, the 1995 Framework Documents, or the 1998 Good Friday Agreement in which the IRA renounced its demand for British withdr awal (Woodwell 2005, 171; Sanchez-Cuenca 2007, 297). Deaths have indeed dropped off si gnificantly since the signing of the Good 97

PAGE 104

Friday Agreement. 46 As long as the killing continues, however, declaring a definitive end to the conflict will remain premature. Organizational history According to Moloney, most members of the IRA joined for a simple set of reasons: they believed that only armed force could force the British from Ireland, and that those who advocate parliamentary methods sell out the nationalist struggle. Their history is full of examples of IRA leaders who turn ed to parliamentarism via Sinn Fein (their political wing) but failed to force the British out of either the North or South of Ireland (Moloney 2002, 55). The only parliament to wh ich they give allegiance is the Irish parliament of 1921, the last gathering of representatives in pre-Treaty Ireland with all thirty two counties. This parliament was com posed of Irish MPs who refused to sit at Westminster, and instead sat in a parallel government in Dublin while claiming to be the only legitimate parliament of Ireland. It has become an article of faith for the IRA that no other parliament or government can claim the legitimacy bestowed by that parliament. In 1938, its dozen or so survivors passed their authority on to the IRAs Army Council to safeguard until all of Ireland could once agai n freely choose their own government. It is on this basis that the IRA leadership frames its claim to be the sovereign government of Ireland (56). The IRAs long history of violence with the British makes it impossible to describe the background of The Troubles without also recounting the IRAs 46 Since the end of 1998 there have been 12 sectarian killings. Fifteen of them were carried out by Loyalist groups, one by the INLA, and one by the Provisionals (McKeown 2009). 98

PAGE 105

organizational history. Thus, much of the IR As organizational history has already been covered. To summarize what has already been stated, the IRA began in 1916, split in the Irish Civil War over the Treaty that ended the Irish-Anglo war, and subsequently engaged in two unsuccessful campaigns against Britai n and Northern Ireland before the outbreak of the Troubles: one from 1939 to 1945 (the Forties Campaign), and one from 1956 to 1962 (the Border Campaign). This section will provide a more detailed history of the IRAs organizational history and provide some detail left out in the exposition above. The IRA was at its peak during the AngloIrish war. It had the backing of the majority of Irish people, it fought a long and difficult campaign, and it forced the British to the negotiating table. Its strength was ruptured by the Ir ish Civil War, and its public support increasingly resembled something mo re like tolerance. In 1939 the IRA was banned in the South of Irela nd after conducting a bombing cam paign against the political leadership of the Irish Free State (Engene 2004, 115). However, it issued an order to its members in 1948 forbidding attacks on the Iris h police, Garda Sioc hana, or any other military force of the Irish free state (Mol oney 2002, 50). The IRA misread public opinion at the start of the Border Campaign, and the support it was expecting from Northern Ireland never materialized. By 1962, the organi zation was reeling from failure, and its members felt demoralized. Cathal Goulding was appointed chief of staff in 1962. Goulding pushed the IRA harder to the left than at any point in its history, embraci ng a rigid Marxist analysis of Northern and Southern Irish politics. Goulding turned to the ideas of intellectuals Roy Johnston and Anthony Coughlan, the latter a young lecturer at Trinity College in Dublin. 99

PAGE 106

Their political program borrowe d Stalins stages theory an d planned for Ireland to pass through three distinct phases before r eaching the goal of a workers republic. First, agitation on civil rights issues would create a liberal parliamentary democracy in the north. This would be fac ilitated by working-clas s cooperation between Protestant and Catholics. Second, revolutiona ry links would be established across the border, and Northern workers would ma ke common cause with their Southern counterparts, already being radicalized by Sinn Fein agitation. The third phase was revolution and victory (Moloney 2002, 57). A nother defining characteristic of the Johnston-Goulding strategy was the cooperati on or merging of the IRA and Sinn Fein with like-minded progressive political parties the classic broad-front strategy (58). The downgrading of armed struggle creat ed rifts in the organization. Critics accused Goulding of deliberately runni ng down the organization, dismantling its command structures, scaling down training, and diminishing the IRAs store of weapons. The ascendance of this new secular leftist ideology was alienating for traditional conservative Catholics (Moloney 2002, 54). The proposal to join with the Communist Party of Northern Ireland, the southern Irish Workers Party, the Connolly Association, and the Connolly Youth movement fixed the idea among conservative dissidents that the IRA was being slowly overtaken by godless Marxists (59). The IRA had always existed to prot ect its communities, but by 1969 it was a shadow of the organization that fought the British to a standstill in 1921. It could now do nothing but watch as the shooting and burning raged on in the Catholic communities. In August of 1969, an entire row of houses called Bombay Street was burned to the ground, and a young boy was shot dead (Moloney 2002, 71). The atrocity was worsened by the 100

PAGE 107

fact that British troops had been present shortly before. Within the IRA there was now outrage at Gouldings neglect towards m ilitary matters, and many felt that the IRA needed to recommit itself to armed struggle. Battles occurred at IRA conventions and at Sinn Feins annual conferences, where Gouldings agenda was blocked by traditi onalists and conservatives. A frustrated Goulding turned to increasingl y confrontational tactics. He began purging his enemies, and, in order to secure a majority on th e Army Council, expanded the number of members from seven to twenty. The once highly secret body that went to great lengths to hide its activities from the aut horities was now nearly forced to hire a hall for its meetings (Moloney 2002, 59). In order to establish a genuine democrat ic republic and connect with the working classes, Goulding proposed an end to abst entionism through Sinn Fein, IRAs political front. 47 This was a very controversial move, as abstention was used to symbolically deny legitimacy to the government of Northern Irel and. Largely because of this conflict, the IRA split in December of 1969, forming the Official IRA and the Provisional IRA (PIRA). In January of 1970, Sinn Fein also split, with Official Si nn Fein breaking from abstentionism and Provisional Sinn Fein maintaining it. Official Sinn Fein would later become The Workers Party, with Provisiona l Sinn Fein becoming Sinn Fein. The new Sinn Fein ended its policy of abstention and began engaging in electoral politics in 1986 (Richard 2001, 76). 47 The terms political front and political wing are used interchangeably. The latter term should not be construed to imply equivalence with the armed wing, however, as political parties usually spring from and are subordinate to terrorist groups This is certainly the case for both the IRA and ETA. In 1974, the Basque Revolutionary Party (EIA) and the Popular United Party (HB) sprung from ETA-pm and ETA-m respectively. Sinn Fein was founded as an Irish separatist organization in 1905, but taken over by the IRA in 1949, when it elected a member of the Army Council as its new President (Richard 2001, 74). 101

PAGE 108

Initially, the remaining members of th e Official IRA competed with the Provisionals, conducting scattered attacks in its first two years. The group ended its militant campaign in May 1972, however, and merg ed with the Workers Party in the late 1970s (Weller 2005, 184). 48 After the OIRA called a ceasefire in 1972, the hardcore militants either left or were forced out. These militants formed the Irish National Liberation Army (INLA). The OIRA attempted to crush this new organization before it could establish itself. Five peopl e were killed in ensuing feud. 49 Though significantly smaller than the IRA, the INLA was the sec ond largest Republican paramilitary group in the late 1970s. The group specia lized mainly in the assassin ation of high-profile targets and was responsible for 117 deaths during the conflict (Weller 2005, 186). After the 1981 hunger strike, in which three of its members died, the organization be gan to break apart. This split culminated in the creation of the Irish Peoples Liberation Organization (IPLO), which developed a reputation for inter-Republican violence and criminality (Engene 2004, 116). The biggest feud among Re publican organizations was between the INLA and the IPLO. 50 The INLA eventually recovered, but the PIRA has remained the pre-eminent Republican terrorist group. In 1998, amid conflict over the peace agreement brokered and accepted by the PIRA, the Real IRA (RIRA) split from the Provisionals. The RIRA has carried out 48 The Workers Party was previously called Sinn Fein The Workers Party, and before that, Official Sinn Fein. 49 The existence of such feuds seems to cast doubt on the practice of consolidating various Republican groups as if they coordinate and are informed by the same strategy. However, the feuds were limited enough that we can view them as su ccessful efforts at control from th e movements center. The two places where lumping Republican groups together does not stand up to scrutiny are the early years of conflict between 1969 and 1972, when the Officials were co mpeting with the Provisionals, and after the Good Friday Agreement of 1998, when the Real IRA broke from the Provisionals and killed 29 people in the Omagh bombing. This attack were widely condemned within the Republican movement. 50 The year 1987 saw the greatest level of feuding between the IPLO and INLA. During this year the IPLO killed five members of the INLA, and the INLA killed three in the IPLO (McKeown 2009). 102

PAGE 109

more violence in the current period than any other Republi can group, and are the biggest threat to the ongoing peace process. Strategic History As the conflict intensified in 1970 the Provisional IRA 51 was under serious pressure to find more modern weaponry. They sent operatives to Europe and the United States, establishing regular contacts with European natio nalist movements like the Basques, Corsicans, and Bretons. In 1972 th e Bretons suggested that the IRA contact Qaddafis government in Tripoli, a regime which had advertised its willingness to assist revolutionary movements that created trouble for the old impe rial powers that once ruled the Middle East (Moloney 2002, 9). The rela tionship proved immensely fruitful for the PIRA. Having just ended centuries of coloni al rule through a bloodless coup, Qaddafi shared the IRAs hatred of th e British. Like the PIRA leadership, he blamed colonialism for his countrys terrible history (Molone y 2002, 8). Qaddafi was willing to give the organization both arms and cash in exchange for the simple promise to use them against the British government. At least $3.5 m illion ($10 million in current prices) was transferred to the PIRA during the Troubles, wh en its campaign was at its height (10). At least one shipment of weapons was inte rcepted, but there are strong indications that as many as three others made the tr ip from Libya to Northern Ireland (Moloney 51 Because the Official IRA gave up violence in 1972 references to the IRA from this point onward will refer exclusively to the PIRA. In the discussion of targ eting strategies the term IRA will refer to both the hegemonic Provisional Irish Republican Army and its numerous satellites such as the INLA, RAF, and IPLO. These organizations have mostly cooperate d with one another despite occasional feuds. 103

PAGE 110

2002, 10). This relationship turned sour in 1974 after the Provisionals middle man in Tripoli took too many liberties re formulating Republican strategy. 52 The relationship was revived in 1981 by the death of ten hunger-striking prisone rs from the IRA and INLA. The prisoners were protesting their revoked political status, which had been granted earlier in the 1970s. The hunger strikes gave th e IRA their greatest propaganda victory to date, bringing attention to th e Republican cause in a way that terrorist attacks never had. The attention gained was ma gnified when Bobby Sands, re presenting Sinn Fein, was elected to the British Parlia ment six weeks into the stri ke (Woodwell 2005, 176). Despite their continued policy of abstention, the electoral success of Sinn Fein candidates increased significantly in Northern Ireland. Qaddafi sent some $1.5 million to the IRA over the next three years, a mere fraction of what they received in the 1970s. But his covert war with the United States intensified in the mid-1980s. In May of 1984 Qa daffi had to fight off a fierce coup, with strong evidence of British and CIA support. In 1986, U.S. warplanes bombed Qaddafis home, killing nearly eighty people, including his adopted daughter (Moloney 2002, 14). The planes had taken off from British air bases with Margaret Thatchers full approval (ibid). Qaddafi was now ready to offer the IRA much more than money. Nasser Ashour, the number three man in the Libyan Intelligence Service, traveled to Ireland and offered the IRA $10 million and three hundred tons of modern weaponry to use against the British government (15). The IRA received 4 shipments of weapons, including automatic 52 The middle man was a schoolteacher named Mister Eddi e. After making a close personal connection to the supreme commander of the Ulster Defense Associatio ns (UDA), a Loyalist terrorist group, and became intrigued by the Loyalists idea for an independent Northern Irish state. He took the liberty of going behind the PIRAs back to invite delegations to Tripoli fr om both the UDA and IRAs political wing, Sinn Fein, to discuss an economic aid package for the UDAs plans. Before it took place, however, the meeting was leaked to the media back in Irela nd, and the Sinn Fein delegation re fused to attend. The UDA returned home triumphant, claiming to have driven a wedge between Qaddafi and the IRA. 104

PAGE 111

pistols, 1,200 AK-47s, SAM-7 missiles, general-purpose machine guns, RPG rocket launchers, over a million rounds of ammunition, and five tons of Semtex, a highly destructive plastic explosive (20). Transfer of the last and largest shipme nt was disrupted by a high-level informer, and intercepted by French customs official s in 1987 (Moloney 2002, 6). As a result, the IRA lost the crucial element of surprise in its planned Tet offensive. The IRA soon learned what it meant to lose the element of surprise. SAM-7 missiles became useless after the British installed elec tronic countermeasures on helico pters two tested in South Armagh flew right past their targets (Moloney 2002, 23). The IRA attempted to compensate by using heavy machine guns agai nst helicopters, but moving them was far too slow. The British also began flying helicop ters in groups up to five, and reinforced the armor on their vehicles (ibid). The Tet Offensive was designed to cr eate tremendous disorder and shock British public opinion into thinking that occupation was unsustainable. IRA leaders calculated that the Br itish would probably respond to the offensive by reintroducing internment and pressing the Irish Republic to do the same. This was not the first time the IRA hoped to provoke the British into potential ly counterproductive security measures the organization had long believed that it thrived on British repression (Moloney 2002, 22). At its first meeting in 1970, the Army Council devised a three-stage strategy that focused on the need to defend Catholic areas of Belfast. The first stage was to build up the capacity of Northern units so that they could withstand Loyalist or British attack. The second phase was a mixture of defense and reta liation designed to deter Loyalist attacks. 105

PAGE 112

Once strong enough, the IRA would launch the th ird phase an offensive war designed to bring Britain to the negotiating table a nd withdraw all of its forces from Ireland (Moloney 2002, 85). Once the Ballymurphy riots broke out duri ng the Easter riots of 1970, the local IRA commander held back IRA members at gu n point. As one of the detainees explained, Adams wanted ordinary people involved in rioting as a way of radicalizing them. (Moloney 2002, 88). Had the IRA intervened the first day, the trouble may have ended in a few hours, leaving the IRA defeated and the people of Ballym urphy unwilling to take up arms. Instead, the rioting lasted for four da ys and affected thousands of people in West Belfast (91). The months of June and July in 1970 provided another opportunity to repeat the exercise, with more historic blunders from the British military and Unionist government. The riots that summer made the Ballymurphy IRA the most militant in Belfast (88). These successes cleared the way for the o ffensive war of attrition phase, which focused on the British Army, the Royal Ul ster Constabulary, the Ulster Defense Regiment, and, when the IRA was constraine d by weakness, the unsuspecting employees of opposing forces. After the Bloody Sunday killings, the PIRA and OIRA made a serious of blunders that angered constituents and created pr essure to enter negotiations (Moloney 2002, 113). These included the Bloody Friday killings and the killing of a well-respected local Catholic (these are discussed in detail on pages 113 and 177). The next significant change occurred in th e late 1970s as the security forces grew more effective, causing IRA killings to fall. There is debate as to whether this constituted a genuine change in strategy, or whether it was simply an acknowledgment that the 106

PAGE 113

British were not easily coerced, and that violen ce would have to be carried on at a lower level due to organizational constr aints. A strategy of attrition was certainly in effect much earlier as soon as the IRA had gained e nough strength defending Catholic communities from Protestant terrorism it launched its o ffensive against British security forces (including the police of Northern Ireland). The strategy was certainly in full force by 1973, when the two IRAs combined killed 103 British soldiers. In 1973 Republican-inflicted fatalities dropped significantly falling almost as rapidly as they climbed the previous year, at the peak year of the Troubles. Violence increased slightly the following year and then dropped during the 1974-1975 ceasefire. The British used this time to vastly increase its intelligence apparatus, contributing to a rapid fall in deaths in 1977, a decline that continued in 1978. According to White, the IRA moved to a long war strategy in 1976-77. The target was still the British presence in Ireland, but in stead of forcing the British out with major disruption and destabilization, the IRA turned to a long te rm strategy of small-scale disruption and waiting the British out. According to White, this shift resulted from a recognition that the IRA had to combine milita ry with political activity in order to achieve its goals (1997, 41). Some authors consider this to be the beginning of the war of attrition strategy. However, declaration of th e long war merely represented an updating of the original war of attrition st rategy in light of new facts. The IRA had wrongly believed in the beginning that a relatively small number of deaths would force the British out. According to one member, The Army Councils first target was to kill thirty-six British soldiersthe same number who died in Aden. The target was reached in early November 1971. But this, the Army Council felt, was not enough: I remember, Dave [OConnell], amongst others, saying: Weve got to get eighty. Once eighty had b een killed, Dave felt, the pressure on the British to negotiate would be imme nse (Sanchez-Cuenca 2007, 296). 107

PAGE 114

In 1972 alone, the IRA killed 235 people includi ng 95 British soldiers, but the British did not withdraw. Up until the truce in 1975, the IRA was enthusiastic that the next year would bring victory. The long wa r doctrine corrected these mistakes and led to deep organizational changes, but did not transform the nature of the war of attrition against the British (Sanchez-Cuenca 2007, 296). By the early 1980s the chances of a quick military victory had receded. The ceasefires in the mid-1970s did not deliver any tangible progress, and the election of Margaret Thatcher in 1979 brought major revers als. There was little indication that the government would compromise on granting politi cal status to IRA pr isoners, and it was clearly prepared for a long-term military commitment to Northern Ireland (McAllister 2004, 127). The hunger strikes in 1981 opened up a significant opportunity for the IRA to break into the electoral arena. At first it had been reluctan t to support the candidacy of Bobby Sands for fear of public humiliation that might undermine the hunger strikes. Sands victory in the election and his succe ss in maintaining the strikes created huge sympathy for the Republican movement far more than its violence ever had (Woodwell 2005, 176). This support encouraged the IRA to participate in the 1982 Northern Ireland Assembly elections, where Sinn Fein ran under its own name for the first time. It won 10.1% of the first preference vote only 8 poi nts less than the long established Social Democratic and Labour Party (SDLP) (McAll ister 2004, 128). Sinn Fein president Gerry Adams was even elected in West Belfast. Neither Adams nor the members of the Assembly took their seats, however. Ending abstentionism would be the key to integrating the Republican movements political and military strategies. This decision officially occurred at Sinn Feins 1986 108

PAGE 115

conference. Sinn Feins Gerry Adams and John Hume of the SDLP then commenced talks to make common cause. The talks ende d in 1988 with disagreement over the IRAs use of violence. By this time, however, the IRA had fundamentally reassessed its strategy and decided on a dual approach : a ballot paper in one hand and an Armalite [assault rifle] in the other (McAllister 2004, 128). Military successes generated public support that translated into votes and seats at th e negotiating table, wh ile electoral success legitimized the continued use or threat of force (ibid). The number of fatalities inflicted by Republican terrorist groups remained relatively constant from 1977 to 1990, when the idea of a more politically oriented strategy began to win support among IRA militant s. Republican terrorist groups preferred to target British soldiers ove r local security forces, because they saw them as a better symbol of colonial rule a nd thought they had a greater im pact on British public opinion (Drake 1998, 112). But from 1970 to 1993, British Army deaths dropped off significantly (118). The rate of RUC and UDR deaths re mained fairly constant until the mid-1980s, and then they also declined (ibid). Killing so ldiers became more difficult, because as the conflict progressed, they made up a smaller prop ortion of the security forces in Northern Ireland. 53 Killing soldiers was also made more difficult by the increas ing use of Kevlar body armor, and the adoption of patrolling methods that made it difficult to attack the army and escape (112-113). In 1977 the PIRA carried out an assa ssination campaign against prominent businessmen, arguing that their very presen ce underpinned the Br itish occupation of 53 Partly for this reason, the IRA has also attacked British soldiers in England and on the European Continent. Between Jan 1, 1988 and Dec. 31, 1990 they killed 14 soldiers in England and seven soldiers and air force members on the continent, compared to 39 soldiers and one naval recruiting officer in Northern Ireland over the same period (Drake 1998, 116). 109

PAGE 116

Northern Ireland. The campaign was extremely unpopular, and was quickly terminated (Drake 1998b, 5). It was resumed in 1985, howev er, with attacks now circumscribed to activities deemed supportive of security forces, such as the maintenance of police stations, doing building work for the British Army, or supplying goods to the security forces (ibid). This campaign a clear example of soft target transference was at least tolerated by the Republican community. It killed 27 people between 1985 and 1993 (118). Due to the increasing use of body armor, the IRA showed a greater tendency to use explosives rather than guns to kill Brit ish soldiers during the 1980s, creating a risk of collateral killings. The most notorious of th ese attacks occurred at the Remembrance Day Parade in Enniskillen in November of 1987, wh en 11 civilians were killed and 63 injured by a bomb targeting members of the Ulster Defense Regiment (Drake 1998, 114). An IRA spokesman later told a journalist that the outer reaches of the Republican support base were totally devastated, but that thei r central base could take a hell of a lot of jolting and crisis (Townshend 1995, 317). The IRA later made a public announcement that one of the units responsible had been disbanded (316). Polls conducted in the aftermath of the Enniskillen bombing show ed a steep reduction of support in Great Britain for withdrawal from Northern Ire land: from 61% in January 1987 to 40% in November 1987, just days afte r the bombing (Drake 1998, 158). Reports also indicated that potential sympathizers in the Republi can of Ireland and elsewhere were more likely to withhold support or actively oppose the Repu blican movement in the aftermath of the bombing (ibid). The attack drew harsh critic ism from the President of Sinn Fein, Gerry Adams (White 1997, 28). 110

PAGE 117

The most disturbing attack from the IRA occurred on October 4, 1990, when three men who were labeled collaborators and supporters of the security forces had their families kidnapped and were forced into suicide missions. One of the drivers killed had been a kitchen assistant for the Ministry of Defense. This method caused revulsion amongst IRA supporters, and was quickly abandoned (Drake 1998, 67; Kalyvas and Sanchez-Cuenca 2005, 220). These a ttacks also solidified suppor t for the peace process. The early 1990s marked the beginning of the end for th e strategy of attrition, and the first part of a Nationalis t Front strategy. The Nationalist Front was to include the SDLP, the Dublin Government of the Irish Republic, and, in some capacity, the Irish lobby in the United States. Aside from its tang ible success in the elect oral arena, the IRA was effectively pressed into institutional politics by a heavily constrained security environment. By the late 1980s and early 1990s it had become clear that the British had successfully placed numerous double-agents in key positions within the organization. At least one of them was in the IRAs security department, and was preventing others from locating informers. A series of operations sabot aged from the inside left the leadership exasperated but keen on keeping the informer a secret from the rank and file. At least in the short term, this situati on rendered an armed strategy impossible. The first big move towards the Nationalist Front strategy was the truce declared in August of 1994. This truce broke down after 17 months. The second tr uce came into effect in July of 1997, and culminated in the signing of the Good Fr iday Agreement in April of 1998. In this agreement, the IRA renounced its demand fo r a British withdrawal from Northern Ireland, and thereby accepted defeat in the war of attrition (Sanchez-Cuenca 2007, 197). 111

PAGE 118

In fact, this defeat had already been adm itted in 1994 in the IRA document Tactical Use of Armed Struggle (ibid). The Question of Sectarianism According to most commentators, Loyalis t terrorist groups in Northern Ireland engage in highly indiscriminate sectarian attacks against Catholics, while Republican groups like the IRA focus on military targets a nd are, for the most part, not sectarian. However, the predominantly Protestant background of the Republicans military targets has motivated some scholars to argue that Re publican terrorist groups are just as sectarian as their Loyalist counterparts, with the only di fference lying in the av ailability of their targets (Bruce 1997, 64). While Republicans can easily find policemen, soldiers, or disliked state officials, they themselves do not hold regular offices hours (Bruce 1997, 62). When Sinn Fein emerged as a political force and began to occupy office buildings and list its addresses, Loyalists attacked them enthusiastically (63). This argument certainly has some merit the availability of targets is obviously an important determinant of a terrorist groups target selection. On the other hand, it completely ignores the importance of ideological motivation, and thus cannot explain why an ostensibly sectarian IRA should choose to attack targets that shoot back (White 1997b, 125). 54 The idea that the IRA is not sectarian is supported by qualit ative evidence: its 54 Furthermore, available evidence indi cates that IRA attacks on the securi ty forces did not select targets based on their religion Catholics in the RUC and UDR did not constitute a particularly greater or smaller proportion of deaths than they constituted in the organizations membership as a whole (White 1997, 46). Between 1969 and 1989, the IRA killed 238 members of the RUC, an organi zation which is 90-93% Protestant. Of 238, 25 (10.5%) were Catholic. This fi nding suggests that the IRA did not select RUC targets on the basis of their religion it neither targeted nor avoided Catholics (ibid). 112

PAGE 119

leadership and rank and file have repeated ly claimed that their organization is not sectarian, while Loyalists freely admit the opposite (White 1997, 29; White 1997b, 123). A further refutation of the argument that the Republican and Loyalist groups are equally sectarian comes from changes in the availability of British soldiers during ceasefires. Between February and August of 1975 the IRA and the British were on a ceasefire. One outcome of this truce was a re duction in the presence of the British Army on the streets of Northern Ireland. With fewe r legitimate targets to shoot at, the IRA could have changed targets and shot at Protestant civilians. Indeed, for the year 1975 this explanation looks plausible. However, if this were the case then the IRA should not have killed even more Protestant civilians in 1976 after the truce had ended. Simplistic notions of larger and smaller targets cannot e xplain this variation (White 1997, 45). To the extent that the conflict in Northern Ireland is characterized by sectarianism, it is because Loyalist terrorist groups attack the entire Catholic population, which sometimes leads to defensive and reta liatory counterattacks. Protestants attack Catholics in this way because they see the Catholic community as supportive of or in some way responsible for the IRAs violence (Drake 1998, 31). They resort to historically effective means of controlling the Catholic popu lation without realizing that the situation has changed dramatically. The IRA killed 345 Protestant civi lians between 1969 and 1993. Of the 307 Protestant civilians killed by the IRA be tween 1969 and 1989, 181 were killed between 1972-76 (White 1997, 42). Throughout the entire conflict, the IRA has maintained its focus on breaking the link between Northern Ireland and the UK. However, after being drawn into serious ethnic conflict for two y ears, it changed its approach so that the 113

PAGE 120

number of Protestant civilians declined dramatically (i bid). Since that time, sectarian violence from the IRA has only been perpetrated by a minority. Both prominent and rank and file Irish Republicans claim that their main opponent is the British presence in Ireland, and that sectarians in th eir ranks are both in the minority and pressured to change their views once they join the movement. White argues that their opinions change because Republicans understand the two liabilities of sectarian attacks they are not only bad pub licity, but they also eliminate the possibility that Republicans might ally with the Protesta nt community to bring about peace in the future (White 1997, 34). According to Mitchel McLoughlin, a member of Sinn Fein since 1966, there has always been an element within the Republican Movement and on the island of Ireland itself who secretly believe that the Protestant population could be coerced into accepting Irish unity and independ ence. That is not acceptable. Neither is it possible, as Republicans should know, because all the might of Britain could not, and cannot, suppress the Irish resistan ce, (White 1997, 33). Mistakes are a very real aspect of the conflict. 55 In 1972 alone, Republican paramilitaries blew up 41 of their own memb ers when bombs went off prematurely. It makes sense to be more suspicious of allege d mistakes when premature explosions or insufficient warnings are directed agains t external targets, however, because organizations may seek out such a grey area so that they can pressure the state without losing constituent support (Bruce 1997, 61). Still, the death of civilians could always be a result of mistakes rather than anti-Protestan t sentiment, and the fact that Catholics are 55 The PIRA has killed a total of 149 people in collateral damage, and 49 due to mistakes (de la Calle and Sanchez-Cuenca 2006, 14). Although a database exists that breaks down yearly casualties in a way that allows us to calculate how many were accidental or collateral, it has not yet been made available to the public. We can find out how many bombs went off prematurely using the Sutton database, but this does not necessarily mean that all fatalities were accidental. Si milarly, this measure does not count those who are accidentally shot when mistaken for another person. In total, Republican s accidentally killed 139 of their own militants due to premature explosions. 55 Of these, 124 deaths occu rred between 1969 and 1979. 114

PAGE 121

also killed in indiscriminate attacks seems to suppor t that these mistakes do occur (White 1997, 38). De la Calle and Sanchez-Cuenca point out that the Catholic community has traditionally rejected sectarian killings by Re publican paramilitaries, but that this norm has been violated 344 out of 1,961 Republican killings are sectarian (2006, 15). These killings could be interpreted as ignoring the support base and thus violating the preference constraint. However, these sectaria n attacks have been concentrated in areas with high levels of religious polarization, wh ere intense violence creates more radical preferences. Fully 41% of Republican sectaria n killings took place in Belfast. Here, the correlation between Loyalist violence and Republican indiscrimi nate killing is .58. Furthermore, seven contiguous wards in West Belfast are home to 42% of all indiscriminate killing. The same seven wards represent only 12% of Belfasts population, and 14% of the number of wards in the city (ibid). Consistent with this distinction, Whites interviews with IRA volunteers show a much clearer acknowledgement of sectarianism from the Belfast respondent (White 1997, 32). From 1969 to 1989 there were more sectarian killings in Belfast than any other location in Northern Ireland (30). History of State and Sectarian Violence State and ethnic violence has a well documented and extensive history in Northern Ireland. Between 1969 and 1993, Briti sh soldiers killed 154 civilians in Northern Ireland with virtual impunity (W hite 1995, 331). The most notorious example 115

PAGE 122

of state brutality occurred on Bloody Sunday (January 30 th 1972), when the British Army killed 14 unarmed civilians and injured another 16. According to Parker, the nature of IRA violence changed dramatically after Bloody Sunday. Another author puts it differen tly: the botched operations followed one upon the other (Moloney 2002, 111). It is diffi cult to disentangle the effects of Bloody Sunday from the British governments subseq uent decision to dissolve the Stormont government and impose direct rule, an event which the IRA described as one of the most momentous in Irish history (112). Th e aftermath of Bloody Sunday saw the first mainland bombing of the Troubles. In May of 1972, with its ranks swollen by anger at Bloody Sunday and other state violence, the IRA conducted 1,200 operations (ibid). On February 22 nd 1972, the Official IRA bombed the officers mess of the Parachute Regiment in Aldershot, Hampshire, accidental ly killing five female kitchen staff, a gardener, and a Catholic Army Chaplain. T hough their deaths were accidental (the real targets were soldiers), an IRA spokesman a nnounced that the attack had been carried out in revenge for the Bloody Sunday killings (Parker 2007, 162). This attack was soon followed by attacks on civilian targets on the British Mainland, including four simultaneous car bombs in London in March 1973, bombs in the mainline London railway stations in Se ptember 1973, and bombs in public houses in Guildford and Birmingham in the autumn of 1974 (Parker 2007, 162). On March 20 th the PIRA loaded two cars full of bombs and parked them in Belfast city center. A number of conflicting phone warnings were given, which caused the police to move crowds in the direction of one of the bombs. When it expl oded, it killed seven people, five of whom 116

PAGE 123

were civilians. The next day the Official IRA kidnapped and killed a local Catholic, producing a strong and hostile publ ic reaction (Moloney 2002, 111). 56 This reaction was precisely the opposite of what the IRA needed. The fall of Stormont had created a rift among nationalists. Moderates wanted to st rike a deal and ask for reform. Hardliners wanted to continue fighting for revolution and a united Ireland (Moloney 2002, 112). The decision to target civilians on the British mainland, combined with the callous killing of a local Catholic a nd poor warnings in Belfast all coalesced to put tremendous pressure on the OIRA and PIRA to call a ceasefire. By June 26 th of 1972 both organizations had entered into a ceasefire. It lasted for just thirteen days (ibid). Data Analysis: The Impact of Repression and Sectarian Violence on the IRAs Target Selection How do terrorist groups change their targ eting strategies over time, and how are these decisions constrained by the support bas e? It is likely that violence suffered by the terrorist groups supporting community at the ha nds of a rival ethnic group, as well as the state it is perceived to be c ontrolling, creates a preference for retaliation which allows the group to use more radical met hods, including attacks on civilians. This strategy satisfies popular desires for revenge while sending the government a costly signal that wears down its will to continue fighting. Ethnic re taliation may be con ceived as deterrence, retaliation, or more costs for the state in a stra tegy of attrition. In any case, it can easily 56 The local Catholic was William Best, a member of th e locally recruited Royal Irish Rangers (a regiment of the British Army barred from serving in Northern Ireland) on leave from Germany. 117

PAGE 124

lead to a security dilemma in which the safety of one group seems to hinge on the domination of the other. We would expect that such retaliation will be strictly limited to the opposition community civilian deaths among the terror ists local community would merely erode local support for violence. To test this, we must break down the religious composition of civilian casualties inflicted by Republican groups, and compare them to the civilian casualties inflicted on the Catholic community. Figure 4.2 below shows the breakdown of civilian casualties inflicted by the three type s of violent actors in Northern Ireland from 1969-2001. Figure 4.3 shows the trends in civilia n deaths inflicted by each actor. Table 4.1 shows the religious breakdown of civilian ca sualties inflicted by each type of actor. As can be seen in table 4.1, Republican pa ramilitaries are clearly discriminating in the religious background of their civilian targets, but thei r killings are no t nearly as onesided as the British security forces or Loya list paramilitaries. Despite sustaining a fairly high level of violence, the Catholic community appears not to have become as radicalized as the Protestant community. The IRA re gularly emphasizes that it does not choose targets based on their religious identity, but instead on thei r support for the status quo in Northern Ireland (Coogan 1994, 290). The IRA also has generally not engaged in deliberate attacks on civilians, although there have been some exceptions, such as its attacks in Britain (Moloney 2002, 126). In 1985, the IRA even began targeting businessmen and their employees who were in volved in construction work or service provision to the security forces. 57 This change can be seen as an example of target 57 Between 1985 and 1993, 27 were killed in this campaign (Drake 1998, 118). 118

PAGE 125

Figure 4.2: Northern Ireland Fatality Trends, 1969. Northern Ireland Fatality trends0 50 100 150 200 250 300 350 400 450 500196 9 1971 1973 197 5 197 7 1979 1981 198 3 198 5 1987 1989 199 1 199 3 1995 1997 199 9 200 1 British Security Republican Paramilitary Loyalist Paramilitary Total Figure 4.3: Northern Ireland Ci vilian fatality trends: 1969 Civilian deaths by actor0 20 40 60 80 100 120 140 160 180 2001969 19 71 19 73 19 75 19 77 19 79 19 81 19 83 1985 1987 1989 1991 1993 1995 19 97 19 99 20 01 Republican paramilitaries Loyalist paramilitaries British security Unknown 119

PAGE 126

Table 4.1: Religious composition of civilian deaths by actor: 196901 substitution caused by the increasing difficulty of killing British soldiers (Drake 1998, 118). If the IRA targets civilians rarely, what explains their occasional forays into sectarian violence and other type s of indiscriminate attacks? Given that the nationalist support base regularly finds itself on the re ceiving end of Loyalist violence, it is reasonable to suspect that changes in the intens ity of that violence crea te varying levels of support for retaliation. More specifically, we will test the hypothesis that Republican paramilitaries respond to civilian deaths in their community by striking back at Protestant communities due to increased public support fo r violence. We can test this hypothesis by examining the six months in the conflict with the highest numbers of Catholic civilian casualties. These months are shown below in table 4.2. Table 4.2: High-Intensity Months for Catholic Civilian Deaths Month Catholic civilian fatalities and actor responsible Protestant civilian casualties August 1971 18, all by British security 2 December 1971 18, 16 by Loyalists 6 July 1972 27, 17 by Loyalists 22 April 1975 18, all by Loyalists 6 January 1976 15, all by Loyalists 14 October 1993 17, all by Loyalists 14 Five of these months seem to suggest a relationship between the killing of civilians in each rival co mmunity: particularly July 1972, January 1976, and October 120

PAGE 127

1993. More restrained cycles of retaliation may have also occurred in December of 1971 and July of 1972. Such restraint would not be surprising given the total differences in religious killings we noted above. However, one of these months seems to show almost no effect on the IRAs target se lection. In order to determine whether a real interaction is occurring we must do two things. First, we must examine the temporal distribution of deaths in each month and the type of attack s that caused them. The distribution will tell us whether the relationships reflect genuine reprisals, or whether they are merely incidental to ongoing violence. If genuine reprisals are occurring, examining their temporal distribution will also reveal their source. Our second ta sk is to examine trends in violence for the months preceding and follo wing each high-intensity month. Comparing these trends will demonstrate the extent to which IRA violence deviated from the norm in the high-intensity months, and will also tell us whether any change in target selection patterns can be detected beyond each month. In terms of our research hypotheses, changes in target selection that extend beyond each month would indicate that the change is more than simply retaliation ( H3 ) and may in fact add a ne w category to the list of legitimate targets ( H1 ). 58 58 The only information summarized in Table 4.3 that is not demonstrated in text is the distribution of killings. These were excluded to save space. 121

PAGE 128

Table 4.3: Analysis of High-Intensity Months August 1971 18 Catholics killed by British security. 2 Protestant civilian casualties Distribution: Catholics: 61% (11) were killed on August 9 th 1 on the 7 th and the rest were about evenly distributed throughout the rest of the month. Protestants: 2 killed, 1 on the 9 th and one on the 25 th Retaliation? No The second death resulted from an inadequate warning. The first victim was a security man killed by a nail bomb. In the 3 mos. No. Most of the attacks in the following months are collateral, and are not after? near the other security forc e attacks in those three months. Dec 1971 18 Catholics killed, 16 by Loya lists. 6 Protestant civilian casualties Distribution: Catholics: Most were killed early in the month: 83% (15) on December 4 th alone. Loyalists killed one more on the 18 th British security killed 1 on the 10 th and one on the 14 th Protestants: Attacks occurred sporadically in the first half of the month. One was killed on the 2 nd one on the 6 th three on the 11 th and one on the 12 th 122

PAGE 129

Retaliation? No. If any attempt at a reprisal was made, it was both late and heavily constrained. Republican attacks on Protest ant civilians seem to be neither the cause nor the effect of Protestant killings of Catholics. In the 3 No The single sectarian attack during the following 3 months was mos. after? very close to the border of the coding rule. July 1972 27 Catholics killed, 17 by Loya lists. 22 Protestants civilian casualties Distribution: Catholics: Concentrated towards the middle a nd end of the month: 48% occurred between the 10 th and the 20 th and 37% occurred between the 20 th and 30 th Protestants : Concentrated towards the end of the month: 41% occurred on the 21 st and 31 st Retaliation? Yes. Throughout the month, indiscriminate attacks from Loyalists and British security forces were tightly correlated wi th attacks on Protestant civilians. Most Protestant civilian deaths caused by the IRA were deliberate (i.e. sectarian). 3 mos. after? Yes. However, Republicans engaged in no sectarian attacks until the second week of August. The next wave of sectarian attacks appears to be better explained by more proximate se ctarian killings from Loyalists. April 1975 18 Catholics killed by Loya lists. 6 Protestant civilian casualties 123

PAGE 130

Distribution: Four Catholics were killed at the beginning of the month, followed by a reprisal that killed four Protestant civilians. Loyalists committed one more sectarian attack the same day. Republicans r esponded with another attack that was followed by a flurry of Catholic deaths throughout the rest of the month. Republicans committed one more sectarian attack at the end of the month. Retaliation? Yes, but relative to Loyalist killings it w as very constrained on all but the first occasion. 3 mos. after? No. The sectarian attacks that follow are too few and occur too late. January 1976 15 Catholics killed by Loyalists, 14 Protestant civilian casualties. Distribution: On Jan. 4th, Loyalists killed 6 Catholics, including 3 members of SDLP. The next day 10 Protestant civilians were st opped with a fake roadblock and shot. Another was killed two days later. Loya lists killed one Catholic civilian three days later, and began again on the 17 th This was only interrupted after it killed two Catholics on the 25th. Two days later Republicans killed two Protestant civilians, and the following day, Loyalists killed two more Catholics. Retaliation? Yes. When an IRA spokesman was asked wh y the killings were ordered on Jan. 5 th his answer was Why not? It stopped the sectarian killings in the area, didnt it? (White 1997, 21). Protestants did not kill another Catholic civilian in Armagh (the location of the first 3 killings) for another 3 months. 3 mos. after? Yes. However, the number of Loyalist killings brings into doubt the explanatory power of the first cycle. October 1993 17 Catholics killed by Loya lists, 10 Protestants civilian casualties Distribution: The month opened with 3 sectarian attacks on Catholics. In response, the IRA attacked the Belfast office of the Ulster Volunteer Association (UVA). The bomb went off prematurely, killing 10 Protestant civilians (all the Protestant civilian 124

PAGE 131

deaths for the month), as well as the IRA member carrying it. No UVA members were killed. Loyalists responded with a massive reprisal, killing 11 Catholic civilians. Retaliation? Not by Republicans. 3 mos. after? No. Table 4.4: Results Summary No evidence was found for hypotheses 2 or 3, which involved collateral killing and a behavior based expansion of targets (see pp. 72-73). Strong support was found for hypothesis 4 in half of the cases Republican paramilitaries reacted to Loyalist violence by attacking Protestant civilians, presumably because supporters developed more radical pref erences and desired repr isals. In all three of these peak months, the targ eting change lasted into the following three months, thus meeting the criteria for satisfying hypothesis 1. However, this hypothesis predicted that the original change would become permanen t. Instead, sectarian reprisals remained tightly linked to one another. Sectarian at tacks from Republicans in later months were better explained by proximate Loyalist sectar ianism than by self-sustaining Republican violence. The three months where reprisals occurred were July 1972, April 1975, and January 1976. All of these months are unique in that the spikes in Catholic civilian deaths occurred in the midst of negotiations between the British and the IRA. Loyalists sought to 125

PAGE 132

dismantle these negotiations because they saw them as a threat to its status (White 1997, 45). The killing of Catholic civilians led to Republican reprisals ag ainst the Protestant community, but not to the replacem ent of nationalism with racism. Of the five months with the highest number of Catholic civilians killed by British security forces, two were coextensive with the six peak months for deaths caused by the British Loyalists combined. The additional three showed no effect on the number of Protestant civilians killed. This result sugge sts that the extent of sectarian violence emanating from Republican paramilitaries was not a function of Catholics killed by British security (see appendix). Public Support for Violence Electoral Analysis In 1968, survey respondents were asked if they believed it was right to take any measures necessary to achieve a political goa l (see Table 4.5 below). For Protestants, the goal was the maintenance of Northern Ireland as a Protestant country, and for Catholics the goal was the end of partition. 51% of Protes tants said yes, as did 13% of Catholics (Hayes and McAllister 2000, 15). In 1973, anot her survey asked respondents if they believed violence was a legitimate way to ach ieve ones goals. This time attitudes had reversed, with 16% of Protestants and 25% of Catholics agreeing. Clearly, the violence experienced by Catholics in the early years of conflict led them to support more violence themselves. Because the Catholic population was subject to violence from Loyalists and British authorities alike, they experienced much more violence than the Protestant population during The Troubles. 126

PAGE 133

After the 1981 hunger strike, Sinn Fein began to take politics much more seriously. In 1986 it officially ended its policy of abstenti on. During the 1980s, Sinn Fein won around 10% of the vote, with a slightly de clining trend that was halted in the 1990s during the peace process. In the Westminster parliament, Sinn Feins share of the vote went from 9.9% in 1992 to 16.1% in 1997 af ter the 1994 truce. After the Good Friday Agreement in 1998, its share of the vote ro se to 21.7% in 2001 and 24.3% in 2005. The party was not simply rewarded because viol ence had ended Sinn Fein played a crucial role in the peace negotiations (Sanchez-Cuenca 2007, 302). Table 4.5: Public Support for the Use of Violence in Northern Ireland, 1968-1978 In terms of electoral support, Sinn Fein has generally had a strong record, and has increased its gains in every type of elec tion as years go by. One exception occurred between 1983 and 1986, when Sinn Fein lost almost 7% of the vote in the Westminster by-elections, and between the 1984 and 1989 Eur opean election, where it lost nearly 4% of the vote. What explains this decline? One plausible explanation is the Brighton hotel bombing that occurred in October 127

PAGE 134

of 1984 4 months after the Westminster by -election. The bomb was set off in the section of a hotel where Margaret Thatcher and many other politicians were staying for the British Conservative Party conference. It killed 5 people and injured 31. Conclusion The years 1975-6 saw 27% of all Protestant civilians killed by the IRA. The three years with by far the most killing of Pr otestant civilians were 1972, 1975 and 1976. Part of the reason for these killings were the cea sefires that were called in July of 1972 and February of 1975. With fewer targets to shoot, IRA militants began to engage in sectarian conflict with Loyalist paramilitaries, and in 1975 they allowed the momentum of sectarianism to carry them through 1976. 59 Group dynamics certainly played a role in the continuation of this violence. However, the degree of sectarianism also reflects the difficulty of running a clandestine military or ganization, with the leadership potentially unable to communicate changing tactics with operatives in the field (White 1997, 54). The violence declined in 1977 for a number of reasons. Among them were the realization by Loyalist paramilitaries that the British government was growing more effective, the IRAs movement toward a long war, and pacts between the IRA and Loyalist paramilitaries in which both sides ag reed to end assassinations (White 1997, 45). Another plausible reason for the drastic d ecline of violence in 1977 was that the IRAs foray with sectarian conflict had alienated much of its soft support. The number of hardcore supporters meant that this loss did no t lead to organizationa l collapse. What it 59 In July 1972 alone, 22 Protestants were killed by Republican paramilitaries. Sectarian killings remained low for the remaining months. Five of these occurred on the day the ceasefire ended or before. By this time Five Catholic civilians had been killed by Loyalists over the same period. 128

PAGE 135

did mean was that the police forces had a much easier time making arrests. Indeed, intelligence gathering during the cease-fire may have been much easier due to discontent created by ethnic violence, which the nati onalist community did not support (Woodwell 2005, 173). The IRA did not kill more civilians as time went by, nor did it become increasingly sectarian. It di d, however, shift towards soft targets when doing so was convenient. Despite the serious setbacks created by the British agents who had infiltrated the organization, the IRA managed to stay re latively strong and continue to hit hard targets. However, the idea that it could win the war of attrition c ould not be sustained, and the turn to politics cu lminated in the 1998 Good Frid ay Agreement. Overall the organization has remained extremely disciplined in its refusal to engage in sectarian violence unlike the Loyalis t terrorist groups who routin ely brutalize the Catholic population. 129

PAGE 136

Chapter 5 Violent Myth Making ETA and Spain The Basque Country (Euzkadi) has been divided between Spain and France since their present boundary was permanently fi xed in 1659 along the Pyrenees mountains (Beck 2005, 46). The Basque Country is about th e size of New Jersey, with about 85% of its territory falling on the Spanish side of the border (Clark 1984, 8). The Basque region of Spain has four provinces: Vizcaya (Bilbao in Spanish) Guipuzcoa (San Sebastian) on the coast and the north, and Navarra and Alava on the south slope of the mountains. Navarra accounts for more than half of this area. Altogether, the Basque region contains about 3.5% of Spanish territory. The Basque independence movement has its ultimate roots in two structural features of the Spanish states development: the historically weak central government, and late development of nationalism. The im portance of these feat ures can be neatly demonstrated by comparing the Basque region of Spain to that of France. When the boundary between the two Basque regions wa s fixed in 1659, Spain was still a loose 130

PAGE 137

confederation (Beck 2005, 46). From 1479 onwards religion was the main unifying force in Spain. More secular perspectives on Spanis h identity began to gain momentum during the rise of liberalism in the nineteenth century, but church and state were only fully separated after the retu rn of parliamentary democracy in 1977 (42). Spains relatively incomplete nation-building impeded state legitimization. Elites in Madrid could not agree on how to concep tualize the essence of the Spanish nationstate. Conflicts over the role of religion and of popular pa rticipation impeded a common conception of the nation-state. These conflicts often became manifest in regime change, which altered the basic idea of Spanish identity and the main features of the Spanish nation (Beck 2005, 75). Between 1876 and 1923, Spain had 27 governments an average of one every 21 months (62). According to Beck, the lack of a historically stable opportunity structure in Spain has led to the radicalization of Basque protest. Spains return to democracy in the mid-1970s has al lowed for the emergence of an opportunity structure for peripheral nationalism (76). By contrast, nation-building in France was rooted in an early French nationalist discourse invented and spread by a core elite and supported by frenchified Basque elites (Beck 2005, 75). The development of a strong stat e in France allowed for a repertoire of contention that could be inst itutionalized into a system of democratic representation, which helped to legitimize the state (ibid). The development of Basque regionalis m in Spain was pushed along by the influx of Castilian workers during the regions tw o waves of industrialization. Regionalist identity then became hardened and radicali zed by the brutal repr ession of Franco, who 131

PAGE 138

tried and failed to stamp out Basque iden tity to create a unified and homogeneous Spanish state. When we compare the longevity of viol ence in the Basque Country relative to Catalonia, it seems as though the weakness ra ther than the strength of Basque national culture was what led to the emergence of a violent nationalist movement (Woodworth 2008, 181). Catalan nationalists could acknowledge that their culture was repressed under Franco without thinking it was on the verge of extinction (ibid). In 1959, millions in Catalonia were using the Catalan langua ge in daily communication. The Catalan language also had a literatary tradition compar able to the best in Europe. Because it had much in common with Castilian Spanish, it was also relatively easy to learn through casual contact. Immigrants were therefore more easily assimilated into Catalonia than the Basque Country (ibid). By comparison, the Basque literary tr adition was weak, or at least it was perceived to be so by the Basques themselv es. Euskera cannot be learned through casual contact. Learning Euskera requires a long period of study and immersion, putting it far out of reach for most immigrants. The difficulty of assimilation stoked Basque nationalist fears that their culture was suffering what ma ny of them described as a form of genocide (Woodworth 2008, 182). According to Conversi, the Basque experience is one where violent conflicts have been revived by weak identities, and weak identities have been rejuvenated through violen t conflict, (ibid). 132

PAGE 139

Conflict Background The Basque region lost its 700 year old autonomy in 1876 for being on the wrong side of the two Carlist wars, which pitted re gionalists against central ists (Clark 1984, 14). Shortly after, Basque indust rialization began drawing in large numbers of foreign workers. The founding of the Basque Nationalist Party ( Partido Nacionalista Vasco or PNV) in 1892 illustrates the rise in Basque nationalism resulting from the perceived threats from Madrid and new groups in the Basque Country. Movements for winning back autonomy met with varying levels of repression. The PNV began to make progress toward regional autonomy in the 1930s. However, the outbreak of the Spanish Civil War in July of 1936 erased this progress. Two of the Basque provinces (Alava and Navarra) chose to side with the rebel forces; the other two remained with the Republic. The Republic formally declared the Basque region autonomous in October, but by June of the following year (1937), the rebel forces had al ready conquered all of the Basque region (Clark 1984, 14). The combination of bloodshed during th e Civil War and the ferocious oppression under the Franco dictatorship raised the salienc e of ethnic heritage as a source of identity for the Basques, and helped bind them togeth er as a political for ce (Clark 1984, 14). As a consequence of their choosing the wrong side in the Spanish Civil War, Franco labeled Vizcaya (Bilbao) and Guipuzcoa (San Sebastia n) traitorous provinces and revoked the last vestiges of their auto nomy, bringing them fully under central control (Beck 2005, 47; Shabad and Ramo 1995, 419). The Franco regi me sought to replace Spanish regionalism with a single Spanish persona lity. Both physical and sy mbolic repression targeted against any outward manifestation of Bas que cultural and political identity. These 133

PAGE 140

policies further entrenched th e perceived boundary between Ba sques and Spaniards in the Basque region (Shabad and Ramo 1995, 419). Spanish policies to discourage the use of Eu skera had started in the last decades of the nineteenth century, but Francos postCivil War measures dwarfed all preceding attempts at cultural repression. By the tim e ETA was founded in 1959, most public use of Euskera was prohibited, and pare nts were not even allowed to give their children Basque first names (Woodworth 2008, 174). Until 1967 it was illegal to teach Euskera in schools (Clark 1984, 12). Following their surrender in the Civil War, Basque nationalists had their property confiscated a nd their church officials denounced and replaced. Use of Euskera was prohibited in all public areas, as was the teaching or use of Euskera in communications media. Random assassinations of Basque citizens were commonplace, and the prisons quickly filled up with politic al prisoners (Clark 1984, 21). Thousands of Basque civil officials and military officers were jailed after the government surrendered in 1937. They were all tried for armed rebel lion. Several thousand we re executed (ibid). According to orthodox Basque nationalists Euskera was the universal language of the Basque region up until the ni neteenth century. The Basque people, they argue, entered into alliance only by their own consent with Castilian kings, each one of whom had to come to the sacred tree of Guernica to swear to uphold their charters of special rights ( fueros ). Among these rights were control ove r their own taxati on, exemption from customs duties, and exemption from military service. Additionally, Basque citizens traveling throughout Spain were afforded sp ecial privileges normally reserved for nobility, such as exemption from arbitr ary arrest and torture (Woodworth 2001, 22). 134

PAGE 141

Critics of Basque nationalism have high lighted its many artificial qualities. For instance, Sabino Arana, the foundi ng father of Basque nationa lism, had to invent both a flag and a name for the country he wished to free from Spanish domination. These critics also point out that the Basques have not been governed by a Basque leader in a single political unit since the eleventh centu ry. Inconsistencies su ch as these are not necessarily damning for the Basque nationa list project, however. As Paddy Woodworth reminds us, Most nations look more than a little ridiculous in the first phase of construction, when the raw mortar of rhetor ic and myth is most apparent (2001, 19). It is indisputable that the Basques have maintained a powerful sense of distinctness. Their fueros were more extensive than t hose of any other region but Catalonia, which lost them 150 years earlier than the Basques. Speci al rights regarding fiscal policy were very consequential for the Basque Country (Woodworth 2001, 24). Under these fueros the Basques had the right to choose their own levels of taxation, were exempt from paying customs duties on goods entering the Basque Country from abroad, and were able to impose duties on goods imported from the rest of Spain. This arrangement gave them the benefit of cheap foreign manufactures while protecting their agricultural sector from Castilian competition (24-5). The development of Basque industry during the nineteenth century was accompanied by conflicts between urban employers and the agricultural sector. The former group saw the customs fuero as an impediment to growth, as it subjected their products to customs duties within the Span ish kingdom, while the latter was in too precarious a position to allow a shift in the customs fron tier (Woodworth 2001, 25). This conflict developed into an urban-rural fission which became highly salient with the 135

PAGE 142

outbreak of the first Carlist war in 1833. Th is conflict was waged between supporters of King Ferdinand VIIs brother, Carlos, and s upporters of his infant daughter, Isabella. Isabellas supporters represented Liberal Sp ain: they favored curbing the powers of the Crown and Church, building a rationalized, cen tralist administration, and advancing free trade. Carlos, on the other hand, represented th e Spain of the Black Legend, absolute monarchy, ecclesiastical dominance, economic protectionism and, the Basques believed, protection of the fueros (ibid). Basque nationalism as an ideology and movement emerged in the 1890s as a result of social developments associated w ith rapid industrializati on. Industrialization led to a spike in population growth and urba nization, creating a new immigrant working class, with thousands of farmers from impoveri shed areas of Spain moving to urban areas of the Basque Country (Shabad and Ramo 1995, 416). The developing working class made Spanish trade unions and anticlerical leftist parties major social and political players in the Basque Country, which create d a serious threat for the traditional urban middle class and for the dominance of the Catholic and rural Basque culture. It was among these latter segments of society that Basque nationalism developed and flourished (ibid). Industrialization also brought social stra tification, with a new class of elite industrialists and financiers who were quickly incorporated into the Spanish oligarchy. The attempts of this elite to integrate the Basque provinces with Spain both economically and politically caused an intens ely hostile reaction from the traditional petite bourgeoisie (Shabad and Ramo 1995, 416). The combined thr eat to the increasingl y idealized Basque 136

PAGE 143

culture posed by new elites and immigrant groups helped to promote ethnic solidarity (417). A similar period of rapid social and ec onomic transformation occurred during the economic boom from 1960 to 1975 as a result of Francos post WWII industrial policy (Shabad and Ramo 1995, 416; Clark 1984, 18). Fr anco shifted the burden of development onto the two major parts of Spanish society ou tside his governing coal ition: the industrial proletariat and the already industrialized Basque and Ca talan regions (Clark 1984, 18). The goal was to increase industrial production in areas which alrea dy had an industrial base and to encourage the migration of workers to these areas. About 20,000 Spaniards migrated to the Basque region each year beginning in the 1950s. Clark estimates that, by the mid-1970s, only about 65% of the population was indigenous (1984, 10). Shabad and Ramo produce an even smaller estimate: 51% in 1975. Residents of urban and metropolitan areas with some Basque ancestry were 60% and 40% respectively, compared to 85% for the rural population (Shabad and Ramo 1995, 416). At the same time, public investment was denied to the Basques and Catalans, and focused instead on Spains poorer regions in order to increase their future industrial potential (Clark 1984, 10). The effects of this industrial policy we re decidedly negative for the Basque provinces. The economic effects were partic ularly harsh on the poor, and led to the radicalization of the Basque working class. Traditional political organizations were completely unable to respond to the challenge s this situation presented (Clark 1984, 20). 137

PAGE 144

Organizational History ETA ( Euskadi Ta Askatasuna or Basque Homeland and Fr eedom) was founded in July of 1959, but its first confirmed killing did not occur until nine years later (for a comprehensive timeline of the conflicts history, ETAs organi zational history, and changes in its strategic orientati on, see figure 5.1 on the following page). 60 In June of 1968, one of ETAs leaders, Javier Eche barrieta, was stopped by the Civil Guard (Spanish military police). Civil Guard Jose Pardines was killed during his escape. Hours later Echebarrieta was killed after the secu rity forces discover ed his location. ETA retaliated two months later (August), killing police commi ssar Meliton Manzanas, a man notorious for torturing political prisoner s (Sanchez-Cuenca 2008, 7). Francos regime responded with a heavy hand, making massive ar rests and declaring a state of emergency in the Basque Country. This move was shortly followed by a state of emergency for all of Spain (ibid). ETA began as a coalition of radical youth groups disgrun tled with the passivity of mainstream nationalist political parties. One of these youth groups was a splinter from the youth wing of the Basque Nationalist Party. 61 Others were former members of the Euzko 60 In June of 1960 a bomb exploded in a train station in San Sebastian, designed to derail a train full of Francos volunteers traveling to a re membrance service for the fascist rebellion. The attack was a failure the only casualty was a one year-old girl. The bombing was never claimed, and there is no hard evidence linking it to ETA. Many attribute this attack to ET A, but others speculate that it was the work of an anarchist group (Sanchez-Cuenca 2008, 6). 61 The splinter group had previously been part of the secret society Ekin, and was pressured to join PNV. From the beginning, the members from Ekin disagree d with the PNV on two fundamental issues. First, Ekin advocated making Euskera the sole official language of a new independent Basque republic, which the PNV found incredibly utopian given how few peop le spoke Euskera at the time. Second, the PNV had made a deal with all the important parties of the Republic for the restoration of the 1936 Statute of Autonomy as soon as Franco was overthrown and the Republic reinstat ed. In exchange, the PNV agreed not to seek further separation from Spain. Ekin did not trust the major parties to fulfill this deal (Clark 1984, 26). 138

PAGE 145


PAGE 146

Ikasle Alkartasuna (EIA), or Society of Basque Students. ETA has had two primary goals throughout its history: independence for the Basque region, and the revival of Basque language and culture. It has also e xpressed varying degrees of socialist ideology at different point s in its history. ETAs first actions were largely symbolic They included the flying of the banned Basque flag (designed by Sabino Arana and his brother Luis), graffiti, and the bombing of Francoist monuments (Woodworth 2008, 175) In the mid-1960s, ETA implemented a strategy of provocation, or action-repression -action. The goal was to goad the dictatorship into indiscriminate repression that would enhance the political capital of ETAs and allow its attacks to become both more frequent and more dramatic. The new wave of attacks would invite more repre ssion and more popular support, until eventually the conditions were laid for a full-scale war of national liberation. This strategy gained significant momentum after the cycle of retalia tion that broke out with the killing of the Civil Guard Pardines and the death of ETA l eader Echebarrieta. The latter gave ETA its first victim and its first martyr. The h uge attendance at Echebarrietas requiem demonstrated that ETA had caught the publics attention. In the ferocious state repression that followed, many uninvolved people were arre sted and tortured. F rancoist security forces would be ETAs best recruiting serg eant, responding to its stimuli like Pavlovs dogs (Woodworth 2008, 176). At its first assembly in 1962, ETA defi ned itself as a Basque revolutionary movement of national liberation (Shaba d and Ramo 1995, 427). Its socioeconomic program was ambiguous and interclassist, but ex pressed in populist rhetoric. This vague opening to leftist ideology was a prelude to major divisions over ETAs identity. The 140

PAGE 147

conflict developed around the extent to wh ich it was an ethnonationalist or leftist movement (Shabad and Ramo 1995, 426). The first major change in ETAs di rection came from the outside. In 1963, Frederico Krutwig published Vasconia Krutwig saw independence as the central objective of Basques, and stressed an extr eme ethnolinguistic definition of Basque identity. Krutwig equated Spanish occupation (earlier invoked by Sabino Arana, the founder of PNV) with colonialism, and he outlined a revolutionary strategy based on Third World models. This comparison helped to legitimate ETAs transition to violence. ETA held its second assembly in 1963, in corporating the notion of colonialism and certain Marxists tenets into its ideology. That year ETA also made its first contacts with the Spanish Communist party. ETAs th ird assembly in 1964 adopted a qualitative change in its ideological expressions, whic h created significant political consequences. During this period the strategy of armed str uggle began to take on sacred overtones, and the struggle was built into an image of lib eration delivered by a vanguard of the enlightened (Shabad and Ramo 1995, 427). Th e fourth assembly, in 1964, sought to organize a leftist nationalist movement and to define the strategy of action-repressionaction that would capitalize on increasing popular and worker mobilization against the Franco regime. After 1964, three distinct tendencies em erged within ETA. The first stressed ethnolinguistic principles and was strongly anti-Spanish. This theme was promoted by founding members exiled in France. The s econd emphasized guerrilla warfare and used anticolonial struggles as its model. The thir d was ideologically pro-worker, replacing the strategy of national liberation wi th Marxist-Leninist class st ruggle in collaboration with 141

PAGE 148

the workers movement. The combination of French action against the anti-Spanish group and Spanish police repression ag ainst the anticolonial group le d to the dominance of the pro-worker group in 1965-1966. However, a frag ile coalition of the two weakened groups held a fifth assembly in late 1966 to denounce the pro-worker strategy, which was seen as a betrayal of the nationalist movement and the result of Spanish infiltration of the organization. The pro-worker faction, in tur n, declared the fifth assembly fraudulent and withdrew. Members of the pro-worker facti on were then expelled from ETA in early 1967 (Shabad and Ramo 1995, 428). Among the new leadership, Marxist anticoloni alists were in the majority. Many founding members saw this ideology as the resu lt of an insidious Spanish influence and left the organization after th e second part of the fifth assembly in March of 1967. This year therefore marked a period of genera tional succession, with contradictory emphases on Marxist class struggle and insurrectionary national liberation based on Third World models (Shabad and Ramo 1995, 428). This was the period in which the strategy of action-repression-action was put into effect. Both the death of ETAs first militant and the assassination of the first policeman cr eated considerable s upport for ETA in the Basque Country (Shabad and Ramo 1995, 428). As a result of growing ETA violence and worker militancy, the Franco regime declared four states of exception between 1967 and 1969. These crackdowns were particularly hars h in the Basque Country, and most of ETAs top leaders were imprisoned. Facing im minent dissolution, ETA had to reconsider its strategy and leadership. Members of the ETA and of the larger social movement were still pushing in many opposing directions. The sixth assembly was held in 1970 to resolve these differences. Instead, it intensified them, and ETA split again. One wing left the 142

PAGE 149

organization and formed its own in the hopes of creating a radical workers party with a new combination of socialism and nationalism less reliant on the petite bourgeoisie. Meanwhile, those who supported armed struggl e and stressed ethnoli nguistic principles left to form ETA-V, sharply criticizing the new leadership for coming under the influence of Spanish Communists and for abandoni ng nationalism (429). The continuity organization became ETA-VI. In December of 1970, shortly after these splits, Franco orchestrated a highly visible trial against several ac tivists accused of killing the to rturing Spanish police officer Meliton Manzanas (Sanchez-Cuenca 2008, 7). The Burges trial was a massive publicity success for ETA, creating solidarity outside of Spain and leading to huge demonstrations, strikes, and church occupations in support of ETAs demands and prisoners (Shabad and Ramo 1995, 429). By the early 1970s, ETA found itself at the forefront of the resistance movement against the Franco regime ( 430). ETA-VI saw this outpouring of public support as a vindication of its plan to form a mass moveme nt based on both national and social liberation. For a short period it seemed that they were correct: ETA-VI was more successful than ETA-V at attracting new me mbers. However, police action following the Burges trial caused many of ETA-VIs leader s to flee into exile. ETA-VIs recruitment was also hurt by its downgrading of armed struggl e, combined with its vacillation on the status of Euskera and of Basque independe nce. By 1972, a majority faction decided to abandon the organization to join the Trotskyite Liga Comunista Revolucionaria (ibid). ETA-Vs radical nationalism and focus on armed struggle proved to be more in tune with ETAs traditional social base than ETA-VIs Marxism. ETA-V grew considerably in 1972 when many radicalized members of the PNVs youth group were 143

PAGE 150

incorporated into the organi zation. In September of 1973 ETA-V held an assembly to endorse its leaders and their deci sion to reorganize into four fronts, of which the military front would dominate. This military front la ter decided, without consultation, to kill instead of kidnap Prime Minister Carrer o Blanco (Shabad and Ramo 1995, 430). The assassination created a new crisis within ETA-V. The disaffected within the workers front left the organization in May 1974 and formed the first radical abertzale (patriotic) party, the Langille Abertzale Iraultzalean Alderdia (Patriotic Revolutionary Workers Party, or LAIA). ETA split into two factions: the first consisting of leaders in France who defended the priority of armed struggled, and the second consisting of activists in Spain who wished to combine military and politic al strategies with high levels of mass mobilization. In late 1974 a definitive split separated these factions into ETA-militar (ETA-m) and ETA-politico military ( ETA-pm). Both sought independence, monolingualism for the Basque Country, a nd socialism through popul ar revolution. The central difference was that ETA-m stressed the primacy and autonomy of the military leadership, while ETA-pm sought a unified leadership combining both political and military struggle (431). The sudden end of the Franco regime in 1975 caught ETA off-guard and in the middle of unresolved internal conflict. ETA -pm leaders increasingly saw violence as a means to achieve lower-level tactical goals such as the release of prisoners, as changes in the political environment created new means to achieve strategic goals that could either complement or supplant armed struggle. ETA -pm sought to participate in the political process and negotiate with the Spanish stat e while producing more violence that would enhance its bargaining position. For ETA-m, violence had acquired a logic of its own, 144

PAGE 151

assuming a more expressive than instrumental role. Changes in the political environment were considered illusory. ETA-m expected poli tical reform to fail, thus leading to a prerevolutionary climate (Shabad and Ramo 1995, 431). At its seventh assembly, ETA-pm decided to foster political connections to create a mass leftist party that could compete in the new pluralistic environm ent. Several radical nationalist groups were formed at this time, ma ny out of the fragment ation of traditional political parties and trade unions (Shabad a nd Ramo 1995, 432). These parties formed the Socialist Patriotic Coordinator (or Koordinadora Abertzale Sozialista KAS) in 1975, composed of both ETAs, LAIA (Patriotic Re volutionary Workers Pa rty), LAB (Patriotic Workers Council), LAK (Patriotic Workers Committee), HAS (Popular Socialist Party), EAS (the Basque Socialist Party), and ELI (F orce of the Basque Workers) (Shabad and Ramo 1995, 432). At the end of 1976, KAS appr oved the KAS Alternative, which had its origins in ETA-pms manifesto of eight poi nts for the Aberri Eguna of 1975. Composed mainly of radical nationalist demands, the KAS alternative program was later put forth by ETA-m (in 1978) as a pre-condition for enteri ng negotiations with the Spanish state. Both ETA-m and ETA-pm rejected th e Reform Law of 1976 and the 1978 Spanish constitution. But by 1979-80, during the establishment of the Basque Autonomous Community, ETA-p m had accepted the supremacy of political action and the leadership of Euskadiko Ezkerra (EE), a coalition that included its own MarxistLeninist Basque Revolutionary Party (EIA), created in 1976, as well as the moderate socialist party Socialist Unif ication for Euskadi (ESEI), Basque National Action (ANV), and factions of the Basque Communist party HASI (Shabad and Ramo 1995, 433). 145

PAGE 152

Shortly after the attempted coup detat by a segment of the military in 1981, ETApm disbanded. Those who wished to continue in armed struggle joined ETA-m. EE evolved into a Basque socialis t party and in 1993 merged with the Basque branch of the Socialist party. The disbanding of ETA-pm in 1981 left ETA-m as the main terrorist group, and sharply reduced the number of violent actions. 62 In 1978 Herri Batasuna was formed as its political wing a coalition c onsisting of HASI, other mass organizations of KAS, and parts of LAIA, ESB and ANV. De spite Basque autonomy, ETA-m and Herri Batasuna continued to see the new Spanish democracy as Francoism in disguise, and perceived an ongoing war between the Spanis h state and the Basque Country. ETAs objective was to de-legitimize existing political institutions established by the Spanish constitution and the Basque Autonomy Statute, while legitimizing itself as the only true representative of the Basque people. Through its network of organizations of the Basque Movement of National Liberation (MVLN), ETA sought to mobilize supporters and public opinion. As late as 1986, internal doc uments planned for violence to accompany mass mobilization until ETAs ultimate objectives were achieved (Shabad and Ramo 1995, 433). ETAs leadership is directly involved in important decisions taken by its affiliated organizations. ETA has been decisive in all three leadership successions of HASI since its founding in 1977. An old member of ETAs executive committee was Herri Batasunas second candidate on a list for th e first European Community elections in Spain. A significant number of the coaliti ons candidates for loca l and regional bodies had been imprisoned ETA member s (Shabad and Ramo 1995, 485). 62 Because it was the only remaining group still claiming the legacy of ETA, from this period onward ETAm will simply be referred to as ETA. 146

PAGE 153

After 1981, ETA evolved into a secret army composed of cells with three to five members with direct connections to the m ilitary leadership in France. The military leadership is at the head of a complex soci al movement called the Basque Movement of National Liberation (MLVN), which divides tasks of political m obilization among legal organizations. KAS is the second in command. Its main political party, HASI, brings together the trade union LAB, the youth organization JARAI, the prisoners support groups, and popular committees (ASK). All of these are part of Herri Batasuna, which competes in elections but does not participat e in representative institutions above the local level. 63 At the periphery, ETA also works th rough specialized organizations dealing with religious issues, education, promotion of the Basque language, mass media, ecology, women, drugs, students, childre n, international solidarity, an d prisoners and refugees (Shabad and Ramo 1995, 435). ETAs 30 year history has been marked by conflict, crisis, and change. Most assemblies have been contested. None of ETAs leaders enjoyed legitimacy for any prolonged period of time, either because of imprisonment, exil, or succession via coup detat. The sources of ETAs factionalism can be attributed to four main issues: the priority given to national versus social liberation; the emphasis placed on mass political activity, and after Francos death; on participation in democratic institutions, relative to armed struggle; the relation between the political and military branches of the organization and the degree of autonomy granted to the military branch; and the 63 This is similar to Sinn Feins abstentionist policy, but different in that it allows for participation in local government. Participation in local government in Spain is not seen as an affirmation of the Spanish states primacy. By contrast, taking seats in either the Republic of Ireland or Northern Ireland was seen by Sinn Fein as an acceptance of partition and theref ore forbidden for most of its history. 147

PAGE 154

desirability of links with S panish political forces. Power struggles have played their part as well, with divisions among the rank a nd file being determined more by friendship patterns and personal ties to l eaders than by ideological and strategic differences (Shabad and Ramo 1995, 436). ETAs emphasis on leftist ideology and discourse at different points in history was mainly the result of changing political c ontext, and it was used to incorporate new social sectors and issues into the nationalis t movement. This gave ETA an increasingly populist image. However, at every critical point the radical nationalists beat the socialists for control of the organization (Shaba d and Ramo 1995, 436; Sanchez-Cuenca 2008, 6). Basque language and occupation remained the primary features of its symbolic capital and the main sources of legitimacy for vi olent struggle and mobilization (Shabad and Ramo 1995, 436). As ETA was reduced to a hard-core vi olent nucleus in th e 1980s, internal conflicts have focused on how to respond to the Spanish state and other political forces in the Basque Country. As its numbers and its ab ility to broaden its base have declined, internal debate over goals and strategies has declined sign ificantly. The primary aim has instead become the very survival of the organization itself (Shabad and Ramo 1995, 488). Herri Batasuna was created in 1977 by former ETA members and sympathizers (Encarnacion 2007, 9). Batasuna has insisted that it is not ETAs political wing. However, in June of 2002, a law was passed that allowed for the banning of Batasuna, and in August of 2002 magistrate Baltasar Garz on suspended Batasuna for three years for its links with ETA. In September of 2002 the government sought a permanent ban, and in March 2003 the Spanish Supreme Court appr oved the permanent ban, simultaneously 148

PAGE 155

banning Batasunas two predecessos, Herri Ba tasuna and Euskal Herritarrok (BBC 2003). This decision marked the first time since 1975 th at any political party had been banned in Spain. After Batasuna was banned, its vot ers voted for marginal party, PCTV (Communist Party of the Basqueland) (Sanchez-Cuenca 2008, 28). Herri Batasuna and Euskal He rritarrok have both glorified etarras (ETA members) as patriots, freedom fighters, a nd idealists. For instance, in March 1997 HB used a propaganda video of masked etarras during the election campaign for the Spanish Parliament. Furthermore, HB and EH have also put imprisoned ETA members on the list of candidates for the Euskadi Parliament a nd some municipal councils. Their political comments on ETA violence also reflect unders tanding or indulgence. The most common practice is to abstain from public disapproval of violent events while other parties openly reject them (Beck 2005, 183). For these reasons both organizations can accurately be seen as political wings of ETA. The relationship between ETA and HB is di fferent from Sinn Feins to the IRA (Woodworth 2008, 182). The leadership of ETA ha ve traditionally lived in exile, which means that their control over HB is necessari ly one step removed. ETAs military leaders ensure loyal supporters in the political leadership, but do not usually participate directly in day-to-day politics, as senior IRA figures like Gerry Adams and Martin McGuinness have always done. As a consequence, ETA leaders have little contact with their social base. In turn, leaders of HB accept orders from those who are considered to be making the real sacrifices (Woodworth 2008, 183). 149

PAGE 156

Strategic History Between 1968 and 2003, ETAs 64 strategic history can be broken down into three phases. The first strategy was one of pr ovocation luring the regime into such indiscriminate repression that the masse s would mobilize and revolt (Sanchez-Cuenca 2008, 17). This strategy was quite successful. In fact, some credit ETA for helping to bring down the Franco government. 65 A state of emergency was declared in the Basque Country after ETAs first killing in 1968. Because the police had no information on the perpetrators, 1,953 people were arrested (San chez-Cuenca and de la Calle 2008, 42). As we will see in the next section, state repressi on during this stage helped to radicalize the Basque population, entrenching th e attitudes that would later allow ETA to take a strong stand against incorporation into a democratic Spanish state. After the first democratic elections in 1977, however, ETA gave up on this strategy, and concl uded that revolution was unlikely in a democratic context. The next strategic period was the war of attrition. No longer expecting the masses to rise up, ETA sought to punish the Spanish st ate until it conceded to Basque territorial demands. The first internal documents articula ting this strategy were circulated in 1975, but it was not clearly put into practice unt il 1978 by ETA-m. In the same year, ETA-m announced that it would stop using violence if the five conditions put forth in the KAS 64 In this section, ETA will designate both the he gemonic organization and its satellite organizations. 65 In 1972 ETA helped provoke a crisis by killing Francos designated succe ssor, the Spanish Prime Minister Carrero Blanco. After discov ering that he attended Mass at the same church at nine oclock every morning, they rigged his car with a bomb that killed him and his driver, throwing them both over a fivestorey building on their way home from Mass (Drake 1998, 60, 160). 150

PAGE 157

alternative were satisfied. 66 Deaths rose dramatically in 1978 the fourth quarter of that year saw the highest number of fatalities in ETAs history (Sanchez-Cuenca 2009, 12). One potential reason for this escalation was the constitutional referendum held in December. Herri Batasuna, Euskadiko Ezkerra (a party associated with ETA-pm), and the Basque Nationalist Party (PNV) the most moderate Basque nationalist group all advocated the rejection of the constituti on via abstention (Shabad and Ramo 1995, 452). ETA violence probably reduced participation in the referendum even further (SanchezCuenca 2009, 12; Sanchez-Cuenca 2008, 18). The resu lt of these efforts was stark: only 45% of eligible Basques voted, compared to a 67.5% Spanish average; 31.3% of eligible Basques supported the constitution, compar ed to 59.4% throughout Spain, and 23.8% of eligible Basques voted no (Shabad and Ramo 1995, 452). Naturally, much has been made of this result by Herri Batasuna and ETA. 67 The PNV and the Basque people appeared to accept the constitution a year late r with their approval of th e Statute of Autonomy, which was framed within constitutional limits Unfortunately, however, the PNV, had consistently hinted that support for the statue was merely a ta ctical move on the road to self-determination (Woodworth 2008, 177). ETAs central offensive began late in the Spanish transition after the first democratic elections in 1977. It did not climax until 1979-80 (Aguilar and SanchezCuenca 2009, 13). This timing is quite surprising in light of the leverage that could have been exerted before the rules of the game had been defined. Why did ETA miss this opportunity? Sanchez-Cuenca sugge sts three answers: organi zational constraints, the 66 The most important of these demands was territorial independence. 67 Most others emphasize the positive side: 68.8% of those who voted in the Basque Country approved the constitution (Linz and Stepan 1996, 100). According to Woodworth, 75% of voting Basques supported the new constitution (2008, 177). 151

PAGE 158

effort to work together with other Basque na tionalist forces, and changes in the level of popular mobilization. The split between ETA-pm and ETA-m in 1974 left the militants in the minority. The sudden death of Franco caught both orga nizations off-guard, and ETA-m logistically unable to launch a massive assault. This constraint was overcome in May 1977 when ETA-pm announced that it would focus on political rather than military struggle, causing its militant wing to walk out and join ETA-m (Sanchez-Cuenca 2009, 17). Furthermore, by this time, the attempt of various Basque na tionalist groups to form a united front in the first general election had thoroughly broken down. 68 The separation of moderate and radical nationalists was well illustrated by the clash among their demonstrators in San Sebastian on September 8 th 1977 (18). The final reason for the late start of nationalist violence was that supporters of Basque separatism were already mobili zed and pressing their demands. Since its foundation ETA had tried to use violence to cr eate revolutionary conditions so the masses would rise up and finish the job; they were not about to risk using too much violence without allowing these newly unleashe d forces to run their course. From 1976 onward, the nationalist left ( ETAs grass-roots support) actively demanded amnesty for all political prisoners The majority of these prisoners were Basques. Particularly in its later stages, th is movement came to endorse explicitly ETAs separatist arguments. The highest number of these demonstra tions occurred in the third quarter of 1977. Not coincidentally, this was also the quarter with the least deadly 68 PNV was in favor of participating in the elections and in convincing the two branches of ETA to do the same. ETA-m demanded total amnesty and legalization for the communist party as a pre-condition for participation. ETA-pm took an intermediate position, agreeing to participate in exchange for partial concessions (Sanchez-Cuenca 2009, 18). 152

PAGE 159

violence perpetrated by ETA for the enti re transition period (Aguilar and SanchezCuenca 2009, 19). Celebration after the first ge neral elections and the granting of full amnesty in October 1977 caused a significant dec line in the number of protests. When the collective action cycle built around amnesty s ubsided, ETA compensated with violence (ibid). In an internal document articulating its new war of attrition against the stage, ETA-m specifically alluded to the decline in popular mobili zation as a reason for action (20). ETAs great offensive hit a lull in 1981, when lethal attacks fell. It is possible to hypothesize that ETA killed less in this period due to the m ilitarys failed coup the reason, after all, that ETA-pm dissolved. Howe ver, the percentage of military casualties among ETAs victims went from about 10% in 1980 to 25% in 1981 (returning to 10% in 1982). If ETA was afraid of destabilizing the nascent democracy, it would not have continued goading the military. The number of fatalities also increased during the second quarter of 1981 (Sanchez-Cuenca 2009, 21). A more plausible reason for the drop in killing is the very high number of arrests that occurred this year. According to the Home Office, 1981 had the highest number of arrests of any year from 1973 to 2003. In all, 714 people were arrested, compared to an averag e of only 222 for all other years. ETA seems not to have recovered from these arrests. In fact, the drop in deaths in 1981 is comparable in size only to what would occur later in 1992 (Sanchez-Cuenca 2009, 22). The available evidence indicates that ETA violence dropped in 1981 due to security i ssues related to the high number of arrests, and not because of their own restraint (Sanchez-Cuenca 2008, 18). 153

PAGE 160

In the late 1970s, ETA joined a popular protest against th e construction of a nuclear power station at Lemoniz in the Basque Country, assassinating managers and engineers involved in its construction (Drake 1998, 21). Between 1978 and 1982, five people were killed in this campaign (de la Calle and Sanchez-Cuenca 2007). During the 1980s ETA also began using car bombs in or der to increase the number of victims and reduce the risk of arrest (S anchez-Cuenca 2009, 8). These indiscriminate killings have been concentrated outside of the Basque Country in order to prevent the deaths of Basque civilians that might erode the support ba se (Sanchez-Cuenca, 2008, 23). The geographic concentration of such indiscriminate atta cks raises the question of whether ETA is attempting to ignite ethnic conflict with the Spanish. The war of attrition changed dramatically in 1992. Starting that year, ETA began to emphasize nonviolent politics in what many have described as a nationalist front strategy. 69 ETA had planned to exert maximum pressu re on the state that year, with Spain hosting both the Olympic Games in Barcelona and the Universal Exhibition in Seville. But in March a number of important activists were arrested in Bidart, southern France, including all of the longest serving leadership. The organi zation nearly collapsed. Its level of violence fell sharply, and never recovered (Sanchez-Cuenca 2009, 6). As a result of this setback ETA was forced to concede that success was impossible through violence alone. In 1993 ETA began seeking ties with moderate nationalist parties with the intention of cr eating a nationalist front (Sanchez-Cuenca 2009, 6). By killing non-nationalists, ETA hoped to further po larize the Basque Country around the nationalist issue, and to force moderate nationalists to negotiate a consensus 69 It is debatable whether this was a genuinely different strategy, or merely the old one under new constraints (Beck 2005 185; Sanchez-Cuenca 2009, 6). 154

PAGE 161

with radicals. Shifting attacks towards state representatives was thought to have a greater impact on public opinion than attacks on the security forces. According to ETA, an increasing reliance on street violence (kale borroka) and a widening of targets would socialize the suffering of Basque fighter s in prison and the Basque population more generally (Beck 2005, 186). The shift toward soft targets clearly reflected the changed security environment more effective polic e work from the Spanish and French police combined with the loss of their leaders fo rced ETA to carry out less risky operations (ibid; Sanchez-Cuenca 2009, 7). ETAs expanding list of legitimate targets came to include local councilors from the PP and PSOE. This strategy was severely divisive for Basque so ciety. The killing of the young PP councilor Miguel Blanco in 1997 sent an unprecedented wave of revulsion through the Basque Country and the rest of Spain (Woodworth 2008, 183). The targeting of councilors al so led to the revival of Spanish nationalism among the PP and PSOE, who came to see Basque nationalism as the root cause of ETAs violence. According to Woodworth, some in both partie s saw ETA as an electoral and ideological bonus that allowed them to revive Spanish nationalism with more fervor than at any time since Franco. Spanish conservatives in pa rticular set out to discredit democratic nationalism in the Basque Country and Catalonia. Some powerful sections of the Spanish media became hysterically anti-Basque, identifying everything from the PNV to Euskera with terrorism. This attack on Basque natio nalism backfired and radicalized moderate nationalists, leading some well-informed observe rs to detect a sinister symbiosis between ETA and the PP (Woodworth 2008, 184). 155

PAGE 162

In September of 1998, ETA and the PNV s ecretly signed the Lizarra pact. In exchange for an indefinite truce, the m oderate nationalists pr omised to fight for independence by politically isolating non-na tionalist parties (Sanchez-Cuenca 2008, 21). ETA declared a ceasefire later that year, but ended it in November of 1999 after disputes with the moderate nationalists, who appare ntly did not perform to ETAs satisfaction. The new cycle of violence that broke out was rather limited and quickly squashed by security forces. With nothing to substitute for the nationalist front strategy, and with a greatly constrained ability to produce violen ce, ETA has arguably entered its terminal phase (Sanchez-Cuenca 2009, 7). The nationalist front strategy created a significant change in ETAs target selection. Although the number of killings declin ed in the 1990s, the deaths of politicians beginning in 1995 had a profound impact on Span ish society, making Spanish citizens far more weary of violence than they had been during the war of attrition. Until 1992, the preferred targets were the army and Civil Gu ard. During the war of attrition (1977-1992), politicians and state officials comprised only 2.6% of all fatalitie s, but from 1992 to 2007 they constituted 21.7% of all fatalities (San chez Cuenca 2009, 6). Shifting targets toward politicians and other noncombatants allowed ETA to maintain its political presence despite the serious setbacks in Bidart. However, this target substitution has been very costly in terms of popular support, tarn ishing ETAs image considerably among both moderate and radical Basque nationa lists (Sanchez-Cuenca 2008, 25-26). In the mid-1990s ETA also began killing those who publicly opposed terrorism. Up to 2006, 19 people were killed in this camp aign. 15 of them were politicians or state officials, 3 were civilians, and 1 was a member of the armed forces. 156

PAGE 163

History of State Violence The Franco regime declared twelve states of exception between 1956 and 1975 (the year of Francos death). Six of these we re directed solely at the Basque provinces, and another five affected all of Spain (C lark 1984, 241). The four states of exception between 1967 and 1969 fell partic ularly hard on the Basque Country (Shabad and Ramo 1992, 428). These states of exception were not fa r removed from pure martial law: police brutality was both protected and encouraged, confessions extracted with torture were common, and people were imprisoned and to rtured for weeks on end without the appearance of any charges, attorneys or j udges (240). An Amnesty International mission on the first state of exception of 1975 found clear evidence that at leas t 295 were tortured. It also found that torture wa s not only used to obtain information and confessions: even when detainees had no knowledge to reveal torture was employed to intimidate the Basque population into submission and to fr ighten them away from the nationalist cause. 70 States of exception also created an at mosphere of impunity for extralegal violence by right-wing vigilante groups, who attacked both opposition leaders and separatists alike (ibid). 71 According to one estimate, approximately 8,500 were directly affected during these states of exception, eith er through arrest, impris onment, torture, or flight into exile (ibid). Within a year of Francos death, the legal structure of his domestic security policy was almost completely dismantled. Unfortunately, Spain only lived under a regime of law and civil rights for about a year and a half (Clark 1984, 248). 70 The techniques of torture included severe and systematic beating, burning with cigarettes, near drowning by being submerged upside-down in water, forced sleeplessness, mock executions, sexual threats, and other forms of psychological torture (Clark 1984, 244). 71 By one account, right-wing groups in the Basque provinces committed eighty-five attacks during the 18 months after the state of exception declared in 1975. Two of these att acks resulted in deaths (Clark 1984, 241). 157

PAGE 164

According to Woodworth, The security forc es used death squads and torture with impunity under the conservative administrations of the transition between 1977 and 1982 (2008, 178). In 1978 the Suarez administration responded to escalating terrorist violence by introducing the democratic peri ods first counterterrorism le gislation. The laws passed in this early democratic period revived many measures that were dismantled immediately after the death of Franco (Clark 1084, 247). Constitutional guarantees against preventive detention and warrantless searches were suspended, and the much hated Civil Guard (responsible for torture under Franco) was given expanded powers to detain and arrest anyone suspected of involvement with terrorism As in the Franco years, an extremely expansive definition of terrorism was employed. The 1981 Law for the Defense of the Constitution defined terrorism as embracing any attack on the integrity of the Spanish nation or as any effort to secure independence of any part of its territory, even if non violent (Encarnacion 2007, 956). Basque politicians of all persuasions strongly condemned these laws. Three prominent Basque leaders decried the passi ng of the special 1978 Decree-Law, arguing that it marked the thirteenth time since 1956 th at the Basque Country was under a state of exception (Clark 1984, 248). These claims were far from unfounded an Amnesty International mission in 1979 confirmed that torture was continuing under the democratic regime, which by now was looking increasing ly unable to rein in the military. For instance, a special parliamentary commission cr eated in 1979 to inve stigate charges of torture was denied access to several Civil Guard and police installations on the grounds that they were the property of th e Ministry of Interior (262). 158

PAGE 165

The continuing autonomy of the military in matters of internal security was an immediate legacy of Spains pacted transi tion. Contrary to the popular notion that the Spanish transition was brilliantly negotiated between former Francoists and democratic parties, it is equally arguable that the constitution was a failed compromise heavily conditioned by the threat of military in tervention (Woodworth 2008, 117). Negotiation prevented the kind of violence and instab ility which normally plague democratic transitions, but also allowed for some parts of the state apparatus to hang on to their authoritarian vices (Encarnacion 2007, 966-7). The militarys continuing autonomy was most dramatically illustrated by the faile d coup in 1981. Of even more importance, however, was the launching of the extra-le gal Dirty War against ETA by Prime Minister Gonzalez and carried out by the Antiterrorist Liberation Group ( Grupo Antiterrorista de Liberacion or GAL). The GAL operated from 1983 to 1987. It killed a total of 27 people, more than a third of whom had no re lation with ETA whatsoever (BBC 1998). Judicial inquiries did not begin to unearth GALs connections to the top of Spains government until 1994. The Dirty War was a huge propaganda gift to ETA, as was the governments increasingly blatant attempts to obstruct media and judicial investigation into GALs government connections. According to Patxo Un zueta, a former ETA leader and current senior commentator for El Pais the GAL was both an ethical and practical disaster: the GAL were a destabilizing factor in the de mocratic system and a catalyst for a new flow of members to ETA, prolonging the prob lem for at least a generation, (Woodworth 2008, 179). 159

PAGE 166

Aside from public pressure to handle terrorism, the Gonzalez government launched the Dirty War for two reasons. First, ETA was killing more generals than Spain had lost in any conventional war, and each death made another coup more likely (Woodworth 2008, 179). In this way the Dirty Wa r acted as something of a safety valve for the anger of right-wing military officers. The second motivation was the French governments refusal to take action agains t well-established ETA sanctuaries on their side of the border, where exiles planned many of their attacks on Sp ain. The core aim of the GAL, according to Woodworth, was to r upture ETAs impunity in France, and to cause such mayhem that French public opinion would dema nd extradition and imprisonment for ETA suspects (ibid). As far as French cooperation was concerned, the strategy was clearly a success. Beginning in 1986, France rapidly incr eased its extradition of terrorists, and cooperation between the two countries has been increasing ever since (Sanchez-Cuenca 2008, 12-13). However, the de mocratic states response to ETA was partly responsible for the Basques persistent support for violence af ter the transition to democracy (Woodworth 2008, 178). The GALs attacks remain the most palpable example of government excess, allowing Basque s to continue seeing themselves at war with a hostile Spanish state. Data Analysis: The Impact of State and Sectarian Violence on Target Selection ETA and all other Basque nationalist or ganizations killed 834 people including 297 civilians between 1960 and 2006 (Sanch ez-Cuenca 2008, 23). Overall, civilians made up 35.6% of all deaths in the conflict. When we in clude politicians and public 160

PAGE 167

officials (49) the percentage rises to 41.3% Geographically, 67% ( 561) of the killings took place within the Basque Country. Outside of the Bas que Country, Madrid is the region with the highest death toll (121), follo wed by Catalonia (54) (ibid). Most of the victims (64%) were killed by shootings, wh ich are generally more selective than bombings and produce fewer collateral deat hs. ETAs 596 lethal attacks produced 834 fatalities, yielding 1.4 average deat hs per lethal attack (24). There is currently no publicly available data that breaks down state-inflicted deaths in the Basque Country in months or even years. The Spanish case therefore does not allow for a micro-level i nvestigation of the relation be tween the level of violence inflicted on a terrorist-supporti ng community (measured in fata lities) and changes in the terrorist groups target selection. Some of the data a ggregated by year is also contradictory. Using secondary sources, de la Calle and Sanchez-Cuenca report that, from 1968 to 2003, fewer than 100 people died as a result of police brutality or anti-ETA paramilitaries (2006, 14). However, in a very detailed dataset later constructed by Sanchez-Cuenca and Aguilar for the period 1975 1982, a tota l of 214 fatalities were attributed to the state, and 76 to fascist terrorist groups (Aguilar and Sanchez-Cuenca 2009, 9). Although the identity of victims attri buted to the state was not presented, 38 of the deaths attributed to fascist groups were directed against people in the Basque Country and southern France (15). Based on the Franco regimes obsession with Basque terrorism, combined with the established tendency of stat es to use more deadly force against leftwing rather than right-wing dissidents, we can re asonably expect that st ate killings in this period were mainly directed at either nationalist or le ftist groups (White 1995, 333). Finally, Shabad and Ramo (1995) present a ye arly breakdown of fatalities inflicted by 161

PAGE 168

extreme-right wing groups in the Basque Country from 1968 to 1991. This data shows a total of 74 fatalities. Combined with the 27 GAL killings, this yields 101 deaths, and thus violates the claim by de la Calle and Sanchez-Cuenca that fewer than 100 people were killed by police mistreatment and the extreme right. Furthermore, this total does not includ e state violence during the transition, which undoubtedly targeted ETA supporters. There are two possibilities: either the new dataset compiled by Sanchez-Cuenca and Augilar c overing the transition had much better coverage than those before it thus invalidating the 100 lim it cited in secondary sources or Shabad and Ramos figures are inflate d. Despite this ambiguity, Shabad and Ramos data must be used by necessity, as it is th e only information available that breaks down the deaths inflicted by extreme right groups into individual years. Due to these data limitations our resear ch hypotheses will only be tested over a very specific time period: from 1975 to 1987. Thanks to a very detailed database on ETAs killings constructed by de la Calle a nd Sanchez-Cuenca, we are able to identify changes in ETAs targeting practices in rela tion to fatalities suffered by Basques. We will do this by using both the GAL killings and the extreme right killings presented by Shabad and Ramo. We will be looking for changes in the percentage of civilian casualties inflicted by ETA, as well as changes in the se lectivity of its violence, measured by its collateral and indisc riminate killings. Table 5.1 presents a breakdown of attacks from the extreme right and GAL along with calculations of the pe rcentage of civilians killed by ETA for the years 1975 1989. 162

PAGE 169

Table 5.1: Basque Country Deaths Inflicted by the Extreme Right and GAL, 1975-1989 Percent civilians (ETA) Killed by extreme right Killed by GAL 1975 26.7 % 1975 3 1976 50 % 1976 6 1977 30 % 1977 8 1978 35 % 1978 22 1979 41.3 % 1979 29 1980 41.7 % 1980 4 1981 34.4 % 1981 1 1982 30.8 % 1983 50 % 1983 4 1984 26.4 % 1984 9 1985 32.4 % 1985 11 1986 17.1 % 1986 2 1987 68 % 1987 1 1988 36.8 % 1989 27.8 % 1989 1 Sources: Shabad and Ramo 1995, 441; Woodworth 2001; de la Calle and Sanchez-Cuenca 2007. These numbers do not suggest any stable relationship between the number killed by GAL or the extreme right and ETAs willingness to target civilians. Because the GAL killings were spread so thin (four occu rred in September 1985 the highest number of any month), there are also no noticeable effect s on target selection for high-intensity months (this relationship was tested by examining the four months with the highest intensity of GAL killings; none showed even the potential for a target shift towards civilians see appendix). 72 Similar results are reported by Sanchez-Cuenca in his analysis of temporal varia tion in lethal ETA violence. The author found that the GAL coefficient was not significant, but the extreme right variable covering 1978-82 was (2009, 13-14). According to his analysis, extr eme-right violence was both a reaction to and a catalyst for ETA violence (ibid). Thes e results both support and contradict the 72 However, between 1978 and 1994 ETA killed 12 civilians for their alleged involvement in the Dirty War. 163

PAGE 170

(non)relationships abov e between the yearly fatalities of ETA, the extreme-right, and the GAL. The discrepancy would lik ely disappear with both more detailed data at the monthly level and with a more soph isticated statistical approach. Can we detect any effect of extreme-ri ght killings on ETAs target selection? Among the 20 killed in 1978 were 13 alleged in formers, 2 accused of being involved in the Dirty War, 6 accused of having a right-wi ng ideology, 2 politicians accused of being informants, and 1 member of ETA killed for being an informant (some of these accusations were coextensive). This last fatali ty is notable given that ETA has only killed 4 of its own members on allegations of inform ing during its entire history. The year 1979 stands out because ETA killed 7 people indisc riminately, in addition to 12 civilians who were accused of being informants. The indisc riminate killings are unusual given that ETA had not carried out any in the previous four years, and has rarely carried them out since (the remaining 39 attacks were all co mmitted within 5 non-contiguous years which display no clear pattern). Th e fact that no such reaction occurred in 1978 seems to invalidate the relationship. However, a full breakdown of these casualties could reveal unique effects for when ETA members are ki lled instead of innocent bystanders. Both years saw unusually high levels of civilian s killed on the allegation that they were informers. Between 1978 and 1979 there was a sl ight shift away from individual attacks and towards collateral killing. There was a str ong turn towards indiscriminate attacks. Before these two peak years no change occurred in the selectivity of ETA violence. If we combine all types of civilian fatalities (indiscriminate, collateral and individual) and subtract thos e who were killed for security reasons (informants, feuds, 164

PAGE 171

extortion, etc), we get the following five years with the highest levels of civilian deaths (see Table 5.2). 73 Table 5.2: Peak Years for ETAs Civilian Fatalities (Adjusted for Security-Related Deaths), 1969 2001 The fact that the GAL killings coinci ded with increased cooperation between French and Spanish police should give us paus e before attributing a nnual relationships to individual GAL killings. In the Northern Irel and case, changes in target selection first occurred during the high-intensity months in which Catholic civilians were killed, and then only occasionally displayed weakened effects beyond that month. The timing of effects should be the same here preferen ces for violence among ETA supporters should begin to manifest within the sa me month and not later. If this assumption is justified, it is clear that our research h ypotheses are not supported by the available data on Spain. Beyond the GAL killings, the repression hypothesi s cannot explain the move towards soft targets that accompanied the policy of socia lizing the suffering in the early 1990s, nor can it explain the numerous spikes in the percentage of civilian casualties caused by ETA over its lifespan (see figure 5.2 below). What does explain this variation? The movement toward indiscriminate attacks in 1987 could li kely have been a result of a constrained security environment. French authoritie s clamped down hard on ETA after the GAL 73 For instance, 1980 (35 civilians) included 31 killed for secur ity reasons (including 23 alleged informers). 1983 (20 civilians) included 12 killed for security r easons (6 alleged informers). Informants were dropped on the assumption that allegations were accurate. This was done so that the attacks under investigation can best reflect the organizations decision to target civilians as a strategic choice. 165

PAGE 172

killings, extraditing 5 suspect s to Spain and deporting an additional 23 from France in 1984 alone. From 1987 onwards France extradited ETA suspects to Spain without requiring the production of an extradition warrant (Sanchez-Cuenca 2008, 12). Indiscriminate killings in 1974 can be explained by the fact that the Basque Country was then under a state of emergency, which must ha ve seriously constrained ETAs ability to carry out discriminate attacks. The killings of 1995 can be explained with regard to the strategy of socializing the suffering a glorified way to frame the fact that ETA was too weak, too desperate, and too afraid to continue focusing on the security forces. Finally, the high level of indiscriminate ki lling in 1979 was probabl y a reflection of the numerous extreme-right killings that occurred that year. Figure 5.2: Percentage of Civilian Ca sualties Caused by ETA Over Time Civilian Fatalities, 1960-2006 0 20 40 60 80 100 1201 9 6 0 196 9 197 3 1 975 1 977 1 9 7 9 1 9 8 1 198 3 198 5 1987 1 989 1 991 1 9 9 3 1 9 9 5 199 7 200 0 2002 2 006 Total fatalities Percentage of civilian casualties Source: de la Calle and Sanchez-Cuenca (2007) The year 1981 saw the highest number of arrests out of the years 1973 to 2003. According to the Home Office, 714 people were arrested. Compared to an annual average 166

PAGE 173

of 222 (Sanchez-Cuenca 2009, 22), 1981 represented almost 9% of all arrests in this period. In terms of the fataliti es it could inflict, it seems that ETA did not recover from these arrests. In fact, the drop in deaths in 1981 is comparable in size only to that which occurred in 1992 after the captu re of ETA leadership at Bi dart (ibid). Surprisingly, despite the fact that the nu mber of annual fatalities dropped from 96 in 1980 to 32 in 1981, the percentage of civilians killed remained relatively constant until 1983, when it rose to 50% before dropping back below the 36% average for the following three years. Despite a weakened organizational capacity, ETA was still able to strike hard targets. ETA detonated its first car bomb in 1982, and began to rely on them heavily in 1984, and again in 1986. ETA decided to us e car bombs from 1986 onward because the security environment had become much less favorable. French cooperation with Spanish security forces, combined with the Dirty Wa r, had made the lives of ETA members much less secure, and led to the arre st of most of their leadership. Car bombs carried much less risk than shooting at the security forces, and killed many more people. In 1987, ETA used eight car bombs to kill 40 people (81% of all fatalities for that year). ETAs reliance on car bombs coincides perfectly with the incr easing percentage of civilian casualties beginning in 1986. One suggested explanation for ETAs willingn ess to target civ ilians is that its members are risk-averse and have poor comm itment to their ideology. Thus, Kalyvas and Sanchez-Cuenca, argue that the communitys living conditions are a more important determinant of the motivation to commit extreme acts than any single individuals experience with political repression or economic deprivation (2005, 228). The Basque Country is a wealthy region with a per capita income which has consistently stayed above 167

PAGE 174

the Spanish average. ETA has tried and faile d to imitate Irish hunger strikes. ETA has also made many attempts to minimize the risk s of being caught or killed in action, even when that meant resorting to tactics that were highly unpopul ar with their community of support (ibid). In 1991 ETA placed a bomb in the car of a policeman that killed his 2-year-old son. This attack was followed by a huge outcry from Spanish society, and even from some imprisoned ETA members. The architect s of the attack quickly defended their decision, arguing that We shouldnt unnecessarily risk the lives of our fighters which are a hundred times more valuable than the life of a txakurras [dogs] son (Kalyvas and Sanchez-Cuenca 2005, 229). Instances such as these have led one sc holar to conclude that members of ETA have never shown suicidal behaviour when preparing their attacks. The rule has been to act under maximum personal security and, when in doubt, not to act (ibid). While the IRA has killed 139 of its own members through accidental bomb explosions, ETA has killed only 7. All were intentional. Woodworth provides an amusing anecdote from a Sinn Fein member visiting the Basque Country in 2001. He witnessed a large street battle in Bilbao between Batasuna demonstrators and the Basque police: He commented to one Batasuna leader that, had this been Belfast and the confrontation equally ferocious, every car and every shop on the street would have been burning by this stage in the riot. The Batasuna leader had th e grace to be a little embarrassed. Well, he replied hesitantly, I suppose that it is different here, because these are our cars, these are our shops (Woodworth 2008, 175). There is some truth to these observati ons. At the risk of oversimplifying they encapsulate the different grievances motiva ting the Northern Irish and the Basques. While the Irish were motivated by economic disc rimination, the denial of civil rights, and by the historical fact that a single nation was split in half by Br itish partitioning, the 168

PAGE 175

Basques are engaged in a battle over their own identity, are economi cally privileged, and enjoy a high degree of autonomy. However, th ese observations lend th emselves more to the individual motivations sustaining ETA viol ence than to the organi zations selection of targets. Changes in ETAs target selec tion are better explai ned by organizational weakness and the perceived failu re of past strategies. For most of its history, ETA has avoide d unnecessary civilian casualties. Almost half of its lethal attacks on civilians (excluding politicians an d state officials) have been related to internal security. Of its 296 civi lian casualties, 44% (131) were targeted for some individual behavior. For instance, 83 were accused of being informants, 35 were accused of being involved in drug trafficking (6 of whom were also accused of being informers), 33 were accused of holding an extreme right-wing ideology (16 were also accused of being informers), and 5 were target ed for their involvement in the construction of the nuclear power plant at Lemoniz. 74 An additional nine were killed because they refused to pay ETAs revolutionary tax. To the extent that these accusations were accurate, at least some of this violence woul d have been necessary for the organizations continued existence. 75 Apart from the 83 alleged informers, prestige killings such as those targeting the extreme right and thos e involved at Lemoniz are undertaken to mobilize popular support. Such security-related violence is designed to control the local population, and is necessary to create a str ong organization that can deter defection and denunciation (de la Calle a nd Sanchez-Cuenca 2006, 4). 74 Many of these accusations were coextensive. 75 It is debatable that control requires violence to take the form of killing. Noting the higher level of control killing in ETA relative to the IRA, de la Calle and Sanchez-Cuenca argue that ETAs more superficial relationship to the Basque community may force it to resort to contro l killings more often because its reputation has not been firmly established. On the other hand, the IRA can resort to lesser punishments such as beating and kneecapping (de la Calle and Sanchez-Cuenca 2006, 16). 169

PAGE 176

Table 5.3 presents a breakdown of the rema ining casualties who were not targeted on the basis of their behavior. Table 5.3: Selectivity of ETAs Non-security Civilian Killings 296 total civilian fatalities -131 individually motivated (security-related) -----------------------46% (76) collateral 28% (46) indiscriminate 10% (17) intra-group killing 9% (15) mistakes 6% (11) state killing 76 In addition to these deaths, ETA has also killed 48 politicians or state officials. Most of these victims we re non-nationalist Basque pol iticians killed since 1995. However, ETA has also killed non-elected mayors from the Franco period, and ETApm killed elected politicians in the early year s of democracy during a campaign against rightwing, non-nationalist Basque politicians. ETA has tried to kill the king of Spain at least five times. Its latest attempt in 1995 nearly succeeded (Kalyvas and Sanchez-Cuenca 2005, 212). ETA also tried to kill the leader of the Popular Party (PP) Jose Maria Aznar in 1995. He became Prime Minister the following year. These were clearly importa nt targets. Nevertheless, ETA rejected the use of a suicide bombing to assassinate the king, despite the fact that a volunteer inside the organization was willing to execute the attack (ibid). In a letter intercepted by the police, one of the leaders explained that in principle we do not agree with the idea of a militant blowing himself up in a car. Yet if the militant is willing to run a high risk, there 76 In this context state killing most likely refers to lo wer level government workers as opposed to politicians and state officials. The Victims of ETA dataset normally uses the state killing variable to describe attacks on politicians, state officials, and members of the secu rity forces. However, the codebook does not state the meaning of this precise combination (civilian deaths motivated by state killing). 170

PAGE 177

could be a chance to carry out an action follo wing a funeral (212). The leader does not elaborate on why the organization rejects suicide missions, but he does distinguish between sending a man to certain death and sending him on a high-risk mission. Suicide missions are a qualitatively uni que form of violence, and represent the most extreme form of violence available to w eak actors. On average, they kill four times as many people as other terrorist attack s (Kalyvas and Sanchez-Cuenca 2005, 216). In Northern Ireland the IRA briefly engaged in forced suicide missions, but quickly abandoned them due to the revulsion they cau sed in the Catholic community. Given its similar support base, ETA may have refrai ned from using this volunteer precisely because of the negative publicity the IRA garn ered for its use of forced suicide missions. Public Support for Violence Electoral Analysis Using every survey he could locate, Sa nchez-Cuenca calculated a mean value of support for ETA. From this mean he found that ETA had significant su pport in the 1980s, and that support peaked in 1987 (Sanchez-Cuenca 2007, 302). C onsistent with the survey evidence, this was the year in which Herri Batasuna reached its electoral peak, winning almost 20% of votes in the Basque Country. Just a few weeks later, however, ETA set off a car bomb in Barcelona that killed 21 civilians. A steady decline in support set in until 1997. ETAs strategy of socia lizing the suffering in the early 1990s probably contributed to this decline, as did the killing of non-nationa list politicians beginning in the mid 1990s. Although there was no electi on between 1994 and 1998 to confirm it, survey evidence shows that ETAs popularity reached a low point in 1997. This decline 171

PAGE 178

was reversed in 1998 when ETA announced a truce. In the 1998 Basque regional elections held during the truce, Batasuna gained 17.7% of th e vote the biggest increase in the partys history. The peak of support in 1999 marked a full year in which ETA had not killed anyone. When ETA ended the truc e later that year, it s popularity once again declined, and its electoral sh are dropped to a mere 10% in the 2001 regional elections (Sanchez-Cuenca 2007, 302). Table 5.4: Electoral Support for ETAs Politic al Wing in the Basque Legislature Were any of these increases in civilian targeting associated with poor electoral performance? HB lost votes in 3 el ections: 1990, 1994, and 2001. The most dramatic change occurred in 2001, when Euskal Herritarrok lost 80,862 votes relatave to the previous election. The other dramatic dec lines were in 1994 and 1990. In each of these elections the percentage of ci vilian casualties was increasing. In 1990 the number of total casualties was also on an upswing. Interestin gly, the two years with the strongest 172

PAGE 179

electoral trend were also the ones which took place r oughly in the middle of the upswing, giving spectators time to acknowledge the current trends in violence. Conclusion ETA has largely avoided civilian casualties in its attacks that aim to influence the state. However, many civilian casualties result from its normal operation as a violent underground group, notably in its killing of in formers. ETA has largely been restrained from using more extreme violence by the pref erences of its supporters. This can be seen by the reaction of Batasuna voters to the start and end of the ceasefire. Repression fed the organization in its infa ncy, but the organization was content to conserve that support rather th an risk it with a wi der conflict among civ ilians. Extra-legal violence against Basque nationalism has waned considerably since the GAL stopped killing in 1987, but the memory of repressi on keeps radical Basque nationalism alive. ETA suffered a major organizational blow in 1992. Its response was to seek a ceasefire, and focus on institutional politics while conti nuing to use violence in the only way that was still possible: by shifting towards softer and less risky targets politicians, state officials, and civilians who denounce ETA. Th e decision to shift towards soft targets alienated soft supporters but was probably n ecessary to maintain the inner core. Starting in the latter half of the 1990s, ETAs collateral killings have made up an increasingly large proportion of its attacks on civilians not because they had increased, but because every other type of attack decreased to nearly zero. In the end, ETAs reputation for hardline nationalism has helped it tremendously in the polls. When they have been willing to 173

PAGE 180

give up violence and enter institutional politics, the voters have rewarded them generously. 174

PAGE 181

Chapter 6 Conclusion: Comparative Insights on the Target Selection Strategies of Terrorist Groups Ideology lays the foundation for a groups strategy and target selection. Within any given ideology, however, the use of violen ce is constrained by th e states security apparatus and the preferences of supporters. The desire to mainta in popular support and to mobilize the ethnic community makes nationa list terrorist groups uniquely inclined to tailor the use of violence to what supporters desire. They are more discriminate than right-wing groups, and perhaps less so than le ftist groups. When religion is mixed with nationalism it tends to produce more indiscriminate violence. This study has examined two cases of nationalist terrorism by focusing on two hegemonic groups: the IRA and ETA. The groups had the same goal, employed very similar tactics and went through very similar evolutionary stages. Their campaigns began in the midst of tremendous social mob ilization, to which the state responded with indiscriminate and ultimately ineffective repression. Provoking such repression was a key goal of ETAs first strategic phase. The IR As political strategy leading up to the 175

PAGE 182

Troubles was based on a similar logic, but, lik e their models in the African-American civil rights movement, they chose nonviolent methods. Repre ssion fed both conflicts in their infancy, and also helped to prolong a nd escalate them. In Northern Ireland, there was also a long history of economic exploi tation. By comparison, the Basque Country has historically had a per capita GDP well above the Spanish average (Sanchez-Cuenca 2008, 3; Clark 1984, 16). Instead, the Basque Country suffered severe physical and cultural repression under Franco, which fostered a militant nationalism that found trust in the Spanish state impossible. The central difference between the two cas es, in terms of conflict dynamics, was that the IRA faced two enemies: Loyalist (Protestant) terrorist groups and the British state (which effectively controlled Northern Ireland). As a result the Catholic population was subjected to a significant amount of violen ce, and came to be highly supportive of armed struggle. The closest equivale nt in the Basque Country we re the right-wing terrorist groups which Franco unleashed on Basque nati onalists during his states of emergency. The repression was similar in that it empl oyed indiscriminate targeting, focusing on nonviolent activists and on Basque citizens in general with the intent of intimidating them against supporting nationalists. The key difference was that most of the repression in Spain was non-lethal, and, at least during the transition, the use of torture in Spain was not as widespread as the use of tortur e in Northern Ireland. The overall scale of repression in Spain, combining both state and right-wing groups, was not nearly as great as it was in Northern Ireland. Both organizations began with an information advantage over their state adversaries. Eventually, however, the states grew more knowledgeable about their habits 176

PAGE 183

and more effective in either killing or arre sting their members. Th e hegemonic terrorist organizations in both countries hit an in termediate stage where violence began to stabilize. This point occurred in 1981 for ETA, and 1977 for the IRA. Fatalities were stabilized largely because of th e security constraint. As the ri sk of participation increased and repression subsided the pool of availa ble recruits shrank. Both the IRA and ETA were able to sustain their attrition strategies during th e 1970s and 1980s, but attrition became unsustainable for both organizations during the 1990s (de la Calle and SanchezCuenca 2006, 20). As security forces becam e more effective, both organizations increasingly turned towards politics and fo rmed alliances with like-minded political parties, hoping to achieve their goals through the creation of a nat ionalist front. The Evolution of Targeting Strategies in Nationalist Terrorist Groups A more difficult security environment caused the number of annual killings inflicted by both to stabilize, and then declin e. As the security environment grew more and more constrained, both organizations were unable to maintain the same level of violence against the states security for ces, which caused their proportion of civilian casualties to grow (see table 6.1). For ETA, this shift was justified as a st rategy of socializing the suffering, and was directed primarily at non-nationalist Basque politicians as part of the nationalist front strategy. The IRA appears to have been much more determined to continue striking hard targets, to the extent that they even expe rimented with non-consensual suicide bombings in the 1990s. However, they also shifted towa rds soft targets, killing businessmen and 177

PAGE 184

their employees who carry out building work or provide services for the security forces. The other benefit of this tactic was that it forced army engineers to complete the work, thus creating a soft military target which makes for better propaganda. Table 6.1: Percentage of Civilian Deaths in Relation to All Deaths Caused by Republican Groups and ETA Source: Sutton 1994 If the experience of repression allows more room for strategic maneuver among a terrorist groups leadership, and shifting towa rds soft targets in times of organizational weakness alienates soft supporters, what expl ains the mismatch? We can think of two sides to this puzzle. On the one hand we can ask why attacks become more indiscriminate over time even when that alienates suppor ters, and on the other hand we can ask why groups do not take advantage of higher suppor t when they have it in order to justify attacks on civilian targets. One potential reason why attacks ma y still turn indiscriminate is that the security environment becomes so constrained that the only way to carry out 178

PAGE 185

attacks is to do so in an very imprecise way. In this situation, the organization must sacrifice soft support to maintain the inne r core. A similar and perhaps more likely possibility would be where a terrorist organi zation becomes extremely risk-averse, to the extent that it passes off all the normal risks i nherent to political violence onto bystanders. Finally, Bell suggests that targ et selection is not a process that is heavily scrutinized. Most revolutionary organizations spend re latively little time on target selection. A target may be rationalized as a fact of a special campaign but the first operation was most often selected because a vulnerability was discoveredthen, if need be, explanation is offered (Bell 2000, 275). The other end of the paradox is why terrori st groups do not attempt to shift targets when they are still popular enough to pull it off. The first possibility is that they simply are not thinking about alternativ e strategies. According to J Bo wyer Bell, an expert on the IRA who has become close friends with many members, life as a militant is far too busy for innovation. If the IRA wanted to destroy information on a computer, Bell tells us, it would sooner bomb it then turn it on and do things manually. No underground has either the time or inclination for much novelty. runn ing in order to stand still, running harder each month, leaves little time for re flection and novelty (Bell 2000, 280). ETAs supporters have a generally lowe r propensity to accept indiscriminate killing. In 1987 ETA planted a car bomb in the basement of a supermarket in Barcelona. Twenty-one people died, making it the bloodiest attack in ETAs history. Months before the attack, Herri Batasuna had won 361,000 votes in the European elections. In the next European election two years later, Herri Ba tasuna lost about 100,000 votes (de la Calle and Sanchez-Cuenca 2006, 15). 179

PAGE 186

The relationship between a terrorist gr oup and its base of support is highly difficult to study, not the least because we lack good proxy variables. 77 The extent to which supporters constrain the leadership is essentially an open ques tion. Terrorist groups must engage in violence to just ify their continued existence an d to win support (as well as to keep pressure on their opponents), but they must also avoid using violence that repels their supporters. On the other hand, terrorist gr oups that take the ini tiative to break into new types of attacks may enhance their st rategic position if they are rewarded by supporters who have radical preferences. As McCormick noted, a terrorist group s principle decision problem is to identify its available room for maneuver, consider how it might increase its room for maneuver down the road, and design and execu te an operational plan that offers the highest expected return (McCormick 2003, 499). Working from this assumption, this study investigated how nationalis t terrorist groups frame the ta rgeting of civilians, and whether they attempt to take advantage of state repression or ethnic violence against their community to target civilians pro-actively, becau se they calculate that it will be effective. Examination of the IRA and ETA revealed that they do not. In fact, the IRA never grew as indiscriminate as ETA because it never suffe red a loss that could parallel the arrest of the entire ETA leadership in 1992. This di stinction suggests a pa rallel to Hultmans finding that rebel groups use i ndiscriminate violence agains t civilians to demonstrate resolve when they are losing battles. The difference between Hultmans claim and the Basque case is that ETA escalated its indi scriminate violence while simultaneously entering peace talks and while its political wing contested elections. 77 For instance, de la Calle and Sanchez-Cuenca (2006) use total fatalities as a proxy measure for support, since killing depends on resources, which in turn depend on popular support. 180

PAGE 187

Nationalist terrorist groups tend to use highly discriminate violence so that they can build a large base of support. Examina tion of the IRA and ETA demonstrates that neither group sought to start a violent conf lict along ethnic lines both made the state their enemy in pursuit of their own home land. However, nationalist terrorism is not always so discriminating. Given a political opportunity structure that rewards sectarian violence, nationalist terrorist groups would certainly use it. De la Calle and Sanchez-Cuenca hypothesi ze that the preferences of supporters determine the selectivity of terrorist violence. As evidence they demonstrate the differences in selectivity between the IR A and ETA, as well as the geographical differences in selectivity within Northern Ireland. The trouble with this hypothesis is that it is too simplistic and leads to an overpredic tion of indiscriminate and ethnic violence. A population that is repeatedly the victim of ethnic violence and that merely develops preferences reactively would undoubtedly come to prefer ethnic warfare. Ethnic warfare has not been the outco me of conflict in Northern Ireland. Instead, we see one population totally devastated by random viol ence while its militant representatives continue targeting the British Army and security forces, incidentally hurting many Protestant civilians along the way. In both cases, indiscriminate violence inflicted on the local ethnic community radicalized the population and made na tionalists both increasingly popular and increasingly militant. An interesting question that arises in these cases, and perhaps the most important question for understanding why civilians do or not become targets of violence, is what allows sectarianism to take hold. What preven ts the Basques from randomly attacking Spaniards, or the Sp aniards from randomly attacking Basques? 181

PAGE 188

Members of the security forces who see themselves as the vanguards of Spanish nationalism have been notorious for their repr ession of Basques, namely through their use of torture. ETAs attacks on the Spanish securi ty forces, particularly the Civil Guard, may potentially be seen as sectaria n given that they have killed only 13 Basque police officers. According to Sanchez-Cuenca, ETA tried to start an ethnic conflic t in the 1990s, but by that point it was significantly weakened. By most accounts ETAs struggle is one of national liberation and is not sectarian. Most commentators describe the conflict in Northern Ireland as sectarian without investigating who is perpetrati ng the sectarian violence. The interesting question here is why the IRA was so much more restra ined than its Loyalist counterparts. 78 One commentator on the conflict is critical of the idea that supporters have ever constrained the IRAs actions. It is to a large extent the public rather than the paramilitaries who ultimately give way. No misconceived or botched operation has yet proved catastrophic for any Republican group (Townshend 1995, 336). This statement is technically true the IRA has a dedicated co re that has not and in all likelihood will not abandon the organization regardless of how indiscriminate it becomes. However, there are supporters who are sensitive to the deploym ent of violence, and their reactions have had powerful effects on the IRA. In the aftermath of Bloody Sunday and the fall of the Stormont parliament the IRA was feeling more vindicated than ever about its violent strategy. However, Republican supporters pushed them into tenuous negotiations with th e British that broke down in just 12 days. In response to the br eakdown of negotiations, the IRA leadership 78 One reason seems to be that the IRA thinks the British want ethnic conflict. One member from Belfast argued that ethnic conflict and the myth of it are both classic tools of British colonialists who wish to maintain control (White 1997, 31). 182

PAGE 189

decided to launch attacks that would demonstr ate the cost of future war (White 1997, 39). Twelve days later, they se nt twenty of their newly crea ted car bombs into Belfast and detonated them in just under an hour, killing 9 (7 civilia ns) and injuring 130. 79 The event came to be known as Bloody Friday. Pictures on television of firemen shoveling human remains into plastic bags compounded the horror (Moloney 2002, 116). Ten days later, the British seized upon hesitation and doubt among the now highly introspective Catholics and reclaimed the no-go areas of Derry and Belfast, an action that would have been unthinkable before Bloody Friday' (117). 80 In a matter of weeks military forts were constructed right on the IRAs doorsteps, seve rely curtailing its freedom of action. The IRAs deadly attacks were dramatically a nd permanently reduced the following year, and its freedom of action would never be as favorable again. The conditions under which the IRA entered into the strongest wave of sectarian conflict in 1975 also merit cl ose attention. Supporters had at least tolerated sectarian reprisals in the past, while the IRA could sti ll point to its violence against British security forces as proof that it was waging a bona-fid e national liberation str uggle. High levels of ethnic conflict in the midst of peace talks with whom most saw as the real enemy, however, may have alienated a good number of soft supporters. Such alienation 79 Many insiders have claimed that the IRA had si mply overestimated the ability of the authorities to respond to warnings. However, the attack fit a clear pattern of sectar ian violence into which the IRA had been pulled. 80 Operation Motorman deployed some 12,000 British troops backed by tanks and bulldozers to reclaim the no-go areas in Derry, Creggan and Andersontown. This represented the most combat units dedicated to a single operation since 1945 even more than during the Suez Crisis of 1956 (Geraghty 2000, 72). Earlier plans to reclaim the no-go areas were shelved due to fear of civilian casualties that might undermine the peace process. After Bloody Friday, however, the mo od had changed significantly. The British Army was even given permission to use rocket launchers agains t IRA. However, the IRA put up no resistance, and only two people were killed one sixteen year-old boy, and one member of the PIRA. Bloody Friday had created the political support to jus tify the operation while simultaneously creating a rift between the PIRA and its supporters. The use of overwhelming force comb ined with popular alienation caused the PIRA to refrain from causing more death among the Catholic co mmunity. In such an unfa vorable political context, blame was much more likely to be placed on the IRA. 183

PAGE 190

interacted with the new push for intelligen ce gathering to produce catastrophic results on the IRAs security environment (Woodwell 2005, 173). The severity of change was well illustrated by the fact that in 1977 the IRA adopt ed a cell structure to avoid infiltration. Furthermore, it is highly unlikely that th e sectarian conflict th at broke out in 1975 was the result of popular demand. A more likely explanation is that IRA militants simply could not sit still. We noted in our analysis of the high-in tensity months for Catholic civilian deaths that reprisals tended not to last longer than a month when retaliation did occur it was tightly correlated with Loyalist attacks. This connection helped prevent the IRAs civilian targeting from creating too much of an outcry from within the nationalist community anything the IRA might do was ea sily dwarfed by the atrocities of the Loyalist paramilitaries and the British security forces. But when the conflict took on an exclusively sectarian tone, many Catholics were likely dumbfounded as to what the armed struggle had become. One can wage a just war against an unjust state, but there is no prestige in two terrorist groups lashi ng out at each others civilian population. Attrition in this context is far less likely than genocide, and particularly for a minority group, this is not a desirable result. The patterns studied here suggest that the radicalizing effect of violence has much to do with how we conceptualize the source of violence and our possible reactions to it. The case studies also show that the support of followers is not simply turned on or off. Defecting supporters lead to a tougher securi ty environment and inte ract with the ongoing work of police forces. Organizations that lose solidarity and turn introspective will often fall in a cascade of arrests. For instance, th e Italian author ities arrested nearly 2,000 Red 184

PAGE 191

Brigades terrorists by 1980, one and a half y ears after the Red Brig ades created a public backlash for killing the former Prime Minister Aldo Moro (Jones and Libicki, 2008, 50). At the three central points where the IRAs violence dropped dramatically, the cause can be traced to its targeting of civi lians and the effect this had on its operating environment. Whether out of misconcei ved strategy or group dynamics, the IRAs military failure was caused by an inability to keep its supporters mobilized. The effects of ETAs indiscriminate vi olence have been even more obvious because its relationship with the community is not as strong. Its popularity went into a permanent decline after a single car bomb in 1987 killed 21 people, all of them civilians. Public opposition to ETA increased, and even led to mass demonstrations against the organization. It was in this context that ETA launched a campa ign against its critics in the mid 1990s. Conclusion This study has examined two hegemonic nationalist terrorist organizations and their satellites. The project sought to explore the extent to which the supporters of terrorist groups constrain their production of violence. More specifically, we wanted to know whether the decision to shift towards so ft targets such as civilians and other noncombatants was the result of rational calculation, organi zational constraints, or emotional reaction. To get at this question we hypothesized that the experience of physical repression in the form of fatalitie s among the terrorist groups base of support would create more radical preferences and le ad to more indiscriminate violence. The 185

PAGE 192

hypothesis worked for a very small range of IRA violence in which the IRA was pulled into sectarian conflict with Loyalist paramilitaries. This shift seems to have occurred because of organizational dynamics. The hypothesis that supporters would become more willing to accept the targeting of civilians, te sted in four different forms, did not work anywhere else for either the IRA or ETA. Instead, organi zational weakness and changes in the security environment were the best explanation for changes in target selection. The constraining effect of constituent support operates quite differently for terrorist groups with different ideologies, as does the decisi on to target civilians. In Northern Ireland we saw that conservative terrorist groups target ed random members of the Catholic community in retaliation for IR A violence. Logistically this could be rationalized by the fact that they do not know where to find members of the IRA. Such attempts at enforcement terror are endemi c to conservative terrorist groups, whose mission is to safeguard the status quo. Clearl y, moral prohibitions against violence can be broken for wildly different reasons. The motiva tions for violence do not lead directly to a strategy, but the structural cons traints facing dissidents tend to create recurring patterns of revolutionary and anti-revo lutionary violence. However, ideology is also subject to change. The IRA was drawn into sectarian conflict for two full years despite an ideological commitment to unify against British occupation. Interestingly, the periods where the IRA engaged in the most sectarian violence were immediately followed by steep declines in its operational capacity. This pattern indicates that targeting civ ilians exacerbates collective action problems for terrorist groups, even when it occurs in periods where it has a relatively greater chance of being accepted by constituents. Minority ethnic populat ions, for example, may be unwilling to 186

PAGE 193

support pure categorical terrorism against th e majority group. The tendency for terrorist groups to target civilians depends not only on their ideology, but also on features of the political opportunity structure. Some of these features incl ude the place of supporters in the population distribution, the likelihood of stat e complicity (an aspect of the security environment), varying levels of repressi on experienced by their supporting group, and the availability of democratic alternatives (w hich tends to diminish support for violence). Ideology may produce some systematic eff ects on the groups po litical opportunity structure (for instance, the probability of state complicity is much higher for right-wing groups than it is for left-wing groups), but these are limited and often overtaken by context-specific factors of the political opportunity structure. Finally, within groups, organizational dynamics have proven to be th e best predictor of whether civilians are chosen as the targets of violence. When this decision is made, however, terrorist groups tend to dissolve quite rapidly. 187

PAGE 194

Appendix Northern Ireland: Top Five Months With the Highest Number of Catholic Civilian Fatalities Caused by British Security Forces Month Catholic civilian fatalities Protestant civilian fatalities August 1971* 18, all by British security 0 October 1971 6, all by Br itish security 1 January 1972 13, all by British security 1 July 1972* 27, 10 by British security 22 Feb 1973 12, 5 by British security 1 *These months were co-extensive with the peak months of Catholic civilian deaths caused by Loyalists and British security combined. Spain: Top Four Months With the Highe st Number of GAL killings: Sept 25 th 1985 4 fatalities ETA killed only twice in Septem ber: one civilian on the 9 th and one Spanish police officers on the 14 th The civilian was a collateral killing. In October ETA killed 4 people: 2 civil guard s and two members of the armed forces. In December ETA killed 3 civil guards. Feb 17 th 1986 2 fatalities ETA only killed once in February, and the at tack took place before the GAL killings. ETA did not kill again until June. June 14 th 1985 2 fatalities ETA killed 7 people in June. Before the GAL k illings, these attacks killed 3 members of the security forces and one civilian bystander. After the attack ETA killed one civil guard and two civilians one on drug allegations and the other an alleged informer. The following month ETA killed only members of the security forces. October 16 th 1983 2 fatalities ETA killed 8 people this month. Before the GAL killings its targets were 1 civilian informer, 1 collateral victim, and two civil gua rds. After the attack its targets were two civilian informers and a member of the armed forces. 188

PAGE 195

Bibliography Unless otherwise stated, all data on ETA deat hs is taken from Luis de la Calle and Ignacio Sanchez-Cuencas Victims of ETA Dataset (2007). Unless otherwise stated, all data on deaths in Northern Ireland are taken from Malcolm Suttons Index of Deaths in Northern Ireland (1994). Abrahms, Max. 2008. What Terrorists Really Want. International Security 32, No. 4 (Spring): 78-105 Abrahms, Max. 2006. Why Terrorism Does Not Work. International Security 31, No. 2 (Fall): 42-78. Aguilar, Paloma and Ignacio Snchez-Cuen ca. 2009. Terrorist Violence and Popular Mobilization: The case of the Spanish Tr ansition to Democracy. Manuscript. Juan March Institute. (accessed 3/12/09). Asal, Victor and R. Karl Rethemeyer. 2008. The Nature of the Beast: Organizational Structures and the Lethality of Terrorist Attacks. The Journal of Politics 70, No. 2 (April): 437-449. Bandura, Albert. 1990. Mechanisms of Moral Disengagement in Terrorism. In Origins of Terrorism: Psychologies, Ideolo gies, Theologies, States of Mind ed. Walter Reich, pp. 161-191. Cambridge: Cambridge University Press. BBC. 2003. Batasuna banned permanently. March 17, 2003 (accessed 6/20/09). Beck, Jan Mansvelt. 2005. Territory and Terror: Conflicting Nationalisms in the Basque Country Routledge. Bell, J. Bowyer. 2000. The IRA: 1968 2000, Analysis of a Secret Army. Frank Cass Publishers Berry, Nicholas O. 1987. Theories on the Efficacy of Terrorism. Conflict Quarterly (Winter): 7-20. Brockett, Charles D. 1993. A Protest-Cycle Resolution of the Repression/PopularProtest Paradox. Social Science History 17, no. 3 (Autumn), 457-484 189

PAGE 196

Bruce, Steve. 1997. Victim Selection in Ethni c Violence: Motives a nd Attitudes in Irish Republicanism. Terrorism and Political Violence 9, no. 1 (Spring 1997), pp. 5671. Published by Frank Cass, London. Brush, Stephen G. 1996. Dynamics of Theory Ch ange in the Social Sciences: Relative Deprivation and Collective Violence. Journal of Conflict Resolution 40, pp. 523545. Clark, Robert P. 1984. The Basque Insurgents The University of Wisconsin Press. Coogan, Tim Pat. 1994. The IRA: A History Roberts Rinehart Publishers Crenshaw, Martha. 2001. Theories of Terrorism. In Inside Terrorist Organizations Ed. David C. Rapoport, pp. 13-31. Frank Cass & Co. Ltd. Crenshaw, Martha. 1995. Terrorism in Context The Pennsylvania State University Press. University Park, Pennsylvania. Crenshaw, Martha. 1991. How Terrorism Declines. In Terrorism Research and Public Policy ed. Clark McCauley, pp. 69-87. Frank Cass. Crenshaw, Martha 1990. The logic of terro rism: Terrorist behavi or as a product of strategic choice. In Origins of Terrorism : Psychologies, Ideologies, Theologies, States of Mind ed. Walter Reich, pp. 7-25. Ca mbridge: Cambridge University Press Davenport, Christian, Hank Johnsto n and Carol Mueller eds. 2005. Repression and Mobilization University of Minnesota Press. De Figueiredo Jr., Rui J. P. and Barry R. Weingast. 2001. Vicious Cycles Endogenous political extremism and political viol ence. Working Paper, published by the Institute for Governmental Studies. (accessed 4/13/09). De la Calle, Luis and Ignacio Snchez-Cuenca. 2006. The Production of Terrorist Violence: Analyzing Target Selection w ithin the IRA and ETA. Working Paper 2006. Instituto Juan March: Madrid. /proyectos/dtv/papers (accessed 3/8/09). De la Calle, Luis and Ignacio Sanchez-Cuenca. 2007. Victims of ETA Dataset yectos/dtv/datasets.asp (accessed 3/25/09). Della Porta, Donatella. 1995. Social Movements, Political Violence, and The State: A Comparative Analysis of Italy and Germany Cambridge University Press. 190

PAGE 197

Della Porta, Donatella. 2009. Leaving left -wing terrorism in Italy: A sociological analysis. In Leaving Terrorism Behind eds. Tore Bjorgo and John Horgan, pp. 66-88. Routledge. DeNardo, James. 1985. Power in numbers: the political st rategy of protest and rebellion Princeton University Press Drake, C. J. M. 1998. Terrorists Target Selection St. Martins Press, Inc. Drake, C. J. M. 1998(b). The Role of Id eology in Terrorists Target Selection. Terrorism and Political Violence 10, No.2 (Summer): 53-85 Encarnacion, Omar G. 2007. Democracy and Dirty Wars in Spain. Human Rights Quarterly 29, pp. 950 Enders, Walter and Todd Sandler. 2005. The Political Economy of Terrorism Cambridge University Press Engene, Jan Oskar. 2004. Terrorism in Western Europe: Explaining the Trends since 1950. Edward Elgar Publishing Limited. Eubank, Willam and Leonard Weinberg, 2001. Te rrorism and Democracy: Perpetrators and Victims. Terrorism and Political Violence 13: 155-64. Fearon, James D. and David D. Laitin. 2000. V iolence and the Social Construction of Ethnic Identity. International Organization 54, no. 4, pp. 845-877. Fearon, James D. and David D. Laitin, 1996. Explaining Interethnic Cooperation. American Political Science Review 90, No. 4. Gambetta, Diego. 2005. Can We Make Sense of Suicide Missions? In Diego Gambetta (ed.) Making Sense of Suicide Missions pp. 259-299. Oxford University Press. Ganor, Boaz. 2002. Defining Terrorism: Is On e Man's Terrorist another Man's Freedom Fighter? Police Practice and Research 3, No. 4: 287-304. Geraghty, Tony. 2000. The Irish War Harper Collings Publishers Ltd. Gibbs, Jack P. 1989. Conceptualization of Terrorism American Sociological Review Vol. 54, No. 3 (Jun., 1989), pp. 329-340 Stable URL: Goodwin, Jeff. 2004. What Must We Explain to Explain Terrorism? (Review Essay). Social Movement Studies, Vo l. 3, No. 2, October 2004, 259-265. Goodwin, Jeff. 2006. A Theory of Categorical Terrorism. Social Forces 84, No. 4 (June): 2027-2046. 191

PAGE 198

Gupta, Dipak K. 2006. Tyranny of Data: Going Beyond Theories. In Tangled Roots: social and psychological facto rs in the genesis of terrorism ed. Jeff Victoroff, 3751. IOS Press Gupta, Dipak K., Harinder Singh, and Tom Sprague. 1993. Journal of Conflict Resolution No. 37, 301-339. Hallett, Brian. 2004. Dishonest Crimes and Language. In Understanding Terrorism eds. Fathali M. Moghaddam and Anthony J. Marsella. American Psychological Association. Hayes, Bernadette C. and Ian McAlllister. 2000. Sowing Dragons Teeth: Public Support for Political Violence and Paramilitarism in Northern Ireland A paper prepared for the meetings of the UK Political Studies Association, London School of Economics and Political Science, 10-13 April 2000. (accessed 7/1/09). Hoffman, Bruce. 1998. Inside Terrorism Columbia University Press. Hultman, Lisa. 2004. Civilians as Pawns in th e Game of Civil War? Paper prepared for the workshop Techniques of Violen ce in Civil War, PRIO, Oslo, 20-21. Hultman, Lisa. 2008. Targeting the Unarmed: St rategic Rebel Violence in Civil War. Report / Uppsala University, Department of Peace and Conf lict Research 81, pp. 1-42. Josiger, William. "Fear Factor: The Impact of Terrori sm on Public Opinion in the United States and Great Britain" Paper presented at the annual meeting of the International Studies Association, Town & Country Resort and Convention Center, San Diego, California, USA Mar 22, 2006. Jones, Edward E. and Keith E. Davis. 1979. F rom Acts to Dispositions: The Attribution process in Person Perception. In Advances in Experimental Social Psychology ed. Leonard Berkowitz, pp.219-236. American Journal of Sociology Jones, Seth and Martin C. Libicki.2008. How Terrorist Groups End: Lessons for Countering al Qaida Published by the Rand Corporation. Kalyvas, Stathis. 2000. The Logic of Viol ence in Civil War. Publishing details unknown. (accessed 11/13/08). Kalyvas, Stathis. 2004. The Paradox of Terrorism in Civil War. The Journal of Ethics 8: 97-138. 192

PAGE 199

Kalyvas, Stathis and Ign acio Sanchez-Cuenca. 2005. K illing Without Dying: the Absence of Suicide Missions. in Making Sense of Suicide Missions m, ed. Diego Gambetta, pp. 209-232. Oxford University Press. Kalyvas, Stathis and Ma tthew Adam Kocher. 2007. How Free is Free Riding in Civil Wars? Violence, Insurgency, a nd the Collective Action Problem World Politics 59, No.2: 177-216 Kalyvas, Stathis and Matthew Adam Kocher 2007. Ethnic Cleavages and Irregular War: Iraq and Vietnam. Politics and Society 35, No. 2: 183-223. Koopmans, Rude. 2005. Repression and the Pub lic Sphere: Discursive Opportunities for Repression against the Extreme Right in Germany in the 1990s. In Repression and Kruglanski, Arie W., Martha Crenshaw, Jerro ld M. Post, and Jeff Victoroff. 2008. What Should This Fight Be Called? Metaph ors of Counterterrorism and Their Implications. Association for Psychological Science 8, No. 3: 97-133 Linz, Juan J. and Alfred Stepan. 1996. Problems of Democratic Transition and Consolidation. The Johns Hopkins University Press. Mobilization editors Christian Davenport, Ha nk Johnston and Carol Mueller. 159-189 University of Minnesota Press. Kruglanski, Arie W. 2006. The Psychology of Terrorism: Syndrome Versus Tool Perspectives. In Tangled Roots: social and psychological factors in the genesis of terrorism ed. Jeff Victoroff, pp. 61-74. IOS Press. Lake, David A. 2002. Rational Extremism: Understanding Terrorism in the TwentyFirst Century. Dialog-IO (Spring): 15-29. Laqueur, Walter. 2003. No End to War The Continuum International Publishing Group Lichbach, Mark Irving. 1987. Deterrence or Escalation? The Puzzle of Aggregate Studies of Repression and Dissent. The Journal of Conflict Resolution 31, no. 2 (June): 266-297 McAllister, Ian. 2004. The Armalit e and the Ballot Box: Sinn Feins electoral strategy in Northern Ireland. Electo ral Studies 23 (2004) 123-142 McCauley, Clark. 2008. Tragic It eration: Terrorism and the Response to Terrorism. Draft contribution for the Searle Center Research Symposium on Bad Public Goods. McCormick, Gordon H. 2003. Terrorist Decision Making. Annual Review of Political Science 6:473-507 193

PAGE 200

McKeown, Michael. 2009. Database of Deaths Associated with Violence in Northern Ireland (accessed 7/0/09). McKeown, Michael. 2001. Post-Mortem tims/mckeown/mckeown01.pdf (accessed 7/8/09). Melaugh, Martin. 2009. Internment Summary of Main Events. (accessed 6/28/09). Melaugh, Martin. 2009b. Collusion Chronology of Events in the Stevens Inquiries. (accessed 7/8/09). Miller, Gregory D. 2007. Confronting Terro risms: Group Motivation and Successful State Policies. Terrorism and Political Violence 19, No. 3: 331-350. Moghaddam, Fathali M., and Anthony J. Marsella. 2004. Understanding Terrorism Moloney, Ed. 2002. A Secret History of the IRA. Norton and Company. Nemeth, Stephen. 2009. Adaptive Tactics: Terrorist Targeting and Regime Type Paper presented at the annual meeting of the Th e Midwest Political Science Association Palmer House Hilton, Chicago, Illinois, 05/25/2009. Neumann, Peter R. and M. L. R. Smith. 2005. Strategic terrorism: The framework and its fallacies. Journal of Strategic Studies 28 No. 4: 571-595 Neumann, Peter R. and M.L.R. Smith, The Strategy of Terrorism : How it Works and Why it Fails 2007. Routledge. Parker, Tom. 2007. Fighting an Antaean Enemy: How Democratic States Unintentionally Sustain the Terrorist Movements They Oppose. Terrorism and Political Violence 19 No. 2: 155-179. Pape, Robert. 2006. Dying to Win: The Strategic Logic of Suicide Terrorism Random House Prorok, Alyssa. 2008. Mixed Regimes and Re pression: A Disaggreg ated Institutional Approach. Paper presented at the annual m eeting of the Mi dwest Political Science Association Annual National Conference Palmer House Hotel, Hilton, Chicago, IL. April 3. (accessed 12/01/08). Rapoport, David C. 2006. The Four Waves of Modern Terrorism in Terrorism: Critical Concepts in Political Science, Vol. 1 ed. David C. Rapoport. Routledge. 194

PAGE 201

Reich, Walter (ed). 1990. Origins of Terrorism University of Cambridge Press. Richards, A. 2001. Terrorist Groups and Political Fronts: The IRA, Sinn Fein, the Peace Process and Democracy. Terrorism and Political Violence, 13:4, 72-89 Richardson, Louise, 2006. The Roots of Terrorism, Routledge Sambanis, Nicholas and Annalisa Zinn, 2005. Fro m Protest to Violence: An Analysis of Conflict Escalation with an Applicati on to Self-Determination Movements. Paper presented at the annual meeti ng of the American Political Science Association, Marriott Wardman Park, Omni Shoreham, Washington Hilton, Washington, DC, Sep. 1, (accessed 12/14/08). Sanchez-Cuenca, Ignacio. 2009. Explaining tempor al variation in the lethality of ETA. Juan March Institute and Complutense University (Madrid) /proyectos/dtv/papers (accessed 3/18/09). Snchez-Cuenca, Ignacio. 2008. The Persistence of nationalist terrorism: the case of ETA. Chapter prepared for Kledja Mulaj (ed) Violent Non-State Actors in Contemporary World Politics os/dtv/pdf/ETA%20(Kledja).pdf (accessed 5/31/09). Sanchez-Cuenca, Ignacio and Luis de la Ca lle. 2008. Domestic Terrorism: The Hidden Side of Political Violence. Annual Review of Political Science 12: 31-49 Sanchez-Cuenca, Ignacio. 2007. Terrorism and the State. Chapter prepared for Handbook of Rational Choice Social Research (forthcoming), eds.Victor Nee, Tom Snijders, and Rafael Wittek. To be published by the Russell Sage Foundation. yectos/dtv/paper s.asp (accessed 3/12/09). Snchez-Cuenca, Ignacio. 2007. The Dynamics of Nationalist Terrorism: ETA and the IRA. Terrorism and Political Violence 19 (3): 289-306. /proyectos/dtv/papers. (accessed 3/8/09). Schmid, Alex. 1983. Political Terrorism: A Research Guide to Concepts, Theories, Data Bases, and Literature Elsevier Science Ltd Schmid, Alex and Ronald Crelinsten. 1993. Western Responses to Terrorism Frank Cass Publishers. Shabad, Goldie and Francisco Jose Llera Ramo. 1995. Political Violence in a Democratic State: Basque Terrorism in Spain. In Terrorism in Context, ed. Martha Crenshaw, pp. 410-469. Pennsyl vania State University Press 195

PAGE 202

Shephard, William. 2002. The ETA: Spain Fi ghts Europes Last Active Terrorist Group. Mediterranean Quar terly: Winter 2002 pp. 54-68 Snow, David A., E. Burke Rochford, Jr., St even K. Worden, and Robert D. Benford, 1986. Frame Alignment Processes, Micromobilization, and Movement Participation American Sociological Review 51, No. 4 (August): 464-481 START National Consortium for the Study of Terrorism and Responses to Terrorism. 2007. Methodology GTD 2. (accessed 3/13/09). Sutton, Malcolm. 1994. Bear in mind these dead... An Index of Deaths from the Conflict in Ireland 1969-1993. Belfast: Beyond the Pale Publications /sutton/index.html (accessed 5/26/09-6/5/09). Tilly, Charles. 2005. Repression, Mob ilization, and Explanation. In Repression and Mobilization eds. Christian Davenport, Hank Johnston and Carol Mueller, 211227 University of Minnesota Press: Tilly, Charles.2004. Terror, Terrorism, Terrorists. Sociological Theory 22, No. 1, Theories of Terrorism: A Symposium. (March): 5-13. (accessed 11/14/08). Townshend, Charles. 1995. The Culture of Paramilitarism in Ireland. In Terrorism in Context, ed. Martha Crenshaw. The Pennsyl vania State University Press. White, Robert W. 2000. Issues in the study of political violence: Understanding the motives of participants in small group polit ical violence. Terrorism and Political Violence 12 no. 1, pp. 95-108. White, Robert W. 1997b. The IRA and S ectarianism: Moving Beyond the Anecdote. Terrorism and Political Violence 9 no. 2. (Summer 1997), pp. 120-131. Published by Frank Cass, London White, Robert W. 1997. The IRA: An Assessment of Sectarianism. Terrorism and Political Violence 9 no. 1 (Spring 1997), pp. 20-55. Published by Frank Cass, London. White, Robert W. and Terry Falkenberg Wh ite. 1995. Repression and the Liberal State: The Case of Northern Ireland, 1969 1972 The Journal of Conflict Resolution 39, No. 2 (June): 330-352. 196

PAGE 203

White, Robert W, 1989. From Peaceful Protes t to Guerrilla War: Micromobilization of the Provisional Irish Republican Army. The American Journal of Sociology 94, No. 6 (May): 1277-1302. Wilkinson, Paul. 2006. Terrorism versus Democracy: The Liberal State Response Routledge Wintrobe, Ronald. 2006. Rational Extremism: The political economy of radicalism Cambridge University Press. Woodworth, Paddy. 2001. Dirty War, Clean Hands. Cork University Press. Woodwell, Douglas. The Troubles of Northern Ireland: Civil Conflict in an Economically Well-Developed State. In Paul Collier and Nicholas Sambanis (eds) Understanding Civil War: Evidence and Analysis, Vol. II pp. 161-190. The World Bank. Wragg, Tony. 2008. An Analysis of the Effect of Id eology on the Life Expectancy of Terrorist Organisations Master of Engineering Thes is. Royal Melbourne Institute of Technology (accessed via google books, 1/17/09). 197