P O L I T I C AL STU D IES: 2001 VO L 49, 901–922

Sowing Dragon’s Teeth: Public Support for Political Violence and Paramilitarism in Northern Ireland Bernadette C. Hayes Queen’s University

Ian McAllister Australian National University While much attention has been devoted to political efforts to solve the Northern Ireland problem, less attention has been given to the role of political violence in sustaining the conflict. We argue that one of the reasons for the intractability of the conflict is widespread exposure to political violence among the civil population. By 1998, thirty years after the conflict started, one in seven of the population reported being a victim of violence; one in five had a family member killed or injured; and one in four had been caught up in an explosion. Such widespread exposure to violence exists alongside latent support for paramilitarism among a significant minority of both communities. Using 1998 survey data, we show that exposure to violence serves to enhance public support for paramilitary groups, as well as to reduce support for the decommissioning of paramilitary weapons. Overall, the results suggest that only a lengthy period without political violence will undermine support for paramilitarism and result in the decommissioning of weapons.

These creatures Aeetes ordered him to yoke and to sow dragon’s teeth; for he had got from Athena half of the dragon’s teeth which Cadmus sowed in Thebes … And when he had sowed the teeth, there rose armed men from the ground … Jason and the Argonauts The most visible and dramatic manifestation of the post-1968 Northern Ireland conflict has been political violence. Comparative studies show that Northern Ireland is easily the most intense violent conflict in Europe, accounting for the majority of terrorist incidents in Europe (US Secretary of State, 2001). The various paramilitary organizations that operate in the province are the most highly organized and equipped in Europe, particularly on the republican side. The statistics of violence suggest that in its duration and intensity relative to population size, the conflict approaches that of a war rather than a local insurgency, with substantial numbers of the population being exposed to many aspects of the violence, from intimidation and physical injury, to being caught up in a bomb explosion or riot. Most of the research on the Northern Ireland problem has assumed that the violence is a consequence of the political problem; once a permanent settlement between the communities is reached, violence will become irrelevant and swiftly disappear. Less attention has been given to analysing how public exposure to violence serves to mould popular attitudes towards the use of violence as a political © Political Studies Association, 2001. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA

902

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

tool, and in turn to engender further political violence. In other words, violence may not simply be a consequence of other (mainly political) things, but it may feed off itself in a continuous and perpetual cycle, as in the proverb of the dragon’s teeth. The apparent widespread support for – or at the very least, ambiguity towards – paramilitary organizations suggests that there is a significant minority prepared both to condone political violence and to participate in it. It is the role of political violence in sustaining paramilitary support that is the topic of this article.

Aggregate Patterns of Violence The political violence that Northern Ireland has experienced since 1968 continues a long tradition of violent conflict dating back several centuries. Secret societies emerged in Ireland in the eighteenth century, using direct action to express agrarian discontent. Most of these secret societies were based in the south, but they occasionally committed outrages, mainly against unpopular landlords, in the north (Williams, 1973). By the end of the century the secret societies were represented in the north by two sectarian groups, the Peep O’Day Boys, who were Protestant, and the Defenders, who were Catholic (Clark and Donnelly, 1983). In turn, these groups were replaced by two organizations with overtly political goals, the Orange Order and the United Irishmen, respectively (Senior, 1966; Elliott, 1982). The loyalist and republican organizations which have employed political violence since 1968 are very much the direct descendants of these agrarian groups, which used similar methods, for similar ends, centuries before. The rapid industrialization of Belfast in the nineteenth century transferred these rural conflicts into an urban setting. Largely based on linen, and later shipbuilding and engineering, Belfast was one of the first industrial cities in the British Isles. The insatiable demand for labour generated by these new industries attracted both Protestants and Catholics from all over the north into Belfast, and their contiguity in working class areas of the city has provided the basis for violent conflict ever since. Throughout the nineteenth century Belfast was the scene of periodic communal rioting, causing around 60 deaths (Budge and O’Leary, 1973). The widespread disturbances that accompanied the 1886 Home Rule Bill resulted in 86 deaths across the province, many of them in Belfast (O’Leary and McGarry, 1993, p. 21). Successive observers of the conflict have noted the importance of the urban interfaces between the two communities as the touchstone for the conflict (Boal et al., 1976); for the most part, these interfaces were established during the nineteenth century. The post-1968 violence dwarfs any previous conflict in scale, intensity and duration. More people have died in communal violence in the past quarter century in Northern Ireland – 3,289 by the end of 19981 – than in any similar period in Ireland over the past two centuries, with the possible exception of the 1922–23 Irish Civil War (Table 1).2 In addition, over 40,000 people have been injured, representing almost 3 per cent of the population. If we extrapolate these figures to Britain, some 111,000 people would have died, with 1.4 million people injured. This represents just under half of all British deaths (265,000) during the Second World War. Further extrapolating the deaths to the United States, some 526,000 would have died, more than died during the Second World War (405,000) and nine times the American war dead in Vietnam.

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

903

Table 1: The Scale of the Political Violence, 1969–98 Estimates (000s)

Deaths Injuries Shooting incidents Bomb explosions Persons charged with terrorist offences

N. Ireland

Britain

USA

3,289 41,837 35,669 15,246 18,258

111 1,406 1,188 503 589

526 6,673 5,161 2,388 2,797

Note: Figures for persons charged with terrorist offences date from 31 July 1972. Source: Elliott and Flackes (1999).

The large number of incidents underlines the intensity of the conflict, with over 35,000 shooting incidents and 15,000 bomb explosions. Many of these bomb explosions have occurred in Belfast or Londonderry, which were the targets of intense and sustained bombing campaigns by the IRA during the 1970s.3 Such levels of violence, maintained over a long period of time, have inevitably drawn many people into the paramilitary organizations. Estimates of paramilitary membership are difficult to make with any accuracy, but police statistics show that since 1972, over 18,000 people have been charged with terrorist offences. It is a reasonable conclusion that more people in Northern Ireland have participated in illegal paramilitary organizations than at any time since the United Irishmen rising of 1798. Once again, extrapolating these figures to Britain or the US show the intensity of the violence; shooting incidents alone would have numbered over 1 million in Britain, and over 5 million in the United States. Around half a million British people would have been charged with a terrorist offence, and nearly 2.8 million Americans. By any standards, what Ulster people euphemistically call ‘the Troubles’ is, in fact, a war. Three main agencies have been responsible, in various ways, for the deaths that have occurred during the course of the conflict.4 The Provisional IRA have been responsible for by far the largest number of deaths – 1,696 by the end of 1998, or just under half of the total (Table 2).5 In the early stages of the conflict, the Official IRA was active in violence, but they regarded most of their operations as ‘defensive’ and did not emulate the Provisionals’ indiscriminate car bombing campaign that resulted in many civilian deaths during the early 1970s. The Official IRA declared a ceasefire in 1972, and thereafter reduced their military capacity, eventually becoming an entirely political organization by the late 1970s (O’Brien, 1999). At the same time, dissidents within the Official IRA who wanted to continue with the use of force formed the Irish National Liberation Army (INLA), which has specialized in the assassination of high profile individuals; for example, they were responsible for the assassination of the Conservative Northern Ireland spokesperson Airey Neave in 1979 (see Taylor, 1997). Of the 335 deaths attributable to non-IRA republicans, the INLA has been responsible for 117, or about one-third.

904

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

Table 2: Organizations Responsible for Deaths from Political Violence, 1969–98 Organization responsible Security forces

Republican

Loyalist

Person’s status

BA RUC UDR IRA Other UDA UVF Other Other Total

Civilian Security forces BA RUC UDR Republican IRA Other Loyalist UDA UVF Other Other

153

33

6

509

212

209

378

273

53

1826

5 2 1

2 1 –

– – –

475 308 222

34 23 10

– 2 1

1 4 2

– 4 1

4 2 1

521 346 238

108 17

12 5

2 –

132 5

4 37

8 3

13 8

8 1

13 10

300 86

6 5 – –

2 1 – –

– – – –

21 11 – 13

6 5 1 3

28 3 – –

6 27 – –

4 3 4 –

1 – – 1

74 55 5 17

Total

297

56

8

1696

335

254

439

298

85

3468

Notes: Memberships combine former as well as current memberships. ‘Civilian’ includes civilian political activist. ‘British Army’ includes RN, RAAF, RIR. ‘RUC’ includes prison officer, Ulster Special Constabulary. ‘Other Republican’ includes INLA, real IRA, Direct Action Against Drugs, People’s Action Force. ‘UDA’ includes UFF. ‘Other’ includes British and Irish police, and Irish Army. Sources: Fay et al., (1999); Royal Ulster Constabulary; Sutton (1994) and updates from http://cain.ulst.ac.uk; authors’ estimates.

The second main agency, the various loyalist organizations, have been responsible for 991 deaths, or 29 per cent of the total. The most active group is the Ulster Volunteer Force (UVF) which, despite its historic name, dates back only to 1966 when it planned and executed a series of sectarian murders in Belfast (Bruce, 1992; Nelson, 1984). The UVF have been responsible for 439 deaths, or more than four out of every 10 deaths caused by loyalists. Although the Ulster Defence Association (UDA) has been responsible for comparatively fewer deaths, they have generally operated through a surrogate organization, the Ulster Freedom Fighters (UFF) in order to avoid becoming a proscribed organization (Guelke, 1999).6 Combining the UDA and UFF results in a total of 254 deaths. The remaining deaths caused by loyalists are attributable to a variety of fringe organizations, or to unaffiliated individuals. The third agency, the security forces – combining the British Army, the Ulster Defence Regiment (UDR) and the Royal Ulster Constabulary (RUC) – have caused the fewest number of deaths.7 The British Army has been responsible for 297 deaths, or 9 per cent of the overall total, and the RUC and the UDR for 64. Most of those who have been killed by the security forces were civilian, although a significant minority were political activists of various kinds. A total of 125 of the

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

905

deaths attributable to the British Army were members of the IRA or other republican organizations. In large measure this supports the view that the British Army has acted to protect the police; although the police have been the target of much violence from the IRA, they themselves have been responsible for only a small proportion of the total deaths attributable to the security forces. The main combatants in wars are generally young males, and the Northern Ireland conflict is no exception to this pattern (Table 3). The two major armed groups, the British Army and the IRA, are remarkably similar in their age and gender profiles. For example, the mean age of British soldiers killed in the conflict is 24.6 years, compared to 24.9 years for IRA members. The average ages of the local security forces who have been killed are substantially higher – between 36 and 37 years. The vast majority of combatants are male, and indeed totally so among the various loyalist organizations. Among the civilians, males constitute 85 per cent of the deaths; the comparatively small number of women killed were mainly caught in explosions or crossfire. Although Catholics are a minority of the population, they have experienced slightly more deaths than Protestants; overall, 54 per cent of those who were from Northern Ireland and were killed there were Catholic. As we would expect, the local security forces’ deaths were overwhelmingly Protestants (although not totally so), while IRA deaths were totally among Catholics. More than nine of out every

Table 3: The Characteristics of those Killed, 1969–98 Organizational affiliation Security forces

Republican

Loyalist

Civilian BA

RUC

UDR IRA Other UDA UVF Other Total

Age (mean years) 35.2 24.6 N 1,821 518 Gender (percent) Male 85 99 N 1,826 521 Religion (percent) Protestant 34 na Catholic 66 na N 1,658 Place killed N Ireland 93 92 Britain 4 8 Elsewhere 3 0 N 1,820 513

36.2 346

36.8 24.9 27.5 238 295 86

98 346

98 238

90 10 345

95 0 3 5 100 97 237 282 79

96 300

97 86

100 100 97 93 0 0 1 0 0.3 0 2 7 346 238 300 86

29.6 30.4 29.0 32.5 74 55 5 3,455 100 100 100 74 55 5

91 3,468

100 100 0 0 74 55

1 0 5

54 46 2,755

100 100 100 0 0 0 0 0 0 74 55 5

94 3 2 3,455

Notes: See Table 2 for details of groups. The numbers vary due to missing values, except for religion, which is estimated only for those who were from Northern Ireland and killed there. Sources: See Table 2.

906

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

10 deaths attributable to the conflict have taken place in Northern Ireland, but it is also notable that 8 per cent of British Army deaths have occurred in Britain, as have 4 per cent of civilian deaths. These are the result of the periodic IRA terrorist campaigns mounted in Britain. Each of the three main agencies in the conflict – the security forces, republicans and loyalists – have been conducting several conflicts which account for the vast majority of deaths (Figure 1). The most destructive conflict has pitched the security forces against republicans, the latter dominated by the IRA (Sullivan, 1998; O’Duffy, 1995). Republicans have killed 1,072 members of the security forces, of whom 475 have been members of the British Army, 308 members of the RUC, and 222 members of the UDR (Figure 2). In turn, the security forces have killed 144 members of the IRA and other republican groups, as well as 192 civilians. This security forcesrepublican conflict has remained a constant backdrop to the Troubles, with the exception of periods when the IRA has maintained a ceasefire, but it was particularly intense in the early 1970s, when the bulk of the deaths occurred. The second dimension to the conflict has been one mounted by the various loyalist organizations against the civilian population, predominantly Catholics. Throughout the 1970s loyalists waged a campaign of sectarian assassination against Catholics in retaliation for IRA violence, and occasionally in response to sectarian attacks mounted by republicans against Protestants (Taylor, 1999). Overall, this campaign resulted in 860 deaths, with a large proportion of them occurring during the mid1970s. In 1975 alone, loyalists accounted for almost as many deaths as republicans. The campaign was largely discontinued after 1976, partly because it was seen to be ineffective in halting IRA violence, and partly because many of the loyalist leaders were arrested and jailed (O’Duffy, 1995). However, the campaign was reactivated in the early 1990s, as Figure 2 shows, albeit on a lesser scale than previously. In 1993 and 1994, loyalists were responsible for more than half of all of the deaths that occurred in those years.

Figure 1: Three Dimensions of the Conflict, 1969–98 1,072 Security forces

Republicans 144

192

Civilians

72

860

Notes: Groups are categorized according to Table 2. Sources: as for Table 2.

1

Loyalists

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

907

Figure 2: Main Agencies Responsible for Deaths, 1969–98 300

Security forces

250

Loyalists Republicans

Number of deaths

200

150

100

50

0

1969

1972

1975

1978

1981

1984

1987

1990

1993

1996

1999

Sources: as for Table 2.

A third dimension has been the campaign the IRA and other republicans have waged against the civilian population. This has taken the form of civilian casualties caused by indiscriminate bombings and shootings; in the early stages of the conflict most civilian deaths were the largely unintended consequences of other activities, such as random car bombings in city and town centres where inadequate or misleading warnings were provided. In recent years the IRA have also engaged in more overtly sectarian acts of violence, often responding to sectarian attacks by loyalists on Catholics, although the extent to which it qualifies as a sectarian organization is open to much dispute (see White (1997) and Bruce (1997) for two competing views in relation to this issue). Two further observations about these patterns of violence are worthy of mention. First, despite their obvious antagonism, republicans and loyalists have conducted only sporadic attacks against one another. Republicans have managed to kill just 44 loyalists during the course of the conflict, or 2 per cent of all of the deaths for which they are responsible. Similarly, loyalists have killed just 41 republicans, or 4 per cent of their total. In large measure, this represents no lack of will – particularly on the loyalist side – but the logistics of such operations often require more daring and planning than each side has access to. Second, significant numbers of

908

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

republicans and loyalists have been killed by their own sides. For example, republicans have killed four times more of their own members than they killed loyalists, and loyalists twice as many of their own as republicans. Some of these deaths have been as the result of accidental shootings and premature bomb explosions (the latter a particular factor among republicans – see National Graves Association (1976)), but a significant number have also been the result of summary executions for informing, corruption, and stemming from personal feuds.

Personal Exposure to Political Violence In most advanced industrial societies, comparatively few people have any direct experience of political violence. The Oklahoma City bombing, while resulting in massive loss of life, directly affected only a small percentage of the US population. Likewise, Arab-inspired terrorist attacks in European airports in the 1970s and 1980s affected only small numbers of people. Since the end of the Second World War, no advanced industrial society has experienced political violence at a remotely comparable level to that of Northern Ireland. The duration, intensity and extent of the Northern Ireland conflict has ensured that large proportions of the population have direct, personal experience of political violence. While many have experienced injury, much larger numbers have witnessed acts of violence, ranging from bomb explosions and shootings, to vehicle hijacking and rioting. There are three forms of political violence that a population might be exposed to.8 First, there is the direct experience of violence – being the victim of a violent event, and perhaps suffering injury. Table 4 shows that by 1998, about one in seven of the adult population reported having been the victim of a violent incident. Threats and intimidation have been a common theme throughout the conflict; in the early years of the disturbances, large population movements occurred as a result of intimidation, with families moving out of religiously mixed areas into areas dominated by their co-religionists where they felt more secure (Whyte, 1990, pp. 33–4). Threats have also been employed to ensure that communities do not inform on paramilitaries, and to impose rudimentary public order on local communities, particularly with regard to drug dealing and vandalism (Thompson and Mulholland, 1995). Starting in 1973 (when the first survey was conducted to ask about exposure to violence), up to one in five of the population report being intimidated at some stage during the Troubles. A second dimension is indirect experience of violence, through having a family member or close relative killed or injured, or knowing someone who has been killed or injured. The nature of Northern Ireland society, with large, extended families and closely-knit communities, means that a death or an injury as a result of the violence has wide social repercussions. By 1998, approximately one in five reported having had a family member or close relative injured or killed, and more than half personally knew someone who had been injured or killed. An additional form of indirect experience of violence is collective exposure to violence, through being caught up in a violent act, either an explosion or a riot. Exactly a quarter of those interviewed in 1998 had been caught up in an explosion, while almost the same number had been caught up in a riot. However, in the early 1970s, when riots were widespread, the levels of exposure were higher, and more than one in three had

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

909

Table 4: Exposure to Political Violence, 1973–98 Percentages reporting

Direct experience Intimidated Victim of violent incident Indirect experience Family member/relative killed or injured Know someone killed/injured Collective exposure Caught up in explosion Caught up in riot

1973

1978

1995

1998

15 3

– –

19 10

18 14

8 26

– –

18 69

21 56

12 34

20 14

21 20

25 23

Notes: The questions were as follows. 1973: Have you been present in a building when a bomb exploded? Have you been present when a riot or confrontation took place? Have you [family/friends] suffered any kind of injury due to the Troubles? Have [family/friends] been killed due to the Troubles? 1978: Have you ever witnessed an act of terrorism, for example, a shooting, an explosion, a hijacking, or rioting? 1995: During the Troubles, were you ever caught up in an [explosion/riot]. Were you intimidated due to the Troubles? Were you a victim of any violent incidents? Were any of your family or close friends killed or injured due to the violence? Did you know anyone (not family or relatives) who was [killed/injured] due to the violence? 1998: During the Troubles, were you ever caught up in an [explosion/riot]. Were you intimidated because of the Troubles? Were you a victim of any violent incidents? Were any of your family or close friends killed or injured because of the violence? Did you know anyone (not family or relatives) who was killed or injured in the violence? Sources: Irish Social Mobility Survey, 1973; Social Attitudes Survey, 1978; Social Identity Survey, 1995; Northern Ireland Referendum and Election Survey, 1998.

experienced a riot.9 This is, then, a conflict whose effects extend far beyond those who participate in the violence or who live in areas where it is mainly conducted; the effects extend to the society as a whole. Exposure to violence is not shared equally between the two communities. Catholics report substantially higher levels of direct experience of violence compared to Protestants; Table 5 shows that Catholics are twice as likely to have been intimidated when compared to Protestants, and they are about one-third more likely to have been the victim of a violent incident. This greater incidence of direct experience of violence is attributable to the nature of the conflict. As the previous section indicated, two of the three main dimensions of activity have involved the conflict between the IRA and the British Army, and between loyalists and the Catholic community. In each case, these activities have exposed Catholics to a higher risk of personal violence than Protestants. Catholics are also more likely than Protestants to have had more indirect experience of violence, again reflecting the nature of the conflict, but the differences are not as pronounced. In 1995, for example, equal proportions of both religions reported knowing someone who had been killed or injured in the Troubles – a reflection of the sometimes random nature of the violence, particularly with regard to bombings which have often targeted pubs, shops and other public places which are not exclusively the preserve of one or other community. Finally, the widespread

910

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

Table 5: Religious Differences in Exposure to Political Violence, 1973–98 Percentages of each religion reporting 1973

1978

1995

1998

Prot Cath

Prot Cath

Prot Cath

Prot Cath

Direct experience Intimidated 9 Victim of violent incident 2 Indirect experience Family/relative killed/injured 7 Know someone killed/injured 21 Collective exposure Caught up in explosion 13 Caught up in riot 31

24 4

– –

– –

12 8

27 11

12 13

23 16

9 34

– –

– –

17 69

22 69

21 54

22 59

10 39

18 12

24 18

22 16

19 26

24 19

26 28

Sources: As for Table 4.

nature of the violence has meant that both communities have been as likely to have been caught up in an explosion. However, the greater incidence of riots in Catholic areas means that they report more exposure to this form of collective violence. Not only is exposure to violence unevenly distributed across the communities, there are significant variations in the groups within each community who are exposed to it. By far the most important factor is geography. A variety of studies have found that urbanization is of prime importance and, net of other factors, living in Belfast or Londonderry is significantly more likely to result in death or injury (Schellenberg, 1977; Murray, 1982; Poole, 1983, 1993). This is confirmed in Table 6, which uses the 1998 Northern Ireland Referendum and Election Survey to run regression analyses predicting the probability of being exposed to all of the six separate types of violence identified in Table 5. For Catholics, there is a particularly strong effect for living in Belfast: compared to Catholics living outside the two major cities, there is a 20 per cent greater probability of someone being exposed to violence, net of other things. Gender and age have also been found to be significant. As in most conflicts, it is young men who are most likely to be caught up in conflict, to be recruited as paramilitary activists and, as a result, the most likely to be killed or injured. The type of violence we are interested in here, ranging across from intimidation to knowing someone who has been killed or injured, should make these relationships less significant. Nevertheless, the results in Table 6 show that men are more likely to be exposed to violence than women, while Protestants aged in their thirties report more exposure than any of the other age groups. Social deprivation does not seem to be important, as least as it is measured by labour force status. However, Protestants who attended a mixed school have a greater probability of exposure. There are two possible explanations for this. First, such schools are often more likely to be located in interface areas between the two communities, where violence is

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

911

Table 6: Factors Affecting Exposure to All Types of Political Violence, 1998 OLS regression estimates Protestant b

Gender Age 18–29 30–39 40–49 50–59 Attended mixed school Labour force Employed Unemployed Residence Belfast Londonderry Constant Adj R-squared (N)

0.08**

beta 0.13**

Catholic b

beta

0.07*

0.13*

–0.03 0.09* 0.06 0.05 0.07*

–0.05 0.14* 0.09 0.07 0.11*

0.03 0.08 0.08 0.05 –0.04

0.05 0.12 0.11 0.06 –0.05

0.05 0.09

0.09 0.07

0.04 –0.08

0.07 –0.05

0.09** 0.11

0.14** 0.08

0.12 0.07 (468)

0.20** 0.16**

0.30** 0.23**

0.13 0.14 (332)

Notes: *= p , 0.05; **= p , 0.01, both two-tailed. Ordinary least squares regression equations predicting the probability of being exposed to political violence. Political violence combines the items listed in Table 5 and is a cumulative scale running from zero to 1. All of the independent variables are scored zero or 1 and the excluded categories are: gender, female; age, 60+; labour force, not in labour force; residence, outside city. Source: Northern Ireland Referendum and Election Survey, 1998.

greater. Second, attending a mixed school brings individuals into contact with those of the opposite religion, and that may make them easier to target for retaliation.10

Public Support for Political Violence Perhaps more than anything else, the Northern Ireland conflict has been sustained by the popular ambiguity that exists towards the use of political violence. Latent support for the use of violence often occurs in societies where political institutions have emerged from war or civil conflict. However, such support is usually transitional, and once the principle of the orderly transfer of political power following democratic elections becomes established, support for violence fades. The Irish state emerged out of a successful war against the British followed by a deeply divisive civil war; nevertheless, by 1932 the republicans who had lost the civil war had been returned to office in a democratic election and the parliamentary tradition securely entrenched (Farrell, 1973; Munger, 1975).

912

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

In Northern Ireland, by contrast, two traditions of achieving political change have been engrained in the political system. The constitutional tradition seeks to attain political change primarily through political parties competing in democratic elections, as well as through pressure and interest group activity. These are the familiar (and exclusive) forms of political activity in the established democracies. The extraconstitutional tradition seeks to achieve political goals through the use of force, either through protest activity (and an implicit threat of physical force) or through the use of armed force itself (the explicit threat of physical force). These two traditions have operated in parallel for two centuries, with each being dominant at particular periods. For example, the Irish Party’s success in bringing the Irish question to the forefront of British politics in the late nineteenth century made the constitutional tradition dominant; with the failure of the third Home Rule Bill, physical force became dominant, leading to the 1916 Easter Rising and the eventual formation of the Irish state in 1921. Two characteristics of how these two traditions have operated in Northern Ireland are important. First, the decision whether or not to use constitutional or extraconstitutional methods is less a moral one than a matter of expediency and practicality; if violence is seen to have the greatest chance of achieving the required political goals, then it will be utilized. Second, while the two traditions are analytically separate, groups and individuals nominally in one tradition may invoke the means of the other in order to advance a political aim. For example, Charles Stewart Parnell recruited a wide and politically heterogeneous following by refusing to define how far he would deviate from constitutional politics to attain his demands (Lyons, 1973, p. 193). In contemporary Northern Ireland a similar strategy is followed by the republican movement, which sustains an electoral organization, Sinn Fein, as well as an armed force, the IRA – in what has been immortalized as the strategy of the ‘armalite and the ballot box’.11 The ambiguity surrounding the use of physical force, which is clearly apparent in many of the main political organizations and leaders in Northern Ireland, is also found within the general population. Using public opinion surveys to gauge public support for political violence is problematic; most respondents are loath to admit their support for physical force in a personal interview and, in any event, such support is usually contingent upon the particular circumstances at the time. A person might be less supportive of force during a ceasefire, when their community was making major gains through constitutional politics, but more supportive if they felt political progress was slow or their community at risk from sectarian attack. A range of questions asked in the 1968, 1973 and 1978 surveys, reported in Table 7, provide different perspectives on the extent of latent public support for the use of physical force. In 1968, on the eve of the Troubles, survey respondents were asked if they believed it right ‘to take any measures necessary’ to achieve a political goal. For Protestants, the political goal was maintaining Northern Ireland as ‘a Protestant country’; the implication of this was the preservation of the Protestant community’s monopoly on political power and the preservation of their British heritage. Moreover, the reference to the use of ‘any measures necessary’ is a direct quote from the 1912 Ulster Covenant, which committed the signatories to use ‘all means that may be

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

913

Table 7: Public Support for the Use of Violence, 1968–78 Percent agree

1968 Right for [Protestant/Catholic] to take up arms 1973 Violence is a legitimate way to achieve one’s goals 1978 The IRA are basically patriots and idealists Loyalist paramilitary actions are justified

Protestant

Catholic

51

13

16

25

35 44

46 25

Notes: The questions were as follows. 1968: ‘Sometimes you hear people say that it would be right to take any measures necessary in order to [keep Northern Ireland a Protestant country/end partition and bring Ulster into the Republic.]’ 1973: ‘Overall, violence is a legitimate way to achieve one’s goals.’ 1978: ‘The IRA are basically patriots and idealists; The actions of loyalist paramilitaries are a justified reaction to what has happened in Northern Ireland.’ Sources: Loyalty Survey, 1968; Irish Social Mobility Survey, 1973; Social Attitudes Survey, 1978.

found necessary’ to defeat Home Rule (Bardon, 1992, p. 438). Given the historical parallels, it is not surprising that half of all of the Protestants interviewed in the survey supported the use of force for that goal. Catholics were less likely to support the use of force to end partition; nevertheless, 13 per cent still agreed with the statement. Among the Catholics supporting the statement, three-quarters said that they did so because of their belief in Irish unity (Rose, 1971, p. 193). A more explicit question about the legitimacy of violence, divorced from historical circumstances, produced different results in 1973. When asked if they believed that violence was a legitimate way to achieve one’s goals, 16 per cent of Protestants agreed, but more significantly, 25 per cent of Catholics took the same view. Clearly the violence of the early years of the conflict, which had disproportionately affected Catholic areas, had caused a shift in opinion towards support for the use of violence in order to defend their areas from Protestant attack. The survey was also conducted after the abolition of the Stormont parliament and the imposition of direct rule from Westminster, which was widely interpreted as having been achieved through sustained republican violence. In 1978 the issue was broached by asking respondents how they felt about the paramilitaries from each of the two communities. Almost half of the Catholics interviewed believed that members of the IRA were ‘patriots and idealists’, while only slightly fewer Protestants believed that loyalist paramilitaries’ actions were justified. The most recent survey question relating to support for violence was asked in 1998, and deals with the levels of sympathy expressed by respondents for each community’s paramilitary organization.12 The question was phrased so as to permit respondents to indicate sympathy, while not at the same time explicitly supporting the use of force. Once again, Table 8 shows that significant minorities within each community lend support to their respective paramilitary groups. A total of 31 per cent of Protestants express some level of sympathy for the loyalist paramilitaries,

914

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

while the figure for Catholics sympathizing with republicans is very similar, at 28 per cent. About one in every 14 people within each community have ‘a lot of sympathy’ for their respective community’s paramilitary organization. By any standards, these are significant numbers of people within a society who have empathy with the methods and goals of terrorist organizations. If we extrapolate these estimates to the general population, around 60,000 Protestants would have a lot of sympathy for loyalist paramilitaries, and about 40,000 Catholics a lot of sympathy for their republican counterparts.13 A notable pattern in both tables is the sympathy expressed for the other community’s paramilitaries. In Table 7, 35 per cent of Protestants agreed that the IRA were ‘patriots and idealists’, and 25 per cent of Catholics that the actions of loyalist paramilitaries was justified. Similarly, in Table 8, although hardly any of the respondents said that they had a ‘lot of sympathy’ for the other side’s paramilitaries, one in five Catholics had ‘some sympathy’ for loyalist paramilitaries, and one in 10 Protestants ‘some sympathy’ for republicans. This pattern is all the more curious when we take into account the ferocity of the conflict between the two main paramilitary groupings. The explanation seems to rest on how the activists on both sides regard the conflict as a war, and their own role in it as one of ‘soldiers’ fighting for a just cause; both paramilitary groupings regard their members serving jail sentences as ‘prisoners of war’. In turn, these ‘soldiers’ and their sympathizers legitimate their own status in the conflict by showing respect for the motives of their opponents (Bairner, 1996, p. 161).

Table 8: Public Support for the Use of Violence, 1998 Percent

Loyalists Lot of sympathy Little sympathy No sympathy Total N Republicans Lot of sympathy Little sympathy No sympathy Total N

Protestant

Catholic

7 24 69

1 20 79

100 462

100 327

0 11 89

7 21 72

100 465

100 326

Notes: The question was : ‘Thinking about the reasons some [Loyalist/Republican] groups have used violence during the Troubles, would you say that you have any sympathy with the reasons for violence – even if you don’t condone the violence itself?’ Source: Northern Ireland Referendum and Election Survey, 1998.

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

915

The opinion poll evidence in both communities about support for physical force, while intermittent, tells a remarkably consistent – and shocking – story. The results show that, however the survey question is phrased, significant minorities within each community support the use of violence for political ends. There is perhaps no other advanced industrial society where such large numbers of people effectively condone paramilitary activity. The explanations can be traced to the complex interaction between constitutional and extra-constitutional politics throughout Irish history, and to the successes of republicans in achieving political independence through the use of force and of unionists in securing exclusion from these new arrangements through the threat of force. For each community, the activities of contemporary paramilitary groups resonate with the iconographic figures of their history.14 Perhaps more pertinently, the message of the Troubles is that the use of physical force does bring political gains, a fact that has not been lost on constitutional politicians.

The Consequences of Political Violence Political violence has a range of consequences for the society within which it takes place. It has a major impact on the everyday lives of individuals, threatening injury or death to themselves and to their relatives and friends; destroying property, jobs and livelihoods; and altering political behaviour and beliefs. Most importantly, we would predict that exposure to violence would engender further violence, in a continuous cycle; as the proverb suggests, those who sow dragon’s teeth reap a harvest of violence in return. Numerous paramilitary activists have cited violent experiences as formative influences in their decision to take up arms. For example, the Sinn Fein president, Gerry Adams, traces his activism to attempts by the RUC to prevent an Irish tricolour being displayed at Sinn Fein’s West Belfast office during the 1964 general election campaign; he says that the experience led him to believe that ‘the north of Ireland was a state based upon the violent suppression of political opposition’ (Adams, 1996, p. 51). Similar experiences have also been recounted by paramilitaries within the loyalist community (see Crawford, 1999; Taylor, 1999). Violence also causes major disruption to democratic political institutions. Although Northern Ireland maintained formal democratic institutions from 1921 to 1972, the province was not a competitive democracy in the conventional sense, since there was no possibility that the Protestant and unionist majority would ever lose an election, nor that the Catholic and nationalist opposition could ever win. It was a hegemonic state where the contiguity of religion and politics was sustained by the ever-present threat of violence (O’Leary and Arthur, 1990). Following direct rule in 1972, even limited local democracy was withdrawn from the province, with the partial (and short-lived) exceptions of the 1973 Assembly, the 1975 Constitutional Convention and the 1982 Assembly, all of which failed in their goal of re-establishing devolution and were undermined by widespread political violence. The current threat to the future of the Northern Ireland executive, caused by delays in the IRA’s commitment to decommission its weapons, shows the ability of physical force groups, even when adhering to a ceasefire, to undermine democratic institutions.

916

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

Disagreement over the decommissioning of paramilitary weapons highlights the ambiguity surrounding the use of political violence. The issue became prominent after the IRA and loyalist ceasefires of August and October 1994, respectively. Constitutional groups argued that the ceasefires could not be translated into a permanent settlement until the paramilitary groups handed in their arms. The British government’s position was that weapons had to be handed in prior to the commencement of talks; republicans argued that decommissioning could only take place once a settlement had been reached. The Independent International Commission on Decommissioning, then chaired by US senator George Mitchell, concluded in January 1996 that while all of the parties should commit themselves to democratic principles and to total disarmament, in practice decommissioning would have to await an agreement (Mitchell, 1999). The British government did not accept this view and called an election as an alternative to decommissioning. The Good Friday Agreement is ambiguous on the decommissioning issue. It commits the signatories ‘to use any influence they may have, to achieve the decommissioning of all paramilitary arms within two years following endorsement in referendums North and South of the agreement and in the context of the implementation of the overall settlement.’ Republicans viewed this goal as an aspiration; once democratic institutions which included Sinn Fein were established and accepted, arms would gradually be decommissioned (Ruane and Todd, 1999, p. 25). By contrast, unionists saw it as binding that decommissioning would be underway prior to the formation of the executive, and that the process would be, at the very least, well advanced by May 2000, as laid out in the Good Friday Agreement. For many unionists, it was an article of faith that they would not share government with an organization that maintained arms. The issue came to a head during George Mitchell’s review of the Good Friday Agreement, conducted between September and November 1999. Following extensive discussions, Mitchell concluded that ‘a basis now exists for devolution to occur, for the institutions to be established, and for decommissioning to take place as soon as “possible”’ (The Times, 19 November 1999). In effect, Mitchell was arguing that both should occur simultaneously. This placed the Ulster Unionists in a quandary: should they adhere to their policy of non-participation without decommissioning, thereby wrecking the Agreement, or should they agree to participate and wait for IRA decommissioning under the terms of the Agreement? The party’s council voted 480 votes to 349 on 27 November 1999 to take the latter path, as advocated by their leader, David Trimble. Trimble swayed party opinion by signing a postdated letter of resignation from the executive, which would be activated if decommissioning had not commenced by February 2000 (Sunday Times, 28 November 1999). The executive was formed on 2 December 1999, but on the clear understanding on the part of the Ulster Unionists that the executive would collapse unless decommissioning commenced immediately. As in the proverb of the dragon’s teeth, we would hypothesize that exposure to political violence fosters support for further violence, expressed through support for paramilitary organizations. In addition, we would hypothesize that violence is deeply intertwined with popular views about the decommissioning of weapons. When asked about their views on this issue, both communities overwhelmingly

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

917

supported decommissioning as one of the main components of the Good Friday Agreement, with 95 per cent of Protestants and 87 per cent of Catholics supporting it (Table 9). However, the figures reveal that Protestants were almost twice as likely to be strongly supportive than Catholics, indicating the depth of Protestant feelings about the issue. By contrast, just 2 per cent of Protestants and 6 per cent of Catholics opposed decommissioning. Indeed, of the eight major proposals contained in the Agreement, decommissioning received the strongest popular endorsement.15 The multivariate results, using exposure to violence to predict sympathy for paramilitary groups and in turn views about the decommissioning of paramilitary weapons, support the hypotheses (Figure 3). In each community the exposure to political violence helps to mould sympathies for paramilitary organizations, and views about the decommissioning of weapons, although there are important differences. Having indirect experience of violence, through knowing someone who has been killed or injured, increases sympathy for paramilitary groups in both communities. By contrast, direct, personal experience of violence decreases sympathies for the loyalist paramilitary groups among Protestants, but increases sympathies for republican groups among Catholics. This may be related to the differing roles of the paramilitaries within each community. For Protestants, redress for being a victim of violence is most likely to be accomplished via the security forces. For Catholics, the role of the IRA as ‘the community’s defender’, coupled with the antipathy of many Catholics towards the security forces, makes them look to republican paramilitaries for retaliation (Brewer et al., 1999). Exposure to violence and support for paramilitary groups also moulds views on the decommissioning of weapons, as we predicted. Within both communities, popular sympathies for the respective paramilitary groups results in weaker support for decommissioning. Paramilitary sympathizers clearly believe that these groups should retain their weapons and their physical force role; the relationship is particularly strong among Catholics. Exposure to collective violence – witnessing a bombing or a riot – also influences views about decommissioning, but again in different ways

Table 9: Support for the Decommissioning of Paramilitary Weapons, 1998 Percent

Protestant

Strongly support Support Neither support nor oppose Oppose Strongly oppose Total N

62 33 3 2 0.3 100 472

Catholic 34 54 7 5 1 100 324

Notes: The question was: ‘Now I would like to ask you about some of the proposals contained in the Good Friday Agreement. Could you tell me how you feel about … decommissioning of paramilitary weapons.’ Source: Northern Ireland Referendum and Election Survey, 1998.

918

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

Figure 3: Exposure to Violence, Paramilitarism, and Decommissioning, 1998 Protestants (n = 468) 0.97 Collective

0.09

(0.43) Indirect

(0.34)

(0.32) Direct

0.98

Supports decommissioning

0.1 5

–0.10

.07

–0

Sympathy for loyalist paramilitaries Catholics (n = 332)

0.93

Collective –0.15

(0.42) (0.45)

0.1

(0.33) Direct

Supports decommissioning

–0.10

Indirect

0.09

2

0.97

–0 .

07

Sympathy for republican paramilitaries

Notes: OLS regression models showing standardized regression coefficients. Figures in parentheses are correlation coefficients. All paths control for age, gender, education and employment status and are statistically significant at p < 0.10, two-tailed. Source: Northern Ireland Referendum and Election Survey, 1998.

within each community. Among Protestants, such exposure serves to foster support for decommissioning but among Catholics exposure weakens it. Once again, these differences may reflect the different role of the security forces in each community, since collective exposure to violent events may bring individuals into contact with the security forces, and reinforce pre-existing views about them. For Catholics, such contact is more likely to be negative. Indirect experience of violence also weakens support for decommissioning, but only among Catholics.

Conclusion The Northern Ireland problem has proved to be one of the most intractable political conflicts of the twentieth century. Scholars have disagreed widely both in their interpretation of its causes and in its possible solutions – a phenomenon that McGarry and O’Leary (1995, p. 355) call ‘the conflict about the conflict’. There is, however, now cautious agreement that it is an internal conflict between the two religious communities, and that outside influences – such as those of the British or Irish governments – have relatively little impact on its trajectory. The conflict is

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

919

perpetuated by a range of social processes, such as family structure, the educational system and social interaction, which reinforce the values and outlooks which are specific to each community. These values and outlooks are transmitted from generation to generation, so that it is the young who are significantly more likely to have been influenced in their political outlooks by growing up amidst widespread violence and civil disturbances (Hayes and McAllister, 1999). One aspect of the conflict that has been consistently neglected in these interpretations is political violence. The violence has been treated as a consequence of the absence of political agreement; once a permanent settlement is found involving all of the main actors, then it is assumed that the violence will disappear. If all parties (and especially the republicans) believe that the Northern Ireland state is legitimate, then there is no need to maintain the armed struggle against it (Todd, 1999, p. 66). This optimistic scenario ignores two factors. First, Northern Ireland maintains two traditions of achieving political change, one constitutional and one extra-constitutional, the latter fostered by an historic communal enmity and, since 1972, by the major political gains that have been delivered by republican violence. Since the decision whether or not to use extra-constitutional methods for political ends is a practical rather than a moral judgement, its future use cannot be excluded. Second, as we have shown in this paper, exposure to political violence is often a cause of further violence, through support for paramilitarism and by a disinclination to support the decommissioning of paramilitary weapons. Only a sustained period of peace is likely to break this vicious cycle. The Irish Republic’s experience in nation building suggests that the transition to an exclusively parliamentary tradition can become established in a decade, although much of that was promoted by the ruthless suppression of the IRA by both pro and anti-treaty governments. Moreover, the large numbers who have been exposed to and directly influenced by political violence in Northern Ireland suggest that the demise of the physical force tradition will take much longer. The current impasse over the decommissioning of paramilitary weapons also indicates that while all parties may have been signatories to what they believe to be a lasting settlement, most (and particularly the republican paramilitaries) wish to maintain their military capacity in the event of a breakdown. Whatever the outcome of this latest phase of the conflict, it suggests that latent support for paramilitary groups in Northern Ireland will continue for some time in the future. (Accepted: 17 May 2001) About the Authors Bernadette C. Hayes, Department of Sociology, Queen’s University, Belfast BT7 1NN, Northern Ireland; email: [email protected] Ian McAllister, Research School of Social Sciences, Australian National University, Canberra ACT 2600, Australia; email: [email protected]

Notes The 1968 Loyalty survey was collected by Richard Rose and funded by the Social Science Research Council; the 1973 Irish Social Mobility Survey was collected by John Jackson and Robert Miller and funded by the Economic and Social Research Council; the 1978 Northern Ireland Attitudes Survey was collected by Edward Moxon-Browne and funded by the Nuffield Foundation; the 1995 Social Identities

920

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

Survey was collected by Karen Trew and Denny Benson and funded by the Northern Ireland Central Community Relations Unit; and the 1998 Northern Ireland Referendum and Election Survey was conducted by John Curtice, Lizanne Dowds, Geoffrey Evans and Bernadette C. Hayes and funded by the Economic and Social Research Council. An earlier version of the paper was presented to the meetings of the UK Political Studies Association, London School of Economics and Political Science, 10–13 April 2000. Our thanks to an anonymous referee from this journal for their constructive comments. 1 Estimates of the numbers of deaths due to political violence vary. The most reliable statistics concerning deaths in Northern Ireland come from the Royal Ulster Constabulary (http://www.ruc.police.uk/main/dept.htm). However they do not include deaths due to the conflict outside Northern Ireland; deaths indirectly the result of violence (such as heart attacks); nor do they provide details of the agencies responsible or the status of the person killed. Sutton (1994, and updates at http://cain.ulst.ac.uk) provides these additional details. 2 Estimates of deaths during the Irish Civil War vary considerably, from 600 to 4,000 (O’Leary and McGarry, 1993, p. 21). 3 In 1984, the British government changed the name of the city council from ‘Londonderry’ to ‘Derry’, although the official name of the city and county remains Londonderry. However, whereas the overwhelming number of nationalists, including John Hume, refer to the city as ‘Derry’, the unionist community continues to use its official title, ‘Londonderry’ (Elliott and Flackes, 1999, p. 229). 4 See also Fay et al. (1999) for a detailed analysis of the patterns of deaths. Qualitative analyses of the deaths and their impacts on families and friends can be found in McKittrick et al., (1998) and Van Voris (1975). 5 The Provisional IRA (PIRA), the dominant element in the Northern Ireland violence, dates effectively from December 1969, when the Republican movement split into two branches, those who wished to maintain the military campaign (Provisional IRA) versus those who favoured moving to a more political strategy, such as the election of party members to the parliaments in Dublin, Belfast and London (Official IRA). Mirroring this event, in January 1970, Sinn Fein also split into Official Sinn Fein and Provisional Sinn Fein. The first president of Provisional Sinn Fein was Ruari O Bradaigh, who, on his resignation in 1983, was replaced by Gerry Adams (see White, 1997). 6 The UDA was eventually declared illegal in August 1992. 7 This is not to deny the view among some that the state perpetuates the conflict. See Ni Aolain (2000) for a comprehensive argument in favour of this position. 8 Two types of political violence, house searches and witnessing car hijacks, are not examined here because corresponding data were not collected in all of the main surveys. 9 The correlations between these items in the 1998 survey are as follows: (Correlations, decimals omitted) 1

2

3

Direct experience 1. Intimidated 2. Victim of violent incident

48

Indirect experience 3. Family/relative killed/injured 4. Know someone killed/injured

16 21

21 27

28

Collective exposure 5. Caught up in explosion 6. Caught up in riot

19 32

29 39

18 23

4

5

27 34

38

6

10 Estimating the models separately for each of the three types of violence shows that the effect comes about through exposure to personal violence, suggesting that the second explanation is the most salient. 11 The actual statement, made by Danny Morrison at the Sinn Fein Ard Fheis in 1981, was ‘Who here really believes we can win the war through the ballot box? But will anyone object if, with a ballot paper in one hand and the armalite in the other, we take power in Ireland?’ (quoted in Taylor, 1997, p. 328). 12 There were no questions about support for the use of violence asked in the opinion surveys between 1978 and 1998. Several surveys did ask about the right to march, but these are not considered comparable to the questions in Tables 7 and 9. 13 This assumes an adult population of about 1,280,000, divided 57 per cent Protestant and 43 per cent Catholic (this is estimated from the 1991 census, attributing the 11.0 per cent who either answered none or were not stated, in proportion to the two communities).

PUB LI C S UPP O R T F OR V I O L E N C E I N N O RTH ERN IRELA N D

921

14 As one Catholic noted, ‘On the one hand, if put to the wall you would disapprove of what they were doing – but the songs allowed you to participate emotionally in some part of their activities. I suppose that’s the purpose of rebel songs: part of your inheritance, the emotional ambiguity that allows you to take part …’ (quoted in O’Connor, 1993, p. 106). 15 The rank ordering of the proposals (together with the per cent supporting) was: decommissioning (92 per cent); the guarantee that Northern Ireland will remain part of the UK for as long as a majority of people in Northern Ireland wish it to be so (89 per cent); the setting up of a Northern Ireland Assembly (87 per cent); the requirement that the new executive is powersharing (77 per cent); the removal of the Republic of Ireland’s constitutional claim to Northern Ireland (69 per cent); the creation of north-south bodies (67 per cent); the creation of a commission into the future of the RUC (54 per cent); and the early release of prisoners (19 per cent).

References Adams, G. (1996) Before the Dawn: an Autobiography. London: Heinemann. Bairner, A. (1996) ‘Paramilitarism’, in Arthur Aughey and Duncan Morrow (eds), Northern Ireland Politics. London: Longman. Bardon, J. (1992) A History of Ulster. Belfast: Blackstaff. Boal, F. W., Murray, R. C. and Poole, M. A. (1976) ‘Belfast: the Urban Encapsulation of a National Conflict’, in S. C. Clarke and J. L. Obler (eds), Urban Ethnic Conflict: a Comparative Perspective. Chapel Hill NC: Institute for Research in Social Science, University of North Carolina. Brewer, J. D., Lockhart, W. and Rodgers, P. (1999) ‘Crime in Ireland 1945–95’, in A. Heath, R. Breen and P. Whelan (eds), Ireland North and South: Perspectives from the Social Sciences. Oxford: Oxford University Press. Bruce S. (1992) The Red Hand: Protestant Paramilitaries in Northern Ireland. Oxford: Oxford University Press. Bruce, S. (1997) ‘Victim selection in ethnic conflict: motives and attitudes in Irish republicanism’, Terrorism and Political Violence, 9, 56–71. Budge, I. and O’Leary, C. (1973) Belfast: Approach to Crisis. London: Macmillan. Clark, S. and Donnelly, J. (eds) (1983) Irish Peasants: Violence and Political Unrest, 1780–1914. Manchester: Manchester University Press. Crawford, C. (1999) Defenders or Criminals? Loyalist Prisoners and Criminalisation. Belfast: Blackstaff. Elliott, M. (1982) Partners in Revolution. New Haven CT: Yale University Press. Elliott, S. and Flackes, W. D. (1999) Northern Ireland: a Political Dictionary, 1968–99. Belfast: Blackstaff Press. Farrell, B. (1973) The Irish Parliamentary Tradition. Dublin: Gill and Macmillan. Fay, M.-T., Morrissey, M. and Smyth, M. (1999) Northern Ireland’s Troubles: the Human Cost. London: Pluto. Guelke, A. (1999) ‘Political Violence and Paramilitaries’, in P. Mitchell and R. Wilford (eds), Politics in Northern Ireland. Boulder CO: Westview. Hayes, B. C. and McAllister, I. (1999) ‘Generations, Prejudice and Politics in Northern Ireland’, in A. Heath, R. Breen and C. Whelan (eds), Ireland North and South: Perspectives from the Social Sciences. Oxford: Oxford University Press. Lyons, F. S. L. (1973) ‘Charles Steward Parnell’, in B. Farrell (ed.), The Irish Parliamentary Tradition. Dublin: Gill and Macmillan. McGarry, J. and O’Leary, B. (1995) Explaining Northern Ireland. Oxford: Blackwell. McKittrick, D., Kelters, S., Feeney, B. and Thornton, C. (1998) Lost Lives: the Stories of the Men, Women and Children Who Died as a Result of the Northern Ireland Troubles. London: Mainstream. Mitchell, G. J. (1999) Making Peace. New York: Knopf. Munger, F. (1975) The Legitimacy of Opposition: the Change of Government in Ireland in 1932. Beverley Hills CA: Sage. Murray, R. (1982) ‘Political Violence in Northern Ireland, 1969–77’, in F. W. Boal and J. N. H. Douglas (eds), Integration and Division: Geographical Perspectives on the Northern Ireland Problem. London: Academic Press. National Graves Association (1976) The Last Post: the Details and Stories of Republican Dead, 1913–75, 2nd ed. Dublin: National Graves Association. Nelson, S. (1984) Ulster’s Uncertain Defenders. Belfast: Appletree.

922

B E R N A DE TTE C. H A YES, IA N M C A LLISTER

Ni Aolain, F. (2000) The Politics of Force: Conflict Management and State Violence in Northern Ireland. Belfast: Blackstaff. O’Brien, B. (1999) The Long War: The IRA and Sinn Fein, 2nd ed. Syracuse NY: Syracuse University Press. O’Connor, F. (1993) In Search of a State: Catholics in Northern Ireland. Belfast: Blackstaff. O’Duffy, B. (1995) ‘Violence in Northern Ireland 1969–94: sectarian or ethno-national’, Ethnic and Racial Studies, 18, 740–71. O’Leary, B. and Arthur, P. (1990) ‘Northern Ireland as a Site of State and National-Building Failures’, in J. McGarry and B. O’Leary (eds), The Future of Northern Ireland. Oxford: Clarendon Press. O’Leary, B. and McGarry, J. (1993) The Politics of Antagonism: Understanding Northern Ireland. London: Athlone Press. Poole, M. (1983) ‘The Demography of Violence’, in J. Darby (ed.), Northern Ireland: Background to the Conflict. Belfast: Appletree. Poole, M. (1993) ‘The Spatial Distribution of Political Violence in Northern Ireland: an Update to 1993’, in A. O’Day (ed.), Terrorism’s Laboratory: the Case of Northern Ireland. Aldershot: Dartmouth. Rose, R. (1971) Governing without Consensus: an Irish Perspective. London: Faber. Ruane, J. and Todd, J. (1999) ‘The Belfast Agreement: Context, Content and Consequences’, in J. Ruane and J. Todd (eds), After the Good Friday Agreement: Analyzing Political Change in Northern Ireland. Dublin: University College Dublin Press. Schellenberg, J. A. (1977) ‘Area variations of violence in Northern Ireland’, Sociological Focus, 10, 69–78. Senior, H. (1966) Orangeism in Ireland and Britain. London: Routledge and Kegan Paul. Sullivan, S. (1998) ‘From Theory to Practice: the Patterns of Violence in Northern Ireland 1969–94’, Irish Political Studies, 13, 76–99. Sutton, M. (1994) An Index of Deaths from the Conflict in Ireland, 1969–93. Belfast: Beyond the Pale. Taylor, P. (1997) Behind the Mask: the IRA and Sinn Fein. New York: TV Books. Taylor, P. (1999) Loyalists: War and Peace in Northern Ireland. New York: TV Books. Thompson, W. and Mulholland, B. (1995) ‘Paramilitary Punishments and Young People in West Belfast’, in Liam Kennedy (ed.), Crime and Punishment in West Belfast. Belfast: The Summer School, West Belfast. Todd, J. (1999) ‘Nationalism, Republicanism and the Good Friday Agreement’, in J. Ruane and J. Todd (eds), After the Good Friday Agreement: Analyzing Political Change in Northern Ireland. Dublin: University College Dublin Press. US Secretary of State (2001) Patterns of Global Terrorism 2000. Washington: Department of State. Van Voris, W. H. (1975) Violence in Ulster: an Oral Documentary. Massachusetts MA: University of Massachusetts Press. White, R. W. (1997) ‘The Irish Republican Army: an Assessment of Sectarianism’, Terrorism and Political Violence, 9, 20–55. Whyte, J. (1990) Interpreting Northern Ireland, Oxford: Oxford University Press. Williams, T. D. (ed.) (1973) Secret Societies in Ireland. Dublin: Gill and Macmillan.