To understand why some groups fighting in civil conflicts target civilians more than others, it is vital to examine the role of ideology.

Recent civil wars in Iraq and Syria underscore the fact that different armed groups fighting in the same conflict can adopt strikingly different approaches to the treatment of civilians. While historically about 40 percent of states and rebels have exercised restraint, others’ victimization of civilians has been routine and manifold. Much of the current understanding attributes these differences to armed groups’ material resources, organization, territorial control, and similar factors. While providing important insights, these accounts are incomplete at best because they either neglect or downplay the critical role that ideology plays in targeting civilians.

Mainstream Explanations of Civilian Victimization

Many analysts share a key assumption with the classical literature on insurgency and counter-insurgency – such as the works of T.E. Lawrence and Mao Tse-Tung – that securing the support of local populations is critical for fighting groups. Several implications are drawn. For example, groups that enjoy local population’s support may be less prone to victimize civilians, particularly in the communities that serve as their home or recruitment base. Conversely, such support can backfire because the enemy forces can attempt to raise its costs for the local population by targeting civilians in this community.

Based primarily on the study of the Greek Civil War (1942–1949), another influential account argues that groups with higher degree of control over a territory are likely to selectively target enemy forces rather than indiscriminately attack civilians. In this view, information flows are pivotal: higher degree of territorial control means better information and this makes it possible and expedient to identify and selectively target specific individuals.

Another assumption is that the fighting groups’ capabilities – their size, training, and experience – relative to their rivals can also affect their targeting patterns. Weak groups can be more prone to victimize civilians than stronger groups as they can fail to limit the collateral damage of their operations to civilians. Alternatively, when they lack resources to secure civilian support through offering benefits, such groups can also deliberately choose to target civilians as an alternative way to coerce such support. An analysis of violence in Afghanistan between 2004 and 2009, disaggregated by province and month, supports this view.

The amount of material resources at the armed group’s disposal and where it obtains them may also affect its targeting patterns. If groups acquire their resources through exploiting natural resources or through foreign sponsorship, they may be more likely to attack civilians than if they depended on the local population for resources. External donor characteristics can matter as well, with a small number of democratic donors believed to have a more restraining effect than either autocratic donors or many donors.

A survey of former fighters in Sierra Leone’s civil war (1991–2002) suggests that organizational characteristics are the pivotal factor. According to this view, civilian abuse is likely to be higher if groups rely on material incentives in their recruitment (thus attracting more opportunists), have an ethnically diverse group of fighters (thus lacking ways to control the fighters’ behavior through social pressure), and lack disciplinary mechanisms. This view resonates with Niccolo Machiavelli’s aversion toward mercenaries and Mao Tse-Tung’s insistence that “it is only undisciplined troops who make the people their enemies.”

More recently, some analysts have drawn attention to the role of political and ethnic cleavages, which had previously been downplayed in large cross-national studies of civil wars. Based on a study of Spanish Civil War (1936–1939), one view maintains that civilians who have mobilized for one belligerent group are likely to be attacked by rivals as they would be considered assets for this group. A study of violence against civilians in African conflicts between 1989 and 2009 holds that ethnic background can serve as a cue for targeting because in an environment of uncertainty about people’s allegiances it serves as a shortcut for identifying potential enemy supporters.

Much of this thinking on civil wars has tended to relegate ideology to a secondary role, if any at all. Yet, a notable resurgence of attention shows just how consequential ideology can be in understanding civil wars.

 Why Ideology Matters

Research shows that in newly democratizing countries, a combination of nationalist ideology and unconsolidated democracy can help ignite internal and international conflict in the first place. Revolutionary ideologies have historically been critical ingredients for a robust insurgency by fostering strong commitment and mobilization. They can enhance fighting capacity by boosting morale. Although often based on case studies of specific groups or lacking in-depth systematic evidence, qualitative literature on civil wars and traditional research on terrorism historically have pointed at the role of ideology in armed groups’ target selection (one solid case study can be found here).

In a recent study our working hypothesis is that far from being a mere rhetorical device, ideology can be the key factor that explains civilian victimization patterns across fighting groups in civil wars. We draw on the established concept of ideology which sees it as “shared framework of mental models that groups of individuals possess that provides both an interpretation of the environment and a prescription as to how that environment should be structured”. We argue that its effect on civilian victimization can work through two channels.

The first is through framing some groups as hostile to the armed groups’ cause. The ideology that an armed group espouses identifies the group’s vision and the sources of threats to achieving this vision. These threat perceptions foster identifying friends and enemies of the cause. The ideology can then frame “enemies” as legitimate targets. Belonging to a certain ethnicity or territory may be a marker, but it need not be just any civilian from within these groups that becomes a legitimate target – only those can be identified as such who are seen through the ideological prism as hostile to the group’s cause.

However, there is nothing automatic between seeing members of a specific group as hostile and victimizing them. The second channel through which ideology can affect civilian victimization is through determining strategies that the group accepts as legitimate in achieving its vision. Of course, in some cases different types of violence may be included or excluded for strategic reasons. But a group’s ideology can also prescribe adopting a strategy that is costly for the group from the material or organizational point of view. That is, some strategies may be filtered out despite presenting strategic or material advantages. This is probably because they go against the group’s vision or its identity as a certain ideological force. Therefore, some ideologies will see civilian victimization as part of their legitimate repertoire of violence to attain its vision, while others will impose constraints on or even exclude it from the group’s approach.

It might be tempting to follow this reasoning by drawing a typology of ideologies by their approach to civilian victimization. However, often broad ideological frameworks are adapted to local conditions – they are crystallized into specific ideologies that different groups adopt. In other words, groups in different contexts that seem to share an ideology may develop different approaches, such as European leftist groups in 1970-1990s like Baader-Meinhof Group in West Germany and the Red Brigades in Italy. Instead, we should understand an armed group’s ideology in its particular context.

Case Study: Northern Ireland

In our study, we examined these ideas using quantitative and qualitative empirical evidence on armed group violence in Northern Ireland’s conflict between 1969 and 2005.  This conflict provides a fertile ground for this study because it involved a number of groups that differed from one another in several ways and because there were considerable differences in civilian killings across groups, locations and time. Unlike many other conflicts, it has also been well documented on the level of individual fatalities, which makes it possible to test our ideas with more nuance than previous studies. Our dataset provides details on almost all fatalities directly attributed to the conflict (3,702) and allows disaggregating them by perpetrator group, location, victim’s identity, etc. Then we try to see whether different perpetrator group characteristics, such as their size, structure, or ideology, consistently predict whether the victim is civilian or combatant as well as the victim’s ethnic identity. We do this in a framework that simultaneously accounts for all suggested factors.

While the two main ideologies embraced by the fighting groups – Irish Republicanism and Unionism – shared similarities, such as the focus on nationalism, historically they developed distinct approaches. Drawing on the civic nationalist ideology of the French Revolution, Irish Republicanism stressed the oppression of all Irish people and adopted an anti-colonialist identity that aimed to end imperial control. This entailed a reluctance to target would-be members of the “imagined community” of free Ireland and instead emphasized focusing on combatants who were viewed as struggling to preserve imperial domination.

Drawing on the historical “Protestant Ascendancy” movement, the ideology of Unionism came to emphasize a defensive settler identity that viewed Catholics as “fifth-columnist” Irish nationalists who intend to dismantle Northern Ireland and its union with Britain. We conjectured that these ideological differences were likely to shape the fighting groups’ targeting patterns. While the group ideologies were further crystallized during the course of the conflict, their key tenets remained.

Our preliminary findings from statistical analysis suggest that the fighting group ideologies were the strongest and most consistent predictors of civilian victimization patterns. Fighting groups that embraced Unionist ideology, such as Ulster Defence Association (UDA) or Ulster Volunteer Force (UVF), were on average more likely to target civilians and launch cross-ethnic attacks on civilians, while Republican fighting groups, such as the Provisional Irish Republican Army (Provisional IRA) or Irish National Liberation Army (INLA), were on average more likely to focus on combatants. These results hold when we account for all other suggested factors, such as group size or resources. Our qualitative historical study suggests that these differences emerged because of differences in previously adopted norms, patterns of recruitment, and relations with the British state forces.

While civilian targeting was prevalent in the initial stages of the conflict in early 1970s, over time it decreased in terms of total numbers. Republican groups were responsible for the largest number of total fatalities and Loyalist paramilitaries were responsible for the largest number of civilian killings. After mid-1970s, all three armed blocs – state forces, Unionist groups, and Irish Republican groups – killed fewer civilians than before, but relative proportions (combatant-civilian) remained the same.

Implications

Naturally, our study may be limited by its focus on civilian killings rather than civilian abuse more generally or our focus on one civil war. Nonetheless, our tentative findings strongly indicate that ideological factors need to be taken much more seriously than before in trying to understand and hopefully prevent civilian victimization by armed groups in civil wars. Neglecting these factors or downplaying their significance is simply dangerous. This is all the more important at the time when transmission of ideas is considerably enhanced by technology, which does not discriminate between benevolent or harmful ideologies. This, for example, is most drastically illustrated by Isil’s media-savvy, effective propaganda.

Anar K. Ahmadov is Assistant Professor of Political Economy at Leiden University.

James Hughes is Professor of Comparative Politics and Director of the Conflict Research Group at the London School of Economics and Political Science (LSE).