| What explains the use of sexual violence during civil wars? In this dissertation, I develop a theory of why some combatant groups rape noncombatants, while others refrain from sexual violence. I argue that combatant groups that recruit new members through forcible means, such as coercion or press-ganging, face a central dilemma; namely, such groups must create a coherent fighting force out of a collection of strangers, many of whom had to be abused in order to compel them to join. Drawing from literatures in economics, sociology and criminology, I argue that sexual violence, and particularly gang rape, enables armed groups with forcibly recruited fighters to create bonds of loyalty and friendship from these initial circumstances of fear and mistrust. The theory was developed during fieldwork in Sierra Leone, where I interviewed both ex-combatants and noncombatants about their experiences with wartime violence. The interviews suggest a set of mechanisms that are distinct from the conventional wisdom on the causes of wartime sexual violence, mainly derived from the patterns of rape in the ethnic wars in Bosnia and Rwanda.;The theory can help to explain some of the most puzzling aspects of rape during conflict. For example, why does wartime rape often take the form of gang rape, when gang rape is a relatively rare form of peacetime violence? If wartime rape is often described as a powerful "tool of war," why is there not stronger evidence that rape is ordered by commanding officers as part of a military strategy? How do we account for the involvement of female perpetrators in the rape of non-combatant victims, as has been documented in a range of recent conflicts? After explicating the theory and its predictions, along with a set of competing hypotheses, I devote a chapter to the methods, challenges and ethics of researching wartime sexual violence.;I explore the observable implications of the theory in several ways. First, I introduce a new cross-national dataset of sexual violence during civil wars between 1980--1999. The dataset enables the documentation of the variation in the incidence of sexual violence across perpetrator groups, time, geographic region and intensity. Using these data, I demonstrate statistically that the argument derived in Sierra Leone also holds for a large cross-national sample. I find that for both insurgent groups and state armed forces, the form of recruitment mechanism has important effects on the perpetration of wartime sexual violence. Specifically, I find that insurgent groups that forcibly recruit their fighters, and states that use press-ganging to garner members, are more likely to commit wartime sexual violence, even controlling for potentially confounding variables. I also find no or limited support for several common explanations of wartime sexual violence, including ethnic war, gender inequality and genocide.;Two additional chapters feature case studies and fieldwork in El Salvador and East Timor that serve to deepen the analysis, and to illustrate the potential utility of the argument across types of conflicts and perpetrator groups. In a final empirical chapter, I explore the participation of women in acts of wartime gang rape, and argue that while the existence of female perpetrators of rape undermines many competing arguments, the theory presented here can help to explain the phenomenon. I conclude with a discussion of the implications of the findings, both for theories of wartime violence and for policymakers. I also briefly address the potential application of the argument to interstate wars. |
