Doctoral Dissertation Research: Capabilities and Constraints: Untangling State and Rebel Strategies of Violence during Civil Conflict

在20世纪，平民占内战伤亡人数的80%以上。虽然其中许多是无意的，但许多冲突的残酷性质表明，政府和叛乱组织经常对平民采取明确和严重的暴力策略。政府军和叛乱组织对非战斗人员实施的强奸、酷刑、失踪和滥杀滥伤似乎是许多现代内战的共同特征。然而，针对平民的规模和范围在武装冲突之间和武装冲突内部差别很大。叛乱分子和现任政府依靠平民的忠诚和支持来实现他们的战争目标。然而，在国内冲突期间，双方有时对非战斗人员进行大规模虐待。然而，以平民为目标可能适得其反，将平民推向对手的怀抱。这就提出了一个重要的经验问题：是什么促使政治行为者以平民为目标，而这样做往往似乎是一种次优策略？本项目通过调查国家、叛乱分子和平民采取的战略演变，研究了内战期间针对非战斗人员的暴力行为。本文认为，叛乱分子和国家支持的暴力行为的频率和严重程度的变化是国家和反叛力量之间战略互动以及这些行动者之间相对能力平衡的函数。简言之，一个行动者的暴力战略与其吸引平民支持的能力有关，而平民支持又取决于其向这些支持者提供利益的能力。弱小（或正在减弱）的行动者不太可能有能力提供足以获得支持的利益。因此，他们更有可能采用暴力作为暂时强制支持和强制忠诚的手段。然而，这种关系取决于对手所采用的策略。对手使用的高度暴力可能会降低行动者吸引支持所需的能力水平。这个项目增加了最近的研究，通过探索如何叛军和国家的暴力战略是构建和这些战略如何演变和变化，随着时间的推移冲突展开解释内战暴力。它还有助于理解这些战略的微观基础，因为它揭示了平民对叛军和政府的支持、行动者的能力以及双方选择的暴力程度之间的联系。此外，该项目将产生迄今为止最详细的内战暴力数据集。本文使用的数据集记录了叛乱分子和政府军每年犯下的暴力行为的水平、频率和类型。确定哪些因素导致了暴力的变化，并揭示行为者之间复杂的相互作用如何影响战略选择，这是设计能够减轻平民受害的政策的必要的第一步。本项目的理论和经验贡献将有助于决策者和学者制定减少内战平民代价的战略。此外，为测试上述假设而构建的数据集将对决策者和学者有用，他们需要关于内战行为者所犯暴力行为的纵向和横向数据

During the 20th century civilians constituted upwards of 80% of the casualties in civil wars. While many of these are unintended, the brutal nature of many conflicts illustrates that governments and insurgent groups often employ strategies of explicit and gross violence against civilians. Rape, torture, disappearances, and indiscriminate killings against non-combatants by both government forces and insurgent organizations appear to be common characteristics of many modern civil wars. Yet the scale and scope of civilian targeting varies tremendously both across and within armed conflicts.Insurgents and incumbent governments depend on civilian loyalty and support to achieve their war objectives. Yet, during civil conflicts both sides at times engage in mass abuses against non-combatants. Targeting civilians, however, may prove counterproductive, driving civilians into the adversary's arms. This raises an important empirical question: what motivates political actors to target civilians when doing so often appears to be a suboptimal strategy? This project examines violence against non-combatants during civil conflict by investigating the evolution of strategies adopted by states, rebels, and civilians. It is argued that changes in the frequency and severity of insurgent- and state-sponsored violence are a function of the strategic interactions between state and rebel forces and the balance of relative capabilities between these actors. In brief, an actor's strategy of violence is related to its ability to entice civilian support, which is in turn contingent upon its ability to provide benefits to these supporters. Weak (or weakening) actors are unlikely to possess the capacity to provide benefits sufficient to elicit support. As a result, they are more likely to adopt violence as a means to temporarily coerce support and enforce loyalty. This relationship, however, is conditioned by the opponent's adopted strategies. High levels of violence employed by adversaries may reduce the level of capabilities necessary for an actor to entice support. This project adds to recent research explaining civil war violence by exploring both how rebel and state strategies of violence are constructed and how these strategies evolve and change as conflict unfolds over time. It also contributes to an understanding of the microfoundations of these strategies by unpacking the linkages between civilian support for rebels and government, actor capacities, and the level of violence chosen by each side. Furthermore, this project will produce the most detailed dataset to date on civil war violence. The dataset used herein captures the level, frequency, and type of violence committed annually by both insurgent and state forces.Determining what factors are responsible for variations in violence and unpacking the manner in which the complex interactions among actors influence strategy choice is a necessary first step to designing policies that can mitigate civilian victimization. The theoretical and empirical contributions of this project will assist policy makers and academics in devising strategies that diminish the civilian cost of civil wars. Furthermore, the dataset constructed to test the hypotheses outlined above will be useful to both policymakers and academics who require longitudinal and cross sectional data on violence committed by civil war actors