The question of how mass atrocities end has been dominated by a normative approach regarding how they ought to end. Arguing that an evidence-based approach to terminate mass atrocities might offer profound insights into theories of mass atrocities as well as policies designed to prevent or end their occurrence, this article outlines the key questions and approaches needed for an evidence-based study of atrocity endings. It draws on theories of genocide, political violence and civil war termination, and presents initial insights from case studies, including the killing of civilians in colonial German Southwest Africa, the Soviet Union, the Nigerian civil war, the Guatemalan civil war, the Nuba Mountains of Sudan and Bosnia-Herzegovina.
IntroductionThe field of genocide and mass atrocities studies has produced significant contributions to knowledge of where, when and why campaigns of large-scale, one-sided violence occur, but offers relatively few explicit examinations of the political, social and military dynamics of the de-escalation of violence. This simple question remains unexplored: how do mass atrocities end?Answering this question will require study of the dynamics of endings in multiple cases and will not likely produce a single theory of termination. 1 This article serves a more modest role; it sketches out the starting point for an evidence-based study of endings. It begins with a brief overview of why both 'mass atrocities' and 'endings' need to be problematized and explores how various theories of genocide and civil war might inform a study of atrocity endings. The second half draws on short case studies to introduce three broad scenarios for endings. (1) Perpetrators change their policies because of outside influence, rise of moderates or resistance. (2) Perpetrators carry their plans to fruition; leaders decrease use of lethal violence in order to normalize conditions for a range of reasons. (3) The most rare ending is that perpetrators are defeated by interested parties or so-called humanitarian interventions.