Accidental Escalation: Missing Mechanism or a Challenge to the Model?

Both the contentious politics and interactionist approaches of explaining political violence arose initially through close analysis of social movement case studies that focused on the specific context in which the violence developed. The emergence of Islamic extremism and the wide diversity of forms in which political violence is enacted today have challenged social movement scholars to move beyond case studies. This has led to the contentious politics causal model based on a sequence of escalating mechanisms that aim to explain the turn to violence among different types of political actors, across organizations, cultures, geographic regions, and time. We argue that a major limitation of understanding political violence through a global causal model is the lack of attention to accidental contingencies, which can lead to misplaced assumptions about intentionality. Using examples from our long-term study of Japanese and American armed clandestine groups that emerged out of the New Left protest cycle of the late 1960s, this study explores the role of accidental occurrences in determining the course of political violence and responses to it. We argue that accidental escalation creates points of no return that affect the behavior of the clandestine group. We examine accidents that produced (1) serious unintended outcomes; (2) alterations in the composition and interpersonal dynamics inside the organization; (3) new strains or new alliances between organizations; (4) dramatic changes in the life course of individuals; (5) and miscalculations by the media, governments, and the public concerning the likelihood and nature of subsequent actions that shape profiles of the group as well as future threat assessment, prosecution, and correctional strategies. Rather than simply constituting another causal mechanism that is added to the contentious politics model, accidental escalation challenges the model both theoretically and methodologically. Methodologically, information about accidental escalation often comes to light only later, when those involved are able to tell their own stories about what happened and how it unfolded over time. Theoretically, even if the causal model incorporates interaction among different actors and adds an error term for accidents, it cannot predict how participants may have responded to unexpected contingencies, especially if these are unknown until much later. Absent the elements of contextual meaning and contingent response that are central to the interactionist approach, the causal model is misleadingly deterministic.