Each year, Protestant organizations in Northern Ireland perform over 2,500 ritual parades to celebrate and commemorate their culture. Many Catholics, however, see parades as triumphalist and hateful. As a result, parades undermine the political peace process and grassroots peace-building by raising interethnic tension and precipitating riots, including significant violence in recent years. This dissertation asks: Why do people participate in these parades? To answer this question, I consider loyalist parading as an example of contentious ritual--symbolic action that makes contested political claims. To understand these parades as ritual actions, I build on two central insights from religious studies, sociology, and anthropology. First, as meaningful and shared practices, rituals provide participants with benefits that are intrinsic to participating in the act itself and do not depend on the achievement of some external outcome. Second, rituals are multi-vocal, meaning that interpretations of the action can vary across actors. Participants need not share the interpretation of their actions held by organizers, rivals, or outside observers. Participants, therefore, may not see the ritual as provocative, aggressive, or even contentious. These arguments stand in contrast to traditional explanations for collective action and ethnic conflict that theorize participation in ethnically polarizing events in terms of the achievement of concrete outcomes, such as selective material benefits, provoking the out-group into overreacting, or intimidating them into quiescence. To test my argument, I conducted fieldwork in Belfast, Northern Ireland. I developed and implemented a household survey to measure mass-level opinion, designed and ran an online survey of all Protestant clergy and elected officials in Northern Ireland to measure elite-level opinion, conducted over 80 semi-structured interviews with parade participants and nonparticipants, and observed dozens of hours of parades and related events. I demonstrate that, as expected by my argument, people approach participation in ritual parades as an end in and of itself. The evidence demonstrates that participants do not view parades instrumentally. This means that people make decisions to participate in contentious behavior without consideration of their actions' profoundly political consequences. The ritual nature of parades severs the expected connection between means (participation) and ends (political consequences), thus creating the environment for sustained conflict. Furthermore, the predictions of influential theories of ethnic conflict--extreme in-group identification or out-group antipathy--and collective action--selective material benefits or sanctions--are not supported by the data.