M.MAS 1/12/17 Finally, as with the other cases, interviewees mentioned a desire to avoid conflict. One person explained “If there is a conflict, there will be many consequences. People will kill each other and die. There will be no benefit. It is worse for people to die than for cows or animals to die.” 44 Conclusions All of these cases represent situations in which there appeared to be a strong possibility that localized conflict between individuals or small groups could escalate into riotous violence between religious communities. Several of them occurred at times when conflicts in other parts of the country had produced heightened tensions—and fears that localized disputes could lead to wider repercussions. It is important to repeat that the factors we have identified in these cases as helping to reduce communal tension and stave off wider violence do not make up a universal “recipe” for responding to religious instigation and potential riots. Even this small number of cases reveal a number of situation-specific factors and those that can be generalized (such as the necessary role of trusted interlocutors) cannot necessarily be directly implemented. We also cannot prove the counterfactual that the actions described in each case prevented violence that would have otherwise occurred. But in each case, those we interviewed believed strongly that such escalation was a serious risk. At a minimum, escalating violence was a distinct possibility in line with the methodological approach of Mahoney and Goertz, explained above. Based on our research into these cases and the analysis offered by those we interviewed, we provide four sets of conclusions that we believe are instructive for anyone interested in promoting peace and preventing inter-religious violence in Myanmar. These conclusions relate to (1) the operational dynamics of each case (the individuals and groups involved and the contestations between them); (2) the broader framing of the conflicts; (3) the kinds of competing arguments made by those attempting to prevent violence; and (4) concerns about forms of violence left unaddressed by these interventions. Contests and actors Each of the four cases discussed in this paper could have escalated into wider violence, but did not. From our perspective, what is most important to highlight is that each of these cases entailed a contest, between and among both individuals and groups operating with their own ideas about what was good and right. Such an observation might border on the banal, but it underpins two corollary observations. First, each case included activities by some organized groups seeking ends that risked escalating the situation into wider violence. Collective violence may not have been the outcome they sought; this paper does not attempt to prove the existence in Myanmar of the “institutionalized riot systems” Brass studied in north India. But, whatever the goals of the groups involved in these four cases, their efforts made such violence probable. In Case Study 1, interviewees indicated that, at least on social media, the rumours of weapons stockpiling and preparation to attack Muslims appeared to be shared and promoted in a systematic fashion. We know little about who might have been organising this information dissemination, though interviewees felt it was clearly more than one individual. In the other three cases, however, the level of organization appears to have been stronger and more transparently institutionalized. In the last two cases, for example, principle protagonists were involved or affiliated with groups such as 969 and Ma Ba Tha. To be clear, if riots had indeed transpired, this paper is not arguing that 969 and Ma Ba Tha members would have been the leaders or that their intentions were to instigate a riot. But, whatever their 44 Female, 31 years old, Christian, local activist. 15/24