22 M. SCHISSLER of Facebook transformed the in-group policing that did occur, with significant conse­ quences: it offered a unique capability to transform threats into public displays. While threatening messages against individuals and their families were conveyed using multiple channels, Meta’s de facto monopoly in Myanmar meant that Facebook was the primary means by which such messages could also address a larger audience. Such displays had to be taken seriously, because there were good reasons to believe both that nation­ alist forces were prepared to act on them and that protection or assistance from state authorities was unlikely. People also had to remain alert to the way such displays could also generate public controversies that would provide state authorities with the impetus, or excuse, to act against them. These concerns were relevant for many more people than just those named in a given post on Facebook: such displays enunciated a warning, backed by clear power, for everyone to heed or ignore at their own peril. This article has argued that limiting concern for the relationship between Facebook and violence to “hate speech” and other content about the Rohingya and other Muslims radically narrows the scope of the problem. Such content is an important matter of concern, based on commitments to values such as dignity, justice, and truth. That alone should be enough to warrant action. The question of what helped cause vio­ lence, on the other hand, should direct attention to a broader range of conduct and content on Facebook. In formalist terms, this range can be differentiated based upon (1) who it referenced, (2) its predication, and (3) its function. In these terms, extant analysis can be said to have focused primarily on (1) content that had the out-group Rohingya and other Muslims as its object of reference, (2) predicated them negatively, and (3) served what could be called a characterization function. It constructed narratives about the Rohingya that were used to justify genocide. This article points towards additional forms of conduct and content that (1) are not necessarily about the out-group at all, but about the in-group and (2) predicates members of the in-group negatively, such as by challenging their loyalty and claiming they are “traitors,” serving (3) the function of in-group policing. To be clear, speech can always serve multiple functions. Posts on Facebook with the out-group as the object of reference, for example, can serve a characterization function while also contributing to in-group policing, such as by signalling affiliation and affirming the existence of a public mobilizing against the Rohingya.89 Sometimes speech about the out-group cleverly avoided the vulgarity and directness of hate speech, which could seem to reiterate claims to moral superiority on behalf of the ingroup.90 Other times, when Facebook content about the out-group was more direct, even vulgar and aggressive, it served as a clear performance of power and impunity, especially in comparison to the negative consequences faced by those who tried to oppose it.91 Other channels and material forms should also be similarly analyzed for their multiple functions, both in-group policing and characterization of the out-group, 89 90 91 Prasse-Freeman, “Hate Bait”; Stefan Bächtold, “Blackouts, Whitelists, and ‘Terrorist Others’: The Role of Socio-Technical Imaginaries in Myanmar,” Journal of Intervention and Statebuilding (2023): 1–21; Gerard McCarthy, “Cyber-Spaces,” in Routledge Handbook of Contemporary Myanmar, ed. Adam Simpson, Ian Holliday, and Nicholas Farrelly (New York: Routledge, 2017), 92–105. Prasse-Freeman, “Hate Bait,” 184. Thant Sin Oo, Zaw Myo Min, and Matt Schissler, “ချစ်စကို ရှည်စေ၊ မုန်းစကို တိုစေ: Structures and Vocabularies of Counter-Speech on Facebook in Myanmar,” Working Paper (Bangkok: Article 19, November 2020), 15–6.