Why do some ethnopolitical minority organizations use violence to achieve their political goals, whereas others eschew force and engage in nonviolence or take part in elections? The literature leads us to expect that the more fragmented the ethnic minority group is, the more likely it is that ethnopolitical minority organizations will use violence against the state. Ethnopolitical minority organizations, however, vary considerably in their strategies. To explain this puzzle, I argue that an under-explored factor - fragmentation within ethnically mobilized groups that control the state - affects how minority organizations select their strategies. Using two original measures of majority fragmentation in combination with existing data on minority strategies in Sri Lanka for 1960-2005, I find that ethnopolitical minority organizations are more likely to use violence when fragmentation within the political majority is relatively low and more likely to engage in nonviolence or to participate in electoral politics when majority fragmentation is relatively high. I also determine that minority organizations are more likely to use mixed strategies of electoral politics and violence and violence and nonviolence as majority fragmentation increases. Finally, I find that majorities are more likely to outbid in positions and policies against minorities when minorities use violence than nonviolence. These results demonstrate that the shadow of majority group fragmentation impacts the nonviolent and violent strategies of ethnic minorities, and introduce a new avenue for research on the role of ethnicity in conflict processes.