Why are some ethnopolitical movements divided while others are relatively unified? A growing literature examines the consequences of internal divisions in ethnopolitical movements – and shows that it matters for a range of conflict outcomes – yet the mechanisms causing such divisions remain poorly understood. Our argument emphasizes competitive dynamics between states and self-determination movements and between rival factions within these movements as key determinants of fragmentation. Drawing from literatures on social movements, contentious politics, and civil war, we situate our argument vis-à-vis three alternative and complementary sets of explanations based on theories emphasizing transnational dimensions, political institutions, and structural factors within ethnopolitical groups. Using an original dataset, we test hypotheses explaining movement fragmentation over time and use a case study of Punjab in India to identify specific causal mechanisms and missing variables. Our findings show some support for three of these theories, suggesting that ethnopolitical movements divide as a result of complex and interactive processes. But our findings also underscore that central to explaining fragmentation dynamics are factors capturing competitive dynamics, including repression, accommodation of movement demands, the turn to violence, and the dynamic and changing nature of ethnopolitical demands.