Why do united rebel fronts emerge in some civil wars, while in other civil wars multiple rebel groups mobilize independently to challenge the state, and often, each other? Scholars have confirmed that, compared with conflicts between a government and a single rebel group, fragmented rebellions last longer, are much more violent, and are more likely to recur after a settlement has been reached. Given the deadly nature of rebel fragmentation, it is critical that we understand it.

Following recent advances in scholarship on violent mobilization processes by, inter alia, Parkinson (2013), Sarbahi (2014), and Cederman, et al. (2013), I have built a diffusion model of rebel fragmentation in which participation in rebellion spreads, completely or incompletely, through networks of civilians and dissidents. Using this theoretical framework I hypothesize that two factors jointly determine whether a rebel movement remains unified or fragments. Rising civilian grievances mobilize heterogeneous dissident networks, tearing rebel movements apart, while cross-cutting broad social networks can bind rebel movements back together by gathering distinct rebel groups under umbrellas. In short, rebel fragmentation reflects the rebellion’s civilian constituency.

My dissertation employs a mixed-methods research design. I first support my theoretical arguments with large-n statistical analysis of rebel fragmentation in 186 internal conflicts from 1946 to 2005. I trace causal mechanisms with an analysis of an original network dataset covering the Frente Sandinista de Liberación Nacional (FSLN) and three splinter groups, gathered through interview and archival research during nine months of fieldwork in Nicaragua. I also illustrate the argument’s application to two cases in Syria, by comparing data gathered from local and international media on early rebel mobilization in the current, radically fragmented Syrian civil war with recently declassified insider accounts of the largely unified 1979 – 1982 Muslim Brotherhood uprising.