Since the end of the Second World War, civil war has been the most frequent and deadly form of armed conflict. This has resulted in an abundance of explanations of the occurrence of civil war. Nevertheless, a complete explanation of any phenomenon must be able to explain the absence as well as the occurrence of said phenomenon. The time has come to turn to the non-events, the dogs that do not bark; the countries where we should expect to see civil war, but where no civil war occurred. One such country is Tanzania. Within Tanzania’s postindependence history, the semi-autonomous region of Zanzibar, and the 1990-2010 period stand out as particularly conflict-prone. Therefore, the research question is: Why was there no civil war in Zanzibar in the period 1990 – 2010? To answer this question, I combine various theories of the causes of civil war into a comprehensive and systematic theoretical framework, which incorporates both structural and process-based theories of civil war onset. The resulting theoretical framework holds that governments and their domestic opposition produce their own civil wars, but not at their own accord; they produce them under certain structural conditions, given and transmitted from the past. This theoretical framework is developed in the theory chapter and operationalized in the methods chapter. I apply this theoretical framework to Zanzibar in the 1990-2010 period, and argue that Zanzibar avoided civil war in this period for three reasons. First, in years preceding elections, escalation to civil war was prevented because of a belief held by the opposition that the next election would lead to substantial political change. Second, regardless of the level of objective opportunity, lack of perceived opportunity has prevented the opposition from initiating a civil war. And third, the leadership of the opposition has acted in a risk-averse manner at key moments. In total, the empirical case study explains the puzzle of Zanzibar’s relative peacefulness and illustrates the advantages of theory synthesis and methodological pluralism in the analysis of civil war risk in specific cases.