In Afghanistan, Libya, Liberia and beyond, armed rebellions have begun when armies fell apart. When does this occur? This paper conducts a large-N analysis of these army-splinter rebellions, distinct from both non-military rebellions from below and from coups, using new data. It finds that they follow a logic of state breakdown focusing on regime characteristics (personalist regimes and the loss of superpower support at the end of the Cold War) rather than drivers of mass mobilization from below. In contrast, these regime-level factors matter much less for the non-military rebellions from below that dominate theorizing about civil war origins. This paper also shows that one option for military rebels lies in not attempting a coup but instead heading straight into a rebellion. This paper thus distinguishes highly different paths to armed conflict, validates the state breakdown approach to why armies fall apart, and extends the well-known tradeoff between coups and civil wars.

在阿富汗、利比亚、利比里亚和其他地区，当军队分崩离析时，武装叛乱已经开始。这发生在什么时候？本文使用新数据对这些军队分裂的叛乱进行了大 N 分析，这些叛乱不同于来自地下的非军事叛乱和政变。它发现他们遵循国家崩溃的逻辑，关注政权特征（个人主义政权和冷战结束时失去超级大国的支持），而不是来自下层的大规模动员的驱动力。相比之下，这些政权层面的因素对于主导关于内战起源理论的非军事叛乱而言，影响要小得多。这篇论文还表明，军事叛乱分子的一个选择是不尝试政变，而是直接发动叛乱。因此，本文区分了武装冲突的截然不同的路径，