CIVIL WAR NON-ONSETS THE CASE OF JAPAN

Abstract

In earlier work, we argued that the pattern of cross-national correlates of civil war onset can be best explained by interpreting several influential variables as indicators of state weakness. This argument was speculative, however, in that partial correlations from a statistical model using country-year data cannot rule out multiple possible mechanisms linking the explanatory variables to civil war onset. To explore mechanisms, we turned to narrative, and justified a procedure for case selection called “random narratives” as having advantages compared to the reliance on convenience samples. This paper, juxtaposing statistical expectations with historical narrative, illustrates what we can learn more generally about the causes of civil war even from Japan, a country that has not experienced a civil war in our period of study, and one of our randomly selected country cases. The narrative reveals that there have been deep grievances in the post WWII era resulting in several uprisings. These were efficiently cauterized and prevented from developing into insurgencies by a security state that had high information and great discipline in its non-use of violence. The narrative evidence is consistent with our interpretation of state weakness as an important variable for explaining cross-sectional and over-time variation in civil war onset. From this case study we cannot rule out that societal grievances are lower on average in Japan than in civil war-afflicted states, but we can rule out that there were no intensely held grievances capable of motivating violent rebellion in post-war Japan, particularly in the late 1940s and 1950s.