CENTER AND PERIPHERIES OF STATE POWER

The previous chapter suggested that Serbian and Israeli methods of violence tended toward police-style efforts where state control was highest, and toward more destructive tactics where the state's grip was weak. This observation prompts elaboration of a general hypothesis of state repression: the more tightly outsider populations are controlled by contemporary, semi-democratic states, the more likely they are to experience police-style repression. The less firmly these regions are controlled by nationalist states, conversely, the more likely they are to experience destructive violence and even ethnic cleansing. Zones of intense state power prevent nationalism from developing to its most virulent proportions, but at the margins of state authority, extremism flourishes. State violence is organized very differently in the core as opposed to the periphery of power.

The rest of this book elaborates and defends this argument for the Serbian and Israeli cases. As such, it represents an effort at theory building, not theory testing, and I make no claim for its unproblematic application elsewhere. Aided by comparisons, theory-building exercises identify important variables, concepts, and arguments that can later be extended to or tested on other cases, a theme I briefly explore

---

in the concluding chapter. In its present form, my explanation provides a reasonable interpretation of empirical variation across and within the Serbian and Israeli cases. By studying failed attempts by some paramilitaries to import specific methods of violence from one institutional setting to another, moreover, I dramatize the importance of context. When nationalist militias try but fail to use Bosnia-style methods in Serbia proper, their aborted trajectory underlines the power of institutional settings. With some modification, this same approach might help explain patterns of state violence in other high-capacity and partially democratic states such as Turkey, apartheid-era South Africa, and India.^[9]