Britain's Campaigns in Kenya (1952–1960) and Cyprus (1955–1959)

This chapter tests two of the four Scenarios in the book’s theoretical framework by comparing Britain’s counterinsurgency against the Mau Mau uprising in Kenya (1952–1960) and the EOKA insurgency in Cyprus (1955–1959). The Cyprus Emergency is an instance of Scenario 2 in which the counterinsurgent uses moderate civilian targeting. The prospect theory logic of Scenario 2 suggests that favourable perceptions of the external threat environment and a neutral local alliance strategy would assuage the counterinsurgent’s fear to lose the occupied territory; that, in turn, would make the counterinsurgent more averse to the numerous risks of killing civilians. The Kenya Emergency, instead, is an instance of Scenario 3 in which the counterinsurgent relies on a high level of civilian targeting. The prospect theory logic of Scenario 3 posits that an alliance with would-be dominant groups as a way to avoid a loss of control over the occupied territory would create incentives to target the insurgent popular base with lethal force, but the perceived absence of external threats will make the counterinsurgent averse to the risks of extreme violence. In studying such cases, the chapter firstly introduces the campaigns and identifies common background factors that could not have produced major cross-case differences in the civilian targeting level. Secondly, it shows that in both cases the British authorities perceived no overwhelming external threats to the British rule over Kenya and Cyprus. Thirdly, the chapter proves that the local alliance strategy of Britain supported would-be dominant groups in Kenya but not in Cyprus. Next, it examines the two expeditionary campaigns and demonstrates that evidence confirms the logic and observable implications of my arguments regarding Scenario 2 and Scenario 3. Finally, the chapter shows how alternative explanations may help cast light on violence and restraint in the cases at hand but do not make my theoretical framework redundant.