During the 1990s, around 3,400 Kurdish villages have been evacuated in Eastern Turkey, resulting in the displacement of 370,000 individuals according to official figures. This paper examines why some villages survived the civil war intact while others were forcibly evacuated by Turkish soldiers or attacked by PKK guerillas. My tentative hypothesis, which relies on anecdotal evidence, is that villages which were able to juggle multiple allegiances--to the state and to the PKK--were able to survive the war intact, while those that unambiguously signaled an exclusive allegiance were either evacuated or suffered high casualties. I hypothesize, moreover, that the ability of villages to "pass" as pro-government when the gendarmerie visited their village, and to camouflage as pro-PKK when the guerillas visited, rested internally on their clan networks and externally on the ability of the villagers to establish ties of trust with outsiders, mainly teachers, judges, and prosecutors, who were sent to their village from the outside, and who were under pressure to inform the government on the allegiances of "the village." The paper seeks to contribute to several broader questions in the study of ethnic conflict: How do social organizations on the ground, such as clan networks, transform the ideological meanings and material incentives associated with the "high" politics of ethnic conflictt How do multiple loyalties--to family, nation, and state--affect survival strategies during civil conflictg And what sorts of order emerge in war zones where states have a tenuous hold? The paper will be based on interviews and data drawn from surveys of the internally displaced population.