Search

This dissertation demonstrates the usefulness of digital open source information (DOSI) for academic research on the causes of war through an in-depth case study of the conflict in eastern Ukraine's Donets Basin (Donbas). It argues that the current social science literature is in need of theoretical and methodological innovation to operate in the abundant but murky information environment that surrounds the Donbas war and other conflicts of the social media age. The result is a deep divide in the academic literature between scholars who emphasize domestic causes of the Donbas war and those who emphasize Russian involvement. To address this shortcoming, my dissertation develops new theoretical and methodological frameworks. My theoretical framework combines conflict escalation theory with the historical institutionalist concept of critical junctures. Based on this framework, I develop an escalation sequence model of the Donbas conflict which divides the formative phase of the war into six critical junctures. Moreover, my theoretical framework draws attention to intervention and delegation as two distinct modes of external actor involvement in these critical junctures. My methodological framework combines process tracing with the journalistic practice of DOSI analysis to shift the methodological focus towards source criticism and probabilistic reasoning. I argue that this shift towards digital forensic process tracing is essential to make social science methodology fit for the social media age. The six empirical chapters of my dissertation apply digital forensic process tracing to the six critical junctures of the Donbas war's escalation sequence. For each critical juncture, they assess the available open source evidence of domestic causes and Russian interference. I argue that there is convincing evidence that Russian involvement was the primary cause of four of the six critical junctures. For this reason, my dissertation concludes that the Donbas war is primarily an interstate war between Russia and Ukraine.