In Greek and Latin historical narratives, Roman armies are repeatedly said to destroy enemy communities, both their physical urban spaces and inhabitant populations. Some ancient authors claim that this conduct was characteristic of the Roman way of war, particularly during the period of the Middle Republic. However, this seemingly prevalent feature of Roman warfare remains poorly understood. Ancient descriptions of urban destruction and mass killing are often vague or formulaic, and rarely indicate how or why this violence took place. Although a few modern studies have examined mass violence in antiquity, the destruction of communities is seldom treated as a distinct category of Roman military action, with its own methods and motives. Furthermore, there has been little effort to explore how ancient armies actually destroyed cities or peoples using pre-modern tools. To redress these gaps in the scholarship, this dissertation aims to demonstrate how and why Roman armies destroyed urban spaces and populations. The project first examines descriptions of urban destruction and mass killing in ancient texts, archaeological and art historical evidence of mass violence, and comparative evidence from other historical periods. The second half of the project investigates individual cases in which Roman commanders attacked and destroyed enemy cities or populations. Case studies allow in-depth examinations of individual events, making it possible to situate episodes of mass violence within a larger set of historical circumstances; this approach highlights the specific causal factors that encouraged Roman military leaders to target enemy communities. Using these methods, this dissertation argues that ancient armies employed demolition and mass arson to destroy urban spaces, and killed populations using cold-blooded mass executions or hot-blooded indiscriminate massacres. Although ancient military forces rarely, if ever, razed entire towns or exterminated whole peoples, even partial destruction required an expenditure of time, labor, and resources. Thus the destruction of communities was not the result of haphazard outburst or violent frenzy, but stemmed from the calculated decisions of military and political leaders. This study further argues that Roman commanders destroyed enemy communities instrumentally, to accomplish a range of goals and objectives. While many Roman commanders employed mass violence strategically, as a response to specific military problems, their political, economic, and personal goals could also motivate destruction and mass killing in war.

This thesis seeks to provide an account of the role played by ideologies in acts of mass violence against civilians, such as genocides, murderous state repression, war crimes, and other ‘atrocities’. Mass violence of this kind has already received extensive study, with scholars frequently emphasising their belief that ideology is important. Until now, however, discussions of ideology have been held back by a lack of conceptual and theoretical development, leading to narrow portrayals of ideology’s role, vagueness over its relevance, and dubious assumptions about its theoretical implications. This thesis addresses these problems by building a more focused and integrative theoretical framework for analysing the ideological dynamics of atrocities. I engage in an extensive conceptual and methodological discussion, to establish the best way of defining and utilising the concept of ideology. In doing so, I emphasise how ideology can be important even for that majority of atrocity perpetrators who do not meet classic but misleading stereotypes of fanatical killers driven by burning hatred. I then detail my actual account of the ideological dynamics of deadly atrocities, which centres around the identification of six ‘justificatory mechanisms’: dehumanisation, guilt-attribution, threat-construction, deagentification, virtuetalk, and future-bias. These justificatory mechanisms describe sets of ideological processes that recur across different cases of violence against civilians, and which make that violence look permissible or even desirable to those who, in a variety of roles, carry it out. I then substantiate this account through three case studies: of Nazi atrocities, Stalinist oppression, and Allied area bombing in World War II. These cases demonstrate the cross-case applicability of the six justificatory mechanisms, and illustrate how the framework I offer allows us to construct more causally explicit, psychologically plausible, and comprehensive pictures of the way key ideologies feed in to the most destructive campaigns of violence against civilians.