Recent civil wars have witnessed a myriad of risks for civilian populations, including attacks against civilians and civilian objects. In spite of the significant risk and long-term human security impacts of attacks on important civilian resources, attacks against civilian objects remain comparatively understudied. This thesis explores the relationship between different modalities of violence against civilians in civil war. It focuses on two categories of violence – 1) violence that attacks civilian bodies, including both lethal attacks and non-lethal attacks, referred to as primary modalities of violence against civilians, and 2) violence that attacks civilian objects, subsequently denying civilians access to important resources and possibly leading to physical harm, referred to as secondary modalities of violence against civilians. To compare these categories of violence, the author developed a framework of violence against civilians and subsequently developed a conflict event dataset on violence against civilian objects through a novel, semi-automated approach using robotic process automation for the identification, collection, and transformation of event data not currently available through existing datasets. Using the South Sudanese civil war as a case study, this thesis identifies important divergences in patterns of modalities of violence against civilians and their relationship to armed actor competition and ethnic identity, highlighting the significance of examining harm beyond violence directed at civilian bodies.