A Crowd-Sourced Collection of Narratives for Studying Conflict

In this paper we have described a corpus that provides many of the details required to understand the context and dynamics that factor into a person’s emotional state, the perception of consequences and their likely selection of a conflict resolution strategy. We believe this information is useful for meta-reasoning about narrative through a deeper knowledge about people’s thought process. It also provides enough detail for attempting a data driven approach for modeling the mental process of conflict resolution in computational agents that respond in a way that people find believable. Finally, we have provided a methodology to extend the corpus, so that over time we may cover a broader spectrum of conflicts and target specific domains when they are needed for applications.