Who Feels What and Why? Annotation of a Literature Corpus with Semantic Roles of Emotions

摘要: Most approaches to emotion analysis in fictional texts focus on detecting the expressed text. We argue that this is a simplification which leads an overgeneralized interpretation of results, as it does not take into account who experiences and why. Emotions play crucial role interaction between characters events they are involved in. Until today, no specific corpora capture such were available for literature. aim at filling gap present publicly corpus based Project Gutenberg, REMAN (Relational EMotion ANnotation), manually annotated spans correspond trigger phrases entities/events roles experiencers, targets, causes emotion. provide baseline results automatic prediction these relational structures show lexicons able encompass high variability expressions demonstrate statistical models benefit from joint modeling emotions with its all subtasks. The we enables future research recognition associated entities It supports qualitative literary studies digital humanities. http://www.ims.uni-stuttgart.de/data/reman .