Moldt, Danielvon Scheve, ChristianHerczeg, MichaelPrinz, WolfgangOberquelle, Horst2017-11-222017-11-2220023-519-00364-3https://dl.gi.de/handle/20.500.12116/6813Research on emotion has just started to investigate emotions on higher levels of social interaction and aggregation, e.g. organizations or distributed work environments. For a long time the focus has been on the interrelation of cognition and emotion in individuals. But as more and more research is conducted on emotional effects in social interaction, aggregation, and emergence, it becomes obvious that the results are also important for emotional agents (both, natural and artificial) in human-computer interaction. Until now, computer scientific studies – mainly inspired by cognitive science – have designed sophisticated emotional architectures for dyadic interactions. But as emotional agents are increasingly required to engage in social interactions within larger aggregates, either as embodied systems or via multimodal interfaces, the need arises to precisely consider the social-structural peculiarities of emotion. Unfortunately, within the social sciences there is no integrative theory of emotion that interrelates various cognitive and sociological aspects and that computer scientists could use to design improved emotional agents and emotion supporting systems. Therefore, we propose a way to integrate sociological and cognitive theories to analyze emotions on three levels of abstraction: cognitive, interactional, and social structural. We illustrate various reciprocal causes and effects of emotion on the different levels and relate them to urging questions in emotional agents design and human-computer interaction.enText/Conference Paper