Modelling group constructions for social analysis

Agent-based modelling is becoming widely used for studies in Social Sciences. However, its application faces limitations coming from its bias to software development, which precludes a more active involvement of social researchers. In order to deal with this problem, this work proposes using domain-specific modelling languages based on the socio-psychological Activity Theory. These languages apply agent research to crystallize that theoretical framework in a formal definition suitable for automated processing but close to Social Sciences. The paper focuses on the language for the specification of group constructions such as organizations, norms and shared knowledge. A case study about contradictory decisions in the space shuttle program illustrates the discussion.