Social network semantics for agent communication

In this paper we introduce a semantics for agent communication languages based on social networks, providing us with a principled way to define and reason about their dynamics. As an instance we consider dependence networks, where the social relations represent that an agent depends on another agent to achieve its intentions. We suggest how FIPA semantics can be reconstructed in this social semantics. Our approach reveals that we need special semantics for relations like ownership, authority or fear: all kinds of interesting social relations, not previously studied by multiagent systems.