Coordination in multi-agent systems requires knowledge sharing, including both normative rules and communication. In this paper we suggest that the Organizational Semiotics stance provides a solid conceptual basis for designing agent systems. We show that semantics and pragmatics are not enough for modeling normative aspects in multi-agent systems. A social world level analysis is also required, regarding issues such as commitments, contracts, and social norms. According to Information Field Theory different norm systems define different information fields, thus providing different semantics, pragmatics and social consequences for agents’ behavior. Since agents may belong simultaneously to different information fields conflicts are bound to occur. Agents need to be equipped with an axiological system to resolve their normative conflicts and direct their behavior in order to maximize individual utility, which we argue is dynamically determined within the active information field.