Information, Meaning and Eigenforms: In the Light of Sociology, Agent-Based Modeling and AI

Abstract: The paper considers the relation of Shannon-type information to those semantic and hermeneutic aspects of communication, which are often referred to as meaning. It builds on considerations of Talcott Parsons, Niklas Luhmann and Robert K. Logan and relates them to an agent-based model that reproduces key aspects of the Talking Head experiment by Luc Steels. The resulting insights seem to give reason to regard information and meaning not as qualitatively different entities, but as interrelated forms of order that emerge in the interaction of autonomous (self-referentially closed) agents. Although on first sight, this way of putting information and meaning into a constructivist framework seems to open possibilities to conceive meaning in terms of Shannon-information, it also suggests a re-conceptualization of information in terms of what cybernetics calls Eigenform in order to do justice to its dynamic interrelation with meaning.