Computing and Informatics

This contribution argues for the proposition that formal models based on the theory of formal grammars and languages are adequate for the study of some computationally relevant properties of agents and multi-agent systems. Some questions are formulated concerning the possibilities to enlarge the universality and realism of such models by considering the possibilities to go with their computing abilities beyond the traditional Turing-computability, and by considering very natural properties of any real (multi-)agent system such as the partially predictable functioning (behavior) of agents, their unreliability, dysfunctions, etc.