Information flow in logic for distributed systems: Extending graded consequence

Abstract In this paper we have proposed a variation of Barwise and Seligman’s proposal for information flow among a distributed network of agents by bringing in a notion of belief structure and a notion of graded inference. In their proposal, belief profile of an agent has played an important role as it is explicitly mentioned in the principles of information flow. In contrary, while developing the formal counterpart of information flow they did not address any connection with the belief profile of an agent. Besides, they have accommodated a non-deterministic notion of derivation to design the local logic of an agent keeping a room open for unsound inferences. Our proposal here is to bring in the notion of belief structure of an agent in the development of the local logic of an agent, and convert the non-deterministic nature of consequence to a deterministic graded (four-valued) notion of consequence. For the time being, we have focused on a specific context of decision making by aggregating opinions of agents based on an approach, known as the theory of graded consequence.