Knowledge, common knowledge and related puzzles (Extended Summary)

  • Authors:

  • Daniel Lehmann

  • Affiliations:

  • -

  • Venue:
  • PODC '84 Proceedings of the third annual ACM symposium on Principles of distributed computing
  • Year:
  • 1984

Quantified Score

Hi-index 0.00

Visualization

Abstract

Many distributed systems, as well as many real life situations, are best described as involving changes in the partial knowledge that components may have about the real state of the whole system. Examples include synchronization and cooperation protocols, cryptographic systems, games, economics and intelligent programs. In such situations the notion of common knowledge has been recognized as of fundamental importance by Lewis [Le] and Aumann [A]. An event is common knowledge if everybody knows it, everybody knows that everybody knows it, and so on. A method for the formal description of such systems and the rigorous proof of certain of their properties is presented. Its limitations are analyzed. As examples, a well-known puzzle and a logical paradox are treated. A propositional language in which one may describe knowledge, common knowledge and their changes with time is defined. In particular one may describe the knowledge that agents may have of the present state of the world, future states of the world and the knowledge that others may or may not have about the present and future states of the world. The language is interpreted in models a la Kripke, where knowledge is interpreted by a binary relation. An axiomatization is given and shown sound and complete with respect to the models. A doubly-exponential deterministic time decision procedure is described.