Mutual knowledge, background knowledge and shared beliefs: Their roles in establishing common ground

This interdisciplinary paper seeks to elucidate the roles played by mutual knowledge, background knowledge and shared beliefs in establishing common ground. We begin our discussion by clarifying the terminological and conceptual confusion associated with the various notions of mutual knowledge, background knowledge and shared belief las used in the philosophy, cognitive psychology and discourse analysis literature) in relation to the philosophical puzzle of the Mutual Knowledge Paradox. After clarifying our use of terminology, we turn our attention to the Mutual Knowledge Paradox which addresses the problem of how an infinite regression of knowledge required for an instance of S-meaning can be achieved in finite processing time in everyday conversation. Two broad approaches to resolving the paradox are examined: (I) those accounts which assume an infinite regression, with a mechanism for inferring mutual knowledge; and (2) those accounts which do not assume an infinite regression, with a truncation heuristic for inferring shared beliefs. We argue that the current impasse in resolving the paradox may be better tackled by a careful analysis of authentic data than by means of introspective theorising based on invented data. The analysis of authentic spoken data demonstrates that a psychologically viable account of how common ground comes to be established between two persons lies in the notion of shared belief, rather than mutual knowledge. We end by proposing a system of classifying various 'categories' of common ground which takes into consideration the intricacies of the distinction we make between shared belief, background knowledge and mutual knowledge above. (C) 2001 Elsevier Science B.V. All rights reserved.