Defining common ground

Stalnaker (Context, Oxford University Press, Oxford, 2014) defends two ideas about common ground. The first is that the common ground of a conversation is definable in terms of an iterated propositional attitude of acceptance, so that p is common ground iff p is commonly accepted. The second is the idea that the “default setting" of conversational acceptance is belief, so that as a default, what is accepted in conversation coincides with what is (commonly) believed. In this paper, I argue that we should favor a pair of contrasting theses instead. First, I argue that we should identify the common ground with what is common knowledge about what is accepted, so that p is common ground iff it is common knowledge that p is accepted. Thus the attitude that is iterated in the definition of common ground is not acceptance but knowledge. Second, I argue that the “default setting" for conversational acceptance is not belief, but knowledge.