Introduction

The semantics of dialogue is a fundamental topic for a number of reasons. First, dialogue is the primary medium of language use, phylogenetically and ontogenetically. Second, studying dialogue forces one to a particularly careful study of the nature of context. The context has a role to play in determining what one can or should say at a given point and also how to say it. Conversely, it affords the interlocutors a very impressive economy of expression – there is much subtlety that can be achieved with relatively little effort drawing simply on material that is in the context.

Consequently, two themes will drive this article, relating to two fundamental problems a semantic analysis in dialogue has to tackle:

Conversational relevance is without doubt a very fundamental and difficult problem. It is closely connected to the Turing test and solving this problem in full generality is, as Turing pointed out, a possible basis for understanding the nature of intelligence (Turing, 1950; see also Łupkowski and Wiśniewski, 2011). However, is it a semantic, as opposed to a pragmatic problem or one connected to generalized notions of cognition? We will not deal with this issue of territorial demarcation in anything but passing (see also Schlenker, Chapter 22). Nonetheless, we will offer a detailed empirical and theoretical analysis of the components of conversational relevance.

Conversational meaning is semantic enough, but the obvious question one might ask is – why conversational? Why should we consider meaning in the context of conversations? One might turn the question on its head and ask, noting, as I already have above, that conversation is the primary linguistic medium in which language evolved and existed for millennia and that it is the setting in which language is acquired: how could we not take conversation as the basic setting for semantic theory?