Quantificational Dynamics via Dialogue
Jonathan Ginzburg · 1995
Much recent work in formal semantics has focussed on developing a dynamic approach to meaning. Stated in its most general form this means that the meaning of a linguistic form is explicated in terms of the effect its use has on existing commonly shared Attention has been focussed on two types of changing resources. First, in order to explicate assertion and presupposition, Stalnaker 1978 and Lewis 1979 urged construing context as a resource that represents the commonly accepted information at any given point in conversation. Second, in order to capture the anaphoric possibilities available at any stage of discourse, Kamp 1981, Heim 1982, Barwise 1987, Groenendijk and Stokhof 1991, proposed keeping track of an additional resource, namely a set of salient (assignments to) variables. The concern of these approaches has been monologue or text. Indeed, these as most formal semantic models, have hitherto abstracted away from issues pertaining to communication, such as the fact that in conversation one participant’s utterances are not automatically and identically comprehended by the other participants. The consequence of this has been that many actually occurring dialogue contributions cannot be analysed, in particular those whose primary function is to indicate comprehension or the need for clarification. In this paper I start out with a framework designed to explicate certain basic features of dialogue interaction and show how subsentential dynamics such as anaphora and ellipsis emerges. More specifically: Ginzburg 1994b,c shows how relativising to individual participants and adding extra structure to the common ground conception of Stalnaker 1978 enables one to formalize the processes of utterance grounding and clarification (e.g. Clark and Schaefer 1989), as well as to offer notions of querying and assertion that capture the interactive nature of these acts in dialogue. On this view, the structure of the common ground, or rather each participant’s version thereof, emerges not solely from facts tha~ accumulate as a conversation evolves, but also from the most recent action performed and from questions that enter into discussion. While facts are, to a first approximation at least, permanent components once introduced, questions remain under discussion only so long as they are unresolved, as long as further information about them can be provided, or indeed until another, typically more specific, question has been introduced. What I will demonstrate in the current paper is that this same notion of context can be exploited to provide an account of sub-sentential dynamics. Consider (la-c):