Abstract:

The classic model of conversation based on the Common Ground (CG), introduced by Karttunen (1974), Lewis (1979) and Stalnaker (1978), was shown to be insufficient for accounting for various conversational phenomena (inter alia Portner 2004, 2007, Farkas & Bruce 2009, Murray 2014). This paper further strengthens this line by analyzing a type of non-truth conditional non-core dative termed the Discursive Dative (DD) as a discourse management device (Krifka 2008, Repp 2013). The DD signals that the asserted proposition p constitutes an exception to a normative generalization believed by the speaker to be shared by the speech event participants. In order to capture the notion of exception we propose to divide the CG into two sets of worlds, those consistent with previous assertions and their presuppositions (CGA) and those consistent with generalizations (CGG). The DD signals a non-inclusion relation between the asserted proposition and the CGG. This enables us to distinguish between different types of mirativity effects, by drawing a distinction between adding a proposition p that was not previously in the speaker’s expectation-set (inter alia DeLancey 1997, 2001, Rett 2009, Peterson 2013, Rett & Murray 2013) and the present case of the DD, where p can very well be in the speaker’s expectation-set, but objectively expected that ~p.