Speech reports and evidence

Abstract

In our communication we often make reference to what other people have said. We do so for a variety of different reasons and we draw upon a variety of different constructions. Following earlier joint work together with Emar Maier, I propose a classification of speech reports along two dimensions: at-issueness and eventivity, resulting in four classes of speech reports. I explore the notion of reportative evidentiality against the background of this landscape. I propose that a speech report is used in an evidential way if the actual speaker seems to intend to convey that the fact that the reported speaker uttered this sentence increases the probability of the reported content being true. I furthermore argue that there is a natural link between the class of non-eventive, not-at-issue speech reports and evidentiality. Non-eventive, not-at-issue speech reports are naturally used to provide evidence because (i) the reported content is grammatically marked as the main point, which will be interpreted as a (weak or hedged) proposal to update the common ground, and (ii) other uses of speech reports (e.g. vivid characterizitions of dialogues between characters in a story) are not available due to their non-eventivity. Interestingly, if we switch to this pragmasemantic perspective, some expressions that are traditionally called ‘evidential’ loose their evidential status.