Technical discourse is fraught with rhetorical devices, such as analogies, illustrations, generalisations and contrasts. These devices carry important information which the listener uses to speed up the comprehension process. In this paper we model the meaning of one rhetorical device, namely Contrastive Utterances, in terms of its anticipated effect on the listener's knowledge, and use these predictions in planning computer generated discourse. In addition, we present an algorithm for the selection of illustrative examples. These ideas are to be implemented in a system which generates commentaries in the domain of high-school algebra.
|Number of pages||13|
|Publication status||Published - 1 Jan 1988|