A TAXONOMY OF STRATEGIES FOR MULTIMODAL PERSUASIVE MESSAGE GENERATION

Abstract

Future intelligent systems will have contextual goals to pursue. As opposed to more traditional scenarios of human computer interaction, intelligent persuasive systems may also aim to induce the user or, in general, the audience, to perform some actions in the real world. Some scenarios of application are dynamic advertisement, preventive medicine, social action, and edutainment. As a step in this direction, a prototype called Promoter was developed for the production of persuasive messages. In modeling persuasion, the cognitive state of the participants (beliefs, desires, and intentions) is taken into account, as well as their social relations, their emotions, and the context of interaction. In this article, a taxonomy of persuasive strategies and the meta-reasoning model that works on this taxonomy is described. The taxonomy is built by taking into consideration studies coming both from social psychology and philosophy, and from the area of natural argumentation. The taxonomy is not domain specific and it helps to bridge persuasion strategies and rhetorical relations, a fundamental element in text planning. The use of this taxonomy also permits reasoning on the basis of emotion expression in accordance with persuasion strategies for multimodal message generation.