Persuasive Dialogue: Changing User Behaviour with a Computer

Dialogue is not new to computational linguistics but, up to now, it has been applied only to very restricted domains. The computer is often either too task oriented and logic or quite mindless: no solution has been found to this problem yet. By applying human computer dialogue to persuasive dialogue, this thesis hopes to take the research in domains that have barely been studied in conjunction with human computer dialogue. First of all the research will examine if, by using the long studied theories of rhetoric in philosophy and linguistics, it will be possible to render automated dialogue systems more natural. Actually, the aspect of argumentation in dialogue has been taken into account only in specialised areas such as law where logic is mostly used. One of the interests of this project is therefore to study how rhetoric could be applied in some less restricted type of dialogue, where other aspects of rhetoric, like emotions will be useful. Another goal of this project is to study the theories of persuasive communication. These theories and the models they propose could be used in conjunction with the theories of argumentation to control the persuasiveness of the dialogue.