A computational model of argumentation in agreement negotiation processes

Abstract

The paper describes a computational model that we have implemented in an experimental dialogue system (DS). Communication in a natural language between two participants A and B is considered, where A has a communicative goal that his/her partner B will make a decision to perform an action D. A argues the usefulness, pleasantness, etc. of D (including its consequences), in order to guide B's reasoning in a desirable direction. A computational model of argumentation is developed, which includes reasoning. Our model is based on the studies in the common-sense conception of how the human mind works in such situations. Theoretical considerations are followed by an analysis of Estonian spoken human–human dialogues. First, calls of clients to travel agencies are studied where a travel agent could use various arguments in order to persuade a client to book a trip. The analysis demonstrates that clients are primarily looking for information; argumentation occurs in a small number of dialogues. Secondly, calls of sales clerks of an educational company to different organisations are analysed where training courses are offered. The sales clerks, unlike the travel agents, try to persuade clients, stressing the usefulness of a course. Finally, face-to-face conversations are studied where one participant is arguing for an action by the partner. The corpus analysis shows that our model covers simple situations occurred in actual dialogues. Therefore, we have chosen a limited application of the model in a DS – communication trainer. The computer can optionally perform A's or B's role and the interaction with the user in Estonian follows norms and rules of human–human communication.

Keywords:

• formal models of argumentation
• argumentation in human–computer interaction
• argumentation support
• conversational agent
• reasoning model
• communicative strategy
• communicative tactics

Acknowledgement

We thank our students for implementing and testing the model.

Funding

This work is supported by the European Regional Development Fund through the Estonian Centre of Excellence in Computer Science (EXCS) and the Estonian Research Council (projects 8558 and 9124).

Notes

1. In the examples, transcription of conversation analysis is used (Hutchby, Wooffitt 1998). Utterances are numerated.