Towards higher impact argumentation

  • Authors:
  • Anthony Hunter

  • Affiliations:
  • Dept of Computer Science, University College London, London, UK

  • Venue:
  • AAAI'04 Proceedings of the 19th national conference on Artifical intelligence
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

There are a number of frameworks for modelling argumentation in logic. They incorporate a formal representation of individual arguments and techniques for comparing conflicting arguments, An example is the framework by Besnard and Hunter that is based on classical logic and in which an argument (obtained from a knowledgebase) is a pair where the first item is a minimal consistent set of formulae that proves the second item (which is a formula). In the framework, the only counter-arguments (defeaters) that need to be taken into account are canonical arguments (a form of minimal undercut). Argumem trees then provide a way of exhaustively collating arguments and counter-arguments. A problem with this set up is that some argument trees may be "too big" to have sufficient impact. In this paper, we address the need to increase the impact of argumentation by using pruned argument trees. We formalize this in terms of how arguments resonate with the intended audience of the arguments. For example, if a politician Wants to make a case for raising taxes, the arguments used would depend on what is important to the audience: Arguments based on increased taxes are needed to pay for improved healthcare would resonate better with an audience of pensioners, whereas arguments based on increased taxes are needtd to pay for improved transport infrastructure would resonate better with an audience of business executives. By analysing the resonance of arguments, we can prune argument trees to raise their impact.