Elements of Argumentation


Elements of Argumentation


By Philippe Besnard and Anthony Hunter.

To be published by MIT Press in early 2008.

Logic-based formalizations of argumentation, that take pros and cons for some conclusion into account, have been extensively studied over a number of years, and some basic principles have now been clearly established. These formalizations assume a set of formulae and then exhaustively lay out arguments and counterarguments. More recently attempts have been to refine these formalizations in order to more closely capture practical argumentation as used in the real world. This has led to techniques for selecting and reforming the more appropriate arguments and counterarguments for use in problem analysis and decision making. These technique identify the better arguments based on (1) taking intrinsic aspects of the arguments into account such as their relative consistency, the exhaustiveness of the consideration of counterarguments, and the relative similarities between arguments; and (2) taking extrinsic factors into account such as impact on the audience, beliefs of the audience, relevance to the audience, etc. The aim of this book is to introduce the background and techniques for formalizing argumentation in artificial intelligence, and in particular to cover the emerging formalizations for practical argumentation.

This book will focus on argumentation by an agent presenting a case for some inference. This involves providing an initiating argument for the inference and then providing undercuts to this argument, and then by recursion, undercuts to undercuts. This approach to argumentation can be described as monological, as opposed to dialectical. In monological argumentation, there is a set of possibly conflicting pieces of information (each piece of information is represented by a formula) that has been collated by some agent, or pooled by a set of agents, and the role of argumentation is to construct a constellation of arguments and counterarguments pertaining to some inference. The presentation of arguments by an agent may then be for the agent to analyze some pieces of information for his/her own purposes (auto-argumentation), or to present to some audience (e.g. politician giving a speech to members of an electorate, a journalist writing an editorial in a newspaper, a clinician explaining a treatment plan to a patient, etc). Dialectical argumentation is conducted between two or more agents and involves more complex issues of multi-agent interaction, dialogue protocols and strategies. This book will not cover dialectical argumentation. However, it is clear that progress in formalizing monological argumentation is necessary for developing better dialectical argumentation systems.

The intended audience for this book are researchers interested in the knowledge representation and reasoning issues surrounding argumentation - either to study existing formalisms or to apply and adapt techniques for real problems. The intended readers are therefore in artificial intelligence and computer science. In addition, it is hoped that the book may be relevant to a wider range of readers in logic, philosophy, linguistics, and cognitive science, and that it could be used as a primary or secondary text for advanced undergraduate and postgraduate courses in logic and argumentation. Readers will be expected to have had some exposure to classical propositional and predicate logic. Though Appendix 1 is a review to remind the reader of the basic concepts and notation, no other prerequisites are assumed.