arrow
Return

Audiences in argumentation frameworks

delete2007-01-01
delete62
PRE
AI
T
Trevor Bench‐Capon
S
Sylvie Doutre
P
Paul E. Dunne *
DOI:10.1016/j.artint.2006.10.013delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
Although reasoning about what is the case has been the historic focus of logic, reasoning about what should be done is an equally important capacity for an intelligent agent. Reasoning about what to do in a given situation-termed practical reasoning in the philosophical literature-has important differences from reasoning about what is the case. The acceptability of an argument for an action turns not only on what is true in the situation, but also on the values and aspirations of the agent to whom the argument is directed. There are three distinctive features of practical reasoning: first, that practical reasoning is situated in a context, directed towards a particular agent at a particular time; second, that since agents differ in their aspirations there is no right answer for all agents, and rational disagreement is always possible; third, that since no agent can specify the relative priority of its aspirations outside of a particular context, such prioritisation must be a product of practical reasoning and cannot be used as an input to it. In this paper we present a framework for practical reasoning which accommodates these three distinctive features. We use the notion of argumentation frameworks to capture the first feature. An extended form of argumentation framework in which values and aspirations can be represented is used to allow divergent opinions for different audiences, and complexity results relating to the extended framework are presented. We address the third feature using a formal description of a dialogue from which preferences over values emerge. Soundness and completeness results for these dialogues are given. (c) 2006 Elsevier B.V. All rights reserved.
Keywords:
argumentation frameworks
practical reasoning
dialogue

Journal

Artificial Intelligence Review cover
Artificial Intelligence Review
IF:
13.9
Papers:
6.1K
Citations:
1.9W

Organization

No organization information available