Prevarication in dispute protocols

Abstract

Models of persuasion, argument, and reasoning motivated by analogies from Law and legal process are now accepted formalisms supporting multi-agent discourse in applications such as contract negotiation and resolving disputed claims. A number of non-classical Logics and proof theories within these have been proposed specifically to deal with the special circumstances wherein propositional theories are not best suited to address modeling issues arising in legal contexts: e.g. exceptions and defaults are treated in a variety of so-called non-monotonic logics; similarly concepts of credulous, cautious and sceptical belief have been developed, partly to reflect differing forms of 'burden of proof' that may apply in various judicial contexts. Our concern in this paper is to consider one aspect of legal argument that appears to have been largely neglected in existing work concerning agent discourse protocols - particularly so in the arenas of persuasion and dispute resolution - the use of legitimate procedural devices to defer 'undesirable' conclusions being finalised and the deployment of such techniques in seeking to have a decision over-ruled. Motivating our study is the contention that individual agents within an 'agent society' could (be programmed to) act in a 'non-cooperative' manner: thus, contesting policies/decisions accepted by other agents in the 'society' in order to improve some national 'individual' utility. Using Dung's argumentation framework, we present various settings in which the use of 'legitimate delay' can be rigorously modeled, formulate some natural decision questions respecting the existence and utility of 'prevaricatory tactics', and, finally, illustrate within a greatly simplified schema, how carefully-chosen devices may greatly increase the length of an apparently 'straightforward' dispute.