The Consensual Dynamics of Debates with Core Updating

Abstract

In the debates we have simulated so far, all sentences were on a par with each other. Proponents didn t consider some of the debates sentences as more and others as less important. If they could, for example, reestablish coherency by changing exactly one truth-value assignment, they were indifferent as to which belief they give up. But this, it seems, doesn t hold in real controversies, where proponents frequently possess some convictions which they are very reluctant to give up as well as other beliefs they are much more willing to modify. In this chapter, we are going to include the proponentsvarying loyalty to different beliefs in our simulations. In order to do so, we assume that there is a subset of the sentence pool which contains the debate score sentences. The proponentspartial positions on these core sentences make up the heart of their belief system, which they are particularly unwilling to modify. In our simulations, this translates into a more sophisticatedupdate mechanism, as explained below. Studying the simulation results, our primary concern is the evolution of the proponentscore positions. As these cores represent partial positions defined on a subset of the sentence pool, we will be able to examine how the robustness of positions, that is, thedegree of justification, influences the debate dynamics.