The Extended Model

Abstract

In this chapter we present a more sophisticated version of the model elaborated in the last two chapters. We call it the extended model, and it is an improvement over the elementary model in two important respects. First of all, the constancy condition is dropped, that is, we allow that members of the group might assign different weights at higher levels of evaluation. Secondly, we propose a method for regimenting discussion at different levels of evaluation to insure proper separation of issues. We then discuss the conditions under which the process of aggregation converges toward a limit, thus yielding a set of consensual weights. We then note that such a set of weights can be computed and justified without the assumption of infinitely iterated aggregation and, therefore, that it is unnecessary to assume an infinite hierarchy of weights. Finally, we turn to the question of whether an individual is rationally committed to the consensus our methods yield. We argue that, given that he assigns weights so consensus is reached, he is committed to the outcome. We note, however, that an individual may escape such a commitment by assigning others a weight of zero at some level, and thus refuse to generate consensus. We conclude with a discussion of the merits of iconoclasm.