Rejoinder to comments on "reasoning with belief functions: An analysis of compatibility"

An earlier position paper has examined the applicability of belief-functions methodology in three reasoning tasks: (1) representation of incomplete knowledge, (2) belief-updating, and (3) evidence pooling. My conclusions were that the use of belief functions encounters basic difficulties along all three tasks, and that extensive experimental and theoretical studies should be undertaken before belief functions could be applied safely. This article responds to the discussion, in this issue, o f my conclusions and the degree to which they affect the applicability o f belief functions in automated reasoning tasks.