>>56
>conglomerability について
>P75-202 に記載があります

conglomerabilityについては、検索ヒットした下記なども、どうぞ
正直、conglomerabilityは難しすぎ。日本ではあまり議論されていない様子。conglomerabilityの訳語もないみたい（＾＾

（参考）
http://www.sipta.org/isipta13/proceedings/papers/s029.pdf
8th International Symposium on Imprecise Probability: Theories and Applications, Compiegne, France, 2013
Two theories of conditional probability and non-conglomerability
Teddy SeidenfeldMark J. SchervishJoseph B. KadaneCarnegie Mellon University

Abstract
Conglomerability of conditional probabilities issuggested by some (e.g., Walley, 1991) as necessary forrational degrees of belief. Here we give sufficientconditions for non-conglomerability of conditionalprobabilities in the de Finetti/Dubins sense. Thesesufficient conditions cover familiar cases where P(?) is acontinuous, countably additive probability. In thisregard, we contrast the de Finetti/Dubins sense ofconditional probability with the more familiar account ofregular conditional distributions, in the fashion ofKolmogorov.

1 Introduction
Consider a finitely, but not necessarily countablyadditive probability P(?) defined on a sigma-field of setsB, each set a subset of the sure-event Ω. In other terms,<Ω, B, P> is a (finitely additive) measure space.We begin by reviewing the theory of conditionalprobability that we associate with de Finetti (1974) andDubins (1975).

This account of conditional probability is not the usualtheory from contemporary Mathematical Probability,which we associate with Kolmogorov (1956).
That theory, instead, defines conditional probability throughregular conditional distributions, as follows.