Clarification of some mathematical misunderstandings about Savage's foundations of statistics, 1954 (Q2366201)

This paper discusses some mathematical misunderstandings about \textit{L. J. Savage's} set-up [in his well-known book ``The Foundations of Statistics''. Wiley (1954; Zbl 0055.12604)], based on finitely additive probability measures which are not necessarily countably additive. The adoption of a finitely additive framework does not prevent from considering, in particular, probabilities which are countably additive, when this turns out to be suitable. On the other hand, finite additivity allows some additional freedom in the probabilistic expression of our knowledge, but this often gives rise to several misunderstandings [see the reviewer, Boll. Unione Mat. Ital., VI. Ser., A 1, 1--33 (1982; Zbl 0484.60003); Statistica 44, 21--45 (1984; Zbl 0557.62004)], since ``probability'' is usually and tacitly meant as ``countably additive probability''. This interesting note points out further specific situations, concerning in particular Savage's model, with the purpose of clarifying some misunderstandings arising in the relevant literature.