Mirowski has done a great service by collecting a number of Edgeworth's most important papers in one book.All of these essays are worth reading again.I will concentrate my review on Edgeworth's 1922 article in Mind, " The Philosophy of Chance " ,which is essentially an extended review of J M Keynes's A Treatise on Probability ,and Edgeworth's formal 1922 Review of the A Treatise on Probability in The Journal of the Royal Statistical Society which hews to the standard line of statisticians,which is to use the Normal distribution without first doing any kind of goodness of fit test first .Edgeworth's overall assessment of the A Treatise on Probability (TP;1921) in his two reviews taken as a whole was, by far, the most erudite and careful examination of the TP that was ever published.It is vastly superior to the reviews of Harold Jeffreys, Frank P Ramsey ,Ronald Fisher,Broad,Joseph,Wilson,etc.It is a great tragedy that the highly defective and error filled Ramsey and Fisher reviews were to become the dominant ones in the literature.In fact,Edgeworth was the ONLY reviewer of the TP who actually had taken the time to read,as best he could,most of the TP.Edgeworth is very forth right in acknowledging that there were parts that he could not report on because he did not have the technical competence to carefully assess their merit.One example would be Keynes's interval estimate approach.However, Edgeworth argued that it would be up to other readers of the journal, Mind,to offer a more detailed critique of a number of Keynes's technical innovations.Unfortunately,this never occurred.Edgeworth realized that Keynes was arguing for the existence of interval estimates,although he could not figure out the Boolean foundation upon which Keynes's interval estimates were erected.Edgeworth never fell for the truly idiotic idea,a commonplace among Post Keynesian economists like J Runde,A Carabelli,R O'Donnell,A Fitzgibbons,D Dequech,etc.,that Keynes's nonnumerical probabilities meant that Keynes did not use any numbers in his assessment of a probability .Edgeworth was the first,and last, economist to suggest that J M Keynes's conventional coefficient of weight and risk,c,where c=p2w/(1+q)(1+w)= p[1/(1+q)][(2w/(1+w)],was worth serious consideration when combined with an outcome A.One modification of this model,c=p/(1+q) that occurs when w=1, explains completely the 1979 Tversky -Kahneman analysis in Econometrica of the certainty ,reflection,translation,and preference reversal effects. The goal is to maximize cA,where A is the outcome,as opposed to the EMV(EU) rule's aim of maximizing pA[(pU(A)].Edgeworth was certainly correct that c should have been seriously examined by readers of Mind.Unfortunately,this never happened in the 20th century with one exception.This reviewer,aided in the beginning by Howard Lee in 1987,is the only researcher who seriously examined the mechanics of Keynes's decision rule.It is only in the published and unpublished papers of Brady and Brady and Lee that Keynes's c coefficient was demonstrated to be the equal of the Tversky -Kahneman Prospect theory or the Gilboa-Schmeidler non additive capacities approach.
Edgeworth's most important essays are worth reading again.I recommend that a reader buy this book .
