No label defined (Q4530613)

The author studies the history of the methods of dealing with uncertainty (p. ix) from antiquity to Huygens and Leibniz (rather than to Pascal) and pays special attention to the relevant qualitative stochastic reasoning. His book contains useful, sometimes hardly known information concerning law, philosophy, medicine, religion, and he argues that the Middle Ages were fruitful and important for the further development of science (and of probability theory in particular). The author also discusses astronomy (Copernicus, Galileo, Kepler), aleatory contracts, dice games and lotteries, again with the least possible use of numbers, and he describes an early solution of the problem of points (ca. 1400). Many shortcomings are conspicuous. Ptolemy's reasonable treatment of direct observations is lamely dealt with; the idea underpinning the law of large numbers (Cardano, Kepler) is neglected and the fundamental problem of separating law from chance mentioned only in passing. Then, passed over are the links between the medieval doctrine of probabilism and nonadditive probabilities (Jakob Bernoulli); between the qualitative approach to decision making and the very nature of ancient science, or the recently introduced assessment of expert estimations. The nonnumerical ``methods'' of dealing with uncertainty are left unsystematized; moreover, they hardly exist, they should have been called principles, and connected strongly, not in the author's feeblest way, with Newton's rules of reasoning in philosophy. Many sources and a host of commentators are quoted but the references are not alphabetically arranged, nor are the pertinent authors included in the index and in many cases the dates of the original publications are not provided. Documentation is often offered only in general, and some specific statements might be mistakenly attributed to Franklin himself.