Bertrand Russell on Keynes's A Treatise on Probability: What One Would Expect from the Twentieth Century's Greatest Philosopher

18 Pages Posted: 22 Jun 2016

See all articles by Michael Emmett Brady

Michael Emmett Brady

California State University, Dominguez Hills

Date Written: June 21, 2016


Bertrand Russell gives an excellent review of Keynes’s A Treatise on Probability. It is what one would expect from the Twentieth Century’s top philosopher. Russell concentrates on Parts I, II and III of the TP and shows that Keynes had done a spectacular piece of path breaking work. Russell places heavy emphasis on the analysis that Keynes presented on pp. 234-37 (also pp. 254-257) of the A Treatise on Probability (1921), which is built on and derived from a modified problem of George Boole’s that Keynes used as his foundation. Russell is even more effusive of praise for Keynes’s achievement in his 1959 My Philosophical Development. The objections of Ramsey and Popper are simply ignored as being insignificant. Thus, Russell handles Ramsey and Popper in the same way that Adam Smith handled Jeremy Bentham -he simply ignores the body of their work dealing with Keynes as being of no consequence. Russell also noticed that an improved relative frequency theory of probability is needed in order to deal with work in sciences like engineering, physics, chemistry and biology and notes that Keynes’s criticisms of these types of theories are not as relevant when dealing with a frequency theory such as Whitehead’s. Russell was also aware of Keynes “weight of the evidence” analysis and found no flaws as weight must be consistent with the relative nature of probability. Thus, the claim the Keynes uses weight in an absolute sense in the A Treatise on Probability is not mentioned by Russell. Finally, Russell does not commit the fatal mistake made by all Post-Keynesian and Institutionalist economists, which is conflating Keynes’s “non numerical" probabilities with ordinal, rank order or comparative probability, as Frank P. Ramsey did. Russell recognizes a possible role for "numerical approximation” and shows that it would have to be an interval. Unfortunately, Russell appears to have overlooked Keynes’s own approximation technique based on Boole’s “indeterminate” probabilities approach.

Keywords: approximation, numerical approximation, chapter 26, weight of the evidence, positive and negative analogy

JEL Classification: B10, B12, B20, B22

Suggested Citation

Brady, Michael Emmett, Bertrand Russell on Keynes's A Treatise on Probability: What One Would Expect from the Twentieth Century's Greatest Philosopher (June 21, 2016). Available at SSRN: or

Michael Emmett Brady (Contact Author)

California State University, Dominguez Hills ( email )

1000 E. Victoria Street, Carson, CA
Carson, CA 90747
United States

Here is the Coronavirus
related research on SSRN

Paper statistics

Abstract Views
PlumX Metrics