J M Keynes's Concept of Uncertainty has Nothing to do with Paul Davidson's Ergodic-Non Ergodic Distinction, which Requires that the Limiting-Relative Frequency Interpretation of Probability be the General Case
22 Pages Posted: 9 Dec 2013 Last revised: 14 Dec 2013
Date Written: December 6, 2013
Paul Davidson has erroneously claimed for 30 years that J M Keynes’s definition and application of his concept of uncertainty in the General Theory is based on the ergodic-non ergodic distinction. It is easy to show that this is false because it would require that J M Keynes was an advocate of the Limiting Frequency or Relative Frequency approaches to probability. Keynes rejected the generality of the Limiting Frequency or Relative Frequency approaches to probability. The Limiting Frequency or Relative Frequency approaches are applicable only in cases where the phenomena under investigation are highly repetitive and repeat continuously over time. Innovations in physical durable capital goods and investment in such goods is either a onetime event or occurs infrequently. There are no repeated series or sequences of continuous innovations over time. Schumpeter correctly pointed out that innovations occur irregularly and in bunches that can’t be replicated.
Keynes recognized that the Limiting Frequency or Relative Frequency approaches to probability could be applied successfully in certain fields in the physical and biological sciences like Physics, Chemistry, and Biology, but that it was not the general case in fields like the Social Sciences and Liberal Arts because it ignored evidence based on infrequent or unique occurrences that took place in historical time. Such evidence had to be analyzed using Keynes’s case based analysis that was founded on an eliminative inductive logic using analogy and degrees of similarity, dissimilarity, and resemblances between single cases or infrequent cases.
The reason for Davidson’s failure to understand and grasp Keynes’s concept of uncertainty was because Davidson, like Shackle, deliberately chose to ignore the clear cut connection that Keynes made in two footnotes in the General Theory that established the link between Uncertainty and the Weight of the argument (evidence). This error was catastrophic, as it completely vitiates all of Davidson’s subsequent work that claims that Keynes was a proponent of the ergodic-non ergodic approach adopted by Davidson. Nowhere in the corpus of either Davidson’s or Shackle’s work is Keynes’s concept of weight discussed, evaluated, or mentioned seriously. This error is fatal to the entire Shackle-Davidson Post Keynesian position.
All of the different neoclassical schools of thought reject the limiting frequency interpretation of probability, advocated by Davidson, for use in economics, finance and business. All of these schools of thought are based on Subjective Expected Utility (SEU) theory, which is based on the Ramsey-De Finetti-Savage-Friedman personalist, psychological, subjectivist, Bayesian approach to probability, which completely rejects Davidson’s totally false claims about probability having to be defined as a limiting frequency in order to calculate numerical probabilities. Davidson’s claims will have no impact or effect on any neoclassical school. All that these claims reveal is Davidson’s own extraordinary ignorance.
Keywords: weight of the argument (evidence), uncertainty, decision making, interval estimates, Logical Theory of Probability
JEL Classification: B30
Suggested Citation: Suggested Citation