Definitions and early twentieth-century foundations

Marc Lavoie (2014: 72) claims that the concept of irreducible or fundamental uncertainty draws a near-universal consensus among post-Keynesian economists, who consider it one of the foundation stones of the school itself and one of the defining characteristics that set it apart from mainstream economics. A consensus may indeed be present as to the overall importance of the concept, but there is nonetheless some debate as to its exact definition. In nearly all accounts of it, the following quotation is often used in the way of a defining preamble:

Some (such as Paul Davidson 1982) have connected this passage to a terminological distinction attributed to Frank Knight (1921) between “risk” and “uncertainty”. A condition of risk is said to arise when the precise future outcome of a process cannot be known, but when the range of its possible outcomes is known fully, and a probability can be assigned to each individual outcome on the basis of observed past realizations. The example of the roulette game, used by Keynes above, can be rationalized this way: where the ball might end up is, obviously, unknown, but all the possible slots in which it can fall are known ex ante. Moreover, a probability can be computed for the ball to fall in each slot, either in theory or in practice, by careful observation of what has actually happened in the past.