Radical Uncertainty Cannot Be Resolved Through Probablistic Thinking

"Radical uncertainty" is a term from Mervyn King and John Kay to describe situations in which there is no means to resolve the uncertainty. In contrast to "resolvable uncertainty," a situation in which answers can be found by looking something up or represented as a known probability, situations of radical uncertainty may produce not only results that we did not expect, but results we could not have imagined.

Such environments cannot be described with probabilistic thinking. Fields characterized by radical uncertainty, including business, politics, and finance, aren't governed by scientific law; they comprise countless individual actions that are neither rational nor optimal. Radical uncertainty is not measurable; there are too many shortcomings in the quantity and quality of the information that is available to us. Moreover, these situations are "non-stationary": they are in constant flux, and may change reflexively in response to our changing believes about them.

They are therefore impossible to model or forecast with any accuracy. Instead, decision-makers should ask "What is going on here?" and look to "thick description" to help make sense of things. (See Sensemaking articulates new perspectives and frames for understanding the world around us in particular Sensemaking produces thick data.)

In the absence of this work, we risk introducing alogrithmic thinking that may reify or amplify underlying structural challenges or inequalities that produced the very condition we hope to resolve. In other words, we risk unwittingly turning "evidence-driven policy" into "policy-driven evidence."



Kay, John, and Mervyn King. Radical Uncertainty: Decision-Making Beyond the Numbers. New York, NY: WW Norton, 2020.