It has long been argued in philosophy and statistics that there exist many facets to uncertainty. First, there is irreducible uncertainty (Risk), one that cannot be avoided even with the best knowledge possible. Second, there are things one knows that one does not know, an hence that one needs to learn — Estimation Uncertainty. Lastly, there are things that one does not know that one does not know, which therefore emerge in complete surprise — Unexpected Uncertainty. Here, we present a new paradigm with which to formalize these facets of uncertainty, and in which all take on equal prominence. We show how to measure the different types of uncertainty, and present behavioral evidence that humans appear to be aware of all of them, selectively reacting to each type. Finally, we discuss potential neurobiological foundations for the human capacity to distinguish between Risk, Estimation Uncertainty and Unexpected Uncertainty.

Comments are closed.