This chapter explores the crucial role of ignorance awareness in economic decisions. It introduces a second-order measure of uncertainty a based on Ignorance I and Knowledge K:

a = I / (K+I))

ranging from 0 (risk, or complete relevant knowledge) to 1 (radical uncertainty, or complete relevant ignorance). This measure of ‘awareness of ignorance’ is similar, but antithetic to, Keynes’s ‘weight of argument’. Alternative decision theories are adequate for different degrees of ignorance awareness: classical expected utility theories (frequentist or subjectivist Bayesian) apply only under risk (a=0); theories based on non-additive probabilities apply under partial ignorance (0<a<1); while under radical uncertainty (a=1) the decision-maker may apply only non-probabilistic choice criteria (like the Maximin). Classical theories treat learning merely as a by-product of optimization. On the contrary, ‘objective’ Bayesian analysis is focused mainly on learning from data. The Rational Expectations Hypothesis (REH), central to New classical macroeconomics, assumes zero ignorance awareness (a=0), justifiable only in familiar, stationary ‘small worlds’, not in complex, evolving economic systems. Finally, this chapter rejects the viability of classical microfoundations based on classical decision theory, arguing for the autonomy of macroeconomic analysis and the need for approaches acknowledging the crucial role of complexity and ignorance awareness.