The urges of human nature constantly seek to compensate for future uncertainty by consciously or subconsciously constructing an edifice of expectations, upon which judgements of relative likelihood are applied and decisions made. Yet lacking future information, choice of action cannot depend on any set of quantitatively established, ranked probabilities; as such, expectations are inevitably constructed by means other than deduction and include “unreasoned elements” (Meeks, 1991). In Keynes' General Theory (1936) and his 1937 publication ‘The General Theory of Employment' in the Quarterly Journal of Economics, Keynes delves into rational choice theory. We can interpret how he essentially deconstructs and separates the Neoclassical assumptions of perfect information and rational expectations and depicts short-term pursuit of economic decisions as falling back upon a rationalised contingency plan or “convention” (Keynes, 1936; Ch.12, IV) when lacking future information. Keynes considers that this type of ‘irrational rationality' is such that the state of long-term expectation is determined on the whole, quite rationally. Given that expectations cannot be purely deductive, in doing so Keynes can be – perhaps legitimately – criticised for stretching the term ‘rational' beyond its scope for the sake of his economics analysis. However the following theoretical paper makes the case that Keynes' argument might be strengthened in impact, coherence and internal consistency if he were to take the term ‘rational' even further in a general theory of economic expectations beyond the remit of investment decision-making.

To make this argument, first the three interdependent sources of irrationality at the most abstract, microeconomic level are identified – namely, the basis for predictability, the basis for constructing expectations, and the basis for forward economic decision-making. Given Keynes' emphasis on irrational psychological influences having associated “factors exert[ing]...compensating effects” (Keynes, 1936, Ch.12, VII), we then illustrate how such factors (including volume of bi-directional fluctuations in decision-making, propensity for herd mentality, trust in personal relationships, and the role of the state) can demonstrate rational agent behaviour. These ignorance-overcoming factors (Carabelli and Cedrini, 2012) are broadly divided into those endogenous to individual decision-making that minimise short-term risk, and those exogenous to individual decision-making that promote long-term stability. Finally, we examine in depth the two conditions of endowment in which Keynes entrenches his argument – on one hand the need to remove dependency of mathematical expectations and on the other hand the “innate urge to activity...[driven by]...spontaneous optimism” (Keynes, 1936, Ch.12, VII). We demonstrate using these conditions how all of the three interdependent sources of irrationality that Keynes' convention introduces can be established in reasonable - or even rational - trails of thought on the microeconomic level.

On this basis, we consider Keynes' argument using non-investment examples make the case that a more general theory of economic expectations could smooth any inconsistencies and imprecisions in Keynes' convention argument. We conclude that the universally relevant relationship between abstract individual behaviour and long-term aggregate behaviour indicates the validity of broadening rather than narrowing Keynes' notion of ‘rationality' in both definition and application. Such an argument further substantiates the Keynesian prescription for greater regulation, lest undesirable activity become justified by 'rationality'.