CONCLUDING REMARKS AND OVERVIEW
The reconciliation of familiar probabilistic notions with normal economic contexts and the visible actions of economic agents is often troublesome. Major difficulties stem from the historical specificity of economic events and from the structural instability produced when market conditions and individual circumstances are amended by unanticipated erratic shocks and frequent qualitative changes, as well as from the placing of probability calculations in recognizable mental processes, basic flows of information, and any dialogues occurring within the principal decision-making groups. Other major difficulties arise from an obvious lack of precision and the costs of both acquiring and processing suitable information to facilitate the use of probabilities - imprecision and costs which are not readily put into the abstract formulations of models with probabilistic elements, and which must effectively narrow the perceived range of applicability for such models to illuminate most economic phenomena. Probability has to be more carefully used or it has to be replaced by alternative notions (such as possibility, belief functions, fuzziness and potential surprise), which could offer a better fit to the economic context. We cannot presume that probability density functions, fixed parameters (e.g. means, variances and correlations) and simple stochastic processes will always provide a suitable means of expressing uncertain outcomes, incomplete information or ignorance. Instead, more scepticism needs to be expressed and alternative approaches vigorously sought and assessed for their potential usefulness.