In industrial countries individuals and collectivities are today more and more faced by the limits of knowledge. They have to accept decisions made with insufficient information — in other words they have to make trial attempts the errors of which must be assumed by those who have made the decision. The trial and error process that one then discovers is no doubt not a novelty in itself, but its requirements, breaking with a great many common practices, seem formidable. When the expert loses his credibility the politician is increasingly isolated with his responsibilities, which are greater than ever before. The people he represents are ever more conscious of participating in decisions the risks of which concern them directly. The political attitude must then both include in its choice factors that until now it had only partially assumed — such as distant deadlines, complexity, the risk of unforeseen factors, the impossibility of going back on decisions, and be an element that encourages the forming of a consensus of opinion in collectivities since without the consent of the latter no attempt can be made. The stress is laid on compromise rather than on technical solutions and this leads to a redistribution of roles among expert, politician and the population in favour of the latter. What conditions enable the maximum to be got out of trials the implementation of which must obviously be accompanied by a minimization which invariably exists of the risks run ?
