Informations and abstract
Political scientists show a very limited confidence in attempts at forecasting. To all hypotheses aiming at the prediction of the future they much prefer statistical projections that are satisfied to reach the smallest possible uncertainty or the realistically highest degree of probability. Among the approaches implying these attitudes one may single out those connected with the institutional as well as the systemic analyses of politics. In no other political science approach but the systemic, induction always ends with a probabilistic conclusion. Therefore, the cognitive outcomes of every forecasting activity present rather different degrees of probability. On its part, the institutional approach indicates that all forecasting inquiries concerning the processes of construction of political institutions appear very complicated. All the changes represented by the weakening of parliaments, the strengthening of the executives, party government and the personalization of leadership are filtered through phenomena of inertia and routine resistances. Hence, when aiming at acquiring knowledge that may identify objective causal relationships, any forecasting effort encounters many a difficulty. The pursuit of this kind of knowledge is then obliged to take into account several and most diversified variables as well as their covariation. The findings are often quite poor. Nonetheless, this article suggests that the risk to make mistakes and to reach often disappointing outcomes should not deter the scholars from trying to engage in a sustained forecasting effort. After all, the findings will still be susceptible to revisions and corrections.