Introduction

Man is concerned about the future and hence ready to listen to anyone claiming to forecast. Foretellers have always had a credulous audience. The rapid development of science and the esteem with which it is regarded have not been enough to make this divinatory activity disappear.

This is because man is in a hurry; he feels a strong urge to know what tomorrow will be like: to understand before forecasting would take much too long. In fact, many believe in forecasts which are not really based on a true understanding of the phenomena in question.

Today, I would like to urge the participants of this symposium to beware of the attitude that treats forecasting as an autonomous discipline. Forecasting would only be truly autonomous if it could be separated from understanding phenomena. Conscious of being provocative, I may say that the main raison d'être of an autonomous science of forecasting, if it was built, could well be to justify resorting to foretellers.

I will on the contrary defend the thesis that a thorough analysis of phenomena is the true source of progress in the art of forecasting. My talk will contain three parts. In the first part I shall claim that forecasting without theory is rarely efficient and therefore studying forecasting methodology as such can only play a rather limited role in improving forecasts.