THIRTY YEARS AGO, when Koopmans began to be well known as an economist, he seemed to be a foreigner in a field where his approach to problems was often so different from the prevailing ones. Today a large group of young economists follows his path to scientific discovery. In this respect the book under review is particularly interesting. Not only does it serve as an easy access to articles that appeared in a large number of journals, but it also gives to the careful reader an opportunity to understand the true personality of Tjalling Koopmans, the scientist, and, beyond him, the methodology of modern economics.

During the last decade, while overlapping generations provided the framework for the study of many theoretical issues, no one seems to have noticed the use of this model in the book published by Maurice Allais in 1947, Economie et intert. The introduction of the model is commonly attributed to a well-known article published in 1958 by Paul Samuelson. This note intends to bring the unnoticed anteriority of Allais' use of the model to the attention of a wide readership. The case may be of interest not only to economic theory but also to the more general historicological study of the process by which science is being built. Indeed, my role in the story is surprising and may reveal something worth remembering. I had closely studied Economie et inWret in 1948, in particular the appendix in which the overlapping generations model is used. I later worked on capital theory and intensively reflected in 1958 on the relationship between Samuelson's findings and my own results. But I did not pay attention to the similarity between the models used independently by Allais and Samuelson. It is only recently, when looking for what might be the subject of my contribution to a festschrift in honor of Allais that I became conscious of this similarity (Malinvaud 1986). In the meantime, like others, I referred to Samuelson's article when introducing overlapping generations in my own writings. '