LAUSR

In this article we follow, over his long career, the evolution of the reflections of Edmond Malinvaud on economic ?science? and its relationship with the ?problem? of statistical induction. To do so we largely rely on his own words by quoting his many articles addressing these issues. We consider in turn his views on statistical methods and structural econometrics, on apriorism versus empiricism, the Koopmans-Vining controversy on measurement with or without theory, econometric methods and the challenge of modelling, the Popperian temptation of a falsifiable economic science, the uneasy alliance between induction and expertise, understanding and advising. We conclude with Malinvaud's views on eclecticism, which he deems to be the only position able to meet, despite its rapid progress, the "advisory imperative" characterizing economics.