If the emergence of physics as a definite academic discipline was a heritage of the late nineteenth century, the emergence of a new theoretical practice, and the settlements of chairs of theoretical physics were the most interesting outcome of that process. The hallmark of the new theoretical practice was the awareness that the alliance between the mathematical language and the experimental practice celebrated by Galileo had to be updated. Besides “definite demonstrations” and “sound experiments” there was a third component, which could be labelled conceptual or theoretical: it dealt with principles, models, and patterns of explanation. That conceptual component, neither formal nor empirical, came to be looked upon as a fundamental component of scientific practice. It is worth remarking that, in that fin de siècle, science had finally managed to realize, at least in part, Bacon’s dream, and the myth of scientific progress emerged. In the debates on science which took place in France from the early 1870s to the early 1890s two main issues were at stake: determinism and reductionism. On the one hand we find some scientists, historians, and philosophers who relied on simplified epistemological and historiographical frameworks, and put forward an optimistic cult of human progress. On the other hand, a sophisticated point of view on science was put forward by scientists and philosophers who did not deny the effectiveness of scientific progress but were able to go beyond the simplified conception of scientific practice as an unproblematic alliance between mathematical and empirical procedures. In 1892 the young physicist Pierre Duhem published the first paper explicitly devoted to meta-theoretical issues or, to make use of a more recent expression, to philosophy of science. At that time he had already published a book on thermodynamic potentials and their applications to different fields of physical sciences, and a demanding paper, where he had put forward an original mathematical approach to thermodynamics on the track of Analytical Mechanics. Theoretical physics, the history of physics, and meta-theoretical remarks on science were mutually interconnected in Duhem’s actual praxis. The historical and epistemological remarks he began to publish systematically in the 1890s were subsequently collected in the book he published in 1906, La théorie physique, son objet, et sa structure. He represented the scientific enterprise as a three-stages task: from the knowledge of “specific facts”, the human mind was able to derive some “experimental laws” by induction, and then create a scientific theory. If the objects of experimental laws were facts, the objects of physical theories were experimental laws. In any case, a theory had nothing to do with the truth: it could not be qualified as true or false, but “suitable or unsuitable, good or bad”. The plurality of theoretical frameworks corresponding to a set of laws was consistent with this essential feature of theories. Moreover he put forward three fundamental theses on experimental physics: first, a physical experiment was not a purely empirical process; second, it could not be so powerful as to lead to the refutation of a single hypothesis; third, it was less reliable, even though more precise, than ordinary experience. After the Second World War some themes which had been put forward in the late XIX-century philosophy of science re-emerged in an unexpected way. In reality, Duhem’s books and papers had almost been forgotten, but a new interest in some of his meta-theoretical theses emerged in the context of a philosophical tradition that was deeply linked to logic. In 1951 Willard van Orman Quine sharply criticised both the dichotomy analytic/synthetic and reductionism, but he neither quoted from nor mentioned Duhem. In 1960 Adolf Grünbaum put forward a refutation of what he called “Duhemian argument”, but the core of Duhem’s meta-theoretical remarks got lost in a net of logical deductions which were extraneous to their context. Only from the 1970s onwards historians, historians of science, and philosophers of science began to be attracted by late XIX-century context in general, and Duhem’s philosophy of science in particular. Late XIX-century philosophy of science stemmed from a remarkable epistemological and historiographical awareness, and that awareness would deserve to be further explored.