The development of statistical methods in the sciences presents an interesting puzzle for philosophers and historians of science. These methods were developed in a context having nothing to do with the natural sciences: namely, the measurement of populations for the purposes of governance and social policy. Very quickly, they were taken up into both biology (a tradition that would become statistical population genetics) and physics (leading to statistical mechanics and thermodynamics). But key figures in this transition are difficult to understand. The leading early proponent of statistics, the Belgian statistician Adolphe Quetelet, offered a methodology for the social sciences that no one ever actually wanted to pursue. And the first person to develop these tools in the life sciences, Francis Galton, was unclear enough about the nature of statistical reasoning that his colleagues (as best as we can tell) failed to fully understand his works. In this talk, I'll offer some first steps towards reconstructing this larger story: first, understanding just what it is that Galton was trying to do, and then pursuing a different path that led to the early statistical works of James Clerk Maxwell, on the Maxwell-Boltzmann distribution for molecular velocities. I hope to offer, via this case study, a window into the kinds of conceptual and historical challenges that make the philosophy of science a fruitful complement to scientific study.
Pence, C. H. (2026, April 28). How the Natural Sciences Became Statistical: From Quetelet to Galton and Maxwell. Alikhanyan National Laboratory, Yerevan, Armenia. https://hdl.handle.net/2078.5/280806