Historical synthesis supporting Duhem's demonstration of the impossibility of the `experimentum crucis' of comparison of the speed of light in air and water

Abstract

In 1850, Foucault was the first to compare experimentally the velocity of light in air to its velocity in water. And as the wave and projectile theories competing at the time predicted inverse results to such a measurement, Foucault ''declared the emission system incompatible with the reality of the facts''. Yet, in his Physical Theory, Duhem uses this very example to illustrate his epistemological demonstration of the impossibility of the 'experimentum crucis' in physics. Then, the ambition of the present article is to augment Duhem's demonstration with historical evidence relating precisely to the question of the speed of light in water. We will review the four major opinions on the nature of light developed between 1637 and 1801; all leading to theories concluding that the ratio of sines is constant and equal to the ratio of the velocities in the two media, but all being defended by at least two authors concluding to inverse ratios of those velocities -- and thus to opposite predictions on the result of an experiment such as Foucault's. The convergence of historical evidence will then confirm that the nature of light, considered in isolation, is in no way constrained by the measurement of its speed in water.