Gibbs, Einstein and the foundations of statistical mechanics

Abstract

It is generally accepted that, around the turn of the century, GIBBS and EINSTEIN independently developed two equivalent formulations of statistical mechanics. GIBBS' version is taken as genuine and rigorous, while EINSTEIN's, despite some features which are characteristic of him, is usually considered a not totally satisfactory attempt. It will be shown in the present work that such a picture is oversimplified and requires further nuancing. In fact, there are significant differences, with important implications which have not been sufficiently examined, between the two formulations. Several issues relating to the impact of the two versions will also be addressed. © Springer-Verlag 1998.