In 1924, Bose sent his rederivation of Planck’s blackbody formula to Einstein, who translated it into German and had it published in the Zeitschrift für Physik. However, he did not stop there. Bose’s main idea was to apply Gibbs’ ensemble construction to an assembly of indistinguishable photons, by a purely combinatorial argument. Einstein realized that the same idea could be applied to a gas of material particles, which he proceeded to do himself, in two papers he published within a year in the Sitzungsberichte der Preussischen Akademie der Wissenschaften zu Berlin. In the second of these, he predicted the phenomenon of Bose-Einstein condensation, which was experimentally observed only in 1995. It was noted by contemporaries that the statistics employed by Bose and Einstein was incompatible with the assumption of independent particles. Indistinguishability causes kinematic correlations even in the absence of interactions, a point not made sufficiently clear in Einstein’s first exposition, as he admitted later. Bose’s innovation was to treat states of motion as independent, rather than particles. He thus inverted standard physical reasoning, which begins with observable particles, in favor of an abstract approach in which states of motion are the primary quantities. This inversion was present already in Gibbs’ phase-space ensembles, but it was latent there, because one could (mostly) leave it undetermined, whether \(\boldsymbol {p}\) and \(\boldsymbol {q}\) in some expression referred to laboratory or phase space. It was arguably Einstein’s reframing of Bose’s work that brought the abstract approach to its first radical conclusion in the realm of quantum mechanics.