In this paper, I accomplish a conceptual task and a historical task. The conceptual task is to argue that (1) Einstein's Principle of Separability (henceforth “separability”) is not a supervenience principle and that (2) separability and entanglement are compatible. I support (1) by showing that the conclusion of Einstein's incompleteness argument would still follow even if one assumes that the state of a composite system does not supervene on the states of the subsystems, and by showing that what Einstein says in “Quantum Mechanics and Reality” (1948) strongly suggests that separability is not a principle about how subsystem states relate to the state of composite systems. I support (2) by showing that if separability was incompatible with entanglement, then Einstein's argument would be incoherent in a trivial way. Thus, by arguing for (1) and (2) I directly challenge what has been, and still is, a very common reading of separability. The historical task is to offer the first detailed review of the different ways in which separability has been defined by physicists and philosophers in the last 60 years. Among other things, such a review distinguishes three different definitions of the principle, and shows that since the 1990s and up until the present date, it became standard to take separability (as presented by Einstein) to be a supervenience principle.