The Genesis of Einstein's Work on the Problem of Motion in General Relativity
There are two equations at the very heart of General Relativity (GR): The Einstein field equation, which governs the dynamics of gravitational fields, and the equation of motion of bodies subject to gravitational fields, the geodesic equation. In his first review paper on GR in 1915, Einstein was very careful to introduce both equations as separate assumptions, as the two pillars on which the rest of GR was to be built.
We know that, from early on, Einstein wondered whether he really had to introduce both equations as separate assumptions; but he never addressed the issue in print until twelve years later. In Volume 15 of The Collected Papers of Albert Einstein, Einstein addresses "the problem of motion" of GR for the first time: the question whether it might be possible to derive the equations of motion from the gravitational field equations after all.
Imagine how amazing that would be: one could derive the motion of bodies from knowing nothing but the gravitational field that surrounds them. Up until our work on volume 15, it was a puzzle as to why Einstein waited twelve years to address this possibility in print. But now we understand that Einstein just did not know a way to address the problem in a way that he found satisfactory.
The key to understanding why all this changed was hitherto unknown correspondence with the mathematician G.Y. Rainich, included in this volume. In Einstein's correspondence with Rainich we see how he changed his mind about what really was, as he put it, the "cardinal question" in gravitational theory, and how, during their correspondence, Einstein finally sees the light: a possible solution to the problem of motion.
Those who want to read more about this should read section II of the introduction to Volume 15, available here, or one of the research papers on the topic that came about as a result of working on the volume, available here.