Recently, I have been discussing Einteinian and Whiteheadian cosmology with Gary Herstein, who is author of Whitehead and the Measurement Problem of Cosmology. (http://www.amazon.com/Whitehead-Measurement-Problem-Cosmology-Process/dp/3937202951) I strongly recommend it for process metaphysics and those concerned with speaking of process metaphysics in terms of contemporary cosmology, although the book is focused on the latter.
I have always struggled to understand Einstein's view of space-time and gravity. I suppose that we're all seen the curved space models, which is what I always imagine. While I get the idea, I never understood why and short of actively studying cosmology--us philosophers already have long reading lists--I wondered if I would ever understand it. I then had an insight.
Gravity is the connectivity of space.
(For "connectivity," see http://mathworld.wolfram.com/SimplyConnected.html and for the "connectivity of space," see http://mathworld.wolfram.com/ConnectedSpace.html.)
I asked Gary and he said that I was on the right track. However, he noted that Whitehead criticized Einstein's theory, and said that the way I framed the question was "making the Einstein & SMC (Standard Model of Cosmology) move of equating the physical with the geometrical." This employs only "one metric tensor," whereas Whitehead's was "bimetric." He also mentions the alternatives, Modified Newtonian Dynamic and Tensor-Vector-Scalar (which I've heard of). Now for his interesting comment.
Part of his critique was that these alternatives to the SMC "do not require the 'ad hoc' assumptions of miracle features like 'dark matter' and 'dark energy.'" This is as I always suspected: these strange phenomenon are likely artifacts of mathematics and if they have any reality are likely misrepresented.
For my part, I was thinking that blackholes might be a case where connectivity fails to be "simple connectivity" (see the link above), which may occur under extreme conditions. That is, parts of space are not longer "geometrically contiguous" even though they are "connected." This would be possible, upon my hypothesis, because gravity is the connectivity of space, and gravity is not strictly geometric. If this doesn't seem plausible, then I ask someone to explain the plausibility of quantum teleportation and entanglement. Gary is right; geometric models appear insufficient.
A side thought. This hypothesis casts "time" as if it were a relative positionality in space or connectivity. I am still musing upon this.