>the interaction happens over a line beam (rather than a spherical surface) so it doesn't drop off as 1/r^2, but as 1 (i.e. constant). Which raised more questions for me (and made me wonder what the MOND case is)
To me the MOND is 1/r as i think the very weak gravity acting only in the plane of the galaxy disk is, very roughly speaking, a result of quantization - i.e. "not enough" gravitons to interact in all spherical directions and thus gravitational field basically exists only in that plane. It is like a mental experiment - say we generated a classic EM spherical wave yet of a very low energy of just one photon, and have several other charges placed at the same distance from the wave source - while the classical 1/r2 would have the wave interacting the same way with all the charges that would be a violation of energy conservation in our low energy "one photon" case where only one charge at best would get interacted with and thus it would look like supposedly violating 1/r2 law of the EM.
I think you're confusing 2 things. The "energy of a field", whether that's a magnetic field or gravitational field is fictional. It's potential energy.
In order to move from A to B you must "pay" the difference in potential energy between the 2 points in space. That payment can be negative (e.g. falling).
So if gravity increased in strength after ~2000 light years (which is the problem dark matter tries to solve) to 1/r instead of 1/r2 that would not represent any energy at all. It would not insert energy anywhere, into any particle, it only changes the "fictional" values of potential energy in a bunch of locations. Therefore it would not violate conservation of gravity.
Oh, and things form discs by default. If things fall into something, they form a disc shape. Round things are only formed once the collisions between stuff in the disc start going over a certain level. Galaxies are so incredibly low-density there are even a few galaxies that have multiple discs, but still very much discs. Only "small" things are ball-shaped, like stars and planets because the particles exert pressure on each other and the third dimension provides a way to relieve the pressure.
To me the MOND is 1/r as i think the very weak gravity acting only in the plane of the galaxy disk is, very roughly speaking, a result of quantization - i.e. "not enough" gravitons to interact in all spherical directions and thus gravitational field basically exists only in that plane. It is like a mental experiment - say we generated a classic EM spherical wave yet of a very low energy of just one photon, and have several other charges placed at the same distance from the wave source - while the classical 1/r2 would have the wave interacting the same way with all the charges that would be a violation of energy conservation in our low energy "one photon" case where only one charge at best would get interacted with and thus it would look like supposedly violating 1/r2 law of the EM.