The momentum paradox you refer to is (possibly?) the Abraham-Minkowski controversy ( https://en.wikipedia.org/wiki/Abraham-Minkowski_controversy ) about electromagnetic momentum in dielectric media. I'm not an expert on this subject, but I would doubt that this new work definitively settles the controversy. Of course this is, no doubt, work towards settling the issue. My (limited) understanding is that the controversy is really about interpreting certain quantities that behave like momentum in certain contexts, and which contexts apply in certain experiments to measure them. I do not believe it constitutes a crisis in our understanding of light; this is a very technical detail.
Of course, using the mechanism of momentum transfer to the transmitting medium as an explanation of redshift - and by doing so, refuting the expansion of the universe - is just yet another "tired light" explanation. (This refers to the idea of explaining redshift through a path-dependent loss of energy for photons traveling from great distances.) This is not a new notion, dating at least to Zwicky in the fifties. This article ( https://arxiv.org/abs/astro-ph/0106566 ) details efforts to demonstrate the reality of this expansion. These efforts do not assume anything about the exact mechanism responsible for tired light, merely the notion that light loses energy as it travels. They refute this to better than 10 sigma.
>with our understanding of light physics in the mid 20th century. >If light behaves even slightly differently to how we currently believe, everything from galactic rotation to expansion goes out of the window.
I... suppose I have to admit that if the current theory of light is wrong, then there may be changes to how we interpret these results. You should know though that this would be extremely unlikely. Generally revolutions in physics tend to subsume the effective results of the theories that are replaced. Quantum mechanics provides a good example: if you take the limit h -> 0 (making Planck's constant zero), you recover classical mechanics. Relativity provides another: if you take the limit c -> infinity in relativity, you also recover classical mechanics. Our understanding of light is quite good.
Forgive me, but I'm detecting a little bit of an "international scientific conspiracy" vibe here. Unpopular work is published all the time, provided that it withstands scientific scrutiny. I know that sounds like I'm dodging the issue, but really, you can apply a cui bono here: what do scientists stand to gain by propping up "wrong" science? We don't get paid a lot, you know.
>Think of epicycles, which for millenia seemed obvious and correct, until heliocentricity (for all bodies, not just earth) became the dominant model due to improved theory - but no new observations.
I don't mean to pick nits, but it was improved observations (Kepler, Brahe, etc.) that drove the acceptance of the heliocentric model. Kepler was famous partly due to his unprecedentedly accurate measurements. Theory was not necessarily rigorous at the time, often referencing theological arguments; heliocentrism was hotly debated, but not novel. Later it was realized that epicycles form a basis set for any trajectory on the surface of a sphere; any trajectory can be reproduced using a sufficiently large number of them. This was a pitfall that astronomers of the time could never have known. Remember that they did not have Newton's insights yet.
On a final note, its great to hear that ATLAS has some good evidence for gamma-gamma interaction!
edit: fix link