In the last several years, much of the time when I was wearing my physicist’s hat I was working on a theory of modified gravity.
Modified gravity theories present an alternative to the hypothetical (but never observed) substance called “dark matter” that supposedly represents more than 80% of the matter content of the Universe. We need either dark matter or modified gravity to explain observations such as the anomalous (too rapid) rotation of spiral galaxies.
Crudely speaking, when we measure the gravitational influence of an object, we measure the product of two numbers: the gravitational constant G and the object’s mass, M. If the gravitational influence is stronger than expected, it can be either because G is bigger (which means modified gravity) or M is bigger (which means extra mass in the form of some unseen, i.e., “dark” matter).
In Einstein’s general theory of relativity, gravity is the curvature of spacetime. Objects that are influenced only by gravity are said to travel along “geodesics”; their trajectory is determined entirely by the geometry of spacetime. On the other hand, objects that are influenced by forces other than Einstein’s gravity have trajectories that deviate from geodesics.
Massless particles, such as photons of light, must travel on geodesics (specifically, “lightlike geodesics”.) Conversely, if an originally massless particle deviates from a lightlike geodesic, it will appear to have acquired mass (yes, photons of light, when they travel through a transparent substance that slows them down, such as water or glass, do appear to have an effective mass.)
Modified gravity theories can change the strength of gravity two ways. They can change the strength of Einstein’s “geometric” gravity (actually, it would be called “metric gravity”); or, they can introduce a non-geometric force in addition to metric gravity.
And herein lies the problem. One important observation is that galaxies bend light, and they bend light more than one would expect without introducing dark matter. If we wish to modify gravity to account for this, it must mean changing the strength of metric gravity.
If metric gravity is different in a galaxy, it would change the dynamics of solar systems in that galaxy. This can be compensated by introducing a non-geometric force that cancels out the increase. This works for slow-moving objects such as planets and moons (or spacecraft) in orbit around a sun. However, stars like our own Sun also bend light. This can be observed very precisely, and we know that our Sun bends light entirely in accordance with Einstein’s general relativity theory. This cannot be explained as the interplay of geometric curvature and a non-geometric force; photons cannot deviate from the lightlike geodesics that are determined in their entirety by geometry alone.
So we arrive at an apparent contradiction: metric gravity must be stronger than Einstein’s prediction in a galaxy to account for how galaxies bend light, but it cannot be stronger in solar systems in that galaxy (or at the very least, in the one solar system we know well, our own), otherwise it could not account for how suns bend light or radio beams.
I have come to the conclusion that it’s not galaxy rotation curves or cosmological structure formation but modeling the bending of light and being able to deal with this apparent paradox is the most important test that a modified gravity theory must pass in order to be considered viable.