We see the presence of giant black holes at the center of galaxies. It can be assumed that these black holes existed before and were the seeds of the galaxy by accreting, through their enormous gravitation, matter around them.

The opposite hypothesis can also be supported: galaxies were formed by an independent mechanism, the black holes at the center of galaxies resulting from the merger of stars or stellar black holes in the bulge of the galaxy where the density of stars is high.

It can also be assumed that the two processes are concomitant.

The thermodynamic approach favors this hypothesis because the gas cloud which, through fragmentation and collapse, will form a galaxy with a complex structure, including many stars, nebulae, clusters, planets, etc., all different, has a much lower entropy [1] than that of a cloud of progenitor gas, which can be described by few parameters (its density, its pressure,  its temperature, etc.., because it probably has a blackbody structure).

Since the entropy of an isolated system cannot decrease, in order for the local entropy in one subsystem to decrease, it is necessary that in this « isolated » system including it, in another subsystem of it, the entropy increases, at least as much.

Black holes are excellent candidates for this because by forming from a complex subsystem with low entropy, they will give it enormous entropy.

Indeed, Kerr black holes, which are the most physical black holes, are defined by only 2 parameters, mass and angular momentum. We ignore the static case and the case with electric charge, which are quite theoretical.

Thus, in the process of fragmentation and collapse of the over-density gas cloud, if some of the fragments collapse into black holes, this allows the other fragments to form complex subsystems with low entropy, respecting the laws of thermodynamics that the entropy of the isolated system cannot decrease.

Notes

In information theory, the amount of information needed to describe a system increases as it becomes more diverse and complex. Its entropy varies in the opposite direction (it decreases). Indeed, entropy is maximum for a system that is as simple as possible, requiring a minimum of parameters to describe it (a minimum of information).