The Aether won't go away and the JWST may be giving the theory a boost
Ever since my research drew me into the alternatives to dark matter of MOdified Newtonian Dynamics (MOND), I have been reading more and more papers on the Aether.
You may recall that the Aether was proposed as the means by which light propagates through space. Intuitively, this makes a lot of sense because all other waves propagate in a medium. Water waves are just water moving up and down. Sound waves are just air or water moving in and out.
The Aether was such an intuitive concept that it was proposed that we could measure its existence because the Earth must be traveling through it. This was known as the Aether wind.
Michelson and Morely, two 19th century experimentalists, constructed several devices to measure the speed of light in multiple directions in order to detect the wind. These devices could precisely measure minute differences in the propagation through the Aether of the Earth itself using interferometry. They found no differences.
It is the most famous failed experiment ever, and their experiment has been repeated numerous time with ever increasing accuracy to this very day. No evidence of the Aether has ever been discovered.
Einstein eventually used their null result to postulate that the speed of light was a universal constant. Indeed, Maxwell’s equations, which govern the propagation of all electromagnetic radiation including light, appeared to suggest that no matter the velocity of the observer, that radiation would always propagate at light speed.
This is not true of any other wave. All other waves have a reference frame. If you were to make a sound, you could catch up to it using a very fast jet or rocket. Not so with light.
With the introduction of relativity, the Aether was scrapped as an idea or, more correctly, replaced with General Relativity, the theory of space and time. We now know that when space and time move as waves, these are gravitational waves which we have only recently been able to detect, not light waves. Light waves are, in fact, a quantum phenomena: the wave functions of many overlapping photons propagating in space.
Einstein-Aether theory was born out of a desire to study Lorentz covariance symmetry breaking. Like the original Aether, it creates a preferred reference frame. Lorentz covariance is, of course, what Einstein used to create relativity, and it is baked into quantum field theory, which is used to understand what is going on in particle accelerators where things are moving at a significant fraction of the speed of light. You can think of Lorentz covariance as the postulate that there is no preferred reference frame in the universe. All measurements and observations are relative to you and what you are measuring or observing.
Lorentz covariance has never been shown to be violated despite many very stringent tests, but, when it comes to quantum theory, it causes problems. Lorentz covariance is at odds with having any cutoff of energy or momentum at the smallest scales, for example, and this is a major roadblock to developing a theory of quantum gravity.
While early versions in the 1980s introduced a scalar field as a universal time, a paper in 2000 introduced the modern version using a vector field. A modern, relativistic vector field is essentially a set of four numbers pointing in a particular direction relative to your coordinate system in space and time defined at every point in the universe. This is the version that everyone uses today in various generalized forms. The key feature of the vector field is that it orients the universe in a particular time direction.
This idea is not new. It goes back to Isaac Newton who believed that the universe had a universal clock that ticked down the seconds. Newton’s idea doesn’t work with Einstein’s theory of general relativity, however. A static external clock structure would require we abandon the whole theory because it would create infinities in the Einstein equations.
Einstein-Aether theory doesn’t do this. Instead, it creates a dynamical universal clock that can bend and warp with matter and gravity. This maintains the structure of the Einstein equations while still breaking Lorentz covariance.
In 2004, Bekenstein (of Bekenstein-Hawking area entropy formula fame) introduced his Tensor-Vector-Scalar (TeVeS) theory as a relativistic version of MOND. While Bekenstein borrowed extensively from Sanders’ earlier stratified theory, Sanders’ own vector field, like that of Newton, was not dynamical.
As Bekenstein said: “Admittedly Sanders’ stratified theory is a preferred frame theory, and as such outside the traditional framework for gravitational theories. But it does point out a trail to further progress.”
In order to create a real theory of gravity out of what Sanders had done, Bekenstein did the same as Einstein-Aether theory. (Although he doesn’t indicate he borrowed from that theory at all, his equations of motion were very similar.) Thus, TeVeS also breaks Lorentz covariance.
Since then, both theories have become generalized and effectively merged with one another. (When I say “generalized” I mean that the theories have had their variations merged together into one theory where various pieces can be zeroed out to arrive at a specific instance or in some cases constants have been promoted to fields.)
While Einstein-Aether theory wasn’t originally introduced as a competitor to dark matter and the vast majority of papers about EA are on Lorentz violations, some researchers recognized, thanks to Bekenstein, that it could be a substitute for dark matter.
There is a lot I could say about EA theory, but I want to focus on one of its lesser known aspects which differentiates itself from straightforward, non-relativistic MOND.
MOND, you may recall, is that theory introduced in 1983 that seems to do so well in explaining the rotation curves of galaxies (known as the Tully-Fisher relation). These rotation curves don’t agree with Newtonian physics unless you add dark matter or modify said physics. MOND modifies it so that when accelerations are very small, the dynamics changes to become nearly constant with increasing distance. This allows the stars in the outer reaches of galaxies that have very weak pull on them to rotate about their galactic centers much faster than they would with Newton’s inverse square law applied to ordinary rotating bodies.
Bekenstein and Milgrom in 1984 turned MOND into a modification of Newton’s law of universal gravitation and called in AQUAL (A QUAdratic Lagrangian). Most people use the term “MOND” when they mean “AQUAL”. The alternative would be to modify the laws of inertia which would affect the conservation of energy, momenta, or angular momentum or modify how we represent material bodies in physics. I will follow the convention, however, that when I say “MOND” I mean a modification of gravity.
One of the problems with MOND is that it couldn’t explain gravitational lensing and, increasingly, astronomers estimate the amount of matter in galaxies and in between galaxy clusters by how light is bent as it comes to us. TeVeS was intended to solve that problem.
Straightforward applications of MOND and TeVeS to lensing run into problems when gravitational centers of galaxy clusters don’t line up with the center of luminous matter.
In this figure above of the Bullet Cluster, you can see the green outline that shows the gravitational centers based on lensing. On the left the outlines are superimposed over an optical image and on the right an X-ray image. It is clear that the gravitational center is pushed out from the luminous centers, going ahead of the baryonic matter as these two clusters collide. (Baryons are simply particles of ordinary matter made up of three quarks like protons and neutrons. We are made of Baryons.)
Intuitively, it looks like there is dark matter where the lensing is happening. This turns out to be deceptive because the most standard model of dark matter, the Λ-CDM, says the Bullet Cluster is highly unlikely.
Nevertheless, if your whole theory to get rid of dark matter depends on luminous matter producing gravity in strange ways, then it would appear that this implies that matter is a kind of ventriloquist, throwing its gravity away from itself into the void. Some naive researchers assumed that this was impossible and concluded that the Bullet cluster was direct evidence of dark matter. Some science writers swallowed this intuition at face value and still churn out faulty criticism based on it.
The death of MOND, however, turned out to be greatly exaggerated.
One counter (due to Milgrom, MOND’s originator) is that MOND only applies when accelerations are small, so we should see effects not uniformly where there is matter but only where accelerations are actually small, well away from the centers of luminous matter. This is what we see in rotating galaxies after all. Gravity only starts acting weird at the edges of the galaxy.
This is an intuitive but incorrect explanation. A more accurate explanation is that the difference between the MOND force and the Newtonian force actually includes not only a multiplicative scale factor based on acceleration but also an additive force field that depends on boundary conditions (basically the shape of the cluster). This force field disappears in most cases that have been considered because of symmetry (like in rotating galaxies).
If this force is zero, MOND cannot explain the Bullet Cluster.
Whether this solves the problem with the Bullet cluster completely is debatable. There are too many unknowns. Studies of the Bullet Cluster using an Einstein-Aether theory (Dai, Matsuo, and Starkman, 2008), however, suggest that there could be even more going on.
One theory goes that self-coupling vector fields in the early universe attracted matter which formed into galaxies. These created weak lensing centers that behave a lot like dark matter halos. Thus, it is the primordial vector field that formed the galaxies, not the one generated from the cluster itself, that is responsible for extra lensing.
This idea has become more important with the discovery just this year, based on data from the JWST, that bright galaxies formed much earlier in the universe than standard theories predicted. Indeed, two separate and now peer reviewed studies were the topic of this recent SciAm post which is, at best, a head scratching article about how nobody could have foreseen this. But some researchers did foresee it: proponents of MOND, TeVeS, and EA theory, none of which are mentioned, had been predicting since about 1998 (Sanders, 1998) that such modifications to gravity would lead to early formation of galaxies.
As Sanders said of MOND in the early universe nearly a quarter century ago:
When matter first dominates the energy density of the Universe, the cosmology
diverges from that of the standard model. Objects of galaxy mass are the first
virialized objects to form (by z=10) and larger structure develops rapidly
In other words, we expect bright galaxies to form in the early universe if MOND is operating. Subsequently, it was shown that with MOND alone, these structures grow too slowly, but TeVeS and EA theory both, in their generalized form, can grow them much faster.
You see, in the early universe when the Cosmic Microwave Background (CMB) was more than 3000 degrees Kelvin, hydrogen ionized. At this time, free electrons glued photons and baryons (matter) together. The photons would diffuse randomly through the mass of particles because of scattering, which occurs within the photon’s wavelength. This fluctuation path is damped exponentially, however, as the path length exceeds the wavelength of the photons. So we call this diffusion or Silk damping, and you can see it in the CMB angular power spectrum. The problem is that if you don’t have some dark matter, you’d expect the damping to be much worse than it is and the CMB to be far smoother and more damped than it is.
At the end of Silk damping, electrons bind to atomic nuclei and photons are able to propagate again. This is when larger scale structures would be able to form. But without dark matter, the damping should have prevented the formation of such structures for a long, long time, and even dark matter models can’t explain the JWST’s observations of bright galaxies at such an early phase of the universe.
Generalized TeVeS has been shown to help overcome diffusion damping in order to match the power spectrum that is attributed to the influence of non-baryonic dark matter. It also creates seeds that attract baryonic matter and form them into large scale structures very quickly after Silk damping ends. This is because the gravitational vector field developed perturbations, basically deviations from homogeneity, that attract baryons. These are like seeds that let galaxies grow.
The vector field perturbations stick around as the universe expands, expanding themselves and weakly interacting gravitationally. Now we see them as dark matter when in fact they are simply left over weak gravity from the early universe.
From this perspective, the vector fields simply become another dark matter candidate. When people do studies, however, trying to prove TeVeS or EA wrong, they don’t take this into account. Instead, they start by assuming that the background gravitational field is smooth and get the result that these theories cannot explain lensing, only dark matter can. But, if the background gravitational field isn’t smooth, then you can end up with almost any kind of lensing effect you want, just like with dark matter.
With the JWST now observing more dark matter than ever, this may be a prime time to further test EA and TeVeS and examine their ability to match lensing observations and form a coherent early universe cosmology that lends itself to galaxy formation, lensing, and intercluster collisions. Given the lack of consensus on what dark matter is, one more theory wouldn’t hurt.