I share the opinion that they're names for holes, not entities; but I'm not qualified to defend that opinion.
And, in order to understand a something, do we always need that something to interact with EM radiation?
Dark energy, in particular, can't be a kind of "stuff".
We could "understand" something that doesn't interact with EM; but we'd have trouble observing it.
If modified Newtonian (MOND) theories won out, the theory would explain this observation by saying that there was no matter there to begin with. If neutrino theories won out, the theory would explain this observation by saying that dark matter is electromagnetically neutral. But those are two entirely different theories, trying to explain the same observations, and trying to fill the hole in our currently accepted models in different ways.
EDIT: Per a child comment, you might mean "theory" as "idea", as in "the [idea] is that it doesn’t react electromagnetically with other matter." In which case, yes!
I usually agree with your understanding of dark matter/dark energy as placeholders for missing knowledge.
There's definitely lots of evidence for dark matter, yes~!
Dark matter has lots of reasons to be assumed but a common one is that when we try to explain the gravitational effects of galaxies based on a quantification of observable matter, there is a huge shortfall. That creates a space for something with a gravitational effect that is unobservable otherwise.
That many of the objects of our physics are theory laden is true, but that they are therefore “unreal” or merely an artefact of theory does not necessarily follow.
Light exiting from stars deep in a galaxy (which is in turn deep in a galaxy cluster) is subject to gravitational redshift. The greater the mass of the galaxy and cluster, the greater the redshift.
If we add transparent, non-radiating mass to a galaxy and its cluster, the mean redshift of starlight from the galaxy will redshift, and we'll see the result on various absorption and emission lines in the spectrum.
Note, we do not need dark matter for this; increasing the central mass of a galaxy cluster will do the trick too. We can do this by adding more stars, or by adding black holes, or by making a central black hole more massive. All of this will cause an increase in gravitational redshift.
However, adding more central mass to galaxies will be reflected in the trajectories of clouds of gas and collections of stars within the galaxies; likewise, adding more concentrated mass to a galaxy cluster will alter the trajectory of galaxies within it. In general, these objects will have faster orbits -- fast enough that a non-gravitational redshift will be apparent in the arrangement of spectral lines. (We would also need to figure out how to keep any central masses from being extremely bright anywhere in the electromagnetic spectrum, and how to keep them from eclipsing parts of their home galaxies and clusters that are behind them from our perspective.)
If we instead redistribute our increase in galactic mass into a gaseous halo-like structure that extends well beyond the outer stars, we get the increased gravitational redshift without requiring stars and gas to travel faster to avoid falling inwards. The halo still has to be electromagnetically quiet, and non-eclipsing, or we would see it. (This all applies to clusters, too).
Stepping away from the "dark matter puzzle" of the spectral lines of molecular and atomic gas clouds in galaxies and clusters of them, we can then look at another aspect of your question.
Galaxy clusters, from their point of view, are just floating passively in space.
The expansion of the universe follows a model wherein if we treat the astrophysical content of galaxies and clusters as sensitive accelerometers that in aggregate show off the peculiar accelerations experienced by these large scale objects, there is no evidence that these accelerometers should report a non-zero magnitude or point in any particular direction. That is, galaxies and galaxy clusters are themselves in free fall, and we can see this because of lines of evidence including expansion-direction-and-magnitude squashings of https://en.wikipedia.org/wiki/Light_echo lines turn out to be null, likewise there are no radio-loud edge shocks for galaxies that appear to be accelerating away from us, there are no shape distortions of the most highly "accelerated" distant galaxies, and so forth. There is simply a spectral redshift.
So, for an isolated galaxy the spectrum it emits depends on a combination of its starshine, any noise from close to its central black hole(s), the noise from any thick gas clouds undergoing gravitational collapse, and the absorption and reemission of diffuse gas clouds heated by these types of emission and absorption-and-reemission sources and the distribution of dark matter within and around the galaxy. The more dark matter, the more redshifted the galaxy's spectral lines will be.
Preserving the Copernican Principle, immersing a set of such isolated galaxies into an expanding cosmos leads to the inhabitants of each galaxy observing a cosmological redshift on the gravitatonally-redshifted spectrum emitted by other sufficiently distant galaxies. The gravitational-redshift depends on the mass of the galaxies including their dark matter. The cosmological redshift only depends on a light-travel distance from the galaxy, no matter how light or how massive that galaxy (or its parent cluster) is.
Next, if we have an arrangement b-->F-->U, where b is a background galaxy, F is a foreground galaxy, and U is an observer, we will that the light from b is cosmologically redshifted before it interacts with the molecular and atomic gas clouds in F, causing a spreading of absorption and reemission lines rather than a thickening. For the very obvious lines from neutral atomic hydrogen, we see this as the https://en.wikipedia.org/wiki/Lyman-alpha_forest which is an important tool in astrophysics and in checking the soundness of our theory of gravitation.
We would get a Lya-forest structure even in the absence of dark matter in b or F or anywhere else; only the fine strucutre would differ. Indeed, we could substitute a diffuse static cloud of atomic hydrogen for F and still see a "forest" structure minus small contributions from the starshine in F and the orbital motions of the gas in F around some central mass. The deep infrared noise from F in that case, however, would not be gravitationally redshifted.
Finally the answer to your question: the cosmological redshift does not in general depend on the "late" distribution of dark matter. (Late as in at least millions of years after the hot big bang). But the gravitational redshift on light emitted by stars and galaxies does depend on dark matter and in general the starshine is reddened by it. The reddening is because of the additional mass of the dark-matter-containing structures, and also because the distribution of the dark matter into halos allows for slower orbits of stars and gas clouds, thus suppressing their non-gravitational non-cosmological redshifts. (In an edge-on spiral galaxy without dark matter we would see a greater difference in the blueshift of one extreme tip of the thin disk and the redshift of the opposite extreme tip of the thin disk, compared to the same spiral galaxy with dark matter; the light-and-radio from the central bulge would be less redshifted in the absence of dark matter).
So, no, the cosmological redshift does not depend on dark matter. The gravitational redshift on light from galaxies, however, does. So does the non-gravitational, non-cosmological redshift on gas clouds whirling around the edges of rotating (e.g. spiral) galaxies, or rising and sinking on radial orbits within elliptical galaxies.
In order to make that measurement, the authors had to combine redshift measurements for ~ 20,000 galaxies from ~ 2,500 different clusters; as they note, it's impossible to detect the effect for just a single cluster: "... even assuming that every galaxy in a cluster could be spectroscopically measured, there are simply too few galaxies to allow the statistical detection of a non-zero gravitational redshift."
So while it's technically true that adding dark matter to a galaxy would increase its gravitational redshift and thus its total observed redshift, this effect is immeasurably small.
Mpetha et al. (2021), "Gravitational redshifting of galaxies in the SPIDERS cluster catalogue" https://arxiv.org/abs/2102.11156
Broadhurst & Scannapieco (2000) https://iopscience.iop.org/article/10.1086/312630 is also relevant here, as it distinguishes between X-Ray studies and optical ones (SPIDERS being an example of the latter) pointedly at §4.2.
I do not disagree that ~ 10-20 km/s is small, but it's the little things that cause "tension" headaches, isn't it?
The Broadhurst & Scannapieco discussion about X-ray studies is interesting and potentially promising, but no one seems to have tried doing for real.
This is the exact opposite of the truth. An edge-on spiral galaxy without dark matter would have less extreme observed rotation velocities (comparing the blueshifted and redshifted sides of the disk), because only the visible mass (stars and gas) is relevant. But in practice we see higher rotation velocities (greater difference between blueshifted and redshifted sides) than can be explained by just the visible mass, and so we have to postulate extra mass in the form of "dark matter" in the galaxy.
(Dark matter halos are not just dark matter outside the visible galaxy, they extend all the way to the centers of galaxies, and in fact become denser the closer to the galaxy center you get. Of course, the visible matter also becomes denser the closer to the center you get, and usually faster than the dark matter does, so the central regions of massive galaxies are dominated by the visible matter. But there's still dark matter there, and in some dwarf galaxies, the dark matter even in the central regions.)
As to your parenthetical, cuspiness is not really relevant to the scale of the points raised in the bulk of my comment, and it is not clear to me why you raise the DM density profile here.
On the one hand, in vacuum, an accelerating expanding universe's geometry breaks time-translation invariance. Let's look at the vacuum geometry with a pair of test photons in the otherwise empty test universe. In the early universe we have the photons widely spacelike-separated but on trajectories where they will ultimately collide in the far future. The photons in the future will be redshifted compared to their wavelengths in the past, with the redshift proportional to the expansion history of the test universe. (In our universe, the cosmic microwave background photons have behaved this way when two CMB photons meet at a detector).
Since photon energy can be calculated E = (hc)/\lambda, where \lambda is the photon wavelength, a redshift corresponds to a decrease in the photon energy.
Where did this energy go? One can simply say that constancy of photon energy is premised on time-translation invariance, which is not a large-scale feature of an accelerating expanding universe.
This position is just that thermodynamics was developed to describe engineering phenomena in the 19th century, and not to describe the motion of clusters of galaxies over the course of billions of years. Since Noether, thermodynamics's first law has its foundations in time-translation invariance, which we only have in some spacetimes (e.g., Minkowski's flat space, the spacetime of special relativity) or in local pieces of general curved spacetime. In the absence of strict time-translation invariance globally, the first law of thermodynamics is just a weak-gravitational-field approximation.
Sean Carroll expands upon some of this here : https://www.preposterousuniverse.com/blog/2010/02/22/energy-... wherein he also discusses an effectively pseudogravitational-field view of our test spacetime's geometry, in order to think of our pair of converging photons donating energy to that (pseudo-)gravitational field. As he says, such a view, which leans on a notion of gravitational potential energy which is simply absent in General Relativity, is often unhelpful.
On the other hand, one can take a different approach and treat \Lambda, the cosmological constant, as a matter field rather than as a geometrical feature of the accelerating expanding spacetime. When we do this, then we are left with either an inertially-expanding spacetime, or with a flat spacetime, and in the stress-energy tensor we have what makes that acceleration or acceleration + expansion field cause everything in the stress-energy tensor (including those fields themselves) to fly apart.
It can be reasonable enough to do this. Einstein and Schrödinger discussed this in correspondence during the first world war : https://arxiv.org/abs/1211.6338 wherein they seem to agree that it is a matter of taste. We know from work in the more modern Hamiltonian formulation of General Relativity that we can't do this generally, but there are large families of spacetimes and matter content that are amenable to this approach of encoding aspects of the dynamical geometry as a form matter we can describe with canonical position & momentum.
We can also note here that the distribution of matter in a spacetime can pick out preferred foliations or even preferred frames of reference.
In our standard cosmology, which you are asking about, we treat the distribution of matter isotropic and homogeneous, where each form of matter is a perfect fluid with a density and a pressure. As the universe expands (with or without acceleration of expansion) the matter and radiation fluids dilute away, their density dropping over time. We encode the acceleration of expansion as an additional fluid which simply does not dilute away: it retains a constant density even as space expands or as the expansion accelerates, and is in effect the energy of the vacuum. "Matter tells spacetime how to curve" means that the matter fields, important when they are dense in the early universe, become less important in the later universe. In the later universe, the additional expansion/acceleration field dominates, causing the spatial distances between inertial observes (we can talk about spatial distances in the preferred frame of reference our isotropic & homogeneous fluids pick out) to widen.
In this view, the expansion of the universe is an adiabatic process on the matter and radiation fields, but the additional field -- a vacuum energy -- retains a constant energy-density.
We can also consider the pressure of these spacetime-filling fluids. All of the (diluting) matter has a positive pressure. If you create a locally-overdense bubble of positive pressure fluid, it will tend to become denser as it collapses gravitationally. Local overdensities of matter (including dark matter, which also has positive pressure, and also dilutes away with expansion) in the early universe are responsible for structure formation, wherein gasses collapse forming the first stars. The (nondiluting) dark energy field by contrast has a negative pressure. Local overdensities of that field tend to anti-collapse gravitationally, becoming less dense. This is peculiar, but you can think of what you see in boiling water, where the readily observed water vapour voids are less dense than the water. Voids from overdensities of dark energy in the early universe may also be important in structure formation, like underdensities of matter in the simpler model in the couple paragraphs above.
This view gets messy when treated formally: you do not need to generate more positive-pressure matter into a collapsing overdensity, it just gathers up into a denser and denser arrangement. (We ignore details of black holes). However, it arguably has a virtue that it at least partially (depends on details of the field dynamics and a time-dependent equation-of-state parameter; for example we may need a decay term where "overdense" dark energy decays into normally-dense dark energy) recovers from the breaking of time-translation symmetry on a cosmological scale. Some of the mess is to keep a "fifth" fundamental force from appearing and causing long-range effects which we simply do not observe.
Carroll (again) discusses some of this here : https://www.preposterousuniverse.com/blog/2013/11/16/why-doe...
Forgive any errors in sign and unclarities. My system has decided to become uncooperative during the editing pass and I'm just hitting reply so the text above doesn't vanish.
We observe with telescopes that the universe is expanding. We don't know why, so we call the mysterious force dark energy. It's just a name for something that we are objectively observing.
(Also this has nothing to do with dark matter, that unrelated to this discussion and you're just getting the names conflated.)
Before criticizing physicists you should keep this in mind.
Not really. It is a hypothesized form of matter. There are competing hypotheses. Dark matter came to popularity in order to explain observations of galaxy rotation curves in the late 1930s. But these galaxies were observed in isolation, giving the illusion of gravitational isolation, which doesn't exist. Once accounting for the tidal effects of nearby galaxies, [1] dark matter becomes unnecessary to explain galaxy rotation curves, all that is needed is Newtonian physics. But in the period since dark matter was hypothesized, other observations that could not be readily explained were lumped into dark matter. Dark matter is nothing if not an amalgamation of unexplainable observations. So while dark matter is unnecessary to explain galaxy rotation curves, it is still somehow useful to explain unrelated observations.
> it isn't a band-aide fix
It really sort of seems like that: can't explain an observation, so invent something out of thin air to explain it. Honestly, an hypothesis that elves are causing the observations has just as much validity (or lack thereof). Now, elves have not been proven to be the cause of these unexplained observations, but neither has dark matter.
> the foundations of the concept of dark matter are fairly reasonable.
See above concerning galaxy rotation curves and endnote. Please understand it is quite possible that dark matter does not exist. The reason why it seems like accepted science has to do with scientific paradigms. Once a notion becomes an accepted part of the science paradigm, it takes years if not decades after disproving it beyond all doubt to dislodge it from the paradigm. If science has a problem, it is paradigms.
"Tidal effects" are a non-starter, and do nothing to explain galaxy rotation curves. Nor do they explain the excess motions of individual galaxies in galaxy groups and clusters, or the excess pressure of hot, X-ray emitting gas in groups and clusters -- all of which requires either some kind of dark matter, or some kind of alternate theory of gravity, or both.
(I notice that I pointed out the ignorant nonsense of that video when you posted a link to it 7 months ago -- https://news.ycombinator.com/item?id=27085979#27090683)
Nice catch.
> "Tidal effects" are a non-starter, and do nothing to explain galaxy rotation curves. Nor do they explain the excess motions of individual galaxies in galaxy groups and clusters, or the excess pressure of hot, X-ray emitting gas in groups and clusters -- all of which requires either some kind of dark matter, or some kind of alternate theory of gravity, or both.
What a well formed and well supported argument, and not just a matter of fact statement that you've left entirely unsupported. I'm totally convinced!