Suppose we made an observation that mercury's orbit precesses by a fraction of a degree every orbit. Heretofore every observation of perturbances has correctly presaged the presence of a large mass (a planet) that was quickly found, but this time we have looked and looked but not found anything. Should we presume that it must be an invisible ball of matter that gravitates but we have no other way of detecting except through the same effect we use to infer its existence, or should we tweak Newton's laws of physics?
> You can postulate an additional separate modification of gravity that comes into play on cluster scales.
Or it might be something else altogether.
Anyways, MOND has made predictions that people set out to disprove MOND went looking for negative evidence and had to turn around (e.g. external field effect), as well as straightforward predictions JWT may have confirmed (early galaxies) and explains many things that dark matter cannot, like how satellite galaxies don't get ripped apart by the dark matter halo.
But you can't tweak Newton's laws, because we know Newton's laws aren't the right laws of gravity. The laws of General Relativity make better predictions. So what you would have to tweak would be the laws of GR.
And the basic problem with doing that is that you can't just tweak a single parameter in GR to get the kind of modified behavior that MOND does with Newtonian dynamics. In GR, there simply isn't any such free parameter to tweak. You have to adjust the whole theory. And any such adjustment ends up looking just like dark matter: you have to invent new fields or new forms of stress-energy that we have no other evidence for. And that destroys the whole point of MOND, which was to just modify the laws of gravity without changing our assumptions about what matter and energy is present in the universe. If you're going to have to add new forms of matter and energy anyway, why not just stick to that and keep the laws of gravity the same? That's what the dark matter hypothesis amounts to, and why it's actually the simpler hypothesis once GR is taken into account.
> But you can't tweak Newton's laws, because we know Newton's laws aren't the right laws of gravity. The laws of General Relativity make better predictions. So what you would have to tweak would be the laws of GR
The invisible planet OP is talking about is Vulkan[0]. We modified Newtonian gravity to explain the orbit of Mercury (instead of discovering Vulkan), the modification is GR.
OP is implying that maybe we have to modify GR like we did Newtonian physics.
I would love to stand corrected, but as far as I understand it, NP->GR modification also instantly explained many same-day and subsequential phenomena, and mostly without issues. GR->? modification still faces some obvious today’s issues like galaxies “almost without dark matter” and galaxies/regions “almost fully consisting of it”.
It doesn’t mean that GR is correct. But GR is a formula that perfectly matches a dataset of local observations. ? should at least match farther ones, but they clump randomly like some non-interacting blobs/streams and do not form a general rule. This new dataset is either still too small or has no clear formula describing it at all. That’s why ? is still ?
> GR is a formula that perfectly matches a dataset of local observations.
More precisely, our best current model of the solar system, which uses GR as its law of gravity, matches our best current dataset of solar system observations.
But similarly, our best current model of the universe, which also uses GR as its law of gravity but is a different model, matches our best current dataset of universe observations--which includes galaxy rotation curves but also includes many other observations.
> they clump randomly like some non-interacting blobs/streams and do not form a general rule. This new dataset is either still too small or has no clear formula describing it at all
I'm not sure what you mean here. Our dataset of universe observations is described by a "clear formula"--our best current model of the universe, which I described above. That model includes "dark matter", i.e., matter that we cannot see, but whose presence is necessary to account for the motion of things that we can see if we use GR as our law of gravity.
MOND proponents criticize the above model because it includes matter that we can't see. But that's not the same as saying the above model doesn't account for observations. It does.
Yeah, I meant that if we “turn a blind eye” on dark matter in GR, then no universal formula can cover it. In the same sense that universal jet aerodynamics formulas don’t account for trees and mountains.
> I meant that if we “turn a blind eye” on dark matter in GR, then no universal formula can cover it. In the same sense that universal jet aerodynamics formulas don’t account for trees and mountains.
I'm still not sure what you mean here. The "universal formula" in GR is the Einstein Field Equation, which covers everything (more precisely, everything in which quantum effects are negligible).
Sometime you have to modify the theory, but sometime the missing mass is really there. For example neutrinos were postulated to exist to explain missing quantities in some interactions. Then they were actually discovered experimentally .
Incidentally neutrinos are dark matter, although not of the right kind.
Slight clarification of that last sentence: neutrinos are "dark" (do not interact with electromagnetism) and "matter" (have mass, take up space), but they are not what we are looking for when we say "dark matter" (the current best explanation for why galaxies move as they do).
In particular, they are weakly-interacting massive particles, but they aren't the Weakly-Interacting Massive Particles (WIMPS) we're looking for. WIMPs have to be slow-moving, or they'd escape the galaxy, and there's no apparent way to make slow-moving neutrinos (almost no mass means very high speed), and we haven't observed any.
It's still not absolutely impossible, but it's very unlikely. Dark matter is more likely to be something else -- though we're still unclear on what, and the most likely theories are looking somewhat less likely recently.
Adiabatic cooling of relic neutrinos (the cosmic neutrino background, CvB).
CvB formed before the cosmic microwave background and so is cooler than the CMB. Massless bosons like CMB photons have their wavelengths stretched through the history of the metric expansion of space; massive (even if the masses are small) fermions (like CvB neutrinos) instead see their speeds drop. The drop is about 161(1+z)/m with m being the neutrino mass and z the redshift; at present times CvB neutrinos are moving nonrelativistically (a couple thousand km/s) and so are certainly cold dark matter.
CvB neutrinos are incredibly abundant and do form a small fraction of CDM in the standard cosmology.
I don't think it's fair to call them WIMPs, though -- one wants invariant masses in the GeV range to gather up sufficient energy-density to drive galaxy cluster dynamics (cf. <https://en.wikipedia.org/wiki/Light_dark_matter>). CvB neutrinos are several orders of magnitude too light (invariant masses of 50-100 meV, kinetic masses of about 250 µeV).
Right, that's why I referred to them as the wrong kind. IANAP, but I believe that neutrinos are in fact sometimes referred as hot dark matter.
Still, I understand that the current best model to explain the (tiny) neutrino mass involves theoretical much more massive "sterile" neutrinos that potentially could be the right kind of dark matter if they indeed exist.
It's ironic that Mercury is the chosen example of a discrepancy in an existing theory requiring a revision, because as I understand it MOND being based on Newtonian gravitation can't explain Mercury's actual orbit either. So it purports to solve one problem (galaxy scale gravitation) while re-introducing one we already solved. This makes it a bit weird that the example of Mercury is so often used in arguments for MOND.
>If you want to merge MOND with Einstein’s General Relativity, it is possible as well, simply by adding in scalar (and possibly vector) terms in addition to the standard metric tensor terms.
There's no issue as long as you think that "adding in scalar (and possibly vector) terms" is somehow a "simpler" change than "adding a new kind of matter".
But when you look at what "adding in scalar (and possibly vector) terms" actually means in GR, it's the same thing as "adding a new kind of matter". You're just calling the new kind of matter a "scalar field" (and possibly also a "vector field", i.e., two new kinds of matter).
In other words, when you take relativity into account, "MOND" is not an alternative to "dark matter"; it's just one particular way of adding "dark matter".
My understanding is that MOND doesn't kick-in in the Solar System at all; the strength of gravitation from the Sun is much greater than the MOND threshold. So I'm surprised that Mercury ever comes into discussions of MOND.
I thought grandparent was saying that GR is wrong, and that Vulcan exists, but it’s made of dark matter. But maybe that’s not the right interpretation ;)
> I thought grandparent was saying that GR is wrong
No. The dark matter hypothesis is part of our best current model based on GR.
That hypothesis does include the belief that, as far as galaxy rotation curves are concerned, GR corrections to Newtonian gravity are negligible. But that is not the case for other observations that dark matter also accounts for (see below).
> and that Vulcan exists, but it’s made of dark matter.
That is basically what the dark matter hypothesis says, yes: that we don't need to modify the laws of gravity to explain galaxy rotation curves, we just have matter there that we can't see.
If this hypothesis were only introduced to explain galaxy rotation curves, that would be one thing. (It still wouldn't be as easily refutable as the Vulcan hypothesis was, since it's a lot easier for us to rule out the presence of significant additional matter in the solar system than it is in distant galaxies. But that's a side issue.) But it isn't. The dark matter hypothesis also explains many other observations (for example, the amount of structure--gravitational clumping of matter into galaxies and galaxy clusters--as compared with the amount of baryonic matter present based on Big Bang nucleosynthesis (which by itself would be much too small to account for the structure that we see). MOND, OTOH, was introduced solely to explain galaxy rotation curves and still leaves all the other observations that dark matter accounts for unexplained.
> Should we presume that it must be an invisible ball of matter that gravitates but we have no other way of detecting except through the same effect we use to infer its existence, or should we tweak Newton's laws of physics?
"we have no other way of detecting except through the same effect we use to infer its existence" is true of everything in the universe. We believe stars exist because we can see the light coming from them, but light is our only way of detecting them.
There's no single answer to this kind of question, we have to actually do science and make observations. We should ask whether it behaves like a ball of matter (e.g. does it have a point location in space? Does it orbit the way a ball of batter does?) or like a modified law of gravity (can we reproduce the same conditions)?
The likes of the Bullet Cluster makes a very compelling case that cold dark matter is real.
> "We have no other way of detecting except through the same effect we use to infer its existence" is true of everything in the universe
No. For example, staying within astronomy, we detect gravitational waves and if we find an optical event that correlates with the gravitational waves, it is multimodal (we are at n=1), and if you don't solicit events (or if you statistically blind it by adding fake events in your solicitation) it's multimodal independent (currently n=0). Going outside of astro, you'll find very rich multimodal studies "I know what this molecule is because I have an IR, UV, NMR, Mass spec, etc etc etc"
Light, in the GP’s example is also traceable back to a source. Observing an effect in a 3rd party isn’t directly attributable to anything without a plausible mechanism of action. Using light to detect stars as defense of dark matter isn’t a very strong argument.
> Heretofore every observation of perturbances has correctly presaged the presence of a large mass (a planet) that was quickly found, but this time we have looked and looked but not found anything. Should we presume that it must be an invisible ball of matter that gravitates but we have no other way of detecting except through the same effect we use to infer its existence, or should we tweak Newton's laws of physics?
Note the leap of faith from "a large planet" to "an invisible ball of matter." As one could imagine, given what we knew in the 19th century, "a small, hard-to-observe planet (or a group of small bodies) orbiting very close to the sun" was a totally reasonable hypothesis. "An invisible ball of matter" wasn't (and still isn't). It was natural that physicists never seriously considered such a possibility.
On the other hand, given what we know now, "a diffuse cloud made of particles that do not interact with light" is a totally reasonable hypothesis. In fact we even know one such "invisible" particle: neutrinos. It's not crazy to assume that there are some other kinds of neutrino-like particles that fit the bill.
> On the other hand, given what we know now, "a diffuse cloud made of particles that do not interact with light" is a totally reasonable hypothesis.
Was a reasonable hypothesis. Given the evidence so far, it's now very debatable. Just like another planet was a reasonable initial hypothesis until it wasn't.
I mean a primordial black hole with the size of a tennis ball would have 5 to 15 earth masses. It would not be invisible, but at that size it would be invisible to us even if it was in our solar vicinity.
If primordial black holes are things that exist, one could reason that they might be an explaination for masses we cannot see maybe?
It's an interesting idea. I suspect such a large population of these objects would generate enough x-rays from interactions with dust and gas that we'd be able to detect their presence. We'd also probably see micro-lensing of various observable phenomena if they were that populous. Just guesses though.
Indeed black holes and other dark stars are a reasonable option for dark matter [1], but apparently observations do not seem to match this model adequately.
[1] they are collectively called MACHOs, as opposed to WIMPs...
The explanatory power of something that has as many free (statistical) parameters as lcdm is so high it's stunning how many things it doesn't explain. You'll surely get some things in there by chance.
It doesn't explain all of cosmology without tuning the distributions of dark matter to reproduce the observations that MOND describes without any parameters aside from visible mass.
There is only one parameter that tunes the distribution of dark matter, for the entire universe: the spectral index.
The spectral index sets the statistics of the overall density distribution in the very early universe. After that, gravity and other physical laws determine the evolution of the universe.
Then this is inconsistent with the recent discovery of galaxies without DM, among other issues. I think you're overstating the consistency of the evidence for DM. I replied with a reference to another of one of your comments.
Dark matter is statistically distributed, the same as baryonic matter. Galaxies with no dark matter or "too much" dark matter is as consistent as galaxies having different baryonic masses or things like https://en.wikipedia.org/wiki/CfA2_Great_Wall
The number of free variables has gone down dramatically due to extensive theoretical and experimental work. And as already mentioned, there are many very different lines of evidence and it's not like different variables are being to fit each one in isolation.
> different variables are being to fit each one in isolation.
For the big picture, phenomena that get "explained" by dark matter are absolutely being fit by different individual variables.
For example, the ratio of dark matter to baryonic matter is different between clusters and individual galaxies by an order of magnitude. I recall the DMD for CMB is also different but I can't substantiate that memory quickly
In the small, too, there are many galaxies where we have no idea what dynamical situation led them to have the dark matter content they exhibit.
> The number of free variables has gone down dramatically due to extensive theoretical and experimental work.
The number of free variables in MOND is one, visible mass, and it's surprising the number of observations it matches despite that. There is almost certainly something there.
The superfluid dark matter idea that incorporated both was a promising unifying idea, but last I checked the proposed models didn't match some observations.
> Should we presume that it must be an invisible ball of matter that gravitates but we have no other way of detecting except through the same effect we use to infer its existence, or should we tweak Newton's laws of physics?
Are these substantially different? Of course using “must” isn’t appropriate, but spouting off the best theory you have at the time seems okay to me. Hopefully no one is saying that we cannot possibly ever find another better theory the same way Einstein’s was better than Newton’s or that we shouldn’t ever look for one.
> Hopefully no one is saying that we cannot possibly ever
The OP article might as well. The thrust of the article's argument is, "heaven forbid we add third parameter to MOND" meanwhile, draco, fornax, and Andromeda have wildly different dark matter densities so that's three parameters right there.
> draco, fornax, and Andromeda have wildly different dark matter densities so that's three parameters right there.
No, it isn't. The distribution of stress-energy is a single free parameter in GR; there is no law in the theory that says the density of any particular kind of stress-energy has to be the same in different galaxies. So adjusting the distribution of stress-energy to match the observed behavior in different galaxies does not require adding a new parameter for each galaxy; it only requires adjusting the single parameter that already exists, in different regions of spacetime.
Ok I concede. It is a single parameter that contains a millions of subparameters, at least one for each galaxy. Against three parameters, each of which is a single number. But it is three parameters!
That's 3x more parameters! Moreover, we had a problem and we had to increase the number of parameters by 50%. So yeah One parameter is much, much better, especially since we have never changed that it's one parameter.
> It is a single parameter that contains a millions of subparameters, at least one for each galaxy.
No, it's a single parameter with a distribution in space and time. It's the same one we already have to vary in space and time anyway based on our observations of light. Dark matter just means adjusting the same distribution further based on observed motions even if we can't see the source of gravity that is producing the motion.
If MOND actually gave the choice of adjusting a single parameter in place of having to adjust the distribution of matter and energy, that would be one thing. But, as I noted in another comment upthread, once you take into account that the theory of gravity we are using is actually General Relativity, not Newtonian gravity, there is no such single parameter anyway. Relativistic versions of MOND actually end up looking just like dark matter: they have to invent new fields, new forms of energy, for which we have no other evidence. There is no way to get the desired effect in GR by just tweaking a parameter in the law of gravity.
This is equally true for baryonic matter. You might counter that baryonic matter is easier to observe/constrain. I would point out that LCDM also has ways to observe and constrain DM, and the universe doesn't owe us perfect legibility. What is it you think you're proving?
What doesn't add up is that MOND can reproduce many of those observations using only the visible mass and a tweak to gravitation. This makes all of these tuned distributions seem very suspicious by comparison. It would be a remarkable coincidence if there wasn't something to MOND.
> MOND can reproduce many of those observations using only the visible mass and a tweak to gravitation
No, it can't. It can only reproduce a subset of galaxy rotation curve observations. It cannot reproduce other observations at all (such as the amount of structure in the universe as compared with the amount of baryonic matter that is consistent with Big Bang nucleosynthesis, which is about the same as the amount of matter we can see, but is much too small to account for the amount of structure--gravitational clumping into galaxies and galaxy clusters--that we see).
> it's not nearly as rosy for DM or dire for MOND as you imply
According to that one paper, perhaps. There are lots of papers published in this area of research. If you want to say the jury is still out on a final resolution, yes, I'll agree with that.
My guess would be something about the dynamics of dark matter in galaxy evolution pushes the distribution into a shape that produces those mathematically simple rotation curves. That seems Occam's-razor compliant to me; it would be weird if rotation curves with dark matter were really complicated and unpredictable. It's also possible that both MOND and LCDM have some truth to them. But usually people are trying to get rid of dark matter, and that's just not happening.
> My guess would be something about the dynamics of dark matter in galaxy evolution pushes the distribution into a shape that produces those mathematically simple rotation curves. That seems Occam's-razor compliant to me;
It's not just rotation curves, there are numerous validated predictions MOND makes based only on the visible matter (see the paper at the bottom). These would all be a set of remarkable coincidences if the current CDM model were true.
> it would be weird if rotation curves with dark matter were really complicated and unpredictable. It's also possible that both MOND and LCDM have some truth to them.
Yes, this was the idea behind "superfluid dark matter", but it doesn't quite match observations either: https://arxiv.org/abs/2201.07282
Maybe another take on this fusion will stick.
> But usually people are trying to get rid of dark matter, and that's just not happening.
I'm not sure "usually" is an accurate summary. It's more like they're claiming that the standard CDM model is refuted or should be very disfavoured, which is not so unreasonable IMO:
IIRC, some MOND proponents suggest that neutrinos can satisfy the dark matter observations that MOND itself doesn't explain. I believe that's discussed in the above paper but it's been awhile since I've read it.
The bulk of the whining I see here about DM is on the lines of "but it's not directly observable, so it can't be real science! Can't you see this is just Big Physics trying to fool you?" These are the ones trying to falsify DM entirely, and having to ignore huge swathes of evidence to do so.
Yes, the paper suggests sterile neutrinos at cluster scale, with MOND at galaxy scale. I'm fine with this, honestly.
The distribution of stress-energy is a single parameter per point in space. That is, I don't think saying "it's only a single parameter" and "it's different in different galaxies" are compatible.
If you want to say it's a "parameter distribution" or something like that, that's fine. It doesn't change the fact that this thingie, whatever you want to call it, is already present in our existing theory of gravity, GR, and is already adjustable in that theory. So we're not adding anything new to the theory if we adjust this thingie to take into account observed motions for which we can't see the gravity source that causes the motion.
I don't think you understand the concept of explanatory power. Because dark matter cannot be observed we must manually tune the dark matter distribution for each galaxy. This creates new parameters that don't exist in the theory that are necessary to properly explain phenomena. Whether or not they exist in the theory is irrelevant to the fact that they are necessary.
Lambda CDM assumes that the initial density distribution is a Gaussian random field with a spectral index close to 1. There is only one free parameter setting the distribution of matter: the deviation of the spectral index from 1.
The detailed distribution of matter that we view in the present-day universe is a consequence of gravity (and other known physical laws) acting on that initial, random (but following a very simple statistical distribution) density distribution.
This is not true. There is not enough explanatory power in LCDM to explain the distributions of dark matter we apply to each specific galaxy, therefore there must be a proliferation of parameters.
I recommend you read about Gaussian random fields in cosmology. Cosmology treats the initial conditions as random, but with a distribution that is describable by only a few parameters. You're confusing the exact realization of the random field with model parameters.
Replying late, because I was on vacation, but just in case anyone reads this:
There is a sense in which what you are saying is true. And yet, when people say "The dark matter distribution in the Bullet Cluster is <some non-trivial distribution>, they are not deriving that from the "few parameters" that describe the cosmological distribution. They're deriving it from what they see of the local situation in the Bullet Cluster. And then they say "dark matter explains what we see in the Bullet Cluster". That's what people are referring to when they say "a tunable parameter at each point in space".
Without that, you have something that explains cosmology, but not the Bullet Cluster. With that, you have something that explains both cosmology and the Bullet Cluster (and a number of other odd things out there), but you no longer have "only a few parameters".
I guess the other side effects might be different right?
E.g. we talk about relativistic mass. But when they did experiments it was found that no other attributes that come with increased mass was found. So the answer is right but there is something off with our understanding.
Notably, Mercury was not a motivation for the development of relativity. Even with the benefit of hindsight, I don't recall ever seeing anyone critisizing contemporary astronomers for looking for the missing planet. GR was developed based on theoretical concerns. The fact that it happended to solve an open issue of Mercury's orbit was a nice confirmation that it is on the right track.
The good news is that we have a lot of physicists looking for an alternative theory of gravity called quantum gravity. It would be great if they come up with something that solves the theoretical problems with GR and happens to resolve dark matter.
> Should we presume that it must be an invisible ball of matter that gravitates but we have no other way of detecting except through the same effect we use to infer its existence, or should we tweak Newton's laws of physics?
I mean this is literally how they discovered Neptune? Dark matter isn’t necessarily some magically invisible particle, we’re just trying to count up all the different types of objects in the universe.
I feel like the author is assuming here that the reader knows this is how Neptune was discovered and that the procession of mercury caused by relativity was long thought to be caused by a hidden planet. A very literal example of something thought to be missing matter that instead was a modified Newtonian theory.
Yes, Vulcan was looked for just like DM has been looked for. The evidence for DM is largely indirect, just like the evidence for Vulcan. The OP's comment is a good summary of where we stand now. I suggest comparing all of the evidence for DM and MOND:
There is definitely something to MOND. It would be a remarkable coincidence if MOND was able to make accurate predictions with only a single parameter, visible mass, what takes DM multiple parameters and a tuned mass distribution specific to each observable galaxy. DM definitely "explains more" in some sense because of how much more work it's received, but that's not saying much.
Sure, I was expecting that when I started reading the comment, but then it pretty much fell apart.
Even if the comment was clear, it still isn't a very good comparison: GR has relatively easy to verify things like the motion of Mercury, and the factor of 2 when light is bent by the Sun. MOND, not so much.
BTW I took a graduate class from one of the people who named WIMPs. It took considerable effort to overcome the objection of ApJ's editor to the cute name. Fun times.
The main argument of this article seems to fall into the counterfactual fallacy category. Even if MOND has problems and ends up being wrong, that doesn't make dark matter exist.
Another great article from Ethan Siegel. I think he makes a pretty strong case for cold dark matter, but it relies pretty heavily on theoretical cosmological arguments. For the pro-mond argument, there’s a recent interview with Milgrom on Dr Brian Keating’s physics podcast Into the Impossible [0]. There is also a talk about all the ways DM does not fit observations from Pavel Kroupa (University of Bonn) on youtube [1]. In particular, Kroupa argues that when galaxies pass through one another, DM should cause them to move slower after the collision, because DM would be drawn behind each system in a turbulent wake; but no such slowdown is observed. MOND plus visible matter predicts the motion of these systems well.
Turbulence entirely due to gravitational interaction? Does that work? I guess I should watch it, but I can't take it seriously with the title "On the non-existence of dark matter" when there's so much evidence for dark matter. That's clearly unwarranted confidence in search of attention.
That said, if true, I think it more likely indicates that both MOND and LCDM have significant truth. The bar for a theory is "as simple as possible, but no simpler." If the evidence truly demands both MOND and DM, we just have to suck it up.
The DM is moved around by the gravity of the moving galaxy. The DM theory is that each galaxy is surrounded by a huge spherical cloud of DM that (by definition) only experiences gravitation and is invisible and insubstantial in every other way. When Voyager gained speed from passing near Jupiter, it caused Jupiter to slow down by a tiny amount. Similarly, when galaxies pass through each other, the DM should take a measurable amount of momentum from the galaxies.
And, yeah, it's a provocative title, but he believes that his measurements falsify DM theory. It's partly a tussle between experimentalists and theoreticians. Kroupa is arguing that the claims of the theory can be tested, and that it doesn't match observations. The theoreticians have a nice story for the big bang, inflation, baryonic nucleosynthesis and the variations seen in the CMB, and the amount of DM it predicts happens to match precisely the amount of DM that we observe.
First: if theory and observation don't match up, that can be due to lacking theory or lacking observations. Both should be investigated.
Second: all candidates for dark matter i know posit an enormous number of places where this dark matter is.
It is not that "dark matter solves this problem". It is a particular distributing of mass that solves the problem - typically, a highly specific distribution of mass. So not only is this theory supposing that most material in the universe is invisible to us, but its properties other than mass are unknown to us except it allows us to fill in the blanks exactly as needed.
E.g, we're not expecting a lot of dark matter black holes, as the hypothesised stuff doesn't even interact with itself and so cannot slow down. Basically, "dark matter" is fairy dust that physicists sprinkle over everything till it works.
And you know what: maybe that is the way to go. Probably it plays at least some part in answering these questions. But until we have a clearer picture, we may as well call it "magical fairy dust". That at least reminds us how much we're fudging.
Until it is experimentally tested, DM is essentially equivalent to the ether. A useful unexplained tool which to date seems to work.
Until another Michelson-Morley experiment comes (and MOND might have something to say, or at least they seem to be testing DM experimentally), yes: as untestable as string theory is.
> It is a particular distributing of mass that solves the problem - typically, a highly specific distribution of mass
It's actually an extremely simple initial distribution of dark matter in the early universe. You can specify the entire distribution with one number: a spectral index. The thing is, we already know that the universe had that distribution of energy density anyways, because of the Cosmic Microwave Background.
If you take that very simple distribution and let the laws of physics run for 13 billion years, you get the required present-day distribution of dark matter. That distribution can be verified in several ways, including gravitational lensing, the rotation curves of galaxies, the motions of galaxies inside clusters, and the temperatures of hot gas inside galaxy clusters.
The fact that such a simple set of initial conditions can explain pretty much everything we can see on scales ranging from stars to the entire visible universe is why cold dark matter is overwhelmingly the favored paradigm.
Yes but dark matter is explanatory while MOND is predictive.
We see a set of problems, we fix them by choosing our unobservable dark matter halo in order to fix those problems. That should automatically be extremely suspicious for any physicist. On the other hand, MOND predicts the correct galaxy rotation curves of galaxies we have not looked at yet.
Sure DM explains more than MOND, but it is also basically not a scientific theory, since it has far too many degrees of freedom that can explain away any problem and is extremely difficult to falsify.
Lambda CDM makes very few assumptions. It only has 6 free parameters.[0] If you start with those 6 parameters, and then let the laws of physics work, you get the universe we see today. There are a huge number of different types of phenomena that the theory has to correctly predict (e.g., the abundance of different elements, the numbers of galaxies with different masses, the rotation curves of galaxies), and it does so extremely well.
> Lambda CDM makes very few assumptions. It only has 6 free parameters.[0] If you start with those 6 parameters, and then let the laws of physics work, you get the universe we see today.
Not including the DM distribution needed to match each galaxy's rotation curve, or external field effects, and dozens more observations:
> Not including the DM distribution needed to match each galaxy's rotation curve
Incorrect. The present-day distribution of dark matter is a consequence of the 6 parameters in Lambda CDM. If you take the Lambda CDM assumptions (with 6 free parameters) and apply the known laws of physics (gravity, hydrodynamics, nuclear fusion, radiative processes, etc.), you get something that looks like the universe that we observe. You get correct rotation curves, correct distributions of galaxy masses, and so on.
If you can find a way in which Lambda CDM does not observe the known universe, then you will invalidate Lambda CDM (and maybe win a Nobel Prize).
> Contrast that with MOND's 1 parameter, visible mass.
There is currently no known theory of modified gravity that has the same predictive power as Lambda CDM. People have been trying very hard to find one, but so far without any luck. Lambda CDM works extremely well, as far as anyone can tell.
> If you can find a way in which Lambda CDM does not observe the known universe, then you will invalidate Lambda CDM
The paper I cited does exactly that, quantifying the extent to which CDM and MOND predict or must be adjusted to fit over 30 different observations via extra parameters.
The general consensus among astrophysicists is that MOND is not able to reproduce the full range of phenomena we observe, which LCDM does correctly reproduce. I'll just state that first, because the paper you're citing is representing an extremely fringe view in astrophysics.
Looking at this paper, the first thing I notice is that it's not published in one of the standard astrophysics journals. There's probably a reason for that, but let's give it a chance. Next, I see that the introduction is making some very strange statements, like the following:
> However, the MW retains a thin disc despite being much older than its dynamical time of a few hundred Myr [17]. For an early review of work on this problem, we refer the reader to [18]. A possible solution is that disc galaxies have a dominant pressure-supported spheroidal halo, even though such a halo is not observed [19]. The currently conventional solution to the above-mentioned issues is still to design invisible pressure-supported DM halos that surround, dominate, and stabilize galaxy discs [20,21].
Dark matter halos aren't specially "designed" to solve a particular problem - they're an inevitable consequence of LCDM. Not only that, but there's direct observational evidence for them: we can see their gravitational lensing signature. Here you see the same theory - cold dark matter - explaining two very different phenomena: galactic rotation curves and gravitational lensing.
There are many different predictions like this, from very different areas of astrophysics, that are all explained by essentially one assumption - that there is some sort of cold matter that does not interact with light (i.e., it has no charge). That's the strength of the theory.
There have been many modified gravity theories (modified Newtonian gravity, or "MOND," is not a serious theory, because it's not relativistic), but they all have the one or more of the following problems:
1. They mess something up that is very firmly established. For example, they might make gravitational waves propagate at less than the speed of light, when we know (since 2017) that they actually propagate at the speed of light.
2. They are crafted to solve one or two particular problems in astrophysics, but they fail to solve all the other problems that cold dark matter solves.
3. They don't actually predict anything new, beyond what vanilla general relativity predicts. They are essentially such slight modifications of relativity that they're experimentally indistinguishable. Such theories obviously can't solve the dark matter problem (because if they did, they'd be experimentally testable).
> Dark matter halos aren't specially "designed" to solve a particular problem - they're an inevitable consequence of LCDM.
You can't take one quote out of context as some meaningful criticism. The point they're making is that the particular distribution of dark matter and the initial conditions at the big bang have to be tuned to match observations because the natural predictions of CDM don't match observations. By contrast, in MOND many observations follow only from the amount of visible matter.
> There are many different predictions like this, from very different areas of astrophysics, that are all explained by essentially one assumption - that there is some sort of cold matter that does not interact with light (i.e., it has no charge). That's the strength of the theory.
This is the lie that's sold. There are numerous other parameters in LCDM (ie. assumptions) that must be tuned just right to match observations, and this paper goes over them and how MOND actually matches those observations without free parameters, or with some additional assumptions of its own.
LCDM itself requires a number of assumptions to match observations, so the question is, given the evidence, how many of those observations are expected outcomes of the model without tuning, how many require tuning and/or specific initial conditions, and how many require additional assumptions?
They do this for over 30+ observations and MOND comes out looking much better than you'd expect, so there's definitely something to it. It would just be a remarkable coincidence for MOND to make so many successful predictions based only on visible matter in cases where the LCDM requires visible
matter + postulating invisible matter in a specific distribution.
> 2. They are crafted to solve one or two particular problems in astrophysics, but they fail to solve all the other problems that cold dark matter solves.
By contrast, most astrophysicists that are gung-ho on LCDM sweep under the rug the inconvenient tuning that's needed to actually match observations. If you actually read the paper they acknowledge that something like sterile neutrinos are needed to explain some observations under MOND, but this is considerably less dark matter than the predominant model suggests is needed.
Nobody serious is claiming that MOND reproduces the full range of phenomena we see, but neither does LCDM, and LCDM without tuning doesn't look like our universe at all.
> The point they're making is that the particular distribution of dark matter and the initial conditions at the big bang have to be tuned to match observations because the natural predictions of CDM don't match observations.
This isn't true. The distribution of dark matter in LCDM matches the observed distribution remarkably well. There are still some uncertainties at the cores of galaxies, where baryonic physics is more complex (for example, star formation and black hole feedback are difficult to simulate accurately), but on larger scales, LCDM does very well without any tuning.
> By contrast, in MOND many observations follow only from the amount of visible matter.
MOND isn't an actual theory. It's an idea of how one might modify Newtonian mechanics, but any fully fleshed-out theory is relativistic. Which theory of modified general relativity are you referring to, specifically?
> LCDM itself requires a number of assumptions to match observations
It requires 6 parameters, as I said earlier, to explain everything in cosmology.
> LCDM requires visible matter + postulating invisible matter in a specific distribution.
You keep saying this, but it's simply wrong. LCDM only requires a Gaussian random field as an initial condition. It has one free parameter to specify that field's density distribution: a scalar spectral index. We actually know for certain that the universe was in such a state early on (from the Cosmic Microwave Background), so this is a very solid assumption. LCDM takes that initial state, applies the known laws of physics, and ends up with the matter density distribution we see today, without any tuning.
> By contrast, most astrophysicists that are gung-ho on LCDM sweep under the rug the inconvenient tuning that's needed to actually match observations.
Astrophysicists are not "gung-ho" about LCDM. They're constantly looking for ways in which it might break. It's actually quite depressing that LCDM does so incredibly well, because it would be exciting to find something new.
Astrophysicists are actually very open-minded about modified gravity. The problem is that, contrary to what the paper you've found is claiming, there's no theory of modified gravity that explains away dark matter, while reproducing all the cosmological phenomena that are observed. Every theory of modified gravity messes up something big, like the halo mass function, the elemental abundances or the speed of gravitational waves. The only theories that don't mess things up are tiny modifications of general relativity, but those theories require dark matter.
If you have to repeatedly modify "predictive" theory to match observation, then it was not predictive in first place. The situation with MOND appears much worse than for example Planck's theory for blackbody radiation. He revised it once and that was it.
Generally I don't get this splitting hairs. Things that we can only observe but can't reproduce in lab are not a science then?
I thought the problem with modifying gravity was that there was no one modification you could make to explain observations. It's plausible that 1000s of different galaxies spin differently because they each have different amount of mass of dark matter in them. But it's not really possible to explain 1000s of different data points without 1000s of different gravity equations. If there were some rule (eg on large scales gravity is actually an exponential) then someone would have found it and we would have a candidate for a new law. But they haven't...
> It's plausible that 1000s of different galaxies spin differently because they each have different amount of mass of dark matter in them.
You also need a plausible explanation why all of these galaxies have such different distributions even though they have similar visible mass. If DM is only supposed to interact gravitationally then there should be a tight correspondence with little deviation, but there are galaxies entirely free of DM and distributions that differ across visible mass. There are also observations that don't even match DM (external field effects, tidal effects, and more). There's clearly a lot more to the story.
Something I've always wondered: Right after the Big Bang there was a massive annihilation of matter and anti-matter, with just a bit of extra matter.
But the energy resulting from that annihilation did not go away. So where is it? No matter what form it takes, energy gravitates, so it should be able to be observed.
Photons. Each annihilation event emitted a photon, and they are still around. IIRC, about 5e9 times as many photons are observed as there are baryons, and that seems to match up well with the calculated imbalance of the baryosynthesis.
A lot of it just went back into producing particle-antiparticle pairs again. Some of it is the CMB, yes, but not all.
Why we didn't just end up with even numbers of particles and antiparticles repeating this dance endlessly is the single biggest open question in particle physics.
Probably people who were slapping more terms to epicycles knew what they were doing is wrong but did it anyway because it allowed them to predict what they needed.
It seems they can account for mass in various places in universe, they can model how things should be moving (being affected by gravity), they can measure how things are actually moving and from this they can infer where this missing mass actually is placed.
This is currently all pretty early stage but the resolution is expected to increase with time.
So no, I don't think dark matter is an epicycle. It is more or less like the case of gravitational disturbance affecting planets in solar system and somebody theorising a planet placed somewhere would explain the effect, we just can't yet measure things accurately enough to tell where the planet is exactly.
Looking back at this episode - after we figure this out - will be similar to looking back at the period before the discovery of general relativity.
Right now, we're really quite clueless.
My fun, science fiction theory to solve the gravity problem is that the universe is actually full of intelligent life mostly billions of years old civilizations and all of the extra mass we can’t see is actually the galactic version of cities built from materials that don’t interact with light much. Likewise we haven’t met any other intelligent species because the growth period before reaching this common technology is short from thousands to millions of years so there just aren’t collisions in space or time where two intelligent species would interact.
It's precisely the reason dark matter tends to not fall into black holes. If you put a particle into an orbit around a black hole, it will basically do so forever as it needs to lose an enormous amount of energy to drop down to a low enough orbit.
Normal matter bumps into other normal matter when orbiting a black hole, heats up, and radiates away it's energy. This effect is most visible in the accretion disk of a black hole.
But dark matter can still fall into a black hole and it still tends to concentrate due to gravitational interactions. If a dark matter particle is on a path unlucky enough to intercept the event horizon, it will of course not orbit. Also, gravitational interactions between dark matter particles or just normal matter can cause an exchange of energy, with one of the particles losing energy.
Oh, and if dark matter falls into a black hole, it just makes the black hole a little bigger. There's no way to externally tell that the black hole contains dark matter.
I was wondering about black holes created from dark matter only.
I mean: if there's 5x as much stuff that has mass (and, thus, can form black holes), there should be 5x as many black holes - at least, that's what I'd expect in the absence of some specific reasons.
The same reason the rest of the matter didn't collapse into black holes, not quite dense enough, or blowing up too fast. And dark matter doesn't have it's own kind of black holes. Related is the idea of "primordial black holes".
In fact I think it is harder for dark matter to collapse into black holes because it can't easily shed momentum (angular of otherwise) and 'spiral down' .
That's what i used to think, but GP alluded to weak interaction between DM particles.
I thinkt interaction consumes energy. If so, DM particles would be able to shed some momentum... but then they (eventually) could clump into black holes.
Again, no more than baryonic matter, which is demonstrably not all wrapped up in black holes (and probably much less, though bear in mind this is all pretty speculative).
Of course any dark matter that happens to be headed straight for a black hole will go in, no problem. Broadly there are a lot more variables in black hole formation than just self-interaction. Remember it's possible in principle for photons to form a black hole, if you get them to converge just right, but that doesn't mean it's common or at all relevant on a cosmological scale.
Maybe dark matter does have electromagnetic interactions with normal matter, but they are at an exceedingly high (or low) frequency. Science has limits in what frequencies it can probe, matter may not have the same limitations. A century long EM wave would look like background noise on a human scale. Petahertz (and beyond) vibrations could have extremely short range interactions that mostly occur in the near field.
My top, crackpot theory is we live in a black hole and dark matter is the accretion disk.
My second crackpot theory is we are four dimensional beings living in a five dimensional universe. Though I also like Cixin Liu’s idea that we are living in a universe that used to be five dimensions and has (mostly) decayed to four, with bits of five dimensional debris holding on.
> we are four dimensional beings living in a five dimensional universe
I think I prescribe to that theory. It feels like it's also a fancy way of saying we live in a simulation. Everything we experience is a puppet/shadow/extension of its true "full" self. Things could be connected in ways we can't measure in our traditional dimensions.
How do you think about yourself in such a way? That you live a limited existence of a bigger mind that experiences more things? Like you'd be living in the "left" brain of a mind and only see the "left brain things?"
If you experience a "shadow", what is the thing/being responsible for the shadow doing?
How does this help you think or rationalize about existence?
I'm asking from a purely curious place. I've come to similar perspectives but I kind of get stuck with "I exist here, and I only experience this reality."
The only thing that has really gone beyond the above line of thinking is when contemplating what the "mind" is. It seems to be the only thing truly beyond "space-time" and limits of "speed of light." Your mind can imagine any point, and travel there one way or another. It's a bit of a logical jump, but it does break the limits of how fast one can travel in some type of definition. So the mind is kind of the "4d/5d being" and everything else is the 3d/4d.
A fun way think about it is pressing on a filled balloon from the outside. Inside the balloon is our whole reality/dimension. Everyone is poking inwards and bending the rubber is what we see and how we interact with each other. But we're completely locked into our current senses.
In some ways, it sounds awfully... "spiritual", right?
The challenge is there's no point in trying to imagine what existence or thought outside of our reality looks like. Carl Sagan's FlatLand helps illustrate how stuck we really might be.
We don’t exactly know what the inside of a black hole looks like. It’s infinitely curved space so it’s potentially quite big. One theory is the space inside a black hole expands at the speed of light. You can’t leave because you can’t go fast enough to get out.
We know that if you have a big enough black hole that the space near the singularity no longer shreds matter. But would it matter if a sun came through intact or a sun’s worth of material came in, condensed, and went supernova?
It depends on what you mean by the "inside" of a black hole. If you mean the event horizon, then for supermassive black holes, you wouldn't even notice that you'd passed it, so at that part of the inside, it looks exactly the same as outside. You might notice a difference if you tried to change your ship's direction and head "away", but you'd see that nothing you could do would let you get away. All geodesics point inwards.
The insides of black holes are not infinitely curved, at least not most of the inside. That might be the singularity, but our understanding of that is definitely incomplete due to a lack of a theory of quantum gravity. And it does seem that nature abhors infinities, so the curvature of spacetime at the singularity is most likely not infinite, but we just don't know what it is.
Black holes have the property that things go from the event horizon to the singularity in a finite amount of proper time. Time reversed, an object shooting out from the singularity will hit the event horizon in a finite proper time. What would happen to us at that point?
All this is not very intuitive to me, but my mind's eye is telling me it would appear as accelerating expansion until you reach the horizon, and reaching the event horizon would feel like being in your own observable universe by yourself.
In GR, space and time form a single geometry called space-time, and gravity is a curvature of this geometry. Within this geometry, there are 3 spacial dimensions, and 1 time dimension. However, there is no universal axis. In the same way that different observers may disagree on which direction x is, they could also disagree on which direction t is.
In the context of a black hole, we can compare the coordinate system that someone far away would use, with the coordinate system of someone near the black hole. For someone far away, the center of the black hole is just that. A point in 3 dimensional space causing a gravitatational field. However as you aproach the black hole, your natural coordinate system starts to warp relative to a distant observer. Your time axis begins to point more towards the center of the black hole.
At the point you cross the event horizon, your perspective of "future" points entirely in what a distant observer would call a spatial direction: towards the center of the black hole. In particular, the black hole singularity is no longer a point in space, but a point in time. Every observer within the event horizon sees a universe where the future is an infinitly dense singularity that all geodesics will arrive at.
Compare this with our current universe. We see a universe where the past is an infinitely dense singularity that all geodesics originate from.
Ehh, read Steven Hawking. Black holes (singularities) are strikingly similar to what we know of the big bang. If we are in a black hole, it could explain the accelerating expansion of the universe.
my pet science fiction theory is that passing the event horizon removes half a dimension. so outside the universe you can move freely back and forth in time, but here in the universe it is constrained you can only go forward. when you pass another event horizon one of what we call a physical dimension gets constrained. you can't move toward the wall of the event horizon, that dimension is gone.
Dimensions do not disappear after you cross the event horizon. In fact, for very very large black holes, you would most likely notice nothing as you crossed the event horizon (barring there being a huge accretion disk spinning around at a significant fraction of the speed of light and spewing X-rays). That's because it's so far away from the singularity that the gravitational field isn't strong enough yet to meaningfully distort space. But all paths at that point forward lead to the singularity.
exactly there is now not only a direction that you cannot go in, you can't detect it is not there any more. the analogy with time still fits. the only way through time is forward toward the singularity but it will take you forever to get there. the best you can do is stop and that would take an infinite amount of energy.
> Mathematically, the universe looks a little bit like a black hole running backwards in time
It really doesn't, not even a little bit.
Schwarzschild has spherical symmetry, Kerr has axisymmetry, neither has homogeneity, and each has isotropy only at a very restrictive set of points in spacetime. These are vacuum solutions of the Einstein Field Equations of General Relativity; comparable metrics that introduce collapsing matter (e.g. Lemaître-Tolman-Bondi in the spherically symmetric case, something like a perturbed LTB or perturbed Oppenheimer-Snyder metric along the lines of Kegeles 1978 https://doi.org/10.1103/PhysRevD.18.1020 in the axisymmetric case; more modern solutions are mostly considered with numerical relativity) have even less isotropy and homogeneity.
Our universe for all practical purposes is homogeneous and looks isotropic from points far outside galaxy clusters, which is most points. There is no universal "centre" around which anything is orbiting. The mathematical description is the expanding Friedmann-Lemaître-Robertson-Walker (FLRW) metric, which is neither axisymmetric nor spherically symmetric, but which is homogeneous and everywhere spatially isotropic.
The Weyl tensor (the tidal part of the Riemann curvature tensor) is very different in the non-vacuum "white hole" solutions from FLRW, with galaxies and galaxy clusters in the former solutions producing a significant redshift-prolation relation for an observer freely-falling away from the white hole. We can also redshift-bucket galaxies and find ourselves (as "Eulerian" observers) collecting sets in planes and annuluses, with a very strongly distinguished set radially through the white hole. It would be attractive to use that set as the cosmological frame, comparable to the attractiveness of the one we use in the standard cosmology.
Time-reversal is interesting. There is a small literature on incoming Vaidya black holes which justifies a tiny handful of papers that treats bulk matter (e.g. galaxy clusters, or the stuff of the Friedmann equations) as the evaporating "shine" on cosmos-originating outgoing Vaiyda white holes. The authors in this area follow some contortions with respect to hiding the central part of such a cosmology (the observable universe must be a truly microscopic wedge in the whole picture in order to hide probably-unphysical observables), and do not appear to have made much progress.
Moreover, this sort of approach approach runs into another problem. "Structure disintegration" at late times (i.e., time reversed structure formation) is tricky, and I think leads you down a path comparable to Einstein-de Sitter 1932, which is now reasonably understood to be a less-than-ideal way of writing down a particular FLRW solution.
Roughly the problem is that you have to evaporate the earliest galaxies (and ultimately the earliest atoms, nuclei, and protons) which is hard to do while preserving the "no drama" condition near a very flat black hole horizon, and even harder to do far from the horizon. Under time-reversed-time-reversal (i.e., back to the white hole spitting out matter in the early universe), how do you recover adiabatic cooling?
We get adiabatic cooling from the metric expansion in FLRW; it's harder to see how BBNS, neutral hydrogen, and the end of the dark ages arise in a white hole solution, at least in the asymptotically flat case. And if you ditch asymptotic flatness for de Sitter (spherical symmetry) or maybe expanding Kasner (axisymmetry), why are you better off than FLRW? Or, back to your point, how does such a universe still look even a little bit like a time-reversed black hole? Accelerated expansion (or accelerated shrinking under time reversal) I think is going to be even harder to lift out of FLRW (with the appropriate value for the cosmological constant) into a cosmology with a central white hole.
> You can postulate an additional separate modification of gravity that comes into play on cluster scales.
Or it might be something else altogether.
Anyways, MOND has made predictions that people set out to disprove MOND went looking for negative evidence and had to turn around (e.g. external field effect), as well as straightforward predictions JWT may have confirmed (early galaxies) and explains many things that dark matter cannot, like how satellite galaxies don't get ripped apart by the dark matter halo.