Why dark matter feels like cheating (and why it isn’t)
4gravitons.com
4gravitons.com
I was pretty skeptical about dark matter until I watched this presentation by Sabine Hossenfelder (she's a great explainer of science, IMO): https://www.youtube.com/watch?v=U4sw3-__pGo
I think that she did a better job of explaining why we should be open to the idea of dark matter, than the original article did.
We very famously don't have a model of gravity that works at quantum scales, and our current model clearly doesn't work at galactic scales.
So our solution to this problem is... Invisible pixie dust. We've just reinvented the aether, and we're even still doing the same experiments to detect dark matter as we tried to detect aether.
I really just don't buy it. Dark matter would require reworking everything from particle physics up to cosmological interactions. How is this more acceptable than our model of gravity being more wrong than it is now?
We're very clearly missing something fundamental in our understanding of gravity. We know that, and we've known it for a long time. The absense of dark matter would mean that our understanding of gravity is still wrong. Having a more incorrect model of gravity involves far fewer variables than an entirely new type of matter, so why is everyone on the dark matter bandwagon?
As stated before, it satisfies the following conditions:
A) It fills in the holes of existing formulas/models
B) It prevents MASSIVE retooling of everything we know
Until a better model comes along to explain the area filled in by dark matter, this is the best solution we have right now. It's the physics equivalent of strcpy() but without any available replacement: It works until the edge case.
Right now the (big J) Job is finding physics' strlcpy(): Something that can be used in place of strcpy().
The only thing we know is that spacetime is curving without any visible/detectable thing causing it (at least not in a way our equations predict.) Calling it “matter” is already making a huge set of assumptions.
I don’t think that’s correct. Most Dark Matter proponents are ardent adherents to general relativity. General Relativity is pretty clear that massless non-matter particles like the photon contribute to the bending of spacetime. So it seems like most Dark Matter proponents would stridently object to the claim that “everything that causes gravity must be matter”.
Given that the only thing we’ve observed is a change in the gravitational field (curvature of spacetime), then why call it matter? Why assume that it’s matter in the first place? This is where my criticism comes from. It seems like they’re eager to assume something that affects gravity must be matter, hence why it’s called dark matter.
The fact that the leading candidate for dark matter is a “weakly interacting massive particle” further cements this criticism IMO… it’s something we’ve never observed, but it’s just assumed to exist because it must be some sort of matter, right? Because we can’t think of anything else that bend spacetime…
My understanding of the position of the various physicists that I follow is that they are *not* eager to assume this. It’s just, after quite a bit of investigation, the best explanation that fits the sum total of information that we have the best. I wouldn’t describe any of the cosmologists that I’ve listened to describe their position as “eager” for it to be matter. The ones I’ve listened to, in fact, would be absolutely thrilled if we had evidence that ruled out the “matter” explanation.
> Why assume that it’s matter in the first place?
I… don’t think physics assume it must be matter. It’s the leading theory largely because we’ve spent a decade ruling out various other possibilities and ideas. Dark Matter is what we have left that best fits the sum of available evidence. But, again, every physicist I’ve heard talk about this subject has acknowledged that it’s basically an area with a lot of question marks, and we don’t really know much about it, so we’ll have to keep studying the question to try and learn more.
> Because we can’t think of anything else that bend spacetime…
Again, we can think of lots of things that can bend spacetime that aren’t matter. That’s trivial. But (again, my understanding from physicists I listen to) is that we’ve thought of lots of those things and ruled many of them out (because the observation we have are more than just “bends spacetime”). Dark Matter is what’s left after ruling lots of our ideas out.
Does that mean DM is definitely the cause of the behaviors we see? No! Of course not. It could absolutely be some matter-less cause that we haven’t ruled out yet.
I would throw out their that I don’t feel like you are able to provide a summary of the Dark Matter proponents position that most Dark Matter proponents would agree with (or, at least the ones I’ve listened to; I’m sure there are some proponents who hold the positions you describe, I’m just not familiar with them).
Not really. "Matter" is not really a formal term of art, but in practice it usually means "not an elementary boson", or at the strictest "fermionic".
So at best their axiom might be that if it has mass it interacts with electromagnetism which we already know not to be be true in all cases.
How gravity works in GR is more like a topological property of space time rather than say like the electromagnetic force which is why I was asking about what exactly they mean by interacting with gravity.
There are detectors that are trying to find weak force interactions. They haven't found anything yet which is constraining paramters of dark matter particles.
They're not literally thinking those thoughts in those terms, it's just what pops out when you try to translate complaints about "tangibility" and "seeing things directly" from folk-physical intuition into something physically meaningful.
Currently I would group the “dark matter” into 3 camps, dark matter as a new form particles, dark matter as a combination of any sort of existing particles and matter e.g. wimps, primordial blackholes, brown dwarfs, cold baryonic matter etc. and dark matter as modified gravity this is MOND and TEVES et al.
None of which as far as I know make a statement around having to have interaction with EM if you are interacting with gravity.
I’ve been following this debate at least casually for a long time, and I’ve never heard that argument. Can you present even one example of any respectable scientists arguing that?
So this normal person looks to see if this has happened, finds there are several rival theories, and no experiment done on Earth had produced results. Naturally, they start to doubt.
There is a reason that scientists communicate with other scientists via papers and a process called peer review and not typically via the media. That is, to distill the topic down to a casual level requires many levels of abstraction and tend to be 'incorrect but interesting' as the media does not sell truth, but instead it sells advertising by making things interesting.
That was the whole thesis. I genuinely described the exersize of looking up a definition.
There are two explanations: there is extra mass in galactic systems that we can't detect, or our understanding of gravity is more incorrect than previously thought.
When you dig in, there's a small amount of evidence that suggests there is extra mass, but also we can modify our equations of gravity to produce similar results.
Dark matter and MOND (one of the leading alternate gravity models) both have compelling cases for them, and both show exceptions that don't match observed data.
I genuinely don't understand why people push dark matter, as a solution it's way too complex for the problem it solves. It makes more sense that our model of gravity (which we know is already incomplete) is more incomplete than it does for there to be an entire class of matter that we can't detect or even construct meaningful theories about.
They like their equations, so they call the difference dark matter (which varies in space and time, as we expect). It is quite ridiculous if you think about it, and also is the fact, that they can get away with it.
I am waiting eagerly how will this turn out. (Well only the "dark matter exist", or the "Einstein equations are so useful that they create a meaningful dark matter concept with properties, even that they are incorrect" cases are interesting, if say MOND wins, that would be less amusing.)
I like my equations, so I call the difference the wall in front of my face.
Now there are some discrepancies between what we expect and what we measure at galactic scale. A very popular hypothesis is that these discrepancies are due to some dark material that we haven’t been able to observe otherwise.
If Einstein's equations didn't match any more observed data than, say, Newton's equations, no-one would have bothered adopting them. But there were lots of observed data at the end of the 1800s that didn't quite fit the models we had. Quite a lot were to do with light, but others were slightly weirder, like the precession of Mercury.
Then Einstein's equations came along and fixed all of them. They even made predictions about observations yet to be made, like the behaviour of clocks on satellites and spaceships, which are moving fast enough for special relativistic effects to be detectable, but also moving in a different-enough gravitational field for general relativistic effects to be detectable.
(Note, GPS relies on knowing and mitigating these effects in order to work accurately. But also, the designers of the first GPS system still weren't 100% sure it was the case, so they made it possible to change whether none, either, or both mitigations were active. It did turn out that both were needed - another win for Einstein.)
I suspect that if the "Einstein's equations not quite correct at intergalactic scale" case ended up being right, we would end up with something that's as different from Einstein as Einstein is from Newton. That is, a new set of equations that is functionally incredibly similar to the old ones, but with an extra term which is very nearly constant under most circumstances, but measurably diverges according to some yet-to-be-discovered criteria.
With Einstein's equations, this is the Lorentz factor γ, where 1/γ = √(1 - v²/c²). When v is only a small fraction of c, this term approaches 1, and Einstein's equations approach Newton's.
Note that in the case of v being small (compared to c), Newton's equations are still useful, and it's perfectly reasonable to use them to calculate motion and energy without including relativistic effects. And similarly, if we were to "replace" relativity with something else, there would be a lot of circumstances where the additional complexity of the replacement wasn't needed, and relativity would still be useful in the scale between Newton and it.
No one is opposed to a better model by the way. MOND is not universally embraced as such because it doesn’t actually get ride of dark matter so there is little point in switching.
Amusingly I think a large part of why some people are so upset about dark matter is that they don’t really understand what a model is and believe that experimental science seeks to convey the truth which is an interesting faith but has little philosophical basis. I generally blame the fact that these points are poorly explained during most people education and most scientists are reluctant to talk about that to no lose the tenuous support of the general population.
The problems don’t indicate that Dark Matter (the theory) is wrong necessarily. MOND is much better at explaining a number of observations of large scale (better meaning simpler in this case), but Dark Matter is better for other observations at smaller scale. Unfortunately it seems that the physicist community (outside of a small subset) is unwilling to research in the MOND space; it has the taint. So physicists just pile on a bunch of extra variables to make Dark Matter fit certain observations, when MOND describes those observations very simply.
Sabine argues long and hard that modern particle physicists have made no fundamental progress for 50 years due to poor scientific method, and she’s acerbic and popular with the plebs (such as me). I’m glad she’s a voice out there, but I’m sure she has put herself offside with a number (maybe most) working particle physicists.
It’s unfortunate. The phenomenon that MOND and Dark Matter seek to explain are really interesting, and the depths have clearly not been plumbed. The continual search and failure to find the Dark Matter particle is not doing physics any favours.
That’s all it is I think. Sabine (and probably others) have done a great job of explaining these concepts for the layman audience.
This article, and so many like it, don’t do that. So the physicist is left shaking their head, and the layman comes away still thinking “what a giant kludge that dark matter is”.
In realm of everyday experience neutrinos and proposed dark matter would be the same ghostly thing.
No, but as a posteriori reasons go "everything else couples to the weak force" is a pretty good one.
> them dark matter is unfalsifiable
Only insofar as all the ontological content of a physical theory are unfalsifiable: you can always shift complexity around between state and dynamics if you're willing to pay the price in elegance. Consider Kaluza-Klein theory, for instance. In some sense this demonstrates that classical electromagnetism is unfalsifiable - but no one actually cares.
For me it's because the explanations don't actually make sense. The classic example to me is the galactic rotation "problem". It gets resolved by suggesting a spherical cloud of dark matter surrounding a galaxy. Great, but why would it affect the galaxy normally but take on such a distribution? The solution raises more questions than the problem. Not to mention it's a whole new kind of matter we can't detect. It's almost magical thinking.
Electromagnetic interactions radiate off energy but not angular momentum, so a spinning sphere of gas will flatten over time. Dark matter doesn't interact electromagnetically and thus sheds energy much more slowly.
We already have examples of particles that do not interact electromagnetically but have mass, and therefore interact gravitationally. They're neutrinos. How is a new theory of gravity simpler than particles that are similar to ones we know exist?
I debunked this idea in the last DM thread: https://news.ycombinator.com/item?id=34365591
I didn’t say it’s better than Dark Matter; it’s not either or. I have no dog in this fight. I simply regard what Sabine has said “combine these two theories you idiots!” (paraphrased), and think “she’s probably got a point”.
I mean it doesn't really. Worse though, it doesn't explain the observations. It's just a model fit post hoc. Everytime they try to explain why gravity would act like that they just end up reinventing dark matter but calling it a field and hoping no one points out particle/field equivalence during peer review.
So it needs more research, but pretending that the model doesn’t fit the observations with a very simple framework, and instead throwing more and more parameters at Dark Matter to get it to fit… At some point it’s starting to look silly?
Why not consider that there is probably more to understand here and try to combine these theories as Sabine is arguing?
What's conspicuous however, is that MOND can make many successful a priori predictions about observations that LCDM needs to fit a posteriori using a "tuned" distribution of dark matter.
So MOND does indeed describe those observations simply even if it doesn't explain why they hold. An example of a better way to handle this than outright dismissing MOND, are recent proposals for superfluid dark matter that reproduce MOND in the right regimes.
those are the same predictions you'd get if you just set dark matter at a single uniform universal distribution. which is again, exactly what every attempt at explaining mond just ends up doing anyway... but then there's also those things that disagree with mond rendering the need for introducing a bunch of new stuff on top of it anyway.
and that's the thing mond leads us to dark matter plus extra stuff. Occam's razor is hardly an absolute truth but we like it and we can explain things without the extra stuff.
it's also like a real fucking brain worm to call "matter follows a statistical distribution": tuning or more parameters like the sibling post does. "Matter follows a statistical distribution" is what we already see in all the matter we see.
Personally, the way I think about dark matter is that it’s a map of the variances we see between the expected behaviour and actual behaviour of galaxies. That’s it. That’s it’s a distribution of gravitating matter is one possibility. That would fit the observations, so it’s a reasonable working hypothesis, that’s all.
MOND simply doesn’t describe the distribution of the effects we see. It does in many common cases, but not all by a long way and the variance is different from case to case, so it still leaves a gap we’d need to fill with something else.
Iirc, LCDM has trouble explaining satellite galaxies. They're traveling through the halo, which should create drag on them. They shouldn't last long enough to exist at all.
Yes that's fine, but it also doesn't necessarily try to [1]. LCDM can't explain linear Tully-fisher, can't explain external field effect, can't explain why satellite galaxies are long lived, can't explain why bullet cluster even exists (galaxies shouldn't be going that fast), can't explain early galaxies, etc.
[1] so maybe it shouldn't. There's a lot of systematic uncertainty in some of those really large scale movements. For example suppose our standard candle was off by half an order of magnitude at the largest scales. For LCDM, not a problem. Reparametrize the distribution and ratios of dark matter. Presto chango theory still works. That this should be possible should scare you.
And LCDM does this too. If you want a comprehensive survey of the tuning done by each model to fit observations, see:
https://arxiv.org/abs/2110.06936v3
It seems clear that neither approach is satisfactory as they stand, but the resistance to recognizing the significant flaws in LCDM is frankly puzzling.
So does Dark Matter
I'm curious about your background: you seem to know both physics and Prolog? How come?
Btw, thanks for your comments in the Prolog thread. Much more level-headed reaction than mine, I have to say with embarrassment.
Just sorta started noodling around with prolog a while back on a whim. Fits my brain well enough and it's kinda fun. Still never really used it in anger but I've been hanging around tech/programming circles for so long i can speak (somewhat) intelligently about languages well before I can do anything with them.
I hope you stick with Prolog. It's a unique thing. If only more people understood it, and used it, in their everyday work.
Thanks!
You'll find the two journal papers in my old profile at the very end of Louise's Readme :)
> and extremely intuitive for the physicist,
No it isn't. We haven't gotten here because it's intuitive, we've gotten here because it's the only hypothesis which is consistent with the data.
> Unfortunately it seems that the physicist community (outside of a small subset) is unwilling to research in the MOND space; it has the taint.
This is false. MOND was a leading theory up until it was falsified by the Bullet Cluster. MOND was later falsified by baryon acoustic oscillations. Then again by ultra diffuse galaxies. Most MOND theories predicted gravitational waves traveling slower than the speed of light, all of which were falsified by GW 170817, the neutron star-neutron star merger a few years ago.
MOND has little support in 'orthodox' physics and cosmology because it keeps getting falsified.
> So physicists just pile on a bunch of extra variables to make Dark Matter fit certain observations, when MOND describes those observations very simply.
MOND is viable because we add more variables to it every time the latest observation comes along that falsifies it. It's true that the original, first incarnation of MOND is simpler than CDM, but every time we learn more about the universe, CDM still stands but MOND is invalidated, and so another variable has to be added to it.
CDM frontloads all its warts and complexity. "We propose new, as yet undetected and possibly undetectable particles." Everything else is just running the numbers. MOND frontloads its simplicity. "F = m a^2 / a_0". Everything else adds more variables. You want it to be compatible with GR? You need to add new variables to Einstein's equations. You want it to have stars in that are stable? You need to add new variables. You want it to be consistent with gravitational waves traveling near or at the speed of light? New variables. You want it to be compatible with BAO observations? New variables.
You want it to be compatible with the bullet cluster and ultra diffuse galaxies? Now you need to add CDM.
We are now at the point where viable MOND theories still include CDM, just less CDM than pure CDM. Ballpark numbers are that pure CDM predicts 80% of galaxy rotation curves are caused by CDM and 20% by normal matter, leading viable MOND predicts 20% by normal unadjusted gravity on normal matter, 40% by CDM, and 40% by the modified gravity equations. If you're just looking for the simpler model, MOND ain't it.
For us to tell that there must be this amount of something that only affects things gravitationally for observations to work, just seems like pure kludge factor to get everything look right.
Then again, I have not studied whole thing deeply.
But then dark energy does feel like a hack, right? Throwing the whole “dark thing” concept under the bus.
So to speak :-)
>But then dark energy does feel like a hack, right? Throwing the whole “dark thing” concept under the bus.
Yes, dark energy really is nothing more than a mathematical hack (at least for now). There's a reason that a certain personality type chooses to go into physics and not marketing, with the result being that the market for physics often leaves something to be desired.
Correction. We don't know this for fact. It's the prevailing theory fitting the phenomena we observe.
That still seems to be a matter of serious dispute.
Dark energy is a whole different story and seems more like the epicycles of the heliocentric model: Adding a mathematical hack in order to fit the model into the data.
I'm remain curios about stationary universe model proposed by Peter Ostermann [1] which can explain the SNe-Ia data with homogeneously distributed dark matter and without a need for dark energy. Also, it is a lot more elegant and "Einsteinian" than the concordance model.
As an explanatory concept is has very low utility: there is not much else you can do with it except plug what you find missing. You can't say, for example, that because of this and this aspect of dark matter I predict this cool effect and then go search for it and either falsify or strengthen the confidence of your thought framework
Example: you observe wonky rotation curves of galaxies. You reckon there might be some extra invisible matter. From that you predict that you should also see this extra matter in lensing observations. You make the lensing observations and lo and behold, you see the same amount of extra matter that is needed to explain the rotation curves.
I wrote a longer comment on this before: https://news.ycombinator.com/item?id=34365591
This is so oversimplified that I'd call it incorrect. The galaxy rotation curves cannot be predicted a priori using dark matter models, the distribution has to be tuned a posteriori to fit the curve. This is in contrast to MOND which can make successful a priori predictions. There's clearly something missing in the dark matter picture.
Dark matter, being (likely?) a low energy kind of thing feels like a genuine gap within the realm we claim as understood. In this sense its more fascinating.
I wonder where that energy went and how they acquired that much energy in the first place.
this has always felt circular to me -- someone drops an astronomy paper like 'we found a galaxy with no dark matter' and I wish someone would rewrite it as 'dear non physicist, here are ten critical takes you just thought of and why we discarded each'
this is mostly a knock on my own knowledge, and slightly a knock on pop science press, but I don't know the steps between 'mass as inferred from light doesn't explain galactic rotation curves' and '80% of mass is ghosts'
This is the root of the disconnect, I think: dark matter isn't ghosts. Or if you're feeling snarky: all matter is ghosts. Any talk of "substances" or "essences" or "intrinsic properties" is entirely outside the scope of physics: the only things we ever see are interactions, and interactions with light don't have any special epistemic status that interactions with gravity lack. If anything, the opposite is true.
unless I'm missing something, this is mysterious in the exact same way as normal galaxies
If it is real and supposedly so common, why doesn't it exist here on Earth? What makes our world so uniquely special? Historically scientific models that treat Earth as unique in the universe always turn out to be wrong.
Hundreds of years of scientific research of all kinds crawling the planet and to date no one has been able to isolate it.
Option 1: Do more research on Gravity to see what we might be missing there. The difficulty in measuring G, and the flyby anomaly would be good places to start.
Option 2: Make up fully unconstrained variable 'X' that can take any value you want at any location or time in the universe[1]. This is amazing because it can perfectly solve so many difficult problems!
Obviously many people will and should be skeptical of throwing full confidence at option 2 in it's current state.
[1] Unlike neutrinos or black holes, there is no theory for what dark matter is made of, and how it is created or destroyed. Thus there are no constraints on how much you can have at any given location or time.
Dark Matter persists precisely because it's the best theory that fits the available observations. Arguments against it are the bits predicated on squishy stuff like "aesthetics".
No one thinks we have all the answers, but demanding we throw out the best model we have is going backwards.
I think what you're asking for is a lab-testable prediction from a dark matter theory. And no, we don't have that. It would be nice if we did, but we don't. We don't have it from MOND either, though, so I still don't get your denialism. If you don't want us to do science to figure this out, what are you asking for?
All those theories were merely fitting to data. Do you genuinely claim they were "unscientific"?
I repeat: the desire to reject a theory for aesthetic reasons is the "faith" side of this argument. It's OK to prefer a better theory if you have one. It's not OK to sling insults at the people who point out that it's the best theory we have.
Cosmology is not amenable to either predictions or engineering. There’s not even anything to deny.
I disagree with that. The biggest hurdle people have (in my experience is) is the claim that 85% of the universe is made of something that no one has ever observed directly. We believe it exists because of the behavior of the matter we can observe and our understanding of gravity. It's the best theory we have, and yet it smells a little like epicycles, phlogiston and ether.
The only thing anyone has ever observed directly is the contents of their own mind. All empirical work, without exception, is mediated by instruments, whether they're retinas or telescopes or 50,000 metric tons of purified water hooked up to a photomultiplier tube. Low-energy electromagnetic interactions happen to be the channel we're naturally best-equipped to use in our daily lives, but they have no special epistemic significance.
But the point you make brings up an interesting analogy though. If we consider optical illusions — where the mental model our brains create at a preconscious level lead us to incorrect conclusions about reality (insofar as an objective reality can exist) — can that inform our understanding of the models we use to interpret our astronomical observations?
The article may claim or argue this, but supports this claim rather poorly.
Look the fact that dark matter is even a topic, is because our models are wrong somehow. Dark matter is a convenient theory, but essentially a non-testable one, and one that has yet to have any real predictive power (outside of it accurately predicting things where our models fail at predicting things - but this isn't a real case, as it's essentially circular reasoning when it's existence is inferred from the errors in our model in the first place).
There’s an explanation for all of this behavior that we are measuring but saying that it’s something we can’t see but it’s there!! is objectively absurd.
It is much the same for dark matter/dark energy. Without the dark matter 'fudge factor' cosmic observations at galactic scales or above are broken. We've also when and by observation removed a number of things that this fudge factor cannot be. And yet in our observations this fudge factor keeps existing and is found groups like the bullet cluster, or in a small number of galaxies that appear to have no dark matter.
There is not a single reputable scientist that says they know what 'dark matter' is. Scientists know that there are knowledge holes that exist here. They do say it is measurable and that you the denier cannot handwave it away without bringing forth a better theory.