Hacker News Reader

Hossenfelder is refering to the ongoing discussion in the physicist community on this topic. Some of the giants of the field (Wald, Ellis) semi-regularly exchange papers on this.

It's not a hunch.

Green and Wald tried to proof that the inhomogeneities don't matter, others disagree that they proofed anything of the sort. It's a fairly robust debate, and I am not qualified to summarize the state of play. Here are some slides:

https://cosmoback.sciencesconf.org/data/program/Ostrowski.pd...

Edit: The rebuttal paper to Green-Wald from three years ago already has 100 citations... https://arxiv.org/abs/1505.07800

It's possible that the question will be settled in the next years through numerical simulations:

https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.11...

Edit 2: All of this is for dark energy, for dark matter I don't know if there is a similar discussion on the role of the post-newtonian approximation that is used there.

“Match up really really well” is not that impressive if you consider that you deduce the distribution of the dark matter from the observations. So it just means that in most cases you can find a distribution of in all other ways unobserved matter that explains your observation. That leaves you quite many degrees of freedom to play with.

Not exactly. Independent experiments that check completely different things find the same total distributions of masss and energy. So the mass-like and energy-like deviations from the observed mass and energy are always pretty much consistent.

They all observe the effects of gravitation, don't they? Whether it is deviations in lensing or the velocity and density distribution or visible matter, they all seem to have the same source, but it doesn't say anything more about the nature of the source of the deviation.

But we also see galaxies whose gravity is much weaker than we’d anticipate based on the objects we can see, impying they have less dark matter.

Are you saying that we observe galaxies whose gravity is weaker than we anticipated from the visible mass? Then that would be very interesting news to me, and something that you can't explain with dark matter.

If you are saying that we observe galaxies whose gravity is weaker than your models for the distribution of dark matter, then that is an argument against your models, not for them.

The point, as I understand it, is that we've found cases of galaxies that look like they should be about the same mass as each other based on the light they're putting out themselves, but gravitational lensing measurements indicate a significant mass discrepancy. That's what's hard to explain without "maybe there's a bunch of matter we can't see" - if it was more of a "maybe gravity just works differently at galactic scales" situation, that scenario would be impossible.

I understand the motivation for dark matter. My point is that as long we have a full density distribution to play with, we can explain almost anything with the model. And the start of this thread was me disagreeing with how much it actually proved that observations matched "really really well " with the model.

I don't see why the scenario "gravity works differently at large scales" couldn't explain the observations until we have some idea how gravity would work differently.

Don't forget that each of these observations rely on a quite large sets of assumptions about everything from how much luminosity from galaxies with a certain (normal) mass can vary to how we estimate distances to very far away objects. We don't have that many observations of e.g. gravitational lensing. And, if we realize we actually don't know how to calculate things correctly with GR, or our theory of GR is wrong, that will have implications for all of those assumptions.

The dark matter hypothesis might very well be right. It is a very reasonable guess. But so far I think the proofs for it have been overstated.

Nobody says it's proven. It's just the best what we have so far. Any theory that would work just as good and explain anything on top would be welcome.

And I can say myself that not everyone agrees with that discovery: http://www.astronomy.com/news/2019/09/astronomers-cant-agree...

Well, people still debate about the Bullet cluster. But if it's correct then it's actually a good (relatively) confirmation for some sort of dark matter. It two similar galaxies look the same but have dramatically different mass distribution then there's some mass we don't see yet.

No, observations of galaxies that seem to have little dark matter are an argument for the existence of dark matter (over the competing theory of a modified gravitational interaction).

And here is why: If you have two galaxies with about the same size and about the same number of observed stars they should have about the same distribution of barionic matter. So gravity should work the same in them, no matter if that gravity is described by Newtonian theory (probably not), General Relativity (I would assume that, but who knows) or your modified theory of gravity (maybe, maybe not). Any difference between them can not be due to the theory of gravity. But it can easily come from different amounts of un-seen dark matter.

Exactly. The only thing that is really certain is that correction terms are required when fitting models to observations and these take on the role of mass and/or energy.

The usual working assumption is that if a well established theory indicates the existence of extra mass or energy, it usually exists in nature and can be found. This is essentially something that has its roots in particle physics: every violation of conservation of energy or momentum turned out to be caused by a hitherto unknown particle.

I think you misread. Sabine is arguing that maybe dark matter is just the error term in our approximations of real matter, and pointing out that no one yet has actually taken the time to disprove that Occam's Razor theory.

I wouldn't presume to argue against her, because she's a physicist and I'm not.

I'm mostly making the case to the peanut gallery here to maintain a state of epistemic noncommittal. It's a very popular intuition to think that "dark matter" is just a weird consequence of modeling things improperly, kind of like how we used to think there was another planet between Mercury and the sun because that was the only way to make Mercury's orbit consistent with Newtonian dynamics. If we are modeling things improperly, I suspect it would have to be a pretty weird bug--and I think the one she's suggesting qualifies as such.

Yes, and she says it quite clearly:

> Now, what we do when we want to explain what a galaxy does, or a galaxy cluster, or even the whole universe, is not to plug the matter and energy of every single planet and star into the equations. This would be computationally unfeasible. Instead, we use an average of matter and energy, and use that as the source for gravity.

> Needless to say, taking an average on one side of the equation requires that you also take an average on the other side. But since the gravitational part is non-linear, this will not give you the same equations that we use for the solar system: The average of a function of a variable is not the same as the function of the average of the variable. We know it’s not. But whenever we use general relativity on large scales, we assume that this is the case.

> So, we know that strictly speaking the equations we use are wrong. The big question is, then, just how wrong are they?

[emphasis added]

It's not a "hunch". It's just how the math works.

average of 0-9 = 4.5

4.5^2 = 20.25

average of squares of 0-9 = 28.5

Thanks for this, that is the best summary I have seen yet in this thread.

It isn't a hunch, though. Physicists have known since the 1930s that the math is wrong because they are using averages in non-linear equations.

The problem is that no one knows how wrong.

> Physicists (including the author here!) are generally smart enough to have thought of those objections themselves

Yet we're no closer to a satisfactory answer. It's not that physics has some unresolved problems, it's that physics has a giant hole through its center.

Einstein got us closer than Newton, but his work is obviously incomplete. The dogma surrounding Einstein's work borders on religion.

But muh experiments!

How many times have Newtonian mechanics been proven? Lots and lots. And at one frame of reference (pun not intended) it's absolutely correct.

How is Einstein's work any different? It's not. Stop pretending only a chosen few can have meaningful ideas about how the universe works. You might need training to translate that idea into a workable theory, but not to generate the ideas themselves.

If I were you (collectively) I'd be taking ideas from everywhere. It's still your job to see if they make sense, so don't worry about the unwashed masses stealing your thunder.