[hutzlibu]

"But now it seems so obviously wrong that it’s almost laughable anyone could ever take it seriously"

Was it really that laughable?

"Whatever difficulties we may have in forming a consistent idea of the constitution of the aether, there can be no doubt that the interplanetary and interstellar spaces are not empty, but are occupied by a material substance or body, which is certainly the largest, and probably the most uniform body of which we have any knowledge.” (James Clerk Maxwell)

We discovered there is no true vacuum. And we still don't know how exactly electromagnetic waves travel, or how gravitaion works. So the old modell of aether is clearly wrong, but before we have not come up with a consistent model explaining all of it, I would not call it laughable. It made sense at the time. And to me there still is some appeall to the basic concept, that there must be something allowing the spreading of the various waves. Or is that question solved by now?

[Swizec]

> Was it really that laughable?

Perhaps laughable is a stretch. I remember aether as an example of the scientific method working in a high school class. We had a theory, it seemed reasonable, it fit observations, and then we disproved it. This is good.

The laughable part of my memory probably comes from being 15 when I learned about this and the whole class thinking “wow look at those fancy scientists, they didn’t even know basic things we all learn in middle school! ha ha”. Dumb kids be like that sometimes :)

BUT aether was also used as an example of failing Occam’s razor. It added weird unmeasurable just-so variables/explanations to existing theories so they could expand to fit new measurements. This rarely leads to a theory that stands the test of time.

In this way dark matter, in my lay-person view, feels similar. We don’t know what’s going on, so we say it’s gotta be some new “thing” that just happens to be invisible and undetectable except by how it just happens to make our existing models/math/explanations work. That seems fishy to me as a non-expert. Kinda like when an engineer says “i’ve tried everything, it’s gotta be a compiler bug” … it usually isn’t a compiler bug.

edit: Point is that when things don't fit together, just adding more <stuff> rarely works long-term. You need a new model. And personally I'm excited to see what we think about dark matter in 20 years.

[nyssos]

> just happens to be invisible and undetectable except by how it just happens to make our existing models/math/explanations work.

That describes all of physics: there's no such thing as "direct" detection in the folk "I know the billiard ball is there because I can see and feel it" sense. It only feels like there are because you've been using your models of human-scale physics for so long that you trust them completely and automatically. The only people operating in everyday life the way astronomers have to professionally are infants: for them, it's not obvious that there are such things as rigid bodies, that sight and touch should correspond in the way that they usually do, and so on. Not because they're stupid, but because if you want to reconstruct the whole world from a bunch of low-rez sense data you're going to need a ton of it, and they don't have that much yet.

> Point is that when things don't fit together, just adding more <stuff> rarely works long-term. You need a new model.

Adding new stuff is coming up with a new model. Particular dark matter candidates aren't things like "there's some stuff out there that makes things happen for no particular reason, case closed": that would get you gently corrected out of an undergrad scientific writing class, let alone a real journal. They're things like

"If we suppose the CP-symmetry violation term of QCD is the result of a spontaneously broken symmetry (resolving the strong CP problem), the allowed masses and EM-coupling strengths of the corresponding pseudo-Nambu-Goldstone boson are consistent with the required properties of cold dark matter, but would give off photons in extremely intense magnetic fields, so here are some proposed experiments:..."

[jancsika]

> This rarely leads to a theory that stands the test of time.

1. The simplest explanation is usually the best one.

2. The simplest explanation is usually the easiest to comprehend, discuss, and-- if applicable-- test. This leads to faster iteration times on the path toward a fuller understanding of the phenomenon in question. Consequently, it's best to tend toward the simplest explanation, even if that choice is no better than a coin's flip over a more complicated competing explanation for a particular paper.

Number 2 seems intuitive and supported by the best practices in programming. Well, let's say the least bad practices in programming. :)

Number 1 seems ambiguous enough to lead to some kind of cargo cult. Perhaps greatly increasing iteration time after the simplest explanations have been ruled out? E.g., reluctance to fund research into a more complicated explanation in such a case?