Hacker News Reader

Comment by cygx | Hacker News Reader

This should, in and of itself, begin to cast down on the hypothesis even lacking a better alternative.

Only if your search was exhaustive: Scooping up a spoonful of ocean is not enough to conclude fish do not exist...

That's really the problem here though. You cannot falsify dark matter, and you cannot prove a negative. Imagine I say there is a 5 headed green deer that weighs 2000 pounds and can speak Gaelic, but it's really good at avoiding being spotted - so good in fact that you can even look at right at it and might not see it. And you can even touch it and not even know you're touching it! What would you could consider an exhaustive proof of its lack of existence?

You can leave carrots everywhere, but perhaps it doesn't like carrots. Or maybe you didn't spot it eating those carrots. Maybe you left them in the wrong spot, or at the wrong time. Maybe you need bigger carrots. Maybe you need smaller carrots. Maybe it likes cooked carrots instead of raw carrots. Maybe you need a better camera. Maybe the reason you camera didn't catch it is because you weren't using a flash, or maybe you need to setup an infrared camera, maybe you need to play a Gaelic greeting on loop, etc, etc. You can compile a literally infinite number of ideas to try to detect our dear deer.

Instead with non falsifiable theories, all you can do is approach them from an abundance of failures. You've tried carrots, you've tried infrared, you've tried ..., and it's still not coming out. Each failure makes it more and more likely that our deer simply does not exist. Of course this does not mean it definitively does not, but nothing will ever mean that.

> You cannot falsify dark matter, and you cannot prove a negative. Imagine I say there is a 5 headed green deer that weighs 2000 pounds and can speak Gaelic, but it's really good at avoiding being spotted - so good in fact that you can even look at right at it and might not see it. And you can even touch it and not even know you're touching it! What would you could consider an exhaustive proof of its lack of existence?

Dark matter isn't like this, though. It's more like an invisible deer that's been eating your grass and shitting on your lawn. At some point you have to say "I might not be able to see it, but the only thing that eats grass and leaves deer shit is ... a deer."

We have the problem that the observable mass of a galaxy is not sufficient to explain the motion of its constituents. As gravity is a pretty well validated theory, it can all be solved if we postulate some matter that we can't see which is still gravitationally affecting those things we can see. Dark matter, in this respect, is no more than the simplest explanation of what's happening via a single item that, if it were true, would explain things.

If you think you have a deer you haven't seen so far eating your lawn, you can put out infrared cameras and all sorts of equipment - and just maybe you might find that someone down the road has been messing with you and is mowing it at midnight. But equally, you might see grass disappear and see shit appear from mid air, and come to the conclusion whatever's doing it is invisible.

Ultimately, it's very conceivable that someone might disprove dark matter. All they have to do is posit some other candidate (eg a modification to gravity), and then predict a result that would follow from that. So far, all the other candidates haven't resulted in a hit, but the field is open for the next great idea to come up with one.

But dark matter no longer is particularly simple. At the time of its initial proposal it was little more than unobserved masses, which could have been something as simple as e.g. red dwarves. And so it made sense to reference dark matter as a place holder. But over time as we've discovered things are even more off, dark matter itself has started to morph into our 2000 pound Gaelic speaking deer. Now it comprises the vast majority of all matter in the world, it can't be baryonic, it mustn't interact with anything except itself and gravity, we need to have entire galaxies made of nothing but dark matter, and so forth, and so on. Add now add in the whole slew of experimental failures to detect it, as null hypotheses, and it features ever more exotic conditions as well.

And on the topic of those test failures, every single test we make to detect dark matter, which everything we hypothesize about it indicates we should be able to measure, says it's not there. This should not be a problem except for the fact that we've performed alot of experiments at this point to try to detect dark matter of all sorts, with increasingly high precision. And they've all returned a strong negative. With Bayesian logic it's almost surprising we haven't had a false positive yet, given all the failures.

I really do think the geocentric model has a lot in common here. Prior to Newton's discovery of strong mathematical indicators for gravitational mechanics, heliocentricism had long been proposed. And it was in many ways a much more simple explanation for many phenomena than geocentricism, which over the centuries became ever more convoluted and complex. The 'petal weaving' geocentric astronomic models are enough to boggle the mind even today. But one of the big problems was not even the church, it that there was an immense amount of academic work based around a geocentric model. As but one example astrology, which was considered a scholarly and academic pursuit in the past, was based heavily on the assumptions of a geocentric universe. For instance Mercury starting to 'go backwards' (which was required to glue together a geocentric astronomic view) is still a part of astrological 'theory' today.

Alternatives to dark matter look increasingly more likely with every failure of experiment designed to test it. And these alternatives need substantial refinement, but I think really the biggest issue with them is that if dark matter is not correct it would basically destroy more than a century of astronomical and cosmological work. Nearly everything we know about the cosmos is predicated on the assumption of dark matter. If it turns out dark matter is not the issue, we're looking at setting astronomical and cosmological work in many areas back a century. It would be, by fair, the biggest failure in modern science. So there is reason that alternatives to dark matter are held to a higher standard than dark matter itself, beyond simplicity - which dark matter no longer really is.

> It would be, by fair, the biggest failure in modern science. So there is reason that alternatives to dark matter are held to a higher standard than dark matter itself, beyond simplicity - which dark matter no longer really is.

I don't think this is anything to do with a higher standard. It's more like Occam's razor. If there's an alternative explanation that will upend our whole understanding of gravity and cosmology, have at it. But such an explanation must come with something testable that shows it's true.

Here's an example of an alternative explanation, "a wizard did it." But that's not science, because there's no way to prove it. As a result, in the absence of a testable hypothesis, you have to go along with the one that does the least violence to all currently existing knowledge, surely?

I.e. what I'm saying is that it isn't on the Dark matter theory to prove itself. We already have observable gravitational issues that need something weird to be true. Dark matter is just the simplest explanation anyone can come up with that fits the facts. It's on any other theory to either throw out less of already-established science by being an even simpler explanation, or to prove it's justified in throwing out more by showing experimental evidence.

Or, to put it another way. I agree that dark matter isn't a simple theory. But there isn't currently a simpler one on the market.

Let's go back to the late 19th century and give the physicists of the time a peek of what would come to be the future of dark matter -- that it would not only end up needing to be a rather exotic form of matter, but that it would need to account for nearly all matter in the universe, along with all the countless negative conditions we can also now apply to its existence as a result of various experimental failures. Of course we would also give them the pros of dark matter in that it also explains, after substantial 'refactoring', some other otherwise difficult to explain issues. The dilemma that they then face is pondering whether this material seems more probable, or whether it might be the case that e.g. Newtonian dynamics is somehow missing something? For brevity, I'm sidestepping relativity.

I think you'll see a lot of the issues with alternatives such as MOND are a 'lack of elegance', yet we've happily accepted that the universe on a very small scale is far from elegant, so it seems peculiar to insist that on the very large scale it must behave with elegance. Another issue is that MOND generally also has a requirement of some dark matter, but the critical difference is it's just that - matter (including baryonic) that cannot yet be observed, not the increasingly exotic material that dark matter has morphed into. And it also only requires a small fraction of the amount that current dark matter does.

I tend to agree that dark matter is the most probable explanation. But, at the same time I think the resource allocation for dark matter as opposed to everything else is probably not the one we would choose if we were being completely impartial.

The idea of non-visible matter as a fix for stellar motion dates back to the 19th century, when our knowledge about the cosmos was far more limited (remember, we only learned about other galaxies in the 1920s - it post-dates general relativity!). That the hypothesis is still viable today with our more elaborate understanding of the universe does validate the idea to some degree: Our cosmological models cut have killed it off easily.

So unless a better model that works equally well comes along, I see no reason to discard the dark matter hypothesis just yet: The universe has no obligation to be filled with a type of matter detectable with the means of today...

Yes, it dates to the XIX Century, and alternative types of it have been proven right or wrong at times.

Postulating unobservable things is still not a good way to do science. The one good thing of the dark matters theories was that physicists were clear in that it's not a done model, just a map of what's wrong with our understanding. This seems to have changed lately.

I'd generally agree with you, except for one thing. I'm not sure our models could have killed it off, excepting finding precisely whatever is causing the apparent mass abnormalities throughout the universe. Have you seen those ancient astronomic models used to illustrate a geocentric Earth? The one with the numerous spinning 'petals' surrounding the Earth? It's interesting imagining the amount of work that went into creating such models, constantly adjusting them over and over each time yet another little problem came up. Happily accepting planets just randomly changing direction, with no analog whatsoever anywhere else in nature, and so on.

When hypotheses are not falsifiable, they just tend to grow ever more elaborate rather than die. Dark matter started off as literally that - a place holder for bodies that did not significantly radiate. It's slowly changed into becoming the vast majority of all matter in the universe, existing practically everywhere, it can't be baryonic, it interacts with nothing except itself and gravity, there must exist near entirely dark matter galaxies, and so forth. So we're left with this ever evolving hypothesis that can't ever really be falsified and which every test, increasingly clever and precise, we create to detect it -- says it's not there.

And it seems that the zeitgeist is starting to increasingly treat dark matter as a fact whose proof is little more than an unscheduled formality. This seems illogical given that we've done more than ever to try to show it exists, and every result is saying - nope. And again this does not mean it does not exist, but each failure makes it more likely that this is the case, and there is an alternative explanation.

Reply on news.ycombinator.com