It obviously doesn't interact with ordinary matter very much, or else it would have been observed non-gravitationally by now. It could interact a little bit, though. Consider neutrinos: they almost don't interact with ordinary matter at all, but they can be observed with enough patience and care, and they are definitely real.

Neutrinos were actually an early candidate for dark matter. The numbers didn't work out, but it's an example of something that would interact weakly.

And sure, dark matter is falsifiable. If observations are made where gravity appears to deviate from what's visible and can't be explained by adding invisible mass, that would falsify it. Or if someone finally comes up with a change to the known laws of physics which explains the observations better than dark matter does, that would do it.

Remember, nothing is truly "directly observed." All we can do to observe stuff is to see how it affects other stuff. Whether it's with photons or with gravity, it's all ultimately indirect.

> And sure, dark matter is falsifiable. If observations are made where gravity appears to deviate from what's visible and can't be explained by adding invisible mass, that would falsify it.

Whats to stop us from "adding epicycles" when that happens?

The same thing that stopped the addition of epicycles to the geocentric model of the solar system: a better theory that explains the observations more naturally.

How do we know dark matter isn't just really transparent matter? Like, we can't see it in space but it's there, not reflecting or deflecting enough light for us to detect it.

Because it would collide with itself. Dark matter isn't just "not visible" - it's non-electromagnetic. As observed it didn't aggregate when colliding with itself, so there aren't big dark matter made black holes around, nor gravity anomalies from dark matter nebula or the like.

Two dark matter particles headed directly at each other would pass through each other and be only slightly deflected by their gravitational interaction.

Not necessarily. Dark matter could be made of small dense objects. Because they are small, we wouldn't see absorption of light. And because they are small, they will very seldom collide. These kinds of object are called MACHOs.

Based on observed dimming events while monitoring many stars, we have ruled out brown dwarfs as more than 20% of the total. Other possibilities still remain, but none have shown evidence of being sufficient to explain dark matter. However we have not ruled out the possibility of it being something that we haven't thought of yet.

Yes, that's what cosmologists mean by "dark": not interacting with light of any observable wavelength, to any detectable degree.

It could be[0]. It's just not very likely, since you'd have to have very special circumstances, but it's not logically impossible, at least in cosmology.

[0] http://abyss.uoregon.edu/~js/cosmo/lectures/lec17.html

From what I understand, if dark matter is stuff that only interacts gravitationally, by definition it is transparent to light. Also collisions, being ultimately electromagnetic interactions, aren't observed. So it's not transparent in the sense of glass. But I'm not a physicist, and this thread is interesting :)

Like some kind of dark matter?

JUST transparent to light and nothing more

The dark matter hypothesis may not actually be falsifiable, however that is not what physicists are working on. Physicists work on dark matter models, precisely defined theories which incorporate dark matter and these are falsifiable. For example roughly ten years ago models of weakly interacting particles (that is, particles that interact via the weak force) were very popular, but they have a influence on the abundance on nucleons in the early universe. Turning the argument around, one gets a interaction probability of the particles, and that part of the parameter space was excluded a few years ago. (You can of course try to mask the signal, but then you are generating different models, the original very simple models were falsified.)

You can observe it via gravitational lensing. For instance: http://scienceblogs.com/startswithabang/2011/04/20/how-gravi... .

We've seen instances already where gravitational lensing was explained through other means (very dim clusters, dust, and even camera issues), we also don't know if the space time on those scales is "smooth", if it isn't then light will not travel in a straight line which would account for the lensing effects.

The spacetime curvature, mass/energy, and gravity have quite a few interesting interactions as in gravity as a force, vs gravity as a feature of spacetime, vs potential other forms of gravity. We aren't even sure at this point if a curved spacetime without mass/energy that would generate that curve, would actually produce gravity or not since we aren't able to observe this nor do we have a complete theory of gravity.

Except we have good evidence that the universe is very close to flat, space-time wise.

A hypothesis that there's local unexplained curvature...Well that's really what dark matter is all about. What's curving it?

edit: Just ignore my comment, it's wrong.

(for posterity) My understanding is that spacetime being flat is akin to it being the surface of the universe's curves. Explain, expand, or correct me if that's wrong; I'm certainly no physicist.

> we also don't know if the space time on those scales is "smooth", if it isn't then light will not travel in a straight line which would account for the lensing effects.

Except space isn't perfectly smooth. It does curve, and that curvature is gravity. That's what general relativity is all about. If you want to postulate some way for the universe to curve absent observable mass, go ahead: you're doing nothing but restating the dark matter problem in a (minimally) different way.

> If you want to postulate some way for the universe to curve absent observable mass

The de Sitter vacuum is an exact solution to the Einstein Field Equations of General Relativity that does exactly that.

Our expanding universe will very likely closely approximate de Sitter space in the extreme far future.

There are many other vacuum solutions; it's normal in General Relativity to think of curvature as being a background laid down by hand (ideally approximating nature) and then perturbing that background with matter or probing it with test particles.

> Except space isn't perfectly smooth.

General Relativity is defined on a smooth (infinitely differentiable) manifold.

This is very important! https://en.wikipedia.org/wiki/Smoothness

There can be arbitrary curvature in the manifold, however, and unfortunately with the Einstein Field Equations you can pretty easily produce manifolds with curvature singularities, and those are manifestly not smooth (and that is one driver for research aiming to extend General Relativity).

But to the first order, your experience of gravitation near the surface of the Earth emerges from the very weak manifold curvature the planet and its microscopic components and their motions source. The nearest strong curvature is a couple of kiloparsecs away.

You can observe nothing but the fact that mass distorts light. Everything else is imagination.

We can observe the effect of gravitational lensing. The claim that dark matter causes it is part of the same claim that dark matter is responsible for galactic rotation curves looking the way they do. Thus it does not provide any further support for the dark matter hypothesis.

There are several ways that dark matter might exhibit detectable interactions. For example, it could undergo weak interactions with itself or with ordinary matter. (See WIMPs.) It could also undergo some antimatter-annihilation interaction, emitting e.g. a Z boson, which could decay to a detectable particle-antiparticle pair. We might be able to observe Z bosons from matter interactions decaying to dark matter pairs, which would manifest as an unexplained shortage of matter after a collision. We're still looking for any such interactions.

That only is if dark matter is indeed "WIMPS" or weakly interactive massive particles.

But then again there are multiple contenders for dark matter.

There is a (waning?) hope that dark matter interacts weakly with ordinary matter, which would make it detectable: https://en.wikipedia.org/wiki/Weakly_interacting_massive_par...