Thread by @seanmcarroll

Dark matter exists -- an occasional reminder.

The first evidence for dark matter came from the dynamics of galaxies and clusters of galaxies. But these days that is not our *best* evidence. 1/n

For galaxies and clusters, you can imagine modifying gravity instead of positing dark matter. Every physicist and astronomer knows this. It is not a radical new idea. A number of people have built explicit models, and hundreds of people have thought about the possibility.

And it could potentially work, because in both cases you need only change the strength of gravity as a function of distance, not the direction of the gravitational force. Not natural from a field-theory point of view, but worth contemplating.

But it's the 21st century now. The best evidence for dark matter comes from anisotropies in the cosmic microwave background (CMB), and there are several independent lines of evidence.
medium.com/starts-with-a-…

We even have maps of where the dark matter is, from gravitational lensing. Check out this beauty!
en.wikipedia.org/wiki/Dark_matt…

Here's the power spectrum of CMB anisotropies. The horizontal axis is angular scale, with smaller scales to the right; the vertical axis is amount of temperature fluctuation. There's a big peak at 1 degree, then smaller peaks dying away.

Kind of like the ringing of a bell, gradually fading. Because that's what it is! These peaks come from acoustic oscillations of matter in the early universe, which fade away over time due to dissipation.

But wait! Wouldn't we expect the peaks to decline uniformly? It looks like the 1st peak is huge, then the 2nd and 3rd peaks are the same, before dropping down. Almost like odd-numbered peaks are being boosted with respect to even-numbered ones.

That's exactly right. Ordinary matter oscillates in waves, but *dark matter* just accumulates in over-dense regions. Dark matter is dark, so it doesn't heat up and bounce back in a pressure wave. Fluctuations in dark matter simply grow.

So the dark-matter theory makes an unambiguous prediction: odd-numbered peaks in the CMB, where ordinary and dark matter march together, should be boosted. Even-numbered peaks, where they are out of step, should be slightly suppressed.

This is exactly what we see.

Without dark matter, you expect the peaks to decline uniformly. This was literally predicted in modified-gravity models before the peaks were observed.
arxiv.org/abs/astro-ph/0…

See the dark curve, with uniformly-fading peaks? That's the modified-gravity model without dark matter. See the data? They don't fit it. It's great to have falsifiable theories, as long as you're willing to accept that sometimes they get falsified.

You can keep trying. It's a free country. But the CMB tells us unambiguously that there is gravity pointing somewhere other than where the ordinary matter is. To explain this you need to have extra fields or degrees of freedom of some sort. We call such fields "dark matter."

Roughly speaking, no alternative-gravity approach fits all of the current constraints at once. To even come close, you need to concoct completely crazy and unnatural models, and still you come up short.
arxiv.org/abs/2007.00555

The lesson is: if someone claims to have a theory that does away with the need for dark matter, but they only consider rotation curves of galaxies, you should feel free to ignore them. Or if you're feeling charitable, tell them to check back with a prediction for the CMB.

This conclusion is what lawyers call a "statement against interest." I would love it if gravity were modified! I've tried to make it work myself. But science says we should listen to the data and update our credences accordingly.

Of course gravity still could be modified on large scales. That would be fun. But you *also* need dark matter, even if gravity is modified. Maybe that's the most fun of all? n/n