death of the universe?!

Chalnoth

What you'd be looking for is areas where there is actually less plasma, and instead is the occasional matter/anti-matter annihilation. This wouldn't be an effect on the scale of the tiny temperature anisotropies. This would be significant compared to the overall 2.7K temperature of the CMB, which is 10,000 times brighter, and much brighter than any other source of light in the sky.

This isn't at all true. Instruments like WMAP and Planck specifically measure relative radiation intensity at different places in the sky at specific wavelengths. Most such instruments aren't very good at measuring the absolute brightness directly: this is inferred through, for example, the dipole in the CMB induced by the motion of the Earth around the Sun. But they are extremely good at measuring relative brightness in different locations. That's what they're built for.

The large error bars you see here are not measurement errors alone. They are measurement errors plus cosmic variance. The measurement errors on low multipoles with WMAP are extremely tiny (typically much smaller than the errors on high multipoles). The cosmic variance errors, however, are based upon our theoretical model of the physics that produce the CMB, a model which makes a probabilistic prediction on that just isn't very precise at low multipoles.

In essence, the theoretical prediction of how the CMB should look given a set of cosmological parameters (e.g. normal matter density, dark matter density, dark energy density, etc.) is not a specific value, but a variance. Since the prediction is only the variance, and since low multipoles have a small number of independent components (e.g. 5 components for ell=2), the measured variance can vary dramatically from the theory variance without being inconsistent with the theory.

At higher multipoles, where you have a lot of independent components, the measured variance and the theoretical variance have to match much more closely for the theory to agree with experiment.

Those aren't particles, though. Those are large collections of particles localized at a specific point.

Anyway, we'll see. But these models seem to me to be highly contrived and thus highly unlikely.

twofish-quant

This is not terribly convincing without even rough numbers. If you presume that matter and anti-matter repel each other, then you have several hundred thousand years for the matter and anti-matter to separate, and you can make the matter/anti-matter annihilation end up as low as you want. That gets rid of the non-thermal spectrum.

I did a quick calculation of gamma ray flux and to make the numbers work, you have to assume a suppression factor of 10^-2 or 10^-3. That's not a crazy number if matter and anti-matter repel.

At that point you'd have much less plasma at the domain walls, but the temperature would have time to thermalize at which point that you'd have a thermal spectrum and no temperature anisotropy.

Right. But if you have domain walls, then the relative brightness over a large chunk of sky is likely to be the same.

Which means that if there is something funny happening at low multipoles, you aren't going to see it.

Also, because of gamma ray flux, I doubt that we are missing anti-matter. However, getting to what the observations show or don't show is interesting because we could be missing something else. Cosmic strings or GUT monopoles would produce similar domain wall effects. For that matter, if you have a model of CMB at low monopoles, you might be able to use it to map nearby voids.

Right, but at high multipoles everything goes thermal so Dirac-Milne gives you the same basic spectrum.

But the topological defect mechanism as far as I can tell could work for the inflaton. Why do we think the inflaton is a massive particle? It's because we need inflation to happen at a specific time and having a massive particle makes the phase transition happen at the right time. Well, what if you have collections of small particles?

Saying that something is unlikely presumes a meta-theory. One problem with meta-theories is that whether something is contrived or not is a matter of taste. One reason Dirac-Milne is interesting is that it seems less contrived than the standard model, but this is a matter of taste, and the problem with aesthetic arguments is that if someone says it "looks contrived" and you disagree, there's no way of easily resolving the argument.

The trouble with "aesthetic arguments" is that all of our standard models are highly contrived. They end up highly contrived because reality is complicated and you have to do messy things to make the models fit reality. For things that we have lots of observations for, it's relatively easy to figure out what those messy things are. For stuff that we don't, it's not.

So we need more data, but then we have to ask what data do we need. It's not a matter of "wait and see" and "wait and see what?"

Chalnoth

This doesn't help, because it would still be much dimmer at the domain walls, which I would be willing to bet would be glaringly obvious in the CMB spectrum. Specifically, it would be glaringly obvious in the spectrum of the anisotropies (because instead of differences in temperature causing the anisotropies, differences in density would cause some of them, which would lead to different spectral effects).

Right, but domain walls would affect multipoles on many scales, because they are linear features.

Why? The optical thickness of the CMB washes out features at high multipoles overall, but the effect of the domain walls should be visible at all scales relative to the CMB anisotropies (which are also washed out at high multipoles due to this effect).

The inflaton is typically modeled as a field, with the quanta of that field being inflatons. I'm not sure a field of solitons makes sense.

I would be willing to bet that Dirac-Milne simply cannot work on purely empirical grounds, just given our current observations of the CMB, regardless of any arguments regarding simplicity.

As for simplicity, however, there are reasonably good measures of simplicity, such as the number of parameters required to describe the model. If a model requires more parameters to describe it, it sure as heck had better explain a lot more experimental evidence than the competing model, or else it's most likely wrong. Even though it's not possible to prove that this is a good way of doing things, and even though there are sometimes arguments about just how simple or complex various theories are, it seems to be a pretty good heuristic that has worked rather well in the past. And there are some rather rough probabilistic justifications for it that at least seem reasonable.

Yes, it is a matter of wait and see, because it takes an overwhelmingly-compelling theory to push people to base new experiments about it.

twofish-quant

A lot depends on the geometry of the domain walls, and on the processing that people do to get the multipoles. If the thickness of the domain walls are large compares to the features that people care about, then the only thing in the higher order multipoles are going to be harmonics and it's not hard for those to get lost.

One thing is that if some says "yes I've actually put in domain walls" here is what they look like, that would convince me, but I think that the Dirac-Milne have put enough of a case that I don't think that it's valid to dismiss their challenges without some numbers.

What if the inflaton is a soliton? The reason you need a high mass particle is so that you get the phase transition at the right time. You can have the inflationary particle be relatively low mass but the phase transition happen because of a soliton.

Also, this is a different argument than Dirac-Milne.

It's not that one is willing to bet but how much. I'd be willing to bet US $25K-$50K that Dirac-Milne is wrong. I wouldn't bet my life on it. As far as primordial baryongenesis. I'd be willing to bet several hundred dollars that primordial baryon number is irrelevant, but I wouldn't bet any more than that.

Or one weird observation. All that has to have happen to have people take Dirac-Milne seriously is to drop some anti-protons and watch them go up.

Chalnoth

It's up to them to put forward their case, not the rest of us to disprove it. And yes, I really think that the domain walls would produce brightness anisotropies that are much, much larger than the temperature anisotropies we see.

Well, right, and there are groups that are trying. The problem is that it's really, really hard given that gravity is some 40 orders of magnitude weaker than electromagnetism, so that the electric charges of the anti-protons tend to react far more strongly than do their masses.