Universe not accidental: Is this Steinhardt statement "rather pathetic"? If so, why?

Chalnoth

No. What you can do is test the predictions of a model of the electron. Many of the models which include a multiverse are testable in the exact same way.

This is the normal way that science is done. Every single prediction of a theory cannot possibly be verified. So we test what we can, and the more ways that we can test a theory, the more confident we are in its overall correctness.

With models of the universe that include a multiverse, even if it turns out to be impossible to observe the multiverse component of those models directly, there are most definitely other aspects of those models that can be tested. And even just the standard model plus unitary quantum mechanics makes a multiverse nearly certain.

And there are many scientists working on creative ways to test for the impact of various multiverse ideas more directly. I don't hold much hope that this will prove to be fruitful. But it does, at least, avoid the useless work of trying to avoid such multiverse hypotheses as a matter of principle.

Chalnoth

Why is it wasteful? What does that even mean?

It's exactly this knee-jerk reaction that I have such a big problem with. No matter which way you slice it, multiverse theories require fewer assumptions. It is easier for a theory to be prolific than not. Demanding that the default hypothesis be proven is ludicrous. The default hypothesis needs to be disproven, not proven.

bapowell

But you still haven't said why you've chosen to accept many-worlds over other interpretations, outside of from what I can tell are purely aesthetic reasons. Unless you have some Bayesian prior on your model space that you aren't telling anyone about.

EDIT: This discussion has actually been helpful because I understand now what we disagree about. It's not so much about the implied correctness of untested predictions or axioms. It's about your staunch acceptance of a version of quantum mechanics based solely on its relative simplicity that I'm not quite in agreement with. Occam's razor is an indispensable guide for selecting the most favored model out of a bunch; but it does not exclude those models it doesn't select.

skydivephil

I think your example above goes back to what I was saying, that really there is a grey area between what is science and not science.
So lets suppose theory A predicts B, C and D. As you say B and C have veen verified. Should we accept D as true without verififcation? I think I would agree that we should not accept it to the same extent as we accept B and C. However neither should we classify it as the same level of non science as something silly like creationism.

Take gravity waves for example, although there has been indirect evidence from binary pulsars there has never been a direct detection despite LIGO being operational for something like 10 years (?). Now lets suppose the pulsar observation had not been made, what should we say about gravity waves? Well I think they would be in this grey area, they are precited by GR and Gr si well verified. Not somehting silly like creastionsim, but neither somehting verified such as time dilation.

Chalnoth

1. Many-worlds makes more predictions as to how the universe behaves. Specifically, it makes definite predictions about the nature of collapse. Most other interpretations sweep the nature of collapse under a rug and make no predictions at all about it. Considering that the nature of collapse is becoming more and more important as we try to take advantage of quantum mechanics for computing, this really is an essential feature and can no longer be considered up to personal choice (not that the nature of reality ever was up to personal choice).
2. Many-worlds makes the fewest assumptions. I don't see how there can possibly be any argument about this point.

Chalnoth

This is kind of off-topic, but I just wanted to point out that the upgrade to advanced LIGO is now under construction, and is expected to be up and running somewhere around 2015. The sensitivity is expected to be great enough that it will be guaranteed to detect gravity waves from a number of known sources, barring some unforseen systematic errors.

martinbn

Why? I mean, the spontaneous symmetry breaking does not say that there are universes where each vacuum state is realized. Or am I wrong? I just don't see how spontaneous symmetry breaking is related to many universes!

marcus

I think this is mistaken. (BTW it's a claim you already were asserting in post#2)
The way I slice it, the appropriate question to be asking at this point is how did the "big bang" come about.
How did the expansion begin and why does it have the observed characteristics?

Bounce theories of how this happened seem to depend on fewer assumptions. They simply have the U extend back further in time, and be in a contracting mode. No different laws from those operating now.

Here's a current survey that briefly describes various approaches to understanding "big bang". It is an invited review for Modern Physics Letters:
http://arxiv.org/abs/1201.4543
Not all the approaches sketched here are "multiverse" and it seems to me some are simpler (as well as more testable.) So I don't think your claim stands.

bapowell

OK. So as I understand it you favor many-worlds due to its parsimony and predictive strength relative to alternatives. This is precisely the way one would go about weighing the relative merits of competing statistical models. But at the end of the day, we are not considering statistical models; we are interpreting the candidate theories as representing objective physical reality. The more complicated, less predictive model may well be correct! Again, these considerations suggest a preference, statistically speaking, for the simpler model. I do not, however, think this is adequate to furnish the kind of certainty and correspondence to objective reality that you are advocating.

marcus

Brian, hope it's OK to interject. I think the main agenda here is to resolve the cosmological singularity and provide either for inflation or for a substitute mechanism.

Renaldi, in the invited review article I mentioned, mentions string cosmology, loop cosmology, Horava, Jacobson's Einstein-aether, and various others. He gives a brief historical account of the earlier attempts which preceded and led up to these approaches---particularly the first two. In one form or another, most of these involve a bounce.

You might want to glance at the relevant section, which is just 2 pages long. It is section 2 "Lines of Research" and begins on page 2. Here's an excerpt.
==Rinaldi review article http://arxiv.org/abs/1201.4543 page 3==
There are several other models that offer alternatives to the direct quantization of gravity. Recently, Horava has proposed a power-counting renormalizable theory of gravity, based on an anisotropic scaling at high energy ²⁰. Essentially, the fundamental hypothesis is that time and space do not scale in the same way, according to the scheme t → b^zt, xⁱ → bxⁱ, where z is called critical (Lifschitz) exponent and b is an arbitrary constant. By adding higher spatial curvature terms to the standard Einstein-Hilbert action, one can construct a model where, at high energy z ≥ 3, which makes the theory power-counting renormalizable, while at low energy z = 1. Local Lorentz invariance is preserved in the infrared (IR), and it is broken in the UV. The original formulation of this model suffered from un unwanted ghost scalar field, that persisted also in the IR ^21,22. To remove this anomalous degree of freedom one needs to add new terms in the action, that are basically formed by combination of a vector field, orthogonal to constant time surfaces, and its derivatives ²³. In this form, the Hoˇrava-Lifschitz theory becomes very similar to the “Einstein-aether” theory proposed by Jacobson many years before as a vector-tensor theory of gravity 24. Both theories offer non-singular solution to the cosmological equations ^25,26 and the horizon problem is solved without recurring to inflation ²⁷.
==endquote==

Chalnoth

Well, there are two ways to look at this. One is that inflation strongly predicts that this and potentially other symmetry breaking events are only local effects, and that they will occur differently in far-away regions. The second is that whatever physical model you have for our early universe, it is highly unlikely that that physical model is a one-off event.

Chalnoth

Two points. First, I find these theories highly unlikely, due to the apparent reversal of entropy at the bounce. Second, even if this isn't a problem, there's still no reason whatsoever to believe it's a one-off event. You still have to assume it's a one-off event separately from the physical model.

A tangential point that I'd make is that the way high-energy physics is progressing, it is seeming increasingly unlikely that you could ever achieve the conditions for life with a model that only started one region of space-time with one set of physical laws.

marcus

Just to be clear, I am addressing the claim of simplicity or fewer assumptions that you made in post #2. It simply is not true.

It is your opinion that the 4 or 5 quantum cosmology approaches discussed briefly in that invited review article are "highly unlikely". Since it happens they are all bounce type. Your opinion could be right or wrong---this is not relevant.

You cannot rightly say that multiverse scenarios require fewer assumptions than other theories being studied that resolve the cosmo singularity.

I don't think you even know what the possible alternative theories are, so it is ridiculous to claim that multiverse theories need fewer assumptions than all the others.

Logically I think what you need to say is that in your opinion the approaches Renaldi covers in his review article (string cosmology, loop, Horava, Einstein-aether...) are "highly unlikely" and if these approaches are excluded then multiverse requires fewer assumptions than whatever theories you know of that resolve the initial singularity.

Fuzzy Logic

I think there is multiple meanings of the word multiverse being used here.
The problem I have with Steinahrdt's statement is the basis of his argument, in that he beleives the universe was not accidental. Other than that statement. I agree with the rest, in so much as to say that it's equally possible that galaxies are the largest structures or that there is possibly more than one isolated 'universe' existing at the same time.

I agree with Chalnoth in that the universe is not a singular event. More time and a bounce scenario can explain the astronomical probabilities for the conditions of life (as we know it!) just as well as more space.

Chalnoth

Except it is true. There is no possible way to have a unique universe without making that an extra, specific assumption in the theory. This is simply because any physical model of the universe which doesn't explicitly mention other regions of space-time also won't explicitly exclude them. It makes the theory more complex to exclude them. Always.

bapowell

But theories are never ruled out on account of their complexity. This is where I disagree with your reasoning. You are essentially performing a Bayesian model selection on your space of competing theories. They all satisfy the data equally well, however, some have additional structure than others that make them either less predictive, more complex, or both. The Bayesian evidence disfavors these models, but it does not exclude them! This is an incorrect interpretation of the statistical method.

marcus

Chalnoth you are not making sense. A bounce cosmology theory does not have any statement in it which says some other bounce in some completely separate realm doesn't exist. It is just a theory whereby the universe that we know and observe, with its physical laws, contracted and rebounded (according to a quantum law of gravity to be tested) resulting in what we now see.

A scientific theory is supposed to explain observations and make testable predictions, this is what we apply the Occam criterion of simplicity to, and fit to data.
A bounce cosmology has no place for some grandiose philosophical speculation about some other completely disconnected realm. Makes no assertion either way.

this is how to get a really simple resolution of the initial singularity.