Sad disconnect at Cosmic Variance blog-how not to discuss

[marcus]

Sad disconnect at Cosmic Variance blog---how not to discuss

Cosmic Variance blog is known for lively discussion but I just noticed a sad disconnect where the key point Lee Smolin was making just slipped by in a "yes it is---no it isn't" kind of way without any reasoned counterargument.

It is really crucial that the CNS hypothesis (cosmic natural selection) is TESTABLE, and in fact could be falsified at any moment by ongoing observation.

and that testability is not some future pie-in-sky promise, it has been ongoing since mid 1990s.

CNS predicts you will never find a neutron star with mass greater than 1.7 solar.

(it also predicts other stuff but that's enough to show falsifiability)

================

this is a general astronomy thing that ties together neutron stars, black holes, and basic physical constants such as the cosmological constant, particle masses and soforth.

================
anyway I found this exchange kind of depressing. Smolin explained why CNS was testable, and Bergman (a young string physicist) said "no it isnt" and he didnt give any reasons for that kind of rude contradiction!

and the discussion went on, and he still didn't give reasons. It was sad.

That was a February thread, and long dead, but maybe we can redress the balance, if only in spirit, by discussing CNS some.
Here is the Cosmic Variance sequence of comment:
http://cosmicvariance.com/2007/02/21/oos-and-bbs/#comment-205299

 

[marcus]

==quote==
Lee Smolin on Feb 22nd, 2007 at 11:49 am

Aaron and Elliot,

That CNS is falsifiable is based on genuine predictions. Here are two:

1) The upper mass limit of neutron stars is less than 1.7 solar masses, due to their cores being kaon condensates.
2) Inflation, if true, is due to a potential governed by a single parameter, so that there is a relation between the number of e-foldings and delta rho/rho.

Both these predictions were published in 92, both could easily have been falsified in the years since, and still could easily be. (See hep-th/0612185 for the details, present status and references.)

Eternal inflation + AP have made no such explicit predictions...
...
Thanks,

Lee
=====
Aaron Bergman on Feb 22nd, 2007 at 12:05 pm

That CNS is falsifiable is based on genuine predictions.

Just because you assert it doesn’t make it so.

...
...
==endquote==

and that was all. They went on talking about other things, others of Smolin's points were discussed, but this main one was simply denied and ignored.

In case anyone wants to check it out, the paper that Smolin referred to in his comment is available online
http://arxiv.org/hep-th/0612185
The status of cosmological natural selection
Lee Smolin
(Submitted on 18 Dec 2006)

"Abstract: The problem of making predictions from theories that have landscapes of possible low energy parameters is reviewed. Conditions for such a theory to yield falsifiable predictions for doable experiments are given. It is shown that the hypothesis of cosmological natural selection satisfies these conditions, thus showing that it is possible to continue to do physics on a landscape without invoking the anthropic principle... A recent criticism of cosmological natural selection made by Vilenkin in hep-th/0610051 is discussed. It is shown to rely on assumptions about both the infrared and ultraviolet behavior of quantum gravity that are very unlikely to be true."

 

[Chronos]

I've followed the CNS conjecture for awhile. It still appears speculative. Neutron stars that appear plausibly 'overweight' [by CNS criteria] remain on the table. I think some might be quark stars, but, that is a horse of a different color. Perhaps a bounce is involved.

 

[marcus]

I've followed the CNS conjecture for awhile. It still appears speculative.

It's a clearly stated testable scientific statement. It has survived testing since 1992 or whenever the challenge to disprove it was first nailed to the mast.

One obvious piece of evidence that it has NOT so far been disproven is that Andrei Vilenkin (a prominent string-friendly multiverse thinker http://en.wikipedia.org/wiki/Alexander_Vilenkin) tried very hard to falsify it in a 2006 paper, and failed.

Smolin has repeatedly said that if anyone ever found a neutron star with mass over 1.7 solar that the hypothesis would be killed. AFAIK the errorbars are too wide on the observations you mention to consider them as reliable disproof.

If those observations were reliable, Vilenkin could simply have cited them. But he didn't, he started talking about quantum fluctuations 50 billion years in the future and other blue sky fantasy.

Neutron stars that appear plausibly 'overweight' [by CNS criteria] remain on the table...

That is what I think is not the case. You could give me URL for the reports of these "overweight" ones. I could check, but I think it is not unlikely that

1. the reports are old and have not been confirmed by other observers
2. the errorbars are too wide for the results to be used as credible disproof
3. there have been no more such reports (of NS massiver than 1.7 solar)

I'd be interested if what I guess turned out not to be right, and if, for instance if you have NEW, narrow-errorbar, observations of massiver neutron stars.
We could email Vilenkin and give him a tip on how he could disprove this hypothesis which he would clearly love to do.

 

[Chronos]

Much depends on how stiff the equation of state is for condensed matter, at least theoretically. Smolin's CNS conjecture relies on a soft EoS prediction [by Brown IIRC] that limits maximum neutron star to ~1.5 solar. There are, however, differing views on this. Some researchers endorse a stiffer EoS permitting masses up to ~2.2 solar. Observational evidence is, of course, the real test. There are candidate NS's that appear to approach the higher limit of 2.2 solar. Most recent notables I am aware of include:

PSR J0751+1807 [2.1 +/- 0.2 solar]
as reported by Nice, et al: http://www.arxiv.org/abs/astro-ph/0508050

and

EXO 0748-676 [2.1 +/- 0.28 solar]
as reported by Ozel: http://arxiv.org/abs/astro-ph/0605106

The error bars are sufficiently small to qualify as challengers to the CNS conjecture, but EoS issues remain and are still actively discussed - eg :

http://arxiv.org/abs/nucl-th/0612021
Neutron stars and quark matter
Authors: Gordon Baym (University of Illinois)
Recent observations of neutron star masses close to the maximum predicted by nucleonic equations of state begin to challenge our understanding of dense matter in neutron stars, and constrain the possible presence of quark matter in their deep interiors.

 

[marcus]

this seems like an open-and-shut case against CNS
if one simply takes Smolin at his own word.
he challenged people to find a NS massiver than 1.7 solar and it looks like they have.

that's curious. I must be missing some point of the argument. why whould Vilenkin not have simply pointed this out

 

[Chronos]

IMO, they do not agree on how 'soft' the EoS might be. Smolin heavily relied on a soft EoS [Brown] to make his case. He assumes a linear [2D] model, again, IMO. Baym raises some very good points to consider, including why binary neutron star masses cluster so close to the Chandy limit [a recoil thing].

 

[Ontoplankton]

and that was all. They went on talking about other things, others of Smolin's points were discussed, but this main one was simply denied and ignored.

No, I don't think it was ignored. Aaron Bergman's point in the CV comment thread, as I understand it, is this: in order to say that "CNS predicts the universe is near a local optimum for black hole generation", you first need to figure out what constitutes a "prediction" if there are actually many different universes (which is true both with and without CNS), and to do so, you need to deal with all the issues of measures and regularization and principles of mediocrity that Smolin thinks are deal-breakers, even if your distribution of universes is strongly (but not perfectly) peaked compared to what the distribution would be without CNS.

 

[Ontoplankton]

Anyway, Smolin's arguments confuse me... why shouldn't the set of universes in CNS converge to some static distribution (probably peaked around the global fitness maximum)? I think the biological evolution analogy confuses more than it clarifies (for example, isn't CNS supposed to go on forever?), and Smolin's CNS views should be able to stand on their own merits if they're sound.

 

[Chronos]

Good point, Onto. If CNS is correct, it suggests our observable universe should be smoother than observed. And possibly too smooth to permit observers [such as ourselves] to arise. This is not an appeal to any 'anthropic' hypotheses - which tend to ignore causality and entropy. I think quantum information theory is the right approach. It is largely background independent.