- #1

- 24,775

- 792

**Max Tegmark, Lee Smolin--polite conversation at Peter's**

interesting exchanges at Woit's blog

http://www.math.columbia.edu/~woit/wordpress/?p=310#comments

raising an important issue with the recent Tegmark et al, and Wilczek, papers.

links to those papers are here:

https://www.physicsforums.com/showthread.php?t=104167

These recent papers touched on multiverse ideas but failed to discuss CNS or justify the omission. Max Tegmark posted a message on Peter's

and the following ensued. Perhaps the most substantive is Smolin's post which is also the latest, so if you wish to cut to the chase just scroll down to the end of this post.

----quoting from Not Even Wrong---

Who Says:

December 17th, 2005 at 2:26 pm

Dear Max,

I thought it was a serious flaw in your paper that it failed to address Smolin’s cosmological natural selection proposal (CNS) and the accompanying arguments presented in

Scientific Alternatives to the Anthropic Principle

http://arxiv.org/abs/hep-th/0407213

—exerpt from abstract—

...We show however that it is still possible to make falsifiable predictions from theories of multiverses, if the ensemble predicted has certain properties specified here. An example of such a falsifiable multiverse theory is cosmological natural selection. It is reviewed here and it is argued that the theory remains unfalsified. But it is very vulnerable to falsification by current observations, which shows that it is a scientific theory...

—endquote—

Since the CNS proposal is the simplest and most directly testable multiverse model, and it has not yet been refuted by observation as far as I know. I think it is incumbent on you to describe CNS and to say why you think it can be ruled out. Or explain why it is not testable by current astronomical observations, if you believe it is not.

It might further the discussion if you did this.

Cheers,

;-)

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

woit Says:

December 17th, 2005 at 3:26 pm

Hi Max,

Thanks a lot for taking the time to write here, I’d certainly enjoy talking to you in person about these issues sometime. Here are some comments, I hope they address the questions you asked.

First, some context. My background is in mathematics and particle physics, not cosmology, and what motivates me is the idea of trying to find an improvement of the standard model, using new mathematical ideas. I think Wilczek would classify me as the sort who is trying to get somewhere by “pure thought”, something which he doesn’t believe will ever work. I’d claim that, given the lack of any new helpful experimental input, particle theorists don’t really have any choice at the moment except to go this admittedly much more difficult route. I’ve never found string theory a particularly appealing idea for unification (although it gives interesting insights into strongly coupled gauge theories and has led to some important new mathematics), and think the way it has completely dominated mathematically-minded particle theory for the last twenty years has been a disaster for the subject. By now it should be clear that it is simply a highly speculative idea that failed, and the field desperately needs to acknowledge this and move on to trying other things. The whole string theory landscape program, especially in its “anthropic” version, seems to me to be a retreat from the very idea of doing science, motivated by a refusal to admit failure. For more about this, read my recent review here of Susskind’s new book.

Like just about any particle theorist, I’ve spent some time looking at what is going on in cosmology and hoping it will provide some new insights into how to get beyond the standard model. So far it seems to me cosmology has provided some interesting hints, but unfortunately they’re no more than hints. We’re agreed that what needs to be done here is to find ways of confronting our mathematical models with the real world. When dealing with models that involve a statistical ensemble of universes and observables that are only probabilistically determined, sure, one has to take into account selection effects.

The scientific part of my critique of your paper was not that I thought any of it was incorrect. The major part of it, your anthropic calculation of the selection effect for the dark energy density was more complicated than I have the time or interest to follow. You end up with a probability distribution for the dark energy, with the observed value near the peak. What was unclear to me was exactly what conclusions you are claiming can be drawn from this. Ignoring the axion cosmology prior, it seems you’re just claiming that there’s an anthropic window in the dark energy density, and observations show we’re in the middle of it.

But, as a particle physicist, what I really want to know is what is left over when you remove the selection effect. Exactly what does your calculation says about the axion cosmology model? Have you provided any new evidence for it other than that it’s not obviously inconsistent? Can you rule out a flat dark energy density prior in favor of the axion cosmology one? Your paper didn’t seem to seriously address these kind of questions, or the more general question of when you can expect to get genuine, falsifiable predictions out of this kind of calculation.

That’s the scientific critique, there’s also a more general critique, one I made in my posting. Especially the early part of your paper seemed to me a heavily ideological push for the idea of anthropic determination of the standard model parameters, without anything to really back this up. Given the on-going disaster in particle theory these days due to the string theory landscape, I think this is really unhelpful.

Anyway, thanks again for writing in. If you’d like to write something more than a short comment in response, I’d be more than happy to put it up as a new posting, so people would be more likely to see it and comment on it if they wish.

Peter

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

Aaron Bergman Says:

December 17th, 2005 at 3:48 pm

For my part, I think probabilistic arguments across multiverses are nonsensical. Otherwise, I don’t see a satisfactory response to the doomsday argument. So, you can file me under (a).

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

Who Says:

December 17th, 2005 at 3:52 pm

Peter to Max: “If you’d like to write something more than a short comment in response, I’d be more than happy to put it up as a new posting, so people would be more likely to see it and comment on it if they wish.”

That would be great! Please consider doing this, Max!

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

Max Tegmark Says:

December 18th, 2005 at 10:29 am

Hi Peter,

Thanks for your thoughful response. I’m glad to hear that you didn’t feel any of it was incorrect.

I agree that it will be interesting of one can make stronger statements about the axion model - as described in the paper, this will require

improved astrophysics calculations that I hope somebody will make. You described the early part of the paper as a heavily ideological push for the idea of anthropic determination of the standard model parameters. The intent was rather to push for an open mindset on the issue, since we frankly don’t know how many parameters will ultimately be computable from others.

Regarding the CNS-critigue by “Who”: The pre-inflationary axion model is a complete physical theory, whereas Smolin’s “Cosmological natural selection” is not (we lack a mathematical description of how black holes spawn universes; see also http://arxiv.org/abs/hep-th/0407266) . Moreover, my guess is that the hypothesis is ruled out by the low observed fluctuation level (~1/10^5), since raising it would lead to more black holes.

Finally and most importantly, I’m glad that many of you disagree with my views! As argued in http://arxiv.org/abs/physics/0510188, I think that diversity in the physics community is more useful than an ideological monoculture, since it motivates physicists to tackle unsolved problems with a wide variety of approaches.

;-)

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

Who Says:

December 18th, 2005 at 12:51 pm

Hi Max,

personally I view your “low observed fluctuation level” not as a fundamental physical constant but more of an ad hoc result. It could be symptomatic of more basic parameters established prior to the beginning of expansion. But this is largely a matter of preference. It is personal opinion on your part (as you say) and equally on mine!

I would have expected you to mention CNS in your paper and to say why, in your opinion, it is probably ruled out. This would have given others a convenient opportunity to argue that it has not yet been ruled out.

There have been quite a few papers in the past year relating to the question of how black holes might spawn universes. Much of the work (QG modeling black hole collapse) was by people who reported in the Friday (14 October) session of Loops ‘05. The main topic was LQC, but the same people work on quantum models of gravitational collapse and, for what it’s worth, the bounce mathematics has turned out to look rather similar in both cases.

You say “we lack a mathematical description” and cite a 2004 paper by Leonard Susskind responding to Smolin, but, as I say, there has been quite a bit of mathematical description since then. I also do not believe Susskind’s rebuttal stands unchallenged.

Be that as it may, my point is that the CNS idea has not been disposed of and should have been discussed. If you think it is not falsifiable, or that it has already been falsified, then you should at least give your reasons.

In case you would like to look up the past year’s QG papers modeling gravitational collapse and bounce, I will post some arxiv numbers later. Some of the relevant authors, if you want to look up their recent work yourself, would be Abhay Ashtekar, Martin Bojowald, Viqar Husain, Oliver Winkler, Leonardo Modesto, Parampreet Singh.

To be fair, one should note that Smolin’s CNS proposal is testable without reference to any specific mathematical description of black hole collapse and bounce. The CNS conjecture is that some reproductive/evolutionary mechanism has fine-tuned the constants for black hole production. CNS challenges us to find even one fundamental constant (hope you will pardon me if I decline to view your “low observed fluctuation level” as a fundamental constant) which if it were better tuned would result in substantially greater black hole abundance.

That is something one can use to test—and possibly falsify—CNS, even before one has a complete mathematical theory of the conjunction of black hole and big bang events.

Thanks,

;-)

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

Lee Smolin Says:

December 18th, 2005 at 1:19 pm

Dear Max and Who

Thanks for mentioning the CNS proposal. The issue of the level of delta rho/rho-the fluctuation level-was discussed in detail and resolved in the first paper published on the model, Did the universe evolve?, Classical and Quantum Gravity 9 (1992) 173-191, summarized in Life of the Cosmos, ps 309-10.

Given that this was the first theory based on the “landscape*” I would hope that people would look at the literature before dismissing it.

The point is that in simple one field inflation models the fluctuation level is proportional to the inflaton self-coupling constant. However the number of e-foldings in inflation is related to (if I recall right) the inverse of the square root of the inflaton self-coupling. The result is that the cost of raising the self-coupling to produce primordial black holes or make more galaxies is that there are fewer e-foldings in inflation and the resulting universe is exponentially smaller. The result is that CNS predicts that the self-coupling should be as small as possible consisent with some steady level of black hole production. i.e. many more black holes are produced by a slow, but steady rate of black hole production in an exponentially larger universe than by a burst of primordial black hole production in an exponentially smaller universe.

This seems to characterize our universe where most black holes are made as supernova remnants. It also leads to a prediction, which is that inflation is governed by one parameter and not by a complicated potential with more than one parameter in which the fluctuation amplitude and number of e-foldings would be independent. So were there any evidence for an inflaton potential governed by more than one parameter the theory would be ruled out. I made this prediction in 1992 and so far it has held up.

I can endorse WHO’s comments: recently we have very good detailed support from the quantum theory of gravity that black hole and cosmological sinigularities bounce. There is not yet a detailed description of the parameters mutate in a bounce, but this is certainly plausible given current views of the “landscape”.

Max, if these were your main objection to CNS, would you now agree that the theory is viable?

I would go further, CNS is the only landscape theory proposed so far that makes falsifiable predictions, Should this not make it the leading

candidate for an explanation of the choices of parameters in the case that the landscape is real?

Thanks,

Lee

*and indeed the origin of the term, which comes from “fitness landscape”.

----------end quote-------

Last edited by a moderator: