Steinhardt's crit of inflation (PIRSA video)

[marcus]

Paul Steinhardt (Princeton cosmology expert) just gave a cogent talk at Perimeter Institute criticizing inflation.
http://pirsa.org/11070029/
The talk clearly lays out the problems with expansion cosmology that inflation was invented to address and again very clearly explains the ways it is unsatisfactory as a solution.
He also explores one or more alternatives, partly for comparison (one of them just as bad but in a different way.)

He does a good job communicating, I think. Not all that hard to understand most of the talk.

It was given at a conference/workshop on this kind of thing, so the audience is experts---but it is still not all that technical or mathy. See what you think.

The conference was on Challenges for Early Universe Cosmology, and all the talks are here:
http://pirsa.org/C11008
Steinhardt's talk was titled Meeting the Challenges.

Roger Penrose also gave one of the talks, and it is here:
http://pirsa.org/11070010/
He also dispenses with the need for inflation.
Penrose's title is
Conformal Cyclic Cosmology: Some Striking New Observational Support
Penrose has 125 slides, and the "new observational support" is not reported until after slide #95. So if you are interested in what new evidence may have accumulated you can scroll quickly down to near the end. It looks like some of it is new, though I can't be sure.

 

[bcrowell]

Steinhardt wrote a recent article in Scientific American about this topic: "The Inflation Debate," Sci Am, Apr 2011, p. 38. I summarized the article here: https://www.physicsforums.com/showpost.php?p=3214329&postcount=4 Of course this is just my quick, non-expert summary of a popular-level article.

 

[marcus]

I think it is part and parcel of the same line of reasoning. I remember that thread. People who read it will be better prepared to watch the conference version.

Steinhardt is not alone. He mentioned a talk at the same conference by Ben Freivogel. You may know of other instances. Here is the link to Freivogel's talk. I haven't watched it but gather it is partially supportive of Steinhardt's stand. It's nice to see a bunch of experts get together and call routinely accepted stuff into question, whether they are right or wrong at least they are putting the ideas on trial.

 

[Chronos]

Even some of the well established theories, like GR and QM, are not above suspicion. We know that GR and QM do not play well together and one or both [I vote both] are at best incomplete.

 

[marcus]

Steinhardt wrote a recent article in Scientific American about this topic: "The Inflation Debate," Sci Am, Apr 2011, p. 38. I summarized the article here: https://www.physicsforums.com/showpost.php?p=3214329&postcount=4 Of course this is just my quick, non-expert summary of a popular-level article.

Thanks for pointing that out, I think it is a good summary. I'd like to quote your post and use it as a basis for discussion, if people want to comment--also to call attention to the PIRSA video, which is well worth watching, I think.

The article consists of two main sections: the case for inflation and the case against.

The case for:

There are lots of fields believed to exist (but not necessarily the ones in the standard model?) that should have acted as inflaton fields. Solves the fine-tuning problems related to observed flatness and uniformity. Makes successful predictions, e.g., scale-invariance of structure.

The case against:

Although it gets rid of some fine-tuning problems, it creates others. E.g., fine-tuning is required to keep temperature variations from being too big. The anthropic principle doesn't help, because larger temperature variations might have made the universe *more* hospitable to life. Penrose argues that obtaining a flat universe with no inflation is higher in probability than obtaining one with inflation.

Eternal inflation robs the theory of its predictive value. Attempts to fix this by applying a measure don't work, because the choice of measure is ad hoc, has no explanatory value, is itself in need of explanation.

He has a quantum-mechanical model which he thinks may work better than inflation.

He refers to some popular-level books plus this paper: http://arxiv.org/abs/hep-th/0609095

 

[bcrowell]

Here are some notes I took from the talk.

Inflation makes no predictions without strong priors about initial conditions or measure.
Inflation is highly unlikely, or of indeterminate likelihood, based on conservative measures.

• ...unless you make certain assumptions
• His opinion is that the number of assumptions is too great.
• Need to find a theoretical way to reduce number of assumptions, or else find new experimental tests.
• Discussion of possible speculative ideas, such as infinite phase space.

Why continue to think only about inflation? Why not alternatives?
Straw-man theory: the original big bang model.

• By Penrose's entropy argument, is more likely than inflation.
• Doesn't require additional degrees of freedom like inflaton.

How might we get out of this predicament?
Inflation rewards "procrastination."

• We made a profound error in judgment in the early days of inflation.
• Didn't take proper accont of how qm and inflation mix.
• A hump in the mexican-hat potential has low-probability tails as it rolls down toward the minimum.
• Inflation rewards "rogue" regions, which are regions in which inflation took an unusually long amount of time to end.
• "anything that can happen will happen: and it will happen an infinite number of times" - Guth 2000

The volume measure fails.

• Predictions will depend entirely on the measure principle -- not inflation, and then we have to consider how a similar measure applied to other theories compares in predictive power.
• Suppose that the big bang could happen many times in separated regions. How would we regularize it?

Penrose: Theoretical max entropy of universe is 10^120, present universe is 10^90, but inflationary big bang is even lower.
Liouville argument

• Gibbons and Turok, 2006
• Turok 2011
• Extrapolate backward. What fraction pass through >60 e-folds of inflation.
• 14:16 "Lenny's janitor."
• Liouville analysis disfavors strong attractors, which we normally consider desirable in a theory.
• In homogeneous case, inflation is favored...just not a lot (<< 60 e-folds)
• Time-reversing an attractor gives a repeller.
• Is the Liouville argument stronger than Penrose's entropy argument?
• Penrose's entropy argument is that if we were going to guess initial conditions, we should guess a typical initial state.
• Standard assumption has been that it's natural to assume KE ~ PE as an initial condition.
• But according to L argument, KE >>> PE much more likely initial condition than KE ~ PE.

23:20 How to fix the problems?

• Smooth and flatten while penalizing rogue regions. $H_{smoothing}<H_o<<H_{after}$
• This means H has to grow, but according to GR, this means you either have to violate the null energy condition or you have to have some kind of singular bounce.
• $\dot{H}=(-3/2)(1+w)H^2$
• Two possibilities: (1) H>0, smooth while expanding, (2) smooth while contracting -- then need information-preserving bounce.

26:56 Contraction has some immediate advantages:

• no causality problem; it's easy to make every part of the universe be in causal contact with every other part
• no flatness problem; for the same reasons that the Friedmann eqns predict that curvature increases with expansion, they predict that it decreases with contraction
• when you look at energy densities and how they scale with contraction, anisotropy grows fastest
• this leads to an anisotropy problem, but it's easy to design high-w fields that take over at high densities and fix this
• avoids problem of rogue regions
• other adantages: higher-entropy initial conditions; possible explanation of the size of the cosmological constant
• picture in which some regions bounce, others don't

38:24 bounce must be unitary/analytic, smoothly matches conditions from before to after

• regions that bounce need not be generic
• sieve out regions with large geometrical entropy (they form black holes)
• passes regions with tiny geometrical entropy and high matter entropy

41:46 generic predictions

• primordial gravity waves 10^50 times weaker than in inflation
• non-gaussian perturbations produced by the high-w field
• either of these is a prediction that is likely to be testable in the foreseeable future, could support or disprove the theory
• potential with two minima; contraction amplifies oscillations until field spills across barrier

 

[bcrowell]

Here are a few impressions I had of the talk. Take these with a grain of salt, since there was a lot of material in the talk that I didn't understand.

He has a more vigorously negative spin on inflation in this than he did in the Sci Am article. In the article, he gave inflation at least some credit for making some successful predictions. In this talk, he says that there are so many assumptions that have to be made to get inflation to work that it essentially has no predictive value.

It's a shame that there was no recorded Q&A period. As a nonspecialist, it leaves me with no real impression of how people in the field see Steinhardt's ideas.

He's trying to sell cyclic models in which there is a contraction and a bounce. He lists a lot of good features of such models. One such model is Penrose's CCC, and although he gives credit to Penrose for a lot of important ideas, he shows no enthusiasm for CCC per se.

His message is cyclic=good, inflation=bad, and he has great hopes that observations will soon confirm that cyclic is what happened. The impression I get is that all such models are failures and none of them make any firm connection with observation.

He compares with the "straw-man" model in which there are multiple big bangs in disconnected regions. I assume this means regions that are not even on the same connected piece of a manifold? This seems silly to me, since the different big bangs would not even be observable from our unverse, so it doesn't sound like real science; but maybe this is his way of pointing out that the contortions needed in order to make inflation work are also not real science...?

 

[marcus]

Here are some notes I took from the talk.

Inflation makes no predictions without strong priors about initial conditions or measure.
Inflation is highly unlikely, or of indeterminate likelihood, based on conservative measures.

• ...unless you make certain assumptions
• His opinion is that the number of assumptions is too great.
• Need to find a theoretical way to reduce number of assumptions, or else find new experimental tests.
• Discussion of possible speculative ideas, such as infinite phase space.

Why continue to think only about inflation? Why not alternatives?
Straw-man theory: the original big bang model.

• By Penrose's entropy argument, is more likely than inflation.
• Doesn't require additional degrees of freedom like inflaton.

How might we get out of this predicament?
Inflation rewards "procrastination."

• We made a profound error in judgment in the early days of inflation.
• Didn't take proper accont of how qm and inflation mix.
• A hump in the mexican-hat potential has low-probability tails as it rolls down toward the minimum.
• Inflation rewards "rogue" regions, which are regions in which inflation took an unusually long amount of time to end.
• "anything that can happen will happen: and it will happen an infinite number of times" - Guth 2000

The volume measure fails.

• Predictions will depend entirely on the measure principle -- not inflation, and then we have to consider how a similar measure applied to other theories compares in predictive power.
• Suppose that the big bang could happen many times in separated regions. How would we regularize it?

...
...

I just saw this post. Admirable notes. They follow the talk closely as far as I can go---up to the part where my understanding gets thin.

 

[skydivephil]

I read Steinhardt's artcile in Sciam , haven't watched the video yet.
But there were a few points that I would love clearing up:
1 Steinhardt claims inflation doesn't predict anything becuase it implies (if eternal inflation is the righrt version of inflation) an infinite multiverse therefore anything that can happen will happen. But it seems to me his alternatrive of a cyclic universe is also a multiverse in another form. In that a cyclic universe is conceptually a multiverse where the different universes exist one after another , rather than at the same time. But I don't see why this is conceptually any different, am I missing something?
2 Steinhardt claims that inflation makes no predictions then said it predicted a strong gravitational wave signature, so which is it?
3. What was his reposnse to the idea that bubble collisions in the inflationary multiverse are potentially detectable? He didnt mention this in his Sciam article, if he think there's a problem with it I think he should explain why. Just ignoring it doesn't seem like good practice. Same with LQC making inflation more likely in the Ashkebar/Sloan paper.

Id also love to hear what the new observational claims of CCC are. I was under the impression the circles claimed by Penrose had not been well received by most CMB experts, so what's new?

 

[marcus]

Phil, with luck (since your question is about the SciAm article) Ben will reply to your questions concerning it.
I don't have a subscription to SciAm and when I first looked for the article I ran into a pay-per-view page, so assumed it was not available except on that basis. I just now discovered what seems to be a free version. Maybe all SciAm articles are free if you know how to look? I'm confused about this.

http://www.scientificamerican.com/article.cfm?id=the-inflation-summer&page=1
http://www.scientificamerican.com/article.cfm?id=the-inflation-summer&page=2
http://www.scientificamerican.com/article.cfm?id=the-inflation-summer&page=3
...
http://www.scientificamerican.com/article.cfm?id=the-inflation-summer&page=6
etc.
Haven't had time to read it yet.

I'm going to try to print the article out and will look it over as time permits later today.

You asked concerning Penrose: I don't know of anybody in Cosmology who goes along with his CCC idea. I may not be adequately informed but my impression is that what he means by "new evidence" is just more data analysis from his Russian guy Gurzadyan. If I am right, it is indeed "new evidence" but that does not mean that the broader community accepts it.

I think it is to the credit of the conference organizers that they invited Penrose to present his idea and what he considers to be new supporting evidence. Science should be open like that. Besides he is a brilliant and entertaining speaker with a huge record of original ideas that (many of them) have worked, so you could say it was a no-brainer to invite him. But it still illustrates the desirable openness of scientific discussion. Penrose idea has been seriously challenged and questioned, but it is not crackpot and it has not been completely ruled out. It has missing pieces, but it is still possible that they will be filled in.

In the meanwhile, my attitude is that I don't want to spend a lot of time studying it but I want to keep it in my peripheral vision, something on the sidelines to be aware of. I certainly watched the talk. He's a great guy, he makes a strong case, it's interesting. I think the odds are against but I totally respect his campaign to get the community to consider it.

I read Steinhardt's artcile in Sciam , haven't watched the video yet.
But there were a few points that I would love clearing up:
1 Steinhardt claims inflation doesn't predict anything becuase it implies (if eternal inflation is the righrt version of inflation) an infinite multiverse therefore anything that can happen will happen. But it seems to me his alternatrive of a cyclic universe is also a multiverse in another form. In that a cyclic universe is conceptually a multiverse where the different universes exist one after another , rather than at the same time. But I don't see why this is conceptually any different, am I missing something?
2 Steinhardt claims that inflation makes no predictions then said it predicted a strong gravitational wave signature, so which is it?
3. What was his reposnse to the idea that bubble collisions in the inflationary multiverse are potentially detectable? He didnt mention this in his Sciam article, if he think there's a problem with it I think he should explain why. Just ignoring it doesn't seem like good practice. Same with LQC making inflation more likely in the Ashkebar/Sloan paper.

Id also love to hear what the new observational claims of CCC are. I was under the impression the circles claimed by Penrose had not been well received by most CMB experts, so what's new?

 

[skydivephil]

I looked at the line up and I agree its great they invited lots of different view points, I will try and watch some of the videos If get a chance, I am keen to see if Penrose has anything new in terms of observational support. In the meantime do you know if anyone is blogging about the conference?

 

[cristo]

I've not watched this talk, but I've been in similar talks by Steinhardt's collaborators in the past. I don't know which model he was talking about, but there are problems with these ekpyrotic models. The objection that crops up the most is how does one get a spectrum of perturbations that is produced in a contracting phase to travel through the 'bounce' (which is really a singularity) and into the expanding phase that we live into map onto the CMB? This is normally just stated without proof, or just hidden under the carpet when people talk about these models but, as far as I am aware, this issue has not been properly addressed.

 

[marcus]

Phil, I can't get the SciAm article to print properly, so i will quote some excerpts:

==sample excerpts from Steinhardt's April SciAm piece==
...The first dictum holds that inflation is inevitable. But if it is, there is an awkward corollary: bad inflation is more likely than good inflation. “Bad inflation” means a period of accelerated expansion whose outcome conflicts with what we observe. For example, the temperature variations might be too large. The difference between good and bad hinges on the precise shape of the potential energy curve, which is controlled by a numerical parameter that could, in principle, take on any value whatsoever. Only an extremely narrow range of values could produce the observed temperature variation. In a typical inflationary model, the value must be near 10–15—that is, zero to 15 decimal places. A less fine-tuned choice, such as zero to only 12 or 10 or eight decimal places, would produce bad inflation: the same degree of accelerated expansion (or more) but with a large temperature variation that is inconsistent with observations.

We could ignore bad inflation if it were incompatible with life. In that case, even if such large temperature variations could arise in principle, we could never observe them. Reasoning of this kind is known as the anthropic principle. Yet it does not apply here. Larger temperature variations would result in more stars and galaxies—the universe would, if anything, be more habitable than it is now.
...
...

Such procrastinating rogue regions are extremely rare, so you might think it safe to ignore them. You cannot, because they are inflating. They continue to grow and, in a matter of instants, dwarf the well-behaved region that ended inflation on time. The result is a sea of inflating space surrounding a little island filled with hot matter and radiation. What is more, rogue regions spawn new rogue regions, as well as new islands of matter—each a self-contained universe. The process continues ad infinitum, creating an unbounded number of islands surrounded by ever more inflating space. If you are not disturbed by this picture, don’t worry—you should not be. The disturbing news comes next.

The islands are not all the same. The inherently random nature of quantum physics ensures that some are highly nonuniform or strongly warped. Their nonuniformity sounds like the problem of bad inflation described earlier, but the cause is different. Bad inflation occurs because the parameters controlling the shape of the potential energy curve are likely to be too large. Here nonuniformity can result from eternal inflation and random quantum fluctuations no matter what values the parameters have.
...
...

An alternative strategy supposes that islands like our observable universe are the most likely outcome of inflation. Proponents of this approach impose a so-called measure, a specific rule for weighting which kinds of islands are most likely—analogous to declaring that we must take three quarters for every five pennies when drawing coins from our sack. The notion of a measure, an ad hoc addition, is an open admission that inflationary theory on its own does not explain or predict anything.

Worse, theorists have come up with many equally reasonable measures that lead to different conclusions. An example is the volume measure, which says that islands should be weighted by their size. At first glance, this choice makes common sense. The intuitive idea underlying inflation is that it explains the uniformity and flatness we observe by creating large volumes of space with those properties. Unfortunately, the volume measure fails. The reason is that it favors procrastination. Consider two kinds of regions: islands like ours and others that formed later, after more inflation. By the power of exponential growth, the latter regions will occupy vastly more total volume. Hence, regions younger than ours are vastly more common. By this measure, it is unlikely we would even exist.
...
==endquote==
Readers who haven't already read the whole article can get more with these and similar links.

http://www.scientificamerican.com/article.cfm?id=the-inflation-summer&page=4
http://www.scientificamerican.com/article.cfm?id=the-inflation-summer&page=6

 

[bcrowell]

I've not watched this talk, but I've been in similar talks by Steinhardt's collaborators in the past. I don't know which model he was talking about, but there are problems with these ekpyrotic models.

The style of the talk is that he's going for generality. He makes general arguments about scaling, time reversal, and attractors that apply to broad classes of models. I didn't get the impression that he was talking about ekpyrotic models at all, although maybe some of his language was broad enough to include those.

 

[marcus]

That's how it struck me too. He wasn't making a specific pitch for the clashing branes picture. He barely mentioned the word "ekpyrotic". It was primarily a critique of inflation.

One can't defend inflation by saying there are problems with some specific alternative (classing branes say) that Steinhardt and Turok happened to propose 10 years ago. One has to listen to the critique in its own right, disconnected from whatever alternative proposal one suspects him of having in mind.

The way I understand it, the cornerstone of his argument is that the original idea of inflation (circa 1980) was based on a mistaken premise---it used classical gravity and not quantum gravity, and assumed that QG effects could be ignored. Then later it was realized that the QG effects were all-important. I'll try to find the slide where he says this.

 

[marcus]

...
He's trying to sell cyclic models in which there is a contraction and a bounce. He lists a lot of good features of such models. One such model is Penrose's CCC, and although he gives credit to Penrose for a lot of important ideas, he shows no enthusiasm for CCC per se.

His message is cyclic=good, inflation=bad, and he has great hopes that observations will soon confirm that cyclic is what happened. The impression I get is that all such models are failures and none of them make any firm connection with observation.
...

What do you think he means by "cyclic"? I suspect he means something more general than the clashing branes scenario he and Turok proposed. Maybe you can point me to a slide where this is clarified.

I'm not sure but I suspect that by "cyclic" he means any model where the beginning of expansion is preceded by a contraction.

But the idea of a contraction is general and might be realized in different ways---not tied to one particular model. He also used the word "bounce", as I recall.

A key slide might be 38/61 (#38 out of 61 in all, about 2/3 of the way down) where he says "smooth while contracting, but then need an information-preserving bounce"

I suspect that "bounce" refers generally to any model where a new expanding classical regime emerges from the collapse of an old classical regime.

He also seems to favor reconsidering the whole thing on the basis of QG rather than classical---viz the key point about 1980s inflation being fundamentally classical and misconceived for that reason.
==================================

EDIT: I watched again, and saw that he is constantly referring to the brane-clash scenario as an EXAMPLE of a bounce. But his treatment is general enough to include other bounce cosmologies.
An important new concept is the SEIVE where badly behaved regions form black holes during the collapse and only well-behaved get through. Those with low geometric entropy and high matter entropy.

He constantly refers to the talk http://pirsa.org/11070028/ by Jean-Luc Lehners which is focused on the usual brane-clash (Steinhardt-Turok) form of cyclic-ekpyrotic. But that is serving simply as an EXAMPLE, I would say. One possible realization of the bounce/smoothing/seive idea.

I may watch the talk again. I think it is a major one and I want to understand accurately. I will also watch some of the one by Jean-Luc. Any other suggestions of PIRSA talks or papers related to this?

 

[marcus]

I just watched about 2/3 of the Jean-Luc talk http://pirsa.org/11070028/
It is one of the clearest presentations of colliding-brane cosmology which I have seen.
The thinking in this line has become quite advanced. Sometimes the branes stick together in patches. The cyclic collision process selects out regions of the branes which are habitable. The cycle tends to take 200 to 300 billion years, but the flapping of one brane against the other can take different times in different regions. The thinking has become intricate and highly imaginative. The model is shown to have many advantages.

I would definitely recommend watching at least a part. Jean-Luc is an excellent speaker and has clear, thoughtfully-organized slides with lots of good diagrams.

This sheds light on Steinhardt's talk. It is, so to speak, the brane-clash PART of Steinhardt's talk which he left out--I think because he wanted to make a general argument criticizing inflation independently of anyone particular alternative model, which could encourage researchers to invent and explore various (bounce cosmology) alternatives to inflation.

In case anyone is coming in new, Steinhardt's talk is here:
http://pirsa.org/11070029/
It is outstanding. He intentionally phrased it in widely understandable language.

 

[skydivephil]

As I recall they renamed their ekpyrotic model as the cyclic model, so I am not sure he's talking as generally as one might imagine when he says cyclic. i will try and find a reference to this.

 

[skydivephil]

I think this is it:
http://arxiv.org/pdf/astro-ph/0204479v1

 

[marcus]

As I recall they renamed their ekpyrotic model as the cyclic model, so I am not sure he's talking as generally as one might imagine when he says cyclic. i will try and find a reference to this.

That's right, as far as I know. For many years Steinhardt and friends were using "cyclic" precisely to refer to their brane-clash model! My point is that I think now he is thinking more generally---not so one-track as before. We'll see.

I think for example Penrose CCC qualifies as cyclic, maybe not. One reason to have a conference is to get a shared terminology and some shared general concepts. Listen to Steinhardt's talk and see what you think.

 

[cristo]

One can't defend inflation by saying there are problems with some specific alternative (classing branes say) that Steinhardt and Turok happened to propose 10 years ago. One has to listen to the critique in its own right, disconnected from whatever alternative proposal one suspects him of having in mind.

But this is coming from someone who clearly has his own theory to push. And, in fact, one can 'defend' inflation in this way, since this is how science works. It is all well and good for people to criticize a theory, but until another theory comes along that fits observations better, inflation is preferred.

The thinking has become intricate and highly imaginative.

To be honest, that sums it up. Intricate, and highly imaginative, but is this really physics? I've seen this talk before and, as I mentioned earlier, the main criticism that was brought up several times is the conjecture that certain parts of the model make it through the bounce such that the amplitude, Q, which is related to some initial conditions in the collapsing phase survives to become a prediction in the expanding phase and thus the CMB. But how is this a prediction if you don't explain *how* such a spectrum can pass through a singularity?

 

[bapowell]

As I recall they renamed their ekpyrotic model as the cyclic model, so I am not sure he's talking as generally as one might imagine when he says cyclic. i will try and find a reference to this.

Originally, ekpyrosis was the early universe scenario that identified the big bang with a brane collision. The cyclic model came later, when it was suggested that these collisions could take place repeatedly. So, the ekpyrotic scenario by itself should not imply anything cyclic.