Susskind and Bousso's incomprehensible new paper

[mitchell porter]

Bousso and Susskind have a http://arxiv.org/abs/1105.3796" .

I have two reasons for creating a thread about a paper which one might otherwise prefer to ignore, as it contains no new quantitative results or ideas for how to obtain them. The first reason is that the paper is bound to get attention because its thesis sounds so simple and yet so incomprehensible, so it might be useful to understand what is actually being asserted, even if it's fallacious (indeed, especially if it's fallacious, for in that case we should want to be able to pinpoint the fallacy). The second reason is that I'm quite interested in holographic cosmology, but I don't like Susskind's version, so I want to get clear on our differences.

In holographic cosmology according to me, we live in something like de Sitter space, ordinary space-time is the bulk, and the boundary is the past conformal boundary and/or the future conformal boundary. In Susskind's version... well, I don't really understand it, that's part of the problem, but he seems to make the observer-dependent cosmological horizon his holographic surface, and he also apparently wants to conceive of all the other inflationary domains in the universe as being holographically emergent from that horizon on the other side from us (in the interior of our Hubble volume), and he also wants to say that there are no extra degrees of freedom associated with these other inflationary domains: that they are literally made up of the same degrees of freedom that make up the galaxies around us, just viewed differently. That sounds crazy, of course, but that is what you get if you apply black hole complementarity to cosmology.

So, my opinion is that these ideas contain a mix of bad philosophy and wrong technical choices. The bad philosophy is the idea that everything outside our Hubble horizon is not as real as we are, because we can't ever see it (this is an anti-realist, instrumentalist approach to the concept of existence, typical of the Copenhagen interpretation), so we can regard it as a mysterious re-description of everything inside the horizon. The wrong technical choice is to develop holographic cosmology in terms of observer-dependent cosmological horizons rather than the observer-independent past and future conformal boundaries - a choice which gets made because of the argument about "meta-observables" in dS/CFT.

Those are my "opinions" - that is, an instant intuitive assessment of where the paper came from intellectually and what's wrong with it - but I haven't yet sat down with it and tested these opinions against the details of the argument. So you could regard it as a "prediction". Now to see if it stands up against the reality of the paper...

 

[MTd2]

In holographic cosmology according to me, we live in something like de Sitter space, ordinary space-time is the bulk, and the boundary is the past conformal boundary and/or the future conformal boundary. In Susskind's version... well, I don't really understand it, that's part of the problem, but he seems to make the observer-dependent cosmological horizon his holographic surface,

That is really the oldest part. And this is the version for which string theoretical papers gave Bousso more than 2000 citations for 4 or 5 papers. Eric Verlinde also used this concept, but people complain abou that. Weird. Odd.

 

[mitchell porter]

In holographic cosmology according to me, we live in something like de Sitter space, ordinary space-time is the bulk, and the boundary is the past conformal boundary and/or the future conformal boundary. In Susskind's version... well, I don't really understand it, that's part of the problem, but he seems to make the observer-dependent cosmological horizon his holographic surface, and he also apparently wants to conceive of all the other inflationary domains in the universe as being holographically emergent from that horizon on the other side from us (in the interior of our Hubble volume), and he also wants to say that there are no extra degrees of freedom associated with these other inflationary domains: that they are literally made up of the same degrees of freedom that make up the galaxies around us, just viewed differently. That sounds crazy, of course, but that is what you get if you apply black hole complementarity to cosmology.

I may need to retract this particular criticism.

Look at figure 13 in the paper (page 40). The causal diamond on the left of the diagram is a very special sort, in Susskind's version of eternal inflation. It's an exactly supersymmetric vacuum and it lasts forever, whereas a space-time region like our own, where geometry is de Sitter and where supersymmetry is broken, is only metastable and within finite time will quantum-mechanically tunnel into a stable, supersymmetric vacuum state.

It seems that what they are really saying is that de Sitter patches like our own are holographically dual to portions of a final eternal flat space in which supersymmetry is restored. Of course, I find that much much more acceptable than what I thought Susskind was saying - that a possibly infinite number of space-time regions beyond our Hubble horizon, including all their galaxies and civilizations, are dual to the ~ 10^80 fermions making up what we can observe. So, I still can't say that they're right, but at least they're not being absurd.

 

[Demystifier]

Lubos Motl has already written a post http://motls.blogspot.com/2011/05/bousso-susskind-hypermultiverse.html" .

It's nice to see that even string theorists may be attacked by Motl. And I must say that I agree with him. Bousso and Susskind seem not to understand what decoherence really is.

 

[Fyzix]

It's nice to see that even string theorists may be attacked by Motl. And I must say that I agree with him. Bousso and Susskind seem not to understand what decoherence really is.

Care to elaborate?

 

[Demystifier]

Motl have already elaborated it.

 

[mitchell porter]

Despite the authors' slogan, I am beginning to think this is not a many-worlds paper at all. Their argument reduces to two steps:

1) Quantum mechanics assigns arbitrarily precise probabilities to experimental outcomes (the probabilities are real numbers). The only way for such quantities to have an operational meaning is for the experiments to be performed an infinite number of times (so that arbitrarily precise asymptotic frequencies of experimental outcomes can be obtained). Furthermore, because of measure problems in inflationary cosmology, these infinite repetitions of the experiment need to be capable of being performed (in principle) by a single observer; you can't just use the fact that they will occur infinitely many times in the separate inflationary regions.

2) But an individual de Sitter geometry is metastable and only lasts for a finite time, therefore it can't contain the necessary infinite number of measurements. Fortunately, according to Susskind's hypothesis of "hat complementarity" or "FRW/CFT", the final ground state of an inflationary region is sometimes a flat, exactly supersymmetric vacuum that lasts forever, and in a region like this, you could have observers that last forever. Also, the individual time slices of these supersymmetric regions are holographically dual to whole de Sitter regions, so stuff that happens in a space-time region like the one we live in is actually equivalent to events happening in one of these terminal space-time regions.

So, the part of the argument that is actually about quantum mechanics (step 1) has nothing to do with many worlds - instead, it's a peculiar form of operationalism, saying that the quantities in a physical theory are meaningless unless they correspond, at least potentially, to observations, therefore the arbitrarily precise predictions of quantum mechanics are meaningless unless the universe potentially allows infinitely many experiments to be performed and their results compared. Logically, this is not too far removed from Frank "Omega Point" Tipler saying that black hole evaporation would violate quantum unitarity, therefore black holes never completely form, therefore intelligent life must conquer the universe and manage the Big Crunch so as to prevent black holes from ever forming, therefore intelligent life lasts forever and produces a God at the end of time. Or, a milder example, it's like another recent paper for which Bousso was a coauthor, which argued that time might end a few billions years from now, because the only way they could solve the measure problem of eternal inflation was by introducing a cutoff on space-time. In all these instances, a relatively technical problem of formalism or of interpretation is being "solved" by a completely over-the-top solution.

I believe Susskind's peculiar variation on holography is of independent interest (and it was already introduced in other papers). That is, one should want to know whether it makes sense as a physical or even mathematical hypothesis. I don't understand it on that level yet, but as I said, it seems to make a single one of these eternal supersymmetric vacua dual to a whole stack of de Sitter space-times on its boundary, so that an individual de Sitter universe is dual to just one slice through the supersymmetric Minkowski space. But in the evaluation of this proposal, the argument from quantum first principles that Bousso and Susskind present is just a big distraction.

 

[mitchell porter]

Gell-Mann and Hartle were feeling left out so they put out a http://arxiv.org/abs/1106.0767" too. I haven't even read it but one sentence leapt out at me and I want to highlight it, because it is entirely characteristic of one approach to the interpretation of QM:

"In this paper we show how to overcome these obstacles. We overcome the first one by extending the notion of probability to include values outside the range [0, 1]."

Well, I have a new interpretation of QM too. It's the What,-Me-Worry? interpretation. It is obtained by extending the notion of logic to include valid contradictions. That way, if anything about QM sounds paradoxical, you don't have to worry - just say it's a valid contradiction. (The idea of a valid contradiction may itself sound like a contradiction, but the obvious response is that it is nonetheless essential for my interpretation to work, so it must itself be a valid contradiction too! Behold the power of illogical logic.)

Sometimes extending a concept beyond its initial domain of validity pays off - non-euclidean geometry, non-commutative geometry. But sometimes it's just a way of hiding conceptual ignorance or incoherence. Let the consumer of quantum explanations beware.

 

[atyy]

Well, I have a new interpretation of QM too. It's the What,-Me-Worry? interpretation. It is obtained by extending the notion of logic to include valid contradictions. That way, if anything about QM sounds paradoxical, you don't have to worry - just say it's a valid contradiction. (The idea of a valid contradiction may itself sound like a contradiction, but the obvious response is that it is nonetheless essential for my interpretation to work, so it must itself be a valid contradiction too! Behold the power of illogical logic.)

That already exists. It's called Copenhagen or Shut up and calculate.

 

[Demystifier]

Well, I have a new interpretation of QM too. It's the What,-Me-Worry? interpretation. It is obtained by extending the notion of logic to include valid contradictions. That way, if anything about QM sounds paradoxical, you don't have to worry - just say it's a valid contradiction. (The idea of a valid contradiction may itself sound like a contradiction, but the obvious response is that it is nonetheless essential for my interpretation to work, so it must itself be a valid contradiction too! Behold the power of illogical logic.)