# People’s evolving opinions about quantum gravity

Luboš Motl, April 25, 2015

About nine years ago, a movement trying to (largely or entirely) «replace» string theory research with would-be «competitors» culminated. Unproductive critics and third-class researchers disconnected from the last 30 years in physics were often marketed as peer of top string theorists – and sometimes as something better.

However, aside from the cheap anti-science populism, there has never been any substance in their claims, and one can’t really run on «promises» indefinitely. For a while, the theory group at the Perimeter Institute operated as a fan club of Lee Smolin’s of a sort – a warrior in the «string wars». Thankfully, «string wars» are over and the crackpots have lost. Unfortunately, they have been replaced by lots of other nonsense. Did this replacement make things better or worse? I don’t know.

To return to the positive news, let me mention that after many decades, Sabine Hossenfelder – once a person working at the periphery of the loop quantum gravity community – has understood a lethal problem with the Poincaré-invariant networks.

The correct insight is that the Minkowski space can’t really be represented by a network or graph that would fill it. The very same argument (with very similar pictures) has been written on this blog many times since 2004 – see e.g. the second myth, «a structure of links or surfaces filling a Minkowski space may be Lorentz-invariant, at least statistically», in the 2009 TRF blog post about myths about the minimal length.

I won’t repeat it here but it’s trivial to see that every quasi-uniform network drawn on a paper inevitably picks a preferred reference frame – the Lorentz boosts acting on the network inevitably make it look «skewed» and «different». So a theory representing the space by the network breaks the Lorentz symmetry – pretty much by 100 percent or so. Consequently, spin foams and all similar discrete theories of quantum gravity may be falsified (by pointing to the successful tests of the Lorentz invariance) within a few seconds, assuming that the falsifier is adequately competent and fast. Now, decades later, people like Hossenfelder started to get the point, too.

(Just to be sure, I don’t claim to be the first one who invented the argument, although I hadn’t heard it in the same clear form before.)

Second. The Perimeter Institute used to be a center of the «bogus research» of quantum gravity. Fortunately, real physics research was gradually strengthening over there in the last 10-15 years. So three days ago, Witten gave a PI lecture on superstring perturbation theory, a way to calculate the stringy S-matrix using the super world sheets (with Grassmannian coordinates).

It is a nice piece of technology but the talk is very technical so I won’t recommend it to non-experts. However, I must mention that I liked his points about the vertex operators’ being localized on divisors etc. because I have spent lots of time by thinking about the mysterious duality – and generalizations of string theory to 2+2-dimensional membranes – and the conclusion that the vertex operators should be associated with divisors seemed to be rather solid. I should write an update about these matters at some point.

**PR, ER=EPR, fuzzballs vs AMPS firewalls**

Meanwhile, a new controversy has erupted in quantum gravity – in the serious (at least before a split?) quantum gravity done by people who are familiar at least with the basic lessons from string/M-theory, and most of them are great string theorists. No, I don’t mean the «gravity as the entropic force» nonsense that is hopefully dead by now as well (much like the spin foams above) – although the dying still took about 50 times longer time than it should have.

I mean the controversies about the «firewall» claims by Polchinski and collaborators, and positive insights that were made as part of the argumentation that the firewalls don’t really exist.

The cover story of the April 2015 issue of Scientific American is hyping firewalls and Joe Polchinski himself wrote this embarrassing, self-glorifying yet completely wrong, article. I’ve written some critiques before. But I await a guest blog about the matters and fuzzballs by Samir Mathur – who is probably also rightfully upset that some of his decade-old claims are being attributed to Polchinski and pals.

But this deeply flawed pro-firewall propaganda makes it to many articles that are otherwise avoiding it, too. So the Quanta Magazine wrote the first article in a series about the ER-EPR correspondence

K.C. Cole wrote a pretty clear exposition of the proposed universal equivalence between the wormholes and the quantum entanglement. The article also contains an amusing historical anecdote about the birth of ER=EPR. How was it born? Well, on one sunny day, the master (Susskind’s nickname for Juan Maldacena) wrote a cryptical e-mail to Susskind that said ER=EPR. Susskind read the e-mail and immediately saw where it was going, he decided it was correct, so he agreed to put his name on the preprint. Nice done, Gentlemen! 😉

I must say that the origin of my much less well-known joint paper with Susskind was the same with one M-name replaced by another. 🙂

There are some reactions to ER=EPR mentioned in Cole’s article. Preskill tries to be characteristically neutral and talks about uncertain smells. Shenker is supportive of the «big new insight». On the other hand, Polchinski and Marolf, the main guys behind the AMPS firewall meme, do their best to kick into Maldacena’s and Susskind’s insight while looking like Gentlemen.

The two last paragraphs of Cole’s article are dedicated to Marolf’s criticism of ER=EPR which I find vague enough not to be «totally sharply and rigorously wrong» but which is still weird and at least morally wrong if you realize what he is saying in between the lines:

To be sure, ER = EPR does not yet apply to just any kind of space, or any kind of entanglement. It takes a special type of entanglement and a special type of wormhole. “Lenny and Juan are completely aware of this,” said Marolf, who recently co-authored a paper describing wormholes with more than two ends. ER = EPR works in very specific situations, he said, but AMPS argues that the firewall presents a much broader challenge.

Like Polchinski and others, Marolf worries that ER = EPR modifies standard quantum mechanics. “A lot of people are really interested in the ER = EPR conjecture,” said Marolf. “But there’s a sense that no one but Lenny and Juan really understand what it is.” Still, “it’s an interesting time to be in the field.”

Please, Don!

First of all, it is complete nonsense that ER=EPR only applies to one special case «so far». ER=EPR is one of the insights that – assuming that they are right, and I think that the case is extremely strong – apply as generally as you can get. The degree of generality is similar to Joule’s equivalence of heat and work – or, indeed, as Noether’s theorem. The claim really is that *any* entanglement is some kind of a wormhole; and any physically allowed wormhole may be equivalently described as a spacetime without a wormhole but with near-maximally entangled local degrees of freedom.

What could have Marolf meant by the wormhole’s being very special? That it is an Einstein-Rosen bridge and not a traversable wormhole? It has to be so. This is a part of the insight. A traversable wormhole would imply massive non-locality (violation of special relativity) – and this is forbidden in both descriptions, one with the wormhole and one with the entanglement. So the choice of the non-traversable wormholes isn’t a sign of any incompleteness of the proposal. It is a detail that is makes Maldacena’s and Susskind’s proposal much more specific and bold. Traversable wormholes are almost certainly prohibited by the laws of physics – and Maldacena and Susskind confirm this expectation while a new argument for the non-existence of traversable wormholes arises as a corollary of their work.

Otherwise, the insight works in many spacetimes where the wormholes have technically different shapes – and different number of dimensions, among other things. They also claim that a description for «excited» wormholes exists on both sides, too. Tiny amounts of entanglement don’t allow the wormholes to be big and smooth. But that’s not a defect of the proposal, either. It’s another bold prediction that follows from the ER=EPR line of reasoning.

If wormholes with many throats etc. are allowed as well, there probably exists some very special kind of entanglement of the three systems that describes the object, too.

But it’s really the comparison of the generality of ER=EPR and of AMPS that sounds crazy. Marolf claims that AMPS is much more general. Well, it’s not. It’s exactly the other way around. ER=EPR is supposed to be relevant and possible for *any* black hole interior – it may be connected elsewhere without spoiling the overall outside appearance of the black hole. On the other hand, AMPS is as special as a physically wrong claim may be. It is extremely special because the firewall is only derived for theories where the black hole complementarity is forbidden as a strict assumption, where the exact locality holds, and where the field operators are state-independent. With these very strong assumptions, you end up with Polchinski-like contradictions. But the assumptions are not obeyed in string theory – in consistent theories of quantum gravity – which makes their range validity pretty much zero for all practical purposes.

The claim that «AMPS is more general than ER=EPR» is at least morally untrue. It is hard to decide how you define the «degree of generality» for two qualitatively different hypotheses that disagree with one another (which implies that at most one of them is actually right) and want to organize quantum gravity in different ways. But if I try to define the «degree of generality» in any sensible way, it’s clear that ER=EPR is much more general.

Marolf’s (and Polchinski’s) assertion that «ER=EPR seems to violate the posulates of quantum mechanics» is preposterous, too. How could it violate the general rules of quantum mechanics? It’s constructed within these rules from the very beginning. One has the quantum mechanical description of the microstates of two identical but separated black holes (pretty much a tensor product of two copies of the same Hilbert space of microstates). And one simply claims that some entangled basis vectors in this product Hilbert space may be given new labels, as «simple» microstates of an Einstein-Rosen bridge. How could it violate anything about quantum mechanics? It’s really just a collection of new labels for some vectors in the Hilbert space. A way to define new observables – field operators in the wormhole’s interior – that are complementary to (i.e. non-commuting with) the usual field operators in two black hole interiors. From the birth of quantum mechanics, its power to allow, encourage, or force us to use superpositions of states (e.g. the entangled states) – and see that many of them are eigenstates of rather natural operators – has been one of the most characteristic changes that quantum mechanics represented. Quantum mechanics is all about the non-commuting operators, stupid!

This ER-EPR correspondence is a non-vacuous hypothesis – there has to exist some Hamiltonian or evolution in both pictures that agrees with both descriptions (that share the idea about the evolution of the exterior of the black holes or the wormhole). But this test is a dynamical question, not one that could change anything about the fact that Maldacena and Susskind operate within the totally standard quantum mechanical framework at every moment of their research.

I think that by now, it should be clear that Raju and Papadodimas’ insights about the unavoidable state dependence are the key – and it is the key that is misunderstood by Marolf, Polchinski, and others. Even in ER=EPR, I just said that what the field operators are depends on which «corner of the Hilbert space» you want to describe. For the non-entangled corners, you pick the field operators describing two interiors. For the maximally entangled corners, field operators within an ER bridge are a better description (in ER=EPR, the state dependence is «manifest» because the two regions of the Hilbert space invite you to use two *inequivalent* sets of field operators because the spacetime topologies are different). The modest claim here is that the local field operators always have a «limited range of validity» on the Hilbert space. But that shouldn’t be controversial. That’s an aspect of backreaction or the impossibility to describe quantum gravity in a manifestly local way.

Finally, Marolf claims that «no one but Juan and Lenny understand ER=EPR». Now, this is just a plain lie. Perhaps a cute lie – because it mimics the bizarre claims that general relativity was only understood by 12 people in the world. But it is a lie, anyway. I surely do claim to understand the content of the claim as well as Maldacena and Susskind, and so do numerous people who have written about it.

In less than 2 years, the ER=EPR paper has 150 citations and most of them seem to understand what Maldacena and Susskind say and why. 150 isn’t astronomical but it’s almost the same rate – 320 citations in 3 years – that the AMPS firewall paper has.

If Marolf was really talking about sociology, I think that if you look at the 320 followups of AMPS, there will be so many that cite it even though they disagree with the firewall claims that the total body of papers citing either AMPS or ER=EPR will have a majority thinking that the firewalls don’t exist.

The truth isn’t decided by polls, however. One may still see that Marolf’s sociological claims painting Maldacena and Susskind as two lonely fenceposts is a complete fabrication – especially if you read this fabrication from someone who implicitly says that his «argument that firewalls exist» is understood and agreed with by almost everyone.

It always bothers me: Couldn’t Erik Verlinde see that gravity as the entropic force has some lethal bugs he may have been (and we may have been) unaware of that kills the whole picture? Isn’t Joe Polchinski, an extremely smart physicist, able to see that similar problems with their AMPS proof (and loopholes) have been found as well – along with some new, much more positive and specific insights – that make the beef of the firewall paper pretty much evaporate? And that make them ignore some nice developments just because they don’t agree with some disproved faith of theirs?

Of course, there are more brutal examples of this. Couldn’t Gerard ‘t Hooft, a more than well-deserved Nobel prize winner, have seen that his claims about hydrodynamical models behind quantum mechanics have turned to a complete self-evident failure after those 20 years? Why do all these men keep on defending something that has become completely indefensible for so many years? Have they really lost the ability to think rationally, or are they afraid to admit that they were wrong and they know that to defend an arbitrarily silly proposition will always be OK in the broader public because most people don’t have a clue, anyway?

In the late 1990s, when ‘t Hooft began with his hydrodynamic things that could have been shown to be wrong, he was ignored, despite his immense aura. But I am worried that similar, demonstrably wrong «research directions» are eating an increasing portion of the researchers, that a majority of the body of researchers is losing their competence. Are we entering a period in which people will defend AMPS-like paradoxes and entropic gravities for centuries even though it should take at most minutes for a competent physicist to understand why they’re not right? Are we returning to the Middle Ages?