## Saturday, December 10, 2016 ... //

### Nature: LIGO may have seen echoes from a black hole firewall

It has become too common for journalists to write crazy stuff that they should be normally ostracized and spitted upon for

It's clear why Joe Polchinski had to win a Milner Prize. The reasons are the D-branes and similar things. However, some people around the prize have suggested that Polchinski could have gotten the award for the proposal that the black hole interior doesn't exist – the infalling observer unavoidably dies in a "firewall" instead of smoothly crossing the event horizon.

Well, whether Milner or Zuckerberg or anyone else made such a comment, I find it silly, and so do almost all fundamental physicists. The firewall paper by Joe – well, it was really Joe with 3 collaborators who couldn't quite be neglected if Joe were getting a prize for firewalls – is a nice piece of quantum information argumentation but like the assumptions, the conclusion is entirely wrong.

They assume that the locality has to be either exact or totally violated at the spacetime background with a black hole. In particular, the black hole complementarity (the failure of the black hole interior degrees of freedom to be independent) has to be banned. Not surprisingly, because the black hole interior wants to exist as a scrambled copy of the exterior and because they assume that it must be totally independent of the exterior if it exists at all, Joe et al. unsurprisingly "derive" that the black hole interior cannot exist at all, so the infalling observers have to be killed by the firewall.

A classic example of the "garbage in, garbage out" (GIGO) rule.

Nature doesn't confirm this conclusion, of course. The black hole interior does exist but its degrees of freedom can't be assumed to be quite independent from the degrees of freedom outside the black hole. The black hole approximately respects the locality but when you measure the information too accurately – when you're looking for correlation functions with $$N$$ field insertions where $$N$$ is comparable to the black hole entropy $$S$$, for example (as articulated in a fresh paper by Ghosh and Raju, but I could have used somewhat different but related ways to argue), one finds out that the locality breaks down and the interior isn't independent at all. But for simple low-energy measurements with a low $$N$$, locality is an excellent approximation even in the presence of a black hole.

That controversial Polchinski et al. paper on the firewalls (AMPS) shows that they are experts in quantum gravity and quantum information but because the assumptions and conclusions are wrong, it's obvious that such a paper can't be said to have a million or multimillion-dollar value from a physics viewpoint. You simply shouldn't be getting physics awards for physically incorrect claims. The whole point of science and physics is to find correct insights about Nature, not just to show that you have mastered some tools and you can design clever, but ultimately irrelevant, arguments that derive incorrect conclusions from invalid assumptions.

But some people surely must believe that the firewalls should be seriously considered. Yesterday, Nature published a truly ambitious article

LIGO black hole echoes hint at general-relativity breakdown
Zeeya Merali mainly discusses a gr-qc preprint submitted a week ago,
Echoes from the Abyss: Evidence for Planck-scale structure at black hole horizons

by Jahed Abedi, Hannah Dykaar, and Niayesh Afshordi who claim that there's some (globally) 2.9-sigma evidence in the three LIGO black hole datasets that supports the existence of some very odd phenomena – an echo coming from the merged black hole. They say that the signals discovered by LIGO tend to repeat many times (as an echo) after the time delay$\Delta t = 8GM \log M.$ The factor $$GM$$ is comparable to the Schwarzschild radius of the resulting black hole. Obviously, they don't want the delay to be comparable to the Schwarzschild radius because the echoes would directly distort the main signal from the merger which is what has been validated so nicely and what made LIGO so celebrated.

They want the time delay for the echoes to be a bit longer – but not too longer – (the ringing fades away before it starts to ring again) which is why they add the logarithmic factor. I think that this extra logarithmic factor is ill-motivated – note that the argument of the logarithm must really be dimensionless so they mean $$\log(M \sqrt{G})$$, a much less natural quantity. But the timescale around $$GM \log M$$ hasn't appeared for the first time in this paper. People have argued that this is the "scrambling time" of the black hole.

OK, so they decide that there could be some "echoes" and they find them, at the 2.9-sigma confidence level after the look-elsewhere effect was taken into account. LIGO folks are going to verify this claim. I know how to do it but I am not motivated to do it because I just don't believe for a picosecond that the laws of physics could predict such a thing. A 2.9-sigma bump for such echoes might be there or not and even if it is there, I would obviously conclude that it is a fluke.

This is the kind of an extraordinary claim that really needs to get above 5 sigma for me to even pay attention to it or start to verify any detailed argument that was proposed. Sorry.

Fine. So these authors obviously have a more bullish attitude, so they accept the existence of these echoes – let's assume that these echoes exist in the data – and propose an interpretation. It's either a fuzzball or a firewall. Or maybe it is a tooth fairy, an iron man, phlogiston, luminiferous aether, or man-made climate change. Pick your favorite superstition, debunked science, or pseudoscience.

Give me a break. Even if the echo of the gravitational waves were physically plausible, and it's not, their interpretation of these echoes would be another masterpiece of sloppiness. Firewalls and fuzzballs are entirely different things. Moreover, the very point of firewalls is that there is no black hole interior. So if black holes had firewalls, they would be even further from the echo chambers – something you need to create the echoes – than the regular black holes.

So all their ambitious firewall-style interpretations of the echoes are an even more outrageous examples of bullšiting than their original claim that the gravitational waves from the mergers arrive with echoes.

Also, Nature conflates the echoes by Abedi et al. (see above) with another paper by Vitor Cardoso et al. which also talks about possible echoes in the LIGO and similar data. I know Vitor Cardoso via the Internet rather well, he's written numerous papers that were related to my work on quasinormal modes. So we were thanking each other etc. This man isn't a nutcase and this paper he wrote with others isn't stupid, either.

The August 2016 paper by Cardoso is a sort of a fine speculation but it talks about echoes that result from some (also implausible but basically classical) corrections at the event horizon. Those hypothetical oscillations may get out and be delayed in the vicinity of the photon sphere in various ways etc. But they still seem to agree with the general assumption that observers outside the place normally known as the event horizon – such as the members of LIGO – simply cannot see any physics inside the event horizon. I would bet that Joe Polchinski would still agree with this claim, too. And he would agree that the firewall only modifies the physics from the event horizon inwards. So echoes envisioned by Cardoso et al. simply cannot have anything to do with alternative theories of the black hole interior such as the firewalls. Despite the similar words, Vitor Cardoso shouldn't be connected with the utter rubbish claims that the firewall has been observed.

There are some technical bits and pieces in all these echo-or-firewall preprints that are fine but I think that the overall quality standards of this sub-industry is simply poor and it isn't just because of the sloppy work of popular science journalists in Nature and elsewhere.