Monday, January 04, 2016

Finding space-time quanta in the cosmic microwave background: Not so simple

“Final theory” is such a misnomer. The long sought-after unification of Einstein’s General Relativity with quantum mechanics would not be an end, it would be a beginning. A beginning to unravel the nature of space and time, and also a beginning to understand our own beginning – the origin of the universe.

The biggest problem physicists face while trying to find such a theory of quantum gravity is the lack of experimental guidance. The energy necessary to directly test quantum gravity is enormous, and far beyond what we can achieve on Earth. But for cosmologists, the universe is the laboratory. And the universe knows how to reach such high energies. It’s been there, it’s done it.

Our universe was born when quantum gravitational effects were strong. Looking back in time for traces of these effects is therefore one of the most promising, if not the most promising, place to find experimental evidence for quantum gravity. But if it was simple, it would already have been done.

The first issue is that, lacking a theory of quantum gravity, nobody knows how to describe the strong quantum gravitational effects in the early universe. This is the area where phenomenological model building becomes important. But this brings up the next difficulty, which is that the realm of strong quantum gravity is even before inflation – the early phase in which the universe blew up exponentially fast – and neither today’s nor tomorrow’s observations will pin down any one particular model.

There is another option though, that is focusing on the regime of where quantum gravitational effects are weak, yet strong enough to still affect matter. In this regime, relevant during and towards the end of inflation, we know how the theory works. The mathematics to treat the quantum properties of space-time during this period is well-understood because such small perturbations can be dealt with almost the same way as with all other quantum fields.

Indeed, the weak quantum gravity approximation is routinely used in the calculation of today’s observables, such as the spectrum of the cosmic microwave background. That is right – cosmologists do actually use quantum gravity. It becomes necessary because, according to the currently most widely accepted models, inflation is driven by a quantum field – the “inflaton” – whose fluctuations go on to seed the structures we observe today. The quantum fluctuations of the inflaton cause quantum fluctuations of space-time. And these, in return, remain visible today in the large-scale distribution of matter and in the cosmic microwave background (CMB).

This is why last year’s claim by the BICEP collaboration that they had observed the CMB imprint left by gravitational waves from the early was claimed by some media outlets to be evidence for quantum gravity. But the situation is so simple not. Let us assume they had indeed measured what they originally claimed. Even then, obtaining correct predictions from a theory that was quantized doesn’t demonstrate the correct theory must have been quantized. To demonstrate that space-time must have had quantum behavior in the early universe, we must instead find an observable that could not have been produced by any unquantized theory.

In the last months, two papers appeared that studied this question and analyzed the prospects of finding evidence for quantum gravity in the CMB. The conclusions, however, are in both cases rather pessimistic.

The first paper is “A model with cosmological Bell inequalities” by Juan Maldacena. Maldacena tries to construct a Bell-type test that could be used to rule out a non-quantum origin of the signatures that are leftover today from the early universe. The problem is that, once inflation ends, only the classical distribution of the, originally quantum, fluctuation goes on to enter the observables, like the CMB temperature fluctuations. This makes any Bell-type setup with detectors in the current era impossible because the signal was long gone.

Maldacena refuses to be discouraged by this and instead tries to find a way in which another field, present during inflation, plays the role of the detector in the Bell-experiment. This additional field could then preserve the information about the quantum-ness of space-time. He explicitly constructs such a model with an additional field that serves as detector, but calls it himself “baroque” and “contrived.” It is a toy-model to demonstrate there exist cases in which a Bell-test can be performed on the CMB, but not a plausible scenario for our universe.

I find the paper nevertheless interesting as it shows what it would take to use this method and also exhibits where the problem lies. I wish there were more papers like this, where theorists come forward with ideas that didn’t work, because these failures are still a valuable basis for further studies.

The second paper is “Quantum Discord of Cosmic Inflation: Can we Show that CMB Anisotropies are of Quantum-Mechanical Origin?” by Jerome Martin and Vincent Vennin. The authors of this paper don’t rely on the Bell-type test specifically, but instead try to measure the “quantum discord” of the CMB temperature fluctuations. The quantum discord, in a nutshell, measures the quantum-ness in the correlations of a system. The observables they look at are firstly the CMB two-point correlations and later also higher correlation functions.

The authors address the question in two steps. In the first step they ask whether the CMB observations can also be reproduced in the standard treatment if the state has little or no quantum correlations, ie if one has a ‘classical state’ (in terms of correlations) in a quantum theory. They find that for what already existing observables are concerned, the modifications due to the lack of quantum correlations are existent but unobservable.
    “[I]n practice, the difference between the quantum and the classical results is tiny and unobservable probably forever.”
They are tentatively hopeful that the two cases might become distinguishable with higher-order correlation functions. On these correlations, experimentalists have so far only very little data, but it is a general topic of interest and future missions will undoubtedly sharpen the existing constraints. In the present work, the authors however do not quantify the predictions, but rather defer to future work: “[I]t remains to generate templates […] to determine whether such a four-point function is already excluded or not.”

The second step is that they study whether the observed correlations could be created by a theory that is classical to begin with, so that the fluctuations are stochastic. They then demonstrate that this can always be achieved, and thus there is no way to distinguish the two cases. To arrive at this conclusion, they first derive the equations for the correlations in the unquantized case, then demand that they reproduce those of the quantized case, and then argue that these equations can be fulfilled.

On the latter point I am, maybe uncharacteristically, less pessimistic than the authors themselves because their general case might be too general. Combining a classical theory with a quantum field gives rise to a semi-classical set of equations that lead to peculiar violations of the uncertainty principle, and an entirely classical theory would need a different mechanism to even create the fluctuations. That is to say, I believe that it might be possible to further constrain the prospects of unquantized fluctuations if one takes into account other properties that such models necessarily must have.

In summary, I have to conclude that we still have a long way to go until we can conclude that space-time must have been quantized in the early universe. Nevertheless, I think it is one of the most promising avenues to pin down the first experimental signature for quantum gravity.


Joy Christian said...

I sent one of my related papers to Maldacena last year after I saw his Bell paper on the arXiv. In my opinion my result completely kills his idea. To my surprise, he kindly replied to my email, but he remains committed to his idea.

This is the paper I had sent him:

Ivan Kaidashenko said...

I think an interesing question arises: Are there macroscopic(!) quantum-gravitational effects? As is well known, there are macroscopic quantum-mechanical effects in the lab, e.g. the magnetic flux quantization is a quantum phenomenon in which the magnetic field is quantized in the units of h/2e. Is there a similar gravitational flux quantization?

Phillip Helbig said...

"But for cosmologists, the universe is the laboratory. And the universe knows how to reach such high energies. It’s been there, it’s done it."

In the words of Zel'dovich, the poor man's accelerator. :-)

" But the situation is so simple not."

Recently Star Wars seen have you?

Uncle Al said...

"The energy necessary to directly test quantum gravity is enormous." 1 joule/gram divergence in ΔH_(fusion) = 110.6 J/g at 95 °C. Two differential scanning calorimeters; 96 benzil 20 mg single crystals, half each in enantiomorphic space groups P3(1)21 and P3(2)21; one day. Run simultaneous ΔΔH_(fusion) every 30 minutes for 24 hours. Test spacetime geometry with geometry.

ΔΔH_(fusion) = 0. Falsified conjecture.
ΔΔH_(fusion) ≠ 0, sinusoidally varying with time of day. Equivalence Principle geometric violation. Spacetime is a left foot toward hadronic matter (embedded opposite shoes then melted into identical socks). Rewrite theory.

"But if it was [were] simple, it would already have been done." Physics postulates failure. Look.

Sabine Hossenfelder said...


It's a good question, people are certainly looking for macroscopic effects. As to the flux quantization, it's an in-medium effect. There isn't any "medium" which could play a similar role for gravity. Topological defects are an option (but haven't been found). Best,


Sabine Hossenfelder said...


I think you are asking a somewhat different question. The papers I wrote about, they ask whether CMB observations can distinguish between a qft origin and some other origin. You are asking a more fundamental question, that is whether quantum mechanical correlelations themselves could have another origin. So I see that it's a relevant question, but I don't think it's what the papers even aimed at studying. Best,


naivetheorist said...
This comment has been removed by the author.
Pentcho Valev said...

Why should Einstein's relativity and quantum mechanics be united if the former harbors a wrong concept of time (spacetime)?
"And by making the clock's tick relative - what happens simultaneously for one observer might seem sequential to another - Einstein's theory of special relativity not only destroyed any notion of absolute time but made time equivalent to a dimension in space: the future is already out there waiting for us; we just can't see it until we get there. This view is a logical and metaphysical dead end, says Smolin."
"Was Einstein wrong? At least in his understanding of time, Smolin argues, the great theorist of relativity was dead wrong. What is worse, by firmly enshrining his error in scientific orthodoxy, Einstein trapped his successors in insoluble dilemmas..."
What scientific idea is ready for retirement? Steve Giddings: "Spacetime. Physics has always been regarded as playing out on an underlying stage of space and time. Special relativity joined these into spacetime... (...) The apparent need to retire classical spacetime as a fundamental concept is profound..."
"Rethinking Einstein: The end of space-time (...) The stumbling block lies with their conflicting views of space and time. As seen by quantum theory, space and time are a static backdrop against which particles move. In Einstein's theories, by contrast, not only are space and time inextricably linked, but the resulting space-time is moulded by the bodies within it. (...) Something has to give in this tussle between general relativity and quantum mechanics, and the smart money says that it's relativity that will be the loser."

Pentcho Valev

Lucy M said...

Bullseye. So you shouldn't be looking for that, or thinking about it, at all.

If the convergences are not there you are not going to find it, and that which you do find will only serve to confuse and misdirect even more than it already is, the scientific instinct and in the process very possibly Science itself and the best hopes of humanity - the very opposite of where your heart is.

You're looking for relativity in the 1680's, or 1790's or 1840's, who knows how far premature, but too far back for now; said differently the convergences are too premature for now. Had they directed the resources of science THEN toward THAT instead of having the good sense (the scientific instinct) to wait, then they would not have found THAT. They would have found.... philosophy. And the philosophers would have rushed back in, just as they are doing now. With the priests, sages & dodgy looking Gandalfs not far behind.

They suspected a fundamentally relative character in Newton's time. Newton almost certainly thought that would eventually be proven true. Thank goodness for Newton's genius that he saw that such knowledge was distant from his time, and far from guaranteed even to future generations. But he saw that he could do something to change the odds for the future.

And that is assuming the quantum gravity thing is even true. Which cannot and must not be assumed and almost certainly is not true. Not in a sense that would leave any part of it, recognizable to me and you.

Lucy M said...

Sorry - the above comment was contextualized to be answering this excerpt: "The biggest problem physicists face while trying to find such a theory of quantum gravity is the lack of experimental guidance"

Doesn't make any sense without that. Which is not to say it makes any with it.

naivetheorist said...

i'm updatng a previous commenrt that i've now deleted.


i just came across an article on Einstein's views contra field theory. it's

"The Other Einstein: Einstein Contra Field Theory" by John Satchel
published in Science in Context 6 (1), 275-290 (1993).

for people who can't access the article themselves, i can send them a pdf of the article if they contact me at

(sabine, i hope its okay to give my email address so others can get this very interesting, almost entirely unknown, article).

for those who would like to see lecture on the subject at the Perimeter Institute, it's available at:

i hope this encourages to think about this - i figure that if einstein thought a non-field approach hade value as early as 1916 and as late as i954, it's worth thinking about. if so, you might also want to look at the causal net model of Rafael Sorkin and the recent blog "what is spacetime, really?" by Stephen Wolfram.


andrew said...

While classical General Relativity should be an approximate limit of quantum gravity, is seems to me that this assumption can be taken too far and that there are probably multiple ways in which quantum gravity could give rise to phenomena quite distinct from GR that is capable of being observed macroscopically, even at significantly less than Big Bang energies.

It would be interesting to know, in general, what sort of phenomenological predictions follow from various alternative quantum gravity theories. Even if there isn't a stark black and white distinction, if two or three independent observations were more probable in a quantum gravity theory than classical GR, that ought to give us considerable comfort that we are on the right track.

Sabine Hossenfelder said...


I'm not sure what you mean with 'taking to far' this assumption. You better reproduce GR in the limits that we've tested it or your theory belongs in the dumpster. Yes, it would be interesting to know, alas, there are only phenomenological models filling the gap between theory and experiment. Which is why I keep repeating that it's important to develop these. Best,