NYU today put out a press release claiming that Physicists ‘Bootstrap’ Validity of String Theory, telling us that
NYU and Caltech scientists develop innovative mathematical approach to back existence of long-held framework explaining all physical reality.
and
String theory, conceptualized more than 50 years ago as a framework to explain the formation of matter, remains elusive as a “provable” phenomenon. But a team of physicists has now taken a significant step forward in validating string theory by using an innovative mathematical method that points to its “inevitability.”
It’s the usual outrageous string theory hype machine in action, with a university press release promoting a PRL paper (this preprint) with hype and misinformation. This has now been going on for decades, clearly is never going to stop, no matter what.
Years ago I used to comment about this kind of thing that it wasn’t helpful for the credibility of physics in particular, but also science in general. Why should you “trust science” when this is what scientists do? At this point though, the damage has now been done. All over social media you’ll find negative attitudes towards science, with “string theory” given as a prime example for why you shouldn’t trust science or scientists.
I took a look at Twitter (which now seems to come up by default featuring lots and lots of Elon Musk) for the first time in a while yesterday. The consensus on Twitter the past few years has been that string theory is an obviously failed research program, and that the failure to acknowledge this is prime evidence that one should not “trust science”. Doing a search on “string theory”, the latest news is that many people are now asking how this could have happened, with the favored explanation: “string theory is a psyop by the deep state, part of a plot to sidetrack physics and keep us all from having free unlimited amounts of energy”. This is quite a bit less compelling than the older explanation that Edward Witten is an alien sent by a more advanced civilization in order to sidetrack physics.
A few other things I learned from Twitter is that Sabine Hossenfelder has a recent Youtube video String Theory Isn’t Dead. This is about the article I discussed here, and Hossenfelder reaches much the same conclusion I reached long ago about the dead/non-dead question
They say that science progresses one funeral at a time. But it’s no longer true. Because the first generation of string theorists has raised their students who are now continuing the same stuff. And why would they not, these are cozy jobs, and there is nothing and no one that could stop them. So yeah, Siegfried is right. String theory is not dead. It’s undead, and now walks around like a zombie eating people’s brains.
If you look at the few string theorists on Twitter, you find that they are outraged about what is going on. Their outrage though is not about their fellow string theorists discrediting the subject and making science look bad, but at Hossenfelder for pointing to the problem. For a very good discussion with Hossenfelder about her views and all of this, see Curt Jaimungal’s podcast What’s Wrong With (Fundamental) Physics?.
One young string theorist (grad student at SUNY Albany) is trying to fight the anti-stringers, in particular with a new podcast where he interviews Zohar Komargodski. The podcast is well-worth listening to, since Komargodski is a good example of the career path of quite a few prominent hep-th theorists these days and he does a good job of explaining the point of view of current leaders of the subject. While he started out as a grad student doing string theory, he soon turned to other topics, and has done excellent work in non-perturbative QFT of various sorts, very little of it involving strings. Despite this, he would often be described as “a string theorist.” The words “string theory” and “string theorist” now have no fixed meaning, making it very hard to have a serious discussion of the topic.
Komargodski does what he can to put a good face on the impact of string theory, but in some ways is not helpful to the anti-anti-string case the podcaster would like to make:
I’m sure that you know people before my time, way before my time in the 80s, people were claiming that soon enough they will find the standard model in some compactification of the heterotic string and this will explain the electron mass everything else and we’ll be done. There were such claims in the 80s, of course that was premature it turned out to be completely false and as far as we understand it’s not the right direction. So of course making preposterous claims is irresponsible and should be avoided by scientists at all costs because we’re supposed to be responsible for what we’re saying and we’re supposed to be rigorous and careful.
Where I strongly disagree with Komargodski is in his argument that all is well, that we’re just in a typical slow period of progress, that the only problem is that “the theory has yet attained its goal”. This is both bad history and an inaccurate characterization of the situation. String theory is not a research program that is slowly advancing towards its goal of a unified theory (or at least a successful theory of 4d quantum gravity). There has been progress, but it has been consistent progress towards understanding that this can’t possibly work. Komargodski sees no particular problem with the job market: 3-5% of theory PhDs may get permanent jobs, the good ones don’t fall through the cracks but do fine.
While both he and the podcaster have a lot of complaints about the critics and their “bad faith”, they don’t seem interested in doing anything at all about the outrageous hype from their own kind that has done such huge damage to the field already, with more to come.
Update:
The absurd hype is just endless.
Update:
John Baez is both more of an optimist and more of a poet than I am:
And yet, despite having installed string theorists in top positions worldwide, string theory is gradually fading. Physics departments are less likely to hire string theorists than they were 10 years ago – and that was also true 10 years ago. So it seems the tree branch is slowly breaking off the tree, and will eventually crash onto the forest floor, opening up a bit more light for new plants to grow.
I found the paper you linked interesting. Why do you think it’s hype? Don’t you think the results are substantial?
PS you may be interested in
https://arxiv.org/abs/2412.13192
They conjecture that isolated highs dimensional reps of the SM are in the swampland. Thus discovering a new particle in a 5d rep of SU(2) would rule out string theory.
fwiw,
Lots of reasons to not waste time explaining why hype is hype. For one thing I find it hard to believe that I’m not being trolled. The paper you link to is a complete joke, surely it’s April 1.
I don’t think it’s a joke! On the other hand, it’s a slippery slope. The LHC could produce a football when we collide two protons together. This is in the swampland. Shall we search for it?
fwiw,
The phenomenon of obviously stupid claims being made by academics supposed to be experts in the most intellectually challenging of subjects is something that I used to find, like Donald Trump, comical. Now it’s also something no longer funny and I’m realizing I’m going to have to live with this depressing crap the rest of my life.
@Peter
“I’m realizing I’m going to have to live with this depressing crap the rest of my life.”
Totally understand that. I feel considerably better since I am spending more time looking at tech startups. It’s real science and real progress. (Some of it at least.)
That string theorists are now basically repeating the same mistakes from 20 years ago is hugely distressing to me. I have seen so many intelligent young people wasting their best years on this pseudo-research, I was hoping it wouldn’t happen again.
I also recently made the mistake to look at gr-qc and it’s now basically 95% calculations with increasingly wilder and ill-motivated modifications of GR. It’s the same paper-production machine that we’ve previously seen in hep-ph where people invented new particles. Now they invent modified gravities.
Pointless to ask, I suppose, why PRL published that piece of garbage. They will publish anything if the dice roll right.
A cynical question:
Are young string theorists learning the tools and methods they need to advance science, or are they learning the techniques of bamboozling people into thinking they have accomplished something when they actually haven’t? If most of these people aren’t able to do productive research, the fact that they have gotten tenure and are supervising graduate students may hamstring high-energy theoretical physics long after it is clear that string theory is a failed direction.
Hin,
Yes, a big part of the string theory hype problem is the editors at PRL. For decades now they have been both allowing publication of papers specifically intended to make a nonsense case for string theory, and at the same time encouraging authors to have their institutions put out press releases of outrageous hype. You’d think seeing this happen once they’d do something about it, instead it happens time and time again, so now not a bug, but a feature of their editorial process.
Peter Shor,
Yes, this is the depressing state the field has got itself into. The way hiring in academia works, this is not fixable and we’re stuck with it for many decades to come. I do think though that the people involved are mainly bamboozling themselves.
I’m curious what the 95% of theory PhDs who don’t get permanent jobs ultimately end up doing. I got my theory PhD nearly 20 years ago and began working at a small tech/data science company that liked hiring theoretical physicists (because that’s what the founders were). The thinking was that theory PhDs may not come in with the necessary skills to write code and work with large data sets, but they could learn relatively quickly, and there wasn’t a huge pool of applicants who already had those skills.
However, as time has gone on, we’ve seen many more applicants from other domains who DO already have the skills (especially astro PhDs, CS majors, and people with degrees specifically in data science) . As a result, we are much less likely to hire theory PhDs anymore.
GS,
From what I’ve seen it is typical for hep-th phds to get hired in the way you explain, but the fields change over time, as people with specific credentials and experience become more available. I’ve seen first finance, then data science, then quantum computing, now AI.
Many years ago Zohar was a student of mine. But don’t count it against him.
You write:
“they don’t seem interested in doing anything at all about the outrageous hype from their own kind”
What can they do about it? They cannot control other peoples actions, even if these people are “their own kind”. The best they can do is to ignore the hype and concentrate on high quality research.
Peter Shor, the decline in the quality of string theory hires is extremely steep. People who got hired in the 1990s at least understand field theory, even if they have veered into various forms of crackpotism in recent years. But the post 2000 crowd does not really understand quantum field theory and the folks who have been hired more recently do not really even understand equations. As one prominent physicist told me recently, the field is stuck in a mode of “zombie physicists training zombies”.
Udi,
I don’t think the excuse “we can’t do anything and are too busy doing serious work” is valid, especially for the small number of people like Komargodski who are leaders of this field. They could do things that would have an effect, have decided not too.
In this particular case he decided to appear on a 3 hour podcast, to help make the case that critics of what is happening in his field are wrong and in bad faith. During those three hours he could have found time to acknowledge that his field has a hype problem, chose not to do so.
@Attendee:
One way to get rid of these people would be for universities to disband their Physics departments, and move all the people they don’t want to get rid of into an Applied Physics department. This would work — it has been done in other disciplines by some universities who wanted to get rid of a few people they thought were a blot on their reputation — but eliminating a physics department would be a truly extreme measure, which would attract substantial criticism from the media, egged on, no doubt, by string theorists.
Peter Shor,
That’s way too extreme and not a viable solution. Most physicists are working in areas that have nothing to do with this problem, which involves a small fraction of the overall field.
In some sense the answer to the problem is just for those with responsibility to start doing their jobs. That PRL is publishing nonsense and encouraging bogus press releases is the fault of PRL editors and the APS. That the NSF and the DOE are funding worthless research is the fault of NSF and DOE program officers and their superiors. That university physics departments are hiring people doing worthless research is the fault of physics department faculty and science deans.
If no reputable journal will publish it, NSF/DOE/Simons won’t fund it and no one will hire people writing it, papers like the ones discussed in this posting won’t get written. Once that happens, you’ll have some number of not very competent tenured people who can join the already in place ranks of university deadwood with no funding, no grad students, and no postdocs that everyone is quite capable of ignoring.
NYU paper: I’ll be impressed when they can bootstrap mutli-loop or M-theory scattering amplitudes (or find that some S-matrix which can’t come from string theory saturates a bootstrap). This crew has been belaboring similar properties of the Veneziano amplitude to death in recent years and this is kind of a minor extension in my opinion.
@fwiw
Seems like a nicely motivated pheno search since after all such high dimensional representations are virtually impossible to cook up in string theory. Definitely way better motivated than other proposals of LHC searches I’ve seen… The provocative title might not be to everyone’s taste though lol.
Having looked at “How to Falsify String Theory at a Collider” paper, I understand your “April 1” reference. One does not have to be a physicist to appreciate the following line from the abstract:
“This scenario is not realized in any known string construction, and we conjecture that this is true of string theory in general. Detection of this scenario would thus amount to falsifying the (known) string landscape.”
Do they really have no idea what’s different between a conjecture and an established fact?..
NoGo,
This joke is now older than my students
https://www.math.columbia.edu/~woit/wordpress/?p=265
Back when Vafa first started this, Jacques Distler explained that string theory is predictive since you couldn’t find “string constructions” with less than three dimensions. Someone quickly found a “string construction” of this kind. Vafa has always been clear that as far as he’s concerned there’s nothing that can show string theory is wrong. If your current “string landscape” get falsified, that just means that you have to find a way to expand the “string landscape”.
“The consensus on Twitter the past few years has been that string theory is an obviously failed research program,”
That’s it. I’m done! No more string theory for me. I bow to the Twitter consensus of the past few years!
On a more serious note, the idea that youtube likes (or similar) represent a way to think about fundamental theoretical physics is indicative of broader cultural trends. How many people with strong views on string theory know any quantum field theory *at all* ?
The vast majority of physicists are engaged in studying phenomena that are tied to experiment and it doesn’t make sense to punish all of them for the actions of the few.
How does String Theory still survive in academia and whose job is it to shut it down?
One reason for String Theory’s survival is that there is the public perception that people who dabble in it are “smart”. This brings them unexpected allies – for example, there are a small number of condensed matter faculty who will support string theory because doing so puts them with the in-crowd. The string crowd also controls a number of science awards – this occasionally results in the condensed matter people winning attention and awards. A prominent example of this is the “SYK” phenomenon – one can now simply claim that studying a model of interacting fermions is the same as studying quantum gravity and once it gets a couple of awards it becomes a field. You can also mumble the words AdS/CFT and engage in fantastica e.g. model electrons in a material with a black hole and believe that this gives “insight”.
I dont think funding agencies/journals are the right gatekeepers to fix the problem – the people who run them mean well and they rarely have the in house knowledge to “fix” science issues. One can indeed say that the problem with string theory is now so obvious that they do have the knowledge to know that it is humbug – I dont disagree with that, but the agencies/journals might feel that this isnt quite their role from a policy perspective. In general, in a healthy community, you would want them to listen to the community and allocate resources on the basis of that. The issue here is that the community is deeply corrupt and cannot be trusted.
In my view, the solution to this problem really lies with the faculty in theoretical particle physics who know this is all nonsense and yet fail to do anything about it. For example, if you look at the jobs in particle theory this year, there is a position at Johns Hopkins that looks tailored to string theory. There are at least two prominent members of the particle theory faculty there who fully understand the silliness of the string enterprise. But yet, they continue supporting hires in the area, presumably for political reasons.
This is a general problem in the field – somehow people place collegiality above the pursuit of the truth.
Peter Woit said:
“During those three hours he could have found time to acknowledge that his field has a hype problem, chose not to do so.”
He does spend a few minutes discussing “misinformation about string theory” (around 1h57m). His view is that hype statements are extremely vague, making them impossible to debunk. He would rather deal with real criticism where someone points to actual errors.
I tend to agree with him. Just read the comment in this blog post by Attendee about string theory. What is the point of arguing with someone who says “the community is deeply corrupt and cannot be trusted”?
Of course Attendee is vary vague. Who are these string theorists? Does he consider Zohar a string theorists? Does he have any evidence that Zohar is corrupt? Did he even listen to Zohar discussion about string thoery? Does he agree with Zohar’s statements?
If you listen carefully, it’s pretty clear that when Komargodski mentions “misinformation” related to string theory he doesn’t mean hype from string theorists but criticism of string theory. Several times he says that the criticism is either bad faith or trivial.
Udi, there isn’t a whole lot of point in me repeating all of the corruption in the field (e.g. do you really believe entanglement leads to wormholes? Should people win fancy awards for work on this topic? Do you believe that this “result” was “mathematically” established ?) .
The corruption in the field is now so obvious that there isn’t much daylight in the criticism of the field by field theorists like me and insiders such as stringking42069.
The king’s opinions are widely shared in the community and a number of junior string theorists are privately openly derisive of the many “stars” of the field that the king regularly attacks.
I think Zohar has done solid work and my criticism of the field wasn’t directed at him.
Udi,
To the previous two comments I’ll just add about the “why don’t you point to actual errors” criticism that the problem with string theory is not that it’s a well-defined theory you can test experimentally or for internal consistency. There is no well-defined theory that agrees with what we observe, just a failed 40 year program of hoping to find one. There’s a reason this blog is called “not even wrong”.
By the way, I recently talked to Curt Jaimungal, and he’s interested in organizing a discussion between me and a serious physicist on the topic of the state of fundamental theory/string theory and the increasingly negative attention this has been getting. If interested, contact him, or me if you want and I’ll forward to him.
Attendee,
Claiming that real wormholes are generated by just simulating QC entanglement on a classical computer is clearly ridiculous. But by your logic, this is a reason shut down all QC research. Your criticism should be focused on the people that create this stupid hype. Otherwise, you are just spreading misinformation about theoretical physicists. Your hype is just as bad as the hype you are trying to criticize.
Peter Woit,
Your criticism of the “string theory program” is all pointed at people trying to do string theory phenomenology and hyping their results with claims that we are about to see experimental evidence of string theory.
String theory, 40 years later, is still the only example we have for a toy model of Quantum Gravity. String theory is also a great laboratory for better understand QFT. String theory might never describe real world physics, but I believe that better understanding of QFT is the most viable research direction towards beyond the standard model physics.
Regarding Curt, I am not sure that I am the right person to discuss the state of fundamental theory considering that I was not active in the field for more than a decade.
Udi,
The problem is that the people generating the ridiculous hype about wormholes include several of the most prominent and influential string theorists (Susskind, IAS faculty), and few other people in the field are willing to publicly challenge this. It’s an extreme example, but the long history of this kind of thing, going on to this day, has been a huge problem for the field and its credibility.
My criticism of the string theory program is much wider than you describe. It’s a failed idea about unification and the hype surrounding it has for forty years made it very difficult for people to get funding/jobs to work on other very different ideas. Lately it’s done something even worse, completely discrediting the idea that it’s possible to get anywhere by thinking about unification, since string theory is the only way forward, and that is not working.
I completely agree with you that “better understanding of QFT is the most viable research direction towards beyond the standard model physics”, but not that string theory is a great laboratory for doing this. The cases of better understanding of QFT coming from string theory (e.g. AdS/CFT) have only given limited insight in one very special direction. We need healthy research programs into a wide variety of questions about QFT, not just the narrow ones that have gotten a lot of hype because of origins in string theory research (by the way, Zohar is a good example of someone working on other such questions).
Udi,
If you think string theory is still worthwhile, could you point to one thread of string theory research in the last five years that seems likely to lead to something interesting in physics?
Peter Woit,
I didn’t follow the wormhole/QC saga too closely, but from my understanding it has not much to do with string theory.
You claim that “it very difficult for people to get funding/jobs to work on other very different ideas”, but here we have Zohar as an example that it is possible. I remember attending a string conference in the early 2000’s and David Gross pointed out that about 90% of the talks were not about string theory. So it seems to me that it is definitely possible to work on different ideas. 90% of “string theorists” work on different ideas.
Peter+Shor,
Like I said, I am not active in the field. Last String conference I visited, I remember that conformal bootstrap was one new subject. Again, conformal bootstrap is not string theory, so I am not sure how you would classify it. This goes back to how vague and meaningless your criticism of string theory is.
Udi, Peter Shor asked you a concrete question – what has happened in string theory in the last few years that you think will lead to something. In response you say that you aren’t active in the field and that some people seemed to be working on the bootstrap which you admit isn’t string theory. You then suggest that Peter Shor is being vague. This is a brilliant Trumpian strategy – accuse your accusers of the things you are being accused of.
One continuing thread of investigation is the gravitational path integral for extremal and near-extremal black holes. One finds large quantum fluctuations of their near-horizon nearly-AdS2 throat geometry, that are needed to make sense of black hole thermodynamics as the Hawking temperature falls to zero; and in the supersymmetric theory, there is a robust gap between the BPS spectrum and the quasi-continuum of excited states (relatedly, the classically infinite length extremal throat is rendered finite in the quantum theory). The highly chaotic properties of black hole dynamics also continue to be revealed.
Also in the supersymmetric context, localization techniques in the gravitational path integral have allowed a precise match to the asymptotic expansion of exact partition functions of black hole microstates constructed via string theory dualities.
The evidence from these controlled examples supports the thesis that the gravitational path integral builds an asymptotic expansion of something more fundamental. The gravitational degrees of freedom are collective variables that capture features of the thermodynamics, but not the underlying microphysics. The underlying microphysics in these examples is string theory.
Of course, if one is demanding a connection to experiment then one can claim that none of this is progress. However, formal theory these days occupies a position between pure math and experimentally-connected physics, and is likely to for some time to come. For instance, investigating issues surrounding the nature, meaning and admissibility of saddle points in the gravitational path integral (particularly in the Euclidean theory). These are physics questions.
Emil Martinec,
I’m all in favor of better understanding of the gravitational path integral (especially the Euclidean version). But the main problem with quantum gravity has always been not having a consistent theory of the physical degrees of freedom and short distance behavior. I don’t see how calculations in 2d spacetime or BPS states of a SUSY theory tells you anything about this.
The idea seems to be that these calculations are telling you nothing about the original problem of what the microphysics is. That’s fine, but I don’t see how you can then claim that for the microphysics, you’re just going to go back to the superstring, where you just have a theory in flat 10d, with a host of problem if you try and get flat 4d out of it.
More simply, you’re claiming great progress based on calculations that have nothing to do with the original problem. Can you write down a consistent theory of the microphysics? Or is the research program now: “quantum gravity is just an emergent phenomenon, we don’t know from what, but we’re sure we don’t want to admit that it isn’t our original ideas about string theory”?
Emil Martinec,
I assume you mean the euclidean gravitational path integral. The euclidean path integral tells you the properties of the ground state of a Hamiltonian system when that ground state is non-degenerate or when there is a suitable projector that can be used to project to specific states of interest. In the case of gravity, the Hamiltonian is unbounded from below and its far from clear that the physical states we are interested in (such as black hole) are actually any kind of ground state of the system. Further, the system has a large number of degeneracies (as a result of the Hamiltonian constraint) and there is no real projection one can do to get states of interest. All of these are well known issues with the Euclidean Path Integral – so just the fact that you can compute an integral isnt all that useful when it doesn’t actually connect to physical states of interest. I fail to see what can be learned from such an ill defined computation.
Moreover, once you go from extremal black holes to near extremal black holes there is an inner horizon and with that comes massive back reactions in the form of mass inflation. This is a huge non perturbative change to the geometry and nobody really knows the actual microphysics of what is going on there – so claims that there is some controlled understanding of black hole entropy or whatnot in near extremal black holes are not believable given the massive back reaction.
I do hold you in considerable respect for being one of the small number of string theorists who have publicly pointed out that the entanglement brigade has not actually solved the hard part of the black hole information problem and I wish you the best in your endeavors on this front.
Peter,
I disagree with the assertion that “the main problem with quantum gravity has always been not having a consistent theory of … short distance behavior”. That is ONE of the central issues, but not the only one.
Gravity has the peculiar property that it knows its thermodynamic equation of state, through the laws of black hole mechanics (as well as subsequent developments regarding geometric entropy). Usually one needs to know the microscopics in order to derive the equation of state, which is then input to the thermodynamics. Gravitational thermodynamics determines the equation of state without knowledge of the microscopics, and in particular the mechanism that tames short distance divergences.
This feature is a double-edged sword. It allows us to make some progress on quantum black hole physics without knowing the underlying theory, because the gravitational effective theory knows a remarkable amount. But gravity doesn’t know everything, and eventually one does need to know the underlying theory. An analogy might be Stirling’s asymptotic series; it tells you a lot about the Gamma function, but doesn’t tell you everything about it. The examples I mentioned have the nice feature that the truncation of gravity to a 2d effective theory of S-wave modes captures remarkably subtle and detailed features of the complete theory, in some cases because of supersymmetric cancellations among the degrees of freedom that have been left out, and in others by virtue of the degrees of freedom being kept being the lowest energy excitations (i.e. one has the correct effective theory). To draw another analogy, in the Callan-Rubakov effect one is not claiming that a truncation to S-wave quantum mechanics captures all of the physics of magnetic monopoles, but it does capture the particular physical phenomenon of interest, by isolating the relevant degrees of freedom.
What gravity captures here is an asymptotic series at large charge, while the exact partition function requires an underlying theory, for instance when the charges are such that the horizon area is not macroscopic. String theory provides such an underlying theory, in particular well-controlled examples.
Often you complain, as in your reply here, about the lack of a complete specification of string theory. Yet there are situations (such as those I described) where one can calculate a result without that. As in any complicated system, it is useful to find situations where you don’t need to solve everything in order to understand something; otherwise one can end up understanding nothing, and endlessly complaining about it.
As to why one is led to the full machinery of string theory from an analysis of e.g. 4d supersymmetric black holes, I gave a rationale some time ago in this blog (https://www.math.columbia.edu/~woit/wordpress/?p=8116):
“[4d] N=8 supergravity has 28 gauge fields. There are black hole solutions in supergravity carrying electric and/or magnetic charges sourcing these gauge fields. The extremal limits of these solutions are … strings, branes, and Kaluza-Klein modes — moving within or wrapped around the extra dimensions; or bound states thereof. We discover this by considering the spectrum as a function of the 70 moduli of N=8 supergravity; in asymptotic limits of that moduli space, the black holes become perturbative objects in string theory. So if there is a consistent theory with only N=8 supergravity in 3+1d without all the extra structure of string theory, it has to come with a reason why ALL the charged black hole solutions, which seem perfectly benign and not all that different from the uncharged black hole solutions, in fact are excluded from the spectrum of the theory. If they are not, then we are led back to the full structure of string theory compactified on a torus. One could go further and think about why these objects indeed cannot be excluded, because they will be pair created in external fields with some small but finite probability, etc…”
Finally, please don’t mischaracterize what I wrote. I did not claim “great progress”. I was responding to the question “could you point to one thread of string theory research in the last five years that seems likely to lead to something interesting in physics?” by pointing out (a) a robust and highly detailed set of results about gravitational path integrals, which (b) reproduce known brane bound state degeneracies in string theory. The latter results extend well beyond the domain of validity of the former. As to your question whether the quantum gravity research program has migrated away from string theory toward some amorphous notion of emergence, perhaps it has for some folks (they can speak for themselves); but for me, string theory remains the underlying microphysics.
Attendee,
I am not entirely referring to the Euclidean path integral; for instance, the results on the gap, and finiteness of the extremal throat depth, stem from an analysis of the Lorentzian theory.
I’m not sure why you say that the gravitational Hamiltonian is unbounded below. In asymptotically flat or AdS geometries, the physical Hamiltonian is conjugate to the asymptotic time and is bounded below – the Minkowski or AdS vacuum is the lowest energy state. Perhaps you are referring to the indefinite sign of the kinetic term in the Hamiltonian *constraint* (associated to the freedom of interior time slicing for fixed asymptotic time); this gauge freedom does not affect physical states, which are by definition gauge invariant. There are indeed subtleties in evaluating the gravitational path integral; part of the progress I referred to has been an analysis of the contour of integration of various collective modes, with the earlier string theory results serving as a guidepost to the right answer.
One of the appealing features of the AdS/CFT correspondence is that it tells us that the objections you have raised can be resolved in these examples; gravity is dual to an ordinary field theory with a standard Hamiltonian quantum mechanics in a Hilbert space, and the goal then becomes to ascertain the prescriptions on the gravity side that match this spectrum and dynamics.
The issue of inner horizon instabilities of near-extremal black holes, is another issue of this sort (and to me, one of the reasons to be interested in this line of research). On the one hand, GR analyses indicate that the inner horizon is unstable to perturbations, and ultimately singular; on the other hand, there is a finite-dimensional Hilbert space of near-extremal states, and a conventional (if strongly coupled) dynamics on that state space. So there is prima facie evidence that the problem is not a problem, and the goal shifts to understanding why from the bulk quantum gravity perspective. Which is in part why I think the gravitational path integral calculations represent interesting progress.
Emil Martinec,
Indeed, I am referring to the fact that there is a negative sign in a part of the gravitational Hamiltonian which means that there are a number of states that have the same total energy (equal to zero). Thus, black holes of arbitrary mass have the same total energy and you have plenty of degeneracy in the states. AdS/CFT does not resolve the issue – all it does is to say that a specific AdS background is dual to some CFT and thus if you are interested in perturbations around a specific AdS background, you can compute those from the CFT side. I agree that the spectrum of gravitons is bounded from below (both in gravity and from the CFT side). But the tricks that you are relying on to relate the results from the Euclidean path integral to some statement of the ground state are not robust in this case.
Similarly, with the inner horizon – you end up with singular dynamics on the GR side or some strongly coupled dynamics on the field theory side. You cannot calculate on either side. Thus, the claim that there is some precise calculation that does entropy matching is very far from robust. Now you may believe that there is reason to pursue this direction, but belief isn’t progress.
To be a bit more explicit about the euclidean path integral, for example, in cosmology you can have two very different states – Minkowski with no gravitons (i.e. no radiation) or an expanding universe with lots of radiation. Both of these have the same total energy and thus these are all degenerate states, even though the radiation itself has positive energy. If you do a cosmological Euclidean path integral, you aren’t going to be able to distinguish contributions to your path integral coming from these huge class of exactly degenerate states. Similarly with the AdS black hole, you can have a black hole with some mass or another black hole with a different mass and lots of radiation.
In fact, I am continually amazed that AdS/CFT is held up as some miraculous insight into quantum gravity – all it shows is that a specific quantum state, namely, a black hole hole in AdS is dual to a specific CFT. An actual quantum theory would have to tell you the behavior of all possible physical states such as Minkowski and FRW (where we live) and AdS CFT does not do that. Moreover, all it does is map the IR of GR (which we know very well) into the UV of the CFT – with zero insight on the UV of GR (the actual problem at hand).
Isn’t it high time to regard the inadequacy of this “miracle” and move on to doing new things?
Emil,
The problem is that string theory has been sold for 40 years as the way to unify fundamental physics, including providing a consistent theory of quantum nature of the gravitational degrees of freedom in the real world. But it has turned out to not do any of that successfully, and now you’re telling me that the argument for it is that it’s useful in understanding some conceptual problems with the thermodynamics of gravitational fields. Whenever I try and learn more about this, I find a set of complex calculations in toy models supposedly relevant to long-standing unresolved debates about issues that seem to have no connection to anything measurable, or to any interesting mathematics, or to anything I care about.
The world is full of people working on complicated things of very narrow interest and there’s nothing wrong with that. But here, the disjunction between what is actually going on and the 40 year old promotional campaign has become extreme.
Attendee,
In GR, the Hamiltonian reduces to boundary terms (eg in the ADM formalism). So while it is true that any classical solution will have zero bulk contribution to the Hamiltonian, it is not true that all states have zero energy, because different mass black holes yield different values for this boundary term. The same will be true if we add some gravitons to the mix. This is why I said that the physical Hamiltonian is conjugate to the asymptotic time. All this is basically the Gauss law of the gravitational field – that you cannot “hide” energy/mass; it is always reflected in the asymptotics of the metric.
Consider asymptotically AdS spacetimes. The laws of black hole mechanics tell us that above the threshold for forming black holes, there is a well-defined density of states given by the exponential of the black hole entropy. If you want to put some gravitons outside the black hole, that’s fine but one is then looking at a highly subdominant region of the phase space because the black hole solution maximizes the entropy for a given energy (having positive specific heat in AdS), and in the process of adding the gravitons you have to reduce the mass of the black hole to keep the total energy fixed. This leading asymptotic of the density of states matches precisely with the density of states of the dual field theory.
As for inner horizons, singular dynamics on the GR side simply indicates that GR is incomplete and one needs its UV completion there. For this to affect the calculations I was describing is equivalent to saying there are some important degrees of freedom that have been left out of the effective action for near-extremal black holes, outside the outer horizon. I don’t believe that’s the case (this has been explored in the literature), but perhaps this is not the best place to have that discussion. As I was saying, part of the magic of black hole thermodynamics is that it doesn’t need to know about what’s going on inside the outer horizon, and whatever UV completion of GR is needed to describe it. In this way, GR outside the horizon provides constraints on that completion, which is in part why the Euclidean path integral (involving the analytic continuation of the black hole exterior) is so useful.
As for empty Minkowski space vs radiation-dominated (open) cosmology, these are vastly different situations, because you have changed the asymptotics of the fields. It is simply not true that they are degenerate states in the same Hilbert space. No finite energy state with Minkowski asymptotics is an expanding cosmology; at any given time in the cosmology, the energy density of the radiation is constant, and the volume of space is infinite.
There are certainly issues for closed cosmologies, where there is no asymptotic time and no corresponding Hamiltonian; string theory so far has little to say about this situation. But again, this is a different topology of solution where the observables are quite different. But as I said, one doesn’t need to solve everything to understand something; it just so happens that the situations where one can say the most currently involve spacetimes with AdS asymptotics.
And no, AdS/CFT is much, much more than a map of the IR of GR to the UV of the CFT. In fact, it is the claim that the UV completion of GR (string theory) is isomorphic to the CFT. It is a duality – a pair of theories claimed to be exactly equivalent in all respects, so that any observable in string theory with AdS asymptotics has an equivalent observable in the dual CFT. Of course the CFT doesn’t contain FRW cosmology, it was only claimed to describe bulk states with AdS asymptotics.
Emil Martinec,
Indeed, if the ADM mass is all that distinguishes different states, you agree that there is a degeneracy equal to the entropy of your gravitational system that all contributes to your path integral. In fact, given the unresolved issues around the inner horizon, the unknown entropy of those degrees of freedom also contribute to your path integral.
Going back to my original point, when you do the Euclidean path integral, you hope to learn something about the ground state of a system when that ground state is not degenerate. Here you have a massive degeneracy, potentially even involving degrees of freedom at the inner horizon. So what exactly are you learning ?
As for black hole thermodynamics : what the Hawking calculation shows is that as long as the black hole geometry is as described by General Relativity, there is an inconsistency with quantum mechanics. I have not seen any actual resolution of this problem, despite claims by the Entanglement Brigade. I find the arguments made by Mathur ( and you) to be clear on this front. So when there is a paradox and if the resolution may be something quite different (fuzzballs, gravastars, firewalls etc.) , what exactly are you learning from a calculation that is based on a geometry that is fundamentally suspect?
With AdS/CFT – the issue is exactly as you say : it is “claimed” that be exactly equivalent on all respects. But the UV of GR is unknown. The IR of the CFT is not calculable. So nothing of interest can actually be calculated with any degree of confidence. The duality so far has been useful for mapping CFT computations to the IR of GR which we all know how to do. It’s been over 26 years since AdS/CFT. What have you learned about the UV of GR from it? In fact, why would you expect the UV to even care about the sign of the cosmological constant?
I am signing off with this comment since I am headed on holiday. It’s likely best to argue this sort of thing in person rather than the internet.