Today’s Washington Post has an opinion piece from Brian Greene, running under the demonstrably false title Decades later, string theory continues its march toward Einstein’s dream.
In the piece, the argument of string theory critics is given as:
Critics argue that the situation is untenable, noting, “If you can’t test a theory, it’s not scientific.” Adherents counter, “String theory is a work in progress; it’s simply too early to pass judgment.” The critics retort, “Forty years is too early?” To which the adherents respond, “We’re developing what could be the most profound physical theory of all time — you can’t seriously cross your arms, tap your foot and suggest that time’s up.”
The problem with the results of forty years of research into string theory is not that progress has been too slow but that it has been dramatically negative. To see this, one can just compare the text of chapter 9 of Greene’s 1999 The Elegant Universe, which has an extensive discussion of prospects for testing string theory by finding superpartners, fractionally charged particles, or cosmic strings. Twenty-five years later, the results of experimental searches are in: no cosmic strings, no fractionally charged particles, and most definitively no evidence of superpartners of any kind from the LHC.
The other sorts of predictions advertised in that chapter are based on the idea that string theorists would better understand the theory and be able to make testable predictions about neutrino masses, proton decay, axions or new long range forces, the nature of dark matter, and the value of the cosmological constant. Instead of progress towards any of these, things have gone in the opposite direction: all evidence from better understanding of string theory is that it either naturally predicts things in violent disagreement with experiment (wrong dimension of space time, huge number of new long-range forces, …) or predicts nothing at all. 25 years later, Greene now goes with the latter:
The challenge for string theory is that it has yet to produce any definitive, testable predictions.
The article goes on to make a different case for string theory:
… string theory continues to captivate seasoned researchers and aspiring students alike because of the remarkable progress that has been made in developing its mathematical framework. This progress has yielded provocative insights into long-standing mysteries and introduced radically new ways of describing physical reality.
For instance, string theory has provided unmatched insights into the surface of black holes, unraveling puzzles that have consumed some of the greatest minds, including Stephen Hawking. It has offered a novel, though controversial, explanation for the observed speedup of the universe’s expansion, proposing that our universe might be just one of many within a larger reality than conventional science ever imagined.
The problem here is that these supposed advances aren’t from advances in string theory. If you follow the link above that justifies “string theory has provided unmatched insights into the surface of black holes”, you’ll find the text:
Most physicists have long assumed it would; that was the upshot of string theory, their leading candidate for a unified theory of nature. But the new calculations, though inspired by string theory, stand on their own, with nary a string in sight. Information gets out through the workings of gravity itself — just ordinary gravity with a single layer of quantum effects.
The string theory “explanation” for the value of the CC is just the “anthropic” explanation, which besides not really being a scientific explanation, has nothing to do with string theory.
The piece ends with something highly speculative and ill-defined (ER=EPR) that has nothing to do with string theory:
Roughly, it’s as if particles are tiny black holes, and the entanglement between two of them is nothing but a connecting wormhole.
If this realization holds up, we will need to shift our thinking about the unification of physics. We have long sought to bring general relativity and quantum mechanics together through a shotgun wedding, fusing the mathematics of the large and the small to yield a formalism that embraces both. But the duality between Einstein’s two 1935 papers would suggest that quantum mechanics and general relativity are already deeply connected — no need for them to marry — so our challenge will be to fully grasp their intrinsic relationship.
Which would mean that Einstein, without realizing it, may have had the key to unification nearly a century ago.
Where string theory research is after 40 years is not on a continuing march forward towards “Einstein’s dream”, but in a state of intellectual collapse with no prospects of any connection to the real world, just more hype about vague hopes for something different, something for which there is no actual theory.
Update: I recommend checking out Sabine Hossenfelder’s latest youtube piece, which is mostly devoted to the Brian Greene wormhole publicity event stunt discussed earlier here. Near the end of the video, she tells a story that explains a lot about why this kind of thing keeps going on (see here). She had been writing a regular column for Quanta Magazine, but they stopped publishing her after she wrote a column in which she argued that physicists should not be misleading the public by claiming that the “black holes” supposedly on the other side of a duality from a given quantum system were actual physical black holes.
What she was warning about in 2019 is an essential part of the wormhole publicity stunt, and of other similar continuing efforts that have been going on for years. One impetus behind this nonsense has always been clear: there’s attention to be gotten, and money to be made. Another factor though is the one Hossenfelder explains here. Press outlets devoted to science want to publish positive news about scientific advances, want nothing to do with authors who explain that this positive news is nothing but hype.
Read this WaPo “column” about 5am local time today. Been waiting for your response. Thank you. Was upset that they allowed such without a reasoned rebuttal.
EPR=ER seems to be the latest rage but as Bob Wald has aptly said all these ideas to save quantum unitary in black hole evolution have a price tag of modifying quantum theory to the point of nonrecognition. ( not his exact words) EPR=ER seem to me to be one of ugliest and ad hoc solutions to a problem it’s not really clear there needs to be a solution for. Much more sensible in my opinion are the viewpoints of Wald, Unruh and Thorne and the late James Hartle who argue(d) that the black hole information problem isn’t a real problem at all. What is the real problem are all the ad hoc theories trying to solve this problem.
What is the duality in Einstein two papers of 1935 that he is talking about ?
martibal,
Einstein Rosen paper on wormholes versus EPR paradox paper.
The problem with string theory is not just that it doesn’t predict anything per se (which it doesn’t) but that in addition it doesn’t make any existing theories more parsimonious or useful. For all kinds of technical reasons the unification part has failed. You are left with the quantum gravity part, which probably has worked better than any other approach, but still not that well at all. If moving away from point particles will always cause this much trouble it would be interesting to reflect upon why that is ..
I find the “we just need more time” argument particularly hard to understand.
There were roughly 50 years between the discovery of QM by Heisenberg and Schrödinger, and ‘t Hooft and Veltman. In those 50 years QM led to many interesting things: the Dirac equation, anti-particles, the neutrino, QED, the development of gauge theories, the description of the Higgs mechanism, the standard model … not to mention the Mössbauer effect, Raman spectroscopy, lasers, developments in solid state physics and so on. The standard model is a success, but it is far from the only success of QM.
It’s now 40 years since the first superstring revolution. Physics’ finest worked on it, uninterrupted by world wars. And as far as I understand, absolutely nothing comparable happened in string theory. In those circumstances, you have to have seriously strong guts to claim that you just need more time and everything will be fine.
I am a retired biophysics professor. This behavior just astounds me. I grew up in an environment where you went overboard to be careful that what you said was either true or you had to carefully qualify it. If I said something in a lecture or paper that turned out to not be quite right, I was mortified. If, for example, I had made the easily checked misstatement about the black hole link that you discuss, I don’t think I could ever live it down. What is happening? Has there been a dramatic change in the academic scientific ethos, or is there something about this field that attracts this kind of behavior.
I think I’ve been following this for long enough that I could simulate the next couple steps of the argument. Greene could say, for example, that the anthropic “explanation” for the CC isn’t completely unrelated to string theory, since string theory gives you a context within which the dimensionless constants of the SM could vary in eternal inflation. He could also say that ER=EPR and the recent progress on the black hole information problem came from the string community, and plausibly wouldn’t have emerged without that path (even if, at the end, the strings can be removed like scaffolds). You, in turn, could reply that it’s nonsense to talk about a “string theory mechanism” for varying constants without a nonperturbative definition of the theory, and also, string theorists said 30 years ago that their goal involved unification of the forces and predictions confirmed by experiment, so why do they get to change the goal midgame to “quantum gravity insights from which the strings can be removed like scaffolding”?
I’m not convinced that the recent “progress” on the black hole information paradox holds water. Does anybody other than the authors really understand it?
If you know the results that you want to come up with, and you do enough non-rigorous manipulations of vaguely related quantities, you can almost always come up with the desired results. How sure are we that this isn’t where these results came from?
Do you think the Washington Post would accept a countering opinion piece?
Bob Zanelli,
My problem with ER=EPR is not that it’s ugly and ad hoc, but that I don’t see any actual calculation of anything that corresponds to what the Washington Post is advertising as the discovery that “particles are tiny black holes, and the entanglement between two of them is nothing but a connecting wormhole.”
As far as I can tell, the problem is not that the theorists doing this have a theory that is ad hoc, but that they have no theory that is supposed to correspond to the real world.
Peter,
I agree, but want to again emphasize that there has been progress in our understanding, but it’s been in the opposite direction of that indicated by the Washington Post headline, which much more accurately should have been
“Decades later, string theory continues its march away from Einstein’s dream.”
David,
I’m also astounded by what is going on here. 25 years ago I understood what Greene and others were doing as understandable enthusiasm for some specific ideas about physics (even though I had my doubts about them). Considering everything that has happened since then, I just don’t understand at all why anyone involved in this would think it’s a good idea to now go to the public claiming all is well and doubling down with even much more outrageous hype than that of 25 years ago.
Scott Aaronson,
Yes, you qualify as a pro at this. That’s exactly right.
Philippe David,
I have no idea if the Washington Post would be interested in publishing such a piece. I do know that I feel I’ve done more than my part trying to deal with this kind of hype, so I’m not interested in being the one to work on arranging for and writing something. Among the many others who know enough about the subject to know what is seriously wrong here, I’d hope there are some who care enough about this to try and do something about it.
It seems increasingly clear that the very foundation of the theory, that fundamental particles are actually made of strings, is wrong or biased from the start. And this would be the reason why this theory is full of oddities or leads nowhere despite all the efforts made.
When you have taken a wrong turn from the start, you must have the humility to retrace your steps and start again in a more promising direction.
When you plug simple numbers in ER=EPR, you see that it cannot work https://arxiv.org/pdf/2002.08178 . A wormhole connecting two entangled electrons would have a mass of 10^27 eV (in order to explain the entanglement entropy). Already grossly excluded with simple experiments.
Scott says,
“…why do they get to change the goal midgame to “quantum gravity insights from which the strings can be removed like scaffolding”?”
“Peter’s” above hypothetical critique would already indeed be a valid one. But now let me add mine: why are we even granting that these are actual “quantum gravity insights” in the first place? When the hypotheses are so wildly speculative and poorly justified, the actual concrete models so far from anything resembling reality, the derivations so full of hand-waving, then, I think, we are more than allowed to be skeptic of even that basic point.
There may be something of interest from the quantum computing and quantum complexity point of view (I don’t know, you tell me), but from the quantum gravity one (the hard problems are well known, are decades old), I see very little substance, as per usual.
We need to start to put the bar much more higher to what we call “quantum gravity” these days, otherwise almost anything is now passing for it. Of course, since there’s no accepted theory or even accepted direction, that’s difficult. My two cents are: let’s stick to the basic and well known issues that appear when you try to apply the current QFT formalism to GR and build from that. What I see now is that people invent their grandiose theories or approaches, with very little justification for their hypotheses and then try to apply those the the mentioned issues; they then proceed to sell this as insights into quantum gravity. Problem with this is that there as many ways of doing that as grandiose theories or approaches the human mind can conceive (let’s not throw LLMs into that mix, please!) Even the black hole information issue (one of the main motivations behind many of these proposals) needs a lot more understanding at the QFT in (classical) curved spacetimes regime before clearing things enough to see if we can even get “QG insights” about it. Bob Wald’s (and also Unruh’s) objections to its status as a supposed paradox that is unresolvable in standard QFT in curved spacetimes deserve more attention, in my opinion. I’m not saying he’s right, but his takes on this are much more technically naunced and well informed than Susskind’s ones, which have been much more influential in some communities in theoretical physics (particularly, the commnunity that gave rise to the subject being discussesed in this blog entry)…
Peter or someone else, how is fractionally charged particles evidence for string theory?
A side note: It feels like this “40 years” time span has been with us for quite a while. The first revolution was 40 years ago but people have been talking about string theories for at least 10 years by then.
I find the objections to this article slightly unreasonable. The argument of Greene’s article is 1) string theory has made no experimental predictions 2) it has made deep and serious mathematical advances 3) one such, still conjectural and not understood even mathematically, are the concepts of duality 4) here is an example of such, which if physically applicable will have physical implications.
As a pure mathematician I can get behind all of those statements, and their consequence — string theory is not nonsense. It has failed, so far, at explaining anything about the physical world. But the many serious physical intuitions that have informed it — intuitions that are (were?) unavailable to pure mathematicians — have produced surprising and deep progress, with consequences in algebraic geometry and representation theory.
It is likely that any description of the physical world that involves mathematics will use these consequences – algebraic geometry (equations) and representation theory (symmetry) are very basic. That is not to say that string theory will be the foundation of any physical understanding, a statement about which I am at best agnostic.
You may find argument (4) overhyped by Greene, but I hope you do not differ in the rest of this assessment.
anon,
You’re making a different argument than the one in article, which is that string theory is making progress towards “Einstein’s Dream” (which was not to do math). My post here is responding to the argument actually made.
I am ignoring the claims in the article about mathematics (which are overhyped) since the relation of string theory and mathematics is a complicated and separate subject.
As far as I can tell, there is no significant new mathematics involved in “ER=EPR”, whatever that is.
Looking at the situation locally at Columbia, there was a period during the 1990s when there was significant joint activity between math and physics here, much of it organized by Brian. That stopped a long time ago, as his interests moved away from mathematics to purely physics issues.
Einstein, Albert, and Nathan Rosen. “The particle problem in the general theory of relativity.” Physical Review 48, no. 1 (1935): 73.
https://www.jp-petit.org/papers/cosmo/1935-Einstein-Rosen.pdf
Einstein, Albert, Boris Podolsky, and Nathan Rosen. “Can quantum-mechanical description of physical reality be considered complete?.” Physical review 47, no. 10 (1935): 777.
https://utf.mff.cuni.cz/virtualni_vyuka/NPOZ008/FJDP_2023/EPR-PhysRev47(1935)777_preklad.pdf
Shantanu,
The idea that string theory models naturally give fractionally charged states goes back to this Wen-Witten paper from the very early days (May 1985) of the string theory unification via Calabi-Yau story, see
https://doi.org/10.1016/0370-2693(92)91557-P
It was prominently featured in The Elegant Universe, but I’ve never seen anyone else in recent decades promoting this as a “prediction” or “test” of string theory.
The WP has a “Today’s Opinions” summary of its op-eds, written by Drew Goins, which described Greene’s editorial as follows:
Frontiers in physics
Math and physics professor Brian Greene’s op-ed isn’t rocket science. At least you can test rocket science.
No, Greene is writing about string theory, the most fantastical and unprovable frontiers of science. As he explains for the uninitiated masses, string theory entirely holds up on its internal mathematics, but to actually test whether the outcomes it predicts are true, experimenters would need to re-create the conditions at the heart of black holes or at the moment of the Big Bang.
Does string theory being untestable mean that string theory is unusable? Absolutely not, Greene writes. Of course, if some flaw were to appear within the theory, or some experiment actually came along to debunk it, Greene would abandon it posthaste: “I’m an advocate for truth, not string theory.”
But, until that moment, string theory is still unlocking fabulous maybe-truths about the universe. Let Greene’s words wash over you: wormholes and black holes, “holographic duality,” quantum-entangled particles that behave the same no matter the distance between them. Even if you don’t fully understand them, they might still inspire awe … and buy-in.
anon,
That this piece is not science but advertising trying to get “buy-in” is right. A good question for the Washington Post is why they are publishing advertising masquerading as science.
Scott,
“He could also say that ER=EPR and the recent progress on the black hole information problem came from the string community, and plausibly wouldn’t have emerged without that path (even if, at the end, the strings can be removed like scaffolds).”
Imagine in 40 years time, Gil Kalai is proven right and QC is still not feasible. This would be like some other member of the QC community rebutting Gil Kalai by saying, “Well, Scott Aaronson’s work on fingerprinting AI outputs is instrumental in AI safety and that came from the QC community and plausibly wouldn’t have emerged without that path (even if, at the end, quantum computing could be removed like scaffolds.” In other words, I give you the non-sequitur.
PS: I’m old enough to remember when ER=EPR was introduced as fanciful big conjecture / brainstorming and even Susskind seemed sheepish in putting it forward given all the handwaving. Now it is treated like a big breakthrough??
@Bob Zanelli, @Alex:
You might be interested in the interview Wald gave in 2020 to the AIP history project where he talks about these issues a bit:
https://www.aip.org/history-programs/niels-bohr-library/oral-histories/47228
“He could also say that Newtons gravitation came from the epicycle community, and plausibly wouldn’t have emerged without that path (even if, at the end, epicycles can be removed like scaffolds).”
Always keep in mind that Editors of Newspapers and Magazines can alter an author’s submitted article. We don’t always know precisely how the sausage was made.
Brathmore,
The points made in the article are substantially the same as the material in the new edition of The Elegant Universe that was just published, see https://www.math.columbia.edu/~woit/wordpress/?p=14092
A logical guess as to why this piece is being published now in the Washington Post is that it’s a byproduct of the publisher’s effort to promote the new book.
An analogy to the wormhole story I have been thinking lately: a lot of integrable systems can be described in terms of vector bundles on twistor space. If I create something in the lab in a setup where it becomes well-approximated by one of these integrable systems, did I just create twistor space in the lab? That would be a good headline: “Physicists create twistor space, a six (real) dimensional space, in the lab.”
One could plausibly say that knot theory came from Kelvin’s vortex model of the atom, which was an attempt at a theory of everything from the 19th-century perspective, unifying electromagnetism, atomic physics, and possibly gravity, in terms of the dynamics of the ether. Ultimately, it failed as a theory of physics, made obsolete by relativity and quantum mechanics. Only knot theory remained after. Maybe string theory will meet a similar fate, in the end.
I could be wrong, but I get the impression that Mr. Greene is trying to protect physics from public backlash. I don’t really know, but it seems like the Physicists are trying to pivot away from string theory without saying so. Break the news to the public that they’ve been BSing you slowly and gently.
Maybe that is for the best. But going forward scientists should be much more cautious in their “science communication” and let the politicians campaign for funding.
Bob,
I don’t think you can create twistor space in the lab. On the other hand I do think we actually live in twistor space and experience it directly (space-time points are indirectly observed quantities, when you open an eye you’re experiencing directly a CP1 in twistor space…).
Robert James Parkinson,
Protection against public backlash is a big goal here, but I don’t think it’s physics in general that he’s trying to protect.
Hi Peter, thanks for the link. I can’t believe we’ve been going on about this for almost 20 years!
Adam Treat: People, including me, HAVE made the argument that quantum computing research can partly be justified by the new things it’s led to in classical theoretical computer science. Watermarking of AI outputs is a terrible example, since that has nothing to do with quantum computing except for the involvement of a couple of the same people (e.g. me), but, say, lower bounds for triangle counting and locally decodable codes, lattice-based cryptosystems, new characterizations of counting classes, advances in lower bounds by polynomials, speedups for classical optimization problems via Quantum Monte Carlo running on classical computers, and much more have come from QC. But I’ve also been very clear that none of this can substitute for the big enchilada, the actual building of a scalable QC and its use to get some benefit over a classical computer for some problem that people care about.
Alex: One could take the position that, e.g., the black-hole information paradox isn’t worth discussing at all without a detailed theory of quantum gravity in which to do calculations. But if one *does* treat it as worth discussing, then I think it’s inarguable that AdS/CFT and “It from Qubit” have now completely changed the terms of discussion. E.g. we now know that the same Euclidean path-integral hocus-pocus that led to the prediction of Hawking radiation in the first place, also leads to the prediction that the radiation encodes the infalling information (even though it doesn’t provide a microscopic description of how). More generally, there’s now a toy setting where (1) spacetime is emergent from a more fundamental description and (2) black holes are understood in the conventional framework of QM.
There’s a difference between praising something that acted as a catalyst to discover other things that ultimately don’t depend on the catalyst, and praising the catalyst as a success on its own terms, when the other discoveries ultimately aren’t contributing to the continuing unsolved problem the catalyst was supposed to solve. A constant argument people make in grant applications here in Australia is for the whole “well, it contributes indirectly to blah blah blah, unforseen consequences etc”. But try telling the ARC that your grant money was deserved because you didn’t achieve anything your research application said you would address, but someone else has later borrowed some techniques and solved some other problem. Research inspiring other research that solved other problems is a basic given. If your research is so insular it can only somehow address one single problem and never connect to anything else, then that is a really risky situation. So praising a program for basically having a bare minimum feature every productive scientific or mathematical theory has, feels like pretty poor change to me.
David Roberts,
In addition, in many cases the supposed new progress is as much or worse hype than the original string theory hype, promoted by the same people. The most egregious example would be the wormhole publicity stunt, promoted as “quantum gravity in the lab”, inspired by ads/cft/string theory. With no consequences for the original string theory hype, there’s now sometimes a doubling down on even more outrageous hype.
Some kind of consequence is the low number of faculty positions in hep-th. This looks like the worst job market in a long time:
https://particle.physics.ucdavis.edu/rumor/doku.php?id=current
Of course this predominantly affects young people. But most departments don’t want to hire in this field anymore (and haven’t for a while).
Alex,
Unfortunately I think that’s right. You can get the public to continue to buy books by doubling down on the hype, but there is then a blowback effect that you lose credibility in your wider academic field, with people being unwilling to hire in your area.
I’m quite serious when I say that I see this subject as now being in a state of intellectual collapse. One is seeing much fewer significant new ideas than at any time in the past. String theorists over the years have often said “you can see the field is healthy because it is attracting talented young people”. But if there are no jobs and nothing is happening, these talented people move on to something else.
I just found the WPO article (FB showed it to me a week late), and as soon as I saw who had written it, I knew what to expect. Your blog is where I went to find some sanity, and did. Along with, of course, Sabine’s video. Just want to say “thank you”, from a retired theoretical nuclear physicist (Iowa State, then JHU/APL).
Have you tried asking Quanta for the reasons why they stopped publishing Sabine, or will you simply take her assumptions and word for it?
Martin G,
I don’t think it’s hard to figure out what Quanta’s point of view is on this, and I don’t see any reason to doubt what Sabine says. If you look at what they do and what she does, they are quite different things. She’s also not the compromising sort who would change what she does to fit with what they want to do.
Their big new project about quantum gravity is a very good example of how different what they do is from what she does. I’ll try and write something about that soon.
Interesting that Quanta is taking sides in the string wars. I still consider what they do extremely valuable in terms of providing access to information about current research for those who aren’t trying to read all the latest journal articles. But if they get too selective it seems that they run the risk of providing a skewed picture of what the research community regards as important. And they also seem to be getting a little sloppy, e.g. a recent, otherwise well written piece on the development of group theory and it’s use in physics, contained the statement “At high energies, group theory can be used to show that electromagnetism and the forces that hold atomic nuclei together and cause radioactivity are all manifestations of a single underlying force.“ It’s a bit ironic that after your nice non-technical description of the experimental failure of GUTs re. the recent podcasts, that they’d be so sloppy. Maybe I’m nit picking but for a publication that aims to provide good information on current developments in science that statement felt problematic.
Karl Young,
That Quanta quote was from someone writing about math, not very well informed about physics. You could probably find similar examples of people writing about physics, saying not very well informed things about math.
Like the Simons Foundation, they’re focused on certain elite institutions, so for instance tend to reflect the point of view of people in Princeton at the IAS, and want to make the scientific community look good. So, rather different than Sabine Hossenfelder.
I wouldn’t describe Quanta at this point as “taking sides in the string wars”. In their latest series of articles about quantum gravity, they more or less ignore string theory. This fits well with what’s going on at places like the IAS, where quantum gravity research has long ago moved away from string theory (even if it’s hard to get them to admit it…).