Another reason is that I didn't see much new that I would overlook during the year – which is probably normal for those of us who diligently follow (not only) hep-th on the daily basis. But after some inspection, it became clear to me that it's not just because of my regular arXiv habits. The conference just didn't really cover most of the stringy craft. The holes were obvious both in the topics and the list of participants.
Look at the list of 494 participants. Search for Stanford which is, you know, a powerful string hotbed in the Bay Area. You will find Shenker and Saad – because Douglas Stanford has Stanford as its last name. Add Santa Barbara. Just two participants, Gross and Maxfield. Zero from Santa Cruz. Just one from MIT, Harlow.
Princeton gives 18 (IAS+Univ) and Harvard 5 hits, Oxford 7 and English Cambridge 7, reasonable numbers, indeed. They make the absence of California even more striking.
But the timetable with talks paints an even more obvious picture of "whole missing branches" of string theory. All of the talks are about some French-speaking style general complexity-thermodynamics-AdS-attempts-on-quantum-cosmology issues, with some CFT rather disconnected from the string vacua. There was also some swampland and the weak gravity conjecture. Some people probably think that I should be happy but I am not.
What seems to be missing is the whole "full-blown string industry". Actual vacua in more than 4 dimensions, their stringy/M-theory origin, higher-dimensional field theories, compactifications geometric and non-geometric, what string theory allows you to do with all these extra dimensions and how, string phenomenology, anything that depends on fancier tools where string theory actually becomes a specific theory allowing something but not something else etc. There are lots of perspectives to take to enumerate what is missing but just to make the hole I perceive obvious, let me say that I see
- no Calabi-Yau or other manifolds
- no type I,IIA,IIB
- zero M-theory, zero F-theory
- one talk with something "heterotic"
- one talk with "D-branes" but the high spacetime dimension etc. isn't the point; no higher-dimensional matrix models
- nothing like Cvetič et al. or Taylor et al. F-theory model building
- well, Vafa wasn't there, and the same holds for "his junior collaborators", I think
- Andy Strominger gave a talk about a self-similar behavior of the photographed M87 black hole, entertaining but clearly not string theory
- nothing like Acharya and his proof of SUSY from classification of Ricci-flat manifolds
- two talks have "super" in the titles but almost anything where supersymmetry really matters in some way are missing
It is just a bad trend. It almost looks like someone was trying to make the conference more compatible with those who aren't really string theorists – or even those who dislike string theory.
Experienced TRF readers must be able to predict that I wouldn't be capable of resisting the analogy with something else linked to Brussels: the "multicultural" immigration. Brussels finds it convenient to import lots of easy-to-manipulate people from the Muslim world who are easier to govern because they don't demand "luxurious" things like freedom and democracy and they aren't attached to any European nation state. So it promotes "multiculturalism" which is a euphemism for an uncivilized "monoculturalism" where everything incompatible with Islam is being rather quickly suppressed.
The composition of the topics at the string conference looks analogous. The full-blown string theory just doesn't seem to be welcome, appreciated etc. – I would like to know whether these omissions are due to the organizers' ignorance of the field or their malicious intent. The increasing fraction of non-string participants and talks suggests that "everyone" has the right to go do a string conference, much like "everyone" has the right to migrate to Europe. It's totally unhealthy and self-harming. It's existentially dangerous. It's extremely harmful to deconstruct the prestigious character of the string conferences.
It almost looks to me as if the nasty anti-string crackpots were co-organizing the conference and could veto talks if not participants.
Now, to make similar points in a more specific context, let me discuss a 26-minute-long video
The general feeling is that the junior people don't have any visions. It may be due to design and pressures, due to the selection, due to something else or a combination, but the gap is clear. But let's not be too quick.
At the beginning, Nima asks about the unitarity in Stanford's talk. Stanford starts to say seemingly complex boring things about "it depends what you mean by unitarity". OK, I am not familiar with every page of Stanford's papers but I just don't believe that the word "unitarity" has become this muddy. Unitarity is the property of an operator producing some evolution or transformation, \(UU^\dagger=1\), usually combined with the assumption of a positive-semidefinite space on which this operator acts. If there's a Hamiltonian, its Hermiticity is equivalent to the unitarity. Unitarity may require appropriate conditions for the normalizability of the external states etc. But I just don't see the room for all this fuzz. "What do you mean by unitarity?" You should damn know what unitarity is.
A minute later, it turns out that Stanford is computing some "averages over theories". You may compute such averages but you cannot live in a world that is averaged over theories. Well, decades ago, Sidney Coleman was trying to create a counterargument with some baby universes etc. but I would insist on my statement. An ensemble is OK but the "theories" must still be considered mutually exclusive and one can still ask whether each of them is unitary or not. So I concluded that the whole paper or talk by Stanford is probably gibberish – assuming some logically inconsistent generalization of quantum mechanics. The "averaging over theories" sounds like some multiculturalism, too. The context where they studied it was apparently the SYK or JT model. They're really toy models where such basic things should be particularly clear. I can't believe that they end up with this fog about the simple question whether the "unitarity" is obeyed by Stanford's talk.
At any rate, this foggy technicality has nothing to do with visions in string theory. Gross surely had the same feelings as I just mentioned so around 4:00, he explicitly asked about the desired future according to practitioners who are young i.e. below 35-40. Harlow who was falling asleep in the recent minute seems surprised or shocked by the question. I mean this is a panel about visions about string theory. Why are you there if you are shocked by such a question? Gross said that he surely had clear dreams – masses of hadrons etc. – but the Millennials don't seem to have anything like that.
Harlow doesn't know what to say, "it's a recipe for saying things that are wrong". "It's a wish," Gross says instead. You can't really do important science without having wishes, or if you're afraid of saying things that will be wrong. Harlow acts like he is calculating what form of self-consorship will be viewed as the most politically correct one. He has nothing authentic to offer here. OK, he ends up with a lukewarm proposition "I still think string theory is a strong candidate" and "it would be nice to get some support, probably from cosmology". If in 40 years, there's no deep advance like "at least calculating the cosmological constant" (wow, that already counts as a modest plan), Harlow "hopes to work on something else". Perhaps authoring petitions to harm careers of ideologically inconvenient people is Harlow's idea about this "something else"? Other junior folks add some "computable early Universe that would be nice" and move their heads in unexcited ways.
Harlow asks Iceberg whether we're just on the tip of a Seiberg, or vice versa, in QFTs, and whether fractons totally revise what QFTs are. Seiberg has lost some sleeps over fractons... but why here and why now? This is a string conference and fractons have been discussed in condensed matter physics at least since 1983. Some people with a more particle physics background may talk about fractons now – because they're doing things closer to condensed matter in general – but why is it interesting and what does it have to do with visions? Some particle physicists are just joining a topic in an adjacent field that is some 35 years old. I don't find it exciting, I don't consider these joiners "pioneers", and I don't think that fractons are likely to be important for the fundamental laws of Nature.
Seiberg doesn't know what fractons are but suggests that several icebergs exist. Seiberg switches to exact solutions of some field theories. Seiberg says one should be as ambitious as possible but also have short-term plans. Alday asks Seiberg whether dualities will be obvious and Seiberg says that dualities must be by definition surprising. Right except that what is surprising to start with may become obvious once you understand more. Seiberg says that a change of variables is by definition "not surprising", I agree with that, too. So the electromagnetic duality is unsurprising in the free Maxwell theory but not in \(\NNN=4\).
What about T-duality, is it surprising? It looks surprising in the spacetime but in string field theory, it is a field redefinition – and it is also a "free electromagnetic duality" on the world sheet. So T-duality probably isn't surprising according to Seiberg. But what about the rest of U-duality group in maximal supergravity (which also includes S-duality subgroups)? It may be fully generated by several subgroups isomorphic to T-dualities. The problem is that T-duality was proven above in a weakly coupled limit only. The full U-duality goes beyond that – but it may still be fully proven e.g. in the BFSS matrix model, up to some number of compact dimensions.
While never considered a full-blown string theorist, Nima actually fixes the discussion and offers some stringy visions. It could be possible – an advance could be around the corner – to figure out whether the large field inflation is possible or impossible within string theory. Well, we were already working on such things 15+ years ago (a full no-go theorem should have been found before BICEP2, Nima insists). Also, maybe someone should have predicted the discrepancies in the value of the Hubble constant, Gross adds while Nima agrees. Existing models implying the numerous values look like contrived mess.
Nima still sees a big gap between "string theory as the unified theory on steroids with a single theory of quantum gravity" and "the picture where every QFT has its dual quantum gravity". We've discussed this dilemma after it was voiced e.g. by Steve Shenker already around 2000. I don't see any big tension here. What is unique is the fully formulated theory but it still has many solutions and the solutions may only become unique or close to unique once you try to minimize the curvature and decompactify as much as possible. So 11D M-theory and five 10D string theories are unique or almost unique but there are many compactifications to 4D etc. – either the regular Calabi-Yau compactifications or theories that may be linked to generic CFTs by AdS/CFT which are generically "Planck radius compactifications". As Nima said in a fresh Wolchover's article, the de Sitter symmetry is an asymptotic one for the future – when things flatten out. Flattened things are like "IR limits" and "universal behavior" and they are almost unique while things with lots of curvature become very non-unique and it's actually a pretty good thing. Doesn't it make sense? If you try to decompactify the AdS theories coming from any CFT via AdS/CFT, you still get one of the well-known maximally dimensional string/M-theories.
Fine. Nima did mention my answer, showing he was aware of it. It's the compactification and/or high, near-Planckian curvature that brings non-uniqueness. But, he adds, take a 9-dimensional theory with the SU(17million) and a horrible matter content. Is it anomalous? Well, Nima, it doesn't have the usual basic types of anomalies but it's almost certainly in the swampland because string theory implies lots of inequalities etc. In some sense, they come from anomaly considerations, too. Take heterotic strings, the maximum rank of the gauge group is bounded from above, right? It comes from the critical dimension – compactified dimensions increase the rank – and the critical dimension comes from the conformal anomaly. So string theory imposes all kinds of inequalities. The matter content is "bounded" in a way that is morally related to the critical dimension – although, in F-theory, the gauge groups may be really large and numerous. But the Hodge numbers of Calabi-Yau three-folds are probably also bounded. It's a tough mathematical question but there may also be a crisp physical derivation of such mathematical inequalities. Lots of extra conditions like that – you know, the swampland conditions – exist to make sure that even in the absence of knowledge about the precise vacuum, string theory is predictive. But without the information about the particular vacuum, string theory doesn't predict "everything".
Gross responds by saying that the Ising model is probably dual to something but who cares? Not all dualities are useful. I agree with that. If we had the full definition of string theory, we could in principle answer this question, but if we don't have the full definition, we can't answer it and I personally don't care because the answer doesn't look important for anything else. The Ising model is either a non-solution of string theory or a solution that is clearly irrelevant for particle physics phenomenology. Is there a big difference between the two? You could add the Standard-Model-like qualitative features as extra conditions on top of equations of string theory, and then the answer would be "it is a non-solution".
OK, Gross would still like to see the QCD string. I don't believe there is an illuminating string theory construction that produces the pure QCD. And if there is one, my emotions remain almost as low as in the Ising model case. Gross probably has some extra emotional attachment to pure QCD but you know, David, it's just another toy model for the rest of us, just like the Ising model!
I personally don't think that a string vacuum exactly equivalent to the Ising model is an exact solution of string theory. But maybe it is one. We don't have the complete equations of string theory, of course. I don't think that the answer is too important because in physics, we also need at least a conceptual agreement with the reality, like 3+1 nearly flat, large dimensions and gravity. It's only with these additional "experimental" constraints when it's important for string theory's space of solution to become semi-unique, and I find it extremely likely by now that the right vacuum won't be quite unique.
The junior people don't seem passionate about any of this.
Questions are asked from the audience. New fog emerges about the phrase "equivalence class of QCD". Like in the case of "unitarity", I find it strange that such basic phrases create so much confusion. "Equivalence" is a strong word so all theories "equivalent" to QCD really are QCD (or its hypothetical dual[s]). The man probably meant a wider class but he should have defined it beyond misleadingly simple adjectives such as "equivalence". The phrase "equivalence class" was apparently started by Seiberg and he meant "trivial to solve". What does "trivial to solve" have to do with "equivalence"? Random confusing things are said about random cousins of QCD or their classes. I have no idea how such a chaotic discussion could be inspiring. Gross is generally annoyed that Seiberg expressed the view that QCD was boring, although he tried to soften that, but I do agree that QCD is boring – relatively to things that would be "visions".
Witten suggests finding the dual of any confining cousin of QCD, or asymptotically solvable generalizations of \(\NNN=4\) etc. Hmm, there are some intermediate softening of the conditions.
An Indian guy asks about hopes to solve the Ramond-Ramond backgrounds of string theory. I personally don't find it conceptually hard. In principle, you may add condensates of all these messy picture-changing operators etc. Also, in Berkovits' pure spinor formalism (and the Green-Schwarz variables in the light-cone gauge, if possible), Ramond-Ramond fields just may be allowed to start with. I don't see RR fields as more conceptual than a technicality that is hard in some particular world sheet variables – but still doable in a weakly coupled string theory. Gross says that we understand the \(\NNN=4\) holographic dual of a Ramond-Ramond background. Right.
A question was asked about non-locality in string theory, mentioning Ashoke's talk. Sen's violation of locality wasn't due to local non-analyticities but due to some behavior in the whole complex plane, Gross argues. Harlow mentions AdS/CFT as a striking non-perturbatively created non-locality of bulk AdS theory – the non-locality coming from the string interactions is even enough to change the target space dimension. I agree with that.
Aside from commenting on some technicalities, the Millennials don't seem to have visions.
A few vague comments about the computation of parameters. Gross thanks for great talks, dance, and the food. Maybe they also drank urine from the little pissing boy of Brussels, it had to be delicious.
OK, to wrap up this display of the murder of the passions and enthusiasm, I think that the field is being systematically diluted, stripped of its concentrated flavor, self-confidence, prestige, renaissance men, and of its dreams about the future. I think it's mostly being done by pressures from outside – combined with the absence of balls inside (especially because I think lots of excited people still do full-blown string work but they weren't participating at the conference at all, so the "organizers" tend to amplify harmful trends) – and it's a model for what is happening with the whole Western civilization. The political correctness is what is ruining both – string theory as well as the Western civilization.