Forty Years of String Theory

The journal Foundations of Physics has been promising a special issue on “Forty Years of String Theory: Reflecting on the Foundations” for quite a while now, with a contribution first appearing back when it really was 40 years since the beginnings (more like 43 now). The final contribution has now appeared, an introductory essay by the editors (’t Hooft, Erik Verlinde, Sebastian de Haro and Dennis Dieks).

The overall tone of the collection is one of defensive promotion of the subject. The fact that string theory’s massively overhyped claims to give a unified theory of particle physics have led to miserable failure is mostly completely ignored. From the introductory essay one would never guess that string theory was ever supposed to have something to do with explaining the Standard Model of particle physics and that there were hopes that it would find some sort of vindication at the LHC, perhaps via the discovery of SUSY (the LHC is not even mentioned in this essay). String theory is presented purely as a theory of quantum gravity that has led to new insights in mathematics and had various other applications through the dualities it has uncovered. It’s main shortcoming is described as

the lack of directly testable experimental predictions that would signal ‘string physics’

which seems to me intentionally misleading, implying that string theory makes indirectly testable predictions. The problem with string theory is that it makes no predictions about anything, not that it only makes indirectly testable ones.

Three of the eleven articles in the collection are described as representing critics of string theory. The first, from Carlo Rovelli, does do a good job of explaining many of the problems of string theory. Lee Smolin’s contribution is not much about string theory, but more an examination of the general issue of the “Landscape problem”, comparing a range of different theories in which the laws of physics are different outside our observable universe.

’t Hooft’s On the Foundations of Superstring theory calls for more attention to the lack of any fundamental description that tells us what string theory really “is”, taking the point of view:

we conjecture that the “true theory” is something totally different from superstring theory (and certainly also different from gravitating quantum field theories), but that string theory may approximate the truth to various degrees of accuracy in one or several of its compactified realizations, just as it does for some condensed matter systems and QCD.

He ends with an argument (which he notes is “one where only few readers will follow me”) that one problem with string theory is that it uses the conventional quantum formalism, which he feels is flawed, needing replacement by an “emergent” version of quantum mechanics. For more about the sort of thing he has in mind, see here.

Two articles by philosophers of science, Dean Rickles and Richard Dawid address the question of how to evaluate a supposedly scientific theory that, like string theory, makes no experimentally testable predictions. Both pieces seem to me to suffer from a rather uncritical attitude towards various forms of string theory hype. For Rickles, the dominance of string theory can be justified by its “mathematical fertility”, for Dawid the justification is “the assessment of scientific underdetermination” (roughly, there aren’t any other good ideas). That it has led to some interesting mathematics and that there’s not a lot of good alternative ideas out there are perhaps the two best arguments for pursuing string theory, but in both cases the situation is far more complicated than string theory advocates would have one believe.

The articles by string theorists (Balasubramanian, Giddings, Gubser. Martinec, Susskind and Duff) have a range of interesting things to say, sometimes amidst large dollops of string theory hype. Almost all evade serious discussion of string theory’s failure to say anything about the Standard Model (although Susskind argues, a la Multiverse, that this a positive feature of string theory). Giddings perhaps makes the most serious criticism of string theory in the entire volume, discussing its problems as a theory of quantum gravity, where other authors see a big success and the theory’s main selling point.

The article by Duff is by far the most bizarre thing in the volume, and I wrote about it extensively a year ago here. As Duff sees it, the problem is just that critics of string theory are misguided and misinformed. He includes a three page denunciation of Garrett Lisi which has nothing to do with string theory, characterizes the major recent research directions in string theory as fluid mechanics and the black hole/qubit correspondence, and has an appendix about the press release Imperial College put out making absurd claims that he had finally figured out how to make predictions from string theory (see here). The editors of the volume seem to be rather defensive about publishing such a thing, noting

Needless to say that the opinions expressed in this paper are entirely the author’s own and that it is not our intent to spark new popular or otherwise heated discussions.

but justifying it as

we are happy to include this paper in our special issue as addressing questions that are important not only to scientists but also to the wider public, which was among our initial intents.

and ending with

We warmly recommend Duff’s very readable and playful contribution.

Nothing about Duff’s piece struck me as “playful”, but that the editors see it as some sort of joke would explain why they thought it worth publishing.

Update: Over at The Browser, Steven Gubser recommends that people should read The Elegant Universe and four string theory textbooks. Asked about the “no predictions problem”, Gubser does his best to mislead, claiming the situation is just like that with QED that Feynman got the Nobel Prize for. As for SUSY, if the LHC finds it, that’s evidence for string theory, if not, no problem. There’s the old favorite “the LHC might produce microscopic black holes”. About whether string theory makes testable predictions about the heavy ion physics the LHC is studying

String theory might predict that such and such number is one, and the experiment might say well it’s about two, but it could instead be one. That’s the kind of accuracy with which things can typically be done.

This entry was posted in Uncategorized. Bookmark the permalink.

31 Responses to Forty Years of String Theory

  1. Bob Jones says:

    “String theory is presented purely as a theory of quantum gravity that has led to new insights in mathematics and had various other applications through the dualities it has uncovered.”

    Most people would say that string theory is an idea about quantum gravity. During the past fifteen years, most string theorists have been using the theory as a tool to study very general conceptual issues in quantum gravity and to understand the relationships between different quantum field theories. You can complain all you want about the lack of testable predictions in particle physics, but these objections seem pretty irrelevant since most string theorists aren’t trying to do phenomenology and since string theory has achieved so much success in other areas…

  2. OMF says:

    I think Theoretical Physicists need to take a time out for a few months studying spinning tops.

  3. Friend says:

    I don’t know how String Theory could have ever been considered as fundamental. It seems to me that a fundamental theory will have to explain why the universe is quantum mechanical to begin with. And ST is only an added layer on top of QM; it does not even attempt to explain where QM came from. It seem that alone should have raised suspicion against its claim of being a fundamental theory of everything.

  4. Girlfriend says:

    It soberly is true, what you say, Friend.

  5. Peter Woit says:

    Sorry, but those who want to discuss why they’re unhappy with quantum theory need to find another place to do this, it’s off topic here, unless it’s specifically about ‘t Hooft and his attempts to bring string theory into it.

  6. Bernhard says:

    One thing I don´t get it is why there was no contribution from Witten? Perhaps he refused?

    String theory is already in pieces and while there is no need to beat a dead horse, the fact that the bunch of crap Duff wrote actually got published as a representative contribution to celebrating 40 years of string theory is good evidence of the moribund state of the theory.

  7. A.J. says:

    Maybe Witten was busy writing the roughly 400 pages of detailed technical notes on superstring perturbation theory that he just put on the arxiv?

  8. Bernhard says:


    Since is Witten we are talking about, I´m sure he could have done both.

  9. no one of consequence says:

    Epicycles were an example of “mathematical fertility”. Yet they were wrong.

    For a supposed whole branch of theoretical physics to consume a generation of minds / carreers, shouldn’t there be something of substance to even hint at relevance in the measurable physical universe?

    Back in the late 70′s in the Berkeley Physics department, I was a dubious observer who inelegantly suggested that studies of this kind belonged elsewhere in the philosophical realm, until such hints presented themselves. Since that time I’ve seen nothing to change that inate skepticism.

    And while I can understand the collective embarassment not wanting to cede ground, perhaps BSM and other grand schemes … should just be pursued when we have a long confirmed discrepancy, as in the past? Rather than assumed as to be present?

  10. fuzzy says:

    dear no one of consequence, i feel that neutrino masses and dark matter are recent achievements that do not belong the standard model. i am not fully sure of what to think about the issue of strong CP, but it is also a stimulating point where we can proceed experimentally. several other appealing issues in cosmology (inflation, present day accelerated expansion, speculations on the origin of the matter etc) have been also clarified since 70′s. i would say, very few of them have received the slightest contribution from the stringy ideologists — an exception is goodman-witten’s contribution on direct dark matter search.

  11. fuzzy says:

    ps perhaps this is the reason why witten is not in the book?

  12. Chris Oakley says:

    These denunciations of epicycles fail to take account of the fact that an ellipse is a circle with a single epicycle

  13. Bob Jones says:

    no one of consequence,

    Has it ever occurred to you that there might be some good reasons for doing string theory? Do you really think the subject has survived for forty years with no results? Do you really think string theorists are that dumb?

  14. Peter Woit says:

    One can argue about what if any role “dumb” plays in the string theory story, but this discussion is fairly deep in the “dumb” category. I realize that this is by now a very tired subject, but still… There’s a lot of material in this volume, if you read some of it and have an interesting comment, please contribute it, otherwise, please spare the rest of us…

  15. Peter,

    It seems to me that String Theory does make at least one definite prediction:
    Like any other mainstream theory of quantum gravity (that I have heard of), it predicts that the gravity field of an object in a quantum superposition is also in a superposition.

    While I don’t see the significance of this pointed out very often, to me that constitutes a profound claim about the world which has as yet not been directly tested. I think the importance of this may be minimized by three factors: First, it is probably at the level of present technology impossible to perform a direct unequivocal experiment to test it; second, there are good indirect reasons (e.g. conservation of momentum coupled to the fact that objects in a quantum superposition are affected by gravity fields in exactly the right way) to expect that if such an experiment could be performed, the prediction will be confirmed and third, there is no mainstream rival hypothesis which makes a different prediction (and any framework which does so would automatically be considered non-mainstream).

    But, given that this claim enjoys such a central position at the core of any quantum theory of gravity, should we not refuse to settle for anything less than a direct test before we dismiss any doubts about its correctness? Rather than testing predictions which rule out regions of the parameter space in which the models based on a framework are still viable, yet leaving virtually infinitely many possibilities open at higher energy scales, would it not be far more definitive to identify a falsifiable prediction which does not permit any such adjustment? Surely, if this central claim were found to be false, it could not be compensated for by modifying string theory because that is part of the very essence of quantum gravity. Changing *that* aspect of string theory is to kill it.

    That means a falsified result would lead to nothing short of a scientific revolution, and it helps to keep in mind that these usually happen when something that was universally regarded as “obvious” turned out to be a false assumption about nature.

    Ideally, recognizing this claim as a prediction should have spurred the development of new experimental techniques by which one might eventually be able to test it. Alas, who is going through all the trouble of tackling a practically impossible experiment for the outcome of which no one expects a surprise?

    Sometimes I think that it sure was a good thing that the Michelson Interferometer was not impossibly difficult to construct, otherwise the aether might have stayed with us much longer than it did. In fact, it does not seem preposterous to me to imagine that parameter adjustments analogous to what one sees today might even have allowed an aether theory to survive (probably in a very complicated form) up to today as the dominant space-time paradigm.

    But, we were lucky that it is in fact relatively easy to build a Michelson Interferometer. We do not appear to be so lucky when it comes to building a device that could measure the existence of gravity fields in a superposition.

    I’d be interested to know whether you consider the superposition of gravity fields a prediction of string theory and if not, why.



  16. Peter Woit says:


    This really has nothing much to do with string theory but is generically about quantum theory, so I don’t want to encourage discussion of this here. Generic questions about quantum gravity are a huge topic, one I’m not very expert in.

    The way you’re trying to relate this to string theory is kind of like the way I’ve seen some prominent string theorists argue with people who say string theory is not falsifiable, by saying that if people observe violations of the axioms of QM in tabletop experiments, that would falsify the current understanding of string theory. First of all, my guess is that if a tabletop QM violation was found, there would quickly be papers out there explaining it with some exotic version of string theory. Secondly, this is all a bit like saying “my theory is falsifiable, because if God emerges from a collision at the LHC with a sign saying my theory is wrong, that would show it was wrong”. The falsifiability criterion is intended to refer to distinctive aspects of a theory that differentiates it from others, not generic properties common to all known theories.

  17. Bob says:

    Yes if QM was violated then string theory would be dead. If local lorentz invariance was violated, then it would be dead too.

    On the latter point, there are alternative theories out there, such as LQG, that predict violations of lorentz invariance, and these claims have been falsified. So string theory certain makes predictions that distinguish it from other theories. Other theories commonly predict a breakdown of lorentz invariance, and some violate quantum mechanics or the equivalence principle badly, and have been ruled out. So certainly the evidence is pointing towards string theory at this stage, but who knows what future experiments will reveal?

    Also, Peter what do your favorite theories of QG predict, other than what I have mentioned so far?

  18. Peter Woit says:


    I don’t have a “favorite” theory of quantum gravity. The LQG vs string theory hype-filled arguments like the ones you are making about “my theory sucks less than your theory” just don’t interest me at all.

  19. Bob says:

    Peter, okay. But just a question: do you know of any predictions coming from any of the QG theories? (I am mainly curious about theories other than string theory here).

  20. Peter Woit says:


    I don’t know of any “prediction” from a quantum gravity theory that isn’t an abuse of the term.

  21. fuzzy says:

    hi bob, a good example of prediction in the field you mention has been done by one contributor to this book: it was shown that neutrinos can be superluminal, and the doubts raised by the other investigators were irrelevant.

    please, try to imagine the impact of this paper on experimentalists and on a wider public, in the moment when the experimental claim was made. then, judge by yourself the scientific value of such a “prediction”.

  22. Yatima says:

    Off-topic but …

    The epitaph of an anomaly which was pretty much improbable to begin with:

    And another interesting anomaly vaporizes into its error bars:

  23. Peter Woit says:


    That’s not the only volume contributor who weighed in on superluminal neutrinos. According to Mike Duff, string theory could explain them (although he did say he didn’t believe the result or that string theory was the explanation, just that it could be…)

  24. fuzzy says:

    the “prediction” i quoted is published: i mean, several colleagues have pondered and decided to leave their findings to posterity, independent editors agreed this was useful, some referee implied in judging, a lot of readers, and all that. but i agree that mod phys lett is less authoritative than prl, thus your duff probably wins the context (even if, counting citations, my smolin should do, and the stringy prediction published on prl concerns something else)

  25. Peter,

    Thank you for your response. I won’t prolong the discussion on this but let me just note that ultimately, when the consistent response to contradictory empirical data is that only a subset of of all possible models based on a framework has been falsified where the entire set may well be infinite, then it seems to me one has to consider falsification at a more generic level, so in that sense I do see this question as quite relevant to string theory.

    And, in my view, your last sentence, that this is an aspect of “all known” theories is an overgeneralization. It is just a feature of all current mainstream approaches to understanding the relation between quantum theory and general relativity.


    It appears to me that you are implying that failure to observe gravity fields in a superposition is on the same footing as a violation of quantum mechanics or local Lorentz Invariance.

    We have overwhelming evidence that the latter two are correct descriptions of nature, but we have never observed a gravity field directly in a superposition. We have indirect arguments suggesting that this may be the case, but the superposition of gravity fields is emphatically not on the same footing as standard quantum mechanics or local Lorentz Invariance.

    I don’t know if you did this on purpose, but in the second paragraph you exactly illustrated the string theorist argument Peter mentioned in his reply to my comment. If his argument, that there are exotic versions of string theory which could explain any violations of QM at all, is true, then it seems to me a rather definitive refutation that this constitutes a falsifiability criterion.

  26. Bob says:

    Armin, my understanding of string theory is that it exactly respects QM. Peter is the only person I’m aware of who appears to be advocating alternative string models that somehow violate QM. It is a weird proposal by Peter, but I can’t comment further on his model.

  27. Dim Reg says:


    LQG does not break local lorentz invariance. It was beleived that since it predicted a minimum length eigenvalue, that length must have been invariant (just like how c is the maximum velocity, so it has to be invariant). But that is wrong because the probabilities are transformed, not eigenvalues, like how you can’t just boost away the vacuum energy. I also beleive that one can spontaneously break lorentz invariance in string theory, so those experimental results have told us very little about these two theories of quantum gravity.


    I seriously doubt it is possible to find an exotic form of string theory that isn’t quantum mechanical. Formulating a theory in terms of non-zero commutators and hilbert spaces builds uncertainties and superpositions into the theory, and this is done when constructing string theory. I don’t see how any clever trick could remove either of these things, though perhaps I’m just not clever enough.

  28. Bob says:

    Dim Reg, I don’t know of any serious proposal to break the local lorentz invariance in string theory. This is especially dangerous as it is needed for the theory to carry the general coordinate invariance in the usual way, and it surely carries this. You refer to spontaneous breaking, but that, by definition is a property of low energy or long distance physics, so can’t be relevant for the local, i.e., short distance, physics.

    With regards to LQG, it was repeatedly claimed by Smolin that it would alter the photon dispersion relations, and this claim was falsified. Maybe there are other proposals to avoid this, I’m not sure, but certainly the evidence is disfavoring LQG at this stage.

  29. Peter Woit says:

    Bob and Dim Reg,

    I don’t even know what it would mean for an experiment to “violate QM”. ‘t Hooft claims to have a possible non-QM foundation for the superstring, and it’s well-known that the experts all say that we don’t know what string theory “is”: the issue of its foundations is still up in the air.

    If tomorrow there’s a solid report of an “experimental violation of QM”, which do you think is more plausible:
    1. string theorists wholesale tell the media the theory has been falsified and stop work on it.
    2. dozens of papers start appearing on the arXiv purporting to explain the experimental anomaly in terms of string theory models or “string-inspired” physics.

    Responding to someone asking if string theory makes predictions by saying “no, not now, but I think for reasons X it is the best thing to work on to try to get to a better theory, one that would make predictions” is honest. Saying “string theory does make predictions” and pulling out something like the QM business just isn’t.

  30. lun says:

    People who talk about “violations of QM in gravitational systems” need to understand two issues: Firstly, gravity is only detectable in macroscopic systems, where in any case QM is in any case”violated” through processes such as decoherence (perfectly consistent with quantum mechanics, producing classical-looking results).
    Secondly, no quantum theory of gravity (and that includes string theory) is sure what either the Hilbert space or the observables of quantum gravity are.
    Therefore, it is wrong to think of “violations of quantum mechanics” as if this was a well-defined experimental signature.

  31. Dim Reg says:


    It looks like I stand corrected on string theory. I’m certainly not an expert in that field and your arguments make sense to me. My point with LQG was that it was never a correct reading of the theory to say that it broke lorentz invariance. Lee Smolin was just making LQG hype, trying to say that his theory was better because it made currently testable predictions. Regardless, I don’t intend to champion LQG (I don’t actually think they are right), and this certainly wouldn’t be the place to do that.