The science magazine Nautilus this month has a profile of me, under the title The Admiral of the String Theory Wars. The writer, Bob Henderson, spent a lot of time talking to me and other people around here, including attending my class. I think he did a very good job of handling the complex topic of the debate over string theory, not so easily done in the context of this kind of personal story. Then again, I guess I would think that since he was mainly hearing my point of view, others may see this very differently…
Talking to Henderson brought back memories of a lot of the very odd things that happened during this period, now nearly ten years ago. I also recently realized that this week is the 10th anniversary of the date on which I sent off the final version of “Not Even Wrong” to the publisher (it then went to copy editing, was published a year later). Looking back at it now, I don’t see much that I’d change. The earliest version of the manuscript written around 2003 didn’t discuss the landscape and multiverse business, which only really got going in 2004. By 2005 I did add a chapter on this, but at the time thought doing this might be a bit unfair to string theorists. Surely they were all aware this was obvious nonsense and would quickly themselves put a stop to it. I was very wrong about that and, oddly, it’s over this argument (where I would have thought my point of view was uncontroversial) that I’ve gotten the most grief from certain prominent string theorists (e.g. Polchinski).
For the latest in the string wars, you can watch last night’s Perimeter Institute public lecture by Amanda Peet. It was a promotional effort that reminded me a lot of the kind of thing you see on late night TV, with an inspirational speaker selling a product to a rapt studio audience. The main selling points for string theory were that it “would blow your mind”, that branes explain black hole entropy (the Strominger/Vafa calculations of 20 years ago), and that the world is a hologram (Maldacena from nearly 20 years ago).
After the talk the first question read to her was the obvious one about testing these ideas experimentally. Remarkably, she claimed that string theory was testable, that published work of 10 years ago showed that it could be tested at accelerators, and so far it had passed the tests. What she was referring to was one of the weirder stories of the string wars. At the time Jacques Distler and collaborators wrote a paper entitled Falsifying String Theory Through WW Scattering, which I discussed here. They also wrote to the Wall Street Journal about this (see here). A few months later their paper was accepted by Physical Review Letters, with their claims about string theory removed, I assume at the insistence of a referee (see here). In a spectacular display of chutzpah, when the PRL paper appeared, the authors had their institutions put out press releases claiming that “Physicists Develop Test for String Theory” (see here), exactly the claim that PRL would not let them make in print. It’s this claim that Peet is now using to mislead the public at Perimeter. I think it’s their responsibility to look into this and issue a public correction. Or do they really feel that it is all right for their public lecture series to be used to mislead the public about science?
Update: I see that the Nautilus piece is part of an issue on “Error”, which also includes interesting articles about the OPERA superluminal neutrino story and about Vladimir Voevodsky.
Update: I probably should have included the relevant question and answer exchange with Peet, here it is:
Q: Any predictions or comments on how and when string theory might be able to be proven experimentally?
A:… there are experiments that can be done, this is published work that has been around for almost ten years now, that you can do experiments in nature in the lab without having to invent equipment that you haven’t invented yet that would test whether or not the assumptions of string theory are true or false. So far none of those experiments has shown a red flag that says string theory is wrong.
Update: Ethan Siegel has a blog entry and live-blogging about the Peet talk. He was hoping that Peet would discuss some sort of connection between string theory and observable physics, didn’t hear any (he seems to have missed the bogus claim about the LHC). Siegel gives an argument that string theory is in principle falsifiable, since it predicts space-time supersymmetry superpartners, although with no prediction for the mass scale. I’ve never heard this before, and I don’t know of any argument that such superpartners must appear in the spectrum of a string theory.
About “It’s this claim that Peet is now using to mislead the public at Perimeter”.
To be honest, at the question session at 1.07.50, she says that “the assumptions of ST are falsifiable”.
That statement is in fact true because we can in principle test the assumptions of any theory if only we had enough energy to test the Planck scale. So, that statement while true is completely useless. The point is not about assumptions but about any falsifiable predictions (that was the question from somebody from twitter). In particular the LHC-predictions (I suppose). I often hear/read from string theorists that they confuse assumptions with predictions. Maybe one should make it clear once and for all that the assumption that particles are extended objects is not a prediction?
She then refers to that PRL paper but still talks about the ST assumptions. In fact she is not answering the question.
Too bad there wasn’t, in the audience, any expert in the field to point out that she is in fact manipulating/misleading the public. Because indeed her answer sounds as if so far ST fulfilled all the experimental tests. But in fact there were no experimental tests of ST so far – to do that you need a prediction first. She “forgot” to add that so far ST made no predictions but merely succeeded in fitting the already obtained experimental results. She also forgot to add that there are so many compactification schemes in ST that they can fit (almost) whatever they want.
Adm. Woit “it’s a trap”
I am the monarch of the sea,
A critic of the theory stringy,
Whose praise Nautilus loudly chants…
Be very wary when advocates claim the “science is settled” or that “xx% of the ‘experts’ have ‘consensus’ ”
re the Nautilus interview in which
“He [Woit] is called an ‘incompetent, power-thirsty … moron’ and a ‘stuttering crackpot-in-chief ‘ ” ,
Falling back onto ad hominem attacks is a sure sign of scoundrels who have no data to justify their beliefs or opinions. You’ll find this behavior in almost every human endeavor (public policy, economics, science, medicine, etc).
I’ve known and worked for a lot of admirals – some good and some not. Henderson seems pretty fair, so I wonder if this is a case where the editor wrote the title rather than the author. It seems like a dose of God Particle sensationalism. As for Dr. Peet, she is a gifted lecturer for a lay audience. She sounds very confident about her subject. She sounds as though she is presenting settled history.
Peter, You have been mentioned on page 8 of the pdf (slides) of the talk Where is Susy? by Howard Baer at Pheno 2015.
Where’s Where is SUSY? — https://indico.cern.ch/event/364031/session/35/contribution/182
Anon,
Thanks. The arguments he’s objecting to though (about how massive gluinos and squarks can be and not violate naturalness) are from Arkani-Hamed and others, I was just reporting them…
nicola,
If we could reach the Planck scale there would still be no unambiguous way to test ST because non-perturbative ST is unknown (see the FAQ). In any case, this is an really old discussion, one that I’m sure Peter is not exactly dying to moderate…
nicola,
Besides the point Bernhard makes about this, if you look at the links where I discussed the Distler et al paper the main point was that the assumptions they claim to test are standard QFT assumptions, not assumptions that have anything to do with strings or string theory. That’s why PRL made them remove the claims about string theory from the paper.
That paper was clearly written exactly to provide someone like Peet the argument she made. But it’s a fact that PRL would not allow that argument to be published, and her claim that it is in a published paper is provably false. Again, I think it’s the responsibility of Perimeter to take some action to correct the false information put out in their public lecture.
Admiral or the admirable kid who pointed out the Emperor’s unclothed state?
String theory is not completely void of predictive power. In fact, I have used string theory to make a falsifiable but confirmed prediction about the LHC. I’m particularly proud that I already in 2007 could predict that poor Lubos would lose his experimental-susy-by-2006 bet.
In my opinion, Bob Henderson’s article could be a beautiful epilogue to any new edition of Peter’s book “’ Not Even Wrong”.
Bernhard,
at Planck scale you would surely test the ST basic assumptions e.g. if particles are indeed stringy. Testing the dynamics and its consequences is another thing.
I am trying to return to old discussions.
Peter,
I perfectly understand what you say. However she seems to refer to the assumptions, not predictions. Maybe she means predictions all the time but says assumptions or maybe she does it on purpose.
nicola,
The point is falsifiability. Of course, if would see stringy behavior at the Planck scale this would certainly be strong evidence for string theory. But since you don’t know the theory non-perturbative behavior not observing it does NOT falsify the theory because stringy behavior is not a necessity. The point is that string theory would still be immune to this test, just as it is immune to the non-SUSY at the LHC test.
“I am trying to return to old discussions. ”
A time-machine to travel to 10 years ago is probably your best option.
nicola,
She is making the same argument as the Distler et al paper which claims to derive “predictions” of string theory testable at the LHC based on “string theory assumptions” about high energy behavior, and explicitly referring to the paper, vetted as a “published” paper. Again, the problem here is that this claim about string theory was removed in the published version since it the “assumptions” are qft assumptions not string theory assumptions.
About the “string theory predicts you see strings at high energy” argument, it’s not the one she is making. If you look at the Nautilus article, you’ll see that Kleban tries to make it, but Michael Dine, a string theorist expert on this, disagrees.
The Admiral that sank the
SpanishString Theory Armada!sorry I meant “I am not trying…”, interesting mistake.
Sir (I suppose I have to address you like that now you’re an Admiral),
Mr. Henderson needs to be congratulated on not mentioning he-who-shall-not-be-named by name, although I see the name crops up in the comment section. Speaking as one of the early followers of the blog (a select group of – according to he-who-shall-not-named – anti-science crackpot losers), I would say that the article is pretty fair, and I am glad that he pointed out the considerable mission creep that has resulted in your blog being one of the prime sources of fundamental physics news irrespective of your views on String Theory. I suppose that there are lot of String Theorists out there who do not appreciate the considerable amount of unpaid work that goes into it, but I do.
“A few months later their paper was accepted by Physical Review Letters, with their claims about string theory removed, I assume at the insistence of a referee.”
It is pretty obvious the change in title refers to a large class of possible ‘new physics’ scenarios at the LHC, not a request from a referee. (This also had to do with the re-doing of the analysis.) The beauty of the relationship discussed in the paper (originally studied by Adams et al. and others in the context of QCD) is rooted in the UV/IR connection: bounds on EFT coefs. that may be derived assuming basic properties of the high-energy theory. For instance, claims have appeared in the literature, e.g. breaking of Lorentz invariance and quantum gravity, which may be then confronted with data (even if somehow hidden in direct low-energy experiments, e.g. due to new symmetries such as SUSY). The existence of a light Higgs changes the tone of the paper, but the premises remain.
The search for UV/IR connections is key to relate measurements we can perform to theories for which we may not have direct access to. This is even more transparent in cosmology, e.g. 1502.07304. Unfortunately you seem to have failed to appreciate the basic idea(s).
“Edgar”,
So, your claim is that the change in title and wording of the paper was because the reaction from the journal was that the “we have found a way to test string theory at the LHC” was too modest? That’s very funny. And your claim that the problem here is that I’m just too ignorant to appreciate the significance of this work does bring back the glory days of the string theory wars. As does the hiding behind anonymity to make arguments you’re too embarrassed to attach your name to…
It seems your point of view is that the real problem here is my raising this issue, not Amanda Peet’s telling the public that Distler et al shows that the LHC can test string theory (as well as claiming that it already has done this, successfully). Do you have any problem with her doing that?
Thanks to Woit for his blog. I have read the blog for many years now. Here is a voice from Norway that you may find nice😊
https://m.youtube.com/watch?v=eCuE7UOP9gw
listen to what she says regarding this point, carefully:
http://www.youtube.com/watch?v=MlDd2HtFfPU&t=68m0s
“experiments […] that would test whether or not the assumptions of string theory are true or false”
This is factually correct. The case of Lorentz invariance for example has been raised many times when the subject of Quantum Gravity has been brought upon, e.g.1106.1417 .
The type of sum rules derived by Distler et al. (and others) does this precisely.
Again, there are measurements (in cosmology) of various kinds which may open a window to physics at high energies, e.g. 1109.0292.
I think this is a more fruitful discussion that nitpicking some specific wording. Unfortunately you stop when the fun/physics begins!
Didn’t occur to you the paper by Distler et al. may have been submitted for the first time with its given title already?
Edgar,
For more context, here’s the question that Peet was asked, followed by the relevant extract from her answer. She is clearly referring to the Distler et al claims that string theory can be tested at the LHC, and claiming that so far it has passed the tests.
Q: Any predictions or comments on how and when string theory might be able to be proven experimentally?
A:… there are experiments that can be done, this is published work that has been around for almost ten years now, that you can do experiments in nature in the lab without having to invent equipment that you haven’t invented yet that would test whether or not the assumptions of string theory are true or false. So far none of those experiments has shown a red flag that says string theory is wrong.
Do you really think that is an honest answer to the question? If it isn’t, why are you writing here to criticize my complaint about it, rather than criticizing her behavior. Is it honest to tell the public that the LHC has successfully tested string theory?
“Factually correct” is dishonest weaseling. It would be factually correct to say that the LHC has tested the “assumptions” of string theory, having in mind that string theory assumes usual mathematical axioms. That was the problem with the Distler et al paper (and press releases). Their assumptions were standard QFT assumptions, having nothing to do with string theory (the day someone finds a violation of Lorentz invariance, do you really believe the reaction of string theorists will be to abandon string theory rather than look at string theory models that can accomodate this?).
You’re correct that I don’t know the details of the publication process of that paper. It sounds like you do and maybe you can enlighten us on what happened between the May and October versions, although your suggestion that the authors changed the title in order to strengthen the claims of the paper doesn’t really pass the laugh test. All I do know is that the preprint carried the bogus title, which I wrote about, the final published version had it removed, then the authors put out press releases using it.
You seem to be too focused on small players in string theory while sparing the big names. In addition you do not report on positive developments such as the workshop that took place at the Simons center this week with diverse topics. One should keep a positive attitude and not to be critical most of the time.
Dundee,
I think I’ve done more than my part as far as offending big names in string theory…
Happy when there’s something positive to report, I’m afraid I can’t agree about the Simons Center talks (other than the Seiberg one, which was much the same as similar talks of his I’ve discussed here before).
Peter: To spend half a post about a talk by Amanda Peet is pointless. Her work for the last ten years have little impact on the string community and as such her talk does not represent them. It is clear that you did not listen to the Simons meeting talks, but only glanced at the titles. Some of the talks like those of Damour and Sunayev, among others, show that there are still serious people who do excellent work without making noise.
Dundee,
I wasn’t discussing Amanda Peet’s own work, but her high profile public talk. You say her public talk doesn’t represent the string theory community, but she was specifically chosen by Perimeter for this purpose (presumably with the participation of the string theorists at Perimeter and who advise Perimeter). Will any string theorists join me in complaining to the people at Perimeter who made this choice, and asking that the institution post a correction concerning the bogus claims she made in her talk? I’ve been considering doing this, but it shouldn’t be me, it is string theorists who should be contacting them.
There’s by now a long tradition of fanatics from the string theory community getting a lot of attention for ridiculous claims for the theory. I can’t think of one example of a string theorist ever publicly raising an objection to any of this.
The Simons Center talks you mention weren’t ones I found compelling, but, tastes differ.
Dear Dr Woit.
The String Theorists at Perimeter may be unaware of what Dr. Peet stated.
Even if they were in attendance, their attention may have been distracted at the time she made her remarkable claims.
In fairness to them, I suggest you let them know of the problem, before criticizing them for not correcting it.
Thank you for linking to the article on Vladimir Voevodsky, Peter.
This resonated with me because it is a scientific issue I’ve always been concerned about — how careful theoretical physicists are in their own work, and how much effort is actually put into checking each other’s work. As with most human behavior, I believe the incentives need to be there to encourage this sort of scrupulousness. However, the terrible academic job market, declining government support for fundamental research, pressure to publish O[10] papers every year, and the (usually) sub-par refereeing process in journals, all point towards the lack of such an incentive structure.
Voevodsky himself has responded by advocating the use of a “proof assistant”. I’m curious, how else has the Mathematics community at large responded? Are there any adjustments being made?
What about theoretical physics? I often wonder if any established theorists share my concern? (I am currently a mere postdoc and have no power whatsoever.) Much more importantly, is there any concrete effort exerted at all to maintain a high standard of intellectual inquiry in theoretical physics?
Yi-Zen Chu,
I don’t personally know anyone in mathematics now using a “proof assistant” in their work (or at least if they are I’m not aware that they’re doing this). It is a huge effort to put a proof in a form that a computer can check, and I’m not even sure that it is possible for lots of areas of mathematics research. One question that came to mind when reading the Voevodsky article was whether the work of his that had an error was actually of the sort that a “proof assistant” could have been used to catch it.
Questions of the reliability of the theoretical physics literature I think are quite different, since there rarely is anything with the level of rigor of a “proof”, so even in principle a “proof assistant” would be of no use.
Tammie Lee Haynes,
The string theorists at Perimeter know Amanda Peet well, so I doubt are surprised she would do this kind of thing. If I thought there was any chance they or Perimeter would do anything about this, I would pursue the matter with them, but there are limits to how much time I want to waste on this, and writing the blog entry probably already went over the limit.
This blog is very widely read by string theorists, so quite a few now know about the Peet issue. If they care about the credibility of their field, they are the ones who should be doing something.
Yi Zen and Peter,
Let me put my 2 cents in as a mathematician on the proof/refereeing question. First, Peter is right most proofs are essentially impossible to check via computer, and I’m guessing they will be for quite some time. There are exceptions, like Hales proof of the Kepler conjecture. But that’s a very special case, and the proof would be impossible without computers, it’s nothing resembling a “normal” proof. As for refereeing, I think it varies, a lot. I think most people take it seriously, but it’s easy enough to miss things. A recent paper of mine had a mistake, not in any of the important parts, just a minor calculation at the end. My co-author and I missed it, basically out of lazy stupidity, we were done with all the big stuff and weren’t paying attention. I’m guessing the referee missed it for the same reason. In that case it didn’t matter, just meant that our estimate for the shortest geodesic was 2.3 and it should have been 2.4. But bigger mistakes have happened. Theres a well known case from the 70’s, where a big time paper was published proving that the bottom of the spectrum of the Laplacian on surfaces was 1/4 (big big deal) – only thing was, it’s not true. Caught pretty quickly. But it happens. And I know of people taking advantage of this. I remember a conversation with Peter Sarnak about some people who were claiming a big result, and he read the preprint and explained to them it was incorrect. So did several other people. They still claimed it was right, and kept sending it to journals until someone finally published it.
I am the monarch of the sea,
A critic of the theory stringy,
Whose praise Nautilus loudly chants…
I rule over all the landscapes
that give multiple answers freely,
And find them most unseemly..
Thanks for the response, Peter.
My questions were meant to be of a more general nature, not just specifically about whether mathematicians are now using a “proof assistant”/computer to check their logic. Perhaps let me give 2 basic examples, to illustrate what I was getting at. It’s very common, nowadays, for theoretical physics papers to contain multiple authors. Most journals do not require authors to describe who contributed what to the research itself. The question here is, do the authors of a given paper even cross-check each other’s work? When physicists write “reviews” of a particular sub-field, they probably do not check — and perhaps do not even read thoroughly! — the papers they cite. How reliable then is the body of work we call theoretical physics, especially given the current climate of an over-saturated job market?
I believe Voevodsky himself, in the article you linked, raised issues of similar spirit — the temptation of publishing a sloppy proof because of the fear of competition, and the consequence of having students learn from their advisers that doing so is alright, etc. I thought this latter point regarding students learning what is OK to do is quite a good one; academia is strongly driven by internal social dynamics and what students assimilate as the “norm” is going to set the climate, and therefore the intellectual standards, for the next generation of scientists.
Jeff M — I was told referees in Mathematics actually step through the proofs in a given paper, so once it passes peer review there is high chance the results are correct? This is one of the reasons why I’ve always held mathematicians in high intellectual regard. Your final story has tarnished that image slightly…
Sorry to tarnish the image, but it does happen. Not often. If you referee something you are supposed to go through it carefully, and if you cannot confirm it you should ask the editor to get the author(s) to explain. Most things in any good math journal are essentially correct, as in the major results are correct. And anything that isn’t correct is discovered quite quickly, unless it’s very obscure. So you can assume most everything you see in good math journals is OK. Physics is different, “correct” should mean “agrees with experiment” and “is mathematically consistent” which is the best you’ll ever do. Peter’s issue, and Lee Smolin’s, and all sorts of other people at this point, is that “correct” no longer has anything to do with the first, and often not the second either. Well, perhaps that’s not quite fair to string theory, it “agrees with experiment” in the sense that some version of it agrees with any experiment you care to do, in any universe you happen to live in 🙂
Yi-Zen (and others),
Sorry, but enough about this. It seems you want to discuss an immense topic (problems with the physics literature and scientific literature in general) that has nothing at all to do with the topic of this posting.
Tho the discussion was off topic….twas also interesting to us listeners.
paddy,
Unfortunately not interesting enough this late this evening to the person who would have to sensibly moderate such a discussion…
@Yi-Zen Chu
The “sloppiness” in higher category theory is of conceptually different nature than the one in theoretical physics. For an example of the kind of things done by a “proof assistant” you can check Carlos Simpson’s paper math/0506471 (“Explaining Gabriel-Zisman localization to the computer”). But the topic of “stability” of results in math and how deep you have to go into someone else’s proof, etc. is huge.
Sorry to go off topic Peter, will be a good boy from now on…