Not Even Wrong

Since every blogger seems to feel it necessary to have a year-in-review posting, I thought it appropriate to point out that 2011 has been a banner year for string theory and related hype, with about twice as many editions of “This Week’s Hype” as in previous years. One reason for this is the LHC. When talking to journalists, string theorists are rarely willing to admit that the hopes of the past couple decades that string theory would make some predictions about LHC energy scale physics turned out to be a dismal failure, and this tends to lead to confused headlines. Besides the LHC though, there’s a huge on-going effort to promote other bogus “tests of string theory”. This has been going on since string theory’s lack of testability problem first started to get a lot of attention a few years ago. I see no reason for either of these two driving forces to weaken in 2012, so expect more editions of “This Week’s Hype” next year.

String theory supported by early LHC heavy ion results

Cosmologists expect the LHC to turn up evidence for the multiverse

M-theory shows that the LHC will be the world’s first time machine

Neutrons could test string theory

Octonions explain string theory

The LHC tests string theory (more heavy ions)

M-theory is a big success, predicts behavior of 4 qubits

String theory and heavy ions, yet again

Multiverse observed in the CMB – Not

String theorists suggest space wormholes possible

String theory testable with black holes and pulsars

String phenomenologists come up with predictions testable at the LHC

Superluminal neutrinos evidence for string theory

Superluminal neutrinos could be explained by string theory

CMS multi-leptons provide evidence for SUSY

A new laser will tear apart the fabric of space

A nuclear clock will test string theory

The LHC will decide between two versions of SUSY

String theory research is going well, only problems are Garrett Lisi, Lee Smolin and Peter Woit.

Gordy Kane predicts the mass of the Higgs using string theory, just days before the announcement

Superstring theory predicts three space dimensions.

The Japanese are getting in on the string theory hype business, with KEK issuing a press release today with the title: The mechanism that explains why our universe was born with 3 dimensions: a 40-year-old puzzle of superstring theory solved by supercomputer. As usual for this kind of press release, the claim is that researchers at the institution issuing the press release have finally solved the age-old problem of string theory predicting nothing. In this case the prediction is that there are 3 dimensions of space (I think that, technically, this is a “post-diction”). According to the press release:

A group of three researchers from KEK, Shizuoka University and Osaka University has for the first time revealed the way our universe was born with 3 spatial dimensions from 10-dimensional superstring theory in which spacetime has 9 spatial directions and 1 temporal direction. This result was obtained by numerical simulation on a supercomputer…

… it is expected that superstring theory allows the investigation of the birth of the universe. However, actual calculation has been intractable because the interaction between strings is strong, so all investigation thus far has been restricted to discussing various models or scenarios…

… It is almost 40 years since superstring theory was proposed as the theory of everything, extending the general theory of relativity to the scale of elementary particles. However, its validity and its usefulness remained unclear due to the difficulty of performing actual calculations. The newly obtained solution to the space-time dimensionality puzzle strongly supports the validity of the theory.

Furthermore, the establishment of a new method to analyze superstring theory using computers opens up the possibility of applying this theory to various problems. For instance, it should now be possible to provide a theoretical understanding of the inflation that is believed to have taken place in the early universe, and also the accelerating expansion of the universe, whose discovery earned the Nobel Prize in Physics this year. It is expected that superstring theory will develop further and play an important role in solving such puzzles in particle physics as the existence of the dark matter that is suggested by cosmological observations, and the Higgs particle, which is expected to be discovered by LHC experiments.

This goes back to the pre-arXiv days, before many of our current graduate students were even born, but some of us are old enough to remember similar claims being made back in the late 1980s. For example there’s the 1989 Brandenberger-Vafa paper claiming that string theory predicts 3 dimensions, using a “string gas” cosmology. I don’t remember if there was a “finally, physicists find a way to make a prediction based on string theory” press release back in 1989 or not.

The announcement at CERN tomorrow of a likely-looking signal for a 125 GeV mass Standard Model Higgs will probably unleash a flood of hype from theorists claiming this as evidence for their favorite Beyond the Standard Model scenario. One obvious problem with any such claim is that the CERN results correspond well so far to the Standard Model with no additions whatsoever, so spinning them as providing support for things like supersymmetry and string theory will require some work.

For the last decade we have known that the Higgs mass is above 114 GeV (from LEP) and unlikely to be very much higher than that (from precision electroweak results). This summer’s LHC results disfavored masses above about 130 GeV, so for the last few months we’ve known that if the Standard Model Higgs is there, it should be between 114 and about 130 GeV. For a couple weeks news has been circulating widely from ATLAS and CMS that they are both seeing something around 125 GeV.

First out of the gate in the hype derby is Gordy Kane, who is quoted by Davide Castelvecchi at Scientific American claiming that string theory predicts the Higgs mass to be between 122 and 129 GeV:

“If it’s in that range it’s an incredible success for connecting string theory to the real world,” Kane says. He says he is confident that the upcoming LHC announcements, if they pan out as predicted, will constitute evidence for string theory. “I don’t think my wife will let us bet our house, but I’ll come close,” he says.

It’s unclear exactly what he’s willing to bet the house on. If it’s just that the Higgs is in that range, this might have something to do with the plots from the experiments widely circulating privately the last few days. In a remarkable coincidence, after more than 25 years of unsuccessfully trying to extract a definite experimental prediction from string theory, Kane and collaborators were able to achieve the holy grail of the subject (a prediction of the one unknown parameter in the SM, the Higgs mass) just a week before the CERN announcement. They submitted their paper to the arXiv the evening of Monday December 5, a few days after rumors of a 125 GeV Higgs were posted on blogs Friday December 2.

The paper deals with a “prediction” you get based on a host of assumptions about which particular class of string theory compactifications to look at. The main result is that in this particular class of models, you can relate the Higgs mass to the parameter tan(β) that occurs in SUSY extensions of the SM. As you increase tan(β) from around 2, the Higgs mass lies in a band, increasing from 105 GeV and a maximum of about 129 GeV:

We will demonstrate that, with some broad and mild assumptions motivated by cosmological constraints, generic compactified string/M -theories with stabilized moduli and low-scale supersymmetry imply a Standard Model-like single Higgs boson with a mass 105 GeV < M_h < 129 GeV if the matter and gauge spectrum surviving below the compactification scale is that of the MSSM, as seen from Figure 1. For an extended gauge and/or matter spectrum, there can be additional contributions to M_h.

This conclusion and Figure 1 correspond closely to what is in the slides of Kane’s talk at String Phenomenology 2011 this past August. The plot of Higgs masses as a function of tan(β) is there, giving a range of 108 GeV to 127 GeV. There is an intriguing comment on the conclusion slide:

Single light Higgs boson, mass about 127 GeV unless gauge group extended.

I can’t tell where the number 127 came from. Since 127 GeV is the top limit in the figure, and the wording “unless gauge group extended” is used, one guess would be that Kane meant that 127 GeV was the upper bound on the Higgs mass in this class of models.

There’s nothing in Kane’s August talk about a 122-129 GeV range for the Higgs mass, but in the December 5 paper it appears explicitly three times:

Kane has chosen Lubos Motl’s blog as the place to guest post and promote this string theory “prediction”, concluding there:

If generic compactified string theories with stabilized moduli correctly predict there is effectively a single Higgs boson and correctly predict its mass, it will be a huge success for the main directions of particle physics beyond the Standard Model, for supersymmetry and for string theory, both of which are crucial for the prediction. It will be a huge success for LHC and the accelerator physicists and experimenters who made the collider and the detectors and the analysis work. It will put us firmly on the path to understanding our own string vacuum, and toward the ultimate underlying theory. The value of the Higgs boson mass not only confirms the approach that predicts it, remarkably depending on its numerical value it may allow an approximate measurement of tan β, the μ parameter, the squark and gravitino masses, that the gauge group and matter content of the theory below the string scale is that of the MSSM, and that light (TeV scale) gluinos and dark matter are likely.

I don’t quite see how finding a SM Higgs at 125 GeV is going to give all these different numbers and pieces of information. Kane claims that these string theory predictions also predict gluinos visible at the LHC within months from now:

Then the gluino should be detected at LHC, and because of the heavy scalars the gluino decays are different from the ones usually discussed, being dominantly to third family quarks, top and bottom quarks. I won’t explain that here because of space and time; we can return to it as the gluinos are being detected in coming months. They have not yet been systematically searched for.

In addition, there’s a dark matter prediction that should be tested in “1-2 years”.

If I were Kane, I wouldn’t bet my house (or go to the press claiming the 125 GeV Higgs as a “huge success” for string theory) just yet. Assuming he’s right, within months the gluinos will be there, and his ticket to Stockholm will be assured.

Update: My prediction in the first paragraph is coming true faster than I thought. In tonight’s hep-ph listings one finds:

http://arxiv.org/abs/1112.2415 (m_H= 127 +/- 5 GeV)

http://arxiv.org/abs/1112.2462 (m_H < 128 GeV)

http://arxiv.org/abs/1112.2659 (m_H =126 +/- 3.5 GeV)

http://arxiv.org/abs/1112.2696 (m_H > 120 GeV)

There are going to be a lot of these…

Update: Over at a Nature live chat, Kane is giving the public some interesting explanations:

“for a higgs to be meaningful it must be part of a supersymmetric theory, so the superpartners should be found. The form it takes implies that gluinos should be found with masses around a TeV, maybe less, by summer, and decaying mainly into topquarks and bottom quarks.”

“Recently we have published string theory calculations that imply the higgs boson mass is 125 GeV so if it is there are strong implications for connecting string theory to the real world, and for what the higgs discovery implies. we did that before the data.”

“string theories are now well enough understood to predict higgs physics”

There will be another live chat involving Kane tomorrow, this one at Science magazine.

Update: Just took a look at the Science chat. Kane seems to have completely lost touch with reality, somehow deciding that the two experiments have reached the 5 sigma level needed to claim a discovery. As far as I know, he’s the only one in the world to think this.

Gordy Kane:
YES. an experimenter from one experiment can’t say that, but theorists see that two different experiments both saw a signal at about the same mass, and also saw additional channels, so it’s a discovery!

…

Gordy Kane:
The 5 sigma is a criteria people have chosen. i think as soon as the data are combined from two detectors, which is entirely legitimate, then the signal will indeed be over 5 sigma.

The following claims make about the same amount of sense as the discovery one:

Comment From Tom
Does the existence of a 125 Gev Higgs give any support to supersymmetry?

Gordy Kane:
Yes. first, for a long time it has been known that the lightest higgs boson of supersymmetry should be lighter than about 135 GeV (actually closer to 140 GeV but people make assumptions), so this is consistent. Then the supersymmetric string theories as i mentioned do predict the 125 number and it is a supersymmetric lightest higgs boson.

Update: The hype goes on, with a column today at Nature.

Mike Duff has a new preprint out, a contribution to the forthcoming Foundations of Physics special issue on “Forty Years of String Theory” entitled String and M-theory: answering the critics. Much of it is the usual case string theorists are trying to make these days, but it also includes vigorous ad hominem attacks on Lee Smolin and me (I’m described as having an “unerring gift for inaccuracy”, and we’re compared to people who campaign against vaccination “in the face of mainstream scientific opinion”). One section consists of a rather strange 3-page rant about Garrett Lisi’s work and the attention it has gotten, a topic that has just about nothing to do with string theory.

Duff explains that his motivation for answering the critics is that we have been successful on the public relations front, supposedly responsible for the British EPSRC “office rejecting” without peer review grant proposals on string theory. I know nothing of this, but I think it’s clear to everyone that the perception of string theory among physicists has changed, and not for the better, over the past decade. One dramatic way to see this is to notice that at this point, US physics departments have essentially stopped hiring string theorists for permanent appointments (i.e. at the tenure-track level).

String theorists have a problem not just with the public, but with their colleagues. The main reason for this is not Smolin or me, but the failure of the string theory research program. Duff’s take on whether the landscape is pseudo-science is that string theory can’t even tell whether there is a landscape, and he is “doubtful whether the kind of issues we are considering here will be resolved any time soon.” On the question of the time scale for possible progress, he invokes the two millennia it took to get from Democritus in 400 BC to quantum theory early last century. His list of greatest achievements of string theory in recent years has just two items: applications to fluid mechanics and his own work on entanglement in quantum information theory. Given this, it’s hard to see why he’s surprised the EPSRC is cutting back on support for string theory.

While Duff has detailed complaints about exactly what Smolin wrote in The Trouble With Physics, he mentions my book without saying anything about what is in it (one suspects his policy of how to deal with it is that of Clifford Johnson and some other string theorists: refuse to read it). He does have some specific complaints about material from my blog:

Duff is much less interested in the virtues of accuracy when he describes my words. I guess I’ve joined Smolin on his list of targets because of what I’ve had to say on the blog (see here, here and here) concerning his publicity campaign claiming a “prediction of string theory” about qubits (recall that he thinks this is one of the two main advances in string theory of this decade). He claims that “falsifiability of string theory is the single issue of Peter Woit’s ‘single-issue protest group'”, and that my argument about the qubit business “may be summarised as (1)It’s wrong (2)It’s trivial (3)Mathematicians thought of it first.” One can read the postings and decide for oneself, but I’d summarise the argument quite differently: Duff has nothing that can possibly be described as a “prediction of string theory” and it’s misleading hype to issue press releases claiming otherwise. The experimentally testable “prediction” is that “four qubits can be entangled in 31 different ways”, but if experimentalists make measurements of four qubits that show something different, one can be sure that the headlines will not be “string theory shown to be wrong in a lab”.

Duff’s article contains an appendix about this, in the form of a “FAQ”, where he explains that he approved the text of the press release headlined “Researchers discover how to conduct first test of ‘untestable’ string theory” which is misleading hype by any standard. Initially someone who was successfully misled in the Imperial media team added the subtitle “New study suggests researchers can now test the ‘theory of everything’”, which was later removed. Duff claims that Shelly Glashow, Edward Witten and Jim Gates told journalists that they didn’t agree with this because of the “theory of everything” subtitle, implying that otherwise they were fine with the “first test of ‘untestable’ string theory” business (except for Gates noting that in any case this is just supergravity, not string theory). It would be interesting to hear from the three of them if they’re really on-board with this “first test of ‘untestable’ string theory”.

What Duff and some other string theorists don’t seem to understand is that this sort of “answering the critics” is exactly what has gone a long way to creating the situation at the EPSRC that he is worried about. Unfortunately it has damaged not just the credibility of string theory, but of mathematically sophisticated work on particle theory in general. According to Duff

Just recently, in fact, EPSRC completely abolished its Mathematical Physics portfolio.

Update: Matin Durrani at Physics World (also a target of Duff’s ire) has a blog entry about this here.

Update: Lee Smolin sent me the following comments on the Duff article:

Maybe it would help if I provide some context for the debate Mike Duff took part in with Nancy Cartwright and myself in London in 2007. The occasion for the debate was the publication of my book TTWP in the UK and the reason for the debate was that I had insisted that, as the point of the book was to explore the role of disagreement and competing research programs in science, the best way to illustrate it was to have a debate. String theory was discussed in the book as a case study illustrating the issues and so it seemed appropriate to have a debate with a string theorist. I also insisted that in each of these debates a philosopher of science would be included to highlight the fact that the main themes of the book were longstanding issues in philosophy of science, having to do with how consensus forms within a scientific community on issues on which there is initially wide disagreement.

There were two such debates in the UK, the other was at Oxford with Philip Candelas and Simon Saunders. That went very well, as Philip gave a strong defence of string theory that stayed focused on the scientific issues.

Duff’s construction of two me’s is, so far as I can tell, a debating tactic to avoid addressing the key issues my book raises. He starts with

“Who can dispute that the ultimate goal of a scientific theory is to make experimentally testable predictions? Who will challenge the need to keep an open mind and listen to unorthodox views? Who can disagree with the assertion that our current understanding is only partial and that the ultimate truth has yet to be uncovered? What Lee Smolin said in the London debate [29] was so uncontroversial that, had I confined my response [11] to these remarks, the evening would have would have fizzled out in a bland exchange of truisms.”

Indeed the constant theme of my book is the development of those “truisms.” What Duff does not explore is that in spite of the agreement there may be over these “truisms”, they have strong consequences for the evaluation of research programs in fundamental physics. Apparently we disagree about the implications for string theory. What Duff could have done is acknowledged these disagreements and explored the reasons for them. Instead he claims to attack my book, but it is striking that he does so, not by criticizing the text I actually wrote-but by attacking first the publicity blurb on the cover and then responses from journalists. As I have stated many times, the material on the cover was neither my text nor my choice and is more strongly worded than anything in the actual book. I hope it is obvious also that you cannot attack a book by pointing out inaccuracies in reviews.

When he finally does get around to quoting from the book, he makes a few good points mixed in with distortions gotten by quoting out of context. Had he stuck to the good points he had we could have had a useful debate that would have shown the audience the role of disagreement among scientists faced with difficult questions. Had he done that, there would have been no need to construct a fiction of two me’s. I am happy to leave it to readers of my book to judge whether its text is or isn’t completely consistent with the “truisms” he asserts we agree about.

There is one aspect of Duff’s rant which deserves correction, which is his attack on me related to Garrett Lisi. What Duff says is, “So when Lee Smolin described him [Lisi] as the next Einstein, the publicity juggernaut moved into overdrive”. There are several untruths in this short sentence.

First, this refers to a Discover article of March 2008 which says, “With Smolin’s aid, DISCOVER has scoured the landscape and found six top candidates who show intriguing signs of that Einsteinian spark” of whom Lisi is one. This was, so far as I recall, based on a phone call with an editor at Discover following a piece I had written for Physics Today on the challenges faced by those who do high risk-high payoff research. I think anyone who looks up the full list of six will see that the editors were aiming to illustrate a wide range of approaches to fundamental research, of which Lisi is at one pole. And as they make clear-the choice of the list was theirs and not mine.

Furthermore, the media attention on Lisi had begun and peaked already in November of 2007, sparked by a New Scientist article, following immediately the posting of his article on arxiv.org. And while there was a very exaggerated media response-which I and others did our best to advise against-there was no “publicity juggernaut” ie no attempts by Lisi or anyone to seek publicity for him, no press releases, no publicist, no calls to journalists except to strongly advise the story was premature. I told everyone who asked not to write a story on Lisi because the preprint had just been uploaded and there had not been time for experts to evaluate it. Indeed, New Scientist had quoted me very much out of context, ignoring emails I sent them advising them not to write a story on Lisi’s paper before the experts could evaluate it. So the reality was the opposite of the impression created by Duff’s sentence.

None of this is new, none of it is said for the first time. It is depressing to revisit these debates from five years ago. Most of us have moved on. At least I have, as readers of my next books, as well as the article I was invited to write for the same special issue, will, I hope, see.

Update: The following is Garrett Lisi’s response to the Duff article. I should note that I’m not complaining about Duff’s listing of my titles. If you want to make up your mind who is right based on titles, Duff’s your man in this argument.

Michael Duff’s article is full of deceptive half-truths. To attack the commentary on Lee’s book, while avoiding Lee’s actual arguments, is just one example of this fundamentally dishonest tactic. A similar example is his reference to Peter as “Computer Administrator and Senior Lecturer in Discipline,” as if Peter was not also a very knowledgable mathematical physicist. Duff then launches an attack on my work, once again focusing on a large volume of commentary by others rather than on my actual arguments. Also, Duff refers only to my first paper, saying it’s never been peer-reviewed and published, avoiding the fact that I’ve since published papers on the theory, including “An Explicit Embedding of Gravity and the Standard Model in E8.”

Scouring Duff’s rhetoric, baseless statements, and ad hominem attacks in search of some factual argument supporting his attack on my work, I can find only this:

“Nature (and the standard model of particle physics) has three chiral families of quarks and leptons. ‘Chiral’ means they distinguish between left and right, as they must to account for such asymmetry in the weak nuclear force. But as rigourously proved by Jacques Distler and Skip Garibaldi, Lisi’s construction permits only one non-chiral family.”

This is, once again, a misleading half-truth, avoiding the fact that a chiral family of quarks and leptons can be part of a non-chiral representation space, as is the case in E8. I cannot credit Duff alone for this deception, as its source is Jacques Distler — a master of the half-truth — but I can blame Duff for supporting it.

For anyone who actually cares about the state of E8 theory, I would recommend my recent papers. Apparently Duff considers the work sufficiently threatening to the string program that he needs to attack it in this dishonest manner. If string theory models are as twisted and misleading as the statements in Michael Duff’s paper, it’s no wonder they’re dying.