This week’s string theory hype is embedded in a story by Michael Schirber about the possibility of variation of fundamental constants that has appeared on msnbc.com, foxnews.com, and Slashdot. According to Schirber:
A popular alternative to relativity, which assumes that sub-atomic particles are vibrating strings and that the universe has 10 or more spatial dimensions, actually predicts inconstant constants.
According to this string theory, the extra dimensions are hidden from us, but the “true” constants of nature are defined on all dimensions. Therefore, if the hidden dimensions expand or contract, we will notice this as a variation in our “local” 3D constants.
It’s kind of funny to hear that string theory “predicts” that constants like the fine structure constant will vary in time. When Michael Douglas was here in New York giving a talk last year and was asked about predictions of the string theory landscape, he said that the best one was that the fine structure constant would NOT vary. His argument was that it couldn’t vary since effective field theory arguments would imply a corresponding variation in the vacuum energy, something inconsistent with observation. So string theory both predicts that the fine structure constant will vary, and predicts that the fine structure constant will not vary.
For more string theory hype, Michio Kaku now has a MySpace site, including a blog. He also has his own web-site, mkaku.org, which has recently been redesigned and now prominently features an offer of signed copies of his (softcover) books for $50.
Update: There’s an informed take on what the data about varying fundamental constants actually says from Rob Knop.
Lubos, following Jacques Distler, wrote about the RHIC and that it already tests string theory. His argument is that one should not distinguish between strings used in AdS/CFT and strings used for a TOE, as you did previously.
What do you think about his arguments or is this just hype again in your opinion?
Peter,
Since you opened the subject wrt to Kaku, will it be possible to get an autographed copy of NEW once it goes on sale in the US? I collect copies of autographed science books.
Best,
David
fp,
And I was trying to ignore Lubos for a while…
Chad Orzel on his blog made the perfectly accurate comment that RHIC would not test string theory, whereupon he was assaulted by the string theory hype and attack squad, including the usual Lubos behavior towards anyone who has anything negative to say about string theory.
The idea that RHIC will “test string theory” is ridiculous hype, and Lubos’s refusal to distinguish between string theory on AdSxS^5 as used in AdS/CFT and string theory on R^4 x Calabi-Yau is just more of his usual nonsense designed to obscure the fact that one of these is a complete failure.
While someday AdS/CFT ideas may give a real calculational method for QCD, and thus for the quark-gluon plasma, perhaps giving some real predictions about RHIC data, the subject is still very far from that point. The main problem is that AdS/CFT doesn’t apply to QCD, but to supersymmetric versions of the theory, and these have different degrees of freedom. Some people hope that even though AdS/CFT gets the spectrum wrong, and has the wrong degrees of freedom, for some aspects of high temperature QCD, there will be some sort of “universal” behavior and the fact that one isn’t dealing with the real QCD won’t matter. While someday this may get somewhere, at the moment as far as I can tell the subject is being seriously overhyped. If someone can point me to an unambiguous AdS/CFT calculation that makes a sharp, testable prediction about RHIC data, I’d be interested to hear about this.
David,
I won’t be selling autographed copies on the web-site, and have no idea how the publisher handles this (I’ve heard frightening rumors of authors locked in rooms with huge stacks of books and told they can’t come out until they’re all signed). On general principles I’ll be happy to sign people’s books though, without charge.
Oh Michio. I can’t believe you have a MySpace page.
On Kaku’s MySpace weblog:
Carpe Diem.
> If someone can point me to an unambiguous AdS/CFT calculation that makes a sharp, testable prediction about RHIC data
What about the ‘ideal liquid’ (low viscosity) results referenced by Clifford?
fp,
I’ve spent some time looking at the papers referenced by Clifford and others, without finding anything that looked like a sharp prediction that could be compared to RHIC data. I didn’t look at all of this literature, and it’s certainly possible I was missing something. If someone can point to a specific place in a specific paper that makes a sharp prediction, and to corresponding RHIC data, I’d be interested to know about this.
Let us assign the abbreviation P to the statement “the fine structure constant will vary.” Since string theory predicts both P and ~P, string theory therefore also predicts Q, where Q is any statement we want. String theory truly predicts everything… remarkable!
Peter wrote:
Some people hope that even though AdS/CFT gets the spectrum wrong, and has the wrong degrees of freedom, for some aspects of high temperature QCD, there will be some sort of “universal” behavior and the fact that one isn’t dealing with the real QCD won’t matter. … If someone can point me to an unambiguous AdS/CFT calculation that makes a sharp, testable prediction about RHIC data, I’d be interested to hear about this.
It’s not a hope, it’s an argument: QCD at high energies and temperatures is approximately conformal, and universal properties of AdS backgrounds should capture relevant information. See hep-th/0405231 for the calculation of viscosity by Kovtun, Son, and Starinets, and hep-ph/0312227 for discussion by Shuryak of the RHIC results and how they show surprisingly low viscosity. You might also want to look at http://www.admin.ias.edu/pitp/2005files/Lecture%20notes/Klebanov-strong.ppt, a lecture by Klebanov that can point to more literature.
Are there string theorists doing this work because they like the opportunity to connect to an experiment, which they otherwise wouldn’t care about? Probably. But there are also real QCD experts who take this stuff seriously, Shuryak being one example.
anon,
Thanks for the references. I’m not claiming there is nothing interesting going on here, just pointing out that the “RHIC tests string theory” claim seems to be overhyped. I understand the argument about approximate conformality, but is there also an argument about why the difference in degrees of freedom doesn’t matter? Or estimates of what the deviation from this “universal” behavior will be for the real theory? That this difference doesn’t matter, and that these deviations are small is what I was referring to as a “hope”. Is there actually more of an argument?
From the references, it looks like AdS/CFT gives a conjectured lower bound on the viscosity, and the claim is that RHIC data can be interpreted as showing that this viscosity is within a factor of two of the lower bound. Is this right, and is this the closest to a confrontation with experiment that people have managed, or is there something better?
There is more and more hilarious stuff going on in physics, not just the string theory stuff. I was reading a book last night debunking “Intelligent Design”. Maybe not such a bad thing to do — but Lee Smolin (whom I generally respect) had an article in this anthology about his ideas about evolution of the multiverse. Black holes “bouncing” into fantastic numbers of new universes, which reproduce themselves somehow with slight “genetic” variations, so the universes most fit for living systems end up reproducing and predominating. He even claims, sort of, that there is experimental evidence for this (other than the fact that we’re here)! He uses the recent great advances supposedly made in quantum gravity in dealing with singularities. I guess this is the same theory that so far, to my knowledge, admittedly sketchy, can’t give us a semiclassical Minkowski space! All of this to explain the embarrassment of the fact that we are here in a determinedly atheistic way.
It’s all very interesting, and clever, and lots of fun, but utterly fantastical. By comparison, my nutty Bible-thumping ID-believing fellow-townsmen may not be as smart, assuredly not when it comes to science, but they sure are models of sobriety, and good sense!
Scientific American August has an article on Alain Connes and his “alternative” to string theory, how his (and Rovelli’s) theory makes predictions at accessible energies whereas string theory does not.
Joe,
I’ll look for the article, haven’t seen it yet. Presumably Connes is talking about his prediction of the Higgs mass. I’ve never found his version of the SM convincing, but at least he’s sticking his neck out and making a prediction. If it turns out to be right, people will pay a lot more attention to his model. If it doesn’t, he’s in trouble.
I didn’t know there was a joint Connes/Rovelli model, maybe this is a reference to purported LQG predictions?
Without seeing the article, I don’t know how much hype it contains. I hope people working on alternatives to string theory will resist following the string theorists and overhyping their results.
Sorry for the confusion — the Higgs mass calculation does not involve Rovelli; his cooperation with Connes concerns quantum gravity and the emergence of time.
To anon.,
First of all, there is no evidence at all that RHIC produced quark-gluon plasma — you seem to be informed enough to read more than overhyped public announcements intended to keep it running (it’s a historic pattern that amazing discoveries are claimed by experimentalists just before shut-downs). So read this abstract from (nucl-ex/0501009):
“However, the measurements themselves do not yet establish unequivocal evidence for a transition to this new form of matter. The theoretical treatment of the collision evolution, despite impressive successes, invokes a suite of distinct models, degrees of freedom and assumptions of as yet unknown quantitative consequence.” What about this “suite of distinct models, degrees of freedom and assumptions”?
If RHIC really found QGP, it would not be canceled, believe me.
Concerning theory, do you claim that High T behavior of QCD is
the same as of Supersymmetric QCD? Ads/CFT and holography are based on symmetries following from supersymmetry. I don’t think that you can make any statement about NON-BPS dynamics of QCD by using Ads/CFT, and I am not the only one.
So both experiments and theory are shaky. Just forget it…
To Peter:
I agree with (some of) your criticisms of ST. However, I am completely puzzled by your neutral position wrt to AdS/CFT. AdS/CFT has always been advertised (and you buy it) as an important step towards solving QCD, a problem that many theorists consider more important than quantizing gravity. A whole generation spent their most productive years working on it. After 9 years, it is clear that it is useful indeed for discussing unrealistic models with extended supersymmetry, but it has nothing to do with the real non-BPS dynamics of QCD. It’s hard to believe that many theorists still consider AdS/CFT as something more than a mathematical curiosity proving once again that extended
SUSY theories have dynamamics completely determined by symmetry principles. After 9 years, don’t you think that it’s time to pull the plug?
Joe Zhou:
The August SA issue is not out yet (neither does anything about it, or more specifically the purported Connes-and-his-“string theory”-alternative, figure on the SA website). Please check your source, and if you can, provide a link, because I am interested to read about that “news” of yours.
For more string theory hype, Michio Kaku now has a MySpace site, including a blog.
I think this can definitely be put down as a watershed moment.
Dear Peter,
-Saying that “string theory both predicts that the fine structure constant will vary, and predicts that the fine structure constant will not vary” is wrong. Simply because Douglas is not identical to string theory (not even as string theory on AdS is identical to a YM theory…).
-And as Rob points out, the data cannot be used to conclude that the fundamental constants vary.
-Concerning the Quark Gluon Plasma, the RHIC fireball can be described as a gravity dual black hole. And then, since properties of black holes can be computed in string theory (in contrast to LQG for example), the shear viscosity divided by the entropy, can be predicted to be \eta/s = 1/(4\pi) = 0.08.
Regards, Kasper
Someone might have commented on this already:
Nova PBS podcasts has two short (4min and 2 min) segments on string theory:
http://www.pbs.org/wgbh/nova/rss/nova-podcast-pb.xml
In the past they would simply interview Brian Greene – their website looks like a promotional vehicle for selling “Elegant Universe”:
http://www.pbs.org/wgbh/nova/elegant/
Recent podcast, however, includes comments from Neil deGrasse Tyson, a director of Hayden planetarium, where he somewhat angrily comments on how string theorists always say that the experiment is just behind the corner – and so they were promising something concrete in 2-3 years for the past 20-30 years. Brian Greene responds.
They will have Glashow in one of the next podcasts – should be interesting….
Kasper,
My point was just that the people hyping string theory are getting their story confused, with some saying it predicts one thing, some the exact opposite. My position on this has always been the same. String theory predicts nothing, nada, zip. Both Douglas’s claim and the story in question are worthless hype.
As for the QGP, the problem is that the prediction you mention is for N=4 SYM, not QCD, and my question is how much of an argument is there that these will be the same. The second part of the question is whether there is a believable measurement of this parameter, and how it compares to the “prediction”. From Klebanov’s vague summary, I gather it looks different by a factor of two, but I’d be curious to hear from someone who knows more about this.
Jean-Paul,
AdS/CFT is certainly over-hyped, but not on the scale of string theory unification, which has conclusively failed after more than 20 years of hype, that continues to this day. It would be a very valuable thing for someone to write up a critical review of exactly what AdS/CFT has accomplished, and what it hasn’t. This would be a very demanding task, since the literature is incredibly voluminous, and rife with mixing of what is a solid result and what is wishful thinking.
You can make a viable case that thinking about AdS/CFT and how to extend it to QCD may lead somewhere (although quite possibly understanding QCD requires completely different methods). This is different than studying complicated string theory “backgrounds”, which is clearly a doomed enterprise, sure to lead nowhere.
Dear Peter,
The same thing could be said about lattice calculations. QCD “is” not lattice-QCD, but the latter is presumably a good approximation to the former. Likewise, N=4 SYM can – in certain respects – be a good approximation to QCD.
But of course, we would prefer to do calculations in a non-supersymmetric background. One step at a time….
Kasper,
You are writing nonsense. Lattice QCD IS QCD. To the extent it’s an approximation, it’s because it’s a cut-off version and you have to take a limit, but this is true for just about any interesting QFT. N=4 SYM is a completely different theory than QCD, it has very different degrees of freedom and physical behavior. There’s no parameter in the the theory that one can take to a particular value and get QCD.
Peter, the “is” above refers to the fact that lattice-QCD is (among other things) a cut-off version of QCD on R^4. And as such an approximation. (Another reason being that most lattice calculations have not included quarks).
As an example, recent lattice calculations give a strange quark mass around 85 MeV as compared to the 150 MeV from phenomenology.
Are you claiming, that taking the continuum limit is trivial? Are you claiming that there are no finite-lattice spacing effects?
Kasper,
To define a QFT like QCD, you have to do it by defining it with a cut-off, then taking a limit. That’s the way an asymptotically free QFT works. Exactly the same thing is true in perturbation theory, or any other way one knows of for defining such a QFT. The cut-off is not an approximation, it’s an intrinsic part of the definition of the theory.
Of course actually doing calculations is non-trivial, and you have to take the lattice spacing to zero carefully, understanding what errors are caused by stopping at a fixed lattice size.
Chiral fermions are tricky, but QCD is not a chiral theory. There are several ways of putting in fermions, this makes numerical calculations more difficult, but many groups are doing such calculations, and getting results consistent with experiment. I’ve never heard before the claim you’re making that lattice calculations give results off by nearly a factor of two, and don’t believe it. Please provide a reference.
According to the PDG, m_s falls between 80 and 130 MeV. If you take the lowest lattice point and the highest phenomenological point, you do get numbers like the ones suggested above. This says two things: (1) the strange quark mass is not very well known, (2) it’s easy to lie by cherry-picking data.
Hello prof. Woit,
Kasper is probably referring to work done by Akira Ukawa, published in late 2004. I remembered running into something like this a few months ago. The relevant quote would be at the end of section four (page 14 of the pdf file accessible from the URL below).
http://www.aapps.org/archive/bulletin/vol14/14_6/14_6_P11-15.pdf
And congratulations on the book. Can’t wait to get a copy of the NA edition. The dust jacket for that edition is particularly elegant, especially with the “mirror image” WRONG in the title.
Gilbert, correct – I didn’t have the PDG at hand and remembered reading such numbers in the review.
I guess, even though you are doing theory, you should still carry it in your pocket. It’s all so good for a late-night conversation at a local bar 😉
Cheers, Kasper
I wanted to mention a few points about lattice QCD, since there seems to be a bit of confusion about it. I am not a computational lattice person myself, but I have a number of friends who are, and I follow developments in the subject.
First, the quenched approximation (in which the dynamical quarks’ contributions to the vacuum polarization are neglected) is pretty much gone. There are now many, many gauge configurations written to disk with 2 or 2+1 fully dynamical quark species, and groups around the world are making use of them. They have been used to make some remarkably precise calculations, but their general applicability is occasionally controversial, because of the second point.
The second point is that the Lagrangian used in lattice QCD is never just the QCG Lagrangian with a cutoff. That could come close to being useful if the lattice spacing was really small–much smaller than any other scale that might arise in a calculation. In order to get useful data at physically realizable lattice spacings (and one must actually use several lattice spacings and extrapolate to zero–this it how the continuim limit is taked), the actions that are used contain all sorts of extra terms and corrections. Generally, these additions are expected to be irrelevant, so that the theory approaches true QCD as the lattice spacing goes to zero, but that has not always been proved, and sometimes the changes to the action are downright controversial. The biggest problem comes from fermion doubling, which is related to the fact that the use of lattice necessarily organizes the momenta into Brillouin zones, and it’s not always clear that the extra fermion “tastes” can be eliminated in a local fashion.
Finally, there are lots of things that lattice QCD cannot calculate usefully at all. I’m not sure what it would mean to “calculate” the strange quark mass, since this is usually an input parameter. (One might fix the up and down quark masses to 1/20-th of the strange quark mass, which is certainly an approximation, but it can be a highly useful one. Of couse, there are complications, since one can distinguish sea quark and valance quark masses, and the whole procedure is entangled with the extrapolation, but the basic idea is right.) A better example of what we cannot calculate on a lattice would be hadron-hadron scattering. The lattices in use today are simply too small to contain external scattering states. So lattice people are very careful to choose things (e.g. single-hadron properties, like masses and decay constants) that can be reliably calculated, although there are still sometimes problems with error estimations.
Chad has (I think) now retracted his statement that “RHIC does not test ST”, and Clifford has thanked him for the retraction. Interesting.
Gilbert,
Thanks for the reference. It seems to me that it shows that all calculations of observable masses are coming out right within expected errors. The strange quark mass is not something observable, so the discrepancy mentioned is not obviously significant unless you look into this more closely and make sure that you’re comparing the same thing on both sides, and understand what the errors are on both sides.
Andy,
To be fair, the string theory attack squad was mainly complaining about Chad’s remark that they weren’t interested in RHIC and the data coming out of it. You could debate this point, but it wouldn’t be useful. They do seem to have done a good job of beating Chad into submission in this case.
Nova PBS series offers short podcasts featuring Brian Greene and Neil deGrasse Tyson, director of Hayden planetarium. Tyson mentions that string theorists tend to promise experimental confirmation that is “just around the corner” – maybe 2-3 years away, but that they have also been doing this for 20-30 years. Greene counters with a claim that he’s never heard anyone say that, and perhaps string theorists were just joking.
It’s a little surprising to hear Nova offer counter-argument, especially since their website looks like a vehicle for selling “Elegant Universe”.
Coming up is the interview with Sheldon Glashow.
More funny is
“A popular alternative to relativity, […] ”
Juan R.
Center for CANONICAL |SCIENCE)
DEAR PETTER,
MICHIO KAKOS BLOG HAS AN ARTICLE ABOUT TESTING OF STRING THEORY USING GRAVITY WAVES AT LISA, DEVIATION OF NEWTON LAW – BUT ARE BOTH THESE NOT CLASSIKAL AFFECTS? IE, BY VARYING CLASSIKAL THEORY WE CAN GET THESE ONES. THERE IS NOTHING THAT STRING THEORY SAYS THAT IS NOT ALREADY SAID. THIS SEEMS TO BE TYPIKAL BEHAVIOR.
ALSO PLEASE TO BE COMMENTING ON THE ARTICLE IN THE AUGUST AMS NOTICES
http://www.ams.org/notices/200607/fea-marateck.pdf
ABOUT DIFFERENTIAL GEOMETRY AND FEYNMANN INTEGRALS. I HATE FEYNMANN INTEGRALS – WHY DO NOT THEY CONVERGE!!
PS I KNOW YOU ARE UPSET BECAUSE I USE ALL BIG LETTERS BUT I MUST FOR TWO REASONS. ONE, IT IS A PART OF MY PERSONALITY AND MAKES ME QUIRKEY AND MISTERIOUS AND ENIGMATIC. ALSO MY KEYBOARD IS BORKEN.
PPS ISRAEL GELFAND FOR ABEL PRIZE 2007!
QWERTY,
Re Kaku and his “predictions”, see
http://www.math.columbia.edu/~woit/wordpress/?p=219
The idea that LISA is going to test string theory is utter nonsense.
The Marateck article is a basic expository article on gauge fields for mathematicians. At the end it has some material about Feynman diagrams, but doesn’t really explain anything much about them, or how they are related to the earlier material on gauge fields.
I see that Feynman is the 37 cent stamp. Personally, I would have preferred it if they had made him a 137.036 cent stamp. OK – it would make the arithmetic more complicated, but we could deal with that, couldn’t we?
The Marateck article is just hot air, the entire thing is filled with factual errors of history, and is presented on the level of an ambitious and green undergraduate. I wish people would refrain from abusing Weyl and Dirac by refusing to comprehend them.
-drl
I could write an article on the lines of the Marateck one, and probably more interesting – i.e. not just cheerleading. Funny that I never get asked to do this.
Peter wrote:
The idea that LISA is going to test string theory is utter nonsense.
Last month I have collected some references related to this matter. I’m too busy to read them, but of course I am curious on them. And much more now, considering Peter’s statement. Here is my selection:
* An interferometric gravitational wave detector as a quantum-gravity apparatus [gr-qc/9808029]
* Accessibility of the Pre-Big-Bang Models to LIGO [astro-ph/0510341]
* Observational test of holographic inflation [PRD, vol. 73, Issue 2, id. 023516]
* Gravity-wave detectors as probes of extra dimensions [astro-ph/0505277]
* The Primordial Gravitational Wave Background in String Cosmology [hep-th/9907185]
I do not know whether Peter Woit wants people to discuss them here, so if that is the case, you are invited to contribute here.
Thank you,
Christine
I just read that Michio Kaku is a comedian. I should have known that.
So Kaku goes to Dr. Woit and says “Doc, my string theory isn’t doing so well” and Woit says, “Yeah, it’s a very sick theory” so Kaku says “can I get a second opinion?”, and Woit says, “Yeah, it’s ugly, too!”
-drl
How many string theorists it takes to change a light bulb?
Two: One to hold the blub, and a second to twist it.
QWERTY writes:
I HATE FEYNMANN INTEGRALS – WHY DO NOT THEY CONVERGE!!
Be careful – don’t forget item 8 of the crackpot index!
PS I KNOW YOU ARE UPSET BECAUSE I USE ALL BIG LETTERS BUT I MUST FOR TWO REASONS. ONE, IT IS A PART OF MY PERSONALITY AND MAKES ME QUIRKEY AND MISTERIOUS AND ENIGMATIC.
No, it just makes you annoying and boosts your score on item 7 of the crackpot index.
ALSO MY KEYBOARD IS BORKEN.
Hmm, well, maybe not.
Q: How many String Theorists does it take to change a light bulb?
A: What are you talking about, you half-wit moron! This room isn’t dark because we have already changed the light bulb, and done a lot more besides! Unfortunately, being a crackpot fool with the intelligence of an amoeba, you don’t realise it!
QWERTY,
Inspired by your fine example, I am now going to call my keyboard “DRELL”.
Re: your crackpot credentials – John Baez makes some valid points, but if there is any doubt, I am prepared to give you the benefit of it as I too don’t like non-convergent Feynman integrals.
Seems like string theory has a high crackpot index…
Looks like we have a real high brow discussion going here of late. Perhaps we could get back to discussing approximations to QCD that were brought up earlier and Brett had a nice post about filling in some of the details. Does the statement that “lattice QCD IS QCD” hold water considering many of the things it does not take into account, and only certain regimes the approximations in lattice QCD hold? For instance what defines the utility or goodness of an approximation since as an example N=4 SUSY YM is more useful to calculating in QCD for the LHC than the lattice albeit our world is certainly not that full of supersymmetries at TeV energies. Of course we have to have some further discussion about N=4 applications in QCD such as what dixon et al do beyond the normal banter about AdS/CFT and whether it is or is not testing QG. Perhaps Peter could make a general post about QCD and various techniques in understanding QCD at a deeper level. Granted this would take us out of the whole joke fest we have going here but it would illustrate an important point many lay people don’t understand. Different theories are useful for different energy regimes as well as different theories have sectors that can be used to calculate quanties better than in the original theory that is the “true” theory. Anyways if this isn’t interesting to the rest of the commenters feel free to go back to the usual.
amazed,
There were quite a few postings and discussions here in the past about twistor methods for doing perturbative QCD calculations, if there is anything new about this subject (and not just hype about N=4 SYM), I’d like to hear about it.
The problem with writing about QCD is that there seems to be relatively little new to say about it. I try to mostly write about news here, I don’t really have the time or energy to do more in the way of making this an educational or expository site. But if there is something new about QCD, I’d be happy to learn more about it and maybe write about it here.