I spent the last two days up in Cambridge, mainly attending the conference in honor of Sidney Coleman. Sadly, Coleman is in poor health, suffering from Parkinson’s disease, and was unfortunately unable to attend the talks in his honor. They were videotaped so that he could watch them later.
For me as for many particle theorists, taking Coleman’s quantum field theory course at Harvard was one of the great intellectual experiences of my life. Another such experience was reading and learning from his great Erice lectures, both the late seventies ones as they came out, as well as going back to his earlier ones that started in 1966. These were collected in 1985 in the book “Aspects of Symmetry”, allowing me and many others to replace a stack of dog-eared Xeroxes with a more durable volume. The fact that Coleman stopped giving these lectures after 1979 was to me one of the first indications that particle theory was entering a much less promising phase of its history. Coleman never really warmed to the topics of supersymmetry and string theory.
For much of his career Coleman played the role of guru for the particle theory community, generously sharing his unmatched insights into quantum field theory. He would sleep through the morning (famously announcing that he couldn’t teach a 9am class because he couldn’t stay up that late), get into his office late in the afternoon, then spend hours dealing with a long line of people waiting to talk to him to try and get some help with whatever problem they were working on. Steven Weinberg spoke for many people at the conference when he said that Coleman was the single person he had learned the most physics from.
The conference was extremely well-attended, with the large lecture hall in the physics building at Harvard overflowing on Saturday. I don’t think I’ve ever seen so many Nobel prize winning particle theorists in one place. They included Gell-Mann, Glashow, Weinberg, ‘t Hooft, Gross, Wilczek, Wilson, as well as Fields medalist Edward Witten. One of the few living Nobel particle theorists who couldn’t make it was David Politzer, who very much directly owes his prize to Coleman.
I won’t describe the talks in detail, this has been done pretty accurately already by Lubos Motl (who I got to meet in person for the first time). Physics weblogging was very well represented at the conference: besides Lubos, Jacques Distler was liveblogging from one corner of Science Center B on Friday, and Serkan Cabi was also there. Sean Carroll also has some comments about Coleman.
Among the more historical talks, perhaps the most interesting was that of Gerard ‘t Hooft (Lubos seems to have missed ‘t Hooft’s comment that he shouldn’t be referred to as “Gerardus”, a formal version of his name that appears on his Nobel citation and his passport, but is otherwise not much used). ‘t Hooft gave his version of the asymptotic freedom story. He said that he had computed the Yang-Mills beta function a couple years before Gross-Wilczek-Politzer, but didn’t realize that this result wasn’t known to the experts. He pointed out that everyone else had experience only in computing the scaling behavior of non-asymptotically free theories, whereas the first theory he did the computation for was an asymptotically free one, so he thought this was unremarkable. He did say that Gross-Wilczek-Politzer deserved the Nobel since (besides being the ones to publish the beta-function result) they had understood how to use this to explain Bjorken scaling, something that he hadn’t known about. He said his advisor Martin Veltman had told him that the Yang-Mills scaling behavior wasn’t relevant to experiment since experimentalists only cared about what happens on mass-shell. Luckily Veltman was one of the few Nobel theorists not in attendance, since he would likely have blown a gasket if he had been there to hear some of the things ‘t Hooft had to say about him. ‘t Hooft went on to say that he had learned one important thing from this episode: always immediately publish any new result you have.
There was significant mention of string theory in only two talks, those of Gross and Witten. Gross gave essentially the same talk he gave last October at the 25th anniversary of the KITP. He joked that he had managed to time the award of the Nobel with the KITP celebration by every year for the last thirty years writing to the Nobel Committee and asking them to wait a while before awarding him the prize, something they had been happy to do. At the point of his talk when he said that the question “What is String Theory” was one of the big questions for the future, he stopped to defensively note that since we don’t know what string theory is, it is an idea that can’t be killed, no matter how much certain members of the audience wanted to do this. He went on to claim that since AdS/CFT kind of connects string theory with QCD, string theory is in some sense part of the standard model, so it’s importance is secure. This argument seemed to me pretty disingenuous, since presumably he’s well aware that the problem most critics have with string theory is not with the idea of using it as a dual representation of QCD, but with the idea of getting a TOE out of it, a project which some have called a “colossal failure”. He didn’t have anything to say either about this failure or about the whole Landscape mania.
The last talk was Witten’s, entitled “Emergent Phenomena in Condensed Matter and Particle Physics”. He started by saying that he was afraid the title of his talk might be more exciting than the talk itself. By “emergent phenomena” he meant roughly non-perturbative phenomena in QFT, where the long distance degrees of freedom one observes are not directly related to the local degrees of freedom. He gave QED as an example of a non-emergent theory, QCD an emergent one, with the nature of the electroweak theory still up in the air until we know more about the origin of electroweak symmetry breaking.
He went on to say that gravity messes up this distinction between local and emergent phenomena, since one doesn’t have diffeomorphism invariant local observables. He then quoted his 1980 work with Weinberg (and with help from Coleman) to the effect that you can’t get a massless spin two bound state in a theory with a local stress-energy tensor, saying that this showed that you can’t start with a local theory in Minkowksi space and generate Einstein gravity as an emergent phenomenon. For him the lesson is that if you want gravity as an emergent phenomenon, you need to find a way to first get space-time as an emergent phenomenon, and he believes that whatever the primoridial M-theory underlying string theory is, it should do this. While such a theory doesn’t now exist, he went on to give the AdS/CFT correspondence as the kind of thing he had in mind. There the Weinberg-Witten argument is evaded since a QFT in 4 dimensions is related to a gravity theory in a different number of dimensions (5).
Maybe I am missing something, but how can QCD be dual to a gravitational theory in AdS_5 when it doesn’t have a massless spin-2 particle?
Isn’t that the same problem which killed string theory as a dual theory to QCD? The absence of a massless spin-2 particle?
Or perhaps the graviton IS a QCD bound state? That would be shocking!
But I guess if you don’t have a Planck brane…
I want to talk about Sidney Coleman. I met him almost 25 years ago. I was a 1st year grad student. He struck me as an “old” man. Now I know that he couldn’t have been older than 45.
Unlike the rest of the Harvard “superstars”, he was not arrogant, he was not patronizing, and he was not abnoxious. He was approachable, and he very patiently listened to questions, then very kindly answered them. Sometimes his answer was an honest “I vaguely know these things. I’m not an expert”.
In those days, there was a joke circulating around the physics dept at Harvard (I think it’s attributed to Claude Bernard who was a grad student of Weinberg’s at the time):
“How do you do physics at Harvard? You go to Witten to give you a problem to work on. You go to Coleman to tell you how to solve it. Then you go to Weinberg to write you a reference letter.”
I’m saddened by his illness, and by the fact that he can longer teach physics and discuss physics.
“That’s only possible for massive particles and the theorem doesn’t apply.”
I thought the purpose of the theorem was to rule out massless bound states; not to assume that all bound states are massive. Anyway, I’m probably missing the point.
Regarding the “gravity as a gauge theory” issue, Deser’s bootstrap approach does give you the Einstein-Hilbert action but I believe it only works for spacetimes that are topologically R^4. Penrose has shown that there is no map from Schwarzschild spacetime to Minkowski spacetime that takes lightcones into lightcones, which jives well with the Witten-Weinberg argument.
P.S. you’re describing a topological quantum field theory.
That’s only possible for massive particles and the theorem doesn’t apply.
Probably a dumb question, but what happens to the Witten-Weinberg theorem if the particular QFT admits of no single-particle states except on length scales very much smaller than the natural scale of the theory?
No, Lee Smolin wasn’t there. But a message from him was read at some point I recall.
Peter, was Lee Smolin there for this conference since he
was Coleman’s graduate student?
I seen this issue has people all tied up in knots?:)
This idea is called emergence. It’s a familiar phenomenon in the theory of condensed matter, which is Laughlin’s background. Solids and liquids sometimes play host to strange entities that bear little resemblance to the atoms making up the substance. For example, in some materials there are things called spin waves. Every atom acts a bit like a small magnet, with a north and a south pole aligned along its spin axis, and spin waves are oscillations in the alignment of these spins. “Somewhat like what would occur if one took a supple picket fence and rapidly twisted one end back and forth,” says Laughlin. Because this is the quantum world, waves can be considered as particles, and vice versa, so spin waves behave like a kind of emergent particle.
Maybe the views expressed a few years ago by Laughlin and Pines, and by Laughlin in his new book, have something to do with Witten’s choice of the word “emergent”. Chris Quigg’s forceful response is contained in this article.
I talked to Lubos for more like 20 seconds, not 20 minutes. I often have perfectly civil e-mail exchanges with him, your guess is as good as mine why he adopts the obnoxious hyper-aggressive persona in his public internet writings.
I kind of agree with you that Witten’s choice of the buzzword “emergent” wasn’t such a great idea. This is a subject that really would benefit from precision of language to combat the wide-spread fuzzy and wishful thinking going on. But his talk did give some insight into hows he thinks about this, which I found interesting.
Peter, a theorem is a theorem is a theorem is a theorem, and I have no problem with the Weinberg-Witten theorem (or any other theorem) IF it is clearly and correctly stated, thus making clear its domain of applicability (and inapplicability).
I DO have a problem with Witten’s use of the popular buzz-word “emergent” in such a restrictive sense that it excludes such well-known formulations of gravity as Deser’s (mentioned by jkg) and MacDowell-Mansouri,
then, having dismissed Deser, MacDowell-Mansouri, etc, by artful language, claiming that “… a theory [of]… gravity as an emergent phenomenon … doesn’t now exist …”
then claiming that HIS pet AdS/CFT model is “… the kind of thing he had in mind …” to describe emergent gravity because in AdS/CFT “… the Weinberg-Witten argument is evaded …”.
Tony Smith http://www.valdostamuseum.org/hamsmith/
To me a bigger mistery than graviton and gravity is what you had to talk to Lubos when you meet. Does he seem like the same mean and aggressive Lubos we see on the blogs? What is your impression of him? You talked for 20 minutes. Is it aboout string theory, none-string theory? Or you could only talk about Boston’s weather and global warming, and nothing in physics? Just curious.
About Deser result, is it true? can we recover full Einstein gravity? There has been a recent paper which discusses that,
Title: From Gravitons to Gravity: Myths and Reality
Where does an emergent spacetime leave perturbative string theory? What is the meaning of maps from the worldsheet into spacetime, if the latter only exists as a long-distance approximation?
By the vague statement of Witten’s I quoted, he was referring to something very precise: his 1980 work with Weinberg. There they show a problem with the idea of getting the graviton to appear as a non-perturbative effect, as a bound state of some more elementary constituents.
In MacDowell-Mansouri, I think the vierbeins are components of the gauge field, the graviton comes from them, not as a non-perturbative bound state. Anyway, you can argue with Witten’s vague statement, but Weinberg-Witten is more or less a theorem.
I am also puzzled by Witten’s comment that “you can’t start with a local theory in Minkowski space and generate Einstein gravity as an emergent phenomenon. …”. Perhaps this is because I do not understand what “emergent” means in this context. I always believed that it was clear since Deser paper of 1970 that if you want to couple spin 2 field to energy momentum, then you must end up with Einstein theory, with possible higher order corrections (that’s why it is not surprising at all that in beta function calculation for heterotic string for example you get einstein action in the leading order.) Perhaps what he has in mind is that graviton is not fundamental and emerges only in low energy approximation (I remember that there were some works on this in 80s). But then again, whatever the fundamental theory is, Einstein gravity must emerge, just as a result of its universality.
Could anyone clarify this.
Peter, you describe “… Witten’s … saying that … you can’t start with a local theory in Minkowksi space and generate Einstein gravity as an emergent phenomenon. …”.
How is that reconciliable with the obvious existence of the MacDowell-Mansouri mechanism that produces the Einstein-Hilbert action from a Spin(2,3) = Sp(2) gauge theory in 4 dimensions ?
References to the MacDowell-Mansouri mechanism include Frank Wilczek’s hep-th/9801184 in which he says that the MacDowell-Mansouri “… approach to casting gravity as a gauge theory was initiated by MacDowell and Mansouri … S. MacDowell and F. Mansouri, Phys. Rev. Lett. 38 739 (1977) … , and independently Chamseddine and West … A. Chamseddine and P. West Nucl. Phys. B 129, 39 (1977); also quite relevant is A. Chamseddine, Ann. Phys. 113, 219 (1978). …”.
Other references include M. Botta Cantcheff who says in gr-qc/0010080 “… MacDowell and Mansouri proposed a gauge theory of gravity based on the group SO(3,2) …”, and a number of other papers and books, including Mohapatra in section 14.6 of his book Unification and Supersymmetry, 2nd edition, Springer-Verlag 1992 and Freund in chapter 21 of his book Introduction to Supersymmetry, Cambridge 1986 and Ne’eman and Regge (Riv. Nuovo Cim. v. 1, n. 5 (1978) 1, at pages 25-28) and Nieto, Obregon, and Socorro in gr-qc/9402029.
It is true that construction of the MacDowell-Mansouri mechanism was motivated by using it in supergravity models, many of which involved dimensions higher than 4, but as some of the references (for example, Freund’s book) explicitly show, the mechanism can
be applied and does work in the context of a 4-dim spacetime.
Am I stupidly missing something basic, or was this an error by Witten? If it were an error, then why was it not caught by Witten or a coworker before he made such a widely attended talk?
Tony Smith http://www.valdostamuseum.org/hamsmith/
Found on the web: “PBS recently aired a series on “string theory” which, to me at least, seemed to lay some groundwork for mathematical proofs of the existence of God – not that God needs such a proof to exist. “
So extra dimensions are but a trick to formulate no local theories. Fine.
Pingback: Not Even Wrong » Blog Archive » Sidney Coleman 1937-2007