This Week’s Hype

Today Slashdot brings us the news that Gamma Ray Anomaly Could Test String Theory. As usual with such media claims about the testability of string theory, this is complete nonsense. The story is based on this Scientific American blog posting, which in turn is based on this paper by the MAGIC gamma ray telescope collaboration.

The claims about testing string theory aren’t in the paper, but appear to come from string theorist Dimitri Nanopoulos who claims that he predicted (or, more accurately, “suggested”) the kind of effect seen by MAGIC using string theory. As far as I can tell though, just about no string theorists except Nanopoulos and his collaborators Nick Mavromatos and John Ellis actually believe this. Mavromatos and Nanopoulos also believe that string theory is responsible for the way that our brains work, here’s the abstract of one of their papers on this:

Microtubule (MT) networks, subneural paracrystalline cytosceletal structures, seem to play a fundamental role in the neurons. We cast here the complicated MT dynamics in the form of a 1+1-dimensional non-critical string theory, thus enabling us to provide a consistent quantum treatment of MTs, including enviromental friction effects. We suggest, thus, that the MTs are the microsites, in the brain, for the emergence of stable, macroscopic quantum coherent states, identifiable with the preconscious states. Quantum space-time effects, as described by non-critical string theory, trigger then an organized collapse of the coherent states down to a specific or conscious state.

Claims have been made by many string theorists that not only does string theory not predict this kind of violation of Lorentz invariance, but exactly the opposite: string theory predicts no such violation. String theorist Jacques Distler earlier this year even went so far as to have the University of Texas issue a press release trumpeting his claims to have shown that string theory is falsifiable, using a calculation based on the assumption that string theory preserves Lorentz invariance (either his colleagues or a PRL referee wouldn’t let him make this claim in the paper the press release was based on, but that’s another story…).

Claims have been made (although there is controversy about this), that the main competing quantum gravity research program, Loop Quantum Gravity, predicts this sort of violation of Lorentz invariance, and this would be one way of distinguishing it from string theory. Lubos Motl has a new posting about the MAGIC result, mainly concerned with knocking it down since he fears that it will be used as evidence for LQG and against string theory.

It seems to me that in any case, the actual experimental evidence here is far too weak to support any claim that a violation of Lorentz invariance has been shown. Among the usual nonsense on Slashdot, there was the following sensible comment about the MAGIC result from an astrophysicist:

What they are saying is that there are still details we don’t understand about AGN [active galactic nuclei] like Markarian 501. So, while this effect could be a first sign of quantum gravity (*not* string theory in particular, as others have pointed out), it could also simply be something going on in the intrinsic spectrum of the flares themselves. I’d personally consider the second explanation more likely at this stage.

As they also point out, one approach to sort out the ambiguity would be to observe other flary AGN at different redshifts (distances). One could then, for example, see if the delay gets shorter or longer as the distance changes, as one would expect with a quantum gravity effect due to propagation to Earth.

Utterly Off-topic, But How Can I Resist Mentioning: According to this blog entry by a USC student, not only am I the “archnemesis” of string theorist blogger Clifford Johnson, but also

If string theory were a vampire, he’d be Buffy.

I’ll have to consult my friends and colleagues on the resemblance to Buffy question, personally I don’t see it.

I don’t know about vampires, but these “tests of string theory” are kind of like the living dead, staggering around trying to get their teeth into people and turn them into string theory partisans. No matter how often you blow their heads off with a shotgun, more keep coming…

Update: Lubos and I seem to be in complete agreement about this experimental result and the Nanopoulos et. al. explanation of it. This situation appears to have driven him over the edge.

Update: See Backreaction for a more detailed posting about the MAGIC result.

This entry was posted in Uncategorized. Bookmark the permalink.

43 Responses to This Week’s Hype

  1. Yatima says:

    Microtubule (MT) networks, subneural paracrystalline cytosceletal structures, seem to play a fundamental role in the neurons.

    Yes, and it’s called scaffolding. “Trigger an organized collapse of the coherent states down to a specific or conscious state”? Where do I find the paper that equates a “collapsed quantum state” to the “conscious state” (aka. the “unicorn state”, often hinted at in dark, confusing tales, never actually photographed in the wild).

    Still, that paper is dated “1995”, and if I remember well, “Microtubules -> Consciousness” inferences, sometimes via Quantum Mechanics, sometimes via old-school Turing Computation made several appearances in Artificial Life proceedings at those time. Even Sir Penrose entered the game. Extremely speculative. Overall, an approach that went precisely nowhere. Anyway, an area to not get into.

    Maybe I should be happy to languish in engineering.

  2. ks says:

    Maybe I should be happy to languish in engineering.

    I guess each discipline has its own no go areas that can be colonialized / made up by members of others as their eccentric hobby. Roger Penrose could do all kinds of speculations about mind, Goedels theorem and quantum gravity. As a cognitive scientists his reputation was done but as a mathematician he won’t endanger his credibility or “core competence” by doing such excursions ( as long as they aren’t too mad ).

  3. Moshe says:

    If a theory violates LI at high energies, say the Planck scale, standard renormalization group suggests that at low energies this will manifest itself by a series of relevant, marginal and irrelevant operators. Bounds on such Lorentz violations at accessible energies are extremely tight, no current experiment will improve those bounds. So, high energy breaking of LI is to my knowledge sufficient grounds for falsifying a theory, or at least casting very strong doubts on it (spontaneous violation is a different story).

    Also to my knowledge LQG does not predict such violation, and Lee does not claim it does. He only claims it might once we understand things better. It escapes me why that should be a good thing, but maybe I am missing something.

  4. Eric says:

    John Ellis and Dimitri Nanopoulos happen to also be the second and fourth most cited high-energy physicists, with Witten first and Weinberg third. Thus, you’re attempt to try to dismiss them as crackpots just isn’t going to fly.

  5. sophia says:

    The link to Motl’s blog leads to something very amusing. Check it out.

  6. bhabha says:


    You are seriously underestimating Peter’s capabilities. Being the second and fourth most highly cited high energy physicist (more precisely, phenomenologist) does not deter Peter from declaring him/her to be a non-sense creating pseudo-scientist. Actually, the more famous person is called a crackpot here the better, because more controversy brings more readers to this blog. It doesn’t work like in science that you have to create quality content in order to get noticed, in the blogosphere the more crazyness you produce the more attention you will get. Pretty much like in mass media which of course includes blogs these days. So don’t be surprised to see more top cited phenomenologists, string theoriests, non-string theorist, etc, etc, getting on Peter’s public enemy list!


  7. Eric says:

    I would never underestimate Peter’s abilities capabilities, especially in regards to using underhanded tactics. 😉

    Regarding the statments about Lorentz violation in string theory, what Lubos and Distler refer to is critical string theory. It’s possible to get this effect (frequency dependent speed of light) in non-critical string theory.

    In regards to the LQG vs. string theory debate, I think this is one more bit of evidence that there is some overlap between the two theories, and they may be part of the same larger theory as Lee has suggested.

  8. Peter Woit says:

    Eric and Bhabha,

    Unlike you I’m not personally attacking anyone, but discussing their scientific arguments. The argument that the MAGIC results give evidence for string theory is, scientifically, nonsense, and it would be hard to find anyone other than Nanopoulos, Mavratos and Ellis who would disagree. I can’t help noticing that string theorists rarely admit that these bogus “tests for string theory” are indefensible, preferring to instead personally attack me for pointing this out, invoking not science but citation counts in their defense.

  9. Eric says:

    First, the MAGIC results are very interesting and cannot be dismissed. Second, such an effect is of interest not just for string theory, but for quantum gravity in general including LQG. Third, Ellis, Nanopoulos, and Mavromatos did predict this effect several years ago, as you may discover on the arXiv.

    Regarding your statement that noone other than ENM takes these results seriously, how do you know? Have you talked with all string theorists and phenomenologists to get there opinion, or are you just relying on what you heard from Lubos and Distler?

    What’s next, trying to undermine their credibility by mentioning that they once wrote papers with Hagelin?

  10. Peter Woit says:


    If you want to provide us with a list of string theorists and phenomenologists who think that the MAGIC results give evidence for string theory and thus that the Slashdot headline is not nonsense, go right ahead.

  11. anon. says:

    John Ellis and Dimitri Nanopoulos happen to also be the second and fourth most cited high-energy physicists

    This isn’t so difficult when John Ellis single-handedly writes as many papers as any random group of ten other physicists. Writing a lot and getting cited a lot does not in itself make one a great physicist.

  12. hack says:

    ” What’s next, trying to undermine their credibility by mentioning that they once wrote papers with Hagelin?”

    In fact their papers with Hagelin were their high point. It’s been all downhill since then.

    I’ve actually read more Nanopoulos/Ellis papers than I care to admit. I have to laugh when I hear Nanopoulos claiming to have predicted something. It’s a million monkeys with typewriters type of situation.

  13. milkshake says:

    Microtubules are huge proteins, maybe 4 orders of magnitude over than the scale where you can observe quantum effects with neutral molecules. (If somebody tells me that it is possible to see interference pattern by shooting tiny molecules like ammonia in vacuo through a double slit, and that scale of the effect can comparable with the actual size of ammonia molecule, I can believe that.)

    Looking for quantum effects in the mechanics of a living cell is completely New Age.

  14. Luboš Motl says:

    I was told about this new article on this spamblog. You may visit my blog to see some clarifications of the statements made by the individual behind this spamblog.

  15. Brett says:

    It’s not clear whether Lorentz violation is a consequence of string theory or if it’s forbidden by string theory. The same holds for loop quantum gravity. However, if you’re going to look for signs of Lorentz violation, there are many ways to go about it, and looking for an energy dependence in the speed of light is not the most sensitive. Because of the vector character of light, the speed of light is such theories generally depends on polarization as well as energy. Searches for this kind of birefringence are much more sensitive than experiments that look for differences in photon arrival times. Indeed, I know that another high-energy telescope experiment already has much better data on photon arrival time differences, but they have refrained from publishing it, in part because it is not competitive with the polarization bounds.

  16. Who says:

    Because of the vector character of light, the speed of light is such theories generally depends on polarization as well as energy.

    Not true Brett, numerous papers about this from non-string QG researchers
    have ruled out polarization dependence.

    Indeed, I know that another high-energy telescope experiment already has much better data on photon arrival time differences, but they have refrained from publishing it, in part because it is not competitive with the polarization bounds.

    In that case they are doing the QG community a disservice by not publishing, because the polarization bounds are irrelevant. You should urge them to publish.

    I assume you mean they have data which would constrain (if not rule out) energy-but-not-polarization dependence, and that is precisely what I see being discussed.

    My sense is we have a ways yet to go with this issue.

  17. Lee Smolin says:

    Brett is right that polarization odd variations in the speed of light are already ruled out at planck scale by observations of polarized radio galaxies, but polarization even variations in the speed of light are not because they cause no bifringence. The latter are, however, a possible consequence of a deformation rather than breaking of Poincare invariance. If an experiment has data on photon arrival time variation with energy it must, because of the limits on parity odd variation, be parity even, and hence, if there is no other explanation, it could be a detection of a deformation of poincare invariance (so called “doubly special relativity”).



  18. Brett says:

    On what grounds can a polarization dependence in the speed of light be ruled out? One can make an assumption that quantum gravity will have certain features, such as no birefringence, but this will limit the terms in the low-energy effective action to a measure zero subset of the full parameter space of Lorentz invariance violation.

    To me, it seems rather wishful thinking to hope that quantum gravity will have such a profound signature as Lorentz violation, while not interacting with the spin structure of the electromagnetic field. There is no compelling reason why this should be the case. The interactions that avoid birefringence deserve to be tested (and I strongly recommended that the data I saw be published), but they are only a peculiar subset of the possible Lorentz-violating interactions. One can always write down a nonrenormalizable interaction which all previous experiments have been insensitive to, but which a new configuration will test; yet selling it as a profound new test of quantum gravity is illogical. (And all renormalizable varying speed of light theories can indeed be bounded by birefringence.) And if you want to make a generic statement about how well Lorentz invariance has been tested, it behooves you to look at the best bounded sectors, not the worst.

  19. Amnetic says:

    You may visit my blog t

    But, Lubos, you have a message in your site saying you don’t want readers from this blog.

  20. sophia says:

    “If string theory were a vampire, he’d be Buffy.”

    Hah. It’s occurred to me that string theory is to real physics as a drag queen is to a real woman. Unlike real women, drag queens are expertly groomed, and beautifully made up, but when the moment of truth arises, who would you rather be with?

    On the other hand, the results of this experiment are quite predictable. That’s one thing drag queens have going for them, unlike string theories.

  21. Moshe says:

    Lee and Brett, regardless of the new and exotic phenomena that you are discussing, I see no reason Lorentz invariance violation at the Planck scale should be automatically a small effect at low energy. Most conservative estimates based on the existence of LV relevant and marginal perturbations to the standard model (I believe the number of those is 46) makes Lorentz violation basically already falsified, based on existing experimental results. Unless one finds a way to fine tune away a lot of really large violations I am not sure why we are discussing those tiny sub-sub-leading effects.

  22. Lee Smolin says:

    Dear Moshe and Brett,

    Parity odd variation of the speed of light with energy is ruled out to at least order 10^-3 [l_{Planck} Energy] by observational limits on bifringence from polarized sources. See gr-qc/0102093. This is a prediction of lorentz symmetry breaking, therefor it is reasonable to infer that lorentz invariance is not broken at order l_{Planck}. But deformation of Poincare symmetry is another thing entirely, there is still a ten parameter global symmetry algebra constraining renormalization effects, so Moshe’s considerations can be answered directly; these are the leading effects of deformed Poincare symmetry. Since the symmetry group is still present it does rule out as many terms as ordinary poincare symmetry, and one of them is a parity odd varation in the speed of light coming from the usual dimension five term seen in lorentz symmetry breaking.

    To be more precise the Casimer invariant of the deformed Poincare algebra is no longer quadratic in energy and momentum, leading to corrections to the speed of light. Thus, deformed Poincare symmetry can imply an order l_{Planck} variation in the speed of light with energy, which is parity even and therefor not ruled out. Therefor if the right interpretation of the observations reported by the MAGIC collaboration is a modification of spacetime symmetry it must be a deformation and not a breaking of Poincare symmetry-because the latter is already ruled out by experiment, but the former is not. The same holds for the observations Brett hinted about.

  23. LDM says:


    Regarding the paper gr-qc/0102093 (You also mention the same on page 226 of TTWP)…the assumption made in the paper

    “If we assume that linearly polarized photons are
    detected, and unambiguously identified with a source at cosmological distance z, without any significant interaction
    in between, we may be immediately sure that (6) is not strongly violated.”

    is too large of an assumption, in my opinion.

    Cosmological data has traditionally been the poorest data of all the sciences, and you want to draw definite and strong conclusions about physics at the Planck scale based upon it? This to me is very wishful thinking.

  24. Brett says:

    I agree with you about naturalness, which is why I think it’s more important to concentrate on renormalizable forms of Lorentz violation. Indeed, there is no known reason why Lorentz violation, were it to exist, should be small. Lorentz violating interactions are, of course, technically natural, since they receive no radiative corrections from Lorentz-invariant physics, but if they actually exist, they must either be finely tuned, or they must be suppressed by some unknown mechanism.

    The number of dimension three and four Lorentz-violating operators that can be constructed out of standard model fields is much larger than 46. With just one generation of fermions and the electromagnetic field, there are about 150. Of course, many of those mix under renormalization. The number of different symmetry types is still greater than 46 though. Actually, how many different physically meaningful operators there are depends on whether you consider only flat spacetime or whether you consider working in a curved background, and exactly how which terms are physical under which circumstances is not completely understood.

  25. Lee Smolin says:


    I understand reservations about claims of positive results because there could be other explanations, but the claim in gr-qc/0102093 is a negative result. Do you believe that there could be leading order lorentz symmetry breaking which produces bifringence, which is then masked by some ordinary astrophsyical effect so that no bifringence is seen? What physics do you have in mind that could reverse rotation in the plane of polarized radio waves so its effects were not seen?

    It seems to me reasonable to infer that lorentz symmetry breaking is not present, in this case the experiment and the more theoretical argument discussed here by Moshe agree.



  26. LDM says:

    Thanks Lee, probably I was a little ambiguous. Sorry. I will see if I can phrase this more precisely…

    I am not arguing either for or against Lorentz violation, I am only arguing against using the cosmological data as you have done…

    In the TTWP, page 226, you mention that the travel time can be “billions of years” for the photons in question. So we are talking about large distances.

    Now, in the paper, we also have the statement
    “comparing the time of arrival of rays at different energies emitted simultaneously from the
    same source, one can test the validity of this prediction”

    I am assuming that time of arrival that is being discussed is based on our distance from the source of photons. (If it is not, and you do not need to know the distance, then I am wrong, and please accept my apologies for wasting your time.) The problem is that cosmological distances are very uncertain, to quote from M Berry “Principles of Cosmology and Gravitation”:

    “How do we know the distances and densities quoted? The Universe is charted by a sequence of techniques, each of which takes us out to a greater range of distances – to the next level of the ‘cosmic distance hierarchy’. Each level is less reliable than the last, so that there is considerable uncertainty about the measurements of very great distances.”

    So, it would seem
    we have considerable uncertainty in our data…but in TTWP, you are talking about measuring differences of 1/1000 of a second, which it seems to me is a fairly precise or certain measurement…The impression I have is we do not have that kind of accuracy. And so you cannot meaningfully use the cosmological data in the way you are attempting to, which is to measure differences of 1/1000 of a second over large cosmic distances.

  27. Moshe says:

    Thanks Brett, the number 46 came from the Coleman-Glashow paper, if the number is bigger I am even more worried…

    Lee, the fact that global symmetries are preserved by renormalization, and therefore can be used to forbid otherwise possible interactions, this fact was established through a series of theorems in the 1960s and 70s, those theorems apply to ordinary global symmetries.

    If there is some deformed version of Poincare symmetry it is then natural to ask if it is preserved by renormalization. If not it doesn’t give any restriction on the form of the low energy EFT. If it is preserved maybe it limits the allowed interactions in some way, but even then it seems to me it will take a miracle to allow small violations of LI while by forbid much the numerous much large effects.

  28. Eric says:

    The different arrival times are for different gamma rays emitted at the same time from the source with different frequencies, and no you don’t need know the cosmological distance precisely. Once only needs the distance to be large so that the difference in arrival times is measurable. Essentially, the higher frequency gammas interact with the vacuum and slow down, just as a light ray does when going through some medium, otherwise known as refraction.

  29. Moshe says:

    garbled the last sentence, the last few words should be “while forbidding the numerous much larger effects”. Those effects refer to all the renormalizable terms Brett discussed, not just the dimension 5 operator discussed by Lee.

  30. anon. says:

    “And so you cannot meaningfully use the cosmological data in the way you are attempting to, which is to measure differences of 1/1000 of a second over large cosmic distances.” – LDM

    I think you completely misunderstand. If two photons of different energies arrive 1 millisecond apart from the same gamma ray burster or whatever, the accuracy of that measurement (relative time of 1 ms difference) is independent of the less accurate cosmological distance ladder which estimates how far the gamma ray burster is from you. You simply don’t need to know exactly how far away the source is, in order to detect that photons are travelling at different speeds…

  31. LDM says:

    Thank you anon. .Yes, perhaps I misunderstand…but let me ask you, the two photons that arive 1 millisecond apart from the GRB, how do you know that the photons did not leave the GRB 1 millisecond apart too?
    And more to the point that is bothering me, what are the error bounds on these measurements?

  32. anon. says:

    LDM, if the 1 ms delay is not caused by differences in photon speeds, then similar delay times whould show up in the time-dependent energy spectra of gamma ray bursters, regardless how far away they are. If the delay is caused by differences in photon speeds, then the furthest sources should show the biggest delays.

  33. Lee Smolin says:

    Dear Moshe,

    I agree, an important question is whether deformed poincare symmetry is preserved by renormalization. My argument assumed yes, but you are right that this needs to be shown, I am not sure of the status of this in various approaches to QFT with deformed poincare symmetry but will check.

    Dear LDM,

    We seem to be at cross purposes, the paper gr-qc/0102093 does not use arrival times, it uses the absence of bifrengence in observations of polarized radio galaxies. The MAGIC claim does use arrival times, as do limits using lower energy gamma ray bursts set to M_QG. There are several ways this is addressed in the literature: 1) as anon mentions one can hope to get redshifts for enough events and see if there is a correlation with distance, 2) by using very short bursts and bounding the dispersion relation by the overall length of the signal, 3) by a better understanding of the source. None of these apply to the MAGIC claim.

    The MAGIC paper uses another argument based on extremizing energy flux. Does anyone know how reliable this kind of argument is? Is it used elsewhere in astrophysics?


  34. Arun says:

    LDM, if the 1 ms delay is not caused by differences in photon speeds, then similar delay times whould show up in the time-dependent energy spectra of gamma ray bursters, regardless how far away they are. If the delay is caused by differences in photon speeds, then the furthest sources should show the biggest delays.

    which would be detectable provided gamma ray bursters at all eras are essentially identical.

  35. Moshe says:

    Thanks Lee.

  36. alex says:

    Hi Lee,
    I’m confused by the terms “deformation” and “violation” of Lorentz symmetry. It seems to me that if the lagrangian contains some terms which are invariant under the “deformed” Lorentz symmetry but NOT invariant under the usual undeformed Lorentz symmetry, such terms would therefore violate the usual Lorentz symmetry since they are not invariant under it, right?


  37. Bee says:

    Thanks for the link!

  38. Cecil Kirkey says:

    This thread is interesting but…Does ST predict a LI violation or not? What is the difference between critical and non-critical ST? Are they both derived from the same assumptions? If they are do we now have two families of ST? And if these ST guys predicting LI violation are not really preaching the true ST why don’t the true ST guys stand up and refute them openly?

  39. Mr. String says:

    Non-critical string theory involves strings propagating in a dimension of spacetime less than the critical dimension of ten. The resulting anomalies are cancelled by exciting the Liouville mode (linear dilaton) of the strings. The statement that string theory strictly obeys Lorentz invariance is true only in the context of critical string theory.

  40. manyoso says:

    So which version of String Theory, critical or non-critical, to String Theorists believe corresponds to the real world? …oh wait

  41. sasay says:

    This situation appears to have driven him over the edge.

    Who is this Lubos that you mention? Is he affiliated to an academic institution? He seems to refer to L. Susskind with strange reverence, is he another squatter at Stanford?

  42. Hendrik says:

    A somewhat belated item has appeared in New Scientist on
    the MAGIC “test” of string theory at:

  43. Pingback: Not Even Wrong » Blog Archive » Various and Sundry

Comments are closed.