Of Particular Significance

Why ICARUS Doesn’t Refute OPERA

POSTED BY Matt Strassler

POSTED BY Matt Strassler

ON 11/21/2011

[If you want to know more about the OPERA experiment’s most recent result on neutrino speeds, and how it is much better than their previous one but still does not settle all the issues, click here.]

The ICARUS experiment, like the Cohen-Glashow paper on which they rely, are claiming again to “refute” the OPERA experiment on superluminal neutrinos.  (I already wrote about that here.)  I’m sympathetic, personally, but although it does provide a very strong constraint on any modification of relativity that could permit what OPERA observes, logically this does not, in my current view, constitute a refutation.  

What Andrew Cohen and Sheldon Glashow of Boston University pointed out (as explained in detail in this post) is that if neutrinos traveled faster than light there would be an effect similar to Cerenkov radiation that would reduce the energy of the neutrinos in OPERA’s neutrino beam — an effect OPERA would have seen, but did not.  ICARUS is simply confirming this: that the beam of neutrinos from CERN, which it too can measure (it is  quite close to OPERA) shows no sign of the neutrinos having lost energy.

The problem is that logically Cohen-Glashow and ICARUS only prove that either OPERA is wrong or there’s a so-far unknown modification of relativity which allows an escape from Cerenkov-type radiation for neutrinos at the energies of OPERA’s beam. It is certainly true that Cohen-Glashow makes OPERA even less plausible than it was before, but it does not settle the issue.  Since OPERA calls a deep theoretical assumption about relativity into question, it is not a great idea to refute it using another set of theoretical assumptions about relativity that might simultaneously come into question.

So until someone proves that there’s absolutely no way to inhibit Cohen-Glashow radiation without violating some other existing experiment, we still, in my view, have to treat OPERA as deeply implausible but not strictly refuted.

Share via:


43 Responses

  1. An awesome starting point for looking for correct number of tones can be a retail store closest people. The key one issue using these shops is they holds the modern varieties of Oakley pigments, however just what for anyone who is looking to return your old number of Oakleys being focused on an individual color and style? Some might definitely not retain the older share all over again, you may have out there their more costly varieties of items

  2. My apologies to Dr. Cohen, it was never my intention to leave him out of my scornful rant. My concern is that given the small size of the violation of the light speed limit ~c(1.00003) at ~ 30 GeV by muon-neutrinos, the opportunity for potential errors that an undedicated experiment like OPERA presents and owing to the stature and authority of Einstein’s Special Relativity it will be all too easy for the physics community to simply call this an ‘anomaly’ and move on. I don’t want that to happen. I believe (though I cannot justify this) that these OPERA results will prove very important to the future.

    Yes, what you have called OPERA II certainly helps. A modified MINOS (201?) may help even more. But what we really need is a dedicated experiment. That will take both time and money. My fear is that the reaction of an already skeptical physics community operating on limited funds will be tempted to hang their hats on theories like Drs. Glashow and Cohen and call it a day. When I hear from people who are interested in physics (one day after OPERA II news release) that “this other (ICARUS) experiment has refuted the faster-than-light results” (when I know it has truly not) I become concerned. When I read that TV/Book Physicist Lawrence Krauss (writer of the ‘Physics of Star Trek’ – of all things) says that the failure by ICARUS to observe the energy profile of muon-neutrinos predicted by the Cohen/Glashow theory “seals the case”, I become concerned.

    Many readers are no doubt thinking; “No the OPERA observations are too big to be swept under the rug and forgotten.” I hope you are right…

    1. I understand your concern. Lawrence Krauss is a smart guy, but it is a shame that he didn’t think the situation through carefully before making such a strong statement. That’s really not good.

      OPERA won’t be swept under the rug, but you are right that time and money are a constraint on being able to check it, and that the fact that everyone knows that many experiments (even by good experimenters) come out wrong, and that this one is particularly implausible, is an issue. But I do think it will happen. What you have to understand, though, is that anything you propose will take years to build, so you are talking about people taking years of their lives to do it, not to mention justifying the expense, attracting collaborators, and convincing students to work on it.

      For this reason I suspect that the best approach is to design either a cheap relatively quick test that takes two years or less to build (which very well might be impossible) or, if it is going to take several years anyway, make the measurement part of a larger experimental program, something to be done at a new still-to-be-built or soon-to-be-upgraded facility. That may sound like an unpleasant compromise, delaying the knowledge by more than necessary, but most particle physics experiments today have to make that compromise. Also, to justify the cost, the measurement should be more precise than OPERA’s by at least a factor of 10, or be able to reach higher energy neutrinos, or something else novel.

      So we may very well have to live with OPERA as an anomaly for some years. You may find that frustrating, but it is very typical; we have been living with DAMA’s claim of dark matter discovery for over ten years, and no one has found their mistake (not that they’ve been very open with their results!) or been able to conclusively confirm it or rule it out (though I’d say we’re closing in on that point.) There are other anomalies out there too. History books sometimes make science sound like a quick process; it’s not.

  3. Wow, just came upon ‘Of Particular Significance’ yesterday and I loved ‘OPERA: Comparing the Two Versions’. Great job! However in terms of the ICARUS publications I must admit my thoughts are more in line with both ppnl and Vladimir Kalitvianski.

    Correct me if I’m wrong, but (prior to dark matter) according to our best theories the galaxies should fly apart – they don’t. In theory (prior to dark energy) the expansion of the universe should be slowing down – it’s not. In theory the vacuum should blow apart with 10^120 times the force it actually is. In theory a muon and an electron should measure the same proton charge radius – they don’t. In theory nothing can exceed the speed of light…

    I can postulate that muon neutrinos will emit p(ositrons)ink ele(ctrons)phants when they exceed c. But what I cannot do is use the failure to observe pink elephants as proof that the muon neutrinos are not exceeding the speed of light. So, while I have the greatest respect for Dr. Glashow, his theoretical expectations cannot trump experimental observations. Not in PHYSICS.

    In PHYSICS one cannot use a theory to dismiss experimental results. Only further experiments, utilizing high energy muon neutrinos timed over a distance (just as they were in the CERN/OPERA observations) can be used to verify or falsify the results. Otherwise, why pretend we are doing physics, we don’t need Dr. Glashow’s pink elephants, just re-publish Einstein’s original 1905 paper and say case closed!

    1. First, you should not call these Glashow’s whatevers. They are Cohen and Glashow’s whatevers; Cohen is a full professor of physics at Boston University, very smart and accomplished, and it is inappropriate not to give him due credit, even though he doesn’t have a Nobel prize.

      I would have said that your point of view is exactly the same as mine, so I’m not sure what your concerns are. Cohen and Glashow give insight into what cannot be true (namely, it cannot be true that OPERA is correct AND that Cerenkov-like radiation is still allowed as would naively be expected.) It makes OPERA much less plausible, but not refuted, as I said. And it also tells us that various kinds of modifications of Einstein’s equations wouldn’t work. This kind of analysis is a very good example of powerful and useful theoretical physics.

    2. Unfortunately, just publishing the 1905 paper won’t work, because it contains an implied “natural” coincidence of points which Einstein made explicit in RELATIVITY: THE SPECIAL AND GENERAL THEORIES. As you may notice, even those who accept the CERN experiments don’t want to address the central question: is there a logical flaw in the relativity of simultaneity? Yes, there is. It is “natural” coincidence (“fallt zwar…zusammen”), an arbitrary insertion which Einstein–following the constructivist program he laid down in “Geometry and Experience”–made in the formulation of the relativity of simultaneity. You see it in its explicit form in the “train” experiment in RELATIVITY. Here is a discussion of that error and the entire constructivist program of the early twentieth-century. You have to realize that the protocol of the CERN experiments is not a Standard Model protocol.http://ssrn.com/abstract=897085

      1. OPERA is not a CERN experiment (it gets its neutrinos from CERN but it is not run by or funded by CERN).

        Ordinary particle accelerators and the GPS system simply would not work as designed if Einstein’s equations were not accurate in most situations and for most particles. So there can’t be something as fundamentally wrong with them as what you suggest.

        1. Well, you should reexamine the train experiment with “natural” coincidence in mind and then let me know what you think. Among the other observations which have been made in connection with the new set theory historiography (and it is from this work that new inquiry has come regarding Einstein’s intellectual orientation), is that Einstein believed that reality is progressive. I don’t think any physicist today would hold that point of view. In general, there are a lot of ideological positions packed into the relativity of simultaneity. It is not correct to assume that we understand the relativity of simultaneity fully. Above all, it is not so important, from the point of view of relativity, whether particles can go faster than the speed of light. What is important is the ROLE the speed of light plays in the logic of the relativity of simultaneity.


  4. The sociology of this is kind of strange. ICARUS’s result came out in October. Then on Nov 17 OPERA announced their follow-up result. Then for the next few days the media (NPR, yahoo.com,…) were full of stories about the ICARUS result from October, but dated in November as if it were new. AFAICT the media is just finding itself confused trying to follow a complex, highly technical, and fast-moving story. Because peer review is becoming irrelevant, this stuff unfolds quickly.

    1. There’s also the fact that ICARUS updated their paper on the 19th, so there was some news there too. But you’re right, the chronology got quite confused. Well… in the end the science will be worked out no matter how it is reported…

  5. Guys, what it nature said “And now for something completely different!”

    FTL neutrinos are completely different and it isn’t clear how current theory will be applicable to something not predicted by current theory.

    But the thing about being completely different is we need extraordinary evidence to believe it. We do not yet have that extraordinary evidence and that is more damming than any argument about an analog to Cerenkov radiation. Get the extraordinary evidence and arguments about Cerenkov type radiation are irrelevant. Fail to get that extraordinary evidence and arguments about Cerenkov type radiation is irrelevant. Either way arguments about Cerenkov type radiation is irrelevant.

    1. The arguments about Cerenkov radiation are irrelevant to the truth of OPERA, but not irrelevant to our understanding of what it might and might not mean. It is still very important for theoretical physicists to go through this sort of reasoning, and for experimentalists to follow up on it.

  6. I see conversational exchange on specific requirements demanded.:) How is time based measure affected?

    I think people have forgotten about the, “first detection of neutrino oscillations in direct appearance mode in the νμ→ντ channel?”- http://estaticos.elmundo.es/documentos/2011/09/23/cern.pdf


    http://3.bp.blogspot.com/-VV60nOVh4nQ/TsuQTQrEWoI/AAAAAAAACzM/ki1AD-ac2xo/s320/Tau-neutrino1.png- Computer reconstruction of the tau candidate event detected in the OPERA experiment. The light blue track is the one likely induced by the decay of a tau lepton produced by a tau-neutrino- http://www.lngs.infn.it/lngs_infn/contents/lngs_en/public/educational/physics/NEWS_docs/OPERA_press_rel_May_2010.pdf

    See Hadron Stop and then muon….muon and then to Gran Sasso?

    What is happening “through too,” Gran Sasso? Can such process in earth medium used in the muon relativistic function(an anomaly in oscillation) be seen on it’s own in vacuum speed of light constant?


  7. In case of super luminal neutrinos, one can advance whatever one wants as a loss mechanism – this regime cannot be tested. CG use “weak decay in vacuum” mechanism as if it was a regular physical situation. But the situation is not as physical as in case of Cherenkov radiation. The region v >c was never tested. Or better, the region v > c has never been reached in experiments. I suspect, because it does not exist in reality. One may say: the region v > c was proven to be impossible. So the corresponding “physics” cannot be but a speculation, however camouflaged it is.

  8. The previous post lost a part of a text. I restore it:

    In Cherenkov radiation, it is a medium who radiates after getting excited with a charged particle. Apart from Cherenkov radiation (collective medium modes), there are always individual atom ionization and excitation losses at any v < c / n for a fast charged and neutral particle. In both cases the interaction with a medium material is the only loss mechanism.

    1. The problem is that your are confusing the medium for the message. (That’s an idiom in English, and therefore sort of a joke — sorry if it isn’t obvious to non-English speakers.) The medium is just a way of getting light to slow down. But once light has been slowed down, Cerenkov radiation from a fast particle occurs for a much deeper reason. If Object A can emit B waves, and Object A travels faster than the maximum velocity of B waves, a shock wave made from B waves always results. That’s a classical-field-theory statement. Go to quantum mechanics, and you have Cerenkov-like radiation of A emitting B (or B + B-bar).

      So I would claim it doesn’t matter how we get neutrinos traveling faster than the maximum speed for light or for electrons. It follows from very deep principles that Cerenkov radiation results. (We just have to calculate it, which Cohen and Glashow did, using basic methods, in regimes where there is no reason they should fail.) So you must find a way to violate those deep principles if OPERA’s result is to be consistent with known physics.

      1. I disagree that the Cherenkov’s radiation is the particle radiation. No, it is the matter (medium) who radiates after having been affected by long-range and fast changing particle field. The energy transferred to matter relaxes then in one or another way, including/excluding EMW. The role of matter is crucial. When v < c / n, the fast charged particle has losses anyway precisely because of interaction with matter. Frankly, it is the change in the medium state that consumes the transferred energy. In case of one atom, you can push it elastically, excite, or ionize it. The latter case is similar to pair production.

  9. Super luminal kinematics is wrong: it is a speculation. One cannot just put non physical data in a formula (or use non physical formula) and say: “Hey, one more channel of losses is kinematically open now”.

    In Cherenkov radiation, it is a medium who radiates after getting excited with a charged particle. Apart from Cherenkov radiation (collective medium modes), there are always individual atom ionization and excitation losses at any v c was never tested. Or better, the region v > c has never been reached in experiments. I suspect, because it does not exist in reality. One may say: the region v > c was proven to be impossible. So the corresponding “physics” cannot be but a speculation, however camouflaged it is.

  10. Just to elaborate on my question: the computation of probabilities in this case, does it use strictly quantum mechanics equations or other more general statistic assumptions are being made?

    … and many thanks for finding time for keeping up this blog.

    1. The calculation is done in the framework of standard quantum field theory, as one would learn from any textbook on the subject. The use of probabilities is the standard one in quantum mechanics and quantum field theory, again as one would find in any textbook. The only assumption being made in that part of the paper is that relativistic quantum field theory is a suitable framework for calculations. One thing I emphasized in my original post on the Cohen-Glashow calculation is that there *is* experimental evidence in favor of this assumption: the decays of W particles involve the same interactions, and similar kinematics, and everything that’s ever been measured about W particles (and there’s a lot) is beautifully predicted by quantum field theory.

  11. “They compute a probability (as one always does in a quantum theory) but (as always in a quantum theory) the emission is occasional.”

    Could you please elaborate on computation of probabilities in this particular case and the assumptions that are being made.

  12. But so energetical neutrinos, even luminal, have always sufficient energy for pair production (pairs are kinematically possible). Experiments show that this mechanism is rather weak; it does not stop neutrinos in a medium.

    If you speak of “neutrino decays” in _vacuum_, kinematically “possible” due to v > c, it is a pure speculation. There is no a single experiment to substantiate this mechanism.

    1. Again: are you saying (a) [simple kinematics] is wrong, or are you saying (b) [the quantum field theory calculation] is wrong?

      Also, your first statement isn’t clear. You don’t mean that. You mean that pair production is possible *IF* the neutrinos hit something. But the point is that if the neutrinos exceed the maximum speed for electrons, then pair production becomes possible *via decay.* That’s just energy and momentum conservation.

      One useful way to understand Cerenkov radiation is to observe that an electron traveling faster than the maximum speed for light can decay to an electron plus a photon.

  13. “pair production, … that is a well-established process that occurs in quantum field theory in a wide variety of contexts…” Then how to distinguish the regular case v = c from v > c if there is no difference in processes?

    1. The point: if particle A has max velocity v1, and other particles B have max velocity v2 which is less than v1, then A decaying to B Bbar A becomes possible when v1 is sufficiently larger than v2. It’s just kinematics. Then it is a matter of calculating the rate for the case where A is a neutrino and B is an electron. Which they did.

      Now you are either saying (a) A decaying to B Bbar A is kinematically impossible, or (b) Cohen and Glashow made a mistake calculating the rate. So is it (a) or (b), and if so, where’s the error?

  14. Thanks, but Satish Ramakrishna’s paper claims, “This paper substantiates the point, as well as introduces a new source of error in the “common-view” method of synchronization of clocks.” And then he proceeds to illustrate where that error is, with the uncanny result of 56 ns . Have you taken a look at that proposal?

      1. Dear Dr. Strassler – I just saw your comment. My paper should be a quick read and I would appreciate your comments. A reputed physics journal is taking an inordinate amount of time to find referees that understand the details of GPS calibration to review my paper, which seems to me hard to understand – since the experiment essentially stands on two legs – a distance and a time measurement. I would think before people go down the route of hypothesizing extra dimensions without a shred of evidence, they would knock off the “easy” questions.

  15. Their assumptions are speculative in a mechanism (pair production) and in supposition that the losses are “continuous” along the trip.

    1. They do not assume they are continuous. That’s not correct. They compute a probability (as one always does in a quantum theory) but (as always in a quantum theory) the emission is occasional. See the figure in my post on the subject: http://profmattstrassler.com/2011/10/06/is-the-opera-speedy-neutrino-experiment-self-contradictory/

      They do use pair production, but that is a well-established process that occurs in quantum field theory in a wide variety of contexts.

      So where is the mistake, precisely?

  16. Cherenkov radiation is a radiation of a transparent medium due to long Coulomb interaction (fast macroscopic polarization of a medium). There is no such a mechanism for a short-range weak interaction. For example, relativistic neutrons do not emit Cherenov-like radiation in a medium. Why neutrino should be different? Due to v>c? But there is no a single experiment to say something certain, so CG’s reasoning is an (erroneous) speculation.

    1. A detailed calculation based on well-established theory and clearly stated assumptions is not the same as a speculation. Point out the mistake in their calculation first; then we’ll discuss it.

  17. As I am reading Icarus paper – pdf – the say explicitly: .. We find that the neutrino energy distribution of the Icarus events in LAr agrees with the expectations for an unperturbed spectrum of the CERN neutrino beam. Our results therefore refute a superluminal interpretation of the Opera results according to the Cohen and Glashow predictions for weak currents analog to Cherenkov radiation… Would they make this statement if they were not reasonably sure of it’s validity ?

    1. The problem is that it isn’t enough to be “reasonably sure” when you say “I refute”. A mathematician or logician who is only reasonably sure won’t say “I have proved…”; he or she will say “I conjecture and give evidence that…”.

      I’m reasonably sure that OPERA already refuted itself (and we didn’t need ICARUS anyway.) That was Cohen and Glashow’s point, in fact; ICARUS is not necessary for this point, although their results drive it home.

      But I’m not absolutely sure.

      The statement you quote from the ICARUS paper is exactly correct as it stands; they “refute” only “according to the Cohen and Glashow predictions”. The problem is that Cohen and Glashow consider a particular class of ways that standard relativity might be modified. Maybe there’s no other way. But if there is one… well, then it might turn out that with that modification, Cohen and Glashow’s calculation is irrelevant. Do I think that’s likely? No. Do I know that it is impossible? No. Not yet, anyway.

      A refutation with a loophole isn’t a refutation.

    1. See the previous comment by Eric Shumard that points out that van Elburg assumed that OPERA used the GPS system in one way, whereas in fact it was used in a different way (Common View mode).

  18. Agree with you 100%. The ICARUS work depends on the Cohen-Glashow paper which is a theoretical speculation. Until Cohen-Glashow paper becomes a fact, the ICARUS work has no power to refute OPERA.

    In my view, OPERA’s result has nothing to do with the Special Relativity (SR), and it is more relating to the issue of General Relativity (GR). The distance between two points measured by us (composed with protons and neutrons) is a bit different from neutrino’s own measurement.

    1. Well … I’m glad you agree with me, but honestly I’m not sure the reverse is true! 🙂

      I would say: Cohen-Glashow is NOT a speculation. It is a carefully worked out calculation under certain ASSUMPTIONS. The assumptions are quite reasonable, and not in any way speculative; they are based on many things we know very well about quantum field theory, particle physics and Cerenkov radiation. In fact, it is a speculation to suggest the assumptions are wrong!

      But of course… speculating, now… they might be wrong, in principle. And if OPERA is right, then what we learn from Cohen-Glashow (and ICARUS) is that they MUST be wrong. Alternatively, and much more likely, the assumptions are correct, and OPERA is wrong.

Leave a Reply


Buy The Book

A decay of a Higgs boson, as reconstructed by the CMS experiment at the LHC


A quick note today, as I am flying to Los Angeles in preparation for and other events next week. I hope many of you were

POSTED BY Matt Strassler

POSTED BY Matt Strassler

ON 04/09/2024

POSTED BY Matt Strassler

POSTED BY Matt Strassler

ON 03/05/2024