Faster-than-light neutrinos observed again

UPDATE (via Matthew Cobb): There’s a good take on this experiment, including what it would mean for physics if the speed of light were exceeded, over at Professor Jim al-Khalili’s site (he doesn’t buy the result and will eat his boxer shorts if it’s true).


According to the New York Times, the same group (“Opera”) that found neutrinos moving faster than the speed of light have repeated the experiment and found the same anomalous result.  And they’ve eliminated one problem that might have made the first observation erroneous:

When these results were presented to a meeting at CERN in September, after a prairie fire of blog rumors, they were greeted by fierce skepticism. Among the problems with the original experiment, scientists said, was that the neutrinos were produced in bursts 10,000 billionths of a second long — much bigger than the discrepancy in arrival time.

Last month CERN retooled so that the neutrinos could be produced in shorter bursts, only 3 billionths of a second long, making it easier to match neutrinos at Gran Sasso with neutrinos at CERN, and the experiment was briefly repeated. The neutrinos still arrived early, about 62 billionths of a second early, in good agreement with the original result and negating the possibility, the Opera team said, that the duration of the neutrino pulse had anything to do with the results.

The details of both the first and second round of experiments are contained in a paper posted on the Internet at and submitted to the Journal of High Energy Physics.

But problems remain, one being how the clocks were synchronized between Geneva and Italy—the 735 km. path taken by the neutrinos. I thought that had already been taken care of, but apparently not.  Nevertheless, physicists, including those who did the study, are still skeptical:

But the group admitted that many questions remain. “This is not the end of the story,” said Antonio Ereditato of the University of Bern in Switzerland, the spokesman for the collaboration, explaining that physicists would not accept the result that neutrinos could go faster than light until other experiments had come up with the same conclusion. “We are convinced, but that is not enough in science,” he said. . . .

I find this a bit weird.  The people who did the experiment should be at least as hard to convince as the physics community, for the possible problems are the same for all of them, and the researchers who publish the result have much more to lose.  Further, as Feynman said, “the first principle is that you not fool yourself, and you are the easiest person to fool.”

Alvaro de Rujula, a CERN theorist, said there were two interpretations of the experiment. “One is that they have stumbled upon a revolutionary discovery; the other, on which I would place my bet, is that they are still making and not finding the very same error.”

With this kind of skepticism properly infusing our community, it’s no surprise that scientists take deep exception to the far less evidenced claims of theology.


  1. Matthew Cobb
    Posted November 19, 2011 at 1:33 pm | Permalink

    It is a bit like finding fossil rabbits in the pre-Cambrian. You’d be rather worried, knowing that it would overthrown everything we’ve known up till now. On the other hand – how exciting would that be!?!

    Here’s what I tweeted earlier today, linking to a post by UK physicist and TV scientist, Jim Al-Khalili:

    @jimalkhalili goes for Occam’s Razor re those pesky neutrinos. I know nothing, but I reckon he’s right.

  2. Gregory Kusnick
    Posted November 19, 2011 at 1:35 pm | Permalink

    The people who did the experiment should be at least as hard to convince as the physics community…

    I agree, and I expect the Opera team would agree too. The difference is that they’ve spent a lot more of their own time and energy checking the results than the rest of the physics community has. I’m not saying they’re right; I’m just saying they may be justified in feeling that they’re farther down the road to conviction than anybody else because they had a head start.

    • Alexander Hellemans
      Posted November 19, 2011 at 2:35 pm | Permalink

      If I’m right, in the first run they detected about 40 neutrinos from CERN at Gran Sasso. One of them seemed to have oscillated into another type of neutrino. So you could say that there where some neutrinos on the verge of oscillation. So I’m just wondering, what happens to the wave function when a neutrino oscillates? Remember Alain Aspect’s experiment whereby one particle knows instantaneously about the other particle’s quantum state (spin) when created by the same event (the wave function spreads out in space). Communication happens there instantaneously, faster than the velocity of light. Any quantum physicists on this list?

      • Alexander Hellemans
        Posted November 19, 2011 at 2:55 pm | Permalink

        Of course read where==were.

      • Torbjorn Larsson, OM
        Posted November 19, 2011 at 8:37 pm | Permalink

        I am not a quantum physicist as such (solid state physics), but I can tell you anyway that it is wrong to claim that entanglement breaks relativity. The whole point with Bell test experiments is that they _confirm relativity_, which means confirming causality and localization.

        It is well known that entanglement carry correlation but not causation (information). You can’t resolve an entangled experiment before you compare the two measurements, which means the information travels by light post experiment.

        Interestingly, it seems not only the quantum field and its particles but also the wavefunction is a real object. This intuitively pleasing and theoretically reasonable result could be tested soon. Meanwhile, here is how to measure it.

        As for Bell test experiments, it is expected that the practical applications will be extensive. “No one foresaw in 1964 that Bell’s theorem would sow the seeds for quantum information theory and quantum cryptography — both of which rely on phenomena that aren’t possible in classical physics. Spekkens thinks this theorem may ultimately have a similar impact. “It’s very important and beautiful in its simplicity,” he says.”

        If the wavefunction is real, classical Copenhagen based on the Born rule statistical interpretation falls. The simplest theory that takes the wavefunction for real is the realistic Many World Theory. It also incorporates the decoherence process, which means there is no “instantaneous” collapse in entanglement experiments.

        Other experiments, which haven’t been able to unambiguously test decoherence, nevertheless suggests quantum systems can move continuously in and out of decoherence. This is also expected from the simple toy model of Hardy. It is pre- and post states of classical physics which is the discontinuous physics, while quantum physics admits a continuous transformation between states.

        What does it all mean? If the wavefunction is real, if decoherence is the process, if relativity holds, then there is no meaning to adhere to the results of measurements that have not been performed. Empirical counterfactuals are as meaningless as counterfactuals in philosophy (“the road not taken”).*

        Which means that it doesn’t matter what time decoherence takes to make another problem for entanglement woo. The idea that there was a physical spin before the measurement is simply mistaken. The wavefunction contains the information that, together with the wavefunction environment, constitutes the spin under an interaction.

        In the MWT I believe the information is parceled away as real spin in inaccessible superpositions between decoherences, spread out over worlds, so you regain its reality in the “bird’s eye” overview. But I’m not conversant with MWT yet.

        * I am sure Coyne is pleased with this problem for philosophical “free will”. =D

        • Torbjorn Larsson, OM
          Posted November 19, 2011 at 8:45 pm | Permalink

          If and when my comment comes out of moderation, I believe there is a link fail. Another try at that part of the text:

          As for Bell test experiments, it is expected that the practical applications will be extensive. “No one foresaw in 1964 that Bell’s theorem would sow the seeds for quantum information theory and quantum cryptography — both of which rely on phenomena that aren’t possible in classical physics. Spekkens thinks this theorem may ultimately have a similar impact. “It’s very important and beautiful in its simplicity,” he says.”

    • jay
      Posted November 19, 2011 at 8:21 pm | Permalink

      They admit to having a bias for their hypothesis. I don’t see that as a bad thing, since they recognize their bias and seem willing to see it challenged.

      Many discoveries (expected and unexpected) result from researchers honestly investigating their personal hunches.

  3. physicalist
    Posted November 19, 2011 at 1:42 pm | Permalink

    I find this really annoying.

  4. daveau
    Posted November 19, 2011 at 1:43 pm | Permalink

    “One is that they have stumbled upon a revolutionary discovery; the other, on which I would place my bet, is that they are still making and not finding the very same error.”

    Still the nuance of an error like that means that they are still learning something

    • daveau
      Posted November 19, 2011 at 1:44 pm | Permalink


      • Posted November 19, 2011 at 2:06 pm | Permalink

        “Hello? Department of Redundancy Department?”

        • Posted November 19, 2011 at 7:00 pm | Permalink

          No, two “still”s works because they have different applications:
          Even so, the nuance of an error like that means that they are even now learning something.

          • Posted November 19, 2011 at 8:22 pm | Permalink

            Well, not to prolong so trivial a quibble about so trivial a matter, but while you’re technically correct that each “still” fulfills a separate function (which would’ve been made more obvious by the use of a comma after the first “still”), from an aesthetic point of view the redundancy is “still” there.

            • Chris Booth
              Posted November 20, 2011 at 7:19 am | Permalink

              Oh, be still. 😉

  5. Posted November 19, 2011 at 1:51 pm | Permalink


  6. DiscoveredJoys
    Posted November 19, 2011 at 2:05 pm | Permalink

    Obvious, innit.


    where n = speed of neutrinos. The speed of light is slightly slower and not the upper limit at all.

    I claim my Nobel prize. {Removes tongue from cheek}

    • Sigmund
      Posted November 21, 2011 at 7:58 am | Permalink

      Wouldn’t it be E=Smn^2?
      Where E = energy, M = mass, n= the speed of neutrinos and S (the Sigmund!) the new constant that converts the speed of neutrinos into the speed of light.

  7. Alexander Hellemans
    Posted November 19, 2011 at 2:42 pm | Permalink


  8. MadScientist
    Posted November 19, 2011 at 2:56 pm | Permalink

    I agree with Al-Khalili. I’m betting that the timing is wrong somewhere. It is very difficult to get the time right when you have an experiment taking place over such distances and with such short transit speeds. A new experiment needs to be planned; there is no value whatsoever in repeating the same experiment with minor fiddling because that does not address the timing issues. A 60ns difference is a matter of a mere few meters of physical separation or a few meters of extra cabling somewhere. Getting a precise enough synchronization is yet another demanding issue. Personally I’m not even convinced that the physical separations are known well enough. You need some impressive custom hardware for the time measurement as well; the latency time of the best computers on the planet are not good enough for the job.

  9. Chris Granger
    Posted November 19, 2011 at 2:57 pm | Permalink

    With the ‘FTL’ neutrinos (as well as the recent claims of ‘cold fusion’) making the headlines, it’s interesting to me as a layman to see how things unfold.

    I’m inclined to say it’s almost certainly a problem with the experiment and not a revolutionary new piece of physics knowledge.

    If neutrinos can travel slightly faster than light, I wonder what would cause them to have this particular speed limit. Would it imply that light is being slowed down by something somehow, even in a vacuum? An effect of pervasive gravity? Is there anywhere in the universe where light can travel in a truly straight line, with no warping of space?

    • Sili
      Posted November 21, 2011 at 6:11 pm | Permalink


      Pons and Fleischmann hit the stage in ’89.

    • Gregory Kusnick
      Posted November 21, 2011 at 7:13 pm | Permalink

      Light does travel in truly straight lines. Einsteinian geodesics are by definition as straight as it gets. There is no objective, unwarped reference frame by which any “truer” straightness could be measured.

  10. Doc Bill
    Posted November 19, 2011 at 5:06 pm | Permalink

    OK, Brighter than Me people, what am I missing?

    This report has been out for a while and has still not been addressed, or if it has, I missed it. Granted, it’s nice to eliminate other potential sources of error, but I would think the new experiment would run afoul of the problem below, too:

    Dutch researcher Ronald A.J. van Elburg lays out the case that the GPS satellite measuring the neutrinos’ movements was also moving relative to the CERN and OPERA facilities as it orbited the Earth. Briefly, van Elburg asserts that the effects of relativity as they pertain to the GPS satellite’s measurements require two corrections to the perceived time of travel.

    Lo and behold, it turns out that applying that double correction shaves 64 nanoseconds off the neutrinos’ travel time, according to van Elburg, “[t]hus bringing the apparent velocities of neutrinos back to a value not significantly different from the speed of light.”

    • Alexander Hellemans
      Posted November 19, 2011 at 5:36 pm | Permalink

      Yes, it appeared on arXiv more than a month ago, but it doesn’t seem to have been taken seriously since CERN published the results of a second run just now. They must have been aware of this paper.

      • DrDroid
        Posted November 20, 2011 at 10:18 am | Permalink

        It is my understanding that when a clock is synchronized to GPS Time via GPS Time Transfer, rotation of the earth is already accounted for in the time transfer calculations. Hence the people running the experiment, who presumably include engineers with GPS expertise, probably think Elburg’s calculation is irrelevant and it has therefore been ignored. But it is interesting to note that the GPS system and GPS receivers assume the correctness of Einstein’s SR and GR and use these as the basis for geolocation and clock synchronization. Yet this theoretical structure is implicitly being used to prove that the theory is wrong! Something circular is going on here…

        • Gregory Kusnick
          Posted November 20, 2011 at 11:57 am | Permalink

          If a theory can be used to prove itself wrong, then surely the problem lies with the theory and not with the proof.

          That said, this experiment is not a proof of anything. It’s just an anomalous measurement that needs explaining. So it makes sense to use the best theoretical tools available in constructing an explanation.

    • PB
      Posted November 19, 2011 at 6:54 pm | Permalink

      Yes. This must be it. See the difference in timing (caused by error in GPS) is basically the same. Wonder why it was not on the big news, party-pooper?

    • Gerdien
      Posted November 20, 2011 at 3:03 am | Permalink

      Van Elburg

    • Eric Shumard
      Posted November 20, 2011 at 10:08 am | Permalink

      The van Elburg paper is wrong and the author does not understand how OPERA is using GPS to synchronize the clocks at the two locations. The OPERA collaboration is aware of this paper and that it is wrong. Many details about the experiment, including clock synchronization, are available in this PhD thesis from one of the collaborators:
      OPERA uses Common View GPS which uses the measured arrival time at two different locations of a signal from a satellite. The fundamental principle of special relativity is that the measured velocity of a light signal by any observer is independent of the motion of the signal source. So knowing the difference in distance from the two locations to the satellite allows one to compare the measured arrival times at the two locations (using atomic clocks) and use that to apply a correction to the difference in times measured at the two clocks.
      There are relativistic corrections that need to be accounted for, which are incorporated into the GPS system, but the effect described in the van Elburg paper is not applicable to Common View GPS. Arxiv publications are not peer reviewed.

      As to the skepticism of the collaboration, a significant fraction of the members refused to agree to publication until this latest test was made. OPERA has been sitting on this result for some time while they performed more checks (the PhD thesis linked to is from May 2011).

    • Posted November 20, 2011 at 10:58 am | Permalink

      The van Elburg calculation was mentioned by the NY Times in an Oct. 25 article. In it, van Elburg says that he has revised his paper (presumably in response to the Opera group’s criticisms), and has submitted it for publication.

  11. Torbjorn Larsson, OM
    Posted November 19, 2011 at 7:47 pm | Permalink

    The problems with this is manyfold.

    – If real particles go faster than light they will destabilise the vacuum by way of special relativity. (Radiate energy indefinitely as they accelerate away from near light speed to infinity.)

    – Ftl particles makes for time travel. That will destabilise the whole universe by way of general relativity. Or even algoritmic theory, as it would mean physical system can solve difficult problems and wouldn’t make for the complexity we see.

    – There is no Cherenkov radiation as expected for particles traveling faster than surrounding photons travel. (Compare nuclear reactors.)

    The people who did the experiment should be at least as hard to convince as the physics community,

    Good point. But note that the author list has changed between v1 (original experiments) and v2 (added experiment). Some has come onboard while some have dropped out.

    Also note that the new of necessity low-intensity experiment has worse timing, 50 ns jitter against the needed resolution of 60 ns delay. The only way to regain the needed resolution was to make a statistical estimate.

    I read somewhere, an anecdotal rumor, that people in the group were dissatisfied and stressed by it all.

    We will have to wait for independent confirmation. Thankfully other groups were getting into this, so it will take but a year or so.

  12. Gerdien
    Posted November 20, 2011 at 2:58 am | Permalink

    see for instance

  13. Posted November 20, 2011 at 8:46 am | Permalink

    I’m sure they will have thought of this, but wouldn’t the fact that the neutrinos are not following the curvature of the earth but going directly between the two locations, make it appear they are travelling faster?

  14. Golkarian
    Posted November 20, 2011 at 9:11 pm | Permalink

    I don’t understand why they don’t just race a light beam with the neutrinos, it would eliminate the timing error, although it might cause problems as the light travels through the atmosphere.

    • Gregory Kusnick
      Posted November 20, 2011 at 10:13 pm | Permalink

      The neutrinos aren’t traveling through the atmosphere. They’re traveling through solid rock miles below the surface.

      In any case, trying to synchronize a light pulse with the neutrino pulse would just introduce a whole different set of timing dependencies that would have to be accounted for.

      • Sigmund
        Posted November 21, 2011 at 8:04 am | Permalink

        Is there any way to set up a detector at a different distance – say ten times further than this initial distance? That way the difference should be ten times what has been observed in this current set of experiments.
        I guess that is partly what will be tested with other laboratories checking the result with their own detectors.

  15. Posted November 23, 2011 at 10:26 am | Permalink

    — James Ph. Kotsybar

    Oh, little neutral one of tiny mass,
    who flies anomolously from the sun,
    you zip through matter photons cannot pass:
    Could this explain the races you have won?

    From Einstein, few believe that it could be
    that any mass can go as fast as light —
    it’s deemed complete impossibility,
    assuming Relativity is right.

    If proved, the implications terrible,
    will give complacent physicists a scare.
    In terms that twist the ancient parable
    it’s you that’s tortoise; the photon’s the hare.

    It seems, though steady, light can’t keep up pace.
    You oscillate, and yet you win the race.

%d bloggers like this: