Particles may not have broken light speed limit
(Note: This is my first post in discussion, instead of just a comment. Feel free to suggest improvements.)
A few weeks ago in Less Wrong discussion, there was a discussion on whether or not we had found neutrinos traveling at superluminal velocities. That discussion can be found here:
Apparently, a paper was recently put into Arxiv from Ronald A.J. van Elburg, which has been popping up in the news as having indicated a possible explanation for the difference. To sum up the paper, instead of superluminal velocities, we may have a possible source of GPS error to compensate for.
Some of the news reports also correctly pointed out that this paper is currently as tentative as the first announcements about the Neutrinos were when OPERA made them.
Arxiv link:
http://arxiv.org/PS_cache/arxiv/pdf/1110/1110.2685v2.pdf
Link to Preprint on Author’s page: (Thanks to Shminux)
http://home.kpn.nl/vanelburg30/Papers/RAJvanElburg_TimeOfFlight_Preprint.pdf
Summary:
The Michelson-Morley experiment shows that the experimental outcome of an interference
experiment does not depend on the constant velocity of the setup with respect to an inertial
frame of reference. From this, one can conclude the existence of an invariant speed of light.
However, it does not follow from their experiment that a time-of-flight is reference frame
independent. In fact, the theory of special relativity predicts that the distance between the
production location of a particle and the detection location will be changed in all reference
frames which have a velocity component parallel to the baseline separating source and detector
in a photon time-of-flight experiment. For the OPERA experiment we find that the
associated correction is approximately 32 ns. Judging from the information provided, the
correction needs to be applied twice in the OPERA experiment. Therefore the total correctiotion
to the final results is 64 ns. Thus bringing the apparent velocities of neutrino’s back
to a value not significantly different from the speed of light. We end this short letter by
suggesting an analysis of the experimental data which would illustrate the effects described.
Hypothesis to test if paper is correct:
We showed that in the OPERA experiment the baseline time-of-flight is incorrectly identified
with the Lorentz transformation corrected time-of-flight as measured from a clock in a non4
stationary orbit and in fact exceeds it by at maximum 64 ns. The calculation presented contains
some simplifying assumptions. A full treatment should take into account the varying angle between
the GPS satellite’s velocity vector and the CERN-Gran Sasso baseline. We expect that such a full
treatment will find a somewhat smaller value for the average correction. In addition such a full
analysis should be able to predict the correlation between the GPS satellite position(s) and the
observed time-of-flight.
References in the news:
http://www.pcmag.com/article2/0,2817,2394747,00.asp#fbid=iQWAnqFuW6P
http://www.geekosystem.com/ftl-not-so-much/
As a side note, if this paper is correct, Eliezer may have won some money, since he made bets in a thread about this from the earlier discussion.
Edit: Fixed three broken links and header formatting.
Edit2: The original author retracted some of his comments on GPS. http://home.kpn.nl//vanelburg30//Publications.html has the link to why. Thanks to Shminux for finding this.
I don’t know why this particular paper has gotten so much attention. Since the initial announcement there have been a lot (I haven’t counted. My guess would be around a hundred) papers on the arXiv trying to explain this, most of which propose mundane explanations. Which papers are getting heavy attention in the general press seems to be pretty random.
This expected arbitrariness of the pick, general unsuitability of LW for usefully judging its relevance, and presence of the statement “if this paper is correct, Eliezer may have won some money” contributed to me downvoting the post...
First of all, I appreciate the feedback.
Upon looking more on arXiv, I agree that there are lots more potential explanations then the one I happened to find out about from reading the news, so I agree the pick is arbitrary. The Wikipedia page on the anomaly, http://en.wikipedia.org/wiki/OPERA_neutrino_anomaly has multiple different explanations. I didn’t think to look harder for multiple explanations, which was incorrect. It appears that I need more training in my ability to gather information.
For construction purposes, I generally followed the template of the previous discussion, since it was rated positively overall: similar topic, An arXiv link, three news sources, quotes from paper, and brief statement from me. I checked after you posted and it appears that in the previous thread there was also some mention of downvoting for topic choice as well, although there were arguments against that as well, and both sides of the argument were upvoted. I feel this is best summed up by something you said earlier in a different thread:
Which is reasonable, and seems to apply to suitability as well as sanity.
So I agree with your first critique, and respect your second. However, there are multiple potential reasons that you might not have liked the statement “if this paper is correct, Eliezer may have won some money.” Can you explain why you added that to the list?
Eliezer’s bet shouldn’t be interesting/relevant in this context. Highlighting this point sets off my cultishness alarm.
Thank you, that makes it more clear.
Except, it would seem, yourself?
It’s true. Folks have been making a bit deal about betting on the subject. I don’t see what is wrong with saying it.
I think that “relevance” means, relevance to the problem of the supposedly faster-than-light particles. He didn’t offer a judgment on that.
Correct.
I would find it quite strange if the original team did not take all of the relativistic effects into account before publishing. If true, however, there are bound to be some embarrassed faces, given that the error is rather elementary in retrospect (as opposed to a hard-to-pinpoint unexpected systematic error in a state-of-the-art one-of-a-kind detector).
We will have to wait for the confirmation from the CERN team that they missed this correction, and if so, what the exact value of the correction is. What I find alarming is that van Elburg, a neuroscientist by training, apparently did not contact the CERN team with his calculations first, but instead rushed to publish them. He certainly does have some credentials in theoretical condensed matter physics, but does not appear to have worked in the area related to the problem in question.
Based on this, I would assign larger than even odds that this FTL refutation will be refuted first, but not large enough to take any bets on it.
From his apology:
Can someone explain why one would even think the experiment was set in the satellite reference frame? As far as I can tell they never used any clocks on satellites. They just synchronized two clocks—in the same reference frame—to an event outside that reference frame. The satellite doesn’t even need to keep time so long as its signals each have a unique identity and we know the location of the satellite when it sends the signal to set the clocks. You can alter the clocks to account for difference in signal travel time—from the satellite to each clock.
I’m glad you found that. He made the following comment that I found very important:
Since the Author is retracting the comments about GPS, I’ll update the original post accordingly.
Presumably you mean not enough to take any bets on it at even odds?
The uncertainty of close-to-even-odds tends to be so high, figuring out comfortable odds for a bet is quite hard (10:1? 100:1?). Thus, I only bet on the things I am extremely confident about, and even then only at even odds. Sorry if it is not overly Bayesian of me.
After replacing “may not have broken” with “almost certainly didn’t break”, that should be everyone’s reaction unless and until additional and more convincing confirmation arrives.
I defy the data.
This post’s title doesn’t seriously annoy me, though. It would take “Did neutrinos break the light speed limit?” to do that. (“Did neutrino’s break the light speed limit?” for triple bonus thermonuclear rage.)
So, I’m not a physicist but I’m pretty sure the clocks aren’t in orbit…
Word on the street (or at least the unofficial Lesswrong IRC channel) is that this dosen’t account for the difference.
Phil Plait thinks this explanation is unlikely to hold water..
The OPERA people have elsewhere said that they calibrated the times using portable atomic clocks. If that is the case then the concern here, coordination from the GPS, shouldn’t be an issue.
I’m also a bit confused by where this paper is getting his 32 ns number. It seems to only work if the satellite is traveling roughly above and parallel to the neutrino beam. Any other orbital path should by his logic have slightly different results. Overall, this seems unsatisfactory.
There should also be General Relativity effects from the Earth’s gravitational field (and its rotation). Whether they would have any impact on the measured time of the neutrino time-of-flight is beyond me, though...
The rotational effects (also known as frame dragging) are probably negligible, given that it took the Gravity Probe B 7 years of data analysis to confirm its detection using dedicated equipment.
The gravitational effect on time of flight, known as the Shapiro time delay is of the order of 2GM/c^3, where M is the mass of the Earth, or about 0.03ns, too small to matter.
Where is the minority now? The so called fringe contrarians?
Which they say, the experiment is good and that the theory (GR) was wrong, all along.
Their number is small, but I think their position is to be considered carefully now, too.
GR is a single point in theory-space. But the position “GR is wrong” is just the set of every other alternate theory, stated or unstated, imagined or yet unimagined. As a position it doesn’t really make predictions, unless it specifies in what particular way GR is wrong.
Could still be wrong, couldn’t it?
You didn’t grasp my point. What I said means that GR probably is wrong in an absolute sense—same way that Newtonian mechanics was wrong in an absolute sense.
But GR also seems to be Less Wrong than any other theory on the topic that we currently have. Which alternate theory do you have that you want us to “carefully consider”?
Not good enough. Is it correct or is not correct? And I mean “absolutely correct”.
I don’t know. Maybe a patch is needed in which case it is technically incorrect. If we have something better than the GR, or we have not—it is just besides the point.
“enough” for what?
What exactly do you believe “the point” is?
The point is, to have a correct theory. This is enough, and nothing else.
I don’t know where and how and if at all, the GR is valid or not. But we have a measurement which might contradict it. And that might not be resolvable without mending the theory.
Okay, but I think you’re saying the obvious. I’m not a physicist but the way I understand it, much stronger evidence for GR’s non-validity would be the fact that it’s inconsistent with quantum mechanics—and since it’s at large distances that we keep getting results we don’t expect and which we need to patch over with hypotheses about dark energy/dark matter/etc, it’s GR that is most likely to be revised first.
Quite a bad omen, yes. Something had to give. Either QM, either GR. Now, we MAY have a contradiction inside the GR and the QM is much less threatened by the GR as it seemingly was, back in 1926.
There were other bad signs for the Relativity before, name a few. Ehrenfest paradox and the already transluminal speed of the Universe expansion, were not helping.
The probability that the Gran Sasso measuring was wrong, is smaller and smaller.
I think you’re giving way too much significance to this latest measuring. Even if GR is shown to be false, my money’s still on the measuring to be false as well.
You are most probably right. But it is not a sure thing yet. With every day passed and with no likely explanation, I update my probability estimation for a major patch needed for GR—up.