Hyperlight Nutrinos Take 2

November 18, 2011 8:44 AM

Neutrino experiment repeat at Cern finds same result "The team which found that neutrinos may travel faster than light has carried out an improved version of their experiment - and confirmed the result."

"If confirmed by other experiments, the find could undermine one of the basic principles of modern physics.

Critics of the first report in September had said that the long bunches of neutrinos (tiny particles) used could introduce an error into the test.

The new work used much shorter bunches."

posted by marienbad (107 comments total)

Good! In fact, great!
posted by jfuller at 8:46 AM on November 18, 2011 [1 favorite]

Scientists Repeat Experiment, Confirm it to be Flawed.
posted by Mr.Encyclopedia at 8:47 AM on November 18, 2011 [11 favorites]

Has the relativistic motion of clocks on board GPS satellites issue been ruled out yet (this doesn't appear to address that concern)
posted by zeoslap at 8:47 AM on November 18, 2011 [5 favorites]

Awesome, but I'm confused. I thought the first result had been shown to be the result of a mistake in not accounting for motion of a GPS satellite clock.
posted by COBRA! at 8:49 AM on November 18, 2011

d'oh
posted by COBRA! at 8:49 AM on November 18, 2011

I would be overjoyed but for the fact in a long-running argument with my 12 year old son on "how can relativity be true?" I've taken the side of Einstein. I'm afraid to tell him about these results.
posted by DU at 8:50 AM on November 18, 2011 [1 favorite]

I thought the first result had been shown to be the result of a mistake in not accounting for motion of a GPS satellite clock.

That's one of the theories. We really won't know what's going on until someone does the experiment in a different location from scratch, instead of re-running this one.
posted by empath at 8:52 AM on November 18, 2011 [3 favorites]

I've taken the side of Einstein. I'm afraid to tell him about these results.

Relativity doesn't change very much with this experiment, even if neutrinos go faster than photons. There still might be a maximum speed -- it's just the speed of neutrinos now, instead of the speed of light.
posted by empath at 8:54 AM on November 18, 2011 [8 favorites]

We really won't know what's going on until someone does the experiment in a different location from scratch

Good lord that would be expensive, wouldn't it?
posted by odinsdream at 8:57 AM on November 18, 2011 [4 favorites]

The find could undermine one of the basic principles of modern physics.

Yeah, but since most of those physicists are science fiction geeks, it confirms what they hoped was true in their hearts.
posted by The 10th Regiment of Foot at 8:57 AM on November 18, 2011 [11 favorites]

The speed of light wasn't called a law of nature because it happened to be the fastest thing. It's called a law of nature because it's derived from Maxwell's Equations. If neutrinos are faster, Maxwell (or that derivation) have to be wrong and there has to be a new derivation before we can call the new c a law of nature, rather than just "the fastest observed speed".
posted by DU at 8:58 AM on November 18, 2011 [11 favorites]

The Guardian says that the GPS clock thingy possible error thingy hasn't been tested (fully?) yet
posted by Bwithh at 8:58 AM on November 18, 2011 [4 favorites]

Uh-oh.
posted by Mental Wimp at 8:59 AM on November 18, 2011

I would be overjoyed but for the fact in a long-running argument with my 12 year old son on "how can relativity be true?"

Tell him it's only relatively true.
posted by goethean at 9:02 AM on November 18, 2011 [7 favorites]

If neutrinos are faster, Maxwell (or that derivation) have to be wrong and there has to be a new derivation before we can call the new c a law of nature, rather than just "the fastest observed speed".

it's derived from the magnetic constant and electrical constant, which are observed from experiment.. Neutrinos might have a different interaction with the electromagnetic field... We don't know a whole lot about them.
posted by empath at 9:02 AM on November 18, 2011 [1 favorite]

oh snap.
posted by sexyrobot at 9:07 AM on November 18, 2011

I look forward to eventually finding out in what way this experiment was flawed or misinterpreted. It seems like they've ruled out all the obvious things, so it's likely to be an interesting mistake.
posted by rusty at 9:08 AM on November 18, 2011 [1 favorite]

If I were a betting man I would bet a lot that neutrinos don't go faster than light. I would also have bet a fair bit this rerun would have not solved the problem.
Also: empath - it's not just electromagnetism that involves c. It is much more than that even if the ε₀ and μ₀ can be used to calculate it.
posted by edd at 9:08 AM on November 18, 2011

Nutrinos ftw! This is way cool!
posted by amazingstill at 9:11 AM on November 18, 2011

The Bartender says "you can't go that fast in here!". A neutrino walks into a bar. (via)
posted by joinks at 9:17 AM on November 18, 2011 [76 favorites]

I LOVE it when one of the laws of the universe is broken! You go, lil' neutrinos!
posted by kinnakeet at 9:17 AM on November 18, 2011

This is interesting, but am I right that we can't really get too excited about this? I mean, if the leading explanation for their observations is a subtle systemic error, then confirmatory observations generated using the same system don't really give us any new information.

It's cool that they've eliminated one of the possible sources of error, but presumably with such a complex experiment they're not going to run out of those any time soon. Hopefully someone who knows more about the engineering of the experiment can correct me, but my impression is that the headline "Flaw in equipment not found yet" can basically be re-applied to all the new data generated by that lab until we see the results confirmed by a different group's equipment, and ideally by taking a very different approach to making the same measurement to guarantee that they're not suffering the same error.

Serious kudos to the lab for going balls-out with "here's some inexplicable data: let's go bug-hunting together!" though, rather than trying to discard, obfuscate or fudge it. Exactly how science is supposed to work. Except there should be more money, funky music and groupies.

Does anyone remember the invocation to summon physicsmatt? I know it starts with sealing a room, drawing a pentagram then placing a running fridge inside it with the door open... but then I think I blacked out before the ritual was complete.
posted by metaBugs at 9:18 AM on November 18, 2011 [9 favorites]

I always suspected that Einstein was just a dumbfuck. . .
posted by Danf at 9:19 AM on November 18, 2011 [1 favorite]

Also: empath - it's not just electromagnetism that involves c. It is much more than that even if the ε0 and μ0 can be used to calculate it.

Every way that it's been calculated has involved photons, AFAIK.
posted by empath at 9:21 AM on November 18, 2011

Hilarious to me that some editor at the Washington Post chose "Odds shrink Einstein was wrong" as the headline for this story, which is exactly wrong.
posted by ericost at 9:22 AM on November 18, 2011

Hoverboards: when?
posted by chavenet at 9:22 AM on November 18, 2011 [4 favorites]

Metabugs: exactly. That's why I would have bet on this result recurring while also betting that it is wrong.
posted by edd at 9:23 AM on November 18, 2011

I'm sure someone with more particle-specific knowledge will come in and clarify, but the main improvement here is that they no longer need to determine arrival times by fitting the distribution of production and arrival events, both of which occurred over time scales larger than they wanted to measure. Instead, they produce a very short (quite a bit less than 20 ns) burst and measure the arrival. Since they are finding an anomalous arrival time of around 60 ns, they have ruled out fitting error as a systematic source. Of course, any possible systematics relating to the distance measurement or the time syntonization remain.

(And meanwhile, the LHC is continuing its slow disproof of the standard model Higgs, but at least shows signs of novel weirdness in complicated flavor physics)
posted by Schismatic at 9:23 AM on November 18, 2011 [2 favorites]

metaBugs, the trick is to put some quality beer in the fridge.

I'm working on writing up some thoughts on this for you guys, but I haven't had a chance to really read the new paper carefully, so it might take awhile.

Meanwhile, here's some light music...
posted by physicsmatt at 9:24 AM on November 18, 2011 [11 favorites]

Empath: well photons are convenient experimentally. But c is in General Relativity for example without electromagnetism getting involved, and speed of gravity constraints are close to c.
posted by edd at 9:25 AM on November 18, 2011

"Syntonization" ought be "synchronization," of course.
posted by Schismatic at 9:25 AM on November 18, 2011

We really won't know what's going on until someone does the experiment in a different location from scratch

Good lord that would be expensive, wouldn't it?

Well, it is the basics of so-called reality we're talking about here. I'll donate $20. Just tell me where to send it.
posted by philip-random at 9:25 AM on November 18, 2011 [2 favorites]

I think that all this may mean for relativity (if it's true-- a big if) is that we may be able to send a super-luminal signal with neutrinos, but it doesn't mean that faster than light travel is possible, it just means that we'll be able to get information about something tiny fractions of a second before it happens (in a relativistic sense). I don't think it's likely to be practical, ever, but you never know.

Though maybe it would have weird implications for neutrino-based computation if such a thing were possible?
posted by empath at 9:25 AM on November 18, 2011 [1 favorite]

Hilarious to me that some editor at the Washington Post chose "Odds shrink Einstein was wrong" as the headline for this story, which is exactly wrong.

It's not only wrong, it's one of the worst sentences written in the English language I've ever read.
posted by griphus at 9:26 AM on November 18, 2011 [9 favorites]

there should be more money, funky music and groupies.

Oh gawds, yes, metaBugs. And not just for the particle physicists!
Oceanographers need love, too.
posted by zomg at 9:26 AM on November 18, 2011

I would be overjoyed but for the fact in a long-running argument with my 12 year old son on "how can relativity be true?"

Just tell him to shut up and stop arguing with people that are smarter than him. Yeah, it'll hurt his feelings but by the time he's in his 20s, it will have morphed into some pivotal life-lesson.
posted by philip-random at 9:27 AM on November 18, 2011

Empath: well photons are convenient experimentally. But c is in General Relativity for example without electromagnetism getting involved, and speed of gravity constraints are close to c.

Right, but these are all experimental measurements. There's definitely new science involved if the new 'c' is the speed of neutrinos, but you don't need to throw out relativity. It just means that there is something that's slowing down everything else slightly that doesn't effect neutrinos, yeah?
posted by empath at 9:27 AM on November 18, 2011

Serious kudos to the lab for going balls-out with "here's some inexplicable data: let's go bug-hunting together!"

It's really good that they're taking the data seriously. On the other hand, the science by press release approach to unpublished work is potentially damaging to scientific credibility in a world with so many people hostile to science. We could easily have one without the other.
posted by howfar at 9:29 AM on November 18, 2011

Philip-random: write to your politician and tell them you'd put in that twenty bucks. Seriously. Fundamental physics research is not in the best situation in a number of countries and politicians knowing that some members of the public support it is a good thing.
(honest declaration: physicsmatt is not the only one here who gets paid by that kind of funding)
posted by edd at 9:29 AM on November 18, 2011 [1 favorite]

it's likely to be an interesting mistake.

I love Science! :)
posted by Celsius1414 at 9:30 AM on November 18, 2011 [1 favorite]

Guys, this is just a Higg's boson that had traveled back in time fucking with us. First it tried to destroy the LHC now it is fucking up our experiments. We must remain vigilant, we have to get them before they get us.

I suggest once we perfect time travel, we travel back in time and kill Higgs to prevent him ever creating the vile things.
posted by Ad hominem at 9:32 AM on November 18, 2011

Can I just hitch a ride on some neutrinos into the future where this has all been not only resolved (as much as something can be said to be "resolved" in science), but translated into a dialect of English that I could understand?
posted by Terminal Verbosity at 9:37 AM on November 18, 2011

I think that all this may mean for relativity (if it's true-- a big if) is that we may be able to send a super-luminal signal with neutrinos, but it doesn't mean that faster than light travel is possible, it just means that we'll be able to get information about something tiny fractions of a second before it happens (in a relativistic sense). I don't think it's likely to be practical, ever, but you never know.

Though maybe it would have weird implications for neutrino-based computation if such a thing were possible?

1) Set up bi-directional version of the experiment
2) Set up automated message relaying back and forth through several (billion) iterations until you're receiving messages several seconds prior to their transmission.
3) Send yourself a stock quote and place an order large enough to change the quoted data
4) Profit! (or reality-ending paradox?)
posted by Ryvar at 9:38 AM on November 18, 2011

Serious kudos to the lab for going balls-out with "here's some inexplicable data: let's go bug-hunting together!" though, rather than trying to discard, obfuscate or fudge it.

Well said.
posted by nebulawindphone at 9:42 AM on November 18, 2011

I am not a physicist (_{and neither are most of you}) but I think it's wonderful to live in a time where we've discovered hundreds of planets in other solar systems and where we may have possibly discovered the first chink in the light speed limitation.. When I was in elementary school extra-solar planets were only a hypothesis at the very best and it was thought that no one could survive a trip to the moon because the trip through the Van Allen belts would be deadly (seriously, that's what was being taught).

A thought that comes to mind is that removing the light speed limitation for some class(es) of particles would help explain the Fermi Paradox in that the reason that we are not receiving transmissions from other civilizations would then be that they are using some non-electromagnetic based FTL means (energetic neutrinos??) to communicate.
posted by Poet_Lariat at 9:49 AM on November 18, 2011

The little engine neutrino that could.
posted by Hairy Lobster at 9:53 AM on November 18, 2011

Are the neutrinos traveling through rock faster than light would travel through a vacuum?
posted by pracowity at 9:53 AM on November 18, 2011

empath - I think that all this may mean for relativity (if it's true-- a big if) is that we may be able to send a super-luminal signal with neutrinos

But isn't that a Really Big Deal? My very limited understanding is that c isn't just the max speed of light, but also the maximum speed of information in the universe. And that all sorts of ideas in physics fall over if it becomes possible for information to travel too fast. This is part of why you can't do anything particularly exciting with quantumn teleportation or quantumn tunelling: it can't actually be used to send information faster than the speed of light.

Mind you, my physics education largely consists of going drinking with high energy physicists (an ironically languid demographic), so I hear and briefly understand lots of fancy ideas, but always filtered through several pints and interspersed with dirty science jokes. Probably the most enjoyable way to talk about it, but perhaps not the most pedagogically sound.

nebulawindphone - Well said.
Thanks. I feel strongly that it's not proper science communication without crude and unnecessary references to genitalia.
posted by metaBugs at 9:57 AM on November 18, 2011 [2 favorites]

I'm working on writing up some thoughts on this for you guys, but I haven't had a chance to really read the new paper carefully, so it might take awhile.

Get that man an Irish coffee. Stat.
posted by clarknova at 9:58 AM on November 18, 2011

This is like in Jaws when Roy's character says, we're going to need a bigger boat right, except what we need is a bigger LHC. Time to dust off the old blueprints and get the SSC built afterall. This time put it in a blue state.
posted by humanfont at 10:00 AM on November 18, 2011 [1 favorite]

I propose the Dark City Hypothesis: At some point after the neutrinos are sent, the aliens put us to sleep while they do their regular maintenance. When we wake up, we notice that the neutrinos have already arrived and we weren't expecting them for another 6 nanoseconds.
posted by -harlequin- at 10:02 AM on November 18, 2011

Are the neutrinos traveling through rock faster than light would travel through a vacuum?

Yes, although I don't kow what your point is :( . BTW do you know that there is really no such thing as a vacuum? Random quantum pair production of so-called "empty space" means that photons still have to occasionally interact with virtual particles in the brief moments that they exist - thus slowing light speed down from some theoretical ultimate speed. I am not a physicist but I would hypothesize that photons traveling great distances (millions/billions of light years) would be more noticeably slowed. Kind of like a "friction" in the fabric of space-time. So whatever the true limitation of light speed is , it's probably something other than the figure that we are quoted (although the difference is likely insignificant for short distances)
posted by Poet_Lariat at 10:02 AM on November 18, 2011

...and we're back after that commercial break.

Still haven't had time to fully digest the update, but I'll give my initial impression. There appears to be a lot of work done in the update on nailing down timing errors, but it’s nasty, complicated, and I’m definitely not the one who will find any mistakes there that the experimenters overlooked. So, while that’s very important, I’ll skip that and go to what was, to me, the most interesting addition.

As you may recall from the end of the last thread on this, I had left with the opinion that the most likely error in the measurement was that the experiment had made the assumption that the shape of the neutrino pulse (the number of neutrinos sent as a function of time) was identical the shape of the proton pulse that created the neutrinos. Let me explain why that might matter, and then what this update means for that explanation.

Basically, you create a neutrino beam by slamming relativistic protons into a stationary target, creating unstable lighter particles called mesons that eventually decay into neutrinos. It is a completely reasonable assumption that the number of neutrinos present at a particular moment in the neutrino beam should be directly proportional to the number of protons that hit at the corresponding moment during the beam creation, but this has never been conclusively proven. Protons are easy to measure, neutrinos are hard, and since there was no near detector (a neutrino detector present at the end of facility that creates the beam), there is no easy way to test that the two shapes are identical, so OPERA was forced to make this assumption.

In the original paper, they only saw one or so neutrino per pulse, and it could come from anywhere in the 10 microsecond-long pulse, which is not enough to say, event by event, that neutrinos were traveling faster than light. However, they repeated the pulses 15,000 times or so, and built up a shape for the number of neutrinos as a function of time (this assumes that every pulse is identical, which apparently is not too crazy, as the accelerator guys are good at their job).

Since the pulse length is so long compared to the time-lag they are measuring (60 ns versus 10 mu-s), and since you can't tell when each individual neutrino was created inside the pulse originally, you can't do a one-to-one comparison and say "this neutrino traveled faster than light." What you can do is fit the built-up shape of the aggregate neutrino pulses and see where the beginning and end points are in time. This is what they did, and the best fit was the 60 ns early arrival of the previous result,

However, coming back to the issue of pulse shapes: if the neutrino beam had a different shape than the proton beam (that is, protons early in the pulse produced more or fewer neutrinos on average than late protons), the fitting procedure could bias you towards early times (or later, for that matter, depending on how the two pulses differ).

Think of the proton shape as a function of time as a picture of a nearly flat mesa: sharp edges (the beam start and stop) and a whole bunch of peaks along the top. The neutrino shape is pretty much that same, but maybe it's slightly tilted in comparison, and it's a whole lot blurrier (since it was determined with fewer events). What you need to do is slide the two pictures back and forth (in this case, that corresponds to figuring out the time of arrival for the neutrino pulse) until they line up as best you can make them. If the two pictures were identical, you'd of course get the best match when the edges of the two were perfectly aligned, and you can use that to determine the arrival time of the neutrinos exactly. But if one is not the same as the other, you might find that the best match does not correspond to the "correct" start time.

This was my operating assumption on the source of the OPERA result, and since the explanations that were floated about the issues with timing coming from the GPS satellites turned out to not understand how the experiment was using that information to maintain synchronization, it seemed like the best game in town. Now, to figure out if that was the source of the error, OPERA got CERN to give them 3 ns pulses of protons. This limits their data-taking, since there aren’t as many neutrinos in their detector in Gran Sasso, after something like 1000 pulses (my estimate), they ended up with 20 neutrino events. However, these events are much cleaner, since the time lag they are looking for is 60 ns and the pulse lasted only 1/20th of that. They say their new measurement is consistent with a 61 ns delay, which matches their previous result. I still have some questions, but if this holds up, it eliminates my favored explanation for the measurement. And I was so sure I was right.

The remaining errors that seem reasonable tend to be much more prosaic, having to do with unaccounted-for lags in the timing. I really hope it doesn’t turn out to be that some poor undergrad accidentally plugged an 18 meter long coax cable in to their timing circuit and forgot about it. (Note: this can’t be true, they’ve done enough cross checks to have seen it). So, seems to be a real puzzle.

However, the arguments against FTL neutrinos are still strong. In particular, the burst of neutrinos from supernova 1987A cannot have travelled faster than light at the speeds measured by OPERA. Those neutrinos were at lower energy than the ones in this experiment though, but you’d need an incredibly strong energy dependence to get the two measurements to agree and that seems really ugly to me.

I'm sure this will be discussed over the next few days here at the lab (where, sadly, alcohol is banned, so no Irish coffee for me). I will try to drop in once I have a better feeling from my colleagues concerning issues I may have overlooked in this initial response to the new paper. Also, at some point I can try to talk about why this would be a bit of a problem for relativity.
posted by physicsmatt at 10:03 AM on November 18, 2011 [78 favorites]

Has the relativistic motion of clocks on board GPS satellites issue been ruled out yet (this doesn't appear to address that concern)

1) GPS already does this, because otherwise, GPS would not work. The entire system is based on know what time it is at the GPS satellites, and knowing what time it is where you are, so you can compare propagation time. Step one of the process is for the local receiver to synthesize local time from four GPS signals.

2) They do not repeat not use GPS sats as the primary clock. They have their own atomic clocks at the sites. These clocks are synced both with each other and various international standard clocks using GPS, in a well known and understood process that's also currently used to help define TAI, which is the basis of UTC, and they're close enough that they will frequently have the ability to sync time off the same set of four satellites.

If it turns out that time syncing via GPS is the cause of this, then things remain *very* interesting, because then it is a fundamental issue with TAI. There are two systems used for the sync, one is based on common GPS sat views, the other is based on TWSTFT (Two Way Sat Time & Frequency Transfer.) Given that TWSTFT is showing results consistent with GPS (though with more precision,) I seriously doubt that this is an issue, and I think there will be more chance of error attempting to change/confirm the time sync -- forex, by a direct fiber connection -- than by simply checking the sync against the associated components of TAI. That would be more than enough to validate the clocks.

If there is an error, it would be because of local gravitational effects at the two sites, which may not be well understood, not at the GPS constellation, which is *very* well understood -- they are some of the most accurately observed devices, if not the most, that humanity has built.

Also, do note that MINOS at Fermilab saw a similar signal, but at a much lower confidence level (~2σ) which they assumed was noise or error and ignored until OPERA posted their initial results. Now both Fermi and T2K in Japan are going to explicitly search for this, however, MINOS is finishing up a detector upgrade for this search and won't have results for a while.

The MINOS collaboration is analyzing some older data to see if they can refute. Because of the lower precision of the original detectors, they won't be able to confirm, but if they show that the signal is not supraliminal, even with the most favorable error, then it would stand as a solid refutation of the result.

it's likely to be an interesting mistake.

Oh, yes -- and it could be a mistake that leads to new science. See the original neutrino experiment that detected only about a third of the expected solar neutrino flux. There was some very interesting hand wringing about that, including "Hmm, maybe we're wrong about how stars work, but that model fit everything else so well" and "Hmm, maybe the Sun is broken?"

End result was that neutrinos oscillate -- then can change between electron, muon and tau neutrinos. The correct number of neutrinos was in the solar flux, but the detector couldn't see the muon and tau neutrinos.

End result? Neutrinos aren't massless, as originally thought.
posted by eriko at 10:03 AM on November 18, 2011 [19 favorites]

Yaayyy for physicsmatt!
posted by Poet_Lariat at 10:05 AM on November 18, 2011

"For most physicists outside the collaboration, however, the key test will be replication by an independent experiment. The one best placed to independently confirm or refute OPERA’s result is MINOS (the Main Injector Neutrino Oscillation Search) at Fermilab in Batavia, Illinois. In response to the latest OPERA result, MINOS issued a statement saying it is upgrading its timing system to match OPERA's precision and might have preliminary results obtained using the existing system that are relevant to assessing OPERA's results as soon as early 2012."
posted by gwint at 10:08 AM on November 18, 2011 [1 favorite]

My very limited understanding is that c isn't just the max speed of light, but also the maximum speed of information in the universe.

The important thing about relativity is that there IS a maximum speed, not what that maximum speed is. If it's different for neutrinos than photons, that's amazing news, but it doesn't mean you toss out relativity.
posted by empath at 10:12 AM on November 18, 2011

While others naysay, I am investing heavily in dilithium.
posted by itstheclamsname at 10:14 AM on November 18, 2011 [1 favorite]

CERN press release on the OPERA results.

The key callout here is that the collaboration has submitted a paper to JHEP, the Journal of High Enerty Physics, for peer review.

Tommaso Dorigo also notes that:

Regardless of all the considerations I made above on systematic uncertainties that might still affect the measurement, I must say I am rather positively impressed by at least a couple of things: first, that Opera has worked like a single man in the attempt of making more solid an already quite scrupulous result; and second, that the OPERA researchers seem now to stand behind the measurement much more united than they were two months ago.

In other words, those in OPERA who did not sign the first preprint - because they probably did not have enough time to scrutinize all the aspects of the measurement - are now apparently all willing to sign the new one, and probably ready to submit the resulting publication to a scientific magazine. Since I know several members of OPERA and I judge them all serious and scrupulous physicists, their willingness to sign the paper means that we need to take it more seriously than (at least a few among us) have so far.

So, OPERA is doubling down on this result. They're submitting for peer review and the collaboration as a whole is standing behind their analysis.
posted by eriko at 10:23 AM on November 18, 2011 [3 favorites]

Dang, I should have taken xkcd's bet!
posted by diogenes at 10:25 AM on November 18, 2011

Meanwhile,Einstein's brain goes on display at Philadelphia's Mutter Museum.
posted by longsleeves at 10:40 AM on November 18, 2011

See also: The Brain of Einstein.
posted by rusty at 10:51 AM on November 18, 2011

It's not only wrong, it's one of the worst sentences written in the English language I've ever read.

No, no, it turns out that a Mr Odds Shrink Einstein had claimed that OPERA wouldn't try to publish their results. Now it seems that Odds Shrink Einstein was, indeed, wrong.
posted by yoink at 10:58 AM on November 18, 2011 [1 favorite]

I think the most important takeaway from physicsmatt's comment is this:

However, coming back to the issue of pulse shapes: if the neutrino beam had a different shape than the proton beam (that is, protons early in the pulse produced more or fewer neutrinos on average than late protons), the fitting procedure could bias you towards early times (or later, for that matter, depending on how the two pulses differ).

And I think this is by far the most likely explanation, rather than "relativity is broken."
posted by chimaera at 11:00 AM on November 18, 2011

Thanks. I feel strongly that it's not proper science communication without crude and unnecessary references to genitalia.

Ah, yes. The McCawley Principle.
posted by nebulawindphone at 11:01 AM on November 18, 2011 [3 favorites]

chimaera, maybe I wasn't clear. That was a completely reasonable explanation for the first OPERA test (with the long proton pulse). It does not appear at first glance to be a possible explanation for the new result, as the new result finds a travel time difference of 60 ns while the beam spill was only 3 ns wide. This by itself is well worth the price of admission for the paper: there was a totally reasonable explanation for their signal in terms of standard physics, they tested it, and that explanation no longer seems to work. My gut reaction is still that this overall result is incorrect, but now we have to scratch off #1 on the list of "most likely to be wrong" and start looking at the rest of the options (and of course, since it's a largish collaboration, they've been multitasking, so they've also apparently addressed issues on the timing circuits at the same time).

Of course, maybe there are more errors with the new experiment they ran, so this conclusion may not be iron-clad, but I'm just giving you my day-zero impression of the paper.
posted by physicsmatt at 11:24 AM on November 18, 2011 [3 favorites]

I think reality is starting to break down, what with pizza becoming a vegetable and all. Expect other laws and constants to get weird soon.
posted by mccarty.tim at 11:27 AM on November 18, 2011

For reference, here's the previous discussion on metafilter: http://www.metafilter.com/107691/Einstein-Wept
posted by physicsmatt at 11:29 AM on November 18, 2011 [1 favorite]

Can we fast forward to the part where we discover all times are essentially the same time and all things are essentially the same thing so speed is irrelevant?
posted by 3FLryan at 11:30 AM on November 18, 2011

3FLryan: "Can we fast forward to the part where we discover all times are essentially the same time and all things are essentially the same thing so speed is irrelevant"

That would certainly help with this work project I'm behind on.
posted by Joakim Ziegler at 11:32 AM on November 18, 2011

That would certainly help with this work project I'm behind on.

Hey, same. But sorry, I lost the remote. Check Einstein's couch?
posted by 3FLryan at 11:37 AM on November 18, 2011

If this result holds, I'd say it likely has implications for the Fermi Paradox as well.
posted by newdaddy at 11:40 AM on November 18, 2011 [1 favorite]

This will finally bring us a step closer to answering that eternal question: If you are in a car traveling at the speed of light and you turn on your headlights, what happens?

Neutrinos! (or something.)

Lame jokes aside, I think this is amazingly cool. I realize that it isn't close to being confirmed, but I was expecting this to be one of those things that was shot down almost immediately upon peer retesting. I love science, even when I don't fully understand it.
posted by quin at 11:48 AM on November 18, 2011 [1 favorite]

So here's my question (IANAP) can you entangle a neutrino? Is it possible to entangle a FTL neutrino with a slower-than-light particle? (OK that was two questions.)
posted by newdaddy at 11:59 AM on November 18, 2011

newdaddy: yes, in principle you can entangle neutrinos with other particles (entangling just means the unmeasured quantum state of one particle is correlated with the unmeasured quantum state of another particle, so by measuring the quantum state of particle A, I've collapsed the wavefunction of particle B). No reason you couldn't do it with neutrinos, though I can't think of obvious examples off the top of my head. That's likely only because neutrinos are a pain in the ass to work with, so if you care about entanglement you'd spend your time with a particle that doesn't require a kiloton of detector mass to see reliably.

As for entangling a superluminal state with a subluminal one, well, I'd see no reason you couldn't, but all the theories I'm relying on to make these statements don't deal well with FTL. Tachyons are not really a sensible idea in quantum field theories, so any answer would be an extrapolation without a good solid foundation.
posted by physicsmatt at 12:08 PM on November 18, 2011

Can we fast forward to the part where we discover all times are essentially the same time and all things are essentially the same thing so speed is irrelevant?

Not irrelevant, merely illusory. But still interesting.
posted by rusty at 12:20 PM on November 18, 2011 [1 favorite]

physicsmatt, I always enjoy reading your comments. Always articulate, they are guaranteed to let me come away better informed. Regarding this comment:

if the neutrino beam had a different shape than the proton beam (that is, protons early in the pulse produced more or fewer neutrinos on average than late protons), the fitting procedure could bias you towards early times (or later, for that matter, depending on how the two pulses differ).

couldn't you assume that the shape of the pulse of neutrino arrivals would be a smeared out version of the originating shape, with a skew toward longer travel times (I'm assuming that conditions could slow them down, but not speed them past their maximum velocity)? Could this help calibrate the calculations?
posted by Mental Wimp at 1:16 PM on November 18, 2011

physicsmatt: thanks for the clarification, i thought the 10ms pulse was in the second experiment as well. With a 3ns pulse of protons and *still* a 60ns-early detection of neutrinos, that definitely wipes out that particular possible source of error.
posted by chimaera at 1:32 PM on November 18, 2011

Mental Wimp, it's likely that doing that would introduce too many extra degrees of freedom into the fit to get sufficient accuracy for the time of flight. That's a fancy way of saying that, if they didn't assume anything about the shape of the neutrino pulse, there's no way to determine the time it took to get to Gran Sasso within the 2 parts in 10^5 that's needed to see the reported effect.

Remember, the measured time lag was 60 ns (that's 6 x 10^-8 s), and the two shapes they are fitting are 10 microseconds (10^-5 s). So overall, that's measuring the start time of the pulse to one part in 167, if you're free to futz with the pulse shape any way you want (or, even within reasonable assumptions), you can get that fit to be arbitrarily good (or bad).

With that in mind, OPERA did exactly the right thing. They can't figure out which neutrinos in that long pulse are coming from where, so they changed the experiment to make the desired measurement possible.

Also, the original assumption was not that neutrinos later in the pulse were moving at different speeds, but that the later protons were creating more or fewer neutrinos than the earlier protons. This would be somewhat odd, but not impossible, as the target is being hit with a shit-load of energy in the pulse, so it's not too crazy to think that the target's material properties were changing over the pulse duration which could plausibly affect the neutrino production. This could still be happening (though it's a bit of a stretch), and MiniBoone should be able to see it with a near and far detector, but it turns out that it's totally irrelevant to the OPERA result.

Also, Metafilter needs a blackboard. You can't explain physics right unless you're drawing a cartoon version of reality on a blackboard.
posted by physicsmatt at 1:41 PM on November 18, 2011 [6 favorites]

> I think reality is starting to break down, what with pizza becoming a vegetable and all.

If there weren't all this unreasoning prejudice against anchovies, pizza could have been a fish.
posted by jfuller at 1:57 PM on November 18, 2011

I am not a physicist but I would hypothesize that photons traveling great distances (millions/billions of light years) would be more noticeably slowed. Kind of like a "friction" in the fabric of space-time. So whatever the true limitation of light speed is , it's probably something other than the figure that we are quoted (although the difference is likely insignificant for short distances)

You're saying that empty space has a non-zero index of refraction. First of all it wouldn't make any difference how long it was, the index of refraction is a linear slowdown, it doesn't decelerate the photons. Once they get out of the medium that was slowing them down, they speed back up.

Anyway, we know from observations of supernovas that at least some neutrons travel at or very close to the speed of light. We see the Neutrinos a couple of days before the light from a supernova, but if the speed difference was this great, they would have been observed much, much earlier.

Maybe it's like: before they turn into neutrinos they spend time as some other kind of particle that moves faster then light, then decays into neutrinos or something weird.

couldn't you assume that the shape of the pulse of neutrino arrivals would be a smeared out version of the originating shape, with a skew toward longer travel times (I'm assuming that conditions could slow them down, but not speed them past their maximum velocity)? Could this help calibrate the calculations?

The new experiment makes that irrelevant.

Also, Metafilter needs a blackboard. You can't explain physics right unless you're drawing a cartoon version of reality on a blackboard.

You could always draw diagrams and link to them. We used to be allowed the img tag but as you can imagine that can lead to a lot of problems.
posted by delmoi at 2:08 PM on November 18, 2011

Assuming this discovery is true, and assuming you could code a message in neutrinos, how could you achieve a time traveling algorithm? I'm imagining a simple search of a hash space, or a search for prime numbers traveling backwards in time with each iteration (neutrino beam), achieving a solution in one perceived step. But I can't visualize how the time-traveling scheme would work. Would the reply arrive before the challenge is sent?
posted by kuatto at 2:25 PM on November 18, 2011

I wonder if they'll find the same results if they don't directly observe the experiment.
posted by pez_LPhiE at 2:27 PM on November 18, 2011

Would the reply arrive before the challenge is sent?

All it would mean is that you receive the signal before you see the signal being sent. So imagine catching a ball before you see someone throw it.
posted by empath at 2:52 PM on November 18, 2011

For people trying to make a time-machine out of this, or a computer that can solve NP problems in P time, I think you'll need either the neutrinos themselves to perform some calculations (i.e., at least part of your computer needs to be FTL), or at least some components will need to be moving at relativistic (but STL) speeds relative to each other. I've sketched some spacetime-diagrams trying to get a working example, but nothing resembling a proof. I'm sure someone else on the internet has thought more about this already.
posted by physicsmatt at 2:58 PM on November 18, 2011 [2 favorites]

As a biologist, the idea that you can just put stuff in Arxiv without any sort of review blows my little mind. I mean, obviously there are things wrong with current methods of peer review and publishing, but I could just write stuff and put it there? Couldn't I completely make stuff up? How do you choose what to read?
posted by maryr at 3:18 PM on November 18, 2011 [1 favorite]

I'm sure someone else on the internet has thought more about this already.

Indeed.
posted by empath at 3:24 PM on November 18, 2011

As a biologist, the idea that you can just put stuff in Arxiv without any sort of review blows my little mind. I mean, obviously there are things wrong with current methods of peer review and publishing, but I could just write stuff and put it there?

From what I understand, it doesn't count as publishing. It's pre-publishing. It's peer review on a massive scale.
posted by empath at 3:25 PM on November 18, 2011 [1 favorite]

Time to get it
posted by kenaldo at 4:12 PM on November 18, 2011

http://cgi.ebay.com/ebaymotors/DeLorean-/200675776659?pt=US_Cars_Trucks&hash=item2eb9355893
posted by kenaldo at 4:14 PM on November 18, 2011

delmoi: You're saying that empty space has a non-zero index of refraction.

Actually, the OPERA neutrinos aren't traveling through empty space, they're traveling through the Earth. We just assume that, being neutrinos, they don't notice the Earth that much, but maybe matter has a negative index of refraction for neutrinos?
posted by localroger at 4:30 PM on November 18, 2011

maryr, arxiv is a preprint server, it grew out an informal system of sharing papers between the theorists of top universities that existed through the 90's. This is a huge improvement, obviously, since it's automatic, fast, and more democratic. That said, as of a 5 or so years ago, if you don't come from approved emails, you need to be invited in by someone already in the system, to cut down on the crackpots.

As empath said, it's not the same as publication, so we still need to go through peer review later. When I write a paper, I tend to post it immediately to arxiv, wait a few days to hear from people about possible errors or missed citations, rewrite and submit to a journal. However, that's just how particle theorists do it; astrophysicists, for example, tend to submit to journals first if I understand correctly. Giant experiments are a different story, since they have more stringent internal reviews than any possible peer review process. Theory is sort of unusual, since we write papers relatively quickly, and use the quick response of arXiv to facilitate rapid movement in the field (at least, that's how I see it).

As for what to read, it's not a huge community, so you can go a lot on name recognition, paper titles, and abstracts to find the good stuff. Every morning, for example, I check hep-ph (30-40 papers usually), hep-ex (20-30), and most of the astro-ph's (100+ papers). Many of these papers are just not interesting to my research interests, so that narrows it down quickly. I don't read hep-th on a regular basis, for example, since I don't do much in pure theory or string theory.

localroger: http://arXiv.org/abs/arXiv:1110.2170
posted by physicsmatt at 4:38 PM on November 18, 2011 [2 favorites]

Ah well, thanks physicsmatt. I figured it would be on the menu but not that it could be so readily checked.
posted by localroger at 4:53 PM on November 18, 2011

Oh, and there were some studies done about paper impact measured by citation count versus location on the arxiv update list. Papers are recorded as they come in every day, up to the cutoff time of 4pm EST. So papers which came in at 4:01 appear at the top of the next day's post. Turns out those papers are cited more than the ones at the bottom.

Though you can argue that it's self-selecting, that people with good papers care enough to submit right after the deadline. Proposals to randomize the list order have not gotten off the ground as of yet.
posted by physicsmatt at 4:53 PM on November 18, 2011

Localroger, that paper was handy, so I figured it post it here. It's a fun idea, but I'm personally not so sure that an index of refraction would be the right answer for this (but then again, I still think OPERA is wrong, so what do I know?). It's a good idea regardless, and I just wanted to let you know that you're thinking along the same lines as the professionals.
posted by physicsmatt at 4:57 PM on November 18, 2011

Am I missing something? If neutrons travel only a little faster than light (Delmoi), not as much as this experiment shows, then there must be an error here? Stellar distances are much more effective at amplifying time differentials?
posted by BadMiker at 5:04 PM on November 18, 2011

Oh, and up a few comments, I said MiniBoone could check this, I meant MINOS.
posted by physicsmatt at 5:05 PM on November 18, 2011

physicsmatt: "metaBugs, the trick is to put some quality beer in the fridge.

I'm working on writing up some thoughts on this for you guys, but I haven't had a chance to really read the new paper carefully, so it might take awhile.

Meanwhile, here's some light music..."

I'm sorry but if Kazumoto Endo is considered light, man... I'd hate to see heavy. Must be like... Higgs heavy.
posted by symbioid at 7:28 PM on November 18, 2011

Ah, so they still haven't found the source of the error. Keep at it, dudes.
posted by Decani at 11:36 PM on November 18, 2011

The Arxiv also works well because it is dominated by theory work that, with the right background, can be reproduced largely by following along with a pen and paper. It isn't like biology where one has to worry about controlling for hard-to-detect interactions between genes or reagents or the like. As one of my string theory friends put it, if he cared about a paper, he would redo its calculations anyway, so the whole business of journals was secondary to the Arxiv, filtering more for impact than correctness. Moreover, having too many errors in what you upload is clearly bad for one's reputation.

The situation is a bit different in other parts of the site, say q-bio. Quantitative biology is still mostly computer models and theory, but not all of the models are actually remotely biologically possible nor the claims actually plausible. (Of course, peer review of biology by physicists doesn't always do that either.) However, I don't think that the researchers those fields use the Arxiv as their prime communication method like the high energy folks.
posted by Schismatic at 1:42 AM on November 19, 2011

Somewhere on that list of authors is the graduate student who spent a year writing the algorithm to match the proton and neutrino pulse shapes, which was apparently done exactly right but no longer needed.
posted by fantabulous timewaster at 9:36 AM on November 19, 2011

But I can't visualize how the time-traveling scheme would work. Would the reply arrive before the challenge is sent?

The reply would have to include the challenge, or an answer packaged with the question, or else is structured in some way that the reply makes the terms of the challenge obvious. How else to know a message from the future is a reply to some query, otherwise?
posted by Blazecock Pileon at 12:34 PM on November 19, 2011

I'd find it interesting to read layman's version of the step by step procedures needed to run such a complex experiment as this -- all the equipment, the synchro, in two countries, and then someone does what? Clicks a red button with a mouse? A talented science writer could create a compelling story.

While on the topic, that writer could easily be physicsmatt, who has quite the talent for clear explanation of complex topics.
posted by Rumple at 3:19 PM on November 19, 2011

Hontas Farmer explains why faster than light neutrinos would not lead to faster than light travel.
posted by Kattullus at 5:01 PM on November 19, 2011 [1 favorite]

Rumple, I really don't know the ins and outs of how big experiments work, since I'm a theorist and thus never had to go through the experience of a grad student in an experiment. Also, experimentalists keep us at arms length for the most part (so that we can't get hints of new results before they've been Blessed - which is the official term for the members of an experiment all agreeing to release a result).

For an somewhat old but very well written book about the structure and personalities of a giant experiment, find a copy of Nobel Dreams by Gary Taubes. It's out of print, but you should be able to find it. Its about the discovery of the W and Z bosons in UA1 at CERN, and the subsequent "discovery" of supersymmetry. It also has hilarious in hindsight passages about the amazing new data entry device in use at CERN: the computer mouse.
posted by physicsmatt at 1:33 PM on November 20, 2011

The ICARUS collaboration, whose detectors are also at Gran Sasso, has posted an article to ArXiv that claim that OPERA's interpretation just doesn't add up:

Our results therefore refute a superluminal interpretation of the OPERA result according to the Cohen and Glashow prediction for a weak currents analog to Cherenkov radiation.

Sounds like there will be some awkward moments in the Gran Sasso cafe.
posted by grouse at 11:13 PM on November 20, 2011

Meanwhile, someone on an obscure corner of reddit explains the experiment in the style of Chaucer.

[...]
My dearest love, your aid I seek, to hide
What we have done; take now your husbands steed
And spur it on with such a speed surmounting
Until the beast it's own dark shadow crowning
Can block no longer sunbeams now outraced.
The past, all shadow, can thereby be erased.

The King this letter read, as Lancelot knew,
And he so loved his bride that would undo
By any means her youthful indiscretion
So he bestrode again his prized possession
And forth he rode- returning nevermore.
For he could not have grasped what heretofore
Had Merlin guessed, the edict of Einstein:
However fast he sped along a line
A sunbeam racing would beat him by a pacing
Of seeming same on every time unfailing.
But proceeding on this errand, trifling hours
to spend, would burn a thousand years and more.

By this proverb shall you understand;
And I should to my tale's major point descend. [...]
posted by Rumple at 4:59 PM on December 10, 2011 [1 favorite]