A Bet Concerning Neutrinos (Part 3)

As you’ve probably heard, an experiment called OPERA measured how fast neutrinos go from a particle accelerator in Switzerland to a detector in Italy. They got a speed slightly faster than light. This got a lot of people excited.

As a conservative old fart, I made a bet with Frederik De Roo saying that no, neutrinos do not go faster than light.

Since then, various reports have been zipping across the internet at near light-speed, claiming that neutrinos don’t go faster than light. But I think they’re a bit premature. Much as I’d like to win, I don’t think I’ve won just yet.

For example, last week someone who works on artificial intelligence at a university in the Netherlands said that the OPERA team made a mistake in their use of special relativity—a mistake that explains away their result:

• Ronald A.J. van Elburg Times of flight between a source and a detector observed from a GPS satellite, 12 October 2011.

Two days later, a pseudonymous blogger who works for MIT’s Technology Review said the argument was “convincing”:

• KentuckyFC, Faster-than-light neutrino puzzle claimed solved by special relativity, The Physics arXiv Blog, 14 October 2011.

The popular news media got all excited! But they may have been getting ahead of themselves. After all, the OPERA team includes a bunch of particle physicists. Special relativity is child’s play for them. Would they really screw up that bad, after years of checking and rechecking their work? Chad Orzel suggests not:

• Chad Orzel, Experimentalists aren’t idiots: The neutrino saga continues, Uncertain Principles, 16 October 2011.

And none of the physicists I know find Elburg’s argument very convincing.

But that’s not all! A couple weeks earlier, Cohen and Glashow did a calculation:

• Andrew G. Cohen, Sheldon L. Glashow, New constraints on neutrino velocities, 29 September 2011.

According to this, faster-than-light neutrinos would lose energy by emitting lots of electron-positron pairs, a bit like how a supersonic jet makes a sonic boom. Two days ago, another team of physicists doing experiments on neutrinos at Gran Sasso claimed that together with their experiment, this result refutes the existence of faster-than-light neutrinos:

• ICARUS team, A search for the analogue to Cherenkov radiation by high energy neutrinos at superluminal speeds in ICARUS, 17 October 2011.

At least one good physics blogger has taken this work as “definitive”:

• Tomasso Dorigo, ICARUS refutes Opera’s superluminal neutrinos, A Quantum Diaries Survivor, 18 October 2011.

He says:

The saga of the superluminal neutrinos took a dramatic turn today, with the publication of a very simple yet definitive study by ICARUS…

And so, the news media are getting excited again, saying that now the OPERA result is really dead, like a vampire with two stakes through its heart.

But how “definitive” is this result, really? I’m a bit disappointed that the Cohen–Glashow paper doesn’t clearly state the assumptions that go into their argument. They zip through the calculation in an offhand way that suggests they’re using standard principles of physics to their heart’s content—in particular, special relativity. Normally that’s fine. But not here. After all, if faster-than-light neutrinos were signalling a breakdown of any of these principles, their calculation might be invalid.

Of course I don’t believe neutrinos are going faster than light: that’s why I made that bet! If you don’t want to believe it either, that’s fine. But if you want to entertain this possibility, in order to disprove it, you’d better be clear on the logic you’re using.

Without actually measuring the speed of neutrinos, the best you can hope for is something like this: “If theoretical principles X and Y and Z are true, then our experiment shows neutrinos don’t go faster than light.” So neutrinos could still go faster than light… but only if X or Y or Z is false.

Maybe X, Y and Z are principles we hold sacred—maybe even more sacred than the principle that nothing goes faster than light! But shocking discoveries can have shocking consequences. Sacred truths can fall like dominoes.

Given this, I think the only truly definitive way to hammer the nail in the coffin of the OPERA experiment is to either

1) find a mistake in the experiment that convincingly explains its result


2) do more measurements of the speed of neutrinos.

And maybe Dorigo acknowledges this, in a way. He says:

So, forget superluminal neutrinos. Or maybe not: what remains to be seen is whether other experiments will find results consistent with v=c or not. That’s right: regardless of the tight ICARUS bound, every nerd with a neutrino detector in his or her garage is already set up to produce an independent confirmation of the startling OPERA result… We’ll soon see measurements by MINOS and Borexino, for instance. Interesting times to be a neutrino expert are these!

(Emphasis mine.)

So, I’m going to wait and see what happens. I want to win my bet fair and square.

10 Responses to A Bet Concerning Neutrinos (Part 3)

  1. Tim van Beek says:

    For some background information have a look at this:

    • Neil Ashby: “Relativity in the Global Positioning System”, online here (living reviews).

  2. Frederik De Roo says:

    Btw, if one follows the link to Chad Orzel’s piece, one of the first comments there reads:

    I have a friend who is an expert in the field (his doctorate is about neutrinos and he’s an experimental physicist in a neutrino-based project currently) suspects that the error is somewhat stupid. He suspects that a confusion of sign convention occurred and in the analysis code something is being added instead of subtracted (or vice versa) yielding an expected delay 60ns off of the proper expected delay. But without access to the analysis code to examine, it’s hard to verify where specifically the error is, if that’s it.

    emphasis mine

    Well, I can’t tell whether this is a fair suggestion (especially if you look at the name of the author of that excerpt) but the reason I included it is only because it directly leads us to the Science Code Manifesto.

  3. linasv says:

    As Fredrick De Roo notes: suppose one has 100 different calibrations to perform, and one is 99.8% sure that each one was correct. This means that one should be about 0.998^100=80% certain in one’s final result! It just takes one flipped sign to blow up the final result.

    On that note, does the bet have any odds or handicap? Say 10 to 1? I remember a table in the original paper, where something like 10 different distances/corrections are totaled up to get the final result. If any one of these are wrong, then ..whooops.

    • John Baez says:

      It would be interesting to study complicated apparatus like the particle detectors at CERN and see how many parts need to work correctly to get an accurate answer, how much redundancy they build in to reduce the effect you mention, and how they test the apparatus to make sure each piece is working!

      Here are the terms of the bet:

      This bet concerns whether neutrinos can go faster than light. John Baez bets they cannot. For the sake of the environment and out of scientific curiosity, Frederik De Roo bets that they can.

      At any time before October 2021, either John or Frederik can claim they have won this bet. When that happens, they will try to agree whether it’s true beyond a reasonable doubt, false beyond a reasonable doubt, or uncertain that neutrinos can (under some conditions) go faster than light. If they cannot agree, the situation counts as uncertain.

      If they decide it’s true, John is only allowed to take one round-trip airplane trip during one of the next 5 years. John is allowed to choose which year this is. He can make his choice at any time (before 4 years have passed).

      If they decide it’s false, Frederik has to produce 10 decent Azimuth Library articles during one of the next 5 years—where ‘decent’ means ‘deserving of three thumbs up emoticons on the Azimuth Forum’. He is allowed to choose which year this is. He can make his choice at any time (before 4 years have passed).

      If they decide it’s uncertain, they can renegotiate the bet (or just decide not to continue it).

  4. Harry Erwin says:

    Neutrinos have mass in some sense. That would suggest their velocity should be less than c. It should also be variable.

    • John Baez says:

      Yes, that’s the usual theory—although neutrino masses are so small they’ve never been directly measured. There are some upper bounds… but all neutrino oscillations let us measure is the differences in the mass squared between various kinds of neutrinos.

      Unfortunately, the neutrino oscillation data seems to be consistent with neutrinos with negative mass squared! These would be tachyons. There are lots of good theoretical reasons why neutrinos should not be tachyons, assuming special relativity and quantum mechanics are true: tachyons cause all sorts of problems. It would be nice to measure the neutrinos masses more directly. But it seems very difficult.

      The particular neutrinos we’re talking about now, emitted by the CERN accelerator, have absurdly high energies: either 17 or 28 GeV, depending on the experiment. By comparison, the mass-energy of a neutrino is at most 0.3 eV, thanks to some bounds from cosmology. So, these particular neutrinos should be going very very near the speed of light… but not faster, unless our theories of physics are wrong!

  5. […] Superluminal Neutrinos,» A Quantum Diaries Survivor, October 18th 2011. En mi opinión (y no soy el único) Tommaso aquí ha pecado de rotundo y poco científico; él está seguro, igual que yo, que los […]

You can use Markdown or HTML in your comments. You can also use LaTeX, like this: $latex E = m c^2 $. The word 'latex' comes right after the first dollar sign, with a space after it.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.