_ring.jpg)
A reader sent me this article over the weekend. The long title indicates something exciting is happening:
A new experiment has broken the known rules of physics, hinting at a mysterious, unknown force that has shaped our universe
I have been sent similar articles by others. Most of them have the same breathless excitement: physicists have found that the laws of physics as we know them can’t be right, because an experiment at Fermilab shows that they are being broken. If true, this is really exciting news. However, much like the “faster-than light neutrino” results that were later found to be incorrect, I remain very skeptical that there is any reason to think that the laws of physics as we understand them are wrong.
So what’s the story here? Twenty years ago, physicists at Brookhaven National Lab were studying muons, which are particles that have the same negative charge as the electron but are significantly heavier. Because they are charged, they produce a magnetic field, just like the electron does. The physics that we know right now (collectively referred to as the “Standard Model”) predicts the behavior of particles that produce magnetic fields, and the way the electron behaves agrees perfectly with the Standard Model’s prediction. Because muons are heavier than electrons, their behavior is more complex, so they can be used as an additional test for the Standard Model. In the Brookhaven experiments, the muon’s behavior differed very slightly from the predictions of the Standard Model. However, because of the limits of the experiment, the physicists couldn’t rule out the idea that the result was a fluke, so the team made no concrete statement about the accuracy of the Standard Model when it comes to muon magnetic fields.
Recently, Fermilab announced that they replicated the Brookhaven experiment, using the same basic setup as what was used at Brookhaven, but they did it with higher precision. They confirmed the Brookhaven experiment’s results, and based on the quality of their data, statistics indicate there is only a 1 in 40,000 chance that the result is a fluke. Because of this, there is a lot of excitement in some parts of the physics community. After all, if the Standard Model can’t correctly predict something, that means there is something wrong with it, and that means there is “new physics” to discover.
Of course, this line of reasoning ignores one inconvenient fact. There is another possible reason the Standard Model’s prediction for the muon is wrong: The prediction itself could be incorrect. It turns out that the muon’s complex behavior causes the math in the Standard Model to be very difficult to solve. As a result, the prediction against which these experiments are compared used a well-accepted shortcut: it incorporated some independent experimental results into the calculations to make things easier. Of course, this leads to a problem. Those experimental results produce uncertainties, because all experiments have error in them. As a result, what’s really going on here is that an uncertain prediction is being compared to an experimental result, which has its own uncertainties. When uncertain things are compared to other uncertain things, it’s not clear what the difference between them means.
To solve this problem, Borsanyi and colleagues did the tough math. They used millions of CPU hours at a supercomputer so that the prediction they produced was purely mathematical. They found that their prediction agreed with the experimental results at both Brookhaven and Fermilab. Thus, as far as they are concerned, there is no discrepancy between the behavior of the muon and the predictions of the Standard Model. Is that the end of the story? Of course not! It could be that Borsanyi and colleagues are wrong. However, I would think their conclusion is more reliable, since a prediction made with pure mathematics has less uncertainty in it.
So my bet is that there is no new physics here, and the Standard Model has been vindicated once again. Of course, only time will tell whether or not I am right. However, there is a lesson to be learned here, and it is an important one. Borsanyi and colleagues’ calculation was published at the same time as the Fermilab results. However, most science “journalists” aren’t bothering to mention their conclusion. Why? Either because it isn’t exciting, or because they haven’t bothered to see what other physicists are saying about the situation. Either way, it’s truly unfortunate, and it confirms what I have said many times before: most “science journalists” know little about science and even less about journalism.
Back when I started grad school I was planning on doing QCD (before my potential advisor left my institution), and I read the most recent Physics Reports review of the muon g-2. They have to do five-loop QED calculations for this thing. And QCD is already difficult enough. When things get this complicated, there are so many opportunities for simple calculational errors it’s almost impossible to know the final number is correct. If “science journalism” bothered to try to explain to people just how difficult these calculations are, it might be useful for once.
I was hoping you would comment on this, since you are more familiar with these kinds of lines of inquiry than I am. Thanks!
I’ll say a bit more, then, since I hadn’t thought much about this before yesterday:
I never got to the point of doing QCD research, so I wouldn’t be able to go look at the relevant papers and provide you a judgment. Thus I haven’t actually looked at the paper on this new result; I have friends in particle physics who told me about it (one was actually interviewed about it by one of the major newspapers; the Wall Street Journal, I think). Among my colleagues in physics we have a tendency to respond to results or propositions with a curt “fine” or “good.” We’d elaborate if necessary, of course, but my immediate attitude to the new result is basically “okay, fine.” They did a good thing; more data helps. But it’s very clearly not enough data to make a conclusion, and that one can tell simply from the little that was said in the announcements of the result.
That review I linked to is on the order of 200 pages long. In general that’s just how difficult research-level science is, but for the muon anomalous magnetic moment I don’t think it’s an exaggeration to say the calculations involved are Herculean. There are so many enormous calculations that go into the theoretical prediction that there could easily be errors we made that we don’t know about. For context, a “simple” one-loop calculation in QED—which we understand!—can take 20 pages front and back to complete. A single mistake in those 20 pages will give you the wrong answer; thus even the easiest calculations in QFT can be an absolute nightmare. As you note in your post, we have yet to nail down the theoretical prediction precisely; a single mistake in the mountain of calculations necessary to solve this problem could shift the magnetic moment enough to change our conclusions about the experimental results. So at this point, what we need is more data and more calculations—which is precisely where we were a decade ago. We have more data than we used to, but not enough to make hard conclusions.
Finally, the experimental particle physics grad students at my institution are so used to seeing potentially promising signals in the data disappear after more careful analysis is done that they speak of the phenomenon as “bump-squashing.” So they’ve developed a strong commitment to dashing their own hopes for discovering something new. I think this is a prudent attitude.
Thanks. I agree that “Herculean” is a good description of these calculations, and I think the experimental particle physics grad students at your institution have the right idea!