It was the science story of the year: Astrophysicists held a news conference at Harvard on March 17 announcing that their South Pole telescope had found evidence of gravity waves from the dawn of time.
Cosmology doesn’t get any bigger than this. The discovery was hailed as confirmation of a mind-boggling addendum to the big-bang theory, something called “cosmic inflation” that describes the universe beginning not in a stately expansion but with a brief, exponentially rapid, inflationary spasm.
Science is a demanding and unforgiving business, and great discoveries are greeted not with parades and champagne but rather with questions, doubts and demands for more data. So it is that, in recent days, scientists in the astrophysics community have been vocalizing their concern that the South Pole experiment, known as BICEP2, may have detected only the signature of dust in our own galaxy.
These doubters say, in effect, that rather than seeing the aftershock of the birth of the universe the scientists may have seen only some schmutz in the foreground, as if they needed to clean their eyeglasses.
This is a delicate issue. Careers and prizes are potentially at stake. So too is the credibility of a field that dares to probe the deepest secrets of the universe no matter where that search may lead.
No one is alleging an outright scientific error. It’s more of a debate about how scientists should communicate their uncertainties when presenting blockbuster findings. This is a case of “extraordinary claims demand extraordinary evidence,” to use the formulation made famous by astronomer Carl Sagan.
The South Pole telescope saw something in the sky — of that there is little doubt, because the team took great care to eliminate systematic errors that could have come from the instruments. But what the telescope saw — polarization of ancient radiation from the early universe — could have been produced by either primordial gravity waves or by foreground dust, or by some combination of both.
“They have very nice measurements of something. We don’t know what that something is,” said Uros Seljak, a professor of physics and astronomy at the University of California at Berkeley. “We can’t tell if BICEP2 has measured dust or has measured gravity waves.”
John Kovac, a Harvard astrophysicist and the principal investigator for BICEP2 — part of a larger collaboration among institutions from coast to coast — stands firmly behind his team’s findings. But he acknowledges that there are lingering uncertainties that will remain until new data is presented, likely this fall, by the European Space Agency’s Planck Space Telescope.
“We are very confident that we have measured B-modes” — polarization of light that can be caused by gravity waves — “with high statistical significance in the sky, and we have looked at them in multiple ways. And the data suggest they are unlikely to be dominated by galactic foregrounds. That is not to say that there is not uncertainty about that,” Kovac said.
Now everyone is waiting for the Planck results. The Planck telescope is scrutinizing the cosmic microwave background (CMB) radiation, a remnant of the moment when the young cosmos became transparent to light. BICEP2 looked at the radiation in a small patch of sky, but Planck is doing an all-sky survey in multiple frequencies and should produce excellent estimates of foreground effects such as dust.
The big announcement March 17 thrilled the cosmology community. People stopped what they were doing to examine the results. Cosmic inflation had been discussed for more than three decades, but this would be the first strong evidence for it.
“What inflation does is pull apart the fabric of space-time much faster than the speed of light,” said Princeton professor of physics Suzanne Staggs, part of a team studying the cosmic background radiation with a telescope in Chile.
“It’s an amazing thing that quantum processes a moment after creation may have been stretched out by inflationary expansion and literally etched across space itself — stretched across the sky,” said Brian Greene, a Columbia University theorist who is not involved in the experiments. “If we are seeing that, it is an monumental moment.”
All the more reason, he said, to ask the hard questions about the discovery, and “hit it with a sledgehammer and see if it survives.”
The BICEP2 dust-up is being played out on, among other places, a Facebook page started by people who couldn’t access the live-streamed Internet feed of the March 17 announcement at the Harvard-Smithsonian Center for Astrophysics. The controversy has popped up in recent days on physics blogs and in several science-oriented publications, including the online news sites of the journals Nature and Science, as well as New Scientist and National Geographic.
Thursday morning, theoretical physicist Raphael Flauger, who has a dual appointment at New York University and the Institute for Advanced Study in Princeton, N.J., gave a presentation at Princeton University outlining his concerns about the BICEP2 conclusions.
“I think foregrounds are maybe larger than they thought, but I’m still hopeful that there’s a signal there,” Flauger said in an interview later. “It will be the biggest discovery maybe in my lifetime. It would be a measurement of the conditions, if you want, when the universe was 10 to the minus-30 seconds old” — a millionth of a trillionth of a trillionth of a second.
Flauger noted that the BICEP2 team estimated the amount of foreground dust by relying, in part, on a slide presented at a conference. The slide gave a visual representation of observations by the Planck spacecraft but did not contain the actual data. Critics say this is not a robust way to make an estimate.
But the BICEP2 scientists say they did their best given that the Planck team has not yet shared the hard data behind that slide. Kovac emphasized that his team used not just one, but six different models to estimate the effects of foreground dust and other factors that might have contributed to polarization of the ancient radiation. The measured polarization was many times greater than would be expected from dust within the galaxy, Kovac said.
“We expressed very carefully, as scientists do, the uncertainties that were inherent in the measurement we made,” Kovac said. “And I think we got it right. It’s unscientific to imagine that one could eliminate all kinds of uncertainty. The credibility of the scientific process doesn’t rest on that. It rests on doing a careful job. In the four years that we analyzed this data set, we did a very careful job.”
Jamie Bock, a Caltech astrophysicist who is one of the principal investigators of the broader collaboration that includes BICEP2, said the objections made by critics about the use of the Planck slide may spur changes in the original paper submitted to a scientific journal for peer review.
“That’s certainly a possibility. We haven’t made a final decision,” Bock said. “We recognize that the models [for the effects of foreground dust] are not very good models.”
But he said this is simply a case of science doing what science does best: pressing a new finding through the fine filter of skepticism.
“This is the way science works,” Bock said. “It’s a very exciting result. We expect that people are going to want to be completely sure. We want to be completely sure.”
Bock said the team was clear about the foreground uncertainties. “When we wrote the paper, when we presented the results,” he said, “we always had this issue that we don’t have the data that directly constrain dust.”
Bock said there were other lines of evidence to bolster the belief that this was, indeed, a signal from the very early universe. His team conveyed plenty of confidence on March 17; the roll-out of the discovery included the distribution of a video, which went viral, showing a scientist informing theorist Andrei Linde that the idea he helped develop more than three decades ago had been confirmed to a high degree of probability.
The final line of the team’s paper, which was submitted for peer review, lightly hedged the bet: “The long search for tensor B-modes is apparently over, and a new era of B-mode cosmology has begun.”
After the March 17 news conference, Kovac told The Washington Post that the observation closely matched what was predicted by the theory of inflation. He added: “But it’s the case that science can never actually prove a theory to be true. There could always be an alternative explanation that we haven’t been clever enough to think of.”
With so much at stake, the small community of scientists who study the cosmic microwave background radiation is on edge, waiting for the discovery to be confirmed by additional data.
“People are uncertain and kind of nervous,” said Philipp Mertsch, a theoretical astrophysicist at Stanford. “I’m still rooting for the BICEP2 team. But I think we need to be really careful.”
Much of the skepticism has emerged from Princeton.
“We need to get it right,” Paul Steinhardt, a theorist who helped develop the theory of cosmic inflation, said when introducing Flauger on Thursday. “It’s not very easy to eke out of nature its secrets. It takes extraordinarily talented people to do that. The challenge we’re talking about here is measuring a signal that is at a level of a few millionths of a degree.”
Among the critics is Bill Jones, a Princeton professor who is leading a rival experiment at the South Pole called SPIDER. It will use a balloon to loft an observational instrument that will search for signs of gravity waves as it circles the pole in conditions akin to outer space.
“The bar of evidence for something like this is extremely high,” Jones said. “It hasn’t met the standard of evidence that’s required, so someone shouldn’t make the claim.”
Jones questioned the decision by the BICEP2 team to announce the discovery in a news conference prior to publication of a peer-reviewed paper. Kovac and Bock defended the news conference as a common procedure.
These are all people who know each other well. Jones had been part of the BICEP1 collaboration but was not part of BICEP2. He said he felt cut out of the loop last year when he was no longer included on e-mail lists dealing with BICEP2. He said the community as a whole suffered a tragic blow with the death of Andrew Lange, a mentor to many young scientists. Lange, a professor at Caltech, took his life in 2010 at the age of 52.
“The deterioration of the group was a very real, if in perspective trivial, consequence of his absence; I am quite sure that we wouldn’t be having this discussion were he still at the helm of the BICEP experiment,” Jones said in an e-mail.
The BICEP2 story provides a reminder that science is a human enterprise, not an automatic and bloodless aggregation of data. The most important science is often conducted in difficult terrain, where data is elusive and the signals are ambiguous. Cosmologists studying the origin of the universe know that they need to be cautious, but what they do is, by its very nature, ambitious in the extreme.
“Scientists are human, and humans are competitive,” Brian Greene said. “That’s true whether you’re a scientist or you play on the Yankees.”