The 4-Percent Universe

by Richard Panek

  But for Perlmutter these results also represented the realization of his dream of using physics to solve the big mysteries. "For the first time," he announced at the AAS press conference, "we're going to actually have data, so that you will go to an experimentalist to find out what the cosmology of the universe is, not to a philosopher." Afterward, he had stayed at a table in the room for an hour, conducting a mini-seminar for the members of the press. They surrounded him, and he held forth. Later, when they played back their cassette tapes, they might think they'd inadvertently hit fast-forward. But no, it was just Saul Perlmutter at regular speed, hyperkinetically trying to convince them that the headline here wasn't just the fate of the universe. It was that we could now know that fate—empirically, scientifically.

  The following day Michael Turner paid Perlmutter a visit in the exhibit hall. The SCP team was part of the AAS meeting's poster sessions that day—dozens of presentations tacked to freestanding corkboards in long lanes, hard by the trade-show booths where representatives from weapons manufacturers sat at white-linen-covered tables and explained why the telescopes on their drawing boards were the best. Perlmutter wanted to show Turner something in the data, something he hadn't mentioned at the press conference.

  Turner liked Perlmutter, and he liked the project; he didn't need to be convinced that the supernova survey was a worthy effort that deserved support from telescope time allocation committees or the National Science Foundation or the Department of Energy. Turner bent close to the panels—eight in all. The first few explained supernova search methodology to the uninitiated. One showed the logistics of the project: the initial observations at the Cerro Tololo 4-meter telescope, the follow-up observations at Cerro Tololo three weeks later, the spectroscopy at Keck, the photometry with telescopes at Kitt Peak and Isaac Newton and, at the highest redshifts, with the Hubble Space Telescope. The second showed light curves from twenty-one of the team's supernovae, the third showed some spectra, the fourth redshifts. The fifth explained how the SCP had calculated the photometry, made some corrections, and applied the stretch method to convert the Type Ia supernovae into calibrated candles, and how that calibration allowed them to plot the supernovae on a redshift-magnitude Hubble diagram. The sixth panel showed the Hubble diagram from the Nature paper, the basis for the claim that the universe will expand forever. Nothing Turner didn't already know.

  But then came the seventh panel. It showed two plots. They were contour plots—plots that take the cumulative statistical effect (rather than individual points) of all the data and plop them on a graph that covers every possible scenario for the life of the universe. If your contour of confidence falls over here, in this region, then you have a universe without a Big Bang. If it falls over there, in that region, then you have a Big Bang universe that expands forever, and if it falls a little bit lower, then you have a Big Bang universe that recollapses eventually.

  The plot on the left was, like the diagram in the sixth panel, from the Nature paper. It reflected the statistical effects of six supernovae, including the 1997 supernova that the SCP team had examined with the Hubble Space Telescope. And it did indicate that the addition of the one HST supernova shifted the likelihood up, into the region corresponding to a universe that expands forever.

  The plot on the right, however, was new. It reflected the statistical effects of the dozens of other supernovae the team had found as well—forty in all. As you might expect, the addition of all that data had tightened the contours, narrowing the confidence regions. Looking at the graph on the left and then at the graph on the right was like putting on glasses; suddenly the fuzzy outlines of the world—the universe—snapped into focus.

  The analysis was preliminary. But the effect so far was arresting. If you knew what you were seeing, you would get it at a glance. Yes, the universe was going to expand forever. But the evidence seemed to be indicating that even in order to exist, the universe couldn't be made up simply of matter, dark or otherwise. It needed something else.

  Turner straightened up. "Dave would have liked that," he said.

  Turner was at the AAS to lead a memorial service for David Schramm. Already he had attended one in Aspen, and he would be leading another later that month at Rockefeller Chapel, on the University of Chicago campus. Here were the supernovae that Schramm had been pestering Turner about, but if the data held up, here too was a hint of a further tragedy. Schramm had spent decades trying to rethink the universe, only to die, like Hubble at Palomar, in sight of the Promised Land—though perhaps even more poignantly. In the half-century since first light, the telescope at Palomar hadn't delivered what Hubble had hoped it would: the two numbers that had kept his protégé Allan Sandage keening all the way into retirement. But if the SCP data held up, then science was entering an era that Schramm had always envisioned: a new cosmology.

  In 1917, in considering the implications of general relativity, Einstein saw that the universe was inherently unstable. Just as Newton had invoked God to keep his version of the universe from collapsing, so Einstein added a symbol to his equations—arbitrarily, the Greek letter lambda, A. Whatever lambda was, it was counteracting gravity, because, in Einstein's idea of a stable universe, something had to be. It was the reason that a universe full of matter attracting other matter through gravity wasn't collapsing. After Hubble's discovery of evidence for the expansion, the universe didn't need lambda, and Einstein abandoned it. Unlike Newton's God, however, you couldn't altogether ignore it. Lambda was, after all, in the equation.

  What you could do instead was set lambda to zero. That's what generations of observers and theorists had done. Sometimes they left the assumption implicit, simply failing to mention lambda. Often they stated the assumption explicitly: "Assume A = 0." For most observers and theorists, lambda was there and it wasn't there. It occupied a parallel existence, like a ghost in the attic.

  Just because you didn't need it, however, didn't mean you couldn't invoke it, and from time to time theorists had done just that. In 1948, when Hermann Bondi and Thomas Gold and, separately, Fred Hoyle were trying to create a new model of the universe that didn't rely on an initial singularity of infinite density but still seemed to be expanding, they invoked what Bondi and Gold called the "hypothetical and much debated cosmological term." Like Einstein, they didn't know what it was, but they set it to non-zero because something had to be fueling the expansion. But then the validation of the Big Bang theory through the discovery of the cosmic background radiation eliminated the need for what had come to be called "the cosmological constant." Lambda didn't exactly die with the Steady State, but it fled the corpse, like a soul escaping.

  It next took up residence in quasars, those mysterious sources of tremendous energy at mystifying distances. In 1967, a trio of Cornell theorists published a paper in the Astrophysical Journal examining, as the title said, "Quasi-Stellar Objects in Universes with Non-Zero Cosmological Constant." They were trying to resolve some possible inconsistencies in the behavior of quasars. But as the understanding of the evolution of quasars became clearer, the need for lambda again receded. Then in 1975 two prominent astronomers argued in Nature that studies of elliptical galaxies as standard candles indicated that "the most plausible cosmological models have a positive cosmological constant." A year later they wrote another paper explaining why elliptical galaxies don't make good standard candles, implicitly undermining their earlier argument.

  Four times now, including Einstein, cosmologists had gone up into the attic, and four times they'd returned with the same report: It was just the wind.

  Then came inflation. It solved problems, the flatness and horizon problems. It explained improbabilities, the homogeneity and isotropy of the universe on the largest scales. And while the participants of the "Very Early Universe" Nuffield workshop at Cambridge in the summer of 1982 didn't agree on a model for inflation, they did, crucially, agree that a model could exist, and in the weeks and months after the workshop they formed a consensus around one model, giving inflati
on a solid basis in mathematics. But most important for its long-term survival or eventual obsolescence, inflation came with a prediction: that the universe was flat. That the amount of matter in the universe was equal to the critical amount that would keep it from collapsing. That omega equaled 1.

  The problem for the inflation theorists, however, was that the observers were consistently finding evidence that the amount of matter in the universe was only 20 percent of the critical amount—that omega equaled 0.2.

  At the final session of the Nuffield workshop, the theoretical physicist Frank Wilczek summarized the conference proceedings, concluding with "A Shopping List of Questions." Among them was whether omega was equal to 1. "If not," he said, "we must give up on inflation." Simple subtraction led you to conclude that for omega to equal 1 while observers were finding evidence that omega equaled 0.2, observers must be missing 0.8, or 80 percent, of the universe.

  This discrepancy wasn't as worrisome as one might imagine. Two options immediately presented themselves. Maybe the rest of the matter was in a form that astronomers hadn't yet detected. The community had only recently conceded that the evidence for dark matter was compelling, and theorists were still working through the implications of dark matter for the structure and evolution of the universe. Or maybe the observers were just wrong, and more precise observations with improved instruments would boost omega and resolve the discrepancy.

  A third option also existed, and if it, too, beckoned immediately, it did so from a distance, or even a different dimension. In any case, it was easy and probably advisable to ignore. Wilczek ended the Nuffield workshop with the last question of his "shopping list":

  What about the cosmological constant?

  "Whereof one cannot speak, thereof one must be silent."

  —Wittgenstein

  Be silent? Be loud! Michael Turner went home from the Nuffield workshop, downed a slice of Primordial Pizza, and, along with fellow theorists Gary Steigman and Lawrence M. Krauss, got to work on a paper that explored the options for making omega equal to 1, entitled "Flatness of the Universe: Reconciling Theoretical Prejudices with Observational Data." Those "theoretical prejudices" referred to inflation's prediction of a flat universe, and the paper explored two ways of reconciling those prejudices with the data. One was a particle of some sort from the era of Big Bang nucleosynthesis—the field that Schramm had pioneered. The other possibility was "a relic cosmological constant."

  "The cosmological constant," Turner liked to say, "is the last refuge of scoundrel cosmologists, beginning with Einstein." He himself, in his "heart of hearts," thought the cosmological constant must be zero. But he also knew that the cosmological constant had "every right to be there." And as he and Rocky Kolb often insisted, their generation wasn't going to make the mistake that Einstein and other twentieth-century cosmologists had made by not taking every remotely serious option seriously.

  If anything, the self-described scientific conservative Jim Peebles took the idea even more seriously than Turner—but then, Peebles prided himself on trusting observations more than most theorists. "What's best," he would say, shrugging, "is what's true." The truth for him had emerged in a 1983 paper he wrote with Marc Davis, a UC Berkeley astronomer, using the latest and largest survey of galaxies to measure their velocities, infer their masses, and derive the mass density of the universe. Peebles looked at their data and thought, "High mass density is dead in the water." Their conclusion: an omega of 0.2.

  The following year, Peebles wrote a paper, "Tests of Cosmological Models Constrained by Inflation," that offered his theoretical interpretation of that data. Maybe omega was indeed 0.2 and lambda equaled 0, he wrote, but in that case "we lose the attractive inflationary explanation for the observed large-scale homogeneity of the universe." He didn't want a cosmological constant. "It's ugly," he often said. "It's an addition." If he were building a universe, he thought, he wouldn't put in a cosmological constant: "No bells and whistles." But perhaps because inflation solved the flatness problem he'd articulated with Dicke, or perhaps because he constitutionally distrusted a simple universe, he accepted the possibility with equanimity. "Considering the observations," he said, "I think the universe might have put in bells and whistles—a cosmological constant."

  The paper met with a lot of resistance, which Peebles sort of enjoyed. He found that he could go to conferences and give a talk, and people would rant at him, and then they apparently forgot, because a few months later he would give the same talk and the same people would rant at him. He realized he didn't even have to write a new talk; he could just give the same one over and over. This went on for a dec ade.

  Theorists are always saying something. That's their job. They don't need to believe what they're saying. The theorist's goal isn't to be right but to be reasonable—to make an internally consistent argument that observers can then go out and reinforce or disprove. For their part, observers regard theorists with patience and exasperation, like a dog that's always depositing gifts at their feet: a stick, a squeaky toy, a dead bird. Often these offerings just lie there. But once in a while the observers will throw them a bone. Go fetch.

  In 1992 observers threw cosmological theorists the biggest bone since the discovery of the cosmic microwave background more than a quarter of a century earlier: the Cosmic Background Explorer results—the ones that said that the universe was flat. The following year, Turner and Kolb added a preface to the paperback edition of The Early Universe reviewing the COBE results and declaring them "a shot in the arm" for a flat universe.

  As other observations accumulated that indicated a universe with a low density of matter—especially the kind of studies of galaxies on the largest scale that had first persuaded Jim Peebles a decade earlier—theorists found themselves increasingly less reluctant to suggest, and observers found themselves increasingly less reluctant to consider, the possibility of a cosmological constant. "why a cosmological constant seems inevitable" read a section heading in one influential paper; "The Observational Case for a Low-Density Universe with a Non-Zero Cosmological Constant" was the title of another paper. And then there was Turner again, again with Lawrence Krauss: "The Cosmological Constant Is Back." The cosmological constant was still the last refuge, but it was a refuge nonetheless.

  Vera Rubin summarized the situation with a joke. There was a wise rabbi, she said, who was trying to mediate a marital dispute. The husband complained about the wife. "You're right," the rabbi said. The wife complained about the husband. "You're right," the rabbi said. Then the rabbi's own wife emerged from behind a curtain, where she had been eavesdropping. "How can you tell them both they're right?" she said to her husband. To which the wise rabbi replied, "You're right too."

  She told this joke at a "Critical Dialogues in Cosmology" conference at Princeton, part of the university's celebration of its 250th anniversary, in the summer of 1996. The purpose of the conference was to bring together the world's leading cosmologists to address the field's greatest challenges. One such event, inevitably, involved the value of omega, and it took the form of a debate. On one side was Avishai Dekel, who had recently measured galaxy motions that were consistent with an omega equal to 1. On the other side was Turner, arguing that the amount of matter in the universe was not enough to nudge omega to 1. But he didn't stop there. Instead he used the forum to argue that omega was indeed equal to 1, because the cosmological constant would close the gap.

  The moderator was none other than Bob Kirshner. At one point in the discussion he turned to Saul Perlmutter, who had arrived at Princeton bearing preliminary results from the SCP's first seven supernovae. Kirshner asked what he thought.

  Like any cosmologists dealing with omega, the SCP had addressed lambda in paper after paper: "(for ? = 0)," "(for Λ = 0)," "If we assume that the cosmological constant = 0." A year earlier, in 1995, Perlmutter and Ariel Goobar had elevated the cosmological constant from its pro forma purgatory, making its existence the subject of a paper in the Astrophysical Journal. Or, rather,
making its nonexistence the subject of the paper, since their assumption while writing it was that matter would indeed account for everything. They figured that they would be explaining how astronomers could use supernovae to show, once and for all, that lambda equals 0.

  And that's what Perlmutter had come to Princeton ready to discuss. Yes, he reported. Now the SCP's first seven supernovae were consistent with a universe where omega equals 1 and lambda equals 0.

  "This could be a lambda killer," Jim Peebles told a journalist.

  Lambda killer. Perlmutter liked the sound of that: Get lambda out of the way so it won't be a spoiler anymore.

  Like his mentor David Schramm, Michael Turner didn't like to lose debates. And like Schramm, he wasn't afraid to practice what the Fermilab and Chicago cosmologists called "jugular science." During a break in the Princeton activities, as various astronomers and cosmologists were climbing a flight of stairs to an auditorium, Turner sent Perlmutter a message. Ostensibly talking to the astronomer walking beside him, Turner raised his voice.

  "I don't think Saul is that stupid," he said.

  Perlmutter didn't appear to hear.

  "I said," Turner repeated, raising his voice, "'I don't think Saul is that stupid.'"

  Turner was slightly more diplomatic during his own talk, but no less needling. "I am anxiously awaiting the results of the two deep searches for supernovae," he said, referring to the rival teams. "I think they're going to shed some important light on this. To draw any conclusion now would be to take away from their thunder later."