This month's provocative results on the acceleration of the universe raise an interesting issue: what can we say about our universe's ultimate fate? In the old days (like, when I was in grad school) we were told a story that was simple, compelling, and wrong. It went like this: matter acts to slow down the expansion of the universe, and also to give it spatial curvature. If there is enough matter, space will be positively curved (like a sphere) and will eventually collapse into a Big Crunch. If there is little matter, space will be negatively curved (like a saddle) and expand forever. And if the matter content is just right, space will be flat and will just barely expand forever, slowing down all the while.
This story is wrong in a couple of important ways. First and foremost, the assumption that the only important component of the universe is "matter" (or radiation, for that matter) is unduly restrictive. Now that we think that there is dark energy, the simple relation between spatial curvature and the ultimate fate of the universe is completely out the window. We can have positively curved universes that expand forever, negatively curved ones that recollapse, or what have you. (See my little article on the cosmological constant.) To determine the ultimate fate of the universe, you need to know both how much dark energy there is, and how it changes with time. (Mark has also written about this with Dragan Huterer and Glenn Starkman.) If we take current observations at face value, and make the economical assumption that the dark energy is strictly constant in density, all indications are that the universe is going to expand forever, never to recollapse. If any of your friends go on a trip that extends beyond the Hubble radius (about ten billion light-years), kiss them goodbye, because they won't ever be able to return -- the space in between you and them will expand so quickly that they couldn't get back to you, even if they were moving at the speed of light. Meanwhile, stars will die out and eventually collapse to black holes. The black holes will ultimately evaporate, leaving nothing in the universe but an increasingly dilute and cold gas of particles. A desolate, quiet, and lonely universe. However, if the dark energy density actually increases with time, as it does with phantom energy, a completely new possibility presents itself: not a Big Crunch, but a Big Rip. Explored by McInnes and by Robert Caldwell, Marc Kamionkowski, and Nevin Weinberg, the Big Rip happens when the universe isn't just accelerating, but super-accelerating -- i.e., the rate of acceleration is perpetually increasing. If that happens, all hell breaks loose. The super-accelerated expansion of spacetime exerts a stretching force on all the galaxies, stars, and atoms in the universe. As it increases in strength, every bound structure in the universe is ultimately ripped apart. Eventually we hit a singularity, but a very different one than in the Big Crunch picture: rather than being squashed together, matter is torn to bits and scattered to infinity in a finite amount of time. Some observations, including the new gamma-ray-burst results, show a tiny preference for an increasing dark energy density; but given the implications of such a result, they are far from meeting the standard for convincing anyone that we've confidently measured any evolution of the dark energy at all. So, it sounds like we'd like to know whether this Big Rip thing is going to happen, right? Yes, but there's bad news: we don't know if we're headed for a Big Rip, and no set of cosmological observations will ever tell us. The point is, observations of the past and present are never by themselves sufficient to predict the future. That can only be done within the framework of a theory in which we have confidence. We can say that the universe will hit a Big Rip in so-and-so many years if the dark energy is increasing in density at a certain rate and we are sure that it will continue to increase at that rate. But how can we ever be sure of what the dark energy will do twenty trillion years from now? Only by actually understanding the nature of the dark energy can we extrapolate from present behavior to the distant future. In fact, it's perfectly straightforward (and arguably more natural) for a phase of super-accelerated expansion to last for a while, before settling down to a more gently accerated phase, avoiding the Big Rip entirely. Truth is, we just don't know. This is one of those problems that ineluctably depends on progress in both observation and theory. Of course, it's thoroughly remarkable that we can even think about these possibilities in a scientific way, even if we haven't yet isolated the ultimate answer. Before the dynamical spacetime of Einstein's general relativity replaced Sir Isaac Newton's view of absolute space and time, we couldn't have been having this conversation. Newtonian spacetime is there once and for all, fixed and unchanging in its structure, only permitting the matter within it to evolve. It's really quite difficult to come up with any sensible long-term history of the universe within Newtonian cosmology; in any finite region of space, the matter will ultimately settle down to some equilibrium -- so why is evolving now, if spacetime is infinitely old? In Einstein's universe, spacetime can evolve and change, and we have grounds for speculating about where it came from and where it's going. In the last ten years we've learned an amazing amount about the universe that bears directly on this question. But the old picture of collapse vs. perpetual expansion is wrong in another way, too: it takes a feature of our observable universe, namely that it is homogeneous and isotropic, and extrapolates it to the entire universe, even the unobservable bits. This is a reasonable first guess on grounds of simplicity, at least if you didn't have any reason to suspect that the ultra-large-scale structure of the universe were wildly different from place to place. But these days we do have such a reason: eternal inflation. The idea of cosmic inflation invokes a period of rapidly accelerated expansion in the very early universe, that takes a tiny patch of space and blows it up to fantastic size. Eternal inflation is simply the realization that this phase doesn't end everywhere at once: inflation typically stops in some part of the universe, but continues on somewhere else -- forever! In other words, somewhere far away, inflation is still going on. And this idea isn't some baroque kind of model that requires a handful of miracles to set in motion; to the best of our current understanding, it's very easy for inflation to be eternal once it begins at all. That means that the ultimate expansion or contraction of our meager patch of universe is just a tiny part of the big picture -- so we shouldn't take the fate of our little neighborhood too seriously. Meanwhile, of course, there's the issue of the distant past of our universe, as well as the distant future. It's becoming more and more popular to contemplate the idea that the Big Bang wasn't the start of it all, but simply a dramatic moment in a much larger picture. For a long time now, Gabrielle Veneziano and others have been investigating the idea of a pre-Big-Bang phase in the context of string theory. Steinhardt and Turok have suggested that the universe is cyclic, repeating an infinite pattern of expansions and collapses. And of course Jennie Chen and I have been arguing (following the exhortations of Huw Price about the arrow of time) that the far past should look like the far future, only backwards. Thinking about what happened before the Big Bang is precisely as respectable as, although admittedly more difficult than, thinking about what the ultimate future of our universe will be. In each case we have to extrapolate into unknown territory, relying on a combination of observational clues and theoretical predictions. And in each case the current state of the art isn't nearly good enough for us to make any definitive statements. But no need to invoke the God of the Gaps just yet! The amazing thing is not that these questions are hard, but that they are legitimate scientific issues, and that we are increasingly able to address them in the context of established (or at least plausible) physical theories. Stick around, we'll figure it out.