For decades, the commonly held view among scientists has been that space and time first emerged about fourteen billion years ago in a big bang. According to this picture, the cosmos transformed from a nearly uniform gas of elementary particles to its current complex hierarchy of structure, ranging from quarks to galaxy superclusters, through an evolutionary process governed by simple, universal physical laws. In the past few years, though, confidence in this point of view has been shaken as physicists have discovered finely tuned features of our universe that seem to defy natural explanation.
The prime culprit is the cosmological constant, which astronomers have measured to be exponentially smaller than naïve estimates would predict. On the one hand, it is crucial that the cosmological constant be so small or else it would cause space to expand so rapidly that galaxies and stars would never form. On the other hand, no theoretical mechanism has been found within the standard Big Bang picture that would explain the tiny value.
Desperation has led to a "dangerous" idea: perhaps we live in an anthropically selected universe. According to this view, we live in a multiverse (a multitude of universes) in which the cosmological constant varies randomly from one universe to the next. In most universes, the value is incompatible with the formation of galaxies, planets, and stars. The reason why our cosmological constant has the value it does is because it it is one of the rare examples in which the value happens to lie in the narrow range compatible with life.
This is the ultimate example of "unintelligent design": the multiverse tries every possibility with reckless abandon and only very rarely gets things "right;" that is, consistent with everything we actually observe. It suggests that the creation of unimaginably enormous volumes of uninhabitable space is essential to obtain a few rare habitable spaces.
I consider this approach to be extremely dangerous for two reasons. First, it relies on complex assumptions about physical conditions far beyond the range of conceivable observation so it is not scientifically verifiable. Secondly, I think it leads inevitably to a depressing end to science. What is the point of exploring further the randomly chosen physical properties in our tiny corner of the multiverse if most of the multiverse is so different. I think it is far too early to be so desperate. This is a dangerous idea that I am simply unwilling to contemplate.
My own "dangerous" idea is more optimistic but precarious because it bucks the current trends in cosmological thinking. I believe that the finely tuned features may be naturally explained by supposing that our universe is much older than we have imagined. With more time, a new possibility emerges. The cosmological "constant" may not be constant after all. Perhaps it is varying so slowly that it only appears to be constant. Originally it had the much larger value that we would naturally estimate, but the universe is so old that its value has had a chance to relax to the tiny value measured today. Furthermore, in several concrete examples, one finds that the evolution of the cosmological constant slows down as its value approaches zero, so most of the history of the universe transpires when its value is tiny, just as we find today.
This idea that the cosmological constant is decreasing has been considered in the past. In fact, physically plausible slow-relaxation mechanisms have been identified. But the timing was thought to be impossible. If the cosmological constant decreases very slowly, it causes the expansion rate to accelerate too early and galaxies never form. If it decreases too quickly, the expansion rate never accelerates, which is inconsistent with recent observations. As long as the cosmological constant has only 14 billion years to evolve, there is no feasible solution.
But, recently, some cosmologists have been exploring the possibility that the universe is exponentially older. In this picture, the evolution of the universe is cyclic. The Big Bang is not the beginning of space and time but, rather, a sudden creation of hot matter and radiation that marks the transition from one period of expansion and cooling to the next cycle of evolution. Each cycle might last a trillion years, say. Fourteen billion years marks the time since the last infusion of matter and radiation, but this is brief compared to the total age of the universe. Each cycle lasts about a trillion years and the number of cycles in the past may have been ten to the googol power or more!
Then, using the slow relaxation mechanisms considered previously, it becomes possible that the cosmological constant decreases steadily from one cycle to the next. Since the number of cycles is likely to be enormous, there is enough time for the cosmological constant to shrink by an exponential factor, even though the decrease over the course of any one cycle is too small to be undetectable. Because the evolution slows down as the cosmological constant decreases, this is the period when most of the cycles take place. There is no multiverse and there is nothing special about our region of space — we live in a typical region at a typical time.
Remarkably, this idea is scientifically testable. The picture makes explicit predictions about the distribution of primordial gravitational waves and variations in temperature and density. Also, if the cosmological constant is evolving at the slow rate suggested, then ongoing attempts to detect a temporal variation should find no change. So, we may enjoy speculating now about which dangerous ideas we prefer, but ultimately it is Nature that will decide if any of them is right. It is just a matter of time.