MARTIN REES: In the last few years the problem of understanding the ultra-early universe has come into focus. We now know the key properties of the universe—its density, its age, and its main constituents. Indeed the last three years will go down as specially remarkable in the annals of cosmology because just within these years we've pinned down the shape and contents of the universe, just as in earlier centuries the pioneer navigators determined the size of the earth and the layout of its continents. The challenge now is to explain how it got that way. The new physics is attempting to understand why it's expanding the way it is, and why it ended up with the content it has. We can trace its history back to about a micro-second after the putative 'big bang' that started it off, but what happened in that first, formative microsecond? The boisterous variety of ideas being discussed—branes, inflation, etc.—makes clear that the issues are fascinating, but also we're still a long way from the right answer. We're at the stage where all possibilities should be explored. It's worthwhile to consider the consequences of even the most flaky ideas, although the chance of any of them actually panning out in the long run is not very high.

In my own work, I try to be open to several ideas at once (even if they're incompatible) because I want to know the answer. If a phenomenon is puzzling, it's a good idea to explore all options: you'll thereby perhaps find new ways to discriminate among them, or else further study may reveal contradictions that rule some of them out. Obviously, the community collectively does that, but individual scientists fall into two categories. Some individuals aren't motivated to work on a theory unless (at least at the time) they feel pretty convinced it's likely to be correct—they put all their money on a particular horse. But other scientists (and I'm in this second category) are happy to spread their bets, and find the wish to clarify the issue itself a sufficient motivation.

I wouldn't claim to be a technical expert in any of the specific theories for the ultra early universe. It seems likely that extra dimensions of space are going to play a role; it's very good that the idea of inflation, which has dominated the field for 20 years, is now being generalized by other concepts that have come from people like Lisa Randall, Neil Turok, and Paul Steinhardt. It's important to explore all of these avenues.

The key goal, of course, is to develop a convincing, all-encompassing theory that describes the early universe and that makes testable predictions about the world today. If we had a theory that gave us a deeper and more specific understanding of the masses of electrons and protons, and of the forces governing them than the so-called 'standard model' does today, then that theory would gain credibility, and we'd take seriously its implications for the ultra-early universe . The hope is that one of the exotic new theories will make testable predictions either about the ordinary world of particles, or about the universe. For instance, some make distinctive predictions about the amount of gravitational radiation filling the universe. We can't yet measure this today but within ten years we might be able to do it. That's one way in which astronomical observations might be able to narrow down the range of options.

The easiest idea to understand conceptually is eternal inflation, which Guth advocates and on which Andrei Linde has done a great deal of detail work. This naturally gives rise to many Big Bangs. Whether or not those Big Bangs will be close replicas of each other, or whether the material in each of them would be governed by different laws is something we don't know. Eternal inflation may bypass the complications of extra dimensions and quantum gravity, because these are relegated to the infinite past.