Abstract: Cosmology is an exploratory quest; many ideas are tentative, and most will surely prove wrong – but it is speculative science, to be distinguished from metaphysics.
The laws of physics established locally seem to apply everywhere we can observe – and also throughout cosmic history back to 1 second, when helium and deuterium form – indeed we can be confident back to a nanosecond. That is when each particle had about 50 Gev of energy – as much as can be achieved in the biggest particle accelerators – and the entire visible universe was squeezed to the size of our solar system. Were there not this uniformity, we would have made little progress in astrophysics.
But fundamental questions like “where did the fluctuations come from?” and “why the early universe contained the actual mix we observe of protons, photons and dark matter?” take us back to the even briefer instants when our visible universe was hugely more compressed still – if inflation theories are correct, down to microscopic size.
Lab experiments offer no direct guide to the relevant physics at these extreme compressions – though inflation is a consequence of specific (though still conjectural) physics and here is another basic question:
How much space is there altogether? How large is physical reality? We can, even in principle, only observe a finite volume – a finite number of galaxies. That is essentially because, like on the ocean, there is a horizon. Unless we take a very non-Copernican view, there are galaxies that are already beyond the horizon – and in our accelerating universe we can never (even in principle) observe them.
There is no perceptible gradient across the visible universe and that suggests it extends hundreds of times further than our horizon. But that is just a minimum. If it stretched far enough, then all combinatorial possibilities could be repeated. Far beyond the horizon, we could all have avatars.
Be that as it may, even conservative astronomers are confident that the volume of space-time we can ever observe even in principle is only a tiny fraction of the aftermath of our big bang.
And there is something else. Some scenarios like Andre Linde’s “eternal inflation” suggest that the aftermath of “our” big bang could be just one island of space-time in a vast archipelago – one big bang among many.
A challenge for 21st century physics is to answer two questions: first, are there many “big bangs” rather than just one? Second, if there are many, are they all governed by the same physics or not?
Many string theorists argue that there could be a huge number of different vacuum states – different vacuum energy, different microphysics. Other big bangs may, when they cool down, be governed by different laws. What we call “laws of nature” may in this grander perspective be local bylaws governing our cosmic patch.
If physical reality is like this, then there is a real motivation for theorists to model “counterfactual” universes, with different gravity, different microphysics, and so forth, to explore what range of parameters would allow complexity to emerge – and which would lead to sterile or “stillborn” cosmos. This is what is called anthropic selection.
Some do not like the concept of such a vast and varied cosmos, because it would render the hope for neat explanations of the fundamental physical numbers as vain as Kepler's numerological quest to relate planetary orbits to nested platonic solids.
But our preferences are irrelevant to the way physical reality actually is – so we should surely be open-minded to the possibility of a 4th and grandest Copernican revolution – we have had the Copernican revolution itself, then the realization that there are zillions of planetary systems in our galaxy; then that there are zillions of galaxies in our observable universe. But we would then realize that our entire observable domain is a tiny part of a far larger and possibly diverse ensemble. We would live not in a typical part of the multiverse – but a typical domain in the subset that allows complexity to develop.
To address that question we have to know the probability distribution – and the correlations between them. That is a can of worms that we cannot yet open – and will have to await huge theoretical advances.
I must re-emphasise that we do not know if there are other big bangs. But they are not just metaphysics. We might one day have reasons to believe that they exist. Specifically, if we had a theory that described physics under the extreme conditions of the ultra-early big bang – and if that theory had been corroborated in other ways [by for instance deriving the unexplained parameters in the ‘standard model’ of particle physics] then if it predicts multiple big bangs we should take that prediction seriously. And there would be no reason to assign them different epistemological status from the un-observable domains beyond our horizon in our own big bang.
We cannot observe the interior of black holes, but we believe what Einstein’s relativity says about what happens there, because his theory has gained credibility by agreeing with data in many contexts that we can observe. Likewise, we would believe the prediction of other big bangs if this was the consequence of applying to the early universe a “battle tested” theory that correctly predicted the values of some basic physical constants of our low-energy world.
It is now an open question – but how would you bet? About 15 years ago I was on a panel at Stanford where we were asked by the chairman, Bob Kirshner, how seriously we took the multiverse concept – on the scale “would you bet your goldfish, your dog, or your life”. I said I was nearly at the dog level. Andrei Linde, who had spent 25 years promoting “eternal inflation” said he had almost bet his life. Later, on being told this, Steven Weinberg said he would happily bet Martin Rees’s dog and Andrei Linde’s life.
Andrei Linde, my dog, and I will surely all be dead before this is settled.
Indeed, some of the crucial physics of unified theories could be just too difficult for human brains – just as quantum theory is too difficult for monkeys.
It is conceivable that machine intelligence could explore the 10-D geometrical intricacies of some string theories, and spew out, for instance, the correct mass of the electron or other key parameters of the standard model. We would then have confidence in the theory and take its other predictions seriously. But we would never have the “aha” insight moment that is the greatest satisfaction for a theorist. Even worse, physical reality at its deepest level could be so profound that its elucidation and comprehension would have to await emergence of a posthuman species, or massive advances in AI.
But let me close with a plug for our field of astronomy. It is the grandest of the environmental sciences. Thanks to improved instruments, on the ground and in space, this subject is becoming broader. For instance, we know that most stars are orbited by retinues of planets. There are billions of planets in the Milky Way that could be abodes of life – but are they?
Astronomy is also a “fundamental” science: to understand the very beginning of our expanding universe will require advances in physics that may take us further from our intuitive concepts than quantum theory and relativity already do – indeed we must be open-minded that these concepts may be too deep for human brains to grasp.
So I end with an admonition from Hubble: “Only when empirical resources are exhausted should we enter the dreamy realm of speculation”.