Britain's most eminent astronomer confronts the question of whether there will ever be an unsolvable problemby Martin Rees / November 13, 2018 / Leave a comment
Published in December 2018 issue of Prospect Magazine
The Sun formed 4.5bn years ago, but it’s got around six billion years more before its fuel runs out. It will then flare up, engulfing the inner planets. Any creatures witnessing the Sun’s demise won’t be human—they’ll be as different from us as we are from a bug. Post-human evolution could be as prolonged as the Darwinian evolution that has led to us.
And evolution is now accelerating; it can happen via intelligent design on a technological time-scale, operating far faster than natural selection and driven by advances in genetics and in artificial intelligence (AI). The long-term future probably lies with electronic rather than organic life.
In cosmological or Darwinian terms, a millennium is but an instant. So let us fast forward not for a few centuries or millennia, but for an astronomical timescale millions of times longer than that. The stellar births and deaths in our galaxy will gradually proceed more slowly, until jolted by the environmental shock of an impact with the Andromeda Galaxy, maybe four billion years hence. The debris of our galaxy, Andromeda and their smaller companions—which now make up what is called the Local Group—will thereafter aggregate into one amorphous swarm of stars. Many billions of years after that, gravitational attraction will be overwhelmed by a mysterious force latent in empty space that pushes galaxies apart from each other. Galaxies accelerate away and disappear over a horizon. All that will be left in view, after 100bn years, will be the dead and dying stars of our Local Group, which could continue for trillions of years. Against the darkening background, sub-atomic particles such as protons may decay, dark matter particles annihilate and black holes evaporate—and then silence.
As we attempt to grapple with this bleak post-human future, we must also confront the question of what humans can hope to understand. Parts of the physical world are understood. They can be observed and described by theories—but much of it cannot. Human observation bumps up against stark limits. Human reasoning is not limitless either, but it does allow us to think through what might in principle be “over the horizon.”
The By-Laws of Physics
For now, we can only see a finite number of galaxies. That’s essentially because there’s a horizon, a shell around us, delineating the greatest distance from which light can reach us. But that shell has no more physical significance than the circle that delineates your horizon if you’re in the middle of the ocean. The volume of space-time within range of our telescopes—what astronomers have traditionally called the universe—is only a tiny fraction of the aftermath of the big bang. We’d expect far more galaxies located beyond the horizon, unobservable, each of which (along with any intelligences it hosts) will evolve rather like our own.
It’s a familiar theorem that if enough monkeys were given enough time to hit random keys on a typewriter, they would eventually type out the complete works of Shakespeare—and all other books, along with every conceivable string of gobbledygook. But when we throw dice (unless they are loaded) we wouldn’t expect to get more than 100 sixes in a row even if we went on for a billion years.
With the monkeys, the number of failures that would precede eventual success is not in the millions—it is a number with perhaps 10m digits. For comparison, the number of atoms in the visible universe has only 80 digits. If all the planets in our galaxy were crawling with monkeys, typing since the first planets formed, the best they would have done is a sonnet (their output would include short coherent stretches from all the world’s literatures, but no single complete work). It’s immensely improbable that one of them would have chanced upon a run of orderly letters as long as a book anywhere within the observable universe.
The more the universe stretches out in time and space, the more likely it is that a certain thing will have occurred. From which it follows that, if the universe stretches far enough, everything could happen—somewhere far beyond our horizon there could be a replica of Earth. This requires space to be very big—described by a number not merely with a million (or “10 to the power 6”) digits, but with 10 to the power of 100 digits.
Given enough space and time, and going far beyond the range of observation, all conceivable chains of events could be played out somewhere—including replicas of ourselves, taking all possible choices. Whenever a choice has to be made, one of the replicas will take each option. If you have a fraught decision to make, it may be a consolation that, somewhere far away you have an avatar who has made the opposite one.
All this could be encompassed within the aftermath of our big bang, which could extend over a stupendous volume. But that’s not all. If we conceive the universe as the entire aftermath of our big bang, it could still be just one island—one patch of space and time, in a perhaps infinite archipelago. There may have been many big bangs, not just one. Each constituent of this multiverse could have cooled down differently, maybe ending up governed by different laws. Just as Earth is a very special planet among zillions of others, so—on a far grander scale—could our big bang have been a special one. In this hugely expanded cosmic perspective, the laws of Einstein and the quantum could be mere parochial by-laws governing our cosmic patch.
At the cutting edge of physics, seemingly foundational concepts like space and time have long been disrupted. But it could be that our current concept of physical reality could be as constricted, in relation to the whole, as the perspective of the Earth available to a plankton whose universe is a spoonful of water.
So a challenge for 21st-century science is to answer two questions. First, are there many big bangs rather than just one? Second—and this is even more interesting—if there are many, are they all governed by the same physics?
Before the beginning
If we’re in a multiverse, it would imply a fourth and grandest Copernican revolution; the first was when Copernicus showed that the planets revolved around the Sun, the second was the realisation that there are billions of planetary systems in our galaxy; and the third, that there are billions of galaxies in our observable universe. But now that’s not all. The entire panorama that astronomers could even in principle observe could themselves be just a tiny part of the aftermath of our big bang, which is itself just one bang among a perhaps infinite ensemble.
“If you have a choice to make, it may be that somewhere far away you have an avatar who made the opposite choice”
Fifty years ago, we weren’t sure whether there had been a big bang. My Cambridge mentor Fred Hoyle contested the concept, favouring a “steady state” cosmos that was eternal and unchanging. (He was never fully converted—in his later years he compromised on something called a “steady bang.”) Now we have enough evidence to delineate cosmic history back to the ultradense first nanosecond—with as much confidence as a geologist inferring the early history of Earth.
In 50 more years, we might have a unified physical theory, corroborated by experiment and observation, that is broad enough to describe what happened in the first trillionth of a trillionth of a trillionth of a second. If that future theory were to predict multiple big bangs we should take that prediction seriously, even though it can’t be directly verified. After all, we give credence to what Einstein’s theory tells us about the unobservable insides of black holes, because the theory has survived many tests in domains we can observe.
My 1997 book, Before the Beginning, speculated about a multiverse. Its arguments were partly motivated by the seemingly biophilic character of our universe—it seemed a bit good to be true that the one and only universe going should just happen to be so neatly fine-tuned to allowing life as we know it. I’ve followed the development of these admittedly speculative ideas closely ever since.
Some years ago, I was on a panel at Stanford University where we were asked by the chairman: “On the scale, would you bet your goldfish, your dog, or your life, how confident are you about the multiverse concept?” I said that I was nearly at the dog level. Andrei Linde, a Russian cosmologist who had spent 25 years promoting a theory of “eternal inflation” said he’d almost bet his life. Later, on being told this, the eminent theorist Steven Weinberg said he’d happily bet Martin Rees’s dog and Linde’s life.
Linde, my dog, and I will all be dead before this is settled. It’s highly speculative. But it’s not metaphysics: it is exciting science. And it may be true. A feature of all science—and certainly astronomy—is that as the frontiers of our knowledge are extended, new mysteries, just beyond the frontiers, come into sharper focus. There will, at every stage, be unknown unknowns. (Donald Rumsfeld was mocked for saying this in a different context—but he was right, and it might have been better for the world had he become a philosopher.)
The twilight zone
But there is a deeper question. Are there things that we’ll never know, because they are beyond the power of human minds to grasp? Are our brains matched to an understanding of all key features of reality? We should actually marvel at how much we have understood. Human intuition evolved to cope with the everyday phenomena that our remote ancestors encountered on the African savanna. Our brains haven’t changed much since that time, so it is remarkable that they can grasp the counterintuitive behaviours of the quantum world and the cosmos.
Iconjectured earlier that answers to many current mysteries will come into focus in the coming decades. But maybe not all of them; some key features of reality may be beyond our conceptual grasp. We may sometime hit the buffers; there may be phenomena, crucial to our long-term destiny and to a full understanding of physical reality, that we are not aware of, any more than a monkey comprehends the nature of stars and galaxies. If aliens exist, some may have brains that structure their consciousness in a fashion that we can’t conceive and that have a quite different perception of reality.
We are already being aided by computational power. In the virtual world inside a computer, astronomers can mimic galaxy formation, or crash another planet into the Earth to see if that’s how the Moon might have formed; meteorologists can simulate the atmosphere, to forecast the weather or long-term climatic trends; brain scientists can simulate how neurons interact. Just as video games get more elaborate, so, as computing power grows, these virtual experiments become more realistic and useful.
There is no reason why computers can’t make discoveries that have eluded unaided human brains—indeed, it’s already on the cusp of happening. For example, some substances are superconductors: perfect conductors of electricity when cooled to very low temperatures. There is a continuing quest to find the recipe for a superconductor that works at ordinary room temperatures (the highest superconducting temperature achieved so far is about minus 135 degrees Celsius at normal pressures). This would allow lossless transcontinental transmission of electricity, and efficient Maglev trains. Such quests traditionally involve a lot of trial and error. But it’s becoming possible to calculate the properties of materials, and to do this so fast that millions of alternatives can be computed, far more quickly than actual experiments could be performed.
“Answers to many mysteries will come into focus, but some key features of reality may be beyond our grasp”
Suppose that a machine came up with a unique and successful recipe for a super-conductor. It would have achieved something that would earn a scientist a Nobel prize. It would have behaved as though it had insight and imagination within its specialised universe. Likewise, searches for the optimal chemical composition for new drugs will increasingly be done by computers rather than by real experiments. Equally important is the capability to discern small trends or correlations by crunching huge data sets. In genetics, machines that are fast enough to scan large samples of genomes to identify small correlations will be able to identify the combinations of genes that are responsible for qualities like intelligence and height.
Any process is in principle computable. However, being able to compute something is not the same as having an insightful comprehension of it. I’m still doubtful that human brains can eventually understand everything. We need to remain open-minded—despite our impressive and in some cases accelerating efforts—to the possibility that some fundamental truths about nature could be too complex for unaided human brains to grasp fully. Perhaps we’ll never understand the mystery of our brains themselves—how atoms can assemble into grey matter that can become aware of itself and ponder its origins. Or perhaps any universe complicated enough to have allowed our emergence is for just that reason too complicated for our minds to understand.
Organic post-humans or intelligent machines might be able to get further than we can in future. But we would be too anthropocentric if we believed that a full understanding of physical reality is within humanity’s grasp, and that no enigmas will remain to challenge our post-human descendants.
We ourselves, our Earth, and the entire cosmic panorama that our telescopes reveal, are the aftermath of a Big Bang 13.8bn years ago. Far longer timespans lie ahead. Nonetheless, the prevailing physical laws predict eventual extinction. But those who find this gloomy may be comforted that other cosmoses may exist where the physics is different. Most of these would be sterile, but some may offer more potential than our own. There could be an overarching theory that allows this multiverse—perhaps beyond our imagination. I’d wager my dog against it, but perhaps not my planet.