It is hard to come to terms with the sheer scale of space: hundreds of billions of stars in our galaxy and, at a minimum, trillions of galaxies in the universe. But to a cosmologist there is something even more intriguing than the boggling numbers themselves, which is the question of how all these stars and galaxies were created over a period of 13.8 billion years. It’s the ultimate prehistoric adventure. Life cannot evolve without a planet, planets do not form without stars, stars must be cradled within galaxies, and galaxies would not exist without a richly structured universe to support them. Our origins are written in the sky, and we are just learning how to read them.
It once seemed that, for all its immensity, the cosmos could be understood through the application of a small number of rigid physical laws. Newton encapsulated this idea, showing how apples falling from trees and planetary orbits around our sun arise from the same force, gravity. This kind of radical unification of earthly and heavenly phenomena survives in modern teaching: all the innumerable molecules, atoms and subatomic particles in the universe are expected to obey the same set of laws. Most of the evidence suggests that this assumption holds true, so it should follow that perfecting our understanding of these laws will resolve any remaining questions about cosmic history.
Yet this is a logical fallacy. Even if we imagine that humanity will ultimately discover a “theory of everything” covering all individual particles and forces, that theory’s explanatory value for the universe as a whole is likely to be marginal. Over the course of the 20th century, even as particle physics revealed the secrets of atoms, it became clear that behaviour at the macro level cannot be understood by focusing exclusively on individual objects.
Social insects here on Earth provide a helpful example. Army ants, for instance, swarm to locate colonies of smaller prey, which they then devour. While swarming, they perform extraordinary feats of cooperation, using their bodies to smooth out the terrain, or even to build bridges over uneven ground.
To human eyes, the collective behaviour of the ants might suggest that an executive within the nest formulates strategies to reach prey efficiently, but there is no such decision-maker. There are just lone ants, following simple unchanging rules, such as joining an ant-bridge if there are many individuals pushing behind, and leaving the structure if no others crawl over. The sophistication emerges from the sheer number of individuals following these rules. As the physicist Philip W Anderson put it: “More is different.”
The solar system, seemingly the epitome of clockwork predictability, has an uncertain long-term future for this reason. In isolation, a single planet around a single star would orbit indefinitely but in reality there are multiple planets and they each tug, albeit very subtly, on the others. Over time a series of tiny nudges can produce a major effect, one that takes an inordinate amount of calculation to predict.
To an extent, computers can take on this challenge, simulating the collective result by adding the individual influences using fast and reliable arithmetic. The problem is that simulations disagree with each other. Some predict that the solar system is stable despite the continual nudging, while others suggest that within a few billion years Mercury might be coaxed on to a collision course with Venus, or even ejected into deep space.
Solar system simulations disagree because no calculation can perfectly account for all the influences, and even the tiniest disagreement about the individual nudges leads eventually to a completely different outcome. It is an example of the phenomenon known as chaos, and it is simultaneously exciting and worrying. Exciting, because it shows that planetary systems can exhibit much richer behaviours than the cold, lifeless law of gravity might suggest. Worrying, because if even the solar system is chaotic and unpredictable, we might fret that attempting to understand the broader universe is a doomed enterprise.
Consider galaxies, on average tens of millions of times larger in extent than the solar system, and lavishly varied in their shapes, colours and sizes. Understanding how galaxies came to be so diverse requires, at a minimum, for us to know how and where the stars formed within them. However star formation is a chaotic process in which diffuse clouds of hydrogen and helium slowly condense under gravity, and no computer is anywhere near able to track all the required atoms (there are around 1057 in our sun alone). Even if the computation were feasible, chaos would magnify exponentially the tiniest uncertainties, forbidding us from obtaining a definitive answer. If we were strict in sticking to traditional laws of physics as an explanation for galaxies, here is the end of the road.
To fit inside computers, a simulation of a galaxy’s formation has to lump together vast numbers of molecules, describing how they move en masse, push on each other, transport energy, react to light and radiation, and so on, all without explicit reference to the innumerable individuals within. This requires us to be creative, finding ways to describe the essence of many different processes, allowing for a range of outcomes without obsessing over the detail, which is anyway unknowable. Our simulations necessarily rely on extrapolations, compromises and all-out speculations developed by experts. The uncertain parts cover not just stars, but black holes, magnetic fields, cosmic rays and the still-to-be-understood “dark matter” and “dark energy” that seemingly govern the overall structure of the universe.
This will never result in a literal digital replica of the universe that we inhabit. Such a recreation is just as impossible as a precise forecast for the future of the solar system. But simulations based even on loose descriptions and best guesses can act as a guide, suggesting how galaxies may have evolved over time, enabling us to interpret results from increasingly sophisticated telescopes, guiding us on how to learn more.
Ultimately, galaxies are less like machines, and more like animals – loosely understandable, rewarding to study, but only partially predictable. Accepting this requires a shift in perspective, but it makes our vision of the universe all the richer.
Prof Andrew Pontzen is the author of The Universe in a Box: A New Cosmic History (Jonathan Cape).
This article was amended on 24 July 2023. Due to a typesetting error, an earlier version said that the number of atoms in the sun was “around 1057”, rather than 1057.
Further reading
The End of Everything by Katie Mack (Penguin, £9.99)
Simulating the Cosmos by Romeel Davé (Reaktion, £15.95)
The Disordered Cosmos by Chanda Prescod-Weinstein (PublicAffairs, £13.99)
