Throw a pebble into a lake and it, obligingly, sinks. Smash particles together and they fall apart in certain patterns. Flick a switch and let there be light. Reality, for all its glory and cosmic drama, seems to operate in a consistent, predictable fashion.

Physicists like me often put this happy fact down to what we call the laws of nature. These laws apply in the same way everywhere: the same force of gravity that bends the light from distant stars also keeps your feet on the ground. Moreover, they don’t change – they are valid from the big bang to happily ever after. All this is assumed in physics to the point that few ever question it.

To be fair, there are good reasons why. Bluntly asking “Where did the laws of physics come from?” can feel like smuggling philosophy into the lab. It can also lead us into risky territory – we might end up expecting stars to burn differently or atoms to fall apart.

But I believe the origin of the laws of nature is a question we can’t ignore, one I’ve been pondering for the past few years. Many previous attempts to explain how these laws arose have foundered because they ended up introducing deeper “meta-laws” along the way. Finally, though, I think I have something better: a framework that explains how the rules of science varied wildly at the start of the universe before settling into what we see today. If I’m right, then the “laws of nature” may not be fundamental at all.

What do I mean by the laws of nature? I am talking about the key equations of physics, like Isaac Newton’s laws of gravity; James Clerk Maxwell’s equations, which govern electricity and magnetism; and Albert Einstein’s field equations, which explain the workings of space-time. These come served with what we call fundamental constants, numbers that are embedded in the equations and describe properties of the universe that we’ve observed – things like the strength of gravity or the charge of an electron. The equations and constants aren’t just convenient summaries of reality, they are the load-bearing beams that support the entire theoretical edifice of physics.

However, if we are going to ask where the laws of nature come from, we must entertain a more unsettling possibility: that once upon a time, there were no laws at all. There was a period before particles, before geometry, before even the notion of time. Reality would have been a chaotic mess.

The visionary physicist John Wheeler called this state of Wild West lawlessness “higgledy-piggledy”. This wasn’t a throwaway line. When I first encountered Wheeler’s remark, my English didn’t quite extend to higgledy-piggledy, so I looked it up. One synonym was “helter-skelter”, which I associated with the Beatles song with that title. That felt about right: a cacophonous universe – guitars badly out of tune, no agreement on rhythm or key.

At the time, I was a cosmologist surrounded by conformist bandwagons – inflation, dark matter and dark energy, which theorists package into the so-called lambda-CDM model of our universe. This doesn’t seek to explain why we have the laws that we do, but says they simply just are and have always been. Perhaps for no other reason than in reaction to this, I was trying to carve out alternatives, like cosmological theories that allow the speed of light to vary in the early moments of the universe. The wilder the idea was, the better.

Wheeler’s higgledy-piggledy idea was intoxicating to me, and immediately alarming. If the laws of physics themselves can change, even chaotically, then what anchors reality at all? Is this even a question that physics can hope to answer, or is this just philosophy, dressed up in a lab coat?

Most physicists prefer not to ask. But there is a sense in which we don’t have that luxury. Physics, at its core, is the attempt to explain why the universe is the way it is, rather than some other way. And that project feels unfinished if we simply take the laws themselves as given. Push this line of questioning far enough, and you are led somewhere more radical: a time when there were no laws at all.

The fundamental laws of nature

Most physicists feel a Pavlovian electric shock when confronted with the concept of a lawless universe. But there are good reasons for this. Part of it, as I’ve already said, is that these laws are integral to the structure of modern physics. Hence, the instinct that they should be eternal, perfect, immutable. I also suspect that some of this dates back to a time when science and religion were deeply intertwined, and ideas about natural laws echoed divine law: timeless, universal and not open to negotiation. Even after science secularised, the reverence remained.

But there is something else at play here that is even more important: symmetry. In geometry, a shape has symmetry if you can carry out some operation, like a rotation, and the shape looks the same. Physics has a similar kind of symmetry. Perform an experiment here or there, today or tomorrow, facing north or south, and the outcome is the same – or at least, we assume it is.

The implications of that assumption are massive. They were first unveiled in 1918 by Emmy Noether, a mathematician whose work permanently rewired theoretical physics, despite her career being obstructed for years by institutional sexism. Noether showed that every continuous symmetry – a symmetry that holds under a smooth shift, such as moving through space or time – comes with a conserved quantity. She showed, mathematically, that if the laws of physics are the same everywhere, it logically means that momentum has to be conserved, meaning the total amount of momentum shared between all objects – for example, between balls in a game of pool – doesn’t change. The conserved quantities can be exchanged between objects, say in a collision, but the sum remains fixed.

One symmetry, however, plays a special role. If the laws of physics are the same from one moment to the next – they obey what we call time-translation invariance – this implies energy can’t be created or destroyed. On its own, that’s fine. The trouble is, the reverse is also true. If you believe that the laws of physics can change with time, then energy conservation is violated. Break one and the other bleeds. This is a big problem for many of my colleagues because, as tenets of physics go, the conservation of energy is just about the most sacred.

Taming a lawless universe

Yet some physicists were undaunted. The forefather of this rebellion was Paul Dirac, best known for unifying quantum mechanics and special relativity. Dirac was famously eccentric and, true to form, he wrote one of the most radical papers of his career while on his honeymoon in Brighton, UK, in 1937.

In the paper, Dirac made an audacious proposal that the constants of nature, those important numbers that appear in our fundamental laws, actually reflect the age of the universe. If this were true, then the constants aren’t “constant” at all and instead themselves evolve in time. The laws of physics, in this view, are no longer timeless.

Decades later, my friend Lee Smolin pushed the idea of evolving laws much further. Lee’s proposal, known as cosmological natural selection, begins with a simple unorthodoxy: black holes might not be cosmic dead ends. Instead, each black hole gives birth to a new expanding universe on the other side of its horizon, a kind of cosmic offspring. This idea isn’t pure fantasy. General relativity allows for extreme rearrangements of space-time inside black holes, and some solutions can be read as bridges to new regions.

Crucially, Smolin suggested that the laws and constants aren’t copied perfectly in this process. Each new universe inherits slightly mutated constants, tiny changes in particle masses or force strengths. Some universes are better than others at seeding black holes, and therefore better at passing on their constants. Over many generations, universes with “successful” constants come to dominate. Strange as it may first sound, this idea still stops short of Wheeler’s total higgledy-piggledy. In Wheeler’s vision, it isn’t simply that the constants plugged into the equations evolve. The equations themselves are in flux, too, if it even makes sense to speak of equations at all.

But wait, what about that major snag that had been staring at us since Noether’s day: the idea that if you allow the laws of nature to change, you give up on energy conservation? For a long time, this was taken as a reason not to pursue evolving laws, but over the past two years, I eventually realised that it was actually the complete opposite. I saw a massive window of opportunity.

How to create a universe

Here’s the thing: in some circumstances, the fact that energy can’t be created or destroyed is troublesome. Take the big bang itself. Currently, physicists are forced to believe that all the matter and energy we see today had to be present at the start of the universe, which forces it into a point of infinite density. But we don’t know how to interpret what this really means and infinities break our equations. If we relax that condition, however, then the creation of matter in the universe can become not an event, but a process: something extended, contingent, fallible.

In principle, that overcomes a big problem. But there is another side to this story. If matter and energy can be created, they can also be destroyed. The same mechanism that giveth with one hand taketh with the other. So, we need some way of explaining how this process leaves us with something – the universe we see around us – rather than nothing.

In a paper published last year, my PhD student Paolo Bassani and I borrowed tools from evolutionary biology and financial mathematics, disciplines that study systems that never sit still and allow genuine randomness within evolving laws. In our picture, nothing is dependable in the earliest phase of the universe before stable laws emerged. Constants fluctuate wildly. Conservation laws rudely fail. Matter is created, but also destroyed, at random. Positive energy is as likely as negative energy, creation as probable as annihilation. Any matter you gain from one throw of the dice may be lost on the very next throw.

The universe is effectively gambling, maybe building up a surplus on a good run, before losing it just as quickly. As long as the laws keep mutating, whatever gains are made in the form of matter are never secure. A single bad fluctuation can wipe everything out. To get anything that lasts, you need a way for the process to stop and for whatever gains the universe has made to be locked in.

Fortunately, random systems like our chaotic universe – but also like many models of genetic mutation, stock markets or chemical reactions – have a built-in feature whereby they can stumble into configurations that they can’t escape from, known as an absorbing state. Take a mutation that has spread through an entire population, a company going bankrupt and vanishing from the market or a chemical reaction that has run to completion. In each case, the system has reached a state from which the dynamics offer no further moves, so the random process of change effectively grinds to a halt.

In our case, the absorbing state is the built-in point in chaotic evolution where the laws are forced to crystallise, their random mutation effectively switching off. With that, the dual hand of creation and destruction is turned off as well. Some universes arrive there empty-handed or in severe debt. Others arrive after a lucky run. And since the dual hand of creation and destruction has now been turned off, they get to keep their gains. We are those gains.

In this picture, order isn’t selected because it is beautiful or true, but because it lasts and holds on to what it has built. Seen this way, several long-standing puzzles, like why our constants take on the values they currently do, look less mystical. The values of the constants need not be unique, only compatible with longevity.

Testing this will be difficult, but far from absurd. The cleanest place to look is in ultra-precise measurements of time. Atomic clocks – incredibly reliable devices that use the vibration of atoms to keep time – are now so stable that they can detect extraordinarily small drifts in fundamental constants. Because different clocks depend on those constants in different ways, any variation would cause them to slowly fall out of sync – a telltale signal that the laws themselves are shifting. So far, current measurements are so precise that any present-day effects must be tiny. But that is precisely what makes this a promising test. With clocks this precise, even a minute residual jitter in the laws would have nowhere to hide.

All this work takes me back to a broader lesson I learned back in 2003 when I shared a house for a couple of years with Lee Smolin. At the time, we were both researchers at the fledgling Perimeter Institute for Theoretical Physics in Waterloo, Canada. Lee kept a remarkable library, with philosophical classics and not-so-classics stacked everywhere.

Before I met Lee, I was a fairly standard-issue scientific philistine when it came to philosophical matters, believing that physicists had no business doing metaphysics – that worrying about the nature of laws or explanations was someone else’s problem.

But it was in Lee’s library that I was first exposed, among much else, to the work of philosopher Paul Feyerabend. He advocated for a demolition of dogmatic science and a pluralism – not only in human culture, but in the methods and theories of science itself. This spirit informed my scientific discussions with Lee, in which we often changed sides at random, simply to see where the reasoning led. This isn’t a weakness but a strength. Proposing scientific theories isn’t like supporting a football team: unruly polygamy is acceptable.

Seen this way, the universe itself may follow a similarly Feyerabendian pluralism, chaotically trying out all theories, backing every football team at random, until some prove stable enough to endure. In the process, it discovers what works, not because it is elegant or ordained, but because it survives long enough to be mistaken for destiny.