麻豆传媒

IMAGINE you are marooned in a vast, featureless expanse. Everywhere you look, no matter how far you travel, it all looks the same. It sounds like a disturbing dream. Believe it or not, though, this is the universe you live in. If you zoom out far enough, past nearby stars, through the Milky Way to clusters of galaxies and the filament-like structures that connect them, and then you keep going, eventually, everything starts to look smooth and uniform wherever you glance.

Or does it? This idea that, on the grandest scales, the cosmos looks largely the same regardless of position or direction is called the cosmological principle, and it underpins our best theory of how the universe evolved. For cosmologists, it is gospel. But some heretics are now calling the principle into question, pointing to fresh evidence that even at its largest scales, the cosmos is not only lumpy, but fundamentally off-kilter.

If they are right, it would upend cosmology. We would have to start our description of the universe鈥檚 evolution from scratch 鈥� and possibly even admit there can be no single model capable of describing it right up to today.

If they are right. Most cosmologists are a long way from convinced. Similar challenges have been seen off before, they shrug. And yet there is a reckoning on the horizon because, one way or another, the cosmological principle 鈥� for so long considered untouchable 鈥� is finally going to run the empirical gauntlet.

The notion that the universe is the same everywhere in all directions grew out of another essential tenet of cosmology: Nicolaus Copernicus鈥檚 argument, made in the 16th century, that Earth doesn鈥檛 occupy a special place in the universe, aka the Copernican principle. But the cosmological principle didn鈥檛 earn its current exalted status until the 1920s, when it made it possible to extract a workable model of the universe from Albert Einstein鈥榮 new theory of gravity 鈥� general relativity.

General relativity is as complex as it is beautiful. Ten intertwined equations balance matter and energy on one side and the warping of space-time on the other. The only way to solve it is to make sweeping assumptions about how matter and energy are distributed and how space-time warps. By assuming the universe is the same everywhere, or homogeneous, and looks the same in every direction, a feature known as isotropy, physicists were able to boil down Einstein鈥檚 equations to extract a simple, evolving universe that matched observations.

This 鈥淔riedmann鈥揕ema卯tre鈥揜obertson鈥揥alker鈥� (FLRW) solution enshrined the cosmological principle, and it remains the foundation of the standard model of cosmology 鈥� our best theory of how the universe evolved. Starting with the big bang, this perfect FLRW cosmos expands symmetrically like a balloon filling with air. Stars, galaxies and clusters then form by allowing small deviations from what remains an otherwise smooth and uniform cosmos.

The standard model works. When we take a detailed map of light from the very early universe, known as the cosmic microwave background (CMB), and extrapolate forward using the model, we get pretty much exactly what we observe today. 鈥淭he same bumps and wiggles that you see in the CMB are present in the large-scale distribution of galaxies,鈥� says at Princeton University, who developed the standard model in the 1970s and 80s. 鈥淭he standard theory works far better than I ever expected. But I never thought of it as a final theory; it was a placeholder.鈥�

The realisation that galaxies seem to be held together by some other kind of invisible matter, known as dark matter, and then the discovery that the universe鈥檚 expansion is accelerating, since attributed to dark energy, meant cosmologists had to add a couple of things to their recipe 鈥� freehand additions, as it were. 鈥淲e just say ad hoc: 鈥淟et there be dark energy! Let there be dark matter!鈥� says Peebles.

Like Peebles, most cosmologists would agree that the standard model is due an upgrade. For starters, we still don鈥檛 have the slightest clue what these mysterious spectres thrown into the pot to make it work are. And even when we incorporate dark energy into the standard model, direct measurements of stars and galaxies tell us that the universe is expanding faster than it should be according to the model.

Another set of observations suggests we may need to do more than just tinker. For years, a small cohort of cosmologists has been pointing out that the vast, bubble-shaped voids around which galaxies are strung in filaments and clusters make the universe look suspiciously inhomogeneous. The same is true of vast walls made of galaxies and galactic superclusters. Then, last year, the discovery of a giant chain of galaxies stretching 3 billion light years across the sky, known as the Giant Arc, gave fresh impetus to questions about the cosmological principle on which the standard model rests.

The argument isn鈥檛 only that such outliers provide more evidence against homogeneity on the biggest scales. It is also that the gravitational pull of such giant structures is what creates streams of matter that drift like the breeze, known as 鈥渂ulk flows鈥�. These can travel at hundreds of kilometres a second and stretch as far as a billion light years. 鈥淭here have also been reports of extreme bulk flows, much larger and much faster, but they have not been confirmed,鈥� says at the Aristotle University of Thessaloniki in Greece, who is building 鈥渢ilted鈥� cosmological models that include such fast-moving streams of matter.

Tsagas says we are in a bulk flow. We live on Earth, which orbits the sun, which orbits the Milky Way, which is pulled towards Andromeda, the nearest large galaxy to Earth, as well as towards clusters of galaxies like Virgo, superclusters like the Great Attractor and perhaps even lumps of matter beyond the horizon of what we can see. Our perspective is tilted relative to the symmetric expansion of the universe as a whole.

As a result, says Tsagas, we could end up with a false impression regarding dark energy. We see the universe鈥檚 expansion as accelerating, and yet its expansion could just as well be decelerating overall. Think of driving along a motorway where all the cars are travelling at the same speed when, suddenly, your engine goes kaput and you start to slow down. From your point of view, it looks like the other cars are accelerating away. In a similar manner, if the bulk flow we are in is decelerating, then the surrounding universe would seem to be accelerating 鈥� creating the illusion of a mysterious expansive force.

That would be a big deal, of course. But at the University of Oxford wants us to go further still. He claims to have found evidence that the entire universe is fundamentally skewed 鈥� and therefore that the cosmological principle is broken.

Again, it starts with the CMB, which more or less glows with the same temperature in every direction. Not quite, though. It appears lopsided to us, mainly because Earth is screeching around the Milky Way at hundreds of kilometres a second. Just as the frequency of an ambulance siren is distorted as it speeds past you, the CMB is distorted by our motion in the universe.

In the 1980s, , now at the University of Cape Town in South Africa, came up with a way to use this lopsidedness to check if the standard model obeys the cosmological principle. The idea was that we should see exactly the same distortion in very distant galaxies because, like the CMB, distant galaxies act like a fixed background against which we can measure our own motion. 鈥淚f those two don鈥檛 match up very precisely, then your standard FLRW model is in deep trouble,鈥� says Ellis.

Getting good data on very distant galaxies is tough, as they are so faint and difficult to distinguish amid astronomical objects closer to home. Astrophysicists were eventually able to perform the test in 2002 thanks to a large catalogue of galaxies, called the NVSS, gathered by the Very Large Array telescope in New Mexico. This showed that our motion through the universe agreed fairly well with the CMB. And yet the relatively small number of galaxies, and the ambiguity in how far away they actually are, left room to speculate.

鈥淲e realised that a better catalogue was required,鈥� says Sarkar, who in 2019 teamed up with Nathan Secrest, an astronomer at the US Naval Observatory in Washington DC. Secrest offered Sarkar and his colleagues a catalogue of 1.4 million quasars, known as catWISE, gathered by the WISE space telescope. Quasars are very bright jets of light powered by supermassive black holes in the centre of galaxies. Compared with the NVSS, catWISE has many more sources spread across the whole sky rather than just a portion of it 鈥� important if you are trying to assess how skewed the universe is overall.

Last year, Sarkar, Secrest and their colleagues reported a distortion in the distribution of quasars that was in a similar direction as the CMB鈥檚 skew, but . The implication is that instead of just bulk flows that drift across large parts of the universe, the entire universe may be drifting too. In other words, according to Sarkar, the underlying scaffold may be fundamentally skewed.

One way to picture it is to imagine that everything is being pulled in a particular direction by something we can鈥檛 see, though the question of how such a skew might have come about is an open one. It might have been etched into the primordial universe, for instance, or appeared much later when matter clumped together 鈥� perhaps on an immense scale outside of our observable universe.

In any case, the implications would be profound. 鈥淲hat we have found in these quasars is questioning the cosmological principle,鈥� says co-author at the Paris Institute of Astrophysics in France.

If Sarkar and his colleagues are right, it would also be ludicrously inconvenient. It would mean going back to Einstein鈥檚 equations to see what other solutions might fit our universe. Other than the smooth and uniform FLRW solution, there are about 20 other options to play with. Some assume the universe is smooth, but not uniform (the Bianchi solutions), some assume it is uniform, but not smooth (the Lema卯tre-Tolman-Bondi solutions) and some assume it is neither. One idea, called the swiss-cheese model, excavates spherical holes dotted across the universe and fills them with black holes.

鈥淭he mathematical scope for looking at these problems is鈥� quite large,鈥� says at the University of Cambridge. 鈥淏ut it鈥檚 a question of whether or not you feel you get a satisfactory picture.鈥� On the one hand, the freedom in these other solutions should offer ample possibility to solve cosmological puzzles such as what dark energy is, or whether it even exists. On the other hand, no one knows if these far more complicated solutions can match up to the intricate map of the universe we already have.

A new cosmology

And therein lies the problem. Switching to an entirely new cosmological framework would require entirely new ways of analysing data that don鈥檛 assume all directions are the same 鈥� which is one way to explain why cosmologists have been so reluctant to go along with any of this. 鈥淲e are in a very unenviable position in that most of our colleagues don鈥檛 even want to hear about it,鈥� says Sarkar.

For their part, sceptics point to conflicting or inconclusive results on the question of a skew in the large-scale structure of the universe. 鈥淚 have not gotten too excited about this,鈥� says at the University of California, Berkeley. Others say that studies of such distant objects as quasars are riddled with potential errors. 鈥淣one of these observations are particularly well verified,鈥� says at Queen Mary University of London.

Mohayaee says that such dismissals are 鈥渦nfair鈥�. 鈥淲e have put all of our data, our codes and simulations online. I would love to see this proven wrong, so please come and show us,鈥� she says. In soon-to-be-published research, the team has cross-checked its analysis of catWISE with a cleaned-up analysis of the NVSS catalogue to build an even stronger case for a skewed universe.

But even people who think the data and analyses are solid see little reason to seriously consider this possibility. Peebles says that instead of hinting at a skewed universe, the data could be explained if quasars clump together much more than other types of matter on those scales. For now, he says, he remains 鈥渄eeply impressed by the tight network of well-checked tests鈥� that the standard model passes. It is possible that there is another overarching theory that agrees with all the measurements as well as the model we have, says Peebles. 鈥淚 can never disprove that, but I can ask myself: 鈥楧oes it seem likely?'鈥�

Sarkar sees no reason why this model, built on the cosmological principle, can鈥檛 sit alongside another one, in which the universe is fundamentally skewed. Rather than a single cosmology that describes everything, he argues that different models could be applied at different epochs and then stitched together.

A new generation of telescopes is poised to offer some clarity. Next year, the Vera C. Rubin Observatory in Chile will start making observations and the Euclid space telescope is due to launch. Meanwhile, the Square Kilometre Array keeps adding to its vast network of radio dishes in South Africa and Australia. Between them, these observatories will scour the night sky, looking at billions of very distant galaxies over a large swathe of sky.

This makes possible a far more precise test of the cosmological principle. at the University of Geneva, Switzerland, proposes using different types of measurement from each telescope to untangle whether the lopsided CMB is entirely due to Earth鈥檚 motion in the universe or whether there is also a fundamental skew. Another approach would be to test the Copernican principle, from which the cosmological principle grew, by envisioning what the CMB looks like to far-off observers. If our vantage from Earth is nothing special, then the CMB should look the same. 鈥淛ust a decade or so ago, people were saying it鈥檚 a philosophical assumption that can鈥檛 be tested, but it really seems that it can,鈥� says Clarkson.

Principles at stake

You do that by studying galaxy clusters, where charged particles divert photons of light from their original path, sometimes sending them hurtling towards Earth. CMB photons that are scattered in this way can give us a rough picture of what the CMB looks like to a hypothetical observer at that galaxy cluster. 鈥淚f a bunch of different observers see an isotropic cosmic microwave background, it really forces you into the cosmological principle,鈥� says Clarkson 鈥� and if they don鈥檛, well, all bets are off.

It isn鈥檛 difficult to see why many cosmologists are reluctant to throw out a very successful, albeit imperfect, model of the universe. 鈥淧eople don鈥檛 like it if you come into their garden, which they鈥檝e planted for years, and then you step in the middle of the rose bed,鈥� says Durrer.

And yet Sarkar insists that cosmology should be led by observations, rather than dogma. 鈥淎ll the great discoveries have been made by simply building an instrument, pointing it at the sky and looking,鈥� he says. 鈥淎stronomy is all about serendipity.鈥�