How did the universe start? Did we begin with a big bang, or was there a bounce? Might the cosmos evolve in a cycle of expansion and collapse, over and over for all eternity? Now, in two papers, researchers have poked holes in different models of a so-called bouncing universe, suggesting the universe we see around us is probably a one-and-done proposition.

Bouncing universe proponents argue that our cosmos didn’t emerge on its own out of nothing. Instead, advocates claim, a prior universe shrunk in on itself and then regrew into the one we live in. This may have happened once or, according to some theories, an infinite number of times.

So which scenario is correct? The most widely accepted explanation for the history of the universe has it beginning with a big bang, followed by a period of rapid expansion known as cosmic inflation. According to that model, the glow left over from when the universe was hot and young, called the cosmic microwave background (CMB), should look pretty much the same no matter which direction you face. But data from the Planck space observatory, which mapped the CMB from 2009 to 2013, showed unexpected variations in the microwave radiation. They could be meaningless statistical fluctuations in the temperature of the universe, or they might be signs of something interesting going on.

One possibility is that the CMB anomalies imply that the universe didn’t emerge out of nothing. Instead it came about after a prior universe collapsed and bounced back to create the space and time we live in today.

Bouncing universe models can explain these CMB patterns as well as account for lingering quibbles about the standard description of the universe’s origin and evolution. In particular, the big bang model of the universe begins with a singularity—a point that appeared out of nothing and contained the precursors of everything in the universe in a region so small that it had essentially no size at all. The idea is that the universe grew from the singularity and, after inflation, settled into the more gradually expanding universe we see today. But singularities are problematic because physics, and math itself, doesn’t make sense when everything is packed into a point that’s infinitely small. Many physicists prefer to avoid singularities.

One bouncing model that averts singularities and makes the CMB anomalies a little less anomalous is known as loop quantum cosmology (LQC). It relies on a bridge between classical physics and quantum mechanics known as loop quantum gravity, which posits that the force of gravity peters out at very small distances rather than increasing to infinity. “Cosmological models inspired by loop quantum gravity can solve some problems,” says University of Geneva cosmologist Ruth Durrer, “especially the singularity problem.” Durrer co-authored one of the two new studies on bouncing universes. In it, she and her colleagues looked for astronomical signs of such models.

In an LQC model, a precursor to our universe might have contracted under the force of gravity until it became extremely compact. Eventually quantum mechanics would have taken over. Instead of collapsing to a singularity, the universe would have started to expand again and may even have gone through an inflationary phase, as many cosmologists believe ours did.

If that happened, says physicist Ivan Agullo of Louisiana State University, it should have left a mark on the universe. Agullo, who was not affiliated with either of the recent analyses, has proposed that the mark would turn up in a feature in the CMB data known as the “bispectrum,” a measure of how different portions of the universe would have interacted in a bouncing scenario. The bispectrum would not be apparent in an image of the CMB, but it would show up in analyses of the frequencies in the ancient CMB microwaves.

“If observed,” Agullo says, the bispectrum “would serve as a smoking gun for the existence of a bounce instead of a bang.” Agullo’s group previously calculated the bispectrum as it would have appeared shortly after a cosmic bounce. Durrer and her colleagues took the calculation further, but when they compared it with the present-day Planck CMB data, there was no significant sign of a bispectrum imprint.

Although lots of other bouncing cosmos models may still be viable, the failure to find a significant bispectrum means that models that rely on LQC to deal with the anomalies in the CMB can be ruled out. It’s a sad result for Agullo, who had high hopes of finding concrete evidence of a bouncing universe. He still considers many bouncing universe models viable, however. And Paola Delgado, a cosmology Ph.D. candidate at Jagiellonian University in Poland, who worked on the new analysis that was co-authored by Durrer, says there’s one potential upside. “I heard for a long time that [attempts to merge quantum physics and cosmology] cannot be tested,” Delgado says. “I think it was really nice to see that for some classes of models, you still have some contact with observations.”*

Ruling out signs of an LQC-driven cosmic bounce in Planck data means the CMB anomalies remain unexplained. But an even larger cosmic issue lingers: Did the universe have a beginning at all? As far as advocates of the big bang are concerned, it did. But that leaves us with the inscrutable singularity that started everything off.

Alternatively, according to theories of so-called cyclic cosmologies, the universe is immortal and is going through endless bounces. Although a bouncing universe may experience one or more cycles, a truly cyclic universe has no beginning and no end. It consists of a series of bounces that go back for an infinite number of cycles and will continue for an infinite number more. And because such a universe doesn’t have a beginning, there’s no big bang and no singularity.

The study that Durrer and Delgado co-authored doesn’t rule out immortal cyclic cosmologies. Plenty of theories describe such a bouncing universe in ways that would be difficult or impossible to distinguish from the “big bang plus inflation” model by looking at Planck CMB data.

But a critical flaw lurks in the idea of an eternally cycling universe, according to physicist William Kinney of the University at Buffalo, who co-authored the second recent analysis. That flaw is entropy, which builds up as a universe bounces. Often thought of as the amount of disorder in a system, entropy is related to the system’s amount of useful energy: the higher the entropy, the less energy available. If the universe increases in entropy and disorder with each bounce, the amount of usable energy available decreases each time. In that case, the cosmos would have had larger amounts of useful energy in earlier epochs. If you extrapolate back far enough, that implies a big bang–like beginning with an infinitely small amount of entropy, even for a universe that subsequently goes through cyclic bounces. (If you’re wondering how this scenario doesn’t violate the law of conservation of energy, we’re talking about available energy. Although the total amount of energy in the cosmos remains static, the amount that can do useful work decreases with increasing entropy.)

New cyclic models get around the problem, Kinney says, by requiring that the universe expands by a lot with each cycle. The expansion allows the universe to smooth out, dissipating the entropy before collapsing again. Although this explanation solves the entropy problem, Kinney and his University at Buffalo co-author Nina Stein calculated in their recent paper that the solution itself ensures that the universe is not immortal. “I feel like we’ve demonstrated something fundamental about the universe,” Kinney says, “which is that it probably had a beginning.” That implies a big bang occurred at some point, even if that event happened many bouncing universes ago, which in turn suggests that it took a singularity to get everything going in the first place.

Kinney’s paper is the latest in the debate over cyclic universes, but proponents of a universe without beginning or end have yet to respond in the scientific literature. Two leading proponents of a cyclic universe, astrophysicists Paul Steinhardt of Princeton University and Anna Ijjas of New York University, declined to comment for this article. If the history of the debate is any indication, though, we may soon hear of a work-around to counter Kinney’s analysis.

Cosmologist Nelson Pinto-Neto of the Brazilian Center for Physics Research, who has studied bouncing and other cyclic models, agrees that the Planck data likely rule out a bounce under loop quantum cosmology, but he’s more sanguine on the question of a cyclic universe. “Existence is a fact. We are all here and now. Nonexistence is an abstraction of the human mind,” Nelson says. “This is the reason I think that a [cyclic universe], which has always existed, is simpler than one that has been created. However, as a scientist, I must be open to both possibilities.”

*Editor’s Note (6/1/23): This paragraph was edited after posting to better clarify Ivan Agullo’s position on bouncing universe models.