Yonsei University: Universe May Not Be Accelerating
The Yonsei University team has responded to a June rebuttal, saying its earlier finding that the universe's expansion is slowing down still holds up.
A team of astronomers from Yonsei University in Seoul published a paper on August 26, 2026, in the Monthly Notices of the Royal Astronomical Society answering a June 2026 rebuttal, in which a group led by the University of Southampton argued that the age of a star that exploded as a supernova has no meaningful effect on conclusions about the universe’s expansion. The paper, whose first author is Chul Chung and whose project is led by professor Young-Wook Lee, carries the digital object identifier 10.1093/mnras/stag1513. Its authors work at Yonsei’s Department of Astronomy and its Center for Galaxy Evolution Research, and their message is clear: the dispute over whether the universe’s expansion is accelerating today has not been settled, it continues.
The story begins in November 2025, when the same Yonsei team reported that the age of a star exploding as a Type Ia supernova affects how bright that explosion appears to an observer on Earth. This concerns the Phillips relation, a formula that adjusts a supernova’s measured brightness based on the shape of its light curve so the explosion can serve as a so-called standard candle for measuring cosmic distances. All of supernova cosmology rests on the assumption that this relation does not depend on the age of the progenitor star. If it does, as the Yonsei team has argued since 2020, then the measurements that led to the discovery of accelerating cosmic expansion carry a systematic error that grows with redshift, meaning with a galaxy’s distance. According to the team’s earlier paper, the corrected data lines up with independent measurements from the Dark Energy Spectroscopic Instrument (DESI), obtained through the baryon acoustic oscillation method, and those measurements point to a universe whose expansion is not accelerating but slowing down.
In June 2026, an international team led by the University of Southampton, which included two Nobel laureates, published a rebuttal. Its authors argued that the measured link between a supernova’s age and its brightness was far weaker than the Yonsei team had reported, and that the age of the progenitor star itself likely depends much less on redshift than the age of its host galaxy does. On that basis, they concluded that the needed correction was negligible and that the universe’s expansion is still accelerating.
The new paper argues that both of those points are methodologically flawed. The first problem is the sample: the Southampton team compared supernovae across an unusually wide redshift range, from 0.04 to 0.42. Host galaxy age shifts by roughly three billion years across that range, so comparing supernovae from such different eras artificially flattens the measured link between age and brightness. The second problem concerns the interstellar dust model the rebuttal used for its correction: that model requires the dependence on host galaxy mass to run in the opposite direction from what large galaxy samples actually show, which, according to the Yonsei team’s findings, makes it unfounded.
According to the paper, the Korean team measured the link between brightness and age at a significance of 5.5 sigma in the Rose et al. sample and 4.8 sigma in the Gupta et al. sample, two independent supernova datasets. The estimated slope of that relation ranges from minus 0.007 to minus 0.034 magnitude per billion years, depending on the processing method used. A key detail is that about 90 percent of the change in host galaxy age tracks the change in the progenitor star’s own age, so separating the two, as the rebuttal insisted on doing, does not change the math much. Once all uncertainties are factored in, the final correction still comes out to about 85 percent of the value the Son et al. team published in 2025.
“The mild slope of the age bias reported in the rebuttal is the result of a clear methodological problem in the analysis,” said Chul Chung, a Yonsei University astronomer and the paper’s first author, in a statement about the published work.
Professor Young-Wook Lee, who leads the project, was more precise in the same statement about the assumption everything hinges on:
Supernova cosmology rests on the key assumption that the calibration relation for brightness, the Phillips relation, does not change with the age of the supernova. Our results consistently show that this assumption does not hold.
Lee also spelled out what that means for the claim of accelerating expansion:
To defend an extraordinary claim such as the accelerating expansion of the universe, that underlying assumption must first be shown to be true, and we have to seriously ask whether the evidence still supports an accelerating universe.
To sidestep further disputes over how a host galaxy’s age translates into the age of the progenitor star itself, the Yonsei team proposes what it calls an evolution-free test, which would rely solely on supernovae in young galaxies of similar ages, drawn from across the full redshift range. Preliminary results from such a test, the authors say, are consistent with their earlier findings. Larger supernova samples, expected from the Vera C. Rubin Observatory, would in their view allow the test to be carried out with far greater precision. The paper first appeared as a preprint on arXiv on May 20, 2026, under the designation 2605.21586 and the title “Still non-accelerating,” and has now passed peer review for the Monthly Notices of the Royal Astronomical Society.
What this paper does not do is settle the question. It is a counter-rebuttal, a fresh analysis of someone else’s data and assumptions, not a new independent measurement. The slope of the age-brightness link still depends on which supernova sample and which dust model are used, and the range the paper itself cites, from minus 0.007 to minus 0.034 magnitude per billion years, shows how sensitive that estimate remains to methodology. The Southampton team has not yet published a new response, so the dispute stays open until independent data, such as what the Vera C. Rubin Observatory is expected to provide, becomes available. Whether the universe is expanding at an accelerating or a slowing rate today remains, for now, a question science has no final answer to.
