The internet spent this year reviving chokers, flower crown filters and Lush Life. Fashion’s revival cycle used to run at twenty years, and a Princeton study has the maths. What collapsed it was not impatience. It was the archive.
On 31 December 2025, a TikTok user posted a montage of 2016. Someone else proposed the first of January as a reset day. Within weeks, searches for 2016 had spiked, millions of videos were running flower crown and puppy dog filters, Zara Larsson’s Lush Life was back on international charts, and a substantial portion of the internet had agreed, apparently without discussion, to spend the year pretending it was nine years ago.
The fashion that came back with it is specific and slightly mortifying. Chokers. Body chains. Flannel shirts over crop tops. Trench coats worn sliding off one shoulder as though the wearer had been interrupted. Cut-crease eyeshadow and highlighter applied with real conviction. Fur-lined Gucci mules from Alessandro Michele’s first proper run at the house, micro bags, and the year Demna arrived at Balenciaga and began dismantling the furniture.
Every account of this has reached the same conclusion, which is that the nostalgia cycle is speeding up. That is true, and it is not the interesting part.
FLJ Insider Pro is open: live fashion law jobs and internships, five career guides and two roadmaps. Early bird USD 12/month. See what's inside
Fashion really does run on a twenty-year clock
The cycle is not folklore. Emma Zajdela, an applied mathematician at Princeton, led a team that fed roughly thirty-seven thousand images into the question, taking sewing patterns from 1869 to 2015 and Vogue runway images from 1988 to 2023, which is about a hundred and fifty years of women getting dressed.
Styles cycle at roughly twenty years. They arrive, they saturate, they become embarrassing, they go quiet, and then somebody young enough to have missed them the first time rediscovers them as new. Hemlines obeyed this beautifully from the 1920s to the 1980s.
Then, from the eighties onwards, the pattern frays. More lengths coexist. The signal gets noisy. The researchers put it down to faster trends and a broader tolerance for people wearing whatever they like, and the model underneath it all borrows from a psychological idea called optimal distinctiveness, which holds that a successful innovation has to be different, but not too different.
Twenty years, it turns out, is roughly how long it takes for a thing to become different enough.
The real reason it collapsed
Something else is going on, and it has very little to do with attention spans.
Every previous revival was a reconstruction. When the nineties did the seventies, nobody was watching the seventies. They were working from photographs, films, a parent’s wardrobe and a general atmospheric impression. The decade had to be rebuilt from fragments, and rebuilding takes time, which is most of what the twenty years were for.
More importantly, memory did the editing. You forgot the parts that were unflattering and kept the parts that photographed well, so what came back was never the decade. It was an improved version of the decade, curated by forgetting.
2016 is the first year that does not need rebuilding, because it never went anywhere.
The filters still run. The videos still play. The Instagram grid from that year is four thousand scrolls down on a phone you still own, in the same app, uncompressed and fully searchable. When somebody posts a flower crown selfie this year, they are not making an homage to a 2016 filter. They are opening the 2016 filter. It is the same software.
That is a different thing from nostalgia, and it deserves a different word. It is playback.
What happens when nothing is forgotten
Which explains why the gap shrank to nine years rather than, say, fifteen.
If revival requires reconstruction, the clock is set by how long it takes to forget enough to be able to rebuild pleasantly. If the archive is complete and permanently available, that clock stops mattering entirely. The only remaining variable is how long a thing needs to sit before it stops being embarrassing.
Nine years, apparently. That is the current half-life of humiliation.
And it is going to keep contracting, because the archive keeps getting better. The material from 2016 was shot on phones with bad sensors and heavy compression, which is why it has a texture at all and why people find it charming. The material from 2022 is in 4K. When 2022 comes back around, and it will, the revival will be indistinguishable from the original, which raises a genuine question about whether it counts as a revival or merely as repetition.
The part everybody skipped
One detail in this is harder to shake, and Roar News put it best.
The 2016 being revived is the filters, the music, the mannequin challenge and the Gucci mules. It is not Brexit. It is not the American election. Both of those also happened in 2016, and neither has made it into the montage.
So even with a complete, searchable, permanently available record of a year, people are still editing it down to the bits that feel nice. The curation did not disappear when memory stopped being necessary. It simply moved. It used to happen slowly, in the forgetting. Now it happens instantly, at the point of retrieval, performed by whoever is assembling the montage.
Which means the twenty-year cycle was never really about memory. It was about the time required to reach a flattering consensus on what a period was, and we have found a way to arrive at that consensus in an afternoon.
The chokers are back. The year they came from is not, and was never invited.
FASHION LAW JOURNAL INSIDER
Join designers, brand founders and fashion lawyers who get the biggest brand battles, IP fights and career moves in fashion law, straight to their inbox.
