60-Year-Old Error in Crystalline Cohomology Found During Fermat’s Last Theorem Formalization

A Hidden Crystalline Cohomology Error Decades in the Making

Mathematicians have uncovered an error in a proof that underpins crystalline cohomology — a significant branch of modern geometry that was also used in the proof of Fermat’s Last Theorem. The mistake, made in 1965, had the potential to threaten thousands of studies built upon that proof.

How the Crystalline Cohomology Error Was Discovered

The error was found by accident while researchers were converting old mathematical papers into a digital, computer-readable format. French mathematician Antoine Chambert-Loir noticed that an article by Norbert Roby was missing a critical symbol, rendering the proof invalid.

But the panic turned out to be unwarranted: Brian Conrad of Stanford University found an independent proof of the same result, saving the situation.

The Fermat’s Last Theorem Formalization Project

What makes this discovery particularly notable is that it emerged from a much larger initiative — the formalization of the proof of Fermat’s Last Theorem, one of the greatest achievements in the history of mathematics. Wiles’s proof draws on numerous complex areas of mathematics, including crystalline cohomology. An error found at its foundations had the potential to cast doubt on the validity of a vast body of research.

Why Modern Mathematics Needs Automated Verification

Modern mathematics has grown too complex for proofs to be verified by hand alone. Many foundational results rely on older papers where, as this case demonstrates, errors can lie hidden for generations.

Formalization and computer-assisted verification help prevent such errors from spreading further. As it turns out, even the most essential branches of mathematics can depend on outdated and insufficiently checked sources.

A Reminder That Mathematical Errors Can Hide for Decades

This case is a reminder: even in mathematics, mistakes can go unnoticed for decades — until someone stumbles upon them by chance.