Everything you need for a story. No paywall, no embargo.
A systematic computational analysis of the Voynich Manuscript across 63 language corpora from 35+ language families has found no natural language that simultaneously matches its full structural profile. The closest match — Tagalog — fails on a key metric and cannot reproduce its result across different texts. The discriminant zone remains empty. The research is published on Zenodo, fully open, and the data is available for independent verification.
Tagalog comes closest on one metric, then fails on two others — and when we test a different Rizal novel under identical conditions, the match disappears entirely.
Copied.The manuscript behaves statistically like language. It's not random noise. But it doesn't behave like any language we've tested.
Copied.Every eliminated hypothesis brings us closer to understanding what structural class the Voynich belongs to — even if we can't read it yet.
Copied.Responds to technical queries within one week. Background briefings available for credentialed journalists — request via email with outlet and deadline.
All data is available for independent verification. No embargo on any published result.