Every Voynich decipherment claim faces the same objection:
“The key was fitted to the manuscript after the fact.”
So I approached mine differently.
I priority-dated the key first, then derived structural predictions from it, then tested those predictions against the public IVTFF EVA corpus maintained by other researchers.
PDF/report link in first comment.
Python reproducibility code in second comment.
That is the point of this report: not “trust my reading,” but “run the test.”
The result is a forward-falsifiable validation layer for the Arabic/WAZN model — morphology, operational clustering, and grammatical position all behaving as the key predicted before verification.
This matters because a real decipherment key should not only translate words.
It should predict hidden structure in the manuscript.
Python file:
You are not allowed to view links.
Register or
Login to view.