23-10-2026, 12:48 PM
(22-10-2026, 07:27 PM)Torsten wrote: You are not allowed to view links. Register or Login to view.He reports the Zipf fit explicitly (parameter 1.854), noting that an order-1 Markov process lands in the same natural-language range, so a Zipf fit is weak evidence either way.Except that simulating an order-1 Markov without a computer is not a trivial task at all. Especially maintaining the digraph and word frequencies consistent for dozens of pages at least. It is extremely unlikely that someone in the 1400s (or even in 1910) would have even thought of using such a method in order to create a manuscript that would fool some rich dude into thinking that it was written in ... phonetic Vietnamese?
Quote: He suggests checking whether the two proposed “authors” differ more than two writers of one language could.If the two writers used different dialects and/or spelling systems for that language (like Old vs New Czech, or Nynorsk vs. Bokmål) the digraph statistics of their outputs could be arbitrarily different, even as different as two distinct languages.
All the best, --stolfi