26-06-2026, 07:48 AM
I think I have the same message for several threads today 
Looking similar to Voynichese and being statistically similar to Voynichese are two different things. It is not hard to invent a scheme that makes a text took somewhat like Voynichese. Your simplified scheme here does that. However, this has never been the problem, as far as I know. The problem is to reproduce the whole collection of paradoxical features, like word/token length distribution, proportion of single token words (hapax), lack of large scale repetitions, line position preferences, etc.
It's not hard to invent a scheme that would explain any single statistic, it's when trying to combine these schemes to reproduce the whole everything usually breaks down.
Looking similar to Voynichese and being statistically similar to Voynichese are two different things. It is not hard to invent a scheme that makes a text took somewhat like Voynichese. Your simplified scheme here does that. However, this has never been the problem, as far as I know. The problem is to reproduce the whole collection of paradoxical features, like word/token length distribution, proportion of single token words (hapax), lack of large scale repetitions, line position preferences, etc.
It's not hard to invent a scheme that would explain any single statistic, it's when trying to combine these schemes to reproduce the whole everything usually breaks down.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)