Latest developments:
- Here are some preliminary results of a little number crunching I did on my theory.
- Why did it not turn out the way I expected? Read a short analysis!
- Switching to Currier transcription brought some relief, but not really a breakthrough.
- Going the other way around: Here I tried to synthesize something which would at least ostensibly look like Voynichese text. It worked somewhat, and has still room for improvement.
Thoughts about the Voynich Manuscript 
I just read your Strokes theory, which is something I tried to look at ( in my own naive and simpleton fashion – not being a Vms intellectual like voi altri ), but one thing important seems wrong to me: The direction the strokes start and finish shouldn’t be taken for granted, because if we are talking, for example, about an ITALIAN author, you should realise that they form their letters ( even today ) very differently to how, let’s say, someone from England would. I have noticed how my friends here in Italy form some of their letters completely back-to-front compared to me. By looking at the angle of the stroke ( you can almost imagine the quill tip) you can see more clearly how the letter was formed, and of course the blob of ink will indicated where the first and last stroke was made.( blob at the beginning, tapered stroke at the end).
Hi Michelle,
There possibly is some confusion here: The Stroke theory assumes that the VM letters represent the strokes of the plaintext letters, not the other way around.
Considering that, there is a huge number of possible plaintext “alphabets” to start from, which would all lead to different encipherings — from Italian or English cursive over various batarde writings to block letters, capitals and minors, and print.
The uncertainty about which plaintext alphabet was used is one thing which makes testing the Strokes so difficult, because each alternative would lead to different encodings.
I’m currently working the other way around, namely, there should be only some 52 “constituents” of all VM words (26 capital and 26 minor letters — give or take some numbers and possibly special characters).
But I haven’t yet found a meaningful way to test whether the whole of the VM corpus can be split up in such a nice manner into a limited number of syllables.
One advantage of the Stroke theory is that it would explain elegantly the different Currier “languages” — these would equate to different sets of plaintext alphabets, or even an “Italian” and an “English” hand, as you suggest.