Hmm... The idea sounds interesting, however I doubt that this would substantially change the second-order character entropy, because with this kind of transformation, essentially we are just shifting spaces to other positions, and nothing more. Furthermore, supposing that line-initial characters are fillers, this does not explain why the writer would always generate fillers in sorted order for the first two lines, almost always - for the first three lines, and far less often - for the subsequent lines.
I have an idea that what we see as lines in the VMS (at least in the Herbal section) may be half-lines of the original text instead, which were subsequently interleaved.
All in all, I would not advise to derive broad conclusions so far from the observations which I described above; what I looked at are just k-initial paragraphs - which is a minority of the overall number of paragraphs (although a substantial one), - and only for the first six quires - which is of course a minority of the whole text.
I have an idea that what we see as lines in the VMS (at least in the Herbal section) may be half-lines of the original text instead, which were subsequently interleaved.
All in all, I would not advise to derive broad conclusions so far from the observations which I described above; what I looked at are just k-initial paragraphs - which is a minority of the overall number of paragraphs (although a substantial one), - and only for the first six quires - which is of course a minority of the whole text.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)