Discussion of "A possible generating algorithm of the Voynich manuscript"

[ReneZ]

Excellent, many thanks Marco.

I didn't intend to experiment myself, initially, but I found a simple implementation of the Levenshtein distance for Fortran, so I have been doing something similar - not yet completed.

Looking at the very first picture, it appears that:
- Quire 20 is not significantly different from the Voynich text as a whole.
- Only the unchanged Latin and Italian plain texts lie above the Voynich MS
- The modified Machiavelli, which was closest to the Voynich text in the other comparison, is actually a little bit below the Voynich MS text in edit distance, so the small edit distance in the Voynich MS is definitely not conclusively a consequence of auto-copying
- Torsten's simulated text seems to be aiming at an average edit distance of 1.5. However, as the comparison text increases down the page, the probability of finding words with smaller differences also increases

I am not so sure how to interpret the second plot with the ratios. I think that the absolute differences between line 2 and line 8 are probably more significant.
Had the algorithm been aiming for an average edit distance of over 2, it would probably be closer to the Voynich MS text. However, an average over 2 basically means a range of 0 to over 4, and this is hardly auto-copying anymore.

The 'absolute drop' of the plain texts is 1.2 - 1.3 (reading from the figure), and this is also true for the Voynich MS text and the modified Macchiavelli.

For the Pliny_mod2 and Torsten's app it is 0.6 - 0.7 . The mod2 text is special in the sense that the words are very short, so edit distances are automatically smaller.

Looking at all this, I suspect that the substitution made on the Macchiavelli text, when applied to Pliny, is likely to be very close to the Voynich text.
Combining this with other observations, I am beginning to wonder if one of the main features of the Voynich MS text is that it is really page-oriented, i.e. properties are quite page-dependent. This would fit with an encyclopedic work (like Pliny).