This general approach reminds me a bit of Schinner (2007): You are not allowed to view links. Register or Login to view.
Instead of mapping Voynichese "letters" to strings of four "bases," Schinner converts the text to a binary string, where 1 EVA "letter" is a unique string of 5 1s or 0s. Binary strings would be better suited to statistical analysis than strings of ATGC "bases" because if you use something like the EVA alphabet as it is, you'd need a string of 3 bases (4x4x4) to give yourself full coverage of EVA letters (q = AAA, o = AAT, etc.). However, that would yield a total of 64 possible base strings, more than twice the coverage you actually need, which would mean that the results of any downstream statistical analysis would depend in part on which specific base strings you randomly assigned to specific EVA letters. By contrast, a string of 5 1s or 0s (q = 00000, o = 00001, etc.) gives full coverage and doesn't overshoot too much, thereby making those random letter-by-letter assignments less of an issue.
The other benefit of converting straight to binary, rather than strings of letter bases as this approach takes, is that you can immediately proceed to numerical analyses of the text, such as a random walk analysis as Schinner (2007) performed. It's on this basis that Schinner (2007) identifies long-range correlations in the Voynich Manuscript that are far stronger than those of the reference texts he analyzed the same way. He takes this to be evidence that the text of the Voynich Manuscript is iteratively generated gibberish, and therefore that the manuscript may have been hoaxed.
My overriding concern with this "DNA-inspired" analysis is that I'm not sure it would yield anything that previous bit-level analyses of the Voynich Manuscript haven't already found. But the fundamental intuition is sound. In fact, the random walk analysis that Schinner (2007) performs was originally developed to study DNA: You are not allowed to view links. Register or Login to view.
Instead of mapping Voynichese "letters" to strings of four "bases," Schinner converts the text to a binary string, where 1 EVA "letter" is a unique string of 5 1s or 0s. Binary strings would be better suited to statistical analysis than strings of ATGC "bases" because if you use something like the EVA alphabet as it is, you'd need a string of 3 bases (4x4x4) to give yourself full coverage of EVA letters (q = AAA, o = AAT, etc.). However, that would yield a total of 64 possible base strings, more than twice the coverage you actually need, which would mean that the results of any downstream statistical analysis would depend in part on which specific base strings you randomly assigned to specific EVA letters. By contrast, a string of 5 1s or 0s (q = 00000, o = 00001, etc.) gives full coverage and doesn't overshoot too much, thereby making those random letter-by-letter assignments less of an issue.
The other benefit of converting straight to binary, rather than strings of letter bases as this approach takes, is that you can immediately proceed to numerical analyses of the text, such as a random walk analysis as Schinner (2007) performed. It's on this basis that Schinner (2007) identifies long-range correlations in the Voynich Manuscript that are far stronger than those of the reference texts he analyzed the same way. He takes this to be evidence that the text of the Voynich Manuscript is iteratively generated gibberish, and therefore that the manuscript may have been hoaxed.
My overriding concern with this "DNA-inspired" analysis is that I'm not sure it would yield anything that previous bit-level analyses of the Voynich Manuscript haven't already found. But the fundamental intuition is sound. In fact, the random walk analysis that Schinner (2007) performs was originally developed to study DNA: You are not allowed to view links. Register or Login to view.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)
