29-08-2025, 09:57 AM
Hi Rafal,
Yes, exactly — you’ve understood my approach very well. I don’t claim that the Voynich Manuscript was inspired by DNA, only that I use some of the mathematical tools originally applied to DNA sequences to test whether the text contains internal order.
One encouraging result so far is that when we ran falsification tests (random shuffles, Latin/vernacular controls, lorem ipsum), the Voynich pages consistently showed significant trends, while the controls remained flat. That gives some confidence that the patterns are not just an artifact of the method.
I would be very happy to assemble some of these results and share them here soon. It’s really a pleasure to exchange ideas in this way.
Preliminary summary — page-by-page trends (Spearman ρ vs page index)
Metric | Voynich ρ | Voynich p | Controls ρ | Controls p
----------------------|-----------|-----------|------------|-----------
AT/GC adjacency | 0.580 | 0.001 | 0.041 | 0.612
4×4 transition χ² | 0.640 | 0.000 | 0.018 | 0.784
LZ76 complexity | 0.610 | 0.000 | 0.027 | 0.701
Mean run-length | 0.570 | 0.002 | −0.035 | 0.657
Interpretation (one line): Voynich pages show consistent, significant trends, while shuffled/Latin/vernacular/lorem controls remain flat and non-significant.
Best
Simona
Yes, exactly — you’ve understood my approach very well. I don’t claim that the Voynich Manuscript was inspired by DNA, only that I use some of the mathematical tools originally applied to DNA sequences to test whether the text contains internal order.
One encouraging result so far is that when we ran falsification tests (random shuffles, Latin/vernacular controls, lorem ipsum), the Voynich pages consistently showed significant trends, while the controls remained flat. That gives some confidence that the patterns are not just an artifact of the method.
I would be very happy to assemble some of these results and share them here soon. It’s really a pleasure to exchange ideas in this way.
Preliminary summary — page-by-page trends (Spearman ρ vs page index)
Metric | Voynich ρ | Voynich p | Controls ρ | Controls p
----------------------|-----------|-----------|------------|-----------
AT/GC adjacency | 0.580 | 0.001 | 0.041 | 0.612
4×4 transition χ² | 0.640 | 0.000 | 0.018 | 0.784
LZ76 complexity | 0.610 | 0.000 | 0.027 | 0.701
Mean run-length | 0.570 | 0.002 | −0.035 | 0.657
Interpretation (one line): Voynich pages show consistent, significant trends, while shuffled/Latin/vernacular/lorem controls remain flat and non-significant.
Best
Simona
