22-03-2026, 12:00 AM
The wheel in action – two sides, two pivot points
No response, which was to be expected. So, to put it more simply: I wrote above that the Voynich Manuscript is the result of a Lullian combinatorial system. I would like to add something concrete, as abstract arguments about entropy and slot structures tend to dissolve into the air of incomprehension. The following makes it clearer:
Folio 56v, lines 7–16:
kchokchy.chol.shol.chol.otchyd
ykchor.chokchy.qokcho.tcheey.kol
shotchot.chokcho.kcho,kaiin.oky
ol,kchy.ksho.shy.ytol.chotor,y,dy
chotchey.keeol.chey.kchol.tchain
qokchey.ctheey.lkeey.kcheeytain
ykor.aiin.chorol.sho,shol.daiin
chol,cheo.kchol.chol.choky.chotor
otchol.chol.chol.daiin.chotaiin
s,o,kchol.chol.chol.daiin
55 word tokens. The occurrences of ‘chol’ and its variants are counted:
chol: 8×
kchol: 3×
otchol: 1×
That is 12 out of 55 tokens – 22% of the entire text – which are “chol” or a direct variant of “chol”, and “shol”, which in the manuscript is simply “chol” with a “macron” above it (the same base form with a single diacritical modification), appears a further 2 times.
No author writing a meaningful text uses the same word in 22% of cases. No cipher produces this, unless the plaintext itself is so repetitive, which would be remarkable. But a combination wheel that has turned to a position where ‘chol’ is the current output – and remains there, thereby generating ‘chol’ and its immediate neighbours (kchol, otchol, shol) – produces exactly that.
Now consider folio 57r, lines 1–5:
poeeockhey.odain.cheop.sheody.shocfhey.dy.sheep.sheody.yodam
daiir.air.chety.cheo.ckhy.chockhy.cheotey.sh.kchey.s.odaiin.shey
qokeeody.cheooky.qokeody.sheey.okeody.cheody.cheeody.cheekeody
dchos.cheocthy.cheody.qoteeody.octhody.okeeody.chteody.cheody.s
qokeeo.daiin.cheeodam
41 word tokens. The linchpin here is not ‘chol’, but ‘cheody’ and its word family:
cheody: 3×
sheody: 2×
qokeeody: 1×
qokeody: 1×
okeody: 1×
cheeody: 1×
cheekeody:1×
qoteeody: 1×
octhody: 1×
okeeody: 1×
chteody: 1×
11 out of 41 words – 27% – are variations of the same base ending in -ody. The prefix changes (qo-, o-, ch-, sh-, cth-, ckh-), the middle vowel lengthens (cheo, chee, cheee), but the ending -ody remains constant. The wheel turns through its prefix positions whilst the root is held steady.
Two pages. Two different fulcrums. On f56v, the wheel is fixed on ‘chol’. On f57r, it is fixed on ‘cheody’. The surrounding material rotates – prefixes, infixes, occasionally other elements – but the core remains stationary.
This is not what a natural language looks like under any form of encryption. In a text in natural language, even in a highly formulaic medieval recipe, no single word or word family dominates 20–27% of a passage. Even the most repetitive recipe texts – and I have spent months of my life studying medieval recipe books – do not produce this. Not to the extent found in the Voynich. There are individual words that occur disproportionately, but this is something entirely different.
So let us be lulled by Llull...
No response, which was to be expected. So, to put it more simply: I wrote above that the Voynich Manuscript is the result of a Lullian combinatorial system. I would like to add something concrete, as abstract arguments about entropy and slot structures tend to dissolve into the air of incomprehension. The following makes it clearer:
Folio 56v, lines 7–16:
kchokchy.chol.shol.chol.otchyd
ykchor.chokchy.qokcho.tcheey.kol
shotchot.chokcho.kcho,kaiin.oky
ol,kchy.ksho.shy.ytol.chotor,y,dy
chotchey.keeol.chey.kchol.tchain
qokchey.ctheey.lkeey.kcheeytain
ykor.aiin.chorol.sho,shol.daiin
chol,cheo.kchol.chol.choky.chotor
otchol.chol.chol.daiin.chotaiin
s,o,kchol.chol.chol.daiin
55 word tokens. The occurrences of ‘chol’ and its variants are counted:
chol: 8×
kchol: 3×
otchol: 1×
That is 12 out of 55 tokens – 22% of the entire text – which are “chol” or a direct variant of “chol”, and “shol”, which in the manuscript is simply “chol” with a “macron” above it (the same base form with a single diacritical modification), appears a further 2 times.
No author writing a meaningful text uses the same word in 22% of cases. No cipher produces this, unless the plaintext itself is so repetitive, which would be remarkable. But a combination wheel that has turned to a position where ‘chol’ is the current output – and remains there, thereby generating ‘chol’ and its immediate neighbours (kchol, otchol, shol) – produces exactly that.
Now consider folio 57r, lines 1–5:
poeeockhey.odain.cheop.sheody.shocfhey.dy.sheep.sheody.yodam
daiir.air.chety.cheo.ckhy.chockhy.cheotey.sh.kchey.s.odaiin.shey
qokeeody.cheooky.qokeody.sheey.okeody.cheody.cheeody.cheekeody
dchos.cheocthy.cheody.qoteeody.octhody.okeeody.chteody.cheody.s
qokeeo.daiin.cheeodam
41 word tokens. The linchpin here is not ‘chol’, but ‘cheody’ and its word family:
cheody: 3×
sheody: 2×
qokeeody: 1×
qokeody: 1×
okeody: 1×
cheeody: 1×
cheekeody:1×
qoteeody: 1×
octhody: 1×
okeeody: 1×
chteody: 1×
11 out of 41 words – 27% – are variations of the same base ending in -ody. The prefix changes (qo-, o-, ch-, sh-, cth-, ckh-), the middle vowel lengthens (cheo, chee, cheee), but the ending -ody remains constant. The wheel turns through its prefix positions whilst the root is held steady.
Two pages. Two different fulcrums. On f56v, the wheel is fixed on ‘chol’. On f57r, it is fixed on ‘cheody’. The surrounding material rotates – prefixes, infixes, occasionally other elements – but the core remains stationary.
This is not what a natural language looks like under any form of encryption. In a text in natural language, even in a highly formulaic medieval recipe, no single word or word family dominates 20–27% of a passage. Even the most repetitive recipe texts – and I have spent months of my life studying medieval recipe books – do not produce this. Not to the extent found in the Voynich. There are individual words that occur disproportionately, but this is something entirely different.
So let us be lulled by Llull...