The Voynich Ninja

(13-01-2026, 12:20 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.The language grammar used in the Rohonc Codex is loose. For me it is actually quite similar to a pidgin. It looks ugly.  But the sentences make sense, make patterns and make a consistent story which agrees with the Bible. So it was most probably grammatical for the author.

I see!  But could it be just a yet unidentified natural language?

Are there identifiable grammatical rules? Like, SVO (subject-verb-object) or SOV order, articles, prepositions ("of", "to", "at", "in", ...), explicit word for "and", pronouns?

Since the writing is ideographic, I suppose that there are no inflections or declensions, no identifiable gender and number, right?

All the best, --stolfi

(13-01-2026, 02:03 AM)magnesium Wrote: You are not allowed to view links. Register or Login to view.Torsten Timm and Andreas Schinner, who have documented large-scale statistical properties of the manuscript

In addition to the long distance effects they discovered there are many more features of the manuscript that point to the text being an artificial fabrication. Here are some that I have found convincing.

Vertical Effects.  [ You are not allowed to view links. Register or Login to view. ]. This seems to show that the first characters of the first words of a line may influence the starting character of the line immediately below. Also vertical pair repeats occur statistically less that would be expected [ You are not allowed to view links. Register or Login to view. ].

Gallows Splitting.  [ You are not allowed to view links. Register or Login to view. ]. This shows that for the majority of gallows words they can be split into initial and final word parts that are independent.

Language Clusters.  In addition to there being A and B each section of the manuscript seems to exhibit its own variations. In particular for the language B sections. [ You are not allowed to view links. Register or Login to view. ].

Positional Preference.  Words in the manuscript seems to show a lot of positional preference. For instance if you generate a spline transform of the line positions of words starting  a in quire 20 you will get something like

[attachment=13426]

An  a word occurs rarely as the first word. They then prefer to be more frequent towards the line ends.


The positional preferences together with the vertical effects and other line first word anomalies in particular show that the text cannot be a continuous narrative where words just wrap to the next line at every line break. This and all the other anomalies are going to be difficult for any natural language / shorthand / cypher hypothesist to explain.

(13-01-2026, 01:49 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.Doctor bullatus, asinus coronatus. That's what you are

I learned that in a famous very very old Italian university - maybe Pisa, Padua, or Bologna -- it is an old tradition that, at the graduation ceremony, as soon as a student receives a Doctoral degree, the audience chants an obscene ditty in dialect,  Dotore, dotore, dotore del ... and the rest would be improper to reproduce here.  With the Distinguished Professors standing there behind the table, impassible.

All the best, --stolfi

(13-01-2026, 01:56 PM)dashstofsk Wrote: You are not allowed to view links. Register or Login to view.Positional Preference.  Words in the manuscript seems to show a lot of positional preference. For instance if you generate a spline transform of the line positions of words starting with  a in quire 20 you will get something like

The frequency of the letter a alone is a bad object of study because it is easily confused with o and y (or even ei), depending on which Scribe wrote it, on how tired/hurried he was, on BEEP!, and on who transcribed that part of the text. 

It would be somewhat more meaningful to plot the occurrences of those three letters added together.  That should largely reduce that source of noise.

However, again, the frequencies of characters are determined by their occurrence in the most common words; and word frequencies are expected to vary a lot with subject matter, style of text, etc.   So, for example, one expects the digraph "rb" to be much more common in a section of an English or Latin book dealing with herbs than in a section about astrology...

All the best, --stolfi

(13-01-2026, 01:56 PM)dashstofsk Wrote: You are not allowed to view links. Register or Login to view.Also vertical pair repeats occur statistically less that would be expected

Some of those "LAAFU" ("line as a functional unit") effects may be due to the word length bias created by the trivial line breaking algorithm.  

Basically, when any text, in any language, is formatted into paragraphs, the first word of each line is noticeably longer than average, while the last few words are noticeably shorter.

If the average word lengths are different, then the word frequencies at those places are different.  

Since character and digraph frequencies depend on their occurrence in the most common words, then character and digraph frequencies in the first word and in the last few words of the line will be different than average too.

This rather unexpected effect of paragraph formatting was observed here only recently.  Does it completely explain the "LAAFU" anomalies, or only part of them?  I don't know; but claims of LAAFU should check for that possiblity.

One possible check is to re-format Voynichese paragraphs with very different margin widths, and see whether the new lines still show the claimed LAAFU anomalies.  If the VMS actually has LAAFU, the anomalies should disappear.

All the best, --stolfi

Quote:I see!  But could it be just a yet unidentified natural language?

Are there identifiable grammatical rules? Like, SVO (subject-verb-object) or SOV order, articles, prepositions ("of", "to", "at", "in", ...), explicit word for "and", pronouns?

Actually I write about all this stuff in my work   I encourage you to read it "from board to board" as we say in Polish (the saying comes of course from wooden covers of old manuscripts).

And yes, it may be some "normal" language unknown to me but it is certainly weird and nothing fits so far.

- there is no declension
- there are no tenses, some words like "happen" may suggest past or future
- there seems to be VSO order for normal sentences and SVO order for passive voice
- there are prepositions but they are used a bit chaotically
- there is no plural case, sometimes repeating of the word is used
- there are no articles like "the"
- there are "and" words
- there is no "to be" word (it is called "zero copula")
- there is one universal pronoun meaning "I", "he", "she", "you", "they" and so on
- people often speak about themselves in the 3rd person, for example Jesus says "Jesus is thirsty"

So my best fit was a sign language of deaf people written down.

Quote:Doctor bullatus asinus coronatus

And it comes from the 17th century Poland. It seems to be an alteration of earlier saying "Rex illiteratus, asinus coronatus".

Actually at some moment I was surprised how many Latin sayings are "local". They are quite well known in some country and practically unknown anywhere else.
You could expect that most Latin sayings are classical, coming from Seneca or Cicero. Actually many of them may come from some 19th century school teacher.

(13-01-2026, 01:56 PM)dashstofsk Wrote: You are not allowed to view links. Register or Login to view.This shows that for the majority of gallows words they can be split into initial and final word parts that are independent.

That linked post claims that "The word counts for the most frequent words are broadly in the range of what would be expected. No significant big swings from parity"  But actually the ratios vary a lot, from zero for lk+aly to 10 for Cheek+y.   Even among the most common prefixes and suffixes we see 0.24 for k+y and 6.39 for Chek+y.  

The relative independence of prefixes and suffixes has long been observed and was the motivation for the various "word paradigms", including mine.   While those statistics and word models may not fit "European" languages well, they are quite consistent with Voynichese words being largely monosyllabic. Or with an "European" lan gua ge   spell ed   as   se pa ra te   syll a bles.

Moreover, all VMS text statistics are contaminated by an unknown amount of errors (by the Author, the Scribe, the BEEP!, the transcriber).  In particular, erratic transcription of dubious spaces (",").  So they must be taken with a spoon of salt (tea- or table-, who knows...)

All the best, --stolfi

(13-01-2026, 02:43 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.Actually I write about all this stuff in my work. I encourage you to read it "from board to board"

Will do! But the link in your profile is broken. What is the best pair of boards?

Quote:Actually many of them may come from some 19th century school teacher.

I barely missed having Latin in high school, but colleagues a bit older than me used to say "tintinnabulus sonabit" for "the bell is ringing" -- a joke that apparently started in their Latin classes, when the school bell rang...

All the best, --stolfi

PS. Since your user name is just "Rafal" and you are here in the Voynich playpen, I always assumed that you were Rafal Prinke, which I knew from the old Voynich mailing list days.  But now I see you are not, correct?

---

(13-01-2026, 01:56 PM)dashstofsk Wrote: You are not allowed to view links. Register or Login to view.In addition to there being A and B each section of the manuscript seems to exhibit its own variations. In particular for the language B sections.

That is what one expects from meaningful text, not from random gibberish!

All the best , --stolfi

The Naibbe cipher shows that, at least in theory, there could be a reasonably meaningful text behind the Voynich manuscript. However, in my opinion, this cipher is somewhat anachronistic.

But there are also simpler ways to recreate some of the peculiarities of the Voynich manuscript that are not anachronistic and are based on ciphers of the time.