(07-06-2025, 09:01 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.For example, it could be possible that okechedy is very rare, because it is preferably written opchedy
I am not saying that I think that that is the case. It is just a valid possibility.
If we assume that okechedy = opchedy, and we assume that the preference is to write it opchedy, wouldn't that then require a reason (e.g. subject matter) for the opchedy/okechedy word type to be so predominantly on the top row? And we would likely have to do the same for the other /pch/ and /kech/ words in the list too.
(07-06-2025, 11:25 AM)tavie Wrote: You are not allowed to view links. Register or Login to view.wouldn't that then require a reason (e.g. subject matter) for the opchedy/okechedy word type to be so predominantly on the top row?
Good point.
Every option has its problems though.
For example, why are only
k or
t that are not followed by
e replaced by
f or
p ?
And why are these predominantly on the first rows?
I don't remember exactly the details of Lisa's statistics, they will play a role here.
(07-06-2025, 11:07 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.a radically new script being developed for an existing language
The issue is not that the script is new but is it for an existing language? Almost every mediaeval european language has been tried as a possible fit to the VMS. We have all seen this lately. It is old Bohemian, Slovenian, Turkish, vowel-less Latin, old French. None has succeeded. But perhaps then a non-european language? Possible, but unlikely.
Are you still of the opinion it might be in some Chinese dialect?
(07-06-2025, 11:34 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view. (07-06-2025, 11:25 AM)tavie Wrote: You are not allowed to view links. Register or Login to view.wouldn't that then require a reason (e.g. subject matter) for the opchedy/okechedy word type to be so predominantly on the top row?
Good point.
Every option has its problems though.
For example, why are only k or t that are not followed by e replaced by f or p ?
And why are these predominantly on the first rows?
I don't remember exactly the details of Lisa's statistics, they will play a role here.
The whole thing is of course quite complex, and that is why I am not in favour of drawing quick conclusions.
Just imagine the possibility that
e is a null character.
Now ignoring the fact that this would bring the A and B languages quite a bit closer together, for the question at hand the whole premise changes, or in fact disappears.
There are just too many variables.
(07-06-2025, 11:38 AM)dashstofsk Wrote: You are not allowed to view links. Register or Login to view.Are you still of the opinion it might be in some Chinese dialect?
Yes! But not specifically Chinese. It can be any language with monosyllabic words -- which AFAIK includes not only the Chinese "dialects" (actually more than a dozen distinct languages) but also Tibetan, Vietnamese, Laotian, Burmese, Thai, Hmong, ... And of course not how they are today, but how they were 500 years ago...
Unfortunately, no cipher is as hard to decipher as a natural language...
All the best, --jorge
(07-06-2025, 02:14 PM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.not only the Chinese "dialects" (actually more than a dozen distinct languages) but also Tibetan, Vietnamese, Laotian, Burmese, Thai, Hmong, ...
I think it is unlikely. I have placed my doubts in a new topic. Perhaps you might care to counter my opinions there.
You are not allowed to view links.
Register or
Login to view.
Instance of 2 hapax matching almost all glyphs except
p ≠
t and
ch ≠
c'h
(and gallow bench left leg
c = i, technically)
[
attachment=10797]
Coincidentally two pairs of glyphs not only resembling each other, but also the two benches long-suspected to be equivalent in some way
The fact that they share the same folio and therefore the same context, provides support for a guess that these two vords (that are nowhere near common in the whole MS) are variations or different encodings of the same/similar plaintxt string, or alternatively, a "self-copying" procedure or folio/context-dependent encoding perhaps
(06-06-2025, 10:53 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.However, in the 1970's Currier already demonstrated that p and f are not alternatives for t and k, and more recently, Lisa did some more stats from which she concluded that the following would explain a lot:
f is an alternative form for te and p is an alternative form for ke.
(There is no typo here).
In my view, the proposed replacements—interpreting
f as an alternative form of
te and
p as an alternative form of
ke—introduce several additional issues rather than resolving existing ones:
1. Increased Repetition:
Substituting rare glyphs (
f and
p) with more common sequences (
te and
ke) increases overall redundancy in the text, pushing it further toward excessive repetition. This could reduce the already limited entropy of the script, weakening arguments for meaningful linguistic content.
2. Glyph System Complexity:
The Voynich script already includes multiple sets of related glyphs with overlapping functions. For example,
ch appears to behave like
sh, and we see patterned series such as
i/ii/iii/iiii,
e/ee/eee,
y/o, r/s etc. This suggests that the gallow glyphs may not be exceptional in function, but instead participate in a broader system of visually similar variants. Replacing
f and
p with distinct multi-character sequences may oversimplify a more nuanced internal logic.
3. Word Length Distortion:
On average, words in the first line of a paragraph are longer (5.45 characters) than those in subsequent lines (5.07 characters). If
f is replaced with
te and
p with
ke, this discrepancy becomes even more pronounced.
4. Contextual Behavior of Gallow Glyphs:
There is also a notable positional pattern: when a gallow glyph follows EVA-
l, the most frequent forms are
k (10%) and
f (7.7%), while
p (2.5%) and
t (1.5%) occur less frequently. This suggests that
f behaves more similarly to
k than to
p, undermining the logic behind grouping
f with
te and
p with
ke.
I hope everyone agrees that the Author (who devised the script, chose or created the information in the book, etc.) did not write directly onto vellum. Vellum was expensive (I have seen an estimate of US$1-2 per folio in today's money) and is a pain to erase. He/she must have written a final draft on paper, and only then got it copied onto vellum.
The author may have done this last "vellification" step him/herself, but I think there are good argments and evidence that he/she instead recruited some Scribe (or Scribes, if you will) to do it. For one thing, that last step was boring and tiring work that he/she surely would rather pay someone to do. Also, the characters of the VMS are very small; writing them requires a very sharp pen and a hand that is used to drive it -- an ability that not every scholar/doctor/whatever would have. And there are many clear mistakes by the Scribe that suggest he/she probably did not understand the contents of the book or even the language.
I believe that the Author taught the Scribe the alphabet, had him/her practice copying text until the results were satisfactory, and then let him/her work mostly on hie/her own. The Author presumably dropped in now and then to check the results, resolve any doubts, and improve the instructions (like "You are wasting too much vellum, write more compactly. And try to make the rs and ss more distinct").
Awareness of this process is important when studying questions like the meaning of p and f. Because I believe that both the Author and the Scribe understood that, in running text (such as the bulk of Herbal, Biological, Pharma, and Starred Parags, excluding titles,labels, text rings, etc.) paragraph breaks were important, but line breaks were not. The Scribe respected the parag breaks of the draft, but disregarded its line breaks, and put a line break in the clean copy whenever he/she reached the right margin.
The Author may have marked the parag breaks in the draft in any of several ways -- by wider line spacing, special marks like ¶ • // , indenting the first line, etc. But the Scribe choose to use the method that he/she routinely used in other Latin or vernacular documents, namely using taller and fancier capitals and ascenders on the first letter or the first line of each paragraph.
However, that was not immediately possible because the original Voynichese scriipt had no capital letters. So I guess that the Author and Scribe agreed to mark the paragraphs by fancifying some of the gallows t and k (or common combinations like te and ke) that occurred on the top line of each parag into p and f (or vice-versa); a conversion which the Scribe did mostly on his/her own.
On the other hand, I see evidence that the Scribe in some places failed to respect the parag breaks intended by the Author, and joined two or more of them into a single block of running text, with each parag starting on the same line that the previous one ended. Or that the Scribe sometimes tried to save vellum by squeezing the text at the end of a parag so that it would end on the right margin (rather than end with a line of only one or two words); and then failed to leave extra space and use p/f on the first line of the next parag. Perhaps assuming that other clues would be sufficient, such as the stars in the Starred Parags section.
Whether or not the details of the process happened as suggested above, it is important to take these possibilities into account in any analysis of p/f use, or of word frequency differences at different positions within lines and parags.
All the best, --jorge
(23-06-2025, 07:17 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.I hope everyone agrees that the Author (who devised the script, chose or created the information in the book, etc.) did not write directly onto vellum. Vellum was expensive (I have seen an estimate of US$1-2 per folio in today's money) and is a pain to erase. He/she must have written a final draft on paper, and only then got it copied onto vellum.
As Currier pointed out back in 1976,
“The ends of the lines contain what seem to be, in many cases, meaningless symbols: little groups of letters which don’t occur anywhere else, and just look as if they were added to fill out the line to the margin.” That alone suggests the text wasn’t neatly prepared in advance and then simply copied onto the vellum — instead, it looks like the layout was adjusted on the fly.
There are other clues that point in the same direction. For example, glyphs like EVA-
y,
o,
d, or
s tend to get added to the first glyph group of a line, and the gallow glyphs often appear at the start of new paragraphs. This inflates the length of the first words in a predictable way.
You are not allowed to view links.
Register or
Login to view. ran some stats on this and found exactly that pattern:
- The first word of a line is longer than average (l₁ > l).
- The second word is shorter than average (l₂ < l).
Note: in nearly half the lines (48%), the second glyph group is shorter than the first, and it’s longer in only 32% of cases. [see Timm 2014].
[
attachment=10884]
Another example of ambiguity can be seen on
folio 105r, specifically in lines 9a and 10. While writing line 10, the scribe left a noticeably larger gap. It appears that the scribe wasn’t entirely satisfied with how the layout turned out. To make the gap less obvious, he used larger gallow glyphs and filled the remaining space between them with additional glyph groups.
So, one reasonable explanation for these findings is that the scribe was adjusting the glyph groups as they went, choosing them to make the lines fit neatly into the available space. In other words, the margins weren’t something the author accounted for separately — the text layout and the writing happened together, right there on the vellum.
That’s why I believe the scribe wasn’t simply copying pre-written content but was actively generating the text during the writing process.