(21-02-2026, 03:19 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.None: glyphs are not letters. Calling them letters presupposes a simple substitution is possible.
No it doesn't? Saying that each symbol represents a letter does not presuppose that the configuration of those letters are the final plaintext.
Mapping
qokeedy -> qokeedy does not presuppose that "qokeedy" is the original, unaltered plaintext. For all we know, the voynich author may have genuinely meant to write "qokeedy" as the encoded ciphertext, but wrote each letter as a different glyph (1-1 substitution).
(24-02-2026, 06:31 PM)eggyk Wrote: You are not allowed to view links. Register or Login to view. (21-02-2026, 03:19 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.None: glyphs are not letters. Calling them letters presupposes a simple substitution is possible.
No it doesn't? Saying that each symbol represents a letter does not presuppose that the configuration of those letters are the final plaintext.
It presupposes that there is a plaintext and that there is a mapping from symbols to letters of the plaintext. The word "symbol" is also problematic, it presupposes that we can decompose Voynichese into a set of symbols with any confidence that the choices we made are correct. This is clearly not true. Some transliteration systems are more analytical than others, some on the contrary group more strokes than others...
(24-02-2026, 07:24 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.It presupposes that there is a plaintext and that there is a mapping from symbols to letters of the plaintext. The word "symbol" is also problematic, it presupposes that we can decompose Voynichese into a set of symbols with any confidence that the choices we made are correct. This is clearly not true.
When you say plaintext, do you mean the original content that the author could freely read? Or are you counting encoded text as "plaintext"?
If I write
ELBL MJ LSII EEZFEIQM, which is a set of voynichese glyphs representing "ELBL MJ LSII EEZFEIQM" (which is ciphertext that would have to be deciphered), is that the "plaintext" in your view?
Because in my view, the voynichese to latin letters in this example is not a mapping of voynichese glyphs to plaintext. Its a mapping of voynichese glyphs to ciphertext. None of this presupposes that "ELBL MJ LSII EEZFEIQM" would then undergo simple substitution to reveal the original plaintext. So why is simple substitution a valid reason to say that they cannot represent letters? A simple substitution to
ciphertext is completely possible with what we know about the VMS, surely.
Also, we clearly are able to seperate out certain voynichese symbols and differentiate at least
some from
some others. Whether we do that completely correctly, or whether we know we have done this completely correctly, has no bearing on whether it is truly the case.
We don't know if they are letters. We can't know if they represent letters. We therefore can't say that they are
definitely not letters.
I am asking people to guess at the abstract representation of whatever underlies the text!
And if your best guess is that there is no underlying representation, then it is fair to say there are no letters and its *exactly* the information I was soliciting
And in that vein let me give my answer:
12 is my "I'm pretty sure these are separate letters" guess
i e
o a
d s
l r
ch
q f p
The arrangement here should hint at my thinking.
When I am feeling ornery I am a radical CLS adherent and think there might only be 2 (e and i) and everything else is just composed of them in some elaborate and possibly flexible system, but idk.
(24-02-2026, 08:16 PM)eggyk Wrote: You are not allowed to view links. Register or Login to view.When you say plaintext, do you mean the original content that the author could freely read?
That anyone who is literate can read without learning a different alphabet or cryptography technique.
(24-02-2026, 09:53 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.That anyone who is literate can read without learning a different alphabet or cryptography technique.
I think that the definition is wider than that, no? I would say that the "plaintext" is more like "the text that was meant to be hidden by the encryption". For example, it could be a list of the coordinates of the enemy's submarines. Only some naval intelligence guy will be able to "read" that, maybe.
And that supposes that the text
was encrypted. No one can read the Linear A tablets, but they are still "plaintext" because the authors were not trying to hide the text.
You know about the "Navajo encryption", right? Would you consider a transcript of one of their messages a "plaintext"? If not, what would be the "plaintext"?
All the best, --stolfi
(24-02-2026, 09:53 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.That anyone who is literate can read without learning a different alphabet or cryptography technique.
Exactly, so if the VMS text is ciphertext, it's entirely possible that each glyph represents a letter in that ciphertext. If that's the case the glyphs are effectively letters. But I suppose its not the point of the thread so i'll leave it.
(24-02-2026, 08:48 PM)rikforto Wrote: You are not allowed to view links. Register or Login to view.When I am feeling ornery I am a radical CLS adherent and think there might only be 2 (e and i) and everything else is just composed of them in some elaborate and possibly flexible system, but idk.
I had similar-ish thoughts of some kind of system when looking at f57v. Seeing how the glyphs are ordered makes me think that some glyphs are "built" from others.
Like:
o +
l =
d
o +
v =
l
Or that some glyphs are capital version of others, like
r r ->
m ->
k
So I don't really have any letters that I definitely think are their own unique thing for sure. There are always plausible variances that you can make that groups them one way or the other and i'm open to a lot of them.
Anyway.. If I were to put what my answer/head canon is:
o,
l,
d,
q, f, k, as unique characters
t ≈
q +
k
p ≈
q +
f
s ≈ variant of
e e (but different, like an accented
e or é)
i =
n
e =
c
a =
y
r ≈
m
h = [
attachment=14369]
I suppose that puts me at 11 seperate characters.
(21-02-2026, 02:12 PM)rikforto Wrote: You are not allowed to view links. Register or Login to view.Feel invited to also expand your answer into a range, explain your cutoff, list your alphabet, and generally explain yourself. I'm curious about all that too!
This is just a casual educated guess, right? In that case, about thirteen common ones:
1. The "curve" [e] -- also first component in [a], [o]; doubled in [ch] etc.
2. The "line" [i]
3. The up-flourish on [s], [r]
4. The loop-flourish on [g], [m]
5. The mid-flourish on [n], [o], and [b]
6. The down-flourish on [y], [l]
7. The "bar" in [ch] etc.
8. The plume on [S]
9. [q] with its forward projection
10. The upper part of [p], [t]
11. The upper part of [f], [k]
12. The lower part of [p], [f]
13. The lower part of [t], [k]
These would be "graphemes" rather than an alphabet. If there's a corresponding "alphabet," I'd suspect that its relationship to component graphemes could be loosely analogous to that of Hangul, Melville Bell's Visible Speech, Roman numerals in the sense that [x] and [xx] and [x̄] have different derived "meanings," or the character sequences of Stenotype -- with emphasis on "loosely."
Of course, I've been thinking about this question a lot over the years, and this thinking has evolved. Each time I wondered how I should answer here, I came to the only valid conclusion that "Indeed I really don't know".
But that isn't very satisfactory.
Trying to separate the question into two aspects: how many different units of writing vs. how many different units of meaning isn't really working. The latter remains out of reach, but the former still cannot be separated entirely. Following reflects some of my current thinking. (Not entirely, and I also have conflicting ideas....)
Even as it is inconsistent, I would treat
e
as a single unit of writing. This means that ee and eee are not units of writing (but they may be units of meaning, or even only parts of units of meaning).
On the other hand, I would treat
i, ii and iii as three different units. I would also consider n a variant of i, so not different.
o, a, and y are three units of wrtiting, but I tend to believe that they mean the same.
That leaves us at 7, for the 'Eva vowels'.
All of the following is referring to units of writing, and my thoughts on units of meaning are quite different.
k and t I consider two different units, but f and p to be variants of 'something'.
ch and Sh: sitting on the fence here. Count as two to keep it simple.
cTh, cKh, cPh, cFh, all combinations, or rather ligatures. True also for the more complicated ones.
That means four more, from the gallows and benches.
d is the most regular character of all, count as one.
s is a bit more mysterious but also count as one.
l and r are two more, but m and g are variants of others.
qo (together) is one more.
That give us 5 more, and we are now at 7 + 4 + 5 = 16.
Remaining characters: j x b v u z are all rare and I would not count them.
16 is a bit low, for example when compared to Latin (21), but not extremely so.
(In Latin, I left out j, k, w, y, z)