Interchangeable ch/sh and k/t characters

[Jorge_Stolfi]

I started simply looking at multiple consecutive repeats and since you mentioned taiin isn’t as popular as kaiin

I have an explanation for that...  but it implies a high rate of errors, which seems to be a unpopular idea among the patrons of this pub.

The non-platform gallows k and t are usually tall, but sometimes they are short, no taller than a d. 

The normal way to write a d would be: write an e, then trace a plume from the top of the e that loops down and closes at the bottom of the e.  But sometimes the plume fails to close, leaving the d open at the bottom.  Sometimes the e comes out straight (j). Sometimes both things happen.

Those variants occur occasionally on the vellum copy.  On the Author's draft, presumably they were more common.  But an open-bottom d is barely different from a short k:

   

So my explanation for the excess of kaiin is that the Author's daiins were often misread by the Scribe as kaiins.  But not as taiin, because the second loop of the later.  The confusion probably happened in other words like chedal <-> chekal, but those do not stand out so much in the statistics because those d-words are nowhere as common as daiin.

All the best, --stolfi

[nablator]

"t" seems to occur significantly less frequently with the “aiin” family than "k".

This is true of almost anything, when numbers are high enough, because there are less t than k.

[MarcoP]

"t" seems to occur significantly less frequently with the “aiin” family than "k".

This is true of almost anything, when numbers are high enough, because there are less t than k.

I agree. It's better to work with percent figures, since character counts differ.

But there’s another piece of “evidence” that, for example, t and k aren’t interchangeable:
k before aiin (20.6 percent) (n=810)
t before aiin   (9.0 percent) (n=356)
k before ain  (38.1 percent) (n=645)
t before ain  (12.4 percent) (n=210)
"t" seems to occur significantly less frequently with the “aiin” family than "k". "t" is more likely to appear with ‘o’ and "ch."

The most informative part of JoJo's post was the last line: to/ko tch/kch have similar counts, though t is only ~60% as frequent as k.

Anyway, I am not sure this means that the two "aren’t interchangeable" - they have different preferences in some contexts, but they still look very much interchangeable to me.

[Mauro]

't' and 'k' are surely the most similar characters for what regards the distribution of the preceding and of the following character. The geometrical distances between the distributions of 'k' and 't' are 0.11 (previous character) and 0.13 (following character) [min possible value = 0, max possible value = 1.414]. Notice this does not consider JoJo's post You are not allowed to view links. Register or Login to view. (there the discrepancies are beyond the immediately following character).

'ch' and 'sh' are rather similar too: 0.18 (previous character) and 0.13 (following character). 'cth' and 'cfh' get 0.16 (previous) and 0.13 (following), while 'ckh' and 'cth' have 0.34 (previous, rarther different) and 0.18 (following).

The remaining 'similar' pairs are almost always only for the following character, and they come from the characters which overwhelmingly sit at the end of words (ie. both 'n' and 'm' are mostly followed by 'space': the distance is 0.03 for the following character, but a whopping 1.18 for the previous character).


As an additional observation, 'ok' and 'ot' are even more similar than 'k' and 't' for what regards the following character (distance = 0.11), but rather distant for the previous (0.31)


PS.: all the statistics calculated on the whole RF1a-n transcroption.

[JoJo_Jost]

Any more stats on k and t? If you're interested...

Token start in Currier B: k 10.9% vs. t 16.6% (t 1.53x k). In Currier A they're still close (0.94x); in B the difference is clear.

That sounds like a lot, but it's tied to a clear difference in where the glyphs occur at the line start:

Line start:
Currier A: k 2.09% t 7.96% -> t/k = 3.8x
Currier B: k 0.77% t 6.53% -> t/k = 8.5x

So the effect is really a line-start phenomenon; if you exclude the line start, the difference collapses. But there, it's large and robust.

By the way: if you include the benched gallows (though no one knows if that's valid), there's an interesting reversal in frequency:

plain: k 10,151 / t 6,047 = 1.679
benched (ckh/cth): ckh 895 / cth 916 = 0.977

Despite all the statistics, it ultimately remains speculation. Every language has glyphs that behave similarly and others that do not. Since we cannot assume a substitution, it is a cipher key anyway (if it is a cipher and not a hoax—though in the case of a hoax, it wouldn’t matter at all). And if it is a cipher, the question becomes even more speculative .

[Aga Tentakulus]

Interchangeable in what sense?
Of course they can be interchanged, but the meaning also changes.
Example in German: Ein/eine, Bein/Beine, sein/seine, mein/meine, dein/deine, kein/keine, … etc.
South German spelling prior to 1960.
‘ain/aine’, ‘sain/saine’, ‘kain/kaine’… and so on.
Of course, I can always swap them around somehow, but why bother? It’s wrong; something’s missing – that was added later.
Why are people working on this when it’s supposedly established that it can’t be a text?

As far as I can tell, the meaning changes; I have plenty of variants, and they’re error-free according to the dictionary.
“chol daiin daiin” = = sind alle/ganz zu verwenden   to be used for ‘all’ or ‘entirely’

Translated with DeepL.com (free version)

[ololololo]

Example in German: Ein/eine, Bein/Beine, sein/seine, mein/meine, dein/deine, kein/keine, … etc.

An interesting example. But in the case of Voynichese, we have pairs of letters that can be replaced. According to the example, I can replace any letter from any word with any other letter, and it will also make a meaningful word, but this is not the case: there is no word aaiin in the manuscript, there is no word fkshy, etc.
It is also necessary to take into account the specifics of Voynichese. As much as we would like to, it's not just German

[Aga Tentakulus]

But in the case of Voynichese, we have pairs of letters that can be replaced.

But unfortunately, it isn’t always a two-letter combination.
But that isn’t all that important. It would still be possible.
Klein/kleine, Stein/Steine, ......
But there is an ‘889’ in the manuscript.
We need to understand what something is and when.

I never said that VM is German.

If ‘ch’ is a combination, then ‘c’ is a single letter.
Both are present.

[ololololo]

But in the case of Voynichese, we have pairs of letters that can be replaced.

But unfortunately, it isn’t always a two-letter combination.
But that isn’t all that important. It would still be possible.
Klein/kleine, Stein/Steine, ......
But there is an ‘889’ in the manuscript.
We need to understand what something is and when.

I never said that VM is German.

If ‘ch’ is a combination, then ‘c’ is a single letter.
Both are present.

Pairs of letters - ch/sh and k/t, not bigrams.

[Aga Tentakulus]

Pairs of letters - ch/sh and k/t, not bigrams.

I don’t understand your point of view.
The way I see it, there are two combinations and two individual characters.
But they aren’t the same as, for example, ‘9’ being written as ‘us’. Here, it’s an abbreviation and not a ‘g’.
You need to understand everything, because nothing is really the same.
So ‘4’ and ‘q’ aren’t the same.
I can already hear you saying, ‘It’s the same thing.’
But it isn’t, because the truth is, it’s already been done that way.
   
That is cryptology.