💾 Archived View for tilde.team › ~aprilnightk › try3301 › 2022_01_09_is_glossary_redtext.gmi captured on 2023-01-29 at 04:30:41. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2022-03-01)
-=-=-=-=-=-=-
This idea struck me in an absolutely random and sudden manner. Felt like an insight, but now I'm a bit less hyped up about it.
Nevertheless, even if it doesn't work out, I need to document this.
The hypothesis G1 is following:
The wordlist at pages no. 71-72 (54-55) is a compilation of words from some subset of the red text from entire LP, with duplicates removed.
W1 will later be shown to not hold.
A weaker hypothesis, G2, is the following:
The wordlist at pages no. 71-72 (54-55) is significantly related to some subset of the red text from entire LP, with duplicates removed.
I will, for the sake of simplicity, refer to the wordlist at pages no. 71-72 (54-55) as "the glossary".
Click here to view the glossary word list, and some other considerations.
This article will make heavy use of ciphertext, and I can't be sure that the Unicode runic symbols are properly shown in all of the reader's clients, so I will make use of my homebrew transliteration for the runic symbols.
This is a transliteration that I call "unilatin", because it provides a strict 1-to-1 transliteration to runes.
Click here to view Gematria table with "unilatin" added.
Also, all ciphertext will be provided in small letters, and all cleartext in capital letters.
The red text of Liber Primus is not totally homogenous, so that gives a leeway for interpretation.
For the sake of clarity, I will refer to certain classes of red text.
Those are the main instances of redtext that start chapters and paragraphs. Those are the ones that start with big red runes. Following is the complete collection of them: those that were already decrypted - in capital cleartext, those that are still unknown - in small unilatin ciphertext.
A WARNŊ WELCOME WISDOM A COAN AN INSTRUCTÄN ÞE LOSS OF DIUINITY SOME WISDOM AN INSTRUCTÄN A COAN AN INSTRUCTÄN sheogmäf syeŋc lj ǫhŋcþtæjtxh tuœ pdþ atxb m gäsb xǫ xixm uxmcþpob luŋn cu bntfŋ ua wŋgxdg ibi ǫtbiy mpy lsoaæhleǫiml lscp dǫ xcfiawhg ŋþǫ ifcœǫ äcs fnm ygdæh a mŋdäþxim AN END PARABLE
Notably, the "a mŋdäþxim" is the header of the very glossary in question. (It could even mean "a glossary").
Those are the ones that look like headers or subtitles of some sort but lack the big red rune. Also, subtitles like "know this" before the magic tables are in this category.
SOME WISDOM CNOW ÞIS CONSUMPTÄN PRESERUATÄN ADHERENCE CNOW ÞIS fulm aąyœa
Some red text are just phrases, mostly at the end of the chapter (four such phrases were found).
FOR ALL IS SACRED AN INSTRUCTÄN COMMAND YOUR OWN SELF do œtþæ cwj xą gąm tpwǫs wbǫþ nhgj rädäþaiǫæxl ǫą äaj hr þpdjŋo gfǫmc hdotc ur pn
For the sake of being pedantic, the only things left that are red in LP are numbers and decorations like punctuation marks. I will disregard the latter. Here are the numbers that are highlighted red in LP:
1 one 2 two 3299 three thousand two hundred ninety nine 1 one 2 two 3 three 4 four 5 five
The idea emerged upon studying the word lengths of the words in the "glossary" in relation to the word lengths in the portions of LP known as "red text".
Following are four strings of integers that represent raw runeword lengths, sorted in descending order.
THE GLOSSARY:
13 11 10 10 8 8 8 8 8 7 7 7 6 6 6 6 6 6 6 6 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1
RED TEXT 1:
12 11 10 10 10 8 8 8 8 8 7 7 6 6 6 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1
RED TEXT 1+2:
12 11 11 10 10 10 10 9 8 8 8 8 8 7 7 6 6 6 6 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1
RED TEXT 1+2+3:
12 11 11 11 10 10 10 10 10 9 8 8 8 8 8 7 7 7 6 6 6 6 6 6 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1
This results in the following distibution of lengths:
LEN || GLO | R-1 | R-12 | R-123 ----- || ----- | ----- | ----- | ----- 13 || 1 | 0 | 0 | 0 12 || 0 | 1 | 1 | 1 11 || 1 | 1 | 2 | 3 10 || 2 | 3 | 4 | 5 9 || 0 | 0 | 1 | 1 8 || 5 | 5 | 5 | 5 7 || 3 | 2 | 2 | 3 6 || 8 | 3 | 4 | 6 5 || 9 | 6 | 7 | 10 4 || 9 | 9 | 13 | 18 3 || 14 | 8 | 10 | 16 2 || 11 | 11 | 11 | 19 1 || 1 | 5 | 5 | 5
Immediately obvious are these facts:
So far, the 6-rune words seem to be the biggest problem.
As stated in hypothesis G1, the duplicates are removed from the red text, and by extension the words in the glossary are unique.
We don't know the whole of redtext yet, but we already have some duplicates we could get rid of.
We will remove the following words (except for their first occurence). We will also remove all 1-letter words but one.
A WISDOM AN INSTRUCTÄN A COAN AN INSTRUCTÄN AN SOME WISDOM CNOW ÞIS AN INSTRUCTÄN
Of course, there may still be duplicates in the redtext that's unknown, but I'm trying to get as close to uniqueness as I can.
LEN || GLO | R-1 | R-12 | R-123 ----- || ----- | ----- | ----- | ----- 13 || 1 | 0 | 0 | 0 12 || 0 | 1 | 1 | 1 11 || 1 | 1 | 2 | 3 10 || 2 | 1 | 2 | 2 9 || 0 | 0 | 1 | 1 8 || 5 | 5 | 5 | 5 7 || 3 | 2 | 2 | 3 6 || 8 | 2 | 2 | 4 5 || 9 | 6 | 7 | 10 4 || 9 | 8 | 10 | 15 3 || 14 | 8 | 9 | 15 2 || 11 | 8 | 8 | 15 1 || 1 | 1 | 1 | 1
13 and 12: situation stays the same.
11, 5, 4, 3, 2: there are less words of these runelengths, which fits the G1.
10, 8, 7, 1: there are exacly the same amount of words of these runelengths in the glossary and in R-12, which fits the G1 perfectly.
9: a 9-rune word is missing from the glossary, which is curious. It's a stretch, but perhaps one of the 8-rune words is also used in plural form somewhere in headers, and thus does not appear in the glossary.
6: there are 8 six-rune words in the glossary, but only 4 of them in entire R-123, which is actually troubling.
I was adviced to look at how this distribution compares against a regular English text.
I took a well-known book in English and yielded some 1976 unique runic representations of English words from it.
I took the percentage distribution of lengths and multiply it by 64 (amount of words in the glossary) to take a comparison. I rounded the fractions to closest integer.
LEN || GLO | R-1 | R-12 | R-123 | BOOK ----- || ----- | ----- | ----- | ----- | ----- 13 || 1 | 0 | 0 | 0 | 0 12 || 0 | 1 | 1 | 1 | 0 11 || 1 | 1 | 2 | 3 | 0 10 || 2 | 1 | 2 | 2 | 1 9 || 0 | 0 | 1 | 1 | 3 8 || 5 | 5 | 5 | 5 | 5 7 || 3 | 2 | 2 | 3 | 9 6 || 8 | 2 | 2 | 4 | 12 5 || 9 | 6 | 7 | 10 | 13 4 || 9 | 8 | 10 | 15 | 13 3 || 14 | 8 | 9 | 15 | 6 2 || 11 | 8 | 8 | 15 | 1 1 || 1 | 1 | 1 | 1 | 0
Runelengths in the glossary visibly peak earlier, which makes sense: in a shorter text, a short word is much more likely to have a larger proportion compared to longer words, because there are simply more longer words than short ones.
I could take random selections of 64 words from the same book, run it several time and figure out the mean runelength distribution in shorter "sample texts", but the result would have ultimately been the same.
All that's possible to do now is to disregard the discrepancies so far (who knows, 3301 could distort the word selecting process?) and take a look at what seems to work.
The solved portion of LP supplies us with all the 7- and 10-letter words we need.
There are no such words left in the yet undecrypted redtext, so we can, for the sake of trying, assume that the glossary contains these tunewords:
PARABLE COMMAND WELCOME INSTRUCTÄN CONSUMPTÄN
If this makes sense then there's only two possibilities to place "instruction" and "consumption" in the glossary:
23 d >> ( 13 p | 16 t ) >> i 10 7 w >> ( 27 io | 2 th ) >> n 9 8 h >> ( 22 oe | 7 w ) >> s 15 17 b >> ( 1 u | 28 ea ) >> t 16 21 ng >> ( 17 b | 12 eo ) >> r 4 0 f >> ( 28 ea | 1 u ) >> u 1 24 a >> ( 19 m | 10 i ) >> c 5 5 c >> ( 18 e | 11 j ) >> t 16 11 j >> ( 13 p | 16 t ) >> io 27 1 u >> ( 21 ng | 8 h ) >> n 9 =========================================== 8 h >> ( 3 o | 26 y ) >> c 5 28 ea >> ( 25 ae | 4 r ) >> o 3 24 a >> ( 15 s | 14 x ) >> n 9 5 c >> ( 19 m | 10 i ) >> s 15 10 i >> ( 9 n | 20 l ) >> u 1 19 m >> ( 0 f | 0 f ) >> m 19 14 x >> ( 1 u | 28 ea ) >> p 13 11 j >> ( 24 a | 5 c ) >> t 16 19 m >> ( 21 ng | 8 h ) >> io 27 18 e >> ( 9 n | 20 l ) >> n 9
23 d >> ( 18 e | 11 j ) >> c 5 7 w >> ( 4 r | 25 ae ) >> o 3 8 h >> ( 28 ea | 1 u ) >> n 9 17 b >> ( 2 th | 27 io ) >> s 15 21 ng >> ( 20 l | 9 n ) >> u 1 0 f >> ( 10 i | 19 m ) >> m 19 24 a >> ( 11 j | 18 e ) >> p 13 5 c >> ( 18 e | 11 j ) >> t 16 11 j >> ( 13 p | 16 t ) >> io 27 1 u >> ( 21 ng | 8 h ) >> n 9 =========================================== 8 h >> ( 27 io | 2 th ) >> i 10 28 ea >> ( 19 m | 10 i ) >> n 9 24 a >> ( 9 n | 20 l ) >> s 15 5 c >> ( 18 e | 11 j ) >> t 16 10 i >> ( 6 g | 23 d ) >> r 4 19 m >> ( 18 e | 11 j ) >> u 1 14 x >> ( 9 n | 20 l ) >> c 5 11 j >> ( 24 a | 5 c ) >> t 16 19 m >> ( 21 ng | 8 h ) >> io 27 18 e >> ( 9 n | 20 l ) >> n 9
Obviously, the next logical step is to try and procrue the possible streamkey from what we have, and that's where I'm stuck for now. Neither of the two placements reveal any immediately obvious key (not that I expected any).
If all this somehow holds true, then decrypting the glossary will give us a lot of insight on what the remaining redtext can be, thus simplifying the cribbing process. It sounds promising, but as was shown there are several shortcomings that could potentially indicate that the whole hypothesis is bogus.
- - - - - - -
Keith Aprilnight (aprilnightk@tilde.team)