08-04-2026, 06:02 PM
Another important point when searching for the core elements is that the gaps in the text are not simply spaces, but, in my opinion, syllabic breaks.
We all know that:
y → q: 3,499 times via the space bar, only 27 times internally!
y → o: 2,625 times via space
n → ch/sh/o/d: almost exclusively via space
y → l/r: almost exclusively via space
l → q: almost exclusively via space
The pattern: y, n, r, l almost always end a token. q, o, ch, sh almost always begin a token. The spaces are thus determined by the glyphs themselves, not by anything else—in this respect, they are structural breaks, not necessarily word boundaries!
What does this mean for the model?
We will be dealing with a two-stage positioning—especially the prefixes, and here particularly the “y,” will represent one or more bi- or trigrams that can appear both within a word and at the end of a word.
In German, for example, the most common final syllables also appear within words, such as: “en” in Middle High German: 22.6% within words, “er”: 30% within words, “el”: 69% within words, etc.
In this respect, Eva y / L / r will likely play a dual role precisely there. Which one?
Yes, that is the question of all questions. While vowels are still relatively easy to identify through frequency analysis, consonants are significantly harder to find. Especially when a VMS glyph absorbs more than one consonant. For the system to work, they must absorb an average of about three. And that is nearly impossible to decipher by simple means.
So we have to find new ways—I’m on it....
We all know that:
y → q: 3,499 times via the space bar, only 27 times internally!
y → o: 2,625 times via space
n → ch/sh/o/d: almost exclusively via space
y → l/r: almost exclusively via space
l → q: almost exclusively via space
The pattern: y, n, r, l almost always end a token. q, o, ch, sh almost always begin a token. The spaces are thus determined by the glyphs themselves, not by anything else—in this respect, they are structural breaks, not necessarily word boundaries!
What does this mean for the model?
We will be dealing with a two-stage positioning—especially the prefixes, and here particularly the “y,” will represent one or more bi- or trigrams that can appear both within a word and at the end of a word.
In German, for example, the most common final syllables also appear within words, such as: “en” in Middle High German: 22.6% within words, “er”: 30% within words, “el”: 69% within words, etc.
In this respect, Eva y / L / r will likely play a dual role precisely there. Which one?
Yes, that is the question of all questions. While vowels are still relatively easy to identify through frequency analysis, consonants are significantly harder to find. Especially when a VMS glyph absorbs more than one consonant. For the system to work, they must absorb an average of about three. And that is nearly impossible to decipher by simple means.
So we have to find new ways—I’m on it....