31-03-2026, 02:00 PM
some findings to check for yourself
Finding 1: Labels are provably identifiers
Across every section — pharmaceutical jars, astronomical stars, zodiac nymphs, herbal plants, Rose foldout circle nodes — the label locus type (L) has dramatically higher vocabulary uniqueness ratios than paragraph text (P). Labels are ~0.82-0.91 unique. Paragraphs are ~0.22-0.41. This is exactly what you'd expect from a system where labels = unique node names, and paragraphs = coordinate descriptions using a constrained vocabulary.
Finding 2: The qok- prefix is the most important structural clue
qok-
prefix words (qokeedy, qokedy, qokain, qokal) dominate paragraph text across sections but are almost entirely absent from pure label loci. The base words appear as labels. The qok+base combinations appear in paragraphs. This is consistent with
qok-
being a relational prefix — something like "belonging to / locating at / in correspondence with" the base concept. It's not a word — it's a grammatical operator on a coordinate system.
Finding 3: Locus type predicts text structure perfectly
The IVTFF format distinguishes P (paragraph), L (label), C (circular), R (radial). These aren't just formatting metadata. Each locus type has its own statistical fingerprint — word length, vocabulary density, daiin frequency, qok- density. The text was written with full awareness that different structural positions carry different information types. That's not a cipher. That's a notation architecture.
Finding 4: daiin is densest in Herbal A, sparse in labels
If daiin were a common grammatical word like "the", it would appear everywhere equally. It doesn't. It's concentrated in paragraph text of certain sections and almost absent from pure label fields. This suggests it marks a specific coordinate axis — probably the primary one — and the sections that deal most with that axis use it most.
Finding 1: Labels are provably identifiers
Across every section — pharmaceutical jars, astronomical stars, zodiac nymphs, herbal plants, Rose foldout circle nodes — the label locus type (L) has dramatically higher vocabulary uniqueness ratios than paragraph text (P). Labels are ~0.82-0.91 unique. Paragraphs are ~0.22-0.41. This is exactly what you'd expect from a system where labels = unique node names, and paragraphs = coordinate descriptions using a constrained vocabulary.
Finding 2: The qok- prefix is the most important structural clue
qok-
prefix words (qokeedy, qokedy, qokain, qokal) dominate paragraph text across sections but are almost entirely absent from pure label loci. The base words appear as labels. The qok+base combinations appear in paragraphs. This is consistent with
qok-
being a relational prefix — something like "belonging to / locating at / in correspondence with" the base concept. It's not a word — it's a grammatical operator on a coordinate system.
Finding 3: Locus type predicts text structure perfectly
The IVTFF format distinguishes P (paragraph), L (label), C (circular), R (radial). These aren't just formatting metadata. Each locus type has its own statistical fingerprint — word length, vocabulary density, daiin frequency, qok- density. The text was written with full awareness that different structural positions carry different information types. That's not a cipher. That's a notation architecture.
Finding 4: daiin is densest in Herbal A, sparse in labels
If daiin were a common grammatical word like "the", it would appear everywhere equally. It doesn't. It's concentrated in paragraph text of certain sections and almost absent from pure label fields. This suggests it marks a specific coordinate axis — probably the primary one — and the sections that deal most with that axis use it most.