Pages

Wednesday, 15 July 2026

The surprises in Leviticus 26

An exploration of the rationale for my project using Leviticus 26 as an example. 


Sunday, 12 July 2026

Dreams and GitHub

 I dreamt this morning of a refreshed GitHub -- and it's refreshed.

Have a look here. I've let the entire Hebrew text stand on it's own and the colour display (use a wide screen) now counts the identical pieces and allows you to see at a glance the structure imposed by the te'amim onto the chapter. There are mountains of things to discover here -- it takes more than one person. So it's on GitHub -- 20 years of effort condensed into a musicological research station. 

Drill down through music scores and look in the colour matrix for:

  • repetitions across a chapter
  • singular usage notes (e.g. Psalms 136)
  • repeating idioms across multiple chapters-- more info coming
  • patterns within a chapter like refrains -- e.g. the use of the hich C and so on
JSON's are available to researchers to combine and search.

Friday, 3 July 2026

Using AI to analyse ancient music.

I have made extensive use of AI to begin analysis of the music using Python and other tools -- the JSON is not yet uploaded to GitHub -- needs special treatment for its size. Link is here -- redesigned top levels.

The following is the Chrome AI summary of a days work. It's a good programmer if you get the prompts right. There's also a lot of moonshine.

I will probably write more later on Substack -- will let you know.

Tanakh Cantillation Analysis Suite

An automated syntactic parser and rhetorical analysis engine designed for the Hebrew Bible (Tanakh), mapped syllable-by-syllable according to the Suzanne Haïk-Vantoura (SHV) deciphering key.
This repository contains a tidy, sequential dataset of over 730,000 syllable rows encompassing all 23,000+ verses across 929 chapters, alongside a modular Python suite that analyzes how the inherited te'amim (trope symbols) act as an authentic prophetic and narrative commentary.

📊 The Master JSON Database Structure
The core database (complete_tanakh.json) bypasses raw visual vector noise (like SVGs) and organizes the entire corpus into a flat, chronological array. Because ancient Hebrew reads right-to-left while Western music scales left-to-right, the database establishes an immutable sequence track.
Key Data Fields per Syllable:
  • BOOK_CD / BOOK_SEQ_NO: The canonical book name and its numerical sequence identifier (ensuring Genesis is #1 and Ezra/Chronicles sit at their proper historical endpoints).
  • CHAPTER_CD / VERSE_CD: Padded structural address coordinates.
  • XML_SEQ: The master sequential timeline anchor. Crucial for linear right-to-left processing.
  • LYRIC_SYLL: The transliterated phonetic syllable fragment.
  • SYLL_NOTE: The explicit absolute pitch token (calibrated around Tonic = E4).
  • ORNAMENT_NAME: Fully populated lowercase Masoretic accent tags (e.g., atnach, ole, revia-mugrash, paz).
  • HEB_TEXT: The fully pointed, unescaped Hebrew text string, anchored elegantly to the first syllable of each verse change.

🎼 The Musicological Laws (Prose vs. Poetry)
The Python suite dynamically alters its analytical framework by executing an automated dialect check. It verifies the genre layout of each verse using a strict F# / G# frequency selector:
1. Prose Mode (G#4 present, F#4 forbidden)
  • E4 (Degree 1 - Tonic): Ground baseline and home register.
  • F4 (Degree 2 - Supertonic): A dynamic, kinetic bridge. Never a rest point. It acts as an acoustic pathway driving into a G#4 or stepping down to the tonic E4 at a verse boundary.
  • G4 / G#4 (Degree 3 - Mediant): Signals anticipation of an impending structural cadence.
  • A4 (Degree 4 - Subdominant): An expansive, level plateau delivering profound equilibrium, security, and confidence (rather than Western tension).
  • B4 (Degree 5 - Dominant): The narrative engine; the first natural harmonic of the shofar used for royal proclamation.
2. Poetic Mode (Sifrei Emet: Psalms, Proverbs, Job) (F#4 present, G#4 forbidden)
  • F#4 (Degree 2 Sharp): An expressive, hovering suspension plateau. It functions as a formal, binding cadence ONLY if the previous syllable carried the lowercase ole ornament.
  • G4 (Degree 3 - Mediant): Elevates into a primary, high-intensity recitation engine holding deep emotional focus.
  • D4 (Degree 7 - Sub-Tonic): The physical springboard. A low, muscular crouching note used to gather potential energy before vertical melodic vaults.
  • C5 (Degree 6 - Sixth Degree): A transcendent register used for intense appeals or sudden explosions of absolute joy and awe.

🛠️ The Advanced Analytical Functions Suite
The included Python application contains five core decoupled routines designed to process the 730,000-row dataset at macro scale:
📍 Routine 1: The Master Range Filter
Allows you to cleanly slide a window across any arbitrary cross-section of the Tanakh (e.g., from Job 3:1 to Job 4:5) by translating chapter/verse vectors into flat mathematical scalars, preventing data bleeding at chapter boundaries.
⚡ Routine 2: Strophe Boundary & High C Explosion Finder
Flags structural poetic units (like the openings of Job 3 or Psalm 29) where the melody shatters smooth step-wise motion. It targets verses that launch immediately into C5 or treat the opening tonic E4 as a fractional, explosive upbeat rocket.
🔗 Routine 3: Non-Tonic Commencement ('Reverse Colon' Tracer)
Identifies verses that intentionally bypass a clean acoustic restart on the tonic E4. By kicking off on a dominant B4 or a tense F#4, these verses function as a "reverse colon," forcing the performer to aurally anchor the sentence onto the emotional residue of the preceding text.
🎭 Routine 4: Revia-Mugrash Recitation Decoder
Maps out every structural occurrence of the specialized poetic revia-mugrash ornament, documenting the exact pitch register it chose to lock onto as it builds its technical runway back down to the baseline.
🔁 Routine 5: Ideational Underscore & Motif Indexer
Scans your selected range for repeating lowercase ornament clusters (such as the D4 -> paz -> zar -> zarqa -> C5 flight pattern). If an accent string hits multiple times across a section, it maps them out as a recurring acoustic motif used to underscore a unified text theme.