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.

---