Critical edition · Formal grammar · Reproducible pipeline

A structural meta-grammar of Indian classical music, extracted from primary sources.

Built around three pilot treatises (Mātaṅga's Brihaddeśī, c. 750 CE ; Joep Bor 1999 ; van der Meer 2000). The pipeline turns OCR'd pages into concepts, rules, and an executable formal grammar — with verbatim evidence at every step. No claim without a sourced quote.

274OCRed pages
1,763concepts
3,187sourced affirmations
1,087generative rules
11formal modules
86operations
Read

Critical editions →

Bilingual rendering (devanāgarī + English Sharma translation), marginal apparatus, hyperlinked concepts.
Explore

The grammar →

From 1,763 concepts through 1,087 rules to an executable formal grammar with 11 modules.
Method

How it was built →

A 30+ step reproducible pipeline : OCR → linguistic validation → semantic tokenisation → generative rules → formal grammar.

The epistemological arc

  1. Pages — raw OCR, two-language facing layout preserved.
  2. Affirmations — every claim in the text, with a verbatim quote.
  3. Concepts — clusters of mentions, the named entities the text talks about.
  4. Generative rules — what the text asserts about each concept, attested by ≥1 affirmation.
  5. Formal grammar — types, constants, operations and constraints, synthesised per domain. Every value carries its evidence rule_id ; unsourced values stay in an explicit UNRESOLVED list.