DNS Bhat Ettuge Project — Recap

Last updated: 2026-03-26 (Phase 34, reverse-chronological)


ಪರಿವಿಡಿ — Contents

Section  
What Is This Project? Overview, About DNS Bhat, Catalogue
Phase History (newest first) PROJECT-PHASES — 5 browsable groups
Eke Romanisation System Rules, consonants, vowels
Current File Status Per-book file inventory
Pending Work TBD items
File Structure Summary Directory layout
Tools & Pipeline OCR, WX decode, CI scripts
Key Intellectual Themes Research themes

What Is This Project?

The ettuge DNS Bhat exercise is a systematic effort to collect, digitise, structure, and make accessible the complete works of Dr. D. N. Shankara Bhat — Kannada’s most important modern linguist.

The work lives at:

/Users/vishwas/code/ettuge/src/main/md/kannada/dnsbhat/

The project has three simultaneous goals:

  1. Text recovery — Get readable Unicode text from all available sources (YouTube, website, PDF, DjVu, blog, OCR)
  2. Structured annotation — Produce per-book files in a consistent format: kn.md (Kannada), en.md (English summary), kn-eke.md (Eke romanisation), claude-prompt.md (AI context primer)
  3. Ellara Kannada (Eke) romanisation — Render Bhat’s Kannada in his own ellara kannaDa system so that non-Kannada readers and AI models can engage with the phonological content directly

About DNS Bhat

Dr. D. N. Shankara Bhat (Dayadhara Nagesha Shankara Bhat) was born in 1931 in the Havyaka Brahmin community of coastal Karnataka.

He held positions at Deccan College (Poona), Trivandrum, Manipur, Leeds, Stanford, Antwerp, the Max Planck Institute, and La Trobe. He won the Pampa Award, Karnataka’s highest literary honour.

His work has two pillars:

Pillar Works Audience
International typology Pronouns (Oxford, 2004), Sound Change (Motilal, 2001), The Koraga Language (1971), An Outline Grammar of Havyaka (1971), Grammatical Relations (journal) Academic, English
Kannada language reform Sollarime (7 vols), Mahaprana Yake Beda, Baasheya Bagge, Hosapadagalannu Kattuva Bage, Nijakku Halegannada Vyakarana, and 20+ more General Kannada readers

His central argument, developed over 50 years: Kannada grammar must be built from Kannada’s own structure, not borrowed from Sanskrit or English frameworks.


The Catalogue: 33 Known Works

The full registry is in README.md. Summary:

# Short Title Language Key Topic Text Status Processed
01 Idu Kannadade Vyakarana Kn Native grammar framework ✅ Transcript ✅ youtube+en
02 Hosapadagalannu Kattuva Bage Kn Word formation ✅ Transcript + 15 blog posts ✅ full
03 Kannada Padagala Olarachane Kn Morphology ✅ Transcript + OCR ✅ full
04 Mathu Matthu Barahada Gondala Kn Speech vs. writing ⚠️ Partial (25/44 parts) ✅ full
05 Mathina Olaguttu Kn Deep structure of language ✅ Transcript (27/37 parts) ✅ full
06 Kalikenudi Matthu Nudikalike Kn Language acquisition ❌ Corrupted ✅ youtube+en
07 Kannada Barahada Sollarime (7 vols) Kn Complete grammar of written Kn ⚠️ OCR vol1–4; vols 5–7 missing ✅ full
08 Kannadakke Mahaprana Yake Beda Kn Script reform (aspirates) ✅ Full text ✅ full
09 Havyaka Kannada Kn Havyaka dialect (popular) ⚠️ Partial (72/88 slots) ✅ full
10 Kannada Nudiya Hinnadavali Kn History of Kannada ⚠️ Corrupted ✅ youtube+en
11 Kannada Barahada Padasamasye Kn Orthographic problems ❌ Corrupted ✅ youtube+en
12 Kannada Bhasheya Kalpita Charitre Kn Reconstructed history ✅ Short excerpt ✅ youtube+en
13 Dharege Doddavaru Kn Old Kannada literature ❌ Corrupted ✅ youtube+en
14 Nijakku Halegannada Vyakarana Entahadu Kn Old Kannada grammar ✅ Full text + 7 blog posts ✅ full
15 Inglish Kannada Padanerake Kn English→Kannada vocabulary ⚠️ Sample only (53p) ✅ full
16 Samskruta Padagalige Kannadade Padagalu Kn Sanskrit→Kannada terms ❌ Not collected
17 Kannada Nudi Nadedu Banda Dari Kn Path of Kannada language ✅ OCR ✅ full
18 Kannada Nudiya Bagege Chintane Kn Reflections on Kannada ✅ 13 blog posts ✅ full
19 The Koraga Language En Koraga (endangered Dravidian) ❌ PDF not extracted
20 An Outline Grammar of Havyaka En Havyaka grammar (academic) ✅ Full DjVu text ✅ en+prompt
21 Pronouns (Oxford) En Pronoun typology ❌ PDF not extracted
22 Sound Change En Historical phonology theory ❌ Not collected
23 A Grammar of Manipuri En+Kn Manipuri grammar ❌ Not collected
24 Grammatical Relations En Syntax theory (journal) ❌ Not collected
25 Kannada Vakyagala Olarachane Kn Sentence structure ✅ OCR ✅ full
26 Uli Marpadina Geregalu Kn Sound change laws ❌ Cloudflare-blocked
27 Baasheya Bagge Kn Introduction to linguistics ✅ OCR ✅ full
28 Kannadakke Beku Kn Why Kannada needs its grammar ✅ OCR ✅ full
29 Kannada Vyakarana Yaake Beku Kn Why grammar matters ✅ OCR ✅ full
30 Kannada Barahavanna Saripadisona Kn Practical script reform guide ✅ Nudi OCR decoded ✅ full
31 Ingliš Padagalige Kannadade Padagalu Kn English→native Kn dictionary (A–Z) ✅ Baraha CID decoded (Phase 27) ✅ raw+en
32 Prominence of Tense, Aspect and Mood En TAM typology (John Benjamins) ✅ Clean English PDF ✅ en+prompt
33 Kannada Sollarime Kn General Kn grammar (YouTube) ✅ Transcript ✅ youtube+en

“Processed” = full means the book has all four key files: en.md + kn-eke.md + claude-prompt.md (+ kn.md where applicable). “en+prompt” = English summary + claude-prompt but no kn-eke. “youtube+en” = YouTube transcript + English summary. “raw+en” = decoded raw.md + English summary (no kn-eke yet).


File Naming Convention

Each book folder follows this pattern:

{NN}-{slug}/
  {NN}-{slug}-book.md      # Raw extracted/OCR text
  {NN}-{slug}-kn.md        # Structured Kannada version with TOC + anchors
  {NN}-{slug}-kn-eke.md    # Eke romanisation companion
  {NN}-{slug}-en.md        # English chapter summaries
  {NN}-{slug}-claude-prompt.md  # AI context primer (key terms, thesis, chapter map)
  {NN}-{slug}-blog.md      # Blog posts (where applicable)
  {NN}-{slug}-djvu.md      # DjVu OCR text (where available)
  {NN}-{slug}-website.md   # Website/online content stub
  {NN}-{slug}-sample.md    # Sample pages

The type suffix taxonomy: book · kn · kn-eke · en · claude-prompt · blog · djvu · website · sample


What Has Been Done

Phase history (34 phases) is in PROJECT-PHASES — split into 5 browsable groups.

Eke Romanisation System

Ellara KannaDa (Eke) is a romanisation of Kannada devised by Vishwas — inspired by HK protocol and DNS Bhat’s ideas, designed to be learnable by any Indian and usable by non-Kannada readers. It is the romanisation used throughout the -kn-eke.md files.

Core Rules

Feature Convention Examples
Short vowels lowercase a i u e o
Long vowels UPPERCASE A I U E O
Aspirates preserved ಭ→bh, ಧ→dh, ಖ→kh, ಥ→th, ಫ→ph bhAShe, adhyayana
Retroflexes UPPERCASE consonant T D N L (ಟ ಡ ಣ ಳ)
S viSwAsa
Sh santOSha
h hesaru
Anusvara assimilation context-sensitive nasal — never M ಂಕ→nk, ಂಗ→ng, ಂತ→nt, ಂದ→nd, ಂಪ→mp, ಂಬ→mb
N exclusively ಣ (retroflex nasal) — never anusvara kaNNu, maNNu
r vs R lowercase r = ರ; uppercase R = ಱ (archaic, rare) hesaru, nuDi — never hesaRu
ಋ / ೃ x (short vocalic ṛ) samskxta, sxSTi, dxSTi
ೠ / ೄ X (long vocalic ṝ, extremely rare)

Key Terminology in Eke

Kannada Eke English
ಹೆಸರು ಪದ hesaru pada noun
ಎಸಕ ಪದ esaka pada verb
ಪರಿಚೆ ಪದ parice pada adjective/qualifier
ಸೊಲ್ಲರಿಮೆ sollarime grammar
ನುಡಿಯರಿಮೆ nuDiyarime linguistics
ಉಲಿ uLi phoneme/sound
ಬರಿಗೆ barige letter/grapheme
ಒಟ್ಟು oTTu affix
ಮುನ್ನೊಟ್ಟು munnoTTu prefix
ಜೋಡುಪದ jODupada compound word
ಕೀಳರಿಮೆ kILarime inferiority complex
ಮೇಲರಿಮೆ mElarime superiority complex
ಎಲ್ಲರ ಕನ್ನಡ ellara kannaDa everyone’s Kannada (Eke)
ಹೊಸ ಬರಹ hosa baraha new orthography (no mahapranas)
ಸೇರಿಕೆ serike sandhi
ವಿಭಕ್ತಿಪಲ್ಲಟ vibaktipaLLaTa case alternation

Current File Status (2026-03-23)

✅ Fully processed (en.md + kn-eke.md + claude-prompt.md)

Book Files present
02 — Hosapadagalannu Kattuva Bage book-website + blog + kn + kn-eke + en + claude-prompt
03 — Padagala Olarachane book + kn (OCR-cleaned, 11,437L) + kn-eke + en + claude-prompt
04 — Mathu Matthu Barahada Gondala transcript + website + kn-eke + en + claude-prompt
05 — Mathina Olaguttu transcript + website + kn-eke + en + claude-prompt
07 — Kannada Barahada Sollarime vol1-book + vol1-kn (20,475L) + vol2-book + vol2-kn (13,928L) + vol3/vol4 raw+full+eke+en + multi-vol index; citation quotes → curly ✅; Nudi-clean ✅
08 — Mahaprana Yake Beda book + djvu + kn + kn-eke + en + claude-prompt
09 — Havyaka Kannada transcript + website + kn-eke + en + claude-prompt
14 — Nijakku Halegannada book + djvu + blog + kn + kn-eke + en + claude-prompt; Nudi-clean ✅
15 — Inglish Kannada Padanerake book (53p sample) + kn-eke + en + claude-prompt
17 — Nudi Nadedu Banda Dari book + kn (OCR-cleaned, 16,883L) + kn-eke + en + claude-prompt; Nudi-clean ✅; citation quotes → curly ✅
18 — Nudiya Bagege Chintane blog + kn-eke + en + claude-prompt
20 — Havyaka Outline Grammar djvu + en + claude-prompt
25 — Vakyagala Olarachane book + kn (OCR-cleaned, 11,676L) + kn-eke + en + claude-prompt; citation quotes → curly ✅; anukaraNike removed from body ✅
27 — Baasheya Bagge book + kn (OCR-cleaned, 8,245L) + kn-eke + en + claude-prompt
28 — Kannadakke Beku book + kn (OCR-cleaned, 9,517L) + kn-eke + en (13 anchors) + claude-prompt; citation quotes → curly ✅
29 — Kannada Vyakarana Yaake Beku book + kn (OCR-cleaned, 11 ch. anchors) + kn-eke + en (12 anchors) + claude-prompt
30 — Kannada Barahavanna Saripadisona book/kn/raw.md + book/kn/full.md (10-ch TOC) + book/eke/full.md + book/en/summary.md + claude-prompt

✅ Partially processed (raw decode done; kn-eke pending)

Book Files present What’s missing
31 — Ingliš Padagalige Kannadade Padagalu book/kn/raw.md + book/kn/full.md (12,121 entries, Baraha CID decoded Phase 27) + book/en/summary.md + claude-prompt book/eke/full.md (Eke for a dictionary is less critical)

✅ YouTube + English summary only

Books 01, 06, 10, 11, 12, 13, 33 — YouTube transcripts + youtube/en/summary.md + claude-prompt. No scanned book text available.

✅ English summary + claude-prompt only

Book Notes
32 — Prominence of Tense, Aspect and Mood English monograph; book/en/summary.md + claude-prompt

❌ PDFs exist in Google Drive but text not yet extracted

Book PDF Pages Notes
19 — The Koraga Language the-koraga-language.pdf 123 English; Unicode-safe — needs pdftotext or OCR
21 — Pronouns (Oxford) DNS-Bhat-Pronouns-Oxford.pdf ~296 English; likely Unicode-safe

❌ Not yet collected (website offline / blocked / unarchived)

Books 16, 22, 23, 24, 26 — website stubs only.


Pending Work

High priority

Task Books What’s needed
Extract Book 19 PDF 19 The Koraga Language — pdftotext or Sarvam OCR; create en.md + claude-prompt.md
OCR text reconstruction multiple Two related issues: (1) word-boundary splits — e.g. meccuge ay should be meccugAy, OCR split at vowel boundaries, audit kn-eke.md files for space-separated suffix artifacts; (2) word-per-line artifact — narrowly-typeset 2-column pages where each line became its own line, join runs of short non-sentence-final lines. Books 17 and others.
Book 15 dictionary entry boundaries 15 WX-decoded OCR has no blank lines between dictionary entries — headwords and Kannada equivalents run together. Fix: identify English headword lines and insert blank lines before each.

Medium priority

Task Notes
Collect missing blog posts Books 02 (parts 1–3), 18 (many parts) — check more Wayback snapshots
Book 26 — Uli Marpadina Geregalu Cloudflare-blocked; try different snapshots or direct Google Cache
Books 16, 22, 23, 24 Not yet collected; website stubs only
Sollarime vol.5–7 Vol.1–4 now processed (Phases 14 + 26); vols 5–7 not found in Google Drive — may be at dnshankarabhat.net (offline)
Book 31 eke/full.md Eke romanisation of the A–Z dictionary entries not yet generated. Lower priority since the dictionary is English-headworded (Eke of Kannada definitions would be secondary use).

Low priority / stretch

Task Notes
Retroflexion 1973 paper Stanford Working Papers; should be in archive.org
Cambridge Dravidian Languages chapter DNS Bhat contributed to Cambridge Language Surveys

File Structure Summary

dnsbhat/
├── README.md                         # Master catalogue — 29 books, annotated
├── PROJECT-RECAP.md                  # This file
├── dns-bhat-analysis.md              # Early analysis document
├── kannada-content-landscape.md      # Content survey
├── kannada-knowledge-gap-analysis.md # Gap analysis
├── swadesh-bhat-prompt-and-analysis.md
├── wiktionary-cost-analysis-kimi-k2.5.md
│
├── 02-kannaDadalle-hosapadagaLannu-kaTTuva-bage/
│   ├── 02-...-blog.md                # 15 blog posts
│   ├── 02-...-kn.md                  # ★ Structured Kannada (547 lines)
│   ├── 02-...-kn-eke.md              # ★ Eke romanisation (830 lines, 43KB)
│   ├── 02-...-en.md                  # ★ English summaries
│   └── 02-...-claude-prompt.md       # ★ AI primer
├── 03-kannaDa-padagaLa-oLaracane/
│   ├── 03-...-book.md                # ★ Sarvam OCR output (239 pages)
│   ├── 03-...-kn.md                  # ★ OCR-cleaned Kannada (11,437L; structural artifact removal)
│   ├── 03-...-kn-eke.md              # ★ Eke romanisation (0 residual Kannada chars)
│   ├── 03-...-en.md                  # ★ English summaries
│   └── 03-...-claude-prompt.md       # ★ AI primer
├── 04-mAtu-mattu-barahada-naDuvina-gondala/
│   ├── 04-....md                     # YouTube transcript (~519 lines, 25/44 parts)
│   ├── 04-...-website.md             # Author's website stub
│   ├── 04-...-kn-eke.md              # ★ Eke romanisation of key passages
│   ├── 04-...-en.md                  # ★ English thematic summary (7 themes)
│   └── 04-...-claude-prompt.md       # ★ AI primer
├── 05-mAtina-oLaguTTu/
│   ├── 05-....md                     # YouTube transcript (~539 lines, 27/37 parts)
│   ├── 05-...-website.md             # Author's website stub
│   ├── 05-...-kn-eke.md              # ★ Eke romanisation of key passages
│   ├── 05-...-en.md                  # ★ English thematic summary (8 themes)
│   └── 05-...-claude-prompt.md       # ★ AI primer
├── 09-havyaka-kannaDa/
│   ├── 09-....md                     # YouTube transcript (~387 lines, 72/88 slots)
│   ├── 09-...-website.md             # Author's website stub
│   ├── 09-...-kn-eke.md              # ★ Eke romanisation of key passages
│   ├── 09-...-en.md                  # ★ English thematic summary (5 themes)
│   └── 09-...-claude-prompt.md       # ★ AI primer
├── 07-kannaDa-barahada-sollarime/
│   ├── 07-...-vol1-book.md           # ★ Sarvam OCR (327 pages)
│   ├── 07-...-vol1-kn.md             # ★ OCR-cleaned Kannada vol1 (20,475L; Ç-fix + arka-ottu)
│   ├── 07-...-vol2-book.md           # ★ Sarvam OCR (301 pages)
│   ├── 07-...-vol2-kn.md             # ★ OCR-cleaned Kannada vol2 (13,928L; Ç-fix + arka-ottu)
│   ├── 07-...-kn-eke.md              # ★ Eke romanisation (9KB)
│   ├── 07-...-en.md                  # ★ English summaries (41KB)
│   └── 07-...-claude-prompt.md       # ★ AI primer (18KB)
├── 08-kannaDakke-mahAprANa-yAke-bEDa/
│   ├── 08-...-book.md                # Full text (archive.org)
│   ├── 08-...-djvu.md                # DjVu OCR
│   ├── 08-...-kn.md                  # ★ Structured Kannada
│   ├── 08-...-kn-eke.md              # ★ Eke romanisation (template book)
│   ├── 08-...-en.md                  # ★ English analysis
│   └── 08-...-claude-prompt.md       # ★ AI primer
├── 14-nijakkU-haLegannaDa-vyAkaraNa-entahadu/
│   ├── nijakku-...-book.md           # Full text (archive.org)
│   ├── 14-...-djvu.md                # DjVu OCR
│   ├── 14-...-blog.md                # 7 blog posts
│   ├── 14-...-kn.md                  # ★ Structured Kannada
│   ├── 14-...-kn-eke.md              # ★ Eke romanisation
│   ├── 14-...-en.md                  # ★ English summary
│   └── 14-...-claude-prompt.md       # ★ AI primer
├── 15-ingliS-kannaDa-padanerake/
│   ├── 15-...-book.md                # ★ Hybrid extraction (53p sample, A–Az)
│   ├── 15-...-kn-eke.md              # ★ Eke romanisation of source text (munnuDi + conventions + A–Az entries)
│   ├── 15-...-en.md                  # ★ English analysis (10 word-formation patterns)
│   └── 15-...-claude-prompt.md       # ★ AI primer (decision tree, cluster tables, 100 entries)
├── 17-kannaDa-nuDi-naDeDu-banda-dAri/
│   ├── 17-...-book.md                # ★ Sarvam OCR (405 pages)
│   ├── 17-...-kn.md                  # ★ OCR-cleaned Kannada (16,883L; ya-fix + Ç-fix + 1,539 zero-KN lines removed)
│   ├── 17-...-kn-eke.md              # ★ Eke romanisation (12KB)
│   ├── 17-...-en.md                  # ★ English summaries (35KB)
│   └── 17-...-claude-prompt.md       # ★ AI primer (20KB)
├── 18-kannaDa-nuDiya-bagege-cintane/
│   ├── 18-...-blog.md                # 13 blog posts
│   ├── 18-...-kn-eke.md              # ★ Eke romanisation
│   ├── 18-...-en.md                  # ★ English summaries (29KB)
│   └── 18-...-claude-prompt.md       # ★ AI primer (7KB)
├── 20-havyaka-outline-grammar/
│   ├── 20-...-djvu.md                # Full DjVu text
│   ├── 20-...-en.md                  # ★ English chapter summaries
│   └── 20-...-claude-prompt.md       # ★ AI primer
├── 25-kannaDa-vAkyagaLa-oLaracane/
│   ├── 25-...-book.md                # ★ Sarvam OCR (289 pages)
│   ├── 25-...-kn.md                  # ★ OCR-cleaned Kannada (11,676L; word-level fixes + zero-KN lines)
│   ├── 25-...-kn-eke.md              # ★ Eke romanisation (19KB)
│   ├── 25-...-en.md                  # ★ English summaries (30KB)
│   └── 25-...-claude-prompt.md       # ★ AI primer (25KB)
├── 27-bhASheya-bagge/
│   ├── 27-...-book.md                # ★ Sarvam OCR (208 pages)
│   ├── 27-...-kn.md                  # ★ OCR-cleaned Kannada (8,245L; structural artifact removal)
│   ├── 27-...-kn-eke.md              # ★ Eke romanisation
│   ├── 27-...-en.md                  # ★ English summaries
│   └── 27-...-claude-prompt.md       # ★ AI primer
├── 28-kannaDakke-bEku/
│   ├── 28-...-book.md                # ★ Sarvam OCR (253 pages, raw WX-decoded)
│   ├── 28-...-kn.md                  # ★ OCR-cleaned Kannada (9,517L; 3-pass OCR fixes + structural artifact removal)
│   ├── 28-...-kn-eke.md              # ★ Eke romanisation (regenerated from kn.md)
│   ├── 28-...-en.md                  # ★ English summaries (25KB; 13 chapter anchors added)
│   └── 28-...-claude-prompt.md       # ★ AI primer (20KB)
├── 29-kannaDa-vyAkaraNa-yAke-bEku/
│   ├── 29-...-book.md                # ★ Sarvam OCR (260 pages, raw WX-decoded)
│   ├── 29-...-kn.md                  # ★ OCR-cleaned Kannada (11 chapter anchors)
│   ├── 29-...-kn-eke.md              # ★ Eke romanisation (12KB)
│   ├── 29-...-en.md                  # ★ English summaries (28KB; 12 chapter anchors added)
│   └── 29-...-claude-prompt.md       # ★ AI primer (19KB)
└── [... 11 more folders with website stubs or transcripts ...]

★ = created/significantly enriched in this project


Tools & Pipeline

Tool Used for
Wayback CDX API Crawling dnshankarabhat.net (91 pages found)
pdfminer PDF text extraction (failed for WX-encoded PDFs)
DjVu text archive.org DjVu OCR (books 08, 14, 20)
Sarvam Vision API OCR for WX-encoded PDFs (kn-IN, 10-page chunks)
pypdf PDF splitting for chunked OCR
sarvamai (PyPI) Python SDK for Sarvam API
wx_decode.py WX→Unicode lookup-table decoder (books 07, 17, 25, 28, 29)
Claude agents Parallel book processing (kn-eke, en.md, claude-prompt)
Jekyll / just-the-docs Static site generator; format-first sidebar via Python script in CI
GitHub Actions Build + deploy pipeline (.github/workflows/pages.yml)
ettuge Scala/SBT project hosting all MD content; public site at https://vwulf.github.io/ettuge/

OCR pipeline location: ettuge/src/main/python/sarvam-ocr/


Key Intellectual Themes

1. Kannada-native grammar (anti-Sanskrit framework)

Bhat’s life work: Kannada has its own grammatical logic. The three base categories are hesaru pada (noun), esaka pada (verb), parice pada (adjective) — not Sanskrit’s verb-root primacy. Books 01, 03, 07, 08, 14, 25, 28, 29.

2. Script reform (mahaprana / aspirated consonants)

Kannada phonology does not distinguish aspirated from unaspirated stops. The mahaprana letters (ಖ ಘ ಛ ಝ ಠ ಢ ಥ ಧ ಫ ಭ) are Sanskrit imports serving no phonological function. Book 08 makes this argument; books 07, 17, 29 are written in hosa baraha (simplified script without mahapranas).

3. Word formation without Sanskrit (native Dravidian roots)

Over 50% of standard Kannada vocabulary is Sanskrit-derived; 80% in scientific writing. Books 02 and 15 document systematic native Dravidian word-formation patterns using roots, affixes, and compounds.

4. Dialect documentation

Academic description of dialects as data: Havyaka (books 09, 20), Koraga (book 19), Belari (in book 19). These are not “corrupted” dialects but independent systems.

5. Sound change as the mechanism of language history

The theoretical underpinning of books 08, 14, 17, 26: sound changes are regular, reconstructable, and explain the Sanskrit/Kannada relationship. Book 22 is the English theoretical monograph; Book 26 is the Kannada popularisation. Book 17 applies historical phonology to trace Kannada from Proto-Dravidian.

6. Applied linguistics (why grammar matters)

Books 27, 28, 29 form Bhat’s popular advocacy trilogy: Baasheya Bagge (intro to linguistics), Kannadakke Beku (why Kannada needs its own grammar), Kannada Vyakarana Yaake Beku (who needs grammar and why). Together they make the case that correct grammatical description has practical consequences for education, translation, and language planning.