| name | kannada-morphology |
| description | Generates Kannada morphological forms from root words using Dr. D.N. Shankara Bhat's grammatical framework. Use this skill whenever someone asks about Kannada suffixes, case forms, verb conjugations, noun declension, dative/accusative/locative forms, verbal noun chains, phonological conditioning in Kannada, or wants to understand why a particular Kannada word takes a particular suffix. Also triggers for: "Kannada suffix for", "conjugate this verb", "case form of", "morphological form", "what is the dative of", "how does -ge work", "human vs non-human distinction in Kannada", "verbal noun", "hesaru pada", "hesaka pada". Do not hesitate to invoke this even for casual phrasing — any question about how Kannada words change form belongs here.
|
Kannada Morphology (DNS Bhat Framework)
You generate Kannada morphological forms — case suffixes, verb conjugations, verbal noun chains — following Dr. D.N. Shankara Bhat's grammatical framework. This is descriptive grammar: it describes the rules native speakers already follow unconsciously, not prescriptive rules imposed from Sanskrit.
The key insight from Bhat: all Kannada speakers already know these rules implicitly. Your job is to make them explicit and apply them systematically.
Use Eke romanization alongside Kannada script in all outputs (see romanization rules at end).
Word Classes (Bhat's Native Terminology)
| Bhat's Term | Sanskrit Term | English |
|---|
| ಹೆಸರುಪದ (hesarupada) | Naama pada (nAmapada) | Noun |
| ಎಸಕಪದ (esakapada) | Kriya pada (kriyApada) | Verb |
| ಪರಿಚೆಪದ (paricepada) | Visheshana (viShESaNa) ) | Adjective/Modifier |
| ಆಡುಪದ (ADupada) | Sarvanama (sarvanAma) | Pronoun |
| ಎಣಿಕೆಪದ (eNikepada) | Sankhya (sankhya) | Numeral |
Case System (ವಿಭಕ್ತಿ / Vibhakti)
The Critical Rational / Non-Rational Split
Kannada divides nouns into rational (ಉತ್ತಮ ಪ್ರಾಣಿ — higher beings) and non-rational (ಕ್ರಿಮಿ/ಜಡ — lower beings / inert) classes. This distinction controls suffix selection throughout the case system.
This is NOT Sanskrit's three-gender system (masculine/feminine/neuter). Book 14 (Nijakku Halegannada Vyakarana Entahadu?) documents how ancient Kannada grammarians like Keshiraja wrongly imposed Sanskrit's three-gender framework on Kannada — a fundamental error that persists in modern grammar textbooks. Old Kannada and modern Kannada both use a binary rational/non-rational split, not Sanskrit's tripartite gender.
- Rational: people, deities, respected beings (controls -anu/-aLu/-aru agreement)
- Non-rational: animals, objects, abstract concepts, plants (controls -itu/-avu agreement)
Dative Case (-ge / "to", "for")
This is Bhat's key example of implicit native speaker knowledge — all Kannada speakers know these rules but cannot typically state them explicitly.
Rule: Three allomorphs — conditioned by stem-final phonology AND human/non-human
| Condition | Suffix | Example |
|---|
| Vowel-final stems (most) | -ge | ರಾಮ+ಗೆ = ರಾಮನಿಗೆ (to Rama), ಹುಡುಗ+ಗೆ = ಹುಡುಗನಿಗೆ |
| Consonant-final, non-human | -ge (with linking) | ಮರ+ಗೆ = ಮರಕ್ಕೆ (to the tree) |
| Consonant-final, certain human nouns | -ige | ನಾಯಿ+ಗೆ = ನಾಯಿಗೆ |
| Words ending in retroflex/special | -kke | ಅವ+ಗೆ = ಅವನಿಗೆ, ಇದ+ಗೆ = ಇದಕ್ಕೆ |
Key dative allomorphs:
- -ge (ಗೆ): vowel-final human nouns, e.g. ರಾಮನಿಗೆ, ಅಮ್ಮನಿಗೆ
- -ige (ಇಗೆ): some consonant-final nouns, e.g. ನಾಯಿಗೆ
- -kke (ಕ್ಕೆ): pronouns and demonstratives, e.g. ಅವನಿಗೆ, ಇದಕ್ಕೆ, ಅವಳಿಗೆ
Other Cases
| Case | English Equivalent | Suffix(es) | Example |
|---|
| Nominative | Subject | — (bare stem + ನು for human) | ರಾಮನು (Rama-NOM) |
| Accusative | Object | -ನ್ನು / -ಅನ್ನು | ರಾಮನನ್ನು (Rama-ACC) |
| Instrumental | by/with | -ಇಂದ | ರಾಮನಿಂದ (by Rama) |
| Locative | in/on/at | -ಅಲ್ಲಿ | ಮನೆಯಲ್ಲಿ (in the house) |
| Ablative | from | -ಇಂದ | ಊರಿಂದ (from the village) |
| Genitive | of/possessive | -ಅ / -ದ | ರಾಮನ (of Rama), ಮನೆಯ (of the house) |
| Vocative | address | -ಏ / -ಅ | ರಾಮ! / ಅಮ್ಮಾ! |
Verbal Noun Chains (ಎಸಕನಾಮ / Esakanama)
One of the most productive patterns in Kannada: verbs become nouns, then take case suffixes.
4-step chain:
Verb root → Verbal noun → Case-marked verbal noun
ಮಾಡು → ಮಾಡುವುದು → ಮಾಡುವುದಕ್ಕೆ (for/to doing)
mADu → mADuvudu → mADuvudakke
Formation rules:
| Step | Operation | Example |
|---|
| 1. Root | Bare verb | ಮಾಡು (mADu) — to do |
| 2. + -vu | Infinitive marker | ಮಾಡುವ (mADuva) |
| 3. + -du | Verbal noun | ಮಾಡುವುದು (mADuvudu) — doing/the act of doing |
| 4. + case suffix | Case-marked noun | ಮಾಡುವುದಕ್ಕೆ (mADuvudakke) — for doing |
Examples of full chains:
| Verb | Verbal Noun | Dative ("for X-ing") | Locative ("in X-ing") |
|---|
| ಓದು (read) | ಓದುವುದು | ಓದುವುದಕ್ಕೆ | ಓದುವುದರಲ್ಲಿ |
| ಬರೆ (write) | ಬರೆಯುವುದು | ಬರೆಯುವುದಕ್ಕೆ | ಬರೆಯುವುದರಲ್ಲಿ |
| ತಿನ್ನು (eat) | ತಿನ್ನುವುದು | ತಿನ್ನುವುದಕ್ಕೆ | ತಿನ್ನುವುದರಲ್ಲಿ |
| ನೋಡು (see) | ನೋಡುವುದು | ನೋಡುವುದಕ್ಕೆ | ನೋಡುವುದರಲ್ಲಿ |
Verb Conjugation
Tense-Aspect-Mood System
Kannada verbs mark tense (past/non-past) and agreement with subject (person-number-gender — PNG).
Past tense (himbottina hesaka): stem + past marker + PNG
| Person | Singular | Plural |
|---|
| 1st | ಮಾಡಿದೆ (mADide) | ಮಾಡಿದೆವು (mADidevu) |
| 2nd | ಮಾಡಿದೆ (mADide) | ಮಾಡಿದಿರಿ (mADidiri) |
| 3rd masc | ಮಾಡಿದನು (mADidanu) | ಮಾಡಿದರು (mADidaru) |
| 3rd fem | ಮಾಡಿದಳು (mADidaLu) | ಮಾಡಿದರು (mADidaru) |
| 3rd neut | ಮಾಡಿತು (mADitu) | ಮಾಡಿದವು (mADidavu) |
Present/Future (mubottina hesaka): stem + -utt- + PNG
| Person | Singular | Plural |
|---|
| 1st | ಮಾಡುತ್ತೇನೆ (mADuttEne) | ಮಾಡುತ್ತೇವೆ (mADuttEve) |
| 2nd | ಮಾಡುತ್ತೀಯ (mADuttIya) | ಮಾಡುತ್ತೀರಿ (mADuttIri) |
| 3rd masc | ಮಾಡುತ್ತಾನೆ (mADuttAne) | ಮಾಡುತ್ತಾರೆ (mADuttAre) |
| 3rd fem | ಮಾಡುತ್ತಾಳೆ (mADuttALe) | ಮಾಡುತ್ತಾರೆ (mADuttAre) |
| 3rd neut | ಮಾಡುತ್ತದೆ (mADuttade) | ಮಾಡುತ್ತವೆ (mADuttave) |
Participial Forms (for adjectives from verbs)
| Form | Suffix | Example | Meaning |
|---|
| Past participial | -da | ಮಾಡಿದ (mADida) | done, the one who did |
| Present participial | -uva | ಮಾಡುವ (mADuva) | doing, the one who does |
| Negative participial | -ada | ಮಾಡದ (mADada) | not done, undone |
Etymology Classification Heuristics
When asked whether a word is native Dravidian or Sanskrit-borrowed:
Sanskrit borrowing signals:
- Contains aspirated consonants: kh, gh, th, dh, ph, bh (e.g., ಭಾಷೆ bhAShe → Sanskrit)
- Repeated consonant clusters from Sanskrit: pr-, tr-, kr-, sth-, etc.
- Retroflex ಷ (Sha) — rare in native Dravidian
- Ends in -a with Sanskrit declension feel
Native Dravidian signals:
- Simple consonant clusters or none
- No aspirates
- Contains retroflex L (ಳ), N (ಣ), D (ಡ), T (ಟ) — these are distinctly Dravidian
- Short, monosyllabic or disyllabic roots
- Found in Tamil/Telugu cognates with regular sound correspondences
Dravidian ↔ Tamil sound correspondences (Bhat's comparative method):
- Kannada k (before front vowels) ↔ Tamil c: ಕೆರಿ/cheri (tank), ಕೆನೆ/chene
- Kannada l ↔ Tamil l/zh in some environments
- These correspondences confirm native Dravidian origin
Key Source: Book 14
Nijakku Halegannada Vyakarana Entahadu? (D.N. Shankara Bhat, 2005) is the most important theoretical source for this skill. It systematically demonstrates — chapter by chapter across phonology, morphology, and syntax — why Sanskrit grammatical categories do not apply to Kannada, and what Kannada's actual Dravidian grammatical principles are.
Key findings relevant to this skill:
- Gender: Kannada has rational/non-rational (2-way), not masculine/feminine/neuter (3-way)
- Case: Old Kannada does not have Sanskrit's 8-case system; suffixes function differently
- Compounds: Sanskrit's tatpurusha/bahuvrihi/dvandva taxonomy doesn't map onto Kannada compounds
- Verb forms: Old Kannada tense/aspect differs fundamentally from Sanskrit's lakara system
- Technical terms: Sanskrit terms like lopa, agama, adesa, karaka misrepresent Kannada morphology
Full text: src/main/md/kannada/dnsbhat/14-nijakku-halegannada-vyakarana-entahadu/
Eke Romanization Quick Reference
- Long vowels → uppercase: a/A, i/I, u/U, e/E, o/O
- Retroflexes → uppercase: T, D, N, L
- Aspirates → preserved with h marker: kh, gh, th, dh, ph, bh (e.g., ಭಾಷೆ → bhAShe, ಅರ್ಥ → artha, ಧರ್ಮ → dharma)
- sh/Sh → S
Output Format
When generating morphological forms:
- State the stem/root in Kannada and Eke
- Identify conditioning factors (human/non-human, vowel/consonant final, tense, person-number-gender)
- Show the suffix selected and why
- Give the complete form in Kannada script + Eke
- For paradigms, use tables
Example:
"Dative of ಮರ (tree, non-human, consonant-final)"
Conditioning: non-human, consonant-final → suffix -kke
Form: ಮರಕ್ಕೆ / marakke — "to the tree"
When asked about a paradigm, give the full set of forms, not just one.
