nijakkU haLegannaDa vyAkaraNa entahadu?
What is Old Kannada Grammar Really Like?
Author: D. N. Shankara Bhat (ಡಿ. ಎನ್. ಶಂಕರ ಭಟ್) Series: DNS Bhat Linguistics Series Published: 2005 (First Edition); Reprinted by Kannada Sahithya Parishath / Prathibha Prakashan Language: Kannada Internet Archive: archive.org/details/arn.nijakkuhaleganna0000dnsh
Book Overview
Nijakku Halegannada Vyakarana Entahadu? (“What is Old Kannada Grammar Really Like?”) is D. N. Shankara Bhat’s systematic scholarly investigation into the grammatical structure of Old Kannada (ಹಳೆಗನ್ನಡ), the classical literary register of Kannada used in inscriptions and literature from approximately the 6th to 12th centuries CE. The book’s central argument is that every grammar of Old Kannada written so far — including Keshiraja’s celebrated Shabdamanidarpana (c. 13th century CE) — is fundamentally flawed, because it was constructed by applying Sanskrit grammatical categories, terminology, and principles to a language whose grammar is of a completely different type.
Old Kannada is a Dravidian language. Its grammatical structure — agglutinative morphology, postpositional case marking, Subject-Object-Verb word order, relative clause formation through participial verbal forms, separate plural suffixes independent of case marking — is typologically opposite to Sanskrit’s inflectional, gender-fused, case-integrated system derived from the Proto-Indo-European tradition. When Keshiraja and other Old Kannada grammarians described Old Kannada using Sanskrit technical terms (lopa, āgama, ādesha, kāraka, vibhakti, liṅga, vacana), they were not simply borrowing convenient labels; they were importing conceptual frameworks that misrepresent how the language actually works, producing grammars that are internally inconsistent and descriptively inaccurate.
Bhat works through every major domain of grammar — phonology, technical terminology, parts of speech, word structure, nouns, gender and number, case suffixes, verb forms, sentence connectives, and embedded clauses — and demonstrates, for each domain, (a) what Old Kannada actually does, (b) what Sanskrit does, (c) how the two differ fundamentally, and (d) how the ancient grammarians went wrong by conflating the two. The book is not polemical but rigorously analytical, using detailed examples from Old Kannada texts throughout. Its conclusion is that modern Kannada grammar has inherited these foundational errors from the ancient grammarians, and that a genuine grammar of Kannada — classical or modern — must be built from Kannada-internal evidence and Dravidian typology.
Table of Contents
- Chapter 1 — Introduction (ಪೀಠಿಕೆ)
- Chapter 2 — Use of Technical Terms (ಪಾರಿಭಾಷಿಕ ಪದಗಳ ಬಳಕೆ)
- Chapter 3 — Use of Phonemes (ವರ್ಣಗಳ ಬಳಕೆ)
- Chapter 4 — Parts of Speech (ಪದವರ್ಗಗಳು)
- Chapter 5 — Internal Structure of Words (ಪದಗಳ ಒಳರಚನೆ)
- Chapter 6 — Nature of Nouns (ನಾಮಪದಗಳ ಸ್ವರೂಪ)
- Chapter 7 — Gender and Number (ಲಿಂಗ ಮತ್ತು ವಚನಗಳು)
- Chapter 8 — Case Suffixes and Semantic Roles (ವಿಭಕ್ತಿಗಳು ಮತ್ತು ಕಾರಕಗಳು)
- Chapter 9 — Verb Forms (ಕ್ರಿಯಾರೂಪಗಳು)
- Chapter 10 — Sentence Connectives (ವಾಕ್ಯಗಳ ಜೋಡಣೆ)
- Chapter 11 — Embedded Clauses (ವಿಷಯವಾಕ್ಯಗಳು)
- Chapter 12 — Conclusion (ಮುಕ್ತಾಯ)
- Key Terms Glossary
Chapter 1 — Introduction
(ಪೀಠಿಕೆ)
1.1 Purpose of Grammar
The primary purpose of grammar (vyākaraṇa) is to describe the structure of a language — the structure of its words, word-forms, clusters of words (padakante), and sentences. This structure differs from language to language; that is precisely why knowing Sanskrit perfectly does not mean one can read Old Kannada texts without separate study. Old Kannada words, word-forms, compounds, and sentences are built on principles fundamentally different from Sanskrit’s.
All languages build words from vowels and consonants, but the inventory of sounds used and the patterns by which they combine differ across languages. Old Kannada lacks the aspirated consonants (mahāprāṇa) and the retroflex sibilant (ṣa) that Sanskrit uses; conversely, Sanskrit lacks the short mid vowels (pracchanna e and o) and the retroflex stops that Old Kannada uses. Equally, the suffixes by which Old Kannada forms derived words and the rules by which words combine differ thoroughly from Sanskrit’s. Bhat opens with this typological grounding to establish that the descriptive tools for Sanskrit grammar cannot simply be transferred to Old Kannada: a language-specific grammar must use language-specific descriptions.
1.2 Two Types of Grammars
Grammars can be prescriptive or descriptive. A prescriptive grammar (vidhāyaka vyākaraṇa) tells speakers what forms they should use; it is a normative manual for correct usage. A descriptive grammar (vivaraṇātmaka vyākaraṇa) records what forms speakers actually do use; it is a scientific account of observed linguistic behaviour. The grammars of Old Kannada written by the classical grammarians were largely prescriptive in orientation — they wanted to produce canonical literary forms — but the model they prescribed was Sanskrit, not the native Dravidian structure of Kannada.
Bhat argues that for a dead or classical literary language like Old Kannada, the only legitimate enterprise is descriptive: to record what the language as actually attested in texts actually does. Prescribing Sanskrit-derived forms was doubly wrong — wrong as linguistics (misrepresenting what the language is) and wrong as literary history (imposing external norms on a tradition that had its own internal coherence). A true grammar of Old Kannada must describe what texts like Vikramarjuna Vijaya (Pampa, 941 CE), Shantipurana (Ponna), and Gadayuddha (Ranna) actually exhibit.
1.3 Grammar of Old Kannada
Old Kannada (roughly 450–1200 CE) is attested in thousands of inscriptions and a rich literary tradition. Its grammatical structure is that of a classical Dravidian language: postpositional case suffixes (rather than prepositional case), a highly agglutinative suffix system, separate markers for plural and for case (unlike Sanskrit’s fused case-number endings), a robust participial system that encodes tense and mood on verb forms used as adjectives, and a sentence-final predicate structure (SOV). It borrowed extensively from Sanskrit in vocabulary — especially in the formal literary registers — but its grammatical skeleton remained Dravidian.
The fact that Sanskrit vocabulary was borrowed in large quantities confused ancient grammarians into thinking that Sanskrit grammar must also apply. But borrowing words does not change the grammatical architecture that those words are used within. Old Kannada speakers inflected Sanskrit nouns using Kannada case suffixes, not Sanskrit ones — precisely because the grammatical system they inhabited was Kannada, even when the lexical material was Sanskrit. Keshiraja and others documented the borrowed vocabulary accurately but misdescribed the grammar it was embedded in.
1.4 Existing Grammars
The major classical grammars of Old Kannada — Keshiraja’s Shabdamanidarpana (c. 1260 CE), and works by Nagavarma, Gunavarma, and others — were composed by scholars who had been trained primarily in Sanskrit grammatical science. The dominant intellectual framework was Paninian grammar, the most sophisticated and systematic linguistic analysis the ancient world produced. When these scholars turned to write grammars of Kannada, they naturally reached for the analytical tools they knew: Panini’s categories of lopa (elision), āgama (insertion), ādesha (substitution), vibhakti (case ending), kāraka (semantic role), liṅga (gender), and the elaborate system of paradigmatic forms.
The result was grammars that documented Kannada data through Sanskrit lenses. Some Sanskrit categories genuinely correspond to Old Kannada phenomena; many do not. Bhat’s thesis is that the resulting confusions and inconsistencies in the classical grammars — which have been noted by later scholars but attributed to Keshiraja’s individual failures — are not personal errors but structural consequences of applying the wrong analytical framework. All existing Old Kannada grammars, both classical and modern, stand on a mistaken foundation.
1.5 Nature and Structure of this Book
This book proceeds chapter by chapter through the major domains of Old Kannada grammar, comparing at each step (a) the Sanskrit pattern, (b) the Old Kannada pattern, (c) the Sanskrit technical term or concept used by the old grammarians, and (d) what went wrong when that concept was applied to Kannada data. The book does not aim to provide a complete grammar of Old Kannada — that would require many volumes — but to demonstrate the nature of the problem clearly enough that future grammarians can build on correct foundations.
Each chapter is structured to present specific examples from attested Old Kannada texts, show how the Sanskrit-trained grammarians treated those examples, and then show how a linguistically accurate description would treat them instead. The book concludes with a call for a complete revision of Old Kannada grammatical description, and by extension of modern Kannada grammar, on the basis of Dravidian typology rather than Sanskrit tradition.
Chapter 2 — Use of Technical Terms
(ಪಾರಿಭಾಷಿಕ ಪದಗಳ ಬಳಕೆ)
2.1 Introduction
Technical terms in grammar (pāribhāṣika padagaḷu) are not neutral labels; they encode theoretical assumptions about how language works. A term like āgama (insertion) presupposes a specific analytical model — one in which underlying forms undergo phonological changes described as additions, deletions, and substitutions. Whether that model accurately represents what is happening in Old Kannada morphophonology is an empirical question, not a matter of terminological convention. Old Kannada grammarians took Sanskrit technical terms wholesale into their analyses without examining whether those terms fit the data they were describing.
The chapter documents the most important cases where Sanskrit terms were misapplied. Some terms were used in contexts where they simply do not apply to Kannada. Others were given new meanings not found in Sanskrit (because the Kannada data did not fit the Sanskrit meaning) — creating terms that are neither accurately Sanskrit nor accurately Kannada. A third set of terms that would have been genuinely useful for describing Kannada were never developed, because the grammarians were working within a borrowed framework rather than generating categories from the data.
2.2 Elision, Insertion, and Substitution (ಲೋಪ, ಆಗಮ, ಆದೇಶ)
The Sanskrit trio lopa (elision), āgama (insertion), and ādesha (substitution) describes phonological changes when roots and suffixes combine. For example: in the Old Kannada form kāḍan (accusative of kāḍu, ‘forest’), the final u of the stem is elided — a lopa. In tujuvan (accusative of tuju), a v is inserted between stem and suffix — an āgama. In māṭāki (from māḍu, ‘to do’, + ki), the final consonant ḍ changes to l — an ādesha.
These Sanskrit terms describe formally accurate phonological alternations. But Bhat shows that applying them to Kannada creates problems when the same terms are applied to morphological phenomena. Keshiraja calls the plural suffix gaḷ an āgama — a meaningless insertion — because in Sanskrit, plural information is encoded in the case suffix itself (there is no separate plural marker). In Kannada, gaḷ carries independent semantic content (plural number); it is a suffix (pratyaya), not an insertion. Calling it an āgama obscures this and creates downstream confusion: since an āgama by definition has no independent meaning, Keshiraja is then forced to say that avu (the plural form used with pronouns and adjectives) is an ādesha substituting for the āgama gaḷ — a rule that is structurally bizarre (an ādesha on an āgama) and descriptively inaccurate.
2.3 Semantic Roles (ಕಾರಕಗಳು)
Sanskrit’s kāraka system distinguishes seven semantic roles (kartā, karman, karaṇa, sampradāna, apādāna, adhikaraṇa, sambandha) and associates each with a grammatical case. In Sanskrit the case-suffix encodes the kāraka directly. In Old Kannada, case suffixes do not map one-to-one onto semantic roles in the same way; the same suffix can mark several different semantic roles, and the same semantic role can be marked by different suffixes in different contexts. Applying the Sanskrit kāraka framework to Old Kannada therefore obscures the actual distribution of Kannada case suffixes and creates apparent anomalies that are artefacts of the wrong framework, not real irregularities.
2.4 Abbreviated Forms (ಸಂಕ್ಷಿಪ್ತ ರೂಪಗಳು)
Old Kannada grammarians used abbreviations and shorthand notations derived from Sanskrit grammatical tradition (particularly Paninian metalanguage) to refer to grammatical categories. These abbreviations presuppose the Sanskrit analytical framework. When applied to Kannada, they import Sanskrit assumptions along with the notation. Bhat discusses how even the notational conventions of existing Old Kannada grammars encode theoretical biases that a genuinely Kannada grammar would need to discard.
2.5 Grammatical Labels (ವ್ಯಾಕರಣ ಸಂಜ್ಞೆಗಳು)
Beyond the major categories of lopa/āgama/ādesha and kāraka, Kannada grammarians used dozens of Sanskrit grammatical labels (sañjñā) — terms like ākbyāta (finite verb), kṛdanta (participial noun), taddhita (secondary derivative), samāsa (compound), sandhi (phonological junction), and so on. Some of these labels genuinely correspond to distinct Kannada phenomena; others do not. The chapter examines which Sanskrit labels are appropriate for Old Kannada grammar and which introduce systematic distortion.
2.6 Unused Labels (ಉಪಯೋಗಿಸದ ಸಂಜ್ಞೆಗಳು)
Conversely, there are grammatical phenomena in Old Kannada that the classical grammarians never described adequately, because those phenomena have no Sanskrit counterpart and no Sanskrit technical term was available. The use of Kannada’s characteristic ā-participial and i-participial forms, the grammar of the quotative endu, the system of converb constructions — these are distinctly Dravidian features that Old Kannada grammarians noticed only partially and described inadequately, because their conceptual toolkit did not include the right categories.
2.7 Kannada-ness in Technical Terms (ಪಾರಿಭಾಷಿಕ ಪದಗಳಲ್ಲಿ ಕನ್ನಡತನ)
Bhat closes the chapter by arguing that a genuine grammar of Old Kannada would need to develop its own technical vocabulary — rooted in Kannada’s own grammatical character and, where possible, using native Kannada words. This is not mere terminological nationalism; it is a practical necessity. Using Sanskrit terms embeds Sanskrit assumptions. Developing Kannada terms (pratyaya → prataya or a native equivalent; kāraka → karalaka or a descriptive native term) would force grammarians to confront what those terms actually mean for Kannada rather than assuming Sanskrit meanings transfer automatically.
Chapter 3 — Use of Phonemes
(ವರ್ಣಗಳ ಬಳಕೆ)
3.1 Introduction
Old Kannada uses a phoneme inventory that partially overlaps with Sanskrit’s but differs in important ways. The overlap exists because Old Kannada borrowed Sanskrit words heavily, and those borrowings brought Sanskrit phonemes into Kannada orthography. But the phonological system of Old Kannada — the system that governs native Kannada words and the adaptation of borrowed words into Kannada speech patterns — is fundamentally different from Sanskrit’s. Existing grammars, which describe Old Kannada phonology entirely through Sanskrit’s classificatory scheme, misrepresent this system.
3.2 List of Phonemes (ವರ್ಣಗಳ ಪಟ್ಟಿ)
Sanskrit’s phoneme inventory includes short and long vowels (a, ā, i, ī, u, ū, ṛ, ṝ), diphthongs (e, ai, o, au), aspirated stops (kh, gh, ch, jh, Th, Dh, th, dh, ph, bh), and the retroflex sibilant ṣa. Old Kannada’s native phoneme inventory lacks the aspirated stops and ṣa altogether. It has short mid vowels (e and o without length distinction in native words) that Sanskrit does not use. The retroflex stop series (ṭa, ḍa, ṇa) exists in both, but Old Kannada adds the retroflex lateral ḷa as a phonologically active consonant. Classical Old Kannada grammarians listed the full Sanskrit phoneme inventory as the inventory of Kannada — because they were describing the literary language that included Sanskrit loanwords — but never distinguished what is native to Kannada from what is borrowed.
3.3 Order of Phonemes (ವರ್ಣಗಳ ಅನುಕ್ರಮ)
The traditional ordering of Kannada phonemes (varṇakrama) follows Sanskrit’s ordering exactly: vowels first, then consonant series arranged by place of articulation (velars, palatals, retroflexes, dentals, labials), then the semivowels and sibilants. This ordering encodes Sanskrit phonological theory — particularly the Maheshvara sutras that underpin Paninian rule ordering. Since Kannada phonology does not work on the same principles as Sanskrit phonology, this ordering is not merely neutral tradition; it imports a Sanskrit theoretical structure into Kannada phonological description and makes it harder to see how Kannada phonological patterns actually work.
3.4 Changes in Compound Forms (ತದ ಓವಗಳಲ್ಲಿ ಕಾಣಿಸುವ ಬದಲಾವಣೆಗಳು)
When words or morphemes combine in Old Kannada, phonological changes occur at junctures — much as they do in Sanskrit (sandhi). However, the Old Kannada rules governing these changes differ substantially from Sanskrit sandhi rules. For example, when a word ending in i is followed by a word beginning with i, Sanskrit merges them into a long ī; Old Kannada drops one i and keeps the other. These are different rules reflecting different phonological systems. Keshiraja describes Old Kannada juncture changes using Sanskrit sandhi categories, producing a description that is at best partial and at worst systematically misleading.
3.5 Pronunciation of Words (ಪದಗಳ ಉಚ್ಚಾರಣೆ)
Old Kannada had pronunciation features that the script, borrowed from Sanskrit-derived orthography, does not capture well. Sanskrit loanwords were pronounced according to Kannada phonological rules — the aspirated consonants were de-aspirated, the retroflex sibilant merged with the palatal sibilant, and long Sanskrit vowels were sometimes shortened in Kannada speech. Grammarians who described the written forms rather than the spoken forms therefore produced phonological descriptions of Old Kannada that were really descriptions of Sanskrit phonology applied to Kannada text.
3.6 Groups of Phonemes (ವರ್ಣಗಳ ಗುಂಪುಗಳು)
This section compares the phoneme groupings used in Sanskrit phonological description with those appropriate for Old Kannada. Sub-sections cover:
3.6.1 Sanskrit vowels: Sanskrit has a more elaborate vowel system (including ṛ, ṝ, and the diphthongs ai and au as distinct phonemes) than Old Kannada. Describing Old Kannada vowels through Sanskrit’s vowel classification misrepresents the Kannada vowel system.
3.6.2 Old Kannada vowels: Old Kannada’s vowel system centers on a, i, u, e, o, and their long counterparts, plus the short mid vowels (pracchanna e and o) not found in Sanskrit. The short mid vowels are phonemically significant in native Kannada words but invisible in Sanskrit-derived phonological description.
3.6.3 Consonant phonemes: Old Kannada’s consonant system is similar to Sanskrit’s in its overall organization (velar, palatal, retroflex, dental, labial stops) but lacks aspiration as a phonemic feature in native words and lacks the retroflex sibilant.
3.6.4 Maheshvara sutras: The Māheśvara sūtras are the foundational phonological index of Paninian grammar, encoding phonological natural classes for rule formulation. Bhat discusses whether these sutras, which structured all Sanskrit phonological analysis, have any legitimate use in describing Old Kannada phonology — and argues they do not.
3.7 Rules of Combination (ಸೇರಿಕೆಯ ನಿಯಮಗಳು)
The rules governing phonological juncture in Old Kannada — what happens when stems and suffixes combine, or when words form compounds — are systematically different from Sanskrit sandhi rules. Bhat surveys the major Old Kannada juncture patterns: medial sandhi (within words) and final sandhi (at word boundaries), and shows that these follow Dravidian phonological patterns, not Sanskrit ones. Applying Sanskrit sandhi categories to describe them produces both missed generalizations (patterns that follow a Dravidian rule but look irregular in Sanskrit terms) and false generalizations (treating Old Kannada forms as if they followed Sanskrit rules they do not).
3.8 Summary (ಸಾರಾಂಶ)
Old Kannada phonology requires description on its own terms. The phoneme inventory, the phonological groupings, and the rules of combination all differ from Sanskrit in ways that have been obscured by the Sanskrit-derived descriptive tradition. A linguistically accurate account of Old Kannada phonology must start from the attested data in Old Kannada texts and develop generalizations from that data, not from Sanskrit theory.
Chapter 4 — Parts of Speech
(ಪದವರ್ಗಗಳು)
4.1 Diversity in Words (ಪದಗಳಲ್ಲಿ ವೈವಿಧ್ಯ)
Languages organize their words into grammatical classes — typically called parts of speech — based on the distributional properties those words have in sentences: which positions they can occupy, which suffixes they can take, and what grammatical relationships they can enter into. Sanskrit’s part-of-speech system (nouns, pronouns, verbs, indeclinables, with elaborate subclassifications) is one solution to this organizational problem; Old Kannada’s part-of-speech system is a different solution suited to its different grammatical structure.
4.2 Structure of Nominal Clusters (ನಾಮಪದಕಂತೆಗಳ ರಚನೆ)
Old Kannada has a distinct word-class that Bhat calls nāmapadakante — nominal clusters formed by combining a noun with participial or adjectival modifiers. The structure of these clusters follows Dravidian head-final patterns (modifiers precede the head noun) rather than Sanskrit’s more flexible arrangement. Sub-sections cover:
4.2.1 Old Kannada extended uses: The distribution of nominal clusters in Old Kannada texts, with examples from literary texts showing their typical syntactic functions and range of usage.
4.2.2 Sanskrit extended uses: The corresponding Sanskrit construction types, and how they differ from Old Kannada’s system.
4.2.3 Problems in Old Kannada grammars: How classical grammarians misdescribed nominal clusters by assimilating them to Sanskrit nominal compound categories.
4.3–4.5 Pronouns and Person Words (ಸರ್ವನಾಮಗಳು, ಪುರುಷಪದಗಳು, ಸರ್ವಪದಗಳು)
Old Kannada has an elaborate system of pronouns and deictic words (puruṣapada) that differs structurally from Sanskrit’s pronoun system. Important differences include: the social/honorific distinctions encoded in the second and third person forms; the demonstrative system (ī, ā, u-deixis); and the interaction between pronoun forms and verb agreement suffixes. Bhat shows that Sanskrit pronoun categories do not map cleanly onto Old Kannada’s demonstrative and personal pronoun system.
4.6 Borrowed Words (ಎರವಲಾಗಿ ಬಂದ ಪದಗಳು)
A large proportion of Old Kannada literary vocabulary consists of Sanskrit loanwords (tatsama — unchanged forms) and nativized Sanskrit loans (tadbhava — phonologically adapted forms). Old Kannada grammarians were well aware of this division but described it using Sanskrit-derived taxonomy. Bhat re-examines the classification: while the tatsama/tadbhava distinction is useful for describing the etymological source, it should not be allowed to drive grammatical description, since both classes of words function grammatically as Kannada words, following Kannada grammatical rules.
4.7 Summary (ಸಾರಾಂಶ)
Old Kannada’s part-of-speech system is a Dravidian system. Its categories — the distinction between finite and non-finite verbs, the behavior of participial adjectives, the pronoun-agreement system — must be described through Dravidian typological categories, not through Sanskrit’s classical parts-of-speech taxonomy.
Chapter 5 — Internal Structure of Words
(ಪದಗಳ ಒಳರಚನೆ)
5.1 Introduction
The internal structure of words — how stems, suffixes, and combining forms relate — is the domain where Dravidian agglutinative morphology differs most starkly from Sanskrit’s fusional inflectional morphology. Sanskrit encodes case and number in a single fused ending that also signals the noun class (liṅga); Old Kannada adds these as separate, independent suffixes stacked onto the stem. Sanskrit attaches prefixes (upasarga) before the root; Old Kannada has no prefix system. The entire architecture of word-building is different.
5.2 Use of Suffixes (ಪ್ರತ್ಯಯಗಳ ಬಳಕೆ)
Old Kannada forms new words and word-forms by adding suffixes to stems, and suffixes can stack (pluralizing suffix + case suffix on the same stem). Sub-sections examine:
5.2.1 Words and word-forms: The distinction between a lexical word (pada) and its inflected forms (padarūpa) in Old Kannada, and how this differs from Sanskrit’s paradigm-based conception of the same distinction.
5.2.2 Diversity in word-classes: How different classes of Old Kannada words take different sets of suffixes, and how the suffix system as a whole reflects the agglutinative character of the language.
5.2.3 Nouns and nominal forms: The morphological behavior of Old Kannada nouns — how stems are formed, what suffixes they take, and how the resulting forms function in sentences.
5.3–5.4 Compounds (ಸಮಾಸಗಳು)
Old Kannada forms compound words (samāsa, samastapada) by combining two or more stems. Sanskrit has an elaborate taxonomy of compound types (dvandva, bahuvrīhi, tatpuruṣa, avyayībhāva, karmadhāraya). Old Kannada uses compounds but its compounding system differs from Sanskrit’s in important respects: Old Kannada uses adjective words (guṇapada) as the first element where Sanskrit would use a noun in an oblique case; the semantic relationships encoded in Old Kannada compounds follow Dravidian patterns not always covered by Sanskrit compound taxonomy; and the phonological junction rules at compound boundaries differ from Sanskrit sandhi.
The section examines the major Sanskrit compound types and asks, for each, whether it applies to Old Kannada data and, if so, whether the Sanskrit analysis produces an accurate description of Old Kannada compound structure or introduces confusion.
5.5 Summary (ಸಾರಾಂಶ)
Old Kannada word structure is agglutinative and suffixing, with no prefix system and a compounding system that follows Dravidian rather than Sanskrit structural principles. Sanskrit’s morphological analysis, designed for a fusional inflecting language with both prefixes and inflectional case, cannot accurately describe Old Kannada word structure.
Chapter 6 — Nature of Nouns
(ನಾಮಪದಗಳ ಸ್ವರೂಪ)
6.1 Classification of Nouns (ನಾಮಪದಗಳ ವಿಭಜನೆ)
Sanskrit classifies nouns primarily by grammatical gender (liṅga): every noun is masculine, feminine, or neuter, and this gender determines the entire paradigm of case and number endings it takes. Old Kannada also assigns gender to nouns but uses a different and simpler system: nouns denoting male animate beings are masculine, nouns denoting female animate beings are feminine, and all other nouns — animate or inanimate, but not specifically gendered — are neuter. This is a natural (semantic) gender system, not the complex lexically-arbitrary gender system of Sanskrit and Latin. Keshiraja and other grammarians imported Sanskrit’s gender taxonomy wholesale, producing a description of Old Kannada noun classes that is structurally inappropriate for the language.
Sub-sections cover: the need (or lack thereof) for the concept of liṅga in describing Kannada nouns; participial nouns (kṛdanta) and derived nouns (taddhitānta) and how they behave in Old Kannada; the class of adjective words (guṇapada) unique to Kannada, with no Sanskrit equivalent.
6.2 Adjectives (ಗುಣಪದಗಳು)
Old Kannada has a distinct word class of adjective words (guṇapada) that do not inflect for case or number — they modify nouns but take no suffixes. This contrasts sharply with Sanskrit, where adjectives inflect in agreement with their head noun in gender, number, and case — a phenomenon Sanskrit grammar encodes systematically in its paradigm tables. The absence of agreeing adjectives in Old Kannada is a major typological fact that Sanskrit-trained grammarians either ignored or described inadequately. Bhat argues that the guṇapada class is one of the clearest markers of Kannada’s Dravidian character and should be the starting point for any description of Old Kannada noun phrase structure.
6.3 Summary (ಸಾರಾಂಶ)
Old Kannada noun grammar centers on a semantic (natural) gender system, a simple agglutinative case-and-number suffix system, and a distinct uninflected adjective class — all of which are Dravidian features systematically misrepresented by the Sanskrit-based grammars.
Chapter 7 — Gender and Number
(ಲಿಂಗ ಮತ್ತು ವಚನಗಳು)
7.1 Introduction
Gender (liṅga) and number (vacana) are encoded differently in Sanskrit and Old Kannada. Sanskrit encodes both in a single case ending that simultaneously marks case, number, and the noun’s gender class. Old Kannada encodes them with separate independent suffixes: number is marked by an optional plural suffix (gaḷ), and gender is relevant primarily to the choice of agreement suffixes on predicates. These are structurally different encoding strategies, and they require structurally different grammatical descriptions.
7.2 Differences in Classification (ವಿಭಜನೆಯಲ್ಲಿ ವ್ಯತ್ಯಾಸ)
Bhat compares Old Kannada and Sanskrit gender classification in detail:
7.2.1 Old Kannada gender classification: Old Kannada distributes nouns into animate-masculine, animate-feminine, and neuter categories based on the natural gender of the referent, with neuter including all non-specifically-gendered entities. Agreement on finite verbs and on certain nominal forms tracks this three-way distinction.
7.2.2 Sanskrit gender classification: Sanskrit’s three-gender system (puṃliṅga, strīliṅga, napuṃsakaliṅga) is partly natural and partly arbitrary, with many nouns assigned to gender classes that have no semantic motivation. The system produces elaborate agreement paradigms (case × number × gender = up to 24 distinct forms per paradigm).
7.2.3 Problems in Old Kannada grammarians: The ancient grammarians attempted to apply Sanskrit’s gender taxonomy to Old Kannada, producing spurious categories and erroneous rules for words that do not fit Sanskrit patterns.
7.3 Indication of Number (ವಚನದ ಸೂಚನೆ)
Old Kannada marks plural with an independent suffix (gaḷ or aru depending on noun class and animacy), which is clearly a separate morpheme with its own meaning. Sanskrit marks number as part of the case ending — there is no separate plural morpheme. Keshiraja’s misclassification of gaḷ as an āgama (discussed in Chapter 2) follows directly from not recognizing this structural difference. Number and case are conceptually fused in Sanskrit grammar; they are conceptually separate in Old Kannada grammar, and treating them as fused produces systematic misdescription.
Chapter 8 — Case Suffixes and Semantic Roles
(ವಿಭಕ್ತಿಗಳು ಮತ್ತು ಕಾರಕಗಳು)
8.1 Introduction
The case system is the domain where the mismatch between Sanskrit and Old Kannada grammar is most consequential. Sanskrit has eight cases (vibhakti) — nominative, accusative, instrumental, dative, ablative, genitive, locative, and vocative — each encoding a specific set of semantic roles (kāraka). Sanskrit case suffixes are fusional: they encode case and number simultaneously. Old Kannada has a smaller set of case suffixes, they are agglutinative (added after the plural marker, not fused with it), and they do not map onto Sanskrit’s eight-case system.
8.2 Non-use of Case Suffixes (ವಿಭಕ್ತಿ ಪ್ರತ್ಯಯವನ್ನು ಬಳಸದಿರುವುದು)
A striking feature of Old Kannada syntax is that nouns can be used in sentences without any case suffix at all — the bare stem form functions as subject, object, or other argument in many contexts. In Sanskrit, all nouns (except indeclinables) must carry a case suffix to appear in a sentence. This is a fundamental typological difference: Sanskrit requires syntactic function marking on nouns; Old Kannada does not. Ancient grammarians, working in a Sanskrit framework where caseless nouns are impossible, had no adequate account of this Kannada feature.
8.3 Number of Case Suffixes (ವಿಭಕ್ತಿ ಪ್ರತ್ಯಯಗಳ ಸಂಖ್ಯೆ)
Old Kannada has a different number of case suffixes from Sanskrit, and its suffixes have different shapes and distributions. A description of Old Kannada case that lists eight cases (because Sanskrit has eight) and maps Old Kannada suffixes onto those eight Sanskrit cases cannot accurately represent either the number of distinct cases in Old Kannada or the meanings/uses of Old Kannada case suffixes.
8.4 Meanings of Case Suffixes (ವಿಭಕ್ತಿ ಪ್ರತ್ಯಯಗಳ ಅರ್ಥ)
Each Old Kannada case suffix has its own range of semantic functions. Bhat examines:
8.4.1 Old Kannada case suffixes: The meaning ranges of each Old Kannada case suffix as attested in actual Old Kannada texts.
8.4.2 Sanskrit case suffixes: The semantic ranges of Sanskrit’s eight cases.
The comparison shows that Old Kannada case suffixes and Sanskrit case endings cover different semantic territories in different ways; trying to map one onto the other produces systematic distortions.
8.5 Extended Uses (ವಿಸ್ತೃತ ಬಳಕೆಗಳು)
Both Sanskrit and Old Kannada case suffixes have “extended uses” where the suffix marks a semantic role beyond its primary meaning. But the extended uses of Old Kannada case suffixes differ from those of Sanskrit case suffixes. Grammarians who described Old Kannada case suffix usage in terms of Sanskrit extended uses therefore mischaracterized which uses are primary and which are extended, and missed the generalizations that actually govern Old Kannada case usage.
8.6 Case Suffixes and Semantic Roles (ವಿಭಕ್ತಿ ಪ್ರತ್ಯಯಗಳು ಮತ್ತು ಕಾರಕಗಳು)
The relationship between case suffixes (formal markers) and semantic roles (meaning categories) is different in Sanskrit and Old Kannada. Sanskrit’s kāraka theory posits a direct relationship between each case suffix and specific semantic roles. Old Kannada’s case suffixes have a more indirect relationship to semantic roles, with more context-dependence and more overlap. Importing Sanskrit’s kāraka theory into Old Kannada grammar creates an analysis that either forces artificial one-to-one mappings or generates large numbers of ad hoc exceptions.
8.7 Problems in Old Kannada Grammarians (ಹಳೆಗನ್ನಡ ವೈಯಾಕರಣಿಗಳ ಸಮಸ್ಯೆಗಳು)
This section documents specific errors in the classical grammarians’ treatment of case, organized by case type:
8.7.1 Nominative (prathamā vibhakti): Keshiraja’s description of the nominative case in Old Kannada, and where it goes wrong.
8.7.2 Instrumental and dative: Old Kannada does not have the same distinction between instrumental and dative that Sanskrit does; collapsing or separating these in the Sanskrit manner misrepresents Old Kannada case usage.
8.7.3 Ablative (pañcamī vibhakti): The Old Kannada ablative suffix and its mismatch with Sanskrit ablative semantics.
8.7.4 Genitive (ṣaṣṭhī vibhakti): Old Kannada genitive usage compared with Sanskrit.
8.7.5 Case shift (vibhakti pallaṭa): A phenomenon in Old Kannada where case suffixes appear in unexpected positions — which the classical grammarians noted but could not explain accurately because they were working with the wrong framework.
8.8 Summary (ಸಾರಾಂಶ)
Old Kannada’s case system is a Dravidian agglutinative case system. It has different cases from Sanskrit (in number, shape, and semantic range), a different relationship between case and number, and a different relationship between case marking and semantic roles. Every major feature of Old Kannada case grammar is misrepresented by the Sanskrit-derived classical grammars.
Chapter 9 — Verb Forms
(ಕ್ರಿಯಾರೂಪಗಳು)
9.1 Internal Structure of Verb Forms (ಕ್ರಿಯಾರೂಪಗಳ ಒಳರಚನೆ)
Old Kannada verb forms are built by adding tense-aspect suffixes, mood suffixes, and agreement suffixes to a verbal root in a left-to-right stack. This agglutinative structure contrasts with Sanskrit’s fusional verb morphology, where the same suffix simultaneously encodes tense, mood, person, and number. The classical grammarians described Old Kannada verb forms using Sanskrit’s conjugational paradigm structure — listing forms by person and number in tables derived from Sanskrit models — which obscures the agglutinative architecture of Old Kannada verb morphology.
9.2 Methods of Indicating Time (ಸಮಯವನ್ನು ಸೂಚಿಸುವ ವಿಧಾನಗಳು)
Old Kannada uses a system of tense-aspect suffixes (da/pa suffix for past, v suffix for future/present) different from Sanskrit’s complex tense-mood system (laṭ, liṭ, luṭ, lṛṭ, leṭ, loṭ, laṅ, liṅ, luṅ, lṛṅ). Sub-sections examine the specific suffixes (da and va; attu; temporal relationship suffixes) and how they encode tense and aspect in Old Kannada.
9.3–9.6 Aspect, Negation, and States
These sections cover: the guṃ/kuṃ suffix and its functions; methods of negation in Old Kannada (which differ from Sanskrit’s negative formations); and the distinction between events (ghaṭane) and states (saṃgati) in Old Kannada verb semantics — a distinction with no direct parallel in Sanskrit grammar.
9.7 Summary (ಸಾರಾಂಶ)
Old Kannada verb morphology is agglutinative, with independent suffixes for tense, aspect, mood, and agreement. Sanskrit’s fusional verbal paradigm structure does not describe it accurately.
Chapter 10 — Sentence Connectives
(ವಾಕ್ಯಗಳ ಜೋಡಣೆ)
10.1 Introduction
Languages encode the relationships between clauses in different ways. Sanskrit uses converb forms (kṛdanta forms: the tvā converb for sequenced events, the tum infinitive for purposive clauses, the satisaptamī for simultaneous events) and relative clauses built on participial forms with extensive agreement morphology. Old Kannada uses a very different system for connecting clauses, based on distinct Old Kannada converb forms and participial constructions that encode temporal and logical relationships between events.
10.2 Old Kannada Connecting Verb Forms (ಹಳೆಗನ್ನಡದ ಜೋಡಿಸುವ ಕ್ರಿಯಾರೂಪಗಳು)
Old Kannada has a rich system of non-finite verb forms used to connect events within complex sentences. These include: the converb (e-form) used for sequential events (‘having done X, did Y’); the al form used for purposive and certain conditional meanings; and forms encoding time, restriction, and condition. Each has specific syntactic distributions and semantic functions that are clearly Dravidian in character.
10.3 Sanskrit Connecting Verb Forms (ಸಂಸ್ಕೃತದಲ್ಲಿ ಜೋಡಿಸುವ ಕ್ರಿಯಾರೂಪಗಳು)
Sanskrit uses tvā and ya converbs, tum infinitives, and the satisaptamī participial construction to connect clauses. These have specific properties tied to Sanskrit’s overall morphological and syntactic architecture — particularly the agreement system that operates on participial forms. Old Kannada’s clause-connecting devices are structurally different.
10.4 Problems in Old Kannada Grammarians
Classical grammarians described Old Kannada converb and participial forms in terms of Sanskrit kṛdanta categories. Since the formal and distributional properties of these forms are different in Kannada, this produced systematic mismatches. In particular, the richness and regularity of Old Kannada’s converb system was not recognized as a unified system because it was described piecemeal using Sanskrit categories that cut across the relevant Kannada generalizations.
10.5 Summary
Old Kannada’s system of clause-connecting verb forms is a coherent Dravidian system. It must be described as such, not as an imperfect version of Sanskrit’s kṛdanta system.
Chapter 11 — Embedded Clauses
(ವಿಷಯವಾಕ್ಯಗಳು)
11.1 Introduction
Embedded clauses (viṣayavākya) — clauses that function as arguments or modifiers of nouns or verbs in a larger sentence — are formed differently in Old Kannada and Sanskrit. In Sanskrit, relative clauses are formed using relative pronouns (ya-, yā-, yad-) that agree in gender, number, and case with their antecedent noun. In Old Kannada, relative clauses are formed using participial verb forms that function as prenominal modifiers — there are no relative pronouns. This is one of the most distinctive typological features of Dravidian syntax, and it requires an entirely different analytical framework from Sanskrit relative clause grammar.
11.2 Classification of Embedded Clauses (ವಿಷಯವಾಕ್ಯಗಳ ವಿಭಜನೆ)
Old Kannada has different types of embedded clauses: relative clauses (modifying nouns), complement clauses (functioning as arguments of verbs), and adverbial clauses (providing circumstantial information). Each type uses specific construction types. A descriptively adequate grammar must distinguish these types and describe each accurately.
11.3 Structure of Relative Clauses (ಗುಣವಾಕ್ಯಗಳ ರಚನೆ)
In Old Kannada, relative clauses are formed by a participial verb form preceding the modified noun. The participial form shows tense (past or non-past) but does not agree in gender or case with the head noun — a striking contrast with Sanskrit’s elaborate relative pronoun agreement. Sub-sections examine: the gender-number agreement suffixes on participial forms; the ākhyāta suffixes; the use of the verbal form in relative clauses.
11.4 Structure of Relational Clauses (ಸಂಬಂಧ ವಾಕ್ಯಗಳ ರಚನೆ)
Relational clauses (sambandha vākya) in Old Kannada express relationships between events or participants. Bhat examines how these differ from Sanskrit’s corresponding constructions.
11.5–11.6 Existential and Sanskrit Embedded Clauses
Existential sentences in Old Kannada and their interaction with embedded clause structures; Sanskrit embedded clause constructions and how they differ from Old Kannada’s.
11.7 Summary
Old Kannada’s embedded clause system — participial relative clauses without relative pronouns, complement clauses using the quotative endu — is a Dravidian system completely unlike Sanskrit’s relative pronoun-based system. Describing it in Sanskrit terms produces both missed generalizations and false analyses.
Chapter 12 — Conclusion
(ಮುಕ್ತಾಯ)
12.1 Critique of Old Kannada Grammars (ಹಳೆಗನ್ನಡ ವ್ಯಾಕರಣಗಳ ವಿಮರ್ಶೆಗಳು)
The conclusion returns to the foundational critique developed across all preceding chapters: Old Kannada grammars, both classical and modern, have been built on a Sanskrit-derived framework that is structurally inappropriate for a Dravidian language. The result is grammars that — regardless of how elaborately they document Kannada forms — produce inaccurate analyses, miss genuine linguistic generalizations, and create artificial complexities where there are none. This is not a criticism of Keshiraja’s intelligence or scholarship; it is a criticism of a framework that was not yet available to be superseded when he wrote. The discipline of comparative linguistics, which would make clear the typological differences between Sanskrit (Indo-Aryan) and Kannada (Dravidian), was not available to medieval grammarians.
12.2 Complexity of Grammar (ವ್ಯಾಕರಣದ ಸಂಕೀರ್ಣತೆ)
The apparent complexity of Old Kannada grammar in existing accounts is largely an artefact of wrong analysis. When grammar is described on the basis of a wrong framework, simple rules look complex (because they are being described in terms that do not fit them) and real regularities are invisible. Bhat argues that Old Kannada grammar, described correctly, is not especially complex; it is, like all natural languages, complex in some domains and simple in others, but consistently Dravidian in its structural character.
12.3 Diversity of Grammar (ವ್ಯಾಕರಣದ ವೈವಿಧ್ಯ)
The final section situates the book within the broader project of Dravidian linguistic research and Kannada language preservation. The diversity of grammatical structures across the world’s languages — Dravidian, Indo-Aryan, Sino-Tibetan, Niger-Congo — is one of the great intellectual resources of humanity. Describing each language on its own terms is both a scientific and a cultural obligation. For Kannada, this means building a grammar that reflects how Kannada actually works — and the first step is recognizing that the existing grammars, however prestigious their history, do not do this.
Key Terms Glossary
| Kannada Term | Romanization | Meaning / Explanation |
|---|---|---|
| ಹಳೆಗನ್ನಡ | haḷegannaḍa | Old Kannada; the classical literary form of Kannada (c. 450–1200 CE) |
| ವ್ಯಾಕರಣ | vyākaraṇa | Grammar |
| ಪಾರಿಭಾಷಿಕ ಪದ | pāribhāṣika pada | Technical term / grammatical term |
| ಲೋಪ | lopa | Elision; deletion of a sound in morphophonological combination |
| ಆಗಮ | āgama | Insertion; addition of a sound at morpheme juncture (Sanskrit: without independent meaning) |
| ಆದೇಶ | ādesha | Substitution; replacement of one form by another |
| ಕಾರಕ | kāraka | Semantic role; the functional relationship between a noun and the verb in a clause (Sanskrit concept) |
| ವಿಭಕ್ತಿ | vibhakti | Case suffix / case ending |
| ಲಿಂಗ | liṅga | Gender (grammatical gender classification of nouns) |
| ವಚನ | vacana | Number (singular/plural distinction) |
| ನಾಮಪದ | nāmapada | Noun |
| ಕ್ರಿಯಾಪದ | kriyāpada | Verb |
| ಗುಣಪದ | guṇapada | Adjective (uninflected; a distinctively Dravidian/Kannada word class) |
| ಪ್ರತ್ಯಯ | pratyaya | Suffix |
| ಸಮಾಸ | samāsa | Compound word |
| ಸಂಧಿ | sandhi | Phonological junction (changes at morpheme/word boundaries) |
| ಪದಕಂತೆ | padakante | Word cluster / phrase |
| ಕೇಶಿರಾಜ | kēśirāja | Keshiraja; author of Shabdamanidarpana (c. 1260 CE), the most important classical Old Kannada grammar |
| ಶಬ್ದಮಣಿದರ್ಪಣ | śabdamaṇidarpaṇa | “Mirror of Jewel-Words”; Keshiraja’s grammar of Old Kannada |
| ತತ್ಸಮ | tatsama | Sanskrit loanwords used unchanged in Kannada |
| ತದ್ಭವ | tadbhava | Nativized Sanskrit loans; Sanskrit words phonologically adapted into Kannada |
| ಕೃದಂತ | kṛdanta | Participial noun; a verbal form functioning as a noun |
| ಆಖ್ಯಾತ | ākhyāta | Finite verb form |
| ಸರ್ವನಾಮ | sarvanāma | Pronoun |
| ದ್ರಾವಿಡ | drāviḍa | Dravidian; the language family to which Kannada belongs |
File Index
| File | Contents |
|---|---|
nijakku-...-book.md | Full Kannada book text (archive.org, 11,791 lines) |
14-...-kn.md | Structured Kannada with TOC and section anchors |
14-...-kn-eke.md | Eke romanisation |
14-...-djvu.md | DjVu OCR text (alternative source) |
14-...-blog.md | 7 blog posts — ಶಬ್ದಮಣಿದರ್ಪಣ critique series |
14-...-claude-prompt.md | AI assistant prompt for Q&A about this book |
| Internet Archive | Original PDF + DjVu source |
Cross-References to Other DNS Bhat Works
| Related Book | Connection |
|---|---|
| 01 — Idu Kannadade Vyakarana | The positive grammar that corrects the errors diagnosed here; synthesises Dravidian-first approach |
| 07 — Kannada Barahada Sollarime | The modern comprehensive grammar reform that Book 14 historically contextualises |
| 08 — Kannadakke Mahaprana Yake Beda | Historical argument: when/why Sanskrit letters entered Kannada writing |
| 03 — Kannada Padagala Olarachane | Word structure / morphology — the Dravidian-correct framework Book 14 argues for |
| 22 — Sound Change | Theoretical backbone: comparative method used to evaluate Old Kannada phonology |
| Blog series in this folder | 7-part companion blog posts specifically on Shabdamani Darpana errors |