1. The Kannada script
Kannada is written in the Kannada script, ಕನ್ನಡ ಲಿಪಿ (kannaḍa lipi, ‘Kannada script’), a Brahmic abugida closely related to the Telugu script. A consonant letter has an inherent vowel, and signs attached to consonants change or suppress that vowel. The resulting written units can combine several letters and signs in a single syllable-sized shape. This chapter describes those units, their representation in Unicode, and their transliteration into Latin script. (Wikibooks contributors 2026, § Alphabet, § Writing system)
The akshara and the inherent vowel
An akshara is a written unit organized around a vowel or a consonant base. A vowel at the start of such a unit can be written with an independent vowel letter. A consonant base can carry an inherent vowel, carry a dependent vowel sign, or take a virama that suppresses its vowel. Several consonants can share a unit when the virama links them into a cluster. Anusvara or visarga can follow a vowel-bearing unit. The term kagunita is used for the series of forms made by pairing a consonant with the vowels. (Wikibooks contributors 2026, § Akshara, § Writing system)
The inherent vowel is the short vowel transliterated a. The bare letter ಕ reads ka. A dependent vowel sign replaces its a, so ಕಿ reads ki. A virama gives ಕ್ k, with the vowel suppressed. The same distinction applies when a consonant is written in isolation or at the end of a word. A bare final consonant letter carries a in the spelling. A final consonant with a virama has no written vowel. These are spelling rules. A transliteration preserves the written vowel distinctions even where speech varies. (Wikibooks contributors 2026, § Akshara)
| Form | Transliteration | Gloss | Translation or value |
|---|---|---|---|
| ಕ | ka | K with inherent A | ka |
| ಕಿ | ki | K with dependent I | ki |
| ಕ್ | k | K with virama | Vowelless k |
| ಅ | a | Independent A | a |
| ಅಂ | aṃ | Independent A with anusvara | a followed by a nasal sign |
The forms in the table are constructed illustrations. Letter-and-sign glosses name the spelling components. The final column gives their reading or function when the form has no lexical translation.
An akshara is a unit of writing. Its graphic divisions follow the rules for dependent vowels and linked consonants. A normal akshara with a vowel corresponds broadly to a syllable, and a cluster written in one unit may contain more than one consonant. A consonant with a visible virama also has a written form, though it has no syllable nucleus of its own. (Wikibooks contributors 2026, § Akshara, § Writing system)
| Form | Transliteration | Gloss | Translation |
|---|---|---|---|
| ಅ · ಕ್ಕ | a · kka | elder.sister | ‘elder sister’ |
| ಬ · ರ | ba · ra | drought | ‘drought’ |
| ಕೆ · ಲ · ಸ | ke · la · sa | work | ‘work’ |
| ಪು · ಸ್ತ · ಕ | pu · sta · ka | book | ‘book’ |
These are constructed illustrations using ordinary words. The centered dots show where the written units divide. The words are spelled ಅಕ್ಕ, ಬರ, ಕೆಲಸ and ಪುಸ್ತಕ without the dots. In ಪುಸ್ತಕ the written unit ಸ್ತ contains both s and t, with a supplied at the end of the cluster. In ಕೆಲಸ the initial e of ke is written as a sign on ಕ. The first vowel of ಅಕ್ಕ is written with an independent letter because it has no preceding consonant in its unit.
Independent vowels and dependent vowel signs
An independent vowel letter supplies the vowel of a unit that begins with a vowel. It can occur at the beginning of a word or later in a word when a vowel starts a new written unit. A vowel after a consonant or consonant cluster within the same unit is written with a dependent sign, except for a, which is inherent. The independent and dependent forms represent corresponding vowels and occupy different positions in the spelling. (Wikibooks contributors 2026, § Vowels, § Akshara)
The ordinary alphabetic count has thirteen vowels. An extended list also includes ೠ ṝ, the long counterpart of ಋ ṛ. Its corresponding dependent sign is ೄ. The thirteen-vowel count and the extended list refer to different inventories. The extended vowel is shown separately in the table so that it can be recognized in text. (Wikibooks contributors 2026, § Vowels, § Akshara, § Writing system)
| Vowel value | Independent letter | Dependent sign | With ದ da | Reading |
|---|---|---|---|---|
| a | ಅ | Inherent | ದ | da |
| ā | ಆ | ◌ಾ | ದಾ | dā |
| i | ಇ | ◌ಿ | ದಿ | di |
| ī | ಈ | ◌ೀ | ದೀ | dī |
| u | ಉ | ◌ು | ದು | du |
| ū | ಊ | ◌ೂ | ದೂ | dū |
| ṛ | ಋ | ◌ೃ | ದೃ | dṛ |
| e | ಎ | ◌ೆ | ದೆ | de |
| ē | ಏ | ◌ೇ | ದೇ | dē |
| ai | ಐ | ◌ೈ | ದೈ | dai |
| o | ಒ | ◌ೊ | ದೊ | do |
| ō | ಓ | ◌ೋ | ದೋ | dō |
| au | ಔ | ◌ೌ | ದೌ | dau |
| ṝ, extended inventory | ೠ | ◌ೄ | ದೄ | dṝ |
The consonant forms in this paradigm are constructed spelling illustrations. The dotted circle isolates a combining sign for display. In a word, the sign attaches to the Kannada consonant or cluster. The table gives the written distinctions. In particular, ಇ and ಈ are i and ī, while ಎ and ಏ are e and ē, following the consonant-and-vowel combinations in the source. The symbols ṛ and ṝ identify the vocalic letters in transliteration. (Wikibooks contributors 2026, § Vowels, § Akshara)
The short and long members of the pairs a–ā, i–ī, u–ū, e–ē and o–ō have distinct spellings. For the a–ā pair after a consonant, the contrast is between the bare base and the base with a sign. For the other pairs, both members have dependent signs. The spellings for ai and au use one independent letter each when they begin a unit and one dependent vowel sign each when they follow a consonant. The two Latin letters in each transliteration represent a single choice in the Kannada vowel paradigm. (Wikibooks contributors 2026, § Vowels, § Akshara)
| Form | Transliteration | Gloss | Translation |
|---|---|---|---|
| ಅಲ್ಲಿ | alli | there | ‘there’ |
| ಆರು | āru | six | ‘six’ |
| ಇದು | idu | this | ‘this’ |
| ಈಜು | īju | swim | ‘swim’ |
| ಎಲೆ | ele | leaf | ‘leaf’ |
| ಏಳು | ēḷu | seven | ‘seven’ |
| ಕೈ | kai | hand | ‘hand’ |
| ಕಾಡು | kāḍu | forest | ‘forest’ |
| ಕೂಡು | kūḍu | join | ‘join’ |
| ಕೌಶಲ | kauśala | skill | ‘skill’ |
These are constructed illustrations. The first six words begin with independent vowels. The other four show a dependent sign on the initial consonant.
The vowel sign belongs to the whole written consonant unit. After a single consonant it supplies the vowel for that consonant. After a cluster it supplies the vowel following the last consonant in reading order. The visible position of the sign can be beside or above the main letter even when the vowel follows a subjoined consonant. The form of the complete unit can depend on the base and the sign chosen. The paradigm uses ದ consistently so that the differences between signs can be compared on one base. (Wikibooks contributors 2026, § Akshara, § Writing system)
Consonant letters and subjoined forms
The ordinary consonant inventory has thirty-four letters. Twenty-five are arranged in five rows according to place of articulation. Each row has the same five positions. The first two positions are voiceless, with the second aspirated. The next two are voiced, with the fourth aspirated. The fifth is nasal. Nine further letters fall outside these five rows. The values in the tables identify the letters in writing and are given without inherent a. (Wikibooks contributors 2026, § Consonants, § Writing system)
| Position in a row | Velar | Palatal | Retroflex | Dental | Labial |
|---|---|---|---|---|---|
| Voiceless unaspirated | ಕ, k | ಚ, c | ಟ, ṭ | ತ, t | ಪ, p |
| Voiceless aspirated | ಖ, kh | ಛ, ch | ಠ, ṭh | ಥ, th | ಫ, ph |
| Voiced unaspirated | ಗ, g | ಜ, j | ಡ, ḍ | ದ, d | ಬ, b |
| Voiced aspirated | ಘ, gh | ಝ, jh | ಢ, ḍh | ಧ, dh | ಭ, bh |
| Nasal | ಙ, ṅ | ಞ, ñ | ಣ, ṇ | ನ, n | ಮ, m |
| Base letter | Consonant value | Bare letter read with its vowel | With virama |
|---|---|---|---|
| ಯ | y | ya | ಯ್, y |
| ರ | r | ra | ರ್, r |
| ಲ | l | la | ಲ್, l |
| ವ | v | va | ವ್, v |
| ಶ | ś | śa | ಶ್, ś |
| ಷ | ṣ | ṣa | ಷ್, ṣ |
| ಸ | s | sa | ಸ್, s |
| ಹ | h | ha | ಹ್, h |
| ಳ | ḷ | ḷa | ಳ್, ḷ |
The palatal series uses c for ಚ and ch for ಛ. The latter therefore represents an aspirated consonant letter. The dots in ṭ, ḍ, ṇ, ṣ, and ḷ distinguish these letters from their undotted counterparts. In particular, ಣ ṇ and ನ n are separate consonant letters, as are ಳ ḷ and ಲ l. The five nasal letters in the first table are consonants in their own right and use the same vowel and virama rules as the other consonant letters. (Wikibooks contributors 2026, § Consonants, § Akshara)
| Historical letter | Consonant value here | Bare letter with its vowel | With virama | Later replacement described in the source |
|---|---|---|---|---|
| ಱ | ṟ | ṟa | ಱ್, ṟ | ರ, r |
| ೞ | ḻ | ḻa | ೞ್, ḻ | ಳ, ḷ |
The two historical letters are outside the count of thirty-four. Their transliterations preserve the distinction when they occur in older writing. The later replacements in the table describe changes in spelling practice. In a transcription of a text that actually uses ಱ or ೞ, the character present in the text determines the transliteration. (Wikibooks contributors 2026, § Consonants, § Obsolete Kannada letters)
A consonant may have a different shape inside a cluster. The first consonant usually provides the main body of the written unit. A later consonant can appear in a smaller or modified form attached to it. Such forms are commonly called subjoined forms, and a joined cluster is an ottakshara. Shape and position vary with the consonants involved. In particular, a form for ರ in a cluster can be positioned away from where its reading order would place a separate letter. The character sequence identifies the consonant and its order. (Wikibooks contributors 2026, § Akshara, § Writing system)
| Form | Transliteration | Gloss | Translation or value |
|---|---|---|---|
| ಕ್ಕ | kka | K joined to K, inherent A | kka |
| ಕ್ರ | kra | K joined to R, inherent A | kra |
| ರ್ಕ | rka | R joined to K, inherent A | rka |
| ಶಕ್ತಿ | śakti | strength | ‘strength’ |
| ಕ್ರಮ | krama | order | ‘order’ |
| ಅರ್ಥ | artha | meaning | ‘meaning’ |
These are constructed illustrations. Compare ಕ್ರ and ರ್ಕ. In ಕ್ರ the reading begins with k and proceeds to r. In ರ್ಕ it begins with r and proceeds to k. Both units end in inherent a. ಶಕ್ತಿ demonstrates a cluster with dependent i in ಕ್ತಿ. The vowels and consonants are identified from the full unit, including its subjoined form.
Virama and consonant clusters
The virama is the sign ್, encoded as U+0CCD. It follows a consonant letter and suppresses the inherent vowel of that letter. When the sequence stops there, the consonant has an explicit virama. When another consonant follows as part of the same written unit, the sequence can render as a conjunct. In a conjunct the virama normally has no visible stroke of its own. The cluster shape represents its effect. (Wikibooks contributors 2026, § Akshara, § Writing system)
The spelling order for a cluster begins with its first consonant. Write a virama between each pair of consonants that are to join without a vowel between them. After the last consonant, use no vowel sign for inherent a, use a dependent sign for another vowel, or use a final virama to suppress the vowel of that consonant too. The final position therefore has the same choices as the vowel position after a single consonant letter. An independent vowel begins its own written unit. (Wikibooks contributors 2026, § Akshara)
| Condition | Form | Transliteration | Gloss | Reading |
|---|---|---|---|---|
| Single consonant with inherent vowel | ತ | ta | T with inherent A | ta |
| Single consonant with suppressed vowel | ತ್ | t | T with virama | t |
| Linked consonants with inherent vowel | ಸ್ತ | sta | S joined to T, inherent A | sta |
| Linked consonants with dependent vowel | ಸ್ತಿ | sti | S joined to T, dependent I | sti |
| Linked consonants with final virama | ಸ್ತ್ | st | S joined to T, final virama | st |
| Full consonant units in succession | ಸತ | sata | S with inherent A, T with inherent A | sa-ta |
These are constructed spelling illustrations. The hyphen in sa-ta indicates a unit boundary. The table shows why both the intervening virama and the final vowel choice matter. ಕ್ಕ kka and ಕ್ಕಿ kki share the same doubled consonant sequence. The latter has a final dependent sign for i, while the former has inherent a.
A doubled consonant is written as a cluster with the same consonant letter on each side of the linking virama. A sequence of different consonants uses the same construction. For example, the spelling of ಅಮ್ಮ amma joins m to m, while the spelling of ಪುಸ್ತಕ pustaka joins s to t. A conjunct with a familiar overall shape can still be analyzed in terms of its consonant letters. In particular, ಕ್ಷ kṣa and ಜ್ಞ jña follow the cluster rule. Their traditional names or prominent shapes leave them outside the inventory of single consonant letters. (Wikibooks contributors 2026, § Consonants, § Akshara, § Writing system)
| Form | Transliteration | Gloss | Translation |
|---|---|---|---|
| ಅಮ್ಮ | amma | mother | ‘mother’ |
| ಕನ್ನಡ | kannaḍa | Kannada | ‘Kannada’ |
| ಪುಸ್ತಕ | pustaka | book | ‘book’ |
| ಕ್ಷೇತ್ರ | kṣētra | field | ‘field’ |
| ಜ್ಞಾನ | jñāna | knowledge | ‘knowledge’ |
These are constructed illustrations. The written units of ಕನ್ನಡ are ಕ ka, ನ್ನ nna and ಡ ḍa. ಕ್ಷೇತ್ರ has two cluster units, ಕ್ಷೇ kṣē and ತ್ರ tra. In ಕ್ಷೇ the sign for long ē supplies the vowel after both consonants. The a in ತ್ರ is inherent. ಜ್ಞಾನ likewise has dependent ā after the jñ cluster and inherent a in the following ನ na.
Anusvara, visarga and other signs
The traditional inventory groups anusvara and visarga together as yogavāhakas. They follow a unit that already has a vowel. The vowel can be an independent letter, an inherent vowel, or a vowel written with a dependent sign. The traditional forms ಅಂ and ಅಃ illustrate the marks on independent ಅ. Each of these forms consists of the vowel letter and a following sign. In the ordinary alphabetic count the two yogavāhakas are listed separately from the thirteen vowels and thirty-four consonants. (Wikibooks contributors 2026, § Yogavaahaka's, § Akshara, § Writing system)
| Sign | Name | Transliteration | With independent A | With a consonant and its inherent vowel |
|---|---|---|---|---|
| ◌ಂ | Anusvara | ṃ | ಅಂ, aṃ | ದಂ, daṃ |
| ◌ಃ | Visarga | ḥ | ಅಃ, aḥ | ದಃ, daḥ |
The forms in the table are constructed spelling illustrations. Anusvara marks a nasal element after the vowel. Before a consonant, the pronunciation of that nasal can depend on the following sound. A spelling-based transliteration uses ṃ for the sign in each environment. A nasal written as an ordinary consonant is transliterated with that letter’s own value. Thus ṃ identifies the sign in the text, whereas ṅ, ñ, ṇ, n and m identify the five consonant letters. (Wikibooks contributors 2026, § Yogavaahaka's, § Consonants, § Writing system)
Visarga represents a postvocalic element conventionally transliterated ḥ. It occurs in forms such as ದುಃಖ duḥkha, ‘sorrow’, and ಪುನಃ punaḥ, ‘again’. Its two dots belong to one Kannada character. The Latin colon is a punctuation character with a separate function. The plain h of transliteration corresponds to the consonant ಹ, which has the usual consonant vowel and virama options. (Wikibooks contributors 2026, § Yogavaahaka's, § Akshara)
| Form | Transliteration | Gloss | Translation |
|---|---|---|---|
| ಅಂಕ | aṃka | mark | ‘mark’ |
| ಸಂಪತ್ತು | saṃpattu | wealth | ‘wealth’ |
| ಸಿಂಹ | siṃha | lion | ‘lion’ |
| ದುಃಖ | duḥkha | sorrow | ‘sorrow’ |
| ಪುನಃ | punaḥ | again | ‘again’ |
These are constructed illustrations. In ಅಂಕ the nasal sign follows an independent vowel. In ಸಂಪತ್ತು it follows the inherent vowel in ಸ sa. In ಸಿಂಹ it follows the dependent i in ಸಿ si. Each condition leaves the spelling with the same anusvara sign. In ದುಃಖ the visarga follows the written u of ದು du. In ಪುನಃ it follows the inherent a of ನ na.
Virama serves a different position in the written unit. It goes after a consonant to suppress its vowel. Anusvara and visarga go after a vowel-bearing unit and retain that vowel. Unicode also has two Kannada length-mark characters, ೕ U+0CD5 and ೖ U+0CD6. They occur in decomposed representations of particular dependent vowel signs. Their combinations are given in the next section. The vowel sign expressed by the combination supplies one vowel value in the akshara.
Unicode character order, normalization and rendering
Unicode represents Kannada letters and signs as characters in logical order. The text sequence begins with the independent vowel or the first consonant of the unit. For a cluster, each linking virama precedes the consonant that it joins. The dependent vowel of the whole cluster comes after the last consonant, and a following anusvara or visarga comes after the vowel-bearing sequence. The font and shaping engine use these characters to produce the visible forms, which can include a subjoined consonant, a repositioned mark or a combined glyph. The source’s vowel and conjunct patterns can therefore be written as sequences of encoded characters. (Wikibooks contributors 2026, § Akshara, § Writing system)
| Form | Unicode character sequence | Transliteration | Gloss and reading |
|---|---|---|---|
| ಅ | U+0C85 | a | Independent A, a |
| ಕ | U+0C95 | ka | K with inherent A, ka |
| ಕಾ | U+0C95 U+0CBE | kā | K and dependent AA, kā |
| ಕಿ | U+0C95 U+0CBF | ki | K and dependent I, ki |
| ಕ್ | U+0C95 U+0CCD | k | K and virama, k |
| ಕ್ಕ | U+0C95 U+0CCD U+0C95 | kka | K joined to K, inherent A, kka |
| ಕ್ತಿ | U+0C95 U+0CCD U+0CA4 U+0CBF | kti | K joined to T, dependent I, kti |
| ಕ್ರ | U+0C95 U+0CCD U+0CB0 | kra | K joined to R, inherent A, kra |
| ರ್ಕ | U+0CB0 U+0CCD U+0C95 | rka | R joined to K, inherent A, rka |
| ಕಂ | U+0C95 U+0C82 | kaṃ | K with inherent A and anusvara, kaṃ |
| ಕಿಂ | U+0C95 U+0CBF U+0C82 | kiṃ | K with dependent I and anusvara, kiṃ |
| ಕಃ | U+0C95 U+0C83 | kaḥ | K with inherent A and visarga, kaḥ |
The table contains constructed spelling illustrations. In ಕ್ತಿ the character for the i sign follows ತ. Its placement on the rendered cluster does not change its position in the character sequence. In ಕ there is no encoded vowel after U+0C95. In ಕಾ U+0CBE is present. This is the encoded distinction between inherent and dependent vowels. ಅ has its own code point and begins a vowel-bearing unit directly.
Some Kannada dependent vowel signs have canonically equivalent Unicode representations. One representation uses a single sign character and the other uses a sequence of sign components. Normalization Form C (NFC) composes the equivalent sequences in the table. Normalization Form D (NFD) expands the compound signs into their canonical components. The full NFD sequence is shown, including the expansion of the short o component in long ō. Both representations carry the same vowel and attach to the same consonant unit.
| Vowel | Sign in NFC | Sign code point in NFC | Sign code points in NFD | Constructed form with ಕ and its reading |
|---|---|---|---|---|
| Long ī | ◌ೀ | U+0CC0 | U+0CBF U+0CD5 | ಕೀ, kī, K with long I |
| Long ē | ◌ೇ | U+0CC7 | U+0CC6 U+0CD5 | ಕೇ, kē, K with long E |
| ai | ◌ೈ | U+0CC8 | U+0CC6 U+0CD6 | ಕೈ, kai, K with AI |
| Short o | ◌ೊ | U+0CCA | U+0CC6 U+0CC2 | ಕೊ, ko, K with short O |
| Long ō | ◌ೋ | U+0CCB | U+0CC6 U+0CC2 U+0CD5 | ಕೋ, kō, K with long O |
For example, ಕೊಡ koḍa, ‘pot’ (constructed), begins with the unit ಕೊ ko. The normalized sequence for that unit in NFC is U+0C95 U+0CCA. In NFD it is U+0C95 U+0CC6 U+0CC2. In the second sequence the dependent e and ū components together encode the sign for short o. They have one vowel reading in this environment. Similarly, the sign sequence for ai encodes a single vowel sign on the consonant. A character-by-character transcription of the components before they are recognized as a compound sign would give the wrong reading.
Canonical equivalence concerns the encoded representation of a spelling. The spellings with short and long vowels remain distinct in both forms. A sequence with an intervening virama also remains distinct from one with an inherent vowel between the consonants. Conjunct rendering forms the visible cluster while the encoded letters and linking virama remain in the text. These distinctions matter when comparing spellings, reading a string’s character sequence, or determining where a written unit ends.
| Comparison | Forms and transliteration | Gloss and translation | Written distinction |
|---|---|---|---|
| Short and long a | ಬರ bara, ಬಾರ bāra | ‘drought’, ‘come’ | Inherent a and dependent ā in the first unit |
| Unlinked and linked consonants | ಕಕ kaka, ಕ್ಕ kka | K.A + K.A, K + K.A | Two vowel-bearing units and one cluster |
| Vowel and nasal sign | ಕಿ ki, ಕಿಂ kiṃ | K.I, K.I.ANUSVARA | The second unit ends with a nasal sign |
| Vowel and virama | ಕ ka, ಕ್ k | K.A, vowelless K | The second unit has an explicit virama |
These are constructed illustrations. One akshara can contain one character, several characters for a base and signs, or several consonants with intervening viramas. Its visual width and the number of visible parts depend on shaping. The spelling can therefore be analyzed by following the character sequence and its vowel and cluster rules even when different fonts lay out the unit differently.
Transliteration and romanized input
The transliteration used here represents Kannada spelling in Latin letters. It writes the inherent a after a bare consonant or vowel-bearing cluster and writes the specified vowel after a dependent sign. It gives independent vowels the same Latin values as their dependent counterparts. A virama has no separate Latin letter in ordinary running transliteration. Its effect is shown by the absence of a vowel after the consonant it follows. The mark values ṃ and ḥ remain visible in transliteration. (Wikibooks contributors 2026, § Vowels, § Consonants, § Yogavaahaka's, § Akshara)
| Kannada spelling pattern | Form | Transliteration | Gloss and reading |
|---|---|---|---|
| Independent vowel | ಏ | ē | Independent long E, ē |
| Corresponding dependent vowel | ಕೇ | kē | K with long E, kē |
| Inherent vowel | ಕ | ka | K with inherent A, ka |
| Final virama | ಕ್ | k | K with suppressed vowel, k |
| Linking virama | ಕ್ರಿ | kri | K joined to R, dependent I, kri |
| Anusvara after a vowel | ಕಂ | kaṃ | K with inherent A and nasal sign, kaṃ |
| Visarga after a vowel | ಕಃ | kaḥ | K with inherent A and visarga, kaḥ |
These are constructed spelling illustrations. The presence or absence of the vowel in the Latin form follows the written structure. For example, kri represents a cluster with no vowel between its consonants. A spelling written with a between the consonants would yield kari.
In this convention the consonant digraphs ending in h are kh, gh, ch, jh, ṭh, ḍh, th, dh, ph and bh. Each corresponds to one aspirated letter in the consonant table. In a cluster containing independent consonantal ಹ h, the h instead represents a second consonant. For example, th in ಥ tha (constructed, gloss TH.A, reading tha) identifies a single aspirated dental letter. The sequence t plus h in ತ್ಹ t-ha (constructed, gloss T joined to H with inherent A, reading tha as two consonants followed by a) represents two letters linked by virama. The explanatory hyphen shows the letter boundary where reading the Latin string alone would be ambiguous. In this convention the vowel digraphs ai and au each represent a single vowel in the spelling. (Wikibooks contributors 2026, § Vowels, § Consonants, § Akshara)
| Form | Transliteration | Gloss | Translation |
|---|---|---|---|
| ಖಾಲಿ | khāli | empty | ‘empty’ |
| ಛತ್ರಿ | chatri | umbrella | ‘umbrella’ |
| ಘಂಟೆ | ghaṇṭe | bell | ‘bell’ |
| ಧನ | dhana | wealth | ‘wealth’ |
| ಕಣ್ಣು | kaṇṇu | eye | ‘eye’ |
| ಹಣ್ಣು | haṇṇu | fruit | ‘fruit’ |
These are constructed illustrations. ಖಾಲಿ, ಛತ್ರಿ, ಘಂಟೆ and ಧನ begin with single aspirated letters. The tr in chatri represents a cluster in the second akshara. The doubled ṇṇ in the last two examples represents two instances of the same retroflex nasal, linked by virama. Each spelling distinction is represented at the appropriate level, whether by a Latin digraph for one consonant or by a sequence of symbols for a Kannada cluster.
Romanized input may omit vowel length and consonant diacritics or use ordinary Latin digraphs for sounds written with diacritics here. Its interpretation then depends on the input convention and on which Kannada word is intended. For example, an input convention that drops the macron and underdot can map all three of the following words to bale.
| Kannada form | Spelling-based transliteration | Possible plain Latin input | Gloss | Translation |
|---|---|---|---|---|
| ಬಲೆ | bale | bale | net | ‘net’ |
| ಬಳೆ | baḷe | bale | bangle | ‘bangle’ |
| ಬಾಳೆ | bāḷe | bale | banana.plant | ‘banana plant’ |
These are constructed illustrations. In the first two spellings the contrast lies between ಲ l and ಳ ḷ. In the third the first vowel is also long. A reader can often recover the intended word from context, but the plain input alone leaves both distinctions unspecified. The full transliteration preserves them. Historical ಱ and ೞ likewise have distinct values here, ṟ and ḻ, so that historical spelling survives transliteration even when the later replacements ರ and ಳ are more familiar. (Wikibooks contributors 2026, § Obsolete Kannada letters)
| Line | Constructed sentence |
|---|---|
| Form | ಅಮ್ಮ ಪುಸ್ತಕ ಓದಿದಳು. |
| Transliteration | amma pustaka ōdidaḷu. |
| Gloss | mother book read.PST.3SG.F |
| Translation | ‘Mother read a book.’ |
The sentence brings together a doubled consonant in ಅಮ್ಮ, a cluster of different consonants in ಪುಸ್ತಕ, an independent long vowel in ಓದಿದಳು, and the retroflex lateral ಳ in the last akshara. Each appears in the Latin line according to its written form.
How Mova handles Kannada script
tools/langseg assigns a language at the sentence and span level. Its script and lexicon coverage leaves Kannada script in other. Its Latin-script coverage does not identify romanized Kannada as kn. The multilingual parser tools/mova-parse needs a level-2 model for each language it parses. There is no such model for kn.
tools/mova-parse reports the absence of a model in the # parse comment. For tokens left unannotated, it puts _ in LEMMA, UPOS, XPOS, FEATS, HEAD and DEPREL, with Parse=None in MISC. Mova holds no annotated Kannada corpus, so it has no measured Kannada treebank baseline.
The Kannada script rules in this chapter have yet to be implemented in Mova. Squid reads each grammatical point as a claim and queues rules and closed lists that can be formalized. Implementation starts with the distinctions needed by a tokenizer, a tagger and a Universal Dependencies treebank baseline.
Sources for this chapter
- Wikibooks contributors 2026. Kannada (English Wikibooks) — CC BY-SA 4.0 (Wikibooks; dump enwikibooks-20261001) — attribution: Wikibooks contributors, page histories at the source URL
This page in Markdown → https://md.mova.elamur.ai/free/en/docs/languages/kn/01-kannada-script/