Same skeleton, different handwriting
Set the Kannada and Devanagari charts side by side and they match row for row, because both descend from Brahmi. Then look closer: Kannada stacks its clusters downward, and keeps two vowel letters Hindi never needed.
Here is a puzzle this project keeps handing people. Hindi and Kannada are unrelated languages, a case made at length in Unrelated languages, familiar words. Yet set the two alphabet charts side by side and they match row for row: ಕ ಖ ಗ ಘ lines up with क ख ग घ, ka kha ga gha, the same sounds in the same order, organized by the same logic. If the languages are strangers, why do their alphabets look like family?
Because they are family. Scripts have their own genealogy, quite separate from the languages they write, and both of these descend from Brahmi, the script of Ashoka’s inscriptions in the third century BCE. Devanagari comes down the northern line of Brahmi’s descendants. Kannada comes down the southern line, and on that side it has a sister rather than a cousin: the Kannada and Telugu scripts were effectively one script for centuries, usually referred to as Kannada-Telugu, and they drift visibly apart only in the later medieval period. Set a Kannada letter beside its Telugu counterpart today and the family resemblance is still hard to miss. One Brahmi skeleton, three handwritings, and between the two southern ones the handwriting is close to the whole of the difference.
Descent from Brahmi is why the two scripts agree on their deep design. Both are abugidas: every consonant letter carries an automatic a, so ಕ is ka and क is ka, and you add a mark to change the vowel or to silence it. Both order their letters by the Sanskrit grammarians’ phonetic grid, vowels first, then consonants row by row according to where in the mouth they are made, from the back of the throat forward to the lips. That chart is not a coincidence, and it is not a borrowing between Hindi and Kannada. It is a shared inheritance from one source, kept on both sides for two thousand years.
A script, though, is clothing and not anatomy. English and Finnish are unrelated languages that both wear the Latin alphabet, and nobody concludes from the shared letters that the languages are kin. The same separation runs straight through South Asia: Brahmi’s descendants write Indo-Aryan languages like Hindi and Marathi and Dravidian languages like Kannada, Telugu, and Tamil. The script family tree records who adapted letterforms from whom. It says nothing at all about the ancestry of the words those letters spell.
Where the two scripts visibly part company is the consonant cluster. Devanagari joins clusters along the line, using half forms and ligatures that run left to right. Kannada writes the second consonant underneath the first, as a small subscript form called an ottu, and the whole apparatus is the ottakshara ಒತ್ತಕ್ಷರ system. The name of the language demonstrates it: the double n in Kannada ಕನ್ನಡ is one ನ with a second tucked below it. Devanagari builds sideways. Kannada builds downward. Learners moving from one to the other usually find this the first genuinely unfamiliar thing, and the last one they stop having to think about. Neither rule is absolute: Devanagari stacks a handful of its own conjuncts vertically, and in Kannada a cluster that begins with r is written as a small mark above the line instead of an ottu below it, which is what sits over the ಚ of ಖರ್ಚು kharchu.
There is also a dot that Kannada never needed. Hindi added a nuqta, a dot below the letter, to write Perso-Arabic sounds Sanskrit did not have, giving ज़ za and फ़ fa and three others. Kannada’s everyday orthography has no nuqta at all. It writes its Perso-Arabic loans with the letters it already had, which is one small reason javaabu and jebu look as thoroughly native as they do. Persian on Kannada ground has the rest of that story.
Both scripts are nonetheless sized for Sanskrit, right down to letters the native vocabulary barely needs. Dravidian words have no aspirated consonants, yet the Kannada script carries the full set of aspirate letters. They earn their keep in the borrowed layers: the ಖ of ಖರ್ಚು kharchu and the ಭ of ಭಾಷೆ bhaashe are spelling sounds that arrived with loanwords, one Perso-Arabic and one Sanskrit. Each script is cut for the most demanding register of its language, and on both sides that register is Sanskrit.
The tailoring runs the other way too, and this is where Kannada has letters Hindi never developed. Kannada contrasts short e and o against long ē and ō, a distinction Hindi simply does not make, so the Kannada chart carries ಎ against ಏ and ಒ against ಓ where Devanagari as used for Hindi has one letter each. Kannada has also retired letters it stopped needing: older Kannada wrote ಱ and ೞ, and the modern language does not. Same workshop, different tailoring, and different centuries of alterations.
Romanization is exactly where that extra precision falls on the floor. Take ಬೆಳೆ and ಬೇಳೆ. The first is a crop standing in a field. The second is the split pulse in the pot. They differ only in the length of the first vowel, and the retroflex ಳ in the middle is a third sound the roman alphabet has no letter for. In the everyday texting roman this app uses for search and display both words come out as bele, which is honest about how people actually type and useless for telling them apart. The script never confuses them for a moment. Bhashegalu’s reading guide writes the pair as beLe and bELe, spending capital letters to carry what lowercase ascii cannot: a capital vowel is long, a capital consonant is retroflex.
So the resemblance between the charts is real, and it means something, just not what it appears to mean. The letters line up because the scripts are relatives. The words they spell do not, because the languages are strangers. Bhashegalu shows the script, the texting roman, and the reading guide together on every Kannada entry for precisely this reason: the roman gets you typing, and the script tells you the truth.
