Devanagari Table
| Letter | IAST | IPA | Articulation | Series |
|---|---|---|---|---|
| क | ka | k | velar | stop 1 (plain) |
| ख | kha | kʰ | velar | stop 2 (aspirated) |
| ग | ga | ɡ | velar | stop 3 (voiced) |
| घ | gha | ɡʱ | velar | stop 4 (voiced asp.) |
| ङ | ṅa | ŋ | velar | nasal of the row |
| ह | ha | ɦ | velar | breathed (aspirate) |
| च | ca | tʃ | palatal | stop 1 (plain) |
| छ | cha | tʃʰ | palatal | stop 2 (aspirated) |
| ज | ja | dʒ | palatal | stop 3 (voiced) |
| झ | jha | dʒʱ | palatal | stop 4 (voiced asp.) |
| ञ | ña | ɲ | palatal | nasal of the row |
| य | ya | j | palatal | glide |
| श | śa | ʃ | palatal | sibilant |
| ट | ṭa | ʈ | retroflex | stop 1 (plain) |
| ठ | ṭha | ʈʰ | retroflex | stop 2 (aspirated) |
| ड | ḍa | ɖ | retroflex | stop 3 (voiced) |
| ढ | ḍha | ɖʱ | retroflex | stop 4 (voiced asp.) |
| ण | ṇa | ɳ | retroflex | nasal of the row |
| र | ra | r | retroflex | liquid |
| ष | ṣa | ʂ | retroflex | sibilant |
| त | ta | t̪ | dental | stop 1 (plain) |
| थ | tha | t̪ʰ | dental | stop 2 (aspirated) |
| द | da | d̪ | dental | stop 3 (voiced) |
| ध | dha | d̪ʱ | dental | stop 4 (voiced asp.) |
| न | na | n | dental | nasal of the row |
| ल | la | l | dental | liquid |
| स | sa | s | dental | sibilant |
| प | pa | p | labial | stop 1 (plain) |
| फ | pha | pʰ | labial | stop 2 (aspirated) |
| ब | ba | b | labial | stop 3 (voiced) |
| भ | bha | bʱ | labial | stop 4 (voiced asp.) |
| म | ma | m | labial | nasal of the row |
| व | va | ʋ | labial | glide |
| Vowel | IAST | After प |
|---|---|---|
| अ | a | प |
| आ | ā | पा |
| इ | i | पि |
| ई | ī | पी |
| उ | u | पु |
| ऊ | ū | पू |
| ऋ | ṛ | पृ |
| ॠ | ṝ | पॄ |
| ऌ | ḷ | पॢ |
| ॡ | ḹ | पॣ |
| ए | e | पे |
| ऐ | ai | पै |
| ओ | o | पो |
| औ | au | पौ |
Devanagari - the "divine city" script - is the writing system of Hindi, Marathi, Nepali and the liturgical texts of Sanskrit, and its consonant chart is the oldest organized sound map still in daily use: the 33 consonants are not listed alphabetically by shape but arranged in a five-by-five grid (the varga) that sorts every stop by where the tongue goes and how the sound is voiced, twenty-five centuries before the IPA drew its first chart. Each row walks the same five-step pattern - unvoiced, unvoiced aspirated, voiced, voiced aspirated, nasal - so learning the grid is learning phonology, not just shapes.
Two mechanical facts make the script work. First, it is an abugida: every consonant letter carries an implied a (क is ka, not k), which is suppressed with a virama or replaced by one of the vowel signs - and in modern Hindi that final a usually goes silent, so मुंबई is Mumbai, not Mumbaia. Second, every letter hangs from the shirorekha, the horizontal headstroke that connects a whole word into one line (देवनागरी reads as a single ribbon), which is why Devanagari text looks striped: the line of the writing is drawn, not implied.
How to use
- Learn the grid row by row, not letter by letter: each varga row is the same five-part chord - क ख ग घ ङ (velar) moves to च छ ज झ ञ (palatal), then the retroflex ट ठ ड ढ ण, the dental त थ द ध न, and the labial प फ ब भ म - so you learn five positions times five voicings, and the sonorants (य र ल व) plus ह and the three sibilants fill in around the frame.
- Practice the vowels on the letter प as your anchor - the Barakhadi drill: प is pa by itself, पा pa, पि pi, पी pi, पु pu, पू pu, and so on - because every vowel in the table exists in two costumes, a standalone letter after another vowel and a diacritic strapped to a consonant, and the Barakhadi is the costume change rehearsed twelve times.
- Watch the headstroke when you write: copy a full word like देवनागरी and draw the shirorekha as one stroke at the end - the vertical strokes rise to meet it, and words that break the line (conjunct consonants like क्त, where letters stack into half-forms) are the exception that proves the rule.
Frequently asked questions
Why are the consonants arranged in a grid?
Because the grid is the point. Sanskrit grammarians - the tradition that runs from Panini's 4th-century-BCE phonology to the alphabet's own ordering - classified consonants by articulation: velar (क-series, tongue root at the soft palate), palatal (च-series, tongue body at the hard palate), retroflex (ट-series, tongue tip curled up), dental (त-series, tongue tip at the teeth) and labial (प-series, the lips). Crossing that with five voicings - plain, aspirated, voiced, voiced-aspirated, nasal - fills twenty-five cells, and the leftover sonorants, sibilants and ह complete the 33. The same five-vowel-by-five-consonant logic later seeded the tabular arrangement of the Japanese kana chart and informed the IPA's own layout: Devanagari's table is not a convenience, it is phonology you can memorize.
Why does every consonant secretly end in a?
Because the script is an abugida, built for Sanskrit where every consonant in the texts really did take a following a - so क is read ka and द da, and no vowel sign means "short a attached", not "no vowel". To strip the vowel you add the virama (्): क् is a bare k. Modern Hindi inherits the spelling but not the pronunciation - word-final inherent a is almost always silent, so कमल is kamal not kamala, and the schwa survives mainly inside words where phonology demands a beat (नमक is na-mak). The table shows every consonant with its a attached because that is the dictionary form; the FAQ is the warning label.
What is the line above the letters?
The shirorekha, literally "head line" - the horizontal bar from which every letter hangs, and the feature that makes Devanagari instantly recognizable: a whole word देवनागरी is drawn as one continuous ribbon with a roof. In handwriting the line often goes on last, one sweep across the word; in some fonts and decorative styles it thins or breaks, but printed Devanagari keeps it because the line is what clusters letters into words - without it, the verticals would float apart. It has no phonetic value at all; it is pure infrastructure, like the spaces between words in Latin script.
Which languages use Devanagari?
Hindi (alongside English, one of India's two official languages), Marathi, Nepali and Sanskrit are the big four, plus Konkani, Bhojpuri and a long tail of regional languages - by speaker count it ranks among the most widely used scripts in the world. The consonant grid is shared, but the languages lean on it differently: Sanskrit uses the full Vedic inventory including the vocalic ऋ; Hindi drops word-final inherent a and pronounces the sibilants as three near-identical s-sounds; Marathi keeps the extra retroflex ळ (listed in the Unicode chart as an additional letter) and rounds its vowels - the ऍ ऑ forms exist mostly for Marathi and English loanwords. Learn the grid once and the adjustments are footnotes, which is why one script serves a subcontinent.
Did our numbers come from Devanagari?
The digits did - the shapes travelled, the script did not. The nine Devanagari digits (० १ २ ३ ४ ५ ६ ७ ८ ९) and zero descend from the Brahmi numerals, were carried west by Arab mathematicians as the "Hindu numerals" (al-Khwarizmi's algebra borrowed them around 820 CE), and reached Europe as Fibonacci's "Arabic numerals" - a thousand-year relay that Devanagari still preserves in near-original form. Read the digits' shapes against 0123456789 and the family resemblance in several (३→3, ६→6, ७→7) is visible. The script itself never left the subcontinent; its numerals conquered the world without it.