L'interface est traduite, mais les explications de cette page sont encore rédigées en anglais — le texte pédagogique n'a pas encore été traduit.

Accueil / Alphabets / Devanagari (Hindi)

Devanagari (Hindi) · देवनागरी

Not an alphabet but an ABUGIDA: each consonant already contains the vowel /ə/, so क is 'kə', not 'k'. To say anything else you add a matra — a vowel sign hung above, below, before or after the letter — and to say nothing at all you add the virama ्, which kills the inherent vowel. The consonants are arranged as a phonetic grid, five places of articulation × five manners, and reading down a column is a pronunciation lesson. Two Hindi habits are not in the script and have to be learned separately: the final inherent vowel goes silent (राम is 'Raam'), and a dot called a nukta turns Sanskrit letters into the Persian and Arabic sounds Hindi borrowed — ज /d͡ʒ/ becomes ज़ /z/.

Brahmi → Gupta → Nagari (11th c.). Cousin to Bengali, Gujarati, Gurmukhi, Tibetan and Thai · s'écrit de gauche à droite · utilisé ici par Hindi

51 signs

#	Signe	Matra	Nom	Romanisation	API	Groupe	Note
1	अ	—	a	`a`	ə	Vowels	The default. Every consonant already contains it — क is /kə/, not /k/.
2	आ	ा	aa	`aa`	aː	Vowels	Long. The contrast with अ is length AND quality: /ə/ vs /aː/.
3	इ	ि	i	`i`	ɪ	Vowels	Short i. Its matra is written BEFORE the consonant it follows — the one real trap.
4	ई	ी	ii	`ii`	iː	Vowels	Long ee.
5	उ	ु	u	`u`	ʊ	Vowels	Short u, as in put.
6	ऊ	ू	uu	`uu`	uː	Vowels	Long oo.
7	ऋ	ृ	ri	`ri`	ɾ · ɪ	Vowels	A Sanskrit vocalic r. In Hindi it is simply said /ɾɪ/.
8	ए	े	e	`e`	eː	Vowels	A pure long e — no glide off the end as in English day.
9	ऐ	ै	ai	`ai`	ɛː	Vowels	Not a diphthong in modern Hindi: one open /ɛː/.
10	ओ	ो	o	`o`	oː	Vowels	A pure long o, again with no glide.
11	औ	ौ	au	`au`	ɔː	Vowels	One open /ɔː/, as in caught.
12	क	—	कk	`k`	k	Velar — क वर्ग	Unvoiced, unaspirated — the k of skin, with no puff.
13	ख	—	खkh	`kh`	kʰ	Velar — क वर्ग	Unvoiced, ASPIRATED — the k of kin, with a strong puff.
14	ग	—	गg	`g`	ɡ	Velar — क वर्ग	Voiced, unaspirated.
15	घ	—	घgh	`gh`	ɡʱ	Velar — क वर्ग	Voiced and BREATHY. English has no such sound; the voicing continues through the puff.
16	ङ	—	ङng	`ng`	ŋ	Velar — क वर्ग	The nasal of this row. Almost never written alone.
17	च	—	चch	`ch`	t͡ʃ	Palatal — च वर्ग	Unaspirated ch.
18	छ	—	छchh	`chh`	t͡ʃʰ	Palatal — च वर्ग	Aspirated ch.
19	ज	—	जj	`j`	d͡ʒ	Palatal — च वर्ग	Voiced j.
20	झ	—	झjh	`jh`	d͡ʒʱ	Palatal — च वर्ग	Voiced breathy j.
21	ञ	—	ञny	`ny`	ɲ	Palatal — च वर्ग	The nasal of this row.
22	ट	—	टt	`T`	ʈ	Retroflex — ट वर्ग	RETROFLEX t — tongue curled back to the roof. Not the English t.
23	ठ	—	ठth	`Th`	ʈʰ	Retroflex — ट वर्ग	Retroflex and aspirated.
24	ड	—	डd	`D`	ɖ	Retroflex — ट वर्ग	Retroflex d.
25	ढ	—	ढdh	`Dh`	ɖʱ	Retroflex — ट वर्ग	Retroflex, voiced, breathy — the hardest sound on this page.
26	ण	—	णn	`N`	ɳ	Retroflex — ट वर्ग	Retroflex nasal.
27	त	—	तt	`t`	t	Dental — त वर्ग	DENTAL t — tongue on the back of the teeth. This is the one closest to French or Spanish t.
28	थ	—	थth	`th`	tʰ	Dental — त वर्ग	Dental and aspirated. Never the th of think.
29	द	—	दd	`d`	d	Dental — त वर्ग	Dental d.
30	ध	—	धdh	`dh`	dʱ	Dental — त वर्ग	Dental, voiced, breathy.
31	न	—	नn	`n`	n	Dental — त वर्ग	The nasal of this row, and the commonest n in Hindi.
32	प	—	पp	`p`	p	Labial — प वर्ग	Unaspirated p — the p of spin.
33	फ	—	फph	`ph`	pʰ	Labial — प वर्ग	Aspirated p. With a nukta (फ़) it becomes /f/.
34	ब	—	बb	`b`	b	Labial — प वर्ग	b as in bed.
35	भ	—	भbh	`bh`	bʱ	Labial — प वर्ग	Voiced breathy b.
36	म	—	मm	`m`	m	Labial — प वर्ग	The nasal of this row.
37	य	—	यy	`y`	j	Approximants — य र ल व	y as in yes.
38	र	—	रr	`r`	ɾ	Approximants — य र ल व	A single tap, as in Spanish pero.
39	ल	—	लl	`l`	l	Approximants — य र ल व	Always clear — never the dark l of English full.
40	व	—	वv	`v`	ʋ	Approximants — य र ल व	Between v and w. Neither English sound is quite right, and Hindi does not distinguish them.
41	श	—	शsh	`sh`	ʃ	Sibilants and ह	sh as in ship.
42	ष	—	षsh	`Sh`	ʃ	Sibilants and ह	Retroflex in Sanskrit; in Hindi said identically to श.
43	स	—	सs	`s`	s	Sibilants and ह	s as in sun.
44	ह	—	हh	`h`	ɦ	Sibilants and ह	VOICED h — the vocal cords keep going. It also lowers the vowel beside it.
45	क़	—	क़q	`q`	q	Nukta letters	क with a dot: the uvular q of Arabic and Persian loanwords.
46	ख़	—	ख़x	`x`	x	Nukta letters	The ch of Bach.
47	ग़	—	ग़g	`G`	ɣ	Nukta letters	A voiced version of the same — French r, roughly.
48	ज़	—	ज़z	`z`	z	Nukta letters	z as in zoo. Very common: ज़्यादा, अंग्रेज़ी.
49	ड़	—	ड़r	`R`	ɽ	Nukta letters	A retroflex FLAP — the tongue curls back and strikes forward. Not in Sanskrit; very much in Hindi.
50	ढ़	—	ढ़rh	`Rh`	ɽʱ	Nukta letters	The same flap, breathy.
51	फ़	—	फ़f	`f`	f	Nukta letters	f as in fish.

Groups: Vowels (independent form, and the matra used after a consonant) · Velar — क वर्ग (back of the tongue against the soft palate) · Palatal — च वर्ग (body of the tongue against the hard palate) · Retroflex — ट वर्ग (tongue curled back. The row English does not have) · Dental — त वर्ग (tongue on the teeth, not the ridge behind them) · Labial — प वर्ग (the lips) · Approximants — य र ल व (the semivowels) · Sibilants and ह (श ष स ह) · Nukta letters (a dot below, for the Persian and Arabic sounds Hindi borrowed)

Les combinaisons les plus fréquentes

Les suites que vous croiserez sans arrêt — apprenez-les comme des blocs.

Suite	Son	Ce qui se passe
क + ् + ष = क्ष	kʃ	A conjunct: the vowel-killer ् (virama) welds two consonants together. क्ष is common enough to be taught as its own letter.
त + ् + र = त्र	tɾ	Another everyday conjunct — त्रिशूल, मित्र.
ज + ् + ञ = ज्ञ	ɡj	Written j + ny, and said /ɡj/ in Hindi. Pure convention: ज्ञान is 'gyaan'.

Les règles du système

Comment les signes se comportent ensemble.

Suite	Son	Ce qui se passe
◌ं (anusvara)	ŋ, n, m	Nasalises, and takes the place of articulation from whatever follows: हिंदी is /ɦɪndiː/.
◌ँ (chandrabindu)	̃	Nasalises the VOWEL itself, with no consonant: हाँ /ɦãː/, 'yes'.
◌् (virama)	—	Kills the inherent /ə/. Needed because every consonant carries a vowel it did not ask for.
final ◌ (schwa)	—	The inherent /ə/ goes silent at the end of a word — राम is 'Raam', not 'Raama'. Sanskrit keeps it; Hindi drops it.