Learning Sanskrit

Sanskrit Pronunciation Guide: How to Say Every Sound Correctly

How to pronounce Sanskrit correctly: vowel length, aspirated and retroflex consonants, the three s-sounds, anusvara and visarga, and common mistakes to avoid.

On this page
  1. Four Rules That Fix Most Mistakes
  2. Vowels
  3. Consonants by Place in the Mouth
  4. Semivowels, Sibilants and h
  5. Anusvāra and Visarga
  6. Doubled Consonants and Conjuncts
  7. Syllable Weight and Rhythm
  8. Common Mistakes for English and Hindi Speakers
  9. How to Practise

Sanskrit spelling is almost perfectly phonetic: each letter has one sound, and every letter is pronounced. Once you know the sounds, you can pronounce any word you can read. This guide uses IAST romanization alongside Devanagari. If you’re not familiar with the diacritics, see our IAST transliteration chart. For the full alphabet layout, see the Sanskrit varnamala.

Pronunciation traditions vary somewhat by region and lineage. The descriptions below follow the widely taught standard and note common variations.

Four Rules That Fix Most Mistakes

  1. Vowel length matters. A long vowel (ā, ī, ū) is held for about twice as long as a short one, and the difference can change meaning.
  2. “h” after a consonant means a puff of air, not a new sound. ph is p + breath (as in “up-hill”), never f. th is t + breath, never the th in “this”.
  3. Dental and retroflex are different sounds. t, d, n touch the teeth; ṭ, ḍ, ṇ curl the tongue back.
  4. Pronounce every vowel, including the final a. It’s Rāma, yoga, Veda, not “Rām”, “yog”, “Ved”.

Vowels

Short and Long Vowels

Letter IAST Approximate sound
अ a the u in “but” (a short, neutral sound)
आ ā the a in “father”, held longer
इ i the i in “sit”
ई ī the ee in “feet”, held longer
उ u the u in “put”
ऊ ū the oo in “food”, held longer

Short a is the most frequent vowel in Sanskrit, and English speakers often get it wrong by saying it like the a in “cat”. Keep it short and relaxed.

ṛ, ṝ and ḷ

ṛ (ऋ) is a vocalic r, an r-sound that acts as the vowel of a syllable, as in kṛṣṇa or ṛṣi. Many teachers recommend a short, lightly tapped r with only a trace of vowel colour. In practice, much of North India says it as “ri” and parts of western and southern India say it as “ru”. Both are common in traditional recitation. ṝ is the long form. ḷ (ऌ), a vocalic l, is rare.

e, ai, o, au

  • e (ए): like the ay in “say”, but held steady without gliding. Always long.
  • ai (ऐ): a diphthong, roughly the ai in “aisle”.
  • o (ओ): like the o in “go”, held steady. Always long.
  • au (औ): a diphthong, roughly the ow in “cow”.

Consonants by Place in the Mouth

The Five Groups

Group Where Sounds
Velar back of tongue against soft palate k, kh, g, gh, ṅ
Palatal middle of tongue against hard palate c, ch, j, jh, ñ
Retroflex tongue tip curled back to roof of mouth ṭ, ṭh, ḍ, ḍh, ṇ
Dental tongue tip against upper teeth t, th, d, dh, n
Labial lips p, ph, b, bh, m

Notes:

  • c is always the ch in “church” (never k or s). ch is the same sound with more breath.
  • ṅ is the ng in “sing”. ñ is the ny in “canyon”.

Aspiration and Voicing

Each group has four stops:

  • Voiceless unaspirated (k, c, ṭ, t, p): softer than English. English p at the start of a word carries a small puff of air, but Sanskrit p doesn’t; it’s closer to the p in “spin”.
  • Voiceless aspirated (kh, ch, ṭh, th, ph): a clear puff of air after the sound.
  • Voiced unaspirated (g, j, ḍ, d, b): as in English.
  • Voiced aspirated (gh, jh, ḍh, dh, bh): English has no exact equivalent. Say the voiced sound, then release it with a breathy h, keeping your voice on, roughly as in “dog-house” or “abhor”. Many English speakers find these the hardest sounds to master.

Retroflex vs. Dental

For dental t, d, n, put the tip of your tongue against the back of your upper front teeth. That’s further forward than English t and d, which are made on the ridge behind the teeth.

For retroflex ṭ, ḍ, ṇ, curl the tip of your tongue back and touch the roof of your mouth behind the ridge. To English ears, retroflex sounds are often closer to English t and d than the dentals are.

Try pairs such as ta / ṭa and na / ṇa until you can hear and feel the difference.

Semivowels, Sibilants and h

  • y: as in “yes”.
  • r: a lightly tapped r, made with a single quick flick of the tongue. Not the English approximant r.
  • l: a clear dental l.
  • v: somewhere between English v and w. Many traditions use a soft v with light lip contact, and after a consonant (as in svāmī) it often sounds closer to w.
  • ś (श): palatal, like sh in “shine”.
  • ṣ (ष): retroflex: make sh with your tongue tip curled back. In much modern usage it sounds very close to ś.
  • s (स): dental, as in “sun”.
  • h (ह): a voiced, breathy h.

Anusvāra and Visarga

Anusvāra (ṃ, ं) is a nasal sound after a vowel. Before a consonant, it’s usually pronounced as the nasal of that consonant’s group: saṃkalpa sounds like “saṅkalpa”, and saṃdhi like “sandhi”. Before y, r, l, v, the sibilants or h, and at the end of a word, it’s pronounced in different ways in different traditions: as a nasalized vowel, an m, or something in between.

Visarga (ḥ, ः) is a breathy h release after a vowel. At the end of a phrase or line, many traditions add a faint echo of the preceding vowel: rāmaḥ sounds like “rāmaha”, hariḥ like “harihi”, guruḥ like “guruhu”. In the middle of a sentence, sandhi rules often change visarga into another sound entirely.

Doubled Consonants and Conjuncts

A doubled consonant is held longer, not said twice with a gap: datta is “dat-ta”, with the t held across the syllable boundary. Consonant clusters (conjuncts) are pronounced as written, with no vowel slipped in between, so agni isn’t “agani”. Our guide to conjunct consonants shows how clusters are written.

Two clusters have well-known regional variations:

  • jñ (ज्ञ), as in jñāna: commonly said as “gy” in North India, “dny” in Marathi-speaking areas, and closer to “j + ny” in some careful recitation.
  • kṣ (क्ष), as in kṣetra: k + retroflex ṣ.

Syllable Weight and Rhythm

Classical verse is built on syllable weight. A syllable is light if it has a short vowel followed by at most one consonant. It’s heavy if it has a long vowel, a diphthong, anusvāra or visarga, or a short vowel followed by two or more consonants. Getting vowel length and consonant clusters right automatically gives chanted verse its proper rhythm. Vedic recitation also uses pitch accents, which are a specialized subject usually learned from a teacher.

Common Mistakes for English and Hindi Speakers

English speakers often:

  • say short a like the a in “cat”
  • turn ph into f and th into the English th
  • miss the difference between dentals and retroflexes
  • use the English r
  • add an extra puff of air to p, t, k

Hindi speakers often:

  • drop final and middle a sounds (schwa deletion), saying “Rām” for Rāma
  • say ṣ and ś as the same sound
  • say ṛ as “ri” in every case, even where a teacher’s tradition differs
  • carry over Hindi nasal habits for anusvāra

How to Practise

  1. Work through the alphabet in order and say each row aloud, feeling where your tongue moves.
  2. Practise minimal pairs, such as kala / kāla or pata / paṭa.
  3. Record yourself reading a short verse and compare it against a recording from a good teacher.
  4. Learn a few well-known verses by heart. Chanting is one of the most effective ways to make pronunciation automatic.
  5. Get feedback from a teacher occasionally, even if you mostly study on your own.

Pronunciation is part of the first stage in our roadmap for learning Sanskrit. It’s much easier to get right at the beginning than to fix later.

The complete guide to learning sanskrit

How to Learn Sanskrit: A Practical Roadmap for Beginners Anywhere
Read the guide

Keep exploring

Guides from other topics

All topics →