Rosetta Stone

Sanskrit and Tamil side by side: the same sounds in two scripts, and where their grammars end up alike even though the languages share no ancestor.

Sanskrit ↔ Tamil / Rosetta Stone

Characters

Sanskrit's Devanagari and Tamil's own script are unrelated in origin (see Tamil Characters). This page doesn't cover either script on its own; it lays the two side by side, sound by sound. Most of it lines up cleanly. Where it doesn't, the gap is usually the interesting part: a sound that one language marks and the other doesn't, or a distinction that one keeps and the other merges. Every correspondence here is already verified elsewhere on this site (Sanskrit in Tamil Script checks it against real chant text). This page puts it all in one table.

Vowels

TamilSoundSanskrit
அ
a
अ
ஆ
ā
आ
இ
i
इ
ஈ
ī
ई
உ
u
उ
ஊ
ū
ऊ
ஐ
ai
ऐ
ஔ
au
औ

Eight of Tamil's twelve vowels map onto Devanagari one letter for one letter, with no adjustment in either direction.

Tamil keeps a distinction Sanskrit doesn't have

எ/ஏ (short/long e) — Sanskrit ஏ has only ए
ஒ/ஓ (short/long o) — Sanskrit ஓ has only ओ

Sanskrit keeps four vowels Tamil doesn't have

ऋ ṛ · ॠ ṝ · ऌ ḷ · ॡ ḹ
syllabic r/l — no Tamil letter is ever a vowel nucleus by itself

Sanskrit's ए and ओ are always long. There is no short e or short o to tell them apart from, so one letter each is enough. Tamil's short e/o are separate sounds from long ē/ō, so it needs both எ and ஏ, and both ஒ and ஓ. On this one point Tamil's vowel inventory is finer than Sanskrit's, which is worth noticing because almost everywhere else on this page it is Sanskrit that distinguishes more.

Where consonants line up exactly

The nasals are the cleanest match on this page. Tamil's mellinam row has five places that match Sanskrit's five nasal consonants one for one, with no changes in either direction:

TamilSoundSanskrit
ங
ṅa
ङ
ஞ
ña
ञ
ண
ṇa
ण
ந
na
न
ம
ma
म

The semivowels and the sibilants (the sibilants via Grantha) match just as cleanly:

TamilSoundSanskrit
ய
ya
य
ர
ra
र
ல
la
ल
வ
va
व
ஶ
Grantha
śa
श
ஷ
Grantha
ṣa
ष
ஸ
Grantha
sa
स
ஹ
Grantha
ha
ह

Where Tamil uses one letter for four Sanskrit sounds

Native Tamil doesn't distinguish voiceless from voiced stops, or unaspirated from aspirated ones (see Tamil Characters). To write Sanskrit precisely, Tamil uses the plain letter for the voiceless sound and adds a superscript number for the other three. Sanskrit in Tamil Script checks this against real chant text:

 Plain+ ²+ ³+ ⁴
Velar
க
ka / क
க²
kha / ख
க³
ga / ग
க⁴
gha / घ
Retroflex
ட
ṭa / ट
ட²
ṭha / ठ
ட³
ḍa / ड
ட⁴
ḍha / ढ
Dental
த
ta / त
த²
tha / थ
த³
da / द
த⁴
dha / ध
Labial
ப
pa / प
ப²
pha / फ
ப³
ba / ब
ப⁴
bha / भ

Palatals are the exception. ja already has its own Grantha letter (ஜ) instead of a modified ச, so the four-way க/க²/க³/க⁴ pattern above doesn't carry over cleanly to ச/छ/ज/झ. They are left out here for that reason, not by oversight.

Sounds only one side has

Beyond the vowels noted above, each script has consonants with no real counterpart in the other:

Tamil only
ழ
ḻa — retroflex approximant, the sound in "Tamiḻ" itself
Tamil only
ள
ḷa — retroflex lateral
Tamil only
ற / ன
ṟa / ṉa — Tamil's 6th ("alveolar") place, alongside velar/palatal/retroflex/dental/labial
Sanskrit only
ळ
ḻa — Vedic-only, the intervocalic allophone of ड (see Sanskrit Characters)

Sanskrit's Vedic-only ळ and Tamil's ழ்/ள் are not the same sound and don't share an origin. It is still curious that both languages ended up with extra retroflex consonants beyond what their core systems predict. Some linguists trace Sanskrit's retroflex series in general to early contact with Dravidian languages, though the details are still debated. Nothing here claims that ळ and ழ்/ள் are historically the same letter.

Grammar & Syntax

Sanskrit (Indo-European) and Tamil (Dravidian) belong to two unrelated language families. Nothing here claims a common ancestor for their grammars, any more than for their scripts. But two languages that have shared a subcontinent for three thousand years pick up real structural similarities, along with a few that look like coincidence and a few real, principled differences. Each language's mechanics are covered in full on Sanskrit Grammar and Tamil Grammar. This section sets the two side by side.

Where the two converge

Same default word order, same reason

Subject – Object – Verb
Both default to subject–object–verb, where English uses subject–verb–object. In both languages it is a default, not a rule. Case endings, not position, show who did what to whom, so word order is free to move for emphasis or meter. That freedom is what lets Vedic chant lines rearrange for the metrical count and stay grammatical.

No verb "to be" in the present

rāmaḥ sundaraḥ  /  kumār māṇavaṉ
"Rama [is] handsome," "Kumar [is] a student". In the present tense both languages put the subject and predicate next to each other and let the sentence stand without a copula. English can't do this. Two unrelated grammars arrived at it independently.

Case does what English prepositions do

8 Sanskrit cases ↔ 9 Tamil cases
Nominative, accusative, instrumental, dative, ablative, genitive, locative and vocative line up almost one to one in the two systems (see Sanskrit Grammar and Tamil Grammar for the full tables). Tamil has one extra, the sociative ("along with"). Sanskrit folds it into the instrumental.

Sandhi happens at every seam, in both

sandhi (संधि) ↔ puṇarcci (புணர்ச்சி)
Both languages smooth the join wherever two sounds meet: at a word boundary in Sanskrit's external sandhi, at a suffix boundary in Tamil's agglutination. They merge vowels, insert a glide or assimilate a consonant so the result is easy to say. The trigger differs but the aim is the same.

One long "word" standing in for a whole clause

nīlotpalam  /  pōkamuṭiyātavarkaḷukkāka
Sanskrit gets there by samāsa, joining independent word stems with no inflection between them (nīla + utpalam → nīlotpalam, "blue lotus"). Tamil gets there by agglutination, gluing one bound suffix after another onto a single root (see the eight-morpheme example on Tamil Grammar, "for the sake of those who cannot go"). The mechanisms are opposite, but the result is the same: one written word doing the work of an English clause.

Relative clauses without a relative pronoun

yaḥ ... saḥ  /  participle + noun
English needs "who" or "which" to embed a relative clause. Sanskrit pairs a correlative pronoun across two clauses (yo dharmaṃ carati, sa sukhī bhavati: "who practices dharma, he becomes happy"). Tamil folds the whole relative clause into a participle placed directly before the noun it modifies, with no pronoun at all. Neither language has a word for "who" here. They fix the same gap in two different ways.

Where they part ways

Two structural differences matter too much to hide behind surface similarity. Each is covered in full on that language's own grammar page and restated here only for contrast:

Fusion vs. agglutination

devāya (one ending)  /  -kk(u) + -āka (stacked)
A single Sanskrit case ending fuses case, number and gender into one piece you can't divide. Tamil's endings stack, and each suffix carries exactly one piece of meaning. See Tamil Grammar.

Gender: grammatical vs. rational/irrational

3 genders  /  5 categories
Sanskrit puts every noun into masculine, feminine or neuter, whatever it refers to. Tamil sorts by personhood first: only humans and deities count as masculine or feminine. It then crosses that with number. So even a named female animal is grammatically "irrational", not feminine.