A dhātu (धातु, "element, support") is a verbal root in Sanskrit. From a comparatively small set of dhātus the entire verbal system of the language is built. Every conjugated verb, every gerund, every participle, every agent noun, every action noun ultimately traces back to a dhātu.
The
The classical inventory of Sanskrit verbal roots is preserved in the Dhātupāṭha (धातुपाठ, "recitation of roots"), traditionally attributed to Pāṇini and recited as an appendix to the Aṣṭādhyāyī. The Dhātupāṭha lists approximately two thousand roots, each accompanied by a brief gloss indicating its meaning. The list is organised into ten conjugational classes (gaṇa).
The Ten Ganas
Each gaṇa is identified by the way it forms the present stem from the root, and is named after a characteristic root:
- Bhvādi (भ्वादि) — class of √bhū "to be"
- Adādi (अदादि) — class of √ad "to eat"
- Juhotyādi (जुहोत्यादि) — class of √hu "to offer"
- Divādi (दिवादि) — class of √div "to play"
- Svādi (स्वादि) — class of √su "to press"
- Tudādi (तुदादि) — class of √tud "to strike"
- Rudhādi (रुधादि) — class of √rudh "to obstruct"
- Tanādi (तनादि) — class of √tan "to stretch"
- Kryādi (क्र्यादि) — class of √krī "to buy"
- Curādi (चुरादि) — class of √cur "to steal"
The first, fourth, sixth, and tenth classes form their present stems by adding a thematic vowel or suffix; the others alter the root in various ways — through reduplication (class 3), nasal infix (class 7), or a special -nu/-no suffix (class 5).
Reading a Entry
A typical Dhātupāṭha entry might read:
bhū sattāyām (भू सत्तायाम्) — "√bhū, in the sense of 'being'"
The first word is the root in its citation form, the second the gloss giving the core meaning. Pāṇini's grammar then specifies how to derive verbs, participles, and nouns from this root.
Roots and Their Family of Words
A single root generates an entire family of related forms. From √jñā ("to know") come:
- jñāna (ज्ञान) — knowledge (noun)
- jñātṛ (ज्ञातृ) — knower (agent noun)
- jñāta (ज्ञात) — known (past participle)
- jñāpaka (ज्ञापक) — informant (causative agent)
- jñātavya (ज्ञातव्य) — to be known (gerundive)
- jñātum (ज्ञातुम्) — to know (infinitive)
- jñātvā (ज्ञात्वा) — having known (gerund)
- jānāti (जानाति) — he knows (present)
- jñāsyati (ज्ञास्यति) — he will know (future)
- jajñau (जज्ञौ) — he knew (perfect)
- prajñā (प्रज्ञा) — wisdom (with prefix pra)
- vijñāna (विज्ञान) — discriminating knowledge (with prefix vi)
The same pattern works for almost any root.
Cognate Roots Across Languages
Many Sanskrit roots have cognates in other Indo-European languages. √vid (to know) is related to Latin videre, English wit. √sthā (to stand) is cognate with Latin stare and English stand. √man (to think) is the source of English mind. The dhātu is thus a deep linguistic unit whose history extends well beyond Sanskrit.
A Generative Lexicon
The dhātu system is what makes Sanskrit's vocabulary so transparently derivable. To know a few thousand roots, the upasargas, and the pratyayas is in effect to know the language's word-formation engine — and through it, to read the literature with comprehension rather than memorisation.