In this post: vrata, mitra, kavi
This word is typically translated as 'vow'. The root is √vṛ of which there are two homonyms meaning either 'to cover' or 'to choose'. The former gives us the word varṇa meaning 'colour' or 'class' (apparently a distinction originally based on outward appearance, ie skin colour). The latter is relevant here. The present form is vṛṇīte or vṛṇāti, and the primary meaning seems to be that one is making a choice, i.e. it refers to the choice that an individual makes. Vrata is a verbal noun meaning 'will, command, law' but because it retains the sense of something an individual does, it comes to mean, in religious contexts, a vow that one takes on. Note it is not something imposed, but chosen.
Part of what makes vrata interesting is that it has an English cognate in the word 'verb'. Verb comes, via Old French, from the Latin 'verbum' which was originally 'a word, verb'. The Online Etymology Dictionary links it to an Indo-European root *were 'to say, to speak' which is directly related also to the Sanskrit vrata. In Greek this root became rhetra 'a covenant or agreement', and gives us the English word 'rhetoric' which is a form of speech designed to persuade. In Gothic *were gave waurd from which, clearly, we get our word for 'word'. Also from the Latin is 'verbose' meaning 'wordy'.
A related Sanskrit word is saṃvara 'restraint, forebearance' (from sam + √vṛ) which in Buddhist contexts can mean something like a vow - a voluntary religious observance which usually consists in not doing something. It is possible that the name śambara is a phonetic variation on saṃvara (perhaps via a Prakrit ṣambara?). Śambara was a demon, who later became a Tantric Dharma-protector; but as a noun the word is also found in the sense of 'a vow'.
Mitra (Pāli mitta) means “friend, companion, asscociate”. PED derives from √med ‘to be fat, to love’. MW derives from √mith ‘alternate’ or √mid which is simply an alternate for √med. Probably both are wrong in this case. Mitra was a Vedic god (paired with Varuṇa) who oversaw the harmonious order of the universe (ṛta), he was concerned with order, and particular moral order and the word mitra was originally associated with a contract, or a formal bond. He appears in Persian myth as Mithra, suggesting an ancient IE lineage. The IE root is *√mei ‘to tie or bind’ (which should be *√mi or *√mī in Sanskrit but is not attested). With an agentive suffix -tṛ; or with an instrumental suffix –tra (similar to mantra, Cf E. ‘meter’ ) mitra means ‘one who makes bonds’ or ‘that which binds’ (i.e. a contract). This seems closer to the ancient function of the god Mitra, to which the sense of ‘friend’ came to be attributed later. The sense of ‘friend’ is restricted to Sanskrit and the word has few English cognates: some words related to threads, and mitre from a band which ties around the head (i.e. a turban).
From mitra we get the word maitrī (P. mettā) which is the feminine form of maitra (a taddhita in -a with vṛddhi of the root vowel) 'of or belonging to Mitra', and in our context 'what comes from a friend' i.e friendship, love, kindness etc. PED suggests that mettā is an abstract noun from mitra, but the Sanskrit morphology argues against this - the form is not mitvā, but maitra/maitrī - though it has an abstract sense.
Kavi means 'wise' or 'a poet', and is related to kāvya 'poetry'. This word is linked by Monier-Williams to the root √kū 'to cry, make a noise' which presumably follows traditional sources. The forms of kavi and kāvya are consistent with √kū: the former being a primary derivative action noun in -i, the latter a taddhitha in -a (with vṛddhi of the root vowel). PED, unusually, gives no etymological information. Whitney tells us that the root √kū means 'design' and casts doubt on kavi deriving from it. There seems to be some confusion in the 19th century Sanskrit reference books, and there are still no signs of replacements!
If we take a step back we find an Indo-European root *√keu. The Online Indo-European Lexicon (OIEL) defines the semantic field as combining 'to see, to think' (I've noted that these two fields overlap in past posts). The OIEL lists kavi as a derivative of *√keu. Some English cognates are (via Germanic) hear, hearken; shine, sheen; (via Greek) acoustic; (via Latin) caution. The Iranian parallel term is kauui (it's worth keeping in mind when making comparisons that Sanskrit 'v' is actually pronounced more like 'w'). Kauui designates a kind of priest in the Avesta, especially priests of myth. These priests, like their brāhmaṇa counterparts in India later became rulers. In Greece the koies or koes were priests of the mysteries of Samothrace. Recall that a wise person in India is one who has 'heard much', the mantras of the Ṛgveda were 'heard' rather than composed by the ṛṣi 'seers'. So it makes sense that the poet is one who has heard, and poetry is what they heard. That fits the requirements of the situation.
It seems as though the IE root *√keu did not come through completely into Sanskrit and most of the forms were lost. So the derivation became unclear. However, because it remained more intact in the European branches of Indo-European we can make more sense of the word kavi in relation to other IE words, than by solely relating it to other Sanskrit words! And this is one of the astonishing and wonderful results of comparative linguistics.
- Philological Odds and Ends I: tathāgata, sūtra, śramaṇa, loka, gahapati/gṛhapati.
- Philological Odds and Ends II: cakravartin, cintāmaṇi, yoniso manasikara, pāramitā, etymology.
- Philological Odds and Ends III: bodhisattva, anagārikā, samyak/mithyā,
(PS Apologies to "gruff" who read and commented on a draft essay that inadvertently went live yesterday. I will save the comment for when the essay is ready.)