Vedā: Svarās – Udātta, Anudātta, Svarita

Before moving to the answer, we must understand that people who lived through the Vedic Sanskrit have already redacted to us the split of verses into words through the śākhās in the form of padapāṭha. We do not need to think of mending those, as we are not in any advantageous position in understanding the Vedic Sanskrit without their effort. So far, the Vedic accentuations have been verified by parallel accents in Greek or other cognates on the same syllables of same meaning, thereby Vedic Sanskrit the best preserving of Indo European tonal accents and the most reliable. Vedic Sanskrit has been instrumental in reconstructing the accents of PIE words.

What we need to know is how accents are affected when two or more words join together, which we do in the form of vikṛti pāṭha exercises like ghanapāṭha or jaṭāpāṭha. The accentuation is a very vast subject, but I will just let you know the most important rules that will enable you to satisfactorily create each of the pāṭhas from the basic padapāṭha form which is traditionally inherited.

About accents in Vedic language

The basic and unchangeable accent is called udātta, which is the principal linguistic accent marker that is unique to each word carrying a particular meaning. Vedic Sanskrit has a pitch-based accent for each word. It is the property of a particular word. (Now just imagine how we have preserved the accent of each and every word with our oral tradition and you will realize the hard work and perfection of the tradition) For example, it is agní (udātta accent on the syllable ni) while it is hótā, with accent on the syllable ho. Henceforth, we will use the accent marker like this to denote udātta. You can understand the other accents based on udātta’s position.

How an accent is realized and chanted depends on individual traditions, it matters less – what matters is the preservation of accent and syllable. If you follow Taittirīya śākhā in Tamil Nadu/Andhra or Śākala śākhā in Karnataka/Tamil Nadu/Andhra of realization of this accent, the udātta syllable is pronounced with the middle and normal pitch. Henceforth, I am speaking from this perspective when I speak of higher pitch or lower pitch, as these śākhās use a relatively simpler and easier system of realization and chanting of accents, and I am knowledgeable of these traditions myself.

As a thumb rule, every udātta syllable has to be preceded by the svara anudātta. An anudātta is marked in the above mentioned system by a low pitch. So, agní becomes practically marked as अ॒ग्नि, the “underline” we mark in this convention for anudātta and a “lower pitch” than what udātta, the normal pitch is pronounced. Since we pronounce the udātta in normal pitch, we don’t usually mark it in devanāgarī or other Indic texts (though udātta is what we should keep in mind while we speak of svara sandhi). This also means that every anudātta is usually followed by udātta. When you see a text or hear the verse “अ॒ग्निमी॑ळे पु॒रोहि॑तं “, you should be suddenly able to locate the udāttas as अ॒ग्निमी॑ळे पु॒रोहि॑तं. (As anudātta or the lower pitch is the marker of the succeeding udātta syllable)

I hope you are clear till this.

Now, there is the third accent, which is called “svarita”, which we realize as “high pitch” or is marked in this kind of writing with a vertical line. (these conventions in writing can change, the accented syllable is what matters) As you might rightly spot, a svarita should have an udātta before it. However, the application of this law is secondary to the first rule. Thus, for example, say the words ágne, yám and yajñá combine together to get ágne yám yajñá. How will you translate this to our three accented chanting? First follow the golden rule of anudātta. Mark every anudātta before the udāttas. And only then mark svaritas after the udātta. So, you will get :- अग्ने॒ यं य॒ज्ञ. We see that there is no need for svarita in this case, as there is no syllable after jña. And there can be no svarita when there is precedence for anudātta, we don’t put svarita at ya of yajña, but put anudātta because jña is udātta. For the sake of interest, you might note that in nominative case agniḥ, udātta is in ni, while in vocative case agne, udātta is in the syllable a(g). You don’t have to bother about this, since this is already redacted to us in padapāṭha.

Now, the problem occurs. After we intonate these accents, we find that there are unmarked syllables which are actually not udātta but which are pronounced in normal pitch in many śākhās, since they are perfectly unaccented syllables. These are said to be of “pracaya” accent.

How do we distinguish between pracaya and udātta in saṃhitāpāṭha? It is easy – the normal pitch (or unmarked syllable) which is preceded by anudātta or succeeded by svarita is udātta, while that which is not so (mostly preceded by a svarita instead of anudātta) will be pracaya.

In padapāṭha, there is no pracaya shown if the word is totally unaccented, also we intonate syllables with anudātta till we see an udātta syllable.

To sum it up, let us consider a real example from Rigveda 1.1.1. Following is the padapāṭha of the words.

agním ǀ īḷe ǀ puráḥ-hitam ǀ yajñásya ǀ devám ǀ ṛtvíjam ǀ

hótāram ǀ ratna-dhā́tamam ǁ

In our three (four) accent system, we will realize this in chanting as :-

अ॒ग्निम् । ई॒ळे॒ । पु॒रः – हि॑तम् । य॒ज्ञस्य॑ । दे॒वम् । ऋ॒त्विज॑म् ।

होता॑रम् । र॒त्न॒ – धात॑मम् ॥

I hope this much is clear.

Now, convert this to saṃhitāpāṭha (alert – you need to be thorough with Sanskrit sandhi of syllables while you combine words). If you find it confusing to begin with three accent system, try it with just udāttas first :-

agnímīḷe puróhitaṃ yajñásya devámṛtvíjam |

hótāram ratnadhā́tamam ||

Now, apply the anudāttas before udāttas. I am going to represent the anudātta syllables in italic. (Note that | indicates a stop, and therefore first half is treated separate from second)

agnimīḷe purohitaṃ yajñasya devamṛtvijam |
hotāram ratnadhātamam.

Now, apply svaritas. We are going to use bold italic to indicate svaritas :-

agnimīḷe purohitaṃ yajñasya devamṛtvijam.
hotāram ratnadhātamam.

Or in Devanāgarī,

अ॒ग्निमी॑ळे पु॒रोहि॑तं य॒ज्ञस्य॑ दे॒वमृ॒त्विजं॑ ।

होता॑रं रत्न॒धात॑मं ॥

Note that while I just illustrated this through scripts for the ease of making the idea clear, in practice this has to spontaneously come from your head while you recite verses. This is achievable by mindful practice. The udātta’s position in a word has to stay in your head and with that you can create any kind of pāṭha if you know Sanskrit sandhis.

A thing to note is that there are a few special accents which are basically modifications of the basic accents. One such is the dīrgha svarita, when long vowel or end nasal stop (ṃ) carries the svarita. It is dependent on individual śākhās on how and what they accentuate as dīrgha svarita. Such accents are helpful to memorize chants and preserve words correctly as they are based on certain sandhis. There is also the accent on pluta (the long vowel) in Rigveda that is particularly unique, and can be gleaned from the tradition.

Also, there are a few noted cases in some words containing clustered syllables which are actually linguistically realized as of more syllables, and thus behave accentually like that. However, I don’t want to confuse you now with them, they rarely enter and you can already understand them from saṃhitāpāṭha and padapāṭha.

I hope I am clear.

pronunciation of तन्वे?

In Rigvedic Sanskrit, tanve is actually realized as three-syllabled word tanuve, usually with ta as anudātta and ve as udātta. The “nu” syllable acts as an almost silent syllable.

The thing you asked is of an advanced level. All svaritas are not dependent on the succeeding or preceding syllable. There are certain svaritas as I indicated in my answer, which are permanent and follow certain exceptional rules. The rules of these are mentioned and regulated in Prātiśākhyas which teach the phonetic aspect of Vedic Sanskrit of a śākhā. Usually, around six kinds of svarita (abhinihita, jātya, kṣaipra etc.) are recognized. The tanve in the verse you quoted, uttānayoś camvoḥ in 1.164, tanvāḥ śumbhamāne in 2.39 etc. are such instances in Rigveda where this particular kind of svarita manifests, which is due to ending “u” of the basic words tanu and camu, ending in svarita which supposedly holds its svarita without any threat from the power of udātta that follows it; this creates a scenario as you hear in the oral tradition of a svarita merged with an anudātta so as to account for the succeeding udātta. This is called kṣaipra svarita.

Is there any way/rule to know which syllable has Udatta or is it random? And what is with all the hand movements when chanting?

Is there any way/rule to know which syllable has Udatta or is it random?

It is a property of the word, and not assigned by us. Vedic Sanskrit is a pitch accented language – when we say that, we mean the presence of accent udātta is the property of a particular word. It is a fundamental accent and not assigned by us. The other accents are based on where udātta comes.

Different recensions have different ways to realize and express the accents through chants. I am well knowledgeable only of certain traditions. There are many other traditions in India, and all of them operate on their own principles of understanding the accents. Though the expression of accents might differ from recension to recension, the accents themselves never change, and that is the greatest proof of the infallibility of oral tradition.

This is a recension-specific thing, detailed in Prātiśākhyas and also the books related to śikṣā.

Pronouncing y like j or ṣ like kh is primarily something that has survived from late Vedic times, as we cannot doubt the recensions have been changed since the time of compilation. So, the recensions reflect the pronunciations of the verses at the time of compilation. Considering that at the time of final compilation and redaction, many of the “Prakritisms” were common and that Vedic Sanskrit had many dialects over a huge area, the recensions reflect the dialect of Vedic Sanskrit spoken by the people of that branch.

The dialectal quality also affects secondary accents like the occurrence of dīrghasvarita and sandhi of anusvāra for example. The pronunciation of anusvāra is a thing unique to many recensions. Almost all Yajurvedic recensions pronounce the anusvāra before an “s”, “ś”, “ṣ”, “h”, “r” or before vowels as a separate syllable “(g)m”, many recensions pronounce the anusvāra before “y”,”v” as a half-nasal.

In Taittirīya Yajurveda recension prevalent in South India, the following rules are followed in pronunciation of anusvāras:-

> The anusvāra is realized as “ṃ” and pronounced as the nasal cognate of the succeeding letter or “m” – This is for cases like:-

amṛtaṃ martyam (pronounced as amṛtam martyam)
agniṃ dūtam (pronounced as agnindūtam or as agnim-dūtam in some recensions)

However, this is not always the case for pronunciation of anusvāra.

> The anusvāra is pronounced as a half-nasal in its conjunction with succeeding syllables like y, v, l. This has been considered to be optional by grammarians, and therefore all recensions don’t necessarily subscribe to this.

Cases would be like ima(ṁ)llokam for imaṃ lokam, śa(ṅ)yyoḥ for saṃyoḥ etc.

> Anusvāra becomes pronounced as (g)ṃ and treated as a separate syllable in accentuation, when followed by r, vowels or fricatives s, ś, ṣ, h.

Examples – a(g)ṃśuś ca me instead of aṃśuś ca me, a(g)ṃhasaḥ instead of aṃhasaḥ, martyā(g)ṃ ā viveśa (martyāṁ ā viveśa), devī(g)ṃ śaraṇam, pratyuṣṭa(g)ṃ rakṣaḥ etc.

> The nasal part of anusvāra gets completely lost and is replaced with a syllable (g) when followed by a cluster headed by any of the earlier mentioned letters that produce gṃ and preceded by a long vowel.

For example, chandāṃsyāpaḥ would be chandā(g)syāpaḥ, jyotīṃṣyāpaḥ would be jyotī(g)ṣyāpaḥ

> The nasal part of anusvāra gets completely lost but gets replaced by syllable cluster (gg) when followed by a cluster that produce gṃ but preceded by a short vowel.

For example, kṛṇvaṃstvā becomes kṛṇva(gg)stvā.

Author/Researcher/Translator: Kiron Krishnan
June 28th, 2020