The Oldest Book in the World Will Not Stop Telling You It Is New

A Poet Says He Has Made Something That Never Existed Before

A man is standing at a soma pressing somewhere in the northwest of the subcontinent, at some point in the second millennium BCE, and he opens his hymn by telling the god exactly what kind of poem he is about to receive:

apūrvyā purutamāny asmai mahe vīrāya tavase turāya | virapśine vajriṇe śaṃtamāni vacāṃsy āsā sthavirāya takṣam

“Words without precedent, most abundant, most benign, I have carved with my mouth for him: for the great hero, strong and swift, for the overflowing one, the mace-bearer, the firm one.”

Rigveda RV 6.32.1, my rendering from the Saṃhitā text.

Apūrvya means “having nothing before it”. The poet is a Bharadvāja, working in the family book that modern scholarship places among the older strata of the collection, and his boast is that his words have no ancestors. The verb he uses for making them, takṣ, is the carpenter’s word, the one used of chariots and wheels, discussed at length in Fashioned Like a Chariot. What he carved is now roughly three thousand years old, and it survives inside the text people reach for when they want a word for ancient.

That gap between what the poets claim and what the text has become is not a one-off curiosity. It runs through the whole Rig Veda. Counting the romanised Saṃhitā stem by stem, about eighty-four verses advertise the poem in progress as navya or navīyas (“newer”), naviṣṭha (“newest”), or apūrvya (“unprecedented”). Eighteen of them manage to call the poem new and the god ancient inside a single stanza. The collection that the later tradition declared beginningless and authorless is, read at close range, a three-thousand-year-old pile of press releases announcing the latest thing.

The question worth asking is not whether these poets were being honest. It is what new could possibly mean to a craftsman working in an inherited formulaic language, performing for gods who had heard the same epithets from his grandfather, and competing with other families for the same patrons. The answer the Rig Veda gives is specific, and it is not the answer a modern reader expects.

The Vocabulary of Being Newer

Vedic has a whole small arsenal for this. The base adjective is nava, “new”, one of the most securely reconstructed words in the Indo-European lexicon: Latin novus, Greek neos, Hittite newa-, Avestan nava-, English new, all from the same ancestor. [1] Off that base the poets built a comparative, navya- and navīyas-, “newer”, and a superlative, naviṣṭha-, “newest”. Beside them sits a quieter word, nūtana- or nūtna-, formed from the particle nū, “now”, and meaning not “new” so much as “of the present day”, “current”, “ours”. And at the far end sits apūrvya-, the flat negative: nothing of this kind has existed.

Table 1. The novelty lexicon and its opposites. Counts are from a stem search over the romanised Śākala Saṃhitā and should be read as floors, not census figures.

Methods note. Counting nava is a trap. Vedic nava means “new” and also “nine”, with no formal difference, so navati “ninety” and navama “ninth” sit in the same search results as navya “newer”; nāvya “navigable” and navanta “they roared” land there too. I therefore counted only the comparatives, superlatives and apūrvya, which are unambiguous, inspected every hit by hand, and discarded two false positives (the dhṛṣṇav īyānaḥ of RV 1.30.14 and a sandhi accident at RV 10.67.3). Because the transmitted text is a continuous stream with words fused at their boundaries, forms hidden inside compounds will have escaped. For figures that will survive a referee, use Grassmann’s 1873 dictionary, which lists every occurrence of every word, or Lubotsky’s two-volume concordance. [2] [3]

The second useful count is what the new thing is called. Vedic has no single word for “poem”; it has a dozen words for different aspects of the same object, and the novelty adjectives attach to all of them.

Table 2. What is being described as new, in the verses that use a comparative or superlative of nava. Several verses carry more than one of these words, so the column sums to more than the verse count.

Notice what is absent. Not one of these words means “words on a surface”. The thing claimed as new is a thought, a voice, an insight, a formulation: something that happens in time, in a particular throat, on a particular afternoon. That alone reframes the claim. A poet saying navyasī manīṣā is not saying “here is a text nobody has written before”. He is saying “here is a thinking that is happening now”.

New in the Manner of the Old

The decisive evidence is the eighteen verses that refuse to choose. A poet of the Bharadvāja family puts the whole programme into eleven syllables:

pratnā navyasā vacasā vivāse

“With newest hymn I call those Wonder-Workers, ancient and brilliant, and exceeding mighty.”

Rigveda RV 6.62.5, third pāda quoted; English of the full verse from Griffith (1896). The Sanskrit says, more flatly: the ancient ones I court with newest speech.

The same family does it again, and this time the adverb is the thing to watch:

taṃ vo dhiyā navyasyā śaviṣṭhaṃ pratnam pratnavat pariṃtaṃsayadhyai

“Him will we fit for you with new devotion, the strongest Ancient One, in ancient manner.”

Rigveda RV 6.22.7, translation Griffith (1896).

Pratnavat, “in the ancient way”. The insight is new, the god is ancient, the procedure is traditional, and all three are asserted without embarrassment in a single line. The ninth maṇḍala, the book of soma hymns, says it outright, and then says it twice:

nū navyase navīyase sūktāya sādhayā pathaḥ | pratnavad rocayā rucaḥ

“Make the paths ready for a hymn newer and newer evermore: Make the lights shine as erst they shone.”

Rigveda RV 9.9.8, translation Griffith (1896).

Newer and newer, shining as before. The companion verse at RV 9.91.5 repeats the formula almost word for word, which is itself the joke: the demand for endless novelty is transmitted as a fixed phrase.

What is going on here is not hypocrisy and not paradox. It is a precise division of labour, and once you see it the Rigvedic poets stop looking like primitives insisting on their originality and start looking like musicians. The form is given. The occasion is not. A jazz standard is not diminished by being a standard, and no listener thinks the pianist is lying when he says he will play something you have not heard. The Vedic poet’s word for his product, sūkta, means “well spoken”, and the praise he most wants is that it was well made, now, for this pressing, for this patron, by him.

The other pillar of the argument is what gets described as new besides the poem. The gods do. Agni, kindled every morning out of dry wood that has been sitting in the yard for a season, is described as being born new in old material:

kūcij jāyate sanayāsu navyo vane tasthau palito dhūmaketuḥ

“He rises ever fresh in ancient fuel: smoke-bannered, gray, he makes the wood his dwelling.”

Rigveda RV 10.4.5, translation Griffith (1896).

Elsewhere Agni is seated as “the ancient priest, from of old, and the new one” in the same half-line (RV 8.11.10). The dawns, which the poets watched arrive one after another in an unbroken series, are called pratnavan navyasīḥ, “newer in the manner of the old” (RV 1.124.9), an observation examined from a different angle in The Fault Called Sameness. The Rigvedic world is structurally like this: fire, dawn, the pressed drink, the year, all of them identical in kind and fresh in instance. A poem that claimed to be timeless would be claiming to be unlike everything else in the cosmos. The hymn is new for the same reason the morning is.

graph LR
    A["Inherited formulas<br/>and family repertoire"] --> B["Poet composes<br/>for this occasion"]
    B --> C["Performance at<br/>the pressing"]
    C --> D["Patron's gift<br/>(danastuti)"]
    C --> E["Verse enters the<br/>family collection"]
    E --> F["Re-performed<br/>as pratna, ancient"]
    F --> A

Figure: how this afternoon’s navyasī becomes next generation’s pratna. The loop, not the line, is the Rigvedic shape of a tradition.

Who Pays for a New Poem

Novelty in the Rig Veda has a price list behind it. The poets worked for patrons, usually chieftains, and were paid in cattle, horses, gold and cloth, a transaction they recorded in the gift-praise verses examined in The Price of Praise. Read the novelty claims beside that economy and their tone changes. One poet of the eighth maṇḍala puts the equation with startling bluntness:

vayaṃ gha te apūrvya indra brahmāṇi vṛtrahan | purūtamāsaḥ puruhūta vajrivo bhṛtiṃ na pra bharāmasi

“We, very many of us, bring forth formulations to you, O unprecedented Indra, smasher of Vṛtra, much-invoked bearer of the mace, the way a man brings a wage.”

Rigveda RV 8.66.11, my rendering from the Saṃhitā text.

Bhṛti is the word for hired payment. The hymn is being handed over the way a day labourer hands over his work, and the adjective apūrvya in that verse is attached not to the poem but to the god, which is the other half of the picture: the same vocabulary of unprecedentedness runs in both directions, because a god who has never been equalled deserves a poem that has never been made.

Where there is a wage there are rivals. The eighth maṇḍala opens an Indra hymn with an instruction that only makes sense in a crowded market:

mā cid anyad vi śaṃsata sakhāyo mā riṣaṇyata

“Praise nothing else at all, friends. Do not go astray.”

Rigveda RV 8.1.1, my rendering from the Saṃhitā text.

Nothing else, nobody else. The verbal contest was a real institution in the Indo-Iranian world, the subject of a classic study by F. B. J. Kuiper, who argued that a formal competition of words lies behind a great deal of Rigvedic phrasing. [4] Competitors need a claim, and “mine is newer” is the one claim available to a poet who cannot claim to be using better words, since the words are inherited by everybody.

Aside. This is where the oral-formulaic question bites. Milman Parry and Albert Lord showed that Homeric diction is built from ready-made phrases sized to fit the metre, and the same analysis applies to Vedic, where whole half-lines recur across families and centuries; the inherited phraseology shared between Vedic and Avestan has been catalogued since Rüdiger Schmitt’s survey of Indo-European poetic language, and Calvert Watkins traced individual formulas from Anatolia to Ireland. [5] [6] A formulaic poet is not in a position to invent his sentences. What he can do is choose, combine, place, and aim, and the Rigvedic evidence suggests that his audience knew the difference and that his pay depended on it.

The Rig Veda’s Own Sense of Its Past

Here is the part that surprised me most. The text does not merely claim novelty; it periodises itself. A Viśvāmitra poet, praising Indra, sorts the praise he has received into three generations:

ya stomebhir vāvṛdhe pūrvyebhir yo madhyamebhir uta nūtanebhiḥ

“Him magnified by ancient songs and praises, by lauds of later time and days yet recent.”

Rigveda RV 3.32.13, translation Griffith (1896).

Former, middle, present. That is a stratigraphy, offered casually in a liturgical hymn, and it is the same three-layer intuition that modern philology arrived at by counting metrical irregularities and linguistic archaisms, the method described in The Syllable Clock and in the account of the collection’s editing in Someone Sorted the Rigveda. The poets knew they stood at the end of a sequence. The second verse of the entire collection says so:

agniḥ pūrvebhir ṛṣibhir īḍyo nūtanair uta

“Worthy is Agni to be praised by living as by ancient seers.”

Rigveda RV 1.1.2, translation Griffith (1896).

Hymn one, verse two, and the subject is already the relationship between the dead poets and the working ones. A Vasiṣṭha hymn names both groups as makers: “the seers who were before and those of now, the inspired ones, who engendered formulations for Indra” (RV 7.22.9). The lineages that produced the text, and whose names are stamped on its final verses as discussed in Fifty-Seven Hymns, One Last Line, understood themselves as a succession of practitioners rather than as recipients of a finished revelation.

Succession brings anxiety. One hymn asks the sharpest question in the collection about its own transmission:

kva ṛtam pūrvyaṃ gataṃ kas tad bibharti nūtanaḥ

“Where is the ancient law divine? Who is its new diffuser now?”

Rigveda RV 1.105.4, translation Griffith (1896).

Where has the old order gone, and which of us present-day people is carrying it? A tradition confident of its own continuity does not ask that. A tradition staffed by people who know that the carrying is done by particular fallible men, one generation at a time, asks it constantly.

graph TD
    P["pūrvya: former praises"] --> M["madhyama: middle praises"]
    M --> N["nūtana: present praises"]
    P -.-> O1["Old family books 2-7"]
    M -.-> O2["Later family hymns"]
    N -.-> O3["Maṇḍalas 1, 8, 10"]
    O1 --> S["Redaction and fixed order"]
    O2 --> S
    O3 --> S

Figure: the three generations of praise named at RV 3.32.13, set beside the layers modern stratigraphy reconstructs. The dotted lines are an analogy, not an equation: the poet was sorting praise, not dating manuscripts.

Where the Claims Cluster, and Where They Do Not

If the cult of the new song were a late fashion, the novelty claims would pile up in the recent layers of the text. They do not.

Table 3. Verses containing a comparative or superlative of nava, or apūrvya, by maṇḍala, over the romanised Śākala text (10,528 verses as searched), with the two false positives discarded.

The overall rate is

$$r = \frac{84}{10{,}528} \approx 8.0 \text{ per } 1{,}000 \text{ verses}$$

and the old family books (2 to 7) sit at 8.8 per thousand against 7.5 for books 1, 8, 9 and 10. The difference is noise. Insisting on your own freshness was not something the Vedic tradition picked up late; it is distributed across the whole corpus, which means it belonged to the craft rather than to a period.

The variation between families is more interesting and much less certain. The Bharadvājas of maṇḍala 6 make the claim at ten times the rate of the Vāmadevas of maṇḍala 4. With fourteen hits against one, that is a difference you should not build a theory on, but it is the kind of signal that would reward a proper study of family style: some workshops advertised, others did not.

Aside. Why the counts here are approximate, in one more way. Griffith’s 1896 translation, still the most widely quoted English version online, numbers the eighth maṇḍala differently from the Sanskrit text, because he inserts the eleven Vālakhilya hymns into the sequence as 8.49 to 8.59. Everything after that in his book 8 is displaced by eleven. A reader who looks up “RV 8.66” in Griffith and finds a hymn about bows and arrows has not caught me in an error; he has caught the Vālakhilya, whose odd status is the subject of The Fallow Strip. Where a verse in this piece comes from the back half of maṇḍala 8, the translation is mine.

Newest Songs in Other Languages

The impulse is not unique to India, and the parallels are close enough to be worth putting side by side.

In the first book of the Odyssey, Penelope asks the bard Phemios to stop singing about the homecomings from Troy, because the subject hurts her. Telemachos overrules his mother with a general principle about audiences: people praise most the song that comes newest to their ears (Odyssey 1.351 to 1.352). [7] Three centuries later Pindar, who was in the business of composing commissioned victory odes much as the Vedic poets composed commissioned praise, built the same preference into an aphorism: praise old wine, but the flowers of newer songs. [8] And in the Hebrew Psalter the instruction to sing to the Lord a new song recurs as a liturgical formula, in Psalms 33, 96, 98 and 149 and at Isaiah 42:10.

Table 4. The new-song topos in four traditions. The Vedic and Greek material shares an inherited word for “new” and an inherited institution of the commissioned praise poet; the Hebrew parallel shares neither, which is the useful part.

Two cautions, because this is exactly the spot where comparison goes wrong. First, a shared word is not a shared topos: Greek neos and Vedic nava descend from the same Indo-European adjective, but Hebrew ḥādāš does not, and the Psalms want a new song anyway. Second, the function differs. The Psalter’s new song is a response to a fresh act of divine deliverance; Pindar’s is a competitive advertisement; the Vedic poet’s is both at once, since the god’s latest intervention and the poet’s latest commission are the same event. What the parallels establish is narrower and still worth having: wherever praise poetry is performed for a living, somebody starts selling freshness. The Rig Veda is simply the earliest surviving place where we can watch it happen in detail.

How a Library of Novelties Became Unalterable

The last turn of the story is the strangest. Everything in these hymns points to individual authorship, dated occasions and competitive innovation. The poets name themselves. They name their patrons and the size of the fee. They sign their families’ hymns with family refrains. They tell you that their composition is more recent than their rivals’.

Then the collection closes. The hymns are arranged by deity, metre and length into the shape studied in Someone Sorted the Rigveda, and the schools begin transmitting them with the machinery described in The Oral Engine: the word-by-word padapāṭha, the permutation recitations, the preserved pitch accents. The apparatus is designed to make change impossible, and it worked: the transmitted text is so uniform that Vedic philology has almost nothing to collate of the kind that fills the critical apparatus of a Greek or a biblical manuscript tradition, which is a strange thing to be able to say about a document this old. [14] [15]

And then the philosophers arrive and finish the job. Mīmāṃsā, the school whose business was the interpretation of Vedic injunction, held that the Veda is apauruṣeya, “not of human origin”: not revealed to a seer at a moment in history, but beginningless, authorless and eternal. Wilhelm Halbfass traced how that doctrine of the Veda’s authority was built and what it did to the idea of the text. [9] Set the doctrine next to RV 1.62.13, where Nodhas son of Gotama “fashioned a new formulation” for Indra, and the distance is almost comic. A corpus of poets shouting that their work is unprecedented became the standard Indian example of words no one ever made.

It is worth being fair to the Mīmāṃsakas. Their doctrine solves a real problem: if the ritual’s authority rests on a human author, it inherits that author’s fallibility, and an eternal text has no such weakness. They were not misreading the hymns so much as needing them to be a different kind of object. But something was lost in the exchange, and it is the thing a modern reader finds most attractive once it is pointed out. The hymns are not the deposit of a revelation. They are the surviving work of a profession.

What the Claim Was Actually For

Pull the threads together and the Rigvedic position on novelty turns out to be more sophisticated than the modern one it superficially resembles. Our instinct is that new and old are rivals, that a thing is valuable either because it is original or because it is traditional. The Vedic poets do not appear to have felt the tension. Their word for what they do, takṣ, is a craft verb; their gods are pratna, ancient, and the sacrifices that reach them must be navyas, newer. Freshness is the condition on which something gets to be effective, and inheritance is the condition on which it gets to be intelligible. The poem has to be both or it is not a poem: a repeat performance is stale offering, and an unprecedented set of words nobody recognises is not a hymn at all but a noise.

Which is why the two verses I would send a new reader to are not the famous ones. Open RV 9.9.8 and read the half-line about making the paths ready for a hymn newer and ever newer, then notice that the very next words are “make the lights shine as they shone before”. Then open RV 1.105.4 and read the question about where the ancient order has gone and who among us present-day people is carrying it. Between those two lines sits the whole working theory: you keep the thing alive by remaking it, and you are never certain you are doing it well enough.

The Rig Veda is usually sold to modern readers on its antiquity, which is the least interesting thing about it. The poets would have found that emphasis bewildering. They were not trying to be old. They were trying to be this morning’s best, in a language they did not invent, for gods who had heard it all before, in front of a patron who was counting cattle. That is a recognisable predicament, and the fact that somebody was in it three thousand years ago, and said so in a single line of verse that outlived him by a hundred generations, is a better reason to open the book than its date.