When, and Whose

The book in front of every other book has no first page.

You can open the Rigveda. You can read its first hymn, the one addressed to Agni, the fire. What you cannot do is the thing you can do with almost any other ancient text. You cannot turn to the back and find a date. There is no colophon. No scribe signed the bottom of a leaf. No king’s regnal year is fixed in the margin. The hymns name many things, rivers and cattle and enemies and gods, but they do not name the century that made them. The most important question a forensic reader can ask, when was this composed, is the one the text refuses to answer in its own voice.

So scholars do what investigators do when there is no confession. They reason from the evidence that the text cannot help leaving behind.

The four kinds of clue

There are roughly four. The first is language. Vedic Sanskrit is an early form of a language whose later forms we can date with more confidence, and languages change at a pace that, while not a clock, is at least a current. The grammar of the oldest hymns is archaic in ways that the grammar of the later Vedic texts is not, and the later texts shade into classical Sanskrit, which we can anchor against datable inscriptions and the grammarian Panini, usually placed somewhere around the fourth century BCE (Jamison and Brereton 2014, Introduction). Working backward from those anchors, through the layers of linguistic change, gives a relative sequence and a rough span.

The second clue is geography. The hymns describe a world, and that world has rivers in it. The poets sing of the Sapta Sindhu, the seven rivers, the land of the northwest of the subcontinent, the Punjab and its neighbours. They name the Sindhu, the Indus, and a river called the Sarasvati, praised as the greatest of rivers, mighty and flowing to the sea (Rigveda 7.95, in Jamison and Brereton 2014). Where the hymns look, and where they do not yet look, is itself a kind of dating. The oldest books know the northwest intimately and the Gangetic east barely at all.

The third clue is what the world contained. The hymns are full of horses and of the spoked-wheel chariot, the ratha, a fast vehicle built of bent wood. The spoked wheel and the light chariot are technologies that the wider archaeological record places in the second millennium BCE (Anthony 2007). A text saturated with chariots is unlikely to be older than the chariot.

The fourth clue is comparison. The Rigveda is not alone in the world. It has a near relative.

The sister text

Across the mountains to the west, in ancient Iran, another body of sacred poetry survives. It is called the Avesta, the scripture associated with Zarathustra, and its oldest layer is composed in a language, Avestan, so close to Vedic Sanskrit that scholars can sometimes convert a line of one into the other by regular sound rules (Jamison and Brereton 2014, Introduction). The two share gods, or rather share the names of powers that became gods in one tradition and something else in the other. They share metres, phrases, the very shape of the sacrifice. They are clearly two branches grown from one trunk.

That shared trunk is the deeper finding. The relationship does not stop at Iran. Sanskrit, Avestan, Greek, Latin, the Celtic and Germanic and Slavic tongues, and others, form a single vast family, the Indo-European languages, whose kinship was first argued in detail in the late eighteenth century and has been the bedrock of comparative linguistics ever since. The word for father runs through it, pitar in Sanskrit, pater in Latin, recognisable across thousands of miles. From the family, scholars reconstruct a hypothetical ancestor, Proto-Indo-European, never written down, inferred entirely from its descendants.

Put the four kinds of clue together and a range emerges. The oldest core of the Rigveda, the so-called family books, books two through seven, is placed by most academic scholars in the second half of the second millennium BCE, given by Witzel as roughly 1500 to 1200 BCE and by Jamison and Brereton as about 1400 to 1000 BCE, with the later books and the other three Vedas trailing after over the following centuries (Witzel 1995; Jamison and Brereton 2014). This is a range, not a date. It is the verdict of inference, and it is held with the appropriate looseness of inference.

And whichever century one chooses, the writing came far later. The hymns were composed for the ear and kept by the voice, transmitted through the memory disciplines described earlier in this volume, for many centuries before anyone set them in ink. The text we date is not the text that was first written. It is the text that was first heard.

The people in the hymns

So much for when. The harder forensic question is who.

The hymns let us reconstruct a society, at least in outline. These were people who counted wealth in cattle and prayed for cows the way a later age would pray for gold. They kept horses and prized them. They drove chariots and raced them and fought from them. They brewed and pressed a sacred plant, soma, into a drink offered to the gods and drunk by the priests. They built no cities that the hymns dwell on; their world is one of herds and rivers and the fire-altar, of clans and their contests. They called themselves arya, a word that in the hymns carries the sense of noble, of one’s own people, of the keepers of the proper rites, and that should not be read through the racial distortions later imposed on it in modern Europe (Bryant 2001).

That much the text supports. The next step, asking where these people came from, is where the ground turns contested, and where this treatise must be most careful.

The contested question

There are two broad positions. This treatise surveys both and adjudicates neither, because the evidence is genuinely disputed and the matter has become politically freighted in ways that reward caution and punish certainty.

The first position is the Indo-Aryan migration framework. Its argument runs roughly as follows. The Indo-European language family is real and requires explanation; languages that close in kinship descend from a common source spoken by a community that then dispersed. On the dominant reconstruction, the ancestral language spread outward from a homeland, often placed on the Eurasian steppe north of the Black and Caspian seas, carried by populations who moved over centuries (Anthony 2007). On this account, speakers of an early Indo-Iranian language entered the northwest of the subcontinent in the second millennium BCE, and the Rigveda is the poetry of that arriving stream and its descendants. The closeness of Vedic and Avestan is read as the recent split of two branches that had only lately parted. In the last decade, studies of ancient DNA have been drawn into this framework, with one large analysis arguing for an influx of steppe-associated ancestry into South Asia in the relevant period (Narasimhan et al. 2019). Proponents present the convergence of linguistics, archaeology, and genetics as mutually reinforcing.

The second position disputes the migration framing. It is sometimes called the indigenous-origins or continuity position, and in its strong form the Out of India thesis, which reverses the arrow and proposes that Indo-European or Indo-Aryan speech spread outward from the subcontinent rather than into it. Its proponents argue several things. They question whether linguistic kinship can establish the direction of movement at all, noting that languages can spread without large movements of people. They point to the archaeological record of the subcontinent, the long and sophisticated Indus or Harappan civilisation, and ask why a major migration left so faint a trace in material culture and in the texts’ own memory, which records no homeland left behind (Bryant 2001; Bryant and Patton 2005). And they make much of the Sarasvati.

The river in the argument

The Sarasvati matters to both sides because the hymns are emphatic about it. The Rigveda praises the Sarasvati as a mighty river flowing from the mountains to the sea (Rigveda 7.95). Yet the great river of that description is not flowing today. There is a seasonal Sarasvati, and there is a long dry palaeochannel system in the northwest, often identified with the Ghaggar-Hakra, which satellite imagery and field survey show once carried far more water than it now does (Danino 2010).

Here the two readings part. The continuity argument runs that a Rigveda which sings of a full and mighty Sarasvati must have been composed while the river still flowed, and that the river’s decline can be dated by geology and archaeology to a period early enough to push the hymns’ world deep into, or behind, the era of the Indus civilisation, implying a population long settled rather than lately arrived (Danino 2010). The migrationist reply runs that praise of a great river need not be literal reportage of the poet’s own day, that Sarasvati may name different rivers in different passages or carry mythic as well as geographic weight, and that the chronology of the river’s drying is itself debated (discussion in Bryant and Patton 2005). The same channel of sand is read by each side as evidence for its own case.

It would be dishonest to pretend the dispute is settled, and dishonest in the other direction to pretend it is merely symmetrical. Within the international academy the migration framework is the majority position, and the ancient-DNA work of the last decade has been taken by many as strengthening it. It is also true that specialists continue to contest the interpretation of that genetic evidence, the dating of the linguistic splits, and the weight the archaeology can bear, and that the indigenous position includes serious scholarship alongside its more polemical forms (Bryant 2001; Bryant and Patton 2005). The honest report is that this is a live question, argued by competent people on more than one side, entangled with modern identity and politics in both India and the West, and not closed. A treatise can lay out the table of evidence. It cannot, in present knowledge, hand down the verdict, and it should distrust anyone who claims to.

What the tradition says about its own age

There is a third account, and it does not sit on the same shelf as the other two, because it answers a different question.

The tradition does not ask when the Veda was composed, because it does not hold that the Veda was composed at all. In the orthodox understanding, refined above all by the Mimamsa school, the Veda is apaurusheya, of no human author, a-paurusheya, not the work of any purusha, any person. It was not written by sages. The sages, the rishis, are called its seers, not its makers; they are those to whom the hymns were revealed, who heard them. This is the meaning folded into the oldest name for the corpus, shruti, that which was heard. And the Veda is held to be without beginning, eternal as sound, so that to ask its date is, within its own terms, a category error, like asking on what afternoon the laws of number were enacted (Mimamsa Sutra 1.1.5, the autpattika sutra, with Shabara’s commentary, in Jha 1933-1936).

This is not a rival carbon date. It is a claim of a different order, a claim about what kind of thing the Veda is. The forensic reader and the devotee are not, on this point, contradicting each other so much as answering two different questions. One asks when the heard words were first uttered by human throats in the world we can excavate. The other asks whether those words began at all. This treatise records both and collapses neither into the other.

The volume’s last turn

We have come, in this volume, the whole length of the heard.

We began with a library that was never written, held for a hundred generations in the trained memory of human beings, more exact for being unwritten than most things are for being inscribed. We followed it into its four collections and its layers, the hymns and the ritual and the forest texts and the parts that turn at last from the altar toward the question behind the altar. We watched the priest tend the fire and the grammarian guard the syllable. And now, at the end, we have stood before the oldest book in the line and found that it will not tell us its age or, with certainty, its makers, and we have learned to sit honestly with that silence rather than fill it.

There is one more thing the heard does, and it is the thing the next volume is about.

At the far edge of the Veda, in the forest, when the fire has been tended and the metre kept and every external rite performed, a few voices stop asking the gods for cattle and rain and begin asking something else. They turn from the offering to the one who offers. They sit down, close, and ask who is listening. That turning has a name, and a literature, and it is where we go next.

The heard becomes the asked. We sit near, and the second volume begins.