What are the Vedas?

A man sits cross-legged before dawn, and from his mouth comes a sequence of sounds older than the Buddha, older than any temple in India still standing. He has never read these words. He learned them from a man who learned them from a man, back along a chain that runs out of sight. The chain has not broken.

This is the strange thing about the Vedas. For most of their existence they were not books at all.

We tend to imagine ancient scripture as something inscribed: clay, papyrus, palm leaf, stone. The Vedas were something else first. They were a sound held in human memory, passed mouth to ear across a span of time that dwarfs most written traditions on earth, and guarded by a discipline so exact that it has few parallels anywhere. To ask what the Vedas are is to ask, before anything else, how a civilisation kept a text perfect without writing it down.

The word that means “knowledge”

The name itself is a clue. Veda (pronounced VAY-da) comes from the Sanskrit root vid, “to know”. It is cousin to the English wit and wisdom through their shared ancestry in the family of languages scholars call Indo-European. So at the most basic level a Veda is a knowing. Not a story, not a scripture in the sense of a single authored book, but a body of knowledge.

There are four of them.

The Rigveda, the Samaveda, the Yajurveda and the Atharvaveda. The first three form the older core, and an old phrase calls them the trayi vidya, the “triple knowledge”. The Atharvaveda joined the canon later, which is why some early sources count three Vedas and others four. The tradition settled on four. We will come to what each one is for.

What unites them is a claim about their nature that is easy to read past and very hard to overstate.

”That which is heard”

In the Hindu reckoning of sacred texts, the Vedas occupy a category of their own. They are shruti (SHROO-ti), a word that means, plainly, “that which is heard”. Everything else, however revered, falls into a second category called smriti (SMRI-ti), “that which is remembered”. The great epics, the Mahabharata and the Ramayana, the Puranas, the law book attributed to Manu: all of these are smriti. The law book itself draws the line, naming shruti and smriti as the twin roots of dharma and placing shruti first (Manusmriti 2.6-2.10, trans. Bühler, 1886).

The distinction is not about age alone. It is about authority, and about origin.

Smriti is human work. It is remembered, handed down, attributed to authors who lived and died. Shruti is held to be different in kind. The tradition teaches that the Vedas were not composed by anyone. They were heard. The sages who stand behind the hymns, the rishis (RISH-ees), are described as seers, not as poets inventing verse: they perceived a sound that was already there, eternal, and gave it human voice. A rishi is one who sees the mantra. He does not make it.

This doctrine has a technical name. Apaurusheya (a-pow-ru-SHAY-ya): “not of human origin”, authorless, without a personal maker. Later philosophers of the Mimamsa school built an entire theory of language and authority on it, arguing that the Vedic word is beginningless and that its sounds and their meanings are bound together eternally. One Upanishad puts the intuition in an image, calling the Vedas the very breath of the Imperishable, breathed out like exhalation (Brihadaranyaka Upanishad 2.4.10, trans. Müller, 1884).

A historian cannot test a claim like that, and should not pretend to. What a historian can say is narrower and still remarkable. Whoever first spoke these hymns, and however they came by them, the people who received them treated them as something they had no right to alter. That conviction is the engine of everything that follows.

Four knowings, four tasks

Set the metaphysics aside for a moment and open the four Vedas. They do different work.

The Rigveda is the oldest and the foundation. It is a collection of suktas, hymns of praise, gathered into ten books called mandalas, “circles”. The tradition and most editions count 1,028 hymns. They are addressed to the gods of the Vedic world: Agni the fire, Indra the storm-king and dragon-slayer, Soma the pressed and deified plant, Ushas the dawn, Varuna the keeper of cosmic order. The Rigveda is poetry of invocation, and at its edges it becomes something stranger, which we will reach.

The Samaveda is the Rigveda set to song. Most of its verses are drawn straight from the Rigveda, rearranged and fitted to melody for chanting at the Soma sacrifice. If the Rigveda is the word, the Samaveda is the music. Indian music traces its own descent from these chants.

The Yajurveda is the priest’s working manual. It supplies the yajus, the prose formulas a priest murmurs while performing the ritual acts, the pouring, the placing, the offering. It comes in two main recensions, the “Black” and the “White”, which differ in how they mix formula and explanation. This is the Veda of the hands as much as the voice.

The Atharvaveda is the outlier, and the most human. Alongside high hymns it preserves spells and charms for ordinary life: to heal a fever, to win love, to protect a house, to curse an enemy, to ease a difficult birth. For a long time it sat uneasily beside the other three. Its eventual acceptance widened the canon downward, toward the village and the sickbed.

Four Vedas. Word, song, act, and the whole tangled remainder of lived want.

A text in layers

There is a second structure cutting across the four, and it confuses many newcomers, so it is worth slowing down for.

Each Veda is not a single document. It grew in layers, deposited over centuries like sediment, and the tradition names four strata.

At the base sits the Samhita (SUNG-hi-ta), the “collection”: the core hymns or formulas themselves, the oldest material. Around it accreted the Brahmana (BRAAH-ma-na), prose manuals that explain the rituals, justify them, and tell the myths that anchor them. Then comes the Aranyaka (AA-ran-ya-ka), the “forest book”, so called, the tradition says, because its teachings were meant for the forest hermit and were too potent or too inward for the village. And at the very end, growing out of the Aranyakas, come the Upanishads.

The Upanishads are where the Vedic corpus turns and looks at itself.

They stop asking which offering pleases which god and begin asking what stands behind the gods and the offerings altogether. What is the self. What is the ground of being. What is the relation between the two. Because they sit at the close of the Vedas, they earned the name Vedanta, “the end of the Veda”, and they became the wellspring of the later schools of Hindu philosophy. A reader who knows only the Upanishads has met the summit of the Vedas without the long climb beneath it.

So the structure runs in two directions at once. Outward, four Vedas. Downward through each, four layers, from ritual hymn to metaphysical inquiry. Knowledge of action at the start, knowledge of meaning at the end.

The machine made of memory

Now to the heart of it. How was any of this kept, for so long, without writing?

The Vedic answer was to refuse the manuscript altogether and build something better: a method of recitation engineered to make error audible. Beyond the plain, flowing recitation (samhita-patha), reciters learned the text word by word (pada-patha), then in overlapping pairs (krama-patha), and then in a series of increasingly elaborate “woven” forms — the most complex, ghana-patha, braiding each run of words forwards and backwards in a fixed knot so intricate that a single wrong syllable shatters it where a trained ear would catch it. A master of the ghana recitation earns the title ghanapathi over years of training. The point was never beauty; it was preservation — turning word order, accent and pitch into a structure that could not be quietly broken. The fidelity this achieved is extraordinary, recitations preserved by communities separated for centuries agreeing almost to the syllable, and UNESCO recognised the living tradition of Vedic chanting in 2003.

It is a wonder in its own right — an error-catching code running on human memory, backed by an entire science of phonetics and metre. We tell that full story, and the theology of why a civilization would refuse to write its own scripture down, in How the Vedas Were Preserved.

How old, and whose

Here the ground turns soft, and an honest article has to say so.

We do not have a date for the Vedas in the way we have a date for, say, a dynasty with inscriptions. The Vedas carry no colophon, no year, no signature. What scholars work with instead is language, geography, internal references, and comparison with related ancient cultures, and from these they reason backward.

The broad scholarly position is this. The oldest hymns of the Rigveda were most likely composed in the second millennium BCE, with a common estimate of roughly 1500 to 1200 BCE for the earliest core, and the rest of the four-Veda corpus built up over the centuries that followed. That is a range, not a fact, and serious scholars argue its edges. The composition was spread across many generations. The writing down came much later still.

The tradition’s own account is different and should be stated as what it is. For the tradition, the Vedas are beginningless and timeless, not datable at all, because they were never made.

And the harder question, who the Vedic people were, is genuinely contested. The hymns describe a society of herders and chariots and fire-altars, speaking an early form of Sanskrit, in the northwest of the subcontinent. Beyond that, the major positions diverge sharply. One body of scholarship connects the Vedic language and culture to a broader Indo-European spread and to movements of peoples into the subcontinent in the second millennium BCE. Another view, advanced by other scholars, argues for far deeper indigenous roots and disputes the migration framing. The evidence is drawn from linguistics, archaeology and, more recently, genetics, and it is read in conflicting ways. This is a live disagreement, freighted with history and politics, and it would be dishonest to hand you a verdict. What can be said plainly is that the hymns themselves are very old, that they are the earliest substantial record of this language and this religious world, and that the rest remains under argument.

We do not know everything we would like to know. That is the truthful state of the record.

When the hymn turns to ask

It would be a mistake to leave the Vedas as a museum of ritual. At their edges they do something that still arrests a reader, and two short passages show it.

Late in the Rigveda comes the hymn called the Nasadiya Sukta, the “Hymn of Creation” (RV 10.129). It looks back to the moment before anything was, and it does not claim to know. Griffith’s old translation renders its opening:

“Then was not non-existent nor existent: there was no realm of air, no sky beyond it.” (RV 10.129.1, trans. Griffith, 1889)

The hymn goes on to wonder where creation came from, and whether even the highest god in the highest heaven knows, and it leaves the question open. A scripture, three thousand years old or more, ending an account of origins in a confessed uncertainty. That alone tells you these texts are stranger and braver than the word “ancient hymn” suggests.

And in another hymn sits a single line that the later tradition would carry like a torch:

“They call him Indra, Mitra, Varuna, Agni… To what is One, sages give many a title.” (RV 1.164.46, trans. Griffith, 1889)

Ekam sad vipra bahudha vadanti. “Truth is one; the wise call it by many names.” Out of a collection of hymns to many gods comes the intuition that the many might be one. It is not the developed monism of the Upanishads. It is the seed of it, already present in the oldest layer.

The Vedas are the floor of Sanatana Dharma. Almost everything that came after, the philosophy, the ritual, the law, the long argument about the self and the absolute, stands on them and refers back to them, even when it moves far beyond them. To meet them is to stand at the source and hear the river before it has named itself.

The man before dawn finishes his recitation. The sounds go quiet, and the chain holds for one more morning.