Skip to content

The Vedas

Lesson 5 of 30. Four collections, and almost everything people repeat about their shape turns out to be checkable — so this lesson checks it.

11 min read

Depth: Seeker

Almost everything people say about the shape of the Vedas is checkable. How many hymns; how much of one is borrowed from another; whether the famous four-layer diagram survives contact with an actual text; where the numbers about lost schools come from. This lesson checks those things rather than repeating them, and several of them come out differently from the summary you will have met.

What the word names

Veda names four collections, together with the literature attached to each. It does not name a book. It names a library, held for most of its life in people rather than on paper.

The four are the Ṛgveda, the Sāmaveda, the Yajurveda and the Atharvaveda. Each has a core collection — its — and around that core a body of further material we will come to.

Why four

The tradition has its own answer, and it is worth hearing in its own terms before any other.

The Viṣṇu Purāṇa says that the Veda is "properly but one", and that observing the limited energy and application of human beings, Viṣṇu in the person of Vyāsa "makes the Veda fourfold, to adapt it to their capacities". The name Veda-vyāsa is given to the one who performs that arrangement.

Two things in that account are usually mislaid. First, the act described is division, not composition. Nothing in it says Vyāsa wrote the Vedas; it says he sorted a single mass into four so that people could carry it. Second, the same passage says the arrangement has been performed twenty-eight times, by twenty-eight different figures, of whom Kṛṣṇa Dvaipāyana is only the latest. Vyāsa, on this account, is an office rather than a man.

Historians ask a different question — how these collections actually came together — and get different-shaped answers. Those are two questions, not one contest, and this lesson keeps them apart.

The Ṛgveda

The Ṛgveda is the oldest of the four and the one everything else leans on.

It is arranged in ten maṇḍalas, circles or books. Counting the verse references straight out of the text gives 1,028 hymns and 10,552 verses — the familiar figures, and they hold up when you actually count.

The arrangement is not random. Books two to seven are the "family books", each associated with a particular lineage of seers. Books one and ten are the largest, with 191 hymns each, and are generally later. And book nine is a curiosity worth knowing: 114 hymns almost entirely to Soma being pressed and clarified. In its 1,108 verses the word agni appears four times. A whole book of the Ṛgveda is about one thing.

What is a Ṛgvedic hymn doing? Praising, mostly — Agni the fire, Indra, Uṣas the dawn, the rivers, and many more — but praising within a rite. That matters for reading. A line that sounds like a free-standing philosophical statement is often sitting inside a sequence of verses addressed to one deity for one ritual purpose.

The Sāmaveda

Here is the fact that reorganises how people think about the Sāmaveda: it is largely not new words.

Its saṃhitā, in the Kauthuma school, runs to 1,875 verses in two parts. Compare its opening lines against the Ṛgveda mechanically and two-thirds match Ṛgvedic lines exactly, and that is a floor rather than a result — because the Sāmaveda deliberately alters wording for singing, stretching vowels and inserting syllables so a melody has something to sit on. Exact matching necessarily under-counts a text whose whole business is modification.

So the Sāmaveda is not a different message from the Ṛgveda. It is very largely the same words made singable. If you want one sentence: the Ṛgveda is what is said, the Sāmaveda is what is sung.

The Yajurveda

The Yajurveda supplies what the priest says while he is doing — the formulae that accompany pouring, placing, kindling, measuring. Words for hands rather than for ears.

It comes down in two families, usually called the Black and the White Yajurveda, and the difference is not doctrinal but editorial. In the Black — the Taittirīya tradition — the formulae and the prose that explains them stand in the same text. Open the Taittirīya Saṃhitā and you find, between the formulae, exactly the kind of prose you would expect to be in a separate commentary: Prajāpati stretching out the sacrifice, "therefore" this and "one who knows thus" that. The White family is described as keeping the formulae in one book and the explanation in another; this course has not read those texts at their own pages and states that half as report.

Hold on to the Black Yajurveda, because it is about to demolish a diagram.

The Atharvaveda

The Atharvaveda is the most varied and the most domestic: healing, protection, blessing, the concerns of a household and a kingdom, alongside speculative hymns.

It is often said to have been "added later". Be careful with that. In the Chāndogya Upaniṣad — itself an early text — a student lists what he has studied and names "the Ātharvaṇa as the fourth", in sequence with the other three. The grammarian Patañjali, discussing the four Vedas, gives the Atharvaveda its own count of branches alongside the rest. By the time of both, it is simply one of the four. What is true is that its relationship to the other three was argued about; what is not established here is any claim that it entered the set late.

Mostly, this is ritual

Before the layers, one proportion that surprises almost everyone.

The Vedic corpus is overwhelmingly about — sacrifice: fire, offerings, exact formulae, exact sequence. You can see it in the shape of the collections themselves. An entire book of the Ṛgveda is given to soma being pressed. The Yajurveda exists to supply words for a person whose hands are occupied. And the Taittirīya Saṃhitā spends its prose explaining why a rite is performed as it is.

Popular presentation reverses this. It treats the philosophical material as the point and the ritual as background. Historically the ritual is the main body of the corpus and the subject of its most technical literature; the material that turns to metaphysical questions is the later and much smaller part.

If you arrive expecting philosophy you will find some, and it is extraordinary. You will find far more ritual, and you will find it treated as serious, exacting, technical work.

The four layers, and why they are not a staircase

Now the diagram everyone is shown: saṃhitā → brāhmaṇa → āraṇyaka → upaniṣad, four neat rungs, ritual at the bottom and philosophy at the top.

The categories are real and useful. The staircase is not.

Two things from texts already opened in this lesson break it.

First, the Taittirīya Saṃhitā — a — is full of brāhmaṇa-style explanatory prose. The first rung already contains the second.

Second, and more decisively: in the third lesson this course quoted the Taittirīya Upaniṣad's charge not to be negligent in study and teaching. That passage is not in a separate book at the top of a staircase. It sits inside the Taittirīya Āraṇyaka, in its seventh section, and runs straight on into "this is the hidden teaching of the Veda". The Upaniṣad is a part of the Āraṇyaka, not a storey above it.

There is a second scheme laid over the same material, and it is worth knowing because it is used constantly. The tradition later divides the Vedic corpus into a karmakāṇḍa, the portion concerning action and ritual, and a jñānakāṇḍa, the portion concerning knowledge. That division is not a neutral description either. Mīmāṃsā and Vedānta both use it, and they use it to opposite ends — one to establish that the Veda's point is injunction, the other to establish that its point is what the last portion teaches. A distinction that two schools deploy against each other is a distinction with an argument inside it.

Nor does every Veda carry all four kinds of text. Different schools organise their material differently, and some have no Āraṇyaka of the ordinary sort at all.

So what are the four words for? They name kinds of material — collection, ritual explanation, material for study away from the settlement, and material turning to other questions — and those kinds do broadly appear in that order of composition. They are not four numbered sections of one book, and they do not sit in four separate boxes.

Śākhā: what a branch actually is

A śākhā is a Vedic school: a lineage that transmits one particular recension of one Veda.

It is worth being exact, because the English word "branch" makes it sound like a sect. It is not. What a śākhā carries is a package: a specific wording of the text, a specific ordering of it, its accents and pronunciation, its attached ritual literature, and the interpretation that goes with all of that. Those things travel together because they are held by the same chain of people — which is why the last lesson's teacher-to-student chain and this lesson's textual differences are the same subject seen from two sides.

That is also why a lost śākhā is not simply a lost book. It is a lost version, a lost way of saying it, and a lost line of people.

Where the numbers come from

You will meet figures for how many śākhās there once were. They are worth tracing rather than repeating.

The classical source is the grammarian Patañjali, who lists the branches: one hundred and one for the Yajurveda, a thousand for the Sāmaveda, twenty-one for the Ṛgveda, nine for the Atharvaveda. That totals 1,131.

But look at what he is doing. He is not conducting a survey of schools. He is arguing with an opponent who claims certain words are never used, and his reply is that the field of Sanskrit usage is enormous — seven continents, three worlds, four Vedas with all their branches, plus legal literature, dialogue, history, old tales and medicine. The śākhā figures are an item in a list demonstrating vastness. And his word for the Sāmaveda's thousand is sahasravartmā, "thousand-pathed", which in Sanskrit is as much an idiom for "very many" as a count.

You will also meet the figure 1,180. This course could not trace it to a source, and does not repeat it.

What can be said, from the other end, is what survives. UNESCO's record of Vedic chanting states that of over a thousand branches of recitation, thirteen survive, and that several of those are considered under imminent threat.

Tradition and history are answering different questions

Dating deserves one careful paragraph rather than a confident sentence.

The relative order is secure: the Ṛgveda's core is older than the ritual prose, which is older than the material that turns to metaphysical questions. That much is visible in the language itself.

Absolute dates are contested, and the disagreement is sharp and sometimes political. This course will not hand you a number as though it were settled.

And the traditional account is not a rival answer to the same question. When the Purāṇa says that Vyāsa divided one Veda into four in every age, it is saying something about how the tradition understands its own transmission and its own periodicity. When a philologist dates a layer by its grammar, they are asking when certain words were composed. Neither statement refutes the other, because they are not answers to the same question — and running them together produces bad history and bad theology at once.

Where the Vedas are now

Not, mostly, on a shelf.

They are recited — at weddings, at funerals, at daily rites, in schools where boys still learn a recension by heart. They are quoted in argument. Portions of them are known by people who have never read a page, because they hear them at every ceremony they attend. And a small number of lineages still carry the full apparatus of the third lesson.

Meanwhile most Hindus, most of the time, encounter their religion through everything except the Vedas — which is not a failure of piety but a fact about how a very large tradition distributes itself.

What to carry out of this

"The Vedas" is a library, not a book. Four collections, each with a layered literature attached, transmitted in schools that differed from each other. The singular makes it sound like something you could open.

The tidy diagram is a teaching aid, not a description. Saṃhitā, brāhmaṇa, āraṇyaka, upaniṣad name kinds of material. Real texts overlap, nest inside each other, and vary by school.

Check the numbers. Some of the famous ones — 1,028 hymns, 10,552 verses — hold up exactly. Others turn out to be a grammarian making a rhetorical point, or to have no traceable source at all. Knowing which is which is most of what it means to read this material well.

Sources & context

These notes point at the texts and traditions a lesson draws on. They are references, not proof.

Ṛgveda

10.90 (Puruṣasūkta)reference being checked

Vedic

The text itself, read directly.

Parmeshwari's readingThe Puruṣasūkta read complete. The lesson uses the hymn's frame — the same grammar yields the varṇas and the moon — and states that verse 12 says rājanya.

Ṛgveda Saṃhitā

maṇḍalas 1–10, complete

Vedic

The text itself, read directly.

Parmeshwari's readingEvery number the lesson gives for the Ṛgveda — ten maṇḍalas, 1,028 hymns, 10,552 verses, maṇḍala 9's shape — was counted out of this text rather than quoted from a summary.

Sāmaveda Saṃhitā, Kauthuma śākhā

Pūrvārcika and Uttarārcika, complete (1,875 verses)

Vedic

The text itself, read directly.

Parmeshwari's readingThe 1,875 verses, and the comparison with the Ṛgveda that the lesson reports as a floor rather than a result.

Taittirīya Saṃhitā (Kṛṣṇa Yajurveda)

kāṇḍa 1 onward, read for its prose

Vedic, Taittirīya śākhā

The text itself, read directly.

Parmeshwari's readingThat brāhmaṇa-type prose sits inside a saṃhitā — the first half of the lesson's case against the four-rung staircase.

Taittirīya Āraṇyaka

praśna 7, at internal reference 7.11reference being checked

Vedic, Taittirīya śākhā

The text itself, read directly.

Parmeshwari's readingThat the Taittirīya Upaniṣad is a section of the Taittirīya Āraṇyaka — the second and decisive half of that case, and a direct link back to Lesson 03, which quoted the same passage under its Upaniṣad name.

Vyākaraṇa-Mahābhāṣya of Patañjali, Paspaśāhnika

Paspaśāhnika, the passage on the extent of śabda-prayoga

Vyākaraṇa

A classical commentator reading the same passage — their view, named as theirs.

Parmeshwari's readingThe śākhā figures, and — more importantly — the argument they occur inside, which is what tells you what kind of claim they are.

Viṣṇu Purāṇa

Book III, on the division of the Veda in every Dvāpara age

Purāṇic

The text itself, read directly.

Parmeshwari's readingThe traditional account of why there are four Vedas: one Veda divided for human capacity, by a Vyāsa who is an office held twenty-eight times. Cited as tradition's own answer, held apart from the philological question.

Tradition of Vedic chanting (UNESCO Representative List of the Intangible Cultural Heritage of Humanity)

inscription file 00062

Modern scholarship on how this developed, rather than on what it means.

Parmeshwari's readingThe surviving-branch figure, set against Patañjali's traditional enumeration so the reader can see the two ends of the same question.

Chāndogya Upaniṣad

7.1.2reference being checked

Vedānta / Upaniṣadic

The text itself, read directly.

Parmeshwari's readingThat an early Upaniṣad already counts “the Ātharvaṇa as the fourth”, which is why the lesson declines to assert that the Atharvaveda was a late addition.

An Introduction to Hinduism

Modern scholarship arguing for a reading. One voice in a live conversation.

Parmeshwari's readingThe declared basis for the one reported claim in this lesson: the later division of the corpus into karmakāṇḍa and jñānakāṇḍa, and its use by Mīmāṃsā and Vedānta to opposite ends.