The Book That Was Never Written

A man sits cross-legged before dawn in a room with no books in it.

He has been awake for some hours. He will recite, without pause and without a page in front of him, a body of text that runs to more than a thousand hymns and roughly ten thousand verses, and he will do it in an order that is not the order in which the hymns are normally arranged. He will say each word, then say it again paired with the word before it, then say it again paired with the word after it, weaving forward and backward through the line like a man braiding rope. If he drops a single syllable, the braid will not close. He has done this since he was a boy. His teacher did it. His teacher’s teacher did it. The chain behind him, if the tradition is right about itself, runs back past the founding of every city he has ever heard of.

There is no book. There has never been a book. That is the point.

This is the central fact of the Vedas, and it is so strange that most accounts hurry past it. A civilization decided to keep its most important text out of writing on purpose, for a very long time, and built a machine out of human beings to do it. The machine worked. The forensic question that opens this volume is simply this: how do you keep a text perfect for three thousand years without writing it down, and what does it mean that a people chose to.

Two words that do all the work

The tradition does not call the Vedas books. It calls them shruti (pronounced roughly shroo-tee), “that which is heard.” The word is precise and it is doing theological work. To the orthodox schools, the Vedas were not written and not even, in the ordinary sense, composed. They were heard. The seers to whom the hymns are attributed, the rishis (ri-shees), are described as the ones who perceived the sounds, the way an astronomer perceives a star that was already there. Patrick Olivelle, introducing his translation of the early Upanishads, notes that this category of “what is heard” is set against smriti (smri-tee), “what is remembered,” a second and frankly more human tier that holds the epics, the Puranas, and the law codes (Olivelle, The Early Upanishads, 1998). Shruti is revelation. Smriti is tradition about the revelation. The line between them is one of the oldest distinctions in Indian thought.

The second word is harder. The Vedas are called apaurusheya (a-pow-rush-eya), “not of human authorship,” sometimes rendered “not man-made.” The Mimamsa school, the school most concerned with the authority of the Veda, pushed this to its limit and argued that the Veda is eternal and uncreated, that it has no beginning and no maker at all, divine or human. On this view the rishis are not authors. They are witnesses.

It is worth being exact about what kind of claim this is. It is a claim made by the tradition about itself, a theological position with a long and sophisticated history of argument behind it. The historian’s account of where the hymns came from is a different account, and Chapter 6 will take up the dating in detail. The two need not be reconciled here. What matters for now is that the people who carried this text believed, and a great many still believe, that they were carrying something they had no right to alter. That belief is not incidental to the preservation. It may be the reason the preservation succeeded.

The size of the impossibility

Set theology aside and look at the engineering problem on its own terms.

The oldest of the four collections, the Rigveda (rig-vay-da), the “Veda of verses,” contains 1,028 hymns gathered into ten books, running to something on the order of ten thousand verses and roughly forty thousand lines. Maurice Bloomfield’s Vedic Concordance of 1906, which indexed the opening words of Vedic verses across the literature, listed tens of thousands of distinct lines. This is a large text by any standard. Now add the constraint that you may not write it down, that you must reproduce it without the loss or substitution of a single sound, that the meaning of the words has in places drifted so far that even the ancient commentators argued over it, and that you must hand it intact to a student who will hand it intact to his, across a span of time longer than most living languages have existed.

The remarkable thing is not that the tradition attempted this. The remarkable thing is that it appears to have succeeded.

When Friedrich Max Müller and the European philologists of the nineteenth century began comparing recitations gathered from reciters in different parts of India, separated by hundreds of miles and many generations, they found the texts agreeing to a degree that startled them. Frits Staal, the Dutch scholar who spent decades studying Vedic recitation in the field, argued that the oral transmission of the Veda is one of the most accurate feats of memory anywhere in human culture, accurate enough that he was willing to call its discipline a forerunner of formal science (Staal, “The Fidelity of Oral Tradition and the Origins of Science,” 1986). That is a strong claim, and not everyone goes as far as Staal does. But the underlying observation is not seriously contested. The text was kept, and it was kept by ear.

A book is the wrong picture

The instinct of a modern reader is to imagine a lost original manuscript, some founding scroll from which all the recitations descend, and to picture the reciters as living photocopiers of that scroll. Reverse the picture. There was no scroll. The recitation was the original. The manuscripts, when they finally came, were the copies, and they came late.

How late is itself a piece of forensic evidence. The Indologist Wilhelm Rau, whose work on the question is summarized by Michael Witzel, pointed out that physical Vedic manuscripts are strikingly recent compared with the antiquity of the text. The surviving Rigveda manuscripts are medieval, the oldest dating to roughly the eleventh century of the common era or after, separated from the likely age of the hymns by something close to two thousand years (Witzel, “Vedas and Upanishads,” 2003). For most of its life the most authoritative text in the tradition existed in no physical form at all. It lived only in trained throats.

And writing, when it was available, was actively distrusted for this purpose. A frequently cited later verse warns that one who has learned the Veda from writing, or who recites without proper command, is among the lowest of reciters. Whatever its exact provenance, the sentiment is well attested in the tradition: a written Veda was a degraded Veda, a crutch for a memory that should not need one. The page was not the safe place. The page was the dangerous place, because a copyist can introduce an error that propagates silently, while a reciter trained in the cross-checking methods will hit a snag the moment a syllable is wrong. The tradition trusted the disciplined human voice over the silent line of ink, and on the evidence of the fidelity, it was not obviously mistaken to do so.

There was no scroll. The recitation was the original. The manuscripts, when they finally came, were the copies.

What the text says about itself

The Rigveda is, in places, unexpectedly self-aware about the act of speech that it is. In the famous hymn to Vac (vaak), Speech personified as a goddess, the verse declares, in Ralph Griffith’s public-domain translation of 1896:

“I travel with the Rudras and the Vasus, with the Adityas and All-Gods I wander. … Through me alone all eat the food that feeds them, each man who sees, breathes, hears the word outspoken.” (Rigveda 10.125, trans. Griffith)

Here the text imagines sound itself as the medium through which the world is held together, which is a striking thing to find inside a text that was itself held together by sound and nothing else.

Elsewhere the hymns reflect on how speech is made. Rigveda 10.71, a hymn on the origin of language, says that the wise “fashioned speech with thought, sifting it as men sift the parched grain with a sieve” (paraphrasing Griffith’s rendering of 10.71). The image is of refinement, of a thing worked over and made exact. That, too, describes the transmission. The text is, in part, a meditation on the very faculty that preserved it.

We should resist over-reading these verses as a how-to manual for memorization. They are hymns, not technical treatises, and the technical methods are a later development that Chapter 5 will examine. But it is worth noticing that the people who carried the Veda did not experience sound as a fragile or secondary thing. Sound was, to them, the firm ground.

Why memory and not paper

It is tempting to assume the Vedas went unwritten simply because writing had not yet arrived. That explanation is too easy, and the chronology does not fully support it. Writing systems were known in the subcontinent in forms whose dating and decipherment remain debated, and by the time of the later Vedic literature, scripts for everyday use were available in the wider world. The point is that even after writing was an option, the tradition kept the Veda oral for a very long time and ranked the oral form above the written one. That is a choice, and the historian has to treat it as one.

Several reasons have been proposed, and they are not mutually exclusive. The first is fidelity, the argument we have already met: a trained oral system with built-in cross-checks may actually be more reliable than hand-copied manuscripts, which accumulate scribal drift. The second is access. To keep the text in memory, transmitted only by initiated teachers to qualified students, is to control who may possess it, and the Vedic tradition did restrict access to its recitation along lines of birth and training. The third is the theology already described: if the Veda is shruti, heard sound, then to fix it on a page is to mistake its nature, to treat the eternal vibration as if it were a perishable object. Witzel and others have laid out these factors; we do not have to choose a single one (Witzel, 2003). Human institutions usually have more than one reason for what they do, and the more durable the institution, the more reasons tend to be braided into it.

What we can say plainly is the consequence. For roughly three millennia, give or take depending on how the dating falls, the most authoritative text of one of the world’s largest religious traditions was a thing you could not hold, could not buy, and could not read. You could only hear it, and to hear it accurately you had to enter a discipline that took years. The text was a practice before it was ever a document.

The wonder, stated plainly

It is easy to lose the strangeness of this under the weight of the scholarship, so let us hold it in front of us for a moment without commentary.

A society looked at its most precious knowledge and decided that the safest place to keep it was inside people. Not in a temple vault, not on stone, not on bark or palm leaf, but in the trained memory of one generation after another, checked against itself by methods that made silent error nearly impossible. It worked. The thing they were protecting reached us, and reached us in a state clean enough that modern philologists can argue about individual syllables with the confidence that the syllables are really there to argue about.

That is the wonder this whole volume is built around. Everything that follows is, in one way or another, an attempt to understand it.

The road ahead

This volume, The Heard, takes the Vedas as its subject, and it will move from the outside in. The chapters to come will lay out the four Vedas and what each is for, the Rigveda of verses, the Yajurveda of sacrificial formulas, the Samaveda of melodies, and the Atharvaveda of the householder and the healer, and they will show how each collection is itself built in layers, from the hymn-collections through the ritual manuals to the philosophical Upanishads at the far end. We will look hard at the rishis, the seers to whom the hymns are credited, and ask what the tradition means when it calls them authors who did not author. We will open up the memory-machine itself, the elaborate recitation methods that made the fidelity possible, and we will face, as squarely as the evidence allows, the contested and unfinished question of when all this began.

But before any of that, we have to do something the modern reader rarely does with a scripture. We have to listen to it the way it was meant to be received, as sound moving through air and through people, and to understand why the tradition insisted that the how of the Veda was inseparable from the what.

For that, we begin where the tradition begins. Not with a book, because there is no book, but with a voice. In the next chapter we walk into the world of shruti itself, and ask what it actually felt like to receive a text by hearing.