Chapter One　The Mechanism of New Sutras: A Media Revolution¶

Around the turn of the Common Era, something unprecedented happened in the history of Buddhist literature. New sutras appeared in batches: vast in scale, accomplished as literature, all beginning with ‘Thus have I heard,’ all claiming to be the Buddha's word, and all absent from the early layer of recitation attested across the schools. Who wrote them, where, and why they dared to use that attribution are questions whose evidence and hypotheses Volume One, Essay Four has already sorted out: the latest possible date established by the physical Gandhāran manuscripts, the chronological anchor supplied by Lokakṣema's Chinese translations (around 179), the evidential standing of the legends of the nāga palace and the three turnings of the Dharma wheel, and the verdicts on the competing origin hypotheses.¹ This chapter will not repeat that work. It asks the question proper to a history of transformation: what mechanism allowed a tradition to add thousands upon thousands of new scrolls of ‘the Buddha's word’ onto itself without splitting, changing its Vinaya, or obtaining certification from any council?

First consider the gate's original design. Admission to the sutras was a content check from the outset. The four great standards in the Mahāparinibbāna Sutta stipulate that a claim to have ‘heard this directly from the Buddha’ is neither to be accepted nor dismissed: take its wording and compare it with the sutras and Vinaya; accept it if it fits, discard it if it does not. The check concerns consistency of content, not a chain of provenance, since on the sutra side there was no chain to inspect in the first place. The Vinaya's chain of living people governs identity alone, and the definition of schism governs procedure alone. Volume One has already established that Mahāyāna founded neither a separate ordination lineage nor a separate poṣadha, and thus never constituted schism in the Vinaya sense. The tradition knew that this gate was soft. The drum-peg analogy had already predicted that later monks would take no pleasure in hearing the profound sutras spoken by the Tathāgata, preferring ornate new sutras made by poets. As the drum's body cracks, pegs are added; by the time the pegging is finished, none of the original body remains.² Nor did attempts to make up for this ever stop: the provenance formula ‘Thus have I heard,’ repeated preservation in lineages of reciters (Essay Two, Chapter Two), and collective recitations held again and again. All of these worked on the content side, and none was strict. The route taken by Mahāyāna sutras bypassed precisely the two locks of identity and procedure and went through the loose gate of the content check: leave identity untouched, leave procedure untouched, and produce texts.

Why did great numbers of new sutras pass this gate at just this moment? The first variable was the medium. In an oral age, ‘publishing’ a sutra was extremely expensive. One had to persuade a guild of reciters to add it to the memorization curriculum and have dozens of people pledge their lifelong memories to preserve it. Guilds were conservative, their capacity finite; admitting something new took space from something old. In the manuscript age, a pen and a stack of birch bark sufficed. The earliest Mahāyāna manuscripts (Gandhāra, first to second centuries) and the earliest Chinese translations (around 179) fall squarely within the window in which writing spread. In almost the same century or two, Sri Lanka fixed its canon in writing at Aluvihāra (Essay Three, Chapter Three). Conservers and innovators used the same new technology: one to fix the old texts, the other to produce new ones.

We can name the earliest new sutras of that window. The physical evidence is the Gandhāran birch-bark scroll of the Aṣṭasāhasrikā Prajñāpāramitā. Radiocarbon dating places it in a range roughly corresponding to the first century CE; its language is the Gāndhārī vernacular, and its text corresponds to the Daoxing bore. This is the earliest surviving Mahāyāna manuscript, and it is a Prajñāpāramitā text. The Chinese anchor is Lokakṣema's chest of translations: the Daoxing, the Pratyutpanna, the Akṣobhya, the old translation of the Kāśyapa Chapter titled Yiri moni bao, and the earliest fragment of the Avataṃsaka, the Dousha Sutra. Internal stratification adds texts such as the Inquiry of Ugra: almost nothing about emptiness, no claim to the name Mahāyāna, only a stringent bodhisattva practice. Volume One, drawing on Nattier, has already noted its very early date. Taken together, these place the composition of the oldest core around the first century BCE. But this batch is not the point of origin; it is only the earliest to leave physical traces. In the lists of meritorious acts, ‘receiving and retaining, reading and reciting’ precede ‘copying’; the verse core represented by the Ratnaguṇasaṃcayagāthā family bears traces of vernacular language and oral recitation. Both point to an earlier oral period of gestation. Beneath that lies the mainstream saṅgha's own oral literature of jātakas, avadānas, and Buddha biographies: ‘bodhisattva’ was a word in jātaka literature to begin with. That oral layer is in principle beyond recovery, for the reason given in the next paragraph.³

Oral transmission operates by majority decision. A guild's memory is a shared resource. For a sutra to live in oral transmission, a community must agree to pledge dozens of people's lifelong memories to it. A minority's new work cannot be transmitted in this system. What writing changed was less the cost of production than the condition for survival: for the first time, a small circle without majority support could let its text outlive its members. This is the point of Gombrich's famous claim that writing was the technological prerequisite for Mahāyāna to begin.⁴ The ‘earliest batch’ we can see is therefore, by definition, the first batch that survived. Mahāyāna's traceable history begins only when a minority's texts become able to survive.

The second variable lies inside the new sutras themselves: they are texts designed for the age of writing. A group of Mahāyāna sutras repeatedly proclaims the same kinds of merit. Receiving and retaining, reading and reciting, copying and making offerings to this sutra bring greater merit than giving the seven treasures throughout worlds as numerous as the sands of the Ganges; wherever this sutra is found, a stupa is present.⁵ Put back into media history, this language is a call to copy. Older scriptures rarely urge people to copy them. The comparison has quantitative readings, and the readings form a gradient: counting terms such as merit, virtuous merit, receiving and retaining, reading and reciting, and four-line verse, the frequencies per ten thousand Chinese characters are one hundred and ninety-four in the Diamond Sutra, seventy in the Lotus, sixteen in the Saṃyukta Āgama, and three in the Middle Treatise. Passages comparing merit make up about three tenths of the Diamond Sutra's main text; see Volume One, Part Two, Essay Twenty-Three, Chapter Two. The Saṃyukta Āgama's figure is sixteen, not zero. Older scriptures had the oral guild to preserve them. New sutras lacked a guild, so they wrote the motive for copying into the text, rewarding copying with merit. A text with a built-in reward for copying competes on unequal terms with one without it. The cult of the book was more than a new form of devotion: it was how texts spread under a new medium. Volume One has already recorded Schopen's name for it, a ‘religion of the book.’

The third variable likewise lay inside the texts: the new sutras supplied themselves with a full set of accounts vouching for their provenance. At the level of provenance, preservation in a nāga palace, direct hearing in Tuṣita, and seeing the Buddha in concentration each explained away the question ‘Why have you never heard of this?’ Volume One identified the reports of seeing the Buddha in pratyutpanna samādhi as the key to understanding ‘how new Buddha-word was produced.’ At the level of testimony, the Lotus had Prabhūtaratna Tathāgata emerge from the earth and attest from within his stupa, ‘So it is, so it is.’ The new sutra supplied its own witness inside its narrative. At the level of defense, a group of sutras incorporated predictions that ‘in the last five hundred years, those who hear will be alarmed, doubt, and slander,’ rewriting future challenges in advance as fulfillments of prophecy.⁶ Together, the three levels supplied exactly all the credibility missing from a text without a guild or a chain. Their very presence shows that the authors knew precisely where the threshold stood.

Combine these variables and consider a single family: why do all the new sutras in the Prajñāpāramitā family speak of prajñā? The answer has four strands. First, literal inheritance. These sutras are one sutra expanded and derived again and again, not many independent revelations that happen to share a topic. The scholarly genealogy is clear: the Eight Thousand Lines is the parent text; the Twenty-Five Thousand and Hundred Thousand are expansions made by ‘list machine’-style enumeration—the same skeleton, each concept multiplied by the eighteen emptinesses and the hundred and eight phrases. The Diamond and Heart are later condensations. Sharing one source, they naturally speak the same language. Second, lineage. The people writing these sutras were Dharma preachers trained in reciting them; a new work continued their own lineage's recited material. Third, a loop of elicited visions. In the Pratyutpanna mode of production, one sees a Buddha in concentration and the Buddha teaches the Dharma, which turns out to be precisely the content already in the practitioner's curriculum. The sutra half acknowledges this: its passage ‘Mind makes the Buddha; mind is the Buddha’ analyzes the concentrated vision as mind-made, then immediately treats that analysis as proof of emptiness. The vision is prearranged by the doctrine it confirms, and the doctrine is then confirmed by the vision. The loop closes, the new text enters the canon, and it becomes what the next meditator has already learned. Fourth, selection effects. Texts carrying rewards for copying were copied most often; today's ‘earliest record’ is dominated by the family best at promoting itself. Together, the four strands make the family's homogeneity the mechanism's necessary result, no longer a mystery.⁷ Why prajñā in particular was selected as the strain to propagate is a question about motives; Chapter Three of this essay will answer it.

The mechanism's deepest level lies in the logic of the check itself. This volume calls it the ratchet of canonical admission. The check instituted by the four great standards tests a new text against a corpus. This kind of check has a structural property: things go in, never out. Every text wrongly admitted becomes tomorrow's standard; a text wrongly rejected does not tighten the standard in return. A step effect thus becomes possible. One cannot pass through the gate directly from the Āgamas to the Lotus, but each new layer need agree only with ‘the corpus as of yesterday.’ Ten steps, no single violation, and the drum's body is entirely pegs. Apologists changed the check once more: ‘Whatever accords with Dharma-nature is the Buddha's word.’ The four great standards were originally a necessary condition—discard what does not fit. This sentence reverses it into a sufficient condition—admit whatever fits. The same check, with its quantifier reversed, turns from exclusion into admission. Volume One, Essay Four, Chapter One has already recorded the doctrinal history of this redefinition of the criterion for ‘the Buddha's word’; here we add its logical form. The ratchet needs no one to act in bad faith. Each admission can be locally honest. Drift is a structural property of the checking method, visible only on a scale of centuries that no participant lives through.

The blue route bypasses the two locks; it does not break the ordination lineage or alter formal communal acts and Uposatha. “Content test” means comparison with the received discourses and Vinaya. “Entry without exit,” the staircase, and “quantifier reversed” state the author’s account of how admission changes the corpus used to judge later texts. The last phrase names the shift from a necessary condition—reject what fails the test—to a sufficient condition—admit whatever passes. The stairs are schematic, not a count of historical stages or a dated series.

The guild’s “majority vote” means communal commitment of memory to preserving a text, not a documented ballot. The English fountain pen is a symbol of writing, not an ancient implement. The diagram presents the chapter’s mechanism, while its note preserves a direct scholarly dispute: Drewes challenges the claim that writing was necessary for early Mahāyāna texts to survive. That disagreement is not merely a division of labor between two complementary accounts. The chapter also retains earlier oral development, expansion of non-Mahāyāna collections, and the possibility that new texts were produced and accepted in good faith.

As usual, the plainer half of the account must be given in full. The schools' own canons were growing too: jātakas, avadānas, and Abhidharma collections expanded layer upon layer. Expansion of the sutra collection was no Mahāyāna monopoly; Mahāyāna merely exceeded the others by an order of magnitude in both scale and self-awareness. Nor were most authors of new sutras thinking like forgers. What they saw in concentration or inferred from the Dharma counted as the Buddha's word in their epistemology, as Volume One's section on its redefinition explains. To accuse individuals of dishonesty is to choose the wrong target. From beginning to end, this chapter's subject is the gate's engineering. One comparison, in broad strokes, with the details left to each tradition's history: traditions that rebuilt their canonical gate as a chain lock do exist. Islam supplied hadith with chains of transmission and criticism of transmitters; Christianity closed its canon through councils. Buddhism put its locks on people, not texts. Three ways of locking, three histories of transformation.

Of the four questions, this chapter gives only brief answers to the last two, leaving the rest to the next three chapters. Those able to write benefited. The cost fell on the term ‘the Buddha's word’: it changed from a historical claim into a standard admitting anything that met the criterion. Once the gate had been altered in this way, what entered depended only on what stood outside. The next chapter considers the first thing through the door: an entirely new ideal person.

References¶