Chapter 2. Councils and the oral transmission of scripture¶

The Buddha's death left the saṅgha with practical questions about what he had taught, the dhamma, and how its members should live, the vinaya. No successor had been appointed, as the preceding chapter explained, and all their material depended on memorization, as discussed in 1.4. The answer had to be collective. They gathered what people remembered, recited it, and established an agreed form. The procedure was called saṅgīti, literally "recitation together," or a council. The present chapter examines several councils, the shape of the scriptures, the languages of transmission, and the date of their writing. As elsewhere, it carefully distinguishes legends that authorize scripture from historical facts that can be inferred.

1. Communal recitation and the First Council¶

A council was a communal recitation, as explained in 1.4. People gathered to recite remembered teachings and disciplinary rules passage by passage, correct one another, and establish an accepted version.

Tradition places the First Council at Rājagṛha in the year of the Buddha's death. Mahākāśyapa assembled five hundred arhats. Ānanda recited the teaching (the discourses), and Upāli recited the discipline.¹ The unresolved question of abandoning the "minor and lesser rules," introduced in the previous chapter, ended here. The Buddha had allowed the saṅgha to abandon them but had not specified which rules qualified. Ānanda could not explain why he had failed to ask, so the assembly agreed to retain every rule.

The scene cannot be accepted literally. Five hundred arhats fixing the whole vast body of later discourses and discipline at a single meeting is an idealized founding narrative. Its purpose is to authorize scripture as having been established in recitation by people who heard the Buddha themselves.² The position taken in the present book is that the historical core of the First Council can amount at most to a smaller gathering to arrange teachings and disciplinary rules after his death. The Sutta Piṭaka we have today accumulated and took shape over a long subsequent period, rather than being settled at one meeting.³ The Vinaya even preserves a discordant account. The elder Purāṇa arrived afterward and was asked by the elders to accept their version. He replied that they had recited well, but he would retain the teaching as he himself had heard it from the Buddha.⁴ His small act of noncompliance discloses the lack of complete uniformity in early recitation.

A narrative observation on the blame assigned to Ānanda

Ānanda receives an unusually large share of negative attention in narratives surrounding the First Council, a feature that deserves separate notice. The Vinaya and related traditions assign him at least the following faults at the council. First, he did not ask the Buddha to remain alive. The Mahāparinibbāna Sutta says that the Buddha repeatedly hinted that he could remain for an aeon if asked. Ānanda failed to respond each time, and the Buddha later departed after Māra came to urge him. Second, Ānanda advocated allowing women to enter monastic life, supposedly shortening the duration of the true teaching by five hundred years. Third, he allowed women to pay homage to the Buddha's body first, and their tears reportedly wet the Buddha's feet. Fourth, he failed to ask which rules were "minor and lesser," leaving the council unable to decide and obliged to retain them all. A passage in the Saṃyutta Nikāya also attributes the following words to Ānanda: "Venerable Kassapa, please be patient; women are foolish" (Khamatha bhante Kassapa bālo mātugāmo). Bhikkhu Bodhi renders the words faithfully in his English translation and defends his rendering in a note. Someone had urged him to translate the passage as "she is a foolish woman" to avoid criticism. He regarded that approach as distorting the Pāli to conform to contemporary views of gender. Whether Ānanda actually spoke the words or the compilers placed them in his mouth was, he said, a question he left to the reader.⁵

Seen from another angle, Ānanda was precisely the person in the early saṅgha most sympathetic to women. He promoted the admission of nuns and attended the Buddha for the longest time. He remembered the most teaching and was the kindest to lay followers and ordinary people. Mahākāśyapa, by contrast, presided over the council and demanded accountability as a representative of the strict tendency. Tension between the two may reflect actual competition within the early saṅgha between a relatively open tendency and a strict, elite tendency. When the oral tradition was fixed in writing, one position had gained an organizational advantage and supplied a narrative framework to justify its choices.

The observation is not a settled conclusion. Whether each alleged fault occurred, what Ānanda actually said, and how the debates at the council unfolded are details that history cannot recover. The present book calls attention to them to remind readers that the surviving accounts were already shaped by their narrators. Their portraits need not mirror the historical individuals.

The techniques of oral recitation

Communal recitation required techniques for transmitting long teachings accurately by word of mouth. They were built into the structure of the discourses and remain visible throughout the surviving texts:

• Standard openings. Every discourse in the Dīgha and Majjhima Nikāyas opens with the same Pāli words, "evaṃ me sutaṃ" ("Thus have I heard"), followed by the location and audience before the main text begins. The Saṃyutta and Aṅguttara Nikāyas do not write out the formula for each discourse. They usually give it once at the beginning of a chapter or group, then substitute "the setting was Sāvatthī" or simply continue with an ellipsis.¹¹ The opening told reciters where they were in the account as soon as they began. Its abbreviated use is itself evidence that it had become a formula which did not need to be repeated in full each time.
• Formulas and repetition. Questions, answers, arguments, and statements follow fixed patterns. The same passage often recurs word for word in different discourses. Long discourses abbreviate substantial passages with the marker peyyāla, or "abridgment," which reciters expand in actual performance. The design serves memory rather than reflecting literary laziness.
• Numerical arrangement. The Aṅguttara Nikāya and the matrices organize teachings as "one item, two items," and so on. Their basic technique attaches teachings to numbers, the most reliable supports for memory.

Recitation also became a specialized task. The Vinaya and later commentaries describe separate groups of reciters (bhāṇaka) responsible for different parts of the Sutta Piṭaka. Reciters of the Dīgha, Majjhima, and Saṃyutta Nikāyas each maintained a distinct collection.⁶ Researchers have not settled how strongly the division shaped textual transmission. Differences in formulaic wording between collections may come from the reciter system, or may reflect differences between manuscript traditions. What is certain is the division of responsibility itself, which separated the complete canon into transmissions maintained by different groups.

The structure explains an apparent contradiction. Independent northern and southern versions agree to a considerable degree at sentence level, but differ substantially in the arrangement, selection, and contextual relations of discourses.⁷ Sentences were the basic units of oral recitation and were relatively stable. The organization of discourses involved editorial choices, which diverged early and quickly.

Further accounts also show how idealized the image of a single meeting establishing the whole canon is.

The first is the "council outside the cave." According to the story, Mahākāśyapa admitted only five hundred arhats to the Seven-Leaf Cave for communal recitation. Large numbers of excluded monks held a separate recitation outside, calling themselves the "great community." One later explanation derives the name Mahāsaṅghika, or "great community," from the group outside the cave. The account coexists with the other explanations of division discussed in 2.3, each an origin narrative with its own commitments. Parallel recitations inside and outside the cave themselves indicate that scripture was not represented by a single version from the beginning.

The second account requires a leap of more than a thousand years. Chan asserted a "separate transmission outside the teachings." It claimed a lineage apart from written scripture, beginning when the Buddha held up a flower and Kāśyapa smiled, transmitted from mind to mind without reliance on words. "Transmission" was thus placed entirely outside the councils and the three baskets. The claim arose much later and belongs to Chinese Chan. Only its existence is noted here. The question it raises, how a transmission outside writing and the canon can count as authoritative, belongs to the same class of problems as the origins of Mahāyāna scripture discussed in Essay 4. The Mahāyāna discussion will address it formally, with Chan treated in its own chapter.

2. The Second Council and the ten practices¶

Tradition places the Second Council at Vesālī about a hundred years after the Buddha's death, prompted by a disciplinary dispute. In the southern account, local Vajjian monks followed ten practices alleged to be unlawful. The most consequential concerned accepting and keeping gold and silver, or whether monks could receive and hold money. The council ruled all ten practices unlawful.¹

The Second Council is much more credible than the first. It describes a specific, actual dispute about strict and lax discipline, with more ordinary details. The next chapter examines the ten practices and the schools' differing accounts of their relation to the root division between Sthaviras and Mahāsāṃghikas. The sources disagree in many places, and the question of who used the ten practices as grounds to break away is much more complicated than it appears. For now, the point to retain is that the first major documented breach within Buddhism concerned disciplinary practice, identified as money by the southern tradition, rather than advanced doctrine.

3. The Third Council in the Theravāda account¶

Tradition also describes a Third Council at Pāṭaliputta under Aśoka, around 250 BCE. Moggaliputta Tissa presided. The council purged divergent views that had entered the saṅgha and compiled the Kathāvatthu to refute them. Missions were sent out afterward, including one to Sri Lanka.¹

The Third Council appears only in the southern Theravāda tradition's own records and is not recognized by the other schools. Its strongly sectarian account serves to establish the southern lineage as orthodox. The judgment of the present book is that it is better understood as an internal reform within the Theravāda line and a starting point for its outward expansion than as a council of the whole Buddhist community. The three councils are compared below:

4. The shape of the three baskets¶

The scriptures eventually took shape within the framework of the "three baskets" (tipiṭaka): the Sutta Piṭaka of the Buddha's teachings, the Vinaya Piṭaka of discipline discussed in the previous chapter, and the Abhidhamma Piṭaka.

The place of the Abhidhamma Piṭaka requires a careful judgment that avoids opposite extremes. One treats it as a "third basket" established by the Buddha himself, equal in status to the discourses and discipline. The other dismisses the entire collection as late, with each school's version unrelated to the others. Neither is accurate.

The evidence has distinct layers. The Abhidharma traditions of the schools are not fundamentally different in every respect. Their analyses of "mind and mental factors" and their methods of classification and enumeration share a substantial core, which may be called a common "phenomenology of mind." Above the core, however, each school added large amounts of plainly later material, including systematic inventories of dharmas and metaphysical assertions such as "real existence in all three times," discussed in Essay 3.⁸ The position taken in the present book is that the Abhidharma collection became a formal "third basket" relatively late, and its metaphysical levels were scholastic products of the individual schools. Its core method, however, had an early source shared across schools, rather than appearing late without antecedents.

The evidence points to the mātṛkā, a "matrix" or "mother" of analytical teaching. Such classified lists condensed teachings into outlines that were easy to memorize and elaborate. The method is already visible within the Sutta Piṭaka. A typical example is the Saṅgīti Sutta in the Dīgha Nikāya (DN 33). In its narrative setting, the Buddha has back pain and lies down to rest, asking Sāriputta to teach in his place. Sāriputta then recites the teachings arranged in groups of one, two, and so on up to ten items.⁹ The discourse is almost an embryonic Abhidharma embedded in the Sutta Piṭaka. There is also a basis for the traditional view of Sāriputta as the disciple most skilled in analysis. The surviving Śāriputrābhidharma, transmitted by the Dharmaguptaka, preserves an older common form and is regarded as a living remnant of early Abhidharma.⁸

The following proposal is explicitly a historical conjecture, rather than a fact. Subsidiary analytical and classificatory teachings of the mātṛkā type may well have begun to form during the Buddha's lifetime. In so large a saṅgha, with able people such as Sāriputta, a complete absence of supplementary or systematic teaching would be surprising. Yet the teachings were disciples' systematic arrangements, rather than discourses delivered by the Buddha, and did not begin with "Thus have I heard." They therefore followed a separate course outside the Sutta Piṭaka, later expanded by each school into a complete Abhidharma collection. We may infer one step further from the substantial shared core and largely later differences between the schools' Abhidharma traditions. Their core analysis of mental phenomena through matrices took shape early, perhaps before the schools separated or even during the Buddha's lifetime. Their differing superstructures were built separately after the divisions.

The Sutta Piṭaka is internally arranged into five "Nikāyas" in the southern tradition and four "Āgamas" in the northern tradition. The collections broadly correspond, though transmitted by different schools. Correspondence is not exact discourse by discourse, and some school affiliations are settled while others remain uncertain. The table gives only the broad pattern:

The fact that different northern and southern schools recited the same body of teachings independently and formed separate collections will be used again in Section 6.

5. Languages of transmission and the move to writing¶

A common misconception needs correction. Pāli was not the Buddha's language. He probably spoke a vernacular of eastern central India close to Old Māgadhī, whereas Pāli is a literary Middle Indo-Aryan language with western features.¹⁰ The Theravāda claim that Pāli is the "original language" and the language of the Buddha is a later expression of devotion, which the present book does not accept.

The early Buddhist attitude toward language is more consequential. The Vinaya tells of two monks from brahmin families who asked to render the Buddha's teaching chandaso. The Buddha refused and instructed people to preserve his words "in their own language" (sakāya niruttiyā).¹⁰ Both key expressions are uncertain. Chandaso has been understood as rendering into Sanskrit, into Vedic, into verse, or as one wishes. Norman rejects the Vedic interpretation and observes that the last two readings would leave the passage saying nothing about the Buddha's language. The passage should therefore not bear the argument on its own. Whether "in their own language" refers to each person's dialect or the Buddha's own speech has also long been disputed. Only the Buddha's rejection of the proposal is certain. More reliable evidence for early Buddhism's preference for local speech over Sanskrit comes from the languages the schools actually used. The Theravāda used Pāli, and some northwestern schools used Gāndhārī. The Sarvāstivāda, Mahāsāṃghika, and later Mahāyāna turned toward Sanskrit. Sanskritization came later, as Buddhism entered circles that regarded Sanskrit as the language of cultivated expression.

As for writing, the account continues from 1.4. Oral transmission lasted three or four hundred years, and the earliest manuscripts appeared only in the first century BCE. The southern tradition says that the Pāli canon was first written down in Sri Lanka. The reasons given in the Mahāvaṃsa and Dīpavaṃsa are that monks "saw the decline of living beings" and wished "to make the true teaching endure." Warfare and the king's exile appear in the same chapter, but neither chronicle names them as the reason for writing, as the note to 1.4 explains. The oldest surviving Buddhist manuscripts are birch-bark fragments found in arid Gandhāra. The sequence of transmission can be shown as follows:

6. What we actually possess¶

Taken together, the preceding sections support a sober conclusion with two sides.

We do not possess the Buddha's exact words. Every surviving version, in any language, stands at several removes. The texts passed through centuries of oral transmission, repeated communal recitation and correction, separate preservation by several schools, and eventual writing. The councils themselves also served, to varying degrees, to authorize particular versions. No single surviving transmission, including the Pāli Tipiṭaka, can be treated as a verbatim record of the Buddha. The paragraph combines several judgments from textual history, rather than resting on one source. Their separate grounds appear earlier in the present chapter and elsewhere in this essay.

Yet the independent preservation of the same teachings by schools that had long been separate gives us the means for comparison. The Pāli Nikāyas, Chinese Āgamas, and Gāndhārī fragments can be checked against one another. Material retained in agreement across independent transmissions very probably derives from a common source preceding the divisions.⁷ The position taken in the present book is that comparison can identify a layer of "common teachings from before the schools divided" with considerable confidence. The method justifies Essay 1's use of early scripture to discuss the "earliest teaching," and accords with the judgment of oral transmission's fidelity in 1.4.

The same body of teachings developed into independent versions because the schools divided. The next chapter examines how the division began, which groups emerged, and what they were actually disputing.

References¶