An image of the Buddha on a red background

The development of the Buddhist 'canon'

The Buddhist ‘canon’ is vast, complex and difficult to define. Here Professor Tim Barrett outlines some of the key works for the different branches of Buddhism.

Like the adherents of many other religious traditions, Buddhists have always considered that some writings are particularly important to them and have taken steps to preserve them as a separate group. Given the long time span and vast geographical range of Buddhist history, it is inevitable that both the structure and the contents of these collections of ‘the word of the Buddha’ have varied markedly.

The Buddhist scripture was first compiled orally, since writing was not in common use in the Buddha’s time in India. Thus his teachings were not written down, and a pattern of memorised discourses and other materials therefore constitutes the oldest layer that was handed down. The origins of this heritage is said to have begun with an assembly of monks immediately after the Buddha’s death, at which his teachings were first recollected and recited. Secondly, everything he had said regarding the rules to be observed by the monastic community were recollected. Added to these was a third type of memorised material consisting of later analyses of the teachings by learned monks. This tripartite structure is central to all branches of Buddhism.

Collection of rules for Buddhist monks

Collection of rules for Buddhist monks

A palm leaf manuscript of the Bhikkhu Patimokkha, the rules for Buddhist monks. Burma, 19th century.

View images from this item  (3)

Usage terms

Public Domain in most countries other than the UK.

When were the Buddhist teachings first written down?

Though the collective memory of the monastic community as a whole seems to have coped with the task of retaining and refining the Buddha’s message, it remained vulnerable in situations where the community was threatened. It was for this reason in the 1st century BCE, when war and famine raised the fear of extinction among Buddhists on the island of Sri Lanka, that the scripture was first codified. Since writing was now available, all that had been inherited was recorded and assigned according to the three types to one of ‘three baskets’ of texts, the ‘Tipiṭaka’. The language used was Pāli, the language in which the heritage had been memorised there. The scripture naturally divided into three: the ‘Sutta Piṭaka’, the discourses and words of the Buddha, the ‘Vinaya Piṭaka’, the monastic regulations and the ‘Abhidhamma Piṭaka’, the scholastic analytical works.

An Abhidhamma commentary

Image of Or 16079, An Abhidhamma commentary folio 001r

Commentary on the Abhidhamma Piṭaka in a folding book in Shan language. Mueang Lakorn, Thailand, 1917 CE.

View images from this item  (6)

Usage terms

Public Domain in most countries other than the UK.

The Pāli canon is thus centred around the Historical Buddha and his life. The ‘Sutta’ category is further divided into five groups called Nikāyas, classified on formal features such as length. Other Buddhist groups on the sub-continent of India also used this tripartite scheme, which must be very ancient, though details of when they put their heritage into writing and in what linguistic form are not clear. In addition to the canonical scriptures in Pāli language, numerous commentaries and sub-commentaries were written by Pāli scholars over the past 2000 years. The works of the 5th-century scholar Buddhaghoṣa for example have been essential for the understanding of Theravāda Buddhist scriptures.

Mahabuddhaguna and other Buddhist texts

Mahābuddhagunā and other Buddhist texts

Manuscript from central Thailand describing the virtues of the Buddha, closely related to descriptions found in the scholar Buddhaghosa’s treatise on the Path of Purification, known as Visuddhimagga. Thailand, 18th century.

View images from this item  (8)

Were other languages used for the Buddhist teachings?

Similar developments to write down the Buddha’s teachings also took place elsewhere. In the 1st century CE they were written down in Sanskrit in India. The Sanskrit Tipiṭaka follows the same structure as the Pāli Tipiṭaka, with the ‘Sūtra Vaibasha’ and the ‘Vinaya Vaibasha’ being similar to the Pāli ‘Sutta Piṭaka’ and ‘Vinaya Piṭaka’. The differences occur in the Sanskrit ‘Abhidhamma Vaibasha’, where scholarly philosophical analysis differs from in the Pāli canon. Both traditions however can be traced back to the original teachings of the Buddha. The Sanskrit canon no longer exists in full in its original language, but we know of important Sanskrit works through multiple translations. Recently discovered manuscripts from the region of the borders of Afghanistan and Pakistan, once known as Gandhāra, probably date to this time.

Gandharan Scrolls

Gandharan Buddhist scroll, first century, f.30v

Fragments of birch bark manuscript scroll from Gandhāra (1st century CE).

View images from this item  (5)

Usage terms Public Domain

Some bear close similarities to works in the Pāli Tipiṭaka. However among these fragments there is also a new type of text that declares itself the key teaching of the Buddha, superior to what had earlier been considered the mainstream. Known in Sanskrit as the ‘Prajñāparāmita sūtra’, the ‘Perfection of Wisdom’, this work is the earliest known of its kind.

Sutra of the Perfection of Wisdom in 100,000 Lines

Sutra of the Perfection of Wisdom in 100,000 Lines, IOL San 1492/15–19, 15a

This is the longest of all the sūtras of the ‘Perfection of Wisdom’, which are a collection of around forty texts that focus on the doctrine of emptiness (śunyatā) and have been popular wherever Mahāyāna Buddhism has flourished.

View images from this item  (10)

Usage terms Public Domain

There are a large number of similar texts making similar claims, which Māhāyana Buddhists accept as the Buddha’s word. The Lotus Sūtra, or the Diamond Sūtra, a short work which is also concerned with the perfection of wisdom, would be another example.

Japanese Lotus Sutra

Start of the Eighth chapter of the 'Lotus Sutra'

The Lotus Sūtra is one of the most influential scriptures of Mahāyāna Buddhism in East Asia and is seen by many of its adherents as the summation of the Buddha’s teachings.

View images from this item  (22)

Usage terms Public Domain

When were the Buddhist works translated into Chinese?

Sanskrit, the standard language that had come to dominate in India, were translated into Chinese. In the 2nd century the ‘Perfection of Wisdom’ was translated from Gandhāran, and gradually more and more examples of this type of text came to be translated into Chinese too, along with some of the older materials similar to those in Pāli. In the 2nd century the rulers of China had undertaken the carving of all the ancient Chinese texts associated with Confucius on stone, thus providing a canon of such writings that defined for everyone what texts were considered important in that heritage. As more and more Buddhist translations were undertaken over the decades, how were Chinese Buddhists to define and organise their own growing heritage? The question became more important as it gradually became apparent to them that, in the Indian language at least, some of the older literature had not just been amassed randomly but had been classified into different categories.

Towards the end of the 4th century in north China these concerns became even more pressing as Buddhists came into a closer relationship with China’s rulers. They needed to know what writings were acceptable as truly Buddhist, so as to guard against imposters. The first definitive catalogues of all Chinese Buddhist works were drawn up, covering all translations that could be verified, and other writings composed in Chinese that were an authentic embodiment of the same teachings. By this time the wave of self-declared superior representations of the Buddha’s teaching, embodying ideas not obvious in the heritage of groups like the transmitters of the Pāli canon, were calling themselves Māhāyana, the ‘Greater Vehicle’ of salvation, and denigrating the older literature as ‘Hinayana’, a ‘Lesser Vehicle’ in which the higher truths of the Buddha’s teaching were not to be found. This distinction, together with the divisions of the canon already established in India, was incorporated into Chinese catalogues from this time onward.

Though the earliest catalogues have not survived, by the 5th century it is clear that these lists already contained over a thousand titles, while a fragment of manuscript preserved in London turns out to be all that remains from one set (out of ten) of the whole mass of these scriptures copied out together on paper at the same time. As more translations were made and the Chinese canon was defined, original Chinese sūtras were also added, composed in Chinese rather than translated. Though they may not be considered as canonical, they played a central part in the local faith. The scripture on the ‘Ten Kings of Hell’ is an example of such texts; it is a guide to the punishments awaiting sinners in the afterlife that shows every sign of being the product of Chinese ways of thinking.

Did other Buddhist civilisations develop their own canons?

Elsewhere the Tibetans had also been translating from Sanskrit and from Chinese since the late 8th century. From the 11th century the Tanguts, living in the north-western borderlands of China, had done likewise, and had used the relatively new technology of Chinese printing. Though Tangut civilisation as an independent force was extinguished by the Mongols in 1227, their conquerors were evidently impressed enough by their achievements to gather their translations and print them in China at the start of the 14th century as an organised canon, though unfortunately this has not survived. A century later, in 1410, the next rulers of China, the Ming dynasty, printed the Tibetan canon, which we know had earlier been organised under the title of ‘Kanjur’ for all the ‘words of the Buddha’, with later scholastic writings in a separate collection, the ‘Tenjur’. This survives, along with several subsequent editions that were published in China and even more from Tibet.

A volume of the Shelkar Kanjur

Volume of the Shelkar Kanjur, f001av

An 18th-century Kanjur, the corpus of collected teachings of the Buddha, translated into Tibetan.

View images from this item  (2)

Usage terms Public Domain

The Narthang print of the Tenjur

14310a RGYUD 1, The Narthang print of the Tenjur, f001v

An 18th-century copy of the Narthang Tenjur, the ‘translated commentaries’ on the teachings of the Buddha.

View images from this item  (5)

Usage terms Public Domain

Taken together, these woodblock editions and manuscripts preserved over four and a half thousand works, and served as the source for the creation of a Mongolian-language Kanjur in the early 17th century. In the late 18th century a Manchu Kanjur was also created by the Manchu emperor in Beijing, drawing on both Tibetan and Chinese materials.

Vajrayana texts

Vajrayana texts draw on the ideas of Māhāyana but present them through meditation practices and visually startling art that often seems to negate basic Buddhist values, though in fact they depend on trust in a teacher (the term guru is used) to give the true meaning of their scripture (the term used is tantra) to those who are initiated into their practices. Vajrayana is said to provide a more direct route to enlightenment, but since it is open to misunderstanding by the uninitiated it is sometimes called ‘esoteric’.

Is there a single ‘Buddhist canon’?

Buddhists have proved adept at preserving their writings over the centuries. However, as we have seen, the idea of a Buddhist ‘canon’ is vast, complex and difficult to define. In fact, the English term may give a misleading impression of a set of writings that all followers of the tradition are expected to know, and clearly the Buddhist case is not like that. There are texts in East Asia of great importance to the daily lives of Buddhists over the centuries that were kept out of the various versions of the canon until modern times because they were – despite claiming to be just like any other sūtra to convey the preaching of the Buddha – not in fact translated but composed in the local language. The Bardo Thodol (the so-called ‘Tibetan Book of the Dead’) is a very popular work that emerged through revelation in the medieval period and therefore was not included in the Tibetan Buddhist canon. Moreover, much besides those works mentioned here has been written by Buddhists in many languages throughout Asia that does not claim any special status at all, but still conveys something of the teachings that can be traced back to ancient India.


Pancaraksa Or 13946

The Goddess Mahāpratisarā, one of the personifications of the five protective mantras (dhāraṇīs). Miniature from a Pañcarakṣā manuscript from Nepal.

View images from this item  (14)

Usage terms Public Domain

There are also texts like the Lotus Sūtra, whose special status is not just proclaimed by the work itself, but also acknowledged in the art and material culture of East Asia in a way that makes the use of the word ‘canonical’ far from inappropriate. The Buddha’s word was for many, in effect, a form of the Buddha’s very presence, still with us in later ages, and still to be treated with every form of respect. In looking at Buddhist books this is certainly something to bear in mind.

  • T H Barrett
  • Professor T. H. Barrett is Professor Emeritus of East Asian History at SOAS, University of London. He works primarily on the religious history of East Asia, and also on British perceptions of East Asia, in both cases dealing mainly with China. Though he has written on may topics from the history of cats in China and Daoism in Japan to the transmission of the ‘Secret History of the Mongols’, he has a particular interest in the history of the book in East Asia, on which he has published The Woman Who Discovered Printing and other studies. He is now working on a survey of Buddhist attitudes to other forms of Chinese religion.

The text in this article is available under the Creative Commons License.